Professional Documents
Culture Documents
DESIGN
HANDBOOK
McGraw-Hill
New York San Francisco Washington, D.C. Auckland Bogota Caracas Lisbon
London Madrid Mexico City Milan Montreal New Delhi San Juan
Singapore Sydney Tokyo Toronto
043210
ISBN 0-07-041152-2
The sponsoring editor for this book was Larry Hager and the production supervisor
was Pamela A. Pelton. It was set in Times Roman by Compuvision.
Printed and bound by R. R. Donnelley & Sons Company.
This book was printed on acid-free paper.
McGraw-Hill books are available at special quantity discounts to use as premiums
and sales promotions, or for use in corporate training programs. For more information, please write to the Director of Special Sales, McGraw-Hill, Inc. Two Penn Plaza,
New York, NY 10121-2298. Or contact your local bookstore.
Information contained in this work has been obtained by The
McGraw-Hill Companies, Inc. ("McGraw-Hill") from sources
believed to be reliable. However, neither McGraw-Hill nor its
authors guarantee the accuracy or completeness of any information
published herein, and neither McGraw-Hill nor its authors shall be
responsible for any errors, omissions, or damages arising out of
use of this information. This work is published with the understanding that McGraw-Hill and its authors are supplying information but are not attempting to render engineering or other professional services. If such services are required, the assistance of an
appropriate professional should be sought.
PREFACE
The Hydraulic Design Handbook, referred to herein as the Handbook, has been an extensive effort to develop a comprehensive reference book on hydraulic design. Over the years
it has become more and more obvious to me the need for such a comprehensive reference
book, particularly for the practicing engineer, who needs a central reference to start to
accumulate knowledge for the hydraulic design process. They need a reference
to find state-of-the-art design procedures, worked out example problems, discussions of
what can go wrong in the design process, up-to-date references for further reading, and
limited discussions of the theoretical aspects as they relate to the design process. Also references and discussions are needed for many different computer codes that are used in
hydraulic design practice. This Handbook is an attempt to fulfill these needs of the practicing engineer.
Also I consider this Handbook to be a valuable reference, if not a text, for use in teaching hydraulic design at both the undergraduate and graduate levels. Basically there are no
published books (or textbooks) that comprehensively cover hydraulic design. Hopefully
the Handbook can be used effectively for such courses because the balance of design concepts and procedures with worked out examples and just enough theoretical coverage
should make this, a valuable book for the hydraulic design classroom.
A large amount of knowledge has accumulated over the years on hydraulic design,
mainly in various government publications and reports, in journals and a limited amount
in textbooks. Unfortunately this knowledge has not been accumulated into one central
location. Over the years many new methods have been developed and many computer
programs have been written for hydraulic design. Engineers need an authoritative
source of guidance as to which methods and computer programs are appropriate for
various given tasks. This Handbook is intended to provide this information in a concise
and accessible form. This Handbook is not, however, an encyclopedia of every known
fact about hydraulic design.
The Handbook is divided into four sections, each consisting of a group of chapters. The
first section on principles for hydraulic design consists of chapters on pressurized flow,
open-channel flow, groundwater flow, environmental hydraulics, sedimentation and erosion hydraulics, and risk-reliability analysis. These are fairly comprehensive chapters that
focus on the concepts of the various types of flow. The chapter on risk-reliability presents
many methodologies for risk-reliability analysis and presents the U.S. Army Corps of
Engineers methodology. The second series of chapters are devoted to hydraulic design for
water supply with the major emphasis on the design of water distribution systems. These
chapters include pump design, system design, pipe design, and hydraulic transients.
A chapter is also presented on the hydraulic design for energy generation. The third series
of chapters are devoted to hydraulic design for water excess management. The chapters
are rather comprehensive ranging from highway drainage, urban drainage, highway crossings, flood control channels, spillways, stilling basins, flow transitions, energy dissipators,
and flood-plain hydraulics. The fourth series of chapters are devoted to environmental considerations with chapters on the hydraulics of water and wastewater treatment systems,
Larry W. Mays
Scottsdale, Arizona
Acknowledgments
I must first acknowledge the authors who made this handbook possible. It has been a sincere privilege to have worked with such an excellent group of dedicated people. They are
all experienced professionals who are among the leading experts in their fields.
References to material in this handbook should be attributed to the respective chapter
authors.
During the past twenty-three years of my academic career I have received help and
encouragement from so many people that it is not possible to name them all. These people represent a wide range of universities, research institutions, government agencies, and
professions. To all of you I express my deepest thanks.
I would like to acknowledge Arizona State University, especially the time afforded me to
pursue this handbook.
I sincerely appreciate the advice and encouragement of Larry Hager of McGraw-Hill
throughout this project.
This handbook has been a part of a personal journey that began years ago when I was a
young boy with a love of water. Books are companions along the journey of learning. I
hope that you will be able to use this handbook in your own journey of learning about
water. Have a happy and wonderful journey.
Larry W. Mays
CHAPTER 1
INTRODUCTION
Larry W. Mays
Department of Civil and Environmental Engineering
Arizona State University
Tempe, Arizona
1.1
OVERVIEW
Since the Egyptian's and Mesopotamian's first successful efforts to control the flow of
water thousands of years ago, a rich history of hydraulics has evolved. Sec. 1.2 contains
a brief description of some ancient hydraulic structures that are found around the world.
During the 20th century, many new developments have occurred in both theoretical and
applied hydraulics. A number of handbooks and textbooks on hydraulics have been published, as indicated in Fig. 1.1. From the viewpoint of hydraulic design, however, only
manuals, reports, monographs, and the like have been published, mostly by government
agencies. Unfortunately, many aspects of hydraulic design have never been published as
a compendium. This Hydraulic Design Handbook is the first effort devoted to producing
a comprehensive handbook for hydraulic design. The book covers many aspects of
hydraulic design, with step-by-step procedures outlined and illustrated by sample design
problems.
Springs
Cisterns
Dams
Wells
Reuse of excrement as fertilizer
Gravity flow supply pipes or channels and drains, pressure
pipes (subsequently forgotten)
8th-6th c. B.C.
Long-distance water supply lines with tunnels and bridges,
as well as intervention in and harnessing of karst water
systems
Public as well as private bathing facilities, consisting of:
6th c. B.C. at the latest
bathtubs or showers, footbaths, washbasins, latrines or
toilets, laundry and dishwashing facilities
Use of definitely two and probably three qualities of water:
6th c. B.C. at the latest
potable, subpotable, and nonpotable, including irrigation
using storm runoff, probably combined with waste waters
6th-3rd c. B.C.
Pressure pipes and siphon systems
*Indicates an element discovered, probably forgotten, and rediscovered later,
indicates an educated guess.
Source: Crouch, 1993.
1.2.2
Irrigation Systems
1.2.2.1 Egypt and Mesopotamia. In ancient Egypt, the construction of canals was a major
endeavor of the Pharaohs beginning in Scorpio's time. Among the first duties of provincial
governors was the digging and repair of canals, which were used to flood large tracts of land
while the Nile was flowing high. The land was checkerboarded with small basins defined by
a system of dikes. Problems associated with the uncertainty of the Nile's flows were recognized. During high flows, the dikes were washed away and villages were flooded, drowning
thousands of people. During low flows, the land was dry and no crops could grow. In areas
where fields were too high to receive water directly from the canals, water was drawn from
the canals or from the Nile by a swape or shaduf(fig. 1.3), which consisted of a bucket on
the end of a cord hung from the long end of a pivoted boom that was counterweighted at the
short end (de Camp, 1963). Canals continued to be built in Egypt throughout the centuries.
The Sumerians in southern Mesopotamia built city walls and temples and dug canals
that were the world's first engineering works. It also is of interest that these people, fought
over water rights from the beginning of recorded history. Irrigation was vital to
Mesopotamia, Greek for "the land between the (Tigris and Euphrates) rivers." An ancient
Babylonian curse was, "May your canal be filled with sand" (de Camp, 1963), and even
their ancient laws dealt with canals and water rights. The following quotation from
approximately the sixth century B.C., illustrates such a law (de Camp, 1963):
"The gentleman who opened his wall for irrigation purposes, but did not make
his dyke strong and hence caused a flood and inundated a field adjoining his,
shall give grain to the owner of the field on the basis of those adjoining"
Because the Tigris and Euphrates carried several times more silt per unit volume of water
than the Nile did, flooding problems were more serious in Mesopotamia than in Egypt. As
a result the rivers in Mesopotamia rose faster and changed course more often.
India
Elarn
Italy
Spain
Writing
Sumer
Indus
Valley
V.cUe
P.ncd
Hyksos
Linear B
Hittite*
Etruscans
Roman Empire
Byzantium
FIGURE 1.2 Chronology and location of different civilizations ranging from India to Western Europe.
(Neugebauer, 1993)
The irrigation systems in both Mesopotamia and the Egyptian Delta were of the basin
type, opened by digging a gap in the embankment and closed by placing mud back into
the gap. (See Fig. 1.4 for a comparison of the irrigation works in Upper Egypt and in
Mesopotamia.) Water was hoisted using the swape, Mesopotamian laws required farmers
to keep their basins and feeder canals in repair; they also required everyone else to wield
hoes and shovels when the rivers flooded or when new canals were required or old ones
needed repair (de Camp, 1963). Some canals may have been used for 1,000 years before
FIGURE 1.3 Shadufs of the Amarna period, from the tomb of Nefer-Hotep at Thebes.
Note irrigation of date palms and other orchard trees and the apparent tank or pool (lower
right). The water pattern in the lowest margin suggests lifting out of an irrigation canal.
(Davies, 1933, pis. 46 and 47). Figure as presented in Butzer (1976).
they were abandoned and others were built. Even today, 4,000 to 5,000 years later, the
embankments of the abandoned canals remain. In fact, these canal systems supported a
larger population than lives there today. Over the centuries, Mesopotamian agriculture
began to decline because of the salty alluvial soil. In 1258, the Mongols conquered
Mesopotamia and destroyed its irrigation systems.
Mesopotamia
A.D. MOO
Upper Egypt
A.O. 1*50
CinalS (.hf-mol
C9tn<itf*t with tf.tct)
SOHAG
Dikes
FIGURE 1.4 Comparative irrigation networks in Upper Egypt and Mesopotamia. A. Example of linear,
basin irrigation in Sohag province, ca. AD 1850. B. Example of radial canalization system in the lower
Nasharawan region southeast of Baghdad, Abbasid (A.D. 883-1150). Modified from R. M. Adams (1965,
(Fig. 9) Same scale as Egyptian counterpart) C. Detail of field canal layout in B. (Simplified from R. M.
Adams, 1965, Fig. 10). Figure as presented in Butzer (1976).
The Assyrians also developed extensive pubic works. When Sargon II invaded Armenia
in 714 B.C., he discovered the qandt (Arabic) or kariz (Persian), a system of tunnels used
to bring water from an underground source in the hills down to the foothills (Fig. 1.5).
Sargon destroyed the system in Armenia but brought the concept back to Assyria. Over the
centuries, this method of irrigation spread across the Near East into North Africa and is
Mountain range
Masonry
lining
Infiltration
to qanat
Groundwater level
Aquifer (alluvial silt)
Rock
Shale (impervious)
still used. Sargon's son Sennacherib also developed waterworks by damming the Tebitu
River and using a canal to bring water to Nineveh, where the water could be used for irrigation without the need for hoisting devices. During high water in the spring, overflows
were handled by a municipal canebrake that was built to develop marshes used as game
preserves for deer, wild boar, and birds. When this system was outgrown, a new canal 30
mi long was built, with an aqueduct that had a layer of concrete or mortar under the upper
layer of stone to prevent leakage.
1.2.2.2 Prehistoric Mexico. During the earliest years of canal irrigation in Mexico,
the technology changed little (Fig. 1.6) and the method of flooding tended to be haphazard. The technological achievements were relatively primitive until about 600 or
500 B.C., and few of the early systems remain. Whereas the earlier systems were constructed of loosely piled rocks, the later ones consisted of storage dams constructed of
blocks that were mortared together. Some spillways were improved, and floodgates
were used in some spillways. (Some dams could be classified as arch dams.) The
canals were modified to an extent during this time: Different cross-sectional areas were
CULTURE PERIOD
Year
VfeMey of Mexico
Oaxaca
TECHNOLOGICAL DEVELOPMENTS
-1600
LATE
POSTCLASSlC
-1400 * Relocation of rivers, Large masonry aqueducts.
POSTCLASSIC
-1200
EARLY
POSTCLASSIC -1000
~ 800 * Advanced relocation of ephemeral streams.
CLASSIC
-600
CLASSIC
LATE
FORMATIVE
MICHXE
FORMATiVE
EARLY
FORMATIVE
TERMINAL
FORMATIVE
LATE
FORMATIVE
MIDDLE
FORMATIVE
EARLY
FORMATIVE
FIGURE 1.6 Regional chronology and dates of developments in various aspects of canal irrigation
technology in Mexico. (Doolittle, 1990)
used, some were lined with stone slabs, and the water for irrigation of crops was more
carefully controlled.
Between 550 and 200 B.C., the irrigation-related features and the entire canal systems
were significantly improved. The channelization of stream beds, the excavation of canals,
and the construction of dams were probably the most significant improvements. However,
the technology stopped improving after 200 B.C., and no significant developments
occurred for approximately 500 years. Around 300 A.D., a few new improvements were
initiated, but the technology remained essentially the same through the classic period
(A.D., 200 - 800/1000) and early postclassical period (A.D. 800/1000-1300). Figure 1.7 is
a map of fossilized canals in the Tehuacan Valley in Mexico.
Tehuacan
San Juan
Ajalpan
San Gabriel
Chilac
San
Sebastian
Zmacatepec
FIGURE 1.7 Map of fossilized canals on the Llano de Ia Taza in the Tehuacan Valley.
(Woodbury and Neely, 1972, as presented in Doolittle, 1990)
1.2.2.3 North America. The canal irrigation systems in the Hohokam and Chaco
regions stand out as two major prehistoric developments in the American Southwest
(Crown and Judge, 1991). The two systems expanded over broad geographic areas of similar size (the Hohokam in Arizona and the Chacoans in New Mexico). Although they were
developed at similar times, they apparently functioned independently. Because the two
systems evolved in different environments, their infrastructures also differed considerably.
The Hohokam Indians inhabited the lower Salt and GiIa River valleys near Phoenix,
Arizona. Although the Indians of Arizona began limited farming nearly 3000 years ago,
construction of the Hohokam irrigation systems probably did not begin until the first few
centuries A.D. Who originated the idea of irrigation in Arizona, whether the technology
was developed locally or it was introduced from Mexico, is unknown. Figure 1.8 illustrates the extensive system in the Phoenix area, and Fig. 1.9 provides a schematic of the
details of its major components.
In approximately 1450 A.D., the Hohokam culture declined, possibly for a combination
of reasons: flooding in the 108Os, hydrologic degradation in the early UOOs, and the
recruitment of laborers by surrounding populations. The major flood in 1358 ultimately
destroyed the canal networks, resulting in movement of the people. Among the Pima
Indians, who were the successors of the Hohokam Indians, use of canals was either
limited or absent. Although the prehistoric people who lived outside the area of Hohokam
FIGURE 1.8 Canal building in the Salt River Valley with a stone hoe held in the hand without a handle. These were the original engineers, the true pioneers who built, used,
and abandoned a canal system when London and Paris were a cluster of wild huts. Turney (1922) (Courtesy of Salt River Project, Phoenix, Arizona)
Diversion weir
Conol intoke oreo
Moin conol
Primory
heodgofe (with
stone poving)
secondory )
(some with stone poving)
Loteroi conol heodgote(topon)
Lateral canal network
Field oreo
Field house
Field turn-out
Junction pool
culture also constructed irrigation systems, none approached the grand scale of the
Hohokam systems.
In the ninth century, the Anasazi people of northwestern New Mexico developed a
cultural phenomenon, the remains of which currently consist of more than 2400
archaeological sites and nine towns, each containing hundreds of rooms, along a 9-mi
stretch. The Chacoan irrigation system is situated in the San Juan Basin in northwestern New Mexico. The basin has limited surface water, most of it discharges from
ephemeral washes and arroyos. Figure 1.10 illustrates the method of collecting and
diverting runoff throughout Chaco Canyon. The water collected from the side canyon
that drained from the top of the upper mesa was diverted into a canal by either an earthen or a masonry dam near the mouth of the side canyon (Vivian, 1990). These canals
averaged 4.5 m in width and 1.4 m in depth; some were lined with stone slabs and others were bordered by masonry walls. The canals ended at a masonry head gate, where
water was then diverted to the fields in small ditches or to overflow ponds and small
reservoirs.
Mesa Top
MULTIPLE HEAOGATE
1.2.3
Dams
The Sadd-el-Kafara dam in Egypt, situated on the eastern bank of the Nile near Heluan
approximately 30 km south of Cairo, in the Wadi Garawi, has been referred to as the world's
oldest large dam (Garbrecht, 1985). The explorer and geographer George Schweinfurth
rediscovered this dam in 1885, and it has been described in a number of publications since
that time (see Garbrecht, 1985). It was built between 2950 and 2690 B.C. Although the Jass
drinking-water reservoir in Jordon and the diversion dams on the Kasakh River in Russia
are probably older, they are much smaller than the Sadd-el-Kafara (Dam of the Pagans).
It is unlikely that the Sadd-el-Kafara dam was built to supply water for drinking or irrigation because the dam lies too far from the alabaster quarries situated upstream to have
supplied the labor force with drinking water. Furthermore, there is a vast supply of water
and fertile land in the nearby Nile valley. The apparent purpose of the dam was to protect
installations in the lower wadi and the Nile valley from frequent, sudden floods. The dam
was destroyed during construction by a flood; consequently, it was never completed. To
date, the dam's abutments still exist.
The dam had an impervious core consisting of rubble, gravel, and weathered material. On both the upstream and downstream sides, the core was bordered by sections of
rockfill that supported and protected the core. The diameter of the stones ranged from 0.1
to 0.6 m. One remarkable construction feature is the facing of the section of rockfill
where parts of the facing on the upstream side are still well preserved. The dam had an
approximate crest length of 348 ft and a base length of 265 ft and was built straight
across the wadi at a suitably narrow point, with a maximum height of 32 ft above the
valley bed. See Smith (1971) and Upton (1975) for more on dams.
Dam building in the Americas began in the pre-Colombian period in the civilizations
of Central and South America: the Aztecs in Mexico, the Mayans in Guatemala and
Yucatan, and the Incas in Peru. Where as old-world civilizations developed in the valleys
of the big rivers, the Nile River, the Euphrates and the Tigris Rivers, the Indus River, and
the Yellow River, most of the early civilizations in the New World were not river civilizations. In South America, the civilizations appeared in the semiarid highlands and the
arid coastal valleys traversed by small rivers. In Central America, the Mayans, the Aztecs,
and the predecessors of the Aztecs were not river civilizations.
The Mayans did not practice irrigation; however, they did provide efficient water
supplies to several of their large cities. They developed the artificial well (cenote\ the underground cistern (chultun\ and the large open reservoir (aguado). The Mayans' failure to
develop irrigation may have accelerated their decline. In the Yucatan, the aguados are still
found in some places, but the cenote was the major source of water for drinking and bathing.
1.2.4
Knossos, approximately 5 km from Herakleion, the modern capital of Crete, was among
the most ancient and unique cities of the Aegean and Europe. The city was first inhabited
shortly after 6000 B.C. and, within 3000 years, it had became the largest Neolithic
Settlement in the Aegean (Neolithic age, circa 5700-2800 B.C). During the Bronze age
(circa 2800-1100 B.C.), the Minoan civilization developed and reached its culmination as
the first Greek cultural miracle of the Aegean world.
The Minoan civilization has been subdivided into four periods: the prepalatial period
(2800-1900 B.C.), the protopalatial period (1900-1200 B.C.), the neopalatial period
(1700-1400 B.C.), and the postpalatial period (1400-1100 B.C.). During the prepalatial
period, a settlement at Knossos; was leveled to erect a palace. Little is known about the
old palace because it was destroyed in approximately 1700 B.C. A new palace was constructed on leveled fill from the old palace. During the neopalatial period, Knossos was at
the height of its splendor. The city covered an area of 75,000 to 125,000 m2 and had a population estimated to be on the order of tens of thousands.
The irrigation and drainage systems at Knossos were most interesting. An aqueduct
supplied water through tubular conduits from the Kounavoi and Archanes regions and
branched out into the city and the palace. Figure 1.11 shows the type of pressure conduits
used within the palace for water distribution. The drainage system consisted of two separate conduits: one to collect the sewage and the other to collect rain water (Fig. 1.12).
Unfortunately, the Mycenean palace was destroyed by an earthquake and fire in approximately 1450 B.C., as were all the palatial cities of Crete.
Anatolia, also called Asia Minor, which is part of the Republic of Turkey, has been the
crossroads of many civilizations during the past 10,000 years. During the last 4000 years,
going back to the Hittite period (2000-200 B.C.) many remains of ancient urban watersupply systems have been found, including pipes, canals, tunnels, inverted siphons, aqueducts, reservoirs, cisterns, and dams, (see Ozis, 1987 and Ozis and Harmancioglu, 1979).
FIGURE 1.11 Water distribution pipe in Knossos, Crete. (Photograph by L.W. Mays)
FIGURE 1.12 Urban drainage system in Knossos, Crete. (Photograph by L.W. Mays)
An example of one such city is Ephesus, which was founded during the 10th century
B.C. as an Ionian city out of the Temple of Artemis. In the sixth century B.C., the city settled near the temple, and subsequently was reestablished at its present site, where it developed further during the Roman period. Water was supplied to Ephesus from springs at different sites. Cisterns also supplied well water to the city. Water for the great fountain, built
between 4 and 14 A.D., was diverted by a small dam at Marnss and was conveyed to the
city by a system 6 km long consisting of one large and two small clay pipe lines. Figure
1.13 shows the type of clay pipes used at Ephesus to distribute water.
The latrine, or public toilet shown in Fig. 1.14, was built in the first century A.D. at
Ephesus. The toilets were placed side by side with no partitions. In the middle was a
square pond, and the floors were paved with mosaics.
The Great Theatre at Ephesus, the city's largest and most impressive building, had a
seating capacity for 24,000 people. Built in the Hellenistic period, the theatre was not only
a monumental masterpiece but during the early days of Christianity, one major confrontation between Artemis and Christ took place there. Of notable interest from a waterresources viewpoint is the theatre's intricate drainage system. Figure 1.15 shows a
drainage channel in the floor of the theatre.
Public baths also were a unique feature in ancient cities: for example, the Skolactica
baths in Ephesus had a salon and central heating; a hot bath (caldariuni), a warm bath
(tepidarium), and a cold bath (frigidarium)', and a dressing room (apodyterium). In the
second century A.D., the first building had three floors. In the fourth century, a woman
named Skolacticia modified the baths, making them accessible to hundreds of people.
There were public rooms and private rooms, and people who wished to could stay for
many days. Hot water was provided by a furnace and a large boiler.
Perge is another ancient city in Anatolia that had a unique urban water infrastructure.
The photographs in Fig. 1.16 illustrate the Majestic Fountain (nymphaion), which consisted of a wide basin and a richly decorated architectural facade. Because of its architecture and statues, the fountain was one of Perge's most magnificent edifices. A water channel ran along the middle, dividing each street and bringing life and coolness to the city.
The baths of Perge were magnificent. The first photograph in Fig. 1.17 shows one of the
baths of Perge; the second photograph illustrates the storage of water under the floor to
keep the water warm. Like the baths in other ancient cities in Anatolia, the baths of Perge
had a caldarium, a tepidarium, and a frigidarium.
FIGURE 1.14 A latrine, or public toilet, built at Ephesus, Turkey, in the first century B.C.
(Photograph by L. W. Mays)
FIGURE 1.15 A drainage channel on the floor of the Great Theater at Ephesus,
Turkey. (Photograph by L. W. Mays)
The early Romans devoted much of their time to useful public works projects, including roads, harbor works, aqueducts, temples, forums, town halls, arenas, baths, and sewers. The prosperous early Roman bourgeois typically had a 12-room house, with a square
hole in the roof to let rain in and a cistern beneath the roof to store the water. Although
the Romans built many aqueducts, they were not the first to do so. King Sennacherio built
aqueducts, as did the Phoenicians and the Helenes. The Romans and Helenes needed
extensive aqueduct systems for their fountains, baths, and gardens. They also realized that
FIGURE 1.16 Two views of the Majestic Fountain (nymphaiori) in Perge, Anatolia, Turkey.
(Photographs by L. W. Mays)
water transported from springs was better for their health than river water and that spring
water did not need to be lifted to street level as did river water. Roman aqueducts were
built on elevated structures to provide the needed slope for water flow. Knowledge of pipe
FIGURE 1.17B View of the baths at Perge, Anatolia, Turkey. (Photographs by L.W. Mays)
water was clear and of sufficient quantity, it was conveyed to the city by aqueduct. Also,
water from several sources was collected in a reservoir, then conveyed by an aqueduct or
a pressure conduit to a distributing reservoir (castellum). Three pipes conveyed the water:
one to pools and fountains, one to the public baths for public revenue, and one to private
houses for revenue to maintain the aqueducts (Rouse and Ince, 1957). Figures 1.18 and
1.19 illustrate the layout of the major aqueducts of ancient Rome. Figure 1.20 shows the
Roman aqueduct in Segovia, Spain, which is probably among the most interesting of
Roman remains in the world. This aqueduct, built during the second half of the first century A.D. or the early years of the second century, has a maximum height of 78.9 m. See
Van Deman (1934) for more details on Roman aqueducts.
Irrigation was not a major concern because of the terrain and the intermittent rivers.
However, the Romans did, drain marshes to obtain more farmland and to eliminate the bad
air, or "harmful spirits," rising from the marshes because they believed it caused disease
(de Camp, 1963). The disease-carrying mechanism was not the air, (or spirits) but the
malaria-carrying mosquito. Empedocles, the leading statesman of Acragas in Sicily during the Persian War in the fifth century B.C., drained the local marshes of Selinus to
improve the people's health (de Camp, 1963).
The fall of the Roman Empire extended over a 1000-year period of transition called
the Dark Ages during which the concepts of science related to water resources probably
retrogressed. After the fall of the Roman Empire, clean water, sanitation, and public health
declined in Europe. Historical accounts tell of incredibly unsanitary conditions: polluted
water, human and animal wastes in the streets, and water thrown out of windows onto
passersby. As a result, various epidemics ravaged Europe. During the same period, the
Islamic cultures on the periphery of Europe religiously mandated high levels of personal
hygiene, highly developed water supplies, and adequate sanitation systems. For furthen
reading see Needham (1959) Payne (1959), Reynolds (1970) Robbins (1946), Sarton
(1952-59) and Wittfogel (1956).
Aqueducts
in
ANCIENT ROME
REGIONES
Porta Capena
IIHII Caelimontium
IsisetSerams
IVV Templum
Esquiliae Pads
VI ViaLata
AltaSemita
.VU
Forum Maximus
Romanum
IXVIII Circus
XII
Piscina
Publica
XIH
Aventinus
XVI Transtiberim
FIGURE 1.18 Termini of the major aqueducts in ancient Rome. (Evans, 1993)
Claudia et
Anio Novus
Marcia
Tepula
Julia
SPES VETUS
Arcus
Caelimontani
Marcia
Tepula
Julia
FIGURE 1.19 The area of Spes Vetus showing the courses of the major aqueducts entering the
city above ground. (From R. Lanciani, Forma Urbis Romae), as presented in Evans (1993).
1.4
Federal legislation contains policies that can affect the design of various types of
hydraulic structures. These policies are listed in Appendix l.A, where they are categorized
into the following sections: environment, health, historic and archeological preservation,
and land and water usage. The appendix also lists the abbreviations used in the policies,
(adapted from AASHTO, 1991).
1.5
Conventional procedures for hydraulic design are basically iterative trial-and-error procedures. The effectiveness of conventional procedures depends on an engineer's intuition,
experience, skill, and knowledge of hydraulic systems. Therefore, conventional procedures
are closely related to the human element, a factor that could lead to inefficient results for the
design and analysis of complex systems. Conventional procedures are typically based on
using simulation models in a process of trial and error to arrive at an optimal solution. Figure
Next Page
CHAPTER 2
HYDRAULICS OF
PRESSURIZED FLOW
Bryan W. Karney
Department of Civil Engineering
University of Toronto,
Toronto, Ontario,
Canada
2.1
INTRODUCTION
The need to provide water to satisfy basic physical and domestic needs; use of maritime and fluvial routes for transportation and travel, crop irrigation, flood protection, development of stream power; all have forced humanity to face water from the
beginning of time. It has not been an easy rapport. City dwellers who day after day
see water flowing from faucets, docile to their needs, have no idea of its idiosyncrasy. They cannot imagine how much patience and cleverness are needed to handle our great friend-enemy; how much insight must be gained in understanding its
arrogant nature in order to tame and subjugate it; how water must be "enticed" to
agree to our will, respecting its own at the same time. That is why a hydraulician
must first be something like a water psychologist, thoroughly knowledgeable of its
nature. (Enzo Levi, The Science of Water: The Foundations of Modern Hydraulics,
ASCE, 1995, p. xiii.)
Understanding the hydraulics of pipeline systems is essential to the rational design,
analysis, implementation, and operation of many water resource projects. This chapter
considers the physical and computational bases of hydraulic calculations in pressurized
pipelines, whether the pipelines are applied to hydroelectric, water supply, or wastewater
systems. The term pressurized pipeline means a pipe system in which a free water surface
is almost never found within the conduit itself. Making this definition more precise is difficult because even in a pressurized pipe system, free surfaces are present within reservoirs and tanks and sometimesfor short intervals of time during transient (i.e.,
unsteady) eventscan occur within the pipeline itself. However, in a pressurized pipeline
system, in contrast to the open-channel systems discussed in Chapter 3, the pressures
within the conveyance system are usually well above atmospheric.
Of central importance to a pressurized pipeline system is its hydraulic capacity: that is,
its ability to pass a design flow. A related issue is the problem of flow control: how design
flows are established, modified, or adjusted. To deal adequately with these
two topics, this chapter considers head-loss calculations in some detail and introduces the
topics of pumping, flow in networks, and unsteady flows. Many of these subjects are treated in greater detail in later chapters, or in references such as Chaudhry and Yevjevich (1981).
Rather than simply providing the key equations and long tabulations of standard
values, this chapter seeks to provide a context and a basis for hydraulic design. In addition to the relations discussed, such issues as why certain relations rather than others are
used, what various equations assume, and what can go wrong if a relation is used incor-
rectly also are considered. Although derivations are not provided, some emphasis is placed
on understanding both the strengths and weaknesses of various approaches. Given the
virtually infinite combinations and arrangements of pipe systems, such information is essential for the pipeline professional.
2.2
IMPORTANCEOFPIPELINESYSTEMS
Over the past several decades, pressurized pipeline systems have become remarkably
competitive as a means of transporting many materials, including water and wastewater.
In fact, pipelines can now be found throughout the world transporting fluids through every
conceivable environment and over every possible terrain.
There are numerous reasons for this increased use. Advances in construction
techniques and manufacturing processes have reduced the cost of pipelines relative to
other alternatives. In addition, increases in both population and population density have
tended to favor the economies of scale that are often associated with pipeline systems. The
need for greater conservation of resources and, in particular, the need to limit losses
caused by evaporation and seepage have often made pipelines attractive relative to openchannel conveyance systems. Moreover, an improved understanding of fluid behavior has
increased the reliability and enhanced the performance of pipeline systems. For all these
reasons, it is now common for long pipelines of large capacity to be built, many of which
carry fluid under high pressure. Some of these systems are relatively simple, composed
only of series-connected pipes; in other systems, the pipes are joined to form complex
networks having thousands of branched and interconnected lines.
Pipelines often form vital links in the process chain, and high penalties may be associated with both the direct costs of failure (pipeline repair, cost of lost fluid, damages
associated with rupture, and so forth) and the interruption of service. This is especially
evident in industrial applications, such as paper mills, mines, and power plants. Yet, even
in municipal systems, a pipe failure can cause considerable property damage. In addition,
the failure may lead indirectly to other kinds of problems. For example, a mainline break
could flood a roadway and cause a traffic accident or might make it difficult to fight a
major fire.
Although pipelines appear to promise an economical and continuous supply of fluid,
they pose critical problems of design, analysis, maintenance, and operation. A successful
design requires the cooperation of hydraulic, structural, construction, survey, geotechnical, and mechanical engineers. In addition, designers and planners often must consider the
social, environmental, and legal implications of pipeline development. This chapter focuses on the hydraulic considerations, but one should remember that these considerations are
not the only, nor necessarily the most critical, issues facing the pipeline engineer. To be
successful, a pipeline must be economically and environmentally viable as well as technically sound. Yet, because technical competence is a necessary requirement for any successful pipeline project, this aspect is the primary focus.
2.3
The designer of a hydraulic system faces many questions. How big should each pipe be
to carry the required flow? How strong must a segment of pipe be to avoid breaking? Are
reservoirs, pumps, or other devices required? If so, how big should they be and where
should they be situated?
There are at least two general ways of resolving this kind of issue. The first way is to
build the pipe system on the basis of our "best guess" design and learn about the system's
performance as we go along. Then, if the original system "as built" is inadequate, successive adjustments can be made to it until a satisfactory solution is found. Historically, a
number of large pipe systems have been built in more or less this way. For example, the
Romans built many impressive water supply systems with little formal knowledge of fluid
mechanics. Even today, many small pipeline systems are still constructed with little or no
analysis. The emphasis in this kind of approach should be to design a system that is both
flexible and robust.
However, there is a second approach. Rather than constructing and experimenting with
the real system, a replacement or model of the system is developed first. This model can
take many forms: from a scaled-down version of the original to a set of mathematical
equations. In fact, currently the most common approach is to construct an abstract numerical representation of the original that is encoded in a computer. Once this model is
"operational," experiments are conducted on it to predict the behavior of the real or proposed system. If the design is inadequate in any predictable way, the parameters of the
model are changed and the system is retested until design conditions are satisfied. Only
once the modeller is reasonably satisfied would the construction of the complete system
be undertaken.
In fact, most modern pipelines systems are modeled quite extensively before they are
built. One reason for this is perhaps surprisingexperiments performed on a model are
sometimes better than those done on the prototype. However, we must be careful here,
because better is a relative word. On the plus side, modeling the behavior of a pipeline
system has a number of intrinsic advantages:
Cost. Constructing and experimenting on the model is often much less expensive than
testing the prototype.
Time. The response rate of the model pipe system may be more rapid and convenient
than the prototype. For example, it may take only a fraction of a second for a
computer program to predict the response of a pipe system after decades of projected
growth in the demand for water.
Safety. Experiments on a real system may be dangerous or risky whereas testing the
model generally involves little or no risk.
Ease of modification: Improvements, adjustments, or modifications in design or
operating rules can be incorporated more easily in a model, usually by simply editing
an input file.
Aid to communication. Models can facilitate communication between individuals and
groups, thereby identifying points of agreement, disagreement, misunderstanding, or
issues requiring clarification. Even simple sketches, such as Fig. 2.1, can aid discussion.
These advantages are often seen as so overwhelming that the fact that alternative
approaches are available is sometimes forgotten. In particular, we must always remember
that the model is not reality. In fact, what makes the model useful is precisely its simplicityit is not as complex or expensive as the original. Stated more forcibly, the model is
useful because it is wrong. Clearly, the model must be sufficiently accurate for its intended purpose or its predictions will be useless. However, the fact that predictions are imperfect should be no surprise.
As a general rule, systems that are large, expensive, complex, and important justify
more complex and expensive models. Similarly, as the sophistication of the pipeline system increases, so do the benefits and advantages of the modeling approach because this
2.4
MODELING APPROACH
If we accept that we are going to construct computer models to predict the performance
of pipeline systems, then how should this be done? What aspects of the prototype can and
should be emphasized in the model? What is the basis of the approximations, and what
principles constrain the approach? These topics are discussed in this section.
Perhaps surprisingly, if we wish to model the behavior of any physical system, a
remarkably small number of fundamental relations are available (or required). In essence,
we seek to answer three simple questions: where?, what? and how? The following sections
provide elaboration.
The first question is resolved most easily. Because flow in a pipe system can almost
always be assumed to be one-dimensional, the question of where is resolved by assuming
a direction of flow in each link of pipe. This assumed direction gives a unique orientation
to the specification of distance, discharge, and velocity. Positive values of these variables
indicate flow in the assumed direction, whereas negative values indicate reverse flows.
The issues of what and how require more careful development.
2.4.1 Properties of Matter (What?)
The question of 'What?' directs our attention to the matter within the control system. In
the case of a hydraulic system, this is the material that makes up the pipe walls, or fills
the interior of a pipe or reservoir, or that flows through a pump. Eventually a modeller
must account for all these issues, but we start with the matter that flows, typically consisting mostly of water with various degrees of impurities.
In fact, water is so much a part of our lives that we seldom question its role. Yet water
possesses a unique combination of chemical, physical, and thermal properties that makes
it ideally suited for many purposes. In addition, although important regional shortages
may exist, water is found in large quantities on the surface of the earth. For both these reasons, water plays a central role in both human activity and natural processes.
One surprising feature of the water molecule is its simplicity, formed as it as from two
diatomic gases, hydrogen (H2) and oxygen (O2). Yet the range and variety of water's
properties are remarkable (Table 2.1 provides a partial list). Some property values in the
TABLE 2.1 Selected Properties of Liquid Water
Physical Properties
1. High densitypliq < 1 000 kg/m3
2. Density maximum at 40Ci.e., above freezing!
3. High viscosity (but a Newtonian fluid)JJL ~ 10~3 N s/m2
4. High surface tensionor = 73 N/m
5. High bulk modulus (usually assumed incompressible)K 2.07 GPa
Thermal Properties
1. Specific Heathighest except for NH3-C 4.187 kJ/(kg-C)
2. High heat of vaporizationcv 2.45 MJ/kg
3. High heat of fusioncf 0.36 MJ/kg
4. Expands on freezingin almost all other compounds, psolid > /9liq
5. High boiling pointc.f., H2 (20 K), O2 (90 K) and H2O (373 K)
6 Good conductor of heat relative to other liquids and nonmetal solids.
Chemical and Other Properties
1. Slightly ionizedwater is a good solvent for electrolytes and nonelectrolytes
2. Transparent to visible light; opaque to near infrared
3. High dielectric constantresponds to microwaves and electromagnetic fields
Note: The values are approximate. All the properties listed are functions of temperature, pressure, water purity, and
other factors that should be known if more exact values are to be assigned. For example, surface tension is greatly
influenced by the presence of soap films, and the boiling point depends on water purity and confining pressure. The
values are generally indicative of conditions near 1O0C and one atmosphere of pressure.
table especially density and viscosity valuesare used regularly by pipeline engineers.
Other properties, such as compressibility and thermal values, are used indirectly, primarily to justify modeling assumptions, such as the flow being isothermal and incompressible. Many properties of water depend on intermolecular forces that create powerful
attractions (cohesion) between water molecules. That is, although a water molecule is
electrically neutral, the two hydrogen atoms are positioned to create a tetrahedral charge
distribution on the water molecule, allowing water molecules to be held strongly together with the aid of electrostatic attractions. These strong internal forcestechnically called
'hydrogen bonds'arise directly from the non-symmetrical distribution of charge.
The chemical behavior of water also is unusual. Water molecules are slightly ionized,
making water an excellent solvent for both electrolytes and nonelectrolytes. In fact, water
is nearly a universal solvent, able to wear away mountains, transport solutes, and support
the biochemistry of life. But the same properties that create so many benefits also create
problems, many of which must be faced by the pipeline engineer. Toxic chemicals, disinfection byproducts, aggressive and corrosive compounds, and many other substances can
be carried by water in a pipeline, possibly causing damage to the pipe and placing consumers at risk.
Other challenges also arise. Water's almost unique property of expanding on freezing
can easily burst pipes. As a result, the pipeline engineer either may have to bury a line or
may need to supply expensive heat-tracing systems on lines exposed to freezing weather,
particularly if there is a risk that standing water may sometimes occur. Water's high viscosity is a direct cause of large friction losses and high energy costs whereas its vapor
properties can create cavitation problems in pumps, valves, and pipes. Furthermore, the
combination of its high density and small compressibility creates potentially dramatic
transient conditions. We return to these important issues after considering how pipeline
flows respond to various physical constraints and influences in the next section.
2.4.2 Laws of Conservation (How?)
Although the implications of the characteristics of water are enormous, no mere list of its
properties will describe a physical problem completely. Whether we are concerned with
water quality in a reservoir or with transient conditions in a pipe, natural phenomena also
obey a set of physical laws that contributes to the character and nature of a system's
response. If engineers are to make quantitative predictions, they must first understand the
physical problem and the mathematical laws that model its behavior.
Basic physical laws must be understood and be applied to a wide variety of applications and in a great many different environments: from flow through a pump to transient
conditions in a channel or pipeline. The derivations of these equations are not provided,
however, because they are widely available and take considerable time and effort to do
properly. Instead, the laws are presented, summarized, and discussed in the pipeline context. More precisely, a quantitative description of fluid behavior requires the application
of three essential relations: (1) a kinematic relation obtained from the law of mass conservation in a control volume, (2) equations of motion provided by both Newton's second
law and the energy equation, and (3) an equation of state adapted from compressibility
considerations, leading to a wavespeed relation in transient flow and justifying the
assumption of an incompressible fluid in most steady flow applications.
A few key facts about mass conservation and Newton's second law are reviewed
briefly in the next section. Consideration of the energy equation is deferred until steady
flow is discussed in more detail, whereas further details about the equation of state are
introduced along with considerations of unsteady flow.
(2.1)
where V1 represents the sum of all the water entering an area, and V0 indicates the total
volume of water leaving the same region. More commonly, however, a budget relation
such as Eq. 2.1 is written as a rate equation. Dividing the "balance" equation by At and
taking the limit as At goes to zero produces
(2-3)
Equation (2.3) states that the rate of flow into a control volume is equal to the rate of
outflow. This result is intuitively satisfying since no accumulation of mass or volume
should occur in any control volume under steady conditions. If the control volume were
taken to be the junction of a number of pipes, this law would take the form of Kirchhoff's
current lawthe sum of the mass flow in all pipes entering the junction equals the sum
of the mass flow of the fluid leaving the junction. For example, in Fig. 2.2, continuity for
the control volume of the junction states that
G1 + Q2 = C3 + Q4
2.4.4
(2-4)
When mass rates of flow are concerned, the focus is on a single component of chemical
species. However, when we introduce a physical law, such as Newton's law of motion, we
obtain something even more profound: a relationship between the apparently unrelated
quantities of force and acceleration.
More specifically, Newton's second law relates the changes in motion of a fluid or solid
to the forces that cause the change. Thus, the statement that the resultant of all external
forces, including body forces, acting on a system is equal to the rate of change of momentum of this system with respect to time. Mathematically, this is expressed as
x^
d(mv)
ix=v
(2 5)
where t is the time and Fext represents the external forces acting on a body of mass m
moving with velocity i). If the mass of the body is constant, Eq. (2.5) becomes
2Fex, = f = mfl
(2.6)
where a is the acceleration of the system (the time rate of change of velocity).
In closed conduits, the primary forces of concern are the result of hydrostatic pressure,
fluid weight, and friction. These forces act at each section of the pipe to produce the net
acceleration. If these forces and the fluid motion are modeled mathematically, the result
is a "dynamic relation" describing the transient response of the pipeline.
For a control volume, if flow properties at a given position are unchanging with time,
the steady form of the moment equation can be written as
2X*,= | PV(V n)dA
./CS
(2.7)
where the force term is the net external force acting on the control volume and the right
hand term gives the net flux of momentum through the control surface. The integral is
taken over the entire surface of the control volume, and the integrand is the incremental
amount of momentum leaving the control volume.
The control surface usually can be oriented to be perpendicular to the flow, and one
can assume that the flow is incompressible and uniform. With this assumption, the
momentum equation can be simplified further as follows:
2> = (P<4w)out - (pAvv)in = PG(vout - vin)
(2.8)
(2.9)
Fx = (PA)1 + PGu1
(2.10)
Thus,
In a similar manner, but taking y as positive upward, direct application of the momentum equation gives
(PA)2 - Fy = pQ( -V2)
(2.11)
(here the outflow gives a positive sign, but the velocity is in the negative direction). Thus,
Fy = (PA)2 + PGu2
(2.12)
In both cases, the reaction forces are increased above what they would be in the static case because the associated momentum must either be established or be eliminated
in the direction shown. Application of this kind of analysis is routine in designing thrust
blocks, which are a kind of anchor used at elbows or bends to restrain the movement of
pipelines.
outflow, the external forces acting on the flow are balanced by the changes in momentum,
and the external work is compensated for by losses of mechanical energy. As a result, the
fluid generally moves down an energy gradient, often visualized as flow in the direction
of decreasing hydraulic grade-line elevations (e.g., Fig. 2.1).
Quasi-steady flow. When the flow becomes unsteady, the resulting model that must
be used depends on how fast the changes occur. When the rate of change is particularly slow, typically over a period of hours or days, the rate of the fluids acceleration is
negligible. However, fluid will accumulate or be depleted at reservoirs, and rates of
demand for water may slowly adjust. This allows the use of a quasi-steady or extended-duration simulation model.
Compressible and Incompressible. If the density of the fluid p is constantboth in
time and throughout the flow fielda flow is said to be incompressible. Thus, p is not a
function of position or time in an incompressible flow. If changes in density are permitted or required, the flow is compressible.
Surge. When the rate of change in flow is moderate, typically occurring over a period
of minutes, a surge model is often used. In North America, the term surge indicates an
analysis of unsteady flow conditions in pipelines when the following assumptions are
made: the fluid is incompressible (thus, its density is constant) and the pipe walls are rigid
and do not deform. These two assumptions imply that fluid velocities are not a function
of position along a pipe of constant cross-section and the flow is uniform. In other words,
no additional fluid is stored in a length of pipe as the pressure changes; because velocities
are uniform, the rate at which fluid enters a pipe is always equal to the rate of discharge.
However, the acceleration of the fluid and its accumulation and depletion from reservoirs
are accounted for in a surge model.
Waterhammer. When rapid unsteady flow occurs in a closed conduit system, the transient condition is sometimes marked by a pinging or hammering noise, appropriately
called waterhammer. However, it is common to refer to all rapidly changing flow conditions by this term, even if no audible shock waves are produced. In waterhammer models,
it is usually assumed that the fluid is slightly compressible, and the pipe walls deform with
changes in the internal pressure. Waterhammer waves propagate with a finite speed equal
to the velocity of sound in the pipeline.
The speed at which a disturbance is assumed to propagate is the primary distinction
between a surge and a waterhammer model. Because the wavespeed parameter a is related to fluid storage, the wavespeed is infinite in surge or quasi-steady models. Thus, in
effect, disturbances are assumed to propagate instantly throughout the pipeline system. Of
course they do no such thing, because the wavespeed is a finite physical property of a pipe
system, much like its diameter, wall thickness, or pipeline material. The implication of
using the surge or quasi-steady approximation is that the unsteady behavior of the pipe
system is controlled or limited by the rate at which the hydraulic boundary conditions
(e.g., pumps, valves, reservoirs) at the ends of the pipe respond to the flow and that the
time required for the pipeline itself to react is negligible by comparison.
Although unsteady or transient analysis is invariably more involved than is steadystate modeling, neglecting these effects in a pipeline can be troublesome for one of two
reasons: the pipeline may not perform as expected, possibly causing large remedial
expenses, or the line may be overdesigned with respect to transient conditions, possibly
causing unnecessarily large capital costs. Thus, it is essential for engineers to have a clear
physical grasp of transient behavior and an ability to use the computer's power to maximum advantage.
One interesting point is that as long as one is prepared to assume the flow is compressible, the importance of compressibility does not need to be known a priori. In fact,
all the incompressible, quasi-steady, and steady equations are special cases of the full transient equations. Thus, if the importance of compressibility or acceleration effects is
unknown, the simulation can correctly assume compressible flow behavior and allow the
analysis to verify or contradict this assumption.
Redistribution of water, whatever model or physical devices are used, requires control
of the fluid and its forces, and control requires an understanding not only of physical law
but also of material properties and their implications. Thus, an attempt to be more specific and quantitative about these matters will be made as this chapter progresses.
In steady flow, the fluid generally moves in the direction of decreasing hydraulic
grade-line elevations. Specific devices, such as valves and transitions, cause local pressure drops and dissipate mechanical energy; operating pumps do work on the fluid and
increase downstream pressures while friction creates head losses more or less uniformly
along the pipe length. Be warned, howeverin transient applications, this orderly situation rarely exists. Instead, large and sudden variations of both discharge and pressure can
occur and propagate in the system, greatly complicating analysis.
2.6
STEADYFLOW
The design of steady flow in pipeline systems has two primary objectives. First, the
hydraulic objective is to secure the desired pressure and flow rate at specific locations in
the system. Second, the economic objective is to meet the hydraulic requirements with the
minimum expense.
When a fluid flows in a closed conduit or open channel, it often experiences a complex interchange of various forms of mechanical energy. In particular, the work that is
associated with moving the fluid through pressure differences is related to changes in both
gravitational potential energy and kinetic energy. In addition, the flow may lose mechanical energy as a result of friction, a loss that is usually accounted for by extremely small
increases in the temperature of the flowing fluid (that is, the mechanical energy is converted to thermal form).
More specifically, these energy exchanges are often accounted for by using an
extended version of Bernoulli's famous relationship. If energy losses resulting from friction are negligible, the Bernoulli equation takes the following form:
H+*-?+!+*
where P1 and/?2 are the pressures at the end points, y is the specific weight of the fluid, V1
and V2 are the average velocities at the end points, and Z1 and Z2 are the elevations of the
end points with respect to an arbitrary vertical datum. Because of their direct graphical
representation, various combinations of terms in this relationship are given special labels,
historically called heads because of their association with vertical distances. Thus,
Head
Definition
Associated with
Pressure head
Elevation head
Velocity head
Piezometric head
Total head
p/j
z
v2/2g
p/j + z
p/j + z + v2/2g
Flow work
Gravitational potential energy
Kinetic energy
Pressure H- elevation head
Pressure + elevation + velocity head
Next Page
CHAPTER 3
HYDRAULICS OF OPEN
CHANNEL FLOW
Richard H. French
Desert Research Institute,
University and Community College System of Nevada
Reno, Nevada
3.7
INTRODUCTION
By definition, an open channel is a flow conduit having a free surface: that is, a boundary
exposed to the atmosphere. The free surface is essentially an interface between two fluids
of different density. Open-channel flows are almost always turbulent, unaffected by surface tension, and the pressure distribution within the fluid is hydrostatic. Open channels
include flows ranging from rivulets flowing across a field to gutters along residential
streets and highways to partially filled closed conduits conveying waste water to irrigation and water supply canals to vital rivers.
In this chapter, the basic principles of open channel hydraulics are presented as an
introduction to subsequent chapters dealing with design. By necessity, the material presented in this chapter is abbreviatedan abstract of the fundamental concepts and
approachesfor a more detailed treatment, the reader is referred to any standard references or texts dealing with the subject: for example, Chow (1959), French (1985),
Henderson (1966), or Chaudhry (1993)
As with any other endeavor, it is important that a common vocabulary be established
and used:
Critical slope (Sc): A longitudinal slope such that uniform flow occurs in a critical
state.
Flow area (A): The flow area is the cross-sectional area of the flow taken normal to
the direction of flow (Table 3.1).
Froude number (Fr): The Froude number is the dimensionless ratio of the inertial and
gravitational forces or
Fr = =
(3.1)
VgD
where V = average velocity of flow, g = gravitational acceleration, and D = hydraulic
depth. When Fr = 1, the flow is in a critical state with the inertial and gravitational forces
in equilibrium; when Fr < 1, the flow is in a subcritical state and the gravitational forces
are dominant; and when Fr > 1, the flow is in a supercritical state and the inertial forces
are dominant. From a practical perspective, sub - and supercritical flow can be differentiated simply by throwing a rock or other object into the flow. If ripples from the rock
Rectangle
Trapezoid
with
equal side
slopes
Trapezoid
with
unequal
side slopes
Triangle
with equal
side slopes
Triangle
with
unequal
side slopes
Circular
Area
Wetted Perimeter
Hydraulic Radius
Top Width
Hydraulic Depth
progress upstream of the point of impact, the flow is subcritical; however, if ripples from
the rock do not progress upstream but are swept downstream, the flow is supercritical.
Hydraulic depth (D). The hydraulic depth is the ratio of the flow area (A) to the top
width (T)OtD= A/T (Table 3.1).
Hydraulic radius (R). The hydraulic radius is the ratio of the flow area (A) to the wetted perimeter (P) or R = A/P (Table 3.1).
Kinetic energy correction factor (a). Since no real open-channel flow is one-dimensional, the true kinetic energy at a cross section is not necessarily equal to the spatially
averaged energy. To account for this, the kinetic energy correction factor is introduced,
or
K^) "-"^*
-^
<
When the flow is uniform, a = 1 and values oc of for various situations are summarized
in Table 3.2.
Momentum correction coefficient (ft): Analogous to the kinetic energy correction factor, the momentum correction factor is given by
PpgV = IJpV2JA
RP = Jlv2JA
-^4-
n
^
(33)
When the flow is uniform, P = I and values of (3 for various situations are summarized
in Table 3.2
Prismatic channel. A prismatic channel has both a constant cross-sectional shape and
bottom slope (S0). Channels not meeting these criteria are termed nonprismatic.
Specific energy (E). The specific energy of an open-channel flow is
E = y + a^
(3.4)
where y = depth of flow and the units of specific energy are length in meters or feet.
TABLE 3.2 Typical Values of a and p for Various Situations
Situation
Min.
Regular channels,
flumes,
spillways
Natural streams and torrents
Rivers under ice cover
River valleys, overflooded
Source: After Chow (1959).
Value of oc
Avg. Max.
Min.
Value of $
Avg. Max.
1.10
1.15
1.20
1.03
1.05
1.07
1.15
1.20
1.50
1.30
1.50
1.75
1.50
2.00
2.00
1.05
1.07
1.17
1.10
1.17
1.25
1.17
1.33
1.33
M = ^- + ~zA
gA
(3.5)
Stage: The stage of a flow is the elevation of the water surface relative to a datum. If
the lowest point of a channel section is taken as the datum, then the stage and depth of
flow (y) are equal if the longitudinal slope (S0) is not steep or cos (0) ~ 1, where 0 is the
longitudinal slope angle. If 0 < 10 or S0 ^ 0.18, where S0 is the longitudinal slope of the
channel, then the slope of the channel can be assumed to be small.
Steady. The depth (y) and velocity of flow (v) at a location do not vary with time; that
is, (dy/dt = O) and (3v/9r = O). In unsteady flow, the depth and velocity of flow at a location vary with time: that is, (3y/9/ ^ O) and (3v/9r ^ O).
Top width (T). The top width of a channel is the width of the channel section at the
water surface (Table 3.1).
Uniform flow. The depth (y). flow area (A), and velocity (V) at every cross section are
constant, and the energy grade line (Sy), water surface, and channel bottom slopes (S0) are
all parallel.
Superelevation (Ay). The rise in the elevation of the water surface at the outer channel boundary above the mean depth of flow in an equivalent straight channel, because of
centrifugal force in a curving channel.
Wetted perimeter (P). The wetted perimeter is the length of the line that is the interface between the fluid and the channel boundary (Table 3.1).
3.2
ENERGYPRINCIPLE
p
V2
H = Z + - + oc
Y
2g
where H = total energy, z = elevation of the streamline above a datum, p = pressure, y
= fluid specific weight, (p/y) = pressure head, V1IZg = velocity head, and g = acceleration of gravity. H defines the elevation of the energy grade line, and the sum [z + (p//)]
defines the elevation of the hydraulic grade line. In most uniform and gradually varied
flows, the pressure distribution is hydrostatic (divergence and curvature of the streamlines
is negligible) and the sum [z + (p/y)] is constant and equal to the depth of flow y if the
datum is taken at the bottom of the channel. The specific energy of an open-channel flow
relative to the channel bottom is
E=
y+ a^ = y + a^v
(3 6)
V=|
(3.7)
Depth of Flow
y
f*-
It can be shown that dA = (T = dy) or (dA/dy = T) (French, 1985). Substituting this
result, using the definition of hydraulic depth and rearranging, Eq. (3.8) becomes
Subcritlcal
Decreasing Increasing
Supercritical
Specific Energy
Specific Momentum
FIGURE 3.1 Specific energy and momentum as a function of depth when the channel geometry and flow rate are specified.
!_
CLf^
_-eLZ: _-^ o
1
gA^dy = 1l gA*A =l 1 gD= U
or
V2
D
Tg = T
<3-9>
and
-= = Fr=l
VgD
(3.10)
which is the definition of critical flow. Therefore, minimum specific energy occurs at the
critical hydraulic depth and is the minimum energy required to pass the flow Q. With this
information, the portion of the curve AC in Fig. 3.1 is interpreted as representing supercritical flows, where as AB represents subcritical flows.
With regard to Fig. 3.1 and Eq. (3.6), the following observations are pertinent. First,
for channels with a steep slope and a * 1, it can be shown that
Fr=
V
IgD COs(Q)
V
a
(3.11)
Second, E - y curves for flow rates greater than Q lie to the right of the plotted curve,
and curves for flow rates less than Q lie to the left of the plotted curve. Third, in a rectangular channel of width b, y = D and the flow per unit width is given by
-
(3-12)
and
V=-
(3.13)
Then, where the subscript c indicates variable values at the critical point,
HT)"
(3 14)
f=y
(3 15)
yc = \Ee
(3.16)
and
In nonrectangular channels when the dimensions of the channel and flow rate are specified, critical depth is calculated either by the trial and error solution of Eqs. (3.8), (3.9),
and (3.10) or by use of the semiempirical equations in Table 3.3.
3.2.3 Variation of Depth with Distance
At any cross section, the total energy is
# = ^ = y+z
(3.17)
Equation for yc
in terms of
V = a Q*/g
(2)
^ t a n g l e f y ] 0.33
(V)
T r a P e Z O l d
27
b
0 8 U-*
l 1 f[ ^0.75*
fcl.25]
J - WZ
Triangle
(WV2Q
U2J
Circle
(MLV*
UH
Source: From Straub (1982).
where y = depth of flow, z = elevation of the channel bottom above a datum, and it is
assumed that a and cos(6) are both equal to 1. Differentiating Eq. (3.17) with respect to
longitudinal distance,
dH _ \2g)
~d^~ dx
dy
~dx
dz
^x
n 1R ,
(J 18)
'
dH
dz
where j = the change of energy with longitudinal distance (Sf), = the channel botx
tom slope (S0), and, for a specified flow rate,
d
(3.19)
which describes the variation of the depth of flow with longitudinal distance in a channel
of arbitrary shape.
Section
(3.20)
A2
where K1 and A1. as follows the conveyance and area of the /th channel subsection, respectively, K and A are conveyance are as follows:
N
K=^K1
i =1
and
N
A = EAI=1
N = number of subsections, and conveyance (K) is defined by Eq. (3.48) in Sec. 3.4.
Equation (3.20) is based on two assumptions: (1) the channel can be divided into subsections by appropriately placed vertical lines (Fig. 3.2) that are lines of zero shear and do
not contribute to the wetted perimeter of the subsection, and (2) the contribution of the
nonuniformity of the velocity within each subsection is negligible in comparison with the
variation in the average velocity among the subsections.
3.3
MOMENTUM
(3.21)
~^)-(K
^)
or
-^ = M1-M2
(3.22)
where
M1 = ^- + Z1A1
(3.23)
&Aand M is known as the specific momentum or force function. In Fig. 3.1, specific momentum is plotted with specific energy for a specified flow rate and channel section as a function of the depth of flow. Note that the point of minimum specific momentum corresponds
to the critical depth of the flow.
The classic application of Eq. (3.22) occurs when Pf = O and the application of the resulting equation to the estimation of the sequent depths of a hydraulic jump. Hydraulic jumps
result when there is a conflict between the upstream and downstream controls that influence
the same reach of channel. For example, if the upstream control causes supercritical flow
while the downstream control dictates subcritical flow, there is a contradiction that can be
resolved only if there is some means to pass the flow from one flow regime to the other.
When hydraulic structures, such as weirs, chute blocks, dentated or solid sills, baffle piers,
and the like, are used to force or control a hydraulic jump, Pf in Eq. (3.22) is not equal to zero.
Finally, the hydraulic jump occurs at the point where Eq. (3.22) is satisfied (French, 1985).
3.3.2 Hydraulic Jumps in Rectangular Channels
In the case of a rectangular channel of width b and Pf = O, it can be shown (French,
1985) that
^- - V0.5 [1 + 8(Fr1)* - 1]
y\
(3.24)
(3.25)
or
3.3.3
(3.27)
yl.8
Fr1 > 1.Iy2 = 73
(3.28)
For horizontal triangular and parabolic prismatic channel sections, Silvester (1964,
1965) presented the following equations.
For triangular channels:
(^J-5- 1 = 1.5 W [l -(U)
(3.29)
Value of Fr1
FIGURE 3.3 Analytic curves for estimating sequent depths in a trapezoidal channel
(From Silvester, 1964)
(3.30)
In the case of trapezoidal channels, Silvester (1964) presented a method for graphical
solution in terms of the parameter
*=
(3.31)
3.4
UNIFORMFLOW
(3.32)
(3.33)
where C is the Chezy resistance coefficient. It can be easily shown that Eq (3.33) is similar in form to the Darcy pipe flow equation. In 1889, Robert Manning, a professor at the
Royal College of Dublin (Levi, 1995) proposed what has become known as Manning's
equation, or
y = i R^Vs
(3.34)
where n is Manning's resistance coefficient and <|> = 1 if SI units are used and <|) = 1.49
if English units are used. The relationship among C, n, and the Darcy-Weisbach friction
factor (D is
C
= jK = y*
(3.35)
At this point, it is pertinent to observe that n is a function of not only boundary roughness and the Reynolds number but also the hydraulic radius, an observation that was made
by Professor Manning (Levi, 1995).
3.4.2 Estimation of Manning's Resistance Coefficient
Of the two equations for estimating the velocity of a uniform flow, Manning's equation is
the more popular one. A number of approaches to estimating the value of n for a channel
are discussed in French (1985) and in other standard references, such as Barnes (1967),
Urquhart (1975), and Arcement and Schneider (1989). Appendix 3.A lists typical values
of n for many types of common channel linings.
In an unvegetated alluvial channel, the total roughness consists of two parts: grain or
skin roughness resulting from the size of the sediment particles and form roughness
because of the existence of bed forms. The total coefficient n can be expressed as
/i = /i' + n"
(3.36)
where n' = portion of Manning's coefficient caused by grain roughness and n" = portion
of Manning's coefficient caused by form roughness. The value of n' is proportional to the
diameter of the sediment particles to the sixth power. For example, Lane and Carlson
(1953) from field experiments in canals paved with cobbles with d15 in inches, developed
/T = 0.026^f
(3.37)
and Meyer-Peter and Muller (1948) for mixtures of bed material with a significant proportion of coarse-grained sizes with J90 in meters developed
n' = 0.038J91O6
(3.38)
In both equations, dxx is the sediment size such that xx percent of the material is smaller
by weight.
Although there is no reliable method of estimating n", an example of the variation of
/for the 0.19 mm sand data collected by Guy et al. (1966) is shown in Fig. 3.4. The n values commonly found for different bed forms are summarized in Table 3.4. The inability
to estimate or determine the variation of form roughness poses a major problem in the
study of alluvial hydraulics (Yang, 1996).
Use of Manning's equation to estimate the velocity of flow in channels where the primary component of resistance is from drag rather than bed roughness has been questioned
(Fischenich, 1996). However, the use of Manning's equation has persisted among engineers
because of its familiarity and the lack of a practical alternative. Jarrett (1984) recognized that
Grain Roughness
guidelines for estimating resistance coefficients for high-gradient streams with stable beds
composed of large cobbles and boulders and minimally vegetated banks (S0 > 0.002) were
based on limited data. Jarrett (1984) examined 21 high-gradient streams in the Rocky
Mountains and developed the following empirical equation relating n to S0 and R (in feet):
= ^^
flO.16
(3.39)
Jarrett (1984) stated the following limitations on the use of Eq. (3.39): First, the equation is applicable to natural main channels with stable bed and bank materials (gravels,
cobbles, boulders) with no backwater. Second, the equation can be used for 0.002 < S0 ^
0.04 and 0.15 < / ? < 2.1 m (0.5 < R < 7.0 ft). Results of the regression analysis indicated that for R > 2.1 m ( 7.0 ft), n did vary significantly with depth; therefore, as long as
the bed and bank material remain stable, extrapolation to larger flows should not result in
significant error. Third, the hydraulic radius does not include the wetted perimeter of the
bed particles. Fourth, the streams used in the analysis had relatively small amounts of suspended sediment.
Vegetated channels present unique challenges from the viewpoint of estimating roughness. In grass-lined channels, the traditional approach assumed that n was a function of
vegetal retardance and VR (CoyIe, 1975). However, there are approaches more firmly
based on the principles of fluid mechanics and the mechanics of materials (Kouwen, 1988;
Kouwen and Li, 1980.) Data also exist that suggest that in such channels flow duration is
not a factor as long as the vegetal elements are not destroyed or removed. Further, inundation times, and/or hydraulic stresses, or both that are sufficient to damage vegetation
have been found, as might be expected, to reduce the resistance to flow (Temple, 1991).
Petryk and Bosmajian (1975) presented a relation for Manning's n in vegetated channels based on a balance of the drag and gravitational forces, or
n = ^[C^]'2
(3.40)
where Cd a coefficient accounting for the drag characteristics of the vegetation and (Veg)rf
the vegetation density. Flippin-Dudley (1997) has developed a rapid and objective
procedure using a horizontal point frame to measure (Veg)d . Equation (3.40) is limited
because there is limited information regarding Cd for vegetation (Flippin-Dudley
et aL, 1997).
3.4.3 Equivalent Roughness Parameter k
In some cases, an equivalent roughness parameter k is used to estimate n. Equivalent
roughness, sometimes called "roughness height," is a measure of the linear dimension
of roughness elements but is not necessarily equal to the actual or even the average height
of these elements. The advantage of using k instead of Manning's n is that k accounts for
changes in the friction factor due to stage, whereas the Manning's n does not. The relationship between n and k for hydraulically rough channels is
*=
6/?1/6
7
TH
YlogJ 12.2K
V
J
<3-41)
k
(ft)
(2)
(1)
k
M
(3)
0.00003048-0.0009
0.0001-0.0024
0.0001-0.0021
0.0002-0.0046
0.0002-0.0055
0.0002-0.0009
0.0004-0.0012
0.0005-0.0030
0.0030-0.0101
0.0006-0.0030
0.0009-0.0091
0.0061
0.0030
0.0305-0.9144
3.4.4
In many designed channels and most natural channels, roughness varies along the perimeter of the channel, and it is necessary to estimate an equivalent value of n for the entire
perimeter. In such cases, the channel is divided into TV parts, each with an associated wetted perimeter (P1), hydraulic radius (R1), and roughness coefficient (n.), and the equivalent
roughness coefficient (ne) is estimated by one of the following methods. Note that the wetted perimeter does not include the imaginary boundaries between the subsections.
1. Horton (1933) and Einstein and Banks (1950) developed methods of estimating ne
assuming that the average velocity in each of the subdivisions is the same as the
average velocity of the total section. Then
ne =
N ,
\ ~|2/3
sM
i^Lj
(3.42)
2. Assuming that the total force resisting motion is equal to the sum of the subsection
resisting forces,
N
~|l/2
E
w>
ne = i^^Jt
(3.43)
3. Assuming that the total discharge of the section is equal to the sum of the subsection discharges,
PR
ne =
(3.44)
Y ^L!
~i "i
4. Weighting of resistance by area (Cox, 1973),
AT
E A
ne =-^~
(3.45)
"^V'V-}
(3 46)
occurs in a critical state can be found: that is, a slope such that normal flow occurs with
Fr = 1. The slope obtained is the critical slope, but it also is a normal slope. The smallest
critical slope, for a specified channel shape, roughness, and discharge is termed the limiting critical slope. The critical slope for a given normal depth is
N
S*c = 8n2
pRjp
(3
49)
(J-W)
where the subscript Af indicates the normal depth value of a variable and, for a wide channel,
gn2
(3 50)
*=^
3.4.6.2 Sheet/low. A special but noteworthy uniform flow condition is that of sheetflow. From the viewpoint of hydraulic engineering, a necessary condition for sheetflow is
that the flow width must be sufficiently wide so that the hydraulic radius approaches the
depth of flow. With this stipulation, the Manning's equation, Eq. (3.48), for a rectangular
channel becomes
Q = ^Ty^ VS
(3.51)
where T = sheetflow width and yN = normal depth of flow. Then, for a specified flow rate
and sheetflow width, Eq. (3.51) can be solved for the depth of flow, or
( nO >/s
(3 52)
^ = UFW/
The condition that the value of the hydraulic radius approaches the depth of flow is not
a sufficient condition. That is, this condition specifies no limit on the depth of flow, and
there is general agreement that sheetflow has a shallow depth of flow. Appendix 3.A summarizes Manning's n values for overland and sheetflow.
3.4.6.3 Superelevation. When a body of water moves along a curved path at constant
velocity, it is acted for a force directed toward the center of the curvature of the path.
When the radius of the curve is much larger than the top width of the water surface, it can
be shown that the rise in the water surface at the outer channel boundary above the mean
depth of flow in a straight channel (or superelevation) is
V2T
A, = _
(3.53)
where r = the radius of the curve (Linsley and Franzini, 1979). It is pertinent to note that
if the effects of the velocity distribution and variations in curvature across the channel are
considered, the superelevation may be as much as 20 percent more than that estimated by
Eq. (3.53) (Linsley and Franzini, 1979). Additional information regarding superelevation
is available in Nagami et al., (1982) and U.S. Army Corps of Engineers (USAGE, 1970).
3.5
3.5.1 Introduction
The gradual variation in the depth of flow with longitudinal distance in an open channel
is given by Eq. (3.19), or
dy _ S0 - Sf
dx
1 - Fr2
and two cases warrant discussion. In the first case, because the distance over which the
change in depth is short it is appropriate to assume that boundary friction losses are small,
or Sf = O. When this is the case, important design questions involve abrupt steps in the
bottom of the channel (Fig. 3.5) and rapid expansions or contractions of the channel
(Fig. 3.6). The second case occurs when Sf# O.
(3 54)
- Fr2)l+ Tx =
'
and the following observations are pertinent (the observations also apply to channels of
arbitrary shape):
1. If dz/dx > O (upward step) and Fr < 1, then dy/dx must be less than zerodepth
of flow decreases as x increases.
2. If dz/dx > O (upward step) and Fr > 1, then dy/dx must be greater than zerodepth
of flow increases as x increases.
Profile
Profile
FIGURE 3.5 Definition of variables for gradually varied flow over positive and negative steps.
Plan
FIGURE 3.6 Definition of variables for gradually varied flow through contracting
and expanding channel sections.
3. If dz/dx < O (downward step) and Fr < 1, then dy/dx must be greater than zero
depth of flow increases as x increases.
4. If dz/dx < O (downward step) and Fr > 1, then dy/dx must be less than zerodepth
of flow decreases as x increases.
In the case of a channel of constant width with a positive or negative step, the relation
between the specific energy upstream of the step and the specific energy downstream of
the step is
E1 = E2 + Az
(3.55)
In the case dz/dx = O, if the channel is rectangular in shape but the width of the channel changes, it can be shown (French, 1985) that the governing equation is
(1 - Fr^ - Fr*J^ = O
dx
b dx
(3.56)
3. If db/dx < O (width decreases) and Fr < 1, then dyldx must be less than zero
depth of flow decreases as x increases.
4. If db/dx < O (width decreases) and Fr > 1, then dyldx must be greater than zero
depth of flow increases as x increases.
In this case, the relation between the specific energy upstream of the contraction
(expansion) and the specific energy downstream of the step contraction (expansion) is
E1 = E2
(3.57)
It is pertinent to note that in the case of supercritical flow, channel expansions and contractions may result in the formation of waves.
Additional information regarding steps, expansions, and contractions can be found in any
standard reference or text on open-channel hydraulics (e.g., French, 1985).
3.5.3 Gradually Varied Flow with Sf O
In the case where Sf cannot be neglected, the water surface profile must estimated. For a
channel of arbitrary shape, Eq. (3.19) becomes
dy
_S0-Sf
_tfQ2p*3
1
~gA3"
^ gA3
For a specified value of Q, Fr and Sf are functions of the depth of flow y. For illustrative purposes, assume a wide channel; in such a channel, Fr and Sf will vary in much the
same way with y since P-T and both Sf and Fr have a strong inverse dependence on the
flow area. In addition, as y increases, both Sf and Fr decrease. By definition, Sf = S0 when
y = yN. Given the foregoing, the following set of inequalities must apply:
Sf > S0
for
y < yN
Fr > 1
for
y < yc
Sf<S0
for
y>yN
Fr < 1
for
y > yc
and
These inequalities divide the channel into three zones in the vertical dimension. By convention, these zones are labeled 1 to 3 starting at the top. Gradually varied flow profiles
are labeled according to the scheme defined in Table 3.5.
For a channel of arbitrary shape, the standard step methodology of calculating the gradually varied flow profile is commonly used: for example, HEC-2 (USAGE, 1990) or HECRAS (USAGE, 1997). The use of this methodology is subject to the following assumptions: (1) steady flow, (2) gradually varied flow, (3) one-dimensional flow with correction
for the horizontal velocity distribution, (4) small channel slope, (5) friction slope (averaged) constant between two adjacent cross sections, and (6) rigid boundary conditions.
The application of the energy equation between the two stations shown in Fig. 3.7 yields
Zone
1
Zone
2
Zone
3
(1)
(2)
(3)
(4)
Mild
O < S0 < Sc
Ml
y>yN>yc
Ml
yN>y>yc
M3
Critical
S0 = SC>0
Cl
yN>yc>y
y>yc = yN
C2
y = yN = yc
C3
Steep
S0 > Sc > O
Relation ofy
to yN and
yc
(5)
Sl
yc = yN>y
y>yc>yN
52
yc>y>yN
Type
of
Curve
(6)
Backwater
(dy/dx > O)
Drawdown
(dy/dx < O)
Backwater
(dy/dx > O)
Backwater
(dy/dx > O)
Parallel to
channel bottom
(dy/dx = O)
Backwater
(dy/dx > O)
Backwater
(dy/dx > O)
Drawdown
(dy/dx < O)
Type
of
Flow
(7)
Subcritical
Subcritical
Supercritical
Subcritical
Uniform
critical
Supercritical
Subcritical
Supercritical
TABLE 3.5:
(Continued)
Channel
Slope
Zone
1
Zone
2
Zone
3
(1)
(2)
(3)
Horizontal
(4)
Relation ofy
to yN and
yc
(5)
Type
of
Curve
(6)
Type
of
Flow
(7)
S3
yc>yN>y
Backwater
(dy/dx > O)
Supercritical
yN>y>yc
Drawdown
(dy/dx < O)
Backwater
(dy/dx > O)
Subcritical
Drawdown
(dy/dx < O)
Backwater
(dy/dx > O)
Subcritical
None
So =
m
m
Adverse
S0 < O
yN>yc>y
Supercritical
None
A2
yN>y>yc
A3
)w > X: > y
Supercritical
Datum
Section 1
Section 2
yi
V2
Z1 + Gi1^- = z2 + a2^ + hf+he
(3.61)
where Z1 and Z2 = elevation of the water surface above a datum at Stations 1 and 2, respectively, he = eddy and other losses incurred in the reach, and hf = reach friction loss.
The friction loss can be obtained by multiplying a representative friction slope, Sf, by
the length of the reach, L. Four equations can be used to approximate the friction loss
between two cross sections:
Sf
( Q. + Q2 >
~v
v~ (average conveyance)
[K1 + K2 )
(3.62)
S +S
Sf = ^ (average friction slope)
(3.63)
2S S
Sf = ^ ,f (harmonic mean friction slope)
Sn + Sj2
(3.64)
(3.65)
and
The selection of a method to estimate the friction slope in a reach is an important decision and has been discussed in the literature. Laurenson (1986) suggested that the "true"
friction slope for an irregular cross section can be approximated by a third-degree poly-
nomial. He concluded that the average friction slope method produces the smallest maximum error, but not always the smallest error, and recommended its general use along with
the systematic location of cross sections. Another investigation based on the analysis of
98 sets of natural channel data showed that there could be significant differences in the
results when different methods of estimating the friction slope were used (USAGE, 1986).
This study also showed that spacing cross sections 15Om (500 ft) a part eliminated the differences.
The eddy loss takes into account cross section contractions and expansions by multiplying the absolute difference in velocity heads between the two sections by a contraction
or expansion coefficient, or
h =c
*^Ts~^Tg
(3 66)
'
There is little generalized information regarding the value of the expansion (C6) or the
contraction coefficient (Cc). When the change in the channel cross section is small, the
coefficients C6 and Cc are typically on the order of 0.3 and 0.1, respectively (USAGE,
1990). However, when the change in the channel cross section is abrupt, such as at
bridges, C6 and Cc may be as high as 1.0 and 0.6, respectively (USAGE, 1990).
With these comments in mind,
H1= Z1+ U1^
(3.67)
H2 = z2 + Oc2 ^J
(3.68)
and
(3.69)
Eq. (3.69) is solved by trial and error: that is, assuming H2 is known and given a longitudinal distance, a water surface elevation at Station 1 is assumed, which allows the
computation of H1 by Eq. (3.67). Then, hf and he are computed and H1 is estimated by
Eq. (3.67). If the two values of H1 agree, then the assumed water surface elevation at
Station 1 is correct.
Gradually varied water surface profiles are often used in conjunction with the peak
flood flows to delineate areas of inundation. The underlying assumption of using a steady
flow approach in an unsteady situation is that flood waves rise and fall gradually. This
assumption is of course not valid in areas subject to flash flooding such as the arid and
semiarid Southwestern United States (French, 1987).
In summary, the following principles regarding gradually varied flow profiles can be
stated:
1. The sign of dy/dx can be determined from Table 3.6.
2. When the water surface profile approaches normal depth, it does so asymptotically.
3. When the water surface profile approaches critical depth, it crosses this depth at a
large but finite angle.
4. If the flow is subcritical upstream but passes through critical depth, then the feature
that produces critical depth determines and locates the complete water surface profile. If the upstream flow is supercritical, then the control cannot come from the
downstream.
5. Every gradually varied flow profile exemplifies the principle that subcritical flows
are controlled from the downstream while supercritical flows are controlled from
upstream. Gradually varied flow profiles would not exist if it were not for the
upstream and downstream controls.
6. In channels with horizontal and adverse slopes, the term "normal depth of flow"
has no meaning because the normal depth of flow is either negative or imaginary.
However, in these cases, the numerator of Eq. (3.60) is negative and the shape of
the profile can be deduced.
Any method of solving a gradually varied flow situation requires that cross sections be
defined. Hoggan (1989) provided the following guidelines regarding the location of cross
sections:
1. They are needed where there is a significant change in flow area, roughness, or longitudinal slope.
2. They should be located normal to the flow.
3. They should be located in detailupstream, within the structure, and downstreamat structures such as bridges and culverts. They are needed at all control structures.
4. They are needed at the beginning and end of reaches with levees.
5. They should be located immediately below a confluence on a main stem and immediately above the confluence on a tributary.
6. More cross sections are needed to define energy losses in urban areas, channels
with steep slopes, and small streams than needed in other situations.
7. In the case of HEC-2, reach lengths should be limited to a maximum distance of
0.5 mi for wide floodplains and for slopes less than 38,550 m (1800 ft) for slopes
equal to or less than 0.00057, and 370 m (1200 ft) for slopes greater than 0.00057
(Beaseley, 1973).
3.6
(3.7Oa)
3v
^v
Tx + 8irx-8(S-s')
+v
=0
(3 71a)
dr
a^) = 0
dJC
and
I*+vav
g 3f
g 3*
*
dx
By rearranging terms, Eq. (3.7Ib) can be written to indicate the significance of each
term for a particular type of flow, or
\
-^
^fS=S ^-"steady
~\ - f ^ l ^
~J lJ steady, nonuniform - f -p1
/)f Iunsteady, nonuniform
(372)
\^''^)
Equations (3.70) and (3.71) compose a group of gradually varied unsteadyflowmodels that are termed complete dynamic models. Being complete, this group of models can
provide accurate results; however, in many applications, simplifying assumptions regarding the relative importance of various terms in the conservation of momentum equation
(Eq. 3.71) leads to other equations, such the kinematic and diffusive wave models (Ponce,
1989).
The governing equation for the kinematic wave model is
f
^f = 0
(3.73)
where e = a coefficient whose value depends on the frictional resistance equation used (e
= 5/3 when Manning's equation is used). The kinematic wave model is based on the equation of continuity and results in a wave being translated downstream. The kinematic wave
approximation is valid when
t SV
-^- > 85
(3.74)
(3 75)
where the left side of the equation is the kinematic wave model and the right side accounts
for the physical diffusion in a natural channel. The diffusion wave approximation is valid
when (Ponce, 1989),
**So (y) 0 ' 5 * !5
(3.76)
If the foregoing dimensionless inequalities ( Eq. 3.74 and 3.76) are not satisfied, then the
complete dynamic wave model must be used. A number of numerical methods can be used
to solve these equations (Chaudhry, 1987; French; 1985, Henderson, 1966; Ponce, 1989).
(3.77)
Since there are unknown losses associated with the wave, the momentum equation
rather than the energy equation is applied between Sections 1 and 2 or
FIGURE 3.8 Definition of variables for simple surges moving in an open channel.
(3.79)
When the slope of a channel becomes very steep, the resulting supercritical flow at
normal depth may develop into a series of shallow water waves known as roll waves. As
these waves progress downstream, they eventually break and form hydraulic bores or
shock waves. When this type of flow occurs, the increased depth of flow requires
increased freeboard, and the concentrated mass of the wavefronts may require additional
structural factors of safety.
Escoffier (1950) and Escoffier and Boyd (1962) considered the theoretical conditions
under which a uniform flow must be considered unstable. Whether roll waves form or not
is a function of the Vedernikov number (Ve), the Montuori number (Mo), and the concentration of sediment in the flow. When the Manning equation is used, the Ve is
Rectangle
Trapezoid
with
unequal side
slopes
Circle
Ve = |rFr
(3.8Ia)
Ve = IfFr
(3.8Ib)
Fr should be computed using Eq. (3.11) and F = a channel shape factor (Table 3.6) or
F = 1 -R^
dA
(3.82)
(3.83)
3.7
CONCLUSION
The foregoing sections provide the basic principles on which the following chapters on
design are based. Two observations are pertinent. First, open-channel hydraulics is incrementally progressing. That is, over the past several decades, there have been incremental
advances that primarily have added details, often important details, but no major new
advances. Second, open-channel hydraulics remains a one-dimensional analytic approach.
However, the assumption of a one-dimensional approach may not be valid in many situations: for example, nonprismatic channels, flow downstream of a partially breached dam,
or lateral flow over a spillway. In some of these cases, the one-dimensional approach may
provide an approximation that is suitable for design. In other cases, however, a two- or
three- dimensional approach should be used. Additional information regarding two- and
three- dimensional aDDroaches can be found in Chaudhrv (1993).
Montuor! Number = *&FIGURE 3.9 Flow stability as a function of the Vedernikov and Montuori numbers for clear water and slurryflow.(Based on data from Montuori, 1963;
Niepelt and Locher, 1989
REFERENCES
Ackers, P., "Resistance to Fluids flowing in Channels and Pipes," Hydraulic Research Paper No. 1,
Her Majety's Stationery Office, London, 1958.
Aisenbrey, A. J., Jr., R. B., Hayes, H. J., Warren, D. L., Winsett, and R. B. Young, Design of Small
Canal Structures, U.S. Department of Interior, Bureau of Reclamation, Washington, DC 1978.
Arcement, G. J., and V. R. Schneider, "Guide for Selecting Manning's Roughness Coefficients for
Natural Channels and Flood Plains," Water Supply Paper 2339, U.S. Geological Survey,
Washington, DC, 1989.
Barnes, H. H., "Roughness Characteristics of Natural Channels," U.S. Geological Survey Water
Supply Paper No. 1849, U.S. Geological Survey, Washington, DC. 1967.
Beasley, J. G., An Investigation of the Data Requirements of Ohio for the HEC-2 Water Surface
Profiles Model, Master's thesis, Ohio State University, Columbus, 1973.
Chaudhry, M. H., Open-Channel Flow, Prentice-Hall, New York 1993.
Chaudhry, M. H., Applied Hydraulic Transients, Van Nostrand Reinhold, New York, 1987.
Chow, V. T., Open-Channel Hydraulics, McGraw-Hill, New York, 1959.
Cox, R. G., "Effective Hydraulic Roughness for Channels Having Bed Roughness Different from
Bank Roughness," Miscellaneous Paper H-73-2, U.S. Army Engineers Waterways Experiment
Station, Vicksburg, MS, 1973.
Coyle, J. J. "Grassed Waterways and Outlets," Engineering Field Manual, U.S. Soil Conservation
Service, Washington, DC, April, 1975, pp. 7-1-7-43.
Einstein, H. A., "The Bed Load Function for Sediment Transport in Open Channel Flows.
Technical Bulletin No. 1026, U.S. Department of Agriculture, Washington, DC, 1950.
Einstein, H. A., and R. B. Banks, "Fluid Resistance of Composite Roughness," Transactions of the
American Geophysical Union, 31(4): 603-610, 1950.
Escoffier, F. F., "A Graphical Method for Investigating the Stability of Flow in Open Channels
or in Closed Conduits Flowing Full," Transactions of the American Geophysical Union, 31(4),
1950.
Escoffier, F. F, and M. B. Boyd, "Stability Aspects of Flow in Open Channels," Journal of the
Hydraulics Division, American Society of Civil Engineers, 88(HY6): 145-166, 1962.
Fischenich, J. C., "Hydraulic Impacts of Riparian Vegetation: Computation of Resistance," EIRP
Technical Report EL-96-XX, U.S. Army Engineers Waterways Experiment Station, Vicksburg,
MS, August 1996.
Flippin-Dudley, S. J., "Vegetation Measurements for Estimating Flow Resistance," Doctoral dissertation, Colorado State University, Fort Collins, 1997.
Flippin-Dudley, S. J., S. R. Abt, C. D. Bonham, C. C. Watson, and J. C. Fischenich, "A Point
Quadrant Method of Vegetation Measurement for Estimating Flow Resistance," Technical Report
No. EL-97-XX, U.S. Army Engineers Waterways Experiment Station, Vicksburg, MS, 1997.
French, R. H., Hydraulic Processes on Alluvial Fans. Elsevier, Amsterdam, 1987.
French, R. H., Open-Channel Hydraulics, McGraw-Hill, New York, 1985.
Guy, H. P., D. B. Simons, and E. V. Richardson, "Summary of Alluvial Channel Data from Flume
Experiments, 1956-61," Professional Paper No. 462-1, U.S. Geological Survey, Washington, DC,
1966.
Henderson, F. M., Open Channel Flow, Macmillan, New York, 1966.
Hoggan, D. H., Computer-Assisted Floodplain Hydrology & Hydraulics, McGraw-Hill, New York,
1989.
Horton, R. E., "Separate Roughness Coefficients for Channel Bottom and Sides," Engineering
News Record, 3(22): 652-653, 1933.
U.S Army Corps Engineers, "Accuracy of Computed Water Surface Profiles," U.S. Army Corps of
Engineers Hydrologic Engineering Center, Davis, CA, 1986.
U.S Army Corps Engineers, "Hydraulic Design of Flood Control Channels," EM 1110-2-1601.
U.S. Army Corps of Engineers, Washington, DC, 1970.
Yang, C. T., Sediment Transport: Theory and Practice, McGraw-Hill, New York, 1996.
Zegzhda, A.P., Theroiia Podobija Metodika Rascheta Gidrotekhnickeskikh Modele (Theory of
Similarity and Methods of Design of Models for Hydraulic Engineering), Gosstroiizdat,
Leningrad, 1938.
APPENDIX 3.A
VALUES OF THE
ROUGHNESS COEFFICIENT n
Minimum
Normal
Maximum
0.010
0.013
0.012
0.016
0.014
0.017
0.013
0.014
0.014
0.016
0.014
0.016
0.015
0.017
0.019
0.024
0.030
0.030
0.009
0.010
0.010
0.013
0.011
0.013
0.013
0.015
0.011
0.013
0.013
0.012
0.014
0.014
0.015
0.013
0.014
0.017
0.014
0.016
0.017
0.020
0.012
0.017
0.014
0.020
Type of Channel
f. Clay
1 . Common drainage tile
2. Vitrified sewer
3. Vitrified sewer with
manholes, inlet, etc.
4. Vitrified subdrain with
open joint
g. Brickwork
1. Glazed
2. Lined with cement mortar
h. Sanitary sewers coated with
sewage slimes with bends and
connections
i. Paved invert, sewer, smooth
bottom
j. Rubble masonry, cemented
k.Polyethylene pipe
1. Polyvinyl chloride
B. Lined or Built-up Channels
B-I Metal
a. Smooth steel surface
1. Unpainted
2. Painted
b. Corrugated
B-2 Nonmetal
a. Cement
1. Neat, surface
2. Mortar
b. Wood
1 . Planed, untreated
2. Planed, creosoted
3. Unplaned
4. Plank with battens
5. Lined with roofing paper
c. Concrete
1. Trowel finish
2. Float finish
3. Finished, with gravel
on bottom
4. Unfinished
5. Gunite, good section
6. Gunite, wavy section
7. On good excavated rock
8. On irregular excavated rock
Minimum
Normal
Maximum
0.011
0.011
0.013
0.014
0.017
0.017
0.013
0.015
0.017
0.014
0.016
0.018
0.011
0.012
0.013
0.015
0.015
0.017
0.012
0.013
0.016
0.016
0.018
0.009
0.010
0.019
0.025
0.020
0.030
0.011
0.012
0.021
0.012
0.013
0.025
0.014
0.017
0.030
0.010
0.011
0.011
0.013
0.013
0.015
0.010
0.011
0.011
0.012
0.010
0.012
0.012
0.013
0.015
0.014
0.014
0.014
0.015
0.018
0.017
0.011
0.013
0.013
0.015
0.015
0.016
0.015
0.014
0.016
0.018
0.017
0.022
0.017
0.017
0.019
0.022
0.020
0.027
0.020
0.020
0.023
0.025
Type of Channel
Minimum
Normal
Maximum
0.015
0.017
0.017
0.020
0.020
0.024
0.016
0.020
0.020
0.020
0.025
0.030
0.024
0.030
0.035
0.017
0.020
0.023
0.020
0.023
0.033
0.025
0.026
0.036
0.011
0.012
0.013
0.015
0.015
0.018
0.017
0.023
0.013
0.025
0.032
0.015
0.030
0.035
0.017
0.013
0.016
0.030
0.013
0.016
0.016
0.018
0.022
0.022
0.018
0.022
0.025
0.027
0.020
0.025
0.030
0.033
0.023
0.025
0.025
0.030
0.030
0.033
0.030
0.035
0.040
0.028
0.030
0.035
0.025
0.035
0.040
0.030
0.040
0.050
0.025
0.035
0.028
0.050
0.033
0.060
0.500
Type of Channel
Minimum
d. Rock cuts
1. Smooth and uniform
0.025
2. Jagged and irregular
0.035
e. Channels not maintained, weeds
and brush uncut
1. Dense weeds, high as
flow depth
0.050
2. Clean bottom, brush on sides 0.040
3. Same, highest stage of flow 0.045
4. Dense brush, high stage
0.080
C.2 Channels with maintained vegetation
and velocities of 2 and 6 ft/s
a. Depth of flow up to 0.7 ft
1. Bermuda grass, Kentucky bluegrass,
buffalo grass
Mowed to 2 in
0.07
Length 4 to 6 in
0.09
2. Good stand, any grass
Length approx. 12 in
0.18
Length approx. 24 in
0.30
3. Fair stand, any grass
Length approx. 12 in
0.014
Length approx. 24 in
0.25
b. Depth of flow up to 0.7-1.5 ft
1. Bermuda grass, Kentucky bluegrass,
buffalo grass
Mowed to 2 in
0.05
Length 4-6 in
0.06
2. Good stand, any grass
Length approx. 12 in
0.12
3. Length approx. 24 in
0.20
Fair stand, any grass
Length approx. 12 in
0.10
Length approx. 24 in
0.17
D Natural streams
D-I Minor streams (top width at
flood stage < 100 ft)
a. Streams on plain
1. Clean, straight, full stage
no rifts or deep pools
0.025
2. Same as above, but with more
stones and weeds
0.030
3. Clean, winding, some pools and
shoals
0.033
4. Same as above, but with some
weeds and stones
0.035
Normal
Maximum
0.035
0.040
0.040
0.050
0.080
0.050
0.070
0.100
0.120
0.080
0.110
0.14
0.045
0.05
0.09
0.15
0.08
0.13
0.035
0.04
0.07
0.10
0.16
0.09
0.030
0.033
0.035
0.040
0.040
0.045
0.045
0.050
Type of Channel
Minimum
Normal
Maximum
0.048
0.050
0.055
0.060
0.070
0.080
0.100
0.150
0.040
0.050
0.050
0.070
0.030
0.035
0.035
0.050
0.030
0.035
0.040
0.040
0.045
0.050
0.050
0.070
0.050
0.060
0.070
0.110
0.070
0.110
0.100
0.160
0.150
0.200
0.040
0.050
0.060
0.080
0.100
0.120
0.120
0.160
Type of Channel
Minimum
Normal
Maximum
0.060
0.100
0.020
0.030
0.040
0.025
0.015
0.020
0.17
0.17
0.20
0.20
0.10
0.30
0.05
0.80
0.48
0.40
0.30
0.20
0.40
0.13
0.09
0.05
0.34
0.25
0.008
0.06
0.06
0.30
0.04
0.07
0.17
0.012
0.22
0.16
0.50
0.10
0.17
0.47
0.10
0.20
0.10
0.08
0.04
0.02
0.15
0.12
0.10
0.05
Source: From Chow (1959), Richardson et al. (1987), Simons, Li, & Associates (SLA),
1982, and others. * The values in bold are recommended for design
CHAPTER 4
SUBSURFACE FLOW
AND TRANSPORT
4.7
INTRODUCTION
This chapter begins with the mathematical description of the constitutive relationships for
flow and transport in porous media. Following this, simple analytical solutions are presented for a variety of subsurface flow and transport problems. The principles of flow and
transport are outlined, and solutions are provided for practical problems of flow and transport in both the saturated and the unsaturated zones. The latter includes problems of
transport in the vapor phase. The major focus is on the processes that are relevant to
subsurface mitigation.
4.2
CONSTITUTIVERELATIONSHIPS
This section presents the basic concepts and laws used to describe flow and transport in the
subsurface. In particular, the constitutive relationships defining the fluid flow in fully and
partially saturated media are given as well as the relationships that describe diffusive and dispersive mass fluxes in porous media. Finally, we show the relations used to describe partitioning of chemicals in the subsurface environment.
4.2.1 Darcy's Law
Consider the flow of a fluid through a pipe filled with a granular material, as shown in
Fig. 4.1. In the figure, Z1 and Z2 represent the elevations of the pipe centerline above a reference level at Sections 1 and 2, respectively, whereas p/y and p2/j represent the water
pressure head at Sections 1 and 2, respectively. We define the piezometric head at any
location in the porous media as
h = z+p/J
(4.1)
where y = specific weight (weight per unit volume) of water, typically, y = 9810 N/m3 or
62.4 lb/ft3. Let q be the average water velocity in the cross section of the pipe: i.e.,
q = Q/A
(4.2)
Arbitrary datum
FIGURE 4.1 Porous media flow.
where Q = volumetric discharge (volume per unit time) and A = total cross-sectional
area of the pipe (including the soil matrix). French hydrologist Henry Darcy discovered
that the average flow velocity could be estimated from
q = KQi1 -H2)IL
(4.3)
where L is the distance, measured along the pipe, between cross sections 1 and 2, and K
is a parameter that depends on the nature of the porous media as well as on the properties
of the transported fluid. For water, K is known as the hydraulic conductivity or the coefficient of permeability. Typical values of K are given in many references (e.g., Bureau of
Reclamation, 1985) see tables 4.1 ard 4.2. Eq. (4.3) is known as Darcy1s law and is commonly used to model the flow of fluids in porous media. Notice that the velocity V is not
the fluid velocity in the soil pores, it is an average velocity calculated over the entire area
of the flow cross section. The average pore velocity is calculated as
v = -j(4.4)
vw
where Qw is the volumetric water moisture content. Note that for saturated flow, Qw = n,
where n is porosity.
Darcy's law (i.e., Eq. 4.3) also can be written more concisely as
q = KI
(4.5)
4.6)
(4.7)
For saturated flow of a constant density fluid in isotopic porous media, the Darcy law
can be written as
*=-*!)
For anisotropic media, the Darcy law is written as
()
(4 9)
* = -*(f|)
where A^ is the conductivity tensor. For saturated flow of a fluid of variable density in
anisotropic porous media, we have
*-fe(t+^)
where ksoilij is the intrinsic permeability tenser. Finally, for unsaturated flow of variabledensity fluid in anisotropic porous media, we have
+
aX
-HrHr
Pfluid \.OXj *-!)
j)
where kr(Q) = relative permeability of the porous media. Relative permeability is a function of soil saturation, which in turn is a function of the capillary pressure. These relationships for partially saturated flow are discussed in the next section.
(4.12)
(m>
"-^-Rj
for > hb and is otherwise (capillary fringe zone),
e. = i
(413)
(4.14)
where n = porosity, G = volumetric moisture content (equal to n SJ, Qe = effective volumetric moisture content, 9r = residual (irreducible) moisture content, X = Brooks-Corey
parameter, and hb = capillary fringe height. The van Genuchten model is
ee = e - -^ - er = [i + (<x)Tw
(4.15)
where: = curve fitting parameter that depends on the type of soil, a = soil property
index, a l/hb, and m = 1 1/rc.
Note that parameter n depends on pore-size distribution. For a well-graded soil (wide
pore-size distribution, which results in a flatter moisture content curve 6(1F)), n is small,
whereas for poorly graded soils (narrow pore-size distribution, which results in a steeper
moisture content curve G(1F), n is large: typically values of n higher than 2.5.
Also note that although the BC and VG models are the most commonly used models
in analysis, there is no restriction on using different mathematical representations to
describe the characteristics of soil-water retention. For example, a simple exponential
model, such as Qe = exp( P *F), is in some cases, sufficient to describe the physics of the
retention.
As the moisture content in partially saturated media decreases, so does the volume of
pores available to fluid flow. Thus, hydraulic conductivity for the partially saturated
media depends on the water content and, in turn, on the metric suction. To describe this
relationship, we modify the value of intrinsic permeability k by the factor of kr(Q) or krC)t
called relative permeability. Several models for kr are shown below:
BC model of relative permeability:
2-3X
fcr = e e
(4.16)
kr = 6e5(l - (1 - 9/T)2
(4.17)
(4.18)
(4 19
- >
In this equation, q = mass flux of solute per unit area per unit time [ML" T"1],
D diffusion coefficient (L2 T -1), C = solute concentration = mass of solute/volume of
solution (M/L3), dC/dx = concentration gradient along the x direction. The minus sign in
Eq. (4.19) indicates that the solute flux will go from regions of larger concentration to
those of lower concentration. Values of the diffusion coefficient, Z), depend on the type of
solute and the type of environmental fluid. For major cations and amnions dissolved in
water, values of D range from 1 X 10~9 to 2 X 10 m2/s (Fetter, 1994).
(4.21)
DT = aT v + D*
(4.22)
or
where OCL, ocr = longitudinal and transverse dispersivities, respectively, and D* = effective porous-media diffusion coefficient.
In addition to the pore-scale dispersion, we also have formation-scale dispersion or,
more accurately, spreading, which is a result of the variability in transport velocity caused
by the heterogeneity of the hydraulic conductivity field. In terms of magnitude, this
microdispersive flux is significantly larger than the one related to mechanical dispersion.
Mathematically, macrodispersive flux is the flux equal to the expected value of the product of Darcian velocity (qf) and of the contaminant concentration (C') fluctuations:
* macrodispersive = < *' C' >
(4.23)
In otherwords, this flux can be described using an expression similar to the Fickian diffusion equation (Eq. 4.19) with a macrodispersivity coefficient. In the most general (threedimensional) case, the equation defining the macrodispersive flux in the flowing fluid is
rlC1
^ = (Ap)SgOXj
(4.24)
where Afj represents the macrodispersivity tensor and the bar indicates averaged quantities. The fluctuations q' and c' result from the heterogeneous nature of the aquifer. This is
expressed in the way the macrodispersivity tensor is estimated. For example, the longitudinal macrodispersivity is estimated using
A11 = ^
(4.25)
where <jf2 = variance of log-conductivity (f= Ln[K]), K1 = correlation scale in the direction of flow, and y is given by
= cx
v = -rT
P\\
KgJ1
Lo J
(4 26)
In summary, the longitudinal and transverse components of the total diffusive and dispersive mass flux (per unit bulk area) in heterogeneous geologic formations is estimated
as follows:
VL= -^FLuiDALvFLUID + D)^
(4.27)
(4-28)
T^C
<7r,vert = -(^FumAj^Fum + W f^
(4-29)
Partitioning
Equilibrium partitioning and sorption are the most common chemical processes that
affect reactive transport. These processes are dealt with by equating the total concentration to the sum of the concentrations in each phase multiplied by their respective volumes.
Furthermore, by equating the concentration in each phase to the concentration in a common phase say, the concentration in water the total concentration can be expressed in
terms of the common phases concentration and a retardation coefficient R
CT = ^WATER^XTER(^ ~*~
T^- /\
2*i "gWATER '= ^WATER^-WATERR
/ * WATER
(4.30)
where: K1 = partitioning coefficient between the 1th phase and the common phase K1= C1/
CWATER, ^i = tne volumetric content of the ith phase, and QWATER = the volumetric content
of the common phase. This type of equilibrium partitioning is frequently used to describe
the relationships between concentrations in the following scenarios: (1) vapor and aqueous phases, (2) soil and vapor phases, (3) soil and aqueous phases, (4) partitioning of a
tracer between aqueous and NAPL or DNAPL phases, and (5) partitioning of a tracer
between vapor and NAPL or DNAPL phases.
(4.31)
where S1 is the solubility of compound 7 in water and X1 is its mole fraction in the mixture.
Finally, for partitioning of compounds present in a NAPL or DNAPL mixture and vapor
phase, use
"w - '^Tf=raSSWir
<4'32'
where V1 is the vapor concentration of compound /, Pvl is the vapor pressure of compound
/, MWI is its molecular weight, R = gas constant, and T = temperature in 0K.
4.2.5
Degradation
In addition to partitioning, degradation of compounds also may affect the fate and
transport of reactive compounds significantly. Typically, degradation is modeled using
either power-order decay models or growth-process-based models. In environmental
subsurface hydrology, three basic power-order models are used: (1) zero-order decay,
(2) first-order decay, and (3) a combination of the first two models. According to these
models, the total decrease of mass in unit bulk volume caused by degradation is
expressed by
(^C1
-^ = - E e/*<y - S eA/c/
<4-33)
/ = PHASES
I = PHASES
where K01 = zero-order degradation rate of the compound in phase 7, K11 = first-order
degradation rate of the compound in phase 7, and C1 = mass/volume concentration of the
compound in phase 7.
In addition to the zero-order and first-order degradation processes, the Monod kinetics
is frequently used to describe oxygen limited aerobic degradation of organic compounds
in the aqueous phase. According to the Monod model, the degradation rate in terms of the
total concentration is expressed by the following system of equations:
C
dCr = _ Q
WATER } ( WATER ]
M1
3f
WATER^N Kr +. Cr
K J- n
( c + WATER)(KO + OWATER)
^
^ ' '
(4
and
CWATER
D
Jvl if
YM
}hr
( 4WATER
}
^t = - 64TK>C
tH r , rC
n
dt
(Kc + WATER )(Ko + OWATER)
(4 35}
^HOj;
where CWATER is the aqueous concentration of the contaminant, OWATER is the dissolved oxygen concentration, M1 is the total concentration of the active microbial biomass, i is the
maximum rate of organic solute utilization, Kc is the concentration of the organic solute
at which the utilization rate is half the maximum, K0 is the electron acceptor (oxygen)
concentration at which the utilization rate is half the maximum, and % is the substrate utilization ratio.
4.3
In this section, we present some solutions to saturated flow and transport problems that
are encountered in the practice of subsurface hydrology. We begin with well hydraulics,
an understanding of which is important to the design of pump-and-treat systems, and discuss several models of transport of soluble plumes.
4.3.1 Flow to a Single Well
Darcy's law describes the average water flux through a porous medium when the local
hydraulic gradient is known. To determine discharges we use the law of conservation of
mass for the water, also known as the continuity equation. For the analysis of wells, it is
assumed that the flow toward the well caused by the well pumping is radially symmetric.
In this situation, it is convenient to use the equation of continuity in radial coordinates
(r, 6). For steady-state state flow toward a single well in a confined aquifer without
recharge, the equation reads
Q = 271/ ^)
(4.36)
where K and B = aquifer conductivity and thickness, respectively. The equation of continuity for well flow in an unconfined aquifer is similar to Eq. (4.19): namely,
Q = 2nrKh(^\
(4.37)
with B replaced by the variable flow depth h. Implicit in Eq. (4.36) and (4.37) is the
Dupuit-Forchheimer assumption, the implication of which is that the flow in the aquifer
can be assumed to be practically horizontal.
Equations (4.36) and (4.37) are used to obtain solutions for the steady-state discharge
to a well in confined and unconfined aquifers, respectively, if the piezometric heads (confined aquifers) or water-table elevations (unconfined aquifers) H1 and h2 are known at two
radial distances T1 and r2, respectively:
Q = 2nKB(h2 - H1)IIn(T2Ir1) = 2nT(h2 - H1)IIn(T2Ir1)
(4.38)
(4.39)
The solutions given in Eqs. (4.38) and (4.39) assume that the well penetrates to the
impermeable bottom of the aquifer and that there is no recharge into the aquifer. They also
assume an infinitely large aquifer with no interaction with surface streams or impermeable boundaries. Well solutions are often given in terms of the drawdown s as a function
of the radial distance r. The drawdown is defined as
s = H- h
(4.40)
where H is the elevation of the original piezometric surface before pumping at the well
starts. The distance R for which h = H and s = O is called the radius of influence of the
well. Using the concepts of drawdown and radius of influence, the steady-state flow equation in a confined aquifer can be rewritten as
,-&L
(4.4,,
By combining the solutions for the steady state flow in confined and unconfined
aquifers, one can derive a relationship between the drawdown calculated from confined
conditions (assuming constant in space and time aquifer thickness equal to //), and that
estimated for unconfined conditions (when the change in aquifer thickness caused by
pumping is taken into account):
SUNC = H- V^ - 2sCONFH
(4.42)
Using this formula, known as Jacob's correction, one can initially assume constant
aquifer thickness B = Hm calculations and use "confined" aquifer solutions to calculate
drawdown, then correct the drawdown using Jacob's correction. This approach is particularly useful when dealing with transient flow. For transient flow conditions in an aquifer
with constant flow thickness, the transient drawdown is given by
<">-&l:s!r**-&'"
where
M(r
' = Wt
(4 44)
'
and S = storativity (for confined aquifers) or porosity (for unconfined aquifers) and W(u)
is known in subsurface hydrology as well function and in mathematics as exponential
integral. This function is tabulated in almost every groundwater hydrology textbook. It
also is available in many engineering mathematics software packages, such as
Mathematica, as a library function.
4.3.2 Superposition and Convolution
For a time-variable pumping rate, the principle of convolution can be used to estimate the
transient drawdown. This approach is strictly valid for linear systems: i.e., systems in
which the response (drawdown) is a linear function of the excitement (pumping rate). The
linearity assumption is strictly valid for confined aquifers only; however, as long as the
drawdowns do not exceed 20% of the initial aquifer thickness, it also may be used for
unconfined aquifers. Using the convolution approach, the transient drawdown for a pumping rate changing in a step-wise fashion is given by
n
S
=
^ -^T
^ (Q* + Q* - 1)^^ - 1
<4'45)
*nik = 1
where the drawdown is estimated at time f, tn<t<tn + l, QK is the pumping rate for
t
K - i < *< * K,to = 0> Qo = 0-0, and &tK _ l = t - tK _ r When several wells are present,
the superposition approach is used to estimate the cumulative drawdown by adding the
drawdown contributions from all the wells:
m
*W = 4Sr 2 QLW(u(rL,t))
(4.46)
where rL is the distance between the point of interest (where the drawdown is estimated) and
well L. When several wells are pumping at variable rates, the superposition and convolution
approaches are used simultaneously. The superposition principle also can be used to superimpose the drawdown on the natural (ambient) flow conditions. Using this principle leads to
(4.47)
where h = transient potentiometric surface that combines ambient conditions and well
impact, H = potentiometric surface under natural (ambient) conditions, and s = transient
drawdown.
4.3.3 Interception Wells
With respect to contaminant transport in the subsurface, interception wells are used to trap
the contaminant plume within the well flow field. It is assumed in this analysis that there
is an ambient steady-state uniform flow through the aquifer. The combination of wellrelated flow and ambient uniform flow satisfies the conditions of two-dimensional potential flow in a horizontal plane, where the discharge described by stream function y is related to potential <|), or the piezometric head h. The extent of the aquifer through which water
travels to the well and is captured by it is called the capture zone. The derivation of the
analytical solution for steady-state flow capture-zone uses the following assumptions:
(1) a homogeneous, isotropic, infinitely large aquifer, (2) uniform flow, (3) no leakage, (4)
aquifer storativity or specific yield neglected, (not relevant for steady-state analysis),
(5) hydrodynamic dispersion neglected, (6) the Dupuit assumption applies, and (7) the
well is fully penetrating and pumping at a constant rate. Three important parameters are
used in delineating the capture zone: namely, the stagnation point, the upgradient maximum width of the capture zone, and the equation for the capture zone boundary.
For a confined aquifer, the distance from the well to the stagnation point (measured in
the direction of the uniform flow) is
^0 = Jf7
(4-48)
(4-49)
(4.50)
/
27tf7y\
tan [n-
\
&I
The procedure for delineating the capture zone consists of the following steps: (1) estimate the location of the stagnation point (XSTAG, O), (2) estimate the maximum width of the
capture zone wD/v, and (3) vary y between zero and wD/v/2 and use the capture zone boundary to estimate the boundary location (;c,v).
4.3.4 Partially Penetrating Wells
Performance of wells that penetrate only partially through the bearing strata is discussed
in this section. The simplest case consists of a well that is barely penetrating into an semi-
infinite porous medium so that the aquifer flow is three-dimensional and spherically symmetric. In this case, the following relationship applies between the flow into the partially
penetrating well Qp and the flow to a fully penetrating one Q:
t-K
where B is the aquifer thickness, rw is the well radius, and R is the radius of influence of
the partially penetrating well. Because, in general, rw B, then the equation above indicates that the spherical flow to a partially penetrating well is highly inefficient compared
with simple radial flow: i.e., for the same drawdown in the well, it results in a significantly
smaller pumping rate.
In the general case of partial penetration, one may consider the total drawdown sr,
which consists of the drawdown equivalent to that of a fully penetrating well s and additional head loss because of the partial penetration of the well As:
ST = s + As
(4.51)
Additional head loss for a well penetrating from the top (or the bottom) of the aquifer
is estimated as follows:
Aj =
CO^)J 0^1
2nTp
I
rw \
(4.53)
(4,4)
Thus, when the pumping rate is defined for a well, we calculate the drawdown correction As and add it to the full penetration drawdown s. However, when the drawdown is
given for a well, we have to recalculate the pumping rate. In this case, the true pumping
rate is given by
c =2
'
TTAJ = 2H^
(455)
where s = drawdown defined at the well, Q = pumping rate estimated using s, and
Qp = actual pumping rate
4.3.5 WellDuplets
Well duplets, each of which consists of one pumping and one recharge well, are frequently used as a means of injecting and removing aquifer mitigation solutes, such as co-solvents, surfactants, or both. Typically, the recharge well is positioned directly upgradient
from the discharge well, and the magnitudes of pumping and injection rates are the same.
In this case, the two wells form a flow circulation cell: i.e., all the injected water is
pumped out by the discharge well. In the case of co-solvent flushing, it is important to
understand what region of the aquifer is subject to the mitigation: i.e., what the boundary
is of the circulation cell. This boundary is defined by the upgradient and downgradient
stagnation points (X STAG , O) and (XSTAG, O) and the cell boundary equation. For the x-axis
parallel to the direction of ambient flow and the origin of the coordinate system located at
the mid-point between the two wells, the two stagnation points, (~XSTAG, O) and (XSTAG, O)
are given as the roots of the quadratic equation
Q0Bd
~27
i ( x/d
_ x/d 2 ] _ Q
^((x/d + l ) 2 (x/d-1)) u
f456 ,
1
<4-57'
The circulation cell is symmetric with respect to the y axis. The cell delineation procedure consists of estimating the locations of the stagnation points and varying x between
zero and XSTAG and using the cell-boundary equation to solve for y. This implicit equation
can be solved by any calculation software, such as Mathematical or MS Excel.
4.3.6 Transport Equations
The following general form of mass transport equation in the saturated zone is derived assuming one-dimensional advective and three-dimensional diffusive-dispersive transport in the
aqueous phase, linear partitioning of a compound between the three phases (water-soil(D)NAPL), and first-order degradation in the aqueous phase. For this conditions, we have
<^H((<""+)t)-t-^
where Cw = aqueous phase concentration, v = pore-water velocity, and 9W = volumetric
moisture content, and oc. = longitudinal, transverse horizontal, and transverse vertical
/^ CQ\
where kSOIL = water-soil partitioning coefficient, kSOIL = S/CW, S = weight/weight concentration of absorbed compound in soil, kNAPL = NAPL - water partitioning coefficient,
^NAPL = CNAPLW> CNAPL = concentration of compound in the NAPL phase, JB = bulk density of dry soil, and QNAPL = volumetric WAPL content.
Assuming that the diffusive fluxes are negligible compared the macrodispersive fluxes, the transport equation can be simplified to yield
3CW
3 (
3CW "1 V dCw
C
-^3r = ^-WVR^3Jt1 ^ l R 3Jt1 Jr *R-^djtj ~ ^RR w
(4-6)
^-Sfc&MsM^])
<"
(4.62)
with a concentration C0 into a uniform flow field from a point (x = O, y = O). Let the
uniform velocity be v^. The asymptotic solution, i.e., for t <*>, is given by
CC*y)
(51 -P
P P^kJ ^ [- + ^]I
^ = UrtVS^J
(-2D1Pr) ]_^l4DL(DL D 7 JJ
(4-63)
where K0 = modified Bessel function of the second kind and of Oth order (Bear, 1972).
The time-dependent solution is
C(X
L Wn n V*/feWmP)] - W,P)
U^VDLDrJ 1,2DJ
(4.64)
where
/ v 2 / r2
V2 \
>- VAt& + )
'4^
W (f, ) = leaky well function (see, for example, Hunt, 1983, p. 100).
4.3.7.4 Point slug injection into a uniform flow field3-D transport and retardation.
In this case, a slug of contaminant of the mass M = C0V is injected at point (0,0,0). The
transient distribution of concentration is described by
C(x y Z t}=
eJ_(*-V)
L
4axvRt
VC
%(nvRt)m(axa^m
2
_^i___jM
4ayv/?r 4a2v/?rJ
(4.66)
4.3.7.5 Continuous injection from a finite-sized source with retardation and degradation. In this case, consider transport from a rectangular source that is perpendicular
to the direction of flow. The source width is F, and its depth below the water table is Z.
The transient concentration distribution in the presence of retardation and degradation is
given by
. C0 \{ x V 1 fi + 4V*H1
cfc^o^m-ji-HH}]
172
j-, / fjc-v^l
A ^ +4^)
Kv' I
l
2(CW)\
ErfC
\y + Y/2]
F_J
y - YI2\
M^^rM^HI
{E*[^}-E*[^}
(4 67)
'
(4.68)
(4.69)
or
Combining Darcy's law with the mass continuity equation, we can write the final flow
equations is
^ = V(KmVh)
at
(4.70)
The flow equations can be simplified for horizontal and vertical flow conditions.
a. One-dimensional horizontal flow:
w = a {K?rg\
Bt dx(
dx)
(4.7i)
where 0 = volumetric water content. In this, the contribution of the elevation head, z,
vanishes, since 3z/3jc = O.
b. One-dimensional vertical flow:
W = JL(Ww + 1I
ot oz {
(oz
Jj
(472)
Note that the flow equations are characterized by the presence of two dependent, albeit
related, variables: namely, 0 and 1F. To simplify this situation, we describe the relationship between 0 and 1F by a term called soil diffusivity D(0) as
=m
where C(0) is called specific moisture capacity and is defined as
C(0) = U
>
(4.74)
(4.75)
(4-76)
As can be seen, we now have only one dependent variable: namely, 6. The only limitation of
this formulation is that specific moisture capacity C(Q) becomes zero in the capillary fringe
zone, thus making the solution impossible. Therefore, this formulation is valid only in the partially saturated zone (water content less than saturated value), not in the capillary fringe.
Another way to solve this problem, which does not have the limitation discussed
above, is to formulate the flow equations in terms of soil suction \(/. For the vertical flow,
we obtain
C(e)
? = i (^(XF)(lr+ l)j
(4 74)
Exercise. Consider steady-state vertical infiltration from the soil surface to the water table
at depth L. The relative hydraulic conductivity of the soil is described by the following
exponential law:
kr(V) = exp[-aV]
(4.78)
Derive an expression for the vertical distribution of h. After the first integration of the
flow equation, we obtain
~\\TJ
\
+1
a- J =
(4-79)
(4.80)
where h = z x|/. Substitution of these relationships into the flow equation leads to
or
(4.81)
exp[ah]dh = -/-exp[ax]dx
^-SAr
After the second integration, we obtain
(4.82)
(4.84)
wW
4=
T?-^fk
+ ^N^l
^ - XC
d/ =3x,
' +6
Qw ) to,J- v &,
W
(4.85)
/^ g^x
where kSOIL = water-soil partitioning coefficient, kSOIL = S/CW, S = weight/weight concentration of absorbed compound in soil, kNAPL = NAPL-water partitioning coefficient,
^VAPL = CNAPI! Cw> CNAPL = concentration of compound in the WAPL phase, kAIR = vaporwater partitioning coefficient, kAIR = V/CW, V = concentration of compound in the vapor
phase, J8 = bulk density of dry soil, QAIR = volumetric vapor content, and QNAPL volumetric NAPL content.
Assuming again that the diffusive fluxes are negligible compared with the macrodispersive and advective fluxes, the transport equation can be simplified to yield
T'l-hfh-f-v.
<">
where VR vIR and KR = K/R. Note that the form of this equation is the same as the form
of the one for transport in the saturated zone; therefore, the analytical solutions presented
for the saturated transport are valid for the unsaturated conditions. This is particularly true
for the one-dimensional (vertical) transport equations, which are of primary interest in the
case of unsaturated fate and transport of compounds.
4.5
This section presents the soil vapor flow equations. This is followed by selected solutions
to vapor flow problems. Finally, we discuss diffusive transport in the vapor phase.
4.5.1 Soil Vapor Flow
The flow of gases in a porous medium can be described by combining the following
equations:
a. The equation of continuity is
V(pv) = -^
(4.88)
with p = vapor density, v = Darcy's flux vector, 9AIR = air-filled porosity, and t = time.
b. The perfect gas law is
P- = -
(4-89)
-Av/?
(4.90)
where k = soil intrinsic permeability and JJL = gas viscosity. The resulting governing
equation is 1953
1V.-2S*
(Bruce et al., 1953).
d. The molar flux [moles/unit area-time] is given by
= ^PVp
(4.92)
'
W = ^l
The pressure p can be expressed in terms of the ambient pressure pAtm and a deviation
/?' from this pressure: p' is equivalent to the vacuum that would be measured in the soil.
If this substitution is used in the flow equation and if we neglect the product p'2 relative
to the product pAtm /?' (linearization), then the resulting equation for radial flow is
k
F^
PATM %- -3-}
r dr( Br )
(4
-99)
"^)-SH5
'"*
as given by
P
' = 4K&W
eXp[
-X]dx=4^)WM
(4 101)
'
where Q is the volumetric flow rate to the vapor well. The well function W(u) is tabulated in almost all groundwater textbooks. The behavior of the integral is such that for
(^A/aM^fcp^t) < 0.001, its value is close to the asymptotic steady-state limit.
Exercise. Given the following parametershydraulic conductivity K= 10~2 cm/sec, air
viscosity |HA//? = 0.018 cp, volumetric air content (equal to porosity) 6 AIR = 0.3, and local
pressure gradient dp/dx = 0.01 atm/cmestimate the pore-vapor velocity. From the conversion table (Domenico and Schwartz, 1990), we have
k[darcy] = K\mls\* 1.04*105
(4.102)
<y
&J
(4.104)
(4105)
where: V = vapor phase concentration, VAIR = pore-air velocity, 0W = volumetric moisture content, QAIR = volumetric vapor content, X = first-order degradation rate in the
aqueous phase, kA _ w = air-water partitioning coefficient, kA_w = C^V, and R = retardation factor,
/? = 1 +
,A ]f\\
where: kSOIL = water-soil partitioning coefficient, kSOIL = SlCw, S = weight/weight concentration of absorbed compound in soil, kNAPL = WAPL-water partitioning coefficient,
^NAPL = CNAPL/CW CNAPL = concentration of compound in the TVAPL phase, kAIR = vaporwater partitioning coefficient, kAIR = Vl Cw, V = concentration of compound in the vapor
phase, Y6 = bulk density of dry soil, QAIR = volumetric vapor content, and QNAPL = volumetric NAPL content.
Division of the transport equation by the retardation factor yields
?--"-**-v
(4 107)
where vAirR = vAIRIR and AT? = kA _ ^^w/C^A//? ^)- Note that the form of this equation is the
same as the form of the one for transport in the saturated zone, except for the absence of
the dispersive term. Therefore, the analytical solutions presented for the saturated transport are valid for the vapor transport conditions.
When there is no advective transport in the vapor phase, the transport equation must
include diffusive fluxes in vapor and aqueous phases to yield
K^V= 3f D vc|V
& dx,( Dx,
avl.x^.*
dx/)
Qm
wv
(4.108)
(4.109)
D = " + ?*-*
K
(4.110)
where
and
*-&
Again, we note the similarity of this fate and transport equation to the one presented
for saturated transport and conclude that all the analytical solutions presented in Sec. 4.2
can, in principle, be used to analyze vapor phase transport.
Exercise. Given that water saturation Sw = 0.20, porosity n = 0.4, compound concentration in soil vapor at depth L = 2 m C0 = 100 mg/L, compound concentration at the soil
surface Cs = 0.01, and molecular diffusion coefficient of the compound D0 =
0.087cm2/sec, estimate the compound mass flux in the vapor phase at the soil surface. The
effective vapor-phase diffusion coefficient is given by
03.33
D = D0 -4= 0.0026 cmVs
nl
and the mass flux of compound A is estimated as follows:
(4.20)
(4.113)
--
where
CT = CSwn + V(I - Sw)n = Vn (1 - SJ S"
+ l\
[K(L bw
)
(4.115)
= *T^J
<4'117>
<>
j
+1
^w
K(I - Sw)
(4.119)
Exercise. Consider diffusive vertical transport of a compound in vapor phase. The compound is subject to first-order degradation in the aqueous phase at rate X and to partitions
between vapor and aqueous phases according to the following relationship:
V=KC
(4.113)
DJfV-I
SC = Q
dx2 n
(4.120)
where
Q3.33
D - D0-^n2
Substituting C = VIK into the mass transport equation leads to
32V
^JL-X2V=O
dr
where
(4.19)
(4.121)
X" = ^
(4.123)
(4.124)
@x = 10Q
(4.125)
REFERENCES
Bear, J., and Y. Bachmat, Introduction to Modeling of Transport Phenomena in Porous Media,
Kluwer Academic, Dordrecht, The Netherlands, 1990.
Bear, J., Dynamics of Fluids in Porous Media, Dover Publications, New York,k 1972.
Bear, J., Hydraulics of Groundwater, McGraw Hill, New York, 1979.
Bouwer H., Groundwater Hydrology, McGraw Hill, New York, 1978.
Bruce, G. H., D. W. Peaceman, and H. H. Rachford, Jr., 1953. "Calculations of Unsteady-State Gas
Flow Through Porous Media," Petroleum Transactions, AIME, 198: 1953.
Bureau of Reclamation, Ground Water Manual (Reprint), U.S. Department of the Interior
Washington, DC, 1995.
Cedergren H. R., Seepage, Drainage, and Flow Nets, 3rd ed., John Wiley & Sons, Inc., New York,
1989.
Charbeneau, R. J., "Kinematic Models for Soil Moisture and Solute Transport," Water Resources
Research, 20: 699-706, June, 1984.
Chirlin, G. R., "A Critique of the Hvorslev Method for Slug Test Analysis: The Fully Penetrating
Well," Ground Water Monitoring Review, 130-138, 1989.
Cho, J. S., 1991. Forced Air Ventilation for Remediation of Unsaturated Soils Contaminated by
VOC., Publication No. EPA/600/S2-91/016, U.S. Environmental Protection Agency, Washington,
D.C.
Dagam, G., "Solute transport in heterogeneous porous formations," Journal of Fluid Mechanics,
145: 151-177, 1984.
De Josselin Jong, G. "Singularity Distribution for the Analysis of Multiple-Fluid Flow Through
Porous Media," Journal of Geophysical Research, 65: 3739-3758, 1960.
De Marsily G., Quantitative HydrogeologyGroundwater Hydrology for Engineers, Academic
Press, San Diego, CA,
De Smedt R, and P. J. Wirenga, "Solute Transport Through Soil With Nonuniform Water Content,"
Soil Science Society of America Journal, 42. (1): 1978.
Domenico, P. A., and F. W. Schwartz, Physical and Chemical Hydrogeology, John Wiley & Sons,
New York, 1990.
Dullien F. A. L., Porous Media: Fluid Transport and Pore Structure, 2nd ed., Academic Press, San
Diego, CA, 1992.
Edelman J. H., Groundwater Hydraulics of Extensive Aquifers, 2nd., International Institute for
Land Reclamation and Improvement, Bulletin No. 13, The Netherland, 1983.
Fetter, C. W., Contaminant Hydrogeology, Macmillan, New York, 1993.
Fetter, C. W., Applied Hydrogeology, Simon & Schuster, Company Englewood, NJ, 1994.
Freeze, R. A., and J. A. Cherry, Groundwater, Prentice-Hall, Englewood Cliffs, NJ, 1979.
Gelhar, L. W. and C. L. Axness. "Three-Dimensional Stochastic Analysis of Macrodispersion in
Aquifers", Water Resources Research, 19 (1): 161-180, 1983.
Germann, P. R, M. S. Smith, and G. W. Thomas, "Kinematic Wave Approximation to the Transport
of Escherichia coli in the Vadose Zone," Water Resources Research. 23 (7), 1281-1287, 1987.
Girinsky, N. K. Determination of the Coefficient of Permeability, Gosgeolizdat, 1950.
Grubb, S. "Analytical Model for Estimation of Steady-state Capture Zones of Pumping Wells in
Confined and Unconfmed Aquifers," Ground Water, 31(1) 27-32, 1993.
Hantush, M. S., Hydraulics of Wells, "in Advances in Hydroscience," V. T. Chow, ed., Academic
Press, New York, 1964.
Hantush, M. S. "Growth and Decay of Groundwater-mounds in Response to Uniform Percolation."
Water Resources Research, 3 (1): 227-234, 1967.
Harr, M. E., Groundwater and Seepage, Dover Publications, New York, 1990.
Haverkmp, R., M. Vauclin, J. Touma, P. J. Wierenga, and G. Vachaud, A "Comparison of
Numerical Simulation Models for One-Dimensional Infiltration," Soil Science Society America
Journal, 41: 285-294, 1977.
Hinchee, R. E. ed., Air Sparging for Site Remediation, Lewis Publishers, Boca Raton, FL, 1994.
Huisman, L., Groundwater Recovery, Winchester Press, New York, 1972.
Hunt, B. "Seepage to Collection Gallery Near Seacoast," Water Resources Research, 21: 311-316,
1985.
Hunt, B., Mathematical Analysis of Groundwater Resources, Butterworths, London, UK, 1983.
Jaffe, P. R., and R. A. Ferrara, "Desorption Kinetics in Modeling of Toxic Chemicals," Journal of
Environmental Engineering, American Society of Civil Engineers, 109: 859-867, 1983.
Javandel, I., and C. F. Tsang, "Capture-zone Type Curves: A Tool for Aquifer Cleanup." Ground
Water, 24: 616-625, 1985.
Parker, J. C., and M. T. van Genuchten, Determining Transport Parameters from Laboratory ad
Field Tracer Experiments, Virginia Agricultural Experiment Station Bulletin No. 84-3, Virginia
Polytechnic Institute and State University, Blacksburg, 1984.
Parker, J. C., and M. T. van Genuchten, "Flux-averaged and Volume-Averaged Concentrations in
Coninuum Approaches to Solute Transport," Water Resources Research, 20: 886-872, 1984.
Parker, J. C., K. UnIu, and M. W. Kemblowski, "A Monte Carlo Model to Assess Effects of Land
Disposed E & P Waste on Groundwater," SPE Annual Technical Conference & Exhibition, 1993.
Philip, J. R. "The theory of infiltration: 1. The infiltration equation and its solution." Soil Science.
83: 345-357, 1957.
Raudkivi, A. J., and R. A. Callander, Analysis of Groundwater Flow, Edward Arnold, London, UK,
1976.
Rosenshein, J., and G. D. Bennet, eds., Groundwater Hydraulics, American Geophysical Union,
Washington, DC, 1984.
Rubin, H., and G. F. Pinder., "Approximate Analysis of Upconing." Advances in Water Resources,
1 (2): 97-101, 1977.
Sallam, A., W. A. Jury, and J. Letey, "Measurement of Gas Diffusion Coefficient Under Relatively
Low Air-Filled Porosity," Soil Science Society of America Journal 48:3-6, 1983.
Schiegg, H. O., "Considerations on Water,. Oil and Air in Porous Media," Water Science
Technology, 17: 467^76. 1984.
Shafer, J. M., "Reverse Pathline Calculation of Time-Related Capture Zones in Nonuniform Flow,"
Ground Water 25: 283-289, 1987.
Sikkema, P. C. and J. C. Van Dam, "Analytical Formulae for the Shape of the Interface in a SemiConfined Aquifer," Journal of Hydrology, 56: 201-220, 1982.
Sposito, G., and W. A. Jury, "Inspectional Analysis in the Theory of Water Flow Through
Unsaturated Soil," Soil Science Society of America Journal, 42 (1): 1985.
Sposito, G., "Chemical Models of Inorganic Pollutants in Soils," CRC Critical Reviews in
Environmental Control, 15 (1): 1-24, undated.
Strausberg, S. I. "Estimating Distances to Hydrologic Boundaries from Discharging Well Data,"
19th Annual Meeting of the Rocky Mountain Section of the Geological Society of America, Las
Vegas, NV, 1966.
Thornton, J. S., and W. L. Wootan, Jr., "Venting for the Removal of Hydrocarbon Vapors from
Gasoline Contaminated Soil," Journal of Environmental Science and Health, All (1), 3144, 1982.
Todd, D. K., Groundwater Hydrology, 2th John Wiley & Sons, New York, 1980.
Todd, D. K., "Salt-Water Intrusion and Its Control." Journal of the American Water Works
Associations, 180-187, 1973.
U. S. Department of Agriculture Agricultural Research Service, Analytical Solutions of the OneDimensional Convective-Dispersive Solute Transport Equation, Technical Bulletin No. 1661,
UnIu, K., M. W. Kemblowski, J. C. Parker, D. Stevens, P. K. Chong, and I. Kamil, "A Screening
Model for Effects of Land-Disposed Wastes on Groundwater Quality" Journal of Contaminant
Hydrology, 11: 27-^9, 1992. Washington, DC, 1982
Van Genuchten M. T. and W. J. Alves, "Analytic Solution of the One-Dimensional Convective
Solute Transport Equation, Technical Bulletin. 1661, U.S. Departament of Agriculture,
Washington D.C., 1982.
Ward, C. H., M. B. Tomson, P. B. Bedient, and M. D. Lee, "Transport and Fate Processes in the
Subsurface," Water Resources Symposium, VoI 13, WARSAG, 1987.
Warrick, A. W. , J. W. Biggar, and D. R. Nielsen, "Simultaneous Solute and Water Transfer for an
Unsaturated Soil," Water Resources Research. 7: 1216-1225, 1971.
Watson, K. K., and M. J. Jones, "Algebraic Equations for Solute Movement During Absorption,"
Water Resources Research, 20: 1131-1136, 1984.
Wilson, J. L., and L. W Gelhar, "Analysis of Longitudinal Dispersion in Unsaturated Flow 1: The
Analytical Method," Water Resources Research, 17 (1): 122-130, 1984.
Wilson, J. L., and P. J. Miller, "Two-Dimensional Plume in Uniform Ground-Water Flow Discussion," Journal of the Hydraulics Division, American Society of Civil Engineers, 103
(HY12): 1567-1570, 1979.
Wilson, J. L., and P. J. Miller., "Two-Dimensional Plume in Uniform Ground-Water Flow," Journal
of the Hydraulics Division, American Society of Civil Engineers, 104 (HY4): 503-514, 1978.
Wirojanagud, P., and R. Charbeneau., "Saltwater Upconing in Unconfined Aquifers," Journal of
Hydraulic Engineering, American Society of Civil Engineers, 111: 417^-34, 1985.
Yeh, G. T., Analytical Transient One-, Two-, and Three-Dimensional Simulation of Waste Transport
in the Aquifer System, Environmental Sciences Division Publication No. 1439, Oak Ridge
National Laboratory, Oak Ridge, TN, 1981.
CHAPTER 5
ENVIRONMENTAL
HYDRAULICS
Richard H. French
Water Resources Center
Desert Research Institute
University and Community College System of Nevada
Reno, Nevada
Steven C. McCutcheon
Ecosystems Research Division
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Athens, Georgia
James L Martin
AScI Corporation
Athens, Georgia
5.1
INTRODUCTION
The thermal, chemical, and biological quality of water in rivers, lakes, reservoirs, and near
coastal areas is inseparable from a consideration of hydraulic engineering principles;
therefore, the term environmental hydraulics. In this chapter we discuss the basic principles of water and thermal budgets as well as mixing and dispersion.
5.2
WATERANDTHERMALBUDGETS
(5.1)
and the expressions of the water budget can range from simple to very complex. For
example, consider the lake or reservoir shown in Figure 5.1. For this situation, a generic
water budget could be written as follows
^ = (/, + / + /, + Pr + Rr) ~(Ev + Tr+Gs + Oc+W) (5.2)
Aquatic
vegetation
Water table
Water table
Storage, 5s
FIGURE 5.1 A hypothetical lake illustrating the variables in the water budget.
where Ic = channel inflow rate, I0 = overland inflow rate, Ig = groundwater inflow rate,
Pr = precipitation rate, Rr = return flow rate, Ev = evaporation rate, Tr = transpiration
rate, G5 = groundwater seepage rate, O0 = channel outflow rate, W = consumptive withdrawal, and Ss = lake/reservoir storage rate at time t (volume).
The solution of Eq (5.2) quantifies the terms, and, in many cases, the goal of the modeling effort is to estimate the value of a single term or group of terms: for example, evapotranspiration (Ev + Tr). The reliability of using a water budget is directly related to the
accuracy of the prediction techniques used, the availability and quality of gauged data,
and the time period involved. Among the methods of evaluating the individual terms in
Eq. (5.2) are the following:
(5.3)
where H = net surface heat flux, Q5 = shortwave radiation incident to the water surface
[30-300 (kcal/m2)/h], Qsr = reflected shortwave radiation [5-25 (kcal-m2)/h], Qa =
incoming longwave radiation from the atmosphere (225-360 kcal/m2/h), Qar = reflected
longwave radiation [5-15 (kcal-m2)/hr], Qbr = longwave back radiation emitted by the
water body [220-345 (kcal-m2)/h], Qe = energy utilized by evaporation [25-900
(kcal-m2)/h], and Q0 = energy converted to or from the body of water (35-50 kcal-m2
/hr). Note that the ranges given are typical for the middle latitudes of the United States
(Bowie et al., 1985).
The equations for estimating the terms of the thermal budget use a mixed set of units,
and appropriate conversions among the different units used are provided in Table 5.1.
5.2.2.1 Net atmospheric shortwave radiation (Qs Qsr). The net shortwave radiation
(Qsn) is that portion of the incident shortwave radiation captured at the water surface, taking into account losses caused by reflection. Although solar radiation can be measured
with specialized meteorological stations equipped with radiometers, these instruments
require painstaking calibration and maintenance. In most cases, measured values of solar
radiation are not available at the location of interest and must be estimated from equations.
Among the formulations for estimating net shortwave solar radiation is
Qsn = QS- Qsr = 0.94fcc(l - 0.65C2C)
(5.4)
where Qsc = clear sky solar radiation [kcal Tn Xh) and C0 = fraction of sky covered by
clouds (Anderson, 1954; Ryan and Harleman, 1973). It is pertinent to note that Eq. (5.4)
=
=
=
=
0.131 W/m2
=
7.61 Btu-ft2)/day =
0.485 W/m2
=
1.16 W/m2
=
0.271 Ly/day
=
2.07 Ly/day
=
3.69 (Btu/-ft2)/day =
2.40 Ly/day
=
0.113 (kcal-m2)/h
0.86 (kcal-m2)/h
0.42 (kcal-m2)/h
8.85 (Btu-ft2)/day
1 kpa
lmb
lmm Hg
I m Hg
=
=
=
10 mb
0.1 kpa
1.3mb
33.0mb
7.69 mm Hg
0.769 mm Hg
0.13 kpa
25.4 mm Hg
0.303 in (Hg)
0.03 in (Hg)
0.039 in (Hg)
3.3 kpa
=
=
=
=
=
=
=
=
assumes average reflectance at the water surface and uses clear sky solar radiation. In
some situations, the effects of atmospheric attenuation are much greater than normal and
more complex equations are required (e.g., 1972). Clear sky radiation (Qsc) can be estimated as a function of calendar month and latitude from Fig. 5.2.
Shortwave solar radiation is absorbed at the water surface and penetrates the water
column, depending on the wavelength of the radiation, the properties of the water, and the
matter suspended in the water. The degree of penetration of shortwave solar radiation
(sunlight) into the water column has a significant effect not only on water temperature but
also on the rate of photosynthesis by aquatic plants and the general clarity, color and aesthetic quality of the water. The penetration of shortwave solar radiation is described by
/ - I0exp (-k e y)
(5.5)
4.61
ye = -jp
(5.7)
5.2.2.2 Net atmospheric long-wave radiation (Qa - Qar). Atmospheric radiation is characterized by much longer wavelengths than solar radiation because the major emitting
elements are water vapor, carbon dioxide, and ozone. The approach generally used to
estimate this flux involves the empirical estimation of an overall atmospheric emissivity
and the use of the Stephan-Boltzman law (Ryan and Harleman, 1973). Swinbank (1963)
developed the following equation, which has been used in many water quality models
Q0n = Qa~ Qar = U6 X 1013(1 + 0.11C2J(T0 + 46O)6
(5.8)
where Qan = net long-wave atmospheric radiation (Btu/ft2/day), Cc = fraction of sky covered by clouds, and T0 = dry bulb air temperature (0F).
