moovvaall ffrroom

Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Arsenic Removal from OU Water

C
Chheem
miiccaall E
Ennggiinneeeerriinngg-- U
Unniivveerrssiittyy ooff O
Okkllaahhoom
maa

R
Roossss CChhaaffffiinn
RRaannddyy G
Goollll
SSaam
mii KKaarraam m
R
Roommaann V Voorroonnoovv

SSpprriinngg 22000033

Page 1 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Executive Summary

In 2000, the EPA mandated that the arsenic level for water distribution systems must be reduced to
10 ppb (parts per billion) by the year 2006. The results of our Senior Capstone Project show how the
University of Oklahoma can save several million dollars over a long term horizon, by using
treatment methods, and only purchasing water from the City of Norman occasionally at peak
demands. In order to comply with the new regulation, our project recommends the construction of an
Ion Exchange (IX) treatment facility as opposed to water purchase from the City of Norman.

The Net Present Cost (NPC) of purchasing water for the next 20 years, that is the total cost at 2003
dollars, is estimated to be approximately $8,000,000 at the current water price of $1.14/1000gals. In
comparison, the NPC of building a treatment facility and treating the water for Arsenic has only a Net
Present Cost of approximately $3,000,000, for the same project lifetime. In fact, in order for the Water
Purchase (WP) option to become more economically attractive than treatment, the water price charged
by the City of Norman would have to be lower than $0.45/1000gals. A price so low could not be
offered by the City, since it costs Norman roughly $0.60/1000 gals to produce it.

Assuming that the University of Oklahoma borrows money to sponsor the project, the money saved
in the first year alone by building an IX treatment facility instead of purchasing water would amount to
$150,000. Furthermore, the savings would increase until the loan is paid off and would roughly double
after the first ten years, when no more payments are made. While the initial capital investment required
for the Ion Exchange facility is substantial, the savings made by implementation of treatment versus
water purchase begin in the first year of the project and greatly outweigh the inconvenience of
borrowing money.

In addition to the substantial economic benefits provided by treatment, construction of a treatment

facility would allow the University of Oklahoma to remain independent from the City of Norman, as
Norman faces water shortages of its own; and may be forced to purchase water on an emergency basis
from Oklahoma City.

Conclusion: Over a twenty-year period, the total savings provided by the construction of a treatment
facility will be in excess of 4.8 million dollars. At the current commercial price of $1.14 per 1000
gallons, the yearly cost of water purchase will be near $500,000, and will increase with the water
demand of the University.

Page 2 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

1.1 Table of Contents

1.3 Table of Figures 4

1.4 Table of Common Abbreviations 5

2. Introduction
2.1 Arsenic Law 7
2.2 OU Arsenic Situation 7
2.3 Water Demand Data 10
2.4 Arsenic and Ion Exchange Chemistry 11

3. Work Previously Done

3.1 Water Treatment Options 14
3.2 Water Purchase 15
3.3 CH2M Hill Report 15
3.4 CE 5244 Arsenic Report 17

4. Treatment Analysis
4.1 Overview 19
4.2 Ion Exchange 20
4.3 Coagulation/Filtration 24
4.4 Nanofiltration 26
4.5 Comparison of Previous Conclusions 31

5. Detailed Ion Exchange Model

5.1 P&ID 33
5.2 Variations with Capacity 34
5.3 Safety 35

6. Planning Model, Results and Recommendations

6.1 Planning Model 39
6.2 Model Results 42
6.3 Financial Analysis 44
6.4 Risk Analysis 47
6.5 Recommendations 50

7. Supporting Documentation
7.1 References 51
7.2 Appendices 52

Page 3 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

1.2 Table of Figures

Tables

2.1 OU Well Data 8

3.1 Plan alternative cost summary – p. 6-16 of CH2M Hill Report 16
3.2 Estimated costs for the Ferric Chloride coagulation 17
4.1 O&M Costs, CF Plant 25
4.2 Nanofiltration Facility Present Worth Results for a 20 yr Project Lifetime 30
6.1 Plant Costs Used in Model 41
6.2 Deterministic Results 42

Figures

2.1 Norman and Campus well field 8

2.2 Arsenic Concentration Gradient 9
2.3 Wells and Piping Layout 10
2.4 Demand By Month 11

4.1 Targeting of Number of Columns 21

4.2 Ion Exchange Capital Costs 22
4.3 Ion Exchange O&M Costs 23
4.4 Capital Costs for Coagulation/Filtration 24
4.5 EPA Preliminary TCI Graph for CF 26
4.6 Nanofiltration PFD 27
4.7 Capital Costs for Nanofiltration 29
4.8 Capital Costs for Nanofiltration, Worst Case 30
4.9 NPC Comparison of Previous Solutions 31

5.1 P&ID 33
5.2 Capital Investment vs. Capacity 34
5.3 Operating Cost vs. Capacity 35
5.4 Bed Volumes treated vs. Breakthrough 36

6.1 Water Consumption Projection 41

6.2 Purchased Water over project lifetime 43
6.3 Loan and Repayment 43
6.4 Yearly Cost with Loan 44
6.5 Yearly Cost without Loan 45
6.6 Savings 45
6.7 NPC at different Prices of Norman Water 46
6.8 Sensitivity of NPC to Water Cost 47
6.9 Sensitivity of NPC to Maximum Field Capacity 48
6.10 Sensitivity of NPC to Capacity Unit Cost 49
6.11 NPC Probability Distribution 50

Page 4 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

1.3 Table of Common Abbreviations

As Arsenic (including all forms)

CE Civil Engineering
cf Cubic Feet
CF Coagulation and Filtration
CI Capital Investment
EPA Environmental Protection Agency
gal Gallon
gpm gallons per minute
IX Ion Exchange
kWh Kilowatt Hours
MCL Maximum Containment Level
MGD Million Gallons per Day
NF Nanofiltration
NPC Net Present Cost
O&M Operating and Maintenance
OC Operating Cost
OKC Oklahoma City
OU The University of Oklahoma
ppb Parts Per Billion (µg / L)
ppm Parts Per Million (mg / L)
USGS United States Geological Survey
WP Water Purchase

Page 5 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Introduction

• Arsenic Law
• OU Arsenic Situation
• Water Demand Data
• Arsenic and Ion Exchange Chemistry

Page 6 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

2.1 Arsenic Law____________________________________________

Under the 1996 amendments to the Safe Drinking Water Act, the EPA reviewed the acceptable arsenic
levels in drinking water. As of January 22, 2001, the previous levels of 50 ppb will be replaced by a 0 ppb
non-enforceable and 10 ppb enforceable levels of Arsenic in all public drinking water systems effective as of
January 23, 2006.

The reason for this recent change from the old 50 ppb rule, which was established in 1975, is that a
March 1999 report by the National Academy of Sciences concluded that the current standard does not
achieve EPA's goal of protecting public health.

Recent studies have linked long-term exposure to Arsenic in drinking water to cancer of the bladder,
lungs, skin, kidney, nasal passages, liver, and prostate. Non-cancer effects of ingesting Arsenic include
cardiovascular, pulmonary, immunological, neurological, and endocrine (e.g., diabetes) effects. Short-term
exposure to high doses of Arsenic can cause other adverse health effects that are reversible in some cases.
(Reference #6).

2.2 OU Arsenic Situation____________________________________

2.2.1 Well Data

With the new Arsenic rule, the OU campus is now faced with the ultimatum to find a solution for its
Arsenic problem. Most of the problem wells in the Norman area are located on campus property and are
used by the University of Oklahoma to supply on-campus buildings with potable water. With an average
Arsenic content of 35 ppb for all the wells being presently used, this automatically makes the whole system
non-compliant with the 2006 deadline.

Norman is located atop of what is known as the “Garber-Wellington” field. This water aquifer is one of
the main sources of drinking water used by the city of Norman and the OU campus. When the Navy
donated all the buildings and water wells to help in the formation to what is now present day OU,
technology wasn’t available to detect the high levels of Arsenic present in the wells drilled.

Figure 2.1 is a USGS map that shows the locations of all of the OU potable water wells. All of the wells
are in or around the Westheimer Airpark/Research Facility, which is the location of the old Navy base
nestled between I-35 and highway 77, just north of Robinson Street.

Page 7 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Figure # 2.1 OU Campus well field/Westheimer Airpark

OU currently operates 11 of those wells (FIGURE 2.1) to feed the Norman campus with potable water.
The OU well data table (TABLE 2.1) illustrates the flow rate variations for each well; this is due to variation
in water consumption on campus. The wells are operated in an automated fashion that regulates the flow
accordingly. Also included in that table are the Arsenic levels observed in each well.

Table#2.1 OU individual well data (2000-2002)

Average Capacity Maximum Average Arsenic

Well #
(g/min) Capacity (g/min) (ppb)

2 110 125 94.1

3 120 200 68.7
4 140 150 30.7
5 --- --- ---
6 160 200 14.8
7 120 160 21.8
8 180 200 22.4
9 120 165 66.2
12 160 200 94.7
13 120 160 32.9
14 150 160 33.9

AVG: 1380 1720 48.02

Page 8 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

The Arsenic average in TABLE #2.1 (48.02 ppb) is not representative of the published average (35ppb)
because readings taken at well source and averaged are different than readings taken at the delivery pipeline
after all well water has been mixed. Automation cause different wells to be operated at different rates thus
affecting the final Arsenic content in the delivered water.

2.2.2 Geological Data

The “Gerber-Wellington” well field is located atop a geological formation that favors the precipitation
of Arsenic in the water aquifer. The underground water moves in South-West current, which explains why
Arsenic concentrations increase in that direction. This is illustrated in the concentration gradient graph
(FIGURE 2.2).

Figure#2.2: Arsenic Concentration Gradient “Gerber-Wellington” Field

2.2.3 Current Piping Layout

Since the new wells can only be drilled on University-owned land, the spatial restrictions of the
drilling locations limit the options available to the University of Oklahoma. Any wells drilled outside
campus property belong to the City of Norman.

All the operational wells are situated in the Westheimer Airpark North of Robinson Street along HWY
77. A piping network connects them all to one main pipeline going down Berry Street towards the main
campus as illustrated in FIGURE 2.3. Some wells (like well #9) are no longer on OU property.

Page 9 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Figure#2.3: Westheimer Airpark/Research Facility, Wells & Piping Layout

Water Collection Pipelines

Well site

As it is seen from Figure #2.3, the geographical restrictions are extremely limiting. This plays an
important part in narrowing down the options available for the OU campus, since the University does not
have the option of simply drilling new wells in the northeast region of the “Gerber Wellington” field.

