Dynamics of Structures Volume 2 - Applications Chowdhury

Dynamics of Structure and
Foundation A Unied Approach

2. Applications
2009 Taylor & Francis Group, London, UK
Dynamics of Structure and
Foundation A Unied Approach
2. Applications
Indrajit Chowdhury
Petrofac International Ltd
Sharjah, United Arab Emirates
Shambhu P. Dasgupta
Department of Civil Engineering
Indian Institute of Technology
Kharagpur, India
Also available:
Dynamics of Structure and Foundation A Unied Approach
1. Fundamentals
Indrajit Chowdhury & Shambhu P. Dasgupta 2009, CRC Press/Balkema
ISBN: 978-0-415-47145-9 (Hbk)
ISBN: 978-0-203-88527-7 (eBook)
CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business
Typeset by Vikatan Publishing Solutions (P) Ltd, Chennai, India.
Printed and bound in Great Britain by Antony Rowe (a CPI Group company),
Chippenham, Wiltshire.
All rights reserved. No part of this publication or the information contained herein
may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, by photocopying, recording or otherwise,
without written prior permission from the publisher.
Although all care is taken to ensure integrity and the quality of this publication
and the information herein, no responsibility is assumed by the publishers nor
the author for any damage to the property or persons as a result of operation
or use of this publication and/or the information contained herein.
Published by: CRC Press/Balkema
P.O. Box 447, 2300 AK Leiden, The Netherlands
e-mail: Pub.NL@taylorandfrancis.com
www.crcpress.com www.taylorandfrancis.co.uk www.balkema.nl
Library of Congress Cataloging-in-Publication Data
Applied for
ISBN: 978-0-415-49223-2 (Hbk)
ISBN: 978-0-203-87922-1 (eBook)
Contents
Preface xiii
1 Dynamic soil structure interaction 1
1.1 Introduction 1
1.1.1 The marriage of soil and structure 1
1.1.2 What does the interaction mean? 2
1.1.3 It is an expensive analysis do we need
to do it? 4
1.1.4 Different soil models and their coupling to
superstructure 6
1.2 Mathematical modeling of soil &
structure 6
1.2.1 Lagrangian formulation for 2D frames or
stick-models 6
1.2.2 What happens if the raft is f lexible? 14
1.3 A generalised model for dynamic soil structure
interaction 28
1.3.1 Dynamic response of a structure with multi
degree of freedom considering the underlying
soil stiffness 28
1.3.2 Extension of the above theory to system
with multi degree of freedom 29
1.3.3 Estimation of damping ratio for the soil
structure system 30
1.3.4 Formulation of damping ratio for single degree
of freedom 31
1.3.5 Extension of the above theory to systems with
multi-degree freedom 32
1.3.6 Some fallacies in coupling of soil and structure 40
1.3.7 What makes the structural response attenuate
or amplify? 41
vi Contents
1.4 The art of modelling 42
1.4.1 Some modelling techniques 42
1.4.2 To sum it up 46
1.5 Geotechnical considerations for dynamic soil
structure interaction 46
1.5.1 What parameters do I look for in the soil
report? 47
1.6 Field tests 49
1.6.1 Block vibration test 49
1.6.2 Seismic cross hole test 50
1.6.3 How do I co-relate dynamic shear modulus
when I do not have data from the dynamic
soil tests? 51
1.7 Theoretical co-relation from other soil parameters 52
1.7.1 Co-relation for sandy and gravelly soil 52
1.7.2 Co-relation for saturated clay 58
1.8 Estimation of material damping of soil 61
1.8.1 Whitmans formula 61
1.8.2 Hardin formula 62
1.8.3 Ishibashi and Zhangs formula 63
1.9 All things said and done how do we estimate
the strain in soil, specially if the strain is large? 65
1.9.1 Estimation of strain in soil for machine
foundation 65
1.9.2 Estimation of soil strain for earthquake
analysis 70
1.9.3 What do we do if the soil is layered with
varying soil property? 77
1.9.4 Checklist of parameters to be looked in the
soil report 79
1.10 Epilogue 80
2 Analysis and design of machine foundations 83
2.1 Introduction 83
2.1.1 Case history #1 83
2.1.2 Case history #2 84
2.2 Different types of foundations 85
2.2.1 Block foundations resting on soil/piles 85
2.2.2 How does a block foundation supporting
rotating machines differ from a normal
foundation? 86
Contents vii
2.2.3 Foundation for centrifugal or rotary type
of machine: Different theoretical methods
for analysis of block foundation 88
2.2.4 Analytical methods 90
2.2.5 Approximate analysis to de-couple equations
with non-proportional damping 99
2.2.6 Alternative formulation of coupled equation
of motion for sliding and rocking mode 105
2.3 Trick to by pass damping Magnif ication factor,
the key to the problem. . . 113
2.4 Effect of embedment on foundation 117
2.4.1 Novak and Beredugos model 119
2.4.2 Wolfs model 119
2.5 Foundation supported on piles 119
2.5.1 Pile and soil modelled as f inite element 121
2.5.2 Piles modelled as beams supported on elastic
springs 123
2.5.3 Novaks (1974) model for equivalent spring
stiffness for piles 124
2.5.4 Equivalent pile springs in vertical direction 125
2.5.5 The group effect on the vertical spring
and damping value of the piles 127
2.5.6 Effect of pile cap on the spring
and damping stiffness 128
2.5.7 Equivalent pile springs and damping
in the horizontal direction 129
in rocking motion 130
2.5.9 Group effect for rotational motion 131
2.5.10 Model for dynamic response of pile 138
2.5.11 Dynamic analysis of laterally loaded piles 162
2.5.12 Partially embedded piles under rocking mode 193
2.5.13 Group effect of pile 201
2.5.14 Comparison of results 203
2.5.15 Practical aspects of design of machine
foundations 205
2.6 Special provisions of IS-code 213
2.6.1 Recommendations on vibration isolation 213
2.6.2 Frequency separation 213