5.2.2.3 Long-wave back radiation (Qbr). The long-wave back radiation from a water
surface in most cases is the largest of all the fluxes in the heat budget (Ryan
and Harleman, 1973). The emissivity of a water surface is well known; therefore, this
flux can be estimated with a high degree of accuracy as a function of the water surface
temperature:
FIGURE 5.2 Clear sky solar radiation. (From Hamon et al. 1954)
Qbr = 0.9707?
(5.9)
where Qbr = long-wave back radiation (cal/m /s), Ts = surface water temperature (K), and
a = Stefan-Boltzman constant (1.357 X IQ-8 cal-m2/s/K4)
5.2.2.4 Evaporative heat flux (Qe). Evaporative heat loss (kcal/m2/s) occurs as a result
of the change of state of water from a liquid form to vapor and is estimated by
Q. = P^A
(5.10)
where Lw = latent heat of vaporization (= 597 Q.51TS, kcal/kg), Ts = surface water temperature (0C), Ev = evaporation rate (m/s), and p = water density (kg/m3).
A standard expression for evaporation from a natural water surface is
Ev = (a + bW)(es - ea)
(5.11)
^^H^Tr^o)
'
where es is in inches of Hg, and Ts = water surface temperature ( F). There are a number
of ways of estimating ea, depending on the available data. For example, if the relative
humidity (Rn) is known, then
RH = ^
(5-13)
and then if the wet bulb temperature and atmospheric pressure are known (Brown and
Barnwell, 1987)
ea = e,- 0.000367Pa(7; - T.'Jl + ^32J
(5.14)
where all pressures are in (in Hg), all temperatures are in (0F), P0 = atmospheric pressure,
and Twb = wet bulb temperature. The relationship among the air and wet bulb temperatures (0F) and relative humidity (Thackston, 1974) is
Twb= (0.655+ 036RH)Ta
(5.15)
There are many equations for estimating the rate of evaporation. For example, Jobson
(1980) developed a modified formula that was used in the temperature modeling of the
San Diego Aqueduct and subsequently was modified for use on the Chattahoochee River
in Georgia (Jobson and Keefer, 1979). McCutcheon (1982) noted that, in many models,
the wind speed function is a catchall term that compensates for many factors, such as (1)
numerical dispersion in some models, (2) the effects of wind direction, fetch, channel
width, sinuosity, bank, and tree height, (3) the effects of depth, turbulence, and lateral
velocity distribution; and (4) the stability of air moving over the stream. (Fetch is the distance over which the wind blows or causes shear over the water's surface.) Finally, it is
important to note that evaporation estimators that work well for lakes or reservoirs will
not necessarily provide the same level of performance when used in streams, rivers, or
constructed open channels.
5.2.2.5 Convective heat flux (Qc). Convective heat is transferred between air and water
by conduction and is transported to or from the air-water interface by convection. The
convective heat flux is related to the evaporative heat flux (Qe) by the Bowen ratio (Bowie
et al., 1985), or
RB = ^JT = (6.19 X 1(M)P, eTs~eTa
Qe
s~ a
(5.16)
where all temperatures are in (0C), all pressures are in (mb), and R8 = Bowen ratio.
5.2.2.6 Conclusion. The foregoing is a brief summary of the approaches used most frequently to estimate surface heat exchange in numerical models. The reader is referred to
other publications for a more detailed discussion of the approaches (Bowie et al., 1985)
and meteorological data requirements (Shanahan, 1984). Note that each situation should
be considered carefully from the viewpoint of specific factors that must be taken into
account. For example, in most lakes, estuaries, and deep rivers, the thermal flux through
the bottom is not significant. However, in water bodies with depths less than 3 m (10 ft),
bed conduction of heat can be significant in determining the diurnal variation of temperatures within the body of water (Jobson, 1980, Jobson and Keefer, 1979).
5.3
5.3.1 Effects
The density of water is strongly affected by temperature and the concentrations of dissolved and suspended solids. Regardless of the cause of differences in water density, water
with the greatest density is found at the bottom, whereas water with the least density resides
at the surface. When density gradients are strong, vertical mixing is inhibited. Stratification
is the establishment of distinct layers of water of different densities (Mills et al., 1982).
Stratification is enhanced by quiescent conditions and is destroyed by wind stress, turbulence caused by large inflows, and destabilizing changes in water temperature. In many
bodies of water (rivers, lakes, and reservoirs), stratification is the single most important
phenomena affecting water quality.
When stratification is absent, the water column is mixed vertically and dissolved oxygen (DO) is present in the vertical water column from the top to the bottom: that is, fully
mixed water columns do not have DO deficit problems. For example, when stratification
occurs, in reservoirs and lakes mixing is limited to the epliminion or surface layer. Since
stratification inhibits, vertical mixing is inhibited by stratification, and reaeration of the
bottom layer (the hypoliminion) is inhibited if not eliminated. The thermocline (the layer
of steep thermal gradient between the epiliminion and hypoliminion) limits not only mixing but also photosynthetic activity as well. The hypolimnion has a base oxygen demand
and benthic matter and the settling of particulate matter, from the epiliminion only adds
to this demand. Therefore, while the demands of DO in the hypoliminion increase during
the period of stratification, inhibition of mixing between the epiliminion and the
hypolimnion and the lack of photosynthetic activity deplete the DO concentrations in the
hypolimnion. Finally, a rule of thumb suggests that when water temperature is the predominant cause of differences in water density a temperature gradient of at least lC/m is
required to define the thermocline (Mills et al., 1982).
The density of water can be estimated by
p = pr + Aps
(5.17)
(5.19)
(Ford and Johnson, 1983), where CTDS = concentration of TDS (g/m3 or mg/L). If the concentration of TDS is specified in terms of salinity (Gill, 1982)
ApSL = CSL(0.824493 - 4.0899 X 10"3 Te + 7.6438 X 10~5 T2
-8.2467 X 10~7 T/ + 5.3875 X 10~9 Te4)
+ C81L (-5.72466 X 10"3 + 1.0277 X 104 Te
-1.6546 X 10"6 Te2) + 4.8314 X 10"4 CSL~2
(5.20)
where CSL = concentration of salinity (kg/m ). The density increment for suspended
solids is
Ap88 = CJl. --1 X l O 3
(5.21)
I
*GJ
where SG = specific gravity of the suspended sediment (Ford and Johnson, 1983).
The total density increment caused by solids is then
Ap8 - (AoTDS or ApSL) + Apss
(5.22)
5.4
MIXINGANDDISPERSIONINOPENCHANNELS
Turbulent diffusion (mixing) refers to the random scattering of particles in a flow by turbulent motions, whereas dispersion is the scattering of particles by the combined effects
of shear and transverse turbulent diffusion. Shear is the advection of a fluid at different
velocities at different positions within the flow.
When a tracer is injected into a homogeneous channel flow, the mixing process can be
viewed as composed of three stages. In the first stage, the tracer is diluted by the flow in the
channel because of its initial momentum. In the second stage, the tracer is mixed throughout the cross section by turbulent transport processes. In the third stage, longitudinal dispersion tends to erase longitudinal variations in the tracer concentration. In some cases, the
second stage is eliminated because the tracer discharge has a significant amount of initial
momentum associated with it; however, in many cases, the tracer flow is small and the
momentum associated with it is insignificant. In the latter case, the first transport stage is
eliminated. In this section, only the second and third transport stages will be treated, with
the implied assumption that if there is a first stage, it can be treated separately. Section 5.6
details how excessive initial momentum must be analyzed.
The reader is cautioned that, in this chapter, y is the vertical coordinate direction and
z is the transverse coordinate direction.
5.4.1 Vertical Turbulent Diffusion
To develop a quantitative expression for the vertical turbulent diffusion
coefficient,
consider a relatively shallow flow in an infinitely wide rectangular channel. It can be
shown that the vertical transport of momentum in such a flow is given by
^ ^ Ty
where T = shear stress at a distance y above the bottom boundary, p = fluid density, ev =
vertical turbulent diffusion coefficient, and v = longitudinal velocity (French, 1985).
Because the one-dimensional vertical velocity profile and shear distribution are known, it
can be shown that
* = 4 v *()N)
(5 24)
(5.25)
When the fluid is stably stratified, mixing in the vertical direction is inhibited, and one
often quoted formula expressing the relationship between the unstratified and stratified
vertical mixing coefficient was provided by Munk and Anderson (1948)
* - = l + 3.MRi)
(5 26)
'
where evs = the stratified vertical mixing coefficient and Ri = the gradient Richardson
number.
(5.27)
where the 50 percent indicates the error incurred in estimating et. In natural channels, e,
is significantly greater than the value estimated by Eq. (5.27). For channels that can be
classified as slowly meandering with only moderate boundary irregularities
e= 0.6Oy^v* 50%
(5.28)
If the channel has curves of small radii, rapid changes in channel geometry, or severe
bank irregularities, then the value of et will be larger than that estimated by Eq. (5.28).
For example, in the case of meanders, Fischer (1969) estimated that
V 2V3J
^.
e =2
~>
(5 29)
(5.30)
and V = average channel velocity, T = channel topwidth, and Rc = radius of the curve.
A comparison of Eqs. (5.25) and (5.27) shows that the rate of transverse mixing is
roughly 10 times greater than the rate of vertical mixing. Thus, the rate at which a plume
of tracer spreads laterally is an order of magnitude larger than is the rate of spread in the
vertical direction. However, most channels are much wider than they are deep. In a typical case, it will take approximately 90 times as long for a plume to spread completely
across the channel as it will take to mix in the vertical dimension. Therefore, in most
applications, it is appropriate to begin by assuming that the tracer is uniformly distributed
over the vertical.
In a diffusional process in which the tracer is added at a constant mass flow rate (M*)
at the center line of a bounded channel (dC/dz = O at z O and 3C/3z = O at z = 7), the
downstream concentration of tracer is given approximately by
^-vfc.iH-^^^)
+ e5p (_<^^2j]
where
r - M*
VTyd
xet
w?
x=
and
, _ z_
T
A reasonable criterion for the distance required for "complete mixing " (where the concentration is within 5 percent of its mean value everywhere in the cross section) from a
center-line discharge is
O IVT 2
L = ^iLi.
(5.32)
$
If the pollutant is discharged at the side of the channel, the width over which the mixing must take place is twice that for center-line injection, but the boundary conditions are
otherwise identical and Eq. (5.32) applies if T is replaced with 2T.
5.4.3 Longitudinal Dispersion
After a tracer becomes mixed across the cross section, the final stage in the mixing
process is the reduction of longitudinal gradients by dispersion. If a conservative tracer is
discharged at a constant rate into a channel, the flow rate of which also is constant, there
is no need to be concerned about dispersion; however, in the case of an accidental release
(spill) of a tracer into a channel or the release is cyclic, dispersion is important. The onedimensional equation governing longitudinal dispersion is
f+=*f+5
(5.35)
2. In the case of a slug of dispersing material (mass M), the longitudinal length of the
cloud after the initial period can be estimated approximately by
O VT^ TCF
24qji!L-o.07
, (Vf
JJ
(5.36)
(5 37)
=~frr
A I/4TrAjC
V~v^
where A = channel cross-sectional area.
Note that the observed value of the peak concentration will generally be less than this
estimate because some of the material is trapped in dead zones and some of the typical
tracers (Martin and McCutcheon, 1999) sorb onto sediment particles.
TRANSITION
LACUSTRINE
RIVERINE
PLUNGE UNE
HYPOLIMNION
active mixing
patch
oss Ii step
structure
boundary
mixing
withdrawal
layer
FIGURE 5.3 Mixing processes in zones of lakes and reservoirs. (Modified from Fischer, et all979)
SEPARATION POINT
Shallow lakes and reservoirs that do not stratify are normally analyzed in the same
fashion as rivers or as a completely mixed body of water. For a completely mixed system,
the residence time (T in seconds or more typically years) or time for an inflow to travel
through the body of water is simply tr = (|)/g, where ty is the volume of the lake (m3) and
Q is the sum of the inflows or the average reservoir discharge (m3/s).
Freshwater lakes tend to stratify when the mean depth exceeds 10 m and the residence
time exceeds 20 days (Ford and Johnson, 1986). The densimetric or internal Froude number Prd (Norton et al., 1968) provides a better indication of the stratification potential of a
reservoir where
Fr, =
^4 = /.y
= Frp
Mp]
/JAH
^ V gp
V gp y
(5.38)
LL = the length of the reservoir (m), yavg = the its mean depth (m), g = gravitational acceleration (m/s2), Ap = the difference in density over the depth for the internal Fr or between
the inflow and surface waters of the lake or reservoir at the plunge point or separation
point (kg /m3), p = average density of the lake for the internal Fr or density of the inflow
(Turner, 1973) at plunge or separation points (kg/m3), V0 = the average velocity of the
inflow (m/s), and y0 = the hydraulic depth or cross-sectional area divided by the top width
of the inflow (m). The Fr at the plunge point Frp, also defined in Eq. (5.38), will be used
in the next section. For design projections, the dimensionless density gradient Ap/(pyavg)
normally is taken to be 10'6 m-1 (Norton et al., 1968). IfFr I/TC, the reservoir is expected to be well mixed. If Fr 1/Ti9 the reservoir is expected to be strongly stratified, and
when Fr 1/Ti9 the reservoir is expected to be weakly or intermittently stratified.
Using the length or depth scales of Sundaram (1973) and Ford and Johnson (1986) the
depth to which wind can mix and destroy the stratification of a lake or reservoir for a particular surface heat flux is
.^A = 5.9 X 109^
(5.39)
^gHn
Hn
where w* = the shear velocity of the wind (m/s), K = an empirical coefficient approximately equal to the von Karman constant of 0.4, a = the volumetric coefficient of thermal
expansion for water (1.8 X 10^/0C), Hn = surface heat flux (W/m2), p = the density of
water ( 1000 kg/m3), and Cp = the specific heat of water (4186J/kgC).
The wind shear velocity is
DI =
(5.40)
where uw = the wind speed (m/s), pa = the density of air (kg/m3), and Cd = the drag coefficient, which usually is taken to be 1.3 X 10~3. (See Martin and McCutcheon (1999) to
estimate the net thermal energy flux.)
5.5.1 Annual Stratification Cycle
In freshwater lakes, stratification results when the sun heats the water faster than wind
shear can mix the heat over the depth. In saline lakes, differences in both temperature and
dissolved solids cause stratification. Stratification involving salinity may persist yearround in deeper saline lakes. The onset of stratification in freshwater lakes occurs in late
spring or early summer and persists into the fall or early winter, depending on latitude.
The surface heats rapidly, becoming less dense than deeper layers and forming stable differences in vertical density that inhibit vertical mixing until the fall overturn. As stratification develops, wind and currents mix the upper layers and tend to deepen the thermocline to form the well-mixed epilimnion. Although storms in late spring and summer
episodically lower the thermocline, the thermocline generally rises as solar heating
increases until midsummer. After late summer cooling begins, the thermocline deepens
until the fall overturn occurs. The decreased difference in temperature in the fall with the
hypolimnion allows more mixing that deepens the epilimnion and thermocline. The variable depth of the thermocline at any time is controlled by seasonal climate, the occurrence
of storms, water temperature, water depth, lake bathymetry, the strength of inflow and
outflow current, and other factors covered in more detail by Chapra and Reckhow (1983),
Ford and Johnson (1986), Hutchinson (1957), and Wetzel (1983).
The onset of cooler fall conditions causes the epilimnion to lose heat to the atmosphere. As heat is lost, mixing tends to become more dominant. The overturning or complete mixing of the reservoir or lake dominates as the epilimnion and hypolimnion
approach the same temperature. During winter, lakes and reservoirs remain unstratified
except in the higher latitudes where the hypolimnion approaches 40C and the surface
approaches O0C. The slight winter stratification of these colder water bodies is the result
of the usual decrease in water density as temperature decreases from 4 to O0C. Ice cover
maximizes and prevents wind mixing and erosion of the mild differences in density.
Stratification is so mild that a distinct thermocline does not form and the epilimnion and
hypolimnion are not well defined. Winter stratification persists until spring warming melts
the ice and heats the surface layer to the temperature of the hypolimnion (usually 40C)
when the spring overturn occurs.
The arrival of spring begins the cycle of heating and stratification anew. A difference
in temperature of just a few degrees results in a difference in density sufficient to inhibit
or prevent most vertical mixing in lakes and reservoirs. Vertical mixing is inhibited almost
completely during summer heating because wind and inflows and outflows do not have
sufficient energy to erode the differences in density that arise. The wind and energy available from wind and currents cannot overcome the potential energy differences that tend to
prevent mixing of the denser hypolimnion and lighter epilimnion. Fresh water flows into
a saline lake cause salinity gradients that have the same damping effect. Density stratification also is caused by suspended sediments, primarily resulting in sediment-laden
underflows. Martin and McCutcheon (1999) have illustrated the stratification cycle for
warm-water lakes and reservoirs.
Run-of-the-river reservoirs and shallow lakes that are weakly stratified because of high
flows or wind mixing follow only the general stratification trend. Complete mixing may
occur during the summer stratification period as a result of wind or runoff events, and the
thermocline may be difficult to define. Fall overturn occurs earlier in these bodies of water
than it does in deeper lakes.
5.5.2 Plunge and Separation Point-End of the Transition
Between Riverine and Lacustrine Conditions
The plunge point or separation point marks the downstream end of the transition zone
defined where buoyancy begins to exceed advective forces. These points move seasonally and, to a limited degree, during the day. Usually distinguished by a line of foam or floating debris across the reservoir or lake, the plunge point occurs when a denser inflow dives
below the lake surface and continues to flow along the bottom as a density current. The
separation point occurs when an underflow has the same density as the lake water at a
given depth, and separates from the bottom to flow into a discrete layer of the lake as an
interflow. Some underflows may be dense enough to flow to the lowest point in a lake or
to the dam that forms a reservoir. If the inflow is less dense than water at the lake surface,
an overflow occurs. Fig. 5.3 illustrates these three types of inflows.
At the plunge or separation point, the internal Fr of the stratified lake Prd is equal to
the Fr of the inflow at that point (Frp), as noted in Eq. (5.38). If the difference in density
between the lake surface and the inflow Ap is positive, an overflow occurs, and if Ap is
negative, an underflow occurs. If the slope of the reservoir bottom, or valley, is mild (S8
< 0.007), then the hydraulic depth (y0) is the normal depth of flow. For steep slopes (SB
> 0.007), the hydraulic depth is the critical depth (Akiyama and Stefan, 1984).
For tributary or river channels that are approximately rectangular or triangular, the
hydraulic depth and location of the plunge point or separation point can be calculated. For
a rectangular cross section of constant width, the hydraulic depth is
1/3
22
q2
=
yo =
kpT
2 |AP| 2
F
^nr*J
l^nr1
1/3
<5-41>
where Q = the riverine inflow rate (mVs) equal to VA, A = the cross-sectional flow in
area of the river (m2), B = the conveyance width (m), and q = the flow per unit width
(m2/s). Similar expressions were proposed by Akiyama and Stefan (1984), Jain (1981),
Singh and Shah (1971), and Wunderlich and Elder (1973), among and others. Savage and
Brimberg (1973) developed an independent expression for the Froude number at the
plunge point or point of separation (Frp) based on the conservation of energy and the theory of two-layered flow in stratified water bodies, which can be expressed as
\).478
f\
h)
(5.42)
where fb = the dimensionless bed friction factor and f. = dimensionless interfacial friction. Martin and McCutcheon (1999) have illustrated the calculations and summarized the
validation of these equations by an example derived from Ford and Johnson (1981, 1983).
For a triangular cross section with an angle 29 between the channel or valley walls, the
hydraulic depth is one-half the total depth. The area of the cross section (m2) is A = y\
tan(9), which, when substituted into the expression for the normal densimetric number Frn
and solved for the hydraulic depth V0 (m), is
2 2
T
Q
11^5
;^r
V0 = 0.5 / g2Ap
\
2
(Fr n^tan (e)J
(5.43)
where the bottom depth (distance between water surface and apex of the triangular cross
section) is twice the hydraulic depth for a triangular cross section. Hebbert et al. (1979)
derived an expression for the downstream densimetric Froude number at the plunge point
or separation point Frp for normal flow (S8 < 0.007) in a triangular cross-section, related
to the reservoir characteristics as
^= -
(5.46)
[w
Lnr-i
In natural settings, overflows are usually dissipated by mixing caused by wind and solar
heating before traveling too far.
Horizontal spreading of an overflow is estimated using the inflow Fr defined by Eq.
(5.38). Safaie (1979, cited in Ford and Johnson, 1983) found that for Fr^ < 3, the flow is
an unsteady, buoyancy-driven spread and can be assumed to be completely mixed laterally except for abrupt changes in the entrance geometry. Typically, reservoirs widen gradually where major tributaries enter, but lakes may have an abrupt widening at the mouth of
tributaries. For Fr^ > 3, the inflow acts like a jet that expands proportionally with distance
B(x) = BQ + ex where B(x) = the overflow width (m) at distance x measured from the
separation point (m), B0 = the width of the riverine or tributary flow at the separation
point (m), and c = a dimensionless empirical constant (Ford and Johnson, 1983). From
laboratory experiments with plane jets, the value of c has been determined to be approximately 0.16 (Fischer et al. 1979; Ford and Johnson, 1983).
5.5.4 Underflow or Density Current Mixing
Underflows are dominated by two mixing processes. First, significant mixing occurs during the plunge beneath the surface. Second, shear at the interface with ambient lake or
reservoir water will result in mixing and entrainment as the underflow moves downward.
The initial turbulent mixing of the plunging flow will increase the total flow rate of the
underflow and reduce the density and concentration gradients. The fraction entrainment
^ caused by plunging is (Qp Q)IQ, where Qp is the flow rate at the plunge point (mVs)
and Q is the river flow-rate (mVs). For mild slopes SB < 0.007, ^ is on the order of 0.15
(Akiyama and Stefan, 1984). The depth of the underflow is the normal depth of flow. For
steep slopes S8 > 0.007, is on the order of 1.18 and the density current depth is the crit-
ical depth (Akiyama and Stefan, 1984). However, the entrained fraction ^ is highly variable. The dilution of concentrations or temperatures resulting from mixing in plunging
flows follows from a simple mass or heat balance
Cp = ^^
(5.47)
1+5
where C is the inflow concentration (g/m3 or mg/L3) or temperature (0C), C0 is the ambient concentration (g/m3 or mg/L3) or temperature (0C) of the lake, and Cp is the concentration (g/m3 or mg/L3) or temperature (0C) of the plunging flow after initial mixing.
The mixing after plunging results from bottom shear as well as shear at the interface
of the underflow with ambient lake water. For a triangular cross section, the entrainment
coefficient is (Imberger and Patterson, 1981)
E = I^CfFr2
(5.48)
where laboratory experiments indicate that Ck is approximately 3.2 (Hebbert et al. 1979),
C0 = the dimensionless bottom drag coefficient defined following Eq (5.42), and Frb the
internal Froude number
Fr, = -^=
VeA
(5.49)
where ub = underflow velocity, hb = underflow depth, and efc = relative density difference. The entrainment coefficient E is a constant for a specific body of water. The depth
or thickness of the underflow (m) is a linear function of the entrainment coefficient
(Hebbert et al. 1979; Imberger and Patterson, 1981)
yuf = (6/5)Ex + y0
where x is the distance downstream from the plunge point (m) and y0 is the initial thickness of the underflow (m) that is approximately equal to the depth at the plunge point. If
entrainment is limited, the depth of the underflow remains approximately constant as long
as the bottom slope remains constant. The increase in flow rate because of entrainment for
an underflow in a triangular cross section is solved iteratively as
IY >/3
1
GCO = G 1LW
HH ""MJ
(5.50)
where Q1 = the discharge (m3/s) and V1 = the depth (m) from the previous calculation
step. For the initial iteration, Q1 = the discharge at the plunge point Qp (m3/s) and V1 =
the plunge point depth y0 (m).
Because of more significant differences in density and less internal mixing contrasted
with the epilimnion, underflows tend to remain more coherent than overflows. Sedimentladen underflows, especially, tend to travel to the lake outlet or dam.
5.5.5 Interflow Mixing
After experiencing approximately 15 percent entrainment at the plunge point (for mild
slopes) and mixing as an underflow, an interflow intrudes into a lake at the depth at which
neutral buoyancy is achieved. The turbulence generated by bottom shear is dissipated
quickly, and entrainment into the interflow is dominated by interfacial shear with ambient
lake water above and below the intrusion layer.
N2L4
Gr = ^
^V
(5.53)
where ev = the vertically averaged diffusivity (m2/s). Generally, if Gr > 1, then an internal wave field will decay slowly, but if Gr < 1 then viscous dissipation damps waves
quickly (Fischer et al. 1979). Imberger and Patterson (1981) also introduced a dimensionless time variable
tN
1
~ Gr1/6
where t = time(s), which, along with the Prandtl number Pr = ev/e,, where etis vertically averaged diffusivity of heat (mVs), is used to define three interflow conditions:
1. If R > 1, the intrusion is governed by a balance of the inertial and buoyancy forces
so that the actual intrusion length L1. is proportional to time, as given by (Ford and
Johnson, 1983; Imberger et al., 1976)
L1 = 0.44L1. VRf
= 0.44 V^7M
(5.54)
If the speed of the intrusion is constant or uniform, the velocity V7 is L/f, so that
( g AD
A Vr V2
,
7
V1 = 0.44 V^V = 0.194 ^p
where pw = the density of the intrusion. The difference in density in the computation
of the buoyancy frequency is that occurring over the thickness of the intrusion hm,
which, along with the relationship um = qjhm, can be substituted into the above equation to yield an alternative formulation for the speed of intrusion. The thickness of the
interflow can be solved by assuming uniform flow (Ford and Johnson, 1983)
fcm = 2.99 Ml;]"3
2
I8P1 J
(5.55)
where hm is generally distributed equally above and below the center line of the
intrusion.
2. If R < t*R < P2/3, then the flow regime is dominated by the balance between viscous and buoyancy forces and the intrusion length becomes
L1. = 0.51 L Rm f516
(5.56)
(5.57)
In this regime, the flow is generally distributed so that 64 percent lies above the center line of the intrusion (Imberger, 1980); thus, the half-thickness (HnJ of the interflow
above the center line is given by
hma = 3.5LJG-^
(5.58)
(5.59)
3. If P2/3< f < R'1 then the flow regime is dominated by viscosity and diffusion and
the intrusion length becomes
L1 = C6LK3" t3/4
(5.60)
UPPER LIMIT
Withdrawal
Zone
VELOCITY
PROFILE
FIGURE 5.4 Reservoir withdrawal. (Adapted from Martin and McCutcheon, 1999)
DENSITY
PROFILE
where Z = distance from the port center line to the upper or lower withdrawal limit; 6 =
the withdrawal angle (radians); and TV = the buoyancy frequency [gAp/(pZ)]1/2, in which
Ap = the difference in density between that at the upper or lower withdrawal limit and at
the port centerline; and p = the density (kg/m3) at the port center line. The convention is
that Ap is positive for stably stratified flows such that Ap = p (upper limit) - p (withdrawal port) or Ap = p (withdrawal port) - p (lower limit). The elevation of the water
surface, the bottom, of the reservoir, and the withdrawal port and the density profile must
be known. The equation must be solved iteratively since both the distance from the port
center line Z and the density as a function of Z are unknown. A typical solution procedure
where the upper and lower withdrawal zones can form freely within the reservoir without
interference at the surface or bottom is as follows:
1. Rearrange the equation as Qout Z3M)/7i = O.
2. Check to see if interference exists by, first, using Z equal to the distance from the
port center line to the surface. Estimate the density at the center line of the withdrawal port and the water surface and substitute the values into the rearranged
equation. If the solution is non-zero and is positive, surface interference exists.
truncated due to
surface interference
interference limit
VELOCITY
free
PROFILE
lower
limit
FIGURE 5.5 Definition of withdrawal characteristics. (From Martin and McCutcheon, 1999)
Similarly, substitute the distance from the port center line to the bottom, along with
the density at the bottom of the reservoir, and determine if a bottom interference
exists.
3. If both of the evaluations from Step 2 are negative, the withdrawal zone forms
freely in the reservoir. The limit of the surface withdrawal zone above the port
can be determined by using iterative estimates of values for Z and the density at
the height above the center line until the equation approaches zero to within some
tolerance. The lower limit of withdrawal below the port center line can be determined in a similar manner.
4. If surface or bottom interference exists, a theoretical withdrawal limit can be
determined using values of Z computed using elevations above the water surface
for surface interference or below the reservoir bottom for bottom interference.
However, this solution requires an estimate of density for regions outside the limits of the reservoir. Davis et al. (1985) estimated these densities by linear interpolation using the density at the port center line and the density at the surface or
bottom of the reservoir.
For the case where one withdrawal limit intersects a boundary and the other does not,
the freely forming withdrawal limit cannot be estimated precisely using the rearranged
equation. Smith et al. (1985) proposed an extension to estimate the limit of the freely
forming layer similar to that described above
O81 = 0.125(D-JPeL
n
N
[
I
n
sin
I D^d } + D--d\
( D
)
D J
where d = the distance from the port center line to the boundary of interference (m) and
D = the distance between the free withdrawal limit and the boundary of interference (m)
shown in Fig. 5.5. The length scale in the buoyancy frequency Af is D in place of Z, and
Ap is the difference in the density between that at the surface for withdrawals that extend
to the surface and between the lower free limit or density at the bottom for withdrawals
that extend to the bottom and upper free limit. For consistency with the definition of stable stratification as positive, the convention is that Ap = p(surface layer) - p(free limit)
or Ap = p(upper free limit) p(bottom layer).
Once the limits of withdrawal are established, the distribution of withdrawal velocity is estimated by dividing the reservoir into layers, the density of which is determined
at the center line of each layer. The computation of the vertical velocity distribution is
based on the location of the maximum velocity, which can be estimated from (Bohan
and Grace, 1973)
(
Y
YL = HsirAL57^\
(5.63)
H
I
)
where Y1 = the distance from the lower limit to the elevation of maximum velocity (m),
H = the vertical distance between the upper and lower withdrawal limits (m), and ZL =
the vertical distance between the outlet center line and the lower withdrawal limit (m)
shown in Fig. 5.4. If the withdrawal intersects a physical boundary, the theoretical withdrawal limit is used, which may be above the water surface or below the reservoir bottom.
Once the location of the maximum velocity Vmax (m/s) is determined, the normalized
velocity VN(I) = V(I)/Vmax in each layer / is estimated for withdrawal zones that intersect
a boundary as (Bohan and Grace, 1973)
-'-(l
or for a withdrawal that does not intersect a boundary
^ = (,-*M>J
(5.65)
where V(T) = the velocity in layer 7 (m/s), y(/) = the vertical distance from the elevation
of maximum velocity to the center line of layer / (m), Y1 = the vertical distance from the
elevation of maximum velocity to the upper or lower withdrawal limit (m) determined by
whether the centerline of layer / is above or below the point of maximum velocity, Ap(T)
= the density difference between the elevation of maximum velocity and the center line
of layer /, and Apmax = the difference in density between the point of maximum velocity
and the upper or lower withdrawal limit.
If the withdrawal intersects the surface or the bottom, velocities are calculated for locations either above the water surface or below the reservoir bottom and the distribution is
truncated at the reservoir boundaries to produce the final velocity distribution. The flow
rate in each layer / is
w = J^
Q~
<5-66)
Ew
i= i
where Qout = the total release rate and m = the number of layers. The quality of the release
can be determined from a simple flow-weighted average or mass balance as
q(I) C(I)
C* = -^2
(5-67)
E 4
/= i
where CR = the concentration or temperature of water quality constituent C in the release
and C(I) = the concentration or temperature in each layer.
For discharge over a weir, the withdrawal limit Z and average velocity in the withdrawal zone Vweir is derived from the densimetric Froude number [Eq. 5.38] as (Grace,
1971, Martin and McCutcheon, 1999),
(5-68)
where Ap = the difference in density between the weir crest and the lower withdrawal
limit, p = the density at the weir crest elevation, Hw = head above the weir crest elevation, Z = distance between the crest elevation and the lower withdrawal limit, and C1 and
C2 are constants, which have values of
(5.69)
ness of the bottom. A spilling breaker tends to form over a gradually shoaling bottom and
tends to break over long distances, with the wave collapsing downward in front of the
wave. Plunging breakers occur when the bottom shoals rapidly or when the direction of
the wind opposes the wave. The plunging breaker begins to curl and then collapses before
the curl is complete. A plunging or surging breaker does not break or collapse but forms
a steep peak as the wave moves up the beach. The type of breaking wave and the associated energy controls beach erosion, aquatic plant growth, surf-zone mixing, and the
exchange of contaminants between surface and ground waters.
After breaking, waves continue to move up a gradually sloping beach until the force
of gravity forces the water back. The extent to which the water runs up the beach is called
the swash zone. The movement of the swash up the beach may result in the deposition of
particles and debris, causing swash marks at the highest point of the zone. Wave run-up
in the swash zone also sets up an imbalance of momentum along the porous beach face
that pumps contaminants into and out of the beach (McCutcheon, 1989).
In large lakes and reservoirs with an extremely long fetch, parallel pairs of large vertical vortices or circulatory cells known as Langmuir circulation develop at an angle of
15 clockwise with the general direction of a sustained wind, when wave and current conditions are favorable. The depth of the vortices depends on stratification and may interact
with internal waves formed on the thermocline, deepening over the troughs of internal
waves. Where the counterrotating Langmuir cells converge, visible streaks or bands form
on the surface that tend to accumulate floating debris. In the convergence zone, downward
velocities of 2-6 cm/s carry surface waters toward the thermocline. These downward currents move in a circular fashion and turn upward into a divergence zone midway between
the Langmuir=streaks. Water near the thermocline moves to a zone near the surface at a
velocity of about 1 to 2 cm/s over a larger area. As first proposed by Langmuir (1938),
this type of large-scale circulation also contributes to the vertical mixing of the epilimnion. Like smaller-scale orbital wave mixing, the effect of Langmuir circulation is
lumped into values selected for the eddy viscosities and eddy diffusivities of the epilimnion.
Because of the smaller differences in density across density interfaces within a body of
water, internal waves travel more slowly than do surface waves, but they achieve greater
wave heights. Internal waves include standing waves, such as seiches (Mortimer, 1974) and
internal hydraulic jumps (French, 1985), but most are progressive waves that radiate energy
from the point at which the waves were generated (Ford and Johnson, 1986). Wind shear,
water withdrawals, hydropower releases, and thermal discharges as well as local disturbances produce internal waves. The most significant mixing between stratified layers occurs
when internal waves break (Turner, 1973). Before breaking, internal waves mix the water
adjacent to the interface and sharpen the density interface to increase the likelihood of breaking. When wave breaking does occur, the entrained water is mixed through the adjacent
layer.
Among the most important internal waves is the seiche. As defined above, seiches are
periodic oscillations of the water surface and density interfaces resulting from a displacement. Displacements are typically caused by large scale wind events or large withdrawals.
Sustained wind across a lake surface increases the elevation of the water surface at the
downwind boundary of the lake, causing wind setup. As the wind subsides, the water surface tilt or displacement results in a sloshing motion, or seiche, of the lake surface and in
thermocline if the lake is stratified. If hydropower operations or reservoir releases change
the net flow toward the dam, the water piles up at the dam and forms a seiche, often resulting in noticeable differences in thermocline depths between periods of operation and nonoperation, such as between weekdays and weekends. More rarely, a seiche may result
from earthquakes or other geologic events. During the rocking or sloshing, potential energy is converted to kinetic energy and is dissipated by bottom friction.
Wind setup in Lake Erie may exceed 2 m during severe storms (Wetzel, 1975), but for
a moderate storm blowing over the long axis of Green Bay, Wisconsin the wind setup has
reached approximately 12 cm (Martin and McCutcheon, 1999). An estimate of wind setup
can be obtained from the onedimensional equation of motion assuming constant depth,
negligible bottom stress, and steady-state conditions in an unstratified lake, or
ai = p^ = l
(5/70)
fa
spy
gy
where = the deviation of the water surface (m), x = the horizontal distance (m), pa =
the density of air (kg/m3), p = the density of water (kg/m3), C0 = the dimensionless drag
coefficient, uw = the wind speed (m/s), y = the water depth (m), and v* = the friction
velocity in water (m/s) or (Ts/p)05, in which T8 is the surface shear stress (kg/m- s2).
The term dfydx is positive in the direction of the wind. The divergence between wind and
shear force is negligible in shallow lakes and reservoirs but not in deep oceans.
5.6
A jet is the discharge of a fluid from an opening into a large body of the same or similar fluid
that is driven by momentum. A plume is a flow that, while resembling a jet, is the result of
an energy source providing the fluid with positive or negative buoyancy rather than momentum relative to its surroundings. Many discharges into the environment are discussed in
terms of negatively or positively buoyant jets, implying that they derive from sources that
provide both momentum and buoyancy. In such cases, the initial flow is driven primarily by
the momentum of the fluid exiting the opening; however, if the exiting fluid is less or more
dense than the surrounding fluid, it is subsequently acted on by buoyancy forces.
Jets and plumes can be classified as either laminar or turbulent, with the difference
between the two being described by a Reynolds number, as with pipe flow. Near the
source of the flow, the flow of a jet or plume is controlled entirely by the primary initial
conditions that include the mean velocity of the jet's exit, the geometry of the exit, and
the initial difference in density between the discharge and the surrounding, or ambient,
fluid. Secondary initial conditions include the intensity of the exiting turbulence and the
distribution of the velocity. Following Fischer et al. (1979), the factors of prime importance to jet dynamics can be defined as follows:
1. Mass flux the mass of fluid passing a jet cross section per unit time
mass flux =
(pu)dA
(5.71)
JA
where A is the cross-sectional area of the jet and u the time-averaged velocity of the
jet in the axial direction.
2. Momentum flux amount of momentum passing a jet cross section per unit time
(pu2)dA
(5.72)
JA
3. Buoyancy flux buoyant or submerged weight of the fluid passing a jet cross section per unit of time
momentum flux =
(5.73)
where Ap = the difference in density between the surrounding fluid and the fluid in the
jet. It is convenient to define g(Ap)/p = g' as the effective gravitational acceleration.
(5.76)
and Ap0 = difference in density between the plume fluid and the ambient fluid and g\ =
apparent gravitational acceleration.
Results similar to those for jets are summarized in Table 5.3, and the numerical constants given are from Chen and Rodi (1976).
Round Jet
Plane Jet
b0y0 V0
-4
^oVo^o
VM
Maximum time-averaged
velocity
ym
Q
^\
vm =- = (7.0 O. Ij -
M
yj
Q
(j\
vw1 -^ = (2.41 0.04) -2"M
\yj
Maximum time-averaged
tracer concentration
Cm
^
fa\
-f
(5.6
O.I)U
C
V)
^
(^ \
-f
=
(2.38
0.04)
-^
C
V)
Mean dilution
Ratio
C1-/^
, o
1.4 * 0.1
-^
M
1.2 0.1
Round Plume
Plane Plume
Maximum time-averaged
velocity
vm
1.66B1'3
Maximum time-average
tracer concentration
C1n
2.38M B1'3 B1
Volume flux
Q
0.34 B^y
1.4 0.2
0.81 0.1
Ratio
Cn/Cavg
Source: After Fischer et al. 1979.
REFERENCES
Akiyama, J. and H. Stefan, "Onset of Underflow in Slightly Diverging Channels," Journal of the
Hydraulics Division, American Society of Civil Engineers 113(HY7), 1987, pp. 825-844.
Akiyama, J., and H. Stefan, "Plunging Flow into a Reservoir, Theory," Journal of the Hydraulics
Division, American Society of Civil Enginners, 110(HY4): 484^99, 1984.
Anderson, E. R. "Energy Budget Studies, Part of Water Loss Investigations Lake Hefner
Studies.", U.S. Geological Survey, Professional Paper No. 269, Washington, DC, 1954.
Beeton, A. M., "Relationship Between Secchi Disk Readings and Light Penetration in Lake
Huron," American Fisheries Society Transactions, 87:73-79, 1958.
Bender, M. D., G. E. Hauser, M. C. Shiao and W. D. Proctor, "BETTER: A Two-Dimensional
Reservoir Water Quality Model, Technical Reference Manual and User's Guide," Report No.
WR28-2-590-152, Tennessee Valley Authority, Engineering Laboratory, Norris, TN, 1990.
Bohan, J. P., and J. L. Grace, Jr., "Selective Withdrawal from Man-Made Lakes: Hydraulics
Laboratory Investigation," Technical Report No. H-73-4, U.S. Army Waterways Experiment
Station, Vicksburg, MS, 1973.
Bowie, G. L., W. B. Mills, D. B. Porcella, C. L. Campbell, J. R. Pagenkopf, G. L. Rupp, K. L.
Johnson, W. H. Chan, S. A. Gherini, and C. E. Chamberlin, "Rates, Constants, and Kinetics
Formulations in Surface Water Quality Modeling", 2nd ed. EPA/600/3-85/040, U.S.
Environmental Protection Agency, Environmental Research Laboratory, Athens, GA 1985.
Brown, L. C. and T. O. Barnwel, The Enhanced Stream Water Quality Models QUAL2E and
QUAL2E-UNCAS: Documentation and User Manual,. EPA/600/3-87/007, U.S. Environmental
Protection Agency, 1987.
Chapra, S. C., and K. H. Reckhow, Engineering Approaches for Lake Management, Butterworth,
Boston, 1983.
Chen, C. J., and W. Rodi, A Review of Experimental Data of Vertical Turbulent Buoyant Jets,
Hydraulic Research Report No. 193, Iowa Institute of Hydraulics, Iowa City, IA, 1976.
Cole, T. M., and E. M. Buchak, CE-QUAL-W2: A Two Dimensional, Laterally Averaged,
Hydrodynamic and Water Quality Model, Version 2.0, User Manual, Instruction Report No.
EL-95-1, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, 1995.
Davis, J. E., J. P. Holland, M. L.Schneider, and S. C. Wilhelms, SELECT, A Numerical, One
Dimensional Model for Selective Withdrawal, Technical Report, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS, 1985.
Environmental Laboratory, CE-QUAL-Rl: A Numerical One-Dimensional Model of Reservoir
Water Quality, User's Manual, Instruction Report No. E-82-1, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS, 1985.
Fischer, H. B., "The Effects of Bends on Dispersion in Streams," Water Resources Research, 5
496-506, 1969.
Fischer, H. B., E. J. List, R. C. Y. Koh, J. Imberger, and N. H. Brooks, Mixing in Inland and
Coastal Waters, Academic Press, New York, 1979.
Ford, D. E., and M. C. Johnson, An Assessment of Reservoir Mixing Processes, Technical Report
No E-86-7, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, 1986.
Ford, D. E., and M. C. Johnson, An Assessment of Reservoir Density Currents and Inflow
Processes, Technical Report No. E-83-7, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS, 1983.
Ford, D. E., and M. C. Johnson, "Field Observations of Density Currents in Impoundments," in H.
G. Stefan, ed. Proceedings Symposium on Surface Water Impoundments, ASCE, New York, 1981.
French, R. H., Open-Channel Hydraulics, McGraw-Hill, New York, 1985.
French, R. H., J. J. Cooper, and S. Vigg, "Secchi Disc Relationships," AWRA, Water Resources
Bulletin, 18 (1):121-123, 1982.
Gill, A. E., "Appendix 3, Properties of Seawater," in Atmosphere-Ocean Dynamics, Academic
Press, New York, 1982, pp. 599-600.
Gu, R., S. C. McCutcheon, and P. F. Wang, "Modeling Reservoir Density Underflow and Interflow
from a Chemical Spill in a River," Water Resources Research, 32:695-705, 1996.
Grace, J. L., Jr., "Selective Withdrawal Characteristics of Weirs: Hydraulics Laboratory
Investigation," Technical Report H-71-4, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS, 1971.
Hamon, R. W, L. L. Weiss, and W. T. Wilson, "Insulation as an Empirical Function of Daily
Sunshine Duration," Monthly Weather Review, 82(6), 1954.
Hebbert, J., J. Imberger, I. Loh, and J. Patterson, "Collie River Flow into Wellington Reservoir,",
Journal of the Hydraulics Division, American Society of Civil Engineers 105(HY(5): 533-545,
1979.
Hutchinson, G. E., A Treatise on Limnology: Vol. 1. Geography, Physics and Chemistry, John
Wiley Sons, New York, 1957.
Imberger, J., "Selective Withdrawal: A Review," in T, Carstens and T. McClimans, eds.,
Proceedings of the 2nd International Symposium on Stratified Flow, International Association for
Hydraulic Research, Tapir, Trondheim, Norway, 1980.
Imberger, J. and J. C. Patterson, "Dynamic Reservoir Simulation Model DYRESM:5," in H.B.
Fischer, ed., Transport Models for Inland and Coastal Waters, Academic Press, Orlando, FL,
1981, pp. 310-561.
Imberger, J., R. T. Thompson, and C. Fandry, "Selective Withdrawal from a Finite Rectangular
Tank," Journal of Fluid Mechanics, 78:489-512, 1976.
Jain, S. C., "Plunging Phenomena in Reservoirs," in H. G. Stefan, ed., Proceedings of the
Symposium on Surface Water Impoundments, American Society of Civil Engineers, new York,
1981.
Jobson, H. E., "Thermal Modeling of Flow in the San Diego Aqueduct, California, and Its Relation
to Evaporation," Professional Paper No. 1122, U.S. Geological Survey, Washington, DC, 1980.
Jobson, H. E., and T. N. Keefer, "Modeling Hight Transient Flow, Mass, and Heat Transport in the
Chattahoochee River Near Atlanta, Georgia," Professional Paper No. 1136, U.S. Geological
Survey, Washington, DC, 1979.
Kent, R. E. and D. W. Pritchard, "A Test of Mixing Length Theories in a Coastal Plain Estuary,"
Journal of Marine Research, 1:456-466, 1957.
Kao, T. W, "Principal State of Wake Collapse in a Stratified Fluid, Two-Dimensional Theory,"
Physics of Fluids, 19:1071-1074, 1976.
Langmuir, L, "Surface Motion of Water Induced by Wind," Science, 87:119-123, 1938.
Lau, Y. L., and B. G. Krishnappen, "Transverse Dispersion in Rectangular Channels," ASCE,
Journal of the Hydraulics Division, American Society of Civil Engineers 103(HYlO): 1173-1189,
1977.
Lerman, A., ed, Lakes: Chemistry, Geology, Physics, Springer-Verlag, New York, 1978.
Martin, J. L. and S. C.McCutcheon, Hydrodynamics and Transport for Water Quality Modeling,
Lewis Publishes-Boca Raton, FL, 1999.
McCutcheon, S. C., Water Quality Modeling: Vol. I. Transport and Surface Exchange in Rivers,
CRC Press, Boca Raton, FL, 1989.
McCutcheon, S. C., "Discussion with Harvey Jobson on Windspeed Coefficients for Stream
Temperature Modeling," Memorandum, U.S. Geological Survey, NSTL Station, MS, March 29,
1982.
Mills, W. B., J. D. Dean, D. B. Porcella, S. A. Gherini, R. J. M. Hudson, W. E. Frick, G. L. Rupp,
and G. L. Bowie, "Water Quality Assessment: A Screening Procedure for Toxic and Conventional
Pollutants," EPA-600/6-82-004b, U.S. Environmental Protection Agency, Athens, GA, 1982.
Mortimer, C. H., "Lake Hydrodynamics," International Association of Applied Limnology
Mitteilungen, 20:124-197, 1974.
Munk, W., and E. R. Anderson, "Notes on a Theory of the Thermocline," Journal of Marine
Research, 7:276-295, 1948.
Norton, W. R., L. A. Roesner, and G. T. Orlob, Mathematical Models for Predicting Thermal
Changes in Impoundments, EPA Water Pollution Control Research Series, U.S. Environmental
Protection Agency, Washington, DC, 1968.
Rossby, G. G. and R. Montgomery, "The Layers of Frictional Influence in Wind and Ocean
Currents," Papers Phy. Oc. Meth., 111(3), 1935.
Ryan, P. J. and D. R. F. Harleman, An Analytical and Experimental Study of Transient Cooling
Pond Behavior. Technical Report No. 161, R.M. Parsons Laboratory, Massachusetts Institute of
Technology, 1973.
Safaie, B."Mixing of Buoyant Surface Jet over Sloping Bottom," ASCE, Journal Waterway,
Port, Coastal and Ocean Engineering Division, 105(WW4): 357-373, 1979.
Savage, S. B. and J. Brimberg, "Analysis of Plunging Phenomenon of Density Currents in
Reservoirs," IAHR, QQ, 13(2): 187-204, 1973.
Shanahan, P., "Water Temperature Modeling: A Practical Guide," in Proceedings of the U.S.
Environmental Protection Agency Stormwater and Water Quality Users Group Meeting, LAMR,
jounal, Hydraulic, Research, , April 1984.
Shuttleworth, W. J., "Evaporation," in D. R. Maidment, ed., Handbook of Hydrology, McGrawHill, New York, 1993.
Singh, B., and C. R. Shah, "Plunging Phenomenon of Density Currents in Reservoirs," La Houille
Blanche, 26(1): 59-64, 1971.
Smith, I. R., Turbulence in Lakes and Rivers, Scientific Publication No. 29, Freshwater Biological
Association, Ambleside, Cumbria, UK, 1975.
Smith, J. A., "Precipitation," in D. R. Maidment, ed., Handbook of Hydrology, McGraw-Hill, New
York, 1993.
Smith, D. R., S. C. Wilhelms, J. P. Holland, M. S. Dortch, and J. E. Davis, Improved Description of
Selective Withdrawal Through Point Sinks, Technical Report No, E-87-2, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS, 1985.
Stefan, H. G., R. B. Ambrose, Jr., and M. S. Dortch, "Formulation of Water Quality Models for
Streams, Lakes and Reservoirs: Modeler's Perspective," Miscellaneous Paper E-89-1, .S. Army
Engineer Waterways Experiment Station, Vicksburg, MS., 1989.
Sundaram, T. R., "A Theoretical Model for Seasonal Thermocline Cycle of Deep Temperate
Lakes," Proceedings of the 16th Conference of Great Lakes Research, 1973, pp. 1009-1025.
Sverdrup, H. U., M. W. Johnson, and R. H. Fleming, The Oceans, Prentice-Hall, Englewood Cliffs,
NJ, 1942.
Swinbank, W. C., "Longwave Radiation from Clear Skies." Quarterly Journal of the Royal
Meteorological Society, 89: 339-348, 1963.
Taylor, G. L, " The Dispersion of Matter in a Turbulent Flow Through a Pipe," Proceedings of the
Royal Society of London, Series A, 223: 446-468, 1954.
Thackston, E. L., Effect of Geographical Variation on Performance of Recirculating Cooling
Ponds, EPA-660/2-74-085, U.S. Environmental Protection Agency, Corvallis, OR, 1974.
Tennessee Valley Authority,(TVA) "Heat and Mass Transfer Between a Water Surface and the
Atmosphere." TN Report No. 14, TVA Water Resources Research Engineering Laboratory,
Norris, TN, 1972.
Turner, J. S., Buoyancy Effects in Fluids, Cambridge University Press, Cambridge, UK, 1973.
Wetzel, R. G., Limnology, Saunders College Publishing, Philadelphia, PA, 1975, and 1983.
Wunderlich, W. O. and R. A. Elder, "Mechanics of Flow Through Man-Made Lakes," in ManMade Lakes: Their Problems and Environmental Effects, W. C. Ackerman, G. F. White, and E. B.
Worthington, eds., American Geophysical Union, Washington, DC, 1973.
CHAPTER 6
SEDIMENTATION AND
EROSION HYDRAULICS
Marcelo H. Garcia
Department of Civil and Environmental Engineering
University of Illinois at Urbana-Champaign
Urbana, IL
6.1
INTRODUCTION
Since the beginning of mankind, sedimentation processes have affected water supplies,
irrigation, agricultural practices, flood control, river migration, hydroelectric projects,
navigation, fisheries, and aquatic habitat. In the last few years, sediment also has been
found to play an important role in the transport and fate of pollutants; thus, sedimentation
control has become an important issue in water quality management. Toxic chemicals can
become attached to, or adsorbed by, sediment particles and then be transported to and
deposited in other areas. By studying the quantity, quality, and characteristics of sediment
in rivers and streams, scientists and engineers can determine the sources of the sediment
and evaluate the impact of pollutants on the aquatic environment. In the United States,
sedimentation control is a multibillion-dollar issue. For example, approximately $500
million are spent every year to dredge waterways and harbors for navigation purposes.
Most of the dredged sediment is the result of substantial soil erosion in watersheds.
Estimates by the U.S. Department of Agriculture indicate that annual offside costs of sediment derived from Copland erosion are on the order of $2 billion to $6 billion, with an
additional $1 billion arising from loss in compared productivity.
The sediment cycle starts with the process of erosion, where by particles or fragments
are weathered from rock material. Action by water, wind, glaciers, and plant and animal
activities all contribute to the erosion of the earth's surface. Fluvial sediment is the term
used to describe the case where water is the key agent for erosion. Natural, or geologic,
erosion takes place slowly, over centuries or millennia. Erosion that occurs as a result of
human activity may take place much faster. It is important to understand the role of each
cause when studying sediment transport.
Any material that can be dislodged is ready to be transported. The transportation
process is initiated on the land surface when raindrops result in sheet erosion. Rills, gullies, streams, and rivers then act as conduits for the movement of sediment. The greater
the discharge, or rate of flow, the higher the capacity for sediment transport.
The final process in the cycle is deposition. When there is not enough energy to transport the sediment, it comes to rest. Sinks, or depositional areas, can be visible as newly
deposited material on a floodplain, on bars and islands in a channel, and on deltas.
Considerable deposition occurs that may not be apparent, as on lake and river beds. A
knowledge of sediment dynamics is an integral part of understanding the aquatic ecosystem.
This chapter presents fundamental aspects of the erosion, transport, and deposition of
sediment in the environment. The emphasis is on the hydraulics of bedload and suspend-
ed load transport in rivers, with the goal of establishing the background needed for sedimentation engineering. Because of their relevance, the hydraulics of both reservoir sedimentation and turbidity currents also is considered. Emphasis is placed on noncohesive
sediment transport, where the material involved can be silt, sand, or gravel. When possible, the behavior of both uniform-sized material and sediment mixtures is analyzed.
Although such topics as cohesive sediment transport, debris and mud flows, alluvial fans,
river meandering, and sediment transport by wave action are not discussed here, it is
hoped that the material covered in this chapter will provide a firm foundation to tackle
problems in those.
For more information on sediment transport and sedimentation engineering, readers
are referred to Allen (1985), Ashworth et al. (1996), Bogardi (1974), Bouvard (1992),
Carling and Dawson (1996), Chang (1988), Coussot (1997), Freds0e and Deigaard
(1992), Garde and Ranga Raju (1985), Graf (1971), Jansen et al. (1979), Julien (1992),
Mehta (1986), Mehta et al. (1989a, 1989b), Morris and Fan (1998), Nakato and Ettema
(1996), National Research Council (1996), Nielsen (1992), National Research council
(1996), Parker and Ikeda (1989), Raudkivi (1990, 1993), Renard et al. (1997), Sieben
(1997), Simons and Senturk (1992), Sloff (1997), van Rijn (1997), Yalin (1972, 1992),
Yang (1996), and Wan and Wang (1994).
6.2 HYDRAULICSFORSEDIMENTTRANSPORT
6.2.1 Flow Velocity Distribution
Consider a steady, turbulent, uniform, open-channel flow having a mean depth H and a
mean flow velocity U (Fig. 6.1). The channel is extremely wide and its bottom has a mean
slope S and a surface roughness that can be characterized by an effective height ks
(Brownlie, 198Ib). When the bottom of the channel is covered with sediment having a
mean size or diameter D, the roughness height ks will be proportional to that diameter.
Because of the weight of the water, the flow exerts on the bottom a tangential force per
unit bed area known as the bed shear stress ib, which can be expressed as:
T* = PgHS
(6.1)
where p is the water density and g is the gravitational acceleration. With the help of the
boundary shear stress, it is possible to define the shear velocity u* as
FLOW
FIGURE 6.1 Definition diagram for open-channel flow over an credible bed.
M. = VVP
(6.2)
The shear velocity, and thus the boundary shear stress, provides a direct measure of the
intensity of flow and its ability to entrain and transport sediment particles. The size of the
sediment particles on the bottom determines the surface roughness, which in turn affects
the flow velocity distribution and its sediment transport capacity. Since flow resistance
and sediment transport rates are interrelated, the ability to determine the role played by
the bottom roughness is important.
Research has shown (Schlichting, 1979) that the flow velocity distribution is well represented by:
JL = I ]nz + const.
(6.3)
where u is the time-averaged flow velocity at distance z above the bed and K is known as
von Karman's constant and is equal to 0.4. For obvious reasons, the above law is known
as the logarithmic law of the wall. It strictly applies only in a thin layer near the bed. It is
empirically found to apply as a reasonable approximation throughout most of the flow in
many rivers.
If the bottom boundary is sufficiently smooth (a condition rarely satisfied in rivers),
turbulence will be drastically suppressed in an extremely thin layer near the bed. In this
region, a linear velocity profile will hold:
^- = ^
(6.4)
W*
V
where v is the kinematic viscosity of water. This law merges with the logarithmic law near
z 5V, where
5 V =11.6-^-
(6.5)
denotes the height of the viscous sublayer. In the logarithmic region, the constant of integration introduced above has been evaluated from data to yield
J. = Ih IWl+5.5
U* K ( V )
(6.6)
Most boundaries in river flow are rough. Let ks denote an effective roughness height.
If kjbv > 1, then no viscous sublayer will exist. The corresponding logarithmic velocity
profile is given by
t=Mt) + 8 - 5 = H 3 t)
(6 7)
As noted above, this relation often holds as a first approximation throughout the flow in
a river. It is by no means exact.
The conditions kjbv 1 for rough turbulent flow and kjbv 1 for smooth turbulent
flow can be rewritten to indicate that u*kjv should be much larger than 11.6 for turbulent
rough flow and much smaller than 11.6 for turbulent smooth flow. A composite form that
represents both ranges, as well as the transitional range between them, can be written as
i = Hf) +B *
with Bs as a function of Re* = u*kjv, which can be estimated with
(6 8)
'
(6.9)
"
(6.12)
This relation is known as Keulegan 's resistance relation for rough flow.
An approximation to Keulegan's relation is the Manning-Strickler power form
^= 8
.
IAJ
(6-13)
Between Eqs. (6.2) and (6.12), a resistance relation can be found for bed shear stress:
T6 = pC/t/*
(6.14)
(6
-i5)
IfEq. (6.13) is used instead of Eq. (6.12), the friction coefficient takes the form
c
,=Kf)T
It is useful to show the relationship between the friction coefficient Cf and the roughness parameters in open-channel flow relations commonly used in practice. Between Eqs.
(6.1) and (6.14), the following form of Chezy's law can be derived:
U=CJI112S112
where the Chezy coefficient Cc is given by the relation
(6.17)
[ fSvS \I2
(6.18)
A specific evaluation of Chezy's coefficient can be obtained by substituting Eq. (6.15) into
Eq. (6.18). It is seen that the coefficient is not constant but varies as the logarithm ofH/ks.
A logarithmic dependence is typically a weak one, partially justifying the common
assumption that Chezy's coefficient in Eq. (6.17) is a constant. Substituting Eq. (6.16) into
Eqs. (6.17) and (6.18), Manning's law is obtained:
U = !#2/351/2
n
(6.19)
^ '
1/6
=8^
(6-2)
(6.21)
Suggested values of OC5, which have appeared in the literature, are listed in Table 6.1 (Yen,
1992). Different sizes of sediment have been suggested for Dx in Eq. (6.21). Statistically, D50
(the grain size for which 50% of the bed material is finer) is most readily available and
meaningful. Physically, a representative size larger than D50 is more meaningful to estimate
TABLE 6.1 Ratio of Nikuradse Equivalent Roughness Size and Sediment Size for Rivers.
Investigator
Ackers and White (1973)
Strickler (1923)
Keulegan (1938)
Meyer-Peter and Muller (1948)
Thompson and Campbell (1979)
Hammond et al. (1984)
Einstein and Barbarossa (1952)
Irmay (1949)
Engelund and Hansen (1967)
Lane and Carlson (1953)
(X^ = ks/Dx
1.23
3.3
1
1
2.0
6.6
1
1.5
2.0
3.2
Gladki (1979)
Leopold et al. (1964)
Limerinos (1970)
Mahmood(1971)
Hey (1979), Bray (1979)
Ikeda(1983)
Colosimo et al. (1986)
Whiting and Dietrich (1990)
Simons and Richardson (1966)
Kamphuis (1974)
van Rijn (1982)
D80
D84
D84
D84
D84
D84
D84
D84
D85
D90
D90
ens = ks/Dx
2.5
3.9
2.8
5.1
3.5
1.5
3~6
2.95
1
2.0
3.0
6.3
SEDIMENTPROPERTIES
important properties of groups of particles include porosity and size distribution. The most
common rock type one is likely to encounter in the river or coastal environment is quartz.
Quartz is a highly resistant rock and can travel long distances or remain in place for long
periods without losing its integrity. Another highly resistant rock type that is often found
together with quartz is feldspar. Other common rock types include limestone, basalt, granite, and more esoteric types, such as magnetite. Limestone is not a resistant rock; it tends
to abrade to silt rather easily. Silt-sized limestone particles are susceptible to solution
unless the water is buffered sufficiently. As a result, limestone typically is not a major
component of sediments at locations distant from its source. On the other hand, it often
can be the dominant rock type in mountain environments.
Basaltic rocks tend to be heavier than most rocks composing the earth's crust and typically are brought to the surface by volcanic activity. Basaltic gravels are relatively common in rivers that derive their sediment supply from areas subjected to vulcanism in
recent geologic history. Basaltic sands are much less common. Regions of weathered
granite often provide copious supplies of sediment. Although the particles produced by
weathering are often in the granule size, they often break down quickly to sand size.
Sediments in the fluvial or coastal environment in the size range of silt, or coarser, are
generally produced by mechanical means, including fracture or abrasion. The clay minerals, on the other hand, are produced by chemical action. As a result, they are fundamentally different from other sediments in many ways. Their ability to absorb water means
that the porosity of clay deposits can vary greatly over time. Clays also display cohesivity, which renders them more resistant to erosion.
6.3.2 Specific Gravity
The specific gravity of sediment is defined as the ratio between the sediment density ps
and the density of water p. Some typical specific gravities for various natural and artificial sediments are listed in Table 6.2.
6.3.3
Size
Herein, the notation D is used to denote sediment size, the typical units of which are
millimeters (mm) for sand and coarser material or microns (JLI) for clay and silt.
Another standard way of classifying grain sizes is the sedimentological <3> scale,
according to which
TABLE 6.2 Specific Gravity of Rock Types
and Artificial Material
Rock type or
materia/
quartz
limestone
basalt
magnetite
plastic
coal
walnut shells
Specific gravity
p/p
2.60 ~ 2.70
2.60 ~ 2.80
2.70 ~ 2.90
3.20 ~ 3.50
1.00 ~ 1.50
1.30-1.50
1.30- 1.40
D = 2*
(6.23)
(6.24)
Note that the size O = O corresponds to D = 1 mm. The usefulness of the O scale will
become apparent upon a consideration of grain size distributions. The minus sign has been
inserted in Eq. (6.24) simply as a matter of convenience to sedimentologists, who are
more accustomed to working with material finer than 1 mm than they are with coarser
material. The reader should always recall that larger O implies finer material. The O scale
provides a simple way of classifying grain sizes into the following size ranges in descending order: boulders, cobbles, gravel, sand, silt, and clay. (Table 6.3).
Note that the definition of clay according to size (D < 2JJL) does not always correspond
to the definition of clay according to mineral. That is, some clay-mineral particles can be
coarser than this limit, and some silt-sized particles produced by grinding can be finer than
that. In general, however, the effect of viscosity makes it difficult to grind up particles in
water to sizes finer than 2 JJL.
In practical terms, there are several ways to determine grain size. The most popular
way for grains ranging from O= -4 to O = -4 (0.0625 to 16 mm) is with the use of
sieves. Each sieve has a square mesh, the gap size of which corresponds to the diameter
of the largest sphere that would fit through it. Thus, the grain size D so measured corresponds exactly to the diameter only in the case of a sphere. In general, the sieve size D
corresponds to the smallest sieve gap size through which a given grain can be fitted.
For coarser grain sizes, it is customary to approximate the grain as an ellipsoid. Three
lengths can be defined. The length along the major (longest) axis is denoted as a, the
length along the intermediate axis is denoted as b, and the length along the minor (smallest) axis is denoted as c. These lengths are typically measured with a caliper. The value b
is then equated to grain size D.
For grains in the silt and clay sizes, many methods (hydrometer, sedigraph, and so
forth) are based on the concept of equivalent fall diameter. That is, the terminal fall velocity vs of a grain in water at a standard temperature is measured. The equivalent fall diameter D is the diameter of the sphere having exactly the same fall velocity under the same
conditions. Sediment fall velocity is discussed in more detail below.
A variety of other more recent methods for sizing fine particles rely on blockage of
light beams. The blocked area can be used to determine the diameter of the equivalent circle: i.e., the projection of the equivalent sphere. It can be seen that all the above methods
can be expected to operate consistently as long as grains shape does not deviate too greatly from a sphere. In general, this turns out to be the case. There are some important exceptions, however. At the fine end of the spectrum, mica particles tend to be platelike; the
same is true of shale grains at the coarser end. Comparison with a sphere is not necessarily an especially useful way to characterize grain size for such materials.
6.3.4 Size Distribution
Any sample of sediment normally contains a range of sizes. An appropriate way to characterize these samples is by grain size distribution. Consider a large bulk sample of sediment of given weight. Let Pj(D)or p/O)denote the fraction by weight of material in
the sample of material finer than size >(O). The customary engineering representation of
Size Range
Millimeters
-9- -8
-8
7
-7
6
-6 --5
-5- -4
-4- -3
-3- -2
-2- -1
-1-0
0-1
1 -2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
Microns
Inches
160 - 80
80-40
40-20
20- 10
10-5
5-2.5
2.5 - 1.3
1.3 - 0.6
0.6 - 0.3
0.3 - 0.16
0.16 - 0.08
2,000 - 1,000
1,000 - 500
500 - 250
250 - 125
125 - 62
62-31
31 - 16
16-8
8-4
4-2
2- 1
1 -0.5
0.5 - 0.24
2- 1/2
5
9
16
32
60
115
250
5
10
18
ac
35
60
120
230
C0
n>Qg
|.
I
2.
3
OI
2
S
the grain size distribution consists of a plot of pf *100 (percentage finer) versus log10(D):
that is, a semilogarithmic plot is used. The same size distribution plotted in sedimentological form would involve plotting /y 100 versus O on a linear plot.
The size distribution p/O) and size density /?(O) by weight can be used to extract useful statistics concerning the sediment in question. Let x denote some percentage, say 50%;
the grain size Ox denotes the size such that x percent of the weight of the sample is composed of finer grains. That is, O^ is defined such that
(6 25)
PW = jfe
It follows that the corresponding grain size of equivalent diameter is given by Dx, where
Dx = 2 -**
(6.26)
The most commonly used grain sizes of this type are the median size D50 and the size
Z)90: i.e., 90% of the sample by weight consists of finer grains. The latter size is especially useful for characterizing bed roughness.
The density p(O) can be used to extract statistical moments. Of these, the most useful
are the mean size Om and the standard deviation a. These are given by the relations.
O1n = /Op(O)JO;
a2 = (O - OJ2p(O)DO
(6.27a, b)
(5g = 2
(6,28a,b)
Note that for a perfectly uniform material, a = O and ag = 1. As a practical matter, a sediment mixture with a value of Gg less than 1.3 is often termed well sorted and can be treated as a uniform material. When the geometric standard deviation exceeds 1.6, the material can be said to be poorly sorted (Diplas and Sutherland, 1988).
In fact, one never has the continuous function p(O) with which to compute the
moments of Eqs. (6.27a, and b). Instead, one must rely on a discretization. To this end, the
size range covered by a given sample of sediment is discretized using n intervals bounded by n + 1 grain sizes O1, O2,..., On + l in ascending order of O. The following definitions are made from / = 1 to n:
> = ^t + /+i)
(6-29a)
P1=Pj(Q)-Pf(^+1)
(6.29b)
(6-31)
Furthermore, it can be demonstrated from a standard table of the Gauss distribution that
the size O displaced one standard deviation larger that Om is accurately given by O84; by
symmetry, the corresponding size that is one standard deviation smaller than Om is O16.
The following relations thus hold:
a = ^(O84 - O16)
(6.32a)
m = |(<*>84 + *16>
(6-32b)
Rearranging the above relations with the aid of Eqs. (6.28a and b) and Eqs. (6.31 and
6.32a),
.=f
Ds = (D84T)16)"*
(6.33b)
It must be emphasized that the above relations are exact only for a gaussian distribution
in O. This is not often the case in nature. As a result, it is strongly recommended that Dg
and Gg be computed from the full size distribution via Eqs. (6.3Oa and b) and (6.28a and
b) rather than the approximate form embodied in the above relations.
6.3.5 Porosity
The porosity Xp quantifies the fraction of a given volume of sediment that is composed of
void space. That is,
, _
volume of voids
P volume of total space
If a given mass of sediment of known density is deposited, the volume of the deposit
must be computed, assuming that at least part of it will consist of voids. In the case of
well-sorted sand, the porosity often can take values between 0.3 and 0.4. Gravels tend to
be more poorly sorted. In this case, finer particles can occupy the spaces between coarser
particles, thus reducing the void ratio to as low as 0.2. Because so-called open-work gravels are essentially devoid of sand and finer material in their interstices, they may have
porosities similar to sand. Freshly deposited clays are notorious for having high porosities. As time passes, the clay deposit tends to consolidate under its own weight so that
porosity slowly decreases.
The issue of porosity becomes of practical importance with regard to salmon spawning grounds in gravel-bed rivers, for example (Diplas and Parker, 1985). The percentage
of sand and silt contained in the sediment is often referred to as the percentage of fines in
the gravel deposit. When this fraction rises above 20 or 26 percent by weight, the deposit
is often rendered unsuitable for spawning. Salmon bury their eggs within the gravel, and
a high fines content implies a low porosity and thus reduced permeability. The flow of
groundwater necessary to carry oxygen to the eggs and remove metabolic waste products
is impeded. In addition, newly hatched fry may encounter difficulty in finding enough
pore space through which to emerge to the surface. All the above factors dictate lowered
survival rates. Chief causes of elevated fines in gravel rivers include road building and
clear-cutting of timber in the basin.
6.3.6
Shape
Grain shape can be classified in a number of ways. One of these, the Zingg classification
scheme, is illustrated here (Vanoni, 1975). According to the definitions introduced earlier,
a simple way to characterize the shape of an irregular clast (stone) is by lengths a, b, and
c of the major, intermediate, and minor axes, respectively. If the three lengths are equal,
the grain can be said to be close to a sphere in shape. If a and b are equal but c is much
larger, the grain should be rodlike. Finally, if c is much smaller than b, which in turn, is
much larger than a, the resulting shape should be bladelike.
6.3.7
Fall Velocity
A fundamental property of sediment particles is their fall velocity. The relation for terminal fall velocity in quiescent fluid vs can be presented as
*-[hy
where
Rff = ^VR^D
(6.35a)
RP=V-^-
(6.35b)
and the functional relation C0 = CD(Rp) denotes the drag curve for spheres. This relation
is not particularly useful because it is not explicit in V5; one must compute fall velocity by
trial and error. One can use the equation for C0 given below
C0 = ^ (1 + 0.152/?/'2 + 0.015 lRp)
P
(6.36)
(6.37)
to obtain an explicit relation for fall velocity in the form of Rf versus Rep. In Fig. 6.2, the
ranges for silt, sand, and gravel are plotted for v = 0.01 cm2/s (clear water at 2O0C) and
R = 1.65 (quartz). A good summary of relations for terminal fall velocity for the case of
nonspherical (natural) particles can be found in Dietrich (1982), who also proposed the
following useful fit:
Rf = expi-b, + b2\n(Rep) - b,[\n(RepW ~ b4[\n(RepW + b5\\n(Re^Y}
(6.38)
In Fig. 6.3, several size distributions from the sand-bed Kankakee River in Illinois, are
shown (Bhowmik et al., 1980). The characteristic S shape suggests that these distributions
might be approximated by a gaussian curve. The median size D50 falls near 0.3 to 0.4 mm.
The distributions are tight, with a near absence of either gravel or silt. For practical purposes, the material can be approximated as uniform.
In Fig. 6.4, several size distributions pertaining to the gravel-bed Oak Creek in
Oregon, are shown (Milhous, 1973). In gravel-bed streams, the surface layer ("armor" or
"pavement") tends to be coarser than the substrate (identified as "subpavement" in the figure). Whether the surface or substrate is considered, it is apparent that the distribution
ranges over a much wider range of grain sizes than is the case in Fig. 6.3. More specifically, in the distributions of the sand-bed Kankakee River, O varies from about O to about
I linois - Indiana
State Line
3 31/79 - 4/21/79
GRAIN SIZE, mm
Illinois - Indiana
State Line
4/29/79 - 9/14/79
Pavement
Sub pavement
Composite of Pavement
and Subpavement
3, whereas in Oak Creek, O varies from about -8 to about 3. In addition, the distribution
of Fig. 6.4 is upward-concave almost everywhere and thus deviates strongly from the
gaussian distribution.
These two examples provide a window toward generalization. A river can be loosely
classified as sand-bed or gravel-bed according to whether the median size D50 of the surface material or substrate is less than or greater than 2 mm. The size distributions of sandbed streams tend to be relatively narrow and also tend to be S shaped. The size distributions of gravel-bed streams tend to be much broader and to display an upward-concave
shape. Of course, there are many exceptions to this behavior, but it is sufficiently general
to warrant emphasis.
More evidence for this behavior is provided in Fig. 6.5. Here, the grain size distributions for a variety of stream reaches have been normalized using the median size D50.
Four sand-bed reaches are included with three gravel-bed reaches. All the sand-bed
distributions are S shaped, and all have a lower spread than the gravel-bed distributions.
The standard deviation is seen to increase systematically with increasing D50(White et al.,
1973).
The three gravel-bed size distributions differ systematically from the sand-bed distributions in a fashion that accurately reflects Oak Creek (Fig. 6.4). The standard deviation
in all cases is markedly larger than any of the sand-bed distributions, and the distributions
are upward-concave except perhaps near the coarsest sizes.
River
Miss, Tarbeii La.
Niobran
Miss, at Sl Louis
Mountain free*
Elbow
Aart
(0,/O3^
FIGURE 6.5 Dimensionless grain-size distribution for different rivers (White et al., 1973)
6A
When a granular bed is subjected to a turbulent flow, virtually no motion of the grains is
observed at some flows, but the bed is mobilized noticeably at other flows. Factors that
affect the mobility of grains subjected to a flow are summarized below:
I grain placement
randomness ^
[turbulence
f fluid 1/lift mean & turbulent
forces on grain <
I drag
I gravity
In the presence of turbulent flow, random fluctuations typically prevent the clear definition of a critical, or threshold condition for motion: The probability for the movement of
a grain is never precisely zero (Lavelle and Mofjeld, 1987). Nevertheless, it is possible to
define a condition below which movement can be neglected for many practical purposes.
6.4.1 Granular Sediment on a Stream Bed
Figure 6.6 is a diagram showing the forces acting on a grain in a bed of other grains. When
critical conditions exist and the grain is on the verge of moving, the moment caused by
the critical shear stress Tc about the point of support is just equal to that of the weight of
the grain. Equating these moments gives (Vanoni, 1975):
Tc = (Js ~ J) Dcos (tan 9 - tanty)
C2a2
(6.39)
in which ys = specific weight of sediment grains, y = specific weight of water, D = diameter of grains, <(> is the slope angle of the stream, 6 = the angle of repose of the sediment, C1
and C2 are dimensionless constants, and av and a2 are lengths shown in Fig. 6.6. Any consistent set of units can be used in Eq. (6.39). For a horizontal bed, Eq. (6.39) reduces to
Tc = (Y, - J)D tan 6
C2#2
(6.40)
-^5-4^)
The Shields values of Tc* are commonly used to denote conditions under which bed
sediments are stable but on the verge of being entrained. Not all workers agree with the
results given by the Shields curve. For example, some workers give Tc* = 0.047 for the
dimensionless critical shear stress for values of R* = u*D/v in excess of 500 instead of
0.06, as shown in Fig. 6.7. Taylor and Vanoni (1972) reported that small but finite amounts
of sediment were transported in flows with values of i* given by the Shields curve.
The value of Tc to be used in design depends on the particular case at hand. If the situation is such that grains that are moved can be replaced by others moving from upstream,
some motion can be tolerated, and the Shields values can be used. On the other hand, if
grains removed cannot be replaced, as on a stream bank, the Shields value of ic are too
large and should be reduced.
The Shields diagram is not especially useful in the form of Fig. 6.7 because to find Tc,
one must know w* = VT/P. The relation can be cast in explicit form by plotting TC* versus Rep, noting the internal relation
_u*D
1_ =
v
u _ VfltfD
U
s D
L
v
VRgD
= (T*)i/2/fc
(6.43)
p p
where R =
is the submerged specific gravity of the sediment. A useful fit is given
byBrownlie(1981a):
T* = 0.22Re- + o.06 exp(- 17.77/te;006)
(6.44)
With this relation, the value of Tc* can be computed readily when the properties of the
water and the sediment are given.
The value of bed-shear stress ib for a wide rectangular channel is given by ib = jHS,
as shown earlier. The average bed-shear stress for any channel is given by ib = JR^S, in
which Rh = the hydraulic radius of the channel cross section.
Amber
** (Shields)
Granite
Barite
Sand (Casey)
Sand (Kramer)
Sand (U.S. WES.)
Sand (Gilbert)
Sand (White)
Sand in air (White)
Steel shot (White)
<6-45>
where (J)1 is the slope of the bank and 0 is the angle of repose for the sediment. Values of 0
are given in Fig. 6.8 after Lane (1955) and also can be found in Simons and Senturk (1976).
6.4.3 Granular Sediment on a Sloping Bed
Equation (6.39) shows that ic diminishes as the slope angle (|> increases. For extremely
small <|)'s, TC is given by Eq. (6.40). Taking the ratio between Eqs. (6.39) and (6.40) yields
<-=$)
T^ is the critical shear stress for sediment on a bed with a slope angle <|), and ico is the critical shear stress for a bed with an extremely small slope. The value of ico can be found
from the Shields diagram or with Eq. (6.44). Equation (6.46) is for positive <|), which is
positive for downward sloping beds. For beds with adverse slope, ty is negative and the
term tan <|)/tan 0 in Eq. (6.46) is positive.
6.4.4 Sediment Mixtures
Several authors have offered empirical or quasi-theoretical extensions of the above relations to the case of mixtures (e.g., Wilcock, 1988). Let D1 denote the characteristic grain
size of the ith size range in a mixture. Furthermore, let Dsg denote the geometric mean size
of the surface (exchange, active) layer. Most of the generalizations can be written in the
following form (Parker, 1990):
^ =^
(6.47)
Here
'--pfe
(6 48a)
*"pfe
(^b)
and
where ifed and ibcsg denote the values of the dimensioned critical shear stress required to
move sediment of sizes D1 and Dsg in the mixture, respectively, and P is an exponent taking a value given below;
P = 0.9
(6.49)
Figure 6.9 shows the similarity between four different published expressions having
the general form given by Eq. (6.47), which is of interest because it includes the effect of
hiding. For uniform material, the critical Shields stress is defined by Eq. (6.44). Consider
two flumes, one with uniform size D0 and the other with uniform size Db. For sufficiently coarse material (*Z)/v 1 or Rep 1), the critical Shields stress must be the same for
both sizes (Fig. 6.7). It follows from Eq. (6.42) that where ibca and ibcb denote the dimensioned boundary shear stresses for cases a and b respectively,
(D,\
*bcb = V^j
(6.50)
For the case of mixtures, on the other hand, it is seen from Eqs. (6.47) and (6.48) that
a
0 1
(6 51)
^1 = *Jff\
\ SgJ "" ^fzr]
\ SgJ '
Comparing Eqs. (6.50) and (6.51), it is seen that a finer particle (Db < Da, or alternatively, D1 < Dsg) is more mobile than a coarser particle. For example, suppose that one grain
size is four times coarser than another. If two uniform sediments are being compared, it
EgiazaroffCl965)
Hayashietal
Andrews
FIGURE 6.9 Critical shear stress for sediment mixture (Source: Misri et
al, 1983)
follows from Eq. (6.50) that the critical shear stress for the coarser material is four times
that of the finer material. In the case of a mixture, however, the critical shear stress for the
coarser material is only about 4-1, or 1.15 times that for the finer material.
A finer particle in a mixture is thus seen to be only a little more mobile than its coarser-sized brethren, where uniform beds of fine material are much more mobile than are uniform beds of coarser material. The reason is that finer particles in a mixture are relatively
less exposed to the flow; they tend to hide in the lee of coarser particles. By the same token,
a particle is relatively more exposed to the flow when most of its neighbors are finer.
A method to calculate the critical shear stress for motion of uniform and heterogeneous
sediments was proposed by Wiberg and Smith (1987) on the basis of the fluid mechanics
of initiation of motion, which takes into account both roughness and hiding effects.
6.5 SEDIMENTTRANSPORT
6.5.1 Sediment Transport Modes
The most common modes of sediment transport in rivers are bedload and suspended load.
In the case of bedload, the particles roll, slide, or saltate over each other, never deviating
too far above the bed. In the case of suspended load, the fluid turbulence comes into play
carrying the particles well up into the water column. In both cases, the driving force for
sediment transport is the action of gravity on the fluid phase; this force is transmitted to
the particles via drag.
The same phenomena of bedload and suspended load transport occur in a variety of
other geophysical contexts. Sediment transport is accomplished in the near-shore lake and
oceanic environment by wave action. Turbidity currents carry sediment into lakes, reservoirs, and the deep sea.
The phenomenon of sediment transport can sometimes be disguised in rather esoteric
phenomena. When water is supercooled, large quantities of particulate frazil ice can form.
As this water moves under a frozen ice cover, the phenomenon of sediment transport in
rivers is stood on its head. The frazil ice particles float rather than sink and thus tend to
accumulate on the bottom side of the ice cover rather than on the river bed. Turbulence
tends to suspend the particles downward rather than upward.
In the case of a powder snow avalanche, the fluid phase is air and the solid phase consists of snow particles. The dominant mode of transport is suspension. These flows are
close analogies of turbidity currents, insofar as the driving force for the flow is the action
of gravity on the solid phase rather than the fluid phase. That is, if all the particles drop
out of suspension, the flow ceases. In the case of sediment transport in rivers, it is accurate to say that the fluid phase drags the solid phase along. In the case of turbidity currents
and powder snow avalanches, the solid phase drags the fluid phase along.
Desert sand dunes provide an example for which the fluid phase is air, but the dominant mode of transport is saltation rather than suspension. Because air is so much lighter
than water, quartz sand particles saltate in long, high trajectories, relatively unaffected by
the direct action of turbulent fluctuations. The dunes themselves are created by the effect
of the fluid phase acting on the solid phase. They, in turn, affect the fluid phase by changing the resistance.
Among the most interesting sediment-transport phenomena are debris flows, slurries,
and hyperconcentrated flows. In all these cases, the solid and fluid phases are present in
similar quantities. A debris flow typically carries a heterogeneous mixture of grain sizes
ranging from boulders to clay. Slurries and hyperconcentrated flows are generally restricted to finer grain sizes. In most cases, it is useful to think of such flows as consisting of a
single phase, the mechanics of which are highly non-Newtonian.
The study of the movement of grains under the influence of fluid drag and gravity
becomes even more interesting when one considers the link between sediment transport
and morphology. In the laboratory, the phenomenon can be studied in the context of a variety of containers, such as channel and wave tanks, specified by the experimentalist. In the
field, however, the fluid-sediment mixture constructs its own container. This new degree
of freedom opens up a variety of intriguing possibilities.
Consider the river. Depending on the existence or lack of a viscous sublayer and the
relative importance of bedload versus suspended load, a variety of rhythmic structures can
form on the river bed. These include ripples, dunes, antidunes, and alternate bars. The first
three of these can have a profound effect on the resistance to flow offered by the river bed.
Thus, they act to control river depth. River banks themselves also can be considered to be
a self-formed morphological feature, thus specifying the entire container.
The container itself can deform in plan. Alternate bars cause rivers to erode their banks
in a rhythmic pattern, thus allowing for the onset of meandering. Fully developed river
meandering implies an intricate balance between sediment erosion and deposition. If a
stream is sufficiently wide, it will braid rather than meander, dividing into several intertwining channels.
Rivers create morphological structures at much larger scales as well. These include
canyons, alluvial fans, and deltas. Turbidity currents create similar structures in the oceanic environment. In the coastal environment, the beach profile itself is created by the interaction of water and sediment. On a larger scale, offshore bars, spits, and capes constitute
rhythmic features created by wave-current-sediment interaction. The boulder levees often
created by debris flows provide another example of a morphologic structure created by a
sediment-bearing flow.
The floodplains of most sand-bed rivers often contain copious amounts of silt and clay
finer than approximately 50 |JL This material is often called wash load because it moves
through the river system without being present in the bed in significant quantities.
Increased wash load does not cause deposition on the bed, and reduced wash load does
not cause erosion because it is transported well below capacity. This is not meant to imply
that the wash load does not interact with the river system. Wash load in the water column
exchanges with the banks and the floodplain rather than the bed. Greatly increased wash
load, for example, can lead to thickened floodplain deposits, with a consequent increase
in bankfull channel depth.
The emphasis here is the understanding of bedload and suspended load transport in
rivers, with the goal of providing the knowledge needed to do sound sedimentation engineering, particularly with problems involving stream restoration and naturalization.
6.5.2
In the context of rivers, it is useful to have a way to determine what kind of sedimenttransport phenomena can be expected for different flow conditions and different characteristics of sediment particles. In Fig. 6.10, the ordinates correspond to bed shear stresses
written in the dimensionless form proposed by Shields
T*
T =
Tfc
pgRD
= HS
RD
(f) 52)
TO
(6
'
and the particle Rep, defined by Eq. (6.37) is used for the abscissa values. There are three
curves in the diagram which make it possible to know, for different values of (T*, Rep), if
the given bed sediment will go into motion, and if this is the case whether or not the prevailing mode of transport will be in suspension or as bedload. The diagram also can be used
to predict what kind of bedforms can be expected. For example, ripples will develop in the
presence of a viscous sublayer and fine-grained sediment. If the viscous sublayer is disrupted by coarse sediment particles, then dunes will be the most common type of bedform.
The Shields regime diagram also shows a clear distinction between the conditions
observed in sand-bed rivers and gravel-bed rivers at bankfull stage. If one wanted to model in the laboratory sediment transport in rivers, the experimental conditions would be different, depending on the river system in question. As could be expected, the diagram also
shows that in gravel-bed rivers, sediment is transported as bedload. In sand-bed rivers, on
the other hand, suspended load and bedload transport coexist most of the time.
The regime diagram is valid for steady, uniform, turbulent flow conditions, where the
bed shear stress lb can be estimated with Eq. (6.1). The ranges for silt, sand, and gravel
also are included. In the diagram, the critical Shields stress for motion was plotted with
the help of Eq. (6.44). The critical condition for suspension is given by the following ratio:
^- = 1
(6.53)
where u* is the shear velocity and vs is the sediment fall velocity. Equation (6.53) can be
transformed into
*; = ^
(6.54)
^ OD
^)
where
Suspension
No Suspension
Field
Sand-Bed
Rivers
Dunes ?
Flat?
Field
Gravel-Bed
Rivers
Motion
No Motion
Gravel
and Rf is given by Eq. (6.35a) and can be computed for different values of Rep with the
help of Eq. (6.38).
Finally, the critical condition for viscous effects (ripples) was obtained with the help
of Eq. (6.5) as follows:
11.6-^-= 1
uD
which in dimensionless form can be written as
(6.56)
*-
Relations (6.44), (6.54), and (6.57) are the ones plotted in Fig. 6.10. The Shields regime
diagram should be useful for studies concerning stream restoration and naturalization
because it provides the range of dimensionless shear stresses corresponding to bankfull
flow conditions for both gravel- and sand-bed streams.
6.6
BEDLOAD TRANSPORT
(6.58)
In general, bedload transport is vectorial, with components qs and qn in the s and n directions, respectively.
6.6.2 Erosion Into and Deposition from Suspension
The volume rate of erosion of bed material into suspension per unit time per unit bed area
is denoted as E. The units of E are lengthVlengthVtime, or velocity. A dimensionless sediment entrainment rate E5 can thus be defined with the sediment fall velocity vs:
E = vfs
(6.59)
In general, E5 can be expected to be a function of boundary shear stress ib and other parameters. Erosion into suspension can be taken to be directed upward normal: i.e., in the
positive z direction.
Let c denote the volume concentration of suspended sediment (m3 of sediment/m3 of
sediment-water mixture), averaged over turbulence. The streamwise volume transport rate
of suspended sediment per unit width is given by
H
qs = J c udz
(6.60)
o
H
<lsn = I cvdz
O
(6.6Ia)
(6.6Ib)
Deposition onto the bed is by means of settling. The rate at which material is fluxed
vertically downward onto the bed (volume/area/time) is given by vscb, where ~cb is a nearbed value of c. The deposition rate D realized at the bed is obtained by computing the
component of this flux that is actually directed normal to the bed:
D = vscb
(6.62)
(6.63)
To solve the Exner equation, it is necessary to have relations to compute bedload transport (i.e., qs and gn), near-bed suspended sediment concentration cb, and sediment entrainment into suspension E5. The basic form of Eq. (6.63) was first proposed by Exner (1925).
6.6.4 Bedload Transport Relations
A large number of bedload relations can be expressed in the general form
q* = <7*(T*, Rep> R)
(6.64)
Here, q* is a dimensionless bedload transport rate known as the Einstein number, first
introduced by H. A. Einstein in 1950 and given by
<f =
V# g D D
The following relations are of interest. In 1972, Ashida and Michiue introduced
<f = 17(T* - ty [(l*)m - (Tyi/2]
(6.65)
(6>66)
and recommend a value of Tc* of 0.05. It has been verified with uniform material ranging
in size from 0.3 mm to 7 mm. Meyer-Peter and Muller (1948) introduced the following:
q* = 8(T* - Tc*)3/2
(6.67)
where T* = 0.047. This formula is empirical in nature and has been verified with data for
uniform gravel.
Engelund and Freds0e (1976) proposed,
q* = 18.74(T* - Tc*) [(f) 1/2 - 0.7(Tp1/2]
(6.68)
where T* = 0.05. This formula resembles that of Ashida and Michiue because the derivation is almost identical.
Fernandez Luque and van Beek (1976) developed the following,
q* = 5.7(T* - TCT'2
(6.69)
where T* varies from 0.05 for 0.9 mm material to 0.058 for 3.3. mm material. The relation
is empirical in nature.
Wilson (1966):
<?* - 12(T* - Tc*)^
(6.70)
where T* was determined from the Shields diagram. This relation is empirical in nature;
most of the data used to fit it pertain to very high rates of bedload transport.
Einstein (1950):
q" = 4*(T*)
(6.71)
1-rV j
*-**=
,
l
f
f
i
s
.
(6-72>
l
V* -,413/,*)-2
+ 43'5<7
Note that this relation contains no critical stress. It has been used for uniform sand and
gravel.
Yalin (1963):
+
* = 0.635,(T*)4l-^
I
a2s -^lJ
(6.73)
where
a2 = 2A5(R + I)0-4 (Tc*)1/2;
s =T ~T*
^C
(6.74)
and T* is evaluated from a standard Shields curve. Two constants in this formula have been
evaluated with the aid of data quoted by Einstein (1950), pertaining to 0.8 mm and 28.6
mm material.
Parker (1978):
(T* - 0.03)45
<7*=11.2 IT ^3 )
(6.75)
developed with data sets pertaining to rough mobile-bed flow over gravel.
Several of these relations are plotted in Fig. 6.11. They tend to be rather similar in
nature. Scores of similar relations could be quoted.
To date, only few research groups have attempted complete derivations of the bedload
function in water. They are Wiberg and Smith (1989), Sekine and Kikkawa (1992), Garcia
and Nino (1992), Nino and Garcia, (1994, 1998), and Nino et al., (1994).
6.6.5
Relatively few bedload relations have been developed specifically in the context of mixtures (e.g., Bridge and Bennett, 1992). One of these is presented below as an example.
The relationship of Parker (1990) applies to gravel-bed streams. The data used to fit
the relation are solely from two natural gravel-bed streams: Oak Creek in Oregon and the
Elbow River in Alberta, Canada. The relation is surface-based; load is specified per unit
of fractional content in the surface layer. The surface layer is divided into N size ranges,
D = 0.5mm
^ D = 28.6mrn
EfNSTEfN (O=O 5mm)WIOERG & SMITH (D-O 5mm)
MEYER-PEIEH A MULLER
*
M
///S
'// /
FERNANDEZ LUOUE
FERNANDEZ LUOUE
EINSTEIN
WIBERG & SMITH
each with a fractional content F1 by volume, and_a mean phi size (J)1; Di = 2~to- The arithmetic mean of the surface size on the phi scale and the corresponding arithmetic standard deviation C^ are given by
<j> = IFfo
Oj = Ufa - ^)2
(6.76a, b)
The corresponding geometric mean size Dsg and the geometric standard deviation Gsg of
the surface layer are given by
A* = 2-* ov = 2*
(6.77a,b)
In the Parker relation, the volume bedload transport per unit width of gravel in the ith
size range is given by the product qf. (no summation), where q. denotes the transport per
unit fraction in the surface layer. The total volume bedload transport rate of gravel per unit
width is qT, where
qT = ^q1F1
(6.78)
The relation does not apply to sand. Thus, before using the relation for a given surface
distribution, the sand content of the grain-size distribution must be removed and F1 must
be renormalized so that it sums to unity over all sizes in excess of 2 mm.
If p. denotes the fraction volume content of material in the ith size range in the bedload, it follows that
P, = ^
(6-79)
<6-80>
A dimensionless Shields stress based on the surface geometric mean size is defined as
follows:
*-&-,
<6TO
where
^ = *
rsgo
T;SO = 0.0386
(6-82)
(6.83)
corresponds to a "near-critical" value of Shields stress. The Parker relation can then be
expressed in the form
Wl1 = 0.0218 G [CD(^(S1)]
(6.84a)
5. = PL
(6.84b)
(6.85)
and is shown in Fig. 6.13. Here, M0 = 14.2 and $ is a dummy variable for the argument
in Eq. (6.84) and is not to be confused with the grain-size scale.
An application of Eq. (6.84) to uniform material with size D results in the relation
* = -0218(TH<*y
(6 86)
'
where
?= VgRDD
^ =-^n
P8RD
(6 87)
and q denotes the volumetric sediment transport per unit width. In Fig. 6.11, Eq. (6.86) is
compared to several other relations and selected laboratory data for uniform material. The
figure is adapted from Figs. 6b and 7 in Wiberg and Smith (1989), where reference to the
data and equations can be found. The data pertain to 0.5 mm sand and 28.6 mm gravel.
Equation (6.86) shows a reasonable correspondence with the data and with several other
relations for uniform material.
The Parker relationship (Eq. 6.84) can be used to predict mobile or static armor in
gravel streams. Note that there is no formal critical stress in the formulation; instead for
$ < 1, the transport rates become extremely small. For the computation of bedload transport in poorly sorted gravel-bed rivers, the above formulation has been used to implement
a series of programs named "ACRONYM" (Parker, 1990). The program "ACRONYMl"
provides an implementation of the surface-based bedload transport equation presented in
Parker (1990). It computes the magnitude and size distribution of bedload transport over
a bed surface of given size distribution, on which a given boundary shear stress is
imposed. The program "ACRONYM2" inverts the same bedload transport equation,
allowing for calculation of the size distribution at a given boundary shear stress. The program was used to compute mobile and static armor size distributions in Parker (1990) and
Parker and Sutherland (1990).
The program "ACRONYM3" allows for the computation of aggradation or degradation to a specified active or static equilibrium final state. To this end, Parker's method
(1990) is combined with a resistance relation of the Keulegan type. In the program, both
constant width and water discharge are assumed.
The program "ACRONYM4" is directed toward the wavelike aggravation of self-similar form discussed in Parker (199Ia, 199Ib). It uses Parker's method and a resistance
relation of the Manning-Strickler type to compute downstream fining and slope concavity caused by selective sorting and abrasion.
FIGURE 6.12 Plots of CO0 and G(J)0 versus (J)^0, the asymptotes are noted on the plot. (Parker, 1990)
Next Page
CHAPTER 7
RISK/RELIABILITY-BASED
HYDRAULIC ENGINEERING
DESIGN
Yeou-Koung Tung
Department of Civil Engineering
Hong Kong University of Science and Technology
Clear Water Bay Kowloon,
Hong Kong
7.1
INTRODUCTION
(7.1)
where P(*) denotes the probability. The failure probability, pf, is the compliment of the
reliability which can be expressed as
pf=P(L>R)= l-ps
(7.2)
In hydraulic engineering system design and analysis, loads generally arise from natural
events, such as floods and storms, which occur randomly in time and in space. A common
practice for determining the reliability of a hydraulic engineering system is to assess the
return period or recurrence interval of the design event. In fact, the return period is equal to
the reciprocal of the probability of the occurrence of the event in any one time interval. For
most engineering applications, the time interval chosen is 1 year so that the probability
associated with the return period is the average annual exceedance probability of the occurrence of the event. Flood frequency analysis, using the annual maximum flow series, is a
typical example of this kind of application. Hence, the determination of return period
depends on the time period chosen (Borgman, 1963). The main disadvantage of using the
return period method is that reliability is measured only in terms of time of occurrence of
loads without considering the interactions with the system resistance (Melchers, 1987).
Two other types of reliability measures that consider the relative magnitudes of resistance and anticipated load (called design load) are frequently used in engineering practice. One is the safety margin (SM), defined as the difference between the resistance (R)
and the anticipated load (L), that is,
SM = R-L
(7.3)
The other is called the safety factor (SF), a ratio of resistance to load, which is
defined as
SF = RIL
(7.4)
Yen (1979) summarized several types of safety factors and discussed their applications to
hydraulic engineering system design.
There are two basic probabilistic approaches to evaluate the reliability of a hydraulic
engineering system. The most direct approach is a statistical analysis of data of past failure
records for similar systems. The other approach is through reliability analysis, which considers and combines the contribution of each factor potentially influencing the failure. The
former is a lumped system approach requiring no knowledge about the behavior of the facility or structure nor its load and resistance. For example, dam failure data show that the overall average failure probability for dams of all types over 15m height is around 10~3 per dam
per year (Cheng, 1993). In many cases, this direct approach is impractical because (1) the
sample size is too small to be statistically reliable, especially for low probability/high consequence events; (2) the sample may not be representative of the structure or of the population; and (3) the physical conditions of the dam may be non-stationary, that is, varying with
respect to time. The average risk of dam failure mentioned above does not differentiate concrete dams from earthfill dams, arch dams from gravity dams, large dams from small dams,
or old dams from new dams. If one wants to know the likelihood of failure of a particular
10 -year-old double-curvature arch concrete high dam, one will most likely find failure
data for only a few similar dams, this is insufficient for any meaningful statistical analysis.
Since no dams are identical and dam conditions change with time, in many circumstances,
it may be more desirable to use the second approach by conducting a reliability analysis.
There are two major steps in reliability analysis: (1) to identify and analyze the uncertainties of each contributing factor; and (2) to combine the uncertainties of the stochastic
factors to determine the overall reliability of the structure. The second step, in turn, may
proceed in two different ways: (1) directly combining the uncertainties of all factors, or
(2) separately combining the uncertainties of the factors belonging to different components or subsystems to evaluate first the respective subsystem reliability and then combining the reliabilities of the different components or subsystems to yield the overall reliability of the structure. The first way applies to very simple structures, whereas the second way is more suitable for complicated systems. For example, to evaluate the reliability of a dam, the hydrologic, hydraulic, geotechnical, structural, and other disciplinary reliabilities could be evaluated separately first and then combined to yield the overall dam
reliability. Or, the component reliabilities could be evaluated first, according to the different failure modes, and then combined. Vrijling (1993) provides an actual example of
the determination and combination of component reliabilities in the design of the Eastern
Scheldt Storm Surge Barrier in The Netherlands.
The main purpose of this chapter is to demonstrate the usage of various practical
uncertainty and reliability analysis techniques through worked examples. Only the essential theories of the techniques are described. For more detailed descriptions of the methods and applications, see Tung (1996).
7.2
In this section, several analytical methods are discussed that would allow an analytical
derivation of the exact PDF and/or statistical moments of a random variable as a function
of several random variables. In theory, the concepts described in this section are straightforward. However, the success of implementing these procedures largely depends on the
functional relation, forms of the PDFs involved, and analyst's mathematical skill.
Analytical methods are powerful tools for problems that are not too complex. Although
their usefulness is restricted in dealing with real life complex problems, situations do exist
in which analytical techniques could be applied to obtain exact uncertainty features of
model outputs without approximation or extensive simulation. However, situations often
arise in which analytical derivations are virtually impossible. It is, then, practical to find
an approximate solution.
7.2.1 Analytical Technique: Fourier and Exponential Transforms
The Fourier and exponential transforms of a PDF, fx(x), of a random variable X are
defined, respectively, as
&x(s) = E[eisx] = I eisxfx(x) dx
(7.5a)
^x(s) = E[e] = I
(7.5b)
and
e**fx(x) dx
where i = V-I and() is the expectation operator; E[eisX] andE[e sX ] are called, respectively, the characteristic function and moment generating function, of the random variable
X. The characteristic function of a random variable always exists for all values of the arguments whereas for the moment generating function this is not necessarily true.
Furthermore, the characteristic function for a random variable under consideration is
unique. In other words, two distribution functions are identical if and only if the corresponding characteristic functions are identical (Patel et al., 1976). Therefore, given a characteristic function of a random variable, its probability density function can be uniquely
determined through the inverse transform as
/ = 2^/1 e-isx9x(s)
ds
(7-6>
The characteristic functions of some commonly used PDFs are shown in Table 7.1.
Furthermore, some useful operational properties of Fourier transforms on a PDF are given
in Table 7.2
Using the characteristic function, the r* order moment about the origin of the random
variable X can be obtained as
-^1,
Fourier and exponential transforms are particularly useful when random variables are
independent and linearly related. In such cases, the convolution property of the Fourier
transform can be applied to derive the characteristic function of the resulting random vari-
PDF,
fyx)
Characteristic Function
Binomial
Poisson
Uniform
Normal
Gamma
Exponential
ExtremeValue I
Standard
fax)
Scaling
f^ax)
Linear
af^x)
Translation 1
^fx(X)
Translation 2
/x(jc - a)
Source: From Springer (1979).
Random
Variable
Fourier
Transform
X
X
X
X
X
&x(s)
a~l&x(s/a)
ax(s)
&x(s + id)
e~ x(s)
(7.8a)
(7.8b)
and
which is the product of the characteristic and moment generating functions of each
individual random variable. The resulting characteristic function or moment generating
function for W can be used in Eq. (7.7) to obtain the statistical moments of any order for
the random variable W. Furthermore, the inverse transform of $y, according to Eq.
(7.6), can be made to derive the PDF of W, if it is analytically possible.
7.2.2 Analytical Technique: Mellin Transform
When the random variables in a function W = g(X) are independent and nonnegative and
the function g(X) has a multiplicative form as
TV
W=g(X) = a0 II X?
(7.9)
i=i
the Mellin transform is particularly attractive for conducting uncertainty analysis (Tung,
1990). The Mellin transform of a PDFAM, where x is positive, is defined as
Mx(S) = M\fx(x)] = I X3^fx(X) dx = E (x*-l\ x>0
JO
(7.10)
where Mx(s) is the Mellin transform of the function fx(x) (Springer 1979). Therefore, the
Mellin transform provides an alternative way to find the moments of any order for nonnegative random variables.
Similar to the convolutional property of the exponential and Fourier transforms, the
Mellin transform of the convolution of the PDFs associated with multiple independent
random variables in a product form is simply equal to the product of the Mellin transforms of individual PDFs. In addition to the convolution property, the Mellin transform
has several useful operational properties as summarized in Tables 7.3 and 7.4.
Furthermore, the Mellin transform of some commonly used distributions are summarized in Table 7.5.
TABLE 7.3 Operation Properties of the Mellin Transform on a PDF
Property
Standard
Scaling
Linear
Translation
Exponentiation
Source: From Park (1987).
PDF
fx(x)
fx(x)
fx(ax)
afx(x)
xafx(x)
AC*0)
Random
Variable
X
X
X
X
X
Mellin
Transform
Mx(s)
a~sMx(s)
a Mx(s)
Mx(a + s)
a~l Mx(sla)
PDF Given
M^s)
W =X
A ( X ) M X
W = Xb
A(JC)
Mx(bs - b + 1)
W = UX
AW
Mx(2-s)
W = XY
A(JC), ,00
Mx(S)M^s)
W = XIY
A(*), S7OO
Mx(S)MfI-S)
W = aXbY<
A(JC), gfy)
a* ~ Mx(bs -b+ \)M^cs -c + 1)
Source: From Park (1987). a, b, c: constants ; X, Y, W: random variables.
TABLE 7.5 Mellin Transforms for Some Commonly Used Probability Density Functions
Probability
PDF, j^x)
Mellin Transform
Uniform
Standard Normal
Lognormal
Exponential
Gamma
Triangular
Weibull
Nonstandard beta
Standard beta
where Mx(k) for standard beta
Example 7.1. Manning's formula is frequently used for determining the flow capacity of storm sewer by
Q = 0.463 n~l D267 S05
where Q is flow rate (ftVs), n is the roughness coefficient (ft176), D is the sewer diameter
(ft), and S is pipe slope (ft/ft). Assume that all three model parameters are independent
random variables with the following statistical properties. Compute the mean and variance of the sewer flow capacity by the Mellin transform.
Parameter
Distribution
D
S
Referring to Table 7.4, the Mellin transform of sewer flow capacity, MQ(s), can be
derived as
MQ(s) = 0.463s ~ l Mn(-s + 2) MD(2.61s - 1.67) Ms(0.5s + 0.5)
For roughness coefficient n having a uniform distribution, from Table 7.5, one obtains
n S+2
b~ ~ na~S +2
Mn(~s + 2) = (_s + 2)(Hb_na)
fe,f5+o-5-(oa5+05
(0:55 + 0 . 5)( ;_ 0
Substituting individual terms into MQ(s) results in the expression of the Mellin transform of sewer flow capacity specifically for the distributions associated with the three stochastic model parameters.
Based on the information given, the Mellin transforms of each stochastic model parameter can be expressed as
,, f . 0.0163* - 0.0137s
^ =
00026^
MD(S)
2
[3.147(3.147-3.000 2.853(3.00* - 2.853*)]
~ (0.294),(5 + 1) [
0-147
~
0.147
\
M( .
M s)
0.00543* - 0.00457*
^
0.00086 5
The computations are shown in the following table:
s 1
0.463 "
M n (-j + 2)
MD(2.61s - 1.67)
Mn(0.55 + 0.5)
s=2
s=3
0.463
Mn(O) = 66.834
MD(3.67) = 18.806
M/1.50) = 0.0707
0.2144
MB(-1) = 4478.080
MD(6.34) = 354.681
Af/2.00) = 0.005
(7.11)
where X = (X1, X2, . . ., XN)1, an TV-dimensional column vector of random variables, the
superscript t represents the transpose of a matrix or vector. The Taylor series expansion of
the function g(X) with respect to the means of random variables X = |U in the parameter
space can be expressed as
- ^ *i=i
iM
* dXjjv
l
[ dAf
>- w 4 i
* zi (
j =f
i [OX
i=
able at (I.
Dropping the higher-order terms represented by e, Eq. (7.12) is a second-order
approximation of the model g(X). The expectation of model output W can be approximated as
N
r
i
E(W] g(n) + I ^ S [S|Cov [X" X;]
(7 13)
'
and
TV r
i
E[W] #00 + U
^SrI R&i ~ M*)2]
2
; = i L dX? I
<7-15)
C-H9
(7.17)
< 7 - 18 >
and
(7.2Oa)
(7.2Ob)
(7.2Oc)
P+x'+-P_x>
_ = Y
(7.2Od)
in which x'_ = \x_ |i|/<7, jc'+ = |*+ - |i|/a, and y is the skew coefficient of the random
variable X. Solving Eqs. (7.20a-d) simultaneously, one obtains
Y
/
/Y\ 2
+
1+
^ =I V
(I)
+
(7 21a)
x'_=x'+-<f
(7.21b)
P+ = -T-V-TJt + + X _
(7.2Ic)
P^ = I-P+
(7.2Id)
When the distribution of the random variable X is symmetric, that is, y = O, then Eqs.
(7.21a-d) are reduced to x'_ = x\ = 1 andp_ = p+ = 0.5. This implies that, for a symmetric random variable, the two points are located at one standard deviation to either side
of the mean with equal probability mass assigned at the two points.
From x'_ and jc' + the two points in the original parameter space, x_ and X+, can respectively be determined as
x_ = \i-x'_a
(7.22a)
X+ = \L + X'+G
(7.22b)
Based on x_ and X+9 the values of the model W = g(X) at the two points can be computed, respectively, as w_ = g(x_) and W + = g(x+). Then, the moments about the origin
ofW = g(X) of any order can be estimated as
E[W-] = [i\m ~ p+ w+- + p_ w_
(7.23)
(7.24)
in which the subscript 8- a sign indicator and can only be + or representing the random variable X1 having the value of Jt1+= p.,-+JtV+ a. or xt_ = ji.-jt'^a,, respectively; the
probability mass at each of the 2N points, p(51t 52,..., 5Ao can be approximated by
N
with
N-I ,
p^...^ =n
1=1 M + E ( E
i = 1 \j = i + 1M /
_
(7 25)
P/ 2N
"JFi]
where p.. is the correlation coefficient between random variables Xi and Xj. The number of
terms in the summation of Eq. (7.24) is 2N which corresponds to the total number of possible combinations of + and for all N random variables.
Example 7.3 Referring to Example 7.2, assume that all three model parameters in
Manning's formula are symmetric random variables. Determine the uncertainty of sewer
flow capacity by Rosenblueth's PE method.
Solution: Based on Manning's formula for the sewer, Q = 0.463 n~l D261 S 05 , the standard deviation of the roughness coefficient, sewer diameter, and pipe slope are
Cn = 0.00075; G0 = 0.06; Gs = 0.00025
With N = 3 random variables, there are a total of 23 = 8 possible points to be considered
by Rosenblueth's method. Since all three random variables are symmetric, their skew
coefficients are equal to zero. Therefore, according to Eqs. (7.21a-b), ri_ n'+ = U _ =
D'+ = S'_ = S1+ = 1 and the corresponding values of roughness coefficient, sewer diameter, and pipe slope are
n+ = JiIn + ri+ Gn = 0.015 + (1)(0.00075) = 0.01575
/i_ = M* ~ ri- Vn = 0.015 - (1X0.00075) = 0.01425
D+ = [I0 + D'+ G0 =3.0 + (1)(0.06) = 3.06 ft
D_ = \iD- U_ G0 = 3.0 - (1)(0.06) = 2.94 ft
S+ = \is + 5"+ G5 = 0.005 + (1)(0.00025) = 0.00525
S_ = \LS- S'_ G8 = 0.005 - (1)(0.00025) = 0.00475
Substituting the values of n_,n + ,>_,>+, 5_, and S+ into Manning's formula to compute the corresponding sewer capacities, one has, for example,
n
+
+
+
+
-
D
+
+
-
S
+
+
+
+
-
Q(ft3s)
42.19
40.14
37.92
36.07
46.64
44.36
41.91
39.87
+
-
p
0.03125
0.03125
0.21875
0.21875
0.21875
0.21875
0.03125
0.03125
Because the roughness coefficient and sewer diameter are symmetric, correlated random variables, the probability masses at 23 = 8 points can be determined, according to
Eqs. (7.25-7.26) as
P+++ = P = d + PnD + Pns +
p++_ = p__+ = (1 + PnD - pns P+^+ = p_ + _ - (1 - pnD + pns /?_ ++ = P+^ = (l~ pnD - Pns +
The values of probability masses also are tabulated in the last column of the above
table. Therefore, the mth order moment about the origin for the sewer flow capacity can be
calculated by Eq. (7.24). The computations of the first two moments about the origin are
shown in the following table in which columns (l)-(3) are extracted from the above table.
Point
(1)
1
2
3
4
5
6
7
8
Sum
<2(ft3)
(2)
42.19
40.14
37.92
36.07
46.64
44.36
41.91
39.87
p
(3)
0.03125
0.03125
0.21875
0.21875
0.21875
0.21875
0.03125
0.03125
1.00000
QXp
(4)
1.318
1.254
8.295
7.890
10.203
9.704
1.310
1.246
41.219
Q2
(5)
1780.00
1611.22
1437.93
1301.04
2175.29
1967.81
1756.45
1589.62
Q2Xp
(6)
55.625
50.351
314.547
284.603
475.845
430.458
54.889
49.676
1715.99
3
From the above table, \IQ = E(Q) = 41.22 ft3 and E(Q2) = 1715.994 /ft
I V
J . Then, the
variance of the sewer flow capacity can be estimated as
\s/
/ft 3 V
Var(0 = E(Q2) - faQ)2 = 1715.994 - (41.219)2 = 16.988
Vs/
Hence, the standard deviation of sewer flow capacity is V16.988 = 4.12 fWs.
Comparing with the results in Example 7.2, one observes that Rosenblueth's PE method
yields higher values of the mean and variance for the sewer flow capacity than those
obtained by the FOVE method.
7.2.5 Approximate Technique: Hair's Probabilistic
Point Estimation (PE) Method
To avoid the computationally intensive nature of Rosenblueth's PE method when the
number of random variables is moderate or large, Harr (1989) proposed an alternative
probabilistic PE method which reduces the required model evaluations from 2N to 2N and
greatly enhances the applicability of the PE method for uncertainty analysis of practical
problems. The method is a second-moment method which is capable of taking into account
of the first two moments (that is, the mean and variance) of the involved random variables
and their correlations. Skew coefficients of the random variables are ignored by the
method. Hence, the method is appropriate for treating normal and other symmetrically
distributed random variables. The theoretical basis of Harr's PE method is built on orthogonal transformations of the correlation matrix.
The orthogonal transformation is an important tool for treating problems with correlated random variables. The main objective of the transformation is to map correlated random variables from their original space to a new domain in which they become uncorrelated. Hence, the analysis is greatly simplified.
Consider Af multivariate random variables X = (X1, X2, ..., XN)1 having a mean vector
JLlx = (JUt1, |I2, ..., jn^)' and the correlation matrix R(X)
1
Pi2 Pi3
Pii * P23
- PIAT
PIN
C (X') = R ( X ) =
PM PATC P*3
Note that the correlation matrix is a symmetric matrix, that is, pj;/. = p{j for i = j.
The orthogonal transformation can be made using the eigenvalue-eigenvector decomposition or spectral decomposition by which R(X) is decomposed as
R(X) = Cf X') = V A V'
(7.27)
(7.28)
(7.29)
in which xi represents the vector of coordinates of the N random variables in the parameter space corresponding to the ith eigenvector V 1 ; |il = (U1, U 2 ,..., uN)r, a vector of means
of TV random variables X.
Based on the 2TV points determined by Eq. (7.29), the function values at each of the 2TV
points can be computed. Then, the mm moment of the model output W about the origin
can be calculated according to the following equations:
wm + wm
pm(x } + Qm (r }
wrn=^i_^_^_ = S (xi+) -r g (x._) ^
fof
.= ^ ^
^ N-9m=i9 2, ...
(7.30)
TV
]>><
E[W] = Un(W) =
, for m = 1, 2,...
(7.31)
V = [V1 V2 V3] =
n V i2 Vi3
0.7071
V
V
2i 22 23 = -0.7071
V
3i V32 V 33j L o.OOOO
0.7071 0.0000
0.7071 0.0000
0.0000 1.0000.
and
<* O O
xi = M-D
\LS
V3
0.015
GD O V1.
O O G5
0.00075 O
= 3.0 V3
0.005
O
O
0.06 O
O
0.00025
v/
, f o r / = 1,2,3
The resulting coordinates at the six intersection points from the above equation are listed in column (2) of the table given below. Substituting the values of x in column (2) into
Manning's formula, the corresponding sewer flow capacities are listed in column (3).
Column (4) lists the value of Q2 for computing the second moment about the origin later.
After columns (3) and (4) are obtained, the averaged value of Q and Q2 along each eigenvector are computed and listed in columns (5) and (6), respectively.
Point
(1)
x - (TI, D, S)
(2)
Q
(3)
Q2
(4)
1+
12+
23+
3-
(0.01592,2.9265,0.00500)
(0.01408,3.0735,0.00500)
(0.01592,3.0735,0.00500)
(0.01408,2.9265,0.00500)
(0.01500,3.00,0.00543)
(0.01500,3.00,0.00457)
36.16
46.61
41.22
40.89
42.74
39.20
1307.82
2172.14
1699.09
1671.99
1826.45
1537.17
~Q
(5)
~Q2
(6)
41.39
1739.98
41.05
1685.54
40.97
1681.81
The mean of the sewer flow capacity can be calculated, according to Eq. (7.31), with
m = 1, as
^
_ A1 Q1 + A 2 Q 2 + A3 Q3
A1 + A2 + A3
=
_ X 1 Qf + ^ 2 Ql + ^ 3 Q ?
K1 + K2 + ^3
^^ ^
16.82
(ft3/s)2.
Hence, the standard deviation of sewer flow capacity is V16.82 = 4.10 ftVs.
Comparing with the results in Examples 7.2 and 7.3, one observes that the mean and variance of the sewer flow capacity computed with Hair's PE method lie between those computed with the FOVE method and Rosenblueth's method.
(7.32)
As discussed in the preceding sections, the natural hydrologic randomness of flow and
precipitation are important parts of the uncertainty in the design of hydraulic structures.
However, other uncertainties also may be significant and should not be ignored.
7.3.1 Performance Functions and Reliability Index
In the reliability analysis, Eq. (7.32) can alternatively be written, in terms of a performance function, W(X) = W(XL, XR)9 as
Next Page
CHAPTER 8
HYDRAULIC DESIGN FOR
ENERGY GENERATION
H. Wayne Coleman
C. Y. Wei
James E. Lindell
Harza Company
Chicago, Illinois
8.1
INTRODUCTION
This chapter describes the design aspects of hydraulic structures related to the production of hydroelectric power. These structures include headrace channels; intakes;
conveyance tunnels; surge tanks; penstocks; penstock manifolds; draft-tube exits; tailtunnels, including tail-tunnel surge tanks and outlets; and tailrace channels. The procedures provided in this chapter are most suitable for developing the preliminary
designs of hydraulic structures related to the development of the hydroelectric projects. To finalize designs, detailed studies must be conducted: for example, economic
analysis for the determination of penstock diameters, computer modeling of hydraulic
transients for surge tank design, and studies of physical models of intake and its
approach.
8.2
HEADRACECHANNEL
Minimum.
Maximum
Unlined rock
0.030
0.035
Shotcrete
0.025
0.030
Formed concrete
0.012
0.016
Grassed earth
0.030
0.100
Headrace channels are generally designed and sized for a velocity of about 2 m/s
(6.6 ft/s) at design flow conditions. Economic considerations may result in some variation
from this velocity, depending on actual project conditions.
Channel sections are normally trapezoidal because this shape is easier to build for
many different geologic conditions. The bottom width should be at least 2 m (6.6 ft) wide.
Side slopes are determined according to geologic stability as follows: earth, 2H: 1V or flatter; and rock, IH: IV or steeper. The channel's proportionsbottom width versus depth
are largely a matter of construction efficiency. In general, the minimum bottom width
reduces excavation, but geologic conditions may require a wider, shallower channel. The
channel slope will result from the conveyance required to produce design velocity for
design flow.
Channel bends should have a center-line radius of 3W to 5W or more, where W is the
water surface width of the design flow. For this radius, head loss and the rise in the water
surface at the outer bank (superelevation) will be minimal. If the radius must be reduced,
the following formula can be used to estimate head loss hL:
hL = K6^
(8.1)
(8.2)
(8.3)
8.3
INTAKES
Most power intakes are horizontal, a few are vertical, and very few are inclined. Figures
8.2, 8.3, and 8.4 are examples of the three types of intakes. The horizontal intake is usually connected to a tunnel or penstock on a relatively small slope (up to 2-3 percent). The
vertical intake is frequently used in pumped-storage projects when the upper reservoir is
on high ground, such as a mountain top, and a vertical shaft-tunnel is the obvious choice.
An inclined intake is used when the topography, geology, or type of dam dictate a steeper slope for the downstream tunnel or penstock. Exhibits 8.2 and 8.3 are examples of
intakes on an arch dam and on a pump-storage upper reservoir.
El 54 O
Max. operating
T.W El 50 O-
Intake
Min. operating
TW El. 43 5 -
Flow
GENERA TOR
HALL
El 3325
Trashrock guide
Emergency gate slot
7 1217.25
e/ 73727
/. 64728
Dasigrt
flood
surcharge
El. 781
Max. power pool
El. 770.0
Normal min. El. 700.0
Extreme min. pool
(Drawdown limit)
El. 6OO O
Trashrack blocKout
Axis of dam
El.785.0
penstock
E1.512
Stair house
Spillway
Powerhouse
Main dam
33 TON CRANE
FOLLOWER
TRAiHWACK
NORMAL MAX
OPERATING WL.EL.20OO
POST'TCNSIONCO
ANCHOM(TYR)
AMCHOMS (TYP.)
TRASHRACK(TYP)
FLOW
HYDRAULIC HOIST
MINIMUM OPERATING
W.L.EL.I8SO
INTAKE OATC
12-0 W 24-0H
ENDOFPIER
BULKHEAD GATE
I4'-OW 32'-OH
FIGURE 8.5 A typical multi-level intake tower structure for selective withdrawal. (Harza Engineering Co.)
hL = K,^
(8.4)
A
(A Y
where Vn = velocity based on the net area, K = 1.45 0.45 -r hp L An = net area of
A
s {AgJ
trashrack and support structure, and Ag = gross area of trashrack and support structure.
An intake gate is generally provided when the power tunnel or penstock is long or
when a short penstock does not have a turbine inlet valve. This gate is provided for emergency closure against flow in case of runaway conditions at the turbine or penstock rupture. The effective area of the gate is usually about the same as that of the power tunnel
or penstock, but it is rectangular in shape, with a height that is the same as the conduit's
diameter and a width that is 0.8 X the conduit diameter. A bulkhead (or stop-logs) is provided upstream of the intake gate for servicing the gate. The trashrack slot might be used
for this function by first pulling the trashrack.
H 737.27
A hydraulic study is generally conducted for emergency closure of the intake gate.
The maximum turbine flow or runaway flow should be considered. The runaway flow
may be 50 percent higher than the normal turbine flow for a propeller turbine. In the
hydraulic study, the water levels and pressures, as well as flow into and from the gate
well, as a function of gate position are investigated (Fig. 8.6). With this information, critical gate loads can be determined for the gate and hoist. The gate also may be used for
penstock filling. A minimum gate opening of 10 to 15 cm (4-6 in) is usually specified
for this, but a special hydraulic study must be made to determine potential gate load and
vibration if the gate opens continuously by accident. In such cases, a generous gate well
or air vent must be provided downstream of the gate to provide relief once the tunnel or
penstock fills.
The head loss for a bulkhead or gate slot, including top opening, is generally about 0.1
of the local velocity head at the slot. The transition length (m or ft) Lt from gate section
to tunnel or penstock should be approximately:
WIRE ROPES
FACE OF
ROLLER TRACK
TRASHRACK
SEAL
EL.545
TOP SEAL
PENSTOCK
EL.5I2
FEET
A =^
(8-5)
8.4
TUNNELS
When the powerhouse is situated a considerable distance from the intake and when geologic conditions permit, a tunnel is often used to convey the flow for power generation.
The size of the tunnel is dictated by economics: that is, construction cost is added to the
cost of head loss (loss of generating revenue) to obtain the minimum combined cost. This
determination is usually obtained by trial and error because the process does not lend itself
Horizontal intake
Vertical intake
Intakes with
vortex problems
Intakes with
no problems
to a simple formula. The resulting tunnel velocity with the economic diameter is usually
in the range of 3 to 5 m/s (10 to 17 ft/s).
The shape of the excavated tunnel normally will approximate a square bottom and a circular top. The diameter of the circular top (or the width of the square bottom) should be larger than the required diameter. If the tunnel is lined with concrete, its cross section is likely to
be circular or have a square or trapezoidal bottom. If it is unlined or lined with shotcrete, the
excavated shape will remain, with some smoothing by filling the larger overbreak sections.
Lining is an economic consideration, balancing the cost of the lining with the power
loss caused by friction. Even an unlined tunnel will have lined sections, such as portals,
and sections where rock needs extra support for geologic stability. Friction factors for
design are as follows:
Manning's n
Lining
Minimum Maximum
Unlined
0.030
0.035
Shotcrete
0.025
0.030
Formed concrete 0.012
0.016
Minimum friction corresponds to new conditions and is used for turbine-rating and
pressure-rise calculations. Maximum friction corresponds to aging and is used for economic-diameter and pressure-drop calculations. Tunnel slope is dictated by construction
suitability and geology, with a minimum of 1:1000 for drainage during dewatered condi-
imping
Bl*2J^o"PO'
Sta'sssSfcr'oiBaisM
200 7. mm.
tions. Tunnel horizontal bends generally have large radii for convenience of construction.
Vertical bends at shafts usually have a minimum radius of 3D to minimize head loss and
to provide constructibility.
8.5
SURGETANKS
Surge tanks generally are used near the downstream end of tunnels or penstocks to reduce
changes in pressure caused by hydraulic transients (waterhammer) resulting from load
changes on the turbines (ASCE, 1989; Chaudhry, 1987; Gulliver and Arndt, 1991; Moffat
et al., 1990; Parmakian, 1955; Rich, 1951; Wylie and Streeter, 1993; Zipparro and Hasen,
1993). A surge tank should be provided if the maximum rise in speed caused by maximum
load rejection cannot be reduced to less than 60 percent of the rated speed by other practical methods, such as increasing the generator's inertia or the penstock's diameter or by
decreasing the effective closing time of the wicket gates. In general, the provision of a
surge tank should be investigated if
I>,jj > 3 to 5 for units in m/s and m or
n
(8.6)
where L1. is the length of a penstock segment and V1 is the velocity for the segment
(Dingman, 1984). The term ^L1V1 is computed from the intake to the turbine and Hn is the
minimum net head.
Surge tanks normally are located as close as possible to the powerhouse for maximum
effectiveness and may be free-standing or excavated in rock. The tanks are usually vented to atmosphere or can be pressurized as air chambers. The latter is not used frequently
because of requirements of size, air compressors, and air tightness. Exhibit 8.4 illustrates
a pumped storage project with a surge tank. Figure 8.8 shows typical installations of surge
tanks for controlling hydraulic transients.
Surge tanks usually are simple cylindrical vertical shafts or towers, but other geometric designs are used when the surge amplitude is to be limited. For instance, an enlarged
chamber can be used at the top if upsurge might cause the water level to rise above the
ground surface. Similarly, an enlargement or lateral tunnel or chamber is sometimes used
near the bottom of the shaft if downsurge would caused the water level to drop below the
tunnel crown. When the geometry is a cylinder, analysis is relatively simple and can be
performed using design charts. If the geometry is more complicated, a hydraulic transient
simulation model is required to carry out the study (Chaudhry, 1987; Wylie and Streeter,
1993; Brater et al., 1996).
Hydraulic stability for a surge tank assures that surging is limited and brief after load
changes (Rich, 1951; Parmakian, 1955; U.S. Bureau of Reclamation 1980; Zipparro and
Hasen, 1993). The minimum cross-sectional area of a simple cylindrical surge tank
required for stability can be determined using the Thoma formula:
AL
^=2^H
(8 7)
'
where AST = minimum tank area, A = tunnel area between reservoir and surge tank, L =
tunnel length between reservoir and surge tank, g = gravitational acceleration, c =
COUNTY FOWCRPUUtT
2100 WBATH
funtf^Ster99
fitapinf t1985)
on Buck Cw*
MNOMU
FIGURE 8.8(a) Typical vented surge tank installation. Bath County powerplant (1985): 2100 MW pumped storage development on Back Creek, Virginia.
HW 1. 10250
r>n/wc*
FlOKf
Cl. 1035.0
EI tee o
0. 940
T.W. El. 900
1 674.1
O. 159.53
A#
1 ( Mi \
head loss coefficient = -F^- = y- T/2/^~~ , where A// = minimum head loss from
reservoir to surge tank, including tunnel velocity head W2g, and H - minimum net operating head on turbine.
For a simple surge tank (without an orifice), increase the diameter obtained from the
Thoma formula by 50 percent. For a typical surge tank with a restricted orifice, increase
the diameter by 25 percent. These increases are necessary to provide damping of the oscillation in a reasonable period of time.
Maximum upsurge in a cylindrical surge tank can be determined from Fig. 8.9. For a
given tank size, the optimum size of the orifice is based on the balanced head design so
that the maximum tunnel pressure below the surge tank equals the maximum upsurge
level.
Maximum downsurge in a cylindrical surge tank can be determined from Fig. 8.10.
Here again, the size of the orifice should be based on balanced head design as a first
attempt. However, since downsurge may differ from upsurge, and the required orifice size
may be different for the two purposes, shaping the orifice (i.e., changing the discharge
coefficient) by rounding the top or bottom may satisfy the two area requirements approximately.
For maximum upsurge, use the maximum normal headwater, minimum head loss
between reservoir and surge tank, and maximum plant flow. Assume full plant load-rejection (tripout) in the shortest reasonable time.
For maximum downsurge, use the minimum normal headwater, maximum head loss,
and accept load from 50 percent to 100 percent in the shortest reasonable time. At some
projects, such as pumped-storage plants, the load acceptance is criterion is more extreme;
full load acceptance, is O percent to 100 percent in the shortest reasonable time. The controlling criterion will be used to design the orifice on downsurge. When the surge tank
geometry is complex (noncylindrical), a computer model should be used to determine the
limiting surge levels. (See Brater et al., 1996, for available computer models). Freeboard
SURGE RATIO
BALANCED DESIGN
SURGE RATIO
BALANCEODESIGN
for the surge tank is 10 percent of the computed rise in the water level in the surge tank
for upsurge and 15 percent of the drop in the water level for downsurge to maintain, submergence of the tank invert or the orifice to avoid admitting air into the penstock.
Pressurized air chambers are often used in pumping plants for surge protection. They
are used occasionally for power plants when the generating flow is not excessive. The
hydraulic characteristics of the chambers are complicated by the compressibility effects
of air and temperature, and the analysis does not lend itself to simple formulas and charts.
A computer model is required to verify performance. Fig. 8.8(B) shows a typical air chamber design for a hydropower plant.
8.6
PENSTOCKS
Cost of pipe
Value of energy loss
Plant efficiency
Minor loss factor
Average head
Waterhammer effect
For the assessment of a preliminary design or a feasibility level, the most economical
diameter can be estimated using the following formula (Moffat et al., 1990).
CP0-43
D. = -JpW
(8.8)
where De = the most economical penstock diameter (m or ft), H = the rated head (m or
ft), P = the rated capacity of the plant (kW or hp), and C = 0.52 (for metric units) or =
3.07 (for English units). If the project is a small hydropower installation, the following
simple equation can be used (Warnick et al., 1984).
D6 = CQ0*
(8.9)
where De = the most economical penstock diameter (m or ft), Q = the design discharge
(m3/s or ftVsec), and C = 0.72 (for metric units) or = 0.40 (for English units).
For large hydroelectric projects with heads varying from approximately 60 m (190 ft)
to 315 m (1,025 ft) and power capacities ranging from 154 MW to 730 MW, the following equation can be used (Warnick et al., 1984).
O?0-43
. = -%r
(8-10)
where D6 = the most economical penstock diameter (m or ft), p = the rated turbine capacity (kW or hp), h = the rated net head (m or ft), and C = 0.72 (for metric units) or = 4.44
(for English units).
The maximum velocity in the penstock is normally kept lower than 10 m/s (33 ft/s).
To determine the minimum thickness of the penstock, based on the need for stiffness, corrosion protection, and handling requirements, the following formula can be used (U.S.
Bureau of Reclamation, 1967; Warnick et al., 1984).
D+K
*min = 400
(8 U)
'
where tmin = the minimum thickness of the penstock (mm or in), D = penstock diameter
(mm or in), and K = 500 (for metric units) or = 20 (for English units). After determining
the economic diameter, check for the operating stability of the generating unit-penstock
combination using the following steps (Chaudhry, 1987; U.S. Bureau of Reclamation,
1980; Warnick et al., 1984).
1. Determine the mechanical starting time in seconds for the unit T1n as
lm
JGD^V-_
36XlO4P
^'^
or
(WR2W
~ 1.6 X 106P7
(8 13)
'
where GD = flywheel effect of the turbine and generator rotating parts used in
metric system (kg-m2), WR2 = flywheel effect of the turbine and generator rotating
parts in English system (lb-ft2) = 5.932 GD2, G = weight of rotating parts (kg), D
= 2 X radius of gyration of the rotating parts (m), W = weight of rotating parts
(Ib), R = radius of gyration of the rotating parts (ft), N = turbine speed (rpm), P =
maximum turbine output (kW), and P1 = maximum turbine output (hp).
Tm is the time for torque to accelerate the rotating mass from zero to rotational
speed. Together, the turbine runner in water, connecting shafts, and the generator
develop the flywheel effect WR2 or GD2. The WR2 can be determined using on the
following formulas:
WRL^ = 23,800 (^)
(8.14)
and
KVA\5/4
-JJST)
'
(8.15)
where Pd = turbine rated output (hp) and kVA = generator rated output (kilovoltamperes).
2. Determine the water column starting time for the penstock Tw as follows:
Z(LV)
= ~W~
<8-16)
where ^LV) = summation of product of length (measured from nearest open water
surface) and velocity for each segment of penstock from intake or surge tank to tailrace (m2/s or ftVsec), g = gravitational acceleration (m/s2 or ft/sec2), and H = minimum net operating head (m or ft).
3. In general, TJTj should be maintained greater than 2 for good operating stability
and to have reasonably good responses to load changes. If Tm/Tw2 is less than 2,
there are three possible solutions:
Increase WR2 or GD2 for the generator; this is relatively inexpensive for
increases of up to 50%.
Increase the penstock diameter; this is probably not economical, except for a
narrow range.
Add a surge tank or move the surge tank closer to the powerhouse.
A combination of these three possible solutions may be the most cost-effective solution. The following friction factors are recommended for designing steel penstocks:
Penstock Age
New
Old
Manning's n
0.012
0.016
Use the value for new penstock to calculating turbine-rating and pressure-rise. To calculate pressure drop use the higher values.
Design pressure is determined on the basis of the turbine's characteristics and the closure rates of the wicket gates or needle valves. For Pelton turbines, closure rates are slow,
and design pressure rise is usually of the order of 20 percent of the static pressure head.
For Francis turbines, design pressure rise is usually 30 to 40 percent of the static pressure head, depending on the cost of steel lining required. A fast closure is desirable to
minimize speed rise and the potential for runaway conditions in the turbine. Detailed
pressure conditions are determined by a computer model that includes the water conductors and surge tank as well as the turbine discharge-speed characteristics and generator inertia. Many computer programs capable of simulating hydraulic transients are
described in Wylie and Streeter, 1993. Such computer simulation studies are often
required of turbine or governor manufacturers now as a part of the specifications.
Ultimately, the predicted pressure conditions are verified in the load rejection tests during unit start-up.
The profile for a free-standing penstock is based on the topographic and geologic conditions of the ground. In other cases, the penstock may consist of shaft and tunnel sections
that are largely lined with concrete, with a relatively short section of steel-lined penstock
near the powerhouse.
If the penstock is free-standing, the risk of penstock rupture is greater than it is for the
shaft and tunnel system. If there is a long tunnel section upstream of the free-standing penstock, an emergency closure valve is often added near the tunnel outlet. A hydraulic transient study is necessary to determine closure conditions (by accident or because of penstock rupture). A vent must be provided to admit air just downstream of the valve for penstock rupture and must be large enough to prevent collapse of the penstock from internal
subatmospheric pressure caused by water-column separation. A free-standing penstock
also requires small air inlet-outlet valves at local high points to remove air during filling
and admit air during dewatering.
8.6.1 Penstock Branches
A penstock often delivers water to more than one turbine. In such cases, the penstock is
branched in various ways to subdivide the flow.Exhibit 8.5 illustrates a typical steel penstock branch structure. When the powerhouse is normal to the penstock, several configurations are possible (Fig. 8.11). If the powerhouse is at an angle with the penstock, a manifold is used (Fig. 8.12).
Head losses in branches and manifolds depend on precise geometry and often are
developed by model studies. However, for a typical well-designed layout, the following
head loss coefficients can be used to estimate the head loss hb from the main into a
branch:
hb = Kb^
(8.17)
where V = branch velocity (m/s or ft/s); g = gravitational acceleration (m/s2 or ft/sec2); and
Kb = head loss coefficient 0.2 for symmetrical bifurcation, 0.3 for symmetrical trifurcation, and 0.2 for manifold branch.
MAIN PENSTOCK
SINGLE
BIFURCATION
MAIN PENSTOCK
TRIFURCATION
BRANCH
(TYPICAL)
2 UNITS
BRANCH
(TYPICAL)
3 UNITS
MAIN PENSTOCK
MAIN PENSTOCK
BOUBLE
Y-BRANCHING
BRANCH
(TYPICAL)
3 UNITS
DOUBLE
BIFURCATION
BRANCH
(TYPICAL)
4 UNITS
FIGURE 8.11(a) Schematic penstock branch configurations for powerhouse normal to the penstock.
The diameters of branched penstocks are usually determined so that the velocity is
increased significantly relative to the main penstock. Here again, the branch size is determined by economics so that construction and material costs added to cost of energy loss
are at a minimum. The lower limit for the size of the branch is the size of the turbine inlet
that is normally provided by the turbine manufacturer. If a turbine inlet valve is provided,
its diameter will either be equal to the inlet diameter or be between the inlet diameter and
the penstock branch diameter. This valve is usually a spherical type, and, as such, no head
loss occurs in the fully open position. Friction losses in the branch penstocks are calculated using the same friction factors used for the main penstock and the conduit lengths
up to the net head taps in the turbine inlet.
SINGLE BIFURCATION
TRIFURCATION
DOUBLE Y-BRANCHING
FIGURE 8.11(b) Configurations for single bifurcated, double y-branching, and trifurcated penstocks.
(Harza Engineering Co.).
8.7
DRAFT-TUBEEXITS
Draft-tubes are designed by considering the turbine's characteristics. The net head for the
turbine is based on pressure taps at the spiral-case inlet and near the draft-tube exit.
Therefore, any head losses which occur after the draft-tube pressure taps are subtracted
from the turbine net head. Because the exit head loss is generally considered to be the
average velocity head at the end of the draft-tube, a longer draft-tube with expansion to a
larger area would, in theory, reduce this loss. In actuality, however the flow is not uniform
MAIN PENSTOCK
MAIN PENSTOCK
BRANCH
(TYPICAL)
BRANCH
(TYPICAL)
3 UNITS
2 UNITS
MAIN PENSTOCK
BRANCH
(TYPICAL)
4 UNITS
FIGURE 8.12(a) Schematic penstock manifold configurations for a powerhouse
oriented at an angle with the penstock.