2.3 Water Demand Data

Though this project is not specifically concerned with the water demand at OU, any analysis of a water
treatment facility should include the parameters of the demand that should be met. OU’s maximum water
consumption peaks toward the spring semester when residence halls and overall demand increases due to
the heat. Currently, peak consumption reaches 1.5 million gallons per day. The maximum capacity of the
water grid is around 2.0 million gallons per day, due to water tower limitations.

Page 10 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

2002 Demand by Month

1600

1400

1200
Demand (1000 gal/day)

1000

800

600

400

200

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure #2.4 Water Demand by month

2.4 Arsenic and Ion Exchange Chemistry

Arsenic is found naturally occurring in ground water; and can be found in two forms, Arsenic (III, or
arsenate, and Arsenic (V), or arsenate. Arsenite (H3AsO3) occurs predominantly in the non-ionic form
when the pH of water is less than 8.01. When the pH is above 9.0, arsenate (HAsO42-) in the ionic form is
more prevalent in water.

The average pH in the OU wells is 8.97, which means that the arsenic in the water is naturally occurring
in the non-ionic arsenite state and the ionic arsenate state. The ion exchange process will not work
effectively unless the arsenic is in a charged state. Arsenic (III) must be oxidized to arsenic (V). Sodium
hypochlorite (NaOCl) will be added for the oxidation reaction to occur. Below is the oxidation reaction that
forms arsenate. The added NaOCl will cause the pH of the water to go down, so sodium hydroxide
(NaOH) is added after the treatment to bring the pH to normal standards.

H3AsO3 + NaOCl ->HAsO42- + NaCl + 2H+ (oxidation reaction arsenite to arsenate)

1
www.hdrinc.com/corporate/JumpStories/WQTCPoster.pdf

Page 11 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

After arsenite is oxidized, arsenate is formed; the water with arsenate anions will be run through the
resin bed. The resin bed is made up of polymeric beads that have been pre-saturated with chloride anions
on their outer layer. When the water is run through the bed, arsenate ions adsorb to the resin beads; and
trades places with the chloride ions, hence the name ion exchange. Below is the reaction where arsenate
binds to the resin.

2RCl + HAsO42- -> R2HAsO4 + 2Cl- (reaction where arsenate binds to the resin)

As can be seen, the chloride ion (Cl-) is bound to the polymeric bead (R). The polymeric bead has a
higher selectivity for arsenate, and knocks the chlorine ion from the polymer.

The resin beads exchange with other ions also. In fact the selectivity for sulfate (SO42-) is higher than
the selectivity for arsenate. Consequently, the sulfate can overtake the bed causing the arsenic ions to be
released back into the water if the ion exchange bed is not monitored properly. If OU chooses to go
through with ion exchange, the arsenic content, although still very important, will no longer be the most
important measurement taken. Sulfate content will need to be monitored to ensure that sulfate is not
overtaking the exchanging of the arsenate ions.

After the process is complete, the bed must be regenerated to remove the arsenate ions from the resin
bed. The bed is then regenerated with concentrated sodium chloride (NaCl) in the reaction as follows.

R2HAsO4 +2NaCl -> 2RCl + Na2HAsO4 (bed regeneration reaction)

The regeneration is driven by concentration. The sodium chloride has such a high concentration the
arsenate ion is knocked off of the resin beads, and the chloride ions are then replaced to the beads. After
this process, the concentrated NaCl with the arsenate ions is sent to a precipitation tank. Sulfuric acid is
added to this stream to bring the pH down, so ferric chloride (FeCl3) can be added to precipitate ferric
hydroxide Fe(OH)3. The arsenic finally adsorbs to the Fe(OH)3, is dried, and sent to a landfill.

Page 12 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Work Previously Done

• Water Treatment Options

• Water Purchase
• CH2M Hill Report
• CE 5244 Arsenic Report Findings

Page 13 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

3.1 Water Treatment Options

Ion Exchange (IX): Simple IX on a basic media is an attractive solution for arsenic removal, due to the
simplicity of the process. Water is run through a resin and Arsenic replaces a nontoxic ion on the resin,
resulting in purified water downstream of the resin. After all adsorption sites are filled, the bed must be
returned to its initial state through regeneration via pH adjustment and exchanging the arsenic ion on the
resin with the initial, non-hazardous ion.

Infinite Brine Recycle: One possible way to avoid consistent replacement of the exchange resin is by
regeneration process via concentrated sulfate solution. The ‘brine’ may be reused, due to its high
concentration and selectivity, and the As may actually remain in the brine for several cycles.

Activated Alumina (AA): Ion Exchange involving activated alumina is highly selective to arsenate ions.
Cost is also very comparable to Ion Exchange. However, three basic problems limit the use of activated
alumina: rapidly diminishing capacity, narrow 5.5-6 pH operation range, hazardous waste with costly
disposal. AA appears most promising for small communities, especially those with a high sulfate
concentration that would make regular IX difficult.

Reverse Osmosis (RO): RO is a membrane separation method allowing the removal of the finest
particles, even ions. The membranes employed by this method have pores on the order of 1 A& and are
operated in Cross Flow mode. Osmotic pressure must be overcome and permeate flow is driven by high
operating pressures (200-400psig).

Nanofiltration (NF): NF is a membrane separation method that has proven to achieve Arsenic
rejection efficiencies in the range of 80-90%. The membranes employed by this method use the principle of
RO to filter out Arsenic and are operated in Cross Flow mode. NF membranes have pores on the order of
1nm.

Point of Entry (POE) and Point of Use (POU): POE/POU technology is a strategy that suggests
treating only the potable water instead of treating all of the water via a centralized water treatment plant.

Coagulation/Filtration (CF): Ferric Chloride reacts with arsenic in water to produce Ferric
Hydroxide, and then forms a precipitate. The water is then run through a very large bed of sand and iron
fillings where the arsenic is precipitated and filtered out of the water. The large bed of sand and iron that is
used as a filter will last for several years.

Lime Softening: Calcium hydroxide is first added to drinking water until the drinking water reaches a
pH of 10, and then sodium carbonate is used to cause the arsenic to precipitate out of the water.
Consequently, Arsenic is removed via simple filtration. Toxicity of the chemicals used in Lime Softening
present a safety issue to its usage.

Electrodialysis Reversal (ER): Salt in water is dissociated and then passes through ion-selective
membranes. Usually used for water desalination, it was found that an 80% efficiency in removing arsenic
can be achieved using this method. As with RO, this process generates a waste stream (20-25% of inflow)
highly concentrated, which means that part of the treated water has to be disposed of, creating a problem
for water starved regions.

Page 14 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

3.2 Water Purchase Option__________________________________

To date the University of Oklahoma does not have a solution to the lowering of Arsenic level limits in
2006. The only solution at this point is buying water from the City of Norman. Buying water from
Norman may not be the least expensive option available to OU. Currently, it is unclear how much Norman
will sell the water for.

The City of Norman currently sells water to OU at a rate of $1.14/1000 gallons. Additionally, OU is
negotiating a deal with Norman involving sewers, waste, and water purchase to buy water at a rate of
$0.85/1000 gallons of water. The rate of $0.85/1000 gallons is the minimum suggested price that the City
of Norman could charge for its water, considering that it spends $0.60/1000 gallons to produce it2.

The current peak water usage for OU in August is about 1.8 million gallons per day of water. For the
economic analysis 1.1 million gallons of water was used. Variations in monthly water usage and growth of
the university will be taken into account. At $1.14/1000 gallons and 1.1 million gallons per day the
university will spend $475,000 per year and $5,565,000 over the next 20 years. At $0.85/1000 gallons the
university will spend about $354,000 per year and $4,150,000 over the next 20 years.

3.3 CH2M Hill Report

In 2002, the City of Norman and the University of Oklahoma hired CH2M Hill, an engineering firm
based in Colorado, to do an analysis of the water supply in the Norman area; and find the best way to
become compliant with the 2006 MCL of 10 ppb. The report treated the well field of Norman and OU as
one entity, and the solution given was the best for the community as a whole. The report is public domain,
and available upon request from the Norman Public Library.

3.3.1 Options Analyzed

Initially, treatments were analyzed separately from the general options, with Ion Exchange,
Nanofiltration, Coagulation/Filtration, and Granular Ferric Hydroxide among the examined. Norman city
officials examined the five treatments assuming that treatment would be used on a large scale, for all of the
contaminated water in the Norman area. Using both economic and social factors, such as waste, Granular
Ferric Hydroxide was chosen as the “best” technology.

The report then analyzed five options: water purchase from Oklahoma City, blending with water
purchase, zonal blending with the creation of new wells, treatment, and a combination of treatment,
blending, and new well creation.

2
Source: Don Carter, OU Physical Plant

Page 15 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Table #3.1 Plan alternative cost summary – p. 6-16 of CH2M Hill Report
Plan Alternative Capital Cost Annual O&M Present Worth
Plan A – Water Purchase $0.7 $4.1 $47
Plan B – Blending and Water
$3.9 $2.8 $37
Purchase
Plan C – New Wells and
$9.2 $1.3 $24
Blending
Plan D - Treatment $17.2 $2.2 $42
Plan E – Treatment,
$12.4 $1.7 $32
Blending and New Wells
Costs in millions of dollars

Plan C was recommended by the report because it provides the following3:

• Independence from relying on OKC treated water

• Non-treatment options for achieving compliance
• Development of non-potable uses for non-compliant groundwater wells, decreasing
demand on OU and City water systems
• Lowest estimated present worth cost

The report recommends the drilling of fifteen new wells over the next 4 years, located in the far
northeastern part of Norman, where current wells display a level of Arsenic below the 2006 MCL.

3.3.2 Limitations

The report seeks – and gives – a feasible solution for the Norman area. However, the report assumes
that the University of Oklahoma and City of Norman operate on the same water system. This is not the
case. The University currently produces its own water. Because the University does not have the land for
drilling new wells, the adoption of the CH2M Hill report’s recommendation would likely cause the
University to have to purchase water from the City of Norman.

When applied to the University alone, only reason two, of the four reasons given above, still holds as
advantages to the chosen plan. The plan will cause the University to have to rely on Norman treated water,
instead of remaining independent. The University already has its own non-potable water system4, used for
irrigation and power supply. Of the five options, water purchase from Oklahoma City had the highest Net
Present Cost. It follows that if OU were to have to purchase water from Norman, the cost may also be
high.

Though the CH2M Hill report is the most rigorous examination of the Arsenic problem in Norman to
date, its recommendation does not apply strictly to the University of Oklahoma. A better option for MCL
compliance may exist.