2.6.3 Permissible amplitudes 214
2.6.4 Permissible stresses 214
2.6.5 Concrete and its placing 214
viii Contents
2.6.6 Reinforcements 214
2.6.7 Cover to concrete 215
2.7 Analysis and design of machine foundation under
impact loading 231
2.7.1 Introduction 231
2.7.2 Mathematical model of a hammer foundation 238
2.8 Design of hammer foundation 248
2.8.1 Design criteria for hammer foundation 248
2.8.2 Discussion on the IS-code method of analysis 252
2.8.3 Check list for analysis of hammer foundation 253
2.8.4 Other techniques of analysis of Hammer
foundation 253
2.9 Design of eccentrically loaded hammer foundation 268
2.9.1 Mathematical formulation of anvil placed
eccentrically on a foundation 268
2.9.2 Damped equation of motion with eccentric
anvil 270
2.10 Details of design 271
2.10.1 Reinforcement detailing 271
2.10.2 Construction procedure 271
2.11 Vibration measuring instruments 272
2.11.1 Some background on vibration measuring
instruments and their application 272
2.11.2 Response due to motion of the support 272
2.11.3 Vibration pick-ups 272
2.12 Evaluation of friction damping from
energy consideration 283
2.13 Vibration isolation 284
2.13.1 Active isolation 285
2.13.2 Passive isolation 287
2.13.3 Isolation by trench 288
2.14 Machine foundation supported on frames 289
2.14.1 Introduction 289
2.14.2 Different types of turbines and the
generation process. . . 290
2.14.3 Layout planning 292
2.14.4 Vibration analysis of turbine foundations 293
2.15 Dynamic soil-structure interaction model for
vibration analysis of turbine foundation 305
2.16 Computer analysis of turbine foundation based on
multi degree of freedom 312
2.17 Analysis of turbine foundation 319
Contents ix
2.17.1 The analysis 319
2.17.2 Calculation of the eigen values 320
2.17.3 So the ground rule is. . . 321
2.17.4 Calculation of amplitude 321
2.17.5 Calculation of moments, shears and torsion 321
2.17.6 Practical aspects of design of Turbine
foundation 322
2.18 Design of turbine foundation 322
2.18.1 Check list for turbine foundation design 322
2.18.2 Spring mounted turbine foundation 330
3 Analytical and design concepts for earthquake
engineering 389
3.1 Introduction 389
3.1.1 Why do earthquakes happen in nature? 390
3.1.2 Essential difference between systems
subjected to earthquake and vibration
from machine 391
3.1.3 Some history of major earthquakes around
the world 392
3.1.4 Intensity 394
3.1.5 Effect of earthquake on soil-foundation
system 395
3.1.6 Liquefaction analysis 395
3.2 Earthquake analysis 412
3.2.1 Seismic coeff icient method 412
3.2.2 Response spectrum method 417
3.2.3 Dynamic analysis under earthquake loading 424
3.2.4 How do we evaluate the earthquake force? 425
3.2.5 Earthquake analysis of systems with multi-
degree of freedom 431
3.2.6 Modal combination of forces 444
3.3 Time history analysis under earthquake force 448
3.3.1 Earthquake analysis of tall chimneys
and stack like structure 456
3.4 Analysis of concrete gravity dams 481
3.4.1 Earthquake analysis of concrete dam 481
3.4.2 A method for dynamic analysis
of concrete dam 485
3.5 Analysis of earth dams and embankments 519
3.5.1 Dynamic earthquake analysis of earth dams 519
x Contents
3.5.2 Mononobes method for analysis of
earth dam 519
3.5.3 Gazetas method for earth dam analysis 522
3.5.4 Makadisi and Seeds method for analysis of
earth dam 523
3.5.5 Calculation of seismic force in dam and
its stability 526
3.6 Analysis of earth retaining structures 526
3.6.1 Earthquake analysis of earth retaining
structures 526
3.6.2 Mononobes method of analysis of
retaining wall 527
3.6.3 Seed and Whitmans method 530
3.6.4 Arangos method 530
3.6.5 Steedman and Zengs method 532
3.6.6 Dynamic analysis of RCC retaining wall 533
3.6.7 Dynamic analysis of cantilever and
counterfort retaining wall 533
3.6.8 Some discussions on the above method 544
3.6.9 Extension to the generic case of soil
at a slope i behind the wall 544
3.6.10 Dynamic analysis of counterfort
retaining wall 547
3.6.11 Soil sloped at an angle i with horizontal 560
3.7 Unyielding earth retaining structures 571
3.7.1 Earthquake Analysis of rigid walls when
the soil does not yield 571
3.7.2 Ostadans method 575
3.8 Earthquake analysis of water tanks 577
3.8.1 Analysis of water tanks under earthquake
force 577
3.8.2 Impulsive time period for non rigid walls 581
3.8.3 Sloshing time period of the vibrating fluid 583
3.8.4 Calculation of horizontal seismic force for
tank resting on ground 583
3.8.5 Calculation of base shear for tanks resting
on ground 584
3.8.6 Calculation of bending moment on the
tank wall resting on the ground 584
3.8.7 Calculation of sloshing height 585
Contents xi
3.9 Mathematical model for overhead tanks under
earthquake 588
3.9.1 Earthquake Analysis for overhead tanks 588
3.9.2 Hydrodynamic pressure on tank wall
and base 592
3.9.3 Hydrodynamic pressure for circular tank 592
3.9.4 Hydrodynamic pressure for rectangular
tank 593
3.9.5 Effect of vertical ground acceleration 593
3.9.6 Pressure due to inertia of the wall 593
3.9.7 Maximum design dynamic pressure 594
3.10 Practical aspects of earthquake engineering 598
3.10.1 Epilogue 603
References 605
Preface
The monograph entitled Dynamics of Structure and Foundation - A Unied
Approach consists of two volumes. While in Volume 1 we dealt with background
theories and formulations that constitute the above subject, this second volume deals
with application of these theories to various aspects of civil engineering problems con-
stituting topics related to dynamic soil-structure interaction, machine foundation and
earthquake engineering.
If we have managed to stir the wrath of the professionals in Volume 1 with mazes