FIGURE 8.12(b) Penstock manifold for an installation with six units. (Harza Engineering Co.)
at this point; it is highly turbulent and swirling, and the true exit loss is difficult to define.
Current thinking is that further extension of the draft-tube is not economical. The rule of
thumb is to end the draft-tube when the mean velocity is about 2m/s and to base the exit
head loss on this velocity.
A trashrack is usually provided at the end of the draft-tube at a pumped-storage project to prevent entry of coarse debris during the pumping mode. However, during the generating mode with the trashrack in place, the trashrack is subject to vibration caused by
the concentration of flow and by swirling. This complicates the design of the trashrack
and increases its cost. The analysis of the rack is a combined hydraulic and structural one.
The hydraulic loadings consist of drag forces on rack bars that are dependent on velocity
patterns along with pulsation of pressure caused by swirling flow. The data on hydraulic
conditions can be obtained from a physical model (usually the model from the pump-turbine manufacturer) because fully developed mathematical models are not readily available to predict these forces. A structural mathematical model is then applied using the
hydraulic loadings obtained from the hydraulic model tests. By trial and error, the
trashrack is designed to withstand the flow-induced vibrations.
8.8
TAIL-TUNNELS
An underground power plant will have a tail-tunnel to deliver the flow to the downstream
river or lake. For a pumped-storage project, this tunnel provides flow both ways, because
it acts as the inlet tunnel during pumping.
For a conventional hydroelectric plant with generating only, the tunnel is usually pressurized.. However, if the turbines are the Pelton type, the tunnel is likely to be free flow
to maintain freeboard on the turbine. For a pumped-storage plant, the tunnel is most likely to be pressurized, because it must deliver water both ways. If the tunnel is pressurized
and is long enough, a surge chamber will be required to prevent large fluctuations of pressure on the turbines during load changes.
The number of tail-tunnels, usually one or two, is based on economics and constructability. From an operational standpoint, two tunnels are desirable to allow partial
operation of the plant even during maintenance or inspection of one of the tunnels.
However, two tunnels are usually more expensive than one, and usually only one will be
used unless its size becomes unmanageable. The limiting size is dictated by available
equipment and tunneling methods. These factors must be evaluated carefully when estimating the costs of one tunnel versus two tunnels.
A manifold is used to collect the flow from the individual draft-tubes and guide the
flow through a transition section to the tail-tunnel proper. This manifold is similar in concept to the penstock manifold, but generally the velocities are much lower. The velocity
at the end of the draft-tube is typically 2 m/s (7 ft/s) and 3 m/s (10 ft/s) at the tail-tunnel.
Therefore, head losses are not significant and the flow conditions are generally acceptable. A typical tail-tunnel manifold design is shown on Fig. 8.13.
8.8.1 Tail-Tunnel Surge Tanks
When an underground power plant has a significant length of pressurized tail-tunnel, a
surge tank is likely to be required. The procedures for sizing and determining extreme
surges are similar to the procedures used for surges in the head-tunnel, using the hydraulic
characteristics of the tail-tunnel instead of the head-tunnel. (Refer to Sec. 8.5). Figure 8.14
shows a typical tail-tunnel surge chamber.
GENERAL PLAN
8.9
TAILRACECHANNELS
If the outlet structure is a significant distance from the receiving waterway, a tailrace
channel will be required (Fig. 8.16). The sizing of the channel will be similar to that of the
headrace channel. (Refer to Sec. 8.2).
FON TRASHRACK
(SLOT
SEE NOTE
31
BULKHEAD SLOT
Extensiondamto main
gravity
Powerhouse No. I
Substoae U
Substag*
Powerhouse Na 2
Final Stage
Ridam
ght (Final
embankmentStage)
Trestle
REFERENCES
American Society of Civil Engineer (ASCE), Civil Engineering Guidelines for Planning and
Designing Hydroelectric Developments: Vol. 2 Waterways, American Society of Civil Engineers,
New York, 1989.
American Society of Civil Engineer (ASCE), Steel Penstock, ASCE Manuals and Reports on
Engineering Practice No. 79, American Society of Civil Engineers, New York, 1993.
Brater, E. R, King,
H. W., J. E. Lindell, and C. Y. Wei, Handbook of Hydraulics, 7th ed., McGraw-Hill, New York,
1996.
Chaudhry, M. H., Applied Hydraulic Transients, 2nd ed., Van Nostrand Reinhold, New York, 1987.
Chow, V. T., Open-Channel Hydraulics, McGraw-Hill, New York, 1959.
Dingman, S. L., Fluvial Hydrology, W. H. Freeman, New York, 1984.
Gulliver, J. S., and R. E. A. Arndt, Hydropower Engineering Handbook, McGraw-Hill, New York,
1991.
Henderson, F. M., Open Channel Flow, Macmillan, New York, 1966.
Moffat, A. I. B., C. Nalluri, and R. Narayanan, Hydraulic Structures, Unwin Hyman, London, UK,
1990.
Parmakian, J., Waterhammer Analysis, Dover Publications, New York, 1955.
Rich, G. R., Hydraulic Transients, Dover Publications, New York, 1951.
U. S. Army Corps of Engineer (USAGE), Hydraulic Design Criteria, U.S. Army Corps of
Engineers, Waterways Experiment Station, Vicksburg, MS, 1988.
U.S. Bureau of Reclamation, Selecting Hydraulic Reaction Turbines, Engineering Monograph
No.20, Department of the Interior, 1980.
U. S. Bureau of Reclamation, Design of Small Dams, U.S. Department of the Interior, Denver, Co,
1987.
U. S. Bureau of Reclamation Welded Steel Penstocks, Engineering Monograph No.3, U.S.
Department of the Interior, Denver, Co, 1967.
Vanoni, V. A., ed., Sedimentation Engineering, American Society of Civil Engineers, New York
1977.
Warnick, C. C., H. A. Mayo Jr., J. L. Carson, and L. H. Sheldon, Hydropower Engineering,
Prentice-Hall, NJ, 1984.
Wylie, E. B., and V. L. Streeter, Fluid Transients in Systems, Prentice-Hall, Englewood Cliffs, NJ,
1993.
Zipparro, V. J., and H. Hasen, Davis' Handbook of Applied Hydraulics, 4th ed., McGraw-Hill, New
York, 1993.
CHAPTER 9
HYDRAULICS OF WATER
DISTRIBUTION SYSTEMS
Kevin Lansey
Department of Civil Engineering and Engineering Mechanics
University of Arizona
Tucson, Arizona
Larry W. Mays
Department of Civil and Enviromental Engineering
Arizona State University
Tempe, Arizona
9.1
INTRODUCTION
In developed countries, water service is generally assumed to be reliable and utility customers expect high-quality service. Design and operation of water systems require an
understanding of the flow in complex systems and the associated energy losses. This
chapter builds on the fundamental flow relationships described in Chap. 2 by applying
them to water distribution systems. Flow in series and parallel pipes is presented first and
is followed by the analysis of pipe networks containing multiple loops. Water-quality
modeling is also presented. Because solving the flow equations by hand for systems
beyond a simple network is not practical, computer models are used. Application of these
models is also discussed.
9.1.1 Configuration and Components of Water Distribution Systems
A water distribution system consists of three major components: pumps, distribution storage, and distribution piping network. Most systems require pumps to supply lift to overcome elevation differences and energy losses due to friction. Pump selection and analysis
is presented in Chap. 10. Storage tanks are included in systems for emergency supply or
for balancing storage to reduce energy costs. Pipes may contain flow-control devices,
such as regulating or pressure-reducing valves. A schematic of a distribution system is
shown in Fig. 9.1.
The purpose of a distribution system is to supply the system's users with the amount
of water demanded under adequate pressure for various loading conditions. A loading
condition is a spatial pattern of demands that defines the users' flow requirements. The
flow rate in individual pipes results from the loading condition and is one variable that
describes the networks hydraulic condition. The piezometric and pressure heads are other
descriptive variables. The piezometric or hydraulic head is the surface of the hydraulic
grade line or the pressure head (p/y) plus the elevation head (z):
Tank 2
Reservoir 1
Pump 9
FIGURE 9.1 Network schematic (from EPANET User's Menual, Rossman, 1994)
h = ?-+z
(9.1)
Because the velocity is relatively small compared to the pressure in these systems, the
velocity head typically is neglected. Heads are usually computed at junction nodes.
A junction node is a connection of two or more pipes or a withdrawal point from the
network. A. fixed-grade node (FGN) is a node for which the total energy is known, such
as a tank.
The loading condition may remain constant or vary over time. A distribution system is in
steady state when a constant loading condition is applied and the system state (the flow in all
pipes and pressure head at all nodes) does not vary in time. Unsteady conditions, on the other
hand, are more common and hold when the system's state varies with time. Extended-period
simulation (EPS) considers time variation in tank elevation conditions or demands in discrete
time periods. However, within each time period, the flow within the network is assumed to
be in steady state. The only variables in the network that are carried between time steps of an
EPS are the tank conditions that are updated by a conservation of mass relationship.
Dynamic modeling refers to unsteady flow conditions that may vary at a point and
between points from instant to instant. Transient analysis is used to evaluate rapidly varying changes in flow, such as a fast valve closure or switching on a pump. Gradually varied
conditions assume that a pipe is rigid and that changes in flow occur instantaneously along
a pipe so that the velocity along a pipe is uniform but may change in time. Steady, extended period simulation, and gradually temporally varied conditions are discussed in this
chapter. Transient analysis is described in Chap. 12.
(9-2)
where Qm and <2out are the pipe flow rates into and out of the node and gext is the external
demand or supply.
Conservation of energy states that the difference in energy between two points is equal
to the frictional and minor losses and the energy added to the flow in components between
these points. An energy balance can be written for paths between the two end points of a
single pipe, between two FGNs through a series of pipes, valves and pumps, or around a
loop that begins and ends at the same point. In a general form for any path,
I hu + S hpj =A
(9-3)
%
where hLi is the head loss across component i along the path, hp . is the head added by
pump J9 and A is the difference in energy between the end points of the path.
Signs are applied to each term in Eq. (9.3) to account for the direction of flow. A common convention is to determine flow directions relative to moving clockwise around the
loop. A pipe or another element of energy loss with flow in the clockwise direction would
be positive in Eq. (9.3), and flows in the counterclockwise direction are given a negative
sign. A pump with flow in the clockwise direction would have a negative sign in Eq. (9.3),
whereas counterclockwise flow in a pump would be given a positive sign. See the Hardy
Cross method in Sec. 9.2.3.1 for an example.
(9.4)
where K is a pipe coefficient that depends on the pipes diameter, length, and material and
n is an exponent in the range of 2. K is a constant in turbulent flow that is commonly
assumed to occur in distribution systems.
In addition to pipes, general distribution systems can contain pumps, control valves,
and regulating valves. Pumps add head hp to flow. As shown in Fig. 9.2, the amount of
pump head decreases with increasing discharge. Common equations for approximating a
pump curve are
hp=AQ> + BQ + hc
(9.5)
Pump head
Pump curve
Horsepower
curve
Flow rate
FIGURE 9.2 Typical pump curve
or
hp = hc- CQ-
(9.6)
where A, B, C, and m are coefficients and hc is the maximum or cutoff head. A pump
curve can also be approximated by the pump horsepower relationship (Fig. 9.2) of
the form
H-&
where Hp is the pump's water horsepower. Further details about pumps and pump selection are discussed in Chap. 10.
Valves and other fittings also appear within pipe networks. Most often, the head loss
in these components is related to the square of the velocity by
*= *%=K^
(9 8)
where hm is the head loss, and Kv is an empirical coefficient. Table 2.2 lists Kv values for
a number of appurtenances.
Pressure-regulating valves (PRVs) are included in many pipe systems to avoid
excessive pressure in networks covering varying topography or to isolate pressure
zones for reliability and maintaining pressures. Pressure regulators maintain a constant pressure at the downstream side of the valve by throttling flow. Mathematical
representation of PRVs may be discontinuous, given that no flow can pass under certain conditions.
Pipel
Pipe 2
Pipe 3
Pipel
Pipe 2
Pipe 3
Problem. For the three pipes in series in Fig. 9.3, (1) find the equivalent pipe coefficient, (2) calculate the discharge in the pipes if the total head loss is 10 ft, and
(3) determine the piezometric head at points b, c, and d if the total energy at the inlet
(pt. a) is 95 ft?
Solution. For English units, the K coefficient for the Hazen-Williams equation is
K
=c^^
<9-12)
where 0 is a unit constant equal to 4.73, L and D are in feet, and C is the Hazen-Williams
coefficient. Substituting the appropriate values gives K1 = 0.229, K2 = 0.970, and K3 =
2.085. The equivalent Ke is the sum of the individual pipes (Eq. 9.11), or K6 = 3.284.
Using the equivalent loss coefficient, the flow rate can be found by Eq. 9.10, or hL =
K6Q1*5. For hL equals 10 feet and K6 equals 3.284, the discharge is 1.83 cfs.
This relationship and K6 can be used for any flow rate and head loss. Thus, if the flow
rate was 2.2 cfs, the head loss by Eq. 9.10 would be hL = 3.284 (2.2)1-85 = 14.1 ft.
The energy at a point in the series pipes can be determined by using a path head-loss
equation of the form of Eq. (9.3). The total energy at Point, b is the total energy at the
source minus the head loss in the first pipe segment, or
H
Similarly, the head losses in the second and third pipes are 2.97 and 6.38 ft., respectively. Thus, the energy at c and d are 91.33 and 84.95 ft, respectively.
Two or more parallel pipes (Fig. 9.3B) can also be reduced to an equivalent pipe with
a similar K6. If the pipes are not identical in size, material, and length, the flow through
each will be different. The energy loss in each pipe, however, must be the same because
they have common end points, or
hA~hB = hL>l = ^2 = hL.
(9.13)
Since flow must be conserved, the flow rate in the upstream and downstream pipes
must be equal to the sum of the flow in the parallel pipes, or
G = d + G2 + - = S Qm
(9-14>
meMp
where pipe m is in the set of parallel pipes, Mp. Manipulating the flow equation (Eq. 9.4),
the flow in an individual pipe can be written in terms of the discharge by Q = (hL/K)1/n
Substituting this in Eq. (9.14) gives
(Hr 1 V1 (hj O VS (hr -I VS
+
+
+
2 HIT
hr
hr
(Kl)
(K2j
(Kl) -
<9-15)
As is noted in Eq. (9.13), the head loss in each parallel pipe is the same. If the same n is
assumed for all pipes, Eq. (9.15) can be simplified to
\( \ \/n ( 1 Vn ( 1 Vn
1
^r-, ( 1 Vn
( 1 Vn
+
+
o-Hfe)
(il nil
-H" sji) -"-(T] "">
^, ( 1 \ln
( 1 \ln
=
n)
SJsJ ra
Because the K values are known for each pipe based on their physical properties,
K6 can be computed, then substituted in Eq. (9.10) to determine the head loss across the
parallel pipes, given the flow in the main pipe.
Problem. Determine the head loss between points A and B for the three parallel pipes.
The total system flow is 0.2 m3/s. Also find the flow in each pipe.
Solution. The head-loss coefficient K for each pipe is computed by Eq. (9.12), with </>
equal to 10.66 for SI units and L and D in meters, or K1 = 218.3, K2 = 27.9, and K3 =
80.1. The equivalent Ke is found from Eq. 9.17:
( 1 ^T85
( 1 ^T85
( 1 ^L85
( 1 "p5
+
+
=
=a313
K
K
hr
hr
hr
hr
1*1 J
( 2j
( 3j
(Ke)
or Ke = 8.58. By Eq. (9.10), the head loss is hL = K6Q^5 = 8.58*(0.2y** = 0.437 m.
The flow in each pipe can be computed using the individual pipe's flow equation and
K. For example, Q1 = (V^i)17185 = (0.437/218.3)1/185 = 0.035 mVs. Similarly, Q2 and Q3
are 0.105 and 0.060 m3/s, respectively. Note that the sum of the flows is 0.2 m3/s, which
satisfies conservation of mass.
9.2.2
The third basic pipe configuration consists of branched pipes connected at a single junction node. As shown in Fig. 9.4, a common layout is three branching pipes. Under steady
conditions, the governing relationship for this system is conservation of mass applied at
the junction. Since no water is stored in the pipes, the flow at the junction must balance
Reservoir 1
H = 100 m
Reservoir 2
H - 60m
Pipel
Pipe 2
Junction
with
pressure, P
Pipe 3
Reservoir 3
H = 40m
G1 + Q2 ~ G3 = O
(9.18)
where the sign on the terms will come from the direction of flow to or from the node. In
addition to satisfying continuity at the junction, the total head at the junction is unique.
Given all the pipe characteristics for each system in Fig. 9.4, the seven possible
unknowns are the total energy at each source (3), the pipe flows (3), and the junction
node's total head P(I). Four equations relating these variables are available: conservation
of mass (Eq. 9.18) and the three energy loss equations. Thus, three of the seven variables
must be known. Two general problems can be posed.
First, if a source energy, the flow from that source and one other flow or source energy is known, all other unknowns can be solved directly. For example, if the flow and
source head for reservoir 1 and pipe 1 are known, the pipe flow equation can be used to
find P by the following equation (when flow is toward the junction):
HA-? = hu = K1Qi"
(9-19)
If a flow is the final known (e.g., Q2), Q3 can be computed using Eq. (9.18). The source
energies can then be computed using the pipe flow equations for Pipes 2 and 3, in the form
of Eq. (9.19), with the computed P.
If the final known is a source head, the discharge in the connecting pipe can be computed using the pipe equation in the form of Eq. (9.19). The steps in the previous paragraph are then repeated for the last pipe.
In all other cases when P is unknown, all unknowns can be determined after P is
computed. P is found most easily by writing Eq. 9.18 in terms of the source heads.
From Eq. (9.19),
/n
= sign(Hsl-p{}-^^}
(9.20)
I Ki J
where a positive sign indicates flow to the node. Substituting Eq. (9.20) for each pipe in
Eq. (9.18) gives
Qj
(9.22)
where Ax is the change in x required to satisfy F(x). Truncating the expansion at the firstorder term gives a linear equation for Ax:
A*=--^
(9.23)
^/3x1,
The estimated x is updated by x = x + Ax. Since the higher order terms were dropped
from Eq. (9.22), the correction may not provide an exact solution to F(JC). So several iter-
ations may be necessary to move to the solution. Thus, if Ax is less than a defined criteria,
the solution has been found and the process ends. If not, the updated x is used in Eq. (9.23)
and another Ax is computed. In the three reservoir case, x=P and the required gradient
3F/3P is:
& = - L r p ^ i - ^ n ' - - 1 ' , P ^ 2 - P I i ' - " 1 ' , ri/^-pn'-- 1 '"!
+
+
3P
n[( K1 }
( K2 J
[ K3 }
\
= -(
\
+
1
+
I
InJT1IGiI11-1 ^2IG2I"-1 W^3IG3I--1 J
(924)}
^
F(P) is computed from Eq. (9.21) using the present estimate of P. AP is then computed
using AP = -F(P)/(dF/dP), and P is updated by adding AP to the previous P. The iterations continue until AP is less than a defined value. The Newton-Raphson method also can
be used for multiple equations, such as the nodal equations (Sec. 9.2.3.3). A matrix is
formed of the derivatives of each equation and the update vector is calculated.
Problem. Determine the flow rates in each pipe for the three-pipe system shown in Fig.
9.4. The friction factors in the table below assume turbulent flow conditions through a
concrete pipe (e = 0.08 cm).
Solution. Using the Darcy-Weisbach equation (n = 2), the K coefficients are computed using
8/L
K_
^2D5g
Pipe
D (cm)
L(m)
/[]
K
Reservoir elevation, H (m)
1
80
1000
0.0195
4.9
100
2
40
600
0.0235
113.8
60
3
40
700
0.0235
132.7
40
Iteration 1
In addition to the three discharges, the energy at the junction P also is unknown. To
begin using the Newton-Raphson method, an initial estimate of P is assumed to be 80 m,
and Eq. (9.21) is evaluated as follows:
F(P = 80.) = L(IOO - 8O)(Ii^f) + L(60 - 8O)MfI
I
^ 4.y
) ype\ i
^ 113.8 ) ypei
(
^
+ (40 - 80)f l 4 ^~^ Q i f
= 2.020 - 0.419 - 0.549= 1.052 mVs
^
I 132.7 ) J,ipe3
which states that flow enters the node at more than 1.052 m3/s, than leaves through pipes
2 and 3 with P = 80 m. Therefore, P must be increased. The correction is computed by
Eq. (9.23) after computing 3F/3P using Eq. (9.24):
aF
(
i
i
3P = " |V2*4.9*I2.020K2-1) + 2*113.8*I-0.419K2-1)
i
^
2*132.7*I-0.549|(2-1)J-
F(P)
1.052
=~W7= -TT00679 = 15'5m3P/P
The P for the next iteration is then P = 80 + 15.5 = 95.5 m.
The following iterations give
AP
j_
The correction for the first iteration is then - (0.514/-0.080) = 6.42 m, and the new
P is 96.42 m. The next two iterations are
Iteration 2: F(P = 96.42 m) = -0.112; ^F/3p|P = 9642m = -0.127; AP = -0.88 m,
P = 95.54 m
Iteration 3: F(P = 95.54 m) = -0.006; ^/3P P = 9554. = -0.115; AP = -0.05 m,
P = 95.49 m
To determine Hs3, the pipe flow equation (Eq. 9.20) is used with the known discharge, or
(\Ha - 95.49Iy2
Q3= -0.4 = sign(Ha - 95.49)1
^ J = HA = 74.26 m
9.2.3
Pipe Networks
A hydraulic model is useful for examining the impact of design and operation decisions.
Simple systems, such as those discussed in Sees. 9.2.1 and 9.2.2, can be solved using a
hand calculator. However, more complex systems, require more effort even for steady
state conditions, but, as in simple systems, the flow and pressure-head distribution
through a water distribution system must satisfy the laws of conservation of mass and
energy (Eqs. 9.2 and 9.3). These relationships have been written in different ways to solve
for different sets of unknowns.
Using the energy loss-gain relationships for the different components, the conservation
equations can be written in three forms: the node, loop, and pipe equations. All are nonlinear and require iterative solution schemes. The form of the equations and their common
solution methods are described in the next four sections. Programs that implement these
solutions are known as network solvers or simulators and are discussed in Sec. 9.5.
9.2.3.1 Hardy Cross method. The Hardy Cross method was developed in 1936 by Cross
before the advent of computers. Therefore, the method is amenable to solution by hand
but, as a result, is not computationally efficient for large systems. Essentially, the method
is an application of Newton's method to the loop equations.
Loop equations. The loop equations express conservation of mass and energy in
terms of the pipe flows. Mass must be conserved at a node, as discussed in Sec. 9.2.2 for
branched pipes. For all Nj junction nodes in a network, it can be written as
2 Qi =
(9-25>
Uj
Conservation of energy (Eq. 9.3) can be written for closed loops that begin and end at
the same point (AE = O) and include pipes and pumps as
E KG" ~ S CAfrC*. + BtPQiP + Clp) = O
(9.26)
ie/L
ip*Jp
This relationship is written for N1 independent closed loops. Because loops can be nested in the system, the smallest loops, known as primary loops, are identified, and each pipe
may appear twice in the set of loops at most. The network in Fig. 9.1 contains 3 primary
loops.
Energy also must be conserved between points of known energy (fixed-grade nodes). If
NfFGNs appear in a network, A^-I independent equations can be written in the form of
E K& - S (**<% + BiPQiP + <V = AEFGN
(9-27)
ieIL
ip*Ip
where A"FGN is the difference in energy between the two FGNs. This set of equations is
solved by the Hardy Cross method (Cross, 1936) by successive corrections to the pipe
flows in loops and by the linear theory method by solving for the pipe flows directly (Sec.
9.2.3.2).
Solution method. To begin the Hardy Cross method, a set of pipe flows is assumed
that satisfies conservation of mass at each node. At each step of the process, a correction
AQL is determined for each loop. The corrections are developed so that they maintain conservation of mass (Eq. 9.25), given the initial set of flows. Since continuity will be preserved, those relationships are not included in the next steps.
The method then focuses on determining pipe flows that satisfy conservation of energy. When the initial flows are substituted in Eqs. (9.26) and (9.27), the equations are not
X dAipQipAQL + BlpAQL) = AE
ipeip
(9.29)
Given Qik the flow estimates at iteration k, Eq. (9.29) can be solved for the correction
for loop L as
AO = - &K>Ql- S C-V&* + BQ,* + cip} - AE)
L
1
^
'^
(9.30)
In this form, the numerator of Eq. (9.30) is the excess head loss in the loop and should
equal zero by conservation of energy. The terms are summed to account for the flow direction and component. The denominator is summed arithmetically without concern for
direction. Most texts present networks with only closed loops and no pumps. Equation
(9.30) simplifies to this case by dropping the pump terms and setting AE to zero, or
-S K0U
-2 hu
WL = -TT
=
!,*,<&-'
n^hu/Qa
C
' L
/e/^
-F(ft)
=~aW
'*Q u
(9-31>
Comparing Eq. (9.31) with Eq. (9.23) shows that the Hardy Cross correction is essentially a Newton's method.
The AQ1 corrections can be computed for each loop in sequence and can be applied
before moving to the next loop (Jeppson, 1974) or corrections for all loops can be determined and applied simultaneously. Once the correction has been computed, the estimates
for the next iteration are computed by
Q*+i = Qi* + ^QL
(9-32)
Qk+l is then used in the next iteration. The process of determining corrections and
updating flows continues until the AQL for each loop is less than some defined value. After
the flows are computed, to determine the nodal heads, head losses or gains are computed
along a path from fixed-grade nodes to junction nodes.
The Hardy Cross method provides an understanding of principles and a tool for solving small networks by hand. However, it is not efficient for large networks compared with
algorithms presented in the following sections.
Problem. List the loop equations for the network shown in Fig. 9.5 using the direction of
flow shown. Then determine the flow in each pipe and the total energy at Nodes 4 and 5.
Solution The loop equations consist of conservation of mass at the five junction nodes
and the loop equations for the two primary loops and one pseudo-loop. In the mass balance equations, inflow to a node is positive and outflow is negative.
Node 1.
Node 2.
NodeS.
Node 4.
Node 5.
Loop I.
Loop II.
Pseudo-loop.
Q1 - Q2 - Q5 = O
Q2 + Q3 - Q6 = 2
-Q3 + Q4 -Q7 = O
Q5 - G8 = 1
Q6 + Q1 + Q, = 2
\2 + hL 6 - hLi8 - hL>5 = O = K2Qi + K6Ql - K8Qi - K5Qi
hLJ - hL>6 - hL 3 = O = K1Qt - K6Ql - K3QI
h L4 - hp + h L3 - h L2 - hLl - EFGN2 + E FGN , = O
= K4Ql - (ApGJ + BpQ4 + Cp + K3QI -K2Ql - K1Gf
~ EpGN,2 + EFGNJ
Because the Darcy-Weisbach equation is used, n equals 2. The loop equations assume
that flow in the clockwise direction is positive. Flow in Pipe 5 is moving counterclockwise and is given a negative sign for loop I. Flow in pipe 6 is moving clockwise relative
to loop I (positive sign) and counter clockwise relative to Loop II (negative sign).
Although flow is moving clockwise through the pump in the pseudo-loop, hp is given a
negative sign because it adds energy to flow. To satisfy conservation of mass, the initial
set of flows given below is assumed, where the values of K for the Darcy-Weisbach equation are given by
K
KDW
- ^ - 8^
~ TtDg ~ VD^;
33)
(933)
factor of 0.032 for turbulent
6
3.222
0.5
7
4.028
1.0
8
2.417
0.5
FIGURE 9.5 Example network (Note all pipes have diameters of 1 ft and friction factors equal to 0.032).
The denominator is
TiK4Q4 + 2ApQ4 + nK3Q3 + nK2Q2 + TiK1Q1
= 2 (1.611(2.5))+I2(-6)2.5I + 2(2.417(1.5)) + 2 |2.417(1.0) +2 1.611(2.5) = 58.20.
Thus, the correction for the pseudo-loop AQPL is
*.--%->
The correction for Loop I is computed next. The numerator of Eq. (9.30) is
K2Ql + K6Gi - K8Gi - K5Q52 =
2.417(1.O) + 3.222(0.5)2-2.417(0.5)2-3.222(1.5)2 = -4.63
2
1
-0.077
2
3 4 a n d pump
5
6
7
8
0.230 0.0770.077
-0.230 0.2300.120
-0.230
-0.077 0.120
0.120
2.42 1.15 1.46 2.58 1.27 0.61 1.12
0.27
Because the flow direction for Pipe 1 is counterclockwise relative to the pseudo-loop,
the correction is given a negative sign. Similarly, Pipe 2 receives a negative correction for
the pseudo-loop. Pipe 2 is also in Loop I and is adjusted with a positive correction for that
loop since flow in the pipe is in the clockwise direction for Loop I. Pipes 3 and 6 also
appear in two loops and receive two corrections.
~899878) = 0.052
~ Solf1 = -065
1
-0.030
2.39
2
3 4 a n d pump
5
0.052+ 0.0300.030
-0.052
(-0.030) 0.065
1.17
1.43
2.61
1.22
6
0.0520.065
0.60
7
0.065
8
-0.052
1.18 0.22
Iteration 3. The corrections for iteration 3 are 0.012, 0.024, and 0.021 for the pseudo-loop, loop I and loop II, respectively. The resulting flows are as follows:
Pipe
AQ
Q
1
-0.012
2.38
2
0.024+
(-0.012)
1.18
3 4 a n d pump
5
0.012- 0.012 -0.024
0.024
1.42
2.62
1.20
6
7
8
0.024- 0.024 -0.024
0.024
0.60
1.20 0.20
After two more iterations, the changes become small, and the resulting pipe flows are
as follows. Note that the nodal mas balance equations are satisfied at each iteration.
Pipe
Q
1
2.37
2
1.19
3
4 and pump
1.41
2.63
5
1.18
6
0.60
7
8
1.21 0.18
The total energy at Nodes 4 and 5 can be computed by path equations from either FGN
to the nodes. For example, paths to Node 4 consist of Pipes 1 and 5 or of pipes 4 (with the
pump), 7, and 8. For the path with pipes 1 and 5, the equation is
100 - K1Q12- K5Q52= 100 - 1.611(2.37)2 - 3.222(1.19)2 = 100 - 9.05 - 4.56 - 86.39m
For the path containing pipes 4, 7 and 8 the result is
10 + (135 - 6(2.63)2) - 1.611(2.63)2 - 4.028(1.22)2 + 2.417(0.19)2 =
10 + 93.50 - 11.14 - 6.00 + 0.09 = 86.45m
This difference can be attributed to rounding errors. Note that pipe 8 received a positive sign in the second path equation. Because the flow in Pipe 8 is the opposite of the path
direction, the energy along the path is increasing from Nodes 5 to 4. The total energy at
Node 5 can be found along pipes 4 and 7 or 86.36 m or along the path of Pipes 1-2-6, giving (100 - 9.05 - 3.42 - 1.16 = 86.37m).
9.2.3.2 Linear theory method. Linear theory solves the loop equations or Q equations
(Eqs. 9.25 to 9.27). Np equations (Nj + N1 + Nf -1) can be written in terms of the Np
unknown pipe flows. Since these equations are nonlinear in terms of Q, an iterative procedure is applied to solve for the flows. Linear theory, as described in Wood and Charles
(1972), linearizes the energy equations (Eqs. 9.26 and 9.27) about Qi>k+1, where the subscript k+\ denotes the current iteration number using the previous iterations Q1k as known
values. Considering only pipes in this derivation, these equations are
and
(9.34)
^
K1QIk1 Qi,k+i = fr al1 N\ dosed loPs
e/
<L
(9-35)
next iterations. Once the pipe flows have been determined, the nodal piezometric heads
can be determined by following a path from a FGN and accounting for losses or gains
to all nodes.
Modified linear theory Newton method. Wood (1980) and his collaborators at the
University of Kentucky developed the KYPIPE program but essentially modified the
original linear theory to a Newton's method. However, rather than solve for the change in
discharge (AQ), Qk+l *s determined.
To form the equations, the energy equations (Eq. 9.3) are written in terms of the
current estimate of Qk, including pipes, minor losses and pumps, as
K
/(ft) - E
Qi+ Ee/ Kim Ql + E < A *G? + B*& + CP - AE
(9-37>
ie/
L
<m m
ip^p
where for simplicity the subscripts /, /m, and ip denoting the pipe, minor loss component,
and pump, respectively, are dropped from the flow terms and k again denotes the iteration
counter. This equation applies to both closed loops (AE = O) and pseudo-loops (AE =
A"FGN), but, in either case, f(Qk) should equal zero at the correct solution.
To move toward the solution, the equations are linearized using a truncated Taylor
series expansion:
f(Qk+l) =f(Qk)
ft)
(9-38)
Note that/and Q are now vectors of the energy equations and pipe flow rates, respectively, and Gk is the matrix of gradients that are evaluated at Qk. Setting Eq. (9.38) to zero
and solving for Qk+l gives
or
0=f(Qk) + Gk(Qk+l-Qk)
GkQk+v = GkQk+f(Qk)
(9.39)
This set of (N1 + Nf-l) equations can be combined with the Adjunction equations in
Eq. (9.34) that also are written in terms of Qk+l to form a set of Np equations. This set of
linear equations is solved for the vector of Np flow rates using a matrix procedure. The
values of Qk+1 are compared with those from the previous iteration. If the largest absolute
difference is below a defined tolerance, the process stops. If not, Eq. (9.39) is formed
using Qk+l and another iteration is completed.
9.2.3.3 Newton-Raphson method and the node equations. The node equations are the
conservation of mass relationships written in terms of the unknown nodal piezometric
heads. This formulation was described in Sec. 9.2.2 for branching pipe system. In Fig. 9.4,
if P and the pipe flows are unknown, the system is essentially a network with one junction node with three FGNs. In a general network, Nj junction equations can be written in
terms of the Nj nodal piezometric heads. Once the heads are known, the pipe flows can be
computed from the pipe's head-loss equation.
Other network components, such as valves and pumps, are included by adding junction nodes at each end of the component. Node equations are then written using the flow
relationship for the component.
Solution method. Martin and Peters (1963) were the first to publish an algorithm
using the Newton-Raphson method for solving the node equations for a pipe network.
Shamir and Howard (1968) showed that pumps and valves could be incorporated and
unknowns other than nodal heads could be deternined by the method. Other articles have
been published that attempt to exploit the sparse matrix structure of this problem.
(9.42)
sign(hl - /J^f
( K2 )
1
***> - 4
( ^
K3 f)
2
***> - 4^f
( A6 ) =
^5 - jyf^f
( K7 J
-^v - 414
( V^f=
K
) o
Node 3:
Sign(h2
- />( 3fv^f
K
3 )
Node 4:
sign(hl - J^f
(. K5 J
Si8n(H5 - J^V^l"
= 1
( As ;
Node 5:
sign(/,2 - hf^f
( K6
2
1.19
3
4
1.41 2.63
5
1.18
6
7
8
Pump
0.60 1.22 0.18 2.63
9.2.3.4 Gradient algorithm Pipe equations. Unlike the node and loop equations, the
pipe equations are solved for Q and h simultaneously. Although this requires a larger set
of equations to be solved, the gradient algorithm by Todini and Pilati (1987) has been
shown to be robust to the extent that this method is used in EPANET (see Sec. 9.5.3).
To form the pipe equations, conservation of energy is written for each network component in the system in terms of the nodal heads. For example, a pipe equation is
ha-hb = KQn
(9.43)
(9.44)
where ha and hb are the nodal heads at the upstream and downstream ends of the component. These equations are combined with the nodal balance relationships (Eq. 9.2) to form
TV7 + Np equations with an equal number of unknowns (nodal heads and pipe flows).
Solution method. Although conservation of mass at a node is linear, the component
flow equations are nonlinear. Therefore, an iterative solution scheme, known as the gradient algorithm, is used. Here the component flow equations are linearized using the previous flow estimates Qk. For pipes,
KQi~l QM + (ha - A4) = O
(9.45)
(9.46)
AnQ - qm = O
(9.47)
and
where Eq. 9.46 is the linearized flow equations for each network component and Eq. 9.47
is the nodal flow balance equations. A12 (= A217) is the incidence matrix of zeros and ones
that identify the nodes connected to a particular component and A10 identifies the
fixed grade nodes. A11 is a diagonal matrix containing the linearization coefficients
(e.g.,\KQs-i DDifferentiating eqs. 9.46 and 9.47 gives:
[JVA11 A 12 IFdBl JdEl
[A21 O J U J [dq\
*'->
where dE and dq are the residuals of equations 9.2 and 9.43-44 evaluated at the present solution, Qk and hk. N is a diagonal matrix of the exponents of the pipe equation
(n). Eq. 9.48 is a set of linear equations in terms of dQ and dh. Once solved Q and h
are updated by
and
G ft+ i = Qk + dQ
(9.49)
hk+l = hk + dh
(9.50)
(9.51)
(9.52)
where A11 is computed at Qk. Note that N and A11 are diagonal matrices so the effort for
inversion is negligible. Yet, one full matrix must be inverted in this scheme.
Problem. Write the pipe equations for the network in Fig. 9.5.
Solution. The pipe equations include mass balance equations for each node in the system. The network contains five junction nodes plus an additional node downstream of the
pump. The pump is considered to be a link and is assumed to be located directly after the
FGN.
Conservation of energy equations are written for each pipe and pump link. Eight pipe
equations and one pump equation are written. The total number of equations is then 15,
which equals the 15 unknowns, including 8 pipe flows, 1 pump flow, and 6 junction node
heads, including the additional nodal head at the pump outlet hp.
Node 1:
Q1 - Q2 - Q5 = O
Node 2:
Node 3:
Node 4:
Node 5:
Pump Node:
Pump:
fcp
Q2 + Q3 - Q6 = 2
- Q3 + Q4 - Q1 = O
5 - G8 = 1
S6 + C7 + G8 = 2
Gp - G4 = O
- 10 = 135 - 6Q2p
Pipe 2:
Pipe 3:
Pipe 4:
Pipe 5:
Pipe 6:
Pipe 7:
Pipe 8:
H 1 - H 2 = K2Qi
H 3 - H 2 = K3Q]
hp-H3 = K4Q4
H 1 - H 4 = K5Qn5
H 2 - H 5 = K6Ql
H 3 - H 5 = K1Qf1
/I4 - /I5 = K,Ql
9.2.3.5 Comparison of solution methods. All four methods are capable of solving the
flow relationships in a system. The loop equations solved by the Hardy Cross method are
inefficient compared with the other methods and are dropped from further discussion. The
Newton-Raphson method is capable of solving all four formulations, but because the node
equations result in the fewest equations, they are likely to take the least amount of per iteration. In applications to the node equations, however, possible convergence problems may
result if poor initial conditions are selected (Jeppson, 1974).
Linear theory is reportedly best for the loop equations and should not be used for the
node or loop equations with the AG corrections, as used in Hardy Cross (Jeppson, 1974).
Linear theory does not require initialization of flows and, according to Wood and Charles
(1972), always converges quickly.
A comparative study of the Newton-Raphson method and the linear theory methods
was reported by Holloway (1985). The Newton-Raphson scheme was programmed in two
codes and compared with KYPIPE that implemented the linear theory. For a 200-pipe network, the three methods converged in eight or nine iterations, with the Newton-Raphson
method requiring the least amount of computation time.
Salgado, Todini, and O'Connell (1987) compared the three methods for simulating
a network under different levels of demand and different system configurations. Four
conditions were analyzed and are summarized in Table 9.1. Example A contains 66
pipes and 41 nodes but no pumps. Example B is similar to Example A, but 6 pumps are
introduced and a branched connection has been added. Example C is the same network
as in Example B with higher consumptions, whereas Example D has the same network
layout but the valves are closed in two pipes. Closing these pipes breaks the network
into two systems. The results demonstrate that all methods can simulate the conditions,
but the gradient method for solving the pipe equations worked best for the conditions
analyzed.
All comparisons and applications in this chapter are made on the basis of assuming
reasonably sized networks. Given the speed and memory available in desktop computers, it is likely that any method is acceptable for these networks. To solve extremely
large systems with several thousand pipes, alternative or tailored methods are necessary. Discussion of these approaches is beyond the scope of this chapter. However,
numerical simulation of these systems will become possible, as discussed in Sec. 9.5
on network calibration, but good representation of the system with accurate parameters may be difficult.
9.2.3.6 Extended-period simulation. As noted earlier, time variation can be considered
in network modeling. The simplest approach is extended-period simulation, in which a
sequence of steady-state simulations are solved using one of the methods described earlier in this section. After each simulation period, the tank levels are updated and demand
and operational changes are introduced.
Special conditions
Low velocities
Closed pipes
Node equations
Solution method:
Loop equations
Converged
Iterations = 16,
T = 70 s
Converged
Iterations =12,
T = 92 s
Converged
Iterations =13,
T=IOOs
Converged
Iterations = 21,
T = 155 s
Some heads not
available
Converged
Iterations =17,
T = 789 s
Slow convergence
Iterations =13,
T = 962 s
Slow convergence
Iterations =15,
T=IIlOs
Converged
Iterations = 21,
T = 1552 s
Some heads not
available
Pipe equations
Converged
Iterations =16,
T = 30 s
Converged
Iterations =10,
T = 34 s
Converged
Iterations =12,
T = 39 s
Converged
Iterations =19,
T = 57 s
Tank levels or water-surface elevations are used as known energy nodes. The levels
change as flow enters or leaves the tank. The change in water height for tanks with constant geometry is the change in volume divided by the area of the tank, or
J. = QI&
T
AT
AT
where QT and VT are the flow rate and volume of flow that entered the tank during the period, respectively; A? is the time increment of the simulation; AT is the tank area; and AHr
is the change in elevation of the water surface during period T. More complex relationships are needed for noncylindrical tanks. With the updated tank levels, the extended-period simulation continues with these levels as known energy nodes for the next time step.
The process continues until all time steps are evaluated. More complex unsteady analysis
are described in the next section.
9.3
In steady state analysis or within an extended-period simulation, changes in the distributions of pressure and flow are assumed to occur instantaneously after a change in external
stimulus is applied. Steady conditions are then reached immediately. In some cases, the
time to reach steady state and the changes during this transition may be important.
Recently, work has proceeded to model rapid and gradual changes in flow conditions.
Rapid changes resulting in transients under elastic column theory are discussed in Chap.
12. Two modeling approaches for gradually varied unsteady flow under a rigid-column
assumption are described in this section.
(9.54)
where F1 and F2 are the forces on the ends of the pipe element, Ff is the force caused by
friction between the water and the pipe, and m and v are the mass and velocity of the fluid
in the pipe element.
The end forces are equal to the force of the pressure plus the equivalent force caused
by gravity or for the left-hand side of the element:
F1 = jA^+z] = JAh1
(9.55)
The friction force is the energy loss times the volume of fluid, or
Ff=jAhL.
(9.56)
JA(H1 - H 2 - H 1 ) = ^-^
(9.58)
Assuming that a steady state friction loss relationship can be substituted for hL and dividing each side by ^A,
*.-*,-*? =
(9-59)
With conservation of mass (Eq. 9.2), this ordinary differential equation and its extensions for loops have been used to solve for time-varying flow conditions.
9.3.2 Solution Methods
9.3.2.1 Loop formulation. Holloway (1985) and Islam and Chaudhry (1998) extended
the momentum equation (Eq. 9.59) to loops as follows:
2 (*-*>-S ^=S ^f1
i/p
ie/p
ielp $ '
Separating variables and integrating over time gives
P+^IV.
1 P+AT^
1
f^+Ar^ L.
1 [s (*-**>}*-i g/4H
|^de'
(9-60)
(9 6i)
At any instant in time, the head loss around a closed loop must equal zero, so the first
term can be dropped. Dropping this term also eliminates the nodal piezometric heads as
unknowns and leaves only the pipe flows.
One of several approximations for the friction loss term can be used:
KQ+"\Q\*-lbt
(9.62)
(9.63)
(9.64)
Holloway (1985) obtained results using Eq. (9.62), known as the integration approximation that compared favorably with the other two nonlinear forms. Using this form in
Eq. (9.61),
(9 65)
^-^T
&-^, K&+"\Q\*-i Ar = X ^T
W
'
8 l
A
MP
Mp
Mp 8 i
This equation is written for each loop and is used with the nodal conservation of mass
equations to given Np equations for the Np unknown pipe flows. Note that these equations
are linear in terms of Qt+At and can be solved at each time step in sequence using the previous time step for the values in the constant terms.
f -3;*^
The left-hand side of Eq. 9.9 is written in terms on the unknowns h and Q at time step
t + At. After substituting and rearranging a general algebraic equation for pipe between
two nodes results in
\Kt\Q^ I"-1 Ar - -^U <2rA'+ iTA'~ ^TA'] & = \ - -^r] Qi
(9-67)
L
&A'J
L A'J
Np equations of this form can be written for each pipe or other component. With the Nj
nodal flow balance equations, a total of Nj + Np equations can be written in terms of an
equal number of unknown pipe flows and nodal heads. Given an initial condition at time
t, the pipe flows and nodal heads at time t + At by solving Eq. 9.67 and Eq. 9.2 The new
values are then used for the next time step until all times have been evaluated. Unlike the
loop formulation, in the form above, Eq. 9.67 is nonlinear with respect to the unknowns.
In addition, like the loop equation, the time step will influence the accuracy of the results.
9.4
WATER-QUALITY MODELING
Interest in water quality in distribution systems heightened with the passing of the 1986
amendment to the Safe Drinking Water Act. This amendment required that standards must
be developed for chlorine levels not only at the point of disinfection but also at the most
distant point of withdrawal. Thus, modeling the fate and transport of dissolved substances
in networks with emphasis on chlorine became necessary. As a result, methods of analysis and computer programs implementing these methods, such as EPANET (Rossman,
1994), have been developed.
Since the velocity in pipes is relatively high, constituents in the water are assumed to
move completely with the flow, that is, by advective transport. This assumption allows the
use of explicit numerical modeling schemes to solve for constituent movement within the
system. As in hydraulic analysis, steady and unsteady transport models have been developed. Both models use conservation of mass as the basic governing equation describing
mixing and movement. Because advective transport dominates, the pipe flow rates are
critical in estimating transport in the system. In most unsteady water-quality models,
extended-period simulation has been used to account for demand and operational changes
(Sec. 9.2.3.6) Although water quality analysis considering slow transients using rigidwater-column theory for the flow analysis has been performed by Chaudhry and Islam
(1998), it will not be discussed here.
As water moves through the network, constituent (with emphasis on chlorine) decay is
generally assumed to follow first-order kinetics, or
ct = c0e-V
(9.68)
where CQ and ct are the constituent concentrations at times O and t, respectively, t is time
and kt is the first-order decay coefficient, which is defined by
"--[^ + ScWl
<*">
where RH is the hydraulic radius of the pipe, and kb, kw, and kf are the bulk flow-decay constant, the wall reaction rate constant, and a mass transfer coefficient that is dependent on
the Reynold's number, respectively.
(9.71)
One constituent mass balance equation can be written for each node. Since the flow
rates are defined by the hydraulic relationships, C8 is known, and the CI for one node is
the outflow from another node, the system of equations can be solved for the Af unknown
c;s.
A steady-state model provides the concentrations at all points in the network under
steady flow and concentrations. By modeling each source concentration independently
in a series of simulations, the model also can be used to determine the relative source
contribution at any point under the same conditions.
9.4.2 Dynamic Analysis
Steady flow conditions for water quality provide information regarding movement of dissolved substances but are likely to be less useful for predicting point concentrations under
normal operations. Unsteady analysis, also known as dynamic modeling, provides a more
realistic picture and better estimates of constituent movement under time-varying flow
conditions.
Dynamic modeling can solve several types of problems. In addition to determining the
variation in concentration at a point over time, it can be used to determine the age of or
average travel time for water at some location and time. Finally, as with steady models,
the relative source contributions providing flow to a point can be computed.
9.4.2.1 Governing equations. To determine the fate and transport of dissolved substances under unsteady conditions, the primary governing equation is the one-dimensional advection equation that is solved in conjunction with the assumption of complete mixing at a node. The advection equation is
=-i,+ *(C'.)
(9.72)
where C1 is the constituent concentration in pipe / at location x and time t\ U1 is the velocity in pipe z, and R(C1) is the reaction/decay function. The decay relationship for firstorder kinetics R = ktc is used when modeling chlorine and possibly other nonconservative substances. For conservative substances, such as fluoride, the reaction relationship
is simply zero. Finally, when modeling water age, R is equal to one and the concentration C is interpreted as the water age with new water entering the system having concentration equal to zero.
Tanks act as sources or sinks in the system with variable water quality, depending on
the history of inflow and outflow as well as on the reactions in the tank. The simplest
water-quality relationship for a tank assumes that the water is mixed completely. In this
case, the variation in constituent concentration is
^^
= 2 G1-CE1. - 2 QjC7 + R(C1)
01
ieIT
j*0T
(9.73)
where VT and CT are the storage volume and constituent concentration within the tank at
time r, respectively. Pipes in the set of 1T provide inflows Q1 to the tank, and pipes in the
set OT receive flows Q. from the tank. CE is the concentration at the exit of the pipe as it
enters the tank. R is the reaction relationship for water in the tank.
9.4.2.2 Solution methods Eulerian methods. Rossman and Boulos (1996) compared
the different solution methods for solving the unsteady water-quality problem. This section generally follows their notation and terminology. Dynamic models can be classified
spatially as Eulerian or Lagrangian models and temporally as time driven or event driven.
Eulerian methods define a grid of either points or volume segments within a pipe. Flow
and the associated constituents are tracked through this fixed grid. Chaudhry and Islam
(1998) used a finite-difference method with a fixed-point grid, and Grayman et al. (1988),
and its extension by Rossman et al. (1993), have developed the discrete-volume method
(DVM). The following discussion focuses on the DVM as it has been implemented in the
EPANET model (Rossman, 1994).
For a given hydraulic condition, the DVM divides each pipe into equally sized,
completely mixed, volume segments. The number of segments for a particular pipe is
computed by
n
'
= J^ = ^L
ut Ar
(9.74)
'
where L1 and u( are the length of and flow velocity in pipe z, respectively; tti is the travel time
for water to pass through pipe /; and Ar is the duration of the water quality time step. A small
Ar provides the highest numerical accuracy at the expense of higher computation times.
When the flow conditions change in the network (i.e., u changes), the grid must be redefined.
At each water-quality step, four operations are completed, as shown in Fig. 9.7. First,
the present constituent masses are reduced to account for the decay reactions. Next, the
elements from each segment are advanced to the next downstream segment. Third, if the
segment is the most downstream in a pipe, the flow is mixed with the flow from other
pipes that enter the node using Eq. (9.71). Finally, the flow from the node is passed to the
first segments of pipes leaving the node.
These operations are repeated for each water-quality time step until the flow distribution
changes. Pipes are then resegmented, and the process is repeated for that hydraulic condi-
Original mass
After reaction
FIGURE 9.7 Computational steps of discrete volume method (From Rossman and Boulos (1996))
tion. When the pipes are divided for different flow conditions, the number of segments may
be different and some numerical blending occurs. As a result, the accuracy of DVM and
finite difference methods depends on the selection of the water-quality time step Af.
Lagrangian methods. Unlike Eulerian methods which use a fixed grid, Lagrangian
methods track segments of water as they move through a network. As the front or leading
edge of the segment reaches a node, it is combined with other incoming segments. The
segments leaving the node are developed with constituent levels determined by Eq. (9.71)
(Fig. 9.8). Two approaches have been used to define when segments are combined and
transported through a pipe.
Liou and Kroon (1987) applied this type of model using a defined time step to
determine when to combine segments. During each time step, the total mass of constituent
and volume of water that reaches a node is computed. The average nodal concentration is
computed, and new segments emanating from the node are introduced. To avoid adding
too many new segments, they are created only when the concentration difference between
the new and the previous segment in a link is above a threshold. When more than one segment in a link reaches a downstream node in one time step, artificial mixing will occur.
Rather than combine segments at defined time intervals, the second Lagrangian
approach is an event-driven method (Boulos et al., 1994, 1995; Hart et al., 1987;
El-Shorbagy and Lansey, 1994; and Shah and Sinai, 1985). Event methods combine segments each time a front reaches a node, thus avoiding artificial mixing. Since defined
times are not used, the projected times when a front reaches a downstream node are corn-
FIGURE 9.8 Water quality transport for the Lagrangian methods for a conservative substance at
three different times. The flowrates in the two inflow pipes are equal and the flowrate in the outgoing pipe is then twice the flow in either inflow pipe. A: water quality at time t: flow is to the left,
and the constituent level equals the average of the inflow concentrations, or (4+l)/2 = 2.5 B: The
water quality at time t + At some time after the front concentration 2 in the vertical pipe reached
the node. For some time, the inflow concentrations were 2 and 4, or an average outflow concentration of 3, C: Water quality at some later time: Two elements have developed downtream. The element with a concentration of 3.5 developed when the inflows of 3 and 4 mixed at node. The final
element closest to the node with concentration 4 developed when the inflow with concentrations of
3 and 5 mixed at the node.
puted for the present flow condition. The water-quality conditions at nodes remain constant until the next segment front reaches a node. At that time, new segments are generated in pipes that carry flow from the node that the first front reaches. The concentration in
these segments is computed by Eq. (9.71) and is recorded with the transition time.
Projection times are then updated, and the process continues when the next closest front
reaches a node or the hydraulic condition changes. If the flow condition changes, new projection times are computed. Event-driven models avoid numerical dispersion; however,
the method can result in a large number of segments. To save computer memory, segments
can be combined according to the difference in concentration between adjacent segments.
Further error may result during flow reversals for reactive constituents.
Comparison of methods. Rossman and Boulos (1996) conducted numerical experiments comparing the alternative methods described in the previous sections, and reached
the following conclusions:
1. The numerical accuracy of all methods is similar, except that the Eulerian methods had
occasional problems. All methods can represent observed behavior adequately in real
systems.
2. Network size is not always an indicator of solution time and computer memory
requirements.
3. Lagrangian methods are more efficient in both time and memory requirements than
Eulerian methods when modeling chemical constituents.
4. The time-dependent Lagrangian method are most efficient in computation time for
modeling water age, whereas the Eulerian methods are the most memory efficient.
Overall, Rossman and Boulos concluded that the time-based Lagrangian method was
the most versatile unless computer memory was limiting for modeling water age for large
networks. In which case, Eulerian methods were preferable.
9.5 COMPUTER MODELING OF WATER
DISTRIBUTION SYSTEMS
Because the numerical approaches for analyzing distribution systems cannot be completed by hand except for the smallest systems, computer-simulation models have been developed. These models solve the system of nonlinear equations for the pipe flows and nodal
heads. In addition to the equation solver, many modeling packages have sophisticated
input preprocessors, which range from spreadsheets to tailored full-page editors, and output postprocessors, including links with computer-aided drafting software and geographic information systems. Although these user interfaces ease the use of the simulation models, a dependable solver and proper modeling are crucial for accurate mathematical models of field systems.
An array of packages is available, and the packages vary in their level of sophistication. The choice of a modeling package depends on the modeling effort. Modeling needs
range from designing subdivisions with fewer than 25 pipes to modeling large water utilities that possibly involve several thousand links and nodes. Users should select the package that best suit their objectives.
match the field observations can be found: that is, a unique solution may not be determined. Each set may give dramatically different results when predicting under other conditions.
During model calibration, field data are compared with model estimates and model
parameters are adjusted so that the model predictions match the field observations. Two
stages of model calibration are desirable. The first stage is a gross study of the data and
the model predictions. The intent is to insure that the data are reasonable and that major
modeling assumptions are valid. For example, this level would determine if valves
assumed to be open are actually closed or if an unexpectedly high withdrawal, possibly
caused by leakage, is occurring. Walski (1990) discussed this level of calibration.
[TITLE]
EPANET Example network 1
[JUNCTIONS]
Elevation
Demand
ID
ft
gpm
10
710
O
11
710
150
12
700
150
13
695
100
21
700
150
22
695
200
23
690
150
31
700
100
32
710
100
[CONTROLS]
Li^~9~6pEiiWN6DE2~Bl^wn6
LINK 9 CLOSED IF NODE 2 ABOVE 140
[PATTERNS]
ID Multipliers
1 1.0 1.2
1 1.0 0.8
1.4
0.6
[QUALITY]
[TANKS]
Elev. Init. Min. Max. Diam
ID
ft Level Level Level
ft
2
850 120 100 150 50.5
9
800
[PIPES]
ID
""To
11
12
21
22
31
110
111
112
113
121
122
Head
Node
To
11
12
21
22
31
2
11
12
13
21
22
Tail
Node
ii
12
13
22
23
32
12
21
22
23
31
32
Length Diam.
ft
in.
io53o""i~8
5280 14
5280 10
5280 10
5280 12
5280 6
200
18
5280 10
5280 12
5280 8
5280 8
5280 6
Rough
Coeff.
Too""
100
100
100
100
100
100
100
100
100
100
100
[PUMPS]
ID
9
Head
Node
9
Tail
Node
10
Design
ft
250
H-Q
gpm
1500
Nodes
2
9
2
Initial
Concen. mg/1
05
1.0
1.0
32
[REACTIONS]
GLOBAL BULK - .5
GLOBAL WALL -1
[TIMES[
DURATION 24
PATTERN TIME STEP
[OPTIONS]
QUALITY
MAP
[END]
FIGURE 9.9 EPANET input file for example network (Figure 9.1)
Chlorine
Net !.map
; Chlorine analysis
; Map coordinates file
EPANET
Hydraulic and Water Quality
Analysis for Pipe Networks
Version 1.0
EPANET Example Network 1
Input data File
Verification File
Hydraulics File
Map File
Number of Pipes
Number of Nodes
Number of Tanks
Number of Pumps
Number of Valves
Headloss Formula
Hydraulic Timestep
Hydraulic accuracy
Maximum Trials
Quality Analysis
Minimum Travel Time
Maximum Segments per Pipe
Specific Gravity
Kinematic Viscosity
Chemical Diffusivity
Total Duration
Reporting Duration
All Nodes
All Links
net 1. inp
Net 1. map
12
11
2
1
O
Hazen-Williams
1.00 hrs
0.001000
40
Chlorine
6.00 min
100
1.00
l.lOe-005 sq ft/sec
1.3e-008 sq ft/sec
24.00 hrs
Grade
ft
1004.50
985.31
970.07
968.86
971.55
969.07
968.63
967.35
965.63
970.00
800.00
Pressure
psi
127.61
119.29
117.02
118.66
117.66
118.75
120.73
115.84
110.77
52.00
0.00
FIGURE 9.10 EPANET output file for example network (figure 9.1)
Chlorine
rng/L
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
1.00
Tank
1.00
Reservoir
End
Node
11
12
13
22
23
32
12
21
22
23
31
32
10
Diameter
in
18.00
14.00
10.00
10.00
12.00
6.00
18.00
10.00
12.00
8.00
8.00
6.00
Flow
gpm
1865.06
1233.57
129.41
190.71
120.59
40.77
-765.06
481.48
189.11
29.41
140.77
59.23
1865.06
Pressure
ft
1006.92
988.05
973.13
971.91
974.49
972.10
971.66
970.32
968.63
973.06
800.00
Diameter
in
18.00
14.00
10.00
10.00
12.00
6.00
18.00
10.00
Velocity Headloss
fps
/1000ft
2.35
1.82
2.57
2.89
0.53
0.23
0.78
0.47
0.34
0.08
0.46
0.33
0.96
0.35
1.97
2.61
0.54
0.19
0.19
0.04
0.90
0.79
0.67
0.65
96 hp -204.50
Chlorine
psi
128.65
120.48
118.35
119.98
118.94
120.07
122.04
117.13
112.06
53.32
0.00
Flow
gpm
1847.49
1219.82
130.19
187.26
119.81
40.42
-747.49
477.68
mg/1
1.00
0.45
0.44
0.44
0.43
0.44
0.45
0.41
0.40
0.97
1.00
Velocity
fps
2.33
2.54
0.53
0.76
0.34
0.46
0.94
1.95
Pump
Tank
Reservoir
Headloss
/1000ft
1.79
2.83
0.23
0.45
0.08
0.32
0.34
2.57
Diameter
in
12.00
8.00
8.00
6.00
Demand
gpm
0.00
180.00
180.00
120.00
180.00
240.00
180.00
120.00
120.00
516.44
-1836.44
End
Node
11
12
13
22
23
32
12
21
22
23
31
32
10
Flow
gpm
192.14
30.19
140.42
59.58
1847.49
Grade
ft
1008.43
989.77
976.09
974.02
975.41
973.81
973.33
969.96
968.13
976.06
800.00
Diameter
in
18.00
14.00
10.00
10.00
12.00
6.00
18.00
10.00
12.00
8.00
8.00
6.00
Flow
gpm
1836.44
1163.77
173.00
150.47
127.00
42.20
-516.44
492.67
294.33
53.00
162.20
77.80
1836.44
Velocity
fps
0.55
0.19
0.90
0.68
97 hp
Pressure
psi
129.31
121.22
119.63
120.90
119.34
120.81
122.77
116.98
111.85
54.62
0.00
Headless
/1000ft
0.20
0.05
0.79
0.66
-206.92
Chlorine
mg/L
1.00
0.87
0.81
0.37
0.76
0.38
0.40
0.34
0.31
0.94
1.00
Velocity
Headless
fps
/1000ft
2.32
1.77
2.43
2.59
0.71
0.39
0.61
0.30
0.36
0.09
0.48
0.35
0.65
0.17
2.01
2.72
0.83
0.43
0.34
0.13
1.04
1.03
0.88
1.08
97 hp -208.43
Pump
Tank
Reservoir
Pump
Adjustment of these terms and others, such as valve settings and pump lifts, can be
made by trial and error or through systematic approaches. Several mathematical modeling methods have been suggested for solving the model calibration problem (Lansey and
Basnet, 1991).
Once a model is believed to be calibrated, an assessment should be completed. The
assessment entails a sensitivity analysis of model parameters to identify which parameters
have a strong impact on model predictions and future collection should emphasize
improving. The assessment also will identify the predictions (nodal pressure heads or tank
levels) that are sensitive to calibrated parameters and forecasted demands. Model assessments can simply be plots of model predictions versus parameter values or demand levels,
or they can be more sophisticated analyses of uncertainty, as discussed in Araujo (1992)
and Xu and Coulter (1998).
9.5.3 Model Results
Water-distribution simulation models require the model parameters, such as pipe and
pump characteristics, nodal demands, and valve settings, to solve the appropriate set of
equations and display the nodal peizometric heads, pipe flow rates water quality predictions, and other results, such as pipe head loss and pipe velocities. No standard format is
used between models. Abbreviated input and output files are shown in Figs. 9.9 and 9.10
for a sample system shown in Fig. 9.1. These files are for the EPANET code and are used
because the EPANET program is in the public domain and it models both flow and water
quality in an extended-period simulation format.
The constituent, chlorine, is reactive and results are shown for a selected subgroup of
nodes. As in most models, the constituent levels along a pipe are not provided. Finally,
tank concentrations, although not shown directly, can be found by examining the concentration closest to the tank node when flow is exiting the tank.
REFERENCES
Ahmed, L, Application of the gradient method for analysis of unsteady flow in water networks,
Master's thesis (Civil Engineering), University of Arizona, Tucson, 1997.
Araujo, J.V., A statistically based procedure for calibration of water systems, Doctoral dissertation,
Oklahoma State University, Stillwater, 1992.
Ahmed, L, and K. Lansey, "Analysis of unsteady flow in networks using a gradient algorithm
based method," ASCE Specialty Conference on Water Resoures, Tempe, AZ, June, 1999.
Boulos, P. E, T. Altman, P. A. Jarrige, and F. Collevati, "An event-driven method for modeling contaminant propagation in water networks," Journal of Applied Mathematical Modeling, 18(2): 8492, 1994.
Boulos, P E, T. Altman, P. A. Jarrige, and E Collevati, "Discrete simulation approach for network
water quality models," Journal of Water Resources Planning and Management, 121(1): 49-60,
1995.
Clark, R., Grayman, W. and R. M. Males, "Contaminant propagation in distribution systems" J. of
Environmental Eng., 114(4): 1988.
Cross, H. "Analysis of flow in networks of conduits or conductors," Bulletin No. 286, University
of Illinois Engineering Experimental Station, Urbana, IL 1936.
El-Shorbagy, W, and K. Lansey, "Non-conservative water quality modeling in water systems,"
Proceedings of the AWWA Specialty Conference on Computers in the Water Industry, Los
Angeles, April 1994.
Grayman, W. M., R.M. Clark, and R.M. Males, "Modeling Distribution System Water Quality:
Dynamic Approach," Journal of Water Resources Planning and Management, 114(3),: 295312, 1988.
Hart, F. L., J. L. Meader, and S.-M. Chiang, "CLNETA simulation model for tracing chlorine
residuals in a potable water distribution network," AWTK4 Distribution System Symposium
Proceedings, American Water Works Association, Denver, CO, 1987.
Holloway, M. B., Dynamic Pipe Network Computer Model, Doctoral dissertation, Washington State
University, Pullman, WA, 1985.
Islam R., and M. H. Chaudhry, "Modeling of constituent transport in unstead flows in pipe networks,"/. of Hydraulics Division, 124 (11): 1115-1124, 1998.
Jeppson, R. W., Analysis of Flow in Pipe Networks, Ann Arbor Science, Ann Arbor, MI, 1974.
Lansey, K., and C. Basnet, "Parameter Estimation for Water Distribution Systems," Journal of
Water Resources Planning and Management, 117(1): 126-144, 1991.
Liou, C. P., and J. R. Kroon, "Modeling the Propagation of Waterborne Substances in Distribution
networks," Journal of the American Water Works Association, 79(11): 54-58, 1987.
Martin, D. W., and G. Peters, "The Application of Newton's Method to Network Analysis by
Digital computer," Journal of the Institute of Water Engineers, 17: 115-129, 1963.
Rossman, L. A., "EPANETUsers manual," EPA-600/R-94/057, U.S. Environmental Protection
Agency, Risk Reduction Engineering Laboratory, Cincinnati, OH, 1994.
Rossman, L., and P. Boulos, "Numerical methods for modeling water quality in distribution systems: A comparison," Journal of Water Resources Planning and Management, 122(2): 137146, 1996.
Rossman, L. A., P. F. Boulos, and T. Altman, "Discrete Volume Element Method for Network
Water Quality Models," Journal of Water Resources Planning and Management, 119(5): 505517,1993.
Rossman, L. A., R. M. Clark, and W. M. Grayman, "Modeling Chlorine Residuals in DrinkingWater Distribution Systems," Journal of Environmental Engineering, American Society of Civil
Engineers, 120(4): 303-320, 1994.
Salgado, R., E. Todini, and P. E. O'Connell, "Comparison of the gradient method with some traditional methods for the analysis of water supply distribution networks," Proceeediigs International
Conference on Computer Applications for Water Supply and Distribution 1987, Leicester
Polytechnic, UK, September 1987.
Shah, M. and G. Sinai, "Steady State Model for Dilution in Water Networks." Journal of
Hydraulics Division, 114(2), 192-206, 1988.
Shamir, U., and C. D. Howard, "Water Distribution System Analysis," Journal of Hydraulics.
Division, 94(1). 219-234, 1965.
Todini, E., and S. Pilati, "A gradient method for the analysis of pipe networks," International
Conference on Computer Applications for Water Supply and Distribution 1987, Leicester
Polytechnic, UK, September 1987.
Walski, T., "Hardy-Cross meets Sherlock Holmes or model calibration in Austin, Texas," Journal
of the American Water Works Association, 82:34-38, March, 1990.
Wood, D. J., User's ManualComputer Analysis of Flow in Pipe Networks Including Extended
Period Simulations, Department of Civil Engineering, University of Kentucky, Lexington,
KY, 1980.
Wood, D., and C. Charles, "Hydraulic Network Analysis Using Llinear Theory," Journal of
Hydraulic Division, 98, (HY7): 1157-1170, 1972.
Xu, C., and I. Goulter, "Probabilistic Model for Water Distribution Reliability," Journal of Water
Resources Planning and Management, 124(4): 218-228, 1998.
CHAPTER 10
PUMP SYSTEM
HYDRAULIC DESIGN
B. E. Bosserman
Boyle Engineering Corporation
Newport Beach, CA
W.1
10.1.1
Pump types are described or defined by various organizations and their respective publications:
Hydraulics Institute (HI), American National Standard for Centrifugal Pumps for
Nomenclature, Definitions, Application and Operation [American National Standards
Institute (ANSI)THI 1.1-1.5-1994]
American Petroleum Institute (API), Centrifugal Pumps for Petroleum, Heavy Duty
Chemical, and Gas Industry Services, Standard 610, 8th ed., August 1995
American Society of Mechanical Engineers (ASME), Centrifugal Pumps,
Performance Test Code PTC 8.2-1990
In addition, there are several American National Standards Institute (ANSI) and
American Water Works Associations (AWWA) standards and specifications pertaining to
centrifugal pumps:
ANSI/ASME B73.1M-1991, Specification for Horizontal End Suction Centrifugal
Pumps for Chemical Process.
ANSI/ASME B73.2M-1991, Specification for Vertical In-Line Centrifugal Pumps for
Chemical Process.
ANSI/ASME B73.5M-1995, Specification for Thermoplastic and Thermoset
Polymer Material Horizontal End Suction Centrifugal Pumps for Chemical
Process.
ANSI/AWWA E 101-88, Standard for Vertical Turbine PumpsLineshaft and
Submersible Types.
HGL
Datum
Gauge
Pump
Gauge
adding weight in one correction plane only. This can be accomplished statically using
balance rails or by spinning.
Two-plane balancing (also called dynamic balancing). Two plane-balancing referes
to correction of residual unbalance to a specified limit by removing or adding weight
in two correction planes. This is accomplished by spinning on appropriate balancing
machines.
Overall efficiency (T|OA). This is the ratio of the energy imparted to the liquid (Pw) by
the pump to the energy supplied to the (Pmo/); that is, the ratio of the water horsepower to
the power input to the primary driver expressed in percent.
Power
Electric motor input power (P1110,). This is the electrical input power to the motor.
Pump input power (Pp). This is the power delivered to the pump shaft at the driver to
pump coupling. It is also called brake horsepower.
Pump output power (PJ. This is the power imparted to the liquid by the pump. It is
also called water horsepower.
P. = ^^
(U-S. units)
(10.1)
"=
Qx
x
^
366 *
<S-L units>
(10-2>
where
Q
H
S
Pw
=
=
=
=
Pump efficiency (r|p). This is the ratio of the energy imported to the liquid (Pw) to the
energy delivered to the pump shaft (Pp) expressed in percent.
Pump pressures
Field test pressure. The maximum static test pressure to be used for leak testing a
closed pumping system in the field if the pumps are not isolated. Generally this is
taken as 125 percent if the maximum allowable casing working pressure. In cases
where mechanical seals are used, this pressure may be limited by the pressure-containing capabilities of the seal.
Note: Seesure of the pump to 125 percent of the maximum allowable casing working
pressure on the suction split=case pumps and certain other pump types.
Maximum allowable casing working pressure. This is the hcase pumps and certain
other pump types.
Maximum allowable casing working pressure. This is the highest pressure at the specified pumping temperature for which the pump casing is designed. This pressure shall
be equal to or greater than the maximum discharge pressure. In the case of some
pumps (double suction, vertical turbine, axial split case can pumps, or multistage, for
example), the maximum allowable casing working pressure on the suction side may be
different from that on the discharge side.
Maximum suction pressure. This is the highest section pressure to which the pump
will be subjected during operation.
Working pressure (pd). This is the maximum discharge pressure that could occur in the
pump, when it is operated at rated speed and suction pressure for the given application.
Shut off. This is the condition of zero flow where no liquid is flowing through the
pump, but the pump is primed and running.
Speed. This is the number of revolutions of the shaft is a given unit of time. Speed is
expressed as revolutions per minute.
Suction conditions
Maximum suction pressure. This is the highest suction pressure to which the pump
will be subjected during operation.
Net positive suction head available (NPSHA). Net positive suction head available is
the total suction head of liquid absolute, determined at the first-stage impeller datum,
less the absolute vapor pressure of the liquid at a specific capacity:
(10.3)
where
hsa = total suction head absolute = hatm + hs
(10.4)
(10.5)
Net positive suction head required (NPSHR). This is the amount of suction head, over
vapor pressure, required to prevent more than 3 percent loss in total head from the first
stage of the pump at a specific capacity.
Static suction lift (Is). Static suction lift is a hydraulic pressure below atmospheric at
the intake port of the pump.
Submerged suction. A submerged suction exists when the centerline of the pump inlet
is below the level of the liquid in the supply tank.
Total discharge head (hd). The total discharge head (hd) is the sum of the discharge
gauge head (hgd) plus the velocity head (hvd) at point of gauge attachment plus the elevation head (Z4) from the discharge gauge centerline to the pump datum:
Total head (H]. This is the measure of energy increase per unit weight of the liquid,
imparted to the liquid by the pump, and is the difference between the total discharge
head and the total suction head.
This is the head normally specified for pumping applications since the complete characteristics of a system determine the total head required.
hd = hgd + hvd + Zd
(10.6)
Total suction head (hs), closed suction test. For closed suction installations, the pump
suction nozzle may be located either above or below grade level.
Total suction head (hs), open suction. For open suction (wet pit) installations, the first
stage impeller of the bowl assembly is submerged in a pit. The total suction head (hs)
at datum is the submergence (Zw). If the average velocity head of the flow in the pit is
small enough to be neglected, then:
hs = Zw
(10.7)
where
Zw = vertical distance in feet from free water surface to datum.
The total suction head (hs), referred to the eye of the first-stage impeller is the algebraic sum of the suction gauge head (hvs) plus the velocity head (hvs) at point of gauge
attachment plus the elevation head (Z5) from the suction gauge centerline (or manometer
zero) to the pump datum:
hs = hgs + hvs + Z3
(10.8)
The suction head (hs) is positive when the suction gauge reading is above atmospheric pressure and negative when the reading is below atmospheric pressure by an amount
exceeding the sum of the elevation head and the velocity head.
Velocity head (hv). This is the kinetic energy of the liquid at a given cross section.
Velocity head is expressed by the following equation:
h, = ~
(10.9)
where v is obtained by dividing the flow by the cross=sectional area at the point of gauge
connection.
10.1.3 Types of Centrifugal Pumps
The HI and API standards do not agree on these definitions of types of centrifugal pumps
(Figs. 10.2 and 10.3). Essentially, the HI standard divides centrifugal pumps into two
types (overhung impeller and impeller between bearings), whereas the API standard
divides them into three types (overhung impeller, impeller between bearings, and vertically suspended). In the HI standard, the "vertically suspended" type is a subclass of the
"overhung impeller" type.
I
Overhung
Impeller
Close coupled
single or two
stage
End suction
(Including submersibies)
In-line
In-line ANSI B73.2
Centrifugal
frame mounted
Separately
coupled
single or
two stage
Centerline support
API-610
Frame mounted
ANSI B73.1
Wet pit volute
Axial flow impeller (propeller)
volute type (horiz. or vertical)
Sealless
Kinetic
Canned motor
Magnetic drive
Impeller
between
bearing
Separately
coupled
single stage
Separately
coupled
multistage
Peripheral
Regenerative turbine
Side channel
Special effect
FIGURE 10.2 Kinetic type pumps per ANSI/HI-1.1-1.5-1994.
Reversible centrifugal
Rotating casing (pitot)
CENTRIFUGAL PUMP
Overhung
Flexibly
Coupled
Horizontal
Foot
Mounted
Rigidly
Coupled
Vertical
In-Line
Bearing
Frame
Centerline
Mounted
OH1 OH2
OH3
TYPES
Vertically Suspended
Between Bearings
Close
Coupled
Vertical
Vertical
Axially
Split
High
In-Line opeea
Integral
Gear
In-Line
OH4 OH5
OH6
Radially
Radially
Split
Split
Axially
Split
Discharge
Thru
Column
Single Double
Casing Casing
Axial
Diffuser
Flow
Volute
BB1
Double
Casing
Single
Casing
Separate
Discharge
(Sump)
Line CantiShaft lever
Diffuser
Volute
VS5 VS6
VS7
NPSH0
Efficiency
Discharge, gal/min
FIGURE 10.4 Typical discharge curves.
10.2
PUMPHYDRAULICS
A system curve describes the relationship between the flow in a pipeline and the head
loss produced; see Fig. 10.5 for an example. The essential elements of a system curve
include:
The static head of the system, as established by the difference in water surface elevations between the reservoir the pump is pumping from and the reservoir the pump is
pumping to,
The friction or head loss in the piping system. Different friction factors representing
the range in age of the pipe from new to old should always be considered.
The system curve is developed by adding the static head to the headlosses that occur
as flow increases. Thus, the system curve is a hyperbola with its origin at the value of the
static head.
The three most commonly used procedures for determining friction in pipelines are the
following:
10.2.2.1 Hazen-Williams equation. The Hazen-Williams procedure is represented by
the equation:
V = 1.318C/?0635054(U.S. units)
(10.1Oa)
where: V= velocity, (ft/s), C= roughness coefficient, R hydraulic radius, (ft), and S = friction head loss per unit length or the slope of the energy grade line (ft/ft).
In SI units, Eq. (10.10a)is
V = 0.849CR0 63S0 54
(10.1Ob)
(10"lla)
Head
Friction, fitting
and valve losses
Total static
head
Discharge
FIGURE 10.5 Typical system head-capacity curves.
where H1= headless, (ft), L = length of pipe, (ft), D = pipe internal diameter, (ft), Q =
flow, (ftVs), and C= roughness coefficient or friction factor.
In SI units, The Hazen-Williams equation is.
H
<-
nftllM
(iailb)
where HL = head loss, (m), Q = flow, (m3/s), D = pipe diameter, (m), and L = pipe
length, (m).
The C coefficient typically has a value of 80 to 150; the higher the value, the smoother
the pipe. C values depend on the type of pipe material, the fluid being conveyed (water or
sewage), the lining material, the age of the pipe or lining material, and the pipe diameter.
Some ranges of values for C are presented below for differing pipe materials in Table
10.1.
TABLE 10.1 Hazen-Williams Coefficents
Pipe Material
PVC
Steel (with mortar lining)
Steel (unlined) 120 to 140
Ductile iron (with mortar lining)
135-150
120-145
110-130
100-140
130-145
120-140
110-130
100-130
120-140
130-140
80-120
110-135
120-130
80-110
AWWA Manual Mil, Steel PipeA Guide for Design and Installation (AWWA,
1989), offers the following relationships between C factors and pipe diameters for water
service:
C
(10.12)
(10.13)
(10.14)
(U.S. units)
(10.15)
. x
/OT
SIumts
(
>
where n=roughness coefficient, H1 = head loss (ft or m), L = length of pipe (ft or m),
D=pipe internal diameter (ft or m), and Q=flow (cuVs or m3/s).
Values of n are typically in the range of 0.010 - 0.016, with n decreasing with
smoother pipes.
(10.16)
where / = friction factor from Moody diagram, g = acceleration due to gravity = 32.2
(ft/s) (U.S. units) = 9.81 m/s2 (SI units), HL = head loss (ft or m), L = length of pipe (ft
or m), D = pipe internal diameter (ft or m), and V = velocity (ft/s or m/sec).
Sanks et al., (1998) discuss empirical equations for determining/values. A disadvantage of using the Darcy-Weisbach equation is that the values for/depend on both roughness (EID) and also on the Reynolds number (Re):
R=
VD
(10.17)
where R = reynolds number (dimensionless), V = fluid velocity in the pipe (ft/s or m/s),
D = pipe inside diameter (ft or m), and v = kinematic viscosity (ft2/s or m2/s)
Values for/as a function of Reynold's number can be determined by the following
equations:
64
R less than 2000:/=
(10.18)
K
R = 2000^000: -= = 2 loglo fe + j^\
(10.19)
0.25
r
(ElD 5.74U
(10.20)
log
+
[ \Tr R55-jJ
where EID = roughness, with E = absolute roughness, feet or meters, and D = pipe diameter, (ft or m).
R greater than 4000: / =
Equation 10.19 is the Colebrook-White equation, and Eq. 10.20 is an empirical equation developed by Swamee and Jain, in Sanks et al., (1998). For practical purposes,/values for water works pipelines typically fall in the range of 0.016 to about 0.020.
10.2.2.4 Comparisons off, C9 and n. The Darcy-Weisbach friction factor can be compared to the Hazen-Williams C factor by solving both equations for the slope of the
hydraulic grade line and equating the two slopes. Rearranging the terms gives, in SI units,
/=(c^)(^i)
(10-21a)
<10-21b)
(10.23)
where hL = minor loss (ft or m), K = minor loss coefficient (dimensionless), V = fluid
velocity (ft/s or m/s), and g = acceleration due to gravity (= 32.2 ft-s/s or 9.81 m-s/s).
Headless or pressure loss through a valve also is determined by the equation
Q = Cv VAP (U.S. units) = 0.3807CV VAP
(S.I. units)
(10.24)
where Q = flow through valve (gal/m or m /s), Cv = valve capacity coefficient, and AP
= pressure loss through the valve (psi or kPa)
The coefficient Cv varies with the position of the valve plug, disc, gate, and so forth.
Cy indicates the flow that will pass through the valve at a pressure drop of 1 psi. Curves
of Cv versus plug or disc position (0-90,with O being in the closed position) must be
obtained from the valve manufacturer's catalogs or literature.
Cv and K are related by the equation
Cv =29.85 -^=
(U.S. units)
(10.25)
The system curve is superimposed over the pump curve; (Fig. 10.6). The pump operating
points occur at the intersections of the system curves with the pump curves. It should be
observed that the operating point will change with time. As the piping ages and becomes
rougher, the system curve will become steeper, and the intersecting point with the pump
curve will move to the left. Also, as the impeller wears, the pump curve moves downward.
Thus, over a period of time, the output capacity of a pump can decrease significantly. See
Fig. 10.7. for a visual depiction of these combined effects.
10.2.5 Pumps Operating in Parallel
To develop a composite pump curve for pumps operating in parallel, add the flows together
that the pumps provide at common heads (Fig. 10.8). This can be done with identical pumps
(those having the same curve individually) as well as with pumps having different curves.
10.2.6 Variable-Speed Pumps
The pump curve at maximum speed is the same as the one described above. The point on
a system-head curve at which a variable=speed pump will operate is similarly determined
by the intersection of the pump curve with the system curve. What are known as the pump
affinity laws or homologous laws must be used to determine the pump curve at reduced
speeds. These affinity laws are described in detail in Chap 12. For the discussion here, the
relevant mathematical relationships are Sanks et al., (1998).
1 = !L.
Q2 n2
(10.26)
IL = (Zk)2
H2
n2
(10.27)
- = (^
(10.28)
H
*2
2
where Q=flow rate, H=head, Ppower, n=rotational speed, and subscripts 1 and 2 are
only for corresponding points. Equations (10.26) and (10.27) must be applied simultaneously to ensure that Point 1 "corresponds" to Point 2. Corresponding points fall on parabolas through the original. They do not fall on system H-Q curves. These relationships,
known collectively as the affinity laws, are used to determine the effect of changes in
speed on the capacity, head, and power of a pump.
The affinity laws for discharge and head are accurate because they are based on actual
tests for all types of centrifugal pumps, including axial-flow pumps. The affinity law for
power is not as accurate because efficiency increases with an increase in the size of the pump.
When applying these relationships, remember that they are based on the assumption that
the efficiency remains the same when transferring from a given point on one pump curve to
a homologous point on another curve. Because the hydraulic and pressure characteristics at
the inlet, at the outlet, and through the pump vary with the flow rate, the errors produced by
Eq. (10.28) may be excessive, although errors produced by Eqs. (10.26) and (10.27) are
extremely small. See Fig. 10.9 for an illustration of the pump curves at different speeds.
Example. Consider a pump operating at a normal maximum speed of 1800 rpm, having a head-capacity curve as described in Table 10.2. Derive the pump curve for operating speeds of 10001600 rpm at 200-rpm increments.
System head
capacity curve
Pump head
capacity curve
Efficiency
Discharge, gpm
FIGURE 10.6 Determining the operating point for a single-speed pump with a fixed value of hstat
Head
Original performance
Performance
after wear
Capacity
FIGURE 10.7 Effect of impeller wear
The resulting new values for capacity (Q) and head (H) are shown in Table 10.2. The
values are derived by taking the Q values for the 1800 rpm speed and multiplying them
by the ratio O1M2) and by taking the H values for the 1800 rpm speed and multiplying
them by the ratio (H1In1)2.
10.3
CONCEPTOFSPECIFICSPEED
Ratio
Vn2
1,800
1,00
1,00
200
1,000
180
2,000
160
3,000
130
1,600
0,889
0,790
158
889
142
1,778
126
2,667
103
1,400
0,777
0,605
121
778
109
1,556
97
2,333
79
1,200
0,667
0,444
89
667
80
1,333
71
2,000
58
1,000
0,555
0,309
62
556
1,111
49
1,667
40
Speed
(rpm)
Ratio
U1Xn2
1,800
1,00
Ratio
(Vn2)2
Speed
(rpm)
Ratio
(Vn2)"
1,00
Point 1
Q (gpm) H(feet)
Point 1
Q (gpm) H(feet)
Point 4
Q (gpm) H(feet)
Point 4
Q (gpm) H(feet)
60
63
55
126
49
189
40
56
44
112
39
168
32
1,600
0,889
0,790
1,400
0,777
0,605
36
49
33
98
30
147
24
1,200
0,667
0,444
27
42
24
84
22
126
18
1,000
0,555
0,309
19
33
17
70
15
105
12
Two pumps
in parallel
Operating
point
One pump
Flow from one pump
with both pumps operating
i
System head <
Capacity (unitless)*
FIGURE 10.8 Pumps operating in parallel
System-head
curve
Head. H
Head Opacity
at
Peed
Head capacity at
speed
Friction
losses
Capacity,
FIGURE 10.9 Typical Discharge Curves for a Variable Speed Pump
impeller designs are as follows:
Type of Impeller
Radial-vane
Mixed-flow
Axial-flow
nQQ.50
NPSH/*
Q0m
Ua3UJ
(1
'31)
Equation (10.31) can be used to determine the approximate maximum allowable pump
speed as a function of net positive suction head available and flow for a given type of
TABLE 10.3 Equivalent Factors for Converting Values of Specific Speed Expressed in One Set
of Units to the Corresponding Values in Another Set of Units
Quantity
N
Q
H
(rev/min,
IJs,
m)
1.0
31.62
0.527
0.612
12.98
Expressed in Units of
(rev/min,
(rev/min,
(rev/min,
gal/mn,
m3/s,
m3/h,
m)
m)
ft)
1.898
1.633
0.0316
51.64
60.0
1.0
0.0167
1.0
0.861
1.162
0.0194
1.0
24.63
0.410
21.19
(rev/min,
ft3/s,
ft)
0.0771
2.437
0.0406
0.0472
1.0
70.4 NETPOSITIVESUCTIONHEAD
Net positive suction head, or NPSH, actually consists of two concepts:
the net positive suction available (NPSHA), and
the net positive suction head required (NPSH^).
The definition of NPSHA and NPSH^, as given by the Hydraulics Institute (1994),
were presented in Sec. 10.1.
10.4.1 Net Positive Suction Head Available
Figure 10.1 visually depicts the concept of NPSHA. Since the NPSH4 is the head available
at the impeller, friction losses in any suction piping must be subtracted when making the
calculation. Thus, the equation for determining NPSH4 becomes
NPSH, = halm + hs- hvp - hL
where:
(10.32)
/ Q at operating point \
\
Q at BEP
)
(10.33)
where the exponent n varies from 1.25 to 3.0, depending on the design of the impeller. In
most water and wastewater pumps, n lies between 1.8 and 2.8. The NPSH7J at the BEP
increases with the specific speed of the pumps. For high-head pumps, it may be necessary
either to limit the speed to obtain the adequate NPSH at the operating point or to lower
the elevation of the pump with respect to the free water surface on the suction side i to
increase the NPSHA.
10.4.3 NPSH Margin or Safety Factor Considerations
Any pump and piping system must be designed such that the net positive suction head
available (NPSHA) is equal to, or exceeds, the net positive suction head required
(NPSH^) by the pump throughout the range of operation. The margin is the amount by
which NPSHA exceeds NPSH^ (Hydraulics Institute, 1994). The amount of margin
required varies, depending on the pump design, the application, and the materials of
construction.
Practical experience over many years has shown that, for the majority of pump
applications and designs, NPSH^ can be used as the lower limit for the NPSH available. However, for high=energy pumps, the NPSH7J may not be sufficient. Therefore,
the designer should consider an appropriate NPSH margin over NPSH7J for high-energy pumps that is sufficient at all flows to protect the pump from damage caused by
cavitation.
Head
NPSH
FIGURE 10.10 Net positive suction head criteria as determined from pump test results.
NPSHR to
eliminate
cavitation
NPSHR to deliver
capacities at 3%
reduction of head
on H- Q
NPSHR, ft
Efficiency
Capacity, % of Q at B.E.P.
FIGURE 10.11 NPSH Required to suppress visible cavitation.
10.4.4 Cavitation
Cavitation begins to develop in a pump as small harmless vapor bubbles, substantially
before any degradation in the developed head can be detected (Hydraulics Institute, 1994).
This is called the point of incipient cavitation (Cavi, 1985; Hydraulics Institute, 1994).
Studies on high-energy applications show that cavitation damage with the NPSHA
greater than the NPSH^ can be substantial. In fact, there are studies on pumps that show
the maximum damage to occur at NPSH4 values somewhere between O and 1 percent head
drop (or two to three times the NPSH^), especially for high suction pressures as required
by pumps with high impeller-eye peripheral speeds. There is no universally accepted relationship between the percentage of head drop and the damage caused by cavitation. There
are too many variables in the specific pump design and materials, properties of the liquid
and system. The pump manufacturer should be consulted about NPSH margins for the
specific pump type and its intended service on high-energy, low-NPSHA applications.
According to a study of data contributed by pump manufacturers, no correlation exists,
between the specific speed, the suction specific speed, or any other simple variable and
the shape of the NPSH curve break-off. The design variables and manufacturing variables
are too great. This means that no standard relationship exists between a 3, 2, 1 or O percent head drop. The ratio between the NPSH required for a O percent head drop and the
NPSH^ is not a constant, but it generally varies over a range from 1.05 to 2.5. NPSH for
a O, 1, or 2 percent head drop cannot be predicted by calculation, given NPSH^.
A pump cannot be constructed to resist cavitation. Although a wealth of literature is
available on the resistance of materials to cavitation erosion, no unique material property
or combination of properties has been found that yields a consistent correlation with
cavitation damage rate (Sanks et al., 1998).
H^)W*
Description
Pipe Size
mm
1
Entrance
300
0.30
2
90 elbow
300
0.30
3
4.5 m of straight pipe
300
0.30
4
30 elbow
300
0.30
5
2 m of straight pipe
300
0.30
6
Butterfly valve
300
0.30
7
1.2m of straight pipe
300
0.30
8
300 mm X 200 mm reducer
200
0.20
9
150 mm X 250 mm increaser
250
0.25
10
1 m of straight pipe
250
0.25
11
Pump control valve
250
0.25
12
1 m of straight pipe
250
0.25
13
Butterfly valve
250
0.25
14
0.60 m of straight pipe
250
0.25
15
90 elbow
250
0.25
0.30
16
1.5m of straight pipe
250
0.25
17
Tee connection
250
0.25
Typical K values. Different publications present other values.
tReasonable value for mortar-lined steel pipe. Value can range from 130 to 145.
Friction Factor
K*
C+
1.0
0.30
140
0.20
140
0.46
140
0.25
0.25
140
0.80
140
0.46
140
140
0.50
HORIZONTAL SPLIT
CASE PUMP
FOREBAY
PLAN VIEW
HIGH WATER
LEVEL
LOW WATER
LEVEL
PUMP
FLOOR
SECTION VIEW
FIGURE 10.12
where K = fitting friction coefficient, V = velocity (m/s), g = acceleration due to gravity (m-s/s)
Sum of K values for various pipe sizes:
K300 = 1.96
K200 = 0.25
^250 = 2.31
Sum of C values for various pipe sizes:
Therefore,
TABLE 10.5 Convert Pump Curve Head Values to Include Minor Piping Losses
Q
Us
O
63
126
189
252
m3/s
O
0.063
0.126
0.189
0.252
H(m)
Uncorrected
60
55
49
40
27
Corrected
60
54.74
47.99
37.74
23.01
2. Modification of pump curve. Using the above equation for H1, a "modified"
pump curve can then be developed (see Table 10.5)
The H values as corrected must then be plotted. The operating point of the pump is
the intersection of the corrected H-Q curve with the system curve.
3. Calculation of NPSHA Using the data developed above for calculating the minor
losses in the piping, it is now possible to calculate the NPSHA for the pump. Only the
minor losses pertaining to the suction piping are considered: items 1-8 in Fig. 10.12. For
this suction piping, we have:
K300,-. = 1-96
Km m = 0.25
Sum of the C values: pipe length for 300-mm pipe is L = 7.7 m.
Determine the headloss in the suction piping:
Flow
(m3/s)
hs
(m)
hatm
(m)
hvp
(m)
9.0
9.0
9.0
9.0
9.0
9.0
1.0
1.0
1.0
1.0
1.0
1.0
10.35
10.35
10.35
10.35
10.35
10.35
10.35
10.35
10.35
10.35
10.35
10.35
0.24
0.24
0.24
0.24
0.24
0.24
0.24
0.24
0.24
0.24
0.24
0.24
19.11
18.99
18.58
17.91
16.99
15.81
11.11
10.99
10.58
9.91
8.99
7.82
10.6
=
=
=
=
10.35m
0.24 mat 150C
9 m maximum
1 m minimum
HYDRAULICCONSIDERATIONSINPUMPSELECTION
3. Fluid separation at high flows. This is one reason why there is also a definite maximum
capacity of a centrifugal pump. The pump components typically are not designed for
continuous operation of flows above about 120 to 130 percent of the flow that occurs
at the BEP.
4. Cavitation due to net positive suction head (NPSH) problems. There is a common misconception that if the net positive suction head available (NPSH4) is equal to or greater
than the net positive suction head required (NPSH^) shown on a pump manufacturer's
pump curve, then there will be no cavitation. This is wrong! As discussed in ANSI/HI
1.1-1.5-1994 (Hydraulics Institute, 1994 and also discussed by Taylor (1987), it takes
a suction head of 2 to 20 times the NPSH^ value to eliminate cavitation completely.
5. Flow disturbances in the pump intake due to improper intake design.
6. Air entrainment or aeration of the liquid.
7. Hydraulic resonance in the piping.
8. Solids contained in the liquids, such as sewage impacting in the pump and causing
momentary unbalance, or wedged in the impeller and causing continuous unbalance.
The HI standard then states:
The pump manufacturer should provide for the first item in the pump design and establish limits for low flow. The system designer is responsible for giving due consideration
to the remaining items.
The practical applications of the above discussion by observing what can happen in a
plot of a pump curve-system head curve as discussed above in Fig. 10.6. If the intersection of the system curve with the pump H-Q curve occurs too far to the left of the BEP
(i.e., at less than about 60 percent of flow at the BEP) or too far to the right of the BEP
(i.e., at more than about 130 percent of the flow at the BEP), then the pump will eventually fail as a result of hydraulically induced mechanical damage.
10.6.2 Causes and Effects of Centrifugal Pumps Operating Outside
Allowable Flow Ranges
As can be seen in Fig. 10.6, a pump always operates at the point of intersection of the system curve with the pump H-Q curve. Consequently, if too conservative a friction factor is
used in determining the system curve, the pump may actually operate much further to the
right of the assumed intersection point so that the pump will operate beyond its allowable
operating range. Similarly, overly conservative assumptions concerning the static head in
the system curve can lead to the pump operating beyond its allowable range. See Fig. 10.13
for an illustration of these effects. The following commentary discusses the significance of
the indicated operating points 1 through 6 and the associated flows Q1 through Q6.
Q1 is the theoretical flow that would occur, ignoring the effects of the minor head losses in the pump suction and discharge piping. See Fig. 10.12 for an example. Q1 is
slightly to the right of the most efficient flow, indicated as 100 units.
Q2 is the actual flow that would occur in this system, with the effects of the pump suction and discharge piping minor losses included in the analysis. Q2 is less than Q1, and
Q2 is also to the left of the point of most efficient flow. As shown in Fig. 10.7, as the
impeller wears, this operating point will move even further to the left and the pump
will become steadily less efficient.
Q1 and Q2 are the flows that would occur assuming that the system head curve that is
depicted is "reasonable," that is, not unrealistically conservative. If, in fact, the system
head curve is flatter (less friction in the system than was assumed), then the operating
point will be Q3 (ignoring the effects of minor losses in the pump suction and discharge
piping). If these minor losses are included in the analysis, then the true operating point
is Q4. At Q3, the pump discharge flow in this example is 130 percent of the flow that
occurs at the BEP. A flow of 130 percent of flow at the BEP is just at edge of, and may
even exceed, the maximum acceptable flow range for pumps (see discussion in Sec.
10.6.1). With most mortar-lined steel or ductile-iron piping systems, concrete pipe, or
with plastic piping, reasonable C values should almost always be in the range of 120
145 for water and wastewater pumping systems. Lower C usually would be used only
when the pumping facility is connected to existing, old unlined piping that may be
rougher.
If the static head assumed was too conservative, then the actual operating points would
be Q5 or Q6. Q5 is 150 percent of the flow at the BEP. Q6 is 135 percent of the flow at
the BEP. In both cases, it is most likely that these flows are outside the allowable range
of the pump. Cavitation, inadequate NPSH4, and excessive hydraulic loads on the
impeller and shaft bearings may likely occur, with resulting poor pump performance
and high maintenance costs.
10.6.3
Pump head^^^
capacity curve,
corrected for minor
losses in pump piping
Theoretical Pump
operating point (T)
Actual Pump
operating point (2)
hstat
hstat
(actual)
System head
capacity curve
(actual friction
value)
(assunned)
Allowable operating
range of pump
conditions normally only occur a minority of the time: the supply reservoir must be
at its low water level and the discharge reservoir must simultaneously be at its maximum water levelconditions that usually do not occur very often. Consequently,
select a pump that can operate properly at this conditionbut also select the pump
that has a BEP which occurs at the flow that will occur most often. See Fig. 10.14
for an example.
4. In multiple-pump operations, check the operating point with each combination of
pumps that may operate. For example, in a two-pump system, one pump operating
alone will produce a flow that is greater than 50 percent of the flow that is produced
with both pumps operating. This situation occurs because of the rising shape of the
system head curve; see Fig. 10.8. Verify that the pump output flows are within the
pump manufacturer's recommended operating range; see Fig. 10.13.
5. Check that NPSHA exceeds the NPSH7J for all the hydraulic considerations and operating points determined in Steps 1 and 3.
Flow
(per Pump)
Minimum
70
Normal 1
100
Normal 2
110
Maximum 1
115
Maximum 2
125
Comments
Maximum static head condition,
all pumps operating
Average or most frequent operating
condition: fewer than all pumps
operating, average static head
condition. Might also be the case
of all pumps operating, minimum
static head condition.
Fewer than all pumps operating,
minimum static head condition
Maximum static head condition,
one pump operating
Minimum static head condition,
one pump operating
System head
capacity curves
Pump head
capacity curve
Operating point at
maximum static
head
hstat
hstat
(minimum)
Operating point
at minimum
static head
Efficiency
Piping
Having selected a pump and determined its operating flows and discharge heads or pressures, it is then desirable to apply this data in the design of the piping. See Fig. 10.12 for
typical piping associated with a horizontal centrifugal pump.
10.7.2.1 Pump suction and discharge piping installation guidelines. Section 1.4 in the
Hydraulic Institute (HI) publication ANSI/HI 1.1-1.5 (1994) and Chap. 6 in API
Recommended Practice 686 (1996) provide considerable discussion and many recommendations on the layout of piping for centrifugal pumps to help avoid the hydraulic problems discussed above.
10.7.2.2 Fluid velocity. The allowable velocities of the fluid in the pump suction and discharge piping are usually in the following ranges:
Suction:
Discharge:
1. Variation in static heads, as the water surface elevations in both the suction and discharge reservoirs vary
2. Long-term variations in pipeline friction factors (Fig. 10.5)
3. Long-term deterioration in impeller (Fig. 10.7)
4. Variation in the number of pumps operating in a multipump system (Fig. 10.8).
A suggested procedure for sizing the suction and discharge piping is as follows:
1. Select an allowable suction pipe fluid velocity of 3-5 ft/s (1.0-1.5 m/s) with all pumps
operating at the minimum static head condition. As fewer pumps are used, the flow
output of each individual pump will increase (typically by about 20 to 40 percent with
one pump operating compared to all pumps operating) with the resulting fluid velocities in the suction piping also increasing to values above the 3-5 ft/s (1.0-1.5 m/s)
nominal criteria;
2. Select an allowable discharge pipe fluid velocity of 5-8 ft/s (1.5-2.4 m/s) also with all
pumps operating at the minimum static head condition. As discussed above, as fewer
pumps are used, the flow output of each individual pump will increase with the resulting fluid velocities in the discharge piping also increasing in values above the 5-8 ft/s
(1.5-2.4 m/s) nominal criteria.
10.7.2.3 Design of pipe wall thickness (pressure design) Metal pipes are designed for
pressure conditions by the equation for hoop tensile strength:
t=
^E
(10 34)
'
where
t = wall thickness, in or mm
D = inside diameter, in or mm (although in practice, the outside diameter is often
conservatively used, partly because the ID is not known initially and because it
is the outside diameter (OD) that is the fixed dimension: ID then varies with the
wall thickness)
P = design pressure (psi or kPa)
S = allowable design circumferential stress (psi or kPa)
E = longitudinal joint efficiency
The design value for S is typically 50 percent of the material yield strength, for "normal" pressures. For surge or transient pressures in steel piping systems, S is typically
allowed to rise to 70 percent of the material yield strength (American Water Works
Association 1989).
The factor E for the longitudinal joint efficiency is associated with the effective
strength of the welded joint. The ANSI B31.1 (American Society for Mechanical
Engineers, 1995) and B31.3 (American Society for Mechanical Engineers, 1996) codes
for pressure piping recommend the values for E given in Table 10.8
Weld Joint
Efficiency Factor (E)
0.80
0.90
1.00
0.85
0.60
0.85
1.0
The wall thickness for plastic pipes [polyvinyl chloride (PVC), high-density polyethylene (HDPE), and FRP] is usually designed in the United States on what is known as the
hydrostatic design basis or HDB:
,,.^.x
where Pt = total system pressure (operating + surge), t = minimum wall thickness (in),
D = average outside diameter (in), HDB = hydrostatic design basis (psi) anh F = factor
of safety (2.50-4.00)
10.7.2.4 Design of pipe wall thickness (vacuum conditions). If the hydraulic transient
or surge analysis (see Chap. 12) indicates that full or partial vacuum conditions may
occur, then the piping must also be designed accordingly. The negative pressure required
to collapse a circular metal pipe is described by the equation:
(10 36)
^ = (T^SP (^
'
where AP = difference between internal and external pipeline pressures (psi or kPa), E =
modulus of elasticity of the pipe material (psi or kPa), fi = Poisson's ratio, SF = safety factor (typically 4.0), e = wall thickness (in or m) anh D = outside diameter (in or m)
Because of factors such as end effects, wall thickness variations, lack of roundness,
and other manufacturing tolerances, Eq. (10.3b) for steel pipe is frequently adjusted in
practice to
AP = 50,000,000 MB
SF
\DJ
(1037)
10.7.2.5 Summary of pipe design criteria. The wall thickness of the pump piping system is determined by consideration of three criteria:
1. Normal operating pressure [Eq. (10.34)], with S = 50 percent of yield strength
2. Maximum pressure due to surge (static + dynamic + transient rise), using Eq. (10.34)
with S = 70 percent of yield strength (in the case of steel pipe)
3. Collapsing pressure, if negative pressures occur due to surge conditions (Eq. 10.36).
The pressures generated due to hydraulics, thus, must be considered in the pipe design,
as was discussed in Sec. 10.7, above.
10.8.1 Effect of Surge on Valve Selection
At its worst, surges in a piping can cause swing check valves to slam closed violently
when the water column in the pipeline reverses direction and flows backward through the
check valve at a significant velocity before the valve closes completely. Consequently, in
pump and piping systems in which significant surge problems are predicted to occur,
check valves or pump control valves are typical means to control the rate of closure of the
valve. Means of controlling this rate of closure include
Using a valve that closes quickly, before the flow in the piping can reverse and attain
a high reverse velocity.
Providing a dashpot or buffer on the valve to allow the valve clapper or disc to close
gently.
Closing the valve with an external hydraulic actuator so that the reverse flowing water
column is gradually brought to a halt. This is frequently done with ball or cone valves
used as pump control valves.
The pressure rating of the valve (both the check valve or pump control valve and the
adjacent isolation) should be selected with a pressure rating to accommodate the predicted surge pressures in the piping system.
10.8.2 Effect of Surge on Pipe Material Selection
Metal piping systems, such as steel and ductile iron, have much better resistance to
surge than do most plastic pipes (PVC, HDPE, ABS, and FRP). The weakness of plastic
pipes with respect to surge pressures is sometimes not adequately appreciated because the
wave velocity (a) and, hence, the resulting surge pressures are significantly lower than is
the case with metal piping systems. Since the surge pressures in plastic piping are lower
than those in metal piping systems, there is sometimes a mistaken belief that the entire
surge problem can then be neglected. However, plastic piping systems inherently offer
less resistance to hydraulic transients than do metal piping systems, even with the lower
pressures. This is particularly the case with solvent or adhesive welded plastic fittings.
HDPE has better resistance to surge pressures than other plastic piping systems. In
addition, the joints are fusion butt welded, not solvent welded, which results in a stronger
joint. However, HDPE is still not as resistant to surge effects as a properly designed steel
or ductile iron piping system.
REFERENCES
American Petroleum Institute, Centrifugal Pumps for Petroleum, Heavy Duty Chemical, and Gas
Industry Services, API Standard 610, 8th ed American Petroleum Institute, Washington, DC.
American Society of Mechanical Engineers (ASME), B31.1, Power Piping, ASME, New York, 1995.
American Society of Mechanical Engineers (ASME), B31.3, Process Piping, ASME, New York, 1996
American
Water Works Association, Steel PipeA Guide for Design and Installation, AWWA Mil,
3rd ed., American Water Works Association, Denver, CO 1989.
Cavi, D., "NPSHR Data and Tests Need Clarification," Power Engineering, 89:47-50, 1985.
Hydraulics Institute, American National Standard for Centrifugal Pumps for Nomenclature,
Definitions, Applications, and Operation, ANSI/HI 1.1-1.5-1994, Hydraulics Institute,
Parsippany, NJ, 1994.
American Petroleum Institute, Recommended Practices for Machinery Installation and Installation
Design, Practice 686, 1st ed. Washington, DC, 1996.
Sanks, R. L., et al., Pumping Station Design, 2nd ed., Butterworths, 1998.
Taylor., "Pump Bypasses Now More Important," Chemical Engineering, May 11, 1987.
APPENDIX 10. A
PUMP SYSTEM
HYDRAULIC DESIGN
Calculation of minor losses and NPSH4 in piping and modification of a pump curve
(U.S. units)
Part 1. Calculation of Minor Losses
Principal headless equations
For straight sections of pipe: H1 = ^jjj ( |
where L = length in feet, D = pipe diameter (ft), Q - flow (ft3/s) anh C = Hazen-Williams friction
factor
For fittings: HL = ^K ^L
where K = fitting friction coefficient, V = velocity in (ft/s), anh g = acceleration due to gravity
[(ft-s)/s]
Sum of K values for various pipe sizes:
Kn = 1.96
KS = 0.25
K10 = 2.31
HL^
H
LM
_ 4.72(26) / 6 V*_ O o m 3
0-0""
(12/12)4.86 ^ 140j
V2
Convert terms to Q2 terms
^g
Friction Factor
Item in
Fig. 10.12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Description
Entrance
90 elbow
15 ft of straight pipe
30 elbow
7 ft of straight pipe
Butterfly valve
4 ft of straight pipe
12 in X 8 in reducer
6 in X 10 in increaser
3 ft of straight pipe
Pump control valve
3 ft of straight pipe
Butterfly valve
2 ft of straight pipe
90 elbow
5 ft of straight pipe
Tee connection
Pipe Size
(in)
12
12
12
12
12
12
12
8
10
10
10
10
10
10
10
10
10
K*
1.0
0.30
140
0.20
140
0.46
Therefore,
C+
140
0.25
0.25
140
0.80
140
0.46
140
0.30
140
0.50
H (ft)
GPM
CFS
Uncorrected
Corrected
O
1000
2000
3000
4000
O
2.228
4.456
6.684
8.912
200
180
160
130
90
200
178.87
151.52
120.0
72.29
The H values as corrected must then be plotted. The operating point of the pump is the
intersection of the corrected H-Q curve with the system curve.
Part 3: Calculation of NPSHA
Using the data developed above for calculating the minor losses in the piping, it is now
possible to calculate the NPSHA for the pump. Only the minor losses pertaining to the suction piping are considered: Items 1-8 in Fig. 10.12. For this suction piping, we have: ^T12
= 1.96, ^T8 = 0.25, sum of C values, Pipe length for 12-in pipe: L = 26 ft.
Determine the headloss in the suction piping
Compute NPSHA:
Condition
Flow
(fP/s)
h,
(ft)
Ji01n
(ft)
hvp
(ft)
O
2
4
6
8
10
29
29
29
29
29
29
33.96
33.96
33.96
33.96
33.96
33.96
0.78
0.78
0.78
0.78
0.78
0.78
0.00
0.37
1.47
3.28
5.81
9.05
62.18
61.81
60.71
58.90
56.37
53.13
O
2
4
6
8
10
5
5
5
5
5
5
33.96
33.96
33.96
33.96
33.96
33.96
0.78
0.78
0.78
0.78
0.78
0.78
0.00
0.37
1.47
3.28
5.81
9.05
38.18
37.81
36.71
34.90
32.37
29.13
33.96 ft
0.78 ft at 6O0F
29 ft maximum
5 ft minimum
HL at Flow NPSHAatlow
(ft)
(ft)
CHAPTER 11
WATER DISTRIBUTION
SYSTEM DESIGN
Mark A. Ysusi
Montgomery Watson
Fresno, California
11.1
INTRODUCTION
The primary purpose of a water distribution system is to deliver water to the individual consumer in the required quantity and at sufficient pressure. Water distribution systems typically carry potable water to residences, institutions, and commercial and industrial establishments. Though a few municipalities have separate distribution systems, such as a highpressure system for fire fighting or a recycled wastewater system for nonpotable uses, most
municipal water distribution systems must be capable of providing water for potable uses
and for nonpotable uses such as fire suppression and irrigation of landscaping.
The proper function of a water distribution system is critical to providing sufficient
drinking water to consumers as well as providing sufficient water for fire protection.
Because these systems must function properly, the principals of their planning, design,
and construction need to be understood. This chapter focuses on the critical elements of
planning and design of a water distribution system. The information presented primarily
discusses typical municipal water distribution systems; however, the hydraulic and design
principles presented can be easily modified for the planning and design of other types of
pressure distribution systems such as fire protection and recycled wastewater.
11.1.1 Overview
Municipal water systems typically consist of one or more sources of supply, appropriate
treatment facilities, and a distribution system. Sources of supply include surface water,
such as rivers or lakes, groundwater, and in some instances, brackish or sea water. The
information contained in this chapter is limited to the planning and design of distribution
systems and does not address issues related to identifying and securing sources of supply
or designing and constructing appropriate water treatment facilities. Water distribution
systems usually consist of a network of interconnected pipes to transport water to the consumer, storage reservoirs to provide for fluctuations in demand, and pumping facilities.
11.1.2 Definitions
Many of the frequently used terms in water distribution system planning and design are
defined here.
Average day demand. The total annual quantity of water production for an agency
or municipality divided by 365.
Maximum day demand. The highest water demand of the year during any 24-h period.
Peak hour demand. The highest water demand of the year during any 1-h period.
Peaking factors. The increase above average annual demand, experienced during a
specified time period. Peaking factors are customarily used as multipliers of average day
demand to express maximum day and peak hour demands.
Distribution pipeline or main. A smaller diameter water distribution pipeline that
serves a relatively small area. Water services to individual consumers are normally placed
on distribution pipelines. Distribution system pipelines are normally between 150 and 400
mm (6-16 in.).
Transmission pipeline or main. A larger diameter pipeline, designed to transport larger quantities of water during peak demand periods. Water services for small individual
consumers are normally not placed on transmission pipelines. Transmission mains are
normally pipelines larger than 400 mm (16 in.).
77.2
DISTRIBUTIONSYSTEMPLANNING
The basic question to be answered by the water distribution system planner/designer is,
"How much water will my system be required to deliver and to where?" The answer to
this question will require the acquisition of basic information about the community
including historical water usage, population trends, planned growth, topography, and
existing system capabilities, to name just a few. This information can then be used to plan
for logical extension of the existing system and to determine improvements necessary to
provide sufficient water at appropriate pressure.
11.2.1 Water Demands
The first step in the design of a water distribution system is the determination of the quantity of water that will be required, with provision for the estimated requirements for the
future.
In terms of the total quantity, the water demand in a community is usually estimated
on the basis of per capita demand. According to a study published by the U.S. Geological
Survey, the average quantity of water withdrawn for public water supplies in 1990 was
estimated to about 397 L per day per capita (Lpdc) or 105 gal per day per capita (gpdc).
The withdrawals by state are summarized in Table 11.1. The reported water usage shown
in Table 11.1 illustrates a wide variation. Per capita water use varies from a low use in
Pennsylvania of just over 60 gpcd to over 200 gpcd in Nevada. These variations depend
on geographic location, climate, size of the community, extent of industrialization, and
other influencing factors unique to most communities. Because of these variations, the
only reliable way to estimate future water demand is to study each community separately, determining exiting water use characteristics and extrapolating future water demand
using population trends.
In terms of how the total water use is distributed within a community throughout the
day, perhaps the best indicator is land use. In a metered community, the best way to determine water demand by land use is to examine actual water usage for the various types of
land uses. The goal of examining actual water usage is to develop water "duties" for the
L/Capita/per day
379
299
568
401
556
549
265
295
678
420
435
450
704
341
288
250
326
265
469
220
397
250
291
560
466
326
488
435
806
269
284
511
450
254
326
189
322
420
235
254
288
307
gal/Capita/day
100
79
150
106
147
145
70
78
179
111
115
119
186
90
76
66
86
70
124
58
105
66
77
148
123
86
129
115
213
71
75
135
119
67
86
50
85
111
62
67
76
81
L/Capita/per day
322
541
825
303
284
522
280
197
617
182
87
397
gal/Capita/day
85
143
218
80
75
138
74
52
163
48
23
105
various types of land uses that can be used for future planning. Water duties are normally
developed for the following land uses:
Single, family residential (some communities have low-medium-and, high-density
zones)
Multifamily residential
Commercial (normally divided into office and retail categories)
Industrial (normally divided into light and heavy categories and separate categories for
very high users
Public (normally divided into park, or open space, and schools)
Water duties are normally expressed in gallons per acre per day. Table 11.2 shows typical
water duties in the western United States. It should be noted that the definitions of land
use terms like "low-density residential," "medium-density residential," and so on, will
vary by community and should be examined carefully.
Another method of distributing water demand is to examine the water usage for individual users. This is particularly the case when an individual customer constitutes a significant portion of the total system demand. Table 11.3 presents water use for many different establishments. Although the rates vary widely, they are useful in estimating total
water use for individual users when no other data are available.
11.2.2 Planning and Design Criteria
To effectively plan and design a water distribution system, criteria must be developed and
adopted against which the adequacy of the existing and planned system can be compared.
Typical criteria elements include the following:
Supply
Storage
Low
Low-density residential
400
3,300
1,670
Medium-density residential
900
3,800
2,610
High-density residential
2,300
12,000
4,160
Single family residential
1,300
2,900
2,300
Multifamily residential
2,600
6,600
4,160
Office commercial
1,100
5,100
2,030
Retail commercial
1,100
5,100
2,040
Light industrial
200
4,700
1,620
Heavy industrial
200
4,800
2,270
Parks
400
3,100
2,020
Schools
400
2,500
1,700
Source: Adapted from Montgomery Watson study of data of 28 western U.S. cities.
Note: gal X 3.7854 = L.
TABLE 11.3 Typical Rates of Water Use for Various Establishments
User
Airport, per passenger
Assembly hall, per seat
Bowling alley, per alley
Camp
Pioneer type
Children's, central toilet and bath
Day, no meals
Luxury, private bath
Labor
Trailer with private toilet and bath,
per unit (2 1/2 persons)
Country clubs
Resident type
Transient type serving meals
Dwelling unit, residential
Apartment house on individual well
Apartment house on public water
supply, unmetered
Boardinghouse
Hotel
Range of Flow
L/person
gal/person
or unit/day
or unit/day
10-20
6-10
60-100
3-5
2-3
16-26
80-120
160-200
40-70
300^00
140-200
500-600
21-32
42-53
11-18
79-106
37-53
132-159
300-600
60-100
79-159
16-26
300-400
300-500
79-106
79-132
150-220
200-400
40-58
53-106
User
Lodging house and tourist home
Motel
Private dwelling on individual well
or metered supply
Private dwelling on public water supply,
unmetered
Factory, sanitary wastes, per shift
Fairground (based on daily attendance)
Institution
Average type
Hospital
Office
Picnic park, with flush toilets
Restaurant (including toilet)
Average
Kitchen wastes only
Short order
Short order, paper service
Bar and cocktail lounge
Average type, per sear
Average type 24 h, per seat
Tavern, per seat
Service area, per counter seat (toll road)
Service area, per table seat (toll road)
School
Day, with cafeteria or lunchroom
Day, with cafeteria and showers
Boarding
Self-service laundry, per machine
Store
First 7.5 m ( 25 ft) of frontage
Each additional 7.5 m of frontage
Swimming pool and beach, toilet and shower
Theater
Indoor, per seat, two showings per day
Outdoor, including food stand, per car
(3 1/3 persons)
Source: Adapted from Metcalf and Eddy (1979).
Range of Flow
L/person
gal/person
or unit/day
or unit/day
120-200
400-600
200-600
32-53
106-159
53-159
400-800
106-211
40-100
2-6
11-26
1-2
400-600
700-1200
40-60
20^0
106-159
185-317
11-16
5-11
25^0
10-20
10-20
4-8
8-12
120-180
160-220
60-100
1000-1600
600-800
7-11
3-5
3-5
1-2
2-3
32^8
42-58
16-26
264-423
159-211
40-60
60-80
200-400
1000-3000
11-16
16-21
53-106
264-793
1600-2000
1400-1600
40-60
423-528
370-423
11-16
10-20
3-5
10-20
3-5
Fire demands
Distribution system analysis
Service pressures
11.2.2.1 Supply. In determining the adequacy of water supply facilities, the source of
supply must be large enough to meet various water demand conditions, and be able to
meet at least a portion of normal demand during emergencies such as power outages and
disasters. At a minimum, the source of supply should be capable of meeting the maximum
day system demand. It is not advisable to rely on storage to make up any shortfall in supply at maximum day demand. The fact that maximum day demand may occur several days
consecutively must be considered by the system planner/designer. It is common for communities to provide a source of supply that meets the maximum day demand, with the
additional supply to meet peak hour demand coming from storage. Some communities
find it more economical to develop a source of supply that not only meets maximum day
but also peak hour demand.
It is also good practice to consider standby capability in the source of supply. If the
system has been designed where the entire capacity of the supply is required to meet the
maximum demand, any portion of the supply that is placed out of service due to malfunction or maintenance will result in a deficient supply. For example, a community that
relies primarily on groundwater for its supply should, at a minimum, be able to meet its
maximum day demand with at least one of its largest wells out of service.
11.2.2.2 Storage. The principal function of storage is to provide reserve supply for (1)
operational equalization, (2) fire suppression reserves, and (3) emergency needs.
Operational storage is directly related to the amount of water necessary to meet peak
demands. The intent of operational storage is to make up the difference between the consumers' peak demands and the system's available supply. It is the amount of desirable
stored water to regulate fluctuations in demand so that extreme variations will not be
imposed on the source of supply. With operational storage, system pressures are typically
improved and stabilized. The volume of operational storage required is a function of the
diurnal demand fluctuation in a community and is commonly estimated at 25 percent of
the total maximum day demand.
Fire storage is typically the amount of stored water required to provide a specified fire
flow for a specified duration. Both the specific fire flow and the specific time duration varies
significantly by community. These values are normally established through the local fire
marshall and are typically based on guidelines established by the Insurance Service Office,
a nonprofit association of insurers that evaluate relative insurance risks in communities.
Emergency storage is the volume of water recommended to meet demand during emergency situations such as source of supply failures, major transmission main failures, pump
failures, electrical power outages, or natural disasters. The amount of emergency storage
included with a particular water system is an owner option, typically based on an assessment
of risk and the desired degree of system dependability. In considering emergency storage, it
is not uncommon to evaluate providing significantly reduced supplies during emergencies.
For example, it is not illogical to assume minimal demand during a natural disaster.
11.2.2.3 Fire demands. The rate of flow to be provided for fire flow is typically dependent on the land use and varies by community. The establishment of fire flow criteria
should always be coordinated with the local fire marshall. Typical fire flow requirements
are shown in Table 11.4.
11.2.2.4 Distribution system analysis. In evaluating an existing system or planning a
proposed system, it is important to establish the criteria of operational scenarios against
Maximum pressure
Minimum pressure during maximum day
Minimum pressure during peak hour
Minimum pressure during fires
65-75
30-40
25-35
20
U.S. Range
Common Range
1.5-3.5:1
2.0-7.0:1
1.8-2.8:1
2.5^.0:1
text editor. Errors were commonplace and just getting a data file that would run could
involve days if not weeks of effort, depending on the size and complexity of the network
being modeled. Model output was usually a voluminous tabular listing of key network
results. Interpretation of the results was time consuming and typically involved hand plotting of pressure contours on system maps.
Because of the widespread use of microcomputers during the last two decades, network modeling has taken on new dimensions. Engineers today rely on computer models
to solve a variety of hydraulic problems. The use of interactive on-screen graphics to enter
and edit network data and to color code and display network maps, attributes, and analysis results has become commonplace in the water industry. This makes it much easier for
the engineer to construct, calibrate, and manipulate the model and visualize what is happening in the network under various situations such as non-compliance with system performance criteria. The engineer is now able to spend more time thinking and evaluating
system improvements and less time on flipping through voluminous pages of computer
printouts, thus leading to improved operation and design recommendations. The new generation of computer models have greatly simplified the formidable task of collecting and
organizing network data and comprehending massive results.
11.2.4.2 Software packages. Many of the software packages available offer additional
capabilities beyond standard hydraulic modeling such as water quality assessment (both
conservative and reactive species), multiquality source blending, travel time determination,
energy and power cost calculation, leakage and pressure management, fire flow modeling,
surge (transient) analysis, system head curve generation, automated network calibration,
real-time simulation, and network optimization. Some sophisticated models can even
accommodate the full library of hydraulic network components including pressure regulating valves, pressure sustaining valves, pressure breaker valves, flow control valves, float
valves, throttle control valves, fixed- and variable speed pumps, turbines, cylindrical and
variable cross-sectional area tanks, variable head reservoirs, and multiple inlet/outlet tanks
and reservoirs. Through their predictive capabilities, computer network models provide a
powerful tool for making informed decisions to support many organization
programs and policies. Modeling is important for gaining proper understanding of system
dynamic behavior, operator training, optimizing the use of existing facilities, reducing operating costs, determining future facility requirements, and addressing water quality distribution issues.
There is an abundance of network modeling software in the marketplace today. Some
are free and others can be purchased at a nominal cost. Costs can vary significantly
between models depending on the range of the features and capabilities provided. The
four major sources of computer models include consulting firms, commercial software
companies, universities, and government agencies. Many of the programs available from
these sources have been on the market for several years with established track records.
Most of the recent computer models however, provide very sophisticated and intuitive
graphical user interfaces and results presentation environment, as well as direct linkages
with information management systems such as relational databases and geographical
information systems. Table 11.7 lists the names, addresses, and phone numbers of network
modeling software vendors, along with their primary modeling product.
11.2.4.3 Development of a system model. As indicated above, the computerized tools
available to the engineer today are impressive and powerful. Once an appropriate software
is selected, data must be then input to the software in order to develop a computer model
of the water system under study. Input data includes the physical attributes of the system
including pipe sizes and lengths, topography, reservoir and pump characteristics, as well
as the anticipated nodal demands.
Vendor
Address
City
State
Country
Zip Code
Telephone
Fax
AQUA
AVWater
BOSS EMS
CYBERNET
EPANET
FAAST
FLOW NETWORK
ANALYSIS
H2ONET
HYDRONET
InWater
KYPIPE
Cuyahoga Falls
Rochester
Madison
Waterbury
Cincinnati
Livermore
Baton Rouge
OH
NY
WI
CT
OH
CA
LA
USA
USA
USA
USA
USA
USA
USA
44221
14614
53705-4200
06708
45268
94550-4738
70816
(330) 929-7886
(716) 232-6998
1-800-488-4775
1-800-727-6555
(513)569-7603
(510)455-8086
(504) 769-6785
(330) 929-2756
(716) 262-2042
(608) 258-9943
(203) 597-1488
(513) 569-7185
(510)455-8087
Pasadena
Truckee
Huntsville
Lexington
CA
CA
AL
KY
USA
USA
USA
USA
91101
92162
35898
40506
(626) 568-6868
(530) 582-1525
1-800-345-4856
(606) 257-3436
(626) 568-6619
(530) 582-8579
(205) 730-6109
(606) 257-8005
France
92007
01133146147181
01133147247202
98503
H7L4S2
60187
V8W 2H7
17013
20707
84322-4100
(360) 412-0702
612-9587-5384
(514)629-8888
(630) 682-4700
(250) 385-0206
(717) 243-1900
(301)490-4515
(801) 797-2943
(360) 412-0672
612-9587-5384
(514) 382-3077
(630) 682-4754
(250) 385-7737
(717) 243-5564
(301)490-4515
(801)750-1185
PICCOLO
PIPE-FLO
PIPES
RINCAD
RJN CASS WORKS
SDP
Stoner Workstation
TDHNET
USU-NETWK
WADISO
Water Distribution
Systems: Simulation
and Sizing
Water Works
WATERAVGRAPH
WaterMax
WATNET
WATSYS
MW Soft, Inc.
Tahoe Designs Software
Intergraph
University of Kentucky
NANTERRE Cedex
Lacey
Bexley, NSW 2217
Laval
Wheaton
Victoria
Carlisle
Laurel
Logan
WA
Quebec
IL
BC
PA
MD
UT
USA
Australia
Canada
USA
Canada
USA
USA
USA
Boca Raton
FL
USA
33431-9868
1-800-272-7737
1-800-374-3401
Vancouver
Ladysmith
Chicago
Trevose
Victoria
B.C.
B.C.
IL
PA
B.C.
Canada
Canada
USA
USA
Canada
V6E 4A2
VOR2EO
60606
19053
V8V 4C9
(604) 688-8271
(250) 722-3810
(312) 236-5655
(215) 244-9972
(250) 384-5955
(604) 688-1286
(250) 722-3088
(312) 580-2691
(215) 244-9977
(250) 383-1692
Development of the nodal demands normally involves distributing the average day
flow throughout the system in proportion to land use. This is commonly accomplished
by determining a demand area for each node, measuring the area of each different land
use within the demand area, multiplying the area of each land use within the demand area
by its respective average day water duty (converted to gal/min or L/sec), and summing
the water duties for each land use within the demand area and applying the sum at the
node. In the past this effort required extensive mapping and determining the land use
areas by planimeter or hand measurement. Today, with the advent of graphical information system software (GIS), development of nodal demands is normally an activity
involving computer based mapping. The elements of the system, the demand areas, and
the land uses are all mapped in separate layers in the GIS software. The GIS software
capability of "polygon processing" intersects the different layers and automatically computes the land use sums with the various demand areas. When the water duties are multiplied by their respective land use, the result is the average day system demand, proportioned to each node by land use. The water system computer model is then used to apply
global peaking factors as described above.
mined and the proposed water pipeline designed accordingly. What is shown on a utility
company plat as a single line representing an electrical conduit may turn out to be a major
electrical line with several conduits encased in concrete having a cross section 2 ft wide
and 4 ft deep! Or, what is shown as a buried 3/4-inch telephone line may turn out to be a
fiber-optic telecommunications cable that, if severed during construction, will result in
exorbitant fines being levied by the communications utility.
Another water pipeline alignment consideration is the lateral separation of the line
from adjacent sanitary sewer lines. Many state and local health officials require a minimum of 10 ft of separation (out-to-out) between potable water and sanitary sewer lines.
11.3.3 Rights-of-Way
The final location of a pipeline can only be selected and construction begun once appropriate rights-of-way are acquired. Adequate right-of-way both for construction and for
future access are necessary for a successful installation. Water lines are commonly located in streets and roadways dedicated to the public use. On occasion, it is necessary to
obtain rights-of-way for transmission-type pipelines across private lands. If this is the
case, it is very important to properly evaluate the width of temporary easement that will
be required during construction and the width of permanent easement that will be required
for future access. If a pipeline is to be installed across private property, it is also very
important for the entity that will own and maintain the pipeline to gain agreements that no
permanent structures will be constructed within the permanent easements and to implement a program of monitoring construction on the private property to ensure that access
to the pipeline is maintained. Otherwise, as the property changes hands in the future, the
pipeline stands a good chance of becoming inaccessible.
77.4 PlPlNGMATERlALS
The types of pipe and fittings commonly used for pressurized water distribution systems are
discussed in this section. The types of pipeline materials are first presented and then factors
effecting the types of materials selected by the designer are presented in Section 11.4.7. The
emphasis throughout this section is on pipe 100 mm (4 in.) in diameter and larger.
References to a standard or to a specification are given here in abbreviated form - code
letters and numbers only such as American National Standards Institute (ANSI) B36.10.
Double designations such as ANSI/AWWA C115/A21.15 indicate that American
Waterworks Association (AWWA) C115 is the same as ANSI A21.15. Most standards are
revised periodically, so it is advisable for the designer to obtain the latest edition.
11.4.1 Ductile Iron Pipe (DIP)
Available in sizes 100-1350 mm (4-54 in.), DIP is widely used throughout the United
States in water distribution systems. On the East Coast and in the Midwest, DIP is
commonly used for both smaller distribution mains and larger transmission mains. On
the West Coast, DIP is generally used for distribution pipelines 40 mm (16 in) and
smaller, with alternative pipeline materials often selected for larger pipelines due to
cost. Detailed descriptions of DIP, fittings, joints, installation, thrust restraint, and
other factors relating to design as well as several important ANSI/AWWA specifications are contained in the Ductile Iron Pipe Research Association (DIPRA) handbook
(DIPRA, 1984).
FIGURE 11.1 Couplings and joints for ductile iron pipe: (a) flexible coupling;
(b) mechanical joint; (c) push-on joint; (d) ball joint. Adapted from Sanks et al,. (1989).
True wye
Wye branch
90 bend
reducing
Cross
45 bend
22J bend
90 bend
Tee
Base tee
I1ibend
Base bend
Reducer
FIGURE 11.2 Ductile iron flanged fittings. Adopted from Sanks et al,. (1989).
Special Fittings
standard, shop-applied mortar linings are relatively thin, some designers prefer to specify
shop linings in double thickness. The designer should also be careful in specifying mortar lining thickness to match the pipe inside diameter (ID) with system valve IDs, particularly with short-body butterfly valves where the valve vane protrudes into the pipe. If the
pipe ID is too small, the valve cannot be fully opened.
Steel Pipe*
mm
in
100-250
4-10
1.6
1/16
6.4
300
1-2
1.6
1/16
7.9
350-550
14-22
2.4
3/32
7.9
600
24
2.4
3/32
9.5
750-900
30-36
3.2
1/8
9.5
1050-1350
42-54
3.2
1/8
12.7
*Single thickness per AWWA C104. Linings of double thickness are also readliy available.
tPerAWWAClOS.
1/4
5/16
5/16
3/8
3/8
1/2
TABLE 11.10 Thickness of Cement-Mortal Linings of Pipe in Place per AWWA C602
Nominal Pipe Diameter
mm
in
100-250
300
350-550
600-900
1050-1350
1500
1650-2250
>2250
4-10
12
14-22
24-36
42-54
60
66-90
>90
1/8
3/16
3/16
3/16
1/4
Steel Pipe
Old Pipe
New Pipe
mm
in
mm
in
6.4
6.4
7.9
9.5
9.5
9.5
12.7
12.7
1/4
1/4
5/16
3/8
3/8
3/8
1/2
1/2
4.8
4.8
6.4
6.4
9.5
9.5
11.1
12.7
3/16
3/16
1/4
1/4
3/8
3/8
7/16
1/2
Reference Standard
Recommended Service
Cement mortar
Glass
Epoxy
Fusion-bonded epoxy
Coal-tar epoxy
Coal-tar enamel
Polyurethane
Polyethylene
None
AWWA C210
AWWA C213
AWWA C210
AWWA C203
None
ASTM D 1248
In corrosive soils, the following coatings may be appropriate for protecting the pipe:
0.88
0.75
0.62
0.50
0.40
0.30
0.22
80
90
100
110
120
130
140
Pressure Class at
Safety Factor = 2.5, psi (kPa)
14
18
25
200 (1380)
150 (1030)
100 (690)
Pressure Rating at
Safety Factor = 2.0 psi (kPa)
18
21
25
26
32.5
41
235*
200
165*
160
125
100
larger than 600 mm (24 in.). The principal advantages of steel pipe include high strength,
the ability to deflect without breaking, the ease of installation, shock resistance, lighter
weight than ductile iron pipe, the ease of fabrication of large pipe, the availability of special configurations by welding, the variety of strengths available, and the ease of field
modification.
11.4.3.1 Materials. Conventional nomenclature refers to two types of steel pipe: (1)
mill pipe and (2) fabricated pipe.
Mill pipe includes steel pipe of any size produced at a steel pipe mill to meet finished
pipe specifications. Mill pipe can be seamless, furnace butt welded, electric resistance
welded, or fusion welded using either a straight or spiral seam. Mill pipe of a given size
is manufactured with a constant outside diameter and a variable internal diameter depending on the required wall thickness.
Fabricated pipe is steel pipe made from plates or sheets. It can be either straight or spiral-seam fusion welded pipe, and it can be specified in either internal or external diameters. Spiral-seam, fusion-welded pipe may either mill pipe or fabricated pipe.
Steel pipe may be manufactured from a number of steel alloys with varying yield and
ultimate tensile strengths. Internal working pressure ratings vary from 690 to 17,000 kPa
(100-2500 lb/in2) depending on alloy, diameter, and wall thickness. Steel piping in water
distribution systems should conform to AWWA C200, in which there are many ASTM
standards for materials (see ANSI B31.1 for the manufacturing processes).
11.4.3.2 Available sizes and thicknesses. Sizes, thicknesses, and working pressures for
pipe used in water distribution systems range from 100 m to 3600 mm (4-144 in) as
shown in Table 4-2 of AWWA Mil (American Water Works Association). The standard
length of steel water distribution pipe is 12.2 m (40 ft).
Manufacturers should be consulted for the availability of sizes and thicknesses of steel
pipe. Table 4-2 of AWWA Mil allows a great variety of sizes and thicknesses.
According to ANSI B36.10,
Standard weight (STD) and Schedule 40 are identical for pipes up to 250 mm (10 in).
Standard weight pipe 300 mm (12 in) and larger have walls 0.5 mm (3/8 in) thick. For
standard weight pipe 300 mm (12 in) and smaller, the ID approximately equals the
nominal pipe diameter. For pipe larger than 300 mm (12 in), the outside diameter (OD)
equals the nominal diameter.
Extra strong (XS) and Schedule 80 are identical for pipes up to 200 mm (8 in). All larger sizes of extra strong-weight pipe have walls 12.7 mm (1/2 in) thick.
Double extra strong (XXS) applies only to steel pipe 300 mm (12 in) and smaller,
There is no correlation between XXS and schedule numbers. For wall thicknesses of
XXS, which (in most cases) is twice that of XS, see ANSI B36.10.
For sizes of 350 mm (14 in) and larger, most pipe manufacturers use spiral welding
machines and, in theory, can fabricate pipe to virtually any desired size. In practice, however, most steel pipe manufacturers have selected and built equipment to produce given
ID sizes. Any deviation from manufacturers' standard practices is expensive so it is
always good practice for the designer to consult pipe manufacturers during the design
process. To avoid confusion, the designer should also show a detail of the specified pipe
size on the plans or tabulate the diameters in the specifications. For cement-mortar lined
steel pipelines, AWWA C200, C205, C207, and C208 apply.
Steel pipe must sometimes either be reinforced at nozzles and openings (tees, wye
branches) or a greater wall thickness must be specified. A detailed procedure for deter-
RUBBERGASKET
CARNEGIESHAPE
I. Carnegie-Shape Rubber Gasket Joint With Weld-On Bell Ring
FIGURE 11.4 Welded and rubber-gasketed joints for steel pipe. AWWA Mil (1989)
11.4.3.3 Joints. For buried service, bell and spigot joints with rubber gaskets or mechanical couplings (with or without thrust harnesses) are common. Welded joints are also common for pipe 600 mm (24 in) and larger. Linings are locally destroyed by the heat of welding, so the ends of the pipe must be bare and the linings field applied at the joints. The
reliability of field welds is questionable without careful inspection, but when properly
made they are stronger than other joints. A steel pipeline project specification involving
field welding should always include a carefully prepared section on quality assurance and
testing of the welds. Different types of steel pipe joints are shown in Fig. 11.4.
11.4.3.4 Gaskets. Gaskets for steel flanges are usually made of cloth-inserted rubber
either 1.6 mm (1/16 in) or 3.2 mm (1/8 in) thick and are of two types:
ring (extending from the ID of the flange to the inside edge of the bolt holes);
full face (extending from the ID of the flange to OD).
Gaskets for mechanical and push-on joints for steel pipe are the same as described in
Section 11.4.1.4 for ductile iron pipe.
11.4.3.5 Fittings. For steel pipe 100 mm (4 in) and larger, specifications for steel fittings
can generally be divided into two classes, depending on the joints used and the pipe size:
Flanged, welded (ANSI B16.9)
Wrought Fittings
Caps
45 elbows
90 elbows, long radius
90 elbows, short radius
90 reducing elbows, long
radius
Multiple-outlet fittings
Blind flanges
Lap joint flanges
Slip-on flanges
Socket-type welding flanges
Reducing flanges
Threaded flanges
Welding neck flanges
Reducers
Eccentric reducers
180 returns, long radius
Saddles
Reducing outlet tees
Split tees
Straight tees
True wyes
Tee
Long elbow
Short elbow
45 bend
Cross
30 bend
Cap
Eccentric
reducer
180 return
90 long
radius elbow
45 elbow
90 short
radius elbow
Concentric
reducer
Straight tee
Reducing elbow,
long radius
Blind flange
FIGURE 11.6 Wrought (forged) steel fittings for use with
welded flanges. Adopted from Sanks et al,. (1989)
Fabricated (AWWA C208).
Fittings larger than 100 mm (4 in) should conform to ANSI B 16.9 ("smooth" or
wrought) or AWWA C208 (mitered). Threaded fittings larger than 100 mm (4 in.) should
be avoided. The ANSI B16.9 fittings are readily available up to 300-400 mm (12-16 in)
in diameter. Mitered fittings are more readily available and cheaper for larger fittings.