3
CH2M Hill Arsenic Report, p. 6-16
4
Source: Don Carter, OU Physical Plant

Page 16 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

3.4 CE 5244 Arsenic Report Findings______________________ ____

An Arsenic study was submitted by a group of graduate civil engineering students at the University of
Oklahoma College of Civil Engineering and Environmental Science, CE 5244 Water and Wastewater
Treatment Course. The Civil Engineering report was submitted on April 22, 2001, to the City of Norman
under the title of “Arsenic Removal in Groundwater”.

3.4.1 Recommended treatment (Coagulation)

Ferric Chloride Coagulation followed by micro-filtration was recommended by the Civil Engineering
study to be chosen as the solution to the Arsenic problem for the University of Oklahoma. Table 3.2 is
summary of the costs associated with the proposed design processes recommended by the study:

Table #3.2: Estimated costs for the Ferric Chloride coagulation

Design Waste
Project
Flow Capital Disposal Operational Waste Disposal
lifetime
Rate Cost ($) Capital Cost ($/yr) ($/yr)
(yr)
(MGD) ($)
4,000,000 500,000 90,000 120,000
20 3
Total 4,500,000 210,000

Note – The actual report is available at http://www.soonercity.ou.edu/ce5244/index.htm

Among the considered options for the University of Oklahoma, the Ferric Chloride Coagulation water
treatment option was found to be most economical solution. However, it required high initial capital costs.

3.4.2 Drilling New Wells

Drilling new deep wells, drilling shallow wells in the Canadian River alluvia, and purchasing water from
the City of Norman were three other alternatives considered by the study. The options involving drilling of
new wells were rejected on the basis of performed geological research. According to the study, drilling new
wells is associated with high risk regarding the concentration of Arsenic in the wells.

3.4.3 Buying from Norman

Purchasing water from the City of Norman was also not recommended due to high long term costs,
estimated at $1.8 million per year. Moreover, the study mentions a potential water shortage for the City of
Norman in the near future, which would force the city to purchase water from elsewhere. This would result
in an even higher purchase price of water for the University of Oklahoma.

Page 17 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Treatment Analysis

• Overview
• Ion Exchange
• Coagulation Filtration
• Nanofiltration

Page 18 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.1 Overview

The Environmental Protection Agency gives a list of possible technologies for the removal of Arsenic
from water. These are detailed in the appendix. After examining these technologies for immediate
applicability to the OU water supply in the next several years, Ion Exchange, Coagulation/Filtration using
ferric chloride, and Nanofiltration were chosen for further study. The following treatments were not
investigated in detail:

4.1.1 Activated Alumina

Activated alumina was not chosen for scale up for 3 simple reasons: Longevity, operating pH, and
waste disposal. Activated alumina is extremely selective to arsenate ions, but only for a short amount of
time. It was decided that a process that had a long life span was necessary. The operating pH for activated
alumina is very low. Although this process is very similar to ion exchange, it needs a pH adjustment to
lower the pH before the process; and a pH adjustment to raise the pH after the process. This would cause
an extra cost that may not be needed for other options such as ion exchange. Waste disposal for activated
alumina would be very costly for the simple fact that the waste would be considered hazardous. Longevity,
operating pH, and waste are the primary reasons for discounting activated alumina.

4.1.2 Lime Softening

Lime softening requires large amounts of calcium hydroxide to be added to raise the pH to 10, and large
amounts of sodium carbonate to lower the pH to normal standards. An operator also needs to monitor this
system 24 hr/day, because if either of these chemicals get out of control, they can be toxic. Adding the high
costs of calcium hydroxide, sodium carbonate, and labor it becomes very obvious that this solution is not
economically feasible.

4.1.3 Electrodialysis Reversal

Electrodialysis reversal involves water being pumped through a membrane. These membranes have a
very large capital cost, and the waste streams are very large.

4.1.4 Reverse Osmosis

Reverse osmosis was not used because of the large energy cost that would be cause from the small pores
in the membrane. Nanofiltration would be more fitting for this problem because it has a much larger pore
diameter.

4.1.5 Pilot Plant Technologies

To date there are several Arsenic removal technologies that are still in developmental stages. Some
involve surfactants, membranes (Ultrafiltration), or even a constant ion exchange recycle process (Infinite
Brine Recycle). These may, in the future, turn out to be potential Arsenic removal options. Unfortunately
they have not yet reached full-scale models. It was decided not to consider such options, because the
Arsenic problem at the University of Oklahoma calls for a fully developed treatment process with existing
designs and infrastructure.

Page 19 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.1.6 Point of Entry (POE) and Point of Use (POU)

Point of Entry (POE) and Point of Use (POU) technology is not an actual large scale treatment, but is
instead a different approach to drinking water sanitation. Specifically, it is a strategy that suggests treating
only the water that is really "important" instead of treating all of the water via a centralized water treatment
plant. Therefore, POE/POU strategies are not considered as a possible solution to the Arsenic problem at
the University of Oklahoma, by this study. However, it is recommended to explore these strategies in future
studies as potential optimization techniques for the water treatment at OU.

4.1.7 Blending with Norman Water

Though not a treatment, it is commonly recommended that Norman’s water simply be blended with
water from OU’s existing well field, minimizing water purchased while requiring no capital investment.
However, EPA regulations require all water entering a distribution system to be under the MCL. Also, due
to the high concentration of arsenic in OU’s water, very little could be used, meaning that the water
purchase price would not be reduced that greatly, anyway.

4.2 Ion Exchange

The design for an Ion Exchange facility is broken up into many parts. The facility will be housed in a
building just north of Robinson Street, between Flood and Berry, on OU’s airport campus. All designs will
be based on the CH2M Hill design capacity of 2.08 million gallons per day. This value is approximately
30% higher than the maximum OU demand, but with OU growing in population and water demand
annually, using this value near to the maximum flow of OU’s well system is prudent.

4.2.1 Booster Pumping

In addition to the pumps already in place at the well sites, pumps must be added to make up for the loss
of head caused by the Ion Exchange system. They must be high capacity. The pumps must handle 1448
gallons per minute at full capacity, and provide a 35 ft head. Providing three 750 gpm pumps in parallel will
do the trick. These pumps are estimated to cost $24,000 each, with a 50% installation premium. Annually,
the pumps require almost 50,000 kWh, easily the highest energy demand in the entire system. This is also
the stage where oxidizing chemicals will be added in very small quantities to adjust the pH and oxidize the
Arsenic (III) to Arsenic (V), which is more easily removed. Chemical costs are covered in a different
section.

4.2.2 Influent Straining

The treatment process calls for a small straining process, to eliminate particulate contaminants before
entering the Ion Exchange columns. Any large particulate matter will contribute to clogging the column or
fouling the resin, reducing efficiency in either case. The cost is small; however, the one 500 micron self-
cleaning strainer costs $17,500.

Page 20 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.2.3 Ion Exchange Facilities

An Ion Exchange column with a diameter of seven feet was chosen using the targeting procedure
shown in Figure 4.1: (Costs are detailed in a supporting spreadsheet)

900000
800000
700000
600000
Total Cost
Cost ($)

500000
Vessel Cost
400000
Resin Cost
300000
200000
100000
0
0 2 4 6 8 10 12
Diameter (ft)

Figure #4.1: Targeting Of Number Of Columns

At seven feet, the resin volume needed and total number of vessels shows a minimum cost. At this
value, four vessels are needed, each with a cost of $70,000. Resin is expensive, at $110 per cubic foot of
packing. A control system should be installed to keep flow uniform in the columns.

Two important practical matters should be discussed. First, there is no reason that this process should
not be readily expandable. Placing valves and connections in such a way that another column could be
added if demand exceeds the process capacity is prudent, even though it is unlikely that the well field could
support more than 2 million gallons per day. Second, with these columns in parallel, the capability exists to
exchange the resin and regenerate the columns individually. Thus, the water supply to OU will not be cut
off the several times a day that regeneration is required.

4.2.4 Brine Regeneration

A brine of Sulfuric Acid is used to strip the resin of the adsorbed Arsenic and return the column to its
default state. The brine must be stored, pumped through the column, and disposed of safely to the filter
press. Pumps, a mixer and a tank must be purchased. This setup, complete with a 2,100 gallon storage
tank, will most likely take up the most space in the facility.

Page 21 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.2.5 Dewatering Facility

The waste will be disposed of as a dry sludge. The concentrated brine must be pressed to remove liquid
(which will be disposed of in the sewer system). Cost will primarily be for the filter press.

4.2.6 Chemical Storage

Iron (III) Chloride and Sodium Hydroxide are used to prepare the feed for Ion Exchange (primarily in
oxidizing the small amount of arsenite to arsenate). A much higher quantity of Sulfuric Acid and Sodium
Hydroxide salt is required for column regeneration. Feed pumps and storage tanks are required for all three
chemicals.

4.2.7 Land and Building

It is estimated that 2,600 square feet of area will be needed. At a crude estimate of $100/square foot,
the process will cost $260,000 to house. The University of Oklahoma already owns the land, which will be
used for the facility.

21000 Chemical Feed

Capital Cost Breakdown Facility
130000 Filter Press

535640 Brine Maker

Brine Regen.
402178 Facility
IX Facility

Influent Straining

34000 Booster Pumping

20000 Allow ances

49200 Contengency
311181
103888 260000 Building

Figure #4.2: Breakdown of Ion Exchange Capital Costs

Operating Costs

4.2.8 Power

Power is required for all mixers, computer systems, pumps, etc. At 60,000 kWh of power per year,
power costs will be near $4,200.

Page 22 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.2.9 Chemicals

The three chemicals discussed before, plus brine salt, must be acquired regularly. At current prices of
$59 per ton of salt, $0.10 per pound of sulfuric acid, $0.13 per pound of sodium hydroxide, and $0.15 per
pound of Iron (III) Chloride, the total chemical cost per year for the operation will be near $60,000, a cost
that is almost entirely made up of the salt and sulfuric acid cost. If the level of sulfate in the water is
significantly higher than estimated, the price to regenerate the column will also become significantly higher,
as more frequent regeneration cycles will become necessary.

4.2.10 Labor

Labor estimations allow for annual equipment equal to 1% of the capital cost, an operator who spends
40% of their time on the process, and a manager for whom the process is 10% of their responsibilities.
Total labor costs, therefore, are close to $35,000. Maintenance and maintenance labor are given as a
percentage of total capital cost.

4.2.11 Resin Replacement

The resin must be replaced every five years. Even with frequent regeneration cycles, the resin will
slowly lose its ability to adsorb Arsenic. The cost of all new resin, roughly $50,000, is linearized over a five-
year period, adding a $10,000 per year cost.