of tensors, differential and integral equations, it is our strong conviction that in this
present volume we will be able to considerably appease this fraternity for it constitutes
of a number of applications that are innovative, easy to apply and solutions to many
practical problems that puts an engineer into considerable difculty and uncertainties
in a design ofce.
We start Volume 2 with the topic of Dynamic Soil Structure Interaction (DSSI). We
believe this topic would play a key role in future and more so with the distinct pos-
sibility of construction of Nuclear power plants (especially in India) globally. A clear
concept on this topic would surely be essential for designing such plants. Though we
have dealt this topic only in terms of fundamental concepts, yet we feel that we have
given sufcient details to eradicate the misnomer from which many engineers suffer
that DSSI is nothing but adding some springs to the boundary of a structure and then
doing the analysis through a computer.
The geotechnical aspects that play an extremely important role in selecting the soil-
spring value, (that are highly inuenced by the strain range) have been dealt in quite
detail. We hope that this section will do away with some of the major blunders that we
make in DSSI analysis, and appreciate howthe results thus obtained become unrealistic
and questionable. We sincerely hope that engineers performing DSSI analysis, would
start paying sufcient attention to some of the key engineering parameters as furnished
in the soil report that are being habitually ignored in design ofces.
Second chapter consists of design and analysis of machine foundations (both block
and frame type). In our collective experience as a consultant and academician we have
seen signicant confusion on this topic as to who is responsible for this hapless orphan,
structural or geotechnical engineers? While people from classical soil mechanics dis-
owns it, as it involves the evaluation of eigen-values and vectors that are far away from
xiv Preface
their traditional failure theories of foundation, structural engineers on the other hand
are equally reluctant to shoulder the guardianship for their inherent apathy towards
what lies beneath the machine foundation. As such, a design involving machine
foundation throws the most challenging and interesting task in the domain of civil
engineering that requires multi-discipline knowledge and should be equally interest-
ing to an engineer having structural or geotechnical background. The matrix analysis
concept that we have introduced herein is quite easy to follow and we hope would
bridge the gap that is still prevalent in academics and practice alike.
We would be looking forward to have some feedback from hardened professionals
who are working in this area, as to how they feel about our representation which we
believe is quite novel and has tried to answer a number of problems that often become
burning issues on which they have spent signicant time on clarifying either to their
Clients or Project Management Consultants.
The last chapter of this volume deals with the most fearful force Mother Nature has
created Earthquake. Earthquake engineering as a topic is so vast, complex and
diverse (and ever changing) that we concede that it did give us some uncomfortable
moments as to what should justiably constitute this chapter? Majority of the books
that address this topic are far too focused on buildings and there are hardly any book
around, that has addressed other specialized structures like chimneys, dams, retaining
walls, water tanks etc (except some very specialized literature). It should be realized
that some of these structures are expensive, important and cannot be ignored while
building an earthquake resistant infrastructure.
Buildings, we concede are the biggest casualties during an earthquake and are
directly related to human life but damages to other structures as mentioned above can
also create havoc especially in the post earthquake relief scenario. The major focus
being still thrust on buildings, we were also quite surprised to nd that there is still
much room for improvement in many of these structures, where technologies which
are as old as 60 years are still in use (for instance earthquake response of retaining
walls). We tried to improve upon many of themand believe that we have brought about
a number of innovative solutions that can be adapted in a design ofce environment
and can also be used as a basis for further research.
While presenting the topic no demarcation is made between geotechnical and
structural earthquake engineering. For, as a seismic specialist our job is to minimize
the destruction of property and save human lives. Thus doing a structural design we
can perform the most sophisticated analysis and provide the most expensive detailing
and our building still fails due to liquefaction killing people__no medals for doing
an excellent structural design!, so if you do something do it in totality and not in
isolation and this has been our major endeavour- that we have tried to communicate
to you through this book.
Indrajit Chowdhury
Shambhu P. Dasgupta
Chapter 1
Dynamic soil structure interaction
1.1 INTRODUCTION
This chapter deals with some of the basic concepts of dynamic soil-structure interaction
analysis. At the advent of this chapter we expect you to have some background on
Static soil structure interaction
Theory of Vibration/structural dynamics
Basic theory of soil dynamics
Based on the above
, we build herein the basic concepts of dynamic soil structure