The radius of a mitered elbow can range from 1 to 4 pipe diameters. The hoop tension
concentration on the inside of elbows with a radius less than 2.5 pipe diameters may
exceed the safe working stress. This tension concentration can be reduced to safe levels
by increasing the wall thickness, as described in ANSI B31, AWWA C208, and Piping
Engineering (Tube Turns Division, 1974). Design procedures for mitered bends are
described in ANSI B31.1 and B31.3. Types of steel fittings are shown in Table 11.15 and
in Figures 11.5 and 11.6.
11.4.3.6 Linings and coatings. Cement mortar is an excellent lining for steel pipe.
Tables 11.9 and 11.10 show required thicknesses for steel pipe.
Steel pipe can also be coated with cement mortar. Recommended mortar coating thicknesses are shown in AWWA C205. These thicknesses, however, are often thinner than
those required to provide adequate protection. Many designers specify a minimum cement
mortar coating thickness of at least 19 mm (3/4 in).
In corrosive soils, the following coatings may be appropriate for protecting steel pipe:
As another alternative, epoxy-lined and -coated steel pipe can be used. Because this
lining is only 0.3-0.6 mm (0.12-0.20 in) thick, the ID of the bare pipe is only slightly
reduced by such linings. Epoxy-lined steel pipe is covered by AWWA C203, C210, and
C213 standards. Before specifying epoxy lining and coating, pipe suppliers must be consulted to determine the limitations of sizes and lengths of pipe that can be lined with
epoxy. Flange faces should not be coated with epoxy if flanges with serrated finish per
AWWA C207 are specified.
11.4.4 Reinforced Concrete Pressure Pipe (RCPP)
Several types of RCPP are manufactured and used in North America. These include steel
cylinder (AWWA C300), prestressed, steel cylinder (AWWA C301), noncylinder (AWWA
AHWA
Standard
Steel
Cylinder
Reinforcement
Design
Basis*
Steel cylinder
Prestressed, steel cylinder
Noncylinder
Pretensioned, steel cylinder
C300
C301
C302
C303
X
X
None
X
Rigid
Rigid
Rigid
Semirigid
"Rigid" and "cemirigid" are terms used in AWWA M9 and are intended to differentiate between two design theories. Rigid pipe does not depend on the passive resistance of the soil adjacent to the pipe for support of vertical
loads. Semirigid pipe requires passive soil resistance for vertical load support. The terms "rigid" and "semirigid"
as used here should not be confused with the definitions stated by Marston in Iowa State Experiment Station
Bulletin No. 96.
Rod Reinforcement or
Welded Wire Fabric
Rubber Gasket
Steel Cylinder and Supplemental
Cement Mortar Placed in Rod or Wire Fabric Interior
Field or Other Protection Reinforcement if Necessary
C302), and pretensioned, steel cylinder (AWWA 303). Some of these types are made for
a specific type of service condition and others are suitable for a broader range of service
conditions. A general description of RCPP types is shown in Table 11.16. The designer
should be aware that not all RCPP manufactures make all of the pipe types listed.
11.4.4.1 Steel cylinder pipe, AWWA C300. Prior to the introduction of prestressed steel
cylinder pipe (AWWA C301), in the early 1940s, most of the RCPP in the United States
was steel cylinder type pipe. New installations of steel cylinder pipe have been declining
over the years as AWWA C301 and C303 pipes have gained acceptance. Steel cylinder
pipe is manufactured in diameters of 750-3600 mm (30-44 in). Standard lengths are
3.6-7.2 meters (12-24 ft).
AWWA C300 limits the reinforcing steel furnished in the cage(s) to no less 40 percent
of the total reinforcing steel in the pipe. The maximum loads and pressures for this type
of pipe depend on the pipe diameter, wall thickness, and strength limitations of the concrete and steel. The designer should be aware that this type of pipe can be designed for
high internal pressure, but is limited in external load capacity.
A cross section of AWWA C300 pipe and a typical joint configuration is shown in
Fig. 11.7.
11.4.4.2 Prestressed steel cylinder pipe, AWWA C301. Prestressed steel cylinder pipe
has been manufactured in the United States since 1942 and is the most widely used type
of concrete pressure pipe, except in the western United States. Due to cost considerations,
AWWA C301 pipe is often used for high-pressure transmission mains, but it also has been
for distribution mains and for many other low and high pressure uses.
A distressing number of failures of this pipe occurred in the United States primarily
during the 1980s. The outer shell of the concrete cracked, allowing the reinforcement to
corrode and subsequently fail. These failures have resulted in significant revisions to the
standards covering this pipe's design. Even so, the designer should not necessarily depend
solely on AWWA specifications or on manufacturers' assurances, but should make a careful analysis of internal pressure (including waterhammer) and external loads. Make certain that the tensile strain in the outer concrete is low enough so that cracking will either
not occur at all or will not penetrate to the steel under the worst combination of external
and internal loading.
Concrete Core
Steel Spigot Ring
Concrete Core
Rubber Gasket
Prestress Wire
Steel Cylinder
Prestressed cylinder pipe has the following two general types of fabrication: (1) a steel
cylinder lined with a concrete core, or (2) a steel cylinder embedded in concrete core.
Lined cylinder pipe is commonly available in IDs from 400 to 1200 mm (16-48 in). Sizes
through 1500 mm (60 in) are available through some manufacturers. Embedded cylinder
pipe is commonly available in inside diameters 1200 mm (48 in) and larger. Lengths are
generally 4.9-7.3 m (16-24 ft), although longer units can be furnished.
AWWA C304, Standard for Design of Prestressed Concrete Cylinder Pipe covers the
design of this pipe. The maximum working pressure for this pipe is normally 2758 kPa
(400 psi). The design method is based on combined loading conditions (the most critical
type of loading for rigid pipe) and includes surge pressure and live loads.
Cross sections of AWWA C301 pipe (lined and embedded) and typical joint configurations are shown in Fig. 11.8.
Steel Cages
Reinforcing Steel Skirt
Grout
InstalJoilantitoAfter
n
Steel Skirt
Rubber Gasket
B. AWWA C302-type pipe with concrete joint rings
FIGURE 11.9 Cross section of AWWA C302 pipe. AWWA M9 (1995)
11.4.4.3 Noncylinder pipe, AWWA C302. The maximum working pressure of noncylinder pipe is 379 kPa (55 psi) and is generally not suitable for typical municipal systems.
Noncylinder pipe is commonly furnished in diameters of 300 to 3600 mm (12-144 in),
but larger diameters can be furnished if shipping limitations permit. Standard lengths are
2.4-7.3 m (8-24 ft) with AWWA C302 limiting the maximum length that can be furnished
for each pipe size.
Cross sections of AWWA C302 pipe with steel and concrete joint ring configurations
are shown in Fig. 11.9.
11.4.4.4 Pretensioned steel cylinder, AWWA C303. Pretensioned steel cylinder, or
commonly called concrete cylinder pipe (CCP), is manufactured in Canada and in the
western and southwestern areas of the United States. It is commonly available in diameters of 300-1350 mm (12-54 in). Standard lengths are generally 7.3 to 12.2 m(24-40
ft). With maximum pressure capability up to 2758 kPa (400 psi), the longer laying
length, and the overall lighter handling weight, AWWA C303 is a popular choice among
many designers for various applications including municipal transmission and distributions mains.
Manufacture of CCP begins with a fabricated steel cylinder with joint rings which is
hydrostatically tested. A cement-mortar lining is then placed by the centrifugal process
inside the cylinder. The nominal lining thickness is 13 mm (1/2 in) for sizes up to and
including 400 mm (16 in), and 19 mm (3/4 in) for larger sizes. After the lining is cured,
the cylinder is wrapped, typically in a helical pattern, with a smooth, hot-rolled steel bar,
using a moderate tension in the bar. The size and spacing of the bar, as well as the thickness of the steel cylinder, are proportioned to provide the required pipe strength. The
cylinder and bar wrapping are then covered with a cement slurry and a dense mortar coating that is rich in cement.
Rod Reinforcement
Grout Joint After
Steel Bell Ring
Installation
Steel Cylinder
Cement Mortar Placed in
Field or Other Protection
Rubber Gasket
Pressure Class,
Safety Factor = 2.0,psi (kPa)
160
130
100
80
(1100)
(900)
(690)
(550)
*These DRs are from the standard dimension ratio series established
by ASTM F 412.
HDPE pipe is rated for pressure capacity at 230C (73.40F). Being a thermoplastic, the
pressure capacity of HDPE pipe is related to its operating temperature. Through the normal range of municipal water system temperatures, 0C-24 (32 F-75 F), the pressure
rating of HDPE remains relatively constant. As the operating temperature rises above
230C (73.40F), however, the pressure capacity of HDPE pipe decreases to a level below
its pressure class. The pressure rating for HDPE pipe at service temperature of 6O0C
(14O0F) would be about half its rating at 230C (73.40F).
11.4.5.2 Available sizes and thicknesses. AWWA C906 covers HDPE pipe in sizes 100
-1600 mm (4-63 in). The design, according to AWWA C906, of HDPE pipe is similar to
AWWA C900 PVC pipe in that HDPE pipe is rated according to "pressure classes." The
pressure classes detailed in AWWA C906 include allowance for pressure rises above
working pressure due to occasional positive pressure transients not exceeding two times
the nominal pressure class and recurring pressure surges not exceeding one and one-half
times the nominal pressure class. AWWA C906 lists HDPE pipe sizes according to the
IPS (steel pipe) and the ISO (metric) sizing systems. Ductile iron pipe sizes are also
available.
As with AWWA C900 and C905 (PVC pipe), AWWA C906 contains dimension ratio
(outside diameter to wall thickness) specifications. Table 11.17 presents dimension ratios
with corresponding pressure classes as defined in AWWA C906 for commonly available
HDPE pipe.
11.4.5.3 Joints. HDPE pipe can be joined by thermal butt-fusion, flange assemblies, or
mechanical methods as may be recommended by the pipe manufacturer. HDPE is not to
be joined by solvent cements, adhesives (such as epoxies), or threaded-type connections.
Thermal butt-fusion is the most widely used method for joining HDPE piping. This procedure uses portable field equipment to hold pipe and/or fittings in close alignment while
the opposing butt-ends are faced, cleaned, heated and melted, fused together, and then
cooled under fusion parameters recommended by the pipe manufacturer and fusion equipment supplier.
For each polyethylene material there exists an optimum range of fusion conditions,
such as fusion temperature, interface pressure, and cooling time. Thermal fusion should
only be conducted by persons who have received training in the use of the fusion equipment according to the recommendations of the pipe manufacturer and fusion equipment
supplier. In situations where different polyethylene piping materials must be joined by
thermal butt-fusion process, both pipe manufacturers should be consulted to determine the
appropriate fusion procedures. ASTM D 2657 covers thermal butt-fusion of HDPE pipe.
HDPE pipe is normally joined above ground and then placed in the pipeline trench. The
thermal butt-fusion joint is not subject to movement due to thrust and does not require
thrust restraint such as thrust blocks.
Flanged and mechanical joint adapters are available for joining HDPE pipe to valves
and ductile iron fittings. The designer should always consult the pipe manufacturer to
insure a proper fit between pipe and fittings. The designer should also make sure that,
when connecting to a butterfly valve, the valve disc will freely swing to the open position
without hitting the face of the stub end or flange adapter.
11.4.5.4 Gaskets. Gaskets are not necessary for HDPE pipe using thermal butt-fusion
joints.
11.4.5.5 Fittings. AWWA C906 HDPE pipe for municipal use is manufactured in ductile
iron pipe OD sizes, so ductile iron fittings, conforming to AWWA CIlO, can be used in
all available sizes. See Sec. 11.4.1.5 for discussion on ductile iron fittings. Although not
widely used, HDPE fittings, in configurations similar to ductile iron fittings, are also
available. AWWA C906 covers HDPE pressure pipe fittings.
11.4.5.6 Linings and coatings. HDPE pipe does not require lining or coating.
11.4.6 Asbestos-Cement Pipe (ACP)
Asbestos-cement pipe (ACP), available in the United States since 1930, is made by mixing Portland cement and asbestos fiber under pressure and heating it to produce a hard,
strong, yet machinable product. It is estimated that over 480,000 km (300,000 mi) of ACP
is now in service in the United States.
In the late 1970s, attention was focused on the hazards of asbestos in the environment
and, particularly, in drinking water. There was significant debate of the issue with one set
of experts advising of the potential dangers and a second set of experts claiming that pipes
made with asbestos do not result in increases in asbestos concentrations in the water.
Studies have shown no association between water delivered by ACP and any general disease, however, the general fear that resulted from the controversy had a tremendous negative impact on ACP use in the United States. The debate on the health concerns of using
ACP along with the introduction of PVC pipe into the municipal water system market, has
reduced the use of ACP significantly in the past several years.
11.4.6.1 Available sizes and thicknesses. ACP is available in diameters of 100-1050 mm
(4-42 in). Refer to ASTM C 296 and AWWA C401, C402, and C403 for thickness and
pressure ratings and AWWA C401 and C403 for detailed design procedures. AWWA C401
for 100-400 mm (4-16 in) pipe is similar to AWWA C403 for 450-1050 mm (18 -42 in)
pipe. The properties of asbestos-cement for distribution pipe (AWWA C400) and transmission pipe (AWWA C402) are identical. However, under AWWA C403 (transmission
pipe) the suggested minimum safety factor is 2.0 for operating pressure and 1.5 for external loads, whereas the safety factors under AWWA C402 (distribution pipe) are 4.0 and
2.5, respectively. So the larger pipe has the smaller safety factors.
Section 4 in AWWA C403 justifies this on the basis of surge pressure in large pipe
tending to be less than those in small pipes. However, surge pressures are not necessarily a function of pipe diameter (Chap. 12). The operating conditions, including surge
pressures, for any proposed pipeline installation should be closely evaluated before the
pipe class is selected. It is the engineer's design prerogative to select which of the safety factors should apply. AWWA C400 specifies that safety factors should be no less than
4.0 and 2.5 if no surge analysis is made. The low safety factors given in AWWA C403
should only be used if all loads (external, internal, and transient) are carefully and accurately evaluated.
11.4.6.2 Joints and fittings. The joints are usually push-on, twin-gasketed couplings,
although mechanical and rubber gasket push-on joints can be used to connect ACP to ductile iron fittings.
Ductile iron fittings conforming to ANSI/AWWA C110/A21.10 are used with ACP,
and adapters are available to connect ACP to flanged or mechanical ductile iron fittings.
Fabricated steel fittings with rubber gasket joints can also be used.
11.4.7 Pipe Material Selection
Buried piping for municipal water transmission and distribution must resist internal pressure, external loads, differential settlement, and corrosive action of both soils and, potentially, the water it carries. General factors to be considered in the selection of pipe include
the following:
Service conditions
-
Availability
- Local availability and experienced installation personnel
- Sizes and thicknesses (pressure ratings and classes)
- Compatibility with available fittings
Properties of the pipe
-
Economics
- Cost (installed cost, including freight to job site and installation)
- Required life
- Cost of maintenance and repairs
The items listed above are general factors relating to pipe selection to be considered during the design of any pipeline. Since most municipal water system projects are
either let out to competitive bid or are installed as a part of private land development,
Advantages
DisavantagesAimitations
2
the designer will find that installed cost, lacking specific service conditions that
require otherwise, will tend to dictate pipe selection. For example, steel pipe and reinforced concrete pressure pipe are both available in 300 mm (12 in) diameter. However,
the installed cost of ductile iron pipe or PVC pipe is typically lower (typical municipal use) in the 300 mm (12 in) size. Therefore, if the service conditions do not require
the high pressure capabilities of steel or reinforced concrete pressure pipe, the logical
choice for 300 mm pipe (12 in) will optionally be ductile iron or PVC. Conversely, if
the proposed pipeline is 900 mm (36 in) in diameter, the installed cost of both steel and
reinforced concrete pressure pipe, depending on location, tend to be much more competitive.
A general comparison of the various types of pipe used in municipal water systems is
shown in Table 11.18.
77.5 PIPELINEDESIGN
This section will address typical issues that are addressed during the design of water distribution and transmission pipelines. Pressure pipelines must primarily be able to resist
internal pressures, external loads (earth and impact loads), forces transferred along the
pipe when pipe-to-soil friction is used for thrust restraint, and handling during construction. Each of these design issues will be discussed and appropriate formulas presented.
11.5.1 Internal Pressures
The internal pressure of a pipeline creates a circumferential tension stress, frequently
termed hoop stress, that governs the pipeline thickness. In other words, the pipe must be
thick enough to withstand the pressure of the fluid within. The internal pressure used in
design should be that to which the pipe may subjected during its lifetime. In a distribution
system this pressure may be the maximum working pressure plus and allowance for surge.
It may be also the pipeline testing pressure or the shutoff head of an adjacent pump. In a
transmission pipeline, the pressure is measured by the vertical distance between the pipe
centerline and the hydraulic grade line. Potential hydraulic grade lines on transmission
pipelines should be carefully considered. The static hydraulic grade line is potentially
much higher than the dynamic grade line if a downstream valve is closed.
Hoop tensile stress is given by the equation
*= ?
<!>
where s = allowable circumferential stress in kPa (lb/in2) p = pressure in kPa (lb/in2) D
= outside diameter of pipe in mm (in) t = thickness of the pipe in mm (in)
It should be noted that this equation is the basis for determining the circumferential stress
in steel and reinforced pressure pipe and for determining the pressure classes and pressure
ratings for virtually all other different types of pressure pipe.
11.5.2 Loads on Buried Pipe
Buried pipes must support external superimposed loads, including the weight of the soil
above plus any live loads such as wheel loads due to vehicles or equipment. The two
broad categories for external structural design are rigid and flexible pipe. Rigid pipe sup-
Next Page
CHAPTER 12
HYDRAULIC TRANSIENT
DESIGN FOR
PIPELINE SYSTEMS
C. Samuel Martin
School of Civil and Environmental Engineering
Georgia Institute of Technology
Atlanta, Georgia
12.1
By definition, waterhammer is a pressure (acoustic) wave phenomenon created by relatively sudden changes in the liquid velocity. In pipelines, sudden changes in the flow
(velocity) can occur as a result of (1) pump and valve operation in pipelines, (2) vapor
pocket collapse, or (3) even the impact of water following the rapid expulsion of air out
of a vent or a partially open valve. Although the name waterhammer may appear to be a
misnomer in that it implies only water and the connotation of a "hammering" noise, it has
become a generic term for pressure wave effects in liquids. Strictly speaking, waterhammer can be directly related to the compressibility of the liquidprimarily water in this
handbook. For slow changes in pipeline flow for which pressure waves have little to no
effect, the unsteady flow phenomenon is called surging.
Potentially, waterhammer can create serious consequences for pipeline designers if not
properly recognized and addressed by analysis and design modifications. There have been
numerous pipeline failures of varying degrees and resulting repercussions of loss of property and life. Three principal design tactics for mitigation of waterhammer are (1) alteration
of pipeline properties such as profile and diameter, (2) implementation of improved valve
and pump control procedures, and (3) design and installation of surge control devices.
In this chapter, waterhammer and surging are defined and discussed in detail with reference to the two dominant sources of waterhammerpump and/or valve operation.
Detailed discussion of the hydraulic aspects of both valves and pumps and their effect on
hydraulic transients will be presented. The undesirable and unwanted, but often potentially possible, event of liquid column separation and rejoining are a common justification for
surge protection devices. Both the beneficial and detrimental effects of free (entrained or
entrapped) air in water pipelines will be discussed with reference to waterhammer and
surging. Finally, the efficacy of various surge protection devices for mitigation of waterhammer is included.
12.2
a0O
/ 1 /^ 1 \
- TT^f- T^IT
V
e E
v
e E
where E is the elastic modulus of the pipe wall, D is the inside diameter of the pipe, e is
the wall thickness, and a0 is the acoustic velocity in the liquid medium. In a very rigid pipe
or in a tank, or in large water bodies, the acoustic velocity a reduces to the well=known
relationship a = a0 = V(/p). For water K = 2.19 GPa (318,000 psi) and p = 998 kg/m3
(1.936 slug/ft3), yielding a value of a0 = 1483 m/sec (4865 ft/sec), a value many times that
of any liquid velocity V.
12.2.3 Joukowsky (Waterhammer) Equation
There is always a pressure change A/? associated with the rapid velocity change AV across
a waterhammer (pressure) wave. The relationship between Ap and AV from the basic
physics of linear momentum yields the well-known Joukowsky equation
Ap = -paAV
(12.2)
where p is the liquid mass density, and a is the sonic velocity of the pressure wave in the
fluid medium in the conduit. Conveniently using the concept of head, the Joukowsky head
rise for instantaneous valve closure is
AH=Ap = _poAV=oK
P
98
g
The compliance of a conduit or pipe wall can have a significant effect on modification
of (1) the acoustic velocity, and (2) any resultant waterhammer, as can be shown from Eq.
(12.1) and Eq. (12.2), respectively. For simple waterhammer waves for which only radial
pipe motion (hoop stress) effects are considered, the germane physical pipe properties are
Young's elastic modulus (E) and Poisson ratio (|n). Table 12.1 summarizes appropriate values of these two physical properties for some common pipe materials.
The effect of the elastic modulus (E) on the acoustic velocity in water-filled circular
pipes for a range of the ratio of internal pipe diameter to wall thickness (Die) is shown in
Fig. 12.1 for various pipe materials.
12.3
Valves are integral elements of any piping system used for the handling and transport of
liquids. Their primary purposes are flow control, energy dissipation, and isolation of portions of the piping system for maintenance. It is important for the purposes of design and
final operation to understand the hydraulic characteristics of valves under both steady and
unsteady flow conditions. Examples of dynamic conditions are direct opening or closing
of valves by a motor, the response of a swing check valve under unsteady conditions, and
the action of hydraulic servovalves. The hydraulic characteristics of valves under either
noncavitating or cavitating conditions vary considerably from one type of valve design to
another. Moreover, valve characteristics also depend upon particular valve design for a
special function, upon absolute size, on manufacturer as well as the type of pipe fitting
employed. In this section the fundamentals of valve hydraulics are presented in terms of
pressure drop (headloss) characteristics. Typical flow characteristics of selected valve
types of controlgate, ball, and butterfly, are presented.
TABLE 12.1 Physical Properties of Common Pipe Materials
Material
Young's Modulus
E (GPa)
Asbestos cement
Cast iron
Concrete
Concrete (reinforced)
Ductile iron
Polyethylene
PVC (polyvinyl chloride)
Steel
23-24
80-170
14-30
30-60
172
0.7-0.8
2.4-3.5
200-207
Poisson's Ratio
|Ll
0.25-0.27
0.10-0.15
0.30
0.46
0.46
0.30
Steel
207
!Al
uminum
: 68.9
Concrete:
27:6 ;
PVC
2.8
Polyethlene
! 0.69 i
Steel
30(10s)
Copper6
18(10 )
Aluminum
10(1O8)
Concrete
4(106)
PVC
0.4(106V
Polyethlene
: 0.1 no8) i
FIGURE 12.1 Effect of wall thickness of various pipe materials on acoustic velocity
in water pipes.
FIGURE 12.2 Cross sections of selected control valves: (From Wood and Jones, 1973).
between upstream and downstream uniform flow sections, the headloss across the valve
is expressed in terms of the pipe velocity and a headloss coefficient K1
*H = KL
(12.4)
(12.5)
where
H = A# + -J(12-6)
2g
Both discharge coefficients are defined in terms of the nominal full-open valve area
Avo and a representative head, A/f for Cf and H for CQJ the latter definition generally
reserved for large valves employed in the hydroelectric industry. The interrelationship
b
e
t
w
e
e
n
C, Cp, and KL is
1
1 C 2
1
^ = i = C^
(12'7>
F
/
Frequently valve characteristics are expressed in terms of a dimensional flow coefficient Cv from the valve industry
Q = CvVZp
(12.8)
where Q is in American flow units of gallons per minute (gpm) and A/? is the pressure loss
in pounds per square inch (psi). In transient analysis it is convenient to relate either the
loss coefficient or the discharge coefficient to the corresponding value at the fully open
valve position, for which Cf = Cfo. Hence,
<L = SL JM-= v fM.
Q0 CfjAH0
^AH0
((i2.9)}
Gate
Area Ratio AyA^1
l
D/DBal
b=0.629
Butterfly
Gate
l
D/DBal
b=0.623
Butterfly
AV = O V0 , where V0 is the initial velocity of liquid in the pipe. Although Eq. (12.2)
applies across every wavelet, the effect of complete valve closure over a period of time
greater than 2LIa, where L is the distance along the pipe from the point of wave creation
to the location of the first pipe area change, can be beneficial. Actually, for a simple
pipeline the maximum head rise remains that from Eq. (12.3) for times of valve closure
tc ^ 2LIa, where L is the length of pipe. If the value of tc > 2LIa, then there can be a considerable reduction of the peak pressure resulting from beneficial effects of negative wave
reflections from the open end or reservoir considered in the analysis. The phenomenon
can still be classified as waterhammer until the time of closure tc 2LIa, beyond which
time there are only inertial or incompressible deceleration effects, referred to as surging,
also known as rigid column analysis. Table 12.2 classifies four types of valve closure,
independent of type of valve.
Type of Closure
O
< 2LIa
> 2LIa
2L/a
Instantaneous
Rapid
Gradual
Slow
Maximum Head
&Hm
a
K/8
aVJg
< aV0/g
aV0/g
Phenomenon
Waterhammer
Waterhammer
Waterhammer
Surging
Using standard waterhammer programs, parametric analyses can be conducted for the
preparation of charts to demonstrate the effect of time of closure, type of valve, and an
indication of the physical processwaterhammer or simply inertia effects of deceleration.
The charts are based on analysis of valve closure for a simple reservoir-pipe-valve
arrangement. For simplicity fluid friction is often neglected, a reasonable assumption for
pipes on the order of hundreds of feet in length.
72.4
\ (P
\ V2
V2
"-(11H-(H+S-S
where subscripts s and d refer to suction and discharge sides of the pump, respectively,
12.4.2 Homologous (Affinity) Laws
Dynamic similitude, or dimensionless representation of test results, has been applied with
perhaps more success in the area of hydraulic machinery than in any other field involving
fluid mechanics. Due to the sheer magnitude of the problem of data handling it is imperative that dimensionless parameters be employed for transient analysis of hydraulic
machines that are continually experiencing changes in speed as well as passing through
several zones of normal and abnormal operation. For liquids for which thermal effects
may be neglected, the remaining fluid-related forces are pressure (head), fluid inertia,
resistance, phase change (cavitation), surface tension, compressibility, and gravity. If the
discussion is limited to single-phase liquid flow, three of the above fluid effectscavitation, surface tension, and gravity (no interfaces within machine)can be eliminated, leaving the forces of pressure, inertia, viscous resistance, and compressibility. For the steady
or even transient behavior of hydraulic machinery conducting liquids the effect of compressibility may be neglected.
In terms of dimensionless ratios the three forces yield an Euler number (ratio of inertia
force to pressure force), which is dependent upon geometry, and a Reynolds number.
For all flowing situations, the viscous force, as represented by the Reynolds number,
is definitely present. If water is the fluid medium, the effect of the Reynolds number
on the characteristics of hydraulic machinery can usually be neglected, the major
exception being the prediction of the performance of a large hydraulic turbine on the
basis of model data. For the transient behavior of a given machine the actual change in
the value of the Reynolds number is usually inconsequential anyway. The elimination
of the viscous force from the original list reduces the number of fluid-type forces from
seven to two-pressure (head) and inertia, as exemplified by the Euler number. The
appellation geometry in the functional relationship in the above equation embodies primarily, first, the shape of the rotating impeller, the entrance and exit flow passages,
including effects of vanes, diffusers, and so on; second, the effect of surface roughness;
and lastly the geometry of the streamline pattern, better known as kinematic similitude
in contrast to the first two, which are related to geometric similarity. Kinematic similarity is invoked on the assumption that similar flow patterns can be specified by congruent velocity triangles composed of peripheral speed U and absolute fluid velocity V
at inlet or exit to the vanes. This allows for the definition of a flow coefficient,
expressed in terms of impeller diameter D1 and angular velocity co:
C
= J?
(12 12)
The reciprocal of the Euler number (ratio of pressure force to inertia force) is the head
coefficient, defined as
Q = J^
(12.13)
"
(1114)
WD?
For transient analysis, the desired parameter for the continuous prediction of pump speed
is the unbalanced torque T. Since T = P/co, the torque coefficient becomes
C
(12 15)
T = WD?
'
JL = ^L = JL??*.
a2 CHR HR co2
p _ cr _ y CQ/
a2 CTR TR co2
Pumping
Turbine (III)
Dissipation (IV)
reversal caused by the positive head on the machine. This purely dissipative mode results
in a negative or zero efficiency. It is important to note that both the head and fluid torque
are positive in Zone H, the only zone in Quadrant IV.
Quadrant III. A machine that passes through Zone H during a pump power failure will
then enter Zone G (normal turbining) provided that reverse shaft rotation is not precluded by a mechanical ratchet. Although a runaway machine rotating freely is not generating
power, Zone G is the precise mode of operation for a hydraulic turbine. Note that the head
and torque are positive, as for a pump but that the flow and speed are negative, opposite
to that for a pump under normal operation (Zone A).
Subsequent to the tripout or load rejection of a hydraulic turbine or the continual operation of a machine that failed earlier as a pump, Zone F (energy dissipation) can be
encountered. The difference between Zones F and G is that the torque has changed sign
for Zone F, resulting in a braking effect, which tends to slow the free=wheeling machine
down. In fact the real runaway condition is attained at the boundary of the two zones, for
which torque J = O .
Quadrant IL The two remaining zonesD and Eare very unusual and infrequently encountered in operation, with the exception of pump/turbines entering Zone E during
transient operation. Again it should be emphasized that both zones can be experienced by
a pump in a test loop, or in practice in the event a machine is inadvertently rotated in the
wrong direction by improper wiring of an electric motor. Zone D is a purely dissipative
mode that normally would not occur in practice unless a pump, which was designed to
increase the flow from a higher to lower reservoir, was rotated in reverse, but did not have
the capacity to reverse the flow (Zone E, mixed or axial flow), resulting in Q > O, TV < O,
T < O, for H < O. Zone E, for which the pump efficiency > O, could occur in practice
under steady flow if the preferred rotation as a pump was reversed. There is always the
question regarding the eventual direction of the flow. A radial-flow machine will produce
positive flow at a much reduced capacity and efficiency compared to TV > O (normal
pumping), yielding of course H > O. On the other hand, mixed and axial-flow machines
create flow in the opposite direction (Quadrant III), and H < O, which corresponds still to
an increase in head across the machine in the direction of flow.
Karman-Knapp Quadrant
Karman-Knapp Quadrant
In computer analysis of pump transients, Figs. 12.7 and 12.8, while meaningful from
the standpoint of physical understanding, are fraught with the difficulty of Iv/ocl becoming
infinite as the unit passes through, or remains at, zero speed (a = O). Some have solved
that problem by switching from /i/oc2 versus via to h/v2 versus o/v, and likewise for (3, for
Iv/ocl > 1. This technique doubles the number of curves on Figs. 12.7 and 12.8, and thereby creates discontinuities in the slopes of the lines at Iv/ocl = 1, in addition to complicating the storing and interpolation of data. Marchal et al. (1965) devised a useful transformation which allowed the complete pump characteris-tics to be represented by two single
curves, as shown for the same pump in Fig. 12.10. The difficulty of via becoming infinite
was eliminated by utilizing the function tan^1 (v/oc) as the abscissa. The eight zones, or
four quadrants can then be connected by the continuous functions. Although some of the
physical interpretation of pump data has been lost in the transformation, Fig. 12.10 is now
a preferred correlation for transient analysis using a digital computer because of function
continuity and ease of numerical interpolation. The singularities in Figs. 12.7 and 12.8 and
the asymptotes in Fig. 12.9 have now been avoided.
12.4.5 Critical Data Required for Hydraulic Analysis
of Systems with Pumps
Regarding data from manufacturers such as pump curves (normal and abnormal), pump and
motor inertia, motor torque-speed curves, and valve curves, probably the most critical for
pumping stations are pump-motor inertia and valve closure time. Normal pump curves are
KARNIAN-KNAPP
QUADRANT
I
Zone
Karman - Knapp
Quadrant
12.5
There are numerous techniques for controlling transients and waterhammer, some
involving design considerations and others the consideration of surge protection
devices. There must be a complete design and operational strategy devised to combat
potential waterhammer in a system. The transient event may either initiate a low-pressure event (downsurge) as in the case of a pump power failure, or a high pressure event
(upsurge) caused by the closure of a downstream valve. It is well known that a downsurge can lead to the undesirable occurrence of water-column separation, which itself
can result in severe pressure rises following the collapse of a vapor cavity. In some systems negative pressures are not even allowed because of (1) possible pipe collapse or
(2) ingress of outside water or air.
The means of controlling the transient will in general vary, depending upon whether
the initiating event results in an upsurge or downsurge. For pumping plants the major
cause of unwanted transients is typically the complete outage of pumps due to loss of electricity to the motor. For full pipelines, pump startup, usually against a closed pump discharge valve for centrifugal pumps, does not normally result in significant pressure transients. The majority of transient problems in pumping installations are associated with the
potential (or realized) occurrence of water-column separation and vapor-pocket collapse,
resulting from the tripout of one or more pumps, with or without valve action. The pumpdischarge valve, if actuated too suddenly, can even aggravate the downsurge problem. To
Steady-State HGL
Compressed Gas
Air Chamber
Steady-State HGL
Dump
Steady-State HGL
FIGURE 12.11 Schematic of various surge protection devices for pumping installations
combat the downsurge problem there are a number of options, mostly involving the design
and installation of one or more surge protection devices. In this section various surge protection techniques will be discussed, followed by an assessment of the virtue of each with
respect to pumping systems in general. The lift systems shown in Fig. 12.11 depict various surge protection schemes.
12.5.1 Critical Parameters for Transients
Before discussing surge protection devices, some comments will be made regarding the
various pipeline, pump and motor, control valve, flow rate, and other parameters that
affect the magnitude of the transient. For a pumping system the four main parameters are
(1) pump flow rate, (2) pump and motor WR2, (3) any valve motion, and (4) pipeline characteristics. The pipeline characteristics include piping layoutboth plan and profile
pipe size and material, and the acoustic velocity. So-called short systems respond differently than long systems. Likewise, valve motion and its effect, whether controlled valves
or check valves, will have different effects on the two types of systems.
The pipeline characteristicsitem number (4)relate to the response of the system to
a transient such as pump power failure. Clearly, the response will be altered by the addition of one or more surge protection device or the change of (1) the flow rate, or (2) the
WR2, or (3) the valve motion. Obviously, for a given pipe network and flow distribution
there are limited means of controlling transients by (2) WR2 and (3) valve actuation. If
these two parameters can not alleviate the problem than the pipeline response needs to be
altered by means of surge protection devices.
HGL
Orifice
Simple
Simple
One-Way
Check Valve
Check
valve
Orifice
Air Chamber
Accumulator
Vacuum Breaker
FIGURE 12.12 Cross sectional view of surge tanks and gas-n related surge protection devices
Check valve slam is also a possibility from stoppage or failure of one pump of several in a parallel system, or resulting from the action of an air chamber close to a pump
undergoing power failure. Check valve slam can be reduced by the proper selection of a
dashpot.
Surge anticipator valves and surge relief valves. A surge anticipation valve, Fig.
12.13c frequently installed at the manifold of the pump station, is designed to open initially under (1) pump power failure, or (2) the sensing of underpressure, or (3) the sensing of overpressure, as described by Lescovitch (1967). On the other hand, the usual type
of surge relief valve opens quickly on sensing an overpressure, then closes slowly, as controlled by pilot valves. The surge anticipation valve is more complicated than a surge
relief valve in that it not only embodies the relief function at the end of the cycle, but also
has the element of anticipation. For systems for which water-column separation will not
occur, the surge anticipation valve can solve the problem of upsurge at the pump due to
reverse flow or wave reflection, as reported in an example by White (1942). An example
of a surge relief valve only is provided by Weaver (1972). For systems for which watercolumn separation will not occur, Lundgren (1961) provides charts for simple pipeline
systems.
As reported by Parmakian (1968,1982a-b) surge anticipation valves can exacerbate the
downsurge problem inasmuch as the opening of the relief valve aggravates the negative
pressure problem. Incidents have occurred involving the malfunctioning of a surge anticipation valve, leading to extreme pressures because the relief valve did not open.
Pump bypass. In shorter low-head systems a pump bypass line (Fig. 12.11) can be
installed in order to allow water to be drawn into the pump discharge line following power
failure and a downsurge. As explained by Wylie and Streeter (1993), there are two possible bypass configurations. The first involves a control valve on the discharge line and a
check valve on the bypass line between the pump suction or wet well and the main line.
The check valve is designed to open subsequent to the downsurge, possibly alleviating
column separation down the main line. The second geometry would reverse the valve
locations, having a control valve in the bypass and a check valve in the main line downstream of the pump. The control valve would open on power failure, again allowing water
to bypass the pump into the main line.
Open (simple) surge tank. A simple on-line surge tank or standpipe (Fig. 12.11) can
be an excellent solution to both upsurge and downsurge problems, These devices are
quite common in hydroelectric systems where suitable topography usually exists. They
are practically maintenance free, available for immediate response as they are on line.
For pumping installations open simple surge tanks are rare because of height considerations and the absence of high points near most pumping stations. As mentioned by
Parmakian (1968) simple surge tanks are the most dependable of all surge protection
devices. One disadvantage is the additional height to allow for pump shutoff head.
Overflowing and spilling must be considered, as well as the inclusion of some damping
to reduce oscillations. As stated by Kroon et al. (1984) the major drawback to simple
surge tanks is their capital expense.
One-way surge tank. The purpose of a one-way surge tank is to prevent initial low
pressures and potential water-column separation by admitting water into the pipeline subsequent to a downsurge. The tank is normally isolated from the pipeline by one or more
lateral pipes in which there ok one or more check valves to allow flow into the pipe if the
HGL is lower in the pipe than the elevation of the water in the open tank. Under normal
operating conditions the higher pressure in the pipeline keeps the check valve closed. The
major advantage of a one-way surge tank over a simple surge tank is that it does not have
to be at the HGL elevation as required by the latter. It has the disadvantage, however, on
only combatting initial downsurges, and not initial upsurges. One-way surge tanks have
been employed extensively by the U.S. Bureau of Reclamation in pump discharge lines,
principally by the instigation of Parmakian (1968), the originator of the concept. Another
example of the effective application of one-way surge tanks in a pumping system was
reported by Martin (1992), to be discussed in section 12.9.1.
Considerations for design are: (1) location of high points or knees of the piping, (2)
check valve and lateral piping redundancy, (3) float control refilling valves and water supply, and other appurtenances. Maintenance is critical to ensure the operation of the check
valve(s) and tank when needed.
Air chamber (hydropneumatic surge tank). If properly designed and maintained, an
air chamber can alleviate both negative and positive pressure problems in pumping systems. They are normally located within or near the pumping station where they would
have the greatest effect. As stated by Fox (1977) and others, an air chamber solution may
be extremely effective in solving the transient problem, but highly expensive. Air chambers have the advantage that the tank-sometimes multiple-can be mounted either vertically or horizontally. The principal criteria are available water volume and air volume for
the task at hand.
For design, consideration must be given to compressed air supply, water level sensing,
sight glass, drains, pressure regulators, and possible freezing. Frequently, a check valve is
installed between the pump and the air chamber. Since the line length between the pump
and air chamber is usually quite short, check valve slamming may occur, necessitating the
consideration of a dashpot on the check valve to cushion closure.
The assurance of the maintenance of air in the tank is essentialusually 50 percent of
tank volume, otherwise the air chamber can be ineffective. An incident occurred at a raw
water pumping plant where an air chamber became waterlogged due to the malfunctioning of the compressed air system. Unfortunately, pump power failure occurred at the same
time, causing water column separation and waterhammer, leading to pipe rupture.
Air vacuum and air release valves. Another method for preventing subatmospheric
pressures and vapor cavity formation is the admittance of air from airvacuum valves
(vacuum breakers) at selected points along the piping system. Proper location and size
of airvacuum valves can prevent water-column separation and reduce waterhammer
effects, as calculated and measured by Martin (1980). The sizing and location of the
valves are critical, as stated by Kroon et al. (1984). In fact, as reported by Parmakian
(1982a,b) the inclusion of air-vacuum valves in a pipeline did not eliminate failures.
Unless the air-vacuum system is properly chosen, substantial pressures can still occur
due to the compression of the air during resurge, especially if the air is at extremely low
pressures within the pipeline when admitted. Moreover, the air must be admitted quickly enough to be effective. Typical designs are shown in Fig. 12.13
As shown by Fleming (1990) vacuum breakers can be a viable solution. The advantage of an air-vacuum breaker system, which is typically less expensive than other measures such as air chambers, must be weighed against the disadvantages of air accumulation along the pipeline and its subsequent removal. Maintenance and operation of valves
is critical in order for assurance of valve opening when needed. Air removal is often
accomplished with a combined airrelease air-vacuum valve. For finished water systems
the admittance of air is not a normal solution and must be evaluated carefully. Moreover,
air must be carefully released so that no additional transient is created.
Flywheel Theoretically, a substantial increase in the rotating inertia (WT?2) of a pumpmotor unit can greatly reduce the downsurge inasmuch as the machine will not decelerate
as rapidly. Typically, the motor may constitute from 75 to 90 percent of the total WR2.
Additional WR2 by the attachment of a flywheel will reduce the downsurge. As stated by
Parmakian (1968), a 100 percent increase in WR2 by the addition of a flywheel may add
up to 20 percent to the motor cost. He further states that a flywheel solution is only economical in some marginal cases. Flywheels are usually an expensive solution, mainly useful only for short systems. A flywheel has the advantage of practically no maintenance,
but the increased torque requirements for starting must be considered.
Uninterrupted power supply (UPS). The availability of large uninterrupted power supply systems are of potential value in preventing the primary source of waterhammer in
pumping; that is, the generation of low pressures due to pump power failure. For pumping stations with multiple parallel pumps, a UPS system could be devised to maintain one
or more motors while allowing the rest to fail, inasmuch as there is a possibility of maintaining sufficient pressure with the remaining operating pump(s). The solution usually
is expensive, however, with few systems installed.
12.6
DESIGNCONSIDERATIONS
enhanced if the input data and network schematization is improved via calibration.
Computer codes have the advantage of investigating a number of options as well as optimizing the sizing of surge protection devices. The ability to calibrate a numerical analysis code to a system certainly improves the determination of the proper surge protection.
Otherwise, if the code does not reasonably well represent a system, surge protection
devices can either be inappropriate or under- or oversized.
Computer codes that do not properly model the formation of vapor pockets and subsequent collapse can cause considerable errors. Moreover, there is also uncertainty regarding any free or evolved gas coming out of solution. The effect on wave speed is known,
but this influence can not be easily addressed in an analysis of the system. It is simply
another possible uncertainty.
Even for complicated systems such as water distribution networks, hydraulic transient calculations can yield reasonable results when compared to actual measurements
provided that the entire system can be properly characterized. In addition to the pump,
motor, and valve characteristics there has to be sufficient knowledge regarding the piping and flow demands. An especially critical factor for a network is the schematization
of the network; that is, how is a network of thousands of pipes simplified to one suitable for computer analysis, say hundreds of pipes, some actual and some equivalent.
According to Thorley (1991), a network with loops tends to be more forgiving regarding waterhammer because of the dispersive effect of many pipes and the associated
reflections. On the other hand, Karney and Mclnnis (1990) show by a simple example
that wave superposition can cause amplification of transients. Since water distribution
networks themselves have not been known to be prone to waterhammer as a rule, there
is meager information as to simplification and means of establishing equivalent pipes
for analysis purposes. Large municipal pipe networks are good examples wherein the
schematization and the selection of pipes characterizing the networks need to be
improved in orde to represent the system better.
72.7 NEGATIVEPRESSURESANDWATER
COLUMN SEPARATION IN NETWORKS
For finished water transmission and distribution systems the application of 138 kPa (20
psig) as a minimum pressure to be maintained under all conditions should prevent column
separation from occurring provided analytical models have sufficient accuracy. Although
water column separation and collapse is not common in large networks, it does not mean
that the event is not possible. The modeling of water column separation is clearly difficult
for a complicated network system. Water column separation has been analytically modeled with moderate success for numerous operating pipelines. Clearly, not only negative
pressures, but also water column separation, are unwanted in pipeline systems, and should
be eliminated by installation of properly designed surge protection devices.
If the criterion of a minimum pressure of 138 kPa (20 psig) is imposed then the issue
of column separation and air-vacuum breakers are irrelevant, except for prediction by
computer codes. Aside from research considerations, column separation is simulated for
engineering situations mainly to assess the potential consequences. If the consequences
are serious, as they often are in general, either operational changes or more likely surge
protection devices are designed to alleviate column separation. For marginal cases of column separation the accuracy of pressure prediction becomes difficult. If column separation is not to be allowed and the occurrence of vapor pressure can be adequately predicted, then the simulation of column separation itself is not necessary.
Some codes do not simulate water column separation, but instead only maintain the
pressure at cavity location at vapor pressure. The results of such an analysis are invalid,
if indeed an actual cavity occurred, at some time subsequent to cavity formation. This
technique is only useful to know if a cavity could have occurred, as there can be no assessment of the consequences of column separation. The inability of any code to model water
column separation has the following implications: (1) the seriousness of any column separation event, if any, can not be determined, and (2) once vapor pressure is attained, the
computation model loses its ability to predict adequately system transients. If negative
pressures below 138 kPa (20 psig) are not to be allowed the inability of a code to assess
the consequences of column separation and its attendant collapse is admittedly not so serious. The code need only flag pressures below 138 kPa (20) psig and negative pressures,
indicating if there is a need for surge protection devices.
The ability of any model to properly simulate water column separation depends upon
a number of factors. The principal ones are
Accurate knowledge of initial flow rates
Proper representation of pumps, valves, and piping system
A vapor pocket allowed to form, grow, and collapse
Maintenance of vapor pressure within cavity while it exists
Determination of volume of cavity at each time step
Collapse of cavity at the instant the cavity volume is reduced to zero
12.8
(12.16)
'/=5f.
<12-17>
tm
(1118)
'.= 2 ^ JWT
(m9)
12.9
CASESTUDIES
For three large water pumping systems with various surge protection devices waterhammer analyses and site measurements have been conducted. The surge protection systems in question are (1) one-way and simple surge tanks, (2) an air chamber, and (3) airvacuum breakers.
12.9.1
A very large pumping station has been installed and commissioned to deliver water
over a distance of over 30 kilometers. Three three-stage centrifugal pumps run at a synchronous speed of 720 rpm, with individual rated capacities of 1.14 mVsec, rated heads of
165 m, and rated power of 2090 kw. Initial surge analysis indicated potential water-column separation. The surge protection system was then designed with one-way and simple
surge tanks as well as air-vacuum valves strategically located.
The efficacy of these various surge protection devices was assessed from site measurements. Measurements of pump speed, discharge valve position, pump flow rate, and
pressure at seven locations were conducted under various transient test conditions. The
site measurements under three-pump operation allowed for improvement of hydraulic
transient calculations for future expansion to four and five pumps. Figure 12.14 illustrates
the profile of the ground and the location of the three pairs of surge tanks. The first and
second pair of surge tanks are of the one-way (feed tank) variety, while the third pair are
simple open on-line tanks.
E l e v a t i o n in m
Qne-Way Tanks
Ohe-Way Tanks
D i s t a n c e along Pipe in km
FIGURE 12.14 Case study of pump power failure at pumping station with three pair of surge tanks
two pair one way and one pair simple surge tanks Martin(1992).
Recorded
Time in seconds
Predicted
Recorded
Time in seconds
Figure 12.15. The test program and transient analysis clearly indicated that the piping system was adequately protected by the array of surge tanks inasmuch as there were no negative pressures.
Measured
Calculated
Time in Seconds
FIGURE 12.16 Case study of air chamber performance for raw water supply.
Pump trip tests were conducted for three-pump operation with cone valves actuated by
the loss of motor power. For numerical analysis a standard computer program applying
the method of characteristics was employed to simulate the transient events. Figure 12.15
shows the transient pressures for three pump power failure. The transient pressures agree
reasonably well for the first 80 seconds. The minimum HGL's in Fig. 12.14 also show
good agreement, as well as the comparison of measured and calculated pump speeds in
12.9.2
Air-inlet valves or air-vacuum breakers are frequently installed on liquid piping systems
and cooling water circuits for the purpose of (1) eliminating the potential of water-column
separation and any associated waterhammer subsequent to vapor pocket collapse; (2) protecting the piping from an external pressure of nearly a complete vacuum; and (3) providing an elastic cushion to absorb the transient pressures.
A schematic of the pumping and piping system subject to the field test program is
shown in Fig. 12.17. This system provides the cooling water to a power plant by pumping water from the lower level to the upper reservoir level. There are five identical vertical pumps in parallel connected to a steel discharge pipe 1524 mm (60 in) in diameter. On
the discharge piping of each pump there are 460 mm (18 in) diameter swing check valves.
Mounted on top of the 1524 mm (60 in) diameter discharge manifold is a 200 mm (8 in)
diameter pipe, in which is installed a swing check valve with a counter weight. Air enters
the vacuum breaker through the tall riser, which extends to the outside of the pump house.
Transient pressures were measured in the discharge header for simultaneous tripout of
three, four, and five pumps. The initial prediction of the downsurge caused by pump
power failure was based on the method of characteristics with a left end boundary condition at the pumps, junction boundary condition at the change in diameter of the piping,
and a constant pressure boundary condition at the right end of the system.
The predicted pressure head variation in the pump discharge line is shown in Fig.
12.17 for a simulated five pump tripout. The predicted peak pressure for the five pump
ELEVATION IN METERS
FREE
DISCHARGE
VACUM
DR BAKER
MVT
PUMPS
DISTANCE IN METERS
Figure 12.17. Case Study of Vacuum Breaker Performance for River Water System of Nuclear Plant,
Martin (1980).
tripout compares favorably with the corresponding measured peak, but the time of occurrence of the peaks and the subsequent phasing vary considerably. Analysis without a vacuum breaker or other protective device in the system predicted waterhammer pressure
caused by collapse of a vapor pocket to exceed 2450 kPa (355 psi). The vacuum breaker
effectively reduced the peak pressure by 60 per cent. Peak pressures can be adequately
predicted by a simplified liquid column, orifice, and air spring system. Water-column separation can be eliminated by air-vacuum breakers of adequate size.
REFERENCES
Chaudhry, M. H., Applied Hydraulic Transients, 2d ed., Van Nostrand Reinhold, New Yorle 1987.
Donsky, B., "Complete Pump Characteristics and the Effects of Specific Speeds on Hydraulic
Transients," Journal of Basic Engineering, Transactions, American Society of Mechanical
Engineers, 83: 685-699, 1961.
Fleming, A. J., "Cost-Effective Solution to a Waterhammer Problem," Public Works, 4244, 1990.
Fox, J. A., Hydraulic Analysis of Unsteady Flow in Pipe Networks, John Wiley & Sons, New Yorle
1977.
Karney, B. W., and Mclnnis, D., "Transient Analysis of Water Distribution Systems," Journal
American Water Works Assoriation, 82: 62-70, 1990.
Kittredge, C. P., "Hydraulic Transients in Centrifugal Pump Systems," Transactions, American
Society of Mechanical Engineers, 78: 1307-1322, 1956.
Knapp, R. T, "Complete Characteristics of Centrifugal Pumps and Their Use in Prediction of
Transient Behavior," Transactions, American Society of Mechanical Engineers , 59:683-689,
1937.
Knapp, R. T, "Centrifugal-Pump Performance Affected by Design Features," Transactions,
American Societiy of Mecharnal Engineers, 63:251-260, 1941.
Kroon, J. R., Stoner, M. A., and Hunt, W. A., "Water Hammer: Causes and Effects," Journal
American Water Works Assoriation, 76:39-45, 1984.
Lescovitch, J. E., "Surge Control of Waterhammer by Automatic Valves," Journal American Water
Works Assoriation, 59:632-644, 1967.
Lundgren, C. W., "Charts for Determining Size of Surge Suppressor for Pump-Discharge Lines,"
Journal of Engineering for Power, Transactions, American Society of Mechanical Engineers,
93:43^7, 1961.
Marchal, M., Flesh, G., and Suter, P., "The Calculation of Waterhammer Problems by Means of the
Digital Computer," Proceedings, International Symposium on Waterhammer in Pumped Storage
Projects, American Society of Mechanical Engineers (ASME), Chicago, 1965.
Martin, C. S., "Entrapped Air in Pipelines," Paper F2 Second BHRA International Conference on
Pressure Surges, The City University, London, September 22-24, 1976.
Martin, C. S., "Transient Performance of Air Vacuum Breakers," Fourth International Conference
on Water Column Separation, Cagliari, November 11-13, 1979. "Transient Performance Air
Vacuum Breakere," L'Energia Elettrica, Proceedings No. 382, 1980, pp. 174-184.
Martin, C. S., "Representation of Pump Characteristics for Transient Analysis," ASME
Symposium on Performance Characteristics of Hydraulic Turbines and Pumps, Winter Annual
Meeting, Boston, November 13-18, pp. 1-13,1983.
Martin, C. S., "Experience with Surge Protection Devices," BHr Group International Conference
on Pipelines, Manchester, England, March pp. 171-178 24-26, 1992.
Martin, C. S., "Hydraulics of Valves," in J. A. Schetz and A. E. Fuhs, eds. Handbook of Fluid
Dynamics and Fluid Machinery, Vol. Ill, McGraw-Hill, pp. 2043-2064. 1996,
Parmakian, J., Water Hammer Analysis, Prentice-Hall, New York, 1955.
Parmakian, J., "Unusual Aspects of Hydraulic Transients in Pumping Plants," Journal of the Boston
Society of Civil Engineers, 55:30^7, 1968.
Parmakian, J., "Surge Control," in M. H. Chaudhry, ed., Proceedings, Unsteady Flow in Conduits,
Colorado State University, pp. 193-207 1982,
Parmakian, J., "Incidents, Accidents and Failures Due to Pressure Surges," in M. H. Chaudhry ed.,
Proceedings, Unsteady Flow in Conduits, Colorado State University, pp. 301-311 1982.
Sanks, R. L., Pumping Station Design, Butterworths, Bestar, 1989.
Stepanoff, A. L, Centrifugal and Axial Flow Pumps, John Wiley & Sons, New York, 1957.
Swanson, WM., "Complete Characteristic Circle Diagrams for Turbomachinery," Transactions,
American Society of Mechanical Engineers, 75:819-826, 1953.
Thorley, A. R. D., Fluid Transients in Pipeline Systems, D. & L. George Ltd., 1991.
Tullis, J. P., Hydraulics of Pipelines, John Wiley & Sons, New Yorhl989.
Watters, G. Z., Modern Analysis and Control of Unsteady Flow in Pipelines, Ann Arbor Science,
Ann Arbor, MI, 1980.
Weaver, D. L., "Surge Control," Journal American Water Works Association, 64: 462-466, 1972.
White, I. M., "Application of the Surge Suppressor in Water Systems," Water Works Engineering,
45,304-306, 1942.
Wood, D.J., and Jones, S.E., "Waterhammer Charts for Various Types of Valves", ASCE, Journal of
Hydraulics Division, HYl, 99:167-178, 1973.
Wylie, E. B., and Streeter, V. L., Fluid Transients in Systems, Prentice-Hall, 1993.
CHAPTER 13
HYDRAULIC DESIGN
OF DRAINAGE FOR
HIGHWAYS
13.1
INTRODUCTION
The objective of this chapter is to provide design guidance for highway hydraulic elements: gutters, roadside conveyance, inlets, and bridge scuppers. Proper hydraulic design
of highway drainage is essential to avoid disruption of the highways transportation function, maintain safe travel conditions, and sustain infrastructure.
When rain falls on a sloped pavement surface, it forms a thin film of water that increases in thickness as it flows to the edge of the pavement and concentrates in gutters or roadside ditches; overflow from gutters and ditches spreads out onto the pavement. Factors
that influence the depth of water on the pavement and its spread in gutters are the length
of the flow path, the texture and slope of the surface, and rainfall intensity. As the depth
of water on the pavement or the gutter spread increases, the potential for vehicular
hydroplaning or disruption of the highway's transportation function increases. With
design methods and guidance, the surface drainage elements are sized to function at predetermined reliability thresholds.
The design of highway surface drainage is a critical component of the highway system
(Johnson and Chang, 1984). The hydraulics of cross drainage handled by culverts and
bridges is not considered in this chapter, and subsurface storm drains that accept surface
water removed at inlets also are excluded. This chapter provides information on the following aspects of highway drainage design: highway geometries that influence hydraulic
design, design event selection, design flow estimation, gutter design, ditch hydraulic and
stability design, inlet design, and bridge deck hydraulics.
The reader also is directed to documents of the Federal Highway Administration,
including Hydraulic Engineering Circular No. 12 (HEC-12), "Drainage of Highway
Pavement" (Johnson and Chang, 1984), HEC-15, "Design of Roadside Channels with
Flexible Linings" (Chen and Cotton, 1986), HEC-21, "Design of Bridge Deck
Drainage" (Young et al., 1993), and HEC-22, "Urban Drainage Design Manual"
(Brown et al., 1996). Information on the physics of surface drainage can be found in
Anderson et al. (1995).
Pavement
Sheet flow on the pavement can cause hydroplaning, and pavement materials can influence sheet flow. Smooth pavements can cause water to flow faster and reduce the thickness of the sheet flow film; however, smooth pavements can decrease the coefficient of
friction of pavement and tires and lead to longer skids as well.
As the depth of water flowing over a roadway's surface increases, the potential for
hydroplaning increases. When a rolling tire encounters a film of water on the roadway, the
water is channeled through the tire tread pattern and through the surface roughness of the
pavement. Hydroplaning occurs when the drainage capacity of the tire tread pattern and
the pavement surface is exceeded and the water begins to build up in front of the tire. As
the water builds up, a water wedge is created, producing a hydrodynamic force that can
lift the tire off the pavement surface. This is considered full dynamic hydroplaning, and,
because the water offers little shear resistance, the tire loses its tractive ability and the driver loses control of the vehicle. Hydroplaning is a function of the water depth, roadway
geometries, vehicle speed, treat depth, tire inflation pressure, and conditions of the pavement surface.
The following are several pavement design factors:
1. Design the highway to reduce the drainage path lengths of the water flowing over the
pavement. Crown sections split flow to both sides and are superior to long sloping runs of
pavement from one side to the other. This will prevent build-up of sheet flow thickness.
2. Increase the pavement surface texture depth by such methods as grooving of Portland
cement concrete. An increase of pavement surface texture will increase the drainage
capacity at the tire-pavement interface and reduce hydroplaning.
3. Use open-graded pavements, which have been shown to reduce greatly the
hydroplaning potential of the roadway surface. This reduction is the result of the
water's ability to be forced through the pavement under the tire and to enter the surbase rather than run off. Surface texture also increases the coefficient of tire-to-pavement friction. Such open grading also involves provision of subsurface drainage
details to the pavement design.
4. Use intercepting drainage structures along the roadway to capture the sheet of water
over the pavement, which will reduce the thickness of the film of water and reduce the
hydroplaning potential of the roadway surface. Long slotted inlets perpendicular to the
flow and open expansion joints on bridges can accomplish this.
13.2.2 Grade
The recommended minimum values of roadway longitudinal slope given in the Policy on
Geometric Design of the American Association of State Highway and Transportation
Officials (AASHTO, 1990) provide safe, acceptable pavement drainage. In addition, the
following general guidelines are presented:
1. A minimum longitudinal gradient is more important for a curbed pavement than for an
uncurved pavement because the water is constrained by the curb on the one hand, and
runs off the shoulder, on the other, with crowned cross sections.
2. Desirable gutter grades should not be less than 0.5 percent for curbed pavements, with
an absolute minimum of 0.3 percent. Minimum grades can be maintained in flat terrain by using of a rolling profile or by warping the cross slope to achieve rolling gutter profiles.
3. In sag vertical curves, a minimum grade of 0.3 percent should be maintained within 15
m (50 ft) of the low point of the curve. This is accomplished where the length of the
curve in meters divided by the algebraic difference in grades in percent (K) is equal to
or greater than 50 (156 in English units). This is represented as
K=
G 2 -rG1
(13.1)
slope of the outermost lanes. The two lanes adjacent to the crown line should be pitched
at the normal slope, and successive lane pairs, or portions thereof outward, should be
increased by approximately 0.5 to 1 percent. The maximum pavement cross slope should
be limited to 4 percent. The following are additional guidelines related to cross slope:
1. Inside lanes can be sloped toward the median if conditions warrant. This is not widely encouraged.
2. Median areas should not be drained across travel lanes.
3. The number and length of flat pavement sections in cross-slope transition areas should
be minimized. Consideration should be given to increasing cross slopes in sag vertical
curves and crest vertical curves and in sections of flat grade.
4. Shoulders should be sloped to drain away from the pavement, except with raised, narrow medians and superelevations.
13.2.4 Safety
The placement of drainage structures in the traveled way can constitute an obstacle to
moving traffic. Median inlets, storm drain inlets, and junctions need to have top elevations that do not project significantly above the surface. Culvert entrances and end walls
need to be situated away from likely trajectories of errant vehicles that have departed
traffic lanes because a driver must maneuver to avoid an accident or is weary.
As a general rule, drainage devices should provide a low silhouette to oncoming traffic.
13.3
The most significant design decisions considered in sizing pavement drainage facilities are
the frequency of the design runoff event and the allowable spread of water on the pavement. A related consideration is the use of an event of lesser frequency (greater storm
magnitude) to check the drainage design (AASHTO, 1991; Johnson and Chang, 1984).
Spread and design frequency are not independent. The implications of the use of a criteria for spread of half a traffic lane is considerably different for one design frequency
than for a lesser frequency. Spread also has different implications for a low-traffic, lowspeed highway than for a higher classification highway. Balancing risks and flooding are
central to the issue of highway pavement drainage and are important to highway safety.
13.3.1 Risk Balancing
The objective of highway storm-drainage design is to provide safe passage for vehicles
during the storm event. The design of a drainage system for a section of highway pavement is to collect runoff in the gutter or roadside ditch and convey it to inlets or culverts
in a manner that provides a reasonable degree of safety for traffic and pedestrians at a reasonable cost. As spread from the curb or roadside ditch increases, the risks of traffic accidents and delays and the nuisance and possible hazard to pedestrian traffic increase.
The recurrence interval and allowable spread for design implies acceptable risks of accidents and traffic delays and acceptable costs for the drainage system. Risks associated with
water on traffic lanes are greater with high-volume traffic, high speeds, and higher highway
classifications than with lower-volumes traffic, speeds, and highway classifications.
The major considerations that enter into the selection of design frequency and design
spread are as follows:
1. Functional classification of the highway drives risks: higher functions imply lower
acceptable risks, which translate to low intervals of recurrence with a typical limit of
1 in 100 years. Because ponding on traffic lanes of high-speed, high-volume highways
is contrary to the public's expectations, the risks of accidents and the costs of traffic
delays are high.
2. Design speed is important to the selection of design criteria. (Also, at speeds greater
than 70 km/h (44 mi/h), water on the pavement can cause hydroplaning; in this case,
rain intensity, is more significant than spread.)
3. Projected traffic volumes are an indicator of the economic importance of keeping the
highway open to traffic. The opportunity costs of lost driver and passenger times associated with traffic delays increase rapidly with increasing volumes of traffic.
4. The intensity of rainfall may significantly affect the selection of design frequency and
spread. Risks associated with the spread of water on pavements may be less in arid
areas that are subject to high-intensity thunderstorms than in areas that have frequent
but less intense storms.
5. Capital costs are the other side of the equation. Cost considerations make it necessary to balance the approach to the selection of design criteria to be essentially risk
based. "Tradeoffs" between desirable and practicable criteria are necessary because
of costs.
Other considerations include inconvenience, hazards, and nuisances to pedestrian traffic. These considerations should not be minimized; in some locations, such as commercial
areas, they may assume major importance. Local design practice also may be a major consideration because it can affect the feasibility of designing to higher standards and it influences the public's perception of acceptable practice.
The relative elevation of the highway and surrounding terrain is an additional consideration where water can be drained only through a storm drainage system, as in underpasses and depressed sections. The potential for ponding to hazardous depths should be
considered when selecting the frequency and spread criteria and when checking the design
against storm runoff events of lesser frequency than in the design event.
Spread on traffic lanes can be tolerated to greater widths when traffic volumes and
speeds are low. Spreads of one-half of a traffic lane or more are usually considered to be
a minimum-type design for low-volume local roads.
The selection of design criteria for intermediate types of facilities can be difficult. For
example, some arterials with relatively high traffic volumes and speeds may not have
shoulders that will convey the design runoff without encroaching on the traffic lanes.
13.3.2 Design Guidance Regarding Frequency
and Spread
Table 13.2 provides suggested minimum design frequencies and spread based on the type
of highway and the traffic speed. The recommended design frequency for depressed sections and underpasses where ponded water can be removed only through the storm
drainage system is a 50-yr frequency event. The use of a lesser frequency event, such as
a 100-yr storm, to assess hazards at critical locations where water can pond to appreciable depths is commonly referred to as a check storm or check event.
Local streets
Design Frequency
Design Spread
10-yr
10-yr
50-yr
10-yr
10-yr
10-yr
5-yr
10-yr
10-yr
Shoulder + 1 m (3 ft)
Shoulder
Shoulder + 1 m (3 ft)
V2 driving lane
Shoulder
V2 driving lane
V2 driving lane
V2 driving lane
V2 driving lane
13.4
The typical highway drainage area is small, and design flows are linked to design rainfall intensity. Three methods are discussed for the selection of design rainfall intensity:
(1) the rational method that pegs rainfall intensity to a design frequency, (2) the avoidance of the hydroplaning method, and (3) driver vision-impairment method. The first
method uses established drainage policy (such as that implied in Table 13.2), to select
a return period and calculate time of concentration to avoid ponding. The second and
third methods consider vehicle safety directly, either from the standpoint of avoidance
of hydroplaning films or from driver vision being impaired because of heavy rain. All
methods lead to designs for which selected spread is a design requirement. Method 1
involves rainfall return period. Methods 2 and 3 select rainfall intensity on the basis of
physical limits of the vehicle (skid avoidance) or the driver's vision; in these two cases,
the design frequency is imputed by the physical limits rather than selected by the analyst or set by the prevailing policy.
13.4.1 Design Flow Calculation
The commonly used equation for the calculation of peak flow from the small areas associated with highway drainage is a simple mass balance which is considered to be valid for
areas less than 80 ha (200 ac) (ASCE, 1992):
Q = A^
c
where
(13.2)
(13.3)
?A'
where / is the subscript designating values for different types of land. Highway rights-ofway typically have C > 0.5 because of high pavement coverage.
TABLE 13.3 Runoff Coefficients
Type of Drainage Area
Business
Downtown areas
Neighborhood areas
Residential
Single-family areas
Multi-units, detached
Multi-units, attached
Suburban
Apartment dwelling areas
Industrial
Light areas
Heavy areas
Parks, cemeteries
Playgrounds
Runoff Coefficient C*
0.70 - 0.95
0.50 - 0.70
0.30 - 0.50
0.40 - 0.60
0.60 - 0.75
0.25 - 0.40
0.50 - 0.70
0.50 - 0.80
0.60 - 0.90
0.10 - 0.25
0.20 - 0.40
Runoff Coefficient C*
0.20 - 0.40
0.10-0.30
0.05 - 0.10
0.10 - 0.15
0.15 - 0.20
0.13 - 0.17
0.18 - 0.22
0.25 - 0.35
0.70 - 0.95
0.80 - 0.95
0.70 - 0.85
0.75 - 0.85
0.75 - 0.95
1"(Vs)
a,4.
(
}
where Ts = sheet flow travel time, min; n = roughness coefficient as given in Table 13.4;
L = flow length, m (ft); 7 = rainfall intensity, mm/h (in/h); S = surface slope, m/m (ft/ft);
and Kc = empirical coefficient equal to 6.943 (0.933 in English units).
Because / depends on Ts and Ts is not initially known, the computation of Ts is an iterative process. An initial estimate of Ts is assumed and used to obtain / from the IDF curve
for the locality. The Ts is then computed from Eq. (13.4) and used to check the initial value
of /. If they are not the same, the process is repeated until two successive Ts estimates are
the same to within a reasonable tolerance (McCuen et al., 1996).
For the following types of segments downhill from the sheet flow, the velocity is calculated and the segment travel time computed. To arrive at the travel time, the segment
length is divided by the velocity.
13.4.2.2 Shallow concentrated flow velocity. After short distances of, at most, 90 m (300
ft), sheet flow tends to concentrate in rills and then gullies of increasing proportions.
Such flow is usually referred to as shallow concentrated flow. The velocity of such flow
can be estimated using a relationship between velocity and slope as follows (McCuen et
al., 1996):
Vc = KckS*
(13.5)
TABLE 13.4 Manning's Roughness Coefficient (n) for Overland Sheet Flow
Surface Description
Smooth asphalt
0.011
Smooth concrete
0.012
Ordinary concrete lining
0.013
Good wood
0.014
Brick with cement mortar
0.014
Vitrified clay
0.015
Cast iron
0.015
Corrugated metal pipe
0.024
Cement rubble surface
0.024
Fallow (no residue)
0.05
Cultivated soils
Residue cover < 20%
0.06
Residue cover > 20%
0.17
Range (natural)
0.13
Grass
Short grass, prairie
0.15
Dense grasses
0.24
Bermuda grass
0.41
Woods*
Light underbrush
0.40
Dense underbrush
0.80
Source Mc Cuen et al 1996*When selecting n, consider cover to a height of about 30 mm. This is the only part of
the plant cover that will obstruct sheet flow.
(13.6)
where Vg = velocity, m/s (ft/s); Kc units conversion equal to 0.752 (1.12 in English
units); G = grade, m/m (ft/ft); Sx = cross slope, m/m (ft/ft); T = spread, m (ft) and, n =
Manning's friction factor (typically taken to be 0.016 for highway gutters).
Thus, if the allowable spread is 5 m, then 0.65 X 5 = 3.25 m of spread is used in Eq. (13.6).
13.4.2.4 Open channel and pipe flow velocity. Flow in gullies empties into channels or
pipes. Open channels are assumed to begin where either the blue streamline shows on the
U.S. Geological Survey quadrangle sheets or the channel is visible on aerial photographs.
Cross-section al geometry and roughness should be obtained for all channel reaches in the
watershed. Manning's equation can be used to estimate average flow velocities in pipes
and open channels as follows:
Vp = #2/3S1/2
(13.7)
where
TABLE 13.5 Intercept Coefficients for Velocity Versus Slope Relationship of Eq. (13.5)
Land Cover/Flow Regime
Forest with heavy ground litter; hay meadow (overland flow)
Trash fallow or minimum tillage cultivation; contour or strip cropped;
woodland (overland flow)
Short grass pasture (overland
flow)
Cultivated straight row (overland
flow)
Nearly bare and untilled (overland flow); alluvial fans in western
mountain regions
Grassed waterway (shallow concentrated
flow)
Unpaved (shallow concentrated
flow)
Paved area (shallow concentrated flow); small upland gullies
Source: McCuen et al., 1996
k
0.076
0.152
0.213
0.274
0.305
0.457
0.491
0.619
TABLE 13.6 Values of Manning Coefficient (n) for Channels and Pipes
Conduit Material
Mannings n*
Closed conduits
Asbestos-cement pipe
0.0110.015
Brick
0.013-0.017
Cast iron pipe: Cement lined & seal coated
0.011-0.015
Concrete (monolithic)
0.012-0.014
Concrete pipe
0.011-0.015
Corrugated-metal pipe (13 mm X 65 mm
[1/2-in X 2 1/2-in corrugations])
Plain
0.022-0.026
Paved invert
0.018-0.022
Spun asphalt lined
0.011-0.015
Plastic pipe (smooth)
0.011-0.015
Vitrified clay
Pipes
0.011-0.015
Liner plates
0.013-0.017
Open Channels
Lined channels
Asphalt
0.013-0.017
Brick
0.012-0.018
Concrete
0.011-0.020
Rubble or riprap
0.020-0.035
Vegetal
0.030-0.40
Excavated or dredged channels
Earth, straight and uniform
0.020-0.030
Earth, winding, fairly uniform
0.025-0.040
Rock
0.030-0.045
Unmaintained
0.050-0.14
Natural channels (minor streams, top width at flood stage
< 30m [100 ft])
Fairly regular section
0.030.07
Irregular section with pools
0.030.10
Source: Jennings et al., 1994.
*Lower values are usually for well-constructed and maintained (smoother) pipes and channels.
r. = T. +2^
(13.8)
where Tc = time of concentration, min; Ts = sheet flow travel time, min; L1 = flow length
for segment /, m (ft); and V1 = velocity for segment /, m/s (ft/s) (V0, Vg, or Vp.
13.4.2.6 Rainfall intensity as a function of duration and return period. Once Tc is
determined, it is assumed to be equal to the rainfall's duration. With the rational method,
the selected return period, as tabulated in the Table 13.2 guidance, has an associated rainfall intensity-versus-duration curve. This curve is used to establish the rational method
rainfall intensity / for use with Eq. (13.2).
13.4.3 Rainfall Intensity by Avoidance of the
Hydroplaning Method
HEC-21 (Young et al., 1993) developed an alternative method for selecting rainfall intensity that is not dependent on rainfall frequency. (Note that once an intensity is selected,
a frequency is implied.) The method assumes that if drivers have a hydroplaning slip sensation at the design speed (plus a reasonable allowance), they will typically slow down to
90 km/h (55 mph) or less and this behavior governs highway function more than traffic
reaction to gutter flooding.
The method selects values of vehicle speed, tire tread depth, pavement texture, and tire
pressure and calculates the thickness of the sheet flow film at incipient hydroplaning
(assumed at 10 percent spindown: e.g., the tire rolls 1.1 times the circumference to move
one circumference).
The empirical equation for the vehicle speed which initiates hydroplaning is
V = K1SEPM(Kf)M
[(K3TD) + K3]
006
(13.9)
where A = Texas Transportation Institute empirical curve fitting relationship which is the
greater of:
A =
Al
10.409
(K4Cf)OM
+ 3 50?
+5
'^
or
A
(13 10)
'
where V = vehicle speed in km/h (mph); TD = tire tread depth in mm (1/32 in); TXD =
pavement texture depth, mm (in); d = water film depth, mm (in); P = tire pressure, kPa
(psi); K1 = unit conversion equal to 0.3048 (1 in English units); K2 unit conversion
equal to 6.894 (1 in English units); K3 = unit conversion equal to 0.794 (1 in English
units) K4 =unit conversion equal to 25.4 (1 in English units) and SD = spindown (percent); hydroplaning is assumed to begin at 10 percent spin down.
This occurs when the tire rolls 1.1 times the circumference to achieve a forward
progress distance equal to one circumference.
The method determines a film depth, d, associated with selected values for V, TD,
TXD, P, and with SD =10 percent, by solving the above equation. An estimate of design
d for V = 88 km/h, TD = 5.55 mm (median tire tread), TXD = 0.97 mm (mean pavement
texture), P = 186 kPA (median tire pressure), and SD =10 percent (by definition) is d =
1.87 mm. This is suggested as a sound design value because it represents the combination
of the mean or median of all the above parameters.
Manipulation of Eq. (13.9), using typical values, gives the following sensitivity information. An increase of 1 percent in pavement texture increases the hydroplaning depth by
1.6 percent, an increase of 1 percent in tread depth increases the hydroplaning depth
by 0.8 percent, and an increase of 1 percent in tire pressure increases the hydroplaning
depth by 2.4 percent.
Study of Eq. (13.9) indicates that 90 km/h (55 mph) is the speed value of concern for
practical control of hydroplaning. At this speed, a decrease of 1 percent in speed increases the hydroplaning depth by 25 percent. Speeds below 90 km/h (55 mph) tend to be safe
from the threat of hydroplaning because heavy rainfall is insufficient to generate
hydroplaning depth. An increase in speed of 1 percent decreases the hydroplaning depth
by 25 percent. Above 90 km/h (55 mph), hydroplaning can occur on extremely thin surface films associated with light rainfall intensities-intensities of 25 mm/h (1 in/h) and
less, which are rainfalls that are usually smaller than those used to design gutters, inlets,
and storm sewers.
Once a design d is determined, it is assumed that the thickness of the water film on the
pavement should be less than d. Water flows in a sheet across the surface to the edge of
the gutter flow. The width of sheet flow is the width W of the deck area minus the design
spread T, or (W-T). At the edge of the gutter flow, the design sheet flow depth d is
obtained from Eq. (13.9), with a suggested default value of d 1.867 mm (0.0735 in).
Using Manning's equation and continuity and solving for / gives the hydroplaning design
rainfall intensity in mm/h (in/h), as
K
s
*}\
*
}\ dl'61
[Cn\ [(S2X + G2)0-25 J [(W - T)
1-\
mm
'
(13.12)
where Sv = driver visibility, m; / = rainfall intensity, mm; V = vehicle speed, km/h; and
K0 = constant equal to 143,587.88 (338,721.26 in English units).
This empirical relationship was developed based on the Texas A&M test. At 90 km/h
(55 mph), the nonpassing minimum stopping sight distance is 15Om (500 ft) (this is the
lower value of a range given by AASHTO). Substituting 90 km/h and 15Om into Eq.
(13.12) gives a rainfall intensity of 125 mm/h. Note that cars in a travel corridor generate
splash and spray that increase the density of water droplets. Therefore, a lower rainfall of
100 mm/h may be more realistic as a threshold value that will cause slight impairment
because spray will increase the density of rain particles at roadway eye level. That is,
design intensities / above 100 mm/h will probably obscure drivers' visibility in heavy traffic and decrease sight distances to less than minimum stopping sight distance recommended by AASHTO.
Therefore, 100 to 150 mm/h is a suggested threshold design rain fall intensity range
for avoiding drivers' vision impairment; this intensity range is determined without respect
to frequency, but, depending on time of concentration, it imputes a frequency. Rainfall
intensities below this range should not obscure a drivers view through a windshield with
functioning windshield wipers. However, night driving in the rain is highly dependent on
vision, and data supporting the predictive relationship were obtained in daylight.
13.5
GUTTERDESIGN
A pavement gutter is adjacent to the roadway and conveys storm water. Gutter spread
(top flow width) may include a portion or all of a travel lane. Gutter sections can be categorized as conventional or as the shallow swale type illustrated in Fig. 13.1. The most representative conventional gutters have a uniform cross slope (Fig. 13.1 Al) or a composite
cross slope, where the gutter slope varies from the pavement cross slope (Fig. 13.1 A2).
To compute triangular, shallow depth, gutter flow, the Manning equation is integrated for an increment of width across the section (Izzard, 1946). The resulting equation is
Q = ^Sx1-61'Go-5P-67
1. Uniform Section
1. *V-Shope Gutter
2. Composite Section
2. "V-Shop* Medion
(13.13)
3. Curved Section
3. Circular
FIGURE 13.1 Typical gutter sections (Source: Brown, Stein and Warner, 1996)
(13.14)
where Qw = flow rate in the depressed section of the gutter, m3/s (ftVs ); Q = gutter flow
rate, m3/s (ft3/s ); and Qs = flow capacity of the gutter section above the depressed section,
calculated using Eq. (13.13), m3/s (ftVs).
The cross-sectional area above the depressed section can be divided into Q8 to get the
gutter flow velocity Vg. Then Qw. is equal to Vg times the cross-sectional area of the
depressed section. The use of depressed sections can put more depth of water above inlet
entrances and increase their efficiency.
When the pavement cross section is curved, gutter capacity varies with the configuration
of the pavement. For this reason, discharge-spread- or discharge-depth-at-the-curb relationships developed for straight cross slopes are not applicable unless approximations are made.
When curbs are not needed for traffic control, a small swale section with a V-shape or
circular shape can be used to convey runoff from the pavement. Also, the control of pavement runoff on fills may be needed to protect the embankment from erosion. Small swale
sections are sized to have sufficient capacity to convey the flow to a location suitable for
interception. Shallow symmetric V-ditches, with side slopes of 10 percent or less, can be
evaluated using Eq. (13.13) and doubling the result.
Flow in shallow, circular gutter sections can be represented by the relationship
d
On
=
o.488
(13 15)
*<[s^H
'
Manning's n
0.012
0.013
0.016
0.013
0.015
0.014
0.016
0.02
where d = depth of flow in circular gutter, m (ft); D = diameter of circular gutter, m (ft);
and Kc = empirical coefficient equal to 1.179 (0.972 in English units). The width of circular gutter section Tw is represented by the chord of the arc which can be computed
using Eq. (13.16):
Tw = 2(r* - (2 - </)2)o.5
(13.16)
where Tw = width of circular gutter section, m (ft), and r = radius of flow in circular gutter, m (ft).
As gutter flow approaches the low point in a sag vertical curve, the flow can exceed
the allowable design spread values as a result of the continually decreasing gutter slope.
The spread in these areas should be checked to insure it remains within allowable limits.
If the computed spread exceeds design values, additional flanking inlets should be provided to reduce the flow as it approaches the low point. Sag vertical curves and inlets are
discussed further in a later section. The effects of spread on gutter capacity are greater
than the effects of cross slope and grade, as would be expected because of the larger exponent of the spread term in Eq. (13.13). The magnitude of the effect is demonstrated by the
gutter capacity of a 3-m (9.8-ft) spread being 18.8 times greater than with a 1-m (3.3-ft)
spread, and three times greater than with a spread of 2 m (6.6 ft).
The effects of cross slope also are relatively great, as illustrated by a comparison of
gutter capacities with different cross slopes. At a cross slope of 4 percent, a gutter has 10
times the capacity of a gutter of 1 percent cross slope. A gutter at 4 percent cross slope has
3.2 times the capacity of a gutter at 2 percent cross slope.
Little latitude is generally available to vary grade to increase the gutter capacity, but
grade changes that change the gutter capacity are frequent. A change from G1 = 0.04
to 0.02 will reduce the gutter capacity to 71 percent of the capacity at G1 = 0.04.
However, grade is not typically within the hydraulic designer's set of responsibilities,
whereas the design problem is: for a given grade, size the gutter and space the inlets to
control spread.
The flow time in gutters is an important component of the time of concentration
for the contributing drainage area to an inlet. To find the gutter flow component of the
time of concentration, Eq. (13.6) for estimating the average velocity in a reach of gutter is used.
13.6 ROADSIDEDITCHDESIGN
Roadside and median channels are open-channel systems that collect and convey storm
water from the pavement surface, roadside, and median areas. These channels may outlet
to a storm drain piping system via a drop inlet, to a detention or retention basin or other
storage component, or to an outfall channel. Roadside and median channels are normally
trapezoidal in cross section and are lined with grass or other protective linings.
The design or analysis of roadside and median channels follows the basic principles of
open channel flow. A more complete coverage of open channel flow concepts can be
found in other chapters in this handbook and in Chow (1959) and Richardson et al.,
(1990). Open channel flow is classified using the following characteristics: (1) steady
(discharge passing a cross section is constant) or unsteady, (2) uniform (flow rate and
depth remain constant along a reach) or varied, and (3) subcritical (Froude number less
than 1) or supercritical-sometimes called tranquil or rapid.
Most natural flow conditions are neither steady nor uniform. However, in most cases
of small element drainage design, it can be assumed that the flow will vary gradually in
time and space and can be described as steady, uniform flow for short periods and
distances. Gradually varied flows are nonuniform flows in which the depth and velocity change gradually enough in the direction of flow that vertical accelerations can be
neglected.
Subcritical flow is distinguished from supercritical flow by a dimensionless number
called the Froude number (Fr), which represents the ratio of inertial forces to gravitational forces and is defined for rectangular channels by the following equation:
Fr
=<^
(m7)
where V = mean velocity, m/s (ft/s); g = acceleration of gravity, 9.8 m/s2 (32.2 ft/s2)' and
y = flow depth, m (ft) Critical flow occurs when the Fr has a value of one (1.0). The flow
depth at critical flow is referred to as critical depth.
Subcritical flow occurs when the F1. is less than one (Fr > 1). Subcritical flow is characterized by slower velocities, deeper depths, and mild slopes, whereas supercritical flow
is represented by faster velocities, shallower depths, and steeper slopes. Supercritical flow
occurs when the Fr is greater than one (Fr > 1). Supercritical flows are apt to be erosive
and may require energy dissipation measures to slow them down to manageable velocities.
Most small-element open-channel flows are Subcritical. However, supercritical flows are
not uncommon for smooth-lined ditches on steep grades; if such elements transition to flat
grades or channels with higher friction, a hydraulic jump may result. It is important to evaluate the Fr in open channel flows to determine how close a particular flow is to the critical
condition. In other words, a hydraulic jump occurs as an abrupt transition from supercritical to subcritical flow. Significant changes in depth and velocity occur in the jump and
energy is dissipated. The potential for a hydraulic jump to occur should be considered in
all cases where the Fr is close to one (1.0) or where the slope of the channel bottom changes
abruptly from steep to mild. The characteristics and analysis of hydraulic jumps are covered in detail in Chow (1959) and in HEC-14 (Federal Highway Administration, 1983).
13.6.1 Steady Uniform Flow Design
Small element design is based on the depth of steady uniform flow. This flow is called
the normal flow and is computed with Manning's equation. The general form of
Manning's equation is:
K A #0.67 C0.5
Q=KnAK^ ^
(13.18)
where Kn = empirical conversion equal to 1.0 (1.486 in English units); Q - discharge rate,
mVs (ftVs); A = cross sectional flow area, m2 (ft2); R = hydraulic radius, m (ft); = A/P, m
(ft); P = wetted perimeter, m (ft); S0 = energy grade line slope, m/m (ft/ft); and n =
Manning's roughness coefficient.
The selection of an appropriate Manning's n value for design purposes is often based
on observation and experience. Manning's n values also vary with normal flow depth
since n is a friction factor and friction is a boundary effect that is averaged into the
water column. Table 13.8 provides a tabulation of Manning's n values for various lining materials for open channels Manning's roughness coefficient for vegetative and
other linings vary significantly, depending on the amount of submergence. The classification of vegetal covers by degree of retardance is provided in Table 13.9; Table 13.10
provides a list of Manning's n relationships for five classes of vegetation defined by
their degree of retardance.
Lining Type
0.015
Concrete
0.013
0.013
0.030
0.040
Grouted riprap
0.028
0.032
0.042
Stone masonry
0.030
0.022
0.025
Soil cement
0.020
0.016
0.016
0.018
Asphalt
0.020
0.020
0.023
Bare soil
Unlined
0.025
0.035
0.045
Rock cut
Woven paper net
Temporary*
0.015
0.015
0.016
0.022
Jute net
0.019
0.028
0.021
0.028
Fiberglass roving
0.019
Straw with net
0.033
0.065
0.025
0.028
0.035
0.066
Curled wood mat
0.021
0.025
0.036
Synthetic mat
0.044
25 mm (1 in), D50
0.030
0.033
Gravel riprap
0.034
0.041
50 mm (2 in), D50
0.066
0.104
150 mm (6 in), Z)50
Rock riprap
0.069
0.035
300 mm (12 in), D50
0.040
0.078
Source: Chen and Cotton, (1986).
Note: Values listed are representative values for the respective depth ranges. Manning's roughness coefficients,(n)
vary with the flow depth.
*Some "temporary" linings become permanent when buried.
Rigid
A
B
Weeping lovegrass
Yellow bluestem schaemum
Kudzu
Bermuda grass
Native grass mixture
(little bluestem, bluestem,
blue gamma, and other long
and short midwest grasses)
Condition
Excellent stand, tall, average 0.76 m (2.5 ft)
Excellent stand, tall, average 0.91 (3.0 ft)
Very dense growth, uncut
Good stand, tall, average 0.30 m (1.0 ft)
Good stand, unmowed
TABLE 13.9
(Continued)
Retardance
Cover
Class
Weeping lovegrass
Lespedeza sericea
Alfalfa
Weeping lovegrass
Kudzu
Blue gamma
Crabgrass
Bermuda grass
Common lespedeza
Grass-legume mixture:
summer (orchard grass,
redtop Italian ryegrass, and
common lespedeza)
Centipedegrass
Kentucky bluegrass
Bermuda grass
Common lespedeza
Buffalo grass
Condition
Good stand, tall, average 0.61 m (2.0 ft)
Good stand, not woody, tall, average 0.48 m (1.6 ft)
Good stand, uncut, average 0.28 m (0.91 ft)
Good stand, unmowed, average 0.33 m (1.1 ft)
Dense growth, uncut
Good stand, uncut, average 0.33 m (1.1 ft)
Fair stand, uncut, avg. 0.25-1.20 m (0.9-4.0 ft)
Good stand, mowed, average 0.15 m (0.5 ft)
Good stand, uncut, average 0.28 m (0.91 ft)
Good stand, uncut average 0.15-0.20 m (0.5-1.5 ft)
ft)
Manning's n Equation*
1.22 R^ I [30.2 + 19.97 log (R^ S00-4)]
1.22 R^ I [37.4 + 19.97 log (R S004)]
1.22 R^ I [44.6 + 19.97 log (R^ S00-4)]
1.22 R^ I [49.0 + 19.97 log (R S00-4)]
1.22 R^ I [52.1 + 19.97 log (R1-4 S00-4)]
* Equations are valid for flows less than 1.42 m3/s (50 ft3/s). Nomograph solutions for these equations are contained
in HEC-15 (Chen and Cotton, 1986).
where Ad = difference in water surface elevation between the inner and outer banks of the
channel in the bend, m (ft); V= average velocity, m/s (ft/s); T= surface width of the channel, m (ft); g = gravitational acceleration, 9.8 m/s2 (32.2 ft/s2); and Rc = radius to the center line of the channel, m (ft).
Equation (13.19) is valid for subcritical flow conditions. The elevation of the water
surface at the outer channel bank will be Ad/2 higher than the center line water surface
elevation (the average water surface elevation immediately before the bend) and the elevation of the water surface at the inner channel bank will be Ad/2 lower than the elevation
of the water surface at the center line. Flow around a channel bend imposes higher shear
stress on the channel's bottom and banks and is discussed in more detail in the next section. The increased stress requires additional design considerations within and downstream of the bend.
13.6.3 Shear Stresses in Open Channels
Stable channel design concepts presented in HEC-15 (Chen and Cotton, 1986) provide a
means of evaluating and defining channel configurations that will perform within acceptable limits of stability. Stability is achieved when the material forming the channel boundary effectively resists the erosive forces of the flow. Principles of rigid boundary
hydraulics can be applied to evaluate this type of system.
A pragmatic approach is to limit channel velocities to the range of 0.3 to 0.6 m/s (1 to
2 ft/s) to avoid erosion. Better estimates are possible by making tractive force calculations
that consider actual physical processes occurring at the channel boundary.
In uniform flow, the shear stress exerted on the bed is equal to the effective component
of the gravitational force acting on the body of water parallel to the channel bottom. The
average shear stress is equal to
T = yxS
(13.20)
where T = average shear stress, Pa (lb/ft ); y = unit weight of water, 9810 N/m3 (62.4
lb/ft2) at 15.6 0C (60 0F); x = height of water column, m (ft); and S = average bed slope
(or energy slope in varied flow conditions) m/m (ft/ft).The maximum shear stress on the
channel bed is computed as follows (Chow, 1959):
id = ydS
(13.21)
where id = maximum shear stress, Pa (lb/ft ) and d = maximum depth of flow, m (ft).
Shear stress in channels is not distributed uniformly along the wetted perimeter of a
channel. A typical distribution of shear stress in a trapezoidal channel tends toward zero at
the corners with a maximum on the bed of the channel at its center line, and the maximum
for the side slopes occurs around the lower third of the slope, as illustrated in Fig. 13.2.
Flow around bends creates higher shear stresses on the channel sides and bottom compared to straight reaches. The maximum shear stress in a bend is a function of the ratio of
the channel's curvature to its bottom width. This ratio increases as the bend becomes
sharper and the maximum shear stress in the bend increases. The bend shear stress can be
computed using the following relationship:
1 = Kfii
2
(13.22)
where ib = bend shear stress, Pa (lb/ft ); Kb = function of RJB as shown in Figure 13.3;
Rc = radius to the centerline of the channel, m (ft); B = bottom width of channel, m (ft);
FIGURE 13.2 Distribution of shear stress. (Source: Brown, Stein, Warner, 1996)
ld = maximum channel shear stress, Pa (lb/ft2) calculated with Eq. (13.21). The increased
shear stress produced by the bend persists downstream of the bend a distance Lp9 which is
computed using the following relationship:
L, = ^-
(13.23)
where Lp = length of protection (length of increased shear stress due to the bend) from
the point of tangency downstream, m (ft); nb = Manning's roughness in the channel bend;
R = hydraulic radius, m (ft); and Kc = coefficient equal to 0.897 (0.736 in English units).
13.6.4 Parameters for Stable Channel Design
Parameters required for the evaluation and design of stable roadside and median channels
include discharge frequency, channel geometry, safety, channel slope, vegetation type,
freeboard, and shear stress. This section provides criteria relative to the selection or computation of these design elements. The applied shear stress is calculated and compared
with the permissible alternative shear stresses in the next section.
Discharge frequency. Roadside and median drainage channels are designed to accommodate the hydrologic design flow. However, when designing temporary channel linings,
a 2-yr return period is appropriate.
Channel geometry. Highway drainage channels are typically trapezoidal in shape.
Several typical shapes with equations for determining channel properties are illustrated in
Fig. 13.4. The depth, bottom width, and top width of the channel must be selected to provide the necessary flow area. The side slopes for triangular and trapezoidal channels
should not exceed the angle of repose of the soil, the lining material, or both and should
generally be 3:1 or flatter (Chen and Cotton, 1986) without consideration of the stability
of the side slopes.
Safety. Design of roadside and median channels should be integrated with the highway's geometric and pavement-design to insure proper consideration of safety and pavement drainage needs. In areas where traffic safety may be a concern, the side slopes of the
channel should be 4:1 or flatter.
Channel slope. Channel bottom slopes are generally dictated by the road profile.
However, if channel stability conditions warrant, it may be feasible to adjust the channel
gradient slightly to achieve a more stable condition. Channel gradients greater than 2 percent may require the use of flexible linings to maintain stability. Most flexible lining materials are suitable for protecting channel gradients of up to 10 percent, with the exception
of some grasses. Linings, such as riprap and wire-enclosed riprap, are more suitable for
protecting steep channels with gradients in excess of 10 percent. Rigid linings, such as
concrete paving, are stable to erosion but can be susceptible to failure from other causes,
such as overtopping, freeze-thaw cycles, swelling, and excessive soil pore water pressure.
Freeboard. The freeboard of a channel is the vertical distance from the water's surface
to the top of the channel. The importance of this factor depends on the consequence of
overflow of the channel bank. At a minimum, the freeboard should be sufficient to prevent waves, changes in superelevation, or fluctuations in water surface that result in the
sides overflowing. In a permanent roadside or median channel, about 150 mm (0.5 ft) of
freeboard is generally considered to be adequate. For temporary channels, no freeboard is
necessary. However, Fr's greater than one warrant freeboard height equal to the flow
depth to compensate for the large variations in flow caused by waves, splashing, and surging. Transitions from steep to mild also should have increased freeboard.
Permissible shear stress. The exerted shear stress is calculated with Eq. (13.21) or
(13.22) using the design parameters. Then, the permissible shear stress is used to find
an acceptable design alternative. The permissible or critical shear stress in a channel
defines the force required to initiate movement of the channel bed or lining material.
Table 13.11 presents permissible values of shear stress for manufactured, vegetative,
and riprap channel lining. The permissible shear stress values for noncohesive soils is a
function of the mean diameter of the channel material (Fig. 13.5). For larger sized
stones not shown in Fig. 13.5 and rock riprap, the permissible shear stress is given by
the following equation:
\ = K1P50
(13.24)
Lining Type
Temporary*
Vegetative
Gravel riprap
Rock riprap
Bare soil
0.15
0.45
28.7
40.7
69.5
74.3
85.7
177.2
100.6
47.9
228.7
16.8
15.7
31.4
95.7
191.5
See Fig. 13.5
See Fig. 13.6
0.60
0.85
1.45
1.55
2.00
3.70
2.10
1.00
0.60
0.35
0.33
0.67
2.00
4.00
Explonolion
"N" Volue
Loose
4-10
Medium Compoct 1 0 - 3 0
Compoct
30 - 50
NEffect
Number
Of Blows OfRequired
12" Penetration
The 2" To
Split-Spoon Sampler Seated To A
Depth Of 6" Driven With A 140 Lb.
Weight Foiling 30".
where ip = permissible shear stress, Pa (lb/ft2); D50 = mean riprap size, m (ft); and Kp =
empirical constant equal to 628 (4.0). For cohesive materials, the plasticity index provides
a good guide for determining the permissible shear stress illustrated in Fig. 13.6.
For trapezoidal channels protected with gravel or riprap having side slopes steeper
than 3:1, stability of the side slopes must be considered. This analysis is performed by
comparing the ratio of the tractive force between side slopes and channel bottom with the
ratio of shear stresses exerted on the channel sides and bottom. The ratio of shear stresses on the sides and bottom of a trapezoidal channel K1 is given in Fig. 13.7, and the tractive force ratio K2 is given in Fig. 13.8. The angle of repose 0 for different rock shapes
and sizes is provided in Fig. 13.9. The required rock size for the side slopes is found using
the following equation:
FIGURE 13.7 Channel side stress to botton shear stress ratio, K1. (Source: Brown,
Stein, Warner, 1996)
FIGURE 13.8 Tractive force ratio, K2. (Source: Brown, Stein, Warner, 1996)
(D50)sUes = ^(D
50)bomm
A
2
(13.25)
where D50 = the mean riprap size, m (ft.); K1 = ratio of shear stresses on the sides and
bottom of a trapezoidal channel as shown in Fig. 13.7; and K2 = ratio of tractive force on
the sides and bottom of a trapezoidal channel shown in Fig. 13.8.
73.7
DRAINAGEINLETDESIGN
As water accumulates and flows in the gutters and ditches, spread increases. At the point
where spread reaches the design value, an inlet is provided to capture some or all of the
flow to control the spread. Letting some gutter flow pass the gutter ("flow by") for subsequent capture can be efficient because it is uneconomical to make the inlets as wide as
the design spread.
The hydraulic capacity of a storm drain inlet depends on its geometry as well as on the
characteristics of the gutter flow. Inlet capacity governs both the rate of water removal
from the gutter and the amount of water that can enter the storm drainage system.
Inadequate inlet capacity or poor inlet location may cause flooding on the roadway, resulting in a hazard to the traveling public.
13.7.1 Inlets
Storm drain inlets are used to collect runoff and discharge it to an underground storm
drainage system. Inlets are typically located in gutter sections, paved medians, and roadside and median ditches. Inlets used for draining of highway surfaces can be divided into
the following four classes: (1) Grate inlets (grate-covered openings); (2) Curb-opening
inlets (vertical curb openings); (3) Combination inlets (grate plus curb opening), and (4)
Slotted inlets (under surface pipe with slotted crown).
Grate inlets, as a class, perform satisfactorily over a wide range of gutter grades but
lose capacity with increase in grade. A disadvantage is clogging. In addition, where bicycle traffic occurs, grates should be bicycle safe (narrow opening widths or transverse grating with the latter being hydraulically inefficient).
Curb-opening inlets are most effective on flatter slopes (less than 3 percent grade), in
sags, and with flows which typically carry significant amounts of floating debris. The interception capacity of curb-opening inlets significantly decreases as the gutter grade steepens.
Combination inlets provide a high-capacity inlet that offers the advantages of both
grate and curb-opening inlets. The curb opening precedes the grate in a "sweeper" configuration, and it acts as a trash interceptor. Used in a sag configuration, the sweeper inlet
can have a curb opening on both sides of the grate.
Slotted inlets can be used in areas where it is desirable to intercept sheet flow before
it crosses onto a section of roadway. Their principal advantage is their ability to intercept sheet flow over a wide section of roadway. Slotted inlets are highly susceptible to
clogging.
Inlet interception capacity Q1 is the flow intercepted by an inlet. The efficiency (E) of
an inlet is the fraction of total flow that the inlet will intercept. The efficiency changes
with changes in cross slope, longitudinal slope, total gutter flow, and, to a lesser extent,
pavement roughness. Efficiency is defined by
E=
(13-26)
where E = inlet efficiency; Q = total gutter flow, mVs (ftVs); and Q1 = intercepted flow,
m3/s (ftVs). Flow that is not intercepted by an inlet is termed carryover, bypass, or flow by
and is equal to
Q1, = Q - Q,
(13.27)
capacity of a grate inlet depends on the amount of water flowing over the gate, the size
and configuration of the grate, and the velocity of flow in the gutter. The interception
capacity of a curb-opening inlet is largely dependent on flow depth at the curb and curbopening length. Flow depth at the curb and, consequently, the interception capacity and
efficiency of the curb-opening inlet is increased by the use of a local gutter depression at
the curb-opening or a continuously depressed gutter to increase the proportion of the head
and flow adjacent to the curb. For curb inlets, top slab supports placed flush with the curb
line can substantially reduce the interception capacity of curb openings and therefore
should be avoided or be recessed and rounded in shape.
Slotted inlets function in essentially the same manner as curb opening inlets: i.e., as
weirs with flow entering from the side. Interception capacity depends on the depth of
flow and the length of the inlet. Efficiency depends on the depth of flow, the inlet length
of the inlet, and the total flow in the gutter.
A combination inlet consisting of a curb-opening inlet placed upstream of a grate inlet
has a capacity equal to that of the curb-opening length upstream of the grate plus that of
the grate, taking into account the reduced spread and depth of flow over the grate because
of the interception by the curb opening. This inlet configuration has the added advantage
of intercepting debris that might otherwise clog the grate and deflect water away from
the inlet.
Grate inlets in sag vertical curves operate as weirs for shallow ponding depths and as
orifices at greater depths. Between weir and orifice flow depths, a transition from weir to
orifice flow occurs. The capacity at a given depth can be affected severely if debris collects on the grate.
Curb-opening inlets operate as weirs in sag vertical curve locations up to a ponding
depth equal to the opening height. At depths above 1.4 times the opening height, the inlet
operates as an orifice and the transition between weir and orifice flow occurs between
these depths. The curb-opening height and length and the water depth at the curb affect
inlet capacity. The effective water depth at the curb can be increased by using a continuously depressed gutter, a locally depressed curb opening, or an increased cross slope, thus
decreasing the width of spread at the inlet. Slotted drains are not recommended in sag
locations because they are susceptible to clogging by debris.
13.7.2 Grate Inlet Design
Grates are effective highway pavement-drainage inlets when clogging by debris is not a
problem. Typical grate configurations are shown in Fig. 13.10. If clogging may be a problem, see Table 13.12, where grates are ranked for their susceptibility to clogging based on
laboratory tests using simulated leaves.
When the velocity approaching the grate is less than the "splash-over" velocity, the
grate will intercept essentially all the frontal flow. Conversely, when the gutter flow velocity exceeds the "splash-over" velocity for the grate, only part of the flow will be intercepted. A part of the flow along the side of the grate will be intercepted, depending on the
cross slope of the pavement, the length of the grate, and the velocity of flow.
The ratio of frontal flow to total gutter flow E0 for a uniform cross slope is expressed by
*. - ^ - 1 - fl - "I
(13.28)
where Q = total gutter flow, mVs (ftVs); Qw = flow in width, W, m3/s (ftVs); W = width
of depressed gutter or grade, m (ft); and T = total spread of water, m (ft). The ratio of side
flow Qs to total gutter flow is
Grate Type
Parallel bars with
1-7/8* spacing
(not bicycle safe)
Description
Comments
Adding tranaverae
) bars
4' o. c. to
1/4 x 4* bars'
on center
3/8 x 4* bars
1-1/8* bar spacing
on center
45 tilt bar
4* vane spacing
1/2 x 2* flat
(See fig. 12, HEC-12)
longitudinal bars
3-7/32* bar spacing
on center
30 tilt bar
1/4 x 4* flat
tranaverae membera
2-9/16* o. c.
3/16 x 2* reticuline performance.
bars
Reticuline
Grate
Curved vane
30-85 tilt bar
45-85 tilt bar
P-50
P-50 X 100
45-60 tilt bar
Reticuline
P-30
Longitudinal Slope
0.005
0.040
46
44
43
32
18
16
12
9
61
55
48
32
28
23
16
20
^ = 1 - ^ = 1 - E0
(13.29)
(13.30)
where K0 = constant equal to 0.295 (0.09 in English units); V = velocity of flow in the gutter, m/s; and V0 = gutter velocity where splash-over first occurs, m/s. (Note: /^cannot exceed
1.0.) The splash-over velocity and /^ factor are solved by the chart in Fig. 13.11 as a function
of the type and length of the grate and average gutter velocity approaching the inlet.
The ratio of side flow intercepted to total side flow Rs, or side flow interception efficiency, is expressed by
*-r^|
(
SJU")
where Kc = a constant equal to 0.0828 (0.15 in English units).
The overall efficiency E of a grate is expressed by the following equation:
E = R1E0 + R1(I- E0)
(13.32)
Flow
EXAMPLE
Given:
W 12 In.; L - 20 In.; T 10 ft.
SX 0.03; Su 4%; n 0.010
(aee Chart 6)
Then:
V 8.2 fpe from Chart 9
E0 0.26 from Chart 7
R, - 0.88 from chart at left
Orate Efficiency E R E
E 0.8S x 0.28 0.21 1 0
Note: Side Inflow to email
rectangular Inlet* Ie negligible.
FIGURE 13.11 Grate inlet frontal flow interception efficiency. Source: HEC-12 (Johnson and
Chang, 1984, English units preserved from source document)
The interception capacity of a grade inlet on a grade is equal to the efficiency of the
grate multiplied by the total gutter flow:
Qi = EQ = Q[RfE0 + Rs(l- E0)]
(13.33)
(13.35)
(13.36)
where S'w = cross slope of the depressed gutter and E0 is the same ratio used to compute
the frontal-flow interception of a grate inlet (Eq. (13.28)). The length of curb opening
required for total interception can be reduced significantly by increasing the cross slope
or the equivalent cross slope. Increasing the equivalent cross slope can be accomplished
by using of a continuously depressed gutter section or locally depressed gutter sections.
To compute efficiency, E for curb inlets shorter than the length needed for 100 percent
capture of flow, Eq. (13.35) is applicable with either straight cross slopes or composite
cross slopes.
13.7.4 Slotted Inlet Design
Placed at right angles to the flow, the slot acts as a short grate. Assuming a splash-over
velocity of 0.3 m/s (1 ft/s) and no side flow, the grate-inlet efficiency equations can
be used. Placed parallel to the flow, in the gutter notch, the flow interception by slotted
inlets and curb-opening inlets is similar. Analysis of data from the Federal Highway
Administration's tests of slotted inlets with slot widths > 45 mm (1.75 in) indicates that
the length of slotted inlets required for total interception can be computed with the curb-
opening equations. It should be noted that it is less expensive to add length to a slotted
inlet to increase its interception capacity than it is to add length to a curb-opening inlet.
Wide experience with the debris-handling capabilities of slotted inlets is not available. Deposition of debris in the pipe is a commonly encountered problem. However,
the configuration of slotted inlets make them accessible for cleaning with a high-pressure water jet.
13.7.5 Combination Inlet Design
Because the interception capacity of a combination inlet consisting of a curb opening
and a grate placed side by side is no greater than that of the grate alone, capacity is computed by neglecting the curb opening. Typical designs place the curb opening upstream
of the grate. The curb opening intercepts debris that might clog the grate and is called
a "sweeper" inlet. A sweeper combination inlet has an interception capacity equal to the
sum of the curb opening upstream of the grate plus the grates capacity, except that the
frontal flow, and thus the interception capacity of the grate, is reduced by interception
by the curb opening.
13.7.6 Design Adjustments for Sag Locations
Inlets in sag locations operate as weirs under low head conditions and as orifices at greater
depths and are more susceptible to clogging. Orifice flow begins at depths dependent on
the grate size, the height of the curb opening, or the slot width of the inlet. At depths
between those at which weir flow definitely prevails and those at which orifice flow prevails, flow is in a transition stage. Grate inlets alone are not recommended for use in sag
locations because of the tendencies of grates to become clogged. Combination inlets or
curb-opening inlets are recommended for use in these locations.
The capacity of grate inlets operating as weirs is
G1- = CJ>d^
(13.37)
where P = perimeter of the grate in m (ft) disregarding the side against the curb, Cw =
weir coefficient = 1.66 (3.0 English units) and d = flow depth, m (ft). The capacity of a
grate inlet operating as an orifice is
Q1 = C^g (2gdy*
(13.38)
where C0 - orifice coefficient = 0.67; Ag = clear opening area of the grate, m2 (ft2); g =
9.90 m/s2 (32.16 ft/s2); and d = flow depth, m (ft).
Use of Eq. (13.38) requires the clear area of opening of the grate. Tests of three grates
for the Federal Highway Administration (Burgi, 1978) showed that for flat bar gates, the
clear opening is equal to the total area of the grate less the area occupied by longitudinal
and lateral bars. However, curved vane grate performed about 10 percent better than one
would expect by calculating a net opening equal to the total area less the area of the bars
projected on a horizontal plane. Tilt bar and curved vane grates are not recommended for
sump locations.
The capacity of a depressed curb-opening inlet operating as a weir (curb water depths
less than opening height) is
G1- = Cw (L + 1.8W)J15
(13.39)
where Cw weir coefficient = 1.25 (2.3 in English units); L = length of curb opening,
m (ft); W = lateral width of depression, m (ft); and d = depth at curb measured from the
normal cross slope, m (ft): i.e., d = TSx. For curb-opening inlets with a continuously
depressed gutter, it is reasonable to expect that the effective weir length would be as great
as that for an inlet in a local depression. The weir equation for curb-opening inlets without any depression simplifies to
Qf = CwLd1-5
(13.40)
Without depression of the gutter section, the weir coefficient Cw becomes 1.60 (3.0,
English system).
At curb-opening lengths greater than 3.6 m (12 ft), Eq. (13.40) for nondepressed inlets
produces captured flows that exceed the values for depressed inlets computed using Eq.
(13.39). Since depressed inlets will perform at least as well as nondepressed inlets of the
same length, Eq. (13.40) should be used for all curb-opening inlets longer than 3.6 m
(12 ft).
Curb-opening inlets operate as orifices at curb depths greater than approximately
1.4 times the opening height. The capacity of curb-opening inlets operating as
orifices is
Qi= C0hL(2gd)*
(13.41)
r (
}a5
Q1 = C0AJ 2 ^ U - I
(13.42)
or
where C0 = orifice coefficient (0.67); d0 = effective head on the center of the orifice
throat, m (ft); L = length of orifice opening, m (ft); A^ = clear area of opening, m2 (ft2);
di depth of lip of curb opening, m (ft); and h = height of curb-opening orifice, m (ft).
The above variables are detailed in Fig. 13.12 as a function of typical variations in geometric configuration. For curb-opening inlets with other than vertical faces (see Fig.
13.12), Eq. (13.41) can be used with h = orifice throat width, m (ft) and d0 = effective
head on the center of the orifice throat, m (ft).
Slotted inlets in sag locations perform as weirs to depths of above 0.06 m (0.2 ft),
depending on slot width. At depths greater than approximately 0.12 m (0.4 ft), they perform as orifices. Between these depths, flow is in a transition stage. The capacity of a slotted inlet operating as a weir can be computed by
Q1 = CJLW*)
(13.43)
where Cw = weir coefficient =1.4 (2.2248 for English units); L = length of slot, m (ft);
and d = depth at curb measured from the normal cross slope, m (ft). The capacity of a
slotted inlet operating as an orifice can be computed by
Q1 = 0.8 L W(2gd)-5
(13.44)
where W = width of slot, m (ft); L = length of slot, m (ft); d = depth of water at slot for
d ^0.12 m (0.4 ft), m (ft); and g = 9.81 m/s2 (32.16 ft/s2).
Combination inlets consisting of a grate and a curb opening are considered advisable
for use in sags. Equal-length inlets refer to a grate inlet placed alongside a curb-opening
a. Horizontal Throat
b. Inclined Throat
c. Vertical Throat
FIGURE 13.12 Curb-opening inlets.
Source: (Brown, Stein, and Warner, 1996)
inlet, both of which have the same length. A sweeper inlet refers to a grate inlet placed at
the downstream end of a curb-opening inlet; the curb-opening inlet is longer than the grate
inlet and intercepts the flow before the flow reaches the grate. The sweeper is more efficient than the equal-length combination inlet and the curb opening and has the ability to
intercept any debris which may clog the grate inlet. The capture capacity of the equallength combination inlet in weir flow is essentially equal to that of a grate alone. In orifice flow, the capacity of the equal-length combination inlet is equal to the capacity of the
grate plus the capacity of the curb opening. A typical assumption is to size sweeper inlets
at the sag, assuming that the grate in between is 100 percent clogged.
13.7.7 Inlet Locations
The location of inlets is determined by geometric controls that require inlets at specific
locations, the use and location of flanking inlets in sag vertical curves, and the criterion
of spread on the pavement at the design rainfall intensity. To design the location of the
inlets adequately for a given project, the following information is needed:
A layout or plan sheet suitable for outlining drainage areas.
Road profiles.
Typical cross sections.
Because the typical inlet is located at a gutter location having sufficient drainage area
to generate the design spread, full or partial removal of spread is accomplished. Bypass of
some of the gutter flow, although leading to more frequent spacing, can result in much
smaller inlets at considerable cost savings.
There are a number of locations where inlets may be necessary with little regard to
contributing drainage area. These locations should be marked on the plans before any
computations regarding discharge, water spread, inlet capacity, or flow bypass. Examples
of such locations follow:
at all sag points in the gutter grade
immediately upstream of median breaks, entrance/exit ramp gores, crosswalks, and
street intersections: i.e., at any location where water could flow onto the travelway
immediately upgrade of bridges (to prevent pavement drainage from flowing onto
bridge decks)
immediately downstream of bridges (to intercept bridge deck drainage)
immediately upgrade of cross-slope reversals
at the end of channels in cut sections
on side streets immediately upgrade from intersections
behind curbs, shoulders, or sidewalks to drain low areas
In addition to the areas identified above, runoff from areas draining toward the highway
pavement should be intercepted by roadside channels or inlets before it reaches the roadway. This applies to drainage from cut slopes, side streets, and other areas alongside the
pavement.
13.8
From a hydraulic standpoint, the ideal bridge has no sag point and bridge deck drainage
is not needed (Young et al., 1993). Such a bridge has off-bridge inlets at the upper end to
remove gutter flow before it reaches to the bridge and at the lower end to collect bridge
deck drainage. If the bridge is too long, spread on the bridge deck may necessitate inlets.
On a long bridge, typical inlets are small and are sometimes called scuppers: a 3-m (10ft) shoulder is desirable for maintenance of scuppers.
The off-bridge inlets are typically curb inlets and are designed using methods presented in Sec. 13.7. The on-bridge inlets, if rectangular, also can be designed using
equations in Sec. 13.7; the dimensions typically are less than 1 m to a side. If the inlets
are circular, their diameter is typically low. Figure 13.13 gives the capture efficiency for
circular inlets.
The dimensions of inlets are relatively constrained for placement and integration into
bridge decks. Many decks are pre- or post-tensioned structural slabs, and inlets are
detailed that may interfere with structural continuity. Thus, the surface grate and the
recessed collection chamber must be considered.
EXAMPLE
Given:
D * 6 in.; T 4 ft.
Su = 2%
Then:
D/TS.125
E * 0.33
FIGURE 13.13 Efficiency curves for circular scuppers. (Source: Young, Walker and Chang, 1993, English units preserved from source
document)
The constraints lead to small inlets and are associated with reinforcing bar schedules
or post-tensioned cable spacings. Typical dimensions need to be less than 0.5 m (18 in);
these details need to be structurally designed to transfer loads into the slab. Large inlet
spans cause a need for special designs and reinforcing details.
The hydraulic problem is that spreads of 2.5 to 3.0 m (8 to 10 ft) of water in gutters
are not reduced effectively with small inlets having capture efficiencies of 5 to 10 percent.
Numerous closely spaced small inlets are necessary to control spread.
Ideally, a bridge should have 3-m (10-ft) shoulders and the inlet boxes should be
placed at the outside edge of the shoulder. In this position, the maintenance crew can
park on the shoulder and work on the side away from traffic in reasonable safety. These
drains have a good chance of being maintained regularly. Unfortunately, accidents may
happen when lanes must be blocked to service the drains or when the maintenance crew
must work on the edge of the stream of traffic. This can result in poorly placed inlets
receiving inadequate maintenance. The larger the inlet, the fewer inlets to maintain.
Hydraulically, larger inlets handle larger volumes of flow and are more apt to clean
themselves. The larger the inlet, the easier it is to clean with a shovel. Inlets should be
sized as large as possible: Practically speaking, 1 m (36 in) is probably an upper limit
for inlets placed within deck slabs.
Round vertical scuppers should not be less than 155 mm (6 in) in diameter, with a
diameter of 200 m (8 in) preferred. Scuppers with a diameter of 100 mm (4-in) are not
uncommon in practice. However, their limited hydraulic capacity, coupled with their tendency to plug with debris, mitigates recommending their use. Although such features are
easy to place, they are relatively ineffective with capture efficiencies on the order of 5 percent. Nonetheless, they may be convenient when drainage can fall directly to underlying
surfaces without under-deck pipe collection and downspouts. With direct fall of water,
their small size is an advantage.
The water collected at inlets either falls directly to surfaces beneath bridges or is collected. Collected storm water is conveyed to bridge support columns and downspouts
that are affixed to vertical bridge members. The collectors are typically cast iron. The
collector pipes should be sloped at least 2 percent (20 mm/m [1/4 in/ft] or more-preferably 8 percent-to provide sufficient velocities at low flow to move silt and small debris
to avoid clogging.
13.8.1 Inlet Design for Constant-Grade Bridges
The general procedure is to start at the high end of the bridge and work downslope from
inlet to inlet. First, the designer selects a rainfall intensity and design spread. General
bridges dimensions, bridge grade, roughness and runoff coefficients, and inlet specifications are assumed to be known. The procedure is as follows:
1. Find the appropriate rainfall intensity. If the rational method is used, assume a time of
concentration of 5 min.
2. Find the flow on the deck Q at design spread T using Eq. (13.13).
3. Start at the high end of the bridge and compute the inlet spacing using Eqs. (13.45) and
(13.46), the derivations of which are given by Young et al. (1993):
L0 = ^j-for the first inlet
and
(13.45)
(13.46)
where / = design rainfall intensity, mm/h (in/h) (Step V)Q = gutter flow, m3/s Lc = constant distance between inlets, m (ft). L0 distance to first inlet, m (ft). C = Rational
runoff coefficient. Wp = width of pavement contributing to gutter flow, m (ft). E = constant, which is equal to 1 for first inlets in all cases and is equal to capture efficiency for
subsequent inlets of constant-slope bridges and Kc = unit conversion equal to 79,662.5
(43,560 in English units).
Because the first inlet receives virtually no bypass from upslope inlets, the constant E
can be assumed to be equal to 1. If the computed distance L0 is greater than the length of the
bridge, then inlets are not needed and only bridge-end treatment design need be considered.
4. If inlets are required, the designer should proceed to calculate the constant inlet spacing Lc for the subsequent inlets.
5. If the bridge slope is nearly flat (less than about 0.003 m/m [0.003 ft/ft]), then the procedures for flat bridges should also be followed as a check.
13.8.2
Four steps are involved when designing inlets for flat bridges:
1. Rainfall intensity, design spread (T), pavement width (Wp), bridge length (LB),
Manning's n (n), rational runoff coefficient (C), and gutter cross slope (Sx) are assumed
to be known. If the rational method is used, assume a time of concentration of 5 min
and determined the intensity from the intensity-duration-frequency (IDF) curves.
2. Constant inlet spacing Lc can then be computed using Equ. (13.47), the derivation of
which is given by Young, et al. (1993):
^(^c^8*-4472'11
(13 47)
H348)
(
;
For nearly flat stations on vertical curves or constant grades, the designer should set spacing assuming tthat he bridge is flat.
13.9
SAMPLEPROBLEMS
Length (m)
Slope (m/m)
1
2
3
4
25
43
79
146
0.005
0.005
0.006
0.008
Segment description
Short-grass prairie
Short-grass pasture
Grassed waterway
380-mm concrete pipe
Find: Rainfall intensity I = 60 mm/hr the time of concentration Tc for the area.
Solution
Step 1. Calculate time of concentration for each segment.
Segment 1: Obtain Manning's n roughness coefficient from Table 13.4: n = 0.15.
then determine the sheet flow travel time using Eq. (13.4):
T+1 = (6.943/104) (/z/S-5)0-6 = [6.943/(6O)04] [(0.15) (25)7(0.005)5]06 - 14.6 min
Segment 2: Obtain intercept coefficient k from Table 13.5: k = 0.213. Then
determine the concentrated flow velocity from Eq. 13.5:
Vc = K1IcSn0-5 = (1)(0.213)(0.5)0.5 - 0.15 m/s
Segment 3: Obtain intercept coefficient k from Table 13.5: k = 0.457. Then
determine the concentrated flow velocity from Eq. (13.5):
Vc= KckSp-5 = (1)(0.457)(0.6)05 - 0.35 m/s
Segment 4: Obtain Manning's n roughness coefficient from Table 13.6: n =
0.011. Then determine the pipe flow velocity from Eq. (13.7):
V= (1.0/0.011)(0.38/4)067 (0.008)05 = 1.7 m/s
Step 2. Determine the total travel time from Eq. 13.8:
Tc = T+1 + T+2 + T+3 + T+4
Tc = 4.6 + 43/[(6O)(0.15)] + 79/[(60)(0.35)] + 146/[(6O)(1.7)] - 24.5 min.
use 25 min.
Area, ha
Runoff coefficient C
8.95
8.60
17.55
0.25
0.22
Area, ha
Paved
Lawn
Unimproved Grass
Grass
Total
Runoff coefficient C
2.20
0.66
7.52
7.17
17.55
0.90
0.15
0.25
0.22
Times of concentration:
Existing condition (unimproved):
Proposed condition (improved):
60
65
70
75
80
85
90
60
57
55
53
51
49
47
Rainfall intensity I
Existing condition (unimproved):
48 mm/hr
Proposed condition (improved):
58 mm/hr
Step 4. Determine peak flow rate Q using equation 13.2:
Existing condition (unimproved):
Q = CIAIKc
= (0.235)(48)(17.55)/360 = 0.55 mVs
Proposed condition (improved):
Q = CIAlKc
= (0.315)(58)(17.55/360 = 0.88 mVs
Example 3
Given: Gutter flow with a spread of 3 m, n = 0.016, Sx = 0.02, and 1 percent grade.
Find: Flow
Solution
Determine the flow using Eq. 13.13:
Q = (0.376/0.016X0.02)167(0.01)05 (3)2'67 = 0.064 mVs
Example 4
Given: A curved vane grate inlet which is 0.3 m wide and 0.5 m long with a full flow of
0.064 mVs and gutter characteristics from Example 3.
Find: The inlet capture efficiency.
Solution
Step 1. Determine the ratio of frontal flow to gutter flow using Eq. 13.28:
E0 = 1 - [1 - (0.3/3)]267 - 0.245
Step 2. Determine the velocity using Eq. 13.6:
Vg = (0.752/0.016) (O.Ol)0'5 (0.02)067 (3)67 = 0.71 m/s
Step 3. Determine the frontal flow interception ratio using Fig. 13.11: For a 0.5 m (1.6 ft)
long grate, the splashover velocity is .4 m/s (4.6 ft/s), which is greater than the
gutter velocity. This implies an Rfof 1.0.
Step 4. Determine the side flow interception ratio using Eq. 13.31:
R5 = 1/(1 + ([(0.0828)(0.71)18]/[(0.02)(0.5)2/3])) - 0.22
Step 5. Calculate the overall efficiency using Eq. 13.32:
E = (1.0X0.245) + (0.22) (1 - 0.245) - 0.411
Example 5
Given: the configuration described in Example 4 and pavement width of 11 m, a rainfall intensity of 75 mm/hr, and a rational runoff coefficient of 0.7 (pavement and median).
REFERENCES
AASHTO, "Storm Drainage Systems", Model Drainage Manual Chapter 13. American Association
of State Highway and Transportation Officials, Washington, DC, 1991.
AASHTO, Policy on Geometric Design of Highways and Streets, American Association of State
Highway and Transportation Officials, Washington, DC, 1990.
Anderson, D. A., J. R., Reed, R. S., Huebner, J. J., Henry, W. R, Kilareski, and J. C., Warner,
Improved Surface Drainage of Pavements, Hydraulic Engineering Circular No. 12, Federal
Highway Administration, U.S. Department of Transportation, Washington, DC, 1995.
ASCE, "Design and Construction of Urban Stormwater Management Systems ", ASCE Manuals
and Reports of Engineering Practice No. 77, WEF Manual of Practice RD-20, American Society
of Civil Engineers, New York, 1992.
ASCE, Design Manual for Storm Drainage, American Society of Civil Engineers, New York, 1960.
Brown, S. A., S. M., Stein, and J. C., Warner, Urban Drainage Design Manual, Hydraulic
Engineering Circular No. 22, FHWA-SA-96-078, Federal Highway Administration, U.S.
Department of Transportation, Washington, DC, 1996.
Burgi, P. H., D. E. Gober, June 1977. Bicycle-Safe Grates Inlets Study, Volume 1 Hydraulic and
Safety Characteristics of Three Selected Grate Inlets on Continous Grades. Report No. FHWARD-77-24, Federal Highway Administration.
Burgi, P. H., May 1978. Bicycle-Safe Grate Inlets Study, Volume 2 Hydraulic Characteristics of
three Selected Grate Inlets on Continous Grades. Report No. FHWA-RD-78-4, Federal Highway
Administration.
Burgi, P. H., Bicycle-Safe Grate Inlets Study: Vol. 3 - Hydraulic Characteristics of Three Selected
Grate Inlets in a Sump Condition, Report No. FHWA-RD-78-70, Federal Highway
Administration, Washington, DC, September 1978.
Chen, Y. H., and G. K.,Cotton, Design of Roadside Channels with Flexible Linings, Hydraulic
Engineering Series No. 15, Publication No. FHWA-IP-87-7, prepared for the Federal Highway
Administration, U.S. Department of Transportation, Washington, DC, 1986.
Chow, V. T, Open-Channel Hydraulics, McGraw-Hill, New York, 1959.
Federal Highway Administration, Design Charts for Open-Channel Flow, Hydraulic Design Series
No. 3, U.S. Department of Transportation, Washington, DC, 1977 (reprint).
Federal Highway Administration, Hydraulic Design of Energy Dissipators, Hydraulic Engineering
Circular No. 14, U.S. Department of Transportation, Washington, DC, 1983.
Gallaway, B. C., et al, December 1979.Pavement and Geometric Design Criteria for Minimizing
Hidroplaning. Texas Transportation Institute, Texas A&M University, Federal highway
Administratuion, Report No. FHWA-RD-79-30, A Technical Summary.
Izzard, C. R, "Hydraulics of Runoff from Developed Surfaces," Proc. Highway Research Board,
Vol. 26, p. 129-150, Highway Research Board, Washington, DC, 1946.
Jennings, M. E., W. O., Thomas, Jr., and H. C., Riggs, Nationwide Summary of U.S. Geological
Survey Regional Regression Equations for Estimating Magnitude and Frequency of Floods for
Ungaged Sites, 1993, U.S. Geological Survey, Water-Resources Investigations Report No. 944002, prepared in cooperation with the Federal Highway Administration and the Federal
Emergency Management Agency, Reston, VA, 1994.
Johnson, F. L., and F. M., Chang, Drainage of Highway Pavements, Hydraulic Engineering Circular
No. 12, FHWA-TS-84-202, Federal Highway Administration, Washington, DC, 1984.
McCuen, R. H., P. A., Johnson, and R. M., Ragan, Hydrology, Hydraulic Design Series No. 2,
FHWA-SA-96-067, Federal Highway Administration, U.S. Department of Transportation,
Washington, DC, 1996.
M. E. Jennings, W. O. Thomas, Jr., and H. C. Riggs, 1994. Nationwide Summary of U.S.
Geological Survey Regional Regression Equations for Estimating Magnitude and Frequency of
Floods for Ungaged Sites, 1993. US Geological Survey, Water-Resources Investigations Report
94-4002, prepared in cooperation with the federal Highway Administration and the Federal
Emergency Managment Agency, Reston, Virginia.
Richardson, E. V., D. B., Simons, and P. Y., Julien, Highways in the River Environment, prepared
for the Federal Highway Administration, Washington, D.C, by the Department of Civil
Engineering, Colorado State University, Fort Collins, Colorado, 1990.
Young, G. K., S. M., Walker, and R, Chang, Design of Bridge Deck Drainage, Hydraulic
Engineering Circular No. 21, FHWA-SA-92-010, Federal Highway Administration, U.S.
Department of Transportation, Washington, D.C, 1993.
V. B. Sauer, W O. Thomas Jr., V. A. Strieker, and K. V. Wilson, 1983. Flood Characteristics of
Urban Watersheds in the United States. Prepared in cooperation with U. S. Department of
Transportation, Federal Highway Administration, U.S. Geological Survey Water-Supply Paper
2207.
V. B. Sauer, 1989. Dimensionless Hydrograph Method of Simulating Flood Hydro graphs. Preprint,
68th Annual Meeting of the Transportation Research Board, WEashington, D.C., (January) pp.
22-26.
CHAPTER 14
HYDRAULIC DESIGN OF
URBAN DRAINAGE SYSTEMS
U.I.
INTRODUCTION
Generally speaking urban drainage systems consist of three parts: the overland surface
flow system, the sewer network, and the underground porous media drainage system.
Some elements of these components are shown schematically in Fig. 14.1. Traditionally
no design is considered for the urban porous media drainage part. Recently porous media
drainage facilities such as infiltration trenches have been designed for flood reduction or
pollution control in cities with high land costs. For example, preliminary work on this
aspect of urban porous media drainage design can be found in Fujita (1987), Morita et al.
(1996), Takaaki and Fujita (1984) and Yen and Akan (1983). Much has yet to be developed to refine and standardize on such designs; no further discussion on this underground
subject will be given in this chapter.
From a hydraulic engineering viewpoint, urban drainage problems can be classified
into two types: (1) design and (2) prediction for forecasting or operation. The required
hydraulic level of the latter is often higher than the former. In design, a drainage facility
is to be built to serve all future events not exceeding a specified design hydrologic level.
Implicitly the size of the apparatus is so determined that all rainstorms equal to and
smaller than the design storm are presumably considered and accounted for. Sewers,
ditches, and channels in a drainage network each has its own time of concentration and
hence its own design storm. In the design of a network all these different rainstorms
should be considered. On the other hand, in runoff prediction the drainage apparatus has
already been built or predetermined, its dimensions known, and simulation of flow from
a particular single rainstorm event is made for the purpose of real-time forecasting to be
used for operation and runoff control, or sometimes for the determination of the flow of
a past event for legal purposes. The hydrologic requirements for these two types of problems are different. In the case of prediction, a given rainstorm with its specific temporal
and spatial distributions is considered. For design purposes, hypothetical rainstorms with
assigned design return period or acceptable risk level and assumed temporal and spatial
FIGURE 14.1 Schematic of components of urban catchment. (From Metcalf & Eddy, Inc. et al., 1971).
distributions of the rainfall are used. Table 14.1 lists some of these two types of design
and prediction problems.
In the case of sanitary sewers, for design purposes the problem becomes the estimation of the critical runoffs in both quantity and quality, from domestic, commercial, and
TABLE 14.1 Types of Urban Drainage Problems (a) Design Problems
Type
Design Purpose
Hydro Information
Sought
Required Hydraulic
Level
Sewers
Low
Low to moderate
Low to moderate
Drainage channels
Detention/retention
storage ponds
Manholes and
junctions
Roadside gutters
Geometric dimensions
(and outlet design)
Geometric dimensions
Geometric dimensions
Pumps
Control gates
or valves
Capacity
Capacity
Geometric dimensions
Low to moderate
Low to moderate
Low
Moderate to high
Moderate to high
TABLE 14.1 (continued) Types of Urban Drainage Problems (b) Prediction Problems
Type
Purpose
Hydro Input
Real-time
operation
Real-time
regulation
of flow
Simulation for
evaluation of
a system
Determination
of runoff at
specific locations
for particular past
or specified events
Determination
of the extent
of flooding
Reduce and
control of water
pollution due to
runoff from
rainstorms
Long-term, usually
large spatial scale
planning for
stormwater
management
Performance
evaluation
Storm event
simulation
Flood level
determination
Storm runoff
quality
control
Storm runoff
master
planning
Hydro Information
Sought
Required Hydraulic
Level
Moderate-high
Specific storm
hyetographs,
netwark data
Event or continuous
rain and pollutant
data, network data
Hydrographs and
stages
High
Long-term data
Runoff volume
Pollutant volume
Low
industrial sources over the service period in the future. For real-time control problems it
involves simulation and prediction of the sanitary runoff in conjunction with the control
measures.
The basic hydraulic principles useful for urban drainage have been presented in
Chapter 3 for free surface flows, Chapters 2 and 12 for pipe flows, and Chapter 10 for
pump systems. In the following, more specific applications of the hydraulics to urban
drainage components will be described. However, the hydraulic design for drainage of
highway and street surfaces, roadside gutters, and inlets has been described in Chapter 13,
design of stable erodible open channels in Chapter 16, and certain flow measurement
structures adaptable to urban drainage in Chapter 21; therefore they are not included in
this chapter.
74.2
Flows in urban drainage channels usually are open-channel flows with a free water surface.
However, sewer pipes, culverts, and similar conduits under high flow conditions could
become surcharged, and pressurized conduit flows do occur. Strictly speaking, the flow is
always unsteady, that is, changing with time. Nevertheless, in a number of situations, such
as in most cases of flow in sanitary sewers and for some rainstorm runoffs, change of flow
with time is slow enough that the flow can be regarded as approximately steady.
14.2.1 Open-Channel Flow
Open-channel flow occurs on overland, ditches, channels, and sewers in urban areas.
Unsteady flow in open channels can be described by a momentum equation given below
in both discharge (conservative) and velocity (nonconservative) forms together with its
various simplified approximate models:
_ L M + i afPfi!]+ i f Uj11*,+ -*. +s,f = o.
gA dt
gAdx(A ) gA Ja ^
dx
(i4.i)
dynamic wave
quasi-steady dynamic wave
noninertia
kinematic wave
lf
(2p-l)^ + ( p - l ) ^ + M + f - 5 . + ^a
(14.2)
where x = flow longitudinal direction measured horizontally (Fig. 14.2); A = flow crosssectional area normal to x\ y = vertical direction; Y = depth of flow of the cross section,
measured vertically; Q = discharge through A; V = QIA, cross-sectional average velocity along x direction; S0 = channel slope, equal to tan 6, 9 = angle between channel bed
and horizontal plane; Sf = friction slope; a = perimeter bounding the cross section A; ql
= lateral flow rate (e.g., rain or infiltration) per unit length of channel and unit length of
perimeter a, being positive for inflow; Ux = ^-component velocity of lateral flow when
(14.3)
f + s (vy) = U-*
(14 5)
dard hydraulics and fluid mechanics references. Its major disadvantage is that for a given
pipe and surface roughness, the value of/varies not merely with the Reynolds number but
also with the flow depth. In other words, as the flow depth in the sewer changes during a
storm runoff,/must be recomputed repeatedly.
Manning's n was originally derived empirically. Its major disadvantage is its troublesome dimension of length to one-sixth power that is often misunderstood. Its main advantage is that for flows with sufficiently high Reynolds number over a rigid boundary with
a given surface roughness in a prismatic channel, the value of n is nearly constant over a
wide range of depth (Yen, 1991). Values of n can be found in Chow (1959) or Chap. 3.
Other resistance coefficients and formulas, such as Chezy's or Hazen-Williams's, have
also been used. They possess neither the direct fluid mechanics justification as/nor independence of depth as n. Therefore, they are not recommended here. In fact, HazenWilliams' may be considered as a special situation of Darcy-Weisbach's formula. A discussion of the preference of the resistance coefficients can be found in Yen (1991).
Equations (14.6) and (14.7) are applicable to both surcharged and open-channel
flows. For the open-channel case, the pipe is flowing partially filled and the geometric
parameters of the flow cross section are computed from the geometry equations given
in Fig. 14.3.
The pair of momentum and continuity equations [Eqs. (14.1) and (14.4) or Eqs. (14.2)
and (14.5)] with ( 3 = 1 and no lateral flow is often referred to as the Saint-Venant equations or full dynamic wave equations. Actually, they are not an exact representation of the
unsteady flow because they involve at least the following assumptions: hydrostatic
pressure distribution over A, uniform velocity distribution over A (hence (3 = 1), and negligible spatial gradient of the force due to internal stresses.
Those interested in the more exact form of the unsteady flow equations should refer to
Yen (1973b, 1975,1996). Conversely, simplified forms of the momentum equation, namely, the noninertia (misnomer diffusion wave) and kinematic wave approximations of the
full momentum equation [Eq. (14.1)] are often used for the analysis of urban drainage
flow problems.
Among the approximations shown in Eqs. (14.1) or (14.2), the quasi-steady dynamic
wave equation is usually less accurate and more costly in computation than the noninertia equation, and hence, is not recommended for sewer flows. Akan and Yen (1981),
among others, compared the application of the dynamic wave, noninertia, and kinematic
wave equations for flow routing in networks and found the noninertia approximation generally agrees well with the dynamic wave solutions, whereas the solution of the kinematic wave approximate is clearly different from the dynamic wave solution, especially when
the downstream backwater effect is important. Table 25.2 of Yen (1996) gives the proper
form of the equations to be used for different flow conditions.
Kinematic
wave
Noninertia
Quasi-steady
dynamic wave
Dynamic
wave
1
No
2
Yes
2
Yes
2
Yes
No
No
Yes
No
Yes
Only
convective
acceleration
Yes
Yes
Analytical solutions do not exist for Eqs. (14.1) and (14.2) or their simplified forms
except for very simple cases of the kinematic wave and noninertia approximations.
Solutions are usually sought numerically as described in Chap. 12. In solving the differential equations, in addition to the initial condition, boundary conditions should also be
properly specified. Table 14.2 shows the boundary conditions required for the different
levels of approximations of the momentum equation. It also shows the abilities of the
approximations in accounting for downstream backwater effects, flood peak attenuation,
and flow acceleration.
For flows that can be considered as invariant with time the steady flow momentum
equations which are simplified from Eq. (14.2) for different conditions are given in Table
14.3. The lateral flow contribution, mq, can be from rainfall (positive) or infiltration (negative) or both. Instead of these equations, the following Bernoulli total head equation is
often used for flow profile computations:
TABLE 14.3 Cross-Section-Averaged One-Dimensional Momentum Equations for Steady Flow of
Incompressible Homogeneous Fluid
Prismatic channel
Constant piezometric
pressure distribution
K = K' = 1
P = constant
K=K' = I
Prismatic or wide
channel
Definitions:
Dh=A/ water surface width;
oc V2
a V2
-^f + Y2 +yb2 = -^+Yl+ybl + he + hq,
(14.8)
where the subscripts 1 and 2 = the cross sections at the two ends of the computational reach,
Ax, of the channel, Y + yb the stage of the water surface where the channel bed elevation
at section 1 is ybl and that at section 2 is yb2 = ybl + S0 Ax; hq is the energy head from the
lateral flow, if any; the energy head loss he = Se Ax where Se is the slope of the energy line;
and a = the Coriolis convective kinetic energy flux correction coefficient due to nonuniform velocity distribution over the cross section (Chow, 1959; Yen, 1973). If there are other
energy losses, they should be added to the right-hand side of the equation. Methods of backwater surface profile computation using these equations are discussed in Chap. 3.