Breakdown of Annual Costs

10120 4144

18671 Power
36664
Chemicals (not Salt)
Salt
Residual Disposal
Labor
16500
Maintenance
Resin Replacement

23904
20759

Figure #4.3: Breakdown of Annual O&M Costs

Page 23 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.3 Coagulation/Filtration

4.3.1 Background

Coagulation/Filtration is used to remove Arsenic from drinking water by adding a coagulant that
precipitates out of water while causing the Arsenic to adsorb to the precipitate. The precipitate is then
filtered or gravity drains from the water. The coagulants that are normally used are iron or aluminum ionic
compounds. These compounds are added to the water in the liquid form. They then hydrolyze into an
insoluble salt that precipitates out and is filtered from the water. The two types of metallic compounds that
are used, aluminum and iron, form aluminum and iron hydroxides that are both insoluble in water. The
aluminum or iron hydroxide is what the Arsenic molecule adsorbs to. The hydroxide precipitate is then
filtered, and discarded in a landfill as long as the Arsenic filled sludge that is produced is dry and has a
concentration of less than 5 mg/L.

4.3.2 Costs
The estimated Present Worth Analysis for the coagulation filtration system was about $5.741 million.
This figure includes initial capital costs to implement this process, and operating and maintenance costs over
the first twenty-year period of the implementation of this process. This analysis was done assuming an
average production of 2.26 MGD with a design capacity of 2.79 MGD. Today OU has an average water
consumption of 1.1 MGD with a peak summer time capacity of 1.5 MGD. This is much less than what was
used for the economic analysis. This was done because of expected growth in the university over the next
20 years. However if an analysis were done at the present flow rates the Present Worth Analysis of this
process would be much less than that of buying water, but would have a great deal of inaccuracy from not
taking into account growth of the university. Below is a graph of the total capital costs, and a table of the
operating and maintenance costs for this process. These numbers are largely taken from the CH2M Hill
Report for the City of Norman.
Pumping Costs
Capital Costs for Coagulation/Filtration

Rapid Mixing
$175,000
Filtration
$560,000 $18,000

Thickener

Dewatering
$1,146,600
$480,000
Chemical Feed System

Building
$400,000
$220,000 Piping, I&C, Electrical, Yard
$270,000 Piping Allowances
$101,000
Contingency 20%

Figure #4.4: CF Capital Costs

Page 24 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Table #4.1: Operating and Maintenance

Costs
Reason for Cost Annual Cost
Power Cost $14,000
FeCl3 Cost $7,700
CO2 Cost $57,500
Labor Cost $16,500
Equipment Cost $33,700
Disposal Cost $3,000
Total O&M Costs $132,000

4.3.3 Explanation of Costs

The filtration facility is the most expensive of the capital costs. This facility takes up $1.1 million of the
estimated $3.4 million of the entire investment. This is a very high because of the elaborateness of the
filtration facility. This high cost is not desired, but it is necessary because of the high flow rates and large of
amount of precipitate that will be filtered. The next highest capital cost was the 20 % contingency that was
implemented to provide an overestimate for this process.

Since Coagulation/Filtration produces a fair amount of waste, there is a price of disposing. There is also
a need of large pumps to pump water through the system. A vessel to mix the water and ferric chloride is
also needed. This vessel contains an impeller to stir the ferric chloride and the water. The water needs a
slight pH adjustment to ensure the highest efficiency possible. Carbon dioxide is added to the water for this
adjustment. Costs for storage of carbon dioxide, along with ferric chloride need to be taken into account.
A building of about 4,000 square feet is needed to house this facility (CH2M Hill Report). Power and
operating & maintenance costs need to be taken into account too.

4.3.4 Analysis

For preliminary cost estimations of total capital costs and operating & maintenance costs an EPA
economic analysis from, Technologies and Costs for Removal of Arsenic from Drinking Water, was used. This graph
calculates total capital cost by using the appropriate equation for the desired design flow rate in million
gallons per day. The total capital cost (y) is found from the equation y = 1030810x + 1067733 (1 < x < 10
MGD), where x is the flow rate in million gallons per day (see figure 4.5 below). The average design flow
rate used in the CH2M Hill report was 2.26 MGD. From this equation 2.26 MGD gives a total capital cost
of $3,397,000. This is not far off from the CH2M Hill report of $3,371,000. The EPA correlation is meant
to be a rough estimate, and the CH2M Hill report is supposed to be very near accuracy. The fact that these
values are within a range of ± $20,000 is evidence that these estimations may be close to accuracy.

Page 25 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Figure #4.5: EPA Preliminary Total Capital Investment Estimation Graph

According to the CH2M Hill report, the total capital cost consists of pumping facilities, mixing facilities,
filtration facilities, thickener facilities, dewatering facilities, chemical storage, I & C electrical, and building
facilities. The pumping facilities consist of 3 pumps with a capacity of 724 gpm. The mixing facilities have
a mixing vessel of 65 ft3, and an impeller of 1.81 ft. in diameter. The filtration facilities have 6 pressure
filters with a vessel diameter of 12 ft. The 4,000-ft.2 building cost close to $400,000. These items bring the
total capital cost to an estimated $3,371,000. Using the O & M costs for 20 years, and the total capital costs
give a Present Worth Analysis of $5.7 million.

4.4 Nanofiltration

Due to a Present Worth larger than that of the other treatment options and due to an uncertainty range
of almost $13,200,000 present in the economic evaluations, the Nanofiltration treatment is not
recommended as a treatment option to the University of Oklahoma. As a result of profitability evaluation
of a Nanofiltration facility the Present Worth for the best and the worst case scenario were calculated to be
$4,100,000 and $17,300,000, respectively. During the economic evaluation it was determined that the
Present Worth is largely a function of membrane cost. The Present Worth of an Ion Exchange treatment
option was used with the membrane, pumping and utility costs specific to the NF processing facility. Costs
specific to Ion Exchange were not considered. The manufacturer quoted membrane costs. Pumping and
utility costs were obtained from a Pro-II simulation and literature correlations.

Page 26 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.4.1 Technology Background

Figure #4.6 illustrates a process flow diagram of a Nanofiltration water treatment facility:

Figure #4.6: Nanofiltration Water Treatment Facility PFD. 5

Nanofiltration (NF) is a membrane separation method that has proven to achieve Arsenic rejection
efficiencies in the range of 80-90%. The membranes employed by this method use the principle of Reverse
Osmosis to filter out Arsenic and are operated in Cross Flow mode (see Appendix A). Larger membrane
pores and lower utility requirements set Nanofiltration apart from Reverse Osmosis. Low feed recovery
rates and high capital cost are a major concern in NF design. A Nanofiltration facility consists of the
membranes themselves, pumps, piping and the controls with instrumentation needed for operation.

Nanofiltration presents an appealing option for the University of Oklahoma because at the given
conditions it can lower the Arsenic content in water from an average of 30ppb to below 1.0 ppb.
Nanofiltration would allow OU to meet the legal Arsenic standards, to be implemented in 2006, by
producing water with constant quality and also by removing other impurities. The NF process is simple,
fully automated and requires little monitoring. The only chemicals requiring storage for the NF treatment
facility are those needed for membrane cleaning, unless it is determined that pretreatment is needed.
Another appealing quality of NF treatment is that no onsite residual processing is required. The concentrate
would be sent directly to the City of Norman water treatment plant, which has processes to treat Arsenic.
5
http://www.membranes.com/docs/trc/flowcon.pdf

Page 27 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.4.2 Economic Analysis

Although, the annual operating and maintenance costs of Nanofiltration are generally minimal, this
separation method is plagued with very high capital costs. A major part of the capital cost is the equipment
associated with Nanofiltration, with the membranes themselves being the most expensive piece of
equipment (approximately 15-40% of the capital cost). Therefore, it is assumed that the major differences
in Present Value of Nanofiltration and that of other methods come from membrane, pumping and
electricity costs. In order to calculate the Present Value for an NF facility, the economic evaluation of an
Ion Exchange facility was used with appropriate changes made to three previously mentioned costs.

4.4.3 Membrane Cost

Initially, an extensive membrane price research was performed and the membrane with the highest
permeate flowrate per dollar of price ratio was isolated from the pool of candidates. A Hydranautics model
ESNA1 membrane turned out to have the largest permeate to cost ratio of 13 gal/$. Since low feed
recovery is a big concern in Nanofiltration design for scarce water regions (see Appendix A), the ESNA 1
presented the optimal choice of membrane for the University of Oklahoma. According to the
manufacturer, a NF facility consisting of ESNA 1 membranes would achieve 95% Arsenic rejection with a
75-80% feed recovery. The total number of membrane units needed for an average production permeate
flow of 2.6 MGD was calculated to be 205 units. The price quoted by the manufacturer for the total
number of membranes required is approximately $172,000. The lifetime of the membranes was taken to be
4 years, as quoted from the manufacturer. Osmotic and operating pressures of 1.37-4psi and 76psi,
respectively, were calculated for the ESNA 1 membrane.

4.4.4 Pumping Costs

The work required for a pump in order to raise the feed to 76 psi was estimated by running a Pro-II
simulation, assuming 70% pump efficiency. The work was determined to be 142kW. Electricity price used
for utility costs was 7 cents per kW hour, which is the same price that was used to evaluate other treatment
options. Pump cost for a centrifugal pump was estimated from Peters and Timmerhaus to be $20,000.

4.4.5 Facility Costs

All costs specific to the Ion Exchange process were removed from the Present Value evaluation, except
for the chemical and chemical storage costs. These costs were assumed to be the same as what was
determined for Ion Exchange by CH2M Hill. This is a good assumption because Nanofiltration would
require less expensive chemicals, in smaller amounts than what would be needed for Ion Exchange.
Therefore, the NF Present Value would not be in danger of underestimation due to this assumption.

Page 28 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Figure 4.7 presents the capital cost break down for the NF facility using the Hydranautics ESNA 1
membrane:

Capital Cost Breakdown For NF Treatment Facility

(Best Case Scenario)
$104,000,
10% $250,000,
23% Pumping

Membrane&Installation

Chemical Feed Facility

$241,000, Building
23%
Contingency

Allow ances

$233,000,
22%
$50,000, 5%

$178,000,
17%

Figure #4.7: Capital Cost Breakdown for NF Treatment Facility for the Best Case Scenario.