interaction, which is slowly and surely gaining its importance in analytical procedure
for important structures.
1.1.1 The marriage of soil and structure
As was stated earlier in Chapter 4 (Vol. 1) even twenty years ago struc-
tures and foundations were dealt in complete isolation where the structural and
geo-technical/foundation engineers hardly interacted
1
.
While the structural engineer was only bothered about the structural configuration
of the systemin hand he hardly cared to knowanything more about soil other than the
allowable bearing capacity and its generic nature, provided of course the foundation
design is within his scope of work. On the other hand the geotechnical engineer only
remained focused on the inherent soil characteristics like (c, , N
c
, N
q
, N
, e
o
, C
c,
G
etc.) and recommending the type of foundation (like isolated footing, raft, pile etc.)
or at best sizing and designing the same.
The crux of this scenario was that nobody got the overall picture, while in reality
under static or dynamic loading the foundation and the structure do behave in tandem.
For theoretical background on these topics please consider Volume 1.

1 Even today there are companies which has divisions like structural and civil engineering!! Where the
responsibility of the structural division is to design the superstructure considering it as fixed base frame,
furnish the results (Axial load, Moments and Shear) and the column layout drawing to the civil division
who releases the foundation drawing based on this input data.
2 Dynamics of Structure and Foundation: 2. Applications
In chapter 4 (Vol. 1), in the problem Example 1.3.1, we have shown how the
soil stiffness can affect the bending moment and shear forces of a bridge girder and
ignoring the same how we can arrive at a result which can be in significant variation
to the reality.
Drawing a similar analogy one can infer that ignoring the soil stiffness in the overall
response (and treating it as a fixed base problem) the dynamic response of structure
(the natural frequencies, amplitude etc.) can be in significant variation to the reality
in certain cases.
This aspect came to the attention of engineers while designing the reactor building
of nuclear power plant for earthquake. Considering its huge mass and stiffness, the
fundamental time period for the fixed base structure came around 0.15 sec while
considering the soil effect the time period increased to 0.5 second giving a completely
different response than the fixed base case.
With the above understanding that underlying soil signicantly affects the response
of a structure, research was focused on this topic way back in 1970, and under the
pioneering effort of academicians and engineers, the two diverging domain of technol-
ogy was brought under a nuptial bond of Dynamic soil structure interaction, where
soil and structure where married off to a unied integrated domain. To our knowl-
edge the first signicant structure where the dynamic effect of soil was considered in
the analysis in Industry in India was the 500 MW turbine foundations for Singrauli
where the underlying soil was modeled as a frequency independent linear spring and
the whole system was analyzed in SAP IV (Ghosh et al. 1984).
1.1.2 What does the interaction mean?
We have seen earlier that considering the soil as a deformable elastic medium the
stiffness of soil gets coupled to the stiffness of the structure and changes it elastic
property. Based on this the characteristic response of the system also gets modified.
This we can consider as the local effect of soil.
On the other hand consider a case of a structure resting on a deep layer of soft
soil underlain by rock. It will be observed that its response is completely differ-
ent than the same system when it is located on soft soil which is of much shallow
depth or resting directly on rock
2
. Moreover the nature of foundation, (isolated
pad, raft, pile), if the foundation is resting or embedded in soil, layering of soil,
type of structure etc. has profound inuence on the over all dynamic response of the
system.
We had shown for static soil-structure interaction (Chapter 4 (Vol. 1)) case that the
soil can be modeled as equivalent springs or as finite elements and are coupled with the
superstructure.
Thus for a simple beam resting on an elastic support can be modeled as shown
in Figure 1.1.1 and an equivalent mathematical model for the same is shown in
Figure 1.1.2.
Based on matrix analysis of structure the element stiffness for this element may be
written as
2 The reason for these effects we will discuss subsequently.
Dynamic soil structure interaction 3
Node ii Node i Node j Node j
Soil Spring K
i
Soil Spring K
i Soil Spring K
j
Soil Spring K
j
Figure 1.1.1 Equivalent beam element connected to soil springs.
2 4
1
1
3
2
Figure 1.1.2 Mathematical model of the equivalent beam element.
[K
beam
] =
EIz
L
3
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
12 6L 0 12 6L 0
6L 4L
2
0 6L 2L
2
0
0 0
IxL
2
2Iz(1 + )
0 0
IxL
2
2Iz(1 + )
0
12 6L 0 12 6L 0
6L 2L
2
0 6L 4L
2
0
0 0
IxL
2
2Iz(1 + )
0 0
IxL
2
2Iz(1 + )
_
_
(1.1.1)
and the displacement vector is given by
{} = <
1

1

2

2

3

4
>
T
(1.1.2)
When the soil springs are added to the nodes, the overall stiffness becomes
[K
beam
] =
EIz
L
3
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
12 +
L
3
Kii
EIz
_
6L 0 12 6L 0
6L 4L
2
0 6L 2L
2
0
0 0
IxL
2
2Iz(1 + )
0 0
IxL
2
2Iz(1 + )
12 6L 0
_
12 +
L
3
Kjj
EIz
_
6L 0
6L 2L
2
0 6L 4L
2
0
0 0
IxL
2
2Iz(1 + )
0 0
IxL
2
2Iz(1 + )
_
_
(1.1.3)
where, [K
beam
] = combined stiffness matrix for the beam and the spring; K
ii
= K
jj
=
spring values of soil at node i and node j of the beam respectively.
The above is a very convenient way of representing the elastic interaction behavior
of the underlying soil and can be very easily adapted in a commercially available finite
element or structural analysis package.
1.1.3 It is an expensive analysis do we need to do it?
This is a common query comes to the mind of an engineer before starting of an analysis.
Based on this fact an engineer do become apprehensive if his/her analysis would suffer
from a cost over run or whether he/she will be able to finish the design within the
allocated time frame.
If he is convinced that soil structure interaction do takes place and the structure is
a crucial one
3
our recommendation would be its worth the effort rather than to be
sorry later. The additional engineering cost incurred is trivial compared to the risk
and cost involved in case of a damage under an earthquake or a machine induced load.
Now the first question is for what soil condition does dynamic soil structure
interaction takes place?
Veletsos and Meek (1974) suggest that chances of dynamic soil structure interaction
can be significant for the expression
V
s
fh
20 (1.1.4)
where V
s
= shear wave velocity of the soil; f = fundamental frequency of the fixed
base structure; h = height of the structure.
Let us now examine what does Equation (1.1.4) signifies?
Knowing the time period T = 1/f , the above expression can be rewritten as
V
s
T
h
20 (1.1.5)
For a normal framed building considering the fixed base time period as (0.1n), where
n is the number of stories and thus, we have
V
s
n
h
200 (1.1.6)
For a normal building the average ratio of h/n (height : storey ratio) is about 3 to
3.3 meter. Thus considering h/n = 3, we have
V
s
600 m/sec. (1.1.7)
3 Like Power House, Turbine foundations, Nuclear reactor Building, Main process piper rack, distillation
columns, bridges, high rise building catering to large number of people etc.
From which we conclude that for ordinary framed structure, when shear wave
velocity is less or equal to 600 meter/sec we can expect dynamic soil structure
interaction between the frame and the soil.
Incidentally, V
s
= 600 m/sec is the shear wave velocity which is associated with rock.
Thus it can be concluded that for all other type of soil, framed structures will behave
differently than a fixed base problem-unless and until it rests on rock. For Cantilever
structures like tall vessels, chimneys etc of uniform cross section fundamental time
period T is given by
T = 1.779
_
mh
4
EI
(1.1.8)
where, m = mass per unit length of the system; h = height of the structure; EI =
flexural stiffness of the system.
Substituting the above value in Equation (1.1.5) we have
V
s
T
h
20; or
V
s
1.779
_
mh
4
EI
h
20; or, V
s