If the flow is steady and uniform, Eqs. (14.1) and (14.4) or Eqs. (14.2) and (14.5)
reduce to S0 = Sf and Q = AV. Hence, for steady uniform flow using Manning's formula,
Q = 0.0496 ^ S0"2 D ^ ~(t)S2^(|))5/3
(14.9)
where <|) is in radians (Fig. 14.3). Correspondingly, the Darcy-Weisbach formula yields
Q = l^ D 5 f l (^n^
Figure 14.4 is a plot of these two equations that can be used to find $.
(141Q)
(14.11)
S /
-^
f
g 8r + ^{
dx ( g + -}=YJ
<14-12>
where P0 = the piezometric pressure of the flow and y = the specfic weight of the fluid.
If the pipe has a constant cross section and is flowing full with an incompressible fluid
throughout its length, then 3V/3jt = O. By further neglecting the spatial variation of P,
integration of Eq. (14.12) over the entire length, L, of the sewer pipe yields
P
exit
-f entrance
V2
=
yi
i av"i
v 5/+ ^
"-*i-^-*'IH
'
(1413)
(Degrees)
< (Radians)
FIGURE 14.4 Central angle cj> of water surface
in circular pipe (from Yen, 1986a).
or
+
T^
A Jr -
p Jr ?
a* =
<14-15)
i a# a# c2 av
----1 + ^-+- -- + ^n 9 = 0
F a/
dx gV dx
(14.16)
i av v av a//
-^- + -^- + - ^ - + &f = 0
S ar
g dx
dx
(14.17)
where p = the bulk density of the fluid, H = PJj = the piezometric head above the reference datum, and c = the celerity of the pressure surge. The fact that Eqs. (14.11) and
(14.12) can be derived from Eqs. (14.2) and (14.5) is the theoretical basis of the
Preissmann hypothetical slot concept, which will be discussed below.
14.2.2.2 Hypothetical slot approach. This approach introduces hypothetically a continuous, narrow, piezometric slot attached to the pipe crown and over the entire length of the
pipe as shown in Fig. 14.6. The idea is to transform the pressurized conduit flow situation
into a conceptual open-channel flow situation by introducing a virtual free surface to the
flow. The idea was suggested by Preissmann (Cunge and Wegner, 1964). The hypothetical open-top slot should be narrow so that it would not introduce appreciable error in the
volume of water. Conversely, the slot cannot be too narrow, with the aim of avoiding the
numerical problem associated with a rapidly moving pressure surge.
A theoretical basis for the determination of the width of the slot is to size the width
such that the wave celerity in the slotted sewer is the same as the surge celerity of the compressible water in the actual elastic pipe. The celerity C1 of the slot pipe is
C1 = ^A/b
(14.18)
where b = the slot width and A = the flow cross-sectional area. Neglecting the area contribution of the slot and hence A = Af = nD2/4 for a circular pipe, and equating C1 to the
pressure wave speed c in the elastic pipe without the hypothetical slot, the theoretical slot
width is
b = ngD2/4c2
(14.19)
The surge speed in a pipe usually ranged from a few hundred feet per second to a few
thousand feet per second. For an elastic pipe with a wall thickness e and Young's modulus of
elasticity EpJ assuming no pressure force from the soil acting on the pipe, the surge speed c is
Hf)('+(^)]
where Ef is the bulk modulus of elasticity and p/ is the bulk density, respectively, of the
flowing water (Wylie and Streeter, 1983). Special conditions of pipe anchoring against
longitudinal expansion or contraction and elasticity relevant to the surge speed c are given
in Table 14.4 where co = Poisson's ratio for the pipe wall material, that is, -co is the ratio
of the lateral unit strain to axial unit strain, and a is a constant to account for the rigidity
with respect to axial expansion of the pipe. For small pipes, Eq. (14.19) may give too
small a slot width, which would cause numerical problems. Cunge et al. (1980) recommend a width of 1 cm or larger.
The transition between part-full pipe flow and slot flow is by no means computationally smooth and easy, and assumptions are necessary (Cunge and Mazadou, 1984). One
approach is to assume a gradual width transition from the pipe to the slot. Sjoberg (1982)
suggested two alternatives for the slot width based on two different values of the wave
speed c in Eq. (14.19). For the alternative applicable to h/D > 0.9999, his suggested slot
width b can be expressed as
L = 10-6 + 0.05423 exp[-(/i/>)24]
(14.21)
He further proposed to compute the flow area A and hydraulic radius R when the depth h
is greater than the pipe diameter D as
TABLE 14.4 Special Conditions of Surge Speed in Full Pipe, Eq. (14.20)
Factor
Pipe Anchor
Condition
T| = ^(l + (D) + a D D+ e
Freedom of pipe
longitudinal
expansion
Only one
Entirely free
(expansion joints at both ends) end anchored
Entire length
anchored
Value of axial
expansion factor a
1 -co 2
Rigid pipe
, = 00
Air entrainment
P7 = pwVw + pfla
c> = Ef/Pf
Ef
1 - 0.5 co
Elasticity E
1 + VJi(EJEJ - 1
Subscript w denotes water (liquid); Subscript a = air; subscript/ = fluid mixture; Y = volume.
No air
P/ = Pw
E
f=
E
W
(14.22)
R = D/4
(14.23)
(14.25)
Because of the lack of reliable data, neither the standard surcharge sewer solution
method nor the Preissmann hypothetical open-slot approach has been verified for a single
pipe or a network of pipes. Past experiences with waterhammer and pressure surge problems in closed conduits may provide some indirect verification of the applicability of the
basic flow equations to unsteady sewer flows. Nevertheless, direct verification is highly
desirable.
Jun and Yen (1985) performed a numerical testing and found there is no clear superiority of one approach over the other. Nevertheless, specific comparison between them is
given in Table 14.5. They suggested that if the sewers in a network are each divided into
many computational reaches and a significant part of the flow duration is under surcharge,
the standard approach saves computer time. Conversely, if transition between open-channel and pressurized conduit flows occurs frequently and the transitional stability problem
is important, the slot model would be preferred.
TABLE 14.5 Comparison Between Standard Surcharge Approach and Slot Approach
Item
Concept
Flow equations
Direct physical
Two different sets, one equation
for surcharge flow, two equations
for open-channel flow
Conceptual
Same set of two equations
(continuity and momentum)
for surcharge and openchannel flows
Divide into AJC' s
Discharge in pipe at
given time
Same
Transition between
open channel flow and
surcharge flow
Specific criteria
Computational effort
74.3
FLOWINASEWER
subcritical
supercritical to surcharge
supercritical
subcritical to surcharge
supercritical to subcritical
surcnarge to supercritical
subcritical to supercritical
surcharge to subcritical
surcharge
easel
(b) case II
(c) case HI
(d) case IV
FIGURE 14.8 Types of sewer entrance flow.
(After Yen, 1986a).
based on whether the flow at a given instant is subcritical, supercritical, or surcharge.
There are four cases of pipe entrance condition, as shown in Fig. 14.8 and below:
Case
I
II
III
IV
Case I is associated with downstream control of the pipe flow. Case II is associated
with upstream control. In Case III, the pipe flow under the air pocket may be subcritical,
supercritical, or transitional. In Case IV, the sewer flow is often controlled by both the
upstream and downstream conditions.
Pipe exit conditions also can be grouped into four cases as shown in Fig. 14.9 and
below:
Case
A
B
C
D
case A
case B
case C
case D
FIGURE 14.9 Types of sewer exit flow. (After
Yen, 1986a).
In Case A, the pipe flow is under exit control. In Case B, the flow is under upstream
control if it is supercritical and downstream control if subcritical. In Case C, the pipe flow
is under upstream control while the junction water surface is under downstream control.
In Case D, the pipe flow is often under downstream control, but it can also be under both
upstream and downstream control.
The possible combinations of the 10 cases of pipe flow with the entrance and exit conditions are shown in Table 14.6 for unsteady nonuniform flow. Some of these 27 possible
combinations are rather rare for unsteady flow and nonexistent for steady flow, for example, Case 6. For steady flow in a single sewer, by considering the different mild-slope M
and steep-slope 5 backwater curves (Chow, 1959) as different cases, there are 27 possible
cases in addition to the uniform flow, of which six types were reported by Bodhaine (1968).
TABLE 14.6 Pipe Flow Conditions
Pipe Flow
Possible
Entrance Conditions
Possible Exit
Conditions
Subcritical
Supercritical
Subcritical > hydraulic drop - supercritical
Supercritical > hydraulic jump > subcritical
Supercritical > hydraulic jump > surcharge
Supercritical surcharge
Subcritical > surcharge
Surcharge -> supercritical
Surcharge > subcritical
Surcharge
I, III
II, III
I, III
II, III
II, III
II, III
I, III
IV
IV
IV
A, B
B, C
B, C
A, B
D
D
D
B, C
A, B
D
Case
1
2
3
4
5
6
7
8
9
10
The nonuniform pipe flows shown in Fig. 14.7 are classified without considering the
different modes of air entrainment. The types of the water surface profile, equivalent to
the M, S, and A (adverse slope) types of backwater curves for steady flow, are also not
taken into account. Additional subcases of the 10 pipe flow cases can also be classified
according to rising, falling, or stationary water surface profiles. For the cases with a
hydraulic jump or drop, subcases can be grouped according to the movement of the jump
or drop, be it moving upstream or downstream or stationary. Furthermore, flow with
adverse sewer slope also exists because of flow reversal.
During runoff, the change in magnitude of the flow in a sewer can range from only a few
times dry weather low flow in a sanitary sewer to as much as manyfold for a heavy rainstorm runoff in a storm sewer. The time variation of storm sewer flow is usually much more
rapid than that of sanitary sewers. Therefore, the approximation of assuming steady flow is
more acceptable for sanitary sewers than for storm and combined sewers.
In the case of a heavy storm runoff entering an initially dry sewer, as the flow enters the
sewer, both the depth and discharge start to increase as illustrated in Fig. 14.10 at times J1,
t2, and t3 for the open-channel phase. As the flow continues to rise, the sewer pipe becomes
completely filled and surcharges as shown at t4 and ^5 in Fig. 14.10. Surcharge flow occurs
when the sewer is underdesigned, when the flood exceeds that of the design return period,
when the sewer is not properly maintained, or when storage and pumping occur.
Under surcharge conditions, the flow-cross-sectional area and depth can no longer
increase because of the sewer pipe boundary. However, as the flood inflow continues
to increase, the discharge in the sewer also increases due to the increasing difference
in head between the upstream and downstream ends of the sewer, as sketched in the
discharge hydrograph in Fig. 14.10. Even under surcharge conditions while the sewer
diameter remains constant, the flow is usually nonuniform. This is due to the effects of
the entrance and exit on the flow inside the sewer, and hence, the streamlines are
not parallel.
As the flood starts to recede, the aforementioned flow process is reversed. The sewer
will return from surcharged pipe flow to open-channel flow, shown at t6 and I1 in Figure
14.1. Since the recession is usuallybut not alwaysmore gradual than the rising of the
flood, the water surface profile in the sewer is usually more gradual during flow recession
than during rising.
The differences in the gradient of the water surface profiles during the rising and recession of the flood bear importance in the self-cleaning and pollutant-transport abilities of
the sewer. During the rising period, with relatively steep gradient, the flow can carry not
only the sediment it brings into the sewer but also erodes the deposit at the sewer bottom
from previous storms. For a given discharge and gradient, the amount of erosion increases with the antecedent duration of wetting and softening of the deposit. During the recession, with a flatter water surface gradient and deceleration of the flow, the sediment being
carried into the sewer by the flow tends to settle onto the sewer bottom.
If the storm is not heavy and the flood is not severe, the rising flow will not reach surcharge state. The flood may rise, for example, to the stage at r3 shown in Fig. 14.10 and
then starts to recede. The sewer remains under open-channel flow throughout the storm
runoff. For such frequent small storms, the flow in the sewer is so small that it is unable
to transport out the sediment it carries into the sewer, resulting in deposition to be cleaned
up by later heavy storms or through artificial means.
For a single-peak flood entering a long circular sewer having a diameter D and pipe
surface roughness k, Yen (1973a) reported that for open-channel flow, the attenuation of
the flood peak, Qpx, at a distance x downstream from the pipe entrance (x = O) and the corresponding occurrence time of this peak, tpx, can be described dimensionlessly as
n
(
( V / H17 (K Vo.42 ( Qn v-16
|-= exp
^
MH=
p -0.0771*- A
Qp0
l
(D][DJ (Dj
(D^ V^)
x
r (f
j?132 ~M
p *-^iM \
<14-26)
v
r
\(o ^
i
}(j V0-11 (R >66
(^-W^={6.03.oglo[{|)-0,8]-52o}(A)
[]
f
x
O4-4 l 0 -^
a0.82~|0.5
(1427)
[slM F ^ - ^ i M '
where Qp0 and tp0 = the peak discharge and its time of occurrence at x = O, respectively;
Qb is the steady base flow rate and Rb = hydraulic radius of the base flow; tg = the time
to the centroid of the inflow hydrograph at x = O above the base flow; g = the gravitational acceleration; and Vw = (QJA1) + (gAJB1)112 = the wave celerity of the base flow,
where Ab = the base flow cross = sectional area and Bb = the corresponding water-surface width. In both equations, the second nondimensional parameter in the right-hand side
klD is a pipe property parameter; the third parameter R1JD is a base flow parameter; the
fourth nondimensional parameter represents the influence of the flood discharge; whereas the fifth and last nondimensional parameter reflects the shape of the inflow hydrograph.
The single - peak hydrograph shown in Fig. 14.10 is an ideal case for the purpose of
illustration. In reality, because the phase shift of the peak flows in upstream sewers and
the time-varying nature of rainfall and inflow, usually the real hydrographs are multipeak.
Piezometric Gradient
Because the flow is nonuniform and unsteady, the depth-discharge relationship, also
known as the rating curve in hydrology, is nonunique. Even if we are willing to consider
the flow to be steady uniform as an approximation, the depth-discharge relation is nonlinear, and within a certain range, nonunique, as shown nondimensionally and ideally in Fig.
14.11 for a circular pipe. The nonunique depth-discharge relationship for nonuniform flow,
aided by the poor quality of the water and restricted access to the sewer, makes it difficult
to measure reliably the time-vary ing flow in sewers. Among the many simple and sophisticated mechanical or electronic measurement devices that have been attempted on sewers
and reported in the literature, the simple, mechanical Venturi-type meter, which has side
constriction instead of bottom constriction to minimize the effect of sediment clogging,
still appears to be the most practical measurement means, that is, if it is properly designed,
constructed, and calibrated and if it is located at a sufficient distance from the entrance and
exit of the sewer. On the other hand, the hydraulic performance graph described in Sec.
14.6.1 can be used to establish the rating curve for a steady nonuniform flow.
Depth-Diameter Ratio,
Area
Discharge
Hidraulic Radius
Velocity
Flow in sewers is perhaps one of the most complicated hydraulic phenomena. Even for
a single sewer, there are a number of transitional flow instability problems. One of them
is the surge instability of the flow in pipes of a network. The other four types of instabilities that could occur in a single sewer pipe are the following: The instability at the
transition between open-channel flow and full conduit flow, the transitional instability
between supercritical flow and subcritical flow in the open-channel phase, the water-surface roll-wave instability of supercritical open-channel flow, and a near dry-bed flow
instability. Further discussion on these instabilities can be found in Yen (1978b, 1986a). It
is important to realize the existence of these instabilities in flow modeling.
14.3.2. Discretization of Space-Time Domain of
a Sewer for Simulation
No analytical solutions are known for the SaintVenant equations or the surcharged
sewer flow equation. Therefore, these equations for sewer flows are solved numerically
with appropriate initial and boundary conditions. The differential terms in the partial differential equations are approximated by finite differences of selected grid points on a
space and time domain, a process often known as discretization. Substitution of the finite
differences into a partial differential equation transforms it into an algebraic equation.
Thus, the original set of differential equations can be transformed into a set of finite difference algebraic equations for numerical solution.
Theoretically, the computational grid of space and time need not be rectangular.
Neither need the space and time differences AJC and Af be kept constant. Nonetheless, it is
usually easier for computer coding to keep AJC and At constant throughout a computation.
For surcharge flow, Eq. (14.14) dictates the application of the equation to the entire length
of the sewer, and the discretization applies only to the time domain. In an open-channel
flow, it is normally advisable to subdivide the length of a sewer into two or three computational reaches of Ax, unless the sewer is unusually long or short. One computational
reach tends to carry significant inaccuracy due to the entrance and exit of the sewer and
is usually incapable of sufficiently reflecting the flow inside the sewer. Conversely, too
many computational reaches would increase the computational complexity and costs
without significant improvement in accuracy.
The selection of the time difference Ar is often an unhappy compromise of three criteria. The first criterion is the physically significant time required for the flow to pass
through the computational reach. Consider a typical range of sewer length between 100
and 1000 ft and divide it into two or three AJC, and a high flow velocity of 5-10 ft/s, a suitable computational time interval would be approximately 0.2-2 min. For a slowly varying unsteady flow, this criterion is not important and larger computational Ar will suffice.
For a rapidly varying unsteady flow, this criterion should be taken into account to ensure
the computation is physically meaningful.
The second criterion is a sufficiently small Ar to ensure numerical stability. An often-used
guide is the Courant criterion
Ac/Ar > V + VgAIB
(14.28)
In sewers, which usually have small AJC compared to rivers and estuaries, this criterion
often requires a Ar less than half a minute and sometimes 1 or 2 s.
The third criterion is the time interval of the available input data. It is rare to have rainfall or corresponding inflow hydrograph data with a time resolution as short as 2, 5, or
even 10 min. Values for Ar smaller than the data time resolution can only be interpolated.
This criterion becomes important if the in-between values cannot be reliably interpolated.
TABLE 14.7 Some Types of Specified Boundary Conditions for Simulation of Exterior
Reaches of Sewers
Location
Subcritical flow
One of
G(O, O
/z(0, f}
V(O, O
One of
/z(L, t)', e.g. ocean tides, lakes
Q(L, t)', release hydrograph
<2(/i); rating curve
V(h); storage- velocity relation
for all t to be simulated
None
Supercritical flow
For a sewer that is divided into M computational reaches and M + 1 stations, there is
a continuity equation and a momentum equation written in finite difference algebraic form
for each reach. There are 2(M + 1) unknowns, namely, the depth and discharge (or velocity) at each station. The 2(M + 1 ) equations required to solve for the unknowns come
from M continuity equations and M momentum equations for the M reaches, plus the two
boundary conditions. If the flow is subcritical, one boundary condition is at the sewer
entrance (x = O) and the other is at the sewer exit (x = L). If the flow is supercritical, both
boundary conditions are at the upstream end, the entrance, one of them often is a critical
depth criterion. If at one instant a hydraulic jump occurs in an interior reach inside the
sewer, two upstream boundary conditions at the sewer entrance and one downstream
boundary condition at the sewer exit should be specified. If a hydraulic drop occurs inside
the sewer, one boundary condition each at the entrance and exit of the sewer is needed;
the drop is described with a critical depth relation as an interior boundary condition.
Handling the moving surface discontinuity, shown schematically in Fig. 14.12, is not a
simple matter. The moving front may travel from reach to reach slowly in different Ar, or
it may move through the entire sewer in one Ar. If, for any reason, it is desired to compute
the velocity of the moving front Vw between two computational stations / and / + 1 in a
sewer, the following equation can be used as an approximation;
Supercriticol to Subcriticol
Subcritical to Supercritical
Supercritical to Subcritical
Subcritical to Supercritical
AV-A V
Vn = ' ' . '^V + '
A- + 1 A
(14-29)
1. Free surface flow exists for the design discharge, that is, the sewer is under "gravity
flow" or open-channel flow. The design discharge used is the peak discharge of the
total inflow hydrograph of the sewer.
2. The sewers are commercially available circular sizes no smaller than, say, 8 in. or 200
mm in diameter. In the United States, the commercial sizes in inches are usually 8, 10,
12, and from 15 to 30 inches with a 3-in. increment, and from 36 to 120 in. with an
increment of 6 in. In SI units, commercial sizes, depending on location, include most
if not all of the following: 150, 175, 200, 250, 300, 400, 500, 600, 750, 1000, 1250,
1500, 1750, 2000, 2500, and 3000 mm.
3. The design diameter is the smallest commercially available pipe that has a flow capacity equal to or greater than the design discharge and satisfies all the appropriate constraints.
4. To prevent or reduce permanent deposition in the sewers, a nominal minimum permissible flow velocity at design discharge or at nearly full-pipe gravity flow is speci-
5.
6.
7.
8.
9.
fied. A minimum full-pipe flow velocity of 2 ft/s or 0.5 m/s at the design discharge is
usually recommended or required.
To prevent the occurrence of scour and other undesirable effects of high velocity =
flow, a maximum permissible flow velocity is also specified. The most commonly
used value is 10 ft/s or 3 m/s. However, recent studies indicate that due to the improved
quality of modern concrete and other sewer pipe materials, the acceptable velocity can
be considerably higher.
Storm sewers must be placed at a depth that will allow sufficient cushioning to prevent
breakage due to ground surface loading and will not be susceptible to frost. Therefore,
minimum cover depths must be specified.
The sewer system is a tree-type network, converging toward downstream.
The sewers are joined at junctions or manholes with specified alignment, for example,
the crowns aligned, the inverts aligned, or the centerlines aligned.
At any junction or manhole, the downstream sewer cannot be smaller than any of the
upstream sewers at that junction, unless the junction has significantly large detention
storage capacity or pumping. There also is evidence that this constraint is unnecessary
for very large sewers.
Various hydrologic and hydraulic methods exist for the determination of the design
discharge Qp. Among them the rational method is perhaps the most widely and simplest
used method for storm sewer design. With this method, each sewer is designed individually and independently, except that the upstream sewer flow time may be used to estimate
the time of concentration. The design peak discharge for a sewer is computed by using the
rational formula
Q11 = I ^ C f J
(14.32)
where i = the intensity of the design rainfall; C = the runoff coefficient (see Chap. 5 for
its values); and a is surface area. The subscript j represents they'th subarea upstream to be
drained. Note that ^a. includes all the subareas upstream of the sewer being designed.
Each sewer has its own design i because each sewer has its own flow time of concentration and design storm. The only information needed from upstream sewers for the design
of a current sewer is the upstream flow time for the determination of the time of concentration.
The rational formula is dimensionally homogenous and is applicable to any consistent
measurement units. The runoff coefficient C is dimensionless. It is a peak discharge coefficient but not a runoff volume fraction coefficient. However, in English units usually the
formula is used with the area aj in acres and rain intensity i in inches per hour. The conversion factor 1.0083 is approximated as unity.
The procedure of the rational method is illustrated in the following in English units for
the design of the sewers of the simple example drainage basin A shown schematically in
Fig. 14.14. The catchment properties are given in Table 14.8. For each catchment, the
length L0 and slope S0 of the longest flow pathor better, the largest LJvS0 should
first be identified. As discussed Sec. 14.7, a number of formulas are available to estimate
the inlet time or time of concentration of the catchment to the inlet. In this example, Eq.
(14.86) is used with K = 0.7 for English units and heavy rain, that is, t0 =
0.7(nL0A/S^)-6. The catchment overland surface texture factor TV is determined from
Table 14.16
The design rainfall intensity is computed from the intensity-duration-frequency relation for this location,
i QQT 0.2
i(in./h) = -^(14.33)
ld T- ZJ
TABLE 14.8 Characteristics of Catchments of Example Drainage Basin A
Catchment
Area
(acres)
I
II
III
IV
V
2
3
3
5
5
0.010
0.0081
0.012
0.010
0.010
0.015
0.016
0.030
0.020
0.021
Inlet Time
t0
(min)
Runoff
Coefficient
C
6.2
9.3
11.7
12.9
13.1
0.8
0.7
0.4
0.6
0.6
Manhole
Ground
(10-YrFlood Level)
Note: All elevations are in feet
(Channel Bed)
FIGURE 14.14 Sewer design example drainage basin A. (a) Layout (b) Profiles.
where td = the rain duration (min) which is assumed equal to the time of concentration,
tc, of the area described, and Tr = the design return period in years. For this example, Tr
= 10 years. Determination of / for the sewers is shown in Table 14.9a. The entries in this
table are explained as follows:
Column L Sewer number identified by the inlet numbers at its two ends.
Column 2. The sewer number immediately upstream, or the number of the catchment
that drains directly through manhole or junction into the sewer being considered.
Column 3. The size of the directly drained catchment.
Column 4. Value of the runoff coefficient for each catchment.
Column 5. Product of C and the corresponding catchment area.
Column 6. Summation of C}a. for all the areas drained by the sewer; it is equal to the
sum of contributing values in Column 5.
Column 7. Values of inlet time to the sewer for the catchments drained, that is, the
overland flow inlet time for directly drained catchments, or the time of
concentration for the immediate upstream connecting sewers.
Column 8. The sewer flow time of the immediate upstream connecting sewer as
given in Column 9 in Table 14.9&.
Column 9. The time of concentration tc for each of the possible critical flow paths,
tc = inlet time (Column 7) + sewer flow time (Column 10) for each
flow path.
Column 10. The design rainfall duration td is assumed equal to the longest of the different times of concentration of different flow paths to arrive at the
entrance of the sewer being considered, for example, for Sewer 31, td is
equal to 13.9 min from Sewer 21, which is longer than that from directly
contributing Catchment V (13.1 min).
Column IL The rainfall intensity i for the duration given in Column 10 is obtained
from the intensity-duration relation for the given location, in this case, Eq.
(14.33) for the 10-year design return period.
Table 14.9/? shows the design of the sewers for which the Manning n = 0.015, minimum soil cover is 4.0 ft, and minimum nominal design velocity is 2.5 ft/s. The exit sewer
of the system (Sewer 31) flows into a creek for which the bottom elevation is 11.90 ft, the
ground elevation of its bank is 21.00 ft, and its 10-year flood water level is 20.00 ft.
Column L
Column 2.
Column 3.
Column 4.
(
o Y8
= [0.0324 ^-j
Runoff
Coefficient
Inlet Time
CA
-L CH
Cj
(4)
<5)
(6)
11-21
I
II
2
3
0.8
0.7
1.6
2.1
12-21
21-31
m
IV
ii
12
3
5
0.4
0.6
1.2
3.0
3.7
1.2
V
21
3.7
1.2
(min)
(7)
Upstream
Time of
Design Rain Design Rain
Sewer Flow Concentration Duration
Intensity
Time
i
(min)
(min)
(min)
(in./h)
(10)
(8)
(U)
(9)
6.2
9.3
6.2
9.3
11.7
12.9
9.3
11.7
1.4
0.9
11.7
12.9
10.7
12.6
1.0
13.1
13.9
7.9
31-41
0.6
3.0
7.9
13.1
12.9
10.9
9.3
11.7
4.62
4.32
12.9
4.18
13.9
4.07
Slope
(1)
Upstream
Manhole
Ground
Elev.
(ft)
(2)
(ft)
(3)
11
12
21
31
(31
35.00
41.50
31.90
28.70
28.70
450
360
400
500
500
Sewer
Sewer
Flow
Time
SL
Upstream
Upstream Downstream Downstream
Crown Elev. Invert Elev. Crown Elev. Invert Elev.
(4)
(ft3/*)
(5)
(ft)
(6)
(ft)
(7)
(ft/*)
(8)
(min)
(9)
(ft)
(10)
(ft)
(U)
(ft)
(12)
(ft)
(13)
(ft)
(14)
0.0081
0.0290
0.0100
0.0144
0.0156
17.1
5.2
33.0
44.4
44.4
1.98
0.99
2.43
2.53
2.50
2.00
1.00
2.50
2.75
2.50
5.4
6.6
6.7
1.4
0.9
1.0
3.65
10.444.00
7.20
7.80
31.00
37.50
27.85
23.85
23.85
29.00
36.50
25.35
21.10
21.35
27.35
27.06
23.85
16.65
16.05
25.35
26.06
21.35
13.90
13.55)
Column 10. Product of Columns 3 and 4; this is the elevation difference between the
two ends of the sewer.
Column 11. The upstream pipe crown elevation of Sewer 11 is computed from the
ground elevation minus the minimum soil cover, 4.0 ft, to save soil excavation cost. In this example, sewers are assumed invert aligned except the
last one (Sewer 31), which is crown aligned at its upstream (23.85 ft for
upstream of Sewer 31 and downstream of Sewer 21) to reduce backwater
influence from the water level at sewer exit.
Column 12. Pipe invert elevation at the upstream end of the sewer, equal to Column
11 minus Column 7.
Column 13. Pipe crown elevation at the downstream end of the sewer, equal to
Column 11 minus Column 10.
Column 14. Pipe invert elevation at the downstream end of the sewer, equal to Column
13 minus Column 7. For the last sewer, the downstream invert elevation
should be higher than the creek bottom elevation, 11.90 ft.
The above example demonstrates that, in the rational method, each sewer is designed
individually and independently, except the computation of sewer flow time for the purpose of rainfall duration determination for the next sewer, that is, the values of tf in
Column 8 of Table 14.9a are taken from those in Column 9 of Table 14.96.
The profile of the example designed sewers are shown as the solid lines in Fig. 14.14.
If the water level of the creek downstream of Sewer 31 is ignored, theoretically a cheaper design could be achieved by putting the exit Sewer 31 on a slightly steeper slope, from
0.0144 to 0.0156 to reduce the pipe diameter from 2.75 to 2.50 ft. The new slope can be
estimated from
(1434)
H^H'*"
This alternative is shown with the parentheses in Table 14.9Z? and as dashed lines in Fig.
14.14b. However, one should be aware that the water level of a 10-year flood in the creek
is 20.00 ft and hence, the last sewer is actually surcharged and its exit is submerged. The
sewer will not achieve the design discharge unless its upstream manhole is surcharged by
almost 4 ft (20.00-16.05). Therefore, the original design of 2.75 ft diameter is a safer and
preferred option in view of the backwater effect from the tailwater level in the creek. In
fact, Sewer 21-31 may also be surcharged due to the downstream backwater effect.
Sometimes, a backwater profile analysis is performed on the sewer network to assess
the degree of surcharge in the sewers and manholes. In such an analysis, energy losses in
the pipes and manholes should be realistically accounted for. However, the intensity-duration-frequency-based design rainfall used in the rational method design is an idealistic,
conceptual, simplistic rain and the probability of its future occurrence is nil. The actual
performance of the sewer system varies with different actual rainstorms, each having different temporal and spatial rain distributions. But it is impossible to know the distributions of these future rainstorms, whereas the ideal rainstorms adopted in the design of the
rational method are used as a consistent measure of protection level. Although designing
sewers using the rational method is a relatively simple and straightforward matter, checking the performance of the sewer system is a far more complex task requiring thorough
understanding of the hydrology and hydraulics of watershed runoff. For instance, checking the network performance by using an unsteady flow simulation model would require
simulation of the unsteady flow in various locations in the network accounting for losses
in sewer pipes as well as in manholes and junctions (the latter will be discussed in the next
section).
Moreover, for a given sewer network layout, by using different sewer slopes, alternative designs of the network sewers can be obtained. A cost analysis should be conducted
to select the most economic feasible design. This can be done with a system optimization
model such as Illinois Least-Cost Sewer System Design Model (ILSD) (Yen et al., 1984).
74.4
HYDRAULICSOFSEWERJUNCTIONS
There are various auxiliary hydraulic structures such as junctions, manholes, weirs,
siphons, pumps, valves, gates, transition structures, outlet controls, and drop shafts in a
sewer network. Information relevant to design of most of these apparatuses are well
described in standard fluid mechanics textbooks and references, particularly in the
German text by Hager (1994) and Federal Highway Administration (FHWA, 1996). In this
section, the most important auxiliary component in modeling the sewer junctions are discussed. For sewers of common size and length, the headloss for the flow through a sewer
is usually two to five times the velocity head. Thus, the head loss through a junction is
comparable to the sewer pipe loss, and is not a minor loss.
14.4.1 Junction Classifications
A sewer junction usually has three or four sewer pipes joined to it. Under normal flow
conditions, one downstream pipe receives the outflow from the junction and other pipes
flow into the junction. However, junctions with only two or more than four joining pipes
are not uncommon. The most upstream junctions of a sewer network are usually one-way
junctions having only one sewer connected to a junction. The horizontal cross section of
the junction can be circular or square or may be another shape. The diameter or horizontal dimension of a junction normally is not smaller than the largest diameter of the joining sewers. To allow the workers room to operate, usually junctions are not smaller than
3 ft (1 m) in diameter. For large sewers, the access to the junction can be smaller than the
diameter of the largest joining sewer.
Sewers may join a junction with different vertical and horizontal alignments, and they
may have different sizes and slopes. Vertically, the pipes may join at the junction with
their centerlines or inverts or crowns aligned, or with any line of alignment in between.
There is no clearly preferred alignment that could simultaneously satisfy the requirements
of good hydraulics at low and high flows without complicating either construction cost or
design. The bottom of the junction is usually at or slightly lower than the lowest invert of
the joining sewers.
In the horizontal alignment, often the outflow sewer is aligned with one (usually the
major) inflow sewer in a straight line with other sewers joining at an angle. For cities with
square blocks, right-angle junctions are most common. Typical sewer benching and flow
guides injunctions are shown in Fig. 14.15.
With the alignment of the joining pipes and the shape and dimensions of junctions not
standardized, the precise, quantitative hydraulic characteristics of the junctions vary considerably. As a result, there are many individual studies of specified junctions, but a general comprehensive quantitative description of junctions is yet to be produced.
For the purpose of hydraulic analysis, junctions can be classified according to the following scheme (Yen, 1986a):
Flat
Bend With Curved Deflector
Full
1. According to the geometry: (a) one-way junction, (b) two-way junction, (c) threeway junctionmerging (two pipes flow into one pipe) or dividing (one pipe flows into
two pipes), and (d) four- or more-way junctionmerging, dividing, or merging and
dividing.
2. According to the flow in the joining pipes: (a) open-channel junction (with openchannel flow in all joining pipes), (b) surcharge junction (with all joining pipes surcharged), and (c) partially surcharged junction (with some, but not all, joining pipes
surcharged).
3. According to the significance of the junction storage on the flow: storage junction or
point junction.
Hydraulically, the most important feature of a junction is that it imposes backwater
effects to the sewers connected to it. A junction provides, in addition to a volumehowever smallof temporal storage, redistribution and dissipation of energy, and mixing and
transfer of momentum of the flow and of the sediments and pollutants it carries. The precise, detailed hydraulic description of the flow in a sewer junction is rather complicated
because of the high degree of mixing, separation, turbulence, and energy losses. However,
correct representation of the junction hydraulics is important in realistic simulation and
reliable computation of the flow in a sewer system (Sevuk and Yen, 1973).
(14.35)
where Q1 = the flow into or out from the junction by the /-th joining sewer, being positive for inflow and negative for outflow; Q. = the direct, temporally variable water inflow
into (positive) or the pumpage or overflow or leakage out from (negative) the junction, if
any; s = the storage in the junction; and t = time. For a two-way junction, the index i =
1, 2; for a three-way junction, / = 1, 2, 3, and so on.
The energy equation in a one-dimensional analysis form is
T/2 p
"N
_i + _i + Z j j + f i ^
fly
T/2
,_ + lG^_i
(14.36)
where Z1., P1, V1 = the pipe invert elevation above the reference datum, piezometric pressure above the pipe invert, and velocity of the flow at the end of the section of the /th
sewer where it meets the junction, respectively; Hj = the net energy input per unit volume
of the direct inflow expressed in water head; K1 = the entrance or exit loss coefficient for
the /th sewer; Y the depth of water in the junction; and g = the gravitational acceleration. The first summation term in Eq. (14.36) is the sum of the energy input and output by
the joining pipes. The second term at the left-hand side of the equation is the net energy
brought in by the direct inflow. The first term to the right of the equal sign is the energy
stored in the junction as its water depth rises. The last term is the energy loss.
The momentum equations for the two horizontal orthogonal directions x and z are
1(2^,) = Jg^JA
(14.37)
2(avfe) = JgydA
(14.38)
and
where px and pz = the x and z components of the pressure acting on the junction boundary, respectively, and A = the solid and water boundary surface of the junction. The direct
flow QJ is assumed entering the junction without horizontal velocity component. The
right-hand side term of Eqs. (14.37 and 14.38) is the x or z component force, where the
integration is over the entire junction boundary surface A. The left-hand side term is the
sum of momentum of the inflow and outflow of the joining pipes. Note that for a threeway merging junction, two of the Q. 's are positive and one Q. is negative, whereas for a
three-way dividing junction, two of the Q. 's are negative.
Joliffe (1982), Kanda and Kitada (1977), Taylor (1944), and others suggested the use
of momentum approach to deal with high velocity situations. To illustrate this approach,
consider the three-way junction shown in Fig. 14.16. The control volume of water at the
junction enclosed by the dashed line is regarded as a point, and there is no volume change
associated with a change of depth within it. One of the two merging sewers is along the
direction of the downstream sewer, whereas the branch sewer makes an angle (p with it.
When one assumes that the pressure distribution is hydrostatic and the flow is steady, the
force-momentum relation can be written as
(14.39)
(14.40)
or
(14.43)
(14.44)
The equation describing the load of sediment or pollutants, expressed in terms of concentration c, can be derived from the principle of conservation as
^J cds + I Qf1 + Qfj = G
S
(14.45)
(14 46)
Type of Junction
Shape of Channels
Channel
Slope
Circular,
D = 3.0 in.3.75 in.
4.75 in. or 5.72 in.
Horizontal
Flushed bottom
Lindvall
(1984, 1987)
Circular,
Dmam = 144 mm,
A>/Anain = 1A
0.686, or 0.389
Horizontal
Round box
Enclosed pipe
junction
Circular,
Horizontal
Anain d = 70 ^Hl,
/U up/Anain,= 0.64,
>b/Dmain, = 0.91
Circular,
DJD^ = 0.25 ~ 1.0
Type of
Flow
Remarks
Steady
Center aligned
Steady
Flushed bottom
Steady
Center or top
aligned
Circular,
(Horizontal) Center aligned
^main = 4441^11'
/V/U, = 0.14, 0.23,
0.34, or 0.46
Enclosed
Rectangular,
(Horizontal) Same height
rectangular
4. 14 mm high,
conduit junction main width 91.5 mm,
branch width
20.4 mm,70.5 mm,
or 91.5 mm
Steady
Reynolds number
effect insignificant
Sangsteretal. (1958)
Lindvall (1987)
Lindvall (1984)
Mil er (1990)
Fri(1989)
ed & ldelchick
Marsalek(1985)M1
Marsalek(1985)M3
deGroot&Boyd(1983)
Sangsteretal. (1958)
Lindvall (1984) Type 1
Mil er (1990)
Fri
ed & ldelchik
(1989)
FIGURE 14.17 Experimental headless coefficients for surcharged 3-way sewer junction with identical
pipe sizes and 90 merging lateral, (a) Mainline loss coefficient K21; (b) Branch loss coefficient K31.
Miller, and Serre et al. However, Fig. 14.18 depicts considerable disagreement among the different sources, indicating the need for more reliable investigations.
The joining angle of the lateral branch is a significant factor affecting the loss coefficients, particular on K^ P The values of the loss coefficients decrease if the joining angle
more or less aligns with the flow direction of the main, and increase if the lateral flow is
directed against the main. The references of Fried and ldelchik (1989) and Miller (1990)
provide some idea on the adjustment needed for the K values due to the joining angle.
Mil er (1990)
Serre etal( 1994)
Sangsteretal (1961)
Mil er
lFridelecdhi&k
Serre
etal
Sangsteretal (1961)
FIGURE 14.18 Headless coefficients for surcharged 3-way junction with 90 merging lateral of different sizes, (a) Mainline loss coefficient K21. (b) Branch loss coefficient K31.
FIGURE 14.19 Headless coefficients for 3-way open-channel sewer junction with identical pipe
sizes and 90 merging lateral, (a) Mainline loss coefficient K21 (B) Branch loss coefficient ^31 (After
Yen, 1987).
Listed in Table 14.11 is a summary of experiments on steady flow in three-way merging open-channel junctions. Most of the studies were done with point-type junctions. The
experimental subcritical flow results of storage-type junctions by Marsalek (1985) and
Townsend and Prins (1978) are plotted in Fig. 14.19 for lateral joining 90 to the same size
mainline pipes. Yevjevich and Barnes (1970) gave the combined main and lateral loss
coefficient but not the separate coefficients, making the result difficult to be used in routing simulation. The points in the figure scatter considerably, but they are generally in the
same range of the loss coefficient values for surcharged three-way 90 merging junction
except K3, for Townsend and Prins' data. It is interesting to note that the most frequently
encountered sewer junctions are three- and four-way box junctions with unsteady subcritical flow in the joining circular sewers. None of the open-channel experiments was
conducted under these conditions. All were tested with steady flow. It is obvious that
existing experimental evidence and theory do not yield reliable quantitative information
on the loss coefficients of three-way sewer junctions. Before more reliable information is
obtained, provincially for design and simulation of three joining identical size sewers, for
K2 j a curve drawn between that of Lindvall and that of Sangster et al. can be used as an
approximation. For ^3}, the curve of Lindvall can be used. For joining pipes of unequal
sizes, the curves of Sangster et al. appear to be tentatively acceptable.
14.4.4 Loss Coefficient for Two-Way Sewer Junctions
Two-way junctions are used for change of pipe slope, pipe alignment, or pipe size.
Experimental studies on two-way, surcharged, top-open sewer junctions are listed in Table
TABLE 14.11 Experimental Studies on Three- Way Junction of Merging Open Channels
Pipe Alignment at Junction
Channel
Slope
Vertical
Longitudinal
Point
Rectangular,
identical width,
B = 4 in.
Horizontal
Flushed
bottom
Trapezoidal,
identical width,
B = 7.2 in.
Behlke and
Pritchett
(1966)
Rectangular or
trapezoidal
(side slope 1:1)
Webber and
Point
Greated (1966)
Rectangular,
B = 5 in.
Horizontal
Flushed
bottom
Yevjevich and
Barnes (1970)
Circular,
Anain = 6'25 inDbT= 1.87 in.
0.00008
0.00054
0.00107
Flushed
bottom
or crown
aligned
Taylor (1944)
Type of
Junction
Type of Flow
Upstream
Downstream Remarks
Pipes
Pipe
Shape of
Channels
References
Point
Square
box
Subcritical
TABLE 14.11
(Continued)
Kanda and
Kitada (1977)
Point
Rectangular,
Horizontal
* main =100,200,
400mm
Bbr= 100mm;
Circular, different
sizes
Circular,
Horizontal
different sizes
Flushed
bottom
Flushed
bottom
Straight through
and one merging
channel
Less than 0.01 Invert drop Straight through
and one merging
across
junction box channel at 45
or 90
Rectangular,
Lin and Soong Point
Invert drop Straight through
Horizontal
B = 457 mm
(15 or
and one 90
(1979)
merging channel
18mm)
Circular,
Joliffe (1982) Point
Flushed
Straight through
Horizontal,
equal diameter,
and one 90
0.0001, 0.0075, bottom
D = 69mm
merging channel
0.005, or 0.01
Rectangular,
Point
Channel slope Not available Straight through
Best and
and one merging
Reid (1984)
identical width,
adjustable
B = 0.5 ft
channel at 15,
to achieve
45, 70, or 90
equilibrium
water depth
Marsalek
Square box Circular,
Flushed
Horizontal
(1985)
or round box identical diameter
bottom
Subcritical
Subcritical
Subcritical
Sub- or
supercritical
Subcritical
Subcritical
Subcritical
TABLE 14.12 Experimental Studies on Straight-Through Two- Way Open-Top Junction of Surcharged Channels
References
Type of
Junction
Shape of Channels
Channel Slope
Type of
Flow
Remarks
Sangster et al.
(1958, 1961)
Square,
rectangular,
or round box
Rectangular
box
Circular,
D = 3.0, 3.75,
4.75, or 5.72 in.
Circular, identical
diameter, D = 6 in.
Horizontal
Flushed
bottom
Steady
0.0094-0.0192
Flushed
bottom
Steady
Rectangular
box or round
box
Rectangular
box or round
box
Round box
Circular, identical
diameter,
D = 102 mm
Circular, identical
diameter,
D = 88 mm
Circular, identical
diameter,
D = 144 mm
Circular, identical
diameter,
D = 6 in
Circular, identical
diameter,
D = 88 mm
Circular, identical
diameter,
D = 90 mm
Flushed
bottom
Horizontal
Flushed
bottom
Horizontal
Center
aligned
Straight through
Steady
Horizontal
Flushed
bottom
Straight through
Steady
Horizontal
Center
aligned
Straight through
Steady
Horizontal
Center
aligned
Straight through
Steady
Ackers (1959)
Archer et al.
(1978)
Howarth and
Saul (1984)
Lindvall
(1984)
Marsalek
(1984)
Square box
or round box
Johnston and
Volker (1990)
Square box
Bo Pedersen
and Mark
(1990)
Round box
Straight
through
or 90 bend
Straight through
or 45 bend in
junction or 52
bend downstream
of junction
Straight through,
or 30 or 60
bend in junction
Straight through
Steady
Steady or
unsteady
14.12. All the experiments were conducted with the same size upstream and downstream
pipes joining the junction. Only Sangster et al. (1958, 1961) tested also the effect of different joining pipe sizes. These experimental results show that for a straight-through, twoway junction, the value of the loss coefficient is usually no higher than 0.2. Alignment of
the joining pipes and benching in the junction are also important factors to determine the
value of the loss coefficient.
Figure 14.2Oa shows the headloss coefficiet of a surcharged two-way open-top junction connecting two pipes of identical diameters aligned centrally given by the experiments of Archer et al. (1978), Howarth and Saul (1984), Johnston and Volker (1990) and
Lindvall (1984). Noticeable is the swirl and instability phenomena when the junction submergence (junction depth to pipe diameter ratio) is close to two and the corresponding
high head loss coefficient. The ranges of loss coefficient given by Ackers (1959),
Marsalek (1984), and Sangster et al. (1958) are also indicated in Fig. 14.2Oa, but the data
on the variation with the pipe-to-junction-size ratio was not given by these investigators.
Sangster et al. (1958) also tested the effect of different sizes of joining pipes for surcharged two-way junction. Some of their results are plotted in Fig. 14.2Ob. They did not
indicate a clear influence of the effect of the size of the junction box. However, Bo
Pedersen and Mark (1990) demonstrated that the loss coefficient of a two-way junction
can be estimated as a combination of the exit headloss due to a submerged discharging jet
and the entrance loss of flow contraction. They suggested that the loss coefficient K
depends mainly on the size ratio between the junction and the joining pipes of identical
size. For an infinitely large storage junction, the theoretical limit of K is 1.5. For the junction-diameter to pipe-diameter ratio, DJD less than 4, they proposed to estimate the K
values according to benching as shown in Fig. 14.21.
14.4.5 Storage Junctions
For a storage- (or reservoir-) type junction, the storage capacity of the junction is relatively large in comparison to the flow volume and hydraulically it behaves like a reservoir. A water surface, and hence, the depth in the junction can be defined without great
difficulty. A significant portion of the energy carried in by the flows from upstream sewers is dissipated in the junction. If the horizontal cross-sectional area of the junction A7
remains constant, independent of the junction depth 7, the storage is s = AjY. Hence, dsldt
= Aj(dY/dt) = Aj(dH/dt), where H = Y + Z = the water surface elevation above the reference datum, and Z = the elevation of the junction bottom. Therefore, from Eq. (14.35),
^Q,+ Qj = Aj ^-
(14.47)
Either the energy equation (Eq. 14.36) or the momentum equations [Eqs. (14.37) and
(14.38)] can be used as the dynamic equation of the junction. If the energy loss coefficient
K1 in Eq. (14.36) can be determined, use of the energy equation is appropriate. On the
other hand, if the pressure on the junction boundary can be determined, the momentum
equation is also applicable. If the junction were truly a large reservoir, both the loss coefficients and the pressure could reasonably be estimated on the basis of information on
steady flow entering or leaving a reservoir.
Customarily for the convenience of computation, instead of Eq. (14.36), the junction
energy relationship is divided for each joining sewer by relating the total head of the sewer
flow to the total head in the junction. Assuming that the energy contribution from the
direct lateral inflow Qj is negligible, the component of Eq. (14.36) for each joining sewer
/ can be written as
Lindvall
Howarth &
Saul
Archer
etal.
Johnston &
Volker
Sangster et al.
(1958, Eq.7)
FIGURE 14.20 Headless coefficient for surcharged 2way open-top straight-through sewer junction,
(a) Same size sewers upstream and downstream, (b)
Different joining pipe sizes. (After Yen, 1987).
Shape
FIGURE 14.21 Effect of benching on loss coefficient of surcharged two-way sewer junction
according to Bo Pedersen and Mark (1990).
H=(I- K1)(VfVg) + (P1Ii) + Z1
(14.48)
For open-channel flow in the joining pipes, the piezometric term P/y is
P/y=ht
(14.49)
where ht is the open-channel flow depth of the ith pipe at the junction. It should be cautioned that Eq. (14.48) is applicable only when there is no free surface discontinuity
between the junction and the sewer. In other words, they are applicable to Cases B and D
in Fig. 14.9 and all four cases in Fig. 14.8. The flow equations for these pipe exit and
entrance cases are given in Table 14.13.
14.4.6 Point Junctions
A point-type junction is the one whose storage capacity is negligible, and the junction is
treated as a single confluence point. Hence, Eq. (14.35) is reduced to
Za + QJ. = O
(14.50)
For subcritical flow in the sewers emptying into the point junction, the flow can discharge freely into and without the influence of the junction only when a free fall exists
over a nonsubmerged drop at the end of the pipe (Case A in Fig. 14.9). Otherwise, the subcritical flow in the sewer is subject to backwater effect from the junction. Since the junction is treated as a point, the dynamic condition of the junction is usually represented by
a kinematic compatibility condition of common water surface at the junctions for all the
joining pipes without a free fall (Harris, 1968; Larson et al., 1971; Roesner et al., 1984;
Sevuk and Yen, 1973; and Yen, 1986a). Thus, by neglecting the junction storage and for
subcritical sewer flow into the junction,
ht = hic
(H51)
Condition
Equation
Remarks
For reservoir
junction K = I
H>Z + D
h = hc
Subcritical flow
in sewer
Supercritical
flow in sewer
H<Z+hc
F< 1
Z + D>H
H>Z + hc
F< 1
Z + D>H
Z + hc>H
F> 1
H>Z+hc
but F > 1
hc> h
Supercritical flow in
sewer, hydraulic jump
in junction
For outflow from junction into sewer
I
Subcritical flow
in sewer
II
Supercritical flow
in sewer
Submerged sewer
entrance, openchannel flow in sewer
Surcharged sewer
III
IV
h>hc
H<Z + D + S(V2 2Kg)
H>Z+hc + (Vc /2g)
H < Z + D + S(Vc2/2)
(h = hc)
H> Z + D + S(V2Kg)
h<D
H> Z + D + 8(V2/2g)
h> D
H=(I- K)(V1IIg) + h + Z
For reservoir
junction K = I
H = h +Z
Upstream control
for sewer
Upstream control for
sewer; H determined
by other sewers
H = (1 - K)(V2SIg) + h + Z
5-1
H = ( I - K)(V2Kg) + hc + Z
8-1
Q = AVVV
Vv = Cv V2g(H - Z)
8-1
8-1
ht + Z. = h0 + Z0
otherwise
where Z0 and h0 = the invert elevation and depth of the flow at the entrance of the downstream sewer taking the outflow from the junction, respectively.
For a supercritical flow in a sewer flowing into a point junction, Case C in Fig. 14.9
would not occur. Only a subcase of Case B in Fig. 14.9 with ht < hic exists where Eq.
(14.51) applies. For Case D of Fig. 14.9 with submerged exit,
Zi + WY) = (PJD + Z0> ZI + D1
(14.52)
The flow in the downstream sewer, which takes water out from the junction, may be
subcritical, supercritical, or submerged, depending on the geometry and flow conditions. The flow equations are the same as the storage junction outflow Cases I-IV given
in Table 14.13.
74.5.
Hydraulically, sewers in a network interact, and the mutual flow interaction must be
accounted for to achieve realistic results. In designing the sewers in a network, the constraints and assumptions on sizing sewers as discussed in Sec. 14.3.4 should be noted.
The rational method is the most commonly and traditionally used method for the
design of sewer sizes. As described in Sec. 14.3.4, each sewer is designed independently
without direct, explicit consideration of the flow in other sewers. This can be done
because to design a sewer, only the peak discharge, not the entire hydrograph, of the
design-storm runoff is required. As previously explained in Sec. 14.3.4, each sewer has its
own design storm. The information needed from upstream sewers is only for the alignment and bury depth of the sewer and the flow time to estimate the time of concentration
for the determination of the rainfall intensity / for the sewer to be designed.
Contrarily, in simulation of flow in an existing or predetermined sewer network for
urban stormwater control and management, often the hydrograph, not merely the peak discharge, is needed, and a higher level of hydraulic analysis that considers the interaction of
the sewers in the network is required. This network system analysis involves combining
the hydraulics of individual sewers as described in Sec. 14.3, together with the hydraulics
of junctions described in Sec. 14.4.
A sewer network can be considered as a number of nodes joined together by a number
of links. The nodes are the manholes, junctions, and network outlets. The links are the
sewer pipes. Depending on their locations in the network, the nodes and links can be classified as exterior or interior. The exterior links are the most upstream sewers or the last
sewer having the network exit at its downstream end. An exterior sewer has only one end
connected to other sewers. Interior links are the sewers inside the network that have both
ends connected to other sewers. Exterior nodes are the junctions or manholes connected
to the upstream end of the most upstream sewers, or the exit node of the network. An exterior node has only one link connected to it. Interior nodes (junctions) inside the network
have more than one link connected to each node.
A systematic numeric representation of the nodes or links is important for computer
simulation of a network. One approach is to number the links (sewers) according to the
branches and the order of the pipes in the branch, similar to Horton's numbering of river
systems. Another approach is to identify the links by the node numbers at the two ends of
the link. Using the node number identification system, a numbering order technique similar to topographic contour lines called the isonodal line method, can be used. This method
Next Page
CHAPTER 15
HYDRAULIC
DESIGN OF
CULVERTS AND
HIGHWAY
STRUCTURES
I. Kaan Tuncok
Stanley Consultants, Inc.
Phoenix, Arizona
Larry W. Mays
Department of Civil and Environmental Engineering
Arizona State University
Tempe, Arizona
15.1
INTRODUCTION
The aim in highway drainage is to prevent on-site water standing on the surface and convey the off-site storm runoff from one side of the roadway to the other. To accomplish the
off-site drainage either a culvert or a bridge can be used. Culverts are closed conduits in
which the top of the structure does not form part of the roadway. Bridges are mainly provided for large streams and the road is practically a part of the span or drainage structure.
National Bridge Inspection Standards (NBIS) define bridges as those structures that have
at least 20 ft of length along the roadway centerline. The main operational differences
between culverts and bridges may be described in terms of:
economics,
hydraulics,
structural aspects,
maintenance attention requirements.
Economically, the initial and operating costs of culverts are considerably less than that
of bridges. Thus, the total investment of public funds for culverts constitutes a substantial
share of the highway budget in relation to the investment for bridges.
The hydraulic properties for both the culverts and bridges can be computed by using
the conservation of mass, energy, and momentum principles defined in Chap. 3. In the
case of bridges the magnitude of energy losses must be carefully computed. One part of
this loss is due to the contraction and expansion that occur in reaches immediately
upstream and downstream from the bridge. The other part is due to the losses at the structure itself and is calculated with either the momentum or energy principles (U.S. Army
Corps of Engineers, 1990). To compute the structural losses, the flow must be identified
as low flow, pressure flow, or weir flow and equations based on the momentum principle,
orifice flow, and a standard weir equation can be used, respectively, to separate the associated losses. The hydraulics of culverts is complicated and the concept of inlet control
and outlet control is commonly used to simplify the analysis. Inlet and outlet control culverts have a barrel capacity larger and smaller than the capacity of the culvert entrance,
respectively. This chapter will focus on the hydraulic design of culverts and stream stability at highway structures.
Structurally, culverts are designed with a heavy dependence on good and proper backfill around the perimeter of the culvert conduit. Often culverts must accommodate significant dead loads (i.e., embankment loads) in addition to the live loads (i.e., vehicles and
pedestrians). The earth load transmitted to a culvert is largely dependent on the type of
installation, and the four common types are:
trench, used in the construction of sewers, drains and water mains,
positive projecting embankment, used in relatively flat streambed or drainage path,
negative projecting embankment, used in relatively narrow and deep streambed or
drainage path,
induced trench, used in the construction of culverts placed under high embankment.
For the details of these different installation types the reader should refer to the
American Concrete Pipe Association (1992). Bridges barely have significant embankment
loads and therefore the live load on bridges is the predominant structural consideration.
Maintenance requirements for culverts are considerable. Constant efforts must be
made to ensure:
Bridges require some of the same items of maintenance attention, but not to the extent
required by culverts.
75.2
DESIGNPARAMETERS
To better understand the design of culverts, basic design parameters must be carefully studied. The headwater, outlet velocity, and tailwater are factors of significant importance.
15.2.1 Headwater and Tailwater
Headwater is the vertical distance from the culvert invert at the entrance, to the energy
line (depth + velocity head) of the headwater. Because of the low velocities in most
entrance pools and the difficulty in determining the velocity head for all flows, the water
surface and the energy line at the entrance are assumed to be coincident (American Iron
and Steel Institute, 1983). Therefore, headwater is the depth of water at the upstream face
of the culvert. The main reason for the accumulation of water is to build up the energy
required to pass the water through the culvert opening and to overcome the friction,
Control Section
FIGURE 15.1 Typical inlet control flow condition. dc = critical depth; HW = headwater; TW = tailwater; WS = water surface. (From Normann et al., 1985)
entrance, and exit losses along the culvert. By this process, the potential energy accumulated in water is transformed into kinetic energy through the culvert. Figure 15.1 shows a
typical inlet control flow condition with the headwater depicted on it.
The designer should not ignore the design limitations on the maximum or allowable
headwater (AHW). AHW is the level to which the culvert headwater may rise before causing an unwanted inundation or damage under the circumstances of the design flood
(Reagan, 1993). The level of AHW may be based on the fill heights, elevation level of
property, or features which would be affected by ponded water due to culvert headwater
levels. In cases with a heavy load of debris in the runoff, provision for its passage (or
retention) must be made. The designer should avoid the use of very high allowable headwater that may result in unacceptable turbulence and objectionable velocities of flow into
and through the culvert. The criteria in establishing the AHW can be summarized as:
(1) AHW should not give damage to upstream properties,
(2) AHW should be below the traffic lanes of interest or lower than the shoulder,
(3) AHW should be lower than the low point in the road grade,
(4) AHW should not be equal to the elevation where flow diverts around the culvert.
The ponding effect at the upstream face of the culvert may cause an attenuation in the
peak discharges, therefore resulting in smaller culvert size requirements for the design.
The tailwater is the depth of water downstream of the culvert measured from the outlet invert. It can be a very significant factor in culvert hydraulic operation. The depth of
the tailwater can affect the outlet velocity, and depending on the character of the culvert
flow it can also affect the operating headwater on the culvert.
15.2.2 Outlet Velocity
For design purposes, the outlet velocity should be similar in magnitude to the velocity in
the channel to provide appropriate protection at the downstream end (American Concrete
Pipe Association, 1992). But generally due to the constriction inside the culvert opening,
the outlet velocity is higher than the channel velocity. High channel velocities can be mitigated in several different ways such as the following.
(1) Stabilization of the channel.
(2) Construction of energy dissipation structures, such as hydraulic jump basin, drop
structure, stilling basin, riprap, and sill. For a detailed description of the individual
methods, the designer should refer to Corry et., al. (1983).
(3) Changing the size or the roughness of the culvert.
A more detailed discussion of outlet velocity will be given in Sec. 15.3.3. Another
important factor is the minimum velocity that must be maintained within the culvert for an
efficient operation. The ideal minimum velocity should be adjusted such that the sediment
particles transported through the culvert should not be allowed to settle.
The culvert barrel should result in a tractive force, T, greater than the critical T of the
transported stream bed material at low flow rates. In case of unknown stream bed material size, a culvert velocity of 2.5 ft/s should be maintained within the culvert. If clogging is
probable, installation of a sediment trap or sizing the culvert to facilitate the cleaning
should be considered.
75.3
CHARACTERISTICSOFFLOW
A culvert is a type of structure that can transmit water as full or partly full. Full flow is not
common for culverts unless governed by a high downstream or upstream water surface elevation. Full flow can be described by the fundamentals of pipe flow. Partly full flow culverts
follow the rules of open-channel flow and need to be classified as either subcritical or supercritical flow to accomplish the design procedure. The reader should refer to Chap: 3 for the
details of open-channel flow and the related terminology that will be used in this chapter.
An exact theoretical analysis of culvert flow is difficult due to the following factors:
(l)the culvert may have both gradually and rapidly varied flow zones and non-uniform
flow conditions must be considered to analyze the flow,
(2) the location of hydraulic jump, if it occurs, must be identified,
(3) the results of hydraulic model studies must be applied,
(4) the change in the flow type as a function of discharge and hydraulic characteristics of
the culvert must be studied.
In the culvert design procedures of this chapter, the concept of control section, as related to the relationship between the flow rate and the upstream water surface elevation will
be used. A culvert can operate either under inlet or outlet control depending on whether
the barrel or the inlet has a greater hydraulic capacity. Under inlet control, the cross sectional area of the culvert barrel, the inlet geometry, and the amount of headwater at the
entrance are of primary importance. Outlet control involves the additional consideration
of tailwater elevation in the outlet channel, the slope, the roughness and the length of the
culvert barrel.
When a culvert operates under inlet control, the barrel will flow partly full and depending on the headwater, the flow may be divided either by using the weir or orifice equations. For outlet control, the culvert barrel is intended to flow full for design conditions.
A more detailed description of both the inlet and outlet control culverts will be described
in the following sections.
15.3.1 Inlet Control
A culvert operates under inlet control when the barrel hydraulic capacity is higher than
that of the inlet. A typical flow condition is critical depth near the inlet and supercritical
flow in the culvert barrel. Depending on the tailwater, a hydraulic jump may occur downstream of the inlet. The barrel geometry and roughness have no direct influence on the
hydraulic characteristics of the culvert.
Due to the severe constriction of the flow at the culvert entrance, the inlet configuration have a significant effect on the hydraulic performance. To increase the capacity of the
culvert the designer can use beveled inlet edges, side-tapered inlet, or slope-tapered inlet.
The details of different inlet conditions can be found in Sec. 15.7.4.
Typically both the inlet and outlet control culverts can be studied in four different types
and can be summarized as:
(1)
(2)
(3)
(4)
Culverts operating under steep slope regimes usually operate with inlet control of the
headwater, and as a result supercritical flow is a very common flow condition. In cases
where the specific energy within the culvert barrel is in the vicinity of minimum specific
energy, the flow depth is sensitive to the changes (i.e., roughness) in the culvert. In application, if the normal depth and critical depth are within 5 percent of each other, hydraulic
computations should be performed separately to identify the worst case for the design. In
the case of inlet control, these four cases are depicted in Fig. 15.2. The flow will be governed by weir flow when the entrance is unsubmerged, and by orifice flow in the case of
a submerged entrance. Laboratory investigations showed that for an entrance, either inlet
or outlet control, to be considered as unsubmerged, the headwater must be less than a critical value, //cr,and can be approximated as (1.2-1.5) times the height of the barrel.
Another important factor is the identification of the culvert as hydraulically short or long
(Chow, 1959), which depends on the length, slope, size, entrance geometry, headwater,
entrance, and outlet conditions.
The most typical case of inlet submerged and outlet unsubmerged condition is shown
in Fig. 15.2c. As soon as the flow enters the culvert, critical depth is observed and the freeflow conditions at the downstream end results in supercritical normal depth. Supercritical
flow is maintained along the culvert barrel. A very similar condition with both inlet and
outlet unsubmerged is shown in Fig. 15.2a. The difference from the previous condition is
the low headwater at the culvert entrance, but again the flow passes through critical depth
at the culvert entrance. In the case of inlet unsubmerged and outlet submerged (Fig.
15.2), even though the high downstream water surface depth cannot enforce outlet control conditions, a hydraulic jump is observed within the culvert. Similar to the previous
cases the flow passes through critical depth at the culvert entrance, but the partly full flow
is maintained only upstream of the hydraulic jump (ARMCO, 1950). An uncommon condition shown in Fig. 15.2J depicts submerged inlet and outlet with partly full flow. Similar
to the previous case, a hydraulic jump will be observed within the culvert.
15.3.2 Outlet Control
The culvert will operate under outlet control conditions if the culvert barrel has a smaller
hydraulic capacity than the inlet does. The flow regime is always subcritical; as a result,
the control of the flow is either at the downstream end of the culvert or further downstream of the culvert outlet, depending on the depth of the tailwater. (Portland Cement
Association, 1964). Typical flow conditions include a full or partially full culvert barrel
for all or part of its length.
Similar to the inlet control culverts, four types of control are available for the outlet
control culverts (Fig. 15.3). In the first flow condition in which both the inlet and the out-
.WATER SURFACE
WATER SURFACE
WATER SURFACE
MEDIAN DRAIN
WATER SURFACE
FIGURE 15.2 Types of inlet control: (A) outlet unsubmerged, (B) outlet submerged, inlet unsubmerged, (C) inlet submerged (D) outlet submerged. (From
Normann et al, 1985)
let are unsubmerged, the flow is subcritical and the culvert barrel is partly full over its
entire length. The unsubmerged flow condition results in critical depth at the culvert outlet. For the second flow condition, in which the inlet is unsubmerged and outlet is submerged, headwater is an important factor. Typically, the water surface will drop at the culvert entrance and experience a contraction through the culvert opening. This is depicted
in Fig. 15.3B and is a result of shallow headwater. The unusual case shown in Fig. 15.3C
can be established with high submergence at the culvert inlet. Although there is no tailwater, the submergence at the inlet maintains the pressure flow along the culvert barrel.
The condition where the inlet is submerged and outlet is unsubmerged (Fig. 15.3D), is a
result of tailwater that does not submerge the culvert depth but is greater than the critical
depth at the outlet. Finally, the common condition where both the inlet and the outlet are
submerged (Fig. 15.3A) results in pressure flow through the culvert opening.
In all of these flow types, in addition to the factors that affect the inlet control culverts, the
barrel characteristics such as roughness, area, length, and slope also play an important role.
Hydraulics To evaluate the outlet control hydraulics the condition of full flow in the
culvert barrel will be used. The energy equation, Eq. (15.1), must incorporate the losses
(15.1)
(15.2)
V =-J
(15.3)
where Qb = discharge per barrel (ft3/s), A= cross-sectional area of flow with the barrel
full (ft2), V0 = approach velocity of flow (ft/s), and g = acceleration due to gravity (32.2
ft/s2).
For design purposes, the approach velocity is usually ignored and the velocity head
can be expressed as
V2
hv = ^
(!5-4)
<15'5)
V2
* = 2
<15'6)
Coefficient ke
Pipe, concrete
Mitered to conform to fill slope
End-section conforming to fill slope*
Projecting from fill, sq. cut end
Headwall or headwall and wingwalls
Square-edge
Rounded (radius = 1/12D)
Socket end of pipe (groove-end)
Projecting from fill, socket end (groove-end)
Beveled edges, 33.7 or 45 bevels
Side- or slope-tapered inlet
0.5
0.2
0.2
0.2
0.2
0.2
0.9
0.7
0.5
0.7
0.5
0.5
Coefficient ke
0.5
0.2
0.2
0.7
0.5
0.5
0.2
0.2
0.2
-\
*f
&
[1.486 A R061I
in American customary units where, n = Manning's roughness coefficient, A = cross-sectional area (ft2), R = hydraulic radius (ft), A and R are based on full-flow conditions.
When the expression for friction slope is inserted into Eq. (15.7), the expression for
friction loss can be written as
HT^)S
The bend losses are also a function of velocity head and can be computed as
(V2}
* = K*]fr)
<15-9>
Angle
90
0.50
0.30
0.25
0.15
0.15
1
2
4
6
8
of
45
0.37
0.22
0.19
0.11
0.11
Bend,0
22.5
0.25
0.15
0.12
0.08
0.08
(15.10)
where /i; = junction loss in the main culvert (ft) hvu = velocity head in the upstream culvert (ft), hvd = velocity head in the downstream culvert (ft),and y = change in hydraulic
grade line through the junction (ft).
The momentum principle can be used to compute y as
TRIBUTARY STREAM
MAIN
STREAM
PIPE 2
PIPE I
y=
Q2V2-Q1V1-Q^CQsQ
0.5(A^A2)S
(15.11)
(15.12)
By using the schematic in Fig. 15.5 that depicts the hydraulic and energy grade lines,
the energy equation with the entrance, exit, and friction loss terms for a full flow, can be
written between the upstream and downstream ends of a culvert system as
HW0 + ^ = TW + ^ + H
(15.13)
where HW0 = headwater depth above the outlet invert, TW = tailwater depth above the
outlet invert, and the subscripts u and d denote the upstream and downstream conditions,
respectively.
Because of the smaller values of upstream and downstream velocities, compared to the
culvert velocity, they can be neglected. Then the energy equation can be expressed as
HW= TW + H-S0L
(15.14)
where
S0L= drop in elevation from the inlet invert to the outlet invert
A practical way to compute the headloss H is to use the outlet control nomographs
(Figs. 15.6 and 15.7). Outlet control nomographs presented for full flow can also be used
SECTION
FIGURE 15.5 Full-flow energy and hydraulic grade lines. (From Normann et al.,
1985)
SECTION
for the partly full flow for the computation of H. Details on the use of outlet control nomographs will be given in Sec. 15.4.2.2.
Equations (15.1-15.14) were developed for full barrel flow in which TW is greater
than or equal to the culvert diameter, D. However in the case of partly full flow, backwater calculations may be required starting at the downstream end of the culvert and proceeding upstream. During the backwater computations if the hydraulic grade line cuts the
top of the barrel, full flow will be observed upstream of this intersection point. FHWA has
also developed an approximate method to overcome the tedious backwater computations
for partly full flow. They found out that the hydraulic grade line at the culvert outlet is at
a point between the critical depth and the culvert diameter, and can be computed as (dc +
D)/2. It was oncluded that the observed TW should only be used if it is higher in magnitude than (dc + D)/2. In such a case the following equation should be used:
HW=h0 + H-S0L
where h0 = max [TW, (dc + D) 12].
(15.15)
The use of these two equations gives reasonable results for HW depths greater than
0.75Z).
FIGURE 15.6 Head for concrete pipe culverts, (n = 0.012) (From Normann
et al., 1985)
FIGURE 15.7 Head for concrete box culverts flowing full (n = 0.012)
(From Normann, et al., 1985)
15.3.3 Outlet Velocity
Outlet velocity is an important factor to define the type of outlet protection. Because of
the decreased area of flow in the culvert, the velocity increases through the culvert barrel
and results in higher velocities than that of the natural stream. Because of the high velocities, outlets must be protected by using riprap or an energy dissipator. To accurately compute the outlet velocity, the type of control must be defined.
For inlet control culverts, either an exact or approximate method can be used. In the
exact method, water surface profile computations start at the upstream and proceed downstream and the velocity is computed by using the cross-sectional area at the exit. For the
approximate method, normal depth is assumed to occur at the culvert outlet as depicted in
Fig. 15.8. This method is more commonly used and results in more conservative estimates
of the outlet velocity. Normal depth can be computed by using Manning's equation.
For outlet control culverts, the outlet velocity depends on the outlet geometry and the
magnitude of tailwater depth with respect to either the critical depth or the barrel diameter. The following procedure can be used to define the tailwater depth and to compute the
outlet velocity with the given geometry.
If
If
If
TW < dc
dc < TW < D
TW > D
The schematic in Fig. 15.9 also summarizes these three flow conditions.
15.3.4 Roadway Overtopping
Roadway overtopping starts soon after the headwater elevation reaches the top of roadway elevation. To compute the magnitude of the overtopping flow, the following broadcrested equation can be used:
Qr=CjL(HW,)1*
(15.16)
where Qr = overtopping flow rate (ftVs), Cd = overtopping discharge coefficient = Ic1 Cr,
the discharge coefficient Cr, and the submergence factor Ic1, are presented in Fig. 15.10; L
= length of roadway crest, (ft), HWr = the upstream depth, measured above the roadway
crest, (ft)
AREA
OF FLOW PRISM BASED
BARREL
DEPTHON
EQUAL TOGEOMETRYANO
NORMAL DEPTH
FIGURE 15.8 Inlet control outlet velocity. (From Normann et al., 1985)
The main reason for this type of equation is the similarity of this flow to the
broad-crested weir flow. One of the most important factors in the computation of Qr is a
good estimate of the length of roadway crest. The length can be established in two different ways. The first method, although time consuming, results in more accurate estimates
of the length, especially if the elevation of the roadway crest varies. It is based on an
approximation of the vertical curve as a series of horizontal segments and can be applied
to cases where the crest is defined by a roadway sag vertical curve (for definition see Sec.
15.7.3) and is depicted in Fig. 15.11. Then the flow over each segment is calculated for a
given headwater and finally the flows for each segment are added together to determine
the total flow. In the second method, which is more practical, the length can be represented by a single horizontal line at a constant roadway elevation. For correct estimates
of Q1, this horizontal line must span the roadway profile accurately. The depth of flow
used in this method is the average depth of upstream pool above the roadway.
Given the headwater, length and the coefficient, Qr can be computed by using Eq.
FIGURE 15.10 Discharge coefficient and submergence factor for roadway overtopping.
(From Flood Control District of Maricopa County, 1996)
ELEVATION Of CREST
15.4
METHODOFCULVERTDESIGN
The culvert design procedures for both the inlet and outlet control culverts must consider
some important factors (AASHTO, 1990) namely,
Establishment of hydrology
Design of downstream channel
Assumption of a trial configuration
Computation of inlet control headwater
Computation of outlet control headwater at inlet
Evaluation of the controlling headwater
Computation of discharge over the roadway and then the total discharge
Computation of outlet velocity and normal depth
Comparison of headwater and velocity to limiting values
Adjustment of configuration (if necessary)
Recomputation of hydraulic characteristics (if necessary)
15.4.1.1 Design equations. The design equations to be used depend on the condition
of the inlet control. A culvert performs as an orifice when the inlet is submerged and
as a weir when it is unsubmerged. The submerged (orifice) equation can be written as
[^] =c [lfc] 2 + r + z
for
[A*H-
<15-17)
where
HWf = headwater depth above the inlet control section invert (ft); D = interior height of
the culvert barrel (ft), Q = discharge (ftVs), A = full cross-sectional area of the culvert
barrel (ft2), c, Y = constants from Table 15.3, and Z = term for culvert barrel slope (ft/ft).
For mitered inlets, Z = 0.7S. For all other conditions, Z = -0.5S.
The unsubmerged (weir) condition can be presented in two forms. The first one is
based on the specific head at critical depth, and can be written as
[^]-BHAN
'[J-H
where Hc = specific head at critical depth (Hc = dc + V*c/ 2g) (ft), and K, M = constants
from Table 15.3.
The second form is similar to a weir equation and has a simpler form:
[H^F
for
[^H5
(15 19)
'
The latter form is the only documented form for some of the design inlet control nomographs in Normann, et al., (1985).
Nomographs As an alternative procedure, the inlet control nomographs presented in
Figs. 15.12 and 15.13 can be used as described in the following:
(1) Identify the culvert size and flow rate to be used. It is important to note that for box
culverts, the flow rate per foot of barrel width is used.
(2) Connect the culvert size and discharge and extend a straight line to cut the HWID axis.
In case HWID is required in another scale, extend the value at the original intersection
horizontally to cut the appropriate scale.
(3) Once the HWID ratio is computed either by the design equations or the nomograph,
compute the inlet control headwater depth, HW1, by multiplying the barrel diameter by
the ratio HWID.
15.4.2 Outlet Control
Similar to the inlet control culverts, the headwater can be determined either by the design
equations or the outlet control nomographs.
15.4.2.1 Design equations. Unlike the inlet control culverts, the headwater required for
an outlet culvert cannot be computed by using a single equation, but Eq. (15.12) can be
used to compute the energy losses. This concept is illustrated in the Example in Sec.
15.4.2.2.
15.4.2.2 Nomographs.
Unsubmerged
Inlet Edge Description Equation
No.
K
Submerged
M
0.0098 2.000
0.0078 2.000
0.0045 2.000
0.0078 2.000
0.0210 1.330
0.0340 1.500
0.0018 2.500
0.0398
0.0292
0.0317
0.0379
0.0463
0.0553
0.0300
0.67
0.74
0.69
0.69
0.75
0.54
0.74
0.0018
2.500
0.0243
0.83
0.0260
1.000
0.0385
0.81
0.0610
0.750
0.0400
0.80
0.0610 0.750
0.5100 0.667
0.4860 0.667
0.0423
0.0309
0.0249
0.82
0.80
0.83
0.5150
0.667
0.0375
0.79
0.7950
0.667
0.0314
0.82
0.4860
0.667
0.0252
0.87
0.5220 0.667
0.0402
0.73
0.5330
0.667
0.0425
0.71
0.5450
0.667
0.0451
0.68
0.4980
0.667
0.0327
0.75
0.4970 0.667
0.0339
0.80
0.4930 0.667
0.0361
0.81
0.4930 0.667
0.0386
0.71
TABLE 15.3
(Continue)
Shape and
Material
Rectangular box,
top bevels
Corrugated metal
boxes
Unsubmerged
Equation
No.
K
M
0.4970 0.667
0.4950 0.667
0.4930 0.667
0.0083 2.000
0.0145 1.750
0.0340 1.500
Submerged
c
0.0302
0.0252
0.0227
0.0379
0.0419
0.0496
0.84
0.88
0.89
0.69
0.64
0.57
Angl*
Wlngwolofl
Flor*
SCALE WINGWALL
FLARE
EXAMPLE
To otosontolMOllyOto(2)tcalo
or (3)(I),pro|thonet
Korl
HM ilroloht lnclinod lino throvoh
OlllMtlrotoO.
ontf O OCoUt1Or rovorto ot
FIGURE 15.12 Headwater depth for box culverts with inlet control. (From
Normann et al., 1985)
TNOTES ON BEVELSI
FACC
DIMENSI
N of ALL
SlSHALL
OC AMD
TOPBEOKVELS
NOT
LESS
THAN
SHOWN.
TO
OBTAI
TEHMI
NATION IN NONEBEVEL
PLANE
IEINTAHERRECTANGULAR
BOX,
4 OAANCLE.
6. OR
DCCMCASEINCREASE
THE BEVEL
Bvl Angle
EXAMPLE
TFT. o-4rr. Q.aoocFs Q/NB .TI.S
HW HW
HM(II O m FMI
Bi AMK
FACE DlMARE
CNSCNS
RELATED
ANO Of TO
BEVELS
NG EACH
DITOMENSI
N AT
RITHEGHTOPENI
ANGLES
THE OEDOC
FIGURE 15.13 Headwater depth for inlet control rectangular box culverts, 90 headwall chamfered or beveled inlet edges. (From Normann et
al, 1985)
(3) Draw a new line that connects the flow rate Q and headless H and goes through point X,
(4) The value read from this nomograph for //, includes the entrance, friction, and exit
losses.
Some of the important steps for design of outlet control culverts can be summarized as
(1) Compute the tailwater depth
(2) Compute the critical depth, and remember the following restrictions;
dc cannot exceed the diameter of the culvert, D.
if dc < 0.9A use Fig. 15.14
if dc > 0.9D, (use the guidelines in Chapter 3 for a more accurate estimate)
(3) Compute (dc + D) / 2
(4) Compute h0
(15.20)
CAXNOTEXCEED
CRITICAL DEPTH
RECTANGULAR SECTION
CANNOT EXCEED O
CRITICAL DEPTH
L1-L^]
(15.21)
where L1 = adjusted culvert length (ft), L = actual culvert length (ft), H1 = desired Manning
n value, and n = Manning n value from the chart,
(8) Compute the outlet control headwater by using the following equation
HW0111 = H + H0-S0L
(15.22)
Once the inlet control headwater, HW1 and the outlet control headwater, HWout are
computed, the controlling headwater is determined by comparing HW1 and HWout,
if HWout > HW1, the culvert is inlet control;
if HWout > HW1, the culvert is outlet control.
Then the discharge over the roadway is computed as given in Sec. 15.3.4. After computing the total discharge, Qt, the outlet velocity, V09 and the normal depth, dn, the results
are evaluated so that the barrels have adequate cover, the headwalls and wingwalls fit to
the site conditions, and the allowable headwater and overtopping flood frequency are not
exceeded.
Example. Design a reinforced concrete box culvert for a roadway crossing to pass a
50-year discharge of 400 ft3/s. The site conditions are as follows:
Shoulder elevation = 155 ft
Streambed elevation at culvert face = 140 ft
Natural stream slope =1.5%
Tail water depth = 3.0 ft
Approximate culvert length = 200 ft
Downstream channel is a 10 X 10 ft rectangular channel
The inlet is not depressed.
Solution. The design will be performed by using the procedures given in Sec. 15.4
Step L The 50-year design discharge is given as 400 ftVs.
Step 2. The geometry for the downstream channel is given (10 ft X 10 ft rectangular).
Step 3. Select a 7 ft X 5 ft reinforced concrete box culvert with 45 beveled edges in a
headwall.
Step 4. Determine inlet control headwater HW1
1. Either by using the inlet control nomograph (Fig. 15.13)
a) D = 5 ft
b) QIB = 400/7 = 57 ft3 / s / ft
c) HWID = (1) 1.80 for 3/4 in chamfer
(2) 1.64 for 45 bevel
d) HW1 = (HWID) D = 1.64 X 5 = 8.2 ft
Neglect the approach velocity.
F]-W"-
HW^
f 400 I2
-ggij = 0.0314 [(35y(5)o.5J + 0.82 - 0.5(0.015)
HWi = 8.20 ft
Step 5. Determine the outlet control headwater depth at inlet,
1. The tailwater depth is specified as 3.0 ft which is obtained either from a backwater computation or normal depth calculation.
2. Compute critical depth, dc
a. Either by using the following equation (Chaudry, 1993)
A - I^
'-VT
total discharge
where q (ft3/s/ft), unit2 discharge = cm*vert width * anc* ^
acceleration, 32.2 ft/s ; then
/(4OO/7) 2
^ = V 32.2
3.
4.
5.
6.
ravitatina^
=4 7ft
'
Since V= 1.5 \ an armoring riprap, as defined in Sec. 15.8.1 can be appropriate for erosion protection at culvert outlet.
Step 10. Review the results: the culvert barrel has 5 ft of cover, which is adequate and
the allowable headwater (10 ft) is greater than HW1 = 8.2 ft, which is appropriate for design.
75.5 PERFORMANCECURVES
A performance curve is a plot of flow rate versus headwater depth or elevation for a culvert. Because a culvert has several possible control sections (inlet, outlet, throat), a given
installation will have a performance curve for each control section and one for roadway
overtopping, as shown in Fig. 15.15. In addition to these individual performance curves,
an overall culvert performance curve can be constructed by the controlling portions of the
individual performance curves for inlet, outlet, and overtopping.
Inlet control. The curves for the inlet control culverts can be plotted either by using the
design equations or the inlet control nomographs described in Sec. 15.4.1.
Outlet control. For the outlet control culverts there are a number of ways to plot the
performance curves; the use of Eqs. (15.1-15.14), outlet control nomographs, or backwater calculations. In the first two cases, the flow rates that will be used for the design of the
culvert are chosen, then the corresponding total losses are computed and are added to the
elevation of the hydraulic grade line at the culvert outlet to obtain the headwater. In the
backwater method the headwater elevations are computed for the corresponding flow
rates by adding the inlet losses to the energy equation. The energy grade line at the culvert inlet should reflect this adjustment.
Roadway overtopping. An overall performance curve is a useful tool to separate the
culvert flow from the roadway overtopping flow and to compute their respective magnitudes. In order to develop a typical overall performance curve as shown in Fig. 15.15, the
step-by-step procedure as defined in the following can be used:
1. A set of discharge values that will be used in the design of the culvert must be selected. Then the corresponding inlet and outlet headwater values must be computed.
2. The inlet and outlet control performance curves must be combined to develop a single
performance curve for the culvert.
3. To compute the flow rates of the overtopping flow; calculate the depth of upstream
water surface above the roadway for each one of the flow rates selected in Step 1, and
use Eq. (15.16) and the depth calculated in Step 3.
4. Then the overall performance curve will be generated by adding the overtopping and
culvert flow at the corresponding backwater elevations.
Example. Develop a performance curve for two 48-in corrugated metal pipe culverts
(Manning's n = 0.024) with metal end sections. The culvert is 150 ft long and on a 0.01
percent slope. The roadway is a 30-ft-wide paved crossing that can be approximated as a
broad-crested weir 130 ft long. The roadway centerline elevation is 109 ft. The culvert
invert elevations are 100 ft. at the inlet, and 99.98 ft at the outlet. Tail water discharges and
depths are listed below:
Q (ft3 /s)
80
120
160
200
240
280
320
360
TW (ft)
102.4
103.0
103.3
103.5
103.7
104.1
104.3
104.5
Step 1. Compute inlet and outlet control headwater elevations as tabulated below.
Step 2. Develop the rating curve for the overtopping flow by using Eq. (15.16)
Qr=CdL(HWr)"
H E A D W A T E R ELEVATION ( f t )
CULVERT FLUS
ROADWAY
OVERTOPPING
ROADWAY CREST
OUTLET
CONTROL
INLET CONTROL
TOP OF
CULVERT
OVERALL
PERFORMACE
CURVE
HWr
C1
kt
Q,
(Overtopping Flow)
O
(Culvert Flow)
Qt
(Total Flow)
0.30
0.50
3.00
3.05
1
1
130
130
64
184
235
240
299
424
Step 3. Draw the performance curve for the inlet control, outlet control, and overtopping
flow, shown in Fig. 15.17, by using the headwater calculations and Step 2.
Total flow Row per barrel
Q^
(1)
80
40
120
60
160
80
200
100
240
120
Inlet Control
HW1TDl H W ~ I ELh1
(2)
(3)
0.68 2.72 102.72
0.88 3.52 103.52
1.08 4.32 104.32
1.30 5.20 105.20
1.58 6.32 106.32
Outlet Control
Tw I dc I (dc+D)/2 I h0 I ke I H I ELh0
ELh
(4)
(5) (6) (7) (8)
(9)
2.40 1.84 2.92 2.92 0.50 0.65 103.55 103.55
3.00 2.31 3.16 3.16 0.50 1.50 104.64 104.64
3.30 2.73 3.37 3.37 0.50 2.60 105.95 105.95
3.50 3.11 3.56 3.56 0.50 4.20 107.74 107.74
3.70 3.47 3.74 3.74 0.50 5.80 109.52 109.52
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
75.6
Culverts are required to serve under various conditions including wide temperature variations, heavy abrasion, erosion and sedimentation within the culvert barrel. Because of
these factors there are a variety of culvert materials used to maintain the long life and
durability of culverts. The three main types used in application are the precast concrete,
monolithic or cast-in-place concrete, and corrugated aluminum and steel. Some other
options are listed in Table 15.4. The culvert must be able to carry loads without large
deflection and efficiently convey the water through the culvert opening. Therefore, the
selection of culvert material should consider the desired hydraulic efficiency, structural,
and operational characteristics.
One of the factors in the physical strength of the culverts is the type of the material used. The two main applications are with rigid and flexible culverts. Rigid culverts
are usually of reinforced concrete pipe and their primary strength is built into the pipe
itself. Special designs are possible that permit the pipe to support any magnitude of
overload. Their durability against freezing, thawing and erosion are major advantages,
combined with good hydraulic efficiency. In design with smaller culvert size requirements and lighter loads, nonreinforced concrete pipe can be used. Flexible culverts,
generally corrugated metal pipe or plastic pipe, provide their strength by the interaction of the pipe and the surrounding backfill material. Corrosion is a concern with the
corrugated metal pipes.
In many applications the operational characteristics of a culvert are as important as
the hydraulic and structural elements. The culvert designer must select the material
with the consideration of hydraulics, structures, maintenance, and geotechnical characteristics of the culvert site. The proper culvert design should optimize the hydraulic
and structural characteristics of the culvert while minimizing the total cost. The culvert
material which directly affects the service life, is a key parameter in the total cost of
the culvert.
Another important factor for the efficient design is the shape of the culvert conduit.
The shape has an impact on the hydraulics and structural characteristics. The type of the
maintenance and accessibility for deterioration and deformation repairs are also affected by the shape. The most common culvert conduit shapes are circular, rectangular,
pipe-arch, and arch.
15.7
The location and alignment has an important effect on the economics of culvert design. As
a general criteria, culvert location should not alter the natural drainage, and be placed at the
ENTRANCE
TYPE
Hatf*aH
Mttr*d conform
t ilop
?r)!ti4
STRUCTURAL PLATE C. M.
DIAMETER OF CULVERT (O) IN INCHES
&~ SCALE
FIGURE 15.16 Headwater depth for CM pipe culverts with inlet control. (From Normann
et al., 1985)
bottom of the ravine. It is almost impossible to find a unique rule that could be applied to
every culvert. The final location and alignment of the culverts depend on the designer's
own experience in hydrology, hydraulics and structural aspects of culvert design. This will
help the designer to achieve maximum economy, utility, and safety (Hendrickson, 1957).
There are different options to locate the culvert; namely, bottom location placement,
and top location placement.
Material
Protective Coating
Circular
Rectangular
Arch
Pipe-arch
Elliptical
Steel
Aluminium
Al/Fe Alloy
Concrete
Plastic
Galvanizing
Aluminized
Bituminous
Polymer
Concrete
FIGURE 15.17 Head for standard CM pipe culverts flowing full, (n = 0.024). (From
Normann et al., 1985)
Wifo Overtopping
ROADWAV GRADE
FILL
BOTTOM
OF RAVINE
SECT/ON A-A
FIGURE 15.19 Bottom location culverts. (From Hendrickson, 1957)
15.7.2 Top Location Placement
In some applications the culvert can be placed near the top of the embankment fill just
beneath the roadway grade. The main advantage over the bottom installation is its shorter length and considerably reduced load that the culvert has to carry. This type of installation forces the embankment to serve as a dam and results in a rise in the level of the
upstream pool until water flows through the culvert. Therefore the embankment material
must ensure stability during flood events. One other concern is the possible accumulation
of sediment and debris due to stagnant ponding on the upstream side when the water level
drops below the level of the culvert invert. Similar to the second solution of the bottom
location culverts, spillways or auxiliary channels might be required to carry water from
the outlet without undercutting or eroding the embankment.
15.7.3
Siphons
The site conditions might require the modification in the horizontal and vertical orientation, and under certain conditions a culvert may become a siphon. A siphon is defined as
a tube by which liquid can be transferred by means of atmospheric pressure from a higher to a lower level over an intermediate elevation. This is the case for a culvert laid on a
broken grade line as shown in Fig. 15.2IA. Due to the change in grade, the hydraulic
grade line drops below the crown of the culvert when the inlet and outlet are submerged.
The negative pressure or vacuum in the center section of the culvert results in a siphon.
In certain cases the designer may have to use the inverted siphon alternative, which is
also named as sag culvert. The inverted siphon is a consequence of a hydraulic grade line
above the crown of the culvert, thus causing the culvert to flow under pressure. These culverts can be constructed under low fills where it is necessary to excavate down to get the
BOTTOM
OF RAVIHE
required cross-sectional area for hydraulic capacity (Fig. 15.21B), or they can result from
the gradual accumulation of sediment and debris at the upstream and downstream ends.
In irrigated areas, inverted siphons are used to carry the water from open ditches and
canals below highways, railroads, and other obstructions. This type of culvert must be
periodically maintained as a protection against sediment and debris accumulation. Also
they are not suggested for intermittent streams, because stagnant water can lay in them for
a long period of time and can result significant health hazards.
15.7.4 End treatments
End treatments are applied to culverts for many purposes, including; retention of roadway embankment, improvement of hydraulic efficiency, protection of highway embankment from discharge momentum, protection against piping, debris control traffic safety.
Headwall end treatments are very common structures standardized with construction
details and are economical and practical. In some instances, hydraulic efficiency can be
improved by minor details such as flared wingwalls on the headwall. Prefabricated*^
sections are very popular due to their convenience of installation and economy, lneir
application is limited to smaller culverts.
The flowline length of a mitered end culvert (Fig. 15.22) corresponds to the distance
between the upstream and downstream roadway toe-of-slope locations. The top length is
measured between the points of interception with the roadway side slopes and the barrel
top. Mitered end treatments are economical and relatively simple in construction.
Projecting end treatments are similar to mitered end treatments except that the upper
portion of the culvert barrel protrudes from the side slopes of the roadway embankment
as shown in Fig. 15.22. Projecting end treatments are convenient and economical.
However, they have serious problems regarding, hydraulics (greater entrance losses),
structural (inadequate perimeter containment), operational (potential for collision with
maintenance operations).
Another alternative is the use of different improved inlet conditions. Since a culvert
usually represents a severe constriction to the stream flow, the physical characteristics of
f/yafr0uJ/c
'/s>e
//ye/r0t///c
/s?e
Grade
Grat/e
FIGURE 15.21 A. Culvert siphons: culvert siphon, B. inverted siphon culvert. (From
Hendrickson, 1957)
the entrance to the culvert have a significant effect on the operational capacity. This
effect is reflected by the headloss coefficient, fce, associated with the entrance type used.
With reference to the table of entrance loss coefficients in Table 15.1, note that some ke
are much lower than others depending on the type of entrance. Often, improvement in
the hydraulic capacity of the culvert can be realized by simple and inexpensive expedients such as, beveling or rounding of the inside perimeter of the entrance to the conduit
and ensuring installation of reinforced concrete pipe with the grooved end upstream.
Beveled edges are commonly used with box culverts and headwall structures. As given
in Normann, et al., (1985), design charts are available for two bevel angles, 45 and 33.7
(Fig. 15.23). Although the 33.7 bevels result in a better inlet performance than the 45
bevel, it requires additional structural modifications. As a result, the 45 bevel is the preferred alternative. Groove ends provide similar hydraulic performance as the bevels. Figs.
15.24 and 15.25 demonstrate bevels with other inlet improvements. Side-tapered inlets
provide enlarged culvert entrance with a transition to the original barrel dimensions. The
PROJECTING BARREL
CAST-IN-PLACECONCRETE
HEADWALL a WINGWALLS
FIGURE 15.22 Four standard inlet types. (From Norman et al., 1985)
test results of FHWA (Harrison et al., 1972) resulted in a section geometry as shown in
Fig. 15.24. The throat section as identified in this figure is used to control the capacity and
maximize the efficiency of side-tapered inlets. Slope-tapered inlets have steeper slopes at
the entrance as compared to the culvert barrel. The steeper slope is provided to increase
the head on the throat section and create additional fall. Depending on the magnitude of
the available fall, the inlet capacities can be significantly higher than the conventional culvert with square edges.
Culvert inlet performance also can be improved cost effectively, by the incorporation
of special entrance geometries. Effectively, such improvements are used to reduce the
headlosses due to flow contractions and to increase capacity of culverts operating under
steep slope regimes.
Expanded entrances have a "funneling" effect on the flow. These flared entrances
allow more discharge into the conduit, thus significantly improving the efficiency of the
conduit. Their main advantage is in the reduction of constriction loss coefficients. While
there may be a minor loss reduction in outlet control culvert operations, they are mostly
effective in situations of steep slope regime. Culverts operating in a supercritical flow
regime often exhibit excessive outlet velocities. The usual effect of an improved entrance
on such a configuration results in more efficient use of the culvert barrel with the further
result of a lower outlet velocity.
Transitions of flow from the relatively wide channel approach to the constricted culvert opening may be useful in some instances. Both expanded entrances and transitions
have the potential of clogging due to debris. Larger debris is allowed to enter the culvert
partially, and subsequently becomes lodged in the smaller opening provided by the main
conduit.
Improved inlets are useful in some situations but the following considerations should
be made in design:
1.5 : I TOP
BEVEL
W/33.70AN GLE
OF
BARREL < Lj
HEIGHT
f
Cl TOP BEVEL
W/45ANGLE .
I in/ft OF
BARREL
HEIGHT ;
1/2 BARREL
in/ft (
OF
HEIGHT
33.7 TOP
AND SIDE
BEVELS
l.5!l SIDE
BEVEL
W/33.7 0 ANGLE
I in/ft OF
BARREL
WIDTH
450TOP
AND SIDE
BEVELS
1/2
in/ft
OF
BARREL
WIDTH
Face Section
Bevel
Throat Section
ELEVATION
Wingwalls
Taper
PLAN
FIGURE 15.24 Typical side-tapered inlet detail. (From State of
Florida Department of Transportation, 1987)
The improved inlet often is a relatively expensive item, sometimes costing more than
the rest of the culvert.
Improved inlets are effective only in supercritical flow situations.
Debris, erosion, and other maintenance problems can represent significant operational
costs when improved inlets are used.
The design charts and methods for improved inlet design are applicable to one or two
barrels only of reinforced concrete box culvert configurations.
The design charts require a culvert face which is perpendicular to the stream flow. This
can be a construction and maintenance problem when the culvert is skewed to the
roadway centerline.
Face Section
Bevel
Throat Sectic
ELEVATION
Wingwalls
Taper
PLAN
FIGURE 15.25 Typical slope-tapered inlet detail. (From State of
Florida Department of Transportation, 1987)
15.8 SPECIALCONSIDERATIONS
15.8.1 Erosion
The high outlet velocities observed at the culvert outlets may result in excessive scour of
the channel in the vicinity of the outlet. The variety in the soil type of natural channels and
varying flow characteristics at the culvert outlet enforces the use of different methods to
Next Page
CHAPTER 16
HYDRAULIC DESIGN OF
FLOOD CONTROL CHANNELS
George K. Cotton
Simons & Associates, Inc.
Fort Collins Colorado
16.1
INTRODUCTION
Flood protection is essential to the economic and environmental integrity of most civil
engineering projects. The reliability of transportation systems, the livability of urban and
suburban developments, the long-term stability of landfills, the operation of mining excavations, the reclamation of disturbed lands, and the management of forest and agricultural lands require careful planning for flood hazards. In nearly all these cases, design of a
flood-control system will include a variety of conveyance channels referred to as floodcontrol channels.
To perform reliably, flood-control channels must behave in a stable, predictable
manner. This ensures that a known flow capacity will be available for a planned flood
event. In most cases, the design goal is a noneroding channel boundary, although, in
certain cases, a dynamic channel is sought. Since most soils erode under a concentrated flow, channel linings are needed either temporarily or permanently to achieve channel stability.
Channel linings can be classified in two broad categories: rigid or flexible. Rigid lining include channel pavements of concrete or asphaltic concrete and a variety of precast
interlocking blocks and articulated mats. Flexible linings include such materials as loose
stone (riprap), vegetation, manufactured mats of light-weight materials, fabrics, or combinations of these materials. The selection of a particular lining is a function of the design
context, involving issues related to the consequences of flooding, the availability of land,
and environmental needs.
A rigid lining is capable of high conveyance and high-velocity flow. Flood-control
channels with rigid linings are often used to reduce the amount of land required for a surface drainage system. When land is costly or unavailable because of restrictions, use of
rigid channel linings is preferred.
Flexible channel linings are distinguished from rigid linings because they can respond
to a change in channel shape. They can sustain some adjustment of the channel's shape
and still maintain their integrity. This would not be the case for a rigid lining, where local
damage to the lining may lead to a general unraveling. Damage to a rigid lining can result
from secondary forces, such as frost heave, piping, or slumping, although uplift and shear
forces are often causes of failure.
Flexible linings are used as temporary channel linings for control of erosion during
construction or reclamation of disturbed areas. Also, when environmental requirements
are part of the design, flexible lining materials have several advantages. Flexible linings
are inexpensive, permit infiltration and exfiltration, and allow growth of vegetation.
Hydraulically, flow conditions in the channel lined with flexible materials generally can
be made to conform to conditions found in a natural channel. This provides better habitat
opportunities for local flora and fauna. By permitting the growth of vegetation in the channel, flexible linings can provide a buffering effect for runoff contaminant's and sediment.
We are fortunate that a large number of innovative and traditional products are currently available for channel linings. In most cases, these products have received extensive
testing, including laboratory flume studies, large-scale prototype modeling, and documentation of field case histories. From this effort has come a better understanding of complicated flow conditions associated with each type of lining and a better understanding of
product performance and of methods needed for design. Construction experience has
resulted in better specifications for channel lining materials and their installation.
The presentation in this chapter covering flexible lining materials is based on work in
preparing Design of Roadside Channels "with Flexible Linings (Hydraulic Engineering
Circular No. 15) for the Federal Highway Administration. Chen and Cotton (1988), Since
1988, when that manual was published, the commercial market for channel-lining products has expanded. Because product testing and performance monitoring. also have
increased, this chapter also serves to update the reader on the status of current practice.
The so-called tractive force (or boundary shear stress) that acts on the channel's
perimeter describes the basic mechanics of channel stability. Design and field methods
based on tractive force are well suited for the evaluation of the stability of smaller channels where the grade of the channel dominates. In these cases, the calculation of shear
stress is simple, the determination of channel slope is the most difficult estimation. For
field observations on a channel of known grade, depth alone needs to be measured to estimate the maximum shear stress. Field estimates of flow velocity require some type of current meter. In design, the performance criteria are simple to recall for a specific type of
lining because it is represented by a single permissible shear stress value. This permissible shear stress value is applicable over a wide range of channel slopes and shapes.
Permissible velocity criteria, on the other hand, are a function of channel slope, lining
roughness, and channel shape.
76.2 DESIGNCONCEPTS
The design methods are based on the concept of maximum permissible tractive force, coupled with the hydraulic resistance of the particular lining material. The method includes
two parts: computation of the flow conditions for a given design discharge and determination of the degree of erosion protection required. The flow conditions are a function of
the channel geometry, design discharge, channel roughness, and channel slope. The erosion protection required can be determined by computing the shear stress on the channel
at the design discharge and by comparing the calculated shear stress to the permissible
value for the type of channel lining used.
16.2.1 Flood-Control Channel Design
The methods given in this chapter are for small (less than 1 m3/s) to medium-sized floodcontrol channels (less than 10 m3/s). The design of larger flood-control channels is
increasingly complex in terms of the physical environment, the nature of the flood risk,
and the number of issues and constraints that a design team must address. Small channels
can often be designed on the basis of simplified hydrologic analysis (i.e., the rational
method), without detailed geotechnical data, and by assuming uniform flow. A mediumsized flood-control channel typically requires hydrologic routing, information on soil
properties, and a hydraulic analysis based on nonuniform flow.
The level of effort for design increases steadily as the size of the channel increases.
It is expected that it may require approximately four times the hours for technical staff
to accomplish a medium-sized flood channel compared with a small channel. Since
costs will increase by a factor of about 10, there is a definite economy of scale in the
design of larger channels. However, the larger a channel becomes, the more tributary
and appurtenant features will be needed for a complete design. Therefore, the design of
a medium-sized channel often includes the design of several small channels and other
features.
16.2.2 Open-Channel Flow
16.2.2.1 Types of flow. Open-channel flow can be classified according to three general
conditions: (1) uniform or nonuniform flow, (2) steady or unsteady flow, and (3) subcritical or supercritical flow. In uniform flow, the depth and discharge remain constant along
the channel. In steady flow, no change in discharge occurs over time. Most natural flows
are unsteady and are described by a runoff hydrograph. One can assume, in most cases,
that the flow will vary gradually and can be described as steady, uniform flow for short
periods. Subcritical flow is distinguished from supercritical flow by a dimensionless number called the Froude number (Fr), which is defined as the ratio of inertial forces to gravitational forces in the flow. Subcritical flow (Fr < 1.0) is characterized as tranquil and has
deep, slower velocity flow. Supercritical flow (Fr > 1.0) is characterized as rapid and
has shallow, high-velocity flow.
For design purposes, uniform flow conditions also are considered to be steady. The
channel grade S09 water-surface grade Sw, and the energy grade Sj are assumed to be equal.
This allows the development of a flow equation based on a friction loss formula (such as
Manning's equation) and channel grade.
Computation of uniform flow is suitable for small channels and a useful approximation for medium-sized channels. In medium channels, steady uniform flow is rare, and a
more complete analysis using gradually varied flow methods is required. In solving gradually varied flow, the designer should rely on a current computer program, the accuracy
of which is well documented.
Design of steep channels with supercritical flow presents a number of special concerns.
Waves can form in a channel bed with a flexible lining (beginning near a Fr of 0.8) that
ultimately approach the depth of flow. In extremely steep channels, the flow may splash
and surge in a violent manner and additional freeboard is required. Ultimately, at Fr's near
2.0, the flow becomes unstable. Channel designs in this range should be avoided.
16.2.2.2 Resistance to flow and boundary shear stress. Flow resistance is the result of
the drag of moving water against the channel boundary. For practical purposes, the flow
in most channels will be fully turbulent and the velocity, V, will be proportional to the
square root of the shear stress, T, on the channel boundary. This gives the following simple formula for estimating channel velocity:
V=CfViTp
where Cf = the conveyance factor and p = the density of water.
(16.1)
(16.2)
where AL = the incremental reach length (Fig. 16.1), P = the wetted perimeter, y = the
specific weight of water, A = the cross sectional area of flow, and Sf = the friction slope
(energy grade).
The conveyance factor in Eq. 16.1 is not constant but is a function of both boundary
and flow conditions. If this were not the case, there would be little difference between a
velocity-based approach and a tractive shear-based approach to channel design. In
Manning's formulation, the conveyance factor varies as the one-sixth power of the
hydraulic radius, R, giving
C7f = R'
n
(16.3)
(16.4)
Not all channel linings behave according to the Manning's formulation for the conveyance factor. Most rigid channel linings have a Manning's n that is approximately constant. For shallow flows, the n value increases in rigid channels, but this effect is often
neglected even for small flood-control channels. However, for flexible types of channel
linings, the conveyance factor is difficult to describe using Eq. 16.3.
A channel lined with a good stand of vegetation cannot be described by a single n
value. The flow resistance is complicated by the mechanics of the vegetation since the
stem of the plant will bend because of the drag force, changing the height of the plant
relative to the flow depth. The Soil Conservation Service (SCS) (USDA, 1954), through
the work of Ree and Palmer (1949), developed standard classifications for vegetative
flow retardance. Grasses are classified into five broad categories of flow retardance:
Class A identifies grasses with the highest flow retardance and Class E identifies the
grasses with the lowest flow retardance. In general, high flow-retardance species of grass
form a dense cover, are tall, and have stiff stems. Sparse cover or short, flexible grasses
have lower flow retardance.
Temple (1980; Temple et al., 1987) and Kouwen, (1988; Kouwen and Li, 1980;
Kouwen and Unny, 1969) have developed modern approaches to the flow resistance of
vegetation. Temples approach is based on a mathematical fit to the SCS retardance charts
and a generalized retardance index that is a function of two properties of gress: stem
length and cover density. Temple compiled standard values of these properties for selected species for reference stem densities representing "good" cover conditions. Adjusting
these reference properties for local conditions and expected growth allows the designer to
determine an appropriate index for specific design conditions.
Kouwen analyzed the biomechanics of vegetation and developed conveyance factors
that are a function of the density, height, and stiffness of the grass. He showed that the
resulting method gives results that are consistent with the SCS retardance curves. Since
Kouwen's method is physically based, it is a useful tool for extending retardance curves
to channel conditions that were not studied by the SCS, including stiff vegetation and mild
channel grades.
16.2.3
A basic flood-control channel consists of a minimum of five components. There are specific design issues associated with the local flow conditions at each component. More
complex flood-control systems also may include additional components for flood storage
(detention basins) and grade control (low dams).
Channel inlet. At the point where water enters a flood-control channel, local shear
stresses can develop as a result of local acceleration of the flow. For small flood-control channels, the inlet may consist of small berms or dikes that collect runoff and direct
this diffuse flow to the channel. A medium-sized flood-control channel will typically be
the result of the combination of several tributaries or the extension of an existing channel. Collection and control of incoming flow is essential to the proper operation of a
larger channel. In most cases, a flood-control channel will discharge as a tributary to a
larger channel.
Reach. The reach component is the length of channel, with only minor variation in the
properties of grade, discharge, cross section, and lining material. A reach may have a
straight or curving alignment. In larger flood-control channels, a reach also may include
a bridge or culvert.
Flow around a bend in an open channel induces centrifugal forces created by the
change in the direction of flow. This results in a superelevation of the water surface, with
its surface being higher at the outside of the bend than at the inside of the bend. Flow
around a channel bend also imposes higher shear stress on the channel bottom and banks.
The increased shear stress requires additional design considerations within and downstream of the bend.
Confluence. A confluence is the site where two or more flows merge without a significant grade difference between each channel. The process of merging flow involve turbulent mixing and a local increase in energy loss and boundary shear stress. Flow separation, wave formation, and bank impingement can occur in a confluence that creates much
higher local boundary shear stresses and a potential for erosion. Control of the velocity
and direction of tributary inflows often is required to prevent damage to the channel within a confluence. Where a large tributary joins the main channel, performance of the confluence can be improved if the direction of the two flows is as nearly parallel as possible.
Side inlet. A side inlet allows flow to enter the channel over the bank. Example of the
need for a side inlet to a channel are flow from a field or a street. Often, the flow enters
the channel from a chute constructed on the channel bank. When the bank channel is at a
mild grade, vegetation may be a sufficient lining for the chute. For steeper bank slopes,
however, riprap or gabion linings are common.
Channel crossing. A channel crossing is required where a private or public road passes over the channel. Structures used for this purpose are stream fords, culverts, and
bridges. Stream fords are best suited for low-traffic-volume roads crossing channels with
an intermittent type of flow. Fords can be hazardous if flow depths of more than 0.30 to
0.45 m. are expected. Culverts of concrete or metal pipe provide an economical crossing
for small channels. Since the shape of the culvert is typically circular, elliptical, or arched,
the culvert will obstruct a portion of the channel section. Bridges of concrete, steel, or timber are used when the channel must remain largely unobstructed.
Transition. A transition is a gradual expansion or contraction between two channel
sections. Transitions can occur between one reach and another or between a reach and the
channel inlet or outlet.
Channel Outlet. At the point where water exits a flood control channel, local shear
stresses can develop as aresult of deceleration of the flow. At the outlet, the flow changes
to match the local velocity and depth of the receiving channel.
16.2.4 Stable Channels
16.2.4.1 Stable channel modes. A stable channel can be either static or dynamic, but
over time the net effect of scour and deposition must be zero. In a static stable channel,
scour and deposition occur within the limits of a channel boundary that effectively resists
the erosive force of the flow. Although the transport of sediment can be significant in a
static channel, the effective change in channel section at any given time is small. This
implies equilibrium between the incoming supply of sediment and the transport of sediment within a channel reach. When the supply of sediment is small, the principles of rigid
boundary hydraulics can be applied to evaluate the channel's capacity and the lining's stability. When the supply of sediment is large, then the determination of the channel's
capacity must consider the transport of both water and sediment. The transport of sediment affects the flow resistance and therefore the shear forces and the stability of the
channel lining.
In a dynamic stable channel, some change in the channel bed and banks are expected.
The channel is considered to be stable if the morphology (i.e., shape) of the channel
remains unchanged over time. Important measures of channel shape include the channel
width, depth, slope, pattern (straight, meandering, or braided), hydraulic variables (flow
velocity and discharge), and sediment variables (particle size, suspended concentration of
sediment). A dynamic channel can be regarded as stable as long as local changes in channel shape do not have adverse consequences. For most small- or moderate-capacity floodcontrol systems, design of a dynamic channel is often too complex or it presents too many
uncertainties. Consequently, development of static stable channels is usually preferable to
using dynamic approaches to the design.
Several empirical methods are commonly applied to determine whether a channel is
stable. These methods are defined as permissible velocity methods. Permissible velocity
approaches were first developed in the early 1920s (Lacey, 1920, Lindley, 1919) and are
related to the successful development of regime theory for large irrigation canals beginning in 1885 with Kennedy's formula. Regime theory continues to be a basic tool for evaluating large rivers and irrigation canals (Blench, 1969; Simons and Albertson, 1963);
however, its empirical nature limits its application in the case of flood control channels.
In the 1950s, under the direction of Lane at the U.S. Bureau of Reclamation, research
was conducted that refined the permissible tractive force method (Glover and Florey,
1951; Lane, 1955). This work and subsequent research clarified the actual physical
processes occurring for flow in stable channels. This method extended the range of stable
channel design beyond the empirical limits of regime theory. In most cases, a more realistic model of channel stability is based on permissible tractive force.
16.2.4.2 Tractive force. The flow resistance of water moving against a channel boundary results in shear force along the channel boundary referred to as the tractive force. As
was discussed in Sec. 16.1.2, there is a close relationship between the tractive force and
the flow velocity. In a uniform flow with a normal (semilog) velocity distribution, the
tractive force is equal to the effective component of the gravitational force acting on the
body of water parallel to the channel bottom (see derivation of Eq. 16.2).
However, shear stress is not uniformly distributed along the wetted perimeter in a
channel, and Eq. 16.2 describes only the average shear force on the channel. For design,
it is necessary to determine the maximum shear along the channel perimeter so that the
lining material is suitable for resisting the shear. At the same time, the stability of a soil
particle decreases as the lateral slope of the channel increases. The relative combination
of particle stability and boundary shear is needed to attain a stable channel. Simply stated, the principle of stable channel design using tractive force is to have the shear strength
of the lining material ip, exceed the boundary shear force ib at every point in the channel's wetted perimeter:
\> \
(16-5)
The permissible shear stress of the lining material lp is the maximum shear stress that
the lining can safely withstand. Using the critical shear stress Tc for a soil particle as a reference value, the permissible shear stress can be defined as
1
P=CA*
(!6-6)
where Ca is the critical shear-stress adjustment factor. The design shear stress is affected
by a number of factors, including: the density of the soil particles, the submerged soil friction coefficient (\JL = tan <]), where is the angle of repose), the lateral (side slope) and
down-slope components of soil particle weight, and the direction of flow.
The boundary shear stress ib varies within a channel section primarily as a function of
depth and also in response to the lateral diffusion of momentum in the channel section.
Using the mean shear stress T0 as a reference value, the local boundary shear stress can be
defined as
Velocity Distribution
Bank Shfrar
Bed Shear
Shear Distribution
FIGURE 16.2 Boundary shear distribution in channel section
\=&a\
(16.7)
where K0 is the boundary shear-stress adjustment factor. Channel alignment, the relative
channel width and depth (aspect ratio), and the channel side slope affect the boundary
shear stress.
Figure 16.2 shows a typical distribution in a parabolic channel. Boundary shear is
greatest at the middle of the channel (maximum flow depth) and tends toward zero
along the channel bank. The distribution shear of in a channel bend is shown in
Fig. 16.3.
16.2.5 Design Parameters
Flood frequency. The most important question facing the designer of a flood-control
channel is - What is the probability of failure? Because flood-control channels are built to
protect valuable facilities, damage should occur only rarely. The probability of flood damage, although never zero, can be low enough to make the risk to the facility acceptable.
Flood Frequency
(Return Period, years)
2 to 5 years
10 to 25 years
25 to 50 years
100 years
When flows are small and potential damages are small, such as in the case of a roadway drainage system, more frequent flooding is allowed. For larger flood-control channels, where flood damage is severe, a low frequency of flooding is desired. Although economic analysis can be used to determine the optimal range of flood frequency for a floodcontrol facility, only larger projects warrant such an effort. In most cases, standard flood
frequencies are used. Table 16.1 provides a brief summary of common flood frequencies
and their use in the design of various types of facilities.
Duration offloading.
Small flood-control channels with intermittent flooding tend
to have short times of flooding, typically on the order of several hours. Larger flood-control channels, or channels with sustained flows, require a design that considers the persistence of boundary stresses. Well-maintained channels that sustain only short periods of
boundary stress often can sustain higher stresses without damage.
Channel profile. Major project features, such as roads generally dictate the slope of
smaller channels. If channel stability cannot be maintained for these conditions, it may be
feasible to reduce the channel gradient slightly relative to other grading. This typically
results in short, steep reaches that are constructed as rigid chutes or low drops.
Channel slope is among the major parameters affecting boundary shear stress. Thus
at a given discharge, the shear stress is less for a mild (subcritical regime) gradient
compared with a steep (supercritical regime) gradient. It is not uncommon for smaller
flood-control channels to operate mainly in supercritical regime. However, extremely
steep channels (above 10 percent) and larger channels require additional design considerations.
Channel section. The most common shapes of drainage channel are trapezoidal or triangular. These shapes are a close approximation of the most efficient hydraulic section, a
semicircle. For smaller channels with flexible linings, these shapes tend to be constructed
with slightly rounded corners, which gives the as-built section a slightly parabolic shape.
The banks are the least stable portions of the channel section. Bank stability increases
as the side slope flattens. For smaller channels, a 3:1 side slope is sufficient to prevent erosion in excess of the channel bed. For larger channels, or for small channels that require
steep side slopes, the stability of the side slope should be checked.
In larger flood-control channels with persistent low flows, a low-flow channel often is
included within the channel section. It is advisable to situate the low-flow channel near
the center of the main channel to avoid creating increased shear near the channel bank.
Low-flow channels are often designed to meander within the larger section. The meander
reduces the gradient of the channel, allowing the enhancement of the channel section for
habitat or appearance. The potential effect at meander bends on the main channel bank
line should be considered.
Although a channel section can be widened to reduce boundary shear, channels with
large width to depth ratios typically are not stable. Over time, a high stress area will develop within the section and cause erosion that will form a narrower, deeper channel, the
channel eventually fails.
16.3
CHANNELLININGS
Considerable research and development has produced a variety of rigid and flexible lining materials. Before the late 1960s, channel linings were constructed primarily of natural materials, such as rock riprap, stone masonry, concrete and vegetation. Material manufactured or fabricated into rolls offered several advantages for erosion control, particularly during construction and during the establishment of vegetation. Use of rolled erosioncontrol products for temporary erosion control assured improved long-term performance
of vegetative linings.
16.3.1 Lining Types
Because of the large number of channel stabilization materials currently available, it is
useful to classify materials according to their performance characteristics. Lining types
are classified as either rigid or flexible. Flexible linings are grouped further as either permanent or temporary.
Rigid linings
Cast-in-place concrete
Cast-in-place asphaltic concrete
Stone masonry
Soil cement
Fabric formed concrete
Grouted riprap
Flexible linings (Long-term
nondegradable)
Riprap
Wire-enclosed stone
Vegetation
Gravel
Synthetic mat
16.3.2
Performance Data
16.3.2.1 Rigid linings. Rigid linings are useful in applications where high shear-stress or
nonuniform flow conditions exist, such as at transitions in channel shape or at an energy
dissipation structure. In areas where loss of water or seepage from the channel is undesir-
able, a rigid lining can provide an impermeable barrier. Because rigid linings are
nonerodible, the designer can use any channel shape that provides adequate conveyance.
Rigid linings may be the best option if right-of-way limitations restrict the channel size.
Despite the nonerodible nature of rigid linings, they are highly susceptible to failure from structural instability. For example, cast-in-place or masonry linings often
break up and deteriorate if the foundation is poor. Once a rigid lining deteriorates, it is
highly susceptible to erosion because it forms large, flat, broken slabs that are easily
moved by flow.
The major causes of structural instability and failure of rigid linings are freeze-thaw,
soil swelling, and excessive soil pore-water pressure. Freeze-thaw and swelling soils exert
upward forces against the lining, and the cyclic nature of these conditions eventually causes failure. Excessive soil-pore pressure may occur when the flow levels in the channel
drop quickly but the soil behind the lining remains saturated. This can result in instability of the bank's slope caused by the high water-table gradients within the channel bank.
Construction of rigid linings requires specialized equipment using relatively costly
materials. As a result, the cost of rigid channel linings is high. Prefabricated linings can
be a less expensive alternative if shipping distances are not excessive.
There often are significant environmental issues associated with rigid channel linings. In
environmentally sensitive areas, replacement of concrete linings with articulated block mats
is a partial solution, because openings between the blocks allows vegetation to take hold.
16.3.2.2 Flexible linings. Riprap and gabion are suitable linings for hydraulic conditions
similar to those requiring rigid linings. Because flexible linings are permeable, they may
require protection of the underlying soil to prevent washout. For example, filter cloth is
often used with riprap to inhibit soil piping.
Vegetative and temporary linings are suited to hydraulic conditions where uniform
flow exists and shear stresses are moderate. Most grass linings cannot survive sustained
flow conditions or long periods of submergence. Grass-lined channels with sustained low
flow and intermittent high flows often are designed with a composite lining of a riprap or
concrete low-flow section.
The primary use of temporary linings is to provide protection from erosion until vegetation is established. In most cases, the lining will deteriorate over the period of one
growing season, which means that successful revegetation is essential to the overall channel stabilization effort. Temporary channel linings can be used without vegetation to control erosion on construction sites temporarily.
16.3.3
The erosion control industry has grown in response to continued infrastructure development and increased awareness of water-quality problems. Traditional methods of erosion
control are well documented, and a variety of specifications exist for these types of linings. In 1997, the Erosion Control Technology Council (ECTC) established standard terminology and testing methods for the variety of manufactured products used for erosion
control. The ECTC terminology for rolled erosion control products (RECPs) is used here
for clarity,
16.3.3.1 Long-term, Nondegradable flexible linings. Vegetation. Vegetative linings
consist of planted or sodded grasses placed in and along the drainage (Fig. 16.4). If planted, grasses are seeded and fertilized according to the requirements for the particular variety or mixture and to soil conditions. Sod is laid parallel to the direction of flow and can
be secured with staples or stakes.
Rock riprap. Rock riprap is dumped or hand placed on prepared ground with a filter
blanket or a prepared bedding material interface (Fig. 16.5). The stone layer is placed to
form a well-graded mass with a minimum of voids. Stones should be hard, durable, prefer-
Gravel. Gravel consists of coarse gravel or crushed stone placed on filter fabric or
prepared bedding material to form a well-graded lining with a minimum of voids (Fig.
16.7). The stones should be hard, durable, and free from overburden, shale, and organic material.
Turf reinforcement mat. Turf reinforcement mat is composed of ultraviolet-stabilized, nondegradable, synthetic fibers, netting, or filaments processed into a three-dimensional reinforcement matrix ranging in thickness from 6 to 20 mm (Fig. 16.8). The mat
provides sufficient thickness, strength, and void space to permit soil filling and development of vegetation within the matrix. The mat is laid parallel to the direction of flow is
stapled or staked to the channel surface, and is anchored into cutoff trenches at regular
intervals along the channel. The mat is top-dressed with fine soil at a depth equal to the
mat's thickness and is seeded and fertilized. Turf reinforcement occurs as root growth penetrates and entangles with the mat.
16.3.3.2 Temporary degradable flexible linings. Woven paper net. Woven paper net is
erosion-control net that consists of knotted plastic netting interwoven with paper strips
(Fig. 16.9). The net is applied evenly on the channel slopes, with the fabric running parallel to the channel's direction of flow. The net is stapled to the ground and placed into
cutoff trenches at regular intervals along the channel. Woven paper net is usually installed
immediately after seeding operations.
Jute net. Jute net is an erosion control mat that consists of jute yarn approximately 6
mm in diameter that is woven into a net with openings approximately 10 to 20 mm (Fig.
16.10). The jute net is loosely laid in the channel parallel to the direction of flow. The net
is secured with staples and is placed int cutoff trenches at regular intervals along the channel. Jute net is usually installed immediately after seeding operations.
Curled wood mat. Curled wood mat is an erosion control blanket that consists of
wood fibers, 80 percent of which are 150 mm or longer, with a consistent thickness and
an even distribution of fiber over the entire mat (Fig. 16.11). The topside of the mat is covered with biodegradable plastic mesh. The mat is placed in the channel parallel to the
direction of flow and is secured with staples and cutoff trenches.
Straw with net. Straw with net can be manufactured as an erosion control blanket
or can be constructed from straw mulch secured with erosion control net. The constructed version consists of plastic mesh with 20-mm2 square openings overlaying
straw mulch (Fig. 16.12). Straw is spread uniformly over the channel surface at a rate
of approximately 4.5 metric tons/hectare and can be crimped into the soil. Plastic mesh
is placed over the straw and stapled to the soil. The manufactured version consists of
straw, either sewn to a layer of erosion control net or sandwiched between to two layers of erosion control net. Straw weight is approximately 2.5 metric tons/hectare. The
mat is placed parallel to the direction of flow and is secured with staples and cutoff
trenches.
Fiberglass roving. Fiberglass roving consists of continuous fibers drawn from molten
glass that is coated and lightly bound together into roving (Fig. 16.13). The roving is ejected by compressed air, forming a random mat of continuos glass fibers. The material is
spread uniformly over the channel and anchored with asphaltic materials.
Next Page
CHAPTER 17
HYDRAULIC DESIGN
OF SPILLWAYS
H. Wayne Coleman
C. Y. Wei
James E. Lindell
Harza Engineering Company
Chicago, Illinois
17.1
INTRODUCTION
The spillway is among the most important structures of a dam project. It provides the project with the ability to release excess or flood water in a controlled or uncontrolled manner to ensure the safety of the project. It is of paramount importance for the spillway facilities to be designed with sufficient capacity to avoid overtopping of the dam, especially
when an earthfill or rockfill type of dam is selected for the project. In cases where safety
of the inhabitants downstream is a key consideration during development of the project,
the spillway should be designed to accommodate the probable maximum flood. Many
types of spillways can be considered with respect to cost, topographic conditions, dam
height, foundation geology, and hydrology. The spillways discussed in this chapter
include overflow, overfall, side-channel, orifice, morning-glory, labyrinth, siphon, tunnel,
and chute spillways A section on design of spillways that considers cavitation and aeration also is included.
77.2 OVERFLOWSPILLWAY
An overflow spillway can be gated or ungated, and it normally provides for flow over a
gravity dam section. The flow remains in contact with the spillways surface (except for
possible aeration ramps) from the crest of the dam to the vicinity of its base. The hydraulic characteristics are defined as follows:
1. Determine design head H0. Normally, H0 is 75 to 80% of maximum head #max.
2. Use the depth from the crest to ground surface P to find the basic discharge coefficient C0 from Fig. 17.1.
3. Find discharge coefficient C for the full range of heads from Fig. 17.2.
4. Correct discharge coefficient C for the sloping upstream face from Fig. 17.3. The
sloping upstream face is normally for structural stability, not hydraulic efficiency.
5. Correct discharge coefficient C for the downstream apron from Fig. 17.4.
6. Correct discharge coefficient C for tail water submergence from Fig. 17.5
7. Define the shape of the pier nose. Normally, use Type 3 or 3A from Fig. 17.6.
8. Check the minimum crest pressure from Fig. 17.7 or 17.8. If the minimum pressure
is below -1/2 atmosphere, increase H0 and start over.
9. Define the crest shape from Fig. 17.9(a) and (b).
10. Determine the effective crest length for the full range of heads from
L = L -2(NKp +K0)H6
(17.1)
K
,
0.02
0.01
0.00
(17.2)
Exhibits 17.1 and 17.2 illustrate overflow spillways for hydroelectric projects.
VALUES OF COEFFICIENT CQ
VALUES OF-H0
FIGURE 17.1 Discharge coefficients for a vertical-faced ogee crest. (USBR, 1987).
Slope
Angle
with
tht vrticol
hrifd
POSITION OF DOWNSTREAM APRON -^-
FIGURE 17.4 Ratio of discharge coefficients associated with the apron effect.
(USER, 1987).
DEGREE OF SUBMERGENCE ^f
FIGURE 17.5 Ratio of discharge coefficients associated with the tailwater effect.
(USER, 1987).
CREST AXIS
TYPE 3
TYPE I
CREST AXIS
TYPES 3 ANO 3A
TYPE 4
PIER NOSE SHAPES
NOTE: PIER
NOSE LOCATED
IN SAME
PLANE
AS UPSTREAM
FACE OF
SPILLWAY.
DIMENSIONS
IN
PARENTHESES
ARE
FOR
TYPE 3A.
PRESSURE jhp\
HORIZONTAL
DISTANCE
DESIGN HEAD
AXIS BOTH QUADRANTS
CCMART MI-S)
CREST
FIGURE 17.7 Crest pressures along piers (Type 3A) for high-overflow dams. (USAGE, 1988)
PRESSURC
CESlGN HEAD
HORIZONTAL DISTANCE
DtSiCN HKAO
AXIS BOTH QUADRANTS
TYPE 3A PIER
(CHART in-4)
CREST
FIGURE 17.8 Crest pressures along the centerline of a pier (Type 3A) bay. (USACE, 1988)
upstream face
VALUES OF K
FIGURE 17.9 (a) Factors for the definition of nappe-shaped crest profiles (USER, 1987)
FIGURE 17.9(b) Factors (xc, yc, R1, and R2) for definition of nappe-shaped crest
profiles.(USBR, 1987)
El 31.50:
El 34.00
BEGINNING OF 1% SLOPE
1% SLOPE
PROTECTION APRON
Exhibit 17.2 Tarbela Hydroelectric Project, Pakistan (Courtesy Water and Power
Development Authority, Pakistan)
(a) General view of the spillway in operation.
(b) Profile of the spillway including the flip bucket.
OVERFALLSPILLWAY
77.3
An overfall spillway can be gated or ungated and provide for flow over an arch or archbuttress dam, wherein the flow free-falls some distance before entering a plunge-pool
energy dissipator in the tailrace. The hydraulic characteristics are defined as follows:
1.
The crest structure and discharge rating are similar to those for the overflow spillway.
2. The flow normally leaves this structure shortly below the crest. The exit structure
is normally some variation of a flip-bucket.
3. The flip-bucket radius for an overfall spillway is normally smaller than the ideal,
which is at least 5d, where d is the flow depth at the bottom of the bucket. The
radius is usually undersized to minimize the size of the overhang, which can destabilize the top of a thin-arch dam. However, the radius should be sufficient to
fullydeflect a significant flood say, the 100-year event.
4. The bucket exit angle is selected to throw the jet to a suitable location in the tailrace. The trajectory can be estimated by
y = * * n g - 36f/;os2,
(17.3)
where: y = vertical distance from the bucket lip, x = horizontal distance from the
bucket lip, 6 = bucket exit angle, and H = depth + velocity head at the bucket lip.
5.
6.
7.
The trajectory can be estimated from Step 4 as long as the bucket radius exceeds
5d. For larger depths, the flow overrides the bucket and will fall short of the maximum trajectory. The trajectory in this range is determined best by a physical
model.
Pressure load on the bucket can be estimated from Fig. 17.10. For larger floods,
where d > R/5, this load is determined best by a physical model.
The energy from an overfall spillway is normally dissipated by a plunge pool,
which can be lined or unlined. If unlined, the scour and the scour rate will be
based on both flow and geology. The scour hole development is usually indeterminate. However, the terminal scour depth for a uniformly credible material can
be estimated from the following empirical formula and from Fig. 17.11
(Coleman, 1982; USER, 1987)
ys = ds 5woc = terminal vertical scour depth
(17.4)
ds = C8 225 q-54
(17.5)
and
where Cs = 1.32 (for units in ft and cfs/ft) 1.90 (for units in m and cms/m), H =
effective head at tailwater level, q = unit discharge = Q/B, B = width of the bucket, and a = average jet entry angle.
The extent of the scour hole is based on judgment of stable slope of material surrounding the deepest hole. A physical model is normally used where topography is
complex and where scour can endanger project structures.
8.
9.
As the jet plunges into the pool, it diffuses almost linearly and entrains air at the surface of the pool and the water from the pool at the boundary of the jet. The behavior
of the plunging jet, including dynamic pressures, can be approximated (Hinchliff and
Houston, 1984) using Fig. 17.12 and Table 17.1 for both rectangular and circular jets.
If the plunge pool is lined because the scour would be unpredictable and unacceptable, the lining must be designed for pressure pulsations from the jet's impact.
O./O.T-'.0(FL/P BUCKETS)
A
X
A
a
DEFiNITION SKETCH
LEGEND
PINE TLAT MODEL
HARTWELL MODEL
PINC TLAT PROTOTYPE
ES aoi (TOE CURVE)
FIGURE 17.10 Pressures for flip-buckets and toe curves of an overflow spillway. (USAGE, 1988).
TRAJECTORY
ORIGINAL GROUND
SCOUR
FIGURE 17.11 Definition sketch of free-jet trajectory and scour depth of an overflow spillway.
(Coleman, 1982)
jet core
velocity head
dynamic pressure
velocity
Circular jet
1 + 0.414 ylyk
1 - 0.184 y/yk
g-n/8(l + x/Bu'y/y/yk-yk/y)2
-1/2(1 + r/R^Yfi-yfip
y<yk
Q
jE
^u
Z.
vz
P_
Pz
y
:r
V
u
%
Pu
e-nl\6(xlBufi
e-l/2(r/Ru)2
^7)7
yJy
yJy
Wyf
^-
1.414 V^
-j-
0.816 V^
Ji
e-rtS(ABu'yk/yP
e-l/2(r/Ru'yk/yfi
JL
e-nl6(x/Eu-y]<!yfi
z
Source: Whittaker and Schleiss (1984).
-l/4(r//eu-y^/y)2
y>yk
2y/yt
0.667 y/y
Flow
Existing masonry
lining (typ.)
Existing masonry
fining <typ>
Existing dam
concrete
Downstream face
of existing dam
(Bond length)
T/rock
varies
(C)
Exhibit 17.3 Boyd's corner dam project, New York (Continued).
(b) A view of the overfall spillway in operation.
(c) Profile of the spillway including the flip bucket.
The design pressures can be determined best from physical model studies.
Exhibit 17.3 illustrates an overfall spillway in operation.
77.4
SIDE-CHANNELSPILLWAY
A side-channel spillway can be gated or ungated and provides for flow into a chute or tunnel at right angles because the abutment topography is not favorable for a normal crest
alignment. Figure 17.13 shows a typical arrangement. Exhibit 17.4 illustrates a side-channel spillway in operation The hydraulic features are defined as follows (USBR, 1987):
1. The crest is often ungated since a long crest may be suitable because of favorable
topography.
2. The crest shape is based on the same criteria as the criteria for an overflow spillway.
3. The trough is sized by trial and error to prevent the maximum discharge water surface
from encroaching on the crest's free-discharge capacity. The trough should be as nearly V-shaped as possible to promote efficient dissipation of energy.
4. The chute crest is proportioned to produce subcritical flow in the trough for all discharges
to dissipate the overflow energy and produce uniform flow into the chute.
5. The trough geometry for the first trial is proportioned to produce an approximately
uniform reduction in trough velocity from downstream to upstream. This will usually
minimize the trough size.
6. The water surface profile in the trough is estimated by the following:
-ff^h-^^]
where AF = change in water level between two sections AX apart, Q2,V2 = discharge
and velocity at the downstream section, and Q1^V1 = discharge and velocity at the
upstream section. The computation begins with a known water level at the downstream
end produced by critical depth control at the chute crest, and proceeds upstream in
increments of AX by trial and error, calculating the water level along the length of the
side channel. If the calculated water profile encroaches on the side channel's capacity,
the geometry of the trough or chute crest or both should be adjusted.
SlOECHANNCl.
CBEST
CHUTE
CNUTE BLOCKS
OENTflTEO SILLS
7. Once an acceptable geometry for the trough and chute crest has been determined, the
chute should be sloped to provide supercritical flow away from the chute crest.
8. Large side-channel spillways are almost always model-tested because of the complex
flow conditions in the trough area.
77.5
ORIFICESPILLWAY
An orifice spillway is normally gated and is used when substantial discharge capacity is
needed at low reservoir levels, as illustrated in Exhibit 17.5. For instance, it is useful when
sediment sluicing is required. It also is useful for diverting flow during construction. Gate
sizes are normally smaller for these spillways, but higher head and sealing details can
make them expensive.
Orifice spillways can be found in gravity dams, adjacent to embankment dams, and in
arch dams. Figs. 17.14 and 17.15 show a typical layout. The proportions of the orifice are
defined as follows:
1. Gates are sized on the basis of discharge requirements at maximum and minimum
headwater levels. The discharge coefficient can be assumed to be 0.90.
2. The roof curve must be shaped carefully to prevent cavitation if the head is high. For
instance, the design in Fig. 17.15 has a roof curve composed of a circular curve with
a large radius, followed by a 1:15 roof slope to the gates top seal. This shape was based
on a physical model test (Fig. 17.14).
Groyne Well
Upper
St.llmg
Bonn
Upocr Ch!t
Chute
PLAN
Scolt of Fttl
SECTION
FIGURE 17.14 Typical arrangement of an orifice spillway. (Institute of Civil Engineers, (1968).
Toilroct
Cltonntl
3. The floor curve is less critical; however, for the design in Figs. 17.14 and 17.15 the
curve radius was 150 ft.
4. Potential formation of vortex is a special concern for orifice spillways. A vortex suppressor will generally be required and can be determined by a physical model.
5. The spillway downstream of the gate structure changes to a chute that may lead to a
stilling basin (see Figs. 17.14 and 17.15) or a flip-bucket.
77.6
MORNINGGLORYSPILLWAY
The morning glory spillway is normally used in conjunction with a tunnel spillway when
the intake is a vertical shaft. Also, because of flow entry from the entire periphery, the
crest capacity is relatively high. Crest gates are not normally used because of access and
cost considerations as well as poor hydraulic conditions in the shaft with partially open
gates Exhibit 17.6 illustrates a morning glory spillway. Figures 17.16 and 17.17 show the
range of possible flow conditions in the crest area for a morning-glory spillway
(USER, 1987). The design procedure is as follows:
1. By trial and error, determine the required design head H0 and the crest radius from Fig.
17.18; H0IRs = 0.3 is recommended. Note that in Fig. 17.18, the discharge coefficient
C0 is for English units. For metric units, the coefficient should be multiplied by a conversion factor of 0.552.
2. Determine the discharge rating curve for the full range of heads from Fig. 17.19.
3. Determine the lower nappe profile from Figs. 17.17 and 17.20 and Tables 17.2, 17.3,
and 17.4.
gl/2
R^
R =C
(17-7)
where: CR = 0.275 for units in m and mVsec, = 0.204 for units in ft and ftVsec, R =
Crest
Transition tub*
Throot of tronsition tub*
inclined shaft
AcceUrotmg
Dc*l*roting flow
CONDITION I. CREST CONTROL
Transition tub*
Oific* control at throat of transition
Hydraulic gradient
Toilwoter.
CONDITION 3. FULL PIPE FLOW
Design head
Point
controlof chong*
to pip* from
flow orifice
Tub* or orifice control, O F(H9'4), condition 2
H*ad
at which tunnel
at downstream
end. flows 0.79 full
Erratic
rqnge flow
FIGURE 17.16 Possible flow conditions for a morning-glory spillway. (USER, 1987).
Noppe-shoped crest
Crotch,where onnulor flow
becomes jet flow.
COEFFICIENT C0
FIGURE 17.18 Circular crest coefficients for a morning-glory spillway with aerated nappe.