In order to observe the effect of the membrane cost on the Present Value, the procedure above was
repeated with the most expensive membrane available. An AMI model M-N2521A membrane turned out
to have the lowest permeate to cost ratio of 1.33 gal/$. A total of 9040 units would be need with an overall
cost of $1,695,000. The life time of the membranes was taken to be 4 years. Figure #4.8 presents the
capital cost break down for the NF facility using the AMI M-N2521A membrane. From comparison of
Figures #4.7 and #4.8, it is apparent that the membrane cost is always one of the major components of the
capital cost.

Page 29 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Capital Cost Breakdown For NF Treatment Facility

(Wost Case Scenario)
$2,373,000,
57%
Pumping

Membrane&Installation
$104,000, 3%
Chemical Feed Facility

Building $250,000, 6%
Contingency
$233,000, 6%
Allow ances

$50,000, 1%

$1,137,000,
27%

Figure #4.8: Capital Cost Breakdown for NF Treatment Facility for The Worst Case Scenario.

4.4.6 Conclusion
It is apparent from the best and the worst case scenarios for the Present Worth that the membrane cost
makes an enormous difference on the profitability of the NF treatment process. This is largely due to the
capital cost of the membranes and their annualized replacement. Table 4.2 is a summary of the Present
Worth results for the best and worst case scenario of the Nanofiltration facility with a project lifetime of 20
years:
Table #4.2 Nanofiltration Facility Present Worth Results for a 20 yr Project Lifetime
Membrane Type Present Value, $ Uncertainty, $
Hydranautics model ESNA1 4,100,000
13,200,00
AMI model M-N2521A 17,300,000

Due to the uncertainty range of $13,200,000, it is not recommended to choose Nanofiltration as the
treatment option for the University of Oklahoma. A chlorine-based pretreatment will most likely be
required in order to convert Arsenite to Arsenate. Chlorine destroys many of the NF membranes available,
including the best-case scenario membrane. Therefore, even though a more detailed analysis might reveal a
Present Worth that is slightly lower than what was estimated in this study, it is highly unlikely. In fact it is
likely to be much higher than the best-case scenario. This treatment option should be investigated in as
much detail as possible, if its utilization is to be considered in the future.

Page 30 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

4.5 Comparison of Previous Conclusions

Compiling the findings of both the CH2M-Hill report and the CE 5244 Group, a comparison can be
made about how the Net Present Cost compares to the finding in this report. Attention should be taken to
the fact that this report’s findings were based solely on using wells already owned and operated by OU.

Previous Conclusions
$16,000,000

$14,000,000
Net Present Cost

$12,000,000

$10,000,000

$8,000,000

$6,000,000

$4,000,000

$2,000,000

$0
WP IX C/F NF
CH2M Hill CE Group Our Group

Figure #4.9: NPC Comparison of Previous Solutions (at different plant capacities)

As is apparent from Figure 4.9, investigating the OU situation independently from the City of Norman
may result in significant savings.
Two of the three proposed treating processes, Ion Exchange and Coagulation Micro-filtration, are more
economically feasible than simply becoming a customer of the City of Norman. According to the
preliminary findings, choosing to build an Ion Exchange treatment facility is the cheapest option available to
the University. Based on this analysis, the Ion Exchange design from section 4.2 will be further specified
and applied to Norman.
It is important to note that the estimates performed by all three groups were made at different plant
capacities: Our Group=1.8MGD, CH2M Hill=2MGD, CE Group=3MGD. However, the graph reveals a
very important point – at the current OU maximum well capacity (used by Our Group), Ion Exchange is the
most economically favorable treatment option. It is also important to note that even though the capacities
were different, for different groups, the NPC values were relatively close. In fact, if the estimates for other
groups were to be rescaled to the plant capacity used in this study, the NPC values would agree much more
favorably.

Page 31 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Detailed Ion Exchange Model

• P&ID
• Specifications
• Variations with Capacity
• Safety

Page 32 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

5.1 P&ID

Figure 5.1: P&ID of Ion Exchange Treatment Facility

5.1.1 Regular Operation

Regular operation of the Ion Exchange plant is given by the solid blue lines. Under regular operation,
water will travel through a sand filter, then mixed with sodium hipochlorite to oxidize the arsenite to
arsenate, making it possible to remove the compound in the ion exchange column. The water then runs
through the primary columns. Sampling ports are set up to measure concentration of sulfate and arsenic, so
that the column may be ‘timed’, and the system assured of no breakthrough of the extra “safety column”.

5.1.2 Regeneration

To remove arsenic and sulfate from the IX resin, replacing it with the original chloride ion, a column is
shut off, then concentrated NaCl brine is run through the column backwards. This allows the resin to shift,
allowing the resin to age evenly. The brine, with sulfate and arsenic in solution, is then mixed with Ferric
Chloride, which causes the arsenate to precipitate out. The waste sludge is then dried and sent to a landfill.
Periodically, sulfuric acid is introduced to act as a cleaning agent for the resin. Regeneration is shown by the
dashed orange lines.

Page 33 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

5.1.3 Controls

Only very simple controls are needed. First, to ensure that NaOH is not wasted, a pH meter determines
how much the pH is adjusted. Second, the level of the brine settling tank is monitored. Green lines
represent the control system.

5.2 Variations with Capacity

The capital investment required for an ion exchange plant has both fixed and variable costs. For the
purposes of a model that is explained later, all variable costs are assumed to vary linearly with respect to the
capacity of the plant. Of the major costs of the facility, only column capacity (and thus the associated
allowances and contingency) is particularly sensitive to changes in the flow rate. A zero capacity portion of
the plant (everything except the ion exchange columns) is still need, and regeneration and chemical feed
facilities must still exist. The cost of this “capacity-less” portion of the plant is 1.1 million dollars.

The plant discussed in section 4.2 had a capacity of 2.08 MGD and a Fixed Capital investment of 2.05
million dollars. This gives the relationship between capital investment and capacity seen in figure 5.2. The
analysis shows that the capital investment increases $450,000 per MGD of capacity.

Capital Investment and capacity

Capital Investment ($MM)

2.5

1.5

1 y = 0.4567x + 1.1

0.5

0
0 0.5 1 1.5 2 2.5
Capacity (MGD)

Figure 5.2: Capital Investment vs. Capacity

Page 34 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

A similar analysis is done with the operating cost. In this case, the labor and maintenance costs for a zero-
capacity plant are the “fixed” values, while the rest (power for pumping, chemicals used, etc) depend directly
on the capacity. Therefore, at zero capacity, the operating cost is near $35,000 per year, while at 2.08 MGD,
the operating cost is 130,000. Figure 5.3 shows the relationship between operating cost and capacity. The
operating cost is increased $43,000 for every MGD of capacity.

Operating Cost and Capacity

140000
Operating Cost ($)

120000
100000
80000
y = 45673x + 35000
60000
40000
20000
0
0 0.5 1 1.5 2 2.5
Capacity (MGD)

Figure 5.3: Capital Investment vs. Capacity

5.3 Safety

5.3.1 Arsenic Spiking

The Ion Exchange process must be over-designed to ensure that all adsorption sites in the column are
never filled. The exchange media is more selective towards sulfate in the water than it is towards Arsenic
compounds. Thus, because sulfate is more plentiful, more sites are actually used to adsorb sulfate than
Arsenic. As the sites begin to fill, the sulfate begins to replace Arsenic on the exchange media. This actually
causes the effluent stream to “peak” in Arsenic, as all the adsorbed Arsenic is pushed from the media by
sulfate at once. The Arsenic level, after mixing in OU’s water tower, will most likely return to pre-treatment
levels at the Point of Use. Figure #5.4 is taken from sulfate breakthrough on a laboratory scale.

Page 35 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Figure #5.4: Pilot experiment on sulfate breakthrough as related to Arsenic peaking.

(http://www.hdrinc.com/engineering/topics/ionexchange.htm)

All care must be taken to ensure that regeneration cycles take place before Arsenic breaks through,
rendering the entire process worthless. It is therefore necessary to sacrifice cost in the form of chemical
cost by reducing total bed volumes between regeneration in order to ensure no breakthrough.

5.3.2 Chemical Safety and Storage

Iron Chloride: Iron Chloride is a light green to yellow crystalline powder at normal conditions. It does
not have an odor and is non-flammable. Iron Chloride is severe irritant and is corrosive upon physical
contact. Chronic effects of exposure are iron poisoning, liver damage, cytogenic and reproductive effects.
Iron Chloride should be stored in a tightly sealed container away from incompatible substances. The
container should be kept in a cool, dry, ventilated area.

Sodium Hydroxide: Sodium Hydroxide has an appearance of white, deliquescent pellets or flakes
under normal conditions. It is an odorless, nonflammable, poisonous substance. Sodium Hydroxide is
severe irritant and is corrosive upon physical contact. Its destructive effects upon tissue include blindness
and severe burns. The Permissible Exposure Limit is 2mg/m3. Scarring of tissue and death are possible
upon ingestion. Protective coating should be worn and vent hoods should be used when handling Sodium
Hydroxide. Sodium Hydroxide should be stored in a tightly sealed container away from sources of heat,
moisture and incompatibles.

Sulfuric Acid: Sulfuric Acid is a white powder, with no strong odor. It is commonly dissolved in
water, in which case it appears as a clear liquid. Burns and tissue necrosis are possible upon contact.
Inhalation may be fatal. Prolonged or repeated inhalation may cause kidney and lung damage. Protective
goggles, gloves and clothing should be worn when handling Sulfuric acid. Sulfuric acid should be stored in a
cool, dry, well-ventilated area away from incompatible substances and combustible materials.

Material Safety Data Sheets are found in Appendix C.

Page 36 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Any arsenic treatment plant that has on-site regeneration facilities (i.e.: Ion Exchange or Activated
Alumina) is classified as a large quantity hazardous waste generator (LQG) if waste brine generated exceeds
1000 kg/month. The handling of these arsenic laden residuals must comply with the provisions of Resource
Conservation and Recovery Act (RCRA).

Arsenic is considered to be at toxic levels if it is found in liquid, sludge and solids with a solid
concentration of more than 0.5% (CFR, Part 261.24).

An LQG, such as in our case, accumulating hazardous waste on site must:

- Comply with the preparedness and prevention procedures (CFR, part 265, Subpart C)
- Develop and maintain a contingency plan on site
- Comply with personnel training requirements
- Before shipping hazardous waste to an RCRA facility, an LQG must comply with several pre-
treatment requirements: EPA ID number, preparing a Uniform Hazardous Waste Manifest
- Comply with Department of Transportation (DOT) requirements

5.3.3 Disposal Options for Arsenic Contaminated Residuals

Waste generated will consist of solid, liquid and sludge:

- Allowable discharge of arsenic contaminated waste liquid will be affected by the ability of the
receiving stream to assimilate the arsenic without exceeding the arsenic standard of the receiving
water
- Disposal of liquid waste to a water treatment facility, generated sludge has to meet the following
requirements: Land disposal 73 mg/kg
Land application (clean sludge) 41 mg/kg

An option to discharge the waste brine into a receiving stream is highly unlikely due to high
concentration of arsenic and sulfates. Disposal into a sanitary sewer is an option considering it meets the
TBLL requirements (Technically Based Local Limits).