11.24
h
_
EI
m
(1.1.9)
Considering, I = Ar
2
and m = A, where A = area of cross section; r = radius of
gyration; = Mass density of the material, we have
V
s

11.24r
h
_
E
(1.1.10)
Shear Wave Velocity for Soil-Structure interaction for
Chimneys
0.00
200.00
400.00
600.00
800.00
1000.00
1200.00
1400.00
1
0
0
1
2
5
1
5
0
1
7
5
2
0
0
2
2
5
2
5
0
2
7
5
3
0
0
Slenderness Ratio
S
h
e
a
r

W
a
v
e

v
e
l
o
c
i
t
y
(
m
/
s
e
c
)
Shear Wave velocity
steel chimney
Shear Wave velocity
concrete chimney
Figure 1.1.3 Chart to assess soil-structure interaction for steel and concrete chimney.
For steel structure the above can be taken as, V
s
57580/ where = h/r, the
slenderness ratio of the structure.
For concrete structure we have
V
s

123970
(1.1.11)
Based on the above expressions one can very easily infer if soil structure interaction
is significant or not.
The chart in Figure 1.1.3 shows limiting shear wave velocity below which soil-
structure interaction could be significant for a steel and concrete chimney.
1.1.4 Different soil models and their coupling
to superstructure
The various types of soil model that are used for comprehensive dynamic analysis are
as follows:
1 Equivalent soil springs connected to foundations modeled as beams, plates, shell
etc.,
2 Finite element models (mostly used in 2D problems),
3 Mixed Finite element and Boundary element a concept which is slowly gaining
popularity.
Of all the options, spring elements connected to superstructure still remain the most
popular model in design practices due to its simplicity and economy in terms of analysis
especially when the superstructure is modeled in 3-dimensions.
It is only in exceptional or very important cases that the Finite elements and Bound-
ary elements are put in to use and that too is mostly restricted to 2 dimensional cases.
1.2 MATHEMATICAL MODELING OF SOIL & STRUCTURE
We present hereafter some techniques that are commonly adopted for coupling the
soil to a structural system.
1.2.1 Lagrangian formulation for 2D frames or stick-models
This formulation is one of the most powerful tool to couple the stiffness of soil to the
superstructure-specially when one is using a stick model or a 2D model.
For the frame shown hereafter we formulate the coupled stiffness and mass matrix
for the soil structure system which can be effectively used for dynamic analysis.
In the system shown in Figure 1.2.1, m
f
, J
= mass and mass moment of inertia

of the foundation; m
1
, J
1
= mass and mass moment of inertia of the 1st story; m
2
,
J
2
= mass and mass moment of inertia of the top story; K
x
, K
= translational and
rotational stiffness of the soil; and k
1
, k
2
= stiffness of the columns.
y
2
m
2
J
2
k
2
y
1
h
2
m
1
, J
1
k
1 h
1
K
x
K m
f
, J
Figure 1.2.1 2D Mathematical model for soil structure interaction.
The equation for kinetic energy of the system may be written as
T =
1
2
m
f
u
2
+
1
2
J
2
+
1
2
m
1
( u +h
1