(USER, 1987)
FIGURE 17.19 Circular crest coefficients of discharge for other than design head. (USER, 1987)
TABLE 17.2 Coordinates of Lower Nappe Surface for Different Values of HJR When PIR = O.15
^
0.20
0.2
0.30
0.35
0.40
0.45
0.50
0.60
0.80
X
Y For portion of the profile above the weir crest
"s
"j
0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000 0.0000
.010
.0120 .0120 .0115 .0115 .0110 .0110 .0105 .0100 .0090
.020
.0210 .0200
.0195 .0190 .0185 .0180 .0170 .0160 .0140
.030
.0285
.0270
.0265
.0260
.0250
.0235
.0225
.0200 .0165
.040
.0345
.0335
.0325
.0310 .0300
.0285
.0265
.0230 .0170
.050
.0405
.0385
.0375
.0360
.0345
.0320
.0300
.0250 .0170
.060
.0450
.0430
.0420
.0400
.0380
.0355
.0330
.0365 .0165
.070
.0495
.0470
.0455
.0430
.0410 .0380
.0350
.0270 .0150
.080
.0525
.0500
.0485
.0460
.0435
.0400
.0365
.0270 .0130
.090
.0560
.0530
.0510 .0480
.0455
.0420
.0370
.0265 .0100
.100
.0590
.0560
.0535
.0500
.0465
.0425
.0375
.0255
.0065
.120
.0630
.0600
.0570
.0520
.0480
.0435
.0365
.0220
.140
.0660
.0620
.0585
.0525
.0475
.0425
.0345 .0175
.160
.0670
.0635
.0590
.0520
.0460
.0400
.0305 .0110
.180
.0675
.0635
.0580
.0500
.0435
.0365
.0260
.0040
.200
.0670
.0625
.0560
.0465
.0395
.0320
.0200
.250
.0615 .0560
.0470
.0360
.0265
.0160 .0015
.300
.0520
.0440
.0330
.0210 .0100
.350
.0380
.0285
.0165 .0030
.400
.0210 .0090
.450 .0015
.500
.550
I
-tH1
0.000
.020
.040
.060
.080
.100
.150
.200
.250
.300
.400
.500
.600
.800
1.000
1.200
1.400
1.600
1.800
2.000
2.500
3.000
3.500
4.000
-4.500
5.000
5.500
6.000
^j-K
0.454
.499
.540
.579
.615
.650
.726
.795
.862
.922
1.029
1.128
1.220
1.380
1.525
1.659
1.780
1.897
2.003
2.104
2.340
2.550
2.740
2.904
3.048
3.169
3.286
3.396
0.20
0.422
.467
.509
.547
.583
.616
.691
.760
.827
.883
.988
1.086
1.177
1.337
1.481
1.610
1.731
1.843
1.947
2.042
2.251
2.414
2.530
2.609
2.671
2.727
2.769
2.800
0.25
0.392
.437
.478
.516
.550
.584
.660
.729
.790
.843
.947
1.040
1.129
1.285
1.420
1.537
1.639
1.729
1.809
1.879
2.017
2.105
2.153
2.180
2.198
2.207
2.210
0.30
0.35
0.40
0.45
0.50
0.60
0.80
TABLE 17.3 Coordinates of Lower Nappe Surface for Different Values of HJR When PIR = 0.30
^-
-jj"s
0000
.010
.020
.030
.040
.050
.060
.070
.080
.090
.100
.120
.140
.160
.180
.200
.250
.300
.350
.400
.450
.500
.550
fj-Y
"s
0.000
.020
.040
.060
.080
.100
.150
.200
.250
.300
.400
.500
.600
.800
1.000
1.200
1.400
1.600
1.800
2.000
2.500
3.000
3.500
-4.000
-4.500
5.000
5.500
6.000
^
0.20
0.2
0.30
OM9
.560
.598
.632
.664
.693
.760
.831
.893
.953
1.060
1.156
1.242
1.403
1.549
1.680
1.800
1.912
2.018
2.120
2.351
2.557
2.748
2.911
3.052
3.173
3.290
3.400
0.20
0488
.528
.566
.601
.634
.664
.734
.799
.860
.918
1.024
1.119
1.203
1.359
1.498
1.622
1.739
1.849
1.951
2.049
2.261
2.423
2.536
2.617
2.677
2.731
2.773
2.808
0.25
O455
.495
.532
.567
.600
.631
.701
.763
.826
.880
.981
1.072
1.153
1.301
1.430
1.543
1.647
1.740
1.821
1.892
2.027
2.113
2.167
2.200
2.217
2.223
2.228
0.30
0.35
0.40
0.45
0.50
0.60
0.80
0.35
0.40
0.45
0.50
0.60
0.80
TABLE 17.4 Coordinates of Lower Nappe Surface for Different Values of HJR When
PIR = 2.00
^
0.00 0.10* 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.60 0.80 1.00 1.20 1.50 2.00
Exhibit 17.6 Big Dalton Dam, California (Courtesy Los Angeles Dept. of Public Works)
(a) General view of the modified morning-glory spillway.
(b) Profile of the spillway.
required shaft radius for throat control (m or ft), Ha = head from headwater level to
throat location (m or ft), and Q = discharge under consideration (mVsec or ftVsec). If
R exceeds the radius of the shaft for a given flow, then throat control exists and the discharge is based on the shaft radius.
5. The downstream tunnel is ordinarily sized to flow no more than three-fourths full with
the maximum value of Manning's roughness coefficient n = 0.016 to avoid potentially unstable flow conditions. If the tunnel flows full, with throat control already developed, the capacity is dictated by full flow throughout.
6. An ideal design has crest control throughout the range of discharge.
77.7
LABYRINTHSPILLWAY
The labyrinth spillway is used to concentrate discharge into a narrow chute, where space
does not permit a linear ungated crest. It generally minimizes approach excavation,
whereas the concrete weir is more complicated to construct. Labyrinth spillways have
been built with a wide range of sizes and discharge capacities and are well suited for rehabilitation of existing spillway structures when increased spillway capacity is needed
(Hinchliff and Houston, 1984; Tacail et al., 1990; Tullis el al., 1995). Labyrinth structures
can be built economically if an adequate foundation is available.
Figure 17.21 is a typical layout (Tacail et al., 1990). The most efficient spillway
entrance for most reservoir applications is a curved approach adjacent to each end cycle
of the spillway, with the approach flow parallel to the centerline of the spillway cycles for
more uniform approach flow. The spillway entrance should be placed as far upstream in
the reservoir as possible to reduce localized upstream head losses. To avoid the submergence
effect, the supercritical flow condition should be maintained at the chute crest downstream
from the labyrinth. The number of spillway cycles should be determined on the basis of the
magnitude of the upstream head, the effect of nappe interference, and the economics of the
design. With normal operating conditions, the vertical aspect ratio for each labyrinth cycle,
w/P (Fig. 17.22), should be 2.5 or greater (Tullis et al., 1995).
Subatmospheric pressures under the nappe will cause nappe oscillation and noise and
should be avoided by venting for structural reasons. Splitter piers can be placed along the
spillway's side walls, or crushed stone can be placed along the downstream edge of the crest.
The piers should be located at a distance equal to 8 to 10 percent of the wall length upstream
of the downstream apexes (Tullis et al., 1995). Although the use of crushed stone may
reduce the spillway's capacity, it can be cost effective. Discharge capacity is a function of
the head over the crest, the height of the crest wall, the shape of the crest, the angle of the
labyrinth, the number of cycles, and the length of the side leg (Tullis et al., 1995). An efficient design will result from the following procedure:
1. Knowing the design discharge and the maximum head, determine the required effective length from
2 = |QLV2^f3/2
(178)
where Q = 0.30 is derived from Fig. 17.23, with a = 8, and H/p = 0.8.
Next Page
CHAPTER 18
HYDRAULIC DESIGN OF
STILLING BASINS AND
ENERGY DISSIPATORS
18.1
INTRODUCTION
Stilling basins and energy dissipators are usually provided in conjunction with development of spillways, outlet works, and canal structures. It is often necessary to perform
hydraulic model studies of individual structures to be certain that these energy dissipating
devices will operate as anticipated. A relatively large volume of data is available from
many laboratory and field studies performed in the past (Blaisdell, 1948; Bowers and
Toso, 1988; Bowers and Tsai, 1969; Chadwick and Morfett, 1986; Chaudhry, 1993; Chow
1959; French, 1985; George, 1978; Hendreson, 1966; International Commission on Large
Dams (ICOLD), 1987; Novak et al., 1990; Peterka, 1964; Robertson et al., 1988; Senturk,
1994; Toso and Bowers, 1988; U.S. Bureau of Reclamation (USBR), 1974, 1987; Vischer
and Hager, 1995, 1998). Based on the results of many intensive studies, in 1958, A. J.
Peterka (1964) of the U.S. Bureau of Reclamation (USBR) published a summary report
of USBR'S studies entitled Hydraulic Design of Stilling Basins and Energy Dissipators,
Engineering Monograph No. 25. Since then, this publication has been referenced widely
within the hydraulic engineering community and still is one of the best references on this
subject available today. Energy dissipators are used to dissipate excess kinetic energy possessed by flowing water. An effective energy dissipator must be able to retard the flow of
fast moving water without damage to the structure or to the channel below the structure.
Vischer (1995) classified various types of energy dissipators by their features as: (1) by
sudden expansions, (2) by abrupt deflections, (3) by counterflows, (4) by rough walls, (5)
by vortex devices, and (6) by spray inducing devices. The stilling basins and energy dissipators discussed in this chapter are related to energy dissipation by expansion and
deflection.
There are two basic types of energy dissipators. They are hydraulic jump-type dissipators and impact-type dissipators. The hydraulic jump type energy dissipators dissipate
excess energy through formation of highly turbulent rollers within the jump. The impacttype dissipators direct the water into an obstruction that diverts the flow in all directions
and generates high levels of turbulence and in this manner dissipates the energy in the
flow. In other cases, the flow is directed to plunge into a pool of water where the energy
Crest EL 5050
Min. T.W. E1481.4
)
KEYPLAN
EL. 456.0
**
WANAPUM SPILLWAY
(State of Washington, USA )
is diffused and dissipated. The impact-type energy dissipators include check drops and
vertical drops, baffled outlets, baffled aprons, and vertical stilling wells. Generally, the use
of an impact-type energy dissipator results in smaller and more economical structures.
18.2 STILLINGBASINS
Using six test flumes, USBR conducted model studies for five stilling basin designs. The
results are summarized and presented in the Engineering Monograph No. 25 mentioned
above (Peterka, 1964). In Basin I tests, all test flumes were used and the test data obtained
provides basic hydraulic information concerning hydraulic jumps on a horizontal apron.
The Type II basin was developed for high dam and earth dam spillways and large canal
structures where the approach velocity is high and the corresponding Froude number
exceeds 4.5. Type III stilling basin is suitable for general canal structures, small outlet
works, and small spillways where the approach velocity is moderate or low and does not
exceed 50-60 ft/s (15-18 m/s) and the unit discharge is less than 200 ftVs/ft (18 mVs). For
smaller canal structures, outlet works, and diversion dams where the approach Froude
number is relatively low (between 2.5 and 4.5) and the heads of the structures do not
exceed 50 ft (15 m), Type IV stilling basin may be used. However, the jumps in the basin
may be unstable and alternative design such as the modified Type IV basin may be considered. To achieve greater structure economy for high dam spillways, Type V stilling
basin with sloping apron may be considered. Photos of several stilling basins in operation
are given in Exhibits 17.2, 17.4, 18.1, and 18.2.
18.2.1 General Hydraulic Jump Basin (Basin I)
The basic elements and characteristics of a hydraulic jump on horizontal aprons (Figs.
18.1 and 18.2) is provided to aid designers in selecting more practical basins such as
Basins II, III, IV, V, and VI. Jump occurs on a flat floor with no chute blocks, baffled piers
or end sill in the basin. Usually, it is not a practical basin because of its excessive length.
For a high-velocity flow down a spillway chute with known terminal velocity (Fig. 18.1)
and depth entering the basin, the required tail water depth, the length of jump, and loss of
energy can be determined based on the curves provided in Fig. I8.2a-e.
18.2.2 Stilling Basins for High Dam and Earth Dam Spillways
and Large Canal Structures (Basin II)
This stilling basin was developed for use on high spillways, large canal structures, and
so forth for approach Froude numbers above 4.5. With chute blocks and dentated end sill,
the jump and basin length can be reduced by about 33 percent. The basic design features
of Basin II stilling basin are given in Fig. 18.3a. For preliminary designs, the curves for
estimating required tailwater depth, and length of jump are given in Figures 18.3Z? and c.
The water surface and pressure profiles can be determined based on Fig. 18.3J and e.
The water surface profile in this basin can be closely approximated by a straight line
making an angle a(jump angle) with the horizontal. This line can also be considered as
a pressure profile. The USBR guidelines for designing this type of stilling basins are
given as follows:
7OThOiSt
Flow
Crest El. 385.0.
Intermediate
. raining wall
KEYPLAN
End training wall
MAYFIELD SPILLWAY
(State of Washington, USA ),
PROTOTYPE TESTS
X Shasta Dam
o Grand Coulee Dam
VA (actual)
VT (theoretical)
FIGURE 18.1 Curves for velocity entering stilling basins from 0.8:1 to 0.6:1 steep slopes.(From
Peterka, 1964)
F.-2.5-4.S, TRANSITION
FIGURE 18.2 Basic hydraulic jump basins on horizontal aprons. (Basin I) (From Peterka, 1964)
OC DEGREES
Slooe 2:l
FROUOE NUMBER
FROUOE NUMBER
MINIMUM TW DEPTHS
(SWEEPOUT)
FROUOE NUMBER
LENGTH OF JUMP
FIGURE 18.3 Stilling basins for high dam and earth dam spillways and large canal structures. (Basin
H)(From Peterka, 1964)
1. Determine velocity V1 of flow entering the jump. Figure 18.1 may be used. This
chart represents a composite of experience, computation, and a limited amount of
experimental information obtained from prototype tests on Shasta and Grand Coulee
Dams. The chart provides a fair degree of accuracy for chute having slopes of 0.8:1
or steeper, where computation is a difficult and arduous procedure. The asymptotic nature of the terminal velocity curves is also depicted in Fig. 18.1. For a constant
head of 2.5 ft (0.8m) on the spillway crest, the terminal velocity does not increase
significantly (from 51 ft/s or 15.5 m/s to 53 ft/s or 16.2 m/s) as the vertical distance
(fall) from the reservoir level to stilling basin floor increases from 200 to 600 ft
(61-183 m).
2. Set apron elevation to utilize full conjugate tail water depth. Add a factor of safety
if needed. A minimum margin of safety of 5 percent of tailwater depth (D2) is recommended.
3. Exercise caution with effectiveness of the basin at lower values of the Froude number
(V1 / (^D1)172) of 4 or lower. D1 is the depth of the flow entering the basin.
4. Determine the length of basin using the curve shown in Fig. 18.3c.
5. Use the depth of flow entering the basin, D1 as the height of chute blocks. The
width and spacing should be equal to approximately D1 but can be varied to avoid
fractional blocks. A space equal to D/2 is preferable along each side of wall to
reduce spray and maintain desirable pressures.
6. As shown in Fig. 18.3a, set the height of the dentated sill equal to 0.2D2 and the
maximum spacing approximately 0.15D2. For narrow basins, the width and spacing may be reduced but they should remain equal.
7. It is not necessary to stagger the chute blocks with respect to the sill dentates.
8. It is recommended that the sharp intersection between chute and basin apron be
replace with a curve of reasonable radius of at least 4D1 when the chute slope is 1:1
or greater. Chute blocks can be incorporated on the curve surface as readily as on
the plane surfaces. The chute slope (0.6:1-2:1) does not have significant effect on
the stilling basin action unless it is nearly horizontal.
Following the above rules should result in a safe, conservative stilling basin for
spillways up to 200 ft (60 ms) high and for flows up to about 500 (ftVs/ft) [46.5
(mVs/m)] basin width, provided that jet entering the basin is reasonably uniform
both as to velocity and depth. For greater falls, larger unit discharges, or possible
asymmetry, a model study of the specific design is recommended.
18.2.3 Short Stilling Basins for Canal Structures, Small Outlet Works,
and Small Spillways [Basin III and the St. Anthony Falls (SAF) Basin]
For structures carrying relatively small discharges at moderate velocities, a shorter
basin having a simpler end sill may be used if baffled piers are placed downstream from
the chute blocks (Fig. 18.4). In this section, stilling basins for smaller structures in
which velocity at the entrance to the basin are moderate or low ( up to 50-60 ft/s or
15-18 m/s) and discharges of up to 200 ft3/s/ft or 18 m3/s/m are discussed. The stilling
basin action is very stable for this design. It has a large factor of safety against sweepout of the jump and operates equally well for all values of the Froude number above 4.0.
Slope 2:1
FROUDE NUMBER
MINIMUM TW DEPTHS
(SWEEPOUT)
FROUOE NUMBER
HEIGHT OF BAFFLE BLOCKS
AND END SILLS
Profile for
conjugate depth
FROUOE NUMBER
LENGTH OF JUMP
FIGURE 18.4 Short stilling basins for canal structures, small outlet works and small spillways.(Basin
III) (From Peterka, 1964)
This basin should not be used for velocities above 50 ft/s or 15 m/s to avoid potential cavitation damages aginst baffle piers. Instead, Basin II type stilling basin should be considered or hydraulic model studies should be performed. The following USBR guidelines
pertain to the design of the Basin III type stilling basin:
1. The stilling basin operates best at full conjugate tail water depth, D2. A reasonable
factor of safety is inherent in the conjugate depth for all values of the Froude number and it is recommended that this margin of safety not be reduced.
2. Determine the length of basin using the design curve given in Fig. ISAc. It is less
than one-half the length of the natural jump. It should be noted that an excess of
tail water depth does not substitute for basin length or vice versa.
3. Exercise caution with effectiveness of the basin at lower values of the approach
Froude number [V11 CgD1)172] of 4.5 or lower.
4. Height, width, and spacing of chute blocks should equal the average depth of flow
entering the basin, or D1. Width of blocks may be decreased, provide spacing is
reduced a like amount. Should D1 proved to be less than 8 in or 20 cm, the blocks
should be made 8 in or 20 cm high.
5. The height of the baffle piers (Fig. 18.4a) varies with the Froude number and is
given in Fig. 18.4d. In narrow structures, block width and spacing may be reduced,
provided both are reduced a like amount. A half space is recommended adjacent to
the walls.
6. The upstream face of the baffle piers should be set at a distance of 0.8D2 from the
downstream face of the chute blocks. This dimension is important.
7. The height of the solid end sill is given in Fig. 18.4d. The slope is 2:1 upward in
the direction of flow.
8. It is undesirable to round or streamline the edges of the chute blocks, end sill, or
baffle piers. It reduces the effectiveness of the energy dissipation. However, small
chamfers on the block edges to prevent chipping of the edges and to reduce cavitation erosion may be used.
9. It is recommended that a radius of reasonable length greater than 4D1 be used at the
intersection of the chute and basin apron for slopes of 45 or greater.
10. As a general rule, the slope of the chute has little effect on the stilling basin action
unless long flat slopes are involved.
11. Experience indicates that the Type III basin works well for flow less than 200
ft3/s/ft or 18 m3/s/m based on basin width and approach velocity at the entrance of
up to 50-60 ft/s or 15-18 m/s.
The St. Anthony Falls (SAF) Hydraulic Laboratory of the University of Minnesota had
also developed a similar basin for small spillways, outlet works, and small canal structures
for approach Froude numbers ranging from 1.7 to 17 (Blaisdell, 1948; Chow, 1959). This
basin was developed to achieve about 70 to 90 percent reduction of the jump lengths.
This basin is commonly known as the SAF stilling basin (Fig. 18.5). Since the basin is
relatively short so that a significant amount of residual energy can still exist downstream
from the end sill, the channel reach downstream from the stilling basin should be allowed
to erode until a stable scour depth is reached. Otherwise riprap protection should be provided to minimize scour (Sec. 18.7). The guidelines for designing this basin are summarized as follows (Blaisdell, 1948; Chow, 1959):
O0Io 90
45preferred
ODfM
IT)
in
Cl
VH
Il
?*
5-3O
9
<*
VMCM
CD
O
^T
O
Chute blocks
Floor blocks
End sill
CENTERLINE SECTION
Vories
Cut-off woll
Side woll
Wing woll
Top slope is1M
'
( U ~ "S)V2
y'2 = 0.8Sy2
y 2 = fl-00 - -^-I2
^
oUUJ
10. The top of the side-wall above the maximum tailwater level to be expected during
the life of the structure is given by z = J2I^11. Wing-walls should be equal in height to the stilling basin side-walls. The top of the
wing-wall should have a slope of IH:IV.
12. The wing-wall should be placed at an angle of 45 to the outlet center line.
13. The stilling basin side-walls may be parallel for a rectangular stilling basin or they
may diverge as an extension of the transition side-walls for a trapezoidal stilling
basin as shown in Fig. 18.5.
14. A cutoff wall of nominal depth should be used at the end of the stilling basin.
15. The effect of entrained air should be neglected in the design of the stilling basin.
18.2.4 Low Froude Number Stilling Basins
(Basin IV and Modified Basin IV)
This stilling basin was developed for canal structures, outlet works, and diversion dams
where the approach Froude number of the basin is relatively low (between 2.5 and 4.5)
Slope 2 - i
FIGURE 18.6 Low Froude number (2.5-4.5) stilling basin design (Basin
IV).(From Peterka, 1964)
and the heads of the structures are about 50 ft (or 15 m). In this case, the jump is not fully
developed and unstable and the methods of design discussed previously do not apply.
Alternative design and/or wave suppressors or Basin VI type stilling basin with a hanging baffle for energy dissipation may be considered. Guidelines for developing
a low Froude number stilling basin (Basin IV) as depicted in Fig. 18.6 are given as
follows (Peterka, 1964):
1. A model study of the stilling basin is imperative.
2. Reduction of excessive waves created in the unstable jump is the main problem
concerning the design of the stilling basin.
3. A tailwater depth of 10 percent greater than the conjugate depth is strongly recommended.
4. Place as few appurtenances as possible in the path of the flow, as volume occupied
by appurtenances helps to create a backwater problem, thus requiring higher
training walls.
5. Use Fig. 18.6 to develop the design of the stilling basin. The number of deflector
blocks shown in the figure is a minimum requirement.
6. The length of basin can be obtained from Fig. 18.2c. No baffle piers are needed in
the basin.
7. The recommended maximum width of the deflector blocks is equal to D1 but
0.75D1 is preferable from a hydraulic standpoint. The ratio of block width to spacing should be maintained as 1:2.5.
8. The extreme tops of the deflector blocks are 2D1 above the floor of the stilling
basin.
9. To accommodate the various slopes of chutes and ogee shapes encountered, the
horizontal top length of the deflector blocks should be at least 2D1. The upper surface of each block is sloped at 5 in a downstream direction for better operation
especially at lower discharges.
10. The addition of a small triangular sill placed at the end of the apron for scour control is desirable. An end sill of the type developed for short stilling basins (Basin
III) can be used. The slope of the upstream face of the sill is 2:1 and the height of
the sill can be determined based on Fig. 18.4J.
11. Basin IV stilling basin is applicable to rectangular cross sections only to minimize
potential wave-related problems.
Type IV stilling basin performs effectively in dissipating the energy at low Froude
number flows for small canals and for structures with small unit discharges. It is also
effective in minimizing wave problems. Based on additional model tests, the U.S. Bureau
of Reclamation (USBR) has developed a modified stilling basin for low Froude number
approach flows (George, 1978). This stilling basin is suitable for approach flows with
Froude numbers ranging from 2.5 to 5.0. The basin is relatively short and is provided with
chute blocks, baffle piers, and a dentated end sill as shown in Fig. 18.7a. The guidelines
for designing Modified Type IV stilling basin are given as follows (George, 1978):
PLAN
Baffle piers
Chute blocks
Chute blocks
Baffle piers
2:1 Slope
Toe of chute
ELEVATION
FIGURE 18.7(a) Low Froude number stilling basin (Modified Basin IV).(From George, 1978)
- Test Data
o -Natural Basin
3.
4.
5.
6.
7.
8.
9.
10.
FIGURE 18.8 Stilling basins with sloping aprons. (From Peterka, 1964)
conjunction with medium or high overfall spillways where the rock foundation is in fairly good condition. When shorter aprons are used, the riverbed downstream must act as
part of the stilling basin. On the other hand, when the quality of foundation material is
questionable, it is desirable to make the apron sufficiently long to confine the entire jump.
The total apron length may range from about 40 to 80 percent of the length of jump.
The hydraulic jump may occur in several ways on a sloping apron, as depicted in
Fig. 18.8. The jump may have its toe form on the slope and the jump itself ends over the
horizontal apron (Case B), or ends at the junction of the slope and the horizontal apron
(Case C), or the entire jumps forms on the slope (Case D). For practical purposes the action
in Cases C and D is the same. Guidelines for the design of sloping aprons are given below:
1. Determine an apron arrangement that will give the best economy for the maximum
discharge condition.
2. The first consideration should be to determine the apron slope that will require the
minimum amount of excavation, the minimum amount of concrete, or both, for the
maximum discharge and tailwater condition.
3. Position the slope so that the front of the jump will form at the upstream end of the
slope for the maximum discharge and tailwater condition (Fig. 18.9). It may be necessary to raise or lower the apron, or change the slope entirely. Data obtained from
13 existing spillways are also shown in Fig. 18.9. Each point in the figure has been
connected with an arrow to the tan(6) curve corresponding to the apron slope. The
adequacy of the tailwater depth of these spillways can then be evaluated.
4. Use Fig. 18.10 to determine the length of the jump for maximum or other
flows. Shorter basins may be used where a solid bed exists. For most installations,
an apron length of about 60 percent of the length of jump for the maximum discharge condition should be sufficient. Longer basins are needed only when the
downstream riverbed is in very poor condition.
5. Ascertain that the tail water and length of basin available for energy dissipation
are sufficient for, say 1/4, 1/2, and 3/4 capacity. If the tailwater depth is deficient,
a different slope or a new position of the sloping portion of the apron should be
considered.
6. Horizontal and sloping aprons will perform equally well for high values of the
Froude number if the proper tail water depth is provided.
7. The slope of the chute upstream from a stilling basin has no significant effect on
the hydraulic jump when the velocity distribution and depth of flow are reasonably
uniform on entering the jump.
8. A small solid triangular sill should be provided at the end of the apron to lift the
flow as it leaves the apron for scour protection. The most effective height is
between 0.05D2 and 0.1OD2 and a slope of 3:1-2:1. Several existing stilling basins
with sloping aprons are shown in Fig. 18.11. All stilling basins shown were
designed with the aid of model studies.
9. The stilling basin should be designed to operate with as nearly symmetrical flow in
the stilling basin as possible to avoid formation of large circulating eddies and
transport of riverbed material into the apron area, and the potential undermining of
the wing walls and riprap.
10. A model study is advisable where the discharge over high spillways exceeds 500
ft3/s/ft or 46.5 m3/s/m based on the apron width, where there is any form of asymmetry involved, and for the high values of the Froude number where stilling basins
become more costly and the performance becomes less acceptable.
FIGURE 18.9 Comparison of existing sloping apron designs with experimental results.
(From Peterka, 1964)
SYMBOL
TAN$
0.050 - .067
0.100
0.135
0.150-.164
0.174
0.185
0.200-.218
0.263 - .280
SOURCE
Flumes A, B and F
Flumes A, B, D and F
Flumes A,
Flumes A, B and F
Flumes E
Flumes A
Flumes A, B and F
Flumes A and F
FIGURE 18.10 Jump length in terms of conjugate depth, D2 for stilling basins with sloping aprons.(From Peterka, 1964)
NORRlS
SHASTA
CANYON FERRY
BHAKRA
( PMELl
MlNAMV )
MAOOEN
BHAKRA
( FINAL )
FIGURE 18.11(a) Existing stilling basins with sloping aprons.(From Peterka, 1964)
FOLSOM
CAPILANO
FRIANT
DICKINSON
FIGURE 18.11(b) Existing stilling basins with sloping aprons.(From Peterka, 1964)
OLYMPUS
RIHAND
KESWICK
at a point on the apron surface, a dangerous local instability may develop with respect to
the uplift pressure at the bottom of the concrete slab. Some projects have experienced high
uplift pressures under large areas of the basin floor and resulted in complete floor concrete
slabs being torn up (Bowers and Toso, 1988; ICOLD, 1987; Toso and Bowers, 1988). In
addition, cavitation, abrasion, and vibration due to intense turbulence and pressure fluctuations can also contribute significantly to the damage of a stilling basin.
Based on the model studies of USBR Type II and Type III stilling basins, Toso and
Bowers (1988) obtained the following useful conclusions:
1. The pressure fluctuations in the jump tend to approach a definite limit, on the order of
80 to 100 percent of the approach velocity head. This is on the order of 10-20 times
the root-mean square (rms) of the pressure fluctuation.
2. Addition of chute blocks, intermediate blocks, and end sills did not result in significantly higher maximum negative and positive deviations than those for basins without
blocks and sills. The energy dissipation was quicker.
3. Side-wall pressure fluctuations are very significant, and peak at one to two inflow
depths above the floor.
4. The longitudinal extent of extreme pressure pulsation in the zone of maximum turbulence
is on the order of eight times the inlet flow depth. The lateral extent of a characteristic
pulse is approximately 1.6 times the longitudinal extent or 13 times the inlet flow-depth.
ICOLD (1987) recommended, as a minimum precaution, that the following two conditions be considered when designing the stilling basin apron:
1. Full downstream uplift pressure applied over the entire area of the floor with basin empty.
2. Full uplift pressure equals 12 percent of the approach velocity head applied under the
whole basin, with the basin full.
If necessary, the basin floor can be strengthened by providing anchors or using thicker slabs which may be held in place by the side walls.
ICOLD (1987) also recommended following structural arrangements to minimize
potential uplift damages due to undesirable turbulent flow induced pressure fluctuations.
1. All contraction joints should be fitted with properly located and embedded seals.
2. There should be no drain openings in the training wall inside the basin. However, drain
outlets in a dentated sill at the beginning of a stilling basin have performed satisfactorily.
3. Keep the areas of the floor slabs as large as possible.
4. Connect slabs by means of dowels, shear keys, and reinforcement across the joints.
5. Keep horizontal construction joints to a minimum, with dowels across them.
6. If drainage is necessary, keep it well away (1-1.5 m at least) from the wetted surfaces
so that abrasion or cavitation erosion will not make it accessible to the turbulent flow.
18.3
For small drops in canals with values of the Froude number between 2.5 and 4.5, a
drop-type energy dissipator which is in the form of a grating is particularly applicable to
reduce wave actions and dissipating energy. The device causes the over falling water jet
to separate into a number of long, thin sheets of water which falls nearly vertical into the
canal below. It has excellent capability in dissipating energy and eliminating wave problems. Guidelines for developing a drop-type energy dissipator are given as follows
(Peterka, 1964):
Next Page
CHAPTER 19
FLOODPLAIN
HYDRAULICS
Roy D. Dodson
Dodson & Associates, Inc.
Houston, Texas.
This chapter deals with the practical considerations involved in identifying the floodplain for
a stream channel, and analyzing the flow characteristics of the channel system, including its
floodplain. In general, any land area that is susceptible to inundation by rising or flowing
flood waters from any source could be considered to be a floodplain area. This would include
areas affected by coastal or lacustrine (lake) flooding. However, the emphasis of this chapter
will be on riverine floodplains, which are affected by flood waters from a stream or river.
The determination of flood elevations and floodplain boundaries along stream channels has become increasingly important as more development has occurred in floodplain
areas, or in areas that have been reclaimed from the floodplain.
A floodplain analysis requires a large amount of data in order to be accurate and
complete. Cross sections of the stream channel and detailed geometric descriptions of
bridges and other structures are required. In addition, experience is usually required to
accurately assess the roughness characteristics of a channel, to properly lay out cross
sections, and to adequately address many other aspects of a good flood plain analysis.
Sophisticated computer programs are available to analyze the data and produce detailed
computations of water surface elevations, floodplain boundaries, and other results.
However, these programs can produce misleading results without a properly planned
and executed analysis.
Because of the subjective judgments required for some of the input data and the uncertainty associated with certain values such as flow rates, some people have the attitude that
"there is no right answer" to the question of determining floodplain elevations and boundaries. However, it is important not to allow this to degenerate into the attitude that "one
answer is as good as another." This is clearly not true. A floodplain study performed well
produces superior results.
This chapter describes an approach toward floodplain studies that emphasizes quality
control and thoroughness in all aspects of the study effort. The effort is focused on those
aspects of the study that will produce the best overall results within given constraints of
time and budget. This chapter provides guidance in the following areas:
Floodplain
Flood
Historic Floods
Streamflow Data
Meteorologic Data
Hydrologic Data
study data typically available from each of the major federal government agencies.
In addition, local and state agencies often have study data available.
The types of information which may be derived from previous studies include
An assessment of the usability and technical accuracy should be made of all available
information, including historical hydrologic data, high-water marks, flooding problems
within the community, flood control measures, hydraulic structures that affect flooding,
available community maps showing and naming all roads in the floodplains, topographic
maps, digital data files, and elevation control data (including consideration of land subsidence* where applicable). Photographs of past major floods, if available, should be obtained.
Within the United States, federal Flood Insurance Studies are often valuable sources
of information on floodplain hydraulics. It is always advisable to obtain the data from previous Flood Insurance Studies whenever it is available. In fact, if a study is being prepared
for submittal to the Federal Emergency Management Agency (FEMA) to replace or revise
an existing Flood Insurance Study, the data from the existing Flood Insurance Study must
be used as much as possible.
Flood Insurance Study information may include hydrologic and hydraulic models, engineering and construction plans, floodplain maps, and flood profiles. In addition, any information should be obtained that may provide data for evaluating changes to the effective
hydrologic or hydraulic models. Rood Insurance Study data may be requested from FEMA.
If the original hydraulic models and work maps used in the Flood Insurance Study are
not available, FEMA requires that the Flood Insurance Study results be recreated from the
information in the Flood Insurance Study report book and the Flood Insurance Rate Maps.
The data available from the published Flood Insurance Study report should be supplemented as needed and entered into the same hydrologic and/or hydraulic model used to
create the original Flood Insurance Study. This model should then be calibrated to the
Flood Insurance Study flood profiles to obtain 100-year water-surf ace elevations within
an 0.1-ft tolerance, if possible. FEMA requires this to ensure a logical transition between
revised and unrevised data.
19.1.1 Sources of Topographic Data
Topographic maps for the area should be obtained whenever available. Although local and
state governments sometimes compile detailed topographic maps, the most widely available topographic maps within the United States are the U.S. Geological Survey's (USGS)
7.5-minute or 15-minute quadrangle maps. USGS quadrangle maps may provide all the
following information:
*Land subsidence is the sinking or settling of land to a lower level in response to various factors, both natural and
of human origin, such as earth movements, lowering of fluid pressure (or lowering of ground water level), removal
of underlying supporting materials by mining or solution of solids, either artificially or from natural causes, compaction caused by wetting, oxidation of organic matter in soils, or added load on the land surface.
1. Survey control points. The locations of permanent vertical and horizontal control
points (monuments or benchmarks) are indicated on the maps.
2. Topography. Topography is prepared to national standards and the accuracy of each
map is keyed to the contour interval. The contour interval will vary depending on the
amount of relief at the individual locations.
3. Ground cover. The color coding indicates the general type of ground cover, which
may be useful in planning survey operations and in confirming estimates of roughness
coefficients.
4. Development features. The locations of road crossings, developments, and other features are indicated on the maps.
In the United States, the USGS is actively converting existing maps to digital form.
Planimetric information is represented using digital raster graphics (DRG), digital
orthophoto quads (DOQ), or digital line graphs (DLG).
A digital raster graphic is a carefully scanned image of a USGS topographic map,
including the borders. The map image is georeferenced to the surface of the Earth, so that
most geographic information system (GIS) software can automatically position the DRG
image correctly with respect to other types of geographic data. Figure 19.1 illustrates a
portion of a USGS DRG file.
A digital orthophoto quad is a digital image of an aerial photograph in which displacements caused by camera orientation and terrain have been removed. Orthophotos
combine the image characteristics of a photograph with the geometric qualities of a map.
The standard digital orthophoto produced by the USGS is a black-and-white or colorinfrared 1-m ground resolution quarter quadrangle (3.75-minute) image. Figure 19.2
illustrates a portion of a USGS DOQ File.
Digital line graphs are vector files containing line data, such as roads and streams, digitized from USGS topographic maps. DLGs offer a full range of attribute codes, are highly accurate, and are topologically structured, which makes them ideal for use in GIS.
A digital elevation model (DEM) is a digital file consisting of terrain elevations for
ground positions at regularly spaced horizontal intervals. DEMs are developed from
stereo models or digital contour line files derived from USGS topographic quadrangle
FIGURE 19.3 Example of elevation contours derived from USGS DEM file.
Canada and most Western European nations are undertaking similar programs for the
development of digital elevation data from existing topographic data. Digital data are also
being developed for other regions of the earth.
Digital terrain data have two major applications to floodplain hydraulics. First, if the
data are sufficiently detailed, they may be used as a basis for the channel and floodplain
cross sections for hydraulic studies. This is emerging as a common practice when digital
terrain models (DTMS) are available. However, the grid-based data available from a digital elevation model may not be suitable for use as channel cross section data, because the
spacing between adjacent points (at least about 30 m for most DEM data) is not dense
enough to provide a sufficiently detailed channel cross section. The use of DEM data for
extending field-surveyed cross sections into floodplain areas is more promising, because
less detail is generally needed in these areas.
Digital topographic data may also be used for floodplain mapping. The results of a
hydraulic model analysis only provides the floodplain boundaries at the locations where
cross sections are available. Between these cross sections, the floodplain must be mapped
using available topographic data. The availability of digital topographic data is dramatically improving the efficiency and accuracy of floodplain mapping, through links between
GIS software and hydraulic analysis software. Figure 19.4 illustrates a stream channel for
which two floodplains have been automatically calculated and displayed using software
that links GIS software with the U.S. Army Corps of Engineers Hydrologic Engineering
Center River Analysis System (HEC-RAS) water surface profiles computer program.
FIGURE 19.4 Example of floodplain computed and mapped using data from digital terrain
model (DTM).
Highway maps from state or local highway agencies or street maps from private map companies may contain information which may be missing from the USGS quadrangle maps,
including recently constructed road crossings or other structures. Like topographic data and
aerial photographs, highway and street data are also becoming available in digital form.
19.1.4 Construction Drawings
Construction drawings may be available for channel modifications, bridges, or other projects affecting the stream channel or floodplain. Local or state transportation agencies and
FIGURE 19.6 Example of digital street map (Census Bureau TIGER File).
railroad companies generally retain the construction drawings for all bridges and many
other types of projects.
As-built construction drawings are those that have been revised to reflect the actual
dimensions of a completed construction project. As-built drawings for existing bridges are
normally available from local or state transportation agencies. Local or federal agencies
normally have as-built construction drawings for channel modification projects.
The use of information from as-built construction drawings can greatly reduce the
amount of survey information required for existing bridges or channel modifications.
However, the survey benchmark used for each set of construction drawings should be
determined, to be sure that all information is based on the same horizontal and vertical datum.
19.1.5 Stream Gage Data
Stream gage data from stream gages operated by the USGS or other agencies can provide
good estimates of the floodplain elevation at a particular point. Many floodplain analysis
computer programs can accept a channel rating curve constructed from the stream gage
data for use in computing the water surface elevation corresponding to a given flow rate.
Figure 19.7 illustrates a channel rating curve.
If the purpose of the analysis is to model a very extreme event, it is important to note
that the rating curve available for the stream gage location may not have any historical
basis for the stages at extremely high flow rates. These stages may simply be extrapolated from the historical record of lower flow rates.
19.2
OBTAININGFIELDSURVEYDATAFOR
FLOODPLAIN STUDIES
Even though considerable information may be available from previous studies and other
sources, it is often necessary to perform field surveys to provide additional information.
Survey operations can be difficult to manage, but they often determine the success of a
floodplain study.
The following steps should be completed before field survey operations are initiated:
1. Channel stationing. The stream length should be accurately determined from the mouth
(or other beginning point). Normally, this is done using a recent aerial photograph or
other map which indicates the current channel morphology. Channel stationing should
begin with 0 + 00 and increase in the upstream direction. Stationing should be measured along the thalweg.*
*The thalweg is the line following the lowest part of a valley, whether under water or not. Usually this is the line
following the deepest part of the bed or channel of a river.
2. Locate structures. The stream stations of all pertinent stream crossings, tributaries, and
all other significant features should be summarized in table form, including those features found on only one map. The stream stations should be transferred to a topographic "work map" for the analysis.
3. Preliminary stream profile. If a stream profile showing the channel flow-line is not
already available from a previous analysis, it may be useful to prepare a preliminary
flowline profile from the contours on the topographic map.
4. Preliminary floodplain map. If a floodplain map is not already available from a previous analysis, the floodplain width at various points along the channel should be estimated. The estimate may be based on the readings of stream flow gages, preliminary
computations using data from the topographic work map, or observations by local residents. For major studies, it may even be advisable to set up a preliminary hydraulic
analysis using data from the topographic work map only.
5. Survey control points. The location and elevation of all known survey benchmarks in
the area of the study should be indicated on the topographic work map.
6. Site reconnaissance. The engineer should then visit the study area, using the work map
as a reference. Each road crossing and other structure should be inspected, photographed, and measured with a tape, if possible. Channel and floodplain conditions
should be noted. Special notes and photograph locations should be recorded on the
topographic work map.
7. Preliminary cross section locations. The proposed location and alignment of all surveyed cross sections should be indicated on the topographic work map.
Finally, the engineer should meet with the survey coordinator to discuss the data
requirements for the project, using the topographic work map, photographs, and the engineer's knowledge of the project area and computer program data requirements. Even after
the survey work begins, the engineer should maintain frequent contact with those performing the survey work to be aware of their progress and any special problems.
19.2.1 Vertical and Horizontal Control for Field Surveys
The Global Positioning System (GPS), when used in the differential mode, is currently
the best method for extending any survey control network unless satellite visibility is
obscured (e.g., because dense forest) or severe radiofrequency disturbances are present.
Differential GPS or third-order leveling* can be used to tie temporary benchmarks to an
established datum; to determine the elevation of high-water marks; and, where needed,
to establish horizontal and vertical control for aerial survey work. Whenever possible,
available benchmarks should be used instead of field surveys, to reduce the survey cost.
In particular, benchmarks shown on official floodplain maps should be used if possible.
As a general rule, there should be approximately two temporary benchmarks per mile
of stream length or four per square mile of floodplain, as appropriate (FEMA, 1995).
Most benchmarks established in North America are referenced to the National
Geodetic Vertical Datum of 1929 (NGVD29), which was formerly referred to as Mean
Sea Level of 1929. This datum was originally established as a vertical reference representing the average sea level as measured at a series of tide gages throughout North
America. As a result of additional work since 1929, a revised vertical reference has been
"Closures within 0.05 ft. X square root of distance in miles.
established which better represents the actual gravitation force of the earth as it exists in
various locations throughout the North American continent. This reference is called the
North American Vertical Datum of 1988 (NAVD88). FEMA and other federal government
agencies are adopting NAVD88 as the standard vertical datum for new studies. The significance of the conversion from NGVD29 to NAVD88 depends on the particular location. The National Geodetic Survey has produced a computer program called VERTCON
which will compute the conversion from NGVD29 to NAVD88 at any location in North
America, given the coordinates of latitude and longitude.
Field surveys should normally be accomplished by differential leveling or differential
GPS methods, with vertical error tolerances of 0.5 ft across the 100-year floodplain.
Horizontal control may be unnecessary since cross sections can be located simply by
visual reference to identifiable points on a map or aerial photograph. However, if crosssectional data will be combined with other data sets, it is necessary to establish and maintain adequate horizontal control.
19.2.2 Cross Section Locations
The survey information required for a floodplain analysis generally consists of cross
sections of the channel and floodplain along the stream. These cross sections may be
field-surveyed or taken from digital terrain models. Some water surface profile computer programs treat each cross-section as representing a reach of the river and use only
one section at the midpoint of the reach to calculate losses through the entire reach.*
However, the most commonly used computer programs use cross sections to define
breakpoints in the geometry, and properties of adjacent sections are averaged to calculate losses through the reach.
The objective of either computation scheme is to describe flow boundaries accurately
enough to predict energy losses due to friction and changes in flow velocity. If only a few
cross sections are available and they are located far apart, a greater amount of engineering judgment is required to satisfactorily analyze the problem. Any deviation from a
smooth profile must be explained, and in some cases it can be traced back to inadequate
cross-sectional data.
A water surface profile in nature is really a curvilinear surface that follows the general slope of the channel. When the water surface profile is computed, the shape of this
curvilinear surface is approximated by a series of straight-line segments. The endpoints of
these line segments are the channel cross section locations. If many channel cross sections
are included, the line segments will be short, and the true curvilinear shape of the natural
water surface profile can be better represented. If only a few channel cross sections are
used, the line segments will be long, and the true curvilinear shape of the natural water
surface profile will not be represented well. Figure 19.8 illustrates a water surface profile
that is well represented (using many channel cross sections) and the same profile that is
poorly represented (using only a few channel cross sections). As indicated, if an adequate
number of cross sections are not provided for the hydraulic analysis, the computed water
surface elevation can vary considerably from the actual value, especially in areas in which
there is a sudden change in the channel slope.
One study indicated that reach lengths should be limited to a maximum of 800 m
(about 0.5 mi) for wide floodplains and for slopes less than about 0.04 percent, 550 m
(about 1800 ft) for slopes equal to or less than 0.06 percent, and 350 m (about 1200 ft) for
*A reach is a length of a channel that is fairly uniform with respect to discharge, depth, area, and slope; more generally, any length of a river or drainage course.
Bends. Additional cross sections may be required in and around channel bends to properly represent the reach lengths in the chanwnel overbank areas.
If an elevated road or dam is found in the field that was not indicated on the cross section layout map, the survey field crew should insert additional cross sections to represent
these additional structures. However, this does not apply to roads that are not elevated and
will not block large amounts of flow.
After cross sections are placed at all of the locations identified above, then additional
cross sections may be placed between these as needed to reduce the spacing between adjacent cross sections to the values recommended above. Laying out cross sections in this
way, rather than at a fixed spacing, generally provides a superior analysis.
19.2.3 Cross Section Alignment and Orientation
Cross sections should tie into high ground, so that the maximum elevation of each end of
a cross section should be at least 1 ft higher than the anticipated maximum water surface
elevation, if possible. In addition, the cross section locations should be accessible and
practical. For example, pipeline and power line crossings make good locations for surveyed cross sections in wooded areas, because the right-of-way is usually cleared.
However, cross sections must represent the average ground profile. Roadside ditches,
washouts, gravel pits, raised areas, and other nonrepresentative conditions should be
avoided. The actual orientation and extent of the surveyed cross section should be indicated on work maps.
Most computer programs used for floodplain analysis require that cross sections
should be surveyed from left to right, with these directions determined while looking
FIGURE 19.9 Example of work map showing cross section locations for field survey.
downstream. The ground elevation as well as the distance from the channel flowline
should be determined and recorded for each point on the cross section.
Negative cross section stationing may be unacceptable to the computer programs
used for floodplain analysis. Therefore, it is advisable to begin cross section stationing
at 10 + OO or 100 + OO so that the cross section can be extended later without having
to restation the entire cross section.
Cross sections should be placed perpendicular to the direction of flow. Cross sections
can and should have angles or "doglegs" in them as needed so that the channel portion of
the cross section can be perpendicular to the flow in the channel while the floodplain portions of the cross section can be perpendicular to the flow in the floodplain. Figure 19.10
provides an example.
In the case of wide floodplains in very flat areas, where flows exceed the channel
capacity and spill into the floodplains, the effective cross section is not easily defined. One
way to solve this problem is to shape the cross section alignment so that the floodplain
portions curve upstream. This curvature will contain the water in a reasonable width since
the cross section elevations on the edge of the flow will be higher than in the case of the
straight cross section alignment.
A minimum of five points is usually required for the channel portion of a surveyed crosssection. This includes one point at the top of each channel bank, one point at the toe of each
side slope, and one point at the channel flowline, as illustrated in Fig. 19.11. Additional
points may be required when discontinuities in channel cross sections are encountered.
The number of cross section points required for floodplain areas is dependent on the
width of the cross section and on the character of the terrain in the floodplain. As a general rule, enough points should be shot to give a true representation of the floodplain terrain and to define any breaks in topography.
Valley cross sections that are intended to show the typical floodplain may be relocated in the field by the survey party to avoid heavy brush or to take advantage of power line
or pipeline easements, fence lines, or pasture lands. Any cross sections that are relocated
must be clearly marked on the work maps.
The channel station at which the cross-section intersects the channel center-line should
be determined for each cross section. The stream stations determined using aerial pho-
FIGURE 19.10 Cross-sectional layout and reach lengths for meandering channel.
Channel Invert
Left Toe of Slope
_
(QLOB) (XLOB) + (QCH) (XLCH) + (QROB) (XLOBR)
L=
Q
(19.1)
where L = the weighted average reach length QLOB = the total flow in the left floodplain XLOB = the specified reach length for the left floodplain QCH = the total flow in
the channel XLCH = the specified reach length for the channel QROB = the total flow in
the right floodplain XLOBR = the specified reach length for the right floodplain Q = the
total flow in the entire cross section.
19.2.4
Field survey costs for a flood plain study may be reduced in several ways. For example,
the spacing of cross sections along the stream channel could be increased so as to reduce
the number of cross sections required. As an alternative, the width of each cross section
could be reduced.
Studies have indicated that one of the best ways to improve the accuracy .of a computed water surface profile is to provide more cross sections (HEC, 1986). This implies
that one of the worst ways to save money on a floodplain study is to reduce the number
of cross sections. In general, it is better to reduce the number of survey points within each
cross section rather than to eliminate some cross sections entirely and thus extend the
reach lengths between adjacent cross sections.
Several methods may be used to maintain the required density of cross sections along
the stream channel. For example, surveyed cross sections could be used more than once
in the analysis. These repeated cross sections may be acceptable if field observations indicate that the channel and floodplain conditions are highly uniform through the stream
reach represented by the repeated cross sections. Intermediate cross sections may also be
synthesized (interpolated) from field-survey cross sections. Interpolated cross sections are
described later in this chapter. During the data collection phase of the project, however,
the most common method of reducing the cost of obtaining cross section data is to obtain
all or part of the data points for the cross-section from aerial topography or contour maps.
In some circumstances, aerial topography or contour maps can be used for the entire
cross section, including the channel portion. Usually, however, the accuracy and level of
detail available from aerial topography or contour maps are not sufficient for channel data.
It is preferable to obtain at least the channel portion of the cross section by field survey
methods, if at all possible. Sometimes, with little or no increase in field survey costs, some
additional ground shots can also be made in the floodplain, using the same setup of the
survey instrument already used for the channel survey. FEMA generally recommends the
use of aerial topography or contour maps for obtaining cross sections for the portion of
the cross section between the 100- and 500-year floodplain boundaries (FEMA, 1995).
Aerial topography is very cost-effective if the size of the project justifies the fixed
costs of mobilizing the aircraft and other related activities. Usually, the mapping data
obtained by aerial photography are useful for many purposes other than the floodplain
analysis, and therefore the full cost may be shared with other projects, departments, or
even other organizational entities. Newer technologies such as LIDAR (Light Imaging
raDAR) may also be very cost-effective,
If a contour map meets commonly accepted accuracy standards, then the contour accuracy is slightly better than half of the contour interval while the accuracy of spot elevations is approximately 20 to 30 percent of the contour interval of the map being prepared.
Therefore, cross sections developed using contours from topographic maps will not be as
accurate as cross sections prepared from spot elevations. Whenever possible, spot elevations along the cross section should be determined directly using the source data set, rather
than indirectly using contours.
19.2.5
Necessary dimensions and elevations of all hydraulic structures and underwater sections
along the streams should be obtained from available sources or by field survey where necessary. Dimensions and elevations of hydraulic structures should not be established by
aerial survey methods.
Most computer programs used for floodplain analysis require four user-defined cross
sections in the computations of energy losses due to the structure. These cross sections
may be identified by number (1 through 4), beginning downstream of the bridge. A plan
view of the basic cross section layout is shown in Fig. 19.13.
Cross section 1 is located sufficiently downstream from the structure so that the
flow is not affected by the structure (i.e., the flow has fully expanded from the narrow
bridge opening). This distance should generally be determined by field investigation
during high flows. If field investigation is not possible, there are two sets of criteria for
locating the downstream section. The USGS has established a criterion for locating
cross section 1 a distance downstream from the bridge equal to one times the bridge
opening width (the distance between points B and C on Fig. 19.13). Traditionally, the
Corps of Engineers criterion has been to locate the downstream cross section about
four times the average length of the side constriction caused by the structure abutments
(the average of the distance from A to B and C to D on Fig. 19.13). The expansion distance will vary depending on the degree of constriction, the shape of the constriction,
the magnitude of the flow, and the velocity of the flow. (A constriction is a local
obstruction narrowing a waterway.) Both criteria should be used as rough guidance for
placing cross section 1. Cross sections 1 and 2 should be close enough to one another
so that friction losses can be adequately modeled. If the expansion reach requires a
long distance, intermediate cross sections should be placed within the expansion reach
to adequately model friction losses.
Cross section 2 is located immediately downstream from the bridge (within a few feet).
This cross section should represent the effective flow area of the natural channel just
below the roadway crossing. It should not include the road fills, road ditches, mounds, or
depressions that are not typical of the floodplain.
Cross section 3 should be located just upstream from the bridge. The distance
between cross section 3 and the bridge should be relatively short. This distance should
only reflect the length required for the abrupt acceleration and contraction of the flow
that occurs in the immediate area of the opening. Cross section 3 should represent the
effective flow area of the natural channel just above the roadway crossing. Like cross
section 2, it should not include the road fills, road ditches, mounds, or depressions that
are not typical of the floodplain.
Cross section 4 is an upstream cross section where the flow lines are approximately parallel and the cross section is fully effective. The USGS recommends that the distance between cross section 3 and 4 should be roughly the same as the average width
of the bridge opening. However, this criteria for locating the upstream cross section
may result in too short a reach length for situations where the width of the bridge opening is very small in comparison to the floodplain. The Corps of Engineers generally
recommends that cross-section 4 be located a distance upstream equal to the average
contraction width (the average of the distance from A to B and C to D on Fig. 19.13).
According to the Corps of Engineers, the distance between cross sections 3 and 4
should be less than the distance between cross sections 1 and 2 for constricted flow
conditions, because flow contractions can occur over a shorter distance than flow
expansions. Both of these recommendations should be considered; the Corps of
Engineers recommendation is probably better when the bridge opening is narrow compared with the width of the floodplain; the USGS recommendation is probably acceptable for all other cases.
19.2.6
The engineer directing the floodplain analysis should determine, during a preliminary
visit to the study area, if the channel is relatively uniform in the area of each road crossing structure. If the channel is highly uniform in the area of a particular road crossing, it
may be possible to survey a single channel and floodplain cross section at the upstream or
downstream face of the bridge, and repeat this cross section for the other three or more
cross sections in the bridge model.
If the channel is relatively uniform through the bridge itself, but changes upstream or
downstream of the structure, a single field surveyed cross section may be used to represent the channel at both faces of the bridge, but different field-surveyed cross sections
may be needed at other locations through the bridge model.
If the channel is not uniform through the bridge, separate cross sections are required
at the downstream face of the bridge and at the upstream face of the bridge. It may even
be necessary to survey a cross section of the channel at the centerline of the bridge
(underneath the bridge superstructure), where special conditions cause the channel
underneath the bridge to differ significantly from the cross section at the upstream or
downstream face.
19.2.7 Obtaining Bridge Survey Data
Plan and profile sketches of the bridge on which all pertinent data are illustrated and identified should always be included in the survey notes. Figure 19.14 is a representative
sketch of the type of data that will be required by the hydraulic engineer.
Several pieces of information about the actual bridge itself are required. Two of the
most important of these are top of road and low chord elevations.
The top of road is the highest part of the roadway that forms a significant obstruction
to flow over the roadway. The top of road can be the crest of the bridge roadway, the top
of a sidewalk, or the crest of a solid bridge railing. When taking a cross-section on a multilane highway, the cross section should be obtained along the higher lane. If the median or
shoulder is higher than the centerline, the high points should be located by side shots on
the median or shoulder. Cross sections on railroads should be obtained on top of the rail.
If the track is superelevated, the highest rail should be shot.
The low chord (or low steel) of a bridge refers to the lowest part of the bridge that
forms a significant obstruction to flow under the bridge. The low chord of most bridges is
formed by the bottom of the bridge girders.
Whenever any doubt exists as to the identification of the top of road and low chord of
a bridge, elevations defining the shapes of all structures which form obstructions to flow
should be determined and recorded. For example, if a water main or other pipe is suspended from the bridge, this condition should be recorded.
(Bridge is Perpendicular
to Channel)
NoteThat Roadway
Embankments Are On
Fill Near the Bridge.
Natural ground elevations should be used instead of top of road elevations when the roadway is in a cut; that is, when natural ground is higher than the top of the roadway. In such
cases, it is necessary to visualize the total obstruction to flood flow at the road. This will be
a combination of the roadway and natural ground. Figure 19.15 illustrates this situation.
Bridge railings or curbs should sometimes be considered when defining the top of
roadway. If a railing or curb forms a substantial obstruction to flow over the bridge, the
top of the rail or curb should be considered as the effective top-of-road. Figures 19.15 and
19.16 illustrate bridge decks with solid and open rails.
Other important bridge data include the following:
Channel station: the channel station at which the bridge center-line (or one of the
bridge faces) intersects the channel flowline.
Natural Ground
Low
^hord
FIGURE 19.15 Defining the bridge deck for roadways in open cuts.
Solid Railing
Low Chord
FIGURE 19.16 Defining the bridge deck for bridges with solid rails.
Open Railing
Top of Curb
Actual Top of Roadway
Low Chord
FIGURE 19.17 Defining the bridge deck for bridges with open rails.
19.3 SELECTINGTHEBESTAPPROACHFORA
FLOODPLAINSTUDY
There are three alternative approaches that are generally available for the computation of
water surface elevations in a stream channel system:
One-dimensional steady, gradually varied flow conditions;
Two-dimensional, steady, gradually varied flow conditions;
One-dimensional unsteady flow conditions.
It is also possible to model two-dimensional unsteady-flow conditions, although this
approach is not common. The following sections describe these various approaches.
19.3.1 One-Dimensional and Two-Dimensional Flows
In a constructed channel with a regular cross section, most of the lines of flow will be
roughly parallel to the channel's longitudinal axis (the axis that follows the alignment of
the channel from upstream to downstream). This is one-dimensional flow. However, in
natural channels with irregular cross-sections, and especially when flow overtops the
channel and enters the floodplain, there may be significant components of flow in other
directions at certain locations. For example, flow may travel in a vertical direction for a
short distance to overtop a roadway surface and spill into a lower area on the downstream
side. Flow may also travel in a lateral direction from the floodplain into the channel to
pass through a narrow culvert opening.
Energy is required to set water in motion in any direction and overcome friction losses. The computer programs used for floodplain studies account for friction losses and
other changes in energy potential. In fact, the computation of water surface elevation is
really a by-product of the computation of the total energy level (often called the "energy
grade line") at each location in the stream channel and floodplain. For those situations
where energy losses are not the primary consideration, such as at a weir or a hydraulic
jump, the programs use alternate solutions. (A hydraulic jump is a sudden transition from
supercritical flow to the complementary subcritical flow, conserving momentum and dissipating energy.)
One-dimensional computer programs used for floodplain studies account for energy
losses in the downstream direction only (longitudinal flows). Two-dimensional computer
programs consider lateral flows as well as longitudinal flows. Therefore, they analyze
flow in both dimensions of the horizontal plane.
A one-dimensional analysis is adequate for most purposes. Two-dimensional computer models are useful at complex roadway crossings, in shallow flooding areas, and whenever there is significant flow in more than one direction.
19.3.2 Changes in Flow Depth With Respect to Time and Distance
Steady flow occurs whenever the depth of flow at a cross section, for a given discharge, is
constant with respect to time. Unsteady flow occurs whenever the depth of flow at a crosssection varies with respect to time, for a given discharge.
During a flood event, the flow depth in a stream channel increases from base flow conditions, reaches a peak value, and declines to the base flow again, generally within the
span of a few hours or less. However, at any one moment, an observer standing on the
edge of the floodplain would see what appears to be a steady flow condition.
Steady flow can further be classified as uniform or nonuniform. Uniform flow occurs
when the depth of flow and quantity of water are constant at every section of the channel
under consideration. Under these conditions, the water surface and flowlines will be parallel to the stream bed and a hydrostatic pressure condition will exist (the pressure at a
given section will vary linearly with depth). Uniform flow conditions are rarely attained
in natural channels.
Gradually varied flow occurs whenever the depth of water changes slowly over the
length of the channel. Under this condition, the streamlines of flow are practically parallel. Therefore, uniform flow principles can be used to analyze the flow conditions, even
though the flow is nonuniform. With rapidly varied flow conditions, there is a pronounced
curvature of the flow streamlines. The assumption of hydrostatic pressure distribution is
not valid. Therefore, uniform-flow principles are not adequate for the analysis of rapidly
varied flow.
Most computer programs used for floodplain analysis are based on the analysis of
steady, gradually varied flow conditions. Some programs have special capabilities to analyze rapidly varied flow conditions, such as hydraulic jumps.
A floodplain analysis that assumes steady flow is a "snapshot" of the channel system, usually at peak flow conditions. However, some channel flow situations are dominated by rapidly changing flow rates or other dynamic conditions. For example, a
floodplain area may have substantial floodplain storage capacity that affects the flow
rate in the stream channel; streams may experience a reversal of flow under certain
conditions; the stream may be subject to sudden changes in flow rate due to the opening or closing of gates or other similar structures; streams may discharge into marine
bodies that are affected by tidal movements; or the stream may be a part of a complex
system of pipes, channels, ponds, and reservoirs. All of these situations may produce
sudden changes in flow rate which are not adequately represented by the assumption
of steady flow. The capability to analyze unsteady flow conditions is sometimes
required.
19.3.3
Critical depth is an important hydraulic parameter because it is always a hydraulic control. (Critical depth occurs when the specific energy is a minimum for a given discharge.)
Hydraulic controls are points along the channel where the water level or depth of flow is
limited to a predetermined level or can be computed directly from the quantity of flow.
When the depth of flow is greater than critical depth, the velocity of flow is less than critical velocity for a given discharge and hence, the flow regime is subcritical. Conversely,
when the depth of flow is less than critical depth, the flow regime is supercritical. Flow
must pass through critical depth in going from a subcritical flow regime to a supercritical
flow regime. Typical locations of critical depth are at
(1). abrupt changes in channel slope when a flat (subcritical) slope is sharply increase
to a steep (supercritical) slope, (2). a channel constriction such as a culvert entrance
under some conditions, (3). the unsubmerged outlet of a culvert on subcritical slope,
discharging into a widechannel or with a free fall at the outlet, and (4). the crest of an
overflow dam or weir.
Critical depth for a given channel is dependent on the channel geometry and discharge
only, and is independent of channel slope and roughness.
19.3.4 Types of Stream Systems
Stream systems may be divided into three main categories according to the complexity of
the interconnections among the various flow paths within the stream system:
Simple channels. These are single-stream reaches with no tributary streams. Therefore,
there are no confluences or junctions. The flow rates may vary from one cross section
to the next along the length of the stream channel, because of local inflows or stream
channel losses. The direction of flow in the channel is known. However, the flow
regime within the channel is not known, and in fact, may vary along the length of the
channel (changing from subcritical to supercritical or from supercritical to subcritical).(Fig. 19.18).
Dendritic channel systems. These include tributary streams that combine at confluences or junctions to form major streams. This process may be repeated through several steps. Streams within a dendritic channel system are sometimes identified by their
"order." First-order streams are the smallest tributaries, while higher-order streams are
larger streams and rivers. In a dendritic system, each of the individual streams is analyzed as a simple channel (as described in the category above) (Fig. 19.19).
Network channel systems. These include all of the features of dendritic channel system. In addition, full network systems (sometimes called fully-looped systems) can
have flow splits, where flow in a single channel (or a group of channels) can be divided among two or more flow paths which may either recombine downstream or discharge at completely different destinations. In some network channel systems, the
direction of flow may be known in advance for each stream; for others, the direction
of flow may be determined during the analysis and may even change during the analysis. A full network model can also include storage areas that can either provide water
to, or divert water from, a channel. A full network model is required to properly deal
with storage areas, because they represent flow splits (since water can be diverted into
the storage area). See Fig. 19.20.
Spruce Creek
Upper Reach
Pottsville
Lower Reach
Upper Spruce
Tusseyville
Bear Run
Middle Spruce
Coburrr
Lower Spruce
FIGURE 19.20 Example of network channel system.
*Backwater is an unnaturally high stage in stream caused by obstruction or confirnement of flow, as by a dam, a
bridge, or a levee. Its measure is the excess of unnatural over natural stage, not the difference in stage upstream and
downstream from its cause.
for the analysis of roadways with multiple bridges and culverts. Variable channel roughness and variable reach length between adjacent cross sections can be accommodated.
HEC-2 (developed by the USAGE Hydrologic Engineering Center) was the predecessor
to HEC-RAS, and has some of the same capabilities (HEC, 199Oa). However, HEC-2 does
not accommodate multiple bridge or culvert openings in most situations, and is restricted to
either subcritical or supercritical flow computations in a single analysis. Because HEC-2
was widely used for almost 30 years, many existing floodplain studies were performed using
HEC-2. HEC-RAS will read HEC-2 input data files and produce very similar results under
most situations, but HEC-2 may still continue to be used for some time because of the need
to match existing study results. HEC-2 is generally restricted to simple channel systems, but
may be used for dendritic channel systems under limited circumstances.
WSPRO Water Surface Profile [developed by the U.S. Geological Survey for the
Federal Highway Administration (FHWA)] is a step backwater program for natural channels with an orientation to bridge constrictions (USGS, 1990). The water surface profile
computation model has been designed to provide a water surface profile for six major
types of open-channel flow situations: unconstricted flow, single opening bridge, bridge
opening(s) with spur dikes, single opening embankment overflow, multiple alternatives
for a single job, and multiple openings. WSPRO was developed for use primarily around
roadway crossings. The bridge analysis routines originally used in WSPRO have now
been incorporated into the HEC-RAS computer program.
WSP2 Water Surface Profile-2 (developed by the USDA Natural Resources
Conservation Service) is a step backwater program for natural channels [Natural Resources
Conservation Service (NRCS, 1993)]. WSP2 estimates head loss at restrictive sections,
including roadways with either bridge openings or culverts.
19.3.6 Two-Dimensional Water-Surface Computer Models
The most commonly used two-dimensional models include TABS-MD (developed by
ACE Waterways Experiment Station) and FESWMS-2DH Two-Dimensional Flow in a
Horizontal Plane computer program (developed by USGS).The TABS-MD
(Multidimensional) Numerical Modeling System is a collection of generalized computer
programs and utility codes, designed for studying multidimensional hydrodynamics in
rivers, reservoirs, bays, and estuaries. The primary computational program in the TABSMD system is the RMA-2 model [Waterways Experiment Station (WES) Hydraulic
Laboratory, 1997]. The TABS-MD system can be used to study project impacts on flows,
sedimentation, constituent transport, and salinity. FESWMS-2DH Finite Element Surface
Water Modeling System-Two-Dimensional Flow in a Horizontal Plane simulates steady
and unsteady flow and is useful for simulating two-dimensional flow at width constrictions and highway crossings of rivers and floodplains (FHWA, 1989). FESWMS-2DH is
based on the momentum balance approach, with consideration of shear stresses due to
friction between the moving water and the fixed boundaries.
19.3.7 One-Dimensional Unsteady-Flow Models
There are several computer programs available for simulating one-dimensional unsteadyflow conditions in stream channels. These include UNET, Storm Water Management
Model (SWMM), DAMBRK, and NETWORK.
UNET, One-Dimensional Unsteady Flow Through a Full Network of Open
Channels (supported and maintained by USAGE HEC is a one-dimensional unsteady
flow model that can simulate flow through a single channel, a dendritic system of channels, or a full network of open channels (HEC, 199Ob). In UNET, storage areas can be the
upstream or downstream boundaries for a river reach. In addition, the river can overflow
laterally into the storage areas over a gated spillway, weir, levee, through a culvert, or a
pumped diversion.
In addition to solving the one-dimensional unsteady flow equations in a network system, UNET provides the user with the ability to apply several external and internal
boundary conditions, including flow and stage hydrographs, gated and uncontrolled spillways, bridges, culverts, and levee systems. UNET can read channel cross sections that are
input in a modified HEC-2 format.
Future versions of the HEC-RAS computer program are expected to incorporate most
of the functions of the UNET computer program, and will likely provide access to these
functions through a user interface that is more convenient to use than the original UNET
user interface, which relies heavily on the HEC-DSS (Data Storage System) for data entry
and output.
Storm Water Management Model (SWMM) (developed by the U.S. Environmental
Protection Agency) is a comprehensive water quantity and quality simulation model
developed primarily for urban areas. SWMM is a complete hydrology and hydraulics
model that is most commonly applied to urban areas with closed storm sewer systems
(Huber and Dickinson, 1988). The model uses different modules (or "blocks") for different types of computations. SWMM is not widely used for floodplain studies on natural
stream channels. However, the Extran and Transport blocks of the SWMM program can
use natural channel cross section data in the same form as required by the HEC-2 computer program. The Transport block provides simple hydraulic computations, but the more
sophisticated Extran block can perform a dynamic backwater analysis (Roesner et al.,
1988). This can be useful when a sophisticated hydraulic analysis is required, particularly when the open channel is a part of an urban drainage network.
Dam Break Flood Forecasting Model (DAMBRK) (developed by the National
Weather Service) is an unsteady flow dynamic routing model which develops an outflow
discharge hydrograph due to spillway and/or dam failure flows (Fread, 1982). This hydrograph is routed through the downstream river valley. This can be a useful method of floodplain analysis for such events.
NETWORK: Enhanced Dynamic Wave Model (developed by the National Weather
Service) is an unsteady flow dynamic routing model for a single channel or network (dendritic and/or bifurcated) of channels for free surface or pressurized flow (Fread & Lewis,
1986, Fread & Lewis, 1988 and Fread, 1993).
19.3.8 Selecting a Computer Program for a
Floodplain Analysis
A computer program that is suitable for one-dimensional steady, gradually varied flow
conditions is adequate for most floodplain studies. The most widely used programs in this
category are the Corps of Engineers HEC-RAS and the older HEC-2 computer programs.
If conditions dictate, a two-dimensional analysis or a dynamic analysis (or, in some cases,
a two-dimensional dynamic analysis) may be required.
If a floodplain study is being done to revise a FEMA flood insurance study, FEMA normally requires that revised study be completed using the same computer program used for
the original study. However, a different computer program may be accepted if documentation can be submitted justifying why a change would be an improvement. It is important to discuss such issues with FEMA in advance (FEMA, 1995).
Next Page
CYLINDRICALQUADRANT
STRAIGHT LINE
WARPED
ABRUPT
WEDGE
FIGURE 20.1 Transition types. (From Corry et al., 1975)
The design procedure for abrupt expansions is as follows (Corry et al., 1975):
1. Determine V0 and ^0 at the culvert outlet using Fig. 20.4 for box culverts or Fig. 20.5
for circular culverts.
o
2. Compute Froude number, Fr = ^-
o/o
3. Determine optimum flare angle (9) using tan 0 = Fr /3 (Blaisdell and Donnelly, 1949).
If the selected wing wall flare (0W) is greater than 0, consider reducing 0W to 6.
4. Use Fig. 20.2 (for box culverts) or Fig. 20.3 (for pipe culverts) to determine the downstream average depth (yA) knowing Fr and the desired distance L (expressed in multiples of D, diameter).
CHAPTER 20
FLOW TRANSITIONS AND
ENERGY DISSIPATORS
FOR CULVERTS
AND
CHANNELS
Larry W. Mays
Department of Civil and Environmental Engineering
Arizona State University
Tempe, Arizona
20.1
FLOWTRANSITIONSFORCULVERTS
Flow transitions are changes in the cross section of an open channel over short distances.
They are designed to have a minimum amount of flow disturbance. Figure 20.1 illustrates
the various types of transitions; the two most common ones are the abrupt (headwall) and
the straight line (wingwall).
Highway culverts typically are designed to operate with an upstream headwater pool
that dissipates the of the channel approach velocity. This type of situation does not require
an approach flow transition. Outlet transitions (expansions) should be considered in the
design of all culverts, channel protection, and energy dissipators. Transition design can be
categorized as
culverts with outlet control (subcritical flow) and
culverts with inlet control (supercritical flow).
20.1.1 Culverts with Outlet Control
Culverts with outlet control can have abrupt expansions or gradual transitions. In an
abrupt expansion, the water surface plunges or drops rapidly and the flow spreads out.
The potential energy stored as depth is converted to kinetic energy with a higher velocity of flow. The transition (apron) end velocity can be determined using the experimental
results of Watts (1968) (Figs. 20.2 and 20.3). Figure 20.2 relates the average depth brink
depth ratio (yA/y0) for a rectangular outlet to the Froude number. Figure 20.3 is a similar
relationship for pipe culverts. These curves in Figs. 20.2 and 20.3 were developed for
Fr's ranging from 1.0 to 2.5, the applicable range for most abrupt outlet transitions.
Usually, a low tailwater is encountered at culvert outlets and the flow is supercritical on
the outlet apron.
CYLINDRICALQUADRANT
STRAIGHT LINE
WARPED
ABRUPT
WEDGE
FIGURE 20.1 Transition types. (From Corry et al., 1975)
The design procedure for abrupt expansions is as follows (Corry et al., 1975):
1. Determine V0 and ^0 at the culvert outlet using Fig. 20.4 for box culverts or Fig. 20.5
for circular culverts.
o
2. Compute Froude number, Fr = ^-
o/o
3. Determine optimum flare angle (9) using tan 0 = Fr /3 (Blaisdell and Donnelly, 1949).
If the selected wing wall flare (0W) is greater than 0, consider reducing 0W to 6.
4. Use Fig. 20.2 (for box culverts) or Fig. 20.3 (for pipe culverts) to determine the downstream average depth (yA) knowing Fr and the desired distance L (expressed in multiples of D, diameter).
B = WIDTH OF CULVERT
= (D) FOR PIPE
L = DISTANCE TO DESIRED DEPT.H
MEASURED IN MULTIPLES OF (B)
Y0 = BRINKDEPTH
Y A = AVERAGE DEPTH OF FLOW
FIGURE 20.2 Average depth for abrupt exapnsion below rectangular culvert outlet. (From Cony et al., 1975)
D = DIAMETER OF CULVERT
L = DISTANCE TO DESIRED DEPTH
MEASURED IN MULTIPLES OF (D)
Y0 = BRINKDEPTH
Y A = AVERAGE DEPTH OF FLOW
FIGURE 20.3 Average depth for abrupt expansion below circular culvert outlet. (From Corry et al., 1975).
Y0 = brink depth
D = height of box culvert
B = width of barrel
Q = discharge
TW = tail water depth
TW/D
FIGURE 20.4 Dimensionless rating curves for the outlets of rectangulars culverts on horizontal and mild slopes (From Simons et al; 1970)
Y0 = brink depth
D = dia. of culvert
TW = tail water depth
TW/D
FIGURE 20.5 Dimensionless rating curve for the outlets of circular culverts on horizontal and
mild slopes (From Simons, et al., 1970)
C ftAand
A T^W ^O
d,c = 5
Solution.
-"
&-&-*<.-<@ti-4ffi-">'
TW
and *** O. From Fig. 20.4, V 0 / D = 0.68; therefore, y0 and V0 can be deter-
PLAN
HYDRAULICGRADELINE
WATERSURFACE
ELEVATION
ATUM
FIGURE 20.6 Definition sketch. (From Cony et al., 1975)
(V2 V 2^
= Cc
C \ 2- L
(2g
2gJ
(20 1 ^
(ml)
(V2
V2}
= CC \^- $- L
'(2g
2g)
(20 2)
2s)
(2
H
HLE
where transition loss coefficients are listed in Table (20.1) for the transition types
(USAGE, 1970).
The design procedure for gradual transitions is as follows (Corry et al 1975):
1. Use Manning's equation to compute y4 = yn and V4, knowing Q, S0, n and the outlet
geometry.
2. Compute the critical depth using yc = ^1 (QIW)m. For box culvert, kv = 0.315 where
Q is in cfs and W is the box culvert width in ft. In SI units ^1 = 0.467 with Q in mVs
and W in m. Compare yn and yc to ensure subcritical flow.
3. For the chosen transition type, use Table 20.1 to obtain the transition loss coefficients
C0 and Ce.
4. Choose V3 = 1.1 yc to have a culvert with a flow depth conservatively above. yc
5. Compute the culvert width using the following energy equation, ignoring the headloss
caused by friction:
V42
PV32 V42 ^l
V3 2
+ C
73 +-f Vv3 ++Z74 + vA;4 ++ +
C ^ - -J-Z
V3 = (W3V3 where V3 = 1.1 yc = IAk1 (QfW3)2 For box culverts and V4 = QtWj4.
Neglecting Z3 - Z4 for a short reach and a mild bottom slope,
^RK^'tei]''--'
><+fei 4*" '" (ir ^ [(^. iut,w-ii ij" -c->
TABLE 20.1 Transition Loss Coefficients
Transition Type
Contraction
Cr
Expansion
Ce
Warped
Cylindrical quadrant
Wedge
Straight line
Square end
1.10
0.15
0.30
0.30
0.30
0.20
0.25
0.50
0.50
0.75
A^-*-^+>$-
n A 1
n2/3
/^
"\ ^ 1 \2/3
^) ^'^-('-"^^s0-*)
x2/3
and
^J
W3 = ([BV[A])
(20.3)
where
=*+][wh]-c>
and
[B] = L 1
2/3+
/i2/3 i
Tfefi ( 1 -^
/y2
y2\
y2
>+y>+t+c^-^rz>+y>+w
(20
-4)
7. If the amount of backwater (yr yn) exceeds a preferred or required freeboard, then
select a larger culvert width and calculate y3 using Eq. (20.3) given in Step 5 (return
to Step 5).
8. If the culvert width is acceptable, use the flow conditions to compute the transition
length using a 4.5:1 flare (USAGE, 1970):
LT = 4.5 (^
(20.5)
LT =4.5^^
(20.6)
or
9. Compute the water surface profile through the structure using a standard step backwater analysis. This will include an evaluation of friction losses.
Example 2. Determine the dimensions for a culvert and gradual transition needed for a 3
m-wide rectangular flood control channel at a slope of 0.001 m/m. The culvert length is
30.5 m, and is to convey 8.5 m3/s. Use n = 0.02.
Solution:
Step 1. Assuming normal depth at Section 4 (refer to Fig. 20.6), then use Manning's
equation to compute V4 = yn = 1.99 m and Vn = 1.42 m/s.
D ^2/3
( S2^3
-^ = 0.467 -^- = 0.935 m.
Step 7. Because ^1 is 1.29 m above the normal depth and because only 0.6 m free board
is available, we go back to Step 5 and compute ^3. Assuming that the culvert
width is 1.45 m: Repeat step 5,
LT
CENTERLINE
ALONGWALL
FLOW
a. SCHEMATICPROFILE
b. PLAN
FIGURE 20.7 Supercritical inlet transition for rectangular channel.
(From USAGE, 1970)
a. Compute L = (W1- W2)/2 tan 9W
b. Find P1, y21 V 1 , and Fr1 from
* *
tan B1 V(I + 8Fr12 sin2 B1 - 3)
1
^= 2 ^ + Vl+8Fr1CV-I
(2
'7)
and
(20.8)
^
I](^+ lfl
2 = f4^ ? -UJL * I2^2JUi
JU
JJ
(20.9)
and
FIGURE 20.8 Supercritical inlet transition design curves for rectangular channels.
(USAGE, 1970)
Fr = -L= =- 19'2
= 3.05m
VgAfT
V 32.2(15.6)712.7
Step 2. Because (T + Bw )/2 = (12.7 + 6) /2 = 9.4 ft, use W1 = 10 ft, a rectangular chan-
nel. Use Manning's equation to compute yn = 1.54 ft and Vn = 19.5 ft/s, then
compute the Fr:
Fr = -^= = . 19'5
=2.77.
Vg^ V32.2(1.54)
Step 3. Assume a trial culvert width W2 = 5 ft.
Step 4. Try 8W = 14 for Fr = 2.8:
L = ( 1 0 - 5 ) / 2 tan 14 = 10ft
P1 = 35, V2^1 = 1.8, Fr2 =1.8
Step 5. Because y3 = 4.4 ft. and W3 = 5 ft, a 5 X 5 box culvert will be satisfactory.
20.2
ENERGYDISSIPATIONFORCULVERTSANDCHANNELS
DATUM
FIGURE 20.10 Definition sketch basin transition. (From Corry et al., 1975)
W <W +
(20 10)
'
'
^T 3Fr0+1
'
where LT = (Z0 Z1)AS7, and the right-hand side is the limit that flares naturally in the
slope distance L.
c. Compute Y1 using the following equation derived from the energy equation from the
culvert outlet brink to the basin (Sec. 1 in Fig. 20.10). Use V1 = Q/ y^W^ to determine^ and then V1:
r
i1'2
Q = J^B [Zg (Z* - ZI + % - *) + ^02J
( 2 O-U)
d. Compute the Fr
Fr 1 =A
vw
e. Compute y2 using Eq. (20.12) for the hydraulic jump basin:
J2 = y[vi + 8 F r J - I J
(20.12)
(20.13)
where
^o
FIGURE 20.11 Length of jump in terms Of^ 1 , rectangular channel. (From Corry et al.,
1975)
FIGURE 20.12 Example problem 18.2.1 hydraulic jump basin. (Not to scale).
Solution:
Step 1. The brink depth and velocity are given: Fr0 = 4.0.
Step 2. The tailwater depth is computed assuming normal depth. Manning's equation
is solved to obtain yn = 0.57 m tailwater depth and Vn = 4.85 m/s. (Refer to
Fig. 20.12)
Step 3. Assuming a basin elevation OfZ 1 = 25.9 m:
a. Select W8 = 3 m and S5 = S7 = 0.5
b. Check W8 using Eq. 20.10, where W0 = 3 m,
W8 = 3 < 3 + [2L7 V5JTTJ(3Fr0I OK.
c. Compute ^1 using Eq. (20.11)
1/2
11.8 = V1(S) 2(9.81)(30.5 - 25.9 + 0.457 - V1) + (8.47)2
Z =
20.2.2.1 Saint Anthony Falls stilling basin. The Saint Anthony Falls, (SAF) stilling basin
is a generalized design based upon model studies conducted by the U.S. Soil Conservation
Service at the St. Anthony Falls Hydraulic Laboratory, University of Minnesota. Figure
20.13 illustrates the SAF stilling basin design which is recommended for small structures
such as spillway outlet works, and for canals where the Fr ranges from 1.7 to 17 (at the dissipator entrance). Through the use of chute blocks, baffle or floor blocks, and an end sill,
the basin length is about 80 percent of the free hydraulic jump length.
The design procedure for SAF basins is as follows (Corry et al 1975):
1. Choose basin configuration and flare dimension, Z. (Refer to Fig. 20.13)
2. Use the design procedure presented in Sec. 20.2.1 for supercritical expansions into
hydraulic jump basins to determine basin width (W^), elevation (Z1), length (Le), total
length (L), incoming depth Cy1), incoming Fr (Fr1), and jump height (y2). Steps 3e and
3f in Sec. 20.2.1 are modified; for Step 3e, determine y2 using the sequent depth y.\
(20.14)
(20.15)
(20.16)
(20.17)
RECTANGULAR BASIN
HALF-PLAN
EQUATION NUMBER
FLARED BASIN
HALF-PLAN
0-90
45 PREFERRED
SIDEWALL
CHUTEBLOCK
FLOORORBAFFLEBLOCKS ENDSILL
VARIES
Height:
H1 = yl
Width:
Number:
Adjusted:
(20.18)
FIGURE 20.14 Example 4 St. Anthony Falls stilling basin. (Not to scale)
b. Next, use Eq. (20.15) for Fr1 = 1.7 to 5.5 to compute J2:
Because y2 > Tw = 0.57, we can lower the elevation of the basin. Use Z1 =
27.9 m with W8 = 3 m and ST = S5 = 0.5. WB is OK and no flare is used.
c. Compute J1 using the energy Eq. (20.11). The basin has been lowered so now
J1 is not J0, the brink depth.
Q = * WB\2g(ZQ - Z1 + J0 - ^1) + V02]
r
nl/2
11.8 = ^(312(9.81X30.5 - 27.9 - 0.457 - J1) + (8.47)2
FREEJUMP
TYPE IV
TYPE Il BASIN
FIGURE 20.15 U.S. Bureau of Reclamation Type II basin. (Corry et al, 1975)
ENDSILL
BAFFLE PIERS
BAFFLE PIERS
END
FIGURE 20.16 Height of baffle piers and end sill (Type III basin). (From U.S. Bureau of Reclamation, 1987)
Use dimensions in Fig. 18.7 to determine chute block dimensions and spacing.
Use Fig. 18,7 to determine the end sill height.
Example 5. Determine the dimensions of U.S. Bureau of Reclamation's Type II basin for
the supercritical flow expansion described in Example 3.
Solution:
Step 1. Use the design procedure for a supercritical flow expansion into a hydraulic jump
basin. (Refer to Examples 3 and 4).
Given V0 = 8.47 m/s, V0 = 0.457 m, and Fr0 = 4.0. (Figure 20.17)
The tailwater depth is Tw = yn = 0.57 m, and Vn = 4.85 m/s.
Assume Z1 = Z0 = 30.5 m, compute y2
y2 = C1J1[Vl + 8Fr^- l]/2
= 1.1(0.457/Vl + 8(4)2 - IJ/2
= 2.6m
y2 > Tw (2.6 > 0.57). Therefore we need to lower elevation Z1 of the basin floor.
a. Use a basin floor elevation of Z1 = Z2 = 25.76 m, with W8 = 3 m, ST =
S5 = 0.5.
b. W8 is OK, no flare.
c. Compute V1 using energy Eq. (20.11).
r
T'2
Q = ViW8^g(Z0 - Z1 + y0 - V1) + V02J
r
i172
11.8 = Vj(3) 2(9.81)(30.5 - 25.76 + 0.457 - V1) + 8.472
Z3
Figure 20.18 illustrates the U.S, Bureau of Reclamation's Type VI impact-type energy dissipator, which can be used with culverts. This basin is contained in boxlike structures
requiring no tail water for operation. The structures can be used for open channels as well
as culverts, and the basin can be used at sites where the entrance velocity to the basin does
not exceed 50 ft/s, and the discharge is less than 400 cfs. This dissipator should not be
used if the buildup of debris or ice can cause substantial clogging.
The design procedure is as follows (Corry et al, 1975).
1. Compute the flow area at the end of the culvert using the maximum design discharge
and velocity. Compute the equivalent depth of flow entering the dissipator from the
culvert as
Je = (f)"2
(20.19)
where A is the cross-sectional area of flow in the culvert. This converts the cross-sectional area of flow of a pipe into an equivalent rectangular cross section with a width
twice the depth of flow. The culvert preceding the dissipator can be open, closed, or
have any cross section. This approach ignores the size and shape of the culvert entirely except for the determination of flow entering the dissipator.
2. Compute Fr and the energy at the end of the culvert H0:
V2
" = ^ + 2 F -
Then use Figure 20.19 to determine the basin width. Enter Figure with Fr to determine
H0IW then W = H0I (H0IW).
3. Use Table 20.2 to determine the dimensions of the dissipator structure.
DATUM
PLAN
SECTION
STILLING BASIN DESIGN
PLAN
BEDDING
SECTION
ALTERNATE ENO SILL
FIGURE 20.18 Baffle-wall energy dissipator of a U.S. Bureau of Reclamation type IV basin. (From Corry et
al., 1975)
FIGURE 20.19 Design curve for a baffle-wall dissipator. (Corry et al., 1975)
D =^
(20.20)
where q is the discharge per unit width of the crest overfall, g is the acceleration caused
by gravity, and h0 is the height of the drop. The dimensions L1, V1, y2, and ^3 in Fig. 20.22
are determined using the following:
^- = 4.3(W*27
AZ0
(20.21)
- = LQN*22
AZ0
(20.22)
TABLE 20.2 Baffle Wall Dissipator: Dimensions of The Basin in Feet and Inches
W
4-0
5-0
6-0
7_0
8-0
9-0
10-0
11-0
12-0
13-0
14-0
15-0
16-0
17_0
18-0
19-0
20-0
H1
3-1 5-5
3-10 6-8
4-7 8-0
5-5 9-5
6-2 10-8
6-11 12-0
7-8 13-5
8-5 14-7
9-2 16-0
10-0 17-4
10-9 18-8
11-6 20-0
12-3 21-4
13-0 22-6
13-9 23-11
14-7 25^
15-4 26-7
H2
1-6
1-11
2-3
2-7
3-0
3-5
3-9
4-2
4-6
4-11
5-3
5-7
6-0
6-4
6-8
7-1
7-6
3-11-8
3-102-1
4-72-6
5-52-11
6-23^
6-113-9
7-84-2
8-54-7
9-25-0
10-05-5
10-95-10
11-66-3
12-36-8
13-07-1
13-97-6
14-77-11
15-48-4
I2
I1
I4
t5
(20.23)
^- = 1.66A#27
(20.24)
/I0
L2 is the length of the jump, L1 can be determined using Fig. 20.22. The sequent depth
and the tailwater depth Tw must be compared to determine whether Tw < y3, or Tw = y3,
or Tw > y3. If Tw < y3, the hydraulic jump moves downstream. In this case, it is necessary
AERATED
FIGURE 20.20 Flow geometry of a straight-drop spillway. (From Corry et al., 1975)
TOP SLOPE I TO I
UPPER NAPPE
SIDE WALL
HEIGHT
FLOOR BLOCKS
LONGITUDINALSILL
(OPTIONAL)
ENDSILL
PLAN
FIGURE 20.21 Straight-drop spillway stilling basin. (From Rand, 1955)
to construct the apron at the bed level and an end sill or baffles or to construct the apron
below the downstream bed level and an end sill. If Tw > y3, the hydraulic jump may
become submerged. If Tw = V3, the hydraulic jump begins at depth V2; there is no supercritical flow on the apron, and L1 is a minimum.
20.2.4.2 Grated energy dissipators. Energy dissipators with grates (Fig. 20.23) also can
be used in conjunction with drop structures. The U.S.Bureau of Reclamation (1987)
developed the following design recommendations for grates for incoming subcritical
flow:
1. Select slot width with a full slot width at each wall.
2. Compute beam length L0 using
LOWER NAPPE
(NOTAILWA1TER)
TAILWATER
LEVEL
TANGENTATPOINT
OFSUBMERGENCE
MEAN
SUBMERGED
TRAJECTORY
NAPPE
TRAJECTORY
FREE NAPPE
TRAJECTORY
FIGURE 20.22 Design chart for determination of L1. (From Corry et al., 1975)
LGG =
,
C(W)(N) V2^~
(20.25)
where C is a coefficient equal to 0.245, W is the width of the slots in feet, and N is the
number of slots or spaces between beams.Then compute the beam width = 1.5 W. The
quantity (W)(N) can be adjusted until an acceptable beam length, L0 is determined.
3. For self-cleaning, the grate can be tilted appoximately 3 in the downstream direction.
20.2.4.3 Straight drop structures. The straight drop structure shown in Fig. 20.21 consists of a horizontal apron with blocks and sills to dissipate energy. This structure is for
drops of less than 15 ft (4.57 m) and for sufficient tailwater. The design parameters
include the length of the basin, the position and size of the floor blocks, the position and
height of the end sill, the position of the wingwalls, and the geometry of the approach
channel. This structure was developed by the Agricultural Research Service at the Saint
Anthony Falls Hydraulic Laboratory.
The design procedure for a straight-drop structure is as follows (Corry et al 1975):
(20.26)
where distances are illustrated in Fig. 20.21. The distance from the headwall to the
point where the surface of the upper nappe strikes the still basin floor L1 is
L 1 =^
(20.27)
(20.28)
where
0.691+0.228 f M - f ^ l L
UJ
UJj
LS
~
I"
1
L
(20.29)
0.185 + 0.456^1
UJJ
and
(20.30)
2.
3.
4.
5.
6.
7.
8.
9.
stilling basin floor to the upstream face of the floor blocks L2 is determined using
L2 = 0.8 yc. The distance between the upstream face of the floor blocks and the end of
the stilling basin, L3, is determined using L3 ^ 1.75yc.
Floor blocks are proportioned as follows:
a) height = 0.8yc;
b) width and spacing should be 0.4yc with a variation of 0.15yc permitted, and
c) blocks should be square in plan, and
d) blocts should occupy between 50 to 60 percent of the stilling basin width.
Compute the end sill height as 0.4yc.
If longitudinal sills (for structural purposes only) are used, they should be constructed
through the floor block, not between the floor blocks.
Compute sidewall height above the tailwater level as 0.85yc.
Wingwalls are constructed at an angle of 45 with the outlet center line with a top slope
of 1:1.
Compute the minimum height of the tailwater surface above the floor of the stilling
basin y3 using y3 = 2.15yc.
Modification, to the approach channel are as follows: The crest of the spillway should
be at the same elevation as the approach channel, the bottom width should be equal to
the spillway notch length W0 at the headwall, and protection with riprap or paving
should be provided for a distance upstream of the headwall of 3yc.
Using the recommendations in Step 8, no special provision for aeration is needed.
Example 6 Determine the dimensions of a straight-drop spillway stilling basin for a discharge of 7.08 m3/s. The downstream trapezoidal channel has a 3:1 side slope with a 3.05
m bottom width S0 = 0.002 m/m, n = 0.03, and a normal depth of 1.024 m. The drop h0 =
1.83 m.
Solution:
Step 1. Determine the minimum basin length L8. The critical depth yc is determined as
yc = 0.655 m. Then hjyc = 1.83/0.655 = 2.79; H2 = h0 -2.15yc = 1.83 2.15(0.655) = 0.422 m; hjyc = 0.422/0.655 = 0.644.
Using Fig. 20.22, L1Jyc = 8.2 for hjyc = 2.79 and hjyc = 0.644.
Then L1 = 8.2(0.655) = 5.37 m; L2 = 0.8yc = 0.8(0.655) = 0.52 m; and L3 = 1.75yc
= 1.75(0.655) = 1.15 m; thus, LB = L1 + L2 + L3 = 5.37 + 0.52 + 1.15 = 7.04 m.
Step 2. Proportions the floor blocks are
height = 0.8vc - 0.8(0.655) = 0.524 m,
width = 0.4yc = 0.262 m, and
spacing = 0.4yc = 0.262 m.
Step 3. Calculate the end sill height = 0.4yc = 0.262 m.
Step 4. Use the longitudinal sill passing through the floor blocks.
Step 5. Calculate the side-wall height above the tailwater = 0.85yc = 0.85(0.655)
= 0.557 m.
Step 6. Local wing walls are at a 45 angle with the outlet center line.
Step 7. Calculate the minimum height of tailwater above the floor of the stilling basin: y3
= 2.15yc - 2.15(0.655) = 1.41 m. The basin must be placed 1.43 - 1.024 =
0.406 m below the downstream bed level.
20.2.4.4 Box-inlet drop structure. The box-inlet drop structure shown in Figure 20.24
consists of two different sections that are effective in controlling flow: the crest of the box
inlet and the opening in the headwall. This structure is based on experiments by the U.S.
Soil Conservation Service at the Saint Anthony Falls Hydraulic Laboratory (Blaisdell and
Donnelly, 1956). The design procedure of box-inlet drop structures is as follows:
(FHWA, 1978):
1. Select/*0
2. Select L1, W2, and Lc where Lc is the length of the box-inlet crest, Lc = W2 + 2L1,
where W2 is the width of the box inlet, and L1 is the length of the box inlet.
3. Compute the head y0 for the crest using the discharge equation for a rectangular weir:
(20 31)
^(dkf
4. Compute -^
and determine the coefficient of discharge C2 from Fig. 20.25.
(W2j
(L,]
5. Compute M- and determine the relative head correction, CH, from Fig. 20.26.
6. Compute y0 for the headwall opening using
y0=
^7=
IC2W2V^]
2/3
-Cn
(20.32)
COEFFICIENT OF DISCHARGE, C2
FIGURE 20.25 Coefficient of discharge, with control at headwall opening. (From Corry et al.,
1975)
FIGURE 20.26 Relative head correction for HJw2 > 1V4 with control at headwall opening.
(From Corry et al., 1975)
e. Determine the adjusted y0:
? -(sW/wJ20
(2033)
'-ml
9. Compute the critical depth at the exit of the stilling basin yc3 using
-tej
f
0.2
L2 y
~ < (LA
IkJ
(20.36)
DISCHARGE COEFFICIENT, C1
W
FIGURE 20.28 Correction for box-inlet shape, with control at the box-inlet crest (-c > 3).
From Corry et al., 1975)
TABLE 20.3 Correction for Dike Efect CE: Control at the Box-Inlet Crest
(Control at Box-Inlet Crest)
W4AV2
L1IW2
OO
0.5
0.90
1.0
.80
1.5
.76
2.0
.76
Source: Corry et al (1975)
Ol
0.96
.88
.83
.83
0.2
1.00
.93
.88
.88
LB
0.3
1.02
.96
.92
.92
0.4
1.04
.98
.94
.94
_(i)
2L
i
(W2)
0.5
1.05
1.00
.96
.96
0.6
1.05
1.01
.97
.97
(20.37)
and
L3 = ^f^
(20.38)
if
W3 < 11.5vc3
(20.39)
or
?3 = yC3 + 0.052PF3
if
W3 > 11.5yc3
(20.40)
20.2.5
RIPRAPBASINS
The riprap basin recommended in Corry et al (1975) for culverts is shown in Fig. 20.30,
which is a preshaped basin lined with riprap. The surface of the riprapped floor of the
energy dissipating pool is at an elevation hs below the culvert invert, where hs is the
approximate depth of scour that would occur in a thick pad of riprap if subjected to the
design discharge (Corry et al 1975). The length of the pool is the larger of I0hs or W0, and
the overall length of the basin is the larger of I5hs or 4W0. The ratio hs/d5Q should be less
than 4 [(HJd50) < 4] .Figure 20.31 provides design detaills for riprapped culvert energy
basins.
NOTEB
IFT
BOARD
NOTEA - SUFFI
IF EXI f VtLOClIY OF BASIONNAL
IS SPECI
EXTENDN A-ABASISUCH
N AS REQUIRED TO OBTAIN
AREAFFIATEIED.DSECTI
SECTIOCNIENTAREACROSS-SECTI
AT SEC. A-A) - SPECI
EXITOVELOCI
TV. THAT O^ /(CROSS
NOTE B - FLOOR
WARP OFBASIBASIN TON SHOULD
CONFORMBE ATTOTHENATURAL
STREAM
CHANNEL.
TOP OF RIPRAP IN
CHANNEL BOTTOM AT SEC. A-A. SAME ELEVATION OR LOWER THAN NATURAL
DlSSlPATOR POOL
Sh1APRON
ORW0MIN.
100,OR3W
0MIN.
TOPOFBERM
TOP OF RIPRAP
CHANNEL
ZONTAL
TO SUPPORT RIPRAP
THI
COPTIONAL
KENEDORSLOPI- CONSTRUCT
NG
TOE
IF DOWNSTREAM
CHANNEL
SECTION
NOTE B
EXCAVATETOTHISLINE,
NOTE:
W0 - PIDIAMETER
FOR
PE CULVERT
W0 - BARREL
WIDTH
FOR
BOX
CULVERT
W = SPAN
OF PIPE-ARCH
CULVERT
SYMM ABOUT
APRON
CULVERT
HORIZONTAL
BERM
AS REQUIPRRAPED
TOSUPPORTRI
FIGURE 20.30 Details of riprapped culvert energy basin. (Corry et al., 1975)
DESIGNDISCHARGE-Q
WETTED AREA AT BRINK OF CULVERT
d50 = THE MEDIAN SIZE OF ROCK
BYWEIGHT.
OR
ANGULARROUNDEDROCK
ROCK.
EQUIVALENT BRINK DEPTH
BRINK DEPTH FOR BOX CULVERT
FOR
NON-RECTANGULAR
SECTIONS
CULVERTBRINK
?. SECTION
FROUDE NUMBER
FIGURE 20.31 Relative depth of scour hole versus Froude number at brink
of culvert with relative size of riprap as a third variable.
REFERENCES
Blaisdell, F. W., Flow Through Diverging Open Channel Transitions at Supercritical Velocities,
SCS Report No. SCS-TR-76, U.S. Department of Agriculture, April 1949.
Blaisdell, F. W., and C. A. Donnelly, "The Box Inlet Drop Spillway and Its Outlet", Transactions,
of the American Society of Civil Engineers, 121:955-986, 1956.
Blaisdell, F. W., The SAF Stilling Basin, U. S. Goverment Printing Office, 1959
Corry, M. L., P. L. Thompson, E. J. Watts, J. S., Jones, and D. L. Richards, Hydraulic Design of
Energy Dissipators for Culverts and Channels, Enqineering Circular 145, Federal Highway
Administration, U.S. Department of Transportation, Washington DC, 1975.
Fedeard Highinay Administratin , Hydraulics of Bridge Waterways, Hydraulic Design Series No. 1,
Federal Highway Administration, U.S. Department of Transportation, Washington, DC, 1978.
Ippen, A. T, "Mechanics of Supercritical Flow", Transactions of the American Society of Civil
Engineers,, lib; 268-295, 1951
Rand, W, "Flow Geometry at Straight Drop Spillways," Paper No. 791, Proceedings of the
American Society of Civil Engineers, VoI, 81, pp. 1-13, September 1955.
Simons, D. B., M. A. Stevens, and F. J. Watts, Flood Protection at Culvert Outlets, CER No.
69-70-DBS-MAS-FJWA, Colorado State University, Fort Collins, CO. 1970.
U.S. Army Corps of Engineers, Hydraulic Design of Flood Control Channels, Engineering and
Design Manual EMl 110-2-1601, pp.2026, July 1970.
U.S. Bureau of Reclamation, (USBR), Design of Small Dams, U.S. Government Printing Office,
Denver, CO, 1987.
Watts, F. J., Hydraulics of Rigid Boundary Basins, doctoral dissertation, Colorado State University,
Fort Collins, CO, 1968.
CHAPTER 21
HYDRAULIC DESIGN OF
FLOW MEASURING
STRUCTURES
21.1
INTRODUCTION
Experienced water providers and users can use this chapter as a quick review of hydraulic
principles related to water measurement and its relation to hydraulic design for environmental considerations.
The hydraulic design of flow measuring structures usually confronts the engineer with
two opportunities. One is the design of measurement structures in a retrofit situation and
the other is in original project design. The retrofit mode is usually difficult and requires
much innovation just to obtain passable function within the space and sizing limitations
and other constraints usually imposed. Because of the increasing emphasis on quantifying
flow rates and volumes in most aspects of water resource planning and management, the
retrofit applications currently dominate the design problems.
Most textbooks deal with recommending ideal installation situations and retrofit projects appear to be unable to comply without great economic impact. This too frequently
can lead to arbitrary compromises that produce poor measurement performance. Even
new installations may be limited by space requirements. This may force design decisions
into the final construction that compromise accuracy. This chapter will strive to show the
design concepts available, particularly those useful for designing both new and retrofit
installations, and will point out measurement behaviors to be expected from various compromises. This chapter suggests those deviations that cause least impact and guides the
designer to choices that may be hydraulically acceptable and still meet structural goals.
Of the numerous flowmetering methods available to the hydraulic engineer, most are
based on well-established hydraulic principles and are amenable to design manipulations
of size, shape, and response. While this aspect of flow measurement is documented in several handbooks and texts, the design and retrofit of sites to accommodate and facilitate
measurement is not as well described or is described in a scattered assortment of books
and articles.
Pipeline flows of water are usually less complicated to measure than open-channel
flows, most obviously because the flow area does not change significantly with flow rate.
Consequently, many applications of pipeline flows are held to stricter accuracy standards
than channel flows can reasonably achieve. Thus, channel flows and their measurement
are usually limited to large delivery volumes and to accuracies acceptable to the related
activities, such as sewer flows and irrigation deliveries.
The purpose of this chapter is to consolidate design information for evaluating a flow
measurement site, selecting a flow measuring system, and adapting the measuring site to
optimize measuring and other functions that may be desired from the site. Emphasis will
be on open-channel flow measurements because that is a likely need of the hydraulic engineer. Pipe flowmeters in water supply will also be discussed in lesser detail because the
major application of the many types of pipe flowmeters is well covered in the chemical
and petroleum industry literature.
Experienced readers may wish to further investigate and seek more advanced references in hydraulics and fluid mechanics. Extensive information on fluid meter theory and
detailed material for determining coefficients for tube-type meters is given in American
Society, of Mechanical Engineers (ASME) (1959, 1971) and revisited with modern
updates in books by Spitzer (1990) and Miller (1996). Brater and King (1982) have a thorough discussion of general critical depth relations and detailed relationships for most
common hydraulic flow section shapes in open channels. Bos (1989) covers a broad segment of open-channel water measurement devices.
dling the equipment and doing the analysis. The most used techniques applicable to openchannel systems, including sewer flows and irrigation canal flow measurements, depend
on exploiting the special velocity properties of critical flow, as discussed in a section
21.2.3.
Continuity equation. The first basic equation for water flowing in either pipes or channels is the continuity equation, which simply states that discharge rate (volumetric flow
per unit time), Q, is equal to flow cross-sectional area, A, times flow mean velocity, V,
through the flow cross section, or
Q = AV
(21.1)
Bernoulli energy equation. Another basic equation involves energy relations and is
also applicable to both pipe and channel flows. The most familiar form is for closed pipe
flow, wherein the basic energy principles are described by the Bernoulli energy equation.
For two locations along a pipe at stations 1 and 2,(Fig. 21.1), the Bernoulli equation can
be expressed as
V2
V2
ZI + h\ + -^- = ^2 + h2 + -^- = constant
(21.2)
where the terms are expressed in length dimensions as z = the height from an arbitrary
reference plane (datum) h = the pressure head V = average velocity through the pipe
cross-section at the designated location V2/2g = the velocity head g = the gravitational
constant i ,2 = subscripts denoting the respective locations along the pipeline.
This equation is based on uniform velocity across the conduit area and no energy losses. However, in real fluid flows, nonuniform velocities exist and friction causes energy
conversion to heat. Typically, these velocities are zero at the walls and reach a maximum
profile velocity near the center of the flow. If the flow is viscous flow in a round pipe, the
flow profile is parabolic, that is, "bullet-shaped." If the velocity is fully turbulent, the bullet-shape is much flattened, with steep velocity gradients near the wall and nearly uniform
profile across the remainder of the pipe. These idealized profiles can be skewed drastically
by regulating valves, structures, conduit bends and other flow obstructions. Therefore,
application of these equations depends on knowing, or controlling, the velocity profile so
that the average velocity in the conduit cross section can be inferred.
Datum
FIGURE 21.1 Energy balance in pipe flow.
Equation (2.2) requires some adjustments to convert it to the energy equation, which
is useful in analyzing flows in pipes or open channels with a small slope (Chow, 1959).
First we introduce correction factors, GL1 and OC2, called the velocity distribution coefficients, to account for the computational expediency of using the average velocities, V1 and
V2, to compute the kinetic energy term, W2g, at the respective locations 1 and 2 along a
channel. These values for the usual range of turbulent flows in water usually range from
about 1.01 to 1.05, although for thick petroleum products in pipe flows and low velocity
flows, the value can approach a value of 2. Second, a term, hp for the loss of energy
between the two points is included. The result is
V2
v2
Z1 + H1 + a, -i = Z2 + /I2 + a2 -^ + hf
(21.3)
5
o
21.2.2 Pipe Hydraulics
Reynolds number. The behavior of flow in pipes is governed primarily by the viscosity of
the fluid. In pipeline flows, the ratio between the dynamic forces and the viscus forces is
important for defining the limits between laminar and turbulent flows and other functions
of pipe flow. This ratio is called the Reynolds number, Rn, and is defined as
Rn = ^-
(21.4)
where V = the velocity of the flow, Lc = characteristic length, typically the pipe diameter, D and v = the kinematic viscosity.
Headloss characteristics in pipes. The Reynolds number, Rn, defined above, represents the effect of viscosity relative to inertia and is used to define appropriate flow ranges
for headloss equations in pipe flow. For example, headloss is proportional to the square of
the velocity, when the velocities and pipe size combinations defined by a pipe-diameterbased Reynolds number, Rn, greater than about 1000. Most of the flows of interest in general hydraulic engineering have Reynolds numbers greater than 1000. Some exceptions
are found in drip or trickle irrigation systems common in agricultural and urban landscape
settings.
The headloss, hf for Rn greater than the minimum value of about 1000 is traditionally
expressed in terms of a friction factor,/, the pipe diameter, D, pipe length, L, and the velocity head, V2/2g, where g is the gravitational constant, and V is the average velocity, as
hf = / I
(21.5)
The value for/is usually obtained from a Moody diagram which is a graphical representation of the/value in terms of the Reynolds number, the roughness height of the pipe
wall material, e, and the pipe diameter, D. The Moody diagram is a graphical solution of
the Colebrook function
1 = -2log^
+ ^=]
Vf
I3-7
RnVf)
(21.6)
The e values range from 0.0000015 m for smooth plastic pipe to 0.00026 m for cast
iron pipe. Concrete pipe ranges from about 0.0003 m to 0.003 m (Daugherty and
Ingersoll, 1954). The equation can be readily solved by iteration techniques using a computer spreadsheet.
(21-7>
For conduits such as pipes flowing nearly full, the surface flow width may be narrow,
and Dm may be a larger value than the physical water depth. For the usual natural channels and most canals, Dn is interchangeable with average depth. Sometimes it is simply
called the hydraulic depth (Chow, 1959).
Froude number. Open-channel flow behavior is governed primarily by gravity
forces. The ratio of the inertial forces to the gravity forces is called the Froude number,
Fr, and is defined by
Fr =
Y
(21.8)
v^:
where V the velocity of the flow, g = the gravitational constant, Dn = the hydraulic
mean depth.
The Froude number applies to most open channel flows and is used for defining model
scale ratios and estimating stable flow characteristics in open channels.
Specific energy. It is useful to define the energy equation in terms of the local channel bottom instead of an arbitrary datum. This is called the specific energy, ", and is given
by:
E = y + ^(21.9)
2g
That is, the specific energy is equal to the sum of the depth of flow y and the velocity
head (Fig. 21.2).
Critical flow and critical depth. In open channels a flow phenomenon occurs that
does not happen in closed pipe flows. The process is called critical flow. Critical flow is
defined for open-channel flows as the maximum discharge for the minimum specific energy, that is, critical flow represents the minimum combination of potential energy (depth
of flow, y) and kinetic energy (velocity head, V2/2g) for the given discharge (Chow, 1959).
The depth of flow then is the critical depth. By virtue of the continuity equation, for a constant discharge at critical flow, an increase in depth must necessarily be accompanied by
a decrease in velocity, which is called subcritical velocity. Conversely, a decrease in depth
for the same flow rate necessarily requires an increase in velocity, which is called supercritical velocity.
When critical flow occurs in an open channel it can be shown (Chow, 1959) that
V2 D
oc^ = -f
(21.10)
EnergyGrade Line
Hydraulic Grade Line
& = AJ%%
<2U1)
In practice, the water surface slope in a contraction is relatively steep and the precise plane of the critical flow section is not easily or reliably located. Thus, the data
for accurately evaluating the hydraulic depth, Dm, is not readily obtained. For critical
flow flumes, the flow depth is therefore not measured at this critical section, but
instead a depth is measured in the upstream channel, where the velocity head is computable or is minimal. The critical depth is then mathematically derived based on energy principles described by Bos et al. (1991). These flumes, sometimes referred to as
the computable flumes that rely on critical flow theory, will be discussed in more detail
in Section 21.7.
For maximum discharge for minimum energy, the condition described above for critical flow in open channels, it can be shown that
af = A = ^
<2U2)
where: A = the channel flow area, T = the top width of the channel flow, Dm = the
hydraulic mean depth, and Vc = the critical velocity.
Thus, the velocity head at critical flow is equal to half the hydraulic mean depth, sometimes called hydraulic depth, Dm = AIT (Chow, 1959).
From the above,
,Vc
= 1 = Fr
(21.13)
VgD m la
where Fr is the Froude number defined above. Thus, at critical flow the Froude number is
unity. Also note that the Froude number can be defined by Fr = VIV0 for velocities other
than critical.
Normal depth. Yet another depth is associated with open channel flows, the normal depth. When the flow in an open channel does not change from station to station,
the flow is said to be uniform and the bottom slope, the hydraulic grade line, and the
energy grade line are all parallel to each other. Figure 21.2 shows the condition when
the flow is not uniform.
Modular limit. If the downstream depth in a channel is too deep, the backwater will
prevent critical depth from occurring. The flow is considered to be submerged whenever
the downstream water surface exceeds the crest elevation of a channel control, such as a
weir or flume. For flumes, particularly, this submergence has little effect on critical depth,
and free flow exists until a certain limiting submergence for that particular flow module
called the modular limit is reached. At some point of submergence, the upstream flow
depth is affected, and the modular limit is exceeded, and free flow does not occur. The
modular limit is defined as that limiting submergence ratio, and is based on the ratio of
the downstream depth to upstream depth. The modular limit occurs when the downstream
backwater causes more than 1 percent change in the calculated discharge in a particular
flow module, or device (Bos, 1989). When the modular limit is exceeded, the flow is
called nonmodular.
21.2.4 Energy Balance Relationships in Channels
Hydraulic problems concerning fluid flow are commonly described in terms of conserving kinetic and potential energy, and are conveniently expressed using the classical
Bernoulli equation in combination with the Continuity equation. The applications of these
equations are generally well documented, particularly for pipe flows, in texts and handbooks and are not repeated here (Brater and King, 1982; Miller, 1996). The case for open
channels is less complete, but is given considerable treatment in Brater and King (1982),
Chow (1959), and Herschy (1985). The computational uncertainties evolve from the
effects of friction and viscosity that distort the classic assumptions of a uniform velocity
profile across the fluid stream. When accountings for friction and flow profile are successfully applied, the results for discharge computations are usually good to excellent for
both pipes and open channels (Bos et al., 1991).
Headloss characteristics in channels. In terms of frictional headlosses, the wetted
perimeter, Pw, of the flow is important. Hydraulic radius, Rh, is defined as the area of the
flow section, A, divided by the wetted perimeter, Pw, or
Rh =
p
(21.14)
Conversely, the wetted perimeter times the hydraulic radius is equal to the area of an
irregular flow section. The hydraulic radius of a channel can be compared to the radius of
a pipe, r, with a cross-sectional area A = nr2 and a circumference or wetted perimeter Pw
= 2 nr. Under these conditions, the hydraulic radius compares to the pipe radius, and to
the pipe diameter, D, as
r D
R
=
(2L15)
=2
^
h
The Manning's formula. Canal and stream discharge rates are usually estimated with
use of the Manning's formula. Many open-channel flow equations have been proposed,
but the most used is the Manning's formula. This expression is partly rational and uses an
empirical coefficient, rc, that is used in both the SI and American unit systems. In general
form it becomes
v=^LR2i3Sm
(2L16)
equal to Se when uniform flow, with the resulting normal depth occurs. As defined above,
normal depth occurs when a channel flow approaches uniformity from station to station
along the channel (Chow, 1959).
For design purposes, the n value for concrete lined canals is usually about 0.014. A
good finish can lower it to 0.012, while concrete in poor condition and channels constructed with shot crete or gunite, usually have n values from 0.016 to 0.018. In some
instances, concrete lined canals, with significant algae growth, have experienced n values
as high as 0.032. This latter value approaches the values usually experienced with unlined
channels, 0.03-0.04. Thus, for reliable application, the use of Manning's formula requires
field experience and on-site inspection of the channel being computed.
21.2.5 Modeling Characteristics for Open Channels
For flowing water in open channels, fluid friction is a factor as well as gravity and inertia. This would seem to present a problem for hydraulic scale modeling, because both
dynamic and kinematic similarity are difficult to achieve simultaneously. Fortunately for
most open-channel flows, there is usually fully developed turbulence. Thus, the fluid friction losses are nearly proportional to V2, and are nearly independent of Reynolds number,
Rn, with rare exceptions.
This means that in open-channel flows, inertia and gravity forces dominate over viscous forces (associated with pipe flows) and are a function of the Froude number, Fn,
alone. Geometric similarity between a model and a prototype then provides kinematic
similarity. For kinematic similarity the ratios of the respective velocities are everywhere
the same. The velocity ratio, Vr, is the velocity in the prototype, Vp, divided by the velocity in the model, Vn, or
Vr = ^(21.17)
m
For Froude modeling, and from the definition for Fn, we note that V is proportional
to the square-root of a length, L (for open channels we used the hydraulic depth, D1n)
with the gravitational constant, g, assumed to be constant. Thus, the above equation can
be written as
"^LT = Y
where Lr = the length ratio between prototype and model dimensions, Lp: Lm
(2L18)
Because the velocity varies as VZ^ and the cross-sectional area as L2 it follows
that
Qp'Qm = LT 1
(21-19)
O
(J Y/2
-SL = U
XZm \.LmJ
(21.20)
or
This equation is valid when all the physical structure dimensions and the heads are of
the same ratio. For example, it can be used to convert a flume rating for one size to that
of a similar flume of another size. Scale modeling works best for determining calibrations
in a range of Lp:Lm less than about 10:1, although ranges exceeding 50:1 have sometimes
been used for studying special situations.
21.3
Flow is usually measured by determining an average flow velocity and using the flow area
to compute the volume discharge. Flow meters then have the function of detecting this
velocity and combining it with the physical information of the conduit to produce a useable readout. This is easily demonstrated for closed conduits. Propeller meters, ultrasonic
meters, laser-Doppler velocimeters, electromagnetic meters, Venturi meters, and orifice
meters all are based on inferring a basic velocity measurement applied to a flow area for
a discharge rate.
For open channels, many flumes depend on determining the velocity based on energy
principles of critical flow. Weirs are usually described in terms of orifice flow integrated
over the weir width and the crest depth. Again these are basically velocity expressions for
flow through a defined area.
Dilution techniques applicable to both closed pipe and open channel flows depend on
detecting the amount of fluid added to a known starting amount of tracer material. The
dilution ratio determines the discharge ratio, in the case of constant injection of a tracer.
The tracer may be a chemical or even injected heat or heated fluid.
Electromagnetic meters depend on generating voltages by flowing a conductive fluid,
usually water, through a magnetic field to produce a velocity indication.
21.3.1 Water Meter Classification
Flow measuring devices are commonly classified into those that are rate meters and measure discharge rate as the primary reported indication and those that are quantity meters
and measure volume as the primary indication. The latter include weighing tanks and
batch volume tanks and are used mostly in laboratory settings as flow rate standards.
Devices in either of these broad classes can again be divided according to the physical
principle that is used to detect that primary indication (ASME, 1959). The meter part that
interacts with the flow to produce the primary indication is referred to as the primary
device. This interaction exploits one or more of a few physical principles, such as pressure force, energy conversion, weight, electrical properties, mixing properties, sonic properties, and so on, to generate a signal. Primary devices are thus limited in number and variety. Secondary devices convert the primary interaction into useable readout. These secondary devices are numerous and relatively unlimited in configuration and variety. The
function of one class can be converted into the response of the other with suitable secondary devices.
Some water measuring devices particularly suitable to municipal water supply, wastewater treatment, agricultural irrigation, and drainage applications are the historical rate
meters that are treated in most hydraulic text. These include (1) weirs, (2) flumes, (3) orifice meters, and (4) Venturi meters.
Head, h, or upstream depth, commonly is used for the open channel devices such as
flumes and weirs. Either pressure, /?, head, /i, or differential head, A/z, or differential pressure, Ap, is used with tube-type devices, such as Venturi meters and orifice meters.
Venturi meters in pipelines and long throated flumes in open-channel flows are examples where the energy principles and the flow accountings mentioned above give good to
excellent computational results with minor dependency on empirical coefficients (Bos,
1989; Bos et al., 1991).
21.3.2 Installation Requirements
Special difficulties arise in applying velocity profile and friction accountings when insufficient pipe or channel exists upstream from a flow measuring device. This is needed to
ensure that predictable and acceptable velocity profiles are presented to the meter.
Frequently pipe or channel lengths can be significantly shortened by special structural
flow conditioners. These structural measures then become a design option. Some of these
are discussed below and in section 21.3.3
Designs for pipe discharges are well described in textbooks and in standard handbooks. The design difficulties center around selecting appropriate metering candidates for
accomplishing the measuring function and in providing an appropriate environment for
economical, accurate, and serviceable operation.
In the case of pipe flows, recommended straight pipe lengths, in terms of pipe diameter, are to be provided upstream of the meter to assure reasonable operating accuracy. These
lengths depend on the flow pattern presented to the meter primarily caused by valves and
pipe elbows upstream from the meter. The number and orientation of elbows greatly influence the circulation patterns and flow profile distortions presented to the meter.
Open-channel flow water measurement generally requires that the Froude number of
the approach flow be less than 0.5 to prevent wave action that would hinder or possibly
prevent an accurate head determination.
Energy concepts are used to describe Venturi meters in pipe flows based on the
Bernoulli equation in which part of the pipe forms a contracted throat that necessarily
changes the flow velocity and hence converts some of the static pressure to velocity head.
The decrease in static pressure is the basis for flow detection. A similar concept can be
applied to open channels. A historical version is the so-called Venturi flume (Brater and
King, 1982) that detects the change in water surface elevation between an upstream station and in a contracted section. However, this small change is difficult to accurately
detect, so the direct concept is not used. Rather, contractions are designed to be severe
enough to force critical flow velocities in the contracted section. Thus, only an upstream
head is needed to define the flow energy and flow area which can be converted to discharge rate. These are generally called critical-flow flumes. The flow condition where
only one head measurement is needed is called freeflow.
The critical-flow flumes themselves consist of those called long-throated flumes that
force parallel flow in the contracted, or control, section, called the throat, and those that
have curvilinear flow in the throat and are called short-throated flumes. The limiting
throat control section is the sharp-crested weir consisting of a thin plate. Thus, for
flumes and weirs one unique head value exists for each discharge, simplifying the calibration procedure.
However, if the downstream flow level submerges critical depth enough to affect the
upstream reading, the modular limit is exceeded, and free flow does not occur. When
exceeded, separate calibrations at many levels of submergence are then required, and two
head measurements are needed to measure flow. This condition generally is to be avoided in meter site design because it reduces the accuracy of the measurement and increases
the difficulty of flow determination. The modular limit for sharp-crested weirs, in practice, is less than zero, requiring full clearance of the overfall nappe of at least 3 cm, while
short-throated flumes can usually tolerate 65 percent to 70 percent submergence. Longthroated flumes can tolerate from 70 percent to 90 percent depending on flow conditions
and flume size.
Designing flumes for submerged flow beyond the modular limit decreases the accuracy of the flow measurement. Sometimes flumes and weirs can be overly submerged unintentionally by poor design, construction errors, structural settling, attempts to supply
increased delivery needs with increasing downstream heads, accumulated sediment
deposits, or weed growths. Sometimes use of the submerged range beyond the modular
limit is an economic compromise.
Approach flow conditions for pipes. Water measurement devices are generally calibrated with certain approach flow conditions. The same approach conditions must be
attained in field applications of measuring devices. Poor flow conditions in the area just
upstream of the measuring device can cause large discharge indication errors. For open
channels, the approaching flow should generally be subcritical. The flow should be fully
developed, mild in slope, and free of curves, projections, and waves.
Pipeline meters commonly require 10 or more diameters of straight pipe approach.
Fittings and combinations of fittings, such as valves and bends, located upstream from a
flow meter can increase the number of required approach diameters. Several references
(ASME, 1971; ISO, 1991) give requirements for many pipeline configurations and
meters. These are discussed in detail by Miller (1996).
Flow conditioning options. Many installations, especially in retrofit situations, do
not provide for sufficient lengths of straight pipe to remove velocity profile distortions
and swirl to an acceptable level. Therefore, the designer may need to use flow conditioners in combination with straight pipe lengths. Swirl sensitivity varies widely. Some meters
are particularly sensitive to swirl, such as the propeller and turbine meters. Magnetic flow
meters are somewhat less sensitive to radial velocities than single-path ultrasonic flow
meters. Venturi meters are less sensitive than orifice meters. For a swirl angle of 20 , the
discharge coefficient changes by about 1 percent for a Venturi meter with p = 0.32 ((3 is
the ratio of meter throat diameter to the pipe diameter) and about +10 percent for a similar orifice. Thus a swirl can increase the discharge through an orifice for the same differential head reading (Miller, 1996).
In pipeline flows, contractions can produce a central jet and also increase an incoming
swirl, while expansions tend to slow swirls and produce enough secondary flow to restore
flow profiles to some semblance of acceptability. These characteristics can modify the
straight pipe lengths needed or the type of flow conditioner to recommend (Miller, 1996).
Rough pipes also tend to reduce a swirl.
For flows, such as that encountered in sewage discharges and irrigation pipeline deliveries that originate from open channels, many of the tube-bundle types of flow conditioners can gather trash and cause maintenance problems. Many meter providers in these situations use fins or vanes that protrude from the wall and have sloped upstream edges that
shed trash. The vanes protrude about one-fourth of the pipe diameter into the flow, leaving the center core of the flow open. While these vanes can vary in number and length,
the logic being that the fewer the vanes the longer they should be in the direction of flow,
common configurations are four vanes that are about two or three pipe diameters long.
Vanes in themselves do not condition wall jets well. Field experience, has shown that troublesome flow profiles can be conditioned significantly by inserting an orifice into the
pipe. The orifice diameter is about 90 percent of the pipe diameter and is used to control
wall jets and force them to mix with the general flow. The orifice in itself tends to crossmix the jets and would appear to reduce spin. However, if the jets are symmetrical and an
initial swirl exists, orifices tend to increase the swirl. Inserting an orifice appears to be
supported by recent recommendations of Miller (1996) where it is stated: "To achieve a
fully developed profile, it is important that the flow be blocked or restricted close to the
wall, with the central core having the larger flow area."
The addition of vanes when space permits is recommended. Because orifices, in general, tend to force the flow to the pipe center while increasing spin, it appears best to place
the vanes upstream from the orifice. If they are placed downstream, the spin not only may
be increased, but the spinning central flow may not be touched by the vanes.
21.3.3
Flow conditioning in an irrigation delivery pipeline. As mentioned previously, measuring devices frequently must be installed in flow situations that are less than optimal. A field
example occurred in Arizona where a large pipe was used as an outlet to a secondary canal
and a single-path ultrasonic meter placed in it was subjected to flow profile distortions. The
pipe was about 0.75 m in diameter and delivered approximately 400 L/s. The flow rate
readout was unstable, with fluctuations varying by about 15 percent. The problem appeared
to be caused by slowly spiraling flow induced by the bottom jet from a partly open pipe
inlet gate and a 45 elbow. This is similar to two closely spaced pipe elbows that are not in
the same plane, which can cause a spiral flow pattern (ASME, 1971).
A successful attempt to modify the jet and cause it to cross mix so that the jet effects
and the strength of the spiral flow were reduced, was accomplished by inserting a large Pratio orifice in the pipe (Fig. 21.3). This consisted of an annular metal ring with the outside radius approximately that of the pipe and an inside diameter about 10 percent less, or
an orifice with (3 = 90 percent. The orifice was installed about three diameters downstream from the elbow. The slight increase in headless was compensated by increasing the
upstream gate opening. The orifice can be constructed by cutting notches from an appropriately sized piece of angle iron or aluminum and bending it to a polygon that approxi-
Control O gate
Ultrasonic flow
meter
Orifice plate
with large
opening
Spiral flow
Improved profile
FIGURE 21.3 An orifice plate with a large opening is used to condition a flow profile.
mates the circle diameter of the pipe interior. Some leakage around the ring is acceptable.
For propeller meters, additional vanes projecting from the walls may be needed to further
reduce spiral flow. These vanes would be placed upstream from the orifice. In this installation, the fluctuation was reduced to within about 3 percent.
Flow conditioning in channels. By analogy and using a minimum of 10 pipe diameters of a straight approach channel, open channel flow would require 40 hydraulic radii of
straight, unobstructed, unaltered approach, based on the calculation of hydraulic radius for
circular pipes being equal to one-fourth the pipe diameter, (Eq. 21.15). This would translate for very wide channels into approximately 40 times the flow depth. For narrow channels that are as deep as they are wide, this would compute to be about 13 channel depths
or top widths.
Other recommendations on approach channel criteria are presented by Bos (1989)
and USBR (1997). Major features of that criteria follow:
If the control width is greater than 50 percent of the approach channel width, 10 average approach flow widths of straight, unobstructed approach are required.
If the control width is less than 50 percent of the approach width, 20 control widths of
straight, unobstructed approach are required.
If upstream flow is below critical depth, a jump should be forced to occur. In this case,
30 measuring heads of straight, unobstructed approach after the jump should be provided.
If baffles are used to correct and smooth out approach flow, then 10 measuring heads
(10 /I1) should be placed between the baffles and the measuring station.
Approach flow conditions should be continually checked for deviation from these conditions as described in Bos (1989) and USBR (1997).
The baffles described above can become unacceptable maintenance problems in open
channels. Some field expediences are therefore described that have been found to work in
specific instances, but have not been studied for assured design generalizations. Nevertheless,
these constructions are but small extensions to currently accepted practices in pipe flows.
Applications for open-channel flow conditioners include abrupt channel turns, sluice
gate outflows, and channels downstream from a hydraulic jump. The abrupt turns may
benefit from floor and wall mounted vanes or fins. Based on pipe flow experience, and
assuming the channel is half of a closed conduit, these fins or vanes would probably be
about 10 percent to 15 percent of the channel depth.
As in pipe flow, wall jets that can develop downstream from sluice gates appear to
need treatment. This can be in the form of a structural angle bolted on the channel floor
and up the walls. Suggested size, based on the pipe flow analogy, is for the angle to be
about 5 percent into the channel flow depth. Whether the sidewalls need larger angles
when the channels are wide has not been tested.
21.3.4 Wave Suppression
Of special concern in open channels is wave suppression downstream from a sluice gate,
hydraulic jump, or an abrupt turn. Thus, the flow conditioners in channels have the additional task not present in pipe flows of surface wave suppression. Excessive waves in irrigation canals make reading sidewall gages difficult. These waves are usually caused by a
jet entry from a sluice gate or by a waterfall situation. The unstable surface can be 10-20
cm high and extend for tens of meters downstream.
Solidly attached
timber or concrete
FLOW
FIGURE 21.4 Wave suppresor design (From USBR, 1997).
them as shown in Fig. 21.5 would accomplish this without excessive obstruction.
Rounding the upstream edges will help shed trash, but may be less effective in suppressing waves. Observe in the sequence of drawings in Fig. 21.5 that the staggering is upward
in the downstream direction. Note that the next baffle slightly overlaps the horizontal flow
lines so that flow passing over the top of one baffle is not allowed to free-fall and start
another wave. Fig. 21.5 a-c illustrate the general behavior as the flow becomes less deep.
21.4
MEASUREMENTACCURACY
Accurate application of water measuring devices generally depends upon standard designs
or careful selection of devices, careful fabrication and installation, good calibration data
and adequate analysis. Also needed is proper user operation with appropriate inspection
and maintenance procedures. During operation, accuracy requires continual verification
that all measuring systems, including the operators, are functioning properly. Thus, good
training and supervision are required to attain measurements within prescribed accuracy
bounds. Accuracy is the degree of conformance of a measurement to a standard or true
value. The standards are selected by users, providers, governments, or compacts between
these entities. All parts of a measuring system, including the user, need to be considered
in accessing the system's total accuracy.
As mentioned above, a measurement system usually consists of a primary element,
which is that part of the system that creates what is sensed, and is measured by a secondary element. For example, weirs and flumes are primary elements. A staff gage is a
secondary element.
Designers, purchasers, and users of water measurement devices generally rely on standard designs and manufacturers to provide calibrations and assurances of accuracy. A few
water users and providers have the facilities to check the condition and accuracy of flow
measuring devices. These facilities have comparison flow meters and/or volumetric tanks
for checking their flow meters. These test systems are used to check devices for compliance with specification and to determine maintenance needs. However, maintaining facilities such as these is not generally practical.
Various disciplines and organizations do not fully agree on some of the definitions
related to measuring device specifications, calibration, and error analysis. Therefore, it is
important to verify that a clear and mutual understanding of the specifications, calibration
terminology, and the error analysis processes is established when discussing these topics
with others.
21.4.1 Definitions of Terms Related to Accuracy
Error. Error is the deviation of a measurement, observation, or calculation from the
truth. The deviation can be small and inherent in the structure and functioning of the system and be within the bounds or limits specified. Lack of care and mistakes during fabrication, installation, and use can often cause large errors well outside expected performance bounds. Because the true value is seldom known, some investigators prefer to use
the term uncertainty. Uncertainty describes the possible error or range of error which
may exist. Investigators often classify errors and uncertainties into spurious, systematic,
and random types.
Precision. Precision is the ability to produce the same measurement value within
given accuracy bounds when successive readings of a specific quantity are measured.
Precision represents the maximum departure of all readings from the mean value of the
readings. Thus, a single observation of a measurement cannot be more accurate than
the inherent precision of the combined primary and secondary precision. It is possible to
have good precision of an inaccurate reading. Thus, precision and accuracy differ.
Spurious errors. Spurious errors are commonly caused by accident, resulting in false
data. Misreading and intermittent mechanical malfunctions can cause discharge readings
well outside of expected random statistical distribution about the mean. Spurious errors
can be minimized by good supervision, maintenance, inspection, and training.
Experienced, well-trained operators are more likely to recognize readings that are significantly out of the expected range of deviation. Unexpected spiral flow and blockages of
flow in the approach or in the device itself can cause spurious errors. Repeating measurements does not provide information on spurious error unless repetitions occur before and
after the introduction of the error. On a statistical basis, spurious errors confound evaluation of accuracy performance.
Systematic errors. Systematic errors are errors that persist and cannot be considered
random. Systematic errors are caused by deviations from standard device dimensions,
anomalies to the particular installation, and possible bias in the calibration. Systematic
errors cannot be removed or detected by repeated measurements. They usually cause persistent error on one side of the true value. The value of a particular systematic error for a
particular device may sometimes be considered as a random error. For example, an installation error in the zero setting for a flume might be + 1 mm for one flume and 2 mm for
another. For each flume the error is systematic, but for a number of flumes it would be a
random error.
Random errors. Random errors are caused by such things as the estimating required
between the smallest division on a head measurement device and water surface waves at
a head measuring device. Loose linkages between parts of flowmeters provide room for
random movement of parts relative to each other, causing subsequent random output
errors. Repeated readings decrease the average expected error resulting from random
errors by a factor of the square root of the number of readings.
Total error. Total error of a measurement is the result of systematic and random errors
caused by component parts and factors related to the entire system. Sometimes, error limits of all component factors are well known. In this case, total limits of simpler systems
can be determined by computation (Bos et al., 1991). In more complicated cases, it may
be difficult to confidently combine the limits. In this case, a thorough calibration of the
entire system as a unit can resolve the difference. In any case, it is better to do error analysis with data where entire system parts are operating simultaneously and compare discharge measurement against an adequate discharge comparison standard.
Expression of errors. Instrument errors are usually expressed by manufacturers as
either a percent of reading or a percent of full scale. The secondary devices based on electronic outputs are more frequently expressed in terms of percent full scale. The designer
must be aware that a probable error value of say 1 percent full-scale can exceed 10
percent for small value readings on the output device. When used with weirs, for example, the head reading of hl 5 in the weir equation can increase this 10 percent head measurement error to a 15 percent flow measurement error.
output. Other possible working standards are weights, volume containers, and stopwatches. More complicated devices are used, such as surveyors' levels, to check weir staff gage
zeros. Dead weight testers and electronic standards are needed to check and maintain
more sophisticated and complicated measuring devices, such as acoustic flow meters and
devices that use pressure cells to measure head.
For further measurement assurance and periodic checking, water users and organizations may keep secondary standards. Secondary standards are used to maintain integrity
and performance of working standards. These secondary standards can be sent to government laboratories, one of which is the National Bureau of Standards in Washington, D.C.,
to be periodically certified after calibration or comparison with accurate replicas of primary standards. Primary standards are defined by international agreement and maintained
at the International Bureau of Weights and Measurements in Paris, France.
Depending on accuracy needs, each organization should trace their measurement performance up to and through the appropriate level of standards. For example, turbine
acceptance testing, such as in the petroleum industry, might justify tracing to the primary
standards level.
I. Operating requirements
j. Ability to pass sediment and debris
k. Longevity of device for given environment
1. Maintenance requirements
m. Construction and installation requirements
n. Device standardization and calibration
o. Field verification, troubleshooting, and repair
p. User acceptance of new methods
q. Vandalism potential
r. Impact on environment
Accuracy requirements. The desired accuracy of the measurement system is an important consideration in the selection of a measurement method. Most water measurement
installations, including the primary and secondary devices, can produce accuracies of 5
percent. Some systems are capable of 1 percent under laboratory settings. However, in
the field, maintaining such accuracies usually requires considerable expense or special
effort in terms of construction, secondary equipment, calibration in-place, and stringent
maintenance. Selecting a device that is not appropriate for the site conditions can result in
a nonstandard installation of reduced accuracy, sometimes exceeding 10 percent.
Accuracies are frequently reported that relate only to the primary measurement method
or device. However, many methods require secondary measurement equipment that produces the actual readout. This readout equipment typically increases the overall error of
the measurement.
Cost. The cost of the measurement method includes the cost of the device itself, the
installation, secondary devices, operation, and maintenance. Measurement methods vary
widely in their cost and in their serviceable life span. Measurement methods are often
selected based on the initial cost of the primary device with insufficient regard for the
additional costs associated with providing the desired records of flows over an extended
period of time.
Legal constraints. Governmental or administrative water board requirements may
dictate the water measurement devices or methods. Water measurement devices that
become a standard in one geographic area may not necessarily be accepted as a standard
elsewhere. In this sense, the term "standard" does not necessarily signify accuracy or
broad legal acceptance. Many water agencies require certain water measurement devices
used within their jurisdiction to conform to their standard for the purpose of simplifying
operation, employee training, and maintenance.
Flow range. Many measurement methods have a limited range of flow conditions for
which they are applicable. This range is usually related to the need for certain prescribed
flow conditions which are assumed in the development of calibrations. Large errors in
measurement can occur when the flow is not within this range. For example, using a bucket and stopwatch for large flows that engulf the bucket is not very accurate. Similarly,
sharp-edged devices, such as sharp-crested weirs, typically do not yield good results with
large channel flows. These are measured better with large flumes or broad-crested weirs,
which in turn are not appropriate for trickle flows.
Certain applications have typical flow ranges. Irrigation supply monitoring seldom
demands a low-flow-to-high-flow range above about 30, while this range on natural
stream flows may exceed 1000.
In some cases, secondary devices can limit the practical range of flow rates. For example, with devices requiring a head measurement, the accuracy of the head measurement
from a visually read wall gage may limit the measurement of low flow rates. For some
devices, accuracy is based on percent of the full-scale value. While the resulting error may
be well within acceptable limits for full flow, at low flows, the resulting error may become
excessive, limiting the usefulness of such measurements. Generally, the device should be
selected to cover the desired range. Choosing a device that can handle an unnecessary
large flow rate may result in compromising measurement capability at low flow rates, and
vice versa. This choice depends on the objective of the measurement. For example, in irrigation practice, usually choose a device that can measure the most common flow range at
the expense of poorly measuring extremes, such as flood flows. For urban drainage, the
flood peak may be important.