Treatment of the brine by adding Ferric Chloride will greatly reduce arsenic concentration. Precipitate
can be dried and taken to municipal waste since limited data shows that it will pass TCLP (Toxicity
Characteristic Leaching Procedure) test. Then with sufficient arsenic removed from the waste brine, disposal
into a sanitary sewer is possible considering the high salt content is acceptable locally.

Page 37 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Planning Model, Results and Recommendations

• Planning Model
• Model Results
• Financial Analysis
• Risk Analysis
• Recommendations

Page 38 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

6.1 Planning Model

6.1.1 Background

The preliminary calculations for the Water Purchase option were based on the current $1.14/1000gal
price of water and the average daily demand of 1.1 MGD. The net present cost for 20 years was calculated
to be $5.6 million, based on these figures. In order to make an economic comparison, the net present cost
for the Ion Exchange option was calculated making similar assumptions that demand, plant capacity and
operating costs remain constant. The net present cost of Ion Exchange was estimated to be $3.6 million.
Although the preliminary estimates show the Ion Exchange option to be more economically attractive, the
previously listed assumptions do not represent well what occurs in the real world.

The water demand for the University of Oklahoma varies from month to month, displaying patterns of
increased demand in the hot months of the school year and decreased demand in the winter and mid-
summer months. Additionally, the water demand displays yearly patterns of growth due to increasing
student body of the University and campus expansions. Other important parameters that introduce a
significant degree of uncertainty to the problem are variation of water prices, operating costs for the IX
facility, unforeseen changes in the well field, inflation, depreciation and the possibility of additions to the
existing plant.

It quickly becomes obvious that the magnitude and complexity of an economic comparison, more
accurate than what was done in the preliminary calculations, quickly gets out of hand and is not feasible to
perform without the use of technological tools. The need for greater accuracy calls for a computer based
simulation that uses a mathematical model in order to describe the Arsenic situation at the University of
Oklahoma. Such high performance numerical simulations are becoming widely used for financial planning
and are already considered to be mainstream engineering tools in the industry.

6.1.2 Planning Model Description

A mathematical model was developed in order to compare the net present cost of building an Ion
Exchange facility to the purchase of water from the City of Norman option. The optimal solution for the
Arsenic problem is determined by the planning model based on the economic comparison results. The
model was solved using a programming interface called Gams® (www.gams.com), which is a high-level
modeling system for mathematical programming problems.

The main objective of the mathematical model is to describe mathematically the Arsenic situation at the
University of Oklahoma and find the most economically attractive solution by minimizing the expected cost,
for a project life time of 20 years:

Total Expected Cost = ∑ p s C s where C is scenario cost, p is probability, and s is scenario.

s
Two types of models were solved : a deterministic model based on one scenario with constant parameters
and a stochastic model based on many scenarios with varying parameters. Aside from minimizing the total
expected cost the mathematical model must:

a) Choose between building a Treatment Plant or Purchase

b) Meet the Water Demand

Page 39 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

c) Decide When/How Much To Build

d) Expand Capacity as Needed
e) Buy Wholesale on Emergency
f) Borrow/Repay Money

6.1.3 Detailed Explanation

a) The model has the capability of choosing between building a treatment facility or purchasing water
from the City of Norman. This means that the model can choose to not ever build a plant and
purchase water all of the time, or vice versa. The optimum choice between the two options is made
based on the option that would give the least total cost for the project lifetime. The model also
explores the possibilities of combining the two options (see part b).
b) One of the constraints used in the model is that it has to satisfy the projected water demand by any
means necessary. This means that the model explores the possibilities of building plants of different
capacities, and then either add-on to the existing plant to increase water production or purchase
water from the City of Norman, all this while minimizing the total cost of the project.
c) If the model makes a decision to build a treatment facility, it has the flexibility of choosing when to
build and what the capacity of the plant should be.
d) The model has the ability to upgrade the existing facility in order to increase the water production
capacity. The upgrade option is available to the model’s discretion in every year, after year one.
e) The price of the water purchased depends on the amount that it purchases. A wholesale or a higher,
emergency-based price is charged for water purchase, based on the amount of water that the model
decides is necessary to purchase.
f) Finally, the mathematical model has the finance equations necessary to consider borrowing money
from a bank in case the specified budget is too small to cover the expenditures. The model
calculates the optimum time to borrow money and the best way to pay off the loan within the
project lifetime. Time value of money, as well as price inflation, are taken into account.

For a complete set of equations used in the mathematical model and a copy of the GAMS program, see
Appendix F.

6.1.4 Parameters

A project life time of 20 years was used in the model, with varying monthly water demand. The water
demand was projected into the future based on the historical consumption data for the University of
Oklahoma and based on 25% growth, over the project lifetime, estimated by the OU Physical plant.
Correlations for the capital investment and operating costs for the Ion Exchange treatment plant were
developed based on the preliminary calculations performed in the section 4.2. Both capital investment and
operating cost correlations depended on a fixed, “must pay” fee and a plant capacity based fee. Table #6.1
summarizes the plant costs used in the Mathematical Model:

Table #6.1 Plant Costs used in the Mathematical Model.

CI ($) CI ($/1000gal) OP ($/yr) OP ($/(yr 1000gal))
563,900.00 635.70 38,000.00 3.72

Page 40 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

The maximum well capacity was determined to be 1.8MGD, based on the existing well data provided by
the OU Physical Plant. A depreciation rate of 6% and an interest rate of 9% were assumed for the financial
purposes of the model. The $1.14/1000gal water price, currently charged by the City of Norman, was used
as the wholesale price that would be charged if the amount of water purchased by the model averaged to be
over 1,000 gallons per month. Conversely, if the model purchased water sporadically, it charges an
Emergency-based price of $3.00/1000 gal. A 2% inflation rate was applied to the water price throughout
the project lifetime. A $500,000 per year budget was specified for the model.

6.1.5 OU Water Consumption Projection

According to the OU physical plant, the demand for water on the OU campus is expected to increase
25% over the next 20 years. The horizon for this increase is shown in Figure 6.1:

W a te r C o n s u m p tio n P r o je c tio n

1800
Average Daily Consumption (1000 gal

1700

1600

1500 Average
C ons um ption
1400
Augus t (P eak)
1300 C ons um ption

1200

1100

1000
0 5 10 15 20
Y e ar

Figure 6.1: Water Consumption Projection

The error bars in figure 6.1 represent the standard deviation, as we predict that the water demand could
vary as much as 5% in either direction. This consumption projection is used later in the report, in order to
determine the future water demands upon the Westheimer airport field. It is important to note that the
black line gives the horizon for average daily water consumption over a year, but in the peak month of
August, the consumption is about 190,000 gallons per day higher.

Page 41 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

6.1.6 Constraints

Physical Constraints:
9 Maximum Capacity - The maximum water production capacity of the treatment plant must be less
than or equal to the maximum production capacity of the existing OU wells.
9 Total Capacity – Total capacity for a given year must equal to the sum of the capacities of the
existing plant and the previously built additions.
9 Meeting Demand – Total capacity for a given year must be greater than or equal to the demand in
any month of the year.
Financial Constraints:
9 Water price - Wholesale price must be charged if the amount of water purchased by the model
averaged to be over 1,000 gallons per month. Otherwise, the Emergency-based price must be
applied.
9 Budget – The total cost for any given year cannot exceed the budget for that year.
9 Debt Payment – The minimum amount of money repaid to the bank must be greater than or equal
to twice the interested accumulated from last year’s debt
9 Repay Loan – The debt in the final year of the project must equal to zero.

6.2 Model Results

6.2.1 Deterministic Model Results:

As a result of the deterministic simulation, the mathematical model produced an optimal solution for the
Arsenic problem at the University of Oklahoma. The model decided to make a one time capital investment,
in order to build a large capacity treatment facility. The model determined that building a plant in year one
of the project lifetime instead of choosing to purchase water would save the most amount of money for the
University of Oklahoma. The detailed results produced by the model are described in Table #6.2:
Table #6.2 Deterministic model results.
# of
Total Cost($) Plant Capacity(MGD) Area(ft2) Labor($/yr) Column Dia (ft)
Columns
2,600,000 1.6 2,600 20,000 4 6

Figure #6.2 displays the water purchased by the model during the peak times of the projected demand.
It can be seen from the figure that if a 1.6 MGD treatment facility is built in the first year of the project, the
University of Oklahoma will not begin purchasing water until approximately the second half of the project
lifetime.

Page 42 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Purchased Water Chart

.6
Purchased Water

.4
(MGD)

Purchased
.2 Water

.0
8
11
14
15
17
18
19
20
Project Lifetime (yr)

Figure #6.2: Purchased Water over lifetime

It is also important to note that the model chose not to build any add-ons to the treatment facility.
Purchasing water on Emergency basis turned out to be more economical.

In order to finance the project the model decided to take out a loan and repay it in the first five years of
the project lifetime. Figure #6.3 displays the financial planning chart produced by the deterministic model.

140 000 0
120 000 0
100 000 0
80 000 0
60 000 0
40 000 0
20 000 0
0
-20 000 0 1 2 3 4 5 6 7
-40 000 0

Figure #6.3: Loan repayment

The loan was repaid in equal amounts of approximately $300,000 per year, for five years and the
remaining sum was repaid in year seven of the project. It is important to note that this result assumes a
budget of near $400,000 per year. This is not meant to represent the way that OU will pay for the capital
investment of the plant, but instead gives one example of how it is possible, and the low price compared to
the water purchase cost of at least $450,000 a year.

Page 43 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

6.3 Financial Analysis

The results obtained from the mathematical model are shown in Figures 6.4-6.5 in terms of the cost per
year under both water purchase and treatment conditions. Figure 6.4, assumes that the capital cost of an
Ion Exchange facility is supplied by a loan taken out in year one, and paid off over the next ten years, in
equal payments of $350,000 per year. When compared to water purchase at $1.14 per 1000 gallons from
Norman, the difference in costs of the two options is apparent in year one of the project lifetime. Initially
the difference in costs is on the order of $150,000, but as time goes on the water purchase option becomes
increasingly more and more expensive, with yearly costs growing to as much as $800,000 in the second half
of the lifetime. The cost of the Ion Exchange treatment option remains constant at $350,000 per year until
the loan is paid off. When no more payments are made, the cost of treatment drops to roughly $100,000
per year, in the second half of the project lifetime.