+ y
1
)
2
+
1
2
J
1

2
+
1
2
m
2
( u + (h
1
+h
2
)
+ y
2
)
2
+
1
2
J
2

2
(1.2.1)
U =
1
2
K
x
u
2
+
1
2
K
2
+
1
2
k
1
y
2
1
+
1
2
k
2
(y
2
y
1
)
2
(1.2.2)
Considering the expression
4
,
d
dt
_
T
q
i
_
+
U
q
i
= 0, we have the free vibration
equation as
_
_
_
_
_
_
m
f
+m
1
+m
2
m
1
h
1
+m
2
H m
1
m
2
m
1
h
1
J +m
1
h
2
1
+m
2
H
2
m
1
h
1
m
2
H
m
1
m
1
h
1
m
1
0
m
2
m
2
H 0 m
2
_
_
_
_
u
y
1
y
2
_
_
+
_
_
_
_
_
_
K
x
0 0 0
0 K
0 0
0 0 k
1
+k
2
k
2
0 0 k
2
k
2
_
_
_
_
u
y
1
y
2
_
_
= 0 (1.2.3)
4 Refer Chapter 2 (Vol. 2) for further application of this formulation where we have derived a 2D soil-
structure interaction model for a Turbine framed foundation.
Figure 1.2.2 Typical finite element mesh with soil springs, for a exible raft.
where
J = J
+J
1
+J
2
sum of all mass moment of inertia;
H = h
1
+h
2
= the total height of the structure.
Above formulation can very well be used in cases the foundation is significantly rigid
and can be modeled as rigid lumped mass having negligible internal deformation
5
.
However for cases where the foundation is more flexible one usually resorts to finite
element modeling of the base raft which is connected to the soil springs as shown in
Figure 1.2.2.
For the problem as shown above irrespective of the raft being modeled as a beam
or a plate the soil stiffness is directly added to the diagonal element K
ii
of the global
stiffness matrix to arrive at the over all stiffness matrix of the system.
Before we proceed further we explain the above assembly by a conceptual problem
hereafter.
Example 1.2.1
For the beamas shown in Figure 1.2.3, compute the global stiffness matrix when
supported on a spring at its mid span. Take EI as the flexural stiffness of the
beam. The spring support has stiffness @ K kN/m.
Solution:
For a beam having two degrees of freedom per node as shown in Figure 1.2.4,
the element stiffness matrix is expressed as follows.
5 A classic example is a turbine frame foundation resting on a bottom raft whose thickness is usually
greater than 2.0 meter.
L L
K
Figure 1.2.3 Spring supported beam.
2
1
4
3
Figure 1.2.4 Two degrees of freedom of a beam element.
The element matrix for such case is given by
1 2 3 4
K
ij
=
_
_
_
_
_
_
_
_
_
_
_
_
_
12EI
L
3
6EI
L
2
12EI
L
3
6EI
L
2
6EI
L
2
4EI
L
6EI
L
2
2EI
L
12EI
L
3
6EI
L
2
12EI
L
3
6EI
L
2
6EI
L
2
2EI
L
6EI
L
2
4EI
L
_
_
Assembling the element matrix for the two beams we have
[K]
g
=
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
12EI
L
3
6EI
L
2
12EI
L
3
6EI
L
2
0 0 0 0
6EI
L
2
4EI
L
6EI
L
2
2EI
L
0 0 0 0
12EI
L
3
6EI
L
2
12EI
L
3
+
12EI
L
3
6EI
L
2
+
6EI
L
2
12EI
L
3
6EI
L
2
0 0
6EI
L
2
2EI
L
6EI
L
2
+
6EI
L
2
4EI
L
+
4EI
L
6EI
L
2
2EI
L
0 0
0 0
12EI
L
3
6EI
L
2
12EI
L
3
6EI
L
2
0 0
0 0
6EI
L
2
2EI
L
6EI
L
2
4EI
L
0 0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
_
_
As Left hand support is fixed hence we have to eliminate row and column 1
and 2.
Similarly, as right hand support is hinged we have to eliminate rowand column
5 from the above when we have
[K]
g
=
_
_
_
_
_
_
_
_
24EI
L
3
0
6EI
L
2
0
8EI
L
2EI
L
6EI
L
2
2EI
L
4EI
L
_
_
with appropriate boundary conditions.
To use the spring support, the spring is now directly added to the diagonal
element of the global matrix.
Thus the combined stiffness matrix is given by
[K
]
g
=
_
_
_
_
_
_
_
_
24EI
L
3
+Ks 0
6EI
L
2
0
8EI
L
2EI
L
6EI
L
2
2EI
L
4EI
L
_
_
The above is the normal practice adapted in global assemblage of soil spring
in a finite element assembly.
We further elaborate the phenomenon with a suitable practical numerical
example.
Example 1.2.2
Shown in Figure 1.2.5 is a bridge girder across a river is resting at points A and B
on rock abutments at ends, and resting on a pier at center of the girder (point C)
A 5.0 m C 5.0 m B
Water Level
Figure 1.2.5 Bridge girder across abutments.
A C
B
1
1
2
2 3
3
4
4
5
Figure 1.2.6 Idealisation of the bridge girder ignoring soil effect.
which is resting on the soil bed of the river. The exural stiffness of the girder is
EI = 100,000 kN m
2
. Area of girder is 5.0 m
2
. The dynamic shear modulus
of soil is G = 2500 kN/m
2
. The bridge pier foundation has plan dimension of
6 m 6 m. Determine the natural frequencies of vibration of the girder consid-
ering with and without soil effect. Unit weight of concrete = 25 kN/m
3
. Mass
moment of inertia per meter run = 30 kN sec
2
m.
Solution:
The bridge girder can be mathematically represented by a continuous beam as
shown in Figure 1.2.6. Here node 2 and 4 are at the center of beam.
Thus, for beam element 1, 2, 3, and 4, we have element stiffness matrix as