For practical reasons, different accuracy requirements for high and low flows may be
chosen. This is reasonable when an annual total is the primary goal and the low flows contribute a small percentage to that total. Also, if the inaccurate low flow readings are truly
a random error then this error approaches zero with large accumulations of readings. Thus
the designer needs to know if management decisions are made from individual readings
or from long term averages.
Headloss. Most water measurement devices require a drop in head. On retrofit installations, for example, to an existing irrigation project, such additional head may not be
available, especially in areas that have relatively flat topography. On new projects, incorporating additional headloss into the design can usually be accomplished at reasonable
cost. However, a tradeoff usually exists between the cost of the device and the amount of
headloss. For example, acoustic flow meters are expensive but require little headloss.
Sharp-crested weirs are inexpensive but require a relatively large headloss. The head loss
required for a particular measuring device usually varies over the range of discharges. In
some cases, head needed by a flow measuring device can reduce the capacity of the channel at that point.
Adaptability to site conditions. The selection of a flow measuring device must
address the site of the proposed measurement. Several potential sites may be available for
obtaining a flow measurement. The particular site chosen may influence the selection of
a measuring device. For example, discharge in a canal system can be measured within a
reach of the channel or at a structure such as a culvert or check structure. A different
device would typically be selected for each site. The device selected ideally should not
alter site hydraulics so as to interfere with normal operation and maintenance. Also, the
shape of the cross-sectional flow area may favor particular devices.
Adaptability to variable operating conditions. Flow demands for most water delivery systems usually vary over a range of flows and flow conditions. The selected device
must accommodate the flow range and changes in operating conditions, such as variations
in upstream and downstream head. Weirs or flumes should be avoided if downstream
water levels can, under some conditions, cause excessive submergence. Also, the information provided by the measuring device should be conveniently useful for the operators
performing their duties. Devices that are difficult and time consuming to operate are less
likely to be used and are more likely to be used incorrectly.
In some cases, water measurement and water level or flow control are desired at the
same site. A few devices are available for accomplishing both (e.g., constant-head orifice,
vertically movable weirs, and Neyrpic flow module; Bos, 1989). However, separate measurement and control devices are typically linked for this purpose and usually can exceed
the performance of combined devices in terms of accuracy and level control, if care is
exercised to assure that the separate devices are compatible and achieve both functions
when used as a system.
Type of measurements and records needed. An accurate measure of instantaneous
flow rate is useful for system operators in setting and verifying flow rate. However,
because flow rates change over time, a single (instantaneous) reading may not accurately
reflect the total volume of water delivered. Where accounting for water volume is desired,
a method of accumulated individual flow measurements is needed. Where flows are
steady, daily measurements may be sufficient to infer total volume. Most deliveries, however, require more frequent measurements. Meters that accumulate total delivered volume
are desirable where water users take water on demand. Totalizing and automatic recording devices are available for many measuring devices. For large structures, the cost for
water-level sensing and recording hardware is small relative to the structure cost. For
small structures, these hardware costs remain about the same and thus become a major
part of the measurement cost, and may often exceed the cost of the primary structure
itself.
Many water measuring methods are suitable for making temporary measurements
(flow surveys) or performing occasional verification checks of other devices. The method
chosen for such a measurement might be quite different from that chosen for continuous
monitoring. Although many of these flow survey methods are suited for temporary operation, the focus here is on methods for permanent installations.
Operating Requirements. Some measurement methods require manual labor to
obtain a measurement. Current metering requires a trained staff with specialized equipment. Pen-and-ink style water-stage recorders need operators to change paper, add ink,
and verify proper functioning. Manual recording of flows may require printed forms to be
manually completed and data to be accumulated for accounting purposes. Devices with
manometers require special care and attention to assure correct differential-head readings.
Automated devices, such as ultrasonic flowmeters and other systems that use transducers
and electronics, require operator training to set up, adjust, and troubleshoot. Setting gatecontrolled flow rates by simple canal level references or by current metering commonly
requires several hours of waiting between gate changes for the downstream canal to fill
and stabilize. However, if a flume or weir is installed near the control gate, that portion of
the canal can be brought to the stable, desired flow level and measured flow rate in a few
minutes, and the canal downstream of the flume or weir can then fill to the correct level
over a longer time without further gate adjustments. Thus, the requirements of the operating personnel in using the devices and techniques for their desired purposes must be
considered in meter selection.
Some measuring devices may inherently serve an additional function applicable to the
operation of a water supply system. For example, weirs and flumes serve to hydraulicalIy isolate upstream parts of a canal system from the influence of downstream parts. This
occurs for free overfall weirs and flumes flowing below their modular limits. Acoustic,
propeller, magnetic, and vortex-shedding flowmeters do not provide this function without
additional structural measures such as a downstream overfall. If these meters are used, and
the isolation function is desired, then the designer should be made aware of the requirement and provide a free overfall. Isolating the influence of upstream changes from affecting downstream channels, is less easily accomplished. However, it can be partly implemented with orifices that have a differential head that is large compared to the upstream
fluctuations.
The designer should be aware that a sharp-crested weir overfall requires a relatively
high head drop and may need to be excessively wide to provide the isolation function
with low absolute head drop. While a long board can be used downstream from a propeller meter to provide the necessary width of flow that will pass a required quantity of
water at small head, that small head, and the crude board would not be well suited for
measuring flow rate.
The designer may wish to take advantage of broad-crested weir behavior and provide
a thick crest that can withstand in excess of about 80 percent submergence, which usually translates into low absolute head loss. When used with a propeller meter, for example, the broad-crested weir need not be well defined and can be economically installed
(Replogle, 1997).
Ability to pass sediment and debris. Canal systems often carry a significant amount
of sediment in the water. Removal of all suspended solids from the water is usually prohibitively expensive. Thus, some sediment will likely be deposited anywhere the velocities are reduced, which typically occurs near flow measuring structures. Whether this sediment causes a problem depends on the specific structure and the volume of sediment in
the water. In some cases, this problem simply requires routine maintenance to remove
accumulated sediment; in others, the accumulation can make the flow measurement inaccurate or the device inoperative. Sediment deposits can affect approach conditions and
increase approach velocity in front of weirs, flumes, and orifices. Floating and suspended
debris such as aquatic plants, washed-out bank plants, and fallen tree leaves and twigs can
plug some flow measurement devices and cause significant flow measurement problems.
Many of the measurement devices which are successfully used in closed conduits (e.g.,
orifices, propeller meters, and so on) are not usable in culverts or inverted siphons because
of debris in the water. Attempting to remove this debris at the entrance to culverts is an
additional maintenance problem.
Flumes, especially long-throated flumes, can be designed to resist sedimentation. The
design options available are to select a structure shape that will maintain velocities that
assure erosion of sediments, or at least continued movement of incoming sediments
through the flume, at important flow rates. In large broad-crested weirs (a class of longthroated flumes) for capacities greater than 1 m3/s per m of flume width, velocities greater
than 1 m/s can be achieved for the upper 75 percent of the flow range, and is usually erosive enough to maintain flume function even for high-sediment bed loads. At the lower
flow ranges and for heavy sediment bed loads, deposition is likely and frequent maintenance may be required.
Trapezoidal sections tend to retain low velocities into the upper ranges of flow and
are less sediment worthy. Long-throated flumes with flat bottoms throughout and side
contractions maintain a high velocity for 0.5 m3/s per m width, and higher, but must
have throat lengths that are 2 to 3 times the throat width in order to be accurately computable. The sediment worthiness of a flume design depends more on these absolute
velocities than on whether the flume floor is flat throughout or raised as in a broadcrested weir. This prompts the designer to select shapes that can provide these velocities. One suggestion for broad-crested weirs in a fixed sized channel is to construct a
false floor in the head gage area to increase the velocity there and prevent changes in
area of flow there. Also, sediments can accumulate in the upstream channel to a depth
of the false floor without affecting the function of the flume. This can extend the time
between mandatory channel cleaning.
Device environment. Any measurement device with moving parts or sensors is subject to failure if it is not compatible with the site environment. Achieving proper operation
and longevity of devices is an important selection factor. Very cold weather can shrink
moving and fixed parts differentially and solidify oil and grease in bearings. Water can
freeze around parts and plug pressure ports and passageways. Acidity and alkalinity in
water can corrode metal parts. Water contaminants such as waste solvents can damage
lubricants, protective coatings, and plastic parts. Mineral encrustation and biological
growths can impair moving parts and plug pressure transmitting ports. Sediment can
abrade parts or consolidate tightly in bearing and runner spaces in devices such as propeller meters.
Measurement of wastewater and high sediment transport flow may preclude the use of
devices that require pressure taps, intrusive sensors, or depend upon clear transmission of
sound through the flow. Water measurement devices that depend on electronic devices and
transducers must have appropriate protective housings for harsh environments. Improper
protection against the site environment can cause equipment failure or loss of accuracy.
Maintenance requirements. The type and amount of maintenance varies widely with
different measurement methods. For example, current metering requires periodic maintenance of the current meter itself and maintenance of the meter site to assure that is has a
known cross section and velocity distribution. When the flow carries sediment or debris,
most weirs, flumes, and orifices require periodic cleaning of the approach channel. As
mentioned above, design and meter selection can mitigate the maintenance problems with
sediments, but are not likely to eliminate them. Electronic sensors need occasional maintenance to ensure that they are performing properly. Regular maintenance programs are
recommended to ensure prolonged measurement quality for all types of devices.
Construction and installation requirements. In addition to installation costs, the difficulty of installation and the need to retrofit parts of the existing conveyance system can
complicate the selection of water measurement devices. Clearly, devices that can be easily retrofitted into the existing canal system are much preferred because they generally
require less down time, and usually present fewer unforeseen problems.
Device standardization and calibration. A standard water measurement device infers
a documented history of performance based on theory, controlled calibration, and use. A
truly standard device has been fully described, accurately calibrated, correctly constructed, properly installed, and sufficiently maintained to fulfill the original installation
requirements and flow condition limitations. Discharge equations and tables for standard
devices should provide accurate calibration. Maintaining a standard device usually only
involves a visual check and measurement of a few specified items or dimensions to ensure
that the measuring device has not departed from the standard. Many standard devices have
a long history of use and calibration, and thus are potentially more reliable. Commercial
availability of a device does not necessarily guarantee that it satisfies the requirements of
a standard device.
When measuring devices are fabricated onsite or are poorly installed, small deviations from the specified dimensions can occur. These deviations may or may not affect
the calibration. The difficulty is that unless an as-built calibration is performed, the
degree to which these errors affect the accuracy of the measurements is largely
unknown. All too frequently, design deviations are made under the misconception that
current metering can be used to provide an accurate field calibration. In practice, calibration by current metering to within 2 percent is difficult to attain. An adequate calibration for free-flow conditions requires many current meter measurements at several
discharges. Changing and maintaining a constant discharge for calibration purposes is
often difficult under field conditions.
Field verification, troubleshooting, and repair. After construction or installation of a
device, some verification of the calibration is generally recommended. Usually, the meth-
ods used to verify a permanent device (e.g., current metering) are less accurate than the
device itself. However, this verification simply serves as a check against gross errors in
construction or calibration. For some devices, errors occur as components wear and the
calibration slowly drifts away from the original. Other devices have components that simply fail, that is, you get the correct reading or no reading at all. The latter is clearly preferred. However, for many devices, occasional checking is required to ensure that they are
still performing as intended. Selection of devices may depend on how they fail and how
easy it is to verify that they are performing properly.
User acceptance of new methods. Selection of a water measurement method must
also consider the past history of the practice at the site. When improved water measurement methods are needed, proposing changes that build on established practice are generally easier to institute than radical changes. It can be beneficial to select a new method
that allows conversion to take place in stages to provide educational examples and demonstrations of the new devices and procedures.
Vandalism potential Instrumentation located near public access is a prime target for
vandalism. Where vandalism is a problem, measurement devices with less instrumentation, or instrumentation that can be easily protected, are preferred. When needed, instrumentation can be placed in a buried vault to minimize visibility.
Impact on environment. During the selection of a water measurement device, consideration must be given to potential environmental impacts. Water measurement devices
vary greatly in the amount of disruption to existing conditions that is needed for installation, operation, and maintenance. For example, installing a weir or flume constricts the
channel, slows upstream flow, and accelerates flow within the structure. These changes in
the flow conditions can alter local channel erosion, local flooding, public safety, local
aquatic habitat, and movement of fish up and down the channel. These factors may alter
the cost and selection of a measurement device.
21.5.3 Selection Guidelines
Selection of a water measurement method can be a difficult, time-consuming design
process if one were to formally evaluate all the factors discussed above for each measuring device. This difficulty is one reason that standardization of measurement devices within water agency jurisdictions is often encouraged by internal administrators. However,
useful devices are sometimes overlooked when devices similar to previous purchases are
automatically selected. Therefore, some preliminary guidance on selection is offered to
the designer so that the number of choices can be narrowed down before a more thorough
design analysis of the tradeoffs between alternatives is performed.
Short list of devices based on application. The list of practical choices for a water
measuring device is quickly narrowed by site conditions because most devices are applicable to a limited range of channel or conduit conditions. Economics also limits applicable
devices. For example, few irrigation deliveries to farms can justify expensive acoustic
meters. Likewise, using current meters for manual flow measurement in a channel is
appropriate for intermittent information but is usually too labor intensive for use on a continuous basis. Table 21.1 provides a list of commonly used measurement methods that are
considered appropriate for each of several applications. Table 21.2 provides an abbreviated table of selection criteria and general compliance for categories of water measurement
devices. The symbols (+), (O), and () are used to indicate relative compliance for each
selection criterion. The (+) symbol indicates positive features that might make the device
Sharp
Crested
Broad
Crested
LongThroated
Short Submerg
Throated Orifice
Criteria
Weirs
Weirs
Flumes
Devise
Current
Metering
Accuracy
O
O
O
O
O
+
Cost
O
O
+
O
O
+
+
+
Flows > 5 m-Vs
O
O
+
+
+
+
Flows < 0.25 mVs
O
O
+
Flow span
+
O
O
O
+
Headloss
+
O
O
Site Condition
+
Lined canal
+
+
O
O
O
Unliner canal O
+
O
O
O
O
O
Short, full pipe NA
NA
+
NA
NA
NA
NA
NA
Closed conduit NA
NA
NA
NA
NA
NA
NA
NA
Measurement Type
+
+
+
Rate
+
+
+
+
+
Volume
O
Sediment
+
+
Sediment pass
O
O
O
O
2
+
+
+
+
Debris pass
+
Longivity
+
+
+
Moving parts +
+
+
O
O
+
+
+
+
Electr. requir. +
O
O
+
+
+
+
Maintenance
O
O
+
+
+
+
+
Construction
O
O
+
+
+
Field verify
O
O
+
+
+
Standarization
O
O
O
Source: Adapted from USER (1997).
Symbols 1, O, 2 are used as relative indicators comparing application of the listed water measuring device to the listed criteria
Symbol V denotes that situability varies widely.
Symbol NA denotes "not applicable" to criteria
"Venturi,
orifice, pilot tube, etc.
f
Propeller meters, turbine meters, paddle wheel meters, etc.
(Pipes)
(Pipes)
4-
O
+
O
O
O
O
+
NA
NA
O
+
V
V
O
O
+
NA
NA
V
NA
NA
O
+
+
V
+
O
O
O
+
+
Acoustic
Doppler
Acoustic Acoustic
Transonic Transon.
Pipe
(Pipes) (Pipe, 1 Path) (Multipath)
+
O
O
+
O
O
O
+
O
+
+
+
NA
NA
NA
NA
NA
NA
+
+
O
O
O
O
+
+
O
O
O
O
+
O
O
O
O
O
+
+
+
O
O
O
O
2
O
O
O
attractive from the standpoint of the associated selection criteria. A (-) symbol indicates
negative aspects that might limit the usefulness of this method based on that criterion. A
(O) indicates no strong positive or negative aspects in general. A (V) means that the suitability varies widely for this class of devices. The letters NA mean that the device is not
applicable for the stated conditions. A single negative value for a device does not mean
that the device is not useful or appropriate, but other devices would be preferred for those
selection criteria.
Vortex-shedding flow meters are not specifically rated in this grouping. They are
expected to compete with orifice meter applications. They generally offer less head loss
that orifice meters and can cover a wider discharge range for a particular installation.
Although they have been around for many years, they have only recently been offered in
a configuration that makes them competitive with orifice meters, which they are generally expected to replace because they can produce less pipe head loss. Open-channel applications for vortex-shedding meters are not considered practical.
Next Page
B = WIDTH OF CULVERT
= (D) FOR PIPE
L = DISTANCE TO DESIRED DEPT.H
MEASURED IN MULTIPLES OF (B)
Y0 = BRINKDEPTH
Y A = AVERAGE DEPTH OF FLOW
FIGURE 20.2 Average depth for abrupt exapnsion below rectangular culvert outlet. (From Cony et al., 1975)
D = DIAMETER OF CULVERT
L = DISTANCE TO DESIRED DEPTH
MEASURED IN MULTIPLES OF (D)
Y0 = BRINKDEPTH
Y A = AVERAGE DEPTH OF FLOW
FIGURE 20.3 Average depth for abrupt expansion below circular culvert outlet. (From Corry et al., 1975).
Y0 = brink depth
D = height of box culvert
B = width of barrel
Q = discharge
TW = tail water depth
TW/D
FIGURE 20.4 Dimensionless rating curves for the outlets of rectangulars culverts on horizontal and mild slopes (From Simons et al; 1970)
Y0 = brink depth
D = dia. of culvert
TW = tail water depth
TW/D
FIGURE 20.5 Dimensionless rating curve for the outlets of circular culverts on horizontal and
mild slopes (From Simons, et al., 1970)
C ftAand
A T^W ^O
d,c = 5
Solution.
-"
&-&-*<.-<@ti-4ffi-">'
TW
and *** O. From Fig. 20.4, V 0 / D = 0.68; therefore, y0 and V0 can be deter-
PLAN
HYDRAULICGRADELINE
WATERSURFACE
ELEVATION
ATUM
FIGURE 20.6 Definition sketch. (From Cony et al., 1975)
(V2 V 2^
= Cc
C \ 2- L
(2g
2gJ
(20 1 ^
(ml)
(V2
V2}
= CC \^- $- L
'(2g
2g)
(20 2)
2s)
(2
H
HLE
where transition loss coefficients are listed in Table (20.1) for the transition types
(USAGE, 1970).
The design procedure for gradual transitions is as follows (Corry et al 1975):
1. Use Manning's equation to compute y4 = yn and V4, knowing Q, S0, n and the outlet
geometry.
2. Compute the critical depth using yc = ^1 (QIW)m. For box culvert, kv = 0.315 where
Q is in cfs and W is the box culvert width in ft. In SI units ^1 = 0.467 with Q in mVs
and W in m. Compare yn and yc to ensure subcritical flow.
3. For the chosen transition type, use Table 20.1 to obtain the transition loss coefficients
C0 and Ce.
4. Choose V3 = 1.1 yc to have a culvert with a flow depth conservatively above. yc
5. Compute the culvert width using the following energy equation, ignoring the headloss
caused by friction:
V42
PV32 V42 ^l
V3 2
+ C
73 +-f Vv3 ++Z74 + vA;4 ++ +
C ^ - -J-Z
V3 = (W3V3 where V3 = 1.1 yc = IAk1 (QfW3)2 For box culverts and V4 = QtWj4.
Neglecting Z3 - Z4 for a short reach and a mild bottom slope,
^RK^'tei]''--'
><+fei 4*" '" (ir ^ [(^. iut,w-ii ij" -c->
TABLE 20.1 Transition Loss Coefficients
Transition Type
Contraction
Cr
Expansion
Ce
Warped
Cylindrical quadrant
Wedge
Straight line
Square end
1.10
0.15
0.30
0.30
0.30
0.20
0.25
0.50
0.50
0.75
A^-*-^+>$-
n A 1
n2/3
/^
"\ ^ 1 \2/3
^) ^'^-('-"^^s0-*)
x2/3
and
^J
W3 = ([BV[A])
(20.3)
where
=*+][wh]-c>
and
[B] = L 1
2/3+
/i2/3 i
Tfefi ( 1 -^
/y2
y2\
y2
>+y>+t+c^-^rz>+y>+w
(20
-4)
7. If the amount of backwater (yr yn) exceeds a preferred or required freeboard, then
select a larger culvert width and calculate y3 using Eq. (20.3) given in Step 5 (return
to Step 5).
8. If the culvert width is acceptable, use the flow conditions to compute the transition
length using a 4.5:1 flare (USAGE, 1970):
LT = 4.5 (^
(20.5)
LT =4.5^^
(20.6)
or
9. Compute the water surface profile through the structure using a standard step backwater analysis. This will include an evaluation of friction losses.
Example 2. Determine the dimensions for a culvert and gradual transition needed for a 3
m-wide rectangular flood control channel at a slope of 0.001 m/m. The culvert length is
30.5 m, and is to convey 8.5 m3/s. Use n = 0.02.
Solution:
Step 1. Assuming normal depth at Section 4 (refer to Fig. 20.6), then use Manning's
equation to compute V4 = yn = 1.99 m and Vn = 1.42 m/s.
D ^2/3
( S2^3
-^ = 0.467 -^- = 0.935 m.
Step 7. Because ^1 is 1.29 m above the normal depth and because only 0.6 m free board
is available, we go back to Step 5 and compute ^3. Assuming that the culvert
width is 1.45 m: Repeat step 5,
LT
CENTERLINE
ALONGWALL
FLOW
a. SCHEMATICPROFILE
b. PLAN
FIGURE 20.7 Supercritical inlet transition for rectangular channel.
(From USAGE, 1970)
a. Compute L = (W1- W2)/2 tan 9W
b. Find P1, y21 V 1 , and Fr1 from
* *
tan B1 V(I + 8Fr12 sin2 B1 - 3)
1
^= 2 ^ + Vl+8Fr1CV-I
(2
'7)
and
(20.8)
^
I](^+ lfl
2 = f4^ ? -UJL * I2^2JUi
JU
JJ
(20.9)
and
FIGURE 20.8 Supercritical inlet transition design curves for rectangular channels.
(USAGE, 1970)
Fr = -L= =- 19'2
= 3.05m
VgAfT
V 32.2(15.6)712.7
Step 2. Because (T + Bw )/2 = (12.7 + 6) /2 = 9.4 ft, use W1 = 10 ft, a rectangular chan-
nel. Use Manning's equation to compute yn = 1.54 ft and Vn = 19.5 ft/s, then
compute the Fr:
Fr = -^= = . 19'5
=2.77.
Vg^ V32.2(1.54)
Step 3. Assume a trial culvert width W2 = 5 ft.
Step 4. Try 8W = 14 for Fr = 2.8:
L = ( 1 0 - 5 ) / 2 tan 14 = 10ft
P1 = 35, V2^1 = 1.8, Fr2 =1.8
Step 5. Because y3 = 4.4 ft. and W3 = 5 ft, a 5 X 5 box culvert will be satisfactory.
20.2
ENERGYDISSIPATIONFORCULVERTSANDCHANNELS
DATUM
FIGURE 20.10 Definition sketch basin transition. (From Corry et al., 1975)
W <W +
(20 10)
'
'
^T 3Fr0+1
'
where LT = (Z0 Z1)AS7, and the right-hand side is the limit that flares naturally in the
slope distance L.
c. Compute Y1 using the following equation derived from the energy equation from the
culvert outlet brink to the basin (Sec. 1 in Fig. 20.10). Use V1 = Q/ y^W^ to determine^ and then V1:
r
i1'2
Q = J^B [Zg (Z* - ZI + % - *) + ^02J
( 2 O-U)
d. Compute the Fr
Fr 1 =A
vw
e. Compute y2 using Eq. (20.12) for the hydraulic jump basin:
J2 = y[vi + 8 F r J - I J
(20.12)
(20.13)
where
^o
FIGURE 20.11 Length of jump in terms Of^ 1 , rectangular channel. (From Corry et al.,
1975)
FIGURE 20.12 Example problem 18.2.1 hydraulic jump basin. (Not to scale).
Solution:
Step 1. The brink depth and velocity are given: Fr0 = 4.0.
Step 2. The tailwater depth is computed assuming normal depth. Manning's equation
is solved to obtain yn = 0.57 m tailwater depth and Vn = 4.85 m/s. (Refer to
Fig. 20.12)
Step 3. Assuming a basin elevation OfZ 1 = 25.9 m:
a. Select W8 = 3 m and S5 = S7 = 0.5
b. Check W8 using Eq. 20.10, where W0 = 3 m,
W8 = 3 < 3 + [2L7 V5JTTJ(3Fr0I OK.
c. Compute ^1 using Eq. (20.11)
1/2
11.8 = V1(S) 2(9.81)(30.5 - 25.9 + 0.457 - V1) + (8.47)2
Z =
20.2.2.1 Saint Anthony Falls stilling basin. The Saint Anthony Falls, (SAF) stilling basin
is a generalized design based upon model studies conducted by the U.S. Soil Conservation
Service at the St. Anthony Falls Hydraulic Laboratory, University of Minnesota. Figure
20.13 illustrates the SAF stilling basin design which is recommended for small structures
such as spillway outlet works, and for canals where the Fr ranges from 1.7 to 17 (at the dissipator entrance). Through the use of chute blocks, baffle or floor blocks, and an end sill,
the basin length is about 80 percent of the free hydraulic jump length.
The design procedure for SAF basins is as follows (Corry et al 1975):
1. Choose basin configuration and flare dimension, Z. (Refer to Fig. 20.13)
2. Use the design procedure presented in Sec. 20.2.1 for supercritical expansions into
hydraulic jump basins to determine basin width (W^), elevation (Z1), length (Le), total
length (L), incoming depth Cy1), incoming Fr (Fr1), and jump height (y2). Steps 3e and
3f in Sec. 20.2.1 are modified; for Step 3e, determine y2 using the sequent depth y.\
(20.14)
(20.15)
(20.16)
(20.17)
RECTANGULAR BASIN
HALF-PLAN
EQUATION NUMBER
FLARED BASIN
HALF-PLAN
0-90
45 PREFERRED
SIDEWALL
CHUTEBLOCK
FLOORORBAFFLEBLOCKS ENDSILL
VARIES
Height:
H1 = yl
Width:
Number:
Adjusted:
(20.18)
FIGURE 20.14 Example 4 St. Anthony Falls stilling basin. (Not to scale)
b. Next, use Eq. (20.15) for Fr1 = 1.7 to 5.5 to compute J2:
Because y2 > Tw = 0.57, we can lower the elevation of the basin. Use Z1 =
27.9 m with W8 = 3 m and ST = S5 = 0.5. WB is OK and no flare is used.
c. Compute J1 using the energy Eq. (20.11). The basin has been lowered so now
J1 is not J0, the brink depth.
Q = * WB\2g(ZQ - Z1 + J0 - ^1) + V02]
r
nl/2
11.8 = ^(312(9.81X30.5 - 27.9 - 0.457 - J1) + (8.47)2
FREEJUMP
TYPE IV
TYPE Il BASIN
FIGURE 20.15 U.S. Bureau of Reclamation Type II basin. (Corry et al, 1975)
ENDSILL
BAFFLE PIERS
BAFFLE PIERS
END
FIGURE 20.16 Height of baffle piers and end sill (Type III basin). (From U.S. Bureau of Reclamation, 1987)
Use dimensions in Fig. 18.7 to determine chute block dimensions and spacing.
Use Fig. 18,7 to determine the end sill height.
Example 5. Determine the dimensions of U.S. Bureau of Reclamation's Type II basin for
the supercritical flow expansion described in Example 3.
Solution:
Step 1. Use the design procedure for a supercritical flow expansion into a hydraulic jump
basin. (Refer to Examples 3 and 4).
Given V0 = 8.47 m/s, V0 = 0.457 m, and Fr0 = 4.0. (Figure 20.17)
The tailwater depth is Tw = yn = 0.57 m, and Vn = 4.85 m/s.
Assume Z1 = Z0 = 30.5 m, compute y2
y2 = C1J1[Vl + 8Fr^- l]/2
= 1.1(0.457/Vl + 8(4)2 - IJ/2
= 2.6m
y2 > Tw (2.6 > 0.57). Therefore we need to lower elevation Z1 of the basin floor.
a. Use a basin floor elevation of Z1 = Z2 = 25.76 m, with W8 = 3 m, ST =
S5 = 0.5.
b. W8 is OK, no flare.
c. Compute V1 using energy Eq. (20.11).
r
T'2
Q = ViW8^g(Z0 - Z1 + y0 - V1) + V02J
r
i172
11.8 = Vj(3) 2(9.81)(30.5 - 25.76 + 0.457 - V1) + 8.472
Z3
Figure 20.18 illustrates the U.S, Bureau of Reclamation's Type VI impact-type energy dissipator, which can be used with culverts. This basin is contained in boxlike structures
requiring no tail water for operation. The structures can be used for open channels as well
as culverts, and the basin can be used at sites where the entrance velocity to the basin does
not exceed 50 ft/s, and the discharge is less than 400 cfs. This dissipator should not be
used if the buildup of debris or ice can cause substantial clogging.
The design procedure is as follows (Corry et al, 1975).
1. Compute the flow area at the end of the culvert using the maximum design discharge
and velocity. Compute the equivalent depth of flow entering the dissipator from the
culvert as
Je = (f)"2
(20.19)
where A is the cross-sectional area of flow in the culvert. This converts the cross-sectional area of flow of a pipe into an equivalent rectangular cross section with a width
twice the depth of flow. The culvert preceding the dissipator can be open, closed, or
have any cross section. This approach ignores the size and shape of the culvert entirely except for the determination of flow entering the dissipator.
2. Compute Fr and the energy at the end of the culvert H0:
V2
" = ^ + 2 F -
Then use Figure 20.19 to determine the basin width. Enter Figure with Fr to determine
H0IW then W = H0I (H0IW).
3. Use Table 20.2 to determine the dimensions of the dissipator structure.
DATUM
PLAN
SECTION
STILLING BASIN DESIGN
PLAN
BEDDING
SECTION
ALTERNATE ENO SILL
FIGURE 20.18 Baffle-wall energy dissipator of a U.S. Bureau of Reclamation type IV basin. (From Corry et
al., 1975)
FIGURE 20.19 Design curve for a baffle-wall dissipator. (Corry et al., 1975)
D =^
(20.20)
where q is the discharge per unit width of the crest overfall, g is the acceleration caused
by gravity, and h0 is the height of the drop. The dimensions L1, V1, y2, and ^3 in Fig. 20.22
are determined using the following:
^- = 4.3(W*27
AZ0
(20.21)
- = LQN*22
AZ0
(20.22)
TABLE 20.2 Baffle Wall Dissipator: Dimensions of The Basin in Feet and Inches
W
4-0
5-0
6-0
7_0
8-0
9-0
10-0
11-0
12-0
13-0
14-0
15-0
16-0
17_0
18-0
19-0
20-0
H1
3-1 5-5
3-10 6-8
4-7 8-0
5-5 9-5
6-2 10-8
6-11 12-0
7-8 13-5
8-5 14-7
9-2 16-0
10-0 17-4
10-9 18-8
11-6 20-0
12-3 21-4
13-0 22-6
13-9 23-11
14-7 25^
15-4 26-7
H2
1-6
1-11
2-3
2-7
3-0
3-5
3-9
4-2
4-6
4-11
5-3
5-7
6-0
6-4
6-8
7-1
7-6
3-11-8
3-102-1
4-72-6
5-52-11
6-23^
6-113-9
7-84-2
8-54-7
9-25-0
10-05-5
10-95-10
11-66-3
12-36-8
13-07-1
13-97-6
14-77-11
15-48-4
I2
I1
I4
t5
(20.23)
^- = 1.66A#27
(20.24)
/I0
L2 is the length of the jump, L1 can be determined using Fig. 20.22. The sequent depth
and the tailwater depth Tw must be compared to determine whether Tw < y3, or Tw = y3,
or Tw > y3. If Tw < y3, the hydraulic jump moves downstream. In this case, it is necessary
AERATED
FIGURE 20.20 Flow geometry of a straight-drop spillway. (From Corry et al., 1975)
TOP SLOPE I TO I
UPPER NAPPE
SIDE WALL
HEIGHT
FLOOR BLOCKS
LONGITUDINALSILL
(OPTIONAL)
ENDSILL
PLAN
FIGURE 20.21 Straight-drop spillway stilling basin. (From Rand, 1955)
to construct the apron at the bed level and an end sill or baffles or to construct the apron
below the downstream bed level and an end sill. If Tw > y3, the hydraulic jump may
become submerged. If Tw = V3, the hydraulic jump begins at depth V2; there is no supercritical flow on the apron, and L1 is a minimum.
20.2.4.2 Grated energy dissipators. Energy dissipators with grates (Fig. 20.23) also can
be used in conjunction with drop structures. The U.S.Bureau of Reclamation (1987)
developed the following design recommendations for grates for incoming subcritical
flow:
1. Select slot width with a full slot width at each wall.
2. Compute beam length L0 using
LOWER NAPPE
(NOTAILWA1TER)
TAILWATER
LEVEL
TANGENTATPOINT
OFSUBMERGENCE
MEAN
SUBMERGED
TRAJECTORY
NAPPE
TRAJECTORY
FREE NAPPE
TRAJECTORY
FIGURE 20.22 Design chart for determination of L1. (From Corry et al., 1975)
LGG =
,
C(W)(N) V2^~
(20.25)
where C is a coefficient equal to 0.245, W is the width of the slots in feet, and N is the
number of slots or spaces between beams.Then compute the beam width = 1.5 W. The
quantity (W)(N) can be adjusted until an acceptable beam length, L0 is determined.
3. For self-cleaning, the grate can be tilted appoximately 3 in the downstream direction.
20.2.4.3 Straight drop structures. The straight drop structure shown in Fig. 20.21 consists of a horizontal apron with blocks and sills to dissipate energy. This structure is for
drops of less than 15 ft (4.57 m) and for sufficient tailwater. The design parameters
include the length of the basin, the position and size of the floor blocks, the position and
height of the end sill, the position of the wingwalls, and the geometry of the approach
channel. This structure was developed by the Agricultural Research Service at the Saint
Anthony Falls Hydraulic Laboratory.
The design procedure for a straight-drop structure is as follows (Corry et al 1975):
(20.26)
where distances are illustrated in Fig. 20.21. The distance from the headwall to the
point where the surface of the upper nappe strikes the still basin floor L1 is
L 1 =^
(20.27)
(20.28)
where
0.691+0.228 f M - f ^ l L
UJ
UJj
LS
~
I"
1
L
(20.29)
0.185 + 0.456^1
UJJ
and
(20.30)
2.
3.
4.
5.
6.
7.
8.
9.
stilling basin floor to the upstream face of the floor blocks L2 is determined using
L2 = 0.8 yc. The distance between the upstream face of the floor blocks and the end of
the stilling basin, L3, is determined using L3 ^ 1.75yc.
Floor blocks are proportioned as follows:
a) height = 0.8yc;
b) width and spacing should be 0.4yc with a variation of 0.15yc permitted, and
c) blocks should be square in plan, and
d) blocts should occupy between 50 to 60 percent of the stilling basin width.
Compute the end sill height as 0.4yc.
If longitudinal sills (for structural purposes only) are used, they should be constructed
through the floor block, not between the floor blocks.
Compute sidewall height above the tailwater level as 0.85yc.
Wingwalls are constructed at an angle of 45 with the outlet center line with a top slope
of 1:1.
Compute the minimum height of the tailwater surface above the floor of the stilling
basin y3 using y3 = 2.15yc.
Modification, to the approach channel are as follows: The crest of the spillway should
be at the same elevation as the approach channel, the bottom width should be equal to
the spillway notch length W0 at the headwall, and protection with riprap or paving
should be provided for a distance upstream of the headwall of 3yc.
Using the recommendations in Step 8, no special provision for aeration is needed.
Example 6 Determine the dimensions of a straight-drop spillway stilling basin for a discharge of 7.08 m3/s. The downstream trapezoidal channel has a 3:1 side slope with a 3.05
m bottom width S0 = 0.002 m/m, n = 0.03, and a normal depth of 1.024 m. The drop h0 =
1.83 m.
Solution:
Step 1. Determine the minimum basin length L8. The critical depth yc is determined as
yc = 0.655 m. Then hjyc = 1.83/0.655 = 2.79; H2 = h0 -2.15yc = 1.83 2.15(0.655) = 0.422 m; hjyc = 0.422/0.655 = 0.644.
Using Fig. 20.22, L1Jyc = 8.2 for hjyc = 2.79 and hjyc = 0.644.
Then L1 = 8.2(0.655) = 5.37 m; L2 = 0.8yc = 0.8(0.655) = 0.52 m; and L3 = 1.75yc
= 1.75(0.655) = 1.15 m; thus, LB = L1 + L2 + L3 = 5.37 + 0.52 + 1.15 = 7.04 m.
Step 2. Proportions the floor blocks are
height = 0.8vc - 0.8(0.655) = 0.524 m,
width = 0.4yc = 0.262 m, and
spacing = 0.4yc = 0.262 m.
Step 3. Calculate the end sill height = 0.4yc = 0.262 m.
Step 4. Use the longitudinal sill passing through the floor blocks.
Step 5. Calculate the side-wall height above the tailwater = 0.85yc = 0.85(0.655)
= 0.557 m.
Step 6. Local wing walls are at a 45 angle with the outlet center line.
Step 7. Calculate the minimum height of tailwater above the floor of the stilling basin: y3
= 2.15yc - 2.15(0.655) = 1.41 m. The basin must be placed 1.43 - 1.024 =
0.406 m below the downstream bed level.
20.2.4.4 Box-inlet drop structure. The box-inlet drop structure shown in Figure 20.24
consists of two different sections that are effective in controlling flow: the crest of the box
inlet and the opening in the headwall. This structure is based on experiments by the U.S.
Soil Conservation Service at the Saint Anthony Falls Hydraulic Laboratory (Blaisdell and
Donnelly, 1956). The design procedure of box-inlet drop structures is as follows:
(FHWA, 1978):
1. Select/*0
2. Select L1, W2, and Lc where Lc is the length of the box-inlet crest, Lc = W2 + 2L1,
where W2 is the width of the box inlet, and L1 is the length of the box inlet.
3. Compute the head y0 for the crest using the discharge equation for a rectangular weir:
(20 31)
^(dkf
4. Compute -^
and determine the coefficient of discharge C2 from Fig. 20.25.
(W2j
(L,]
5. Compute M- and determine the relative head correction, CH, from Fig. 20.26.
6. Compute y0 for the headwall opening using
y0=
^7=
IC2W2V^]
2/3
-Cn
(20.32)
COEFFICIENT OF DISCHARGE, C2
FIGURE 20.25 Coefficient of discharge, with control at headwall opening. (From Corry et al.,
1975)
FIGURE 20.26 Relative head correction for HJw2 > 1V4 with control at headwall opening.
(From Corry et al., 1975)
e. Determine the adjusted y0:
? -(sW/wJ20
(2033)
'-ml
9. Compute the critical depth at the exit of the stilling basin yc3 using
-tej
f
0.2
L2 y
~ < (LA
IkJ
(20.36)
DISCHARGE COEFFICIENT, C1
W
FIGURE 20.28 Correction for box-inlet shape, with control at the box-inlet crest (-c > 3).
From Corry et al., 1975)
TABLE 20.3 Correction for Dike Efect CE: Control at the Box-Inlet Crest
(Control at Box-Inlet Crest)
W4AV2
L1IW2
OO
0.5
0.90
1.0
.80
1.5
.76
2.0
.76
Source: Corry et al (1975)
Ol
0.96
.88
.83
.83
0.2
1.00
.93
.88
.88
LB
0.3
1.02
.96
.92
.92
0.4
1.04
.98
.94
.94
_(i)
2L
i
(W2)
0.5
1.05
1.00
.96
.96
0.6
1.05
1.01
.97
.97
(20.37)
and
L3 = ^f^
(20.38)
if
W3 < 11.5vc3
(20.39)
or
?3 = yC3 + 0.052PF3
if
W3 > 11.5yc3
(20.40)
20.2.5
RIPRAPBASINS
The riprap basin recommended in Corry et al (1975) for culverts is shown in Fig. 20.30,
which is a preshaped basin lined with riprap. The surface of the riprapped floor of the
energy dissipating pool is at an elevation hs below the culvert invert, where hs is the
approximate depth of scour that would occur in a thick pad of riprap if subjected to the
design discharge (Corry et al 1975). The length of the pool is the larger of I0hs or W0, and
the overall length of the basin is the larger of I5hs or 4W0. The ratio hs/d5Q should be less
than 4 [(HJd50) < 4] .Figure 20.31 provides design detaills for riprapped culvert energy
basins.
NOTEB
IFT
BOARD
NOTEA - SUFFI
IF EXI f VtLOClIY OF BASIONNAL
IS SPECI
EXTENDN A-ABASISUCH
N AS REQUIRED TO OBTAIN
AREAFFIATEIED.DSECTI
SECTIOCNIENTAREACROSS-SECTI
AT SEC. A-A) - SPECI
EXITOVELOCI
TV. THAT O^ /(CROSS
NOTE B - FLOOR
WARP OFBASIBASIN TON SHOULD
CONFORMBE ATTOTHENATURAL
STREAM
CHANNEL.
TOP OF RIPRAP IN
CHANNEL BOTTOM AT SEC. A-A. SAME ELEVATION OR LOWER THAN NATURAL
DlSSlPATOR POOL
Sh1APRON
ORW0MIN.
100,OR3W
0MIN.
TOPOFBERM
TOP OF RIPRAP
CHANNEL
ZONTAL
TO SUPPORT RIPRAP
THI
COPTIONAL
KENEDORSLOPI- CONSTRUCT
NG
TOE
IF DOWNSTREAM
CHANNEL
SECTION
NOTE B
EXCAVATETOTHISLINE,
NOTE:
W0 - PIDIAMETER
FOR
PE CULVERT
W0 - BARREL
WIDTH
FOR
BOX
CULVERT
W = SPAN
OF PIPE-ARCH
CULVERT
SYMM ABOUT
APRON
CULVERT
HORIZONTAL
BERM
AS REQUIPRRAPED
TOSUPPORTRI
FIGURE 20.30 Details of riprapped culvert energy basin. (Corry et al., 1975)
DESIGNDISCHARGE-Q
WETTED AREA AT BRINK OF CULVERT
d50 = THE MEDIAN SIZE OF ROCK
BYWEIGHT.
OR
ANGULARROUNDEDROCK
ROCK.
EQUIVALENT BRINK DEPTH
BRINK DEPTH FOR BOX CULVERT
FOR
NON-RECTANGULAR
SECTIONS
CULVERTBRINK
?. SECTION
FROUDE NUMBER
FIGURE 20.31 Relative depth of scour hole versus Froude number at brink
of culvert with relative size of riprap as a third variable.
REFERENCES
Blaisdell, F. W., Flow Through Diverging Open Channel Transitions at Supercritical Velocities,
SCS Report No. SCS-TR-76, U.S. Department of Agriculture, April 1949.
Blaisdell, F. W., and C. A. Donnelly, "The Box Inlet Drop Spillway and Its Outlet", Transactions,
of the American Society of Civil Engineers, 121:955-986, 1956.
Blaisdell, F. W., The SAF Stilling Basin, U. S. Goverment Printing Office, 1959
Corry, M. L., P. L. Thompson, E. J. Watts, J. S., Jones, and D. L. Richards, Hydraulic Design of
Energy Dissipators for Culverts and Channels, Enqineering Circular 145, Federal Highway
Administration, U.S. Department of Transportation, Washington DC, 1975.
Fedeard Highinay Administratin , Hydraulics of Bridge Waterways, Hydraulic Design Series No. 1,
Federal Highway Administration, U.S. Department of Transportation, Washington, DC, 1978.
Ippen, A. T, "Mechanics of Supercritical Flow", Transactions of the American Society of Civil
Engineers,, lib; 268-295, 1951
Rand, W, "Flow Geometry at Straight Drop Spillways," Paper No. 791, Proceedings of the
American Society of Civil Engineers, VoI, 81, pp. 1-13, September 1955.
CHAPTER 20
FLOW TRANSITIONS AND
ENERGY DISSIPATORS
FOR CULVERTS
AND
CHANNELS
Larry W. Mays
Department of Civil and Environmental Engineering
Arizona State University
Tempe, Arizona
20.1
FLOWTRANSITIONSFORCULVERTS
Flow transitions are changes in the cross section of an open channel over short distances.
They are designed to have a minimum amount of flow disturbance. Figure 20.1 illustrates
the various types of transitions; the two most common ones are the abrupt (headwall) and
the straight line (wingwall).
Highway culverts typically are designed to operate with an upstream headwater pool
that dissipates the of the channel approach velocity. This type of situation does not require
an approach flow transition. Outlet transitions (expansions) should be considered in the
design of all culverts, channel protection, and energy dissipators. Transition design can be
categorized as
culverts with outlet control (subcritical flow) and
culverts with inlet control (supercritical flow).
20.1.1 Culverts with Outlet Control
Culverts with outlet control can have abrupt expansions or gradual transitions. In an
abrupt expansion, the water surface plunges or drops rapidly and the flow spreads out.
The potential energy stored as depth is converted to kinetic energy with a higher velocity of flow. The transition (apron) end velocity can be determined using the experimental
results of Watts (1968) (Figs. 20.2 and 20.3). Figure 20.2 relates the average depth brink
depth ratio (yA/y0) for a rectangular outlet to the Froude number. Figure 20.3 is a similar
relationship for pipe culverts. These curves in Figs. 20.2 and 20.3 were developed for
Fr's ranging from 1.0 to 2.5, the applicable range for most abrupt outlet transitions.
Usually, a low tailwater is encountered at culvert outlets and the flow is supercritical on
the outlet apron.
Simons, D. B., M. A. Stevens, and F. J. Watts, Flood Protection at Culvert Outlets, CER No.
69-70-DBS-MAS-FJWA, Colorado State University, Fort Collins, CO. 1970.
U.S. Army Corps of Engineers, Hydraulic Design of Flood Control Channels, Engineering and
Design Manual EMl 110-2-1601, pp.2026, July 1970.
U.S. Bureau of Reclamation, (USBR), Design of Small Dams, U.S. Government Printing Office,
Denver, CO, 1987.
Watts, F. J., Hydraulics of Rigid Boundary Basins, doctoral dissertation, Colorado State University,
Fort Collins, CO, 1968.
CHAPTER 21
HYDRAULIC DESIGN OF
FLOW MEASURING
STRUCTURES
21.1
INTRODUCTION
Experienced water providers and users can use this chapter as a quick review of hydraulic
principles related to water measurement and its relation to hydraulic design for environmental considerations.
The hydraulic design of flow measuring structures usually confronts the engineer with
two opportunities. One is the design of measurement structures in a retrofit situation and
the other is in original project design. The retrofit mode is usually difficult and requires
much innovation just to obtain passable function within the space and sizing limitations
and other constraints usually imposed. Because of the increasing emphasis on quantifying
flow rates and volumes in most aspects of water resource planning and management, the
retrofit applications currently dominate the design problems.
Most textbooks deal with recommending ideal installation situations and retrofit projects appear to be unable to comply without great economic impact. This too frequently
can lead to arbitrary compromises that produce poor measurement performance. Even
new installations may be limited by space requirements. This may force design decisions
into the final construction that compromise accuracy. This chapter will strive to show the
design concepts available, particularly those useful for designing both new and retrofit
installations, and will point out measurement behaviors to be expected from various compromises. This chapter suggests those deviations that cause least impact and guides the
designer to choices that may be hydraulically acceptable and still meet structural goals.
Of the numerous flowmetering methods available to the hydraulic engineer, most are
based on well-established hydraulic principles and are amenable to design manipulations
of size, shape, and response. While this aspect of flow measurement is documented in several handbooks and texts, the design and retrofit of sites to accommodate and facilitate
measurement is not as well described or is described in a scattered assortment of books
and articles.
Pipeline flows of water are usually less complicated to measure than open-channel
flows, most obviously because the flow area does not change significantly with flow rate.
Consequently, many applications of pipeline flows are held to stricter accuracy standards
than channel flows can reasonably achieve. Thus, channel flows and their measurement
are usually limited to large delivery volumes and to accuracies acceptable to the related
activities, such as sewer flows and irrigation deliveries.
The purpose of this chapter is to consolidate design information for evaluating a flow
measurement site, selecting a flow measuring system, and adapting the measuring site to
optimize measuring and other functions that may be desired from the site. Emphasis will
be on open-channel flow measurements because that is a likely need of the hydraulic engineer. Pipe flowmeters in water supply will also be discussed in lesser detail because the
major application of the many types of pipe flowmeters is well covered in the chemical
and petroleum industry literature.
Experienced readers may wish to further investigate and seek more advanced references in hydraulics and fluid mechanics. Extensive information on fluid meter theory and
detailed material for determining coefficients for tube-type meters is given in American
Society, of Mechanical Engineers (ASME) (1959, 1971) and revisited with modern
updates in books by Spitzer (1990) and Miller (1996). Brater and King (1982) have a thorough discussion of general critical depth relations and detailed relationships for most
common hydraulic flow section shapes in open channels. Bos (1989) covers a broad segment of open-channel water measurement devices.
dling the equipment and doing the analysis. The most used techniques applicable to openchannel systems, including sewer flows and irrigation canal flow measurements, depend
on exploiting the special velocity properties of critical flow, as discussed in a section
21.2.3.
Continuity equation. The first basic equation for water flowing in either pipes or channels is the continuity equation, which simply states that discharge rate (volumetric flow
per unit time), Q, is equal to flow cross-sectional area, A, times flow mean velocity, V,
through the flow cross section, or
Q = AV
(21.1)
Bernoulli energy equation. Another basic equation involves energy relations and is
also applicable to both pipe and channel flows. The most familiar form is for closed pipe
flow, wherein the basic energy principles are described by the Bernoulli energy equation.
For two locations along a pipe at stations 1 and 2,(Fig. 21.1), the Bernoulli equation can
be expressed as
V2
V2
ZI + h\ + -^- = ^2 + h2 + -^- = constant
(21.2)
where the terms are expressed in length dimensions as z = the height from an arbitrary
reference plane (datum) h = the pressure head V = average velocity through the pipe
cross-section at the designated location V2/2g = the velocity head g = the gravitational
constant i ,2 = subscripts denoting the respective locations along the pipeline.
This equation is based on uniform velocity across the conduit area and no energy losses. However, in real fluid flows, nonuniform velocities exist and friction causes energy
conversion to heat. Typically, these velocities are zero at the walls and reach a maximum
profile velocity near the center of the flow. If the flow is viscous flow in a round pipe, the
flow profile is parabolic, that is, "bullet-shaped." If the velocity is fully turbulent, the bullet-shape is much flattened, with steep velocity gradients near the wall and nearly uniform
profile across the remainder of the pipe. These idealized profiles can be skewed drastically
by regulating valves, structures, conduit bends and other flow obstructions. Therefore,
application of these equations depends on knowing, or controlling, the velocity profile so
that the average velocity in the conduit cross section can be inferred.
Datum
FIGURE 21.1 Energy balance in pipe flow.
Equation (2.2) requires some adjustments to convert it to the energy equation, which
is useful in analyzing flows in pipes or open channels with a small slope (Chow, 1959).
First we introduce correction factors, GL1 and OC2, called the velocity distribution coefficients, to account for the computational expediency of using the average velocities, V1 and
V2, to compute the kinetic energy term, W2g, at the respective locations 1 and 2 along a
channel. These values for the usual range of turbulent flows in water usually range from
about 1.01 to 1.05, although for thick petroleum products in pipe flows and low velocity
flows, the value can approach a value of 2. Second, a term, hp for the loss of energy
between the two points is included. The result is
V2
v2
Z1 + H1 + a, -i = Z2 + /I2 + a2 -^ + hf
(21.3)
5
o
21.2.2 Pipe Hydraulics
Reynolds number. The behavior of flow in pipes is governed primarily by the viscosity of
the fluid. In pipeline flows, the ratio between the dynamic forces and the viscus forces is
important for defining the limits between laminar and turbulent flows and other functions
of pipe flow. This ratio is called the Reynolds number, Rn, and is defined as
Rn = ^-
(21.4)
where V = the velocity of the flow, Lc = characteristic length, typically the pipe diameter, D and v = the kinematic viscosity.
Headloss characteristics in pipes. The Reynolds number, Rn, defined above, represents the effect of viscosity relative to inertia and is used to define appropriate flow ranges
for headloss equations in pipe flow. For example, headloss is proportional to the square of
the velocity, when the velocities and pipe size combinations defined by a pipe-diameterbased Reynolds number, Rn, greater than about 1000. Most of the flows of interest in general hydraulic engineering have Reynolds numbers greater than 1000. Some exceptions
are found in drip or trickle irrigation systems common in agricultural and urban landscape
settings.
The headloss, hf for Rn greater than the minimum value of about 1000 is traditionally
expressed in terms of a friction factor,/, the pipe diameter, D, pipe length, L, and the velocity head, V2/2g, where g is the gravitational constant, and V is the average velocity, as
hf = / I
(21.5)
The value for/is usually obtained from a Moody diagram which is a graphical representation of the/value in terms of the Reynolds number, the roughness height of the pipe
wall material, e, and the pipe diameter, D. The Moody diagram is a graphical solution of
the Colebrook function
1 = -2log^
+ ^=]
Vf
I3-7
RnVf)
(21.6)
The e values range from 0.0000015 m for smooth plastic pipe to 0.00026 m for cast
iron pipe. Concrete pipe ranges from about 0.0003 m to 0.003 m (Daugherty and
Ingersoll, 1954). The equation can be readily solved by iteration techniques using a computer spreadsheet.
(21-7>
For conduits such as pipes flowing nearly full, the surface flow width may be narrow,
and Dm may be a larger value than the physical water depth. For the usual natural channels and most canals, Dn is interchangeable with average depth. Sometimes it is simply
called the hydraulic depth (Chow, 1959).
Froude number. Open-channel flow behavior is governed primarily by gravity
forces. The ratio of the inertial forces to the gravity forces is called the Froude number,
Fr, and is defined by
Fr =
Y
(21.8)
v^:
where V the velocity of the flow, g = the gravitational constant, Dn = the hydraulic
mean depth.
The Froude number applies to most open channel flows and is used for defining model
scale ratios and estimating stable flow characteristics in open channels.
Specific energy. It is useful to define the energy equation in terms of the local channel bottom instead of an arbitrary datum. This is called the specific energy, ", and is given
by:
E = y + ^(21.9)
2g
That is, the specific energy is equal to the sum of the depth of flow y and the velocity
head (Fig. 21.2).
Critical flow and critical depth. In open channels a flow phenomenon occurs that
does not happen in closed pipe flows. The process is called critical flow. Critical flow is
defined for open-channel flows as the maximum discharge for the minimum specific energy, that is, critical flow represents the minimum combination of potential energy (depth
of flow, y) and kinetic energy (velocity head, V2/2g) for the given discharge (Chow, 1959).
The depth of flow then is the critical depth. By virtue of the continuity equation, for a constant discharge at critical flow, an increase in depth must necessarily be accompanied by
a decrease in velocity, which is called subcritical velocity. Conversely, a decrease in depth
for the same flow rate necessarily requires an increase in velocity, which is called supercritical velocity.
When critical flow occurs in an open channel it can be shown (Chow, 1959) that
V2 D
oc^ = -f
(21.10)
EnergyGrade Line
Hydraulic Grade Line
& = AJ%%
<2U1)
In practice, the water surface slope in a contraction is relatively steep and the precise plane of the critical flow section is not easily or reliably located. Thus, the data
for accurately evaluating the hydraulic depth, Dm, is not readily obtained. For critical
flow flumes, the flow depth is therefore not measured at this critical section, but
instead a depth is measured in the upstream channel, where the velocity head is computable or is minimal. The critical depth is then mathematically derived based on energy principles described by Bos et al. (1991). These flumes, sometimes referred to as
the computable flumes that rely on critical flow theory, will be discussed in more detail
in Section 21.7.
For maximum discharge for minimum energy, the condition described above for critical flow in open channels, it can be shown that
af = A = ^
<2U2)
where: A = the channel flow area, T = the top width of the channel flow, Dm = the
hydraulic mean depth, and Vc = the critical velocity.
Thus, the velocity head at critical flow is equal to half the hydraulic mean depth, sometimes called hydraulic depth, Dm = AIT (Chow, 1959).
From the above,
,Vc
= 1 = Fr
(21.13)
VgD m la
where Fr is the Froude number defined above. Thus, at critical flow the Froude number is
unity. Also note that the Froude number can be defined by Fr = VIV0 for velocities other
than critical.
Normal depth. Yet another depth is associated with open channel flows, the normal depth. When the flow in an open channel does not change from station to station,
the flow is said to be uniform and the bottom slope, the hydraulic grade line, and the
energy grade line are all parallel to each other. Figure 21.2 shows the condition when
the flow is not uniform.
Modular limit. If the downstream depth in a channel is too deep, the backwater will
prevent critical depth from occurring. The flow is considered to be submerged whenever
the downstream water surface exceeds the crest elevation of a channel control, such as a
weir or flume. For flumes, particularly, this submergence has little effect on critical depth,
and free flow exists until a certain limiting submergence for that particular flow module
called the modular limit is reached. At some point of submergence, the upstream flow
depth is affected, and the modular limit is exceeded, and free flow does not occur. The
modular limit is defined as that limiting submergence ratio, and is based on the ratio of
the downstream depth to upstream depth. The modular limit occurs when the downstream
backwater causes more than 1 percent change in the calculated discharge in a particular
flow module, or device (Bos, 1989). When the modular limit is exceeded, the flow is
called nonmodular.
21.2.4 Energy Balance Relationships in Channels
Hydraulic problems concerning fluid flow are commonly described in terms of conserving kinetic and potential energy, and are conveniently expressed using the classical
Bernoulli equation in combination with the Continuity equation. The applications of these
equations are generally well documented, particularly for pipe flows, in texts and handbooks and are not repeated here (Brater and King, 1982; Miller, 1996). The case for open
channels is less complete, but is given considerable treatment in Brater and King (1982),
Chow (1959), and Herschy (1985). The computational uncertainties evolve from the
effects of friction and viscosity that distort the classic assumptions of a uniform velocity
profile across the fluid stream. When accountings for friction and flow profile are successfully applied, the results for discharge computations are usually good to excellent for
both pipes and open channels (Bos et al., 1991).
Headloss characteristics in channels. In terms of frictional headlosses, the wetted
perimeter, Pw, of the flow is important. Hydraulic radius, Rh, is defined as the area of the
flow section, A, divided by the wetted perimeter, Pw, or
Rh =
p
(21.14)
Conversely, the wetted perimeter times the hydraulic radius is equal to the area of an
irregular flow section. The hydraulic radius of a channel can be compared to the radius of
a pipe, r, with a cross-sectional area A = nr2 and a circumference or wetted perimeter Pw
= 2 nr. Under these conditions, the hydraulic radius compares to the pipe radius, and to
the pipe diameter, D, as
r D
R
=
(2L15)
=2
^
h
The Manning's formula. Canal and stream discharge rates are usually estimated with
use of the Manning's formula. Many open-channel flow equations have been proposed,
but the most used is the Manning's formula. This expression is partly rational and uses an
empirical coefficient, rc, that is used in both the SI and American unit systems. In general
form it becomes
v=^LR2i3Sm
(2L16)
equal to Se when uniform flow, with the resulting normal depth occurs. As defined above,
normal depth occurs when a channel flow approaches uniformity from station to station
along the channel (Chow, 1959).
For design purposes, the n value for concrete lined canals is usually about 0.014. A
good finish can lower it to 0.012, while concrete in poor condition and channels constructed with shot crete or gunite, usually have n values from 0.016 to 0.018. In some
instances, concrete lined canals, with significant algae growth, have experienced n values
as high as 0.032. This latter value approaches the values usually experienced with unlined
channels, 0.03-0.04. Thus, for reliable application, the use of Manning's formula requires
field experience and on-site inspection of the channel being computed.
21.2.5 Modeling Characteristics for Open Channels
For flowing water in open channels, fluid friction is a factor as well as gravity and inertia. This would seem to present a problem for hydraulic scale modeling, because both
dynamic and kinematic similarity are difficult to achieve simultaneously. Fortunately for
most open-channel flows, there is usually fully developed turbulence. Thus, the fluid friction losses are nearly proportional to V2, and are nearly independent of Reynolds number,
Rn, with rare exceptions.
This means that in open-channel flows, inertia and gravity forces dominate over viscous forces (associated with pipe flows) and are a function of the Froude number, Fn,
alone. Geometric similarity between a model and a prototype then provides kinematic
similarity. For kinematic similarity the ratios of the respective velocities are everywhere
the same. The velocity ratio, Vr, is the velocity in the prototype, Vp, divided by the velocity in the model, Vn, or
Vr = ^(21.17)
m
For Froude modeling, and from the definition for Fn, we note that V is proportional
to the square-root of a length, L (for open channels we used the hydraulic depth, D1n)
with the gravitational constant, g, assumed to be constant. Thus, the above equation can
be written as
"^LT = Y
where Lr = the length ratio between prototype and model dimensions, Lp: Lm
(2L18)
Because the velocity varies as VZ^ and the cross-sectional area as L2 it follows
that
Qp'Qm = LT 1
(21-19)
O
(J Y/2
-SL = U
XZm \.LmJ
(21.20)
or
This equation is valid when all the physical structure dimensions and the heads are of
the same ratio. For example, it can be used to convert a flume rating for one size to that
of a similar flume of another size. Scale modeling works best for determining calibrations
in a range of Lp:Lm less than about 10:1, although ranges exceeding 50:1 have sometimes
been used for studying special situations.
21.3
Flow is usually measured by determining an average flow velocity and using the flow area
to compute the volume discharge. Flow meters then have the function of detecting this
velocity and combining it with the physical information of the conduit to produce a useable readout. This is easily demonstrated for closed conduits. Propeller meters, ultrasonic
meters, laser-Doppler velocimeters, electromagnetic meters, Venturi meters, and orifice
meters all are based on inferring a basic velocity measurement applied to a flow area for
a discharge rate.
For open channels, many flumes depend on determining the velocity based on energy
principles of critical flow. Weirs are usually described in terms of orifice flow integrated
over the weir width and the crest depth. Again these are basically velocity expressions for
flow through a defined area.
Dilution techniques applicable to both closed pipe and open channel flows depend on
detecting the amount of fluid added to a known starting amount of tracer material. The
dilution ratio determines the discharge ratio, in the case of constant injection of a tracer.
The tracer may be a chemical or even injected heat or heated fluid.
Electromagnetic meters depend on generating voltages by flowing a conductive fluid,
usually water, through a magnetic field to produce a velocity indication.
21.3.1 Water Meter Classification
Flow measuring devices are commonly classified into those that are rate meters and measure discharge rate as the primary reported indication and those that are quantity meters
and measure volume as the primary indication. The latter include weighing tanks and
batch volume tanks and are used mostly in laboratory settings as flow rate standards.
Devices in either of these broad classes can again be divided according to the physical
principle that is used to detect that primary indication (ASME, 1959). The meter part that
interacts with the flow to produce the primary indication is referred to as the primary
device. This interaction exploits one or more of a few physical principles, such as pressure force, energy conversion, weight, electrical properties, mixing properties, sonic properties, and so on, to generate a signal. Primary devices are thus limited in number and variety. Secondary devices convert the primary interaction into useable readout. These secondary devices are numerous and relatively unlimited in configuration and variety. The
function of one class can be converted into the response of the other with suitable secondary devices.
Some water measuring devices particularly suitable to municipal water supply, wastewater treatment, agricultural irrigation, and drainage applications are the historical rate
meters that are treated in most hydraulic text. These include (1) weirs, (2) flumes, (3) orifice meters, and (4) Venturi meters.
Head, h, or upstream depth, commonly is used for the open channel devices such as
flumes and weirs. Either pressure, /?, head, /i, or differential head, A/z, or differential pressure, Ap, is used with tube-type devices, such as Venturi meters and orifice meters.
Venturi meters in pipelines and long throated flumes in open-channel flows are examples where the energy principles and the flow accountings mentioned above give good to
excellent computational results with minor dependency on empirical coefficients (Bos,
1989; Bos et al., 1991).
21.3.2 Installation Requirements
Special difficulties arise in applying velocity profile and friction accountings when insufficient pipe or channel exists upstream from a flow measuring device. This is needed to
ensure that predictable and acceptable velocity profiles are presented to the meter.
Frequently pipe or channel lengths can be significantly shortened by special structural
flow conditioners. These structural measures then become a design option. Some of these
are discussed below and in section 21.3.3
Designs for pipe discharges are well described in textbooks and in standard handbooks. The design difficulties center around selecting appropriate metering candidates for
accomplishing the measuring function and in providing an appropriate environment for
economical, accurate, and serviceable operation.
In the case of pipe flows, recommended straight pipe lengths, in terms of pipe diameter, are to be provided upstream of the meter to assure reasonable operating accuracy. These
lengths depend on the flow pattern presented to the meter primarily caused by valves and
pipe elbows upstream from the meter. The number and orientation of elbows greatly influence the circulation patterns and flow profile distortions presented to the meter.
Open-channel flow water measurement generally requires that the Froude number of
the approach flow be less than 0.5 to prevent wave action that would hinder or possibly
prevent an accurate head determination.
Energy concepts are used to describe Venturi meters in pipe flows based on the
Bernoulli equation in which part of the pipe forms a contracted throat that necessarily
changes the flow velocity and hence converts some of the static pressure to velocity head.
The decrease in static pressure is the basis for flow detection. A similar concept can be
applied to open channels. A historical version is the so-called Venturi flume (Brater and
King, 1982) that detects the change in water surface elevation between an upstream station and in a contracted section. However, this small change is difficult to accurately
detect, so the direct concept is not used. Rather, contractions are designed to be severe
enough to force critical flow velocities in the contracted section. Thus, only an upstream
head is needed to define the flow energy and flow area which can be converted to discharge rate. These are generally called critical-flow flumes. The flow condition where
only one head measurement is needed is called freeflow.
The critical-flow flumes themselves consist of those called long-throated flumes that
force parallel flow in the contracted, or control, section, called the throat, and those that
have curvilinear flow in the throat and are called short-throated flumes. The limiting
throat control section is the sharp-crested weir consisting of a thin plate. Thus, for
flumes and weirs one unique head value exists for each discharge, simplifying the calibration procedure.
However, if the downstream flow level submerges critical depth enough to affect the
upstream reading, the modular limit is exceeded, and free flow does not occur. When
exceeded, separate calibrations at many levels of submergence are then required, and two
head measurements are needed to measure flow. This condition generally is to be avoided in meter site design because it reduces the accuracy of the measurement and increases
the difficulty of flow determination. The modular limit for sharp-crested weirs, in practice, is less than zero, requiring full clearance of the overfall nappe of at least 3 cm, while
short-throated flumes can usually tolerate 65 percent to 70 percent submergence. Longthroated flumes can tolerate from 70 percent to 90 percent depending on flow conditions
and flume size.
Designing flumes for submerged flow beyond the modular limit decreases the accuracy of the flow measurement. Sometimes flumes and weirs can be overly submerged unintentionally by poor design, construction errors, structural settling, attempts to supply
increased delivery needs with increasing downstream heads, accumulated sediment
deposits, or weed growths. Sometimes use of the submerged range beyond the modular
limit is an economic compromise.
Approach flow conditions for pipes. Water measurement devices are generally calibrated with certain approach flow conditions. The same approach conditions must be
attained in field applications of measuring devices. Poor flow conditions in the area just
upstream of the measuring device can cause large discharge indication errors. For open
channels, the approaching flow should generally be subcritical. The flow should be fully
developed, mild in slope, and free of curves, projections, and waves.
Pipeline meters commonly require 10 or more diameters of straight pipe approach.
Fittings and combinations of fittings, such as valves and bends, located upstream from a
flow meter can increase the number of required approach diameters. Several references
(ASME, 1971; ISO, 1991) give requirements for many pipeline configurations and
meters. These are discussed in detail by Miller (1996).
Flow conditioning options. Many installations, especially in retrofit situations, do
not provide for sufficient lengths of straight pipe to remove velocity profile distortions
and swirl to an acceptable level. Therefore, the designer may need to use flow conditioners in combination with straight pipe lengths. Swirl sensitivity varies widely. Some meters
are particularly sensitive to swirl, such as the propeller and turbine meters. Magnetic flow
meters are somewhat less sensitive to radial velocities than single-path ultrasonic flow
meters. Venturi meters are less sensitive than orifice meters. For a swirl angle of 20 , the
discharge coefficient changes by about 1 percent for a Venturi meter with p = 0.32 ((3 is
the ratio of meter throat diameter to the pipe diameter) and about +10 percent for a similar orifice. Thus a swirl can increase the discharge through an orifice for the same differential head reading (Miller, 1996).
In pipeline flows, contractions can produce a central jet and also increase an incoming
swirl, while expansions tend to slow swirls and produce enough secondary flow to restore
flow profiles to some semblance of acceptability. These characteristics can modify the
straight pipe lengths needed or the type of flow conditioner to recommend (Miller, 1996).
Rough pipes also tend to reduce a swirl.
For flows, such as that encountered in sewage discharges and irrigation pipeline deliveries that originate from open channels, many of the tube-bundle types of flow conditioners can gather trash and cause maintenance problems. Many meter providers in these situations use fins or vanes that protrude from the wall and have sloped upstream edges that
shed trash. The vanes protrude about one-fourth of the pipe diameter into the flow, leaving the center core of the flow open. While these vanes can vary in number and length,
the logic being that the fewer the vanes the longer they should be in the direction of flow,
common configurations are four vanes that are about two or three pipe diameters long.
Vanes in themselves do not condition wall jets well. Field experience, has shown that troublesome flow profiles can be conditioned significantly by inserting an orifice into the
pipe. The orifice diameter is about 90 percent of the pipe diameter and is used to control
wall jets and force them to mix with the general flow. The orifice in itself tends to crossmix the jets and would appear to reduce spin. However, if the jets are symmetrical and an
initial swirl exists, orifices tend to increase the swirl. Inserting an orifice appears to be
supported by recent recommendations of Miller (1996) where it is stated: "To achieve a
fully developed profile, it is important that the flow be blocked or restricted close to the
wall, with the central core having the larger flow area."
The addition of vanes when space permits is recommended. Because orifices, in general, tend to force the flow to the pipe center while increasing spin, it appears best to place
the vanes upstream from the orifice. If they are placed downstream, the spin not only may
be increased, but the spinning central flow may not be touched by the vanes.
21.3.3
Flow conditioning in an irrigation delivery pipeline. As mentioned previously, measuring devices frequently must be installed in flow situations that are less than optimal. A field
example occurred in Arizona where a large pipe was used as an outlet to a secondary canal
and a single-path ultrasonic meter placed in it was subjected to flow profile distortions. The
pipe was about 0.75 m in diameter and delivered approximately 400 L/s. The flow rate
readout was unstable, with fluctuations varying by about 15 percent. The problem appeared
to be caused by slowly spiraling flow induced by the bottom jet from a partly open pipe
inlet gate and a 45 elbow. This is similar to two closely spaced pipe elbows that are not in
the same plane, which can cause a spiral flow pattern (ASME, 1971).
A successful attempt to modify the jet and cause it to cross mix so that the jet effects
and the strength of the spiral flow were reduced, was accomplished by inserting a large Pratio orifice in the pipe (Fig. 21.3). This consisted of an annular metal ring with the outside radius approximately that of the pipe and an inside diameter about 10 percent less, or
an orifice with (3 = 90 percent. The orifice was installed about three diameters downstream from the elbow. The slight increase in headless was compensated by increasing the
upstream gate opening. The orifice can be constructed by cutting notches from an appropriately sized piece of angle iron or aluminum and bending it to a polygon that approxi-
Control O gate
Ultrasonic flow
meter
Orifice plate
with large
opening
Spiral flow
Improved profile
FIGURE 21.3 An orifice plate with a large opening is used to condition a flow profile.
mates the circle diameter of the pipe interior. Some leakage around the ring is acceptable.
For propeller meters, additional vanes projecting from the walls may be needed to further
reduce spiral flow. These vanes would be placed upstream from the orifice. In this installation, the fluctuation was reduced to within about 3 percent.
Flow conditioning in channels. By analogy and using a minimum of 10 pipe diameters of a straight approach channel, open channel flow would require 40 hydraulic radii of
straight, unobstructed, unaltered approach, based on the calculation of hydraulic radius for
circular pipes being equal to one-fourth the pipe diameter, (Eq. 21.15). This would translate for very wide channels into approximately 40 times the flow depth. For narrow channels that are as deep as they are wide, this would compute to be about 13 channel depths
or top widths.
Other recommendations on approach channel criteria are presented by Bos (1989)
and USBR (1997). Major features of that criteria follow:
If the control width is greater than 50 percent of the approach channel width, 10 average approach flow widths of straight, unobstructed approach are required.
If the control width is less than 50 percent of the approach width, 20 control widths of
straight, unobstructed approach are required.
If upstream flow is below critical depth, a jump should be forced to occur. In this case,
30 measuring heads of straight, unobstructed approach after the jump should be provided.
If baffles are used to correct and smooth out approach flow, then 10 measuring heads
(10 /I1) should be placed between the baffles and the measuring station.
Approach flow conditions should be continually checked for deviation from these conditions as described in Bos (1989) and USBR (1997).
The baffles described above can become unacceptable maintenance problems in open
channels. Some field expediences are therefore described that have been found to work in
specific instances, but have not been studied for assured design generalizations. Nevertheless,
these constructions are but small extensions to currently accepted practices in pipe flows.
Applications for open-channel flow conditioners include abrupt channel turns, sluice
gate outflows, and channels downstream from a hydraulic jump. The abrupt turns may
benefit from floor and wall mounted vanes or fins. Based on pipe flow experience, and
assuming the channel is half of a closed conduit, these fins or vanes would probably be
about 10 percent to 15 percent of the channel depth.
As in pipe flow, wall jets that can develop downstream from sluice gates appear to
need treatment. This can be in the form of a structural angle bolted on the channel floor
and up the walls. Suggested size, based on the pipe flow analogy, is for the angle to be
about 5 percent into the channel flow depth. Whether the sidewalls need larger angles
when the channels are wide has not been tested.
21.3.4 Wave Suppression
Of special concern in open channels is wave suppression downstream from a sluice gate,
hydraulic jump, or an abrupt turn. Thus, the flow conditioners in channels have the additional task not present in pipe flows of surface wave suppression. Excessive waves in irrigation canals make reading sidewall gages difficult. These waves are usually caused by a
jet entry from a sluice gate or by a waterfall situation. The unstable surface can be 10-20
cm high and extend for tens of meters downstream.
Solidly attached
timber or concrete
FLOW
FIGURE 21.4 Wave suppresor design (From USBR, 1997).
them as shown in Fig. 21.5 would accomplish this without excessive obstruction.
Rounding the upstream edges will help shed trash, but may be less effective in suppressing waves. Observe in the sequence of drawings in Fig. 21.5 that the staggering is upward
in the downstream direction. Note that the next baffle slightly overlaps the horizontal flow
lines so that flow passing over the top of one baffle is not allowed to free-fall and start
another wave. Fig. 21.5 a-c illustrate the general behavior as the flow becomes less deep.
21.4
MEASUREMENTACCURACY
Accurate application of water measuring devices generally depends upon standard designs
or careful selection of devices, careful fabrication and installation, good calibration data
and adequate analysis. Also needed is proper user operation with appropriate inspection
and maintenance procedures. During operation, accuracy requires continual verification
that all measuring systems, including the operators, are functioning properly. Thus, good
training and supervision are required to attain measurements within prescribed accuracy
bounds. Accuracy is the degree of conformance of a measurement to a standard or true
value. The standards are selected by users, providers, governments, or compacts between
these entities. All parts of a measuring system, including the user, need to be considered
in accessing the system's total accuracy.
As mentioned above, a measurement system usually consists of a primary element,
which is that part of the system that creates what is sensed, and is measured by a secondary element. For example, weirs and flumes are primary elements. A staff gage is a
secondary element.
Designers, purchasers, and users of water measurement devices generally rely on standard designs and manufacturers to provide calibrations and assurances of accuracy. A few
water users and providers have the facilities to check the condition and accuracy of flow
measuring devices. These facilities have comparison flow meters and/or volumetric tanks
for checking their flow meters. These test systems are used to check devices for compliance with specification and to determine maintenance needs. However, maintaining facilities such as these is not generally practical.
Various disciplines and organizations do not fully agree on some of the definitions
related to measuring device specifications, calibration, and error analysis. Therefore, it is
important to verify that a clear and mutual understanding of the specifications, calibration
terminology, and the error analysis processes is established when discussing these topics
with others.
21.4.1 Definitions of Terms Related to Accuracy
Error. Error is the deviation of a measurement, observation, or calculation from the
truth. The deviation can be small and inherent in the structure and functioning of the system and be within the bounds or limits specified. Lack of care and mistakes during fabrication, installation, and use can often cause large errors well outside expected performance bounds. Because the true value is seldom known, some investigators prefer to use
the term uncertainty. Uncertainty describes the possible error or range of error which
may exist. Investigators often classify errors and uncertainties into spurious, systematic,
and random types.
Precision. Precision is the ability to produce the same measurement value within
given accuracy bounds when successive readings of a specific quantity are measured.
Precision represents the maximum departure of all readings from the mean value of the
readings. Thus, a single observation of a measurement cannot be more accurate than
the inherent precision of the combined primary and secondary precision. It is possible to
have good precision of an inaccurate reading. Thus, precision and accuracy differ.
Spurious errors. Spurious errors are commonly caused by accident, resulting in false
data. Misreading and intermittent mechanical malfunctions can cause discharge readings
well outside of expected random statistical distribution about the mean. Spurious errors
can be minimized by good supervision, maintenance, inspection, and training.
Experienced, well-trained operators are more likely to recognize readings that are significantly out of the expected range of deviation. Unexpected spiral flow and blockages of
flow in the approach or in the device itself can cause spurious errors. Repeating measurements does not provide information on spurious error unless repetitions occur before and
after the introduction of the error. On a statistical basis, spurious errors confound evaluation of accuracy performance.
Systematic errors. Systematic errors are errors that persist and cannot be considered
random. Systematic errors are caused by deviations from standard device dimensions,
anomalies to the particular installation, and possible bias in the calibration. Systematic
errors cannot be removed or detected by repeated measurements. They usually cause persistent error on one side of the true value. The value of a particular systematic error for a
particular device may sometimes be considered as a random error. For example, an installation error in the zero setting for a flume might be + 1 mm for one flume and 2 mm for
another. For each flume the error is systematic, but for a number of flumes it would be a
random error.
Random errors. Random errors are caused by such things as the estimating required
between the smallest division on a head measurement device and water surface waves at
a head measuring device. Loose linkages between parts of flowmeters provide room for
random movement of parts relative to each other, causing subsequent random output
errors. Repeated readings decrease the average expected error resulting from random
errors by a factor of the square root of the number of readings.
Total error. Total error of a measurement is the result of systematic and random errors
caused by component parts and factors related to the entire system. Sometimes, error limits of all component factors are well known. In this case, total limits of simpler systems
can be determined by computation (Bos et al., 1991). In more complicated cases, it may
be difficult to confidently combine the limits. In this case, a thorough calibration of the
entire system as a unit can resolve the difference. In any case, it is better to do error analysis with data where entire system parts are operating simultaneously and compare discharge measurement against an adequate discharge comparison standard.
Expression of errors. Instrument errors are usually expressed by manufacturers as
either a percent of reading or a percent of full scale. The secondary devices based on electronic outputs are more frequently expressed in terms of percent full scale. The designer
must be aware that a probable error value of say 1 percent full-scale can exceed 10
percent for small value readings on the output device. When used with weirs, for example, the head reading of hl 5 in the weir equation can increase this 10 percent head measurement error to a 15 percent flow measurement error.
output. Other possible working standards are weights, volume containers, and stopwatches. More complicated devices are used, such as surveyors' levels, to check weir staff gage
zeros. Dead weight testers and electronic standards are needed to check and maintain
more sophisticated and complicated measuring devices, such as acoustic flow meters and
devices that use pressure cells to measure head.
For further measurement assurance and periodic checking, water users and organizations may keep secondary standards. Secondary standards are used to maintain integrity
and performance of working standards. These secondary standards can be sent to government laboratories, one of which is the National Bureau of Standards in Washington, D.C.,
to be periodically certified after calibration or comparison with accurate replicas of primary standards. Primary standards are defined by international agreement and maintained
at the International Bureau of Weights and Measurements in Paris, France.
Depending on accuracy needs, each organization should trace their measurement performance up to and through the appropriate level of standards. For example, turbine
acceptance testing, such as in the petroleum industry, might justify tracing to the primary
standards level.
I. Operating requirements
j. Ability to pass sediment and debris
k. Longevity of device for given environment
1. Maintenance requirements
m. Construction and installation requirements
n. Device standardization and calibration
o. Field verification, troubleshooting, and repair
p. User acceptance of new methods
q. Vandalism potential
r. Impact on environment
Accuracy requirements. The desired accuracy of the measurement system is an important consideration in the selection of a measurement method. Most water measurement
installations, including the primary and secondary devices, can produce accuracies of 5
percent. Some systems are capable of 1 percent under laboratory settings. However, in
the field, maintaining such accuracies usually requires considerable expense or special
effort in terms of construction, secondary equipment, calibration in-place, and stringent
maintenance. Selecting a device that is not appropriate for the site conditions can result in
a nonstandard installation of reduced accuracy, sometimes exceeding 10 percent.
Accuracies are frequently reported that relate only to the primary measurement method
or device. However, many methods require secondary measurement equipment that produces the actual readout. This readout equipment typically increases the overall error of
the measurement.
Cost. The cost of the measurement method includes the cost of the device itself, the
installation, secondary devices, operation, and maintenance. Measurement methods vary
widely in their cost and in their serviceable life span. Measurement methods are often
selected based on the initial cost of the primary device with insufficient regard for the
additional costs associated with providing the desired records of flows over an extended
period of time.
Legal constraints. Governmental or administrative water board requirements may
dictate the water measurement devices or methods. Water measurement devices that
become a standard in one geographic area may not necessarily be accepted as a standard
elsewhere. In this sense, the term "standard" does not necessarily signify accuracy or
broad legal acceptance. Many water agencies require certain water measurement devices
used within their jurisdiction to conform to their standard for the purpose of simplifying
operation, employee training, and maintenance.
Flow range. Many measurement methods have a limited range of flow conditions for
which they are applicable. This range is usually related to the need for certain prescribed
flow conditions which are assumed in the development of calibrations. Large errors in
measurement can occur when the flow is not within this range. For example, using a bucket and stopwatch for large flows that engulf the bucket is not very accurate. Similarly,
sharp-edged devices, such as sharp-crested weirs, typically do not yield good results with
large channel flows. These are measured better with large flumes or broad-crested weirs,
which in turn are not appropriate for trickle flows.
Certain applications have typical flow ranges. Irrigation supply monitoring seldom
demands a low-flow-to-high-flow range above about 30, while this range on natural
stream flows may exceed 1000.
In some cases, secondary devices can limit the practical range of flow rates. For example, with devices requiring a head measurement, the accuracy of the head measurement
from a visually read wall gage may limit the measurement of low flow rates. For some
devices, accuracy is based on percent of the full-scale value. While the resulting error may
be well within acceptable limits for full flow, at low flows, the resulting error may become
excessive, limiting the usefulness of such measurements. Generally, the device should be
selected to cover the desired range. Choosing a device that can handle an unnecessary
large flow rate may result in compromising measurement capability at low flow rates, and
vice versa. This choice depends on the objective of the measurement. For example, in irrigation practice, usually choose a device that can measure the most common flow range at
the expense of poorly measuring extremes, such as flood flows. For urban drainage, the
flood peak may be important.
For practical reasons, different accuracy requirements for high and low flows may be
chosen. This is reasonable when an annual total is the primary goal and the low flows contribute a small percentage to that total. Also, if the inaccurate low flow readings are truly
a random error then this error approaches zero with large accumulations of readings. Thus
the designer needs to know if management decisions are made from individual readings
or from long term averages.
Headloss. Most water measurement devices require a drop in head. On retrofit installations, for example, to an existing irrigation project, such additional head may not be
available, especially in areas that have relatively flat topography. On new projects, incorporating additional headloss into the design can usually be accomplished at reasonable
cost. However, a tradeoff usually exists between the cost of the device and the amount of
headloss. For example, acoustic flow meters are expensive but require little headloss.
Sharp-crested weirs are inexpensive but require a relatively large headloss. The head loss
required for a particular measuring device usually varies over the range of discharges. In
some cases, head needed by a flow measuring device can reduce the capacity of the channel at that point.
Adaptability to site conditions. The selection of a flow measuring device must
address the site of the proposed measurement. Several potential sites may be available for
obtaining a flow measurement. The particular site chosen may influence the selection of
a measuring device. For example, discharge in a canal system can be measured within a
reach of the channel or at a structure such as a culvert or check structure. A different
device would typically be selected for each site. The device selected ideally should not
alter site hydraulics so as to interfere with normal operation and maintenance. Also, the
shape of the cross-sectional flow area may favor particular devices.
Adaptability to variable operating conditions. Flow demands for most water delivery systems usually vary over a range of flows and flow conditions. The selected device
must accommodate the flow range and changes in operating conditions, such as variations
in upstream and downstream head. Weirs or flumes should be avoided if downstream
water levels can, under some conditions, cause excessive submergence. Also, the information provided by the measuring device should be conveniently useful for the operators
performing their duties. Devices that are difficult and time consuming to operate are less
likely to be used and are more likely to be used incorrectly.
In some cases, water measurement and water level or flow control are desired at the
same site. A few devices are available for accomplishing both (e.g., constant-head orifice,
vertically movable weirs, and Neyrpic flow module; Bos, 1989). However, separate measurement and control devices are typically linked for this purpose and usually can exceed
the performance of combined devices in terms of accuracy and level control, if care is
exercised to assure that the separate devices are compatible and achieve both functions
when used as a system.
Type of measurements and records needed. An accurate measure of instantaneous
flow rate is useful for system operators in setting and verifying flow rate. However,
because flow rates change over time, a single (instantaneous) reading may not accurately
reflect the total volume of water delivered. Where accounting for water volume is desired,
a method of accumulated individual flow measurements is needed. Where flows are
steady, daily measurements may be sufficient to infer total volume. Most deliveries, however, require more frequent measurements. Meters that accumulate total delivered volume
are desirable where water users take water on demand. Totalizing and automatic recording devices are available for many measuring devices. For large structures, the cost for
water-level sensing and recording hardware is small relative to the structure cost. For
small structures, these hardware costs remain about the same and thus become a major
part of the measurement cost, and may often exceed the cost of the primary structure
itself.
Many water measuring methods are suitable for making temporary measurements
(flow surveys) or performing occasional verification checks of other devices. The method
chosen for such a measurement might be quite different from that chosen for continuous
monitoring. Although many of these flow survey methods are suited for temporary operation, the focus here is on methods for permanent installations.
Operating Requirements. Some measurement methods require manual labor to
obtain a measurement. Current metering requires a trained staff with specialized equipment. Pen-and-ink style water-stage recorders need operators to change paper, add ink,
and verify proper functioning. Manual recording of flows may require printed forms to be
manually completed and data to be accumulated for accounting purposes. Devices with
manometers require special care and attention to assure correct differential-head readings.
Automated devices, such as ultrasonic flowmeters and other systems that use transducers
and electronics, require operator training to set up, adjust, and troubleshoot. Setting gatecontrolled flow rates by simple canal level references or by current metering commonly
requires several hours of waiting between gate changes for the downstream canal to fill
and stabilize. However, if a flume or weir is installed near the control gate, that portion of
the canal can be brought to the stable, desired flow level and measured flow rate in a few
minutes, and the canal downstream of the flume or weir can then fill to the correct level
over a longer time without further gate adjustments. Thus, the requirements of the operating personnel in using the devices and techniques for their desired purposes must be
considered in meter selection.
Some measuring devices may inherently serve an additional function applicable to the
operation of a water supply system. For example, weirs and flumes serve to hydraulicalIy isolate upstream parts of a canal system from the influence of downstream parts. This
occurs for free overfall weirs and flumes flowing below their modular limits. Acoustic,
propeller, magnetic, and vortex-shedding flowmeters do not provide this function without
additional structural measures such as a downstream overfall. If these meters are used, and
the isolation function is desired, then the designer should be made aware of the requirement and provide a free overfall. Isolating the influence of upstream changes from affecting downstream channels, is less easily accomplished. However, it can be partly implemented with orifices that have a differential head that is large compared to the upstream
fluctuations.
The designer should be aware that a sharp-crested weir overfall requires a relatively
high head drop and may need to be excessively wide to provide the isolation function
with low absolute head drop. While a long board can be used downstream from a propeller meter to provide the necessary width of flow that will pass a required quantity of
water at small head, that small head, and the crude board would not be well suited for
measuring flow rate.
The designer may wish to take advantage of broad-crested weir behavior and provide
a thick crest that can withstand in excess of about 80 percent submergence, which usually translates into low absolute head loss. When used with a propeller meter, for example, the broad-crested weir need not be well defined and can be economically installed
(Replogle, 1997).
Ability to pass sediment and debris. Canal systems often carry a significant amount
of sediment in the water. Removal of all suspended solids from the water is usually prohibitively expensive. Thus, some sediment will likely be deposited anywhere the velocities are reduced, which typically occurs near flow measuring structures. Whether this sediment causes a problem depends on the specific structure and the volume of sediment in
the water. In some cases, this problem simply requires routine maintenance to remove
accumulated sediment; in others, the accumulation can make the flow measurement inaccurate or the device inoperative. Sediment deposits can affect approach conditions and
increase approach velocity in front of weirs, flumes, and orifices. Floating and suspended
debris such as aquatic plants, washed-out bank plants, and fallen tree leaves and twigs can
plug some flow measurement devices and cause significant flow measurement problems.
Many of the measurement devices which are successfully used in closed conduits (e.g.,
orifices, propeller meters, and so on) are not usable in culverts or inverted siphons because
of debris in the water. Attempting to remove this debris at the entrance to culverts is an
additional maintenance problem.
Flumes, especially long-throated flumes, can be designed to resist sedimentation. The
design options available are to select a structure shape that will maintain velocities that
assure erosion of sediments, or at least continued movement of incoming sediments
through the flume, at important flow rates. In large broad-crested weirs (a class of longthroated flumes) for capacities greater than 1 m3/s per m of flume width, velocities greater
than 1 m/s can be achieved for the upper 75 percent of the flow range, and is usually erosive enough to maintain flume function even for high-sediment bed loads. At the lower
flow ranges and for heavy sediment bed loads, deposition is likely and frequent maintenance may be required.
Trapezoidal sections tend to retain low velocities into the upper ranges of flow and
are less sediment worthy. Long-throated flumes with flat bottoms throughout and side
contractions maintain a high velocity for 0.5 m3/s per m width, and higher, but must
have throat lengths that are 2 to 3 times the throat width in order to be accurately computable. The sediment worthiness of a flume design depends more on these absolute
velocities than on whether the flume floor is flat throughout or raised as in a broadcrested weir. This prompts the designer to select shapes that can provide these velocities. One suggestion for broad-crested weirs in a fixed sized channel is to construct a
false floor in the head gage area to increase the velocity there and prevent changes in
area of flow there. Also, sediments can accumulate in the upstream channel to a depth
of the false floor without affecting the function of the flume. This can extend the time
between mandatory channel cleaning.
Device environment. Any measurement device with moving parts or sensors is subject to failure if it is not compatible with the site environment. Achieving proper operation
and longevity of devices is an important selection factor. Very cold weather can shrink
moving and fixed parts differentially and solidify oil and grease in bearings. Water can
freeze around parts and plug pressure ports and passageways. Acidity and alkalinity in
water can corrode metal parts. Water contaminants such as waste solvents can damage
lubricants, protective coatings, and plastic parts. Mineral encrustation and biological
growths can impair moving parts and plug pressure transmitting ports. Sediment can
abrade parts or consolidate tightly in bearing and runner spaces in devices such as propeller meters.
Measurement of wastewater and high sediment transport flow may preclude the use of
devices that require pressure taps, intrusive sensors, or depend upon clear transmission of
sound through the flow. Water measurement devices that depend on electronic devices and
transducers must have appropriate protective housings for harsh environments. Improper
protection against the site environment can cause equipment failure or loss of accuracy.
Maintenance requirements. The type and amount of maintenance varies widely with
different measurement methods. For example, current metering requires periodic maintenance of the current meter itself and maintenance of the meter site to assure that is has a
known cross section and velocity distribution. When the flow carries sediment or debris,
most weirs, flumes, and orifices require periodic cleaning of the approach channel. As
mentioned above, design and meter selection can mitigate the maintenance problems with
sediments, but are not likely to eliminate them. Electronic sensors need occasional maintenance to ensure that they are performing properly. Regular maintenance programs are
recommended to ensure prolonged measurement quality for all types of devices.
Construction and installation requirements. In addition to installation costs, the difficulty of installation and the need to retrofit parts of the existing conveyance system can
complicate the selection of water measurement devices. Clearly, devices that can be easily retrofitted into the existing canal system are much preferred because they generally
require less down time, and usually present fewer unforeseen problems.
Device standardization and calibration. A standard water measurement device infers
a documented history of performance based on theory, controlled calibration, and use. A
truly standard device has been fully described, accurately calibrated, correctly constructed, properly installed, and sufficiently maintained to fulfill the original installation
requirements and flow condition limitations. Discharge equations and tables for standard
devices should provide accurate calibration. Maintaining a standard device usually only
involves a visual check and measurement of a few specified items or dimensions to ensure
that the measuring device has not departed from the standard. Many standard devices have
a long history of use and calibration, and thus are potentially more reliable. Commercial
availability of a device does not necessarily guarantee that it satisfies the requirements of
a standard device.
When measuring devices are fabricated onsite or are poorly installed, small deviations from the specified dimensions can occur. These deviations may or may not affect
the calibration. The difficulty is that unless an as-built calibration is performed, the
degree to which these errors affect the accuracy of the measurements is largely
unknown. All too frequently, design deviations are made under the misconception that
current metering can be used to provide an accurate field calibration. In practice, calibration by current metering to within 2 percent is difficult to attain. An adequate calibration for free-flow conditions requires many current meter measurements at several
discharges. Changing and maintaining a constant discharge for calibration purposes is
often difficult under field conditions.
Field verification, troubleshooting, and repair. After construction or installation of a
device, some verification of the calibration is generally recommended. Usually, the meth-
ods used to verify a permanent device (e.g., current metering) are less accurate than the
device itself. However, this verification simply serves as a check against gross errors in
construction or calibration. For some devices, errors occur as components wear and the
calibration slowly drifts away from the original. Other devices have components that simply fail, that is, you get the correct reading or no reading at all. The latter is clearly preferred. However, for many devices, occasional checking is required to ensure that they are
still performing as intended. Selection of devices may depend on how they fail and how
easy it is to verify that they are performing properly.
User acceptance of new methods. Selection of a water measurement method must
also consider the past history of the practice at the site. When improved water measurement methods are needed, proposing changes that build on established practice are generally easier to institute than radical changes. It can be beneficial to select a new method
that allows conversion to take place in stages to provide educational examples and demonstrations of the new devices and procedures.
Vandalism potential Instrumentation located near public access is a prime target for
vandalism. Where vandalism is a problem, measurement devices with less instrumentation, or instrumentation that can be easily protected, are preferred. When needed, instrumentation can be placed in a buried vault to minimize visibility.
Impact on environment. During the selection of a water measurement device, consideration must be given to potential environmental impacts. Water measurement devices
vary greatly in the amount of disruption to existing conditions that is needed for installation, operation, and maintenance. For example, installing a weir or flume constricts the
channel, slows upstream flow, and accelerates flow within the structure. These changes in
the flow conditions can alter local channel erosion, local flooding, public safety, local
aquatic habitat, and movement of fish up and down the channel. These factors may alter
the cost and selection of a measurement device.
21.5.3 Selection Guidelines
Selection of a water measurement method can be a difficult, time-consuming design
process if one were to formally evaluate all the factors discussed above for each measuring device. This difficulty is one reason that standardization of measurement devices within water agency jurisdictions is often encouraged by internal administrators. However,
useful devices are sometimes overlooked when devices similar to previous purchases are
automatically selected. Therefore, some preliminary guidance on selection is offered to
the designer so that the number of choices can be narrowed down before a more thorough
design analysis of the tradeoffs between alternatives is performed.
Short list of devices based on application. The list of practical choices for a water
measuring device is quickly narrowed by site conditions because most devices are applicable to a limited range of channel or conduit conditions. Economics also limits applicable
devices. For example, few irrigation deliveries to farms can justify expensive acoustic
meters. Likewise, using current meters for manual flow measurement in a channel is
appropriate for intermittent information but is usually too labor intensive for use on a continuous basis. Table 21.1 provides a list of commonly used measurement methods that are
considered appropriate for each of several applications. Table 21.2 provides an abbreviated table of selection criteria and general compliance for categories of water measurement
devices. The symbols (+), (O), and () are used to indicate relative compliance for each
selection criterion. The (+) symbol indicates positive features that might make the device
Sharp
Crested
Broad
Crested
LongThroated
Short Submerg
Throated Orifice
Criteria
Weirs
Weirs
Flumes
Devise
Current
Metering
Accuracy
O
O
O
O
O
+
Cost
O
O
+
O
O
+
+
+
Flows > 5 m-Vs
O
O
+
+
+
+
Flows < 0.25 mVs
O
O
+
Flow span
+
O
O
O
+
Headloss
+
O
O
Site Condition
+
Lined canal
+
+
O
O
O
Unliner canal O
+
O
O
O
O
O
Short, full pipe NA
NA
+
NA
NA
NA
NA
NA
Closed conduit NA
NA
NA
NA
NA
NA
NA
NA
Measurement Type
+
+
+
Rate
+
+
+
+
+
Volume
O
Sediment
+
+
Sediment pass
O
O
O
O
2
+
+
+
+
Debris pass
+
Longivity
+
+
+
Moving parts +
+
+
O
O
+
+
+
+
Electr. requir. +
O
O
+
+
+
+
Maintenance
O
O
+
+
+
+
+
Construction
O
O
+
+
+
Field verify
O
O
+
+
+
Standarization
O
O
O
Source: Adapted from USER (1997).
Symbols 1, O, 2 are used as relative indicators comparing application of the listed water measuring device to the listed criteria
Symbol V denotes that situability varies widely.
Symbol NA denotes "not applicable" to criteria
"Venturi,
orifice, pilot tube, etc.
f
Propeller meters, turbine meters, paddle wheel meters, etc.
(Pipes)
(Pipes)
4-
O
+
O
O
O
O
+
NA
NA
O
+
V
V
O
O
+
NA
NA
V
NA
NA
O
+
+
V
+
O
O
O
+
+
Acoustic
Doppler
Acoustic Acoustic
Transonic Transon.
Pipe
(Pipes) (Pipe, 1 Path) (Multipath)
+
O
O
+
O
O
O
+
O
+
+
+
NA
NA
NA
NA
NA
NA
+
+
O
O
O
O
+
+
O
O
O
O
+
O
O
O
O
O
+
+
+
O
O
O
O
2
O
O
O
attractive from the standpoint of the associated selection criteria. A (-) symbol indicates
negative aspects that might limit the usefulness of this method based on that criterion. A
(O) indicates no strong positive or negative aspects in general. A (V) means that the suitability varies widely for this class of devices. The letters NA mean that the device is not
applicable for the stated conditions. A single negative value for a device does not mean
that the device is not useful or appropriate, but other devices would be preferred for those
selection criteria.
Vortex-shedding flow meters are not specifically rated in this grouping. They are
expected to compete with orifice meter applications. They generally offer less head loss
that orifice meters and can cover a wider discharge range for a particular installation.
Although they have been around for many years, they have only recently been offered in
a configuration that makes them competitive with orifice meters, which they are generally expected to replace because they can produce less pipe head loss. Open-channel applications for vortex-shedding meters are not considered practical.
Next Page
CHAPTER 22
WATER AND WASTEWATER
TREATMENT PLANT
HYDRAULICS
INTRODUCTION
Designers of water treatment plants and wastewater treatment plants are faced with the
need to design treatment processes which must meet the following general hydraulic
requirements:
Water treatment plants. Provide the head required to allow the water to flow through
the treatment processes and to be delivered to the transmission/distribution system in
the flow rates and at the pressures required for delivery to the users.
Wastewater treatment plants. Provide the head required to raise the flow of wastewater
from the sewer system to a level which allows the flow to proceed through the treatment processes and be delivered to the receiving body of water.
The above requires knowledge of open-channel, closed-conduit, and hydraulic
machine flow principles. It also requires an understanding of the interaction between these
elements and their impact on the overall plant (site) hydraulics. Head is either available
through the difference in elevation (gravity) or it has to be converted from mechanical
energy using hydraulic machinery. Distribution of flows using open channels or closed
conduit is critical for proper hydraulic loading and process performance.
This chapter brings together information on commonly used hydraulic elements and
specific applications to water treatment plants and wastewater treatment plants. The
development of hydraulic profiles through the entire treatment process with examples for
water treatment plants and wastewater treatment is also presented.
Many processes and flow control devices are similar in both water treatment plants and
wastewater treatment plants. Both types of plants require flow distribution devices, gates
and valves, and flowmeters. These devices are discussed in Section 22.2. The development of water treatment plant hydraulics, including examples from in-place facilities, are
presented in Section 22.3. Wastewater treatment plant hydraulics are discussed in Section
22.4, and Section. 22.5 is devoted to non-Newtonian flow principles.
22.2
GENERAL
22.2.1 Introduction
This section addresses some elements which are common to both water treatment plants
and wastewater treatment plants including:
Flow distribution-manifolds
Gates and valves
Flowmeters
Local losses
22.2.2
Flow distribution-manifolds
In the design of water and wastewater treatment plants, proper flow distribution can be as
critical as process design considerations, which typically receive much more attention.
Plant failures resulting from unequal and unmanageable flow distribution are possibly as
common and as serious as those resulting from errors in process design.
Flow distribution devices, such as distribution channels, pipe manifolds or distribution
boxes, are commonly used to distribute or equalize flow to parallel treatment units, such
as flocculation tanks, sedimentation basins, aeration tanks, or filters.
22.2.2.1 Distribution boxes. The simplest of these devices, the distribution box, typically consists of a structure arranged to provide a common water surface as the supply to two
or more outlets. The outlets are typically over weirs and the key to equal flow distribution
is to provide independent hydraulic characteristics between the downstream system and
the water level in the distribution box. In other words, provide a free discharge weir (nonsubmerged under all conditions) for each outlet to eliminate the impact of downstream
physical system differences on the flow distribution. Velocity gradients across the distribution box must be nearly zero to equalize flow conditions over each outfall weir. Weirs
clearly should be of uniform design in terms of physical arrangement length and materials of construction. They should also be adjustable to account for any minor flow differences noted in actual operation. The same principles apply if the designer wishes to distribute flows in specific proportions which are not necessarily equal. In this case the
designer could control the proportions of flow distribution by varying the relative geometry of the weirs (i.e., change the width or invert of each weir to achieve a desired flow
distibution). The specifics of weir hydraulics are covered in various texts in the literature.
Attention should always be paid to the selection of the proper coefficients to model the
specific weir geometry and the geometry of the approach flow.
22.2.2.2 Distribution channels and pipe manifolds. Distribution channels and manifolds
are also common in plant design but a bit more complex in their function and design. The
distribution of flow in these devices is impacted by the flow distribution itself. Since a
portion of the flow leaves the channel or manifold along the length of the device, the
velocity of flow and, therefore, the relationship of energy grade line, velocity head and
hydraulic grade line varies along the length of the device. This is more clearly visible in
a distribution channel of uniform cross section, using side weirs along its length for flow
distribution. At each weir, flow leaves the channel, resulting in less velocity head in the
channel and possibly a higher water surface at each ensuing weir. Chao and Trussell
(1980), Camp and Graber (1968), and Yao (1972) have presented comprehensive
approaches for the design of distribution channels and manifolds and should be reviewed
for details of design.
As in distribution boxes, the most important consideration to achieving equalized flow
distribution is to minimize the effects of unequal hydraulic conditions relative to each
point of distribution. In channels this can be accomplished by tapering the channel cross
section, varying weir elevations, making the channel large enough to cause velocity head
changes to be insignificant or a combination of these. Similar considerations may be
applied to manifolds with submerged orifice outlets. A reliable approach here is to provide
a large enough manifold, resulting in a total headloss along the length of the distribution
of less than one tenth the loss through any individual orifice. This approach essentially
results in the orifices becoming the only hydraulic control and the accuracy of the flow
distribution is then dependent on the uniformity of the orifices themselves.
22.2.3
Gates and valves generally serve to either control the rate of flow or to start/stop flow.
Gates and valves in treatment plants are typically subjected to much lower pressures
than those in water distribution systems or sewage force mains and can be of lighter
construction.
22.2.3.1 Gates. Gates are typically used in channels or in structures to start and stop flow
or to provide a hydraulic control point which is seldom adjusted. Because of the time and
effort required to operate gates, they are not suited for controlling flow when rapid
response, frequent variation, or delicate adjustments are needed. Primary design considerations when using gates are the type of gate fabrication and the installation conditions
during construction.
There are many fabrication details including materials used, bottom arrangement, and
stem arrangement. For instance, for solids bearing flows, a flush bottom, rising stem gate
can be used to avoid creating a point of solids deposition and to minimize solids contact
with the threaded stem. Gate manufacturers are a good source of information for gate fabrication details and can assist with advice regarding specific applications.
Most commonly used gates are designed to stop flow in a single direction. They may
use upstream water pressure to assist in achieving a seal (seating head), but typically also
must be designed to resist static water pressure from downstream (unseating head). Both
seating and unseating heads must be evaluated in design of a gate application. For most
manufacturers, the seating or unseating head is expressed as the pressure relative to the
center line of the gate.
22.2.3.2 Valves. Table 22.1 provides a summary of several types of valves and their
applications. Valves are used to either throttle (control) flow or start/stop flow.
Start/stop valves are intended to be fully open or fully closed and nonthrottling. They
should present minimum resistance to flow when fully open and should be intended for
infrequent operation.
Gate valves, plug valves, cone valves, ball valves, and butterfly valves are the most
common start/stop valve selections. Butterfly valves have a center stem, are most common
in clean water applications and should not be used in applications including materials that
could hang-up on the stem. Therefore, they are seldom used at wastewater plants prior to
achieving a filter effluent water quality.
Open/Close
Sluicegate
Slide gate
Gate valve
Plug valve
Cone valve
Ball valve
Butterfly valve
Swing check
Lift check
Ball check
Spring check
Globe valve
Needle valve
Angle valve
Pinch/diaphragm
X
X
X
X
X
X
X
X
X
X
X
X
Throttling
X
X
X
X
X
X
X
Water
Wastewater
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
*Typical applications-exceptions are possible, but consultation with valve manufacturers is recommended.
Check valves are a special case of a start/stop valve application. Check valves offer
quick, automatic reaction to flow changes and are intended to stop flow direction reversal. Typical configurations include swing check, lift check, ball check and spring loaded.
These valves are typically used on pump discharge piping and are opened by the pressure
of the flowing liquid and close automatically if pressure drops and flow attempts to
reverse direction. The rapid closure of these valves can result in unacceptable "waterhammer" pressures with the potential to damage the system. A detailed surge analysis may
be required for many check valve applications (see Chapter 12). At times, mechanically
operating check valves should be avoided in favor of electrically or pneumatically operated valves (typically plug, ball, or cone valves) to provide a mechanism to control time
of closing and reduce surge pressure peaks.
Throttling valves are used to control rate of flow and are designed for frequent or nearly continuous operation depending on whether they are manually operated or electronically controlled. Typical throttling valve types include globe valves, needle valves, and
angle valves in smaller sizes, and ball, plug, cone, butterfly, and pinch/diaphragm valves
in larger sizes. Throttling valves are typically most effective in the mid-range of loose line
open/close travel and for best flow control should not be routinely operated nearly fully
closed or nearly fully open.
22.2.4 Flow meters
The most common types of flow meters used in water and wastewater treatment plants are
summarized in Table 22.2 and fall into the following categories:
Headloss
Cost
WW
1-120 in
Any size
Low
Medium
Medium
Low
X
X
X
X
Pitottube
0.5-5% of scale
Parshall flume 5% of rate
1/2-96 in
Wide range
Low
Low
Low
Medium
X
X
Magnetic
Doppler
Propeller
Turbine
1/10-120 in
1/8-120 in
Up to 24 in
Up to 24 in
None
None
High
High
High
High
High
High
X
X
X
X
Venturi
Orifice plate
Typical Accuracy
0.75% of rate
2% of scale
0.5% of rate
1-2.5% of rate
2% of rate
0.5-2% of rate
Size Range
X
X
Pressure differential/pressure measuring meters (e.g., Venturi, orifice plate, pitot tube,
and Parshall flume meters)
Magnetic meters
Doppler (ultrasonic) meters
Mechanical meters (e.g., propeller and turbine meters)
Accurate flow measurements require uniform flow patterns. Most meters are
significantly impacted by adjacent piping configurations. Typically a specific number of
straight pipe diameters is required both upstream and downstream of a meter to obtain
reliable measurements. In some cases, 15 straight pipe diameters upstream and 5 straight
pipe diameters downstream are recommended. However, different types of meters have
varying levels of susceptibility to the uniformity of the flow pattern. Meter manufacturers
should be consulted.
22.2.4.1 Pressure differential/pressure measuring meters. Pressure differential/pressure
measuring flow meters include Venturi meters, orifice plates, averaging pitot meters, and
Parshall flume s. These meters measure the change in pressure through a known flow cross
section-or in the case of the pitot meter, measure the difference in pressure at a point in
the flow versus static pressure just downstream in a uniform section of pipe.
Venturi meters and orifice plates are commonly used in water and wastewater. Solids
in wastewater could plug the openings of a pitot tube meter-limiting their use to relatively clean liquids. The Venturi meter and orifice plate meter use pressure taps at the wall of
the device and can be arranged to minimize potential for debris from clogging the taps.
The Parshall flume can be arranged with a side stilling well and level measuring float system or an ultrasonic level sensing device to measure water level.
22.2.4.2 Magnetic meters. In a magnetic flowmeter, a magnetic field is generated around
a section of pipe. Water passing through the field induces a small electric current proportional to the velocity of flow. Because a magnetic meter imposes no obstruction to the
flow, it is well suited to measuring solids bearing liquids as well as clean liquids and produces no headless in addition to the normal pipe loss. Magnetic meters are among the least
susceptible to the uniformity of the stream lines in the approaching flow.
22.2.4.3 Ultrasonic meters. In an ultrasonic flow meter, a pair of transceivers (transmitter/receiver) are positioned diagonally across from each other on the pipe wall. The
transmitter sends out a signal which is affected by the speed of the flow. The receiver measures the difference between the speed of the signal when directed counter to the flow
(slowed by the flow) and when directed with the flow (speeded up by the flow). The time
difference is a function of fluid velocity, which is used to compute the flow. As with magnetic meters, no flow obstruction is imposed resulting in no headloss in addition to the
normal pipe loss. Ultrasonic meters are also available for open-channel applications.
22.2.4.4 Mechanical meters. Mechanical meters include propeller and turbine-type
equipment. The two meters are similar in function in that in each a device is inserted into
the flowpath. The device is rotated by the flow and the speed of rotation is used to compute rate of flow. These devices impose an obstruction to flow, are recommended for clean
water only, and generally result in significant headloss.
22.2.5
Local Losses
In any piping system as flow travels along the pipe, pressure drops as a result of headloss
due to friction along the pipe and local losses at bends, fittings, and valves. The local
losses at bends, fittings, and valves are least significant in long, straight piping systems
and most significant at treatment plants where the length of straight pipe is relatively short
and therefore, the frictional pipe losses comprise a smaller fraction of the total losses
when compared to the summation of all local losses. A term often used to refer to local
losses is "minor losses," however, because of the later consideration the term "minor losses" can be misleading.
Traditionally, local losses have been computed in terms of "equivalent length" of
straight pipe or in terms of multiples of velocity head. The "equivalent length" or loss factor K methods attempt to estimate the local losses based on the characteristic of the specific bend, fitting or valve. The K loss factor method is discussed here. Essentially, a local
loss is computed as follows:
L =K^
(22-D
Gate valve
100% open
0.39
75% open
1.1
50% open
4.8
25% open
27
Globe valve-open
10
Angle valve^open
4.3
Check valve-ball
4.5
Swing check
Butterfly valve-open
1.2
Foot valve-hinged
2.2
Foot valve-poppet
12.5
Elbows
45 regular
45 long radius
90 regular
90 long radius
180 regular
180 long radius (flanged)
Tees
Std. teee-flowthrough run 0.6
Std. teee-flow-through branch 1.8
Return bend
1.5
Mitre bend
90
1.8
60
0.75
30
0.25
Expansion
d/D = 0.75
0.18
d/D = 0.5
0.55
d/D = 0.25
0.88
Contraction
d/D = 0.75
0.18
d/D = 0.5
0.33
d/D = 0.25
0.43
Entrancee-projecting
0.78
Entrancee-sharp
0.5
Entrancee-well rounded
0.04
Exit
1.0
10
5 2.1-3.1
65-70
0.6-2.3
0.19
1.15
5.6
24
0.30-0.42
0.18-0.20
0.21-0.3
0.14-0.23
0.38
0.25
0.19
0.1-0.3
10
5
4.0-6.0
1.8-2.9
0.42
0.6
1.8
2.2
0.6
1.8 1.8
2.2
1.129-1.265
0.471-0.684
0.130-0.165
0.78
0.5
0.04
1.0
Sanks (1989)
Simon (1986)
Daugherty (1977)
Cameron Hydraulic Data
Walski (1992)
0.2
1.2
5.6
24
10
2.5
06-2.2
0.16-0.35
1.0-1.4
5.0-14.0
0.2
1.2
5.6
25
10
5
5
0.6-2.5 2.5
0.5
2.2
14
0.18
0.25
0.18
0.5
0.7
0.6
0.3
0.75
0.4
1.8
0.8
0.35
0.1
0.42
0.2
0.25
0.19
0.38
0.25
0.6
1.8
2.2
1.3
0.6
0.16
0.19
0.56
0.92
0.2
0.6
0.9
0.2
0.6
0.9
0.19
0.33
0.42
0.83
0.5
0.2
0.3
0.4
0.78
0.5
0.04
1.0
0.2
0.33
0.43
0.8
0.5
0.04
1.0
0.8
0.5
0.040.04
0.8
0.5
0.25
1.0
22.3
22.3.1 Introduction
Water treatment comprises the withdrawal of water from a source of supply and the
treatment of raw water through a series of unit processes for the beneficial use of the
system customers. Raw water quality can vary widely. The ultimate uses of water by the
system customer (e.g., drinking, fire protection, irrigation, aquifer recharge, etc.) can also
vary and be subject to different treatment level requirements and regulations. Therefore,
the selected treatment processes vary widely over a multitude of treatment technologies
in use. Water treatment consists of a series of chemical, biological, and physical processes connected by channels and pipelines. Figures 22.1 and 22.2 illustrate process
flow diagrams (flowsheets) for typical surface water and groundwater treatment plants,
respectively. The designer of the water treatment process must carefully evaluate source
water characteristics and desired water quality characteristics of the treated water to
design treatment processes capable of purifying the source water to water suitable for the
system customers. The objective of this chapter is to review the hydraulic considerations
required to convey water through the treatment process.
Design of a plant's treatment process is closely linked with the hydraulic design of the
treatment plant. This chapter presumes that the designer has evaluated and selected treatment processes for the water treatment plant. Although design flows are discussed below,
we have also assumed that the designer has chosen a design flow requirement for the treatment process. For municipal treatment plants, design flows are based on the service area
FLOCCULATION/COAGULATION
RAPID MIX
CLEARWELL
TO
DISTRIBUTION
RAW
WATER
SEDIMENTATION
FILTRATION
- RAW WATER CONTROL
CHAMBER/RAPID MIX
FINISHED
WATER
PUMPING
FIGURE 22.1 Typical surface water treatment plant process flow diagram.
PRESSURE FILTRATION
TO
DISTRIBUTION
WELL
PUMP
WELL
CHLORINE
CONTACT
CHAMBER
population and the per capita use of water by the population served. The per capita use of
water can be obtained from literature sources as an initial approximation. However, these
initial estimations must be corroborated with actual site specific population counts and
water usage. For nonmunicipal treatment facilities, treated water needs of the service area
must be individually evaluated.
22.3.1.1 Sources of supply. Natural sources of supply include groundwater and surface
water supplies. Groundwater supplies typically are smaller in daily delivery but serve
more systems than surface water supplies. Groundwater supplies normally come from
wells, springs, or infiltration galleries.
Wells constitute the largest source of groundwater. Except in rare circumstances of
artesian wells (wells under the influence of a confined aquifer) and springs, groundwater
collection involves pumping facilities. Hydraulics of groundwater treatment plants are
frequently based on hydraulics of conduits under pressure, such as pipelines, pressure
filters, and pressure tanks. Raw water characteristics of groundwaters are uniform in
quality compared with surface supplies.
Surface water supplies are normally larger in daily delivery. Surface supplies are used
to service larger population centers and industrial centers. In areas where groundwater
supplies are limited in yield or where groundwater supplies contain undesirable chemical
characteristics, smaller surface water treatment plants may be utilized. Surface water
sources of supply include rivers, lakes, impoundments, streams, and ponds. The treatment
processes chosen in plants treating surface water favor nonpressurized systems such as
gravity sedimentation. The larger flow volumes characteristic of surface water supplies
also favor open channel hydraulic structures for conveying water through the treatment
process. Raw water characteristics of surface supplies can vary rapidly over short periods
of time and also experience seasonal variation.
22.3.1.2 Treatment requirements. Treatment requirements for municipal water treatment plants are normally defined by regulatory agencies having authority over the plant's
service area. In the United States, regulatory agencies include national government regulations promulgated through the Environmental Protection Agency and state government
regulations. Water treatment plants are designed to meet these regulations. Treatment regulations change through improved knowledge of health effects of water constituents and
through identification of possible new water-borne threats. The designer therefore should
attempt to select treatment processes which will also meet treatment requirements which
are expected to be promulgated over the next few years. To the extent possible, treatment
plant process design should provide flexibility for future plant expansions or for possible
additional treatment processes. Because hydraulic design of plants must go hand-in-hand
with the process selection, plant hydraulic design should provide for the flexibility to add
future plant facilities.
Treatment requirements for industrial water treatment plants are dictated by process
needs of the industry and less by regulatory agency requirements. Industrial water treatment plants that result in contact between or ingestion of the treated water by humans
must conform to the local regulatory requirements.
22.3.1.3 General design philosophy. Effective design of water treatment plant
hydraulics requires that the hydraulic designer have a thorough knowledge of all aspects
of the water system. The overall treatment system hydraulic design must be integrated and
coordinated including the treatment plant, the raw water intake and pumping facilities, the
treated water storage, and treated water pressure/head requirements. The design within the
water treatment plant must also be integrated between the various treatment processes.
Unit Process
Head Requirement
at Rated Capacity, m (ft)
0.3-0.9 (1-3)
0.15-0.30 (0.5-1)
0.06-0.15 (0.2-0.5)
0.6-2.4 (2-8)
3-4.6 (10-15)
3-7.6 (10-25)
0.15-0.6 (0.5-2)
3-4.6 (10-15)
3-4.6 (10-15)
0.15-0.6 (0.5-2)
0.15-0.6(0.5-2)
0.6-1.5 (2-5)
0.6-1.5 (2-5)
POINT
RAW WATER
CONTROL CHAMBER
MIXING
CHAMBER
FLOCCULATION/
SEDIMENTATION
BASINS
CLEARWELL
FINISHED WATER
PUMPING STATION
WATER SURFACE AT DESIGN
MAXIMUM HOUR FLOW
FIGURE 22.3 Hydraulic profile.
After the plant site has been identified, the plant design may be arranged for optimal
hydraulic benefit. In particular, arrangement of treatment processes to allow flow to move
down gradient minimizes excavation needs for structures. Arrangements which are
designed for future expansion should consider the hydraulic needs of the expanded plant
as well as the process needs. Grouping of processes together facilitates movement of
water through the treatment process train.
The designer should also consider secondary hydraulic systems for optimal design.
Chemical feed systems and dewatering systems are examples of secondary hydraulic
systems which must be coordinated with the treatment flow system. Normally it is
desirable to minimize the length of chemical piping systems. Dewatering systems are usually based on gravity drainage of basins and conduits.
22.3.5
After evaluation and selection of a source of supply and development of the treatment
plant process train, the designer is prepared to develop the plant Bases for Design. The
Bases for Design is a summary of design flow and capacity, and proposed treatment
processes, including the chemical storage and feed facilities.
22.3.5.1 Design flows. Design flows for water treatment plants serving municipalities
are typically based on the projected population within the water service area for the
design life of the treatment facilities. Population data is normally determined from
census records, land use zoning information, and studies of existing and projected
population densities. Service area per capita demands are affected by the mix of
domestic, commercial, and industrial water users which are unique to each service
area.Typically water consumption records are available for water service areas. For new
facilities, the use of generalized water consumption data may be needed. In the United
States, water consumption varies widely but generally ranges between 100-200 gallons per capita per day.
From studies of projected population and per capita demand, planned design flows for
the water treatment facilities may be developed. These demands include the following:
Annual average demand. The average daily water consumption for the water service
areas, generally computed by multiplying the average daily consumption (gallons per
capita) by the projected population of the service area.
Maximum demand. Maximum demand experienced by the water plant throughout its
service life. The maximum hour demand is generally 200 to 300 percent of the average demand but numerous factors affect the peak demand experienced by water treatment plants. These factors include seasonal demands (particularly for plants where service areas are located in extremes of hot and cold temperatures), normal daily flow
variations, the community size, industrial usage, and system storage. Normally system
storage is provided to service peak hour demands, allowing the treatment facilities to
be designed on peak day demands. Peak day demands generally range between 125
and 200 percent of the average demand.
Minimum flow. As the name suggests, the minimum flow expected to be processed
through the treatment facilities. Minimum flow depends upon system operations. In
general, minimum flows for municipal plants may be estimated as 50 percent of the
average demand, but range between 25 and 75 percent of the average demand.
22.3.5.2 Rated treatment capacity. The rated treatment capacity of a plant is that capacity for which each of the unit processes are designed. For municipal treatment plants with
adequate system storage, the rated treatment capacity is the system's maximum day
demand. Where storage is limited, the rated treatment capacity may be greater, for example, the system maximum hour demand or greater. Smaller systems may be designed to
produce the rated treatment capacity in one or two 8-h shifts rather than over the entire
24-h day.
22.3.5.3 Hydraulic treatment capacity. Treatment plants are normally designed for a
hydraulic capacity greater than the rated treatment capacity. Hydraulic treatment capacities are normally equal to 125 to 150 percent of the rated treatment capacity. The hydraulic
treatment capacity provides flexibility for future process changes or alternative flow routings through the plant. Hydraulic capacities in excess of the rated treatment capacity provide some margin of safety for operations which may not be optimal (e.g., control gates
inadvertently left partially open).
22.3.5.4 Treatment process bases for design. The development of the water treatment
plant's "Bases for Design" is a key step in establishing the criteria to which the plant will
be designed. This document must be reviewed carefully with the water treatment plant
owner representatives and understood and agreed to by all before the final design proceeds. The Bases for Design presents a summary of each treatment process including
design flows (minimum, average, rated capacity), specification of dimension of major elements (e.g., tanks, pumps), both hydraulic and process loading characteristics, required
performance, and design data for the chemical storage and feed system. Table 22.4 presents an example of the bases for design for sedimentation basins (one of the many unit
processes in a water treatment plant).
22.3.6 Plant Hydraulic Design
As noted above, a water treatment plant consists of a series of treatment processes
connected by free surface flow channels and pipelines. During development of the plant's
Number of basins
Basin characteristics
Plan-75 ft X 230-6 in
Nominal side water
depth-12 ft (SWD)
Surface area/basin- 17, 28 8 ft2
Volume/basin-207,456 f3
Channels/basin-2
L:Wratio-6.1:l
Displacement time (h)
Surface loading [(gal'm)/ft2]
Flowthrough velocity (ft/min)
Sludge collectors
Longitudinal collectors
Type: chain flight
Number per basin
Cross collectors
Type: chain flight
Number per basin
Settled sludge pumps
Type: progressive cavity
Number:
100 gal/min capacity
400 gal/min capacity
200 gal/min capacity
Capacity (gal/min)
Installed
Firm
Stage I
Stage II
StageIII
Annual
Average
4
Maximum
Day
4
Annual
Average
8
Maximum
Day
8
Annual
Average
12
Maximum
Day
12
3.17
0.47
1.21
1.99
0.75
1.93
3.17
0.47
1.21
1.99
0.75
1.93
3.17
0.47
1.21
1.99
0.75
1.93
4
4
4
4
4
4
8
4
4
8
4
4
16
4
4
16
2000
1600
2000
1600
3600
3200
3600
3200
5200
4800
5200
4800
Bases for Design, the designer determines the rated treatment capacity, average flow,
minimum flow and hydraulic capacity of the plant.
Following development of the Bases for Design, the designer must evaluate plant
operating modes to develop a detailed plant flow diagram and hydraulic profile
through the plant.
22.3.6.1 Plant operating modes. Operating modes describe the sequence of treatment
processes the water goes through to achieve the required level of purification. Operational
modes are normally presented in the form of simplified block diagrams which illustrate
the flow path through the plant from one process to the next. These operational mode
block diagrams are useful in visualizing stages during construction, future planned plant
expansions or simply alternative operating modes.
Figures 22.4 through 22.9 show an example of a sequence of plant operating modes
for a surface water treatment plant which illustrate three stages of a plant expansion program with alternatives for the flocculation and sedimentation basins to work in series or
in parallel. Plant processes proposed include raw water control chambers, rapid mix
chambers, flocculation/sedimentation basins, ozone contact chambers, and filters. In this
example, the raw water control chambers are used to split flow between plant process
groups and also as a rapid mix chamber for chemical addition.
STAGE I
FLOC/SED
STAGE I
FILTER
FIGURE 22.4 Stage Ioperational mode diagram.
(FUTURE)
STAGE I
FLOC/SED
- NEW OZONE
CONTACT CHAMBER
STAGE Il
FLOC/SED
STAGE Il OZONE
CONTACT CHAMBER
STAGE
Il
RLTER
STAGE I
FILTER
RLTER BUILDINGS
STAGE I
FLOC/SED
STAGE Il OZONE
CONTACT CHAMBER
STAGE Il
FLOC/SED
NEW OZONE
CONTACT CHAMBER
STAGE
FILTER
STAGE I
FILTER
FILTER BUILDINGS
STAGE I
RAW WATER CONTROL
CHAMBERS/RAPID MIX
(FUTURE)
STAGE I
FLOC/SED
NEW OZONE
CONTACT CHAMBER
STAGE
Il
RLTER
STAGE I
RLTER
STAGE Il
FLOC/SED
STAGE Il OZONE
CONTACT CHAMBER
FILTER BUILDINGS
STAGE I
RAW WATER CONTROL
CHAMBER/RAPID MIX-
STAGE I
FLOC/SED
STAGE III
FLOC/SED
NEW OZONE
CONTACT CHAMBER
RLTER BUILDINGS
- STAGE HI
RAW WATER CONTROL
CHAMBER/RAPID MIX
STAGE
III
FILTER
STAGE
I
RLTER
STAGE Il
FLOC/SED
STAGE Il OZONE
CONTACT'CHAMBER
STAGE I
RLTER
STAGE III
RAW WATER CONTROL
CHAMBER/RAPID MIX
STAGE I
FLOC/SED
STAGE III
FLOC/SED
NEW OZONE
CONTACT CHAMBER
RLTER BUILDINGS
STAGE
III
FILTER
STAGE
I
RLTER
STAGE I
RLTER
STAGE Il
FLOC/SED
STAGE Il OZONE
CONTACT CHAMBER
plant. Under this mode, twice as much flow is routed to each basin and the flow pattern
is longer, since the settled water from the first sedimentation stage must be returned to the
influent of the second sedimentation stage.
Operational mode block diagrams are also a convenient means to illustrate the effect
of side stream flows which may impact the overall plant flow. For example, removal of
sludge from the sedimentation basins is accompanied by a decrease in flow leaving the
basins compared with flow entering the basins. In a similar manner, filter backwash water
removes a certain amount of flow. A plant designed to produce a certain rated capacity
may have to treat more than the rated capacity through certain processes. The impact of
these side stream flows must be evaluated on an individual basis. In many treatment
plants, backwash water treatment facilities are installed to recycle backwash water to the
head of the plant.
22.3.6.2 Plant flow diagrams. After establishing plant operating modes, more detailed
flow diagrams are developed by the designer. The diagrams normally start with possible
valving and gating arrangements and are then expanded with tentative valve, sluice gate,
pipeline, and conduit sizes.
Valving arrangements are designed to enable any of the major operational units (e.g.,
sedimentation basin, ozone contact chamber) to be removed from service. The arrangement may include design of temporary flow stop devices, such as stop logs (sectional barriers which were originally constructed of logs but are now commonly metal plates). The
arrangement should be designed to permit maintenance work on major valves and sluice
gates while minimizing the impact on plant process. Major channel sections should be
designed so they can be removed from service and dewatered while minimizing impacts
on the rest of the plant.
The designer should distinguish between units taken out of service frequently (such
as filters), periodically (such as sedimentation basins), or rarely (such as conduits).
Filter backwashing occurs so frequently that the rated treatment capacity can be met
with one filter out for backwashing. Sedimentation basins may be removed from service
once or twice per year for equipment maintenance. Since the basins outages occur at
widely scattered intervals, it is reasonable to design the units to be removed from service during lower flow periods. Conduits and pipelines are rarely removed from service,
but the hydraulic impacts can be significant. Depending on the conduit location,
removal of a conduit can remove a portion of the plant from service. Effective design
will provide redundant conduits so that a portion of the plant can remain in service during conduit dewatering.
The focus of this section has been on the main plant hydraulics, but the hydraulic
designer must also design for hydraulic subsystems. An important group of these subsystems include dewatering of all basins and conduits. Where plant elevations will allow,
gravity de watering is recommended. In most cases, dewatering pumps are necessary.
These pumps may be located in the unit being dewatered or may be located in a separate
structure connected to the process unit by de watering pipelines.
22.3.6.3 Hydraulic Profile. One of the most important tools in the hydraulic design of a
water treatment plant is the development of a hydraulic profile. The hydraulic profile is a
diagram showing the energy grade line (EGL) at each unit process. For open tanks with
flows at minimal velocities, which is the case in most water treatment plants, the velocity head is negligible and the hydraulic grade line (HGL) or water surface elevation
(WSEL) provide an adequate representation of the EGL. Profiles normally include critical
structural elevations of processes and conduits. The profile may also include ground surface profiles and other site information.
Hydraulic profiles are developed for each of the design flows. In the case of water
treatment plants, the design flows may include rated treatment capacity, hydraulic capacity, average flow, and minimum flow. Hydraulic profiles should also take into consideration unit processes or conduits which may be taken out of service. Hydraulic profiles are
valuable design and operational tools to assist in scheduling routine maintenance activities and for evaluating the impact to the treatment plant capacity during outages of process
units or conduits.
Computations of hydraulic profiles begin at control points where there is a definite
relationship between the plant flow and water surface depth. For gravity flow plants, the
most common forms of control points are weirs and tank water surface elevations (e.g.,
clear well water surface elevations), but other types of control points may be used. From
each control point, head losses associated with local losses, plant piping, and open channel flow are added to the control water surface. Since flows in water treatment plant's are
mostly in the subcritical regime (Froude number < 1), most hydraulic designers will work
upstream from the control point. For pressure plants, control points are typically pressure
regulating or pressure control points, frequently in the service area distribution system.
From these control points and knowledge of the flow velocity, both the EGL and HGL
may be computed back to the treatment facilities.
Hydraulic profiles are valuable design tools to identify overall losses through the plant.
Profiles are also valuable to identify units with excessive losses. Since total head available is normally limited, units with excessive losses should be considered for redesign to
reduce local loss coefficients or to reduce velocities.
Figure 22.3 is an example hydraulic profile for a gravity surface water treatment plant
with conventional treatment processes. The method of computing headlosses is presented
in Section 22.3.7.
22.3.7 Water Treatment Plant Process Hydraulics
In this section calculations required to establish the WSEL through a medium-sized water
treatment plant will be presented. A schematic of the water treatment plant is shown in
Fig. 22.10. Notice that future growth has been considered in the initial design. Three
examples are included which illustrate typical hydraulic calculations. The first example
calculates the WSEL from the sedimentation basin effluent chamber back through the
flocculation/sedimentation basins to the Raw Water Control Chamber. The second follows
the flow from the clear well back through the filters. Filter hydraulics are illustrated in the
third example. All examples are presented in a spreadsheet format which is designed to
facilitate calculating the EGL, HGL, and WSEL at various points through the treatment
process and for multiple flow rates (i.e., minimum, daily average, peak hour, future
conditions).
22.3.7.1 Coagulation. Process criteria and key hydraulic design parameters. The coagulation process, used to reduce particulates and turbidity, is carried out in three steps: mixing (often referred to as rapid or flash mixing), flocculation, and sedimentation. Each of
these steps is briefly discussed below.
Rapid mixing. The mixing process imparts energy to increase contact between
existing solids and added coagulants. Possible mixer types include turbine, propeller,
pneumatic, and hydraulic. Headless that occurs in mixing chambers depends on the chosen mixing device. Most mechanical mixers do not create significant head losses. The
headloss coefficient (K) associated with a specific mixer can be obtained from the manufacturer. Pneumatic mixing, which is not common, has associated losses similar to those
RAW WATER
CONTROL
CHAMBER
RAPID MIX
CHAMBER
FLOCCULATION/SEDIMENTATION
BASINS
BASIN BASIN BASIN BASIN
1
4
3
2
FILTER BOX
LEGEND
INITIAL FACILITIES
FUTURE FACILITIES
CLEARWELL
EFFLUENT
PUMPING STATION
FIGURE 22.10 Schematic of a water treatment plant.
for aeration (see table in Section 22.3.2.2, above). Hydraulic mixing takes place using
weirs, swirl chambers, throttled valves, Parshall flumes, or other devices to induce turbulence. Head loss coefficients for these devices can be obtained from the manufacturer.
Important considerations during the initial design of a mixing chamber include:
Velocity gradient. This is mixerspecific information and can be obtained from the
manufacturer. The system should be designed to provide a velocity gradient that is
optimal for the coagulation process taking place.
Dead spots and short circuiting. An ideal mixing system will have minimal dead spots
and short circuiting. These can be avoided with proper sizing and placement of
mixers.
Flocculation. Coagulated particles form larger particles (floes) during the gentle mixing of flocculation, where the flow travels slowly through a series of flocculator paddles,
baffles, or conduits. Inlets and weirs are designed to provide low turbulence for protection
of the floes. The energy provided to the system by the flocculators (manufacturer-specific) or baffling is decreased as the flow approaches the sedimentation basins.
Sedimentation. Gravity sedimentation removes coagulated solids prior to filtration.
There are four zones in a clarifier as shown in Fig. 22.11 and listed below:
Inlet zonewhere upstream flow conditions transition smoothly to uniform flow settling conditions
Sedimentation zonewhere sedimentation takes place
Sludge zonewhere solids collect and are removed
Outlet zonewhere settling conditions smoothly transition to downstream flow
conditions
OUTLET ZONE
INLET ZONE
SEDIMENTATION ZONE
SLUDGE ZONE
Initial Operation
Design Operation
Min Day Avg Day Avg Day Max Hour
2.19
109.73
3.06
109.73
3.28
109.74
4.38
109.74
1.44
3.13
16.05
0.09
1.41
2.01
3.13
16.06
0.13
1.41
2.15
3.13
16.07
0.13
1.41
2.87
3.14
16.10
0.18
1.41
0.00
109.73
0.00
109.73
0.00
109.74
0.00
109.74
0.68
3.13
16.05
0.04
1.41
0.96
3.13
16.06
0.06
1.41
1.03
3.13
16.07
0.06
1.41
1.37
3.14
16.10
0.08
1.41
0.00
109.73
0.00
109.73
0.00
109.74
0.00
109.74
0.27
3.13
16.05
0.38
3.13
16.06
0.41
3.13
16.07
0.55
3.14
16.10
(1) The Energy Grade Line (EGL) is equal to the Hydraulic Grade Line (HGL) + the Velocity Head (W2g). At
points in the system where the velocities are very low (i.e., V => O) the velocity head approaches O (i.e., W2g =>
O). Consequently, EGL => HGL => WSEL. Where: WSEL = Water Surface Elevation.
TABLE 22.5
(Continued)
Parameter
Initial Operation
Design Operation
Min Day Avg Day Avg Day Max Hour
0.02
1.41
0.03
1.41
0.03
1.41
0.00
109.73
0.00
109.74
0.00
109.74
0.10
0.23
0.41
0.01
109.74
0.10
0.23
0.44
0.01
109.74
0.14
0.23
0.59
0.01
109.76
23.16
0.77
105.97
3.77
87.36
23.16
0.82
105.97
3.77
87.41
23.16
1.09
105.97
3.79
87.68
9.71
0.01
9.71
0.01
9.74
0.01
0.08
0.08
0.11
0.00
109.74
0.00
109.74
0.00
109.76
0.01
0.04
109.78
0.03
0.09
109.83
0.05
0.15
109.91
0.13
0.14
0.92
0.05
109.83
0.14
0.14
0.98
0.06
109.89
0.18
0.14
1.31
0.11
110.02
0.13
0.72
0.65
0.19
0.28
0.14
0.77
0.71
0.19
0.29
0.18
0.90
0.82
0.22
0.30
0.00
109.83
0.00
109.89
0.00
110.02
0.26
0.27
0.36
TABLE 22.5
(Continued)
Parameter
Initial Operation
Min Day Avg Day
Design Operation
Avg Day Max Hour
0.68
0.62
0.29
0.27
0.72
0.65
0.39
0.28
0.77
0.71
0.39
0.29
0.90
0.82
0.44
0.30
0.00
109.80
0.00
109.84
0.00
109.89
0.00
110.02
0.27
97.34
89.01
0.00
0.00
109.80
0.38
97.38
89.04
0.00
0.00
109.84
0.41
97.44
89.09
0.00
0.00
109.89
0.55
97.56
89.21
0.01
0.00
110.02
0.55
0.68
1.04
0.52
0.00
0.36
0.77
0.72
1.09
0.70
0.00
0.37
0.82
0.78
1.18
0.69
0.00
0.38
1.09
0.90
1.37
0.80
0.00
0.41
0.00
109.81
0.00
109.84
0.00
109.90
0.00
110.03
0.55
0.69
1.05
0.52
0.01
109.82
0.77
0.72
1.10
0.69
0.02
109.87
0.82
0.78
1.19
0.69
0.02
109.92
1.09
0.91
1.38
0.79
0.03
110.06
0.55
3.34
0.16
0.00
109.82
0.77
3.34
0.23
0.01
109.87
0.82
3.34
0.25
0.01
109.93
1.09
3.34
0.33
0.02
110.07
1.09
2.79
0.39
0.40
1.53
2.79
0.55
0.40
1.64
2.79
0.59
0.40
2.19
2.79
0.78
0.40
0.00
0.01
0.01
0.02
0.01
109.83
0.01
109.89
0.01
109.95
0.02
110.11
TABLE 22.5
(Continued)
Parameter
Initial Operation
Design Operation
Min Day Avg Day Avg Day Max Hour
109.82
109.88
109.94
110.08
109.83
109.89
109.95
110.10
2.19
2.79
0.78
0.40
3.06
2.79
1.10
0.40
3.28
2.79
1.18
0.40
4.38
2.79
1.57
0.40
0.04
109.87
0.08
109.97
0.10
110.04
0.16
110.26
2.19
2.79
2.81
2.80
0.78
0.40
0.42
0.41
3.06
2.79
2.81
2.80
1.09
0.40
0.42
0.41
3.28
2.79
2.81
2.80
1.17
0.40
0.42
0.41
4.38
2.79
2.81
2.80
1.56
0.40
0.42
0.41
0.00
109.88
0.01
109.98
0.01
110.05
0.01
110.27
2.19
3.06
3.28
4.38
2.81
0.78
2.21
0.99
0.02
109.90
2.81
1.09
2.21
1.39
0.04
110.01
2.81
1.17
2.21
1.49
0.04
110.09
2.81
1.56
2.21
1.98
0.8
110.35
0.42
0.42
0.42
0.42
0.06
0.12
0.14
0.25
0.07
110.03
0.13
110.27
0.15
110.39
0.26
110.86
Initial Operation
Design Operation
MinDay Avg Day Avg Day Max Hour
109.73
109.73
109.73
109.73
0.30
2.74
0.54
2.74
0.66
2.74
1.13
2.74
0.60
1.84
0.66
2.18
0.90
3.14
0.89
3.07
0.95
3.12
0.97
3.27
1.35
4.21
1.37
4.42
0.66
110.39
0.89
110.62
0.97
110.70
1.37
111.10
2.19
1.88
1.16
0.10
3.06
1.88
1.63
0.20
3.28
1.88
1.74
0.23
4.38
1.88
2.33
0.41
110.49
110.82
110.93
0.54
0.67
0.70
111.51
0.85
hydraulic reaches analyzed in the example. The circled numbers indicate points at which
the WSEL is calculated. Hydraulic calculations start downstream of the sedimentation
basins (Fig. 22.12) and proceed upstream through the mixing chamber (Fig. 22.13) and
the Raw Water Control Chamber (Fig. 22.14). Mechanical mixers and mechanical flocculators are used. Conduit losses between the rapid mix chambers and the Raw Water
Control Chamber are also calculated in the example. Three different flow rates (i.e., minimum day, average day, and, maximum hour) are used in the calculations. This is a range
of design flow conditions that a design engineer would typically take into consideration.