Yearly Cost With Loan

$1,000,000

$800,000
Cost($/yr)

$600,000

$400,000

$200,000

$0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Project Lifetime(yr)
Water Purchase Ion Exchange

Figure #6.4: Yearly cost with a loan

If sufficient capital is made available so the University can build the facility with no loan or bond sale
needed, the savings after year one would be substantially greater. Figure 6.5 is a graph analogous to Figure
6.4, specific to this situation.

Page 44 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Yearly Cost Without Loan

$1,800,000
$1,600,000
$1,400,000
Cost ($/yr)

$1,200,000
$1,000,000
$800,000
$600,000
$400,000
$200,000
$-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Project Lifetime(yr)
Water Purchase Ion Exchange

Figure #6.5: Yearly Cost without A Loan

Therefore, substantial savings are obvious right away. Two price levels have been proposed for water sold
by the City of Norman. The first is the current commercial price level, $1.14 per 1000 gallons, upon which
most of the previous calculations were based. However it is possible that the University and the City of
Norman will agree upon a lower price. The model was given the water price of $0.85 per 1000 gallons as a
low margin estimate. Figure 6.6 shows that the savings are substantial from year one even at this low price
level.

Savings per year Current Dollars

900000
800000
700000
600000
Savings ($)

500000 $1.14/1000 gallons

400000 $0.85/1000 gallons
300000
200000
100000
0
1

11

13

15

17

19

Project Lifetime (yr)

Figure #6.6: Savings At Both Price Levels

Page 45 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

As can be seen from Figure 6, the ion exchange (IX) plant option results in significant savings for every
year of the project lifetime, when compared to the Water Purchase option. For approximately the first half
of the project lifetime the savings start out low, $50,000 for the $0.85 price and $150,000 for the $1.14 price,
and increase steadily until the money borrowed for the capital investment is paid off. This occurs
approximately midway through the project lifetime, at which point the savings roughly double and stay
constant until the last year.
Figure 6.7 is a Net Present Cost comparison of the Ion Exchange treatment and the water purchase options.

9
8
Net Present Cost($MM)

7
6
5 Ion Exchange Plant
4 Water Purchase
3
2
1

$1.14 $0.85 $0.45

Price of water ($/1000 gal)

Figure #6.7: Comparison Of NPC For Ion Exchange And Water Purchase At Different Price Levels

Assuming that the price of water charged by the City of Norman remains constant at the current value
of $1.14/1000 gal, for the next 20 years, the Ion Exchange facility option will result in approximately
$5,000,000 of Net Present Cost savings. It is also apparent from Figure 6.7 that the minimum allowable
price of water for the WP option to match the economic attractiveness of the IX option is an unrealistic
price of $0.45/1000 gals. A price so low could not be offered by the City of Norman, since it costs the city
roughly $0.60/1000 gals to produce it. Any price charged by the city of Norman that is above this number,
results in loss of money if the WP option is chosen. It should also be taken into consideration that the price
of water is not expected to stay constant or drop. Conversely, it is expected to increase, because of
Norman’s overwhelming growth that is projected to result in water shortage by year 2010.

Page 46 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

6.4 Risk Analysis

The mathematical model is also able to show how the results respond to uncertainty in the
values that were used to estimate Net Present Cost. For example, the price of “emergency”-
based water purchase was varied from the wholesale commercial value of $1.14 per 1000
gallons to $7.00, which represents a value higher than is realistically possible. Figure 6.8
shows how the Net Present Cost of building the plant reacts to such large variations.

3.3

3.25
Present Cost ($MM)

3.2

3.15

3.1

3.05

3
0 1 2 3 4 5 6 NPC7
Price of high-rate water ($/1000 gal) Linear (NPC)

Figure #6.8: Sensitivity of NPC to water cost

The price of emergency water, as seen, has very little effect on the overall NPC of the
plant. Even if the water price were to be doubled, the cost of the plant would only increase
by approximately 3%.

The well field’s maximum output capability could also vary. Currently, the field’s
maximum sustainable output is 1.8 MGD. Figure 6.9 shows how the mathematical model
responds to changes in the maximum output, which is essentially an indicator of how NPC
changes if a well is added or shut down. The line planes at two points, one at low
production, where it becomes more cost effective to simply purchase all the water (at slightly
more than .5 MGD) and the point where even increasing the output does not cause the
model to want to build a bigger plant (approximately 1600 MGD).

Page 47 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

10
Net Present Cost ($million)

6
NPC
4

0
0 500 1000 1500 2000 2500
Maximum Field Capacity (1000 gal/day)

Figure #6.9: Maximum Field Capacity vs. NPC

Operating cost per unit capacity and capital investment per unit capacity were some of
the parameters estimated by this study. It is conceivable that, due to rising chemical prices
or unforeseen complications, uncertainty could exist in these numbers. Figure 6.10 plots
Operating Cost vs. Capital Investment and shows the ‘threshold’ between Water Purchase
and IX Treatment options. This is the point at which raising the values causes the plant to
become economically infeasible, causing the model to recommend Water Purchase as a
solution. Areas below and to the left of each black line are areas where Treatment is
favorable at a given price level, while areas to the right and above the black line are those at
which Water Purchase is the best choice.

As Figure 6.10 shows, the costs estimated for the Ion Exchange treatment facility
(marked by the blue diamond) are well within the area of treatment, at either $0.85 per 1000
gallons or $1.14 per 1000 gallons. In fact, one value would have to be off by 400% for
treatment not to be economically favorable when compared to water purchase. Therefore, it
is safe to conclude that the Ion Exchange treatment will be less expensive than purchasing
water from the city of Norman, even if some uncertainty in the estimates is encountered.

Page 48 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

35
Operating Cost $/1000 gallons
30
Area of Water Purchase
25
capacity

20
15
$1.14
10
5 Area of Treatment $0.85
0
0 1000 2000 3000 4000 5000
Capital Investment $/1000 gal capacity

Figure #6.10: Sensitivity of NPC to capacity unit cost

The probability distribution displayed in Figure 6.11 quantifies the amount of risk involved
in building the Ion Exchange treatment facility. This distribution was obtained by running
the stochastic model. The average value of the produced Net Present Cost distribution
occurs at approximately $3,150,000, which compares favorably with the preliminary estimate
of the study and the results previously generated by the model. The shape of the probability
distribution is similar to that of the normal distribution, with the exception of its tails. The
probability distribution has a short left tail, because there is less probability of the project
being cheaper than estimated. This makes sense, since there is a minimum capital
investment associated with building a treatment facility. The right tail of the distribution
quantifies the probability of the Net Present Cost being higher than what was estimated by
the model. The fact that the right tail does not extend far from the mean value signifies that
the risk associated with the Net Present Cost estimate is minimal. In fact, there the
probability of the Net Present Cost exceeding $3,600,000 is estimated to be less than 5%!

Page 49 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

Net Present Cost Probability Distribution

25
20
Frequency

15
10
5
0

e
,8 00

0
00

or
$2 ,00

$3 ,00

$3 ,00

$3 ,00

$3 ,00

$3 ,00

$3 ,00

$3 ,0 0
,0

,0

M
00

00

00

00

00

00

00

00

00
00
,7

,9

,0

,1

,2

,3

,4

,6
,5
$2

$2

Net Present Cost

Figure #6.11: NPC Probability Distribution

6.5 Recommendations

As a result of the planning model, our group recommends that OU construct an ion
exchange plant with a capacity of 1.6 MGD at the Westheimer airport facility. No additions
to the capacity of the plant are needed. The plant’s capital investment is expected to be less
than $2,000,000, with a total net present cost of near $3,000,000.
While this is a substantial investment, the yearly cost of buying water from Norman
starts at $500,000 and increases, with an expected 20-year net present cost of $8,100,000.
There is little doubt that treatment is the best solution for the University of Oklahoma.
However, before implementation, we recommend further study of the ion exchange process
itself, in order to more accurately design the treatment facility.

Page 50 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

7.1 References

7.1.1 Online References

These References can all be found in appendix G.

Reference.1.html
Key parameters for Arsenic removal, HDR, inc.
Reference.2.ppt
Civil Engineering Arsenic Project
Reference.3.html
EPA Regulations
Reference.4.html
MIT Synopsis of Arsenic Removal Processes
Reference.5.html
Online pH monitor
Reference.6.html
Effects of Arsenic Consumption by Humans
Reference.7.doc
Document exploring the effects of Arsenic treatment on the environment

7.1.2 Other References

Bryan Mitchell, City of Norman

Don Carter, OU Physical Plant
CH2M Hill: Joe Chwirka, Murray Flemming
Trenton Brown, OU Environmental Health and Safety
CE: Dr. Sabatini, Daniel Gallo (graduate asst.)

Page 51 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

7.2 Appendix 1
Mathematical Model

Equations6:
TotalCost = ∑ p s C s
s

(
C s = ∑ C yr =∑ CI yr + OPyr + Pr ice yr * WPyr * (1 + i ) yr −1 − Borrowed yr + Repaid yr * df yr )
yr yr

CI yr = a * z yr + b * Cap yr
yr
OPyr = α F ∑ zξ + β ∑ Q yr ,mo , s
ξ =1 mo

Demand yr ,mo , s = Q yr ,mo , s + WPyr ,mo , s

ztot yr = ∑ zξ if ξ ≤ yr
ξ

Finance Equations:
Pr ice yr , s = WholeSale Pr ice * y yr , s + Emergency Pr ice * (1 − y yr , s )
Debt yr , s = (1 + i ) * Debt yr −1, s + Borrowed yr , s − Repaid yr , s

Constraints:

Physical Constraints:
CapTot yr = ∑ (Capξ + CapAdd ξ ) if ξ ≤ yr
ξ

CapTot yr ≥ Q yr ,mo , s
Cap yr ≤ MaxCap * z yr

Finance Constraints:
∑WP mo
yr , mo , s − 1000 * y yr , s ≥ 0

C yr , s ≤ Budget
Debt 20, s = 0
Repaid yr ,s ≥ 2 * i * Debt yr −1, s

6
Subscripts: s-scenario; yr-year; ζ-year alias; mo-month.

Page 52 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033

GAMS Model

$Title STOCHASTIC MODEL, BUDGET FIXED FOR S, PRICE FIXED FOR S.