[K
ij
] =
EI
L
3
_
_
_
_
_
12 6L 12 6L
6L 4L
2
6L 2L
2
12 6L 12 6L
6L 2L
2
6L 4L
2
_
_
The unconstrained combined stiffness matrix as
[K
ij
]
=
EI
L
3
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
12 6L 12 6L 0 0 0 0 0 0
6L 4L
2
6L 2L
2
0 0 0 0 0 0
12 6L 24 0 12 6L 0 0 0 0
6L 2L
2
0 8L
2
6L 2L
2
0 0 0 0
0 0 12 6L 24 0 12 6L 0 0
0 0 6L 2L
2
0 8L
2
6L 2L
2
0 0
0 0 0 0 12 6L 24 0 12 6L
0 0 0 0 6L 2L
2
0 8L
2
6L 2L
2
0 0 0 0 0 0 12 6L 12 6L
0 0 0 0 0 0 6L 2L
2
6L 4L
2
_
_
Substituting the values we have
[K] =
76800 96000 76800 96000 0 0 0 0 0 0
96000 160000 96000 80000 0 0 0 0 0 0
76800 96000 153600 0 76800 96000 0 0 0 0
96000 80000 0 320000 96000 80000 0 0 0 0
0 0 76800 96000 153600 0 76800 96000 0 0
0 0 96000 80000 0 320000 96000 80000 0 0
0 0 0 0 76800 96000 153600 0 76800 96000
0 0 0 0 96000 80000 0 320000 96000 80000
0 0 0 0 0 0 76800 96000 76800 96000
0 0 0 0 0 0 96000 80000 96000 160000
Now imposing the boundary condition that vertical displacement are zero at
1, 3, 5,
6
we have
[K] =
160000 96000 80000 0 0 0 0
96000 153600 0 96000 0 0 0
80000 0 320000 80000 0 0 0
0 96000 80000 320000 96000 80000 0
0 0 0 96000 153600 0 96000
0 0 0 80000 0 320000 80000
0 0 0 0 96000 80000 160000
Lumped mass at each node is given by M
ii
= 25 5 2.5/9.81 = 31.85
kN sec
2
/m.
Mass moment of inertia at each node is given by J
ii
= 30 1.25 = 37.5.
Thus combined mass matrix is given by
[M] =
_
_
_
_
_
_
_
_
_
_
_
_
_
_
37.5 0 0 0 0 0 0 0
0 31.85 0 0 0 0 0 0
0 0 65 0 0 0 0 0
0 0 0 31.85 0 0 0 0
0 0 0 0 37.5 0 0 0
0 0 0 0 0 31.85 0 0
0 0 0 0 0 0 65 0
0 0 0 0 0 0 0 37.5
_
_
6 We assume that since the bridge is supported on hard rock at ends, displacement at node 1 and 5
are zero.
A C
B
1
2
3
4
5
K
z
Figure 1.2.7 Idealisation of the bridge girder considering soil effect.
Considering the equation
[K] [M]
2
= 0 we have
MODE 1 2 3 4 5 6 7
Eigen value 692 1328 2684 4897 7448 7787 11722
Natural 26.30 36.44 51.80 69.97 86.59926 88.24996 108.26855
frequency
(rad/sec)
Considering the effect of soil we can construct the model as in Figure 1.2.7.
Here K
z
=
4Gr
0
1
where r
0
=
_
LxB
, Here L = B = 6.0 m
Here r
0
= 3.38 m and for G = 2500 kN/m
2
and = 0.3 K
z
= 48285.71
kN/m.
Now imposing the boundary condition that vertical amplitude at node 1 and
5 are zero (node 3 is not zero) we have
[K] =
160000 96000 80000 0 0 0 0 0
96000 153600 0 76800 96000 0 0 0
80000 0 320000 96000 80000 0 0 0
0 76800 96000 201959.1 0 76800 96000 0
0 96000 80000 0 320000 96000 80000 0
0 0 0 76800 96000 153600 0 96000
0 0 0 96000 80000 0 320000 80000
0 0 0 0 0 96000 80000 160000
The Mass matrix remains same as derived earlier.
Performing the eigen value solution we have
Modes 1 2 3 4 5 6 7 8
Eigen-values 75 692 2684 3045 7067 7448 9489 11722
Natural 8.660 26.30 51.80 56.18 84.06 86.30 97.41 108.27
frequency
(rad/sec)
Having established the fact as to how soil affects the dynamic response let us
see further what different type of soil model is possible. For design office practices
spring values considered are usually based on Richart/Wolfs model which are effec-
tively combined with structure as shown above to find out the overall response of a
system.
The example above, though it has been worked out based on beam the theory, it is
effective for any kind of structural elements like plates, shells, 8-nodded brick element
etc. Thus implementing the above in a general purpose Finite element package is
quite straight forward. For raft modeled as beam with underlain spring, the essence
of arriving at individual springs at each node is same as shown in the case of static
analysis based on influence zone
7
.
The only difference being that the nodal influence area is to be converted into an
equivalent circular area to arrive at vertical spring values. The horizontal springs are
based on the full area and are divided equally at the end.
1.2.2 What happens if the raft is f lexible?
Methodology described in previous section is usually adapted when the raft is uncon-
ditionally rigid. However there could be cases where the raft could be perfectly exible
or intermediate (i.e. somewhere between perfectly rigid and perfectly exible) when
the calculation of spring values is different than what has been mentioned in the
preceding.
Before we get into this issue the obvious query would be what is the boundary
condition for raft rigidity in terms of dynamic loading?
Unfortunately there is none, and the condition pertaining to static load still applies
8
.
Thus as explained in Chapter 4 (Vol. 1), if L is the c/c distance between the columns,
then for
L