The longest path through the flocculation and sedimentation processes, through Basin
No. 4, is followed (Points 1 through 15). Although not shown, losses along the shortest
path have also been calculated. As would be expected, the calculated head loss is smaller
for the shorter path. The actual losses are equal for each path. The flows through each path
naturally adjust to equalize losses. The flow through the longest path is slightly smaller
than the flow through the shortest path. In the example, the WSEL at Point 15 is adjusted
to reflect the average losses through the basins. The WSEL calculations upstream of Point
FLOW
INFLUENT PORT
(SEE DETAIL BELOW)
(TYP OF 24)
FLOCCULATION
ZONE BAFFLES
SEDIMENTATION
ZONE BAFFLE
TANK 4
TANK 3
TANK 2
TANK 1
EFFLUENT PORT
FLOCCULATION/SEDIMENTATION BASIN
MIXED WATER
CHANNEL
BOT EL 109.12mBAFFLE
DETAIL
INLET PORT
BETWEEN (7) AND (S)
EL = 108.20m-
SECTION A-A
INLET PORT
TANK 4
TANK 3
FLOCCULATION BASIN
INFLUENT DETAIL
MIXED WATER CHANNEL
BOT. EL=109.12m
MIXING
CHAMBER
NO. 2
MIXING
CHAMBER
NO. 1
1.68m DIAM
RAW WATER
CONDUIT
RAW WATER CONTROL CHAMBER
AND CONDUITS
However, during backwash, the high rate of flow expands the filter bed and, over time,
the media are regraded so that the more coarsely graded grains are located at the bottom
and the fines are located at the top. To benefit from the coarse-to-fine grading, an upward
flow pattern can be used, but is very uncommon. More often the filter media are selected
such that the fine media have a higher specific gravity than the coarse media to maintain
the course-to-fine gradation during backwash. The most commonly used filter media are
natural silica sand and crushed anthracite coal; however garnet and ilmenite are used in
mixed media beds. Granular carbon is often used if taste and odor control is desired.
The terminal headloss is determined by a combination of factors including filter breakthrough (when the filter bed loses its adsorptive capacity), available static head, and outlet pressure required. The filter should be designed so that the headloss in any level of the
filter bed does not exceed the static pressure. A negative head can result in air binding in
the filter which will, in turn, further increase headloss.
Filter influent piping is sized to limit velocities to about 0.6 m/s (2 ft/s). Washwater and effluent piping flow velocities are kept below 1.8 m/s (6 ft/s) so that
hydraulic transients (waterhammer) and excessive headlosses are minimized and controlled to within tolerable limits.
Hydraulic design example. Table 22.6 illustrates the calculation of the WSEL from the
clear well back upstream to the Sedimentation Basin effluent at the medium-sized water
treatment plant shown in Fig. 22.10. Figures 22.15 and 22.16 show details of the hydraulic
reaches analyzed in the example. Table 22.7 illustrates the filter hydraulic calculation, the
details of which are shown in Figs. 22.17 and 22.18.
The hydraulic profile of the plant (based on hydraulic calculations done in Tables 22.5,
22.6 and 22.7) is shown in Figure 22.3.
TABLE 22.6 Hydraulic Calculations in a Medium-Sized Water Treatment Plant from the Filter
Effluent to the Effluent Clearwell
Parameter
1. Flow(m3/s)
Note: for Points 22 through 28, see Figure 22.15
2. Point 22 to Point 23
Maximum water level in Clearwell (Point 22) (m)
Invert in Clearwell (m)
Flow = 0/2 (mVs)
Stop logs @ A
Flow area (2 openings,
1.52 m wide,
3.66 m deep) (m2)
Velocity = flow/area (m/s)
Loss = 0.5 V2/2g (m)
Baffles
Flow area (3.05 m wide, 3.66 m deep) (m2)
Velocity = flow/area
(m/s)
Loss = 1.0 V2/2g (m)
Stop logs @ B and C
Same as the losses @ A, times 2 (m)
WSEL (see Note 1) at Point 23 (m)
Initial Operation
Min Day Avg Day
Design Operation
Avg Day Max Hour
2.19
3.06
3.28
4.38
105.16
101.50
1.09
105.16
101.50
1.53
105.16
101.50
1.64
105.16
101.50
2.19
11.15
0.20
0.00
11.15
0.27
0.00
11.15
0.29
0.00
11.15
0.39
0.00
11.15
0.20
0.00
11.15
0.27
0.00
11.15
0.29
0.00
11.15
0.39
0.01
0.00
105.16
0.00
105.17
0.00
105.17
0.01
105.18
(1) The Energy Grade Line (EGL) is equal to the Hydraulic Grade Line (HGL) + the Velocity Head (W2g). At
points in the system where the velocities are very low (i.e., V => O) the velocity head approaches O (i.e., W2g =>
O). Consequently, EGL => HGL => WSEL. Where: WSEL = Water Surface Elevation.
TABLE 22.6
(Continued)
Parameter
3. Point 23 to Point 24
Flow = QI2 (mVs)
1.68 (m) diameter pipe
Flow area = d2/4 X TT (m2)
Velocity = flow/area (m/s)2
Exit loss @ clearwell = V /2g (m)
Loss @ 2 - 90bends = (0.25 V2/2g) X 22 (m)
Entrance loss @ Filter Building
= 0.5 V /2g (m)
Pipe loss
= (3.022
X V1-85 X L)/
L8S
1 165
(C X D - ) where C = 120 and
L = 57.91 m (m)
WSEL at Point 24 (m)
4. Point 24 to Point 25
Flow = Q/4 (mVs)
Flow area = 1.52m X 1.52 (m2)
Velocity = Ql'A (m/s)
Loss as flows merge = 1.0 2/3
V2/2g2 (m)
Conduit loss = [(V X n)/(/? )] X L (m)
where n = 0.013, L = 16.76 m and R = AIP
(P = 6.1Om)
WSEL at Point 25 (m)
5. Point 25 to Point 26
Sluice Gate No. 1 flow area = 1.22 m X 0.91 m (m2)
Velocity = QIA
(m/s)
Loss = 0.5 V2/2g (m)
WSEL at Point 26 (m)
6. Point 26 to Point 27
Sluice Gate No. 2 Loss = 0.8 V2/2g (m)
WSEL at Point 27 (m)
7. Point 27 to Point 28
Port to Filter Clearwell: Calculate losses through port
as if were a weir when depth of flow is below top
of port. Port dimmensions = 2.74 m wide
by 0.813 m deep. Flow = Q/4 (m's)
Weir (bottom of port) elevation (m)
Depth of flow over weir =
(WSEL @ 27 - weir elevation) (m)
Flow over submergedweir = q 1.71 X hm
X [1 - (d/hf2]03*5 X L
Note: Rather than solve for h, find an h, by trial
and error, that gives a q equal to the flow for the
given flow scenario
assume h (m) =
then q (mVs) =
assume h (m) =
then?(m3/s)=
Note: These q's equal the flows for the given
scenarios
h(m)
Initial Operation
Design Operation
Min Day Avg Day Avg Day Max Hour
1.09
1.53
1.64
2.19
2.21
0.50
0.01
0.01
0.01
2.21
0.69
0.02
0.01
0.01
2.21
0.74
0.03
0.01
0.01
2.21
0.99
0.05
0.03
0.03
0.00
105.19
0.00
105.22
0.00
105.23
0.00
105.28
0.55
2.32
0.24
0.00
0.77
2.32
0.33
0.01
0.82
2.32
0.35
0.01
1.09
2.32
0.47
0.01
0.00
0.00
0.00
0.00
1.11
0.49
0.01
105.20
1.11
0.69
0.01
105.24
1.11
0.74
0.01
105.24
1.11
0.98
0.02
105.32
0.01
105.21
0.02
105.25
0.02
105.27
0.04
105.36
0.55
104.85
0.77
104.85
0.82
104.85
1.09
104.85
0.36
0.40
0.42
0.51
0.40
0.59
0.39
0.52
0.45
0.69
0.46
0.76
0.50
0.95
0.48
0.82
0.60
1.23
0.58
1.09
0.39
0.46
0.48
0.58
TABLE 22.6
(Continued)
Parameter
Initial Operation
Design Operation
Min Day Avg Day Avg Day Max Hour
105.24
105.31
105.33
105.43
109.73
109.73
109.73
109.73
0.27
0.38
0.41
0.55
0.18
0.00
0.26
0.00
0.28
0.00
0.37
0.01
0.23
0.00
0.00
109.73
0.33
0.00
0.00
109.73
0.35
0.00
0.00
109.73
0.47
0.00
0.00
109.74
2.44
4.46
0.06
0.67
2.44
4.47
0.09
0.67
2.44
4.47
0.09
0.67
2.45
4.48
0.12
0.67
0.00
109.73
0.00
109.73
0.00
109.73
0.00
109.74
0.55
2.44
4.46
0.12
0.67
0.77
2.44
4.47
0.17
0.67
0.82
2.44
4.47
0.18
0.67
1.09
2.45
4.48
0.24
0.67
0.00
109.73
0.00
109.73
0.00
109.73
0.00
109.74
0.82
2.44
4.46
0.18
0.67
1.15
2.44
4.47
0.26
0.67
1.23
2.44
4.47
0.28
0.67
1.64
2.45
4.48
0.37
0.67
0.00
109.73
0.00
109.73
0.00
109.73
0.00
109.74
1.09
2.44
4.46
0.24
0.67
1.53
2.44
4.47
0.34
0.67
1.64
2.45
4.47
0.37
0.67
2.19
2.45
4.48
0.49
0.67
0.00
109.73
0.00
109.73
0.00
109.74
0.00
109.74
FILTER
EFFLUENT
CONDUIT
-CONTROL WEIR
SLUICE GATE *2
FILTER
BUILDING
SLUICE GATE #1
JO FINISHED WATER
"PUMPING STATION
CLEARWELL TO FILTER EFFLUENT
FIGURE 22.15 Clearwell to filter effluent
SEDIMENTATION BASIN
EFFLUENT CONDUIT
OPENING(TYP)
1.22m BUTTERFLY
VALVE(TYP)
FILTER
INFLUENT
Initial Operation
Min Day Avg Day
Design Operation
Avg Day Max Hour
NORTH HALF
FILTER
SOUTH HALF
FILTER
W.S. EL 109.72mPLAN AT 100.58m
WEIR
CLEAR
WELL
SECTION A-A
SECTION B-B
Head, meters
22.3.8
Membrane Technology
Membranes are synthetic filtering media manufactured from a variety of materials including polypropylene, polyamide, polysulfone, and cellulose acetate. The membrane material can be arranged in various configurations, including the following:
Spiral wound
Hollow fiber
Tubular
Plate frame
Examples of these configurations are presented in Fig. 22.19. In water and wastewater
treatment applications, the most common configurations are spiral wound and hollow fiber.
In general, there are four classes of membranes: microfllters (MF), ultrafilters (UF),
nanofilters (NF), and hyperfilters. Treatment through hyperfilters is referred to as hyperfiltration, or reverse osmosis (RO).
The hydraulics associated with membranes are membrane-specific and can be obtained
from the manufacturer. This section presents general considerations pertinent to flow
through membranes.
As with natural particle media filters, clean membranes have a specific headless and,
over time, as the membranes become covered with a cake buildup, the effectiveness of the
membrane decreases and headloss increases. Fouling (excessive buildup) may damage the
membrane.
The need for pretreatment ahead of membranes is determined by the raw water quality and the membrane type. In general, microfilters and ultrafilters do not require pretreatment for treating surface or groundwater. Nanofilters and reverse osmosis membranes may require pretreatment depending on the type of fouling. Membrane fouling
can result from particulate blocking, chemical scaling, and biological growth within the
membranes.
An estimate of particulate blocking can be made using indices such as the Silt Density
Index (SDI) and the Modified Fouling Index (MFI). These fouling indices are determined
from simple bench membrane tests using 0.45 micron Millipore filters and monitoring
flow through the filter at a given pressure, usually 30 psig. Approximate values of suitable SDIs for nanofiltration are 0-3 units, and for reverse osmosis, 0-2 units.
Corresponding values of MFI are, for nanofiltration O to 10 s/L2, and for RO, 0-2 s/L2.
Scaling control is essential in RO and nanofilter membrane filtration, especially when
the filtration provides water softening. Controlling precipitation or scaling within the
membrane element requires identification of limiting salt, acid addition for prevention of
calcium carbonate precipitation within the membrane, and/or the addition of an
antiscalant. The amount of antiscalant or acid addition is determined by the limiting salt.
A diffusion controlled membrane process will naturally concentrate salts on the feed side
of the membrane. As water is passed through the membrane, this concentration process
will continue until a salt precipitates and scaling occurs. Scaling will reduce membrane
productivity and, consequently, recovery is limited by the allowable recovery just before
the limiting salt precipitates. The limiting salt can be determined from the solubility products of potential limiting salts and the actual feed stream water quality. Ionic strength must
also be considered in these calculations as the natural concentration of the feed stream
during the membrane process increases the ionic strength, allowable solubility and recovery. Calcium carbonate scaling is commonly controlled by sulfuric acid addition, although
sulfate salts, such as barium sulfate and strontium sulfate, are often the limiting salt.
Commercially available antiscalants can be used to control scaling by complexing the
FIGURE 22.19 Membrane configurations, (a) Spiral wound, (b) hollow fiber, (c) tubular, (d) plate
and frame.
metal ions in the feed stream and preventing precipitation. Equilibrium constants for these
antiscalants are not available which prohibits direct calculation. However, some manufacturers provide computer programs for estimating the required antiscalant dose for a
given recovery, water quality, and membrane.
Biological fouling is controlled with some membranes such, as cellulose acetate, by
maintaining a free chlorine residual of not more than 1 mg/L. Other membranes, such as
the thin-film composites, are not chlorine tolerant and must rely on upstream disinfection
by, for example, ultraviolet disinfection or chlorination-dechlorination. The extent of fouling for a specific application and its influence in the design of nanofiltration and RO
membrane systems is best determined by pilot studies.
It has been suggested that some buildup on the membrane may be beneficial to treatment by providing an additional filtering layer. At facilities operated by the Metropolitan
Water District of Southern California (MWD), removal rates of 1.7-2.9 logs were
observed for seeded virus MS2 bacteriophage through microfliters that had a pore size an
order of magnitude larger than the nominal size of MS2 (1).
The microfiltration system used by MWD utilizes an air backwash procedure whereby compressed air at 90-100 psig is introduced into the filtrate side of the hollow fiber
membranes. Accumulated particulates dislodged by the compressed air are swept away
by raw water introduced to the feed side of the membranes. The backwash sequence is
carried out automatically at preset time intervals. MWD found the best interval to be
every 18 minutes. The total volume of backwash represents approximately 5-7 percent
of influent flow.
The difference between influent and effluent pressure across the membrane is termed
the transmembrane pressure (TMP). Despite the frequent air and water backwashes, the
TMP gradually increases over time. Generally, when the TMP reaches approximately
15 psig, chemical cleaning of the membranes is carried out. If the TMP is allowed
to increase beyond 15 psig, particulates can become deeply lodged within the lattice structure of the membranes and will not be removed, even by chemical cleaning. Chemical
cleaning typically lasts 2-3 hours and involves circulating a solution of sodium hydroxide and a surfactant through the membranes, and soaking them in the solution.
The membranes at the MWD microfilter plants have a surface loading rate of 40-67
gal/day ft2. The lower flux rate of 40 ft2 has the advantage that the rate of increase of
TMP is reduced and the interval between chemical cleanings is increased. A possible
explanation for this is that particulates are not forced as deeply into the lattice structure of
the membranes, thereby allowing the air-water backwash to clean the membranes more
effectively. By reducing the flux rate from 67-40 gal/day ft2, the interval between chemical cleanings was increased from 2 to 3 weeks to almost 20 weeks. However, MWD has
instituted a maximum run time of 3 months between chemical cleanings to ensure the
long-term integrity of the membranes.
Nanofiltration is widely used for softening groundwaters in Florida. A typical nanofiltration plant would include antiscalant for scale control added to the raw water. Cartridge
filters, usually rated at 5 microns, remove particles that may foul the membrane system.
Feed water pumps boost the pretreated water pressure to about 90-130 pounds per square
inch (psi) before entering the membrane system. The membranes typically are spiral
wound nanofiltration membranes generally with molecular weight cutoff values in the
200-500 dalton range.
22.4
WASTEWATER TREATMENT
Many factors and considerations influence the hydraulic design of a wastewater treatment
plant. This section describes typical phases of wastewater treatment planning required for
design of new plants or additions to existing plants, and then presents typical unit process
hydraulic computations.
22.4.1 Wastewater Treatment Planning
Hydraulic decision making for a new wastewater treatment plant or expansion of an existing plant involves several planning phases. Typical planning phases are presented below
in their common order of consideration.
22 A.I.I Service area and flows. More than 15,000 municipal wastewater treatment plants
are in operation in the United States today. The plants are designed to treat a total of about
Next Page
CHAPTER 23
HYDRAULIC DESIGN
FOR GROUNDWATER
CONTAMINATION
Paul C. Johnson
Department of Civil and Environmental Engineering
Arizona State University
Tempe, Arizona
23.1
INTRODUCTION
This chapter focuses on in situ treatment systems designed to eliminate or minimize the
impacts of hazardous chemicals on groundwater quality. This chapter builds on the principles introduced in Chapter 4. First, typical soil and groundwater contamination scenarios of interest are introduced. Then a discussion of general remediation design principles
and strategies follows. Finally, the design, monitoring, and refinement of selected in situ
remediation technologies is presented. These include both conventional and developing
technologies, and include containment, reaction barrier, active remediation, and passive
remediation systems.
23.1.1 Unique Features of In Situ Treatment Technology Design
The design of in situ treatment systems for groundwater contamination is uniquely different from the design of above-ground treatment and hydraulic conveyance systems. The
subsurface is an environment that is difficult to describe precisely; it is naturally heterogeneous and, due to practical constraints, characterization data are generally limited.
Thus, treatment systems must be designed under conditions of great uncertainty-usually
the amount of chemical spilled is unknown, the timing of the release is unknown, and the
system into which the spill occurred is poorly characterized. For this reason, the design of
in situ treatment systems relies heavily on
conceptual models,
screening-level calculations,
empiricism, heuristics, and experience, and
monitoring and refinement of the design.
A conceptual model is a realization of how the subsurface might look and how contaminants might move, based on the available data. Often, there are several plausible conceptual models, and the treatment system designer must allow for a range of possibilities
in the design. It is also recognized that the conceptual model is continuously refined as
new data become available.
Screening level calculations are used to estimate treatment performance limits. They
are generally biased toward estimating best-case performance, and are useful for making
rough estimates of treatment times and treatment limits. They are primarily used for feasibility screening and creating the initial treatment system design. More sophisticated
models are available and can be used; however, given the typical level of site characterization data and gross approximations that must be made, their use is rarely warranted,
except at those sites where treatment costs are projected to be very high.
Empiricism and experience play a key role in the design as well. In many ways, each
site is unique; yet, the collective experience from many sites is invaluable. This design
field is still relatively young and most experiences are not well-documented. As is often
the case, successful applications are more likely to be reported than failures, although the
knowledge from both would be equally valuable. In practice, there are many "rules of
thumb" that designers use and some are well founded in theory, while others are indefensible (yet they still persist in practice). In any case, it is useful to be aware of these rulesof-thumb even if their bases are not well documented in the literature.
Finally, given the inherent large initial uncertainty, it is important to recognize that the
design phase continues well past the construction, installation, and operation of the initial
system. Following sound design practices does not guarantee success; it does, however,
provide a higher probability of success at most sites. Appropriate monitoring is essential
to verify assumptions built into the conceptual model and initial design, and system
design refinement should be anticipated in the initial design. For this reason, it is important to build robust systems that are both flexible and expandable.
23.1.2
Overview
It has been estimated that costs associated with the remediation of known hazardous
waste sites in the United States will exceed $700 billion (all costs are quoted in U.S. dollars) and that the current average rate of expenditure is about $20 billion (Bredehoeft,
1994). Despite the fact that society has been tackling this difficult problem since the
1970s, the environmental profession still struggles to find cost-effective remediation
processes. Remediation costs for service station sites now range from $100,000 to
$500,000 per site, while costs for the larger landfill and manufacturing sites often exceed
$10,000,000 per site. Costs associated with sites where groundwater is contaminated
generally exceed those where only soil is impacted by at least an order of magnitude.
Clearly, these are not trivial expenses; thus, appropriate technology selection, design, and
optimization are important activities.
23.1.3 Groundwater Contamination Scenarios-Point Versus
Area Sources
Groundwater contamination is prevalent in most industrial societies. Leaking fuel tanks,
petroleum storage and refining, solvent degreasing activities, chemical manufacturing
and storage, landfills, and mining activities have all contributed to groundwater contamination problems. Groundwater contamination is also prevalent in most modern agricultural areas, where the combination of irrigation and fertilizer, pesticide, and herbicide use
have contributed to many large-scale groundwater contamination problems. Fig. 23.1
[US Enviromental Protection Agency (USEPA), 1990] summarizes the more common
sources of groundwater contamination and the relative frequency at which they are
expected to cause groundwater impacts. Septic tanks and leaking underground fuel storage tanks are considered to be the largest single contributor to the number of groundwa-
ter contamination sources. Agricultural activities are likely the largest contributor to
groundwater contamination in terms of the actual extent of impact.
For the purposes of this chapter, we first segregate the universe of groundwater contamination scenarios into two main categories:
Point sources,
Area, or distributed sources
Underground storage tanks, above-ground storage tanks, landfills, pipeline releases,
chemical manufacturing and petroleum refining locations, wood treating facilities, and so
forth are all considered to be point sources. Agricultural activities are often considered to
be area, or distributed sources.
By virtue of their large areal extent, the environmental engineering profession has yet
to develop effective solutions for the remediation of area source contamination problems.
These often involve widespread region- or basin-scale nitrate or pesticide contamination.
Currently, the only practicable solution for these problems appears to be above-ground
water blending and direct well-head treatment. In other words, contaminated groundwater is either diluted with water of higher quality as it is pumped from the aquifer, or else
it is treated as necessary above-ground as it is removed from the aquifer. As stated, area
source problems are rarely addressed with in situ treatment systems, and therefore, these
will not be considered further in this chapter.
23.1.4 Groundwater Contamination Scenarios-Segregation
by Contaminant Type
Point source groundwater contamination problems are divided here into three main categories according to the contaminant properties. The three categories are
Light nonaqueous phase liquids (LNAPLs)
Dense nonaqueous phase liquids (DNAPLs)
Inorganics and other dissolved constituents
LNAPLs, in their pure liquid form, are less dense than water (p < 1 g/cm3). Most
LNAPL sites involve the release of petroleum products or crude oil (e.g., service stations,
refineries, pipeline spills). DNAPLs, in their pure liquid form, are more dense than water
(p > 1 g/cm3). DNAPL contamination is generally found near sites where dry cleaning and
aviation, automobile, and electronic circuit board degreasing operations took place.
Historically these used chlorinated solvents, such as trichloroethylene (TCE) and perchloroethylene (PCE). Metals and salts fall into the category of inorganics and other dissolved constituents. Mining operations, electroplating operations, leaking wastewater
treatment facilities, and landfills are examples of sources that historically have added contaminants to groundwater in dissolved form.
23.1.5 Groundwater Contamination Scenarios-Subsurface
Contaminant Distributions
Figures 23.2a, 23.2&, and 23.2c schematically illustrates release scenarios involving
LNAPLs, DNAPLs, and inorganics. While very simplistic, these figures highlight key
characteristics that impact subsurface characterization activities, remediation design,
and monitoring.
Underground
Storage Tank
free-product (mobile)
Jiy&ocwbon ItiyM
dry cleaners
dissolved
hydrocarbon
plume
dissolved
hydrocarbon
plume
FIGURE 23.2 Example contaminant release scenarios: (a) LNAPL release.(b)
DNAPL release.
land fill
leachate containing
dissolved inorganic
chemicals
vadose zone
dissolved
inorganic
plume
FIGURE 23.2 (Continued) (c) Example contaminant release scenarios: (c) dissolved
contaminant release.
As their density is less than that of water, one can expect to find LNAPLs in the general vicinity of the water table. Provided that a given spill or leak is of sufficient size and
that there are not any impermeable vertical flow barriers, LNAPLs travel downward and
pool on top of the water table. Then over time, water table fluctuations cause a vertical
smearing so that immiscible-phase LNAPL can be found trapped by capillary forces in the
soil pores above and below the current ground water table level. LNAPL solubilities are
generally so low that dissolution (and other natural loss processes) occurs very slowly and
pools of LNAPL will persist for decades.
Dissolution of LNAPLs into groundwater leads to the formation of dissolved contaminant plumes. Aquifer characteristics and chemical properties control the growth and
extent of the dissolved contaminant plume. Fortunately, most LNAPL sites involve the
release of petroleum hydrocarbons and some of these are aerobically degradable by
indigenous microorganisms. The result is that natural chemical degradation often limits
the growth of the dissolved LNAPL plume. For example, two survey studies have shown
that most benzene, toluene, ethylbenzene, and xylenes (BTEX) dissolved plumes at
leaking underground storage tank (LUST) sites do not extend more than 300 ft (100 m)
from the downgradient edge of the source zone (Mace et al., 1997; Rice et al., 1995). It is
important to note, however, that not all fuel-related contaminants degrade at an appreciable rate. For example, some oxygenates (e.g., MTBE), are highly soluble and resistant to
degradation. Consequently, these compounds generally migrate much further distances
than the more degradable monoaromatic chemicals (e.g., benzene, toluene, xylenes). The
behavior of these compounds is currently the focus of intense research.
DNAPL sites are generally more complex and challenging. Provided that a given
spill or leak is of sufficient size and that there are not any impermeable vertical flow
barriers, DNAPLs travel downward to the water table. Initially, they too will pool on
top of the water table; but, provided that the spill or leak is large enough, the weight
of the DNAPL pool will exceed the entry pressure of the medium and at that point the
DNAPL will penetrate through the groundwater table. It will continue to migrate
downward until a vertical flow barrier is encountered, or until all of the DNAPL liquid is bound by capillary forces. DNAPL vertical migration in aquifers appears to be
very sensitive to small changes in soil structure, and as a DNAPL migrates downward,
it may also spread horizontally along thin, more permeable, soil layers. DNAPL solubilities are also generally low so that dissolution (and other natural loss processes)
occurs very slowly and DNAPL found below the water table will persist for decades.
The assessment and remediation of DNAPL sites is generally much more challenging
than that of LNAPL sites.
Most DNAPLs of interest are chlorinated solvents (e.g., TCE, PCE, and so on) and
these degrade slowly, if at all, under natural conditions. This results in dissolved plumes
that are much longer than the dissolved LNAPL plumes that might be found in similar settings, and plumes > 1500 ft (500 m) in length are not uncommon; in fact, plumes several
miles (kilometers) in length can be found in regions of concentrated historical aviation or
electronics manufacturing.
In the case of inorganics, such as those originating from the landfill in Fig. 23.2, the
contaminants are already dissolved in solution when they reach the water table. Thus,
these chemicals follow groundwater flow once they reach the aquifer. Dissolved inorganic chemical migration is governed by complex geochemical interactions; however,
in some cases of interest (e.g., nitrate contamination), the contaminants typically are not
transformed or retarded, and may also form long dissolved plumes that persist for
extended periods of time. These solutions generally behave as neutrally buoyant liquids
(p ~ 1 g/cm3); however, if the total concentration of inorganics exceeds about 10,000
mg/L H2O, then density gradients will cause these solutions to migrate vertically downward within aquifers.
23.2
REMEDIATIONGOALS
When subsurface soil and/or groundwater contamination have been identified, decisions
must be made as to whether or not treatment is necessary, and if so, to what levels. Often
regulatory guidelines prescribe the actions that must be taken. In most cases, contaminant
concentrations in soil and groundwater are compared with levels that have been deemed
acceptable. These levels are sometimes referred to as target levels. If exceedences are
noted, then treatment or further study is required; otherwise, no action other than monitoring is typically required.
While the development and enforcement of target levels is not the focus of this chapter, it is important to recognize that target levels play a significant role in the selection and
design of in situ treatment systems. There is no universal set of target levels, and they vary
from state to state and country to country. Thus, it is worthwhile to briefly discuss the
range of situations that might be encountered in practice.
Target levels may be prescribed in terms of acceptable groundwater concentrations,
or in terms of soil concentrations that are expected to be sufficiently protective of groundwater quality. In the United States, target levels are often linked to
maximum contaminant levels (MCLs) or maximum contaminant level goals (MCLGs)
risk-based considerations, or
resource protection goals
In some areas aquifers are considered to be valued resources, and therefore must be protected from any impacts. In such areas, no level of impact is acceptable. These anti degradation policies require that all contamination must be remediated and the resource must
be restored to pristine, or background, conditions. Of the three approaches, this is typically the most stringent as the cleanup goal is either zero, or the background concentration. It is also typically the most difficult goal to achieve.
23.2.4 Application of the Target LevelsRemediation,
Points of Compliance, and Containment
Once target levels are selected, one must choose how and where to apply them. Two obvious choices are
apply target levels everywhere in the soil and groundwater, or
apply the target levels to the boundary of a compliance zone.
While total cleanup of soils and groundwater is always preferred, it is not always practicable. For this reason, a compliance zone is sometimes negotiated. Contaminant concentrations must not exceed the target levels outside of this compliance zone, but may
exceed them within the compliance zone. This approach is similar to that taken when
permitting mixing, or dilution zones for surface water discharges.
The allowance of a compliance zone greatly impacts the range of technologies that can
be selected at a site. In this case complete remediation of the source zone and groundwater is not always necessary and this opens the door for consideration of contaminant con-
TABLE 23.1 Regulatory and Human Health Benchmarks Used for Soil and Groundwater Target
Cleanup Levels
Chemical
Maximum
Contaminant
Level Goal
(MCLG)
WQ
Acenapthene
Acetone
Aldrin
Anthracene
6
Antimony
Arsenic
Barium
2000
Benz(a)anthracene
Benzene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzoic acid
Benzo(a)pyrene
4
Beryllium
Bis (2-chloroethyl) ether
Bis (2-ethylhexyl) phthalate
Bromodichloromethane
Bromoform (tribromomethane)
Butanol
Butyl benzyl phthalate
Cadmium
5
Carbazole
Carbon disulfide
Carbon tetrachloride
Chlordane
p-Chloroaniline
100
Chlorobenzene
Chlorodibromomethane
60
Chloroform
2-Chlorophenol
Chromium
100
Chromium (III)
Chromium (VI)
Chrysene
Cyanide (amenable)
200
DDD
Maximum
Contaminant
Level
(MCL)
(g/L)
2,000
4,000
0.005
10,000
6
50
2,000
0.1
0.005
0.1
1.0
100,000
0.2
4
0.08
6
100*
100*
4,000
7,000
5
4
4,000
5
2
100
100
100*
100*
200
100
40,000
100
10
200
0.4
TABLE 23.1.
(Continued)
Chemical
DDE
DDT
Dibenze(a,h)anthracene
Di-n-butyl phthalate
1 ,2-Dichlorobenzene
1 ,4-Dichlorobenzene
3 ,3-Dichlorbenzidene
1 , 1 -Dichloroethane
1 ,2-Dichloroethane
1 , 1 -Dichloroethylene
cis-1 ,2-Dichloroethylene
trans- 1 ,2-Dichloroethylene
2,4-Dichlorophenol
1 ,2-Dichloropropane
1 ,3-Dichloropropene
Dieldrin
Diethylphthalate
2,4-Dimethylphenol
2,4-Dinitrophenol
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Di-rc-octyl phthalate
Endosulfan
Endrin
Ethylbenzene
Fluoranthene
Fluorene
Heptachlor
Heptachlor epoxide
Hexachlorobenzene
Hexachloro- 1 ,3-butadiene
a-HCH (a-BCH)
P-HCH O-BCH)
HCH (Lindane)
Hexachlorocyclopentadiene
Hexachloroethane
Indeno( 1 ,2,3-cd)pyrene
Maximum
Contaminant
Level Goal
(MCLG)
WU)
Maximum
Contaminant
Level
(MCL)
WV
0.3
0.3
0.01
4,000
600
75
600
75
0.2
4,000
7
70
100
5
7
70
100
100
5
0.5
0.005
30,000
700
40
0.1
0.1
700
200
2
700
2
700
1,000
1,000
0.4
0.2
1
1
0.2
50
1
0.01
0.05
0.2
50
6
0.1
TABLE 23.1
(Continued)
Chemical
Maximum
Contaminant
Level Goal
(MCLG)
WQ
Maximum
Contaminant
Level
(MCL)
(V8/L)
90
Isophorone
2
2
Mercury
40
40
Methoxychlor
50
Methylbromide
Mthylene Chloride
5
2,000
2-Methylphenol (0-cresol)
1,000
Napthalene
100
Nickel
20
Nitrobenzene
20
N-Nitrosodiphenylamine
0.01
N-Nitrosodi-n-propylamine
1
Pentachlorophenol
20
Phenol
1,000
Pyrene
Selenium
50
50
200
Silver
100
100
Styrene
0.4
1 , 1 ,2,2-Tetrachloroethane
5
Tetrachloroethylene
2
Thallium
0.5
Toluene
1,000
1,000
3
Toxaphene
70
70
1 ,2,4-Trichlorobenzene
200
200
1,1,1-Trichloroethane
3
5
1 , 1 ,2-Trichloroethane
Trichloroethylene
O
5
4,000
2,4,5-Trichlorophenol
8
2,4,6-Trichlorophenol
300
Vandium
40,000
Vinyl acetate
2
Vinyl chloride (chloroethene)
10,000
10,000
w-, o, and p-Xylene
Zinc
10,000
proposed MCL=SO ng/L, Drinking Water Regulations and Health Advisories, USEPA (1995a)
tainment systems. In summary, prior to selecting and designing in situ treatment technologies, the user must
define the target treatment levels (both in soil and ground water),
define the points of compliance where these goals are to be applied, and
define the time frame within which compliance must be achieved.
23.3 INSITU TREATMENT TECHNOLOGIES-GENERAL
CLASSIFICATIONS
In situ treatment technologies are generally designed to perform one, or more of the following functions:
Contaminant source zone removal
Aquifer restoration
Prevent, or minimize continued contaminant migration
Rarely does a single technology accomplish all three of these goals, and so combinations and sequences of complementary technologies are often used.
23.3.1 Source Zone Treatment Technologies
Source zone treatment technologies target removal and destruction of the residual contaminants in soil that serve as the source for groundwater contamination. The goal here is
to treat the cause of the groundwater contamination, rather than the symptom. Example
source zone treatment technologies include
Free-product recovery
Excavation and disposal or above-ground treatment
In situ soil venting
Bioventing
In situ air sparging
Enhanced in situ soil venting with soil heating and/or soil fracturing
In situ vitrification
Phytoremediation
Groundwater pump and treat systems
Fig 23.3 depicts simple process schematics of the most common of these source zone
treatment technologies.
23.3.2 Aquifer Restoration Technologies
Aquifer restoration technologies target treatment of dissolved contaminant plumes. These
technologies may be employed before, during, or after source zone treatment. Examples
of aquifer restoration technologies include
blower
water
treatment
oil-water
separator
oil
collection
compressor
air
vapor
treatment
LNAPL
Free
Product
blower
blower
vapor
treatment
source zone
sotttte gone
FIGURE 23.3. Example source zone treatment technology processes schematic: (a) free-product recovery, (b) in situ air sparging, (c)
soil vapor extraction, and (d) bioventing.
Fig. 23.4 displays simple process schematics of the first three of these aquifer restoration technologies. Restoration of groundwater quality and use of these technologies is
only feasible if a complementary source zone treatment technology is also utilized.
23.3.3 Contaminant Migration Prevention
Contaminant migration prevention technologies are designed to minimize future impacts
of contaminants on groundwater. These technologies are often employed at sites where:
a) the source zone location is not known and/or b) there are no practicable source zone
and aquifer restoration options. Examples of these technologies, or strategies, include
Natural attenuation
Groundwater pump and treat systems
In situ reaction walls
In situ air sparging curtains
Infiltration barriers
In situ containment options (grout walls, sheet piling walls, and so on)
blower
discharge or
reinjection
compressor
water
treatment
dissolved
plume
monitoring wells
FIGURE 23.4. Example aquifer restoration process schematics: (a) groundwater pump and treat, (b) in situ air sparging, (c) natural attenuation.
compressor
sheet piling
sheet piling
source zone
source zone
FIGURE 23.5 Example contaminant migration technology processes schematics: (a) physical barriers, (b) in situ air sparging curtain, (c)
reaction barrier.
No Further Action
Remediation Necessary?
Compliance Monitoring
Yes
Treatment Goals Met?
Refine Design
Technology Appropriate?
Operate &
Monitor
Pilot Test
Select Alternate
Technology or
Change Remedial
Goals
Initial Design
Technology Appropriate
FIGURE 23.6 Generalized technology selection and design flowchart.
A historic assessment
A geologic/hydrogeologic assessment
A contaminant distribution assessment
A receptor assessment
A political assessment
The historic assessment reviews the history of activities at the site, including chemical inventory records, plot plans (locations of structures, pipelines, rail roads, storage
areas, maintenance areas, and so on); records of known spills, leaks, and other releases, and any existing site assessment data and reports for this site and any other nearby
sites.
The geologic/hydrogeologic assessment often involves collecting soil cores, installing
groundwater wells, performing aquifer characterization tests, monitoring groundwater
elevations, and using other geotechnical tools (e.g., cone penetrometer) to characterize the
subsurface.
The contaminant distribution assessment involves conducting chemical analyses of
soil, water, groundwater, and soil gas samples, and measuring free-product levels in
monitoring wells.
The receptor assessment involves visiting the site and nearby area and reviewing water
well inventory records for the immediate vicinity. The goal is to identify persons or
resources that have been, or could be impacted by the contamination.
Often the time of the release and the quantity released are unknown, unless a single
catastrophic well-documented event took place (many slow leaks from underground storage tanks go undetected for years). However, if they are known, then this information can
be used to better choose the number and locations of soil and groundwater samples. For
example, for a spill size of volume Vspill (m3 or ft3), the maximum depth dspm (m or ft) penetrated by the spill can be estimated by
4- = A^fV
^spiii ^R VT
dspin =
Vspin =
A spin =
SR
=
(23 1}
'
TABLE 23.2 Ranges of Residual Liquid Hydrocabon Concentrations in the Unsaturated Zone.
Residual Liquid Hydrocarbon Concentrations in Unsaturated Zone Soils (= SR$T)
Gasolines
Soil
Middle Distillates
(gal/ft )
(Um ) (mg/kg)
Fuel Oils
3
(gal/ft )
(Um3)
(mg/kg)
Coarse
0.02 2.5
950
0.04
5
2200
0.07
10
4,800
gravel
Coarse
0.06 7.5 2800
0.1
15
6500
0.22 30 15,000
sand
Fine
sand/
0.15 20.0 7500
0.3
40
17000
0.6
80 39,000
silts
Abreviations: (gal/ft3) = gallons of liquid per ft3 of soil; L/m3] b = L3 of liquid/m3 of soil, (mg/kg) = mg of liq- 3
uid/kg of soil; (mg/kg) concentrations based on psoil = 1.85 g soil/cm soil, pliquid = 0.7, 0.8, and 0.9 g,-liquid/cm
liquid for gasoline, middle distillates, and fuel oils, respectively.
Source: API, 1989.
In the case of single-component spills, the maximum dissolved concentration Cwmax
[mg/LH20] is the pure component solubility S (mg/L H2O), while the maximum vapor concentration Cvmax (mg/Lvapor) is derived from tabulated values of vapor pressure Pv (torr),
the molecular weight Mw (g/mol), and the Ideal Gas Law. In summary, for single-component spills:
cL, (mg/LH20] = S
(Pv/160 torr I aim) X (Mw X 103 mg/g]
CL [mg/Lvapor] =
0.0821 L - atm lmol IK X 293 K
coarse fine
sands' sands silts
finer soils
FIGURE 23.7 Depth of spill penetration vs. spill size and infiltration area.
(23.2)
(23.3)
Table 23.3 contains Cy10x and CJJ0x for a range of chemicals of interest. For chemicalspecific properties of other chemicals, the reader is referred to the USEPA Superfund
Chemical Data Matrix databases (USEPA, 1996b).
For the case of mixtures, the appropriate equations for estimating the maximum dissolved and vapor concentrations of each compound are
<, i (ngfrnj
=
c
max, i (m8'LvaPor)
(23.4)
>
TABLE 23.3 Chemical Properties of Selected Soil and Groundwater Contaminants of Interest
(derived from USEPA SCDM 1996b)
Chemical
Acetone
Aldrin
Anthracene
Atrazine
Benz(a)anthracene
Benzene
Benzo(a)pyrene
Benzo(b)fluoranthene
1,3 -Butadiene
Butanol
Carbon tetrachloride
Chlordane
Chlorobenzene
Chloroform
Chloromethane
Chrysene
m-Cresol
Cyclohexane
DDD
DDE
DDT
1,3 Dichlorobenzene
1 ,2-Dichloroethane
trans 1,2-Dichloroethylene
1 ,2-Dichloropropane
Dieldrin
Molecular
Weight
(g/mole)
58
365
178
216
238
78
252
252
54
74
154
410
113
119
51
228
108
84
320
318
354
147
99
97
113
381
Pure
Component
Solubility
(mg/LH20)
Vapor
Pressure
(torr)
106
0.017
0.043
70
0.009
1800
0.002
0.002
740
106
790
0.056
470
7900
5300
0.006
23,000
55
0.09
0.12
0.025
130
8500
6300
2800
0.2
23
730
6 X 10-6
0.0001
3 X 10-6
0.00003
0.000004
3 X 10-7
1 X 10-7
0.000001
9
405
6 X 10-9
0.00000008
5 X 10-7
0.000007
6200
2100
74,000
300,000
1010
120
5 X 10-5
0.001
12
14
200
1300
4300
12,000
9.5 X 10-5
0.001
0.14
0.83
97
450
7 X 10-7
0.00001
6 X 10-6
0.0001
2 X 10-7
0.000004
2.2
18
79
430
1800
330
52
320
5 X 10-6
0.0001
Maximun
Vapor
Cone.
(m/Lvapor)
Chemical
Molecular
Weight
(g/mole)
Endrin
Ethylbenzene
Ethylene glycol
Formaldehyde
Heptachlor
Hexane
Isobutanol
Lindane
Methanol
Methylethylketone
Napthalene
PCBs
Phenol
Styrene
Tetrachloroethylene
Toluene
Toxaphene
Trichlorbenzene
1 , 1 ,2-Trichlorethane
Trichloroethylene
Vinyl chloride
m-xylene
Gasoline (fresh)*
Gasoline (weathered) *
381
106
62
30
373
86
74
291
32
72
128
356
94
104
166
92
413
181
133
131
63
106
95
111
Pure
Component
Solubility
(mg/LH20)
0.2
170
106
550,000
0.18
12
85,000
7.3
106
220,000
31
0.07
83,000
310
200
530
0.74
49
4400
1500
8800
160
Vapor
Pressure
(Torr)
Maximun
Vapor
Cone.
(m8/Lvapor)
3 X 10-6
9.6
0.092
5200
0.0004
152
10
0.0004
130
95
0.085
7.7 X 10-5
0.28
6.1
19
28
9.8 X 10-7
0.43
23
73
3000
8.5
0.00006
0.056
0.31
8500
0.008
720
40
0.006
230
370
0.60
0.0015
1.4
35
170
141
0.00002
4.3
170
523
10,000
49
1300
220
(23.6)
For example, if a spill of benzene liquid occurred, we would expect to see dissolved
and soil gas vapor benzene concentrations approaching 1800 wg/LH20 and 400 mg/Lvapor,
respectively, near the source zone. If, on the other hand, a gasoline spill occurred, then
we would expect to see dissolved and soil gas vapor benzene concentrations approaching 18 mglLmo and 4 mglLvapor, respectively, for a gasoline containing 1 percent by
weight benzene.
For the site assessment, how much data collection is enough? Clearly, it is desirable to
be able to completely characterize a site, but this option is cost-prohibitive. There are no
universal guidelines that define minimum numbers of soil borings, samples, and so on;
however, some useful questions to consider are
If treatment is necessary, a gross pre-screening of technologies takes place. This prescreening considers the site conditions, documented performance of the available technologies at similar sites, the state of understanding of the technology, economic factors,
and the treatment goals (target levels and target time frame). Table 23.4 contains information that can be used to provide a rough screen of technologies.
At the end of this step the user is generally considering one to three treatment options.
23.4.4 Screening Level Calculations
At this stage a more refined screening of technologies occurs based on basic fundamental
screening calculations. These calculations are generally biased toward predicting optimal
system performance. Typical calculations involve estimating maximum treatment rates.
Examples of these screening level calculations are presented in Sees. 23.5 and 23.7.
Application
Robustness
Typical Settings
Excavation and
above-ground
treatment
Groundwater
pump and treat
Source treatment
Short
Containment
Long
Bioventing
Source treatment
Medium
Soil vapor
extraction
Source treatment
Medium
Physical barriers
Containment
Long
Natural
attenuation
Dissolved plume
management
Long
Petroleum
release sites
Free-product
recovery
systems
Source zone
treatment
Medium
Sites with
mobile and
pumpable freeproduct liquid
In situ air
sparging
Source treatment
Medium
Petroleum fuels
and solvents
trapped below
the water table
in permeable soils
Chemical
migration
/reaction
barriers
Dissolved plume
containment
Long
Chlorinated
solvent spill sites
Shallow
contamination
(<20 ft BGS)
All, except very
low permeability
settings
Petroleum
release sites and
permeable soils
Petroleum fuel
and solvent
release sites and
permeable soils
Mixed waste
sites involving
very hazardous
chemicals, low
permeability
settings
23.4.5
The user then compares the output of the screening level calculations with treatment
expectations. If, even under optimal conditions, the performance predicted by the
screening-level calculations does not meet treatment expectations, then the user has
two options:
eliminate that technology as an option, or
revise treatment expectations (target levels or treatment time) and retain the technology.
23.4.6 Pilot Testing
In situ pilot tests are usually conducted for the most promising treatment technologies.
Often these are designed to test critical conceptual model assumptions, such as achievable
flow rates, maximum effluent concentrations, and so on. Convention and economic constraints dictate that most pilot tests will be less than 1 week in duration.
Key guidelines for planning and conducting pilot tests include
Design the pilot test so that the equipment used is easily integrated into any full-scale
system that might be installed
As it is very unlikely that the short-term performance of a pilot test can be accurately
extrapolated to long-term full-scale system performance, pilot tests should be designed
to look for infeasibility. In other words, the pilot test should be designed to identify
behavior or fatal flaws that suggest that the technology would not work at the site
(e.g., maximum achievable flow is too low, geology prohibits access to the contaminants, and so on).
23.4.7 Decision PointIs the Technology Appropriate?
The user then compares the pilot test results with treatment expectations. If the performance does not meet treatment expectations, then the user has two options:
eliminate that technology as an option or,
revise treatment expectations (target levels or treatment time) and retain the technology
23.4.8 Initial Design
Based on all available data, the site conceptual model is revised, and an initial system
design is created and installed. The design process often relies on a combination of screening level models, heuristics, and sometimes more sophisticated computational design
tools. It is important at this stage to
Design a robust system that can succeed under a range of actual site conditions
Design a system that can easily be modified and expanded
Design the system to be easily monitored
Design Refinement
As stated above, given the inherent large initial uncertainty, the design phase continues
well past the construction, installation, and operation of the initial system. It should be
recognized that following sound design practices does not guarantee success; it does,
however, provide a higher probability of success. For this reason, it is likely that refinements will be made to the original system design based on the performance monitoring
data. These refinements may include changes in operating conditions (e.g., flow rates) or
the addition of new hardware (e.g., additional wells).
23.4.11 Decision Point-Have the Treatment Goals Been Met?
To determine when it is appropriate to terminate the treatment process, the following
questions must be addressed:
Has the treatment system achieved the treatment goals?
If the system is turned off, will treatment goals continue to be met?
If the treatment goals have not yet been met, is there another technology that can more
effectively be used to progress toward the treatment goals?
If the answer to any of these questions is yes, then operation of the current treatment
system should be terminated. If treatment goals have yet to be met, it is time to transition
to the next technology in the treatment train. For example, at petroleum fuel release sites,
a typical sequence would involve free-product liquid recovery soil vapor
extraction/bioventing > natural attenuation.
23.5
SOURCEZONETREATMENT
In this section, general design principles are presented for common source zone treatment
technologies, including free-product recovery, pump and treat, soil vapor extraction,
bio venting and in situ air sparging.
23.5.1 Free-Product Recovery
Most regulatory programs require removal of any mobile and pumpable immiscible freeproduct liquid that can be collected in monitoring wells. In the case of LNAPLs, free
product is found floating on top of groundwater in a monitoring well; in the case of
DNAPLs, free=product liquid is found at the bottom of the well. This of course assumes
that the well is screened across the interval containing the free=product liquid. The following discussion is specific to LNAPL sites, as our understanding of these sites greatly
exceeds our knowledge of free-product recovery at DNAPL sites.
Next Page
CHAPTER 24
ARTIFICIAL RECHARGE
OF GROUNDWATER:
SYSTEMS, DESIGN,
AND MANAGEMENT
Herman Bouwer
USDA-ARS Water Conservation Laboratory
Phoenix, Arizona.
24.1
ABSTRACT
The type of system to be selected for artificial recharge of groundwater and how it should
be designed and managed for optimum performance depend entirely on local conditions
of soil, hydrogeology, topography, water availability (quality, continuous, or interrupted
supply), and climate. Unlike water or wastewater treatment plants, guidelines, standards,
and "blueprints" for artificial recharge systems cannot be given. Rather, the responsible
planner should have extensive knowledge of the various types of recharge systems that
can be used, how they work, the processes, cause and effect relations, what experiences
have been obtained elsewhere, and how they should be adapted to local conditions. Once
the type system has been selected, a protocol of site investigations should be followed for
preliminary design. Key factors in the design and management of successful artificial
recharge systems are site and system selection, maintenance of adequate infiltration rates,
hydraulic continuity between the recharge system and the aquifer (no clay layers in the
vadose zone), and groundwater control for effective water recovery and prevention of
undue groundwater rises in the recharge area. Another important factor in the selection of
the type of recharge system is the pretreatment of the water required before recharge to
minimize physical, chemical, or biological clogging of the infiltrating surface (bottom in
basins and walls in trenches, shafts, and wells). Since clogging layers are easier to control and remove in surface infiltration systems, proper pretreatment is especially important for trenches, shafts, and wells. Since trenches and shafts are relatively inexpensive,
they can be replaced when their useful life has come to an end. On the other hand,
recharge wells, while much more expensive than trenches and shafts, are more amenable
to clogging prevention and control by frequent pumping and periodic redevelopment. All
these aspects have to be factored in, however, when selecting the type system that gives
recharge at minimum cost per unit volume of water put into the aquifer. While presedimentation in settling ponds may be adequate pretreatment for surface infiltration systems,
recharge with trenches, shafts, and wells may require coagulation and sand filtration to
remove suspended solids to acceptable levels. Where sewage effluent is used for groundwater recharge, conventional primary and secondary treatment may be sufficient for surface infiltration systems. However, recharge with effluent through wells may require
advanced treatment processes including nitrogen and phosphorus removal, sand filtration
for removal of suspended solids and parasitic organisms like helminths, Giardia, and
Cryptosporidium, disinfection for removal of viruses and bacteria, and reverse osmosis
(RO) for removal of organic compounds and other chemicals.
If the site investigations indicate no fatal flaws, a small pilot or test project should be
installed to gain a better understanding of how the system behaves and how the full-scale
system should be designed and managed for best performance. Even then, a phased construction should be used for the full-scale system, so that operating experience is obtained
and refinements can be made as the rest of the system is designed and constructed. This
is especially true for "new" areas where artificial recharge has not yet been practiced.
Where local recharge experience already is available, additional systems then are essentially expansions of existing systems and the normal protocol of preinvestigations and
pilot testing can be reduced. The objective of this chapter is to provide information on the
types of recharge systems, how they work, and how they should be designed and managed, so that planners, designers, and managers can make best available decisions.
24.2
INTRODUCTION
Artificial recharge of groundwater is accomplished by applying water to the soil for infiltration and downward movement through the unsaturated or "vadose" zone to the groundwater. Such "surface systems" (Fig. 24.1A) require soils that are sufficiently permeable,
vadose zones that have no clay or other restricting layers, and aquifers that are unconfined. Where permeable surface soils are not available but sufficiently permeable material is found at relatively small depth (about 1 m, for example), artificial recharge can be
achieved with excavated basins that are sufficiently deep to reach the permeable material
(Fig. 24.1B). If the permeable material is too deep for removal of overlying material, but
is within trenchable depth (less than about 7 m, for example) seepage trenches can be used
(Fig. 24.1C). Such trenches are also suitable where soils are highly stratified with alternating layers of fine and coarse materials. Where permeable material is too deep for
trenches, large-diameter wells, pits, or shafts in the vadose zone can be used (Fig. 24.1D).
Such shafts can be drilled with bucket augers to a depth of about 50 m with a diameter of
about 1 m. Both trenches and shafts are backfilled with coarse sand or fine gravel to prevent caving. Where permeable surface soils are not available, vadose zones are not sufficiently permeable to transmit water to the underlying aquifer, or aquifers are confined,
artificial recharge can be achieved by applying water to recharge wells penetrating the
aquifer (Fig. 24.1E).
24.3
SYSTEMS
The first step in planning and designing an artificial recharge system is selecting the type
of system to be used, based on soil and hydrogeologic information. Often, the choice is
obvious or determined by other factors such as high cost of land in urban areas which preclude the use of surface systems or excavated basins. Another important factor in the
selection of the type of recharge system is the pretreatment of the water required before
recharge to minimize physical, chemical, or biological clogging of the infiltrating soil surface in basins or of the walls in trenches, shafts, and wells. Since clogging layers are easier to control and remove in surface systems, proper pretreatment is especially important
for trenches, shafts, and wells. Since trenches and shafts are relatively inexpensive, they
FIGURE 24.1 Recharge systems for increasingly deep permeable materials: surface basin (A), excavated basin (B),trench (C), shaft or vadose zone well (D),
and aquifer well (E).
can be replaced when their useful life has come to an end. On the other hand, recharge
wells, while much more expensive than trenches and shafts, are more amenable to clogging prevention and control by frequent pumping and periodic redevelopment. All these
aspects have to be factored in, however, when selecting the type of recharge system.
While presedimentation in settling ponds may be adequate pretreatment for surface infiltration systems, recharge with trenches, shafts, and wells may require coagulation and
sand filtration to remove suspended solids to low enough levels. Where sewage effluent
is used for groundwater recharge, conventional primary and secondary treatment may be
sufficient for surface infiltration systems. However, effluent recharge with wells may
require advanced treatment processes (nitrogen and phosphorus removal, sand filtration,
removal of suspended solids and parasitic organisms like helminths, Giardia, and
Cryptosporidium, disinfection for removal of viruses and bacteria, and RO for removal of
organic compounds and other chemicals).
While conjunctive use of surface and groundwater often is the main objective of artificial recharge, recharge is also used for other purposes such as creating hydraulic barriers
against seawater intrusion, stopping land subsidence, stopping spread of contaminated
groundwater (plumes) in the aquifer from point sources of pollution, raising groundwater
levels for reduction of pumping costs, protecting streams, wetlands or other ecosystems,
protecting wooden building foundations that would decay when above groundwater, etc.
Dams for surface storage of water have a number of disadvantages, such as evaporation
losses (about 2 m/year in warm, dry climates), sediment accumulation, potential of structural failure, and increased malaria, schistosomiasis, and other diseases in the local population (Devine, 1995; Knoppers and van Hulst, 1995; Pearce, 1992). New dams often are
more and more difficult to build because of poor economics and public opposition. They
interfere with the river ecology and may flood sensitive areas (cultural, religious, archeological, environmental, and recreational, scenic, and so on.). People living on the reservoir
site have to relocate and good dam sites are becoming scarcer. Dams are not sustainable
because eventually they all silt up, and they are not effective for long-term storage of water
(years or decades), which may become increasingly necessary as increases of carbon dioxide and other greenhouse gases in the atmosphere may cause global climatic changes and
increase the probability of extremes in weather such as more prolonged droughts and periods of excessive rain. This increases the need for storing excess water in wet periods to
meet water demands in dry periods. Underground storage via artificial recharge then is preferred because of essentially zero evaporation, economics, and other factors. The water
sources for artificial recharge include water from perennial or intermittent streams that may
or may not be regulated with dams, storm runoff (including from urban areas), aqueducts
or other water conveyance facilities, irrigation districts, drinking water treatment plants,
and sewage treatment plants. Of all the water in the world, 97 percent is salt water in the
oceans (Bouwer, 1978). Of the remaining fresh water, two-thirds is in the form of ice in arctic and mountainous regions. Of the remaining liquid fresh water, less than 2 percent is surface water in streams and lakes, and much of that is fed by groundwater. More than 98 percent of the world's liquid fresh water thus is groundwater. Not only is groundwater the dominant water resource, but the potential for underground storage of water also vastly exceeds
that for surface storage.
24.3.1 Surface Systems
Surface infiltration systems for artificial recharge of underlying unconfined aquifers can
be constructed in streambeds (in-channel systems) or outside streambeds (off-channel systems). In-channel systems (Fig. 24.2) typically consist of weirs or dams across the
FIGURE 24.2 Schematic of in-channel infiltration systems consisting of low weirs in narrow steep channel (upper left), bigger dams in wider, milder sloping channel (upper right),
and T-levees in wide, flat channels (bottom).
streambed to back the water up and spread it over the entire width of the streambed to
increase the water depth and the wetted area and, hence, the infiltration of water. The
weirs or dams are small and closely spaced where channel slopes are steep, and larger and
more widely spaced where slopes are flat. The groundwater table for these systems must
be well below the streambeds so that the streams are "losing." The weirs or dams are made
of sheet metal (low weirs), earth, concrete, or inflatable rubber (fabridam). The dams
should have adequate spillway capacity or structural washout sections to handle large
flows. Such flows can clean the channel from fine sediment that may have accumulated
above the dam. Otherwise, the channel periodically must be mechanically cleaned to
remove accumulated sediment. Such sediment can be a "layercake" of coarse and fine
material deposited during high and low flows, respectively. This sediment greatly restricts
infiltration of water and storage of water behind the dams when flood waters recede and,
hence, must be removed. In coastal plains or other areas of mild channel slopes, dikes or
berms about 1 m high are pushed up by bulldozers to make T- or L-dikes in the streambed
that force the water into circuitous paths that cover the entire streambed (Fig. 24.2) and,
hence, increase the infiltration and groundwater recharge. Periodic high flows can destroy
the dikes but also clean the bottom from fine sediment and other clogging material. For a
system in southern California, this happens at least once a year. When the high flows have
receded, bulldozers restore the dikes in a few days.
In-channel systems are most effective if the stream is dammed higher up in the watershed; for example, in the mountains. This removes sediment from the water and also regulates the flow by storing water during high-flow events and releasing it slowly for
recharge. Storage dams themselves are, of course, ineffective for groundwater recharge
because of the sediment that accumulates on the bottom of the reservoirs.
Off-channel systems consist of basins constructed by building berms from soil that is
normally excavated from the basin areas. They can be square or rectangular, or of irregular form to adapt to existing topographies or for environmental enhancement. The latter
may include nature and wildlife benefits, scenic enhancement (landscaping, and the like),
trails for walking and bicycle riding, recreational facilities (picnic tables, playgrounds),
and so on. Each basin should have its own water supply. Deeper basins may also need a
drainage facility so that water can be drained out for drying and maintenance where dry-
ing by natural infiltration would take too long. Each basin is then hydraulically independent and can be operated according to its best schedule of flooding, drying, and cleaning.
The inflow facility should be carefully designed to dissipate the energy of the inflowing
water to avoid soil erosion in the basin. In finer textured soils such erosion will cause the
water to become "muddy" which will lead to clogging as the fine particles settle out in the
rest of the basin. Basin banks also need to be protected against erosion by rainfall and
wave action.
Occasionally, agricultural irrigation-type systems are also used for recharge, such as
borders, furrows, and terraces. On very permeable soil in irregular topographies like
kames, eskers, and other glacial deposits, low-pressure perforated pipe (hole diameter
about 5 mm) can be used to apply the water so as to avoid construction of basins and
destruction of trees or other natural vegetation.
24.3.2 Trenches, Shafts, and Wells.
Trenches (Fig. 24.3) are constructed with backhoes or other trenching equipment.
Vadose zone shafts (Fig. 24.3) are drilled with bucket augers. Trenches and shafts are
backfilled with fine gravel or coarse sand. For trenches, water is supplied with a perforated pipe on top of the backfill and the trench can be covered to avoid exposure to
sunlight and public access. For shafts, water is supplied with a vertical perforated pipe
or well screen in the center of the backfill. Recharge wells in aquifers are drilled with
conventional well drilling techniques and are constructed much like pumped wells,
including sand or gravel envelopes in unconsolidated materials. Careful grouting is
necessary where large injection pressures are used as, for example, in fractured rock
aquifers. Injection water flows into the well with one or more pipes inside the well.
These pipes should terminate some distance below the water level in the well to avoid
free falling water that can cause air entrainment and subsequent air blocking of pores
in the aquifer. For the same reason, the pipe or pipes themselves also should have full
flow. This can be achieved by using several pipes with different diameters so that the
small flows can go through the smaller pipe to maintain full-pipe flow. Alternatively,
one pipe can be used if it has an exit valve at the bottom that can be adjusted to always
FIGURE 24.3 Vadose zone recharge well with sand or gravel fill and perforated supply pipe
(left) and recharge trench with sand or gravel fill, supply pipe, and cover (right).
create full-pipe flow. Special valves are available for this purpose. Recharge wells
should be equipped with permanent, dedicated pumps to allow periodic pumping of the
well for clogging control.
24.4
INFILTRATION
WETTED
ZONE
yardstick, clock time is recorded, and the cylinder is filled back to the top. This is repeated for about 6 hours or when the accumulated infiltration has reached about 20 in,
whichever comes first. The last drop yn is measured and clock time is recorded to obtain
the time increment Arn for yn. A shovel is used to dig outside the cylinder to determine
the distance x of lateral wetting or divergence (Fig. 24.4). The infiltration rate In inside
the cylinder during the last water level drop is calculated as yjtn. The corresponding
downward flow rate or flux iw in the wetted area below a cylinder or radius r is then calculated as
'.-^
The depth L of the wet front at the end of the test is calculated from the accumulated
drop yt of the water level in the cylinder as
-^s?
where n is the tillable porosity of the soil. The value of n is estimated from soil texture and
initial water content. For example, n may be about 0.3 for dry uniform soils, 0.2 for moderately moist soils, and 0.1 for relatively wet soils. Well-graded soils would have lower values of n than uniform soils. The value of L can also be determined by digging down with a
shovel right after the test and seeing how deep the soil has been wetted. This works best if
the soil initially is fairly dry, there is good contrast between wet and dry soil, and there are
no rocks. Applying Darcy's law (see section 4.2.1) to the downward flow in the wetted zone
then yields
( . = K^L-HJ
(243)
where z is the average depth of water in the cylinder during the last water level drop. The
term hwe is the water-entry value of the soil and it is used to estimate the suction at the wet
front as it moves downward. The water-entry value is about half the air-entry value (due
to hysteresis) and may be estimated as follows (in inches of water):
Coarse sands
Medium sands
Fine sands
Loamy sands-sandy loams
Loams
Structured clays (aggregated)
Dispersed clays
2
4
6
-10
14
14
-40
Since K is now the only unknown in Eq. (24.3), it can be solved for K as
i*L
)
(24.4)
(z + L-hwe
The calculated value of K can be used as an estimate of long-term infiltration rates in
large and shallow inundated areas, without clogging of the surface and without restricting
layers deeper down. Because of entrapped air, K of the wetted zone is less than K at saturation, for example, about 0.5 of K at saturation for sandy soils, and about 0.25 of K at
saturation for finer soils. If the lvalues calculated with Eq. (24.4) look promising for an
infiltration system, the next step is to put in some test basins of about 1 acre for long-term
K=
flooding to evaluate clogging effects and potential for infiltration reduction by restricting
layers deeper down. Good agreement has been obtained between predicted infiltration
rates (K in Eq. 24.4) and those of larger basins. Six infiltrometers were installed in a field
west of Phoenix, Arizona. They gave an average K of 40 cm/day. Two test basins of 0.75
acre each in the same field yielded final infiltration rates of 30 and 35 cm/day (Bouwer
et al, 1997), which is a good agreement. If the infiltrometer tests give infiltration rates
that are too low for surface infiltration systems, alternate systems such as excavated
basins, recharge trenches, recharge shafts or vadose zone wells, or aquifer wells can be
considered.
The above procedure is by no means exact. However, in view of spatial variability
(vertical as well as horizontal) of soil properties, exact procedures and measuring water
level drops with vernier equipped hook gages are not necessary. The main idea is to
account somehow for divergence and limited depth of wetting, rather than applying a flat
reduction percentage to go from cylinder infiltration rates to long-term large area infiltration rates. Because of spatial variability, cylinder infiltration tests should be carried out at
various locations within a given site. Finally, the resulting infiltration rates should never
be expressed in more than two significant figures!
Hypothetical Example: A cylinder 24 in in diameter and 12 in high was driven lin into
a relatively dry sandy loam soil. The following time and drop-of-water level data were
recorded:
0800
0830
0900
1000
1100
Digging with a shovel showed a lateral wetting of 12 in outside the cylinder, n was
taken as 0.2 and hwe as - 10 in. The value of in is 4/1 = 4 in/h which when substituted
into Eq. (24.1) yields iw = 4 n 122In (12 + 12)2 = 1 in/h. The value of yt is 16 in, which
when substituted into Eq. (24.2) yields L = I6nl22/Q.2n (12 + 12)2 = 20 in. Since the
average water depth in the cylinder during the last measured water level drop was 11
4/2 = 9 in, /Hs calculated with Eq. (24.4) as 1 X 20/(9 + 20 + 10) = 0.5 in/h. Thus, the
last measured cylinder infiltration rate of 4 in/h is reduced to 0.5 in/h to eliminate effects
of divergence, limited depth of wetting, and water depth in the cylinder to produce a value
that can be used to predict potential infiltration rates for extended inundation of large
areas.
24.4.3
Clogging
The main enemy of infiltration systems for artificial recharge is clogging of the infiltrating surface (bottoms of basins or other surface infiltration systems, and walls of trenches,
shafts, and wells). Clogging is caused by inorganic (clay, silt) and organic (algae, sludge)
suspended solids in the water that accumulate on the infiltrating surface, and by microorganisms that grow on the soil particles (biofilms) and produce polysaccharides and other
insoluble metabolites to form a soil-clogging biomass. Bacteria also can produce gases
(nitrogen, methane, carbon dioxide) that can block soil pores. Gas also can be formed in
aquifers when water from recharge wells contains entrained air or is cooler than the
groundwater itself. As the recharge water then warms up in the aquifer, dissolved air may
go out of solution to form pore blocking air pockets in the aquifer (called air binding).
Also, well recharge can lead to precipitation of iron and manganese oxides or hydroxides
as dissolved oxygen levels change, and to solution or precipitation of calcium carbonates
due to changes in pH and carbon dioxide levels.
Since clogging layers are much less permeable than the natural soil material, they
reduce infiltration rates and become the controlling factor or "bottleneck" in the infiltration process (Fig. 24.5). When the infiltration rate in surface systems has become
less than the hydraulic conductivity of the soil below the clogging layer, this soil
becomes unsaturated to a water content whereby the corresponding unsaturated
hydraulic conductivity becomes numerically equal to the infiltration rate (Bouwer,
1982). The resulting unsaturated downward flow then is entirely due to gravity with a
hydraulic gradient of 1. The thickness of clogging layers may range from 1 mm or less
(biofilms, thin clay, and silt layers or "blankets") to several cm or more for thick sediment deposits.
Clogging is best controlled by prevention; that is, by removing the parameters that
cause clogging. For surface water, this typically means presedimentation to settle clay,
silt, and other suspended solids. This can be accomplished by dams in the river or aqueduct system (which would also regulate the flow) or by passing the water through dedicated presedimentation basins prior to recharge. Coagulants like alum and organic polymers can be used to accelerate settling. For well recharge, filtration may also be necessary. Algae growth and other biological clogging in basins can be reduced by removing
nutrients (nitrogen and phosphorus) from the water. This is also important where trenches, shafts, or wells are used for recharge with sewage effluent or effluent contaminated
water. Organic carbon levels then must also be reduced, using activated carbon filtration
and/or possibly reverse osmosis or other membrane filtration. Disinfection with chlorine
or other disinfectants with residual effects reduces biological activity on and near the
walls of the trenches, shafts, or wells and, hence, clogging. Clogging rates increase with
increasing infiltration rates, because of the increased loading rates of suspended solids,
nutrients, and organic carbon on the surface. Because of this, increasing the injection pres-
CLOGGING LAYER
UNSATURATED
CAPILLARY FRINGE
WATER TABLE
AQUIFER
FIGURE 24.5 Schematic of infiltration basin with clogging layer, unsaturated flow to aquifer, and capillary fringe above water table.
sures in recharge wells that show signs of clogging can actually hasten the clogging
process. Regular pumping of recharge wells and periodic redevelopment of the wells can
control and delay clogging, but probably not "forever." Increasing the water depth in
recharge basins or the injection pressure in recharge wells can compress the clogging layer
which reduces its permeability and, hence, the infiltration rates (see Sec. 24.4.4).
For surface infiltration systems, clogging is controlled by periodically drying the
basins or other infiltration facility, and letting the clogging layer dry, decompose, shrink,
crack, and curl up. This may be sufficient to restore infiltration rates to satisfactory values. If clogging materials continue to accumulate, they must be periodically removed at
the end of a drying period. This can be done mechanically with scrapers, front-end loaders, graders, or manually with rakes. After removal of the clogging material, the soil
should be disked or harrowed to break up any crusting that may have developed at or near
the surface. Disking or plowing clogging layers as such into the soil gives short-term
relief, but eventually fines and other clogging materials will accumulate in the topsoil and
the entire disk or plow layer must be removed. For good quality surface water with very
low suspended solids contents, drying and cleaning may be necessary once a year or
maybe every few years. Where soils are relatively fine textured or have many stones, clogging control becomes a major challenge (see Sec. 24.6). Where the water is extremely
muddy or where inadequately treated sewage effluent is used, drying and cleaning may be
needed after each flooding period which may then be as short as 1 or 2 days.
24.4.4 Effect of Water Depth on Infiltration
If there is no clogging layer on the bottom, the groundwater is relatively deep, and the
vadose zone is uniform, the downward flow in the vadose zone is governed by gravity and
water depth in recharge basins only has a minor effect on infiltration (Bouwer and Rice,
1989). For example, if the general groundwater table away from the infiltration system is
at a depth of 10 m, increasing the water depth in the basin from 0.1 to 1.1 m (about 10fold) would increase the infiltration rate only by about 10 percent. If, on the other hand,
infiltration is controlled by a clogging layer on the bottom and the underlying material is
unsaturated, the infiltration rate theoretically increases almost linearly with water depth,
so that doubling the water depth would double the infiltration rate (Bouwer, 1982). In reality, however, this seldom occurs because clogging layers often are loose deposits that are
quite compressible. Since increasing water depths produce increasing intergranular pressures in the clogging layer, this layer becomes compressed as the water depth in the basin
increases and, hence, becomes denser and less permeable (Bouwer and Rice, 1989). Thus,
in practice, increasing the water depth in an infiltration basin with a clogging layer will
cause a less than linear increase and maybe even a decrease (if the clogging layer is very
compressible) in infiltration rate. If the infiltration rate does not increase linearly with
water depth, increasing the water depth will then reduce the rate of turnover or the
replacement rate of the water in the basin. This increases the time that water is exposed to
sunlight which, for eutrophic water with relatively high concentrations of nitrogen, phosphorus, and organic carbon, means additional algae growth and especially the single-cell
type algae like Carteria klebsii that give the water a green appearance. On infiltration,
these algae form an algae filter cake on the bottom which increases the clogging process.
Also, photosynthezing algae absorb dissolved carbon dioxide from the water, which at
high algae concentrations increases the pH of the water to 9, 10, or maybe even 11. This
in turn causes precipitation of dissolved CaCO3, which further aggravates the clogging
process and, hence, decreases infiltration rates. This explains why increasing water depths
in infiltration basins has actually produced decreases in infiltration rates, to the surprise
and dismay of operators who wanted to increase infiltration rates!
For this reason, and also to allow quick draining of basins by natural infiltration when
a drying period is needed, shallow basins (water depth of about 20 cm) are preferred. If
plant growth becomes a problem, water depths may periodically have to be increased to
kill the plants before they emerge above the water level. Deep basins (10 m or more), like
old gravel pits, are not very good for recharge because clogging layers become compressed and infiltration rates are so low that the pits need to be pumped out to start a drying period so that clogging layers can be removed. This is expensive and slow, so that drying often is delayed and infiltration rates remain disappointingly low. An advantage of
deep basins is, however, that they can store more water in times of high runoff events. For
subsequent recharge, however, this water should then be conveyed to shallow basins
where clogging is less and easier controlled by drying and cleaning.
Where groundwater levels are so high that they are close to the bottom of the basin
or even above the bottom, the flow in the aquifer is mainly in the horizontal direction
so that infiltration rates are controlled by the slope of the water table away from the
basins. Under these conditions, increasing the water depth in the basins will increase
infiltration rates in direct proportion to the vertical distance between the water surface
in the basin and the groundwater level some distance away from the basin (Bouwer,
1990; Bouwer and Rice, 1989).
If the groundwater table is above the bottom of a recharge basin, the basin can never
be dried for infiltration recovery and removal of the clogging layer. Thus, sediment continues to accumulate in those basins and infiltration occurs mainly through the sides of the
basins which are then made and kept as steeply as possible to minimize sediment accumulation on the basin banks. Under those conditions, increasing the water depth in the
basin will also increase the infiltration or recharge rate.
24.4.5 Effect of Groundwater Depth on Infiltration Rate
Usually, bottoms and banks of infiltration basins are covered with a clogging layer that
controls and reduces infiltration rates so that the underlying soil material is unsaturated
(Fig. 24.5). The water content in the unsaturated zone then is at a value whereby the corresponding unsaturated hydraulic conductivity is numerically equal to the infiltration rate,
since the downward flow is due to gravity alone with a hydraulic gradient of 1 (Bouwer,
1982). The unsaturated zone breaks the hydraulic continuity between the basin and the
aquifer, so that infiltration rates are independent of the depth to groundwater, as long as
the water table is deep enough so that the top of the capillary fringe above the water table
is below the bottom of the basin. This capillary fringe may be about 1 ft (30 cm) thick for
medium sands, more for finer sands or soils, and less for coarse sands. Thus, a conservative conclusion is that as long as the groundwater table is more than 3 ft (about 1 m) below
the bottom of a basin where infiltration is controlled by a clogging layer on the bottom,
infiltration rates are unaffected by changes in groundwater levels. If the groundwater table
rises, infiltration rates will start to decrease when the capillary fringe reaches the bottom
of the basin, and continue to decrease linearly with depth of groundwater below the water
level in the basin, until they become zero when the groundwater table has risen to the
water surface in the basin.
Where the water for recharge is exceptionally clear and free from nutrients and organic carbon, temperatures are low, and soils are relatively coarse, infiltration can go on for
considerable time without development of a clogging layer on the bottom. In that case,
there is direct hydraulic continuity between the basin and the aquifer with the groundwater table joining the water surface in the basin (Fig. 24.6; Bouwer, 1969, 1978).
Groundwater levels are then characterized by the depth Dw of the groundwater table below
the water surface in the basin at sufficient distance from the basin to be relatively unaffected by the recharge flow system (Fig. 24.6). If Dw is relatively large, the flow below the
recharge basin is mainly downward and gravity controlled (Bouwer 1969, 1978) with a
hydraulic gradient of about 1 (Fig. 24.6). In that case, infiltration rates are essentially
unaffected by depth to groundwater. However, if groundwater levels rise and Dw decreases, the flow from the recharge basin becomes more and more lateral until it is controlled
by the slope of the water table away from the basin (Fig. 24.6; Bouwer 1969, 1978).
Modeling these flow systems on an electrical resistance network analog has shown that
the change from gravity controlled flow to slope-of-the-water-table-controlled flow
occurs when Dw is about twice the width W (or diameter) of the recharge system (Bouwer,
1990). This relation is shown in Fig. 24.7 where I is the infiltration rate per unit area of
water surface in the basin and K is the hydraulic conductivity in the wetted zone or
aquifer. Thus, as long as Dw < 2K, infiltration rates decrease linearly with decreasing Dw
and reach zero when Dw = O (Fig. 24.7). These relations apply to uniform, isotropic underground formations. Anisotropic or stratified situations need to be considered on a case-bycase basis.
Infiltration rates in clean basins (no clogging layers) thus are more sensitive to depth
to groundwater than in clogged basins. Clogged basins are the rule rather than the exception and groundwater mounds can rise much higher there than below clean basins before
reductions in infiltration rates occur. Sometimes maximum mound heights are dictated by
circumstances other than their effect on infiltration rates, such as presence of sanitary
landfills, underground sewers or other pipelines, basements (especially deep basements of
commercial buildings), cemeteries, and deep = rooted vegetation like old trees that may
die when groundwater levels rise too high.
24.4.6 Induced Recharge
One method for increasing recharge of groundwater from losing streams is to lower
groundwater levels near the stream by groundwater pumping. As described in the previ-
FIGURE 24.7 Dimensionless plot of seepage (expressed as I/K) versus depth to groundwater (expressed as DJW) for clean stream channel or basin (no clogging layer on bottom).
ous section, this will only increase recharge or seepage losses from the stream if the
stream bottom is not covered by a clogging layer and the groundwater table is relatively
high so that the depth to groundwater below the water level in the stream is less than two
stream widths (Fig. 24.7). If the groundwater level is already more than two stream widths
below the water surface in the stream, recharge is already near maximum and further lowering of the groundwater will have little or no effect on recharge (Fig. 24.7). If the stream
wetted perimeter is covered by a clogging layer, lowering of groundwater levels will only
increase recharge if the groundwater table is above the stream bottom or only a small distance below the stream bottom so that the capillary fringe still touches the bottom. If the
groundwater level already is deep enough for the top of the capillary fringe to be below
the stream bottom, the soil material below the stream bottom is unsaturated. There is then
no direct hydraulic connection between the stream and the aquifer, so that further lowering of the groundwater level by pumping from wells will not increase the recharge or
seepage losses from the stream.
Groundwater pumping can decrease streamflow in two ways. If the groundwater table
is above the water surface in the stream, the stream is gaining and groundwater moves into
the stream and supports the base flow in the stream. Any pumping of groundwater will
then diminish the flow of groundwater into the stream. If the groundwater table is below
the water surface in the stream, lowering the groundwater table will increase recharge or
seepage losses from the stream when the groundwater table is high, but not when it is relatively low, as explained above. These relationships can then be used in resolving situations where there are real or perceived conflicts between owners of surface water rights
and holders of groundwater pumping rights (Bouwer and Maddock, 1997).
Where groundwater levels are high enough to support growth of phreatophytes (plants
and trees that live off groundwater), a lowering of groundwater levels can reduce the
water uptake by phreatophytes (Bouwer, 1975) and, hence, increase the capture of
groundwater by the aquifer (Bouwer and Maddock, 1997). This means that groundwater
pumping can then lessen the reduction in flow of tributary groundwater to gaining
streams, and reduce water losses from losing streams.
24.5
GROUNDWATERMOUNDING
flow through the restricting layer (Fig. 24.8). If the material below the restricting layer is
relatively coarse, the pressure head of the water at the bottom of the restricting layer can
be taken as zero. For finer materials below the restricting layer, the water-entry values listed in Sec. 24.4.2 be taken as a first estimate. The resulting equation for the perched
mound, assuming zero pressure head for the water at the bottom of the restricting layer, is
then (Bouwer et al., 1997).
F' 1
Lp = I^-r
(24.5)
K.
where Lp = height of perching mound above restricting layer, Lr = thickness of restricting layer, / = infiltration rate, Kr = hydraulic conductivity of restricting layer, and K5 =
hydraulic conductivity of soil above restricting layer.
Often, / will be much smaller than Ks because surface soils are finer textured than
deeper soils, or there is a clogging layer on the surface soil that reduces infiltration.
Also, i often will be much larger than Kr. For these conditions, Eq. 24.5 can be simplified to
^ = I-TJT
Ar
(24.6)
Lp should be small enough so that the top of the perched mound is deep enough to
avoid reductions in infiltration rates, as discussed in the Sec. 24.4.5. Perched mounds
above noncontinuous perching layers (lenses or strips) are not as high as above continuous perching layers with the same Lr and Kr because there is then also lateral flow in the
perched mound. Thus, Eqs. 24.5 or 24.6 will then overestimate the height of the perched
mound. Where perched mounds would be too high, recharge may be achieved with long
narrow basins that allow lateral flow in the mound and reduce its height as the perched
mound spreads laterally on the restricting layer until all the recharge water will pass
through the restricting layer (Bouwer, 1978). Another possibility is to use infiltration
basins that are excavated through the restricting layer (Fig. 24.1B). If the restricting layer
is too deep, recharge trenches (Fig. 24.1C) or recharge shafts (large diameter holes in the
vadose zone, also called dry wells or vadose zone wells) can be used (Fig. 24.1D).
24.5.2 Groundwater Mounds
Numerous papers have been published on the rise of groundwater mounds on the aquifer
in response to infiltration from a recharge system, and some also on the fall of the mound
after infiltration has stopped (Glover, 1964: Hantush, 1967; Marino, 1975a, b; Warner et
al., 1989). The usual assumption is a uniform aquifer of infinite extent with no other
recharges or discharges. One of the difficulties in getting meaningful results from the equations is getting a representative value of aquifer transmissivity. The most reliable transmissivity data come from calibrated models. Next come Theis-type pumping tests, step-drawdown and other pumped well tests, and slug tests, in decreasing order of sampling size.
Slug tests, while simple to carry out, always have the problem of how to get representative
areal values from essentially point measurements. Averages from various tests often seriously underestimate more regional values (Bouwer, 1996, and references therein).
Piezometers at two different depths in the center of a mound enable the determination of
both vertical and horizontal hydraulic conductivity of an aquifer already being recharged
with an infiltration system (Bouwer et al. 1974). In deep or thick unconfined aquifers,
streamlines of recharge flow systems are concentrated in the upper or "active" portion of
the aquifer, with much less flow and almost stagnant water in the deeper or "passive" portion of the aquifer. Use of transmissivities of the entire aquifer between the water table and
the impermeable lower boundary for mound calculations can then seriously underestimate
the rise of the mound. Older work (Bouwer, 1962) with resistance network analog modeling showed that for rectangular recharge areas, the thickness of the active, upper portion of
the aquifer is about equal to the width of the recharge area. This thickness should then be
multiplied by K of the upper aquifer to get an "effective" transmissivity for mounding predictions. Also, if the Hantush or other equation is used to calculate long-term mound formation, as for water banking in areas with deep groundwater levels, larger transmissivity
values should be used to reflect the increase in transmissivity as the groundwater level
rises. Otherwise, the Hantush equation will seriously overestimate the mound rise.
The best way to get representative transmissivity values for artificial recharge systems
is to have a large enough infiltration area that produces a groundwater mound, and to calculate the transmissivity from the rise of the groundwater mound using, for example,
Hantush's equation (Eq. 24.7). The fillable porosity to be used in the equations for mound
rise usually is larger than the specific yield of the aquifer, because vadose zones often are
relatively dry, especially in dry climates and if they consist of coarse materials like sands
and gravels. The fillable porosity should be taken as the difference between existing and
saturated water contents of the material outside the wetted zone below the infiltration system. The Hantush equation is
hx,y,t ~H=V~^
(24.7)
where hx t = height of water table above impermeable layer at x, v, and time (Fig. 24.9),
H = original height of water table above impermeable layer, va = arrival rate at water
table of water from infiltration basin, t = time since start of recharge, / = fillable porosi-
ty (1 >/> 0)> = length of recharge basin (in y direction), W = width of recharge basin
(in x direction), n = (4t 77/)-1/2, and F(OC, p) = /1Q erf(ai - 1/2) erf($i - m)dx [which was
tabulated by Hantush (Table 24.1)]
where a = (W/2 + x) n or (W/2 - x)n and (3 = (L/2 + y)w or (1/2 - y)n.
Of most interest to operators and managers often, is the long-term effect of recharge
on groundwater; that is, where the groundwater mound will be 10, 20, or 50 years from
now, how much water can be stored or "banked" underground, will the whole area become
waterlogged, and how must the water be recovered from the aquifer to prevent waterlogging? Computer models can be used to simulate regional recharge inputs and pumped
well outputs for the aquifer. However, a quick idea about ultimate or quasi-equilibrium
mound heights can be obtained from a steady-state analysis where the mound is considered to be in equilibrium with a constant water table at some depth and at some distance
from the infiltration system. The constant "faraway " water table is obtained by groundwater pumping, discharge into surface water like rivers or lakes, or some other "control."
Usually, recharge systems consist of a number of basins or other infiltration facilities.
Steady-state equations were developed for two general geometries of the entire recharge
area: (1) the basins form a long strip with a length of at least 5 times the width so that after
long times it still performs about the same as an infinitely long strip (Glover, 1964), and
(2) the basins are in a round or square or irregular area that can be handled as an equivalent circular area (Bouwer et al., 1999b). For the long strip (Fig. 24.10) the groundwater
flow away from the strip was taken as linear horizontal flow (Dupuit-Forchheimerflow).
Below the infiltration area, the lateral flow was assumed to increase linearly with distance
from the center. The lateral flow was then assumed to be constant between the edge of the
recharge system at distance W/2 from the center and the constant control water table at
distance Ln from the edge (Fig. 24.10). This yielded the equation
FIGURE 24.9 Geometry and symbols for recharge system and groundwater mound.
"<-"- = f (T+ L)
(24.8)
for the ultimate rise of the groundwater mound below the center of the recharge strip when
there is equilibrium between recharge and pumping from the aquifer (Bouwer et al.,
1999b). The symbols in this equation are (Fig. 24.10):
H0 = height of groundwater mound in center of recharge area,
Hn = height of groundwater table at control area,
/ = average infiltration rate in recharge area (total recharge divided by total area),
W = width of recharge area,
Ln = distance between edge of recharge area and control area, and
T = transmissivity of aquifer.
For a round or square recharge area (Fig. 24.11), the groundwater flow will be radially away from the area. The equilibrium height of the mound below the center of the
recharge system above the constant groundwater table at distance Rn from the center
of the recharge system then can be calculated with radial flow theory (Bouwer et al.,
1999b) as
H0-Hn = ^ [l + 2/n^]
(24.9)
where R is the radius or equivalent radius of the recharge area, Rn is the distance from the
center of the recharge area to the water-table control area (Fig. 24.11) and the other symbols are as defined above.
Equations (24.8) and (24.9) thus can be used to predict the final mound height below
a recharge area for a given elevation of the control water table at distance Rn or Ln from
the recharge area. As indicated for the Hantush equation, the value of Tin Eq. (24.8) and
(24.9) must reflect the average transmissivity of the aquifer at the ultimate equilibrium
mound height. If the ultimate mound height is too high, Rn or Ln must be reduced by
groundwater pumping from wells closer to the recharge area or Hn must be reduced by
pumping more groundwater. Equations (24.8) and (24.9) can then indicate where groundwater should be recovered and to what depth groundwater levels should be pumped to prevent groundwater tables below the recharge areas from rising too high. Ultimate mound
heights can also be reduced by making the recharge area longer and narrower, or by reducing recharge rates as can be achieved by using less water for recharge or by spreading the
infiltration facilities out over a larger area.
24.6
CHALLENGINGSOILS
Index
Index terms
Links
A
Acoustic (sonic) velocity
12.2
Aguado
1.14
Ancient dams
1.13
1.14
Egyptian systems
1.7f
Kariz
Mesopotamian systems
Mexican systems
North American systems
1.7
1.7f
1.8
1.9f
1.10
Chaco systems
1.13f
Hohokam systems
1.11f
1.12f
Qanat
1.7
1.8f
Shaduf
1.4
1.6f
Swape
1.4
1.14
Minoan systems
1.14
1.15f
Roman aqueducts
1.20f
1.22f
Roman systems
1.16
1.19f
1.17f
1.20f
1.10f
1.18f
1.21f
7.47
1.17
23.15f
23.15f
23.15f
Natural attention
23.15f
I.1
I.2
Index terms
Aquifer transmissivity
Links
4.8
Artificial recharge,
Subsurface systems, dry wells
Aquifer wells
Aquifer storage and recovery wells
24.7
24.31
24.3
Recharge shafts
24.29
24.6f
24.29
Surface systems
24.4
In-channel systems
24.4
24.5f
Off-channel systems
24.4
24.7
5.3
Atmospheric radiation
5.4
B
Benefit-cost analysis
1.27
Benefit-cost ratio
1.28
Flowchart of
1.29f
1.28
Marginal values
1.28
Bernoulli's equation
2.13
Bioventing
23.51
Design calculations
23.55
Design principles
23.53
23.53
23.54
System schematic
23.52
23.53
2.14
16.29
21.3
16.29
I.3
Index terms
Links
16.32
16.29
16.29
Bowen ratio
Bridge-deck drainage design
5.7
13.36
13.38
13.37
13.39
Off-bridge inlets
13.36
Scuppers
13.36
13.37f
2.36
C
Caldarium
Capillary fringe
Castellum
Catchment runoff modeling
1.17
24.13
1.21
14.67
Calibration
14.71
Modeling procedure
14.69
Models
14.72t
14.75t
Verification
14.71
Cavitation
10.22
Cavitation index
17.40
Cavitation potential
17.40
Cavitation number
17.40
Cenote
1.14
10.6f
10.6f
10.7f
10.8
I.4
Index terms
Links
Channel bends
8.2
Channel linings
16.1
Flexible linings
16.1
Rigid linings
16.1
19.9
Channel stationing
19.9
Characteristic curves
2.27
Chezy equation
3.11
Chezy's law
6.4
3.11
Chultun
1.14
Clogging of,
Aquifer wells
24.31
24.31
24.31
24.31
Infiltration basins
Colebrook-White equation
Contaminant migration barriers
24.10
2.17
23.61
23.65
23.61
Design process
23.61
Recirculation wells
23.62
10.12
23.63f
23.62
23.63
23.63
23.64f
23.61
I.5
Index terms
Contaminant mitigation prevention technologies
Links
23.14
23.16f
Physical barriers
23.16f
Reaction barrier
23.16f
23.5f
23.6f
DNAPL release
23.5f
LNAPL release
23.5f
23.15f
23.6f
2.6
Chemical species
2.7
Continuity
14.32
Energy
9.3
14.32
Mass
9.3
14.32
Molecular species
2.7
Momentum
9.24
Control volume
2.7
5.7
Critical depth
3.5
Critical flow
21.5
15.16
15.66
15.67
CDS
15.67
CULVERT2
15.67
HY8
15.66
15.21f
Design example
15.22
Inlet control
15.17
15.21f
6.6
CAP
Critical depth
14.32
15.18t
21.5
I.6
Index terms
Links
15.17
15.17
15.17
15.34f
15.30
End treatments
15.32
Inlet types
15.33
Side-tapered inlet
15.35f
Slope-tapered inlet
15.36f
Siphon
15.32
15.31
Special considerations
15.37
Debris control
15.38
Erosion
15.37
Sedimentation
15.38
Bend losses
Culvert design
15.9
15.10t
15.16
15.2
Flow characteristics
15.4
Headwater
15.2
Allowable
15.3
Maximum
15.3
Outlet velocity
15.31f
15.1
Design parameters
Inlet control
15.20f
15.30
Culverts
15.19f
15.3f
15.4
15.25
15.5
15.25
15.14f
Types of
15.6f
Junction losses
15.10
Nomographs
15.12
Outlet control
15.4
I.7
Index terms
Links
Culverts (Continued)
Entrance loss coefficients for
15.8t
Hydraulics of
15.6
Outlet velocity
15.14f
Types of
15.7f
Outlet velocity
15.13
Performance curves
15.23
Roadway overtopping
15.14
15.15f
15.15f
15.16f
Tailwater
Culvert flow transitions
Abrupt expansions
15.3
20.1
20.2
Design procedure
20.2
Example design
20.7
Gradual transitions
20.9
Design procedure
20.9
Example
20.10
Inlet control
20.13
Design procedure
20.13
Example
20.16
20.13f
20.1
20.2f
15.25
15.25
D
DAMBRK
19.26
Darcy's law
4.1
Anisotropic media
4.3
I.8
Index terms
Links
4.3
4.3
4.3
Vapor flow
4.18
Wetted zone
24.9
Darcy-Weisbach equation
Sewer pipe
2.16
14.5
14.8
3.12
18.40
Desander
8.3
Desanding basin
8.3
Desilting basin
8.3
Detention storage
Computer models for design
14.76
14.91
BASINOPT
14.91
PONDOPT
14.91
14.91
Detention basins
14.76
Basic elements
14.77f
14.82t
Design guidelines
14.79
Design procedure
14.81
14.78f
14.80f
14.79
14.81
14.70
Stage-storage parameters
Stage-storage relationships
9.9
21.4
14.81f
14.80
10.12
I.9
Index terms
Links
14.79
14.85
Design considerations
14.88
14.86
14.87f
14.88
14.89f
Retention basins
Design considerations
14.89
14.90
Outlet structures
Permanent pool volume
14.90f
14.88
19.5
19.4
19.4
19.4
19.6
Dispersion,
Density current mixing
5.17
Entrainment coefficient
5.18
Fraction entrainment
5.17
Interflow mixing
5.18
Bouyancy frequency
5.19
Grashof number
5.19
Intrusion length
5.20
Prandtl number
5.19
5.19
5.12
Longitudinal
5.11
Equation for
Meteorological forces
5.13f
5.11
5.26
5.27
Internal waves
5.27
I.10
Index terms
Links
Dispersion (Continued)
Langmuir circulation
5.27
Plunging breaker
5.27
Shortwave motion
5.26
Spilling breaker
5.27
Swash zone
5.27
Waves
5.26
5.26
Mixing dispersion
5.12
Plunge point
5.14
Separation point
5.15
5.12
Open channels
5.9
Outflow mixing
5.21
BETTER model
5.22
CE-QUAL-R1
5.22
CE-QUAL-W2
5.22
SELECT model
5.22
Withdrawal characteristics
5.23f
Overflow
5.17
Underflow
5.17
Draft tubes
Dupuit-Forchheimer flow
5.15
8.26
24.19
E
Energy dissipation for culverts and channels
Drop structures
20.17
20.33
20.38
20.35
20.44f
20.45
20.36
I.11
Index terms
Links
20.37
20.21
20.21
Design procedure
20.22
Example design
20.23
20.26
20.26
20.29
20.17
Definition sketch
20.17f
Design procedure
20.17
Example design
20.19
Energy dissipators
18.1
Channels
20.17
Culverts
20.17
18.25
Guidelines
18.25
18.1
Impact-type dissipators
18.1
Engineering economics
1.25
1.27
1.28
1.29f
1.28
Marginal values
1.28
Discount factors
1.25
1.28
1.29t
Energy gradeline
2.14
22.18
Entrapped air
12.2
EPANET
9.33
I.12
Index terms
Euler number
Evaporation
From natural water surface
Links
12.10
5.6
5.6
5.7
5.6
Bowen ratio
5.7
Exner equation
6.27
Failure probability
7.2
FESWMS-2DH
19.26
FLDWAV
19.26
Fick's law
Flip bucket
Fix grade node
Flood control channel
Components
5.7
4.4
17.38
9.2
16.1
16.5
Channel crossing
16.6
Channel inlet
15.5
Channel outlet
16.6
Channel reach
16.5
Confluence
16.6
Side inlet
16.6
Transition
16.6
Design
16.2
Design parameters
16.8
Channel profiles
16.9
Channel section
16.9
Channel slope
16.9
Duration of flooding
16.8
16.9
Flood frequency
16.8
16.9t
Linings
Curled mat
16.10
16.17
I.13
Index terms
Links
16.17
Flexible
16.11
Fiberglass roving
16.19f
Gravel
16.13f
Jute net
16.16f
Performance data
Rock channel
16.10
16.12f
Rock riprap
16.12
Turf-reinforcement mat
16.14
Types
16.10
Vegetative
16.12f
16.13f
16.15f
Stable channels
Boundary shear stress
16.16
16.14
16.6
16.8
Bends
16.8f
Channel sections
16.8f
Channel alignment
16.8
Dynamic-stable channels
16.6
Modes
16.6
16.7
Static-stable channel
16.6
Tractive force
16.7
F
Flood control channel design, composite sections
16.38
Manning's roughness
16.39
Special considerations
16.39
16.35
16.35
I.14
Index terms
Links
16.36
16.36
Resistance to flow
16.35
Bathhurst equation
16.35
16.36
16.36
16.19
16.29
16.29
16.32
16.29
16.29
16.26
16.27
16.21
16.21
16.22
Manning's roughness
16.21
Mild gradients
16.20t
Vegetative classes
16.22
16.29f
16.25f
Lining materials
16.24t
Noncohesive soils
16.24f
Retardance classes
Manning's roughness for
16.21
16.22
16.24
16.21
I.15
Index terms
Flood control channel design, with sediment transport
Links
16.39
Aggradtion
16.4
Bed forms
16.42
Degradation
16.41
16.40
Resistance to flow
16.42
16.40
Sediment discharge
16.41
16.42
Sediment supply
16.39
Sediment transport
16.40
Bed load
16.40
16.40
16.40
Floodplain area
19.1
Floodplain studies
19.9
Approaches
19.21
Critical depth
19.23
Critical flow
19.23
19.22
Hydraulic controls
19.23
One-dimensional flow
19.22
19.22
Steady flow
19.22
Subcritical flow
19.23
Supercritical flow
19.23
Two-dimensional flow
19.22
Uniform flow
19.22
Aerial topography
19.16
19.19
Channel structures
19.21
19.20f
I.16
Index terms
Links
19.25
DAMBRK
19.26
FESWMS-2DH
19.25
FLDWAV
19.26
HEC-2
19.25
HEC-RAS
19.25
NETWORK
19.26
RMA-2
19.25
SWMM
19.26
TABS-MD
19.26
UNET
19.26
WSPRO
19.25
WSP-2
19.25
19.16
Cross-section alignment
19.13
Cross-section interpolation
19.37
Cross-section locations
19.11
19.14f
19.17
Cross-section orientation
19.13
Culvert data
19.21
Field surveys
19.10
Floodway
19.6
Floodway determination
19.45
Performing studies
19.28
Starting conditions
19.28
19.30
Quality of analysis
19.38
19.17
19.31
19.23
19.15f
I.17
Index terms
Links
19.23
First-order streams
19.23
Higher-order streams
19.23
19.23
19.25f
Simple channels
19.23
19.24f
Floodway
19.45
Floodway determination
19.45
Flow control
19.24f
2.1
Flow measurement,
Flow conditioning
21.11
Channels
21.12
21.12
Options
21.11
Orifice plates
21.12
Flumes,
Critical flow flumes
21.10
Long-throated flumes
21.10
Short-throated flumes
21.10
Venturi flume
21.10
Installation requirements
21.10
Measurement accuracy
21.14
Accuracy terms
21.15
Comparison standards
21.17
21.16
21.9
Primary device
21.9
Quantity meters
21.9
Rate meters
21.9
Secondary devices
21.9
Wave suppression
22.5
21.13
I.18
Index terms
Links
21.14
21.14f
21.44
21.47f
21.48
21.46
21.45
Rising bubbles
21.46
Surface floats
21.45
21.44
21.29
21.30t
Circular channels
21.41
21.34f
Example calculations
21.36
21.39t
21.32t
21.30t
21.38
21.18
21.18
21.25
21.26t
Quality assurance
21.29
Selection guide
21.28t
Flow meters
22.4
Magnetic meters
22.5
Mechanical meters
22.6
I.19
Index terms
Links
22.5
Orifice plates
22.5
Parashall flumes
22.5
Pitot meters
22.5
Venturi meters
22.5
Ultrasonic meters
22.6
21.8
Forebay
8.2
Freeboard
8.2
Friction slope
3.22
3.22
19.34
3.22
19.34
3.22
19.34
3.22
19.34
Frigidarium
1.17
Froude number
3.1
16.3
5.14
5.14
5.16
Plunge point
5.16
Separation point
5.16
3.6
5.14
5.18
5.19
G
Gates
22.3
Seating head
22.3
Unseating head
22.3
13.2
Cross slope
13.3t
Grade
13.2
I.20
Index terms
Links
13.2
Hydroplaning
13.2
13.2
Sheetflow
13.2
Geometric similarity
Global positioning system
Gradually varied flow
21.8
19.10
3.16
Profile classification
3.20t
3.19
3.23
3.25
3.25
Dynamic model
3.25
3.25
13.29
Efficiency
13.29
13.31
13.31f
Interception capacity
13.32
Overall efficiency
13.31
13.31
Splash-over velocity
13.29
13.31
13.30t
Groundwater contamination,
Aquifer restoration process schematics
23.15f
23.15f
23.15f
Natural attention
23.15f
Bioventing
23.51
I.21
Index terms
Links
23.55
Design principles
23.53
23.53
23.54
System schematic
23.52
23.53
23.20t
Common sources
23.3f
23.61
23.65
23.61
Design process
23.61
Recirculation wells
23.62
23.63f
23.62
23.63
23.63
23.64f
23.61
23.14
23.16f
Physical barriers
23.16f
Reaction barrier
23.16f
23.5f
23.6f
DNAPL release
23.5f
LNAPL release
23.5f
Design flowchart
23.15f
23.6f
23.14f
23.59
Natural attenuation
23.60
I.22
Index terms
Links
23.25
23.31
Interceptor trench
23.29f
Liquid monitoring
23.26
23.27
23.32f
Skimming system
23.30f
23.30
Trench systems
23.28
23.28
23.29f
23.50
23.55
23.57
Biosparging
23.58
Blowers
23.58
Compressors
23.58
Design parameters
23.57t
Design principles
23.56
Pilot tests
23.58
Process principles
23.56
23.24
23.12
23.12
23.12
23.7
23.9t
23.7
23.9t
23.18
23.50
I.23
Index terms
Links
23.7
Scenarios
23.2
Site characterization
23.16
23.17
Geologic/hydrogeologic assessment
23.17
Historic assessment
23.17
Political assessment
23.17
Receptor assessment
23.17
23.34
Efficiency
23.39
23.35
23.50
23.49
Extraction pump/blower
23.45
Feasibility assessment
23.35
Injection blower
23.45
Instrumentation
23.46
23.39
Manifolding
23.46
23.36
23.38
Pilot tests
23.40
23.34
System design
23.42
Vapor treatment
23.45
Well construction
23.43
23.44f
23.12
23.25
23.13f
23.13f
23.13f
23.41f
I.24
Index terms
Links
23.23
23.14f
Groundwater flow:
Capture zone
4.10
Darcy's law
4.1
Interception wells
4.10
Capture zone
4.10
Dividing streamline
4.10
Stagnation point
4.10
Upgradient divide
4.10
Jacob's correction
Partially penetrating wells
4.9
4.11
24.20
Retardation factor
4.13
Saturated flow
Convolution
Mass transport equations
4.9
4.9
4.12
4.8
4.8
Superposition
4.9
Transient flow
4.9
Transmissivity
Well duplets
Groundwater mounding
Groundwater mounds
4.8
4.12
24.16
24.18
24.19f
24.24f
Hantush equation
24.18
Perched mounds
24.16
24.17
Restricting layer
24.16
24.17f
24.23f
I.25
Index terms
Links
4.13
4.14
4.14
4.13
4.13
4.14
4.12
Retardation factor
4.13
4.15
4.15
Continuity equation
4.15
Flow equations
4.15
4.16
4.15
4.16
4.17
4.17
Darcy's law
4.18
Discharge potential
4.18
Molar flux
4.18
4.18
Radial flow
4.18
4.18
4.19
4.19
4.19
Retardation factor
4.19
I.26
Index terms
Links
13.10
Gutter design
13.14
Depressed flow
13.15
13.15t
Manning's equation
13.14
Gutter design
13.14
Depressed flow
13.15
13.15t
Manning's equation
13.14
13.14f
13.14f
H
Hantush equation
Hardy Cross method
Hazen-Williams equation
24.18
9.11
2.18
10.10
9.6
22.84
10.9
3.19
15.39
19.25
3.19
19.25
15.39
19.6
Coefficients for
10.11t
22.86t
Headrace channel
8.1
Heat flux
5.6
5.7
5.6
HEC-2
HEC-RAS
Highway drainage
Bridge-deck drainage design
13.36
13.38
13.37
13.39
I.27
Index terms
Links
13.36
Scuppers
13.36
13.4
Check storm
13.5
Major considerations
13.5
13.6t
Minimum spread
13.6t
13.27
Classes of inlets
13.28
13.33
13.32
Efficiency
13.32
13.32
Length
13.32
Equal-length inlets
13.34
13.33
13.34
13.33
Slotted inlets
13.34
Orifices
13.33
Weirs
13.33
13.35f
13.29
Efficiency
13.29
13.31
13.37f
13.31f
Interception capacity
13.32
Overall efficiency
13.31
13.31
Splash-over velocity
13.29
13.31
I.28
Index terms
Links
13.30t
Inlets in sag
13.33
Inlet locations
13.35
13.40
13.32
Sweeper inlet
13.33
13.2
Cross slope
13.3t
Grade
13.2
Pavement
13.2
Hydroplaning
13.2
13.2
Sheetflow
13.2
Gutter design
13.14
Depressed flow
13.15
13.15t
Manning's equation
13.14
13.14f
13.6
13.12
13.13
13.12
13.6
Avoidance of hydroplaning
13.6
13.6
Rational method
13.6
13.13
13.10
Intensity-duration-frequency curves
13.8
13.9
13.12
I.29
Index terms
Links
13.9t
Rational method
13.6
Runoff coefficients
13.7t
Sheetflow
13.8
13.8
Time of concentration
13.7
13.16
13.20
13.20
13.21f
13.18t
13.19t
Manning's equation
13.17
13.20
13.23
13.25f
13.24f
13.27f
Shear stresses
13.20
13.22
Superelevation in bends
13.18
13.18t
13.24t
15.48
Aggradation
15.48
Contraction scour
15.48
Degradation
15.48
Design approach
15.48
15.53
15.61
15.58
15.57
15.64
I.30
Index terms
Links
15.60f
15.60
15.59f
15.64f
15.39
Armoring
15.40
Armoring layer
15.41
15.46
Check dams
15.46
15.43
Deflector spurs
15.44
Engineering analysis
15.39
Example calculations
15.47
15.43f
Geomorphic factors
15.40f
Guide banks
Hydraulic factors
Incipient motion analysis
Location factors
15.44
15.41f
15.39
15.40f
Retarder/deflector spurs
15.44
Retarder spurs
15.44
15.43
Spur dikes
15.44
Spurs
15.44
HYDRAIN
13.8
Hydraulic capacity
2.1
Hydraulic conductivity
4.2
Intrinsic permeability
Relative hydraulic conductivity
Relative permeability
15.47f
4.2
4.16
4.2
I.31
Index terms
Links
24.28
1.24f
2.14
3.9
3.10
3.9
20.17
Definition sketch
20.17f
Design procedure
20.17
Example design
20.19
21.5
Hydraulic radius
Hydraulic scale modeling
3.3
21.8
Kinematic similarity
21.8
24.26
13.6
13.12
13.13
13.12
13.6
Avoidance of hydroplaning
13.6
13.6
Rational method
13.6
13.13
13.10
Intensity-duration-frequency curves
13.8
13.9
21.7
21.8
Geometric similarity
Hydro compaction
22.18
13.12
13.9t
13.6
I.32
Index terms
Links
13.7t
Sheetflow
13.8
13.8
Time of concentration
13.7
Hydroplaning
13.2
I
Impact-type stilling basin, pipes and channel outlets
Basic design
Design guidelines
18.30
18.30f
18.30
Energy losses
18.31f
Width selection
18.31f
Infiltration
24.7
Capillary fringe
24.13
Clogging
24.10
24.2
Rates
24.7
Infiltration basins
24.11
Clogging of
24.10
24.26
Micro-clogging
24.25
Pilot basin
24.27
Wash-out wash-in
24.25
Infiltrometers
24.8
Buffered
24.8
Double ring
24.8
Single ring
24.8f
23.55
23.57
Biosparging
23.58
I.33
Index terms
Links
23.58
Compressors
23.58
Design parameters
23.57t
Design principles
23.56
Pilot tests
23.58
Process principles
23.56
Interception wells
Intrinsic soil permeability
4.10
4.2
J
Jet
5.28
5.28
Prime factors
Buoyancy flux
5.28
5.28
Plane jet
5.29
Simple jets
5.29
Mass flux
5.28
Momentum flux
5.28
5.29
5.30t
Joukowsky equation
12.2
12.3
K
Kariz
1.7
6.4
14.64
14.99
14.64
14.65
Linear approximation
14.64
I.34
Index terms
Links
14.64
14.65
14.99
ILSD-B2, 14.100t
14.101
ILSD-B3, 14.100t
14.103
SWMM-TRANSPORT, 14.100t
14.102
USGS, 14.100t
14.101
WASSP-SIM, 14.100t
14.103
3.3t
Kinetic similarity
21.8
L
Labyrinth spillway
17.30
Chute slope
17.30
Design procedure
17.30
Nappe oscillation
17.30
Subatmospheric pressures
17.30
Laminar flow
2.16
12.2
6.3
M
Macrodispersive flux
4.5
4.6
Equation for
4.6
Longitudinal macrodispersivity
4.6
Manifold
Manning's conveyance
8.28
19.45
I.35
Index terms
Manning's equation
Gutter flow
Manning's resistance coefficient
Links
3.11
3.15
10.11
13.10
16.4
13.14
21.7
14.5
13.14
3.11
Composite sections
16.39
Compound channels
3.15
3.13
19.35t
Floodplains
19.34t
19.36t
19.35t
Open channels
13.18t
13.9t
13.11t
13.15t
8.15t
13.19t
Vegetation
16.21
3.14t
8.1t
Tunnels
16.39
14.62
Excavated channels
Linings
19.34
16.22
6.4
5.10
4.5
4.5
Formation-scale dispersion
4.5
Longitudinal dispersivity
4.5
Macrodispersive flux
4.5
Spreading
4.5
Transverse dispersivity
4.5
I.36
Index terms
Molecular diffusion
Links
4.4
4.5
Fisk's law
4.4
In porous medium
4.5
Millington-Quirk equation
4.5
3.3t
Momentum equation
2.8
Monad kinetics
4.7
Montuori number
Morning glory spillway
Circular crest coefficients
3.28
17.22
17.24f
Coordinates
17.26
Design procedure
17.22
Flow conditions
17.23f
17.24f
Throat control
Muncipal water systems
17.25f
17.22
11.1
Muskingum-Cunge routing
14.104
14.103
N
Net positive suction head
10.20
Available
10.20
Cavitation
10.22
10.22f
Incipient cavitation
10.23
Major considerations
10.22
Required
10.21
10.22
10.23f
NETWORK
19.26
I.37
Index terms
Links
Newton-Rhapson method
9.8
2.8
6.5t
Non-Newtonian flow
22.80
Apparent viscosity
22.80
Critical velocities
22.81
Flow in suspension
22.81
Headloss computation
22.85
Examples
Turbulent flow
22.85
22.88f
Plastic flow
22.81
22.82
22.82
Reynold's number
22.81
Rigidity
22.81
22.83f
Suspension/laminar stage
22.82
Suspension/turbulent stage
22.82
Yield stress
22.81
Sewage sludges
22.84f
O
Open-channel flow computer programs
19.25
DAMBRK
19.26
FESWMS-2DH
19.25
FLDWAV
19.26
HEC-2
19.25
HEC-RAS
19.25
NETWORK
19.26
RMA-2
19.25
SWMM
19.26
I.38
Index terms
Links
19.26
UNET
19.26
WSPRO
19.25
WSP-2
19.25
3.1
Alternate depths
3.5
6.2
Bernoulli equation
Boundary shear stress
Mean
Bousinesq momentum flux correction coefficient
Celerity
14.8
16.2
16.24
6.4
16.4
14.19
Critical slope
3.1
Critical state
3.1
19.34
14.10
Chezy's law
Eddy loss
16.7
14.5
3.2
Courant criterion
16.4f
16.4
6.39
3.23
16.21
3.4
6.27
3.1
Flow resistance
16.3
Friction slope
3.22
14.5
3.1
21.4
Froude number
Full-dynamic wave equation
14.6
3.16
Hydraulic depth
3.3
21.4
I.39
Index terms
Links
3.4
Hydraulic jumps
3.9
Hydraulic radius
3.3
6.4
Kinematic-wave approximation
14.6
3.3t
6.3
Manning's law
6.5
19.34t
21.6
3.3t
Normal depth
21.6
16.2
14.6
Saint-Venant equations
14.6
Shield's diagram
Shield's regime diagram
16.7
16.24
6.19
6.25f
3.3
Specific momentum
3.4
Stage
3.4
Steady flow
3.4
Subcritical flow
16.3
Supercritical flow
16.3
Superelevation
3.4
Surcharge flow
14.8
Tractive force
19.36t
6.2
Specific energy
Top width
19.35t
3.3
21.7
6.4
Modular limit
Prismatic channel
21.6
21.4
3.4
16.2
16.7
16.24
I.40
Index terms
Links
6.97
Uniform flow
3.4
3.11
Unsteady flow
3.4
14.4
16.5
3.4
1.24
Constraints
1.30
Dynamic programming
1.30
Feasible solution
1.30
Linear programming
1.30
Nonlinear programming
1.30
Objective function
1.30
Orifice plate
21.12
Orifice spillway
17.20
Proportions of
17.20
Overfall spillways
17.13
17.15f
17.13
Hydraulic characteristics
17.13
17.16t
Plunge pools
17.13
17.14f
Scour hole
17.13
Trajectory
17.13
Overflow spillways
Abutment contraction coefficients
Crest pressures
17.21f
17.13
1.30
17.1
17.2
17.6f
Discharge coefficients
17.3
Hydraulic characteristics
17.1
17.7f
I.41
Index terms
Links
17.8f
17.2
17.5f
14.61
14.64
14.65
Linear approximation
14.64
Nonlinear approximation
14.64
14.65
14.62
14.61f
Time of concentration
14.66
Akan's equation
14.66
Izzard's equation
14.66
14.66
14.68t
14.67
P
Penstocks
8.21
8.22
Penstock branches
8.24
Thickness
8.22
8.23
Penstock manifolds
Perfect gas law
Permeability
8.27f
4.17
4.2
Intrinsic
4.2
Relative
4.2
Piezometric head
8.25f
4.1
9.1
8.26f
I.42
Index terms
Pipe flow
Links
2.18
9.12
9.7
Flow in parallel
2.18f
9.5
Flow in series
2.18f
9.5
Local losses
2.21
9.5
K factors
2.23t
Networks
9.11
Gradient method
9.20
9.11
9.17
Newton-Rhapson method
9.18
Unsteady flow
9.23
Pipeline models,
Modeling approach
2.4
Numerical models
2.2
Operational models
2.4
Planning models
2.4
12.3
12.4f
Hoop stress
12.3
Physical properties
12.3
Poisson ratio
12.3
Young's modulous
12.3
Pipe systems
Branching systems
Networks
9.5
9.7
9.11
Parallel systems
9.5
Series systems
9.5
22.6
I.43
Index terms
Plume
Links
5.28
Properties of
5.30t
Simple plumes
5.29
Plume hydraulics
5.28
4.2
8.1
Power intakes
8.4
Gate arrangement
8.13f
Horizontal intake
8.4
8.5f
Inclined intake
8.4
8.7f
Vertical intake
8.4
8.6f
Vortex formation
8.14
9.4
Pressurized pipeline
2.1
Pump definitions
10.2
10.2
Atmospheric head
10.2
Centrifugal point
10.2
Condition points
10.2
10.2
10.2
10.2
Specified point
10.2
Cutoff head
9.4
Datum
10.2
Elevation head
10.2
Friction head
10.2
Gauge head
10.2
Head
10.2
High-energy pump
10.2
Impeller balancing
10.2
I.44
Index terms
Links
10.2
Static balancing
10.2
Two-phase balancing
10.3
Overall efficiency
10.2
Power
10.3
Brake horsepower
10.3
Pump efficiency
10.4
10.3
10.3
Water horsepower
10.3
Pump pressures
10.4
10.4
10.4
10.4
Shutoff
10.4
Speed
10.4
Suction conditions
10.4
10.4
10.4
10.5
10.5
10.5
Total head
10.5
10.4
10.4
10.4
Velocity head
10.4
Pump characteristics
12.9
12.11
10.24
I.45
Index terms
Links
10.24
Minor losses
10.24t
Design point
12.11
Duty
12.11
12.17f
Negative rotation
12.16f
Positive rotation
12.15f
Suter diagram
12.17f
12.14
Nameplate
12.11
Rated conditions
12.11
Pump hydraulics
Hydraulics of valves
Impeller wear, effect of
Operation
10.9
10.13
10.16f
10.14
Affinity laws
10.14
In parallel
10.14
Homologous laws
10.14
Pipeline hydraulics
10.9
Colebrook-White equation
10.12
Darcy-Weisbach equation
10.12
Hazen-Williams equation
10.9
Manning's equation
10.11
Roughness factors
10.12
10.18f
9.3
10.8f
10.9
10.14
10.15f
10.16
10.18
System curves
10.9
I.46
Index terms
Links
10.14
10.18f
Pump performance
Homologous (affinity) laws
10.14
Dynamic similarity
12.10
Kinematic similarity
12.10
Power coefficient
12.1
Torque coefficient
12.11
Performance parameters
12.9
Flow coefficient
12.10
Flowrate
12.10
Head coefficient
12.10
Impeller diameter
12.9
Power coefficient
12.10
Rotational speed
12.9
12.10
12.10
12.11
12.11
12.13f
10.28
10.28
Steps
10.30
10.16
10.18
Pump standards
Pump station design
Hydraulic transients
10.1
10.32
10.37
10.37
Valve selection
10.37
I.47
Index terms
Links
10.34
Piping
10.34
Piping design
10.35
Criteria
10.36
Pressure design
10.35
Vacuum condition
10.36
10.32t
10.32f
1.7
1.8f
2.12
2.31
Q
Qanat
Quasi-steady flow
R
Radiation
5.3
5.4
5.4
5.3
5.4
Stephan-Boltzman law
5.4
3.26
Montuori number
3.28
Surges
3.26
Vedernikov number
3.28
Rational method
13.6
Assumptions
13.7
Intensity-duration-frequency curves
13.8
13.8
Rational formula
13.7
I.48
Index terms
Links
13.7t
Sheetflow
13.8
13.8
13.7
13.8t
Recharge basins,
Artificial
24.2
Induced
24.14
With clogging
24.11f
24.14f
Relative humidity
5.6
Relative permeability
4.4
Brooks-Corey model
4.4
Mualem model
4.4
4.4
Relative transport
4.6
Degradation
4.7
Equilibrium partitioning
4.6
Monad kinetics
4.7
Partitioning of chemicals
4.6
Sorption
4.6
Reliability
Anticipated load
1.31
7.3
Definition of
7.19
Design load
7.3
Equation for
7.2
Failure probability
7.2
Load-resistance interference
Diagram of
Performance functions
Probabilistic approaches
7.1
7.21
7.23f
7.19
7.3
I.49
Index terms
Links
Reliability (Continued)
Reliability formulas
7.22t
Reliability index
7.19
Reliability measures
7.3
Safety factor
7.3
Safety margin
7.3
Resistance
Reliability analysis methods
7.2
7.19
7.25
7.21
7.24
7.38
6.89
6.90
Exner equation
6.94
Governing equations
6.92
6.91
6.90
Theoretical considerations
6.89
2.14
22.81
12.7
Risk
1.31
Failure probability
Risk function accounting
Risk-based design
Annual expected flood damage cost
21.4
7.2
7.38
7.38
7.47
Conventional approach
7.47
7.49
7.49
22.80
I.50
Index terms
Links
7.48
7.46
7.45
7.45
Tangible costs
7.46
7.44
7.50
Risk-reliability evaluation
1.32
Composite risk
1.33
Failure
1.33
Load
1.32
1.34
Resistance
1.32
Risk assessment
1.33
Risk-based design
1.33
Safety factor
1.33
Roman aqueducts
1.20f
7.2
1.22f
S
Sadd-el-Kafara dam
1.13
Saint-Venant equations
3.22
Safety factor
1.33
Safety margin
7.3
9.26
5.6
7.3
15.68
BRI-STARS
15.69
GSTARS
15.68
HEC-6
15.68
I.51
Index terms
Links
15.69
HY9
15.69
RMA
15.69
SMS
15.69
Sediment bedforms
6.34
Antidunes
6.36
Bars
6.36
Bedform charts
6.42
Bedform classification
6.41
Bedform criteria
6.41
6.37
6.45
Bedform predictor
6.40
Bedform progression
6.37
Bedform types
6.39
6.39f
6.39
Dunes
6.34
6.44
6.52f
Form drag
6.47
Ripples
6.36
River stage
6.46
Schematic of bedforms
6.35
6.47
Einstein partition
6.47
6.50
Nelson-Smith partition
6.49
Skin friction
6.47
Stage-discharge relation
6.51
Brownlie method
6.53
I.52
Index terms
Links
6.51
Englund-Hansen method
6.52
6.26
ACRONYM
6.31
6.26
6.27
6.27
Einstein
6.28
6.28
6.28
6.28
Parker
6.28
Wilson
6.28
Yalin
6.28
6.26
6.30
Einstein number
6.27
6.26
Exner equation
6.27
Mixtures
6.27
6.29f
6.16
6.20
6.17f
6.17
Bank
6.20
Sloping bed
6.20
Streambed
6.17
Mobility of grain
6.16
Sediment mixtures
6.21
Shield's diagram
6.19
I.53
Index terms
Sediment properties
Links
6.6
Armor
6.14
Fall velocity
6.12
6.13f
6.10
6.10
Porosity
6.11
Rock types
6.6
6.9t
Sedimentation scale
6.7
Shape
6.12
Size
6.7
Size distribution
6.8
6.14f
6.16f
Specific gravity of sediments
6.7
Subpavement
6.14
6.12
6.55
Boundary conditions
6.55
6.60
Eddy diffusivity
6.57
6.66
Equilibrium suspension
6.57
6.7t
6.68
6.66f
6.58
Mass conservation
6.55
6.59
Prandtl analogy
6.57
6.59
6.15f
I.54
Index terms
Links
6.58
6.57f
6.63
6.60
6.64t
6.65t
6.74
Ackers-White relation
6.79
6.85f
6.86f
Brownlie relations
6.78
6.83f
6.84f
6.81f
6.82f
6.87f
6.88
6.79
Hydraulic resistance
6.78
Normal flow
6.79
Sediment transport
6.78
Dimensionless relations
6.74
Englund-Hansen relation
6.77
6.77
Hydraulic resistance
6.77
Normal flow
6.77
Sediment transport
6.77
Yang relation
Sediment transport
6.80
6.22
6.24
6.24
Transport modes
6.22
6.2
6.6
6.2
6.4
Chezy's law
6.4
6.4
Manning's law
6.5
6.4
I.55
Index terms
Links
6.5
6.3
6.3
Movable-bed roughness
6.5
6.5t
Shear velocity
6.2
Velocity distribution
6.2
24.30
Shaduf
1.4
1.6f
Shear
5.9
13.20
Shear stress
Bed shear stress
Boundary shear stress
13.20
6.2
16.29
16.29
16.32
16.29
16.29
15.50
Shear velocity
Sheetflow
6.2
3.16
Shield's diagram
6.19f
6.25f
Side-channel spillways
17.18
Hydraulic characteristics
Siphon spillway
13.2
17.18
17.32
Air-regulated siphon
17.34
Blackwater siphon
17.32
Design procedure
17.32
Standard siphon
17.32
24.34
17.33f
13.8
I.56
Index terms
Soil matrix suction
Soil vapor extraction
Links
4.3
23.34
Efficiency
23.39
23.35
23.50
23.49
Extraction pump/blower
23.45
Feasibility assessment
23.35
Injection blower
23.45
Instrumentation
23.46
23.39
Manifolding
23.46
23.36
23.38
Pilot tests
23.40
23.34
System design
23.42
Vapor treatment
23.45
Well construction
23.43
4.3
Specific energy
3.3
Specific momentum
3.9
17.40
Cavitation index
17.40
Cavitation characteristics
17.43
Design procedure
17.44
Spillway basins
18.24
Baffled apron
18.33
Pier heights
23.44f
17.41f
18.3
Baffle-block basin
Design procedure
23.41f
18.34f
18.34
18.35f
I.57
Index terms
Links
18.24
Basin I
18.7
18.3
18.8f
18.9
18.10f
18.13
18.13f
Design examples
18.4
18.24
18.30
18.17
Negative-step basin
18.24
Positive-step basin
18.24
Riprap
18.35
18.11
18.12f
Sloping aprons
18.19
18.20f
Design guidelines
18.19
Examples
18.22f
Jump lengths
18.21f
Structural arrangements
18.25
18.40
Velocities entering
Spillway chutes
18.17f
18.23f
18.6
17.38
Side-wall freeboard
17.38
Smooth chutes
17.38
Stepped chutes
17.38
Design procedure
17.38
Headless
17.40
17.49
17.39f
Spillway types,
Labyrinth
17.30
Morning glory
17.22
I.58
Index terms
Links
17.20
Overall
17.13
Side-channel
17.18
Siphon
17.32
Air-regulated
17.34
Standard siphon
17.32
Spillway chute
17.38
Smooth chute
17.38
Stepped chute
17.38
Tunnel spillway
17.36
Flat-tunnel
17.36
Flip-bucket
17.38
Inclined tunnel
17.36
Steady flow
Quasi-steady flow
Stephan-Boltzman law
Storm sewer design
2.11
2.13
2.12
2.31
5.4
14.22
14.22
Example design
14.24
Rational formula
14.24
Rational method
14.22
14.91
Classification of models
14.91
Dynamic-wave models
14.93
CARESDAS
14.97
DAGVL-A
14.94
HYDROWORKS
14.97
ISS
14.96
MOUSE
14.95
SURDYN
14.97
I.59
Index terms
Links
14.94t
SWMM-EXTRAN
14.93
UNSTDY
14.97
Noninertia models
14.98t
DAGVL-DIFF
14.98T
DHI
14.99
HVM
14.98T
MOUSE
14.98T
NISN
14.98T
14.99
ILSD-B2, 14.100t
14.101
ILSD-B3, 14.100t
14.103
SWMM-TRANSPORT, 14.100t
14.102
USGS, 14.100t
14.101
WASSP-SIM, 14.100t
14.103
14.30
14.33f
14.32
14.31f
Junction classification
14.30
Point junctions
14.44
Storage junctions
14.42
Flow equations
14.99
14.43t
14.34
Experimental studies
14.35t
14.38t
Headloss coefficients
14.36f
14.37f
14.34
14.37
Benching
14.42f
Experimental studies
14.40t
I.60
Index terms
Links
14.41
14.44
Bottleneck determination
14.56
14.56
14.50
Major characteristics
14.50
Procedure
14.52
14.91
14.45
14.47
Cascade method
14.47
14.47
14.48
14.48
Stratification of reservoirs
Annual cycle
5.7
5.14
5.7
Epiliminion
5.7
Hypoliminio
5.7
Metaliminion
5.12
Thermocline
5.7
5.12
19.23
19.23
First-order streams
19.23
Higher-order streams
19.23
19.23
19.25f
Simple channels
19.23
19.24f
Subcritical flow
13.17
Subcritical state
3.1
Surge
19.24f
2.12
This page has been reformatted by Knovel to provide easier navigation.
I.61
Index terms
Surge tanks
Links
8.17
Hydraulic stability
8.17
8.19
Maximum upsurge
8.19
Thoma formula
8.17
Surging
12.2
Supercritical flow
13.17
Supercritical state
3.1
Superelevation
Swape
3.16
8.2
1.4
T
TABS-MD
19.26
Tailrace channels
8.29
Tail-tunnels
8.28
Manifold
8.28
Outlet structures
8.29
Surge tanks
8.28
Tepidarium
8.32f
1.17
Thermal budget
5.3
Thoma formula
8.17
Tractive force
2.22
Time of concentration
13.7
Transients
2.11
2.36
Causes
2.34
Joukowsky relations
2.39
Kinetic energy
2.35
Physical nature
2.35
Wavespeed
2.38
Valves
2.40
13.18
I.62
Index terms
Transitions
Local loss coefficients
Links
2.23t
2.23
Transmissivity
4.8
Trashracks
8.9
Tower structures
8.9
8.10f
8.14
Manning's n values
Tunnel spillway
8.15
17.36
Entrance structures
17.36
17.36
Flip bucket
17.38
Inclined tunnels
17.36
Turbidity currents
Bed friction coefficient
6.97
6.102f
6.100
6.99f
6.104
Lake Superior
6.105
Plunging flow
6.100
6.103f
Shape factor
6.103f
6.101f
5.9
5.10
5.9
2.15
I.63
Index terms
Links
U
Uncertainties
1.31
Data uncertainties
1.31
Economic uncertainties
1.31
7.1
Hydraulic uncertainties
1.32
7.1
Hydrologic uncertainties
1.32
7.1
Model uncertainties
1.31
7.1
Parameter
1.31
Structural
1.31
Natural uncertainties
1.31
Structural uncertainties
1.31
7.4
7.4
Exponential transforms
7.4
Fourier transforms
7.4
Mellin transform
7.6
Approximate techniques
7.9
7.1
7.7t
7.9
7.16
7.12
19.26
Uniform flow
3.4
Chezy equation
3.11
Manning's equation
3.11
3.11
Sheetflow
3.16
Special cases
3.15
Superelevation
3.16
3.11
15.39
I.64
Index terms
Unsteady flow
Saint-Venant equations
Links
2.11
2.32
3.25
14.3
14.5
Darcy-Weisbach formula
14.5
Friction slope
14.5
Manning's formula
14.5
Preissman slot
14.8
14.6
Surcharge flow
14.8
14.11
Unsteady flow
14.4
14.13
Boundary conditions
14.20
Classification of flow
14.13f
Courant criteria
14.19
Discretization
14.19
14.14f
14.15f
Initial conditions
14.20
14.21
14.6f
Rating curves
14.18f
Single sewer
14.13
14.16
14.8
14.10
Surge speed
14.8
14.21t
19.7
3.24
I.65
Index terms
Links
V
Valves
11.11
11.47
11.47
Altitude valves
11.47
Angle valve
22.4
Ball valves
12.5f
Blow-offs
11.47
Butterfly valves
11.45
Check valves
22.4
12.5f
22.4
Lift check
22.4
Swing check
22.4
Spring check
22.4
22.4
Control valves
11.46
11.47
12.8
Gate valves
11.44
12.5f
Geometric characteristics
12.5
12.8f
Cross-sections of
12.5f
Globe valve
12.5f
Headlosses
12.6
Equation for
22.4
12.7
Hydraulic characteristics
12.8f
Isolation valves
11.44
Pinch/diaphragm
22.4
Plug valve
22.4
11.46
11.47
11.46
Manning's equation
22.3
22.4
Ball check
Cone valves
22.3
3.11
22.3
I.66
Index terms
Links
Valves (Continued)
Manning's resistance coefficient
Needle valve
3.11
12.5f
Throttling valves
22.4
22.4t
Valve closure
Classification of
12.7
12.9t
Valve operation
12.8
Vedernikov number
3.28
Venturi condition
Wastewater treatment plant
2.22f
22.1
Distribution boxes
22.2
Distribution channels
22.2
Flowmeters
22.4
Gates
22.3
Seating head
22.3
Unseating head
22.3
Pipe manifolds
Process flow diagram
Valves
22.4
22.2
22.40
22.3
Butterfly valves
22.3
Check valves
22.4
Gate valves
22.3
22.4
Throttling valves
22.4
W
Wastewater treatment plant hydraulic design
Aeration tanks
Cascade system
Configuration
22.36
22.53
22.77f
22.55
I.67
Index terms
Links
22.65
Distribution box
22.58
Effluent channel
22.64
Effluent weir
22.64
Inlet baffles
22.64
Inlet channel
22.58
22.58
22.53
22.53
Process configuration
22.53
22.53
Schematic diagram
22.65
22.53
22.53
Bar screens
22.41
Approach channel
22.41
Design example
22.43
22.43
Operating headloss
22.43
Schematic diagram
22.44
22.67f
22.45t
22.38
22.65
Backwash requirements
22.66
Collection systems
22.65
Design example
22.71
Filter operation
22.66
Grit tanks
22.74t
22.43
22.43
Design example
22.47
22.49t
I.68
Index terms
Links
22.43
Effluent weir
22.47
22.43
22.45
Hydroclone
22.43
Inlet baffle
22.45
Inlet channel
22.45
Vortex type
22.43
22.49f
Hydraulic profile
22.41
22.42f
22.71t
Design example
22.72
Design parameters
22.71
Diffusers
22.73
Submerged discharge
22.73
Planning phases
22.36
Plant layout
22.39
Process selection
22.38
22.77
Sedimentation tanks
22.47
22.78t
22.53
22.55t
22.53
22.60t
Effluent launder
22.52
22.48
Inlet baffle
22.48
Inlet channels
22.48
22.48
Outlet conditions
22.48
Overflow weir
22.52
Primary tanks
22.48
22.54f
22.48
22.59f
I.69
Index terms
Water budget
Links
5.1
5.2f
5.8
Dissolved solids
5.8
5.8
Temperature
5.8
2.18
9.22
Flow in parallel
2.18f
9.5
Flow in series
2.18f
9.5
Local losses
2.21
9.5
Networks
9.11
Gradient method
9.20
9.11
9.17
Newton-Rhapson method
9.18
Unsteady flow
9.23
9.31
Application
9.32
Calibration
9.32
Calibration process
9.32
EPANET
9.33
History of
11.9
Model development
11.10
Model selection
9.32
Network representation
9.32
Problem definition
9.32
Software packages
11.10
Verification
9.32
9.26
Conservation of energy
11.9
11.11t
9.3
I.70
Index terms
Links
9.3
9.2
Hydraulic head
9.1
Loading condition
9.1
Hardy-Cross method
9.11
Piezometric head
9.1
9.4
Pump curve
9.4
11.34
Flexible pipe
11.37
Internal pressures
11.34
Loads
11.34
Rigid pipes
11.36
Thrust restraints
11.38
Preliminary design
11.12
Alignment
11.12
Rights-of-way
11.13
Subsurface conflicts
11.12
11.13
11.31
11.13
11.29
11.18
11.25
Steel pipe
11.20
11.1
11.2
Fire demands
11.7
11.2
I.71
Index terms
Links
11.8
Peaking factors
11.2
11.2
11.4
Service pressures
11.8
Storage
11.7
Emergency storage
11.7
Fire storage
11.7
Operational storage
11.7
Supply
11.7
Water demands
11.3t
11.9t
11.9t
Water duties
11.2
11.5t
Water hammer
2.12
12.1
Case studies
12.24
Definition of
12.2
Entrapped air
12.2
Joukowsky equation
12.2
2.39
12.2
Mitigation of
12.1
Time constants
12.24
12.24
12.24
12.24
12.24
Water properties
12.3
2.5t
Water reuse,
Constructed aquifers
24.39
Potable reuse
24.40
Restricted irrigation
24.33
I.72
Index terms
Links
24.33
22.1
Distribution boxes
22.2
Distribution channels
22.2
Flowmeters
22.4
Gates
22.3
Seating head
22.3
Unseating head
22.3
Pipe manifolds
Process flow diagram
Valves
22.2
22.40
22.3
Butterfly valves
22.3
Check valves
22.4
Gate valves
22.3
22.4
Throttling valves
22.4
22.10
22.12
Design flows
22.12
22.13
Maximum demand
22.13
Minimum demand
22.13
Hydraulic profile
22.3f
22.13
22.13
Plant layout
22.12
22.15
Plant siting
22.12
22.10
Head available
22.18
22.10
I.73
Index terms
Links
22.10t
22.13
22.13
22.15
Coagulation
22.19
Design example
22.21
Flocculation
22.20
Flocculation/sedimentation basin
22.16f
Hydraulic mixing
22.20
Pneumatic mixing
22.19
Rapid mixing
22.19
Sedimentation
22.20
Filtration
22.26
22.28t
22.26
Process criteria
22.26
Membranes
Biological fouling
22.32t
22.34
22.35
Configuration
22.35f
Hyperfiltration
22.34
Hyperfilters
22.34
Microfilters
22.34
22.34
Nanofilters
22.34
Reverse osmosis
22.34
Scaling control
22.34
22.34
22.36
Ultrafilters
22.34
Water use
22.14t
11.5t
I.74
Index terms
Wavespeed
Wave suppressors
Raft-type
Links
2.37
18.26
18.27f
Underpass-type
18.28
Weisbach coefficient
14.66f
Well duplets
4.12
5.14
WSP-2
19.35
WSPRO
15.39
18.29f
19.25
Z
Zangar's equation
24.30