$offlisting
$offsymxref
$offsymlist
$offuelxref
$offuellist
options iterlim=1000000;
*DECLARING VARIABLES===================================
Sets
yr project lifetime /1*20/
mo month of each year /1*12/
s scenario number /1*1/
* zeta number of facilities in that year? /0*10/;

*ENTERING DATA=============================================
Scalar
*CI=(a+b(capacity))*z
a capital cost of facility addon /1163900/
b slope_extra price for capacity /428/
* b slope_extra price for capacity /635.7/
maxcap maximum capacity /1800/
* beta OpCost varying costs fudge /3.72/
alpha OpCost fixed costs fudge /3800/
df depreciation factor /.06/
* price water price each year(CONST) /34.2/
budget budget /500000/
i interest rate /0.09/
aa capital investment of added /300000/
* bb capital investment of added /500/
inflate inflation rate /.02/;

Table demand(yr,mo,s) projected monthly water demand for each year

1.1*1 2.1*1 3.1*1 4.1*1 5.1*1
6.1*1 7.1*1 8.1*1 9.1*1 10.1*1 11.1*1
12.1*1
1 1146 1093 1242 1348 1322
1200 1332 1408 1261 1256 1175
1067
2 1056 1102 1236 1336 1338
1199 1336 1402 1418 1170 1129
973
3 1160 1084 1299 1406 1358
1320 1355 1396 1440 1276 1171
1112
4 1207 1144 1202 1320 1457
1222 1451 1539 1345 1275 1218
1029
5 1297 1148 1318 1314 1324
1153 1249 1434 1293 1304 1251
1129
6 1221 1242 1319 1525 1436
1355 1324 1525 1426 1398 1282
1121

Page 53 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033
7 1360 1140 1257 1376 1494
1369 1463 1634 1334 1288 1239
1230
8 1304 1105 1348 1437 1413
1391 1419 1610 1462 1446 1331
1216
9 1224 1255 1357 1510 1384
1317 1571 1712 1371 1433 1252
1213
10 1360 1330 1426 1647 1518
1278 1534 1588 1481 1343 1302
1148
11 1260 1261 1405 1362 1553
1343 1608 1568 1339 1513 1252
1249
12 1207 1272 1374 1511 1520
1384 1400 1651 1602 1417 1373
1185
13 1383 1143 1549 1568 1543
1434 1590 1815 1500 1404 1344
1244
14 1357 1279 1541 1537 1507
1429 1432 1607 1375 1519 1461
1397
15 1405 1441 1553 1511 1616
1345 1790 1717 1261 1514 1350
1229
16 1451 1505 1592 1518 1574
1326 1343 1914 1574 1539 1357
1310
17 1552 1186 1474 1672 1548
1446 1571 1639 1492 1458 1426
1220
18 1394 1420 1552 1599 1720
1521 1462 1680 1647 1557 1444
1168
19 1417 1312 1369 1706 1444
1538 1664 1695 1487 1533 1474
1262
20 1371 1302 1537 1593 1657
1461 1552 1848 1595 1647 1495
1327;

*display demand;

*VARYING PARAMETERS FOR RISK ANALYSIS BASED ON SCENARIO

Parameter beta
demands(yr,mo,s) varied scenario specific demand;
beta(s) = normal(3.72,0);
demands(yr,mo,s)=normal(demand(yr,mo,s),0)
*display demands;

*DECLARING VARIABLES======================================

Page 54 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033
Variables
x (yr) secondary binary
y(yr,s) price binary
m(yr,s) price dummy varible
z(yr) plant binary
ztot(yr) z sum
capadd(yr) capacity added
cap(yr) capacity for given year
q(yr,mo,s) water
wp(yr,mo,s) water purchased
ci(yr) capital investment in given year
ciadd(yr) capital investment added
op(yr,s) operating cost in given year
c TOTAL COST
captot(yr) sum of capacities for all years
borrowed(yr,s) Money borrowed in given year
repayed(yr,s) Money repayed in given year
debt(yr,s) debt in given year
yrcost(yr,s) Cost in given year
ns sum of scenarios
scost(s) total cost in scenario s
wcost(yr,s) water purchasing cost
* price(yr,s) price of water(whole sale or demand based);
Binary variable z,y,x;
Positive variable cap;
Positive variable capadd;
Positive variable wp;
Positive variable q;
Positive variable repayed;
Positive variable borrowed;
Positive variable debt;
Positive variable wcost;

*EQUATIONS=================================================
alias (yrr, yr)
Equations
capacity(yr) capacity eqn constraint
capinv(yr) capital investment for given year
capinvadd(yr) capitalinvestment added
opcost(yr,s) operating cost for given year
cost Total cost expectation value(BIG CAHUNA)
watcost(yr,s) Water cost each month
scenariocost(s) total cost for all years for scenario
s
qcap(yr,mo,s) water capacity in yr&month less than year
capacity
waterdemand(yr,mo,s) total water demand
processed+purchased
totfac(yr) total facility
captotal(yr) sum of capacities
capaddd(yr) capacity added
constr(yr) constraint of adding capacity

* scenario
*FINANCE STUFF
yearcost(yr,s)

Page 55 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033
budgetcost(yr,s) budget =g= cost
debtfinal(s) last year no debt
debteqn(yr,s) defines debt
repay(yr,s) pay interest minimum
*WATER PRICE STUFF
* waterprice(yr,s) defines water price(wholesale or not)
* pricedisplay(yr,s)
chooseprice(yr,s)
pone(yr,s) system of equations that makes price linear
operation
ptwo(yr,s) -----------"------------"-----------
pthree(yr,s) -----------"------------"-----------
pfour(yr,s) -----------"------------"-----------
;

*=====================EQUATIONS===================================
*Scenario Probability

*CAPITAL INVESTMENT CALC

captotal(yr).. captot(yr) =e= sum(yrr$(ord (yrr) le
ord(yr)),cap(yrr)+capadd(yrr));
qcap(yr,mo,s).. captot(yr) =g= q(yr,mo,s);
capacity(yr).. cap(yr) =l= maxcap*z(yr);
capinv(yr).. ci(yr) =e= (a*z(yr)+b*cap(yr))*(1/(1+df)**(ord(yr)-1));
waterdemand(yr,mo,s).. demands(yr,mo,s) =e= q(yr,mo,s)+wp(yr,mo,s);
totfac(yr).. ztot(yr) =e= sum(yrr$(ord (yrr) le ord(yr)),z(yrr));
opcost(yr,s).. op(yr,s) =e= (alpha * ztot(yr) + beta(s) *
sum(mo,q(yr,mo,s)))*(1/(1+df)**(ord(yr)-1));
watcost(yr,s).. wcost(yr,s) =e= yrcost(yr,s)-op(yr,s)-
ci(yr)+borrowed(yr,s)-repayed(yr,s);
Capaddd(yr).. capadd(yr) =l= maxcap * x(yr);
constr(yr).. x(yr) =l= ztot(yr-1);
capinvadd(yr).. ciadd(yr) =e=
(aa*x(yr)+1.1*b*capadd(yr))*(1/(1+df)**(ord(yr)-1));
*VARIABLE WATER PRICE
chooseprice(yr,s).. sum(mo,wp(yr,mo,s))-1000*y(yr,s) =g= 0;
*waterprice(yr,s).. price(yr,s) =e= (34.2*y(yr,s)+45*(1-y(yr,s)));
*making price linear
pone(yr,s).. m(yr,s)-y(yr,s)*40000 =l=0;
ptwo(yr,s).. m(yr,s) =g= 0;
pthree(yr,s).. (sum(mo,wp(yr,mo,s))- m(yr,s))-(1-y(yr,s))*40000 =l=0;
pfour(yr,s).. sum(mo,wp(yr,mo,s))-m(yr,s) =g=0;

*pricedisplay(yr,s)..
price(yr,s)=e=(34.2*m(yr,s)+45*(sum(mo,wp(yr,mo,s))-m(yr,s)));

*FINANCE BUDGET
yearcost(yr,s).. yrcost(yr,s) =e=
op(yr,s)+ci(yr)+(1+inflate)**(ord(yr)-
1)*(25.5*m(yr,s)+90*(sum(mo,wp(yr,mo,s))-m(yr,s)))-
borrowed(yr,s)+repayed(yr,s);
budgetcost(yr,s).. yrcost(yr,s) =l= budget * (1/(1+df))**(ord(yr)-1);
debteqn(yr,s).. debt(yr,s) =e= (1+i)*debt(yr-1,s)+borrowed(yr,s)-
repayed(yr,s);
debtfinal(s).. debt('20',s) =e= 0;

Page 56 of 57
 moovvaall ffrroom
Reem
AArrsseenniicc R OU
mO Waatteerr
UW
55..22..22000033
repay(yr,s).. repayed(yr,s) =g= 2*i*debt(yr-1,s);
*repay(yr,s).. repayed(yr,s) =g= Sum(yrr$(ord (yrr) le
ord(yr)),i*(1+i)**(20-ord(yrr))*debt(yrr,s)/((1+i)**(20-ord(yrr)))-1);
*TOTAL COST
scenariocost(s).. scost(s) =e=
sum(yr,op(yr,s))+sum(yr,ci(yr))+sum(yr,ciadd(yr))+sum(yr,(1+inflate)**(
ord(yr)-1)*(25.5*m(yr,s)+90*(sum(mo,wp(yr,mo,s))-
m(yr,s)))*(1/(1+df)**(ord(yr)-1)))
-sum(yr,borrowed(yr,s)*(1/(1+df)**(ord(yr)-
1)))+sum(yr,repayed(yr,s)*(1/(1+df)**(ord(yr)-1)));
cost.. c =e=sum(s,(1/1)*(
sum(yr,op(yr,s))+sum(yr,ci(yr))+sum(yr,ciadd(yr))+sum(yr,(1+inflate)**(
ord(yr)-1)*(25.5*m(yr,s)+90*(sum(mo,wp(yr,mo,s))-
m(yr,s)))*(1/(1+df)**(ord(yr)-1))))
-sum(yr,borrowed(yr,s)*(1/(1+df)**(ord(yr)-
1)))+sum(yr,repayed(yr,s)*(1/(1+df)**(ord(yr)-1))));

*OUTPUT TO EXCEL
*$libinclude XLEXPORT scost.l c:\results.xls c_ /m
file out /m:\output.txt/;
put out;
*put c.l;
*CALL PROGRAM
model totcost /all/;
solve totcost using mip minimizing c;
display cap.l, z.l,x.l, ci.l,ciadd.l, beta,
wp.l,op.l,ci.l,scost.l,borrowed.l, repayed.l, captot.l, c.l, yrcost.l;

Page 57 of 57