4
the raft will behave as rigid raft
For L the raft will behave as flexible raft
For all values between /4 L , the slab behave in between rigid/flexible
in which =
4
_
kB
4E
c
I
, k = modulus of sub-grade reaction, (in kN/m
3
); B = width
of raft in meter; E
c
= modulus of elasticity of concrete (in kN/m
2
); I = moment of
inertia of the raft (in m
4
).
1.2.2.1 Calculation of spring constant for rigid raft
The rigidity of raft plays a significant role in the soil spring values connected to the
plate elements as mentioned above.
7 Refer Example 4.6.1 in Chapter 4 (Vol. 1) for further details.
8 This is not illogical for dynamic load can be conceived as a system under static equilibrium at a time t.
Thus condition of rigidity as explained in Chapter 4 (Vol. 1) should hold good.
When the raft is rigid the gross spring value is obtained based on the full raft
dimension and then are broken up into discrete values
k
z
= K
z
_
A
p
A
G
_
(1.2.4)
where, k
z
= value of discrete spring for the rigid finite element; K
z
= value of gross
spring considering the overall dimension of the raft; Ap = area of the finite element
plate, and A
G
= gross area of the raft.
1.2.2.2 Calculation of spring constant for f lexible raft
When the raft is flexible an equivalent radius within which the load gets dispersed is
first obtained from the formula
r
0
= 0.8t
s
__
E
c
G
s
_
1
1
2
c
_1
3
(1.2.5)
The gross spring value is then obtained based on this equation. Finally the discrete
spring for the finite element is obtained as
k
z
= K
z
_
A
p
r
2
0
_
(1.2.6)
where, r
0
= equivalent radius within which the load gets dispersed; E
c
= dynamic
modulus of the concrete raft; G
s
= dynamic shear modulus of the soil; = Poissons
ratio of soil;
c
= Poissons ratio of the raft, and t
s
= thickness of the raft.
A suitable problem cited hereafter elaborates the above more clearly.
Example 1.2.3
A raft of dimension 30 m 15 m is resting on a soil having dynamic shear
modulus of 35000 kN/m
2
and Poissons ratio of soil = 0.4. Determine the soil
springs for plate elements of size 2.0 m 2.0 m for finite element analysis
considering,
The raft as rigid
Considering the raft as flexible.
The thickness of the raft is 1.8 m.
Solution:
Considering the raft as rigid:
r
0
=
_
30 15
= 11.968 meter;
K
z
=
4Gr
0
1
=
4 35000 11.96
0.6
= 2790666.67 kN/m
For finite element of size 2 m 2 m discrete spring value will be
k
z
= K
z
_
A
p
A
G
_
k
z
= 2790666.67
_
2 2
30 15
_
= 24806kN/m
Thus spring values at four nodes are 6201 kN/m i.e 1/4th of the above
calculated value. When the raft is considered exible, we have:
r
0
= 0.8t
s
__
E
c
G
s
_
1
1
2
c
_
1/3
Here E
c
= 3 10
8
kN/m
2
;
c
= 0.25(say),
then r
0
= 0.8 1.8
__
3 10
8
35000
_
1 0.4
1 0.25
2
_
1/3
= 25.39 m
Thus K
z
=
4Gr
0
1
=
4 35000 25.39
0.6
= 5924333.333 kN/m
Thus for finite element of size 2 m 2 m the discrete spring value is
k
z
= 5924333.333
_
2 2
25.39
2
_
= 11701 kN/m
Thus spring values at four nodes are 2925 kN/m
It will be observed that the spring values vary considerably for the two different
approach.
1.2.2.3 What sin thou make in treating foundation
& the structure separately?
Difficult to pass a sweeping judgment for depending on the situation, the sin could be
cardinal or even trivial.
Based on a number of analysis carried out it can be stated that treating them
in isolation can result in conservative design
9
or dangerously un-conservative, thus
resulting in an unsafe structure which could be a danger to human life and property.
Having made the above statement a number of questions obviously come to
mind
10
like
1 How conservative or how susceptible the system can be ignoring the soil effect?
2 Considering soil effect (specially for FEM analysis) makes the analysis more
laborious and time consuming thus more costly is it worth?
3 My boss is a traditionalist and under project time pressure can I convince him
it is worth the effort.
4 Before doing the detailed analysis itself can I come up with a quantitative value
based on which I can assess how far this effect will be (for good or worse) and
thus convince my boss on the value addition to this effort?
5 What is the risk in terms of cost and safety if I do not do this analysis?
The questions are surely pertinent and not always very easy to answer. However with
a little bit of intelligent analysis it is not difficult to come up with a logical conclusion
on this issue.
We try to explain. . .
The obvious answer is it essentially could modify the natural frequency/time period
of the system
11
.
What needs to be evaluated is what is the effect of this modified time period
on the system compared to, if the soil is ignored (i.e. it is considered a fixed base
problem).
The two classes of problems under which dynamic soil structure interaction plays a
significant role are
Systems subjected to vibration from machines like block foundations (machine
foundations for pumps, compressors, gas turbines etc), frame foundations (turbine
foundations, compressor foundations, boiler feed pump foundations)
Structures subjected to earthquake.
For the machine foundation source of disturbance is the machine mounted on the
system the dynamic waves generated are transferred from the machine via structure
to the surrounding soil-which is an infinite elastic half space.
While for earthquake the source of disturbance is the ground itself where elastic
waves generate within the soil mass due to the tectonic movement/rupture of the rock
mass (geologically known as faults).
It is obvious that soil will affect these two classes of problem in different ways.
For instance a machine supported on a frame- the frame is usually made signi-
cantly stiff to ensure stress induced in it are not signicant and are generally made
9 For big projects which could mean a cost over run.
10 Speciallyfor freshmannewtothetopicwhohas got aleadengineer andadepartmental HEADtoanswer to.
11 We say the word could as because the extent of modification will depend upon the shear wave velocity
of the soil. We had shown previously the boundary limits within which it can have a significant effect.
over tuned for medium or low frequency machine when considered as a fixed based
problem.
But in reality considering the soil effect, the foundation may actually be under tuned
or even hover near the resonance zone when the underlying soil participates in the
vibration process. Thus the amplitude of vibration could significantly vary than the
calculated one.
Generically, considering the soil stiffness will make the system more flexible then
a fixed base problem and it can be intuitively deduced that though the stress might
remain within the acceptable level the amplitude of vibration will be more and could
well exceed the acceptable limit which might have secondary damaging effect to the
machine and its appurtenances.
For earthquake the effect is quite different. In this case the structure resting on the
site can be visualized as a body resting on an infinite elastic space (similar to a ship
floating in sea). Due to rupture in the fault as waves dissipate in all direction the soil
mass starts vibrating at its own fundamental frequency known as the free field time
period of the site.
In such case the earthquake acts as an electronic lter and tries to excite the super-
structure resting on it to its own fundamental frequency and suppressing or even
eliminating other modal frequencies
12
. Thus if the fixed base frequency of the struc-
ture matches the fundamental frequency of the soil strata on which it is resting, they
are in resonance and catastrophe could well be a reality.
Before dwelling into the mathematical aspect of it we further substantiate the above
statement by some real life facts and observations.
Dowrick (2003) reports that in the Mexico earthquake in 1957 extensive damage
occurred to the buildings that were tall and were found to be resting on alluvium soil
of depth >1000 m. In 1967, the Caracas earthquake showed identical result where the
tall structures underwent extensive damage and those were resting on deep alluvium
soil overlying bedrock. In 1970 earthquake at Gediz in Turkey a part of a factory
was demolished in a town about 140 Km from the epicenter while no other build-
ings in the town underwent any damage! Subsequent investigation revealed that the
fundamental period of the building matched the free field time period of the site. The
Caracas earthquake as cited earlier also showed a distinctive pattern where medium
rise buildings (59 storeys) underwent extensive damage where depth to bedrock was
less than 100 m, while buildings over 14 stories were damaged where the depth to
bedrock was greater than 150 meters.
Let us see why such thing happened and how does it substantiate the free field time
period phenomenon as stated earlier.
The free field time period of a site is given by the equation
T
n
=
4H
(2n 1)V
s
(1.2.7)
12 It can be visualized as a giant hand trying to shake a small body resting on it. Since the body is much
weaker to the giant it tries to follow the same phase of vibration as the soil medium.
0
20
40
60
80
100
120
0
0
.
1
5
0
.
3
0
.
4
5
0
.
6
0
.
7
5
0
.
9
1
.
0
5
1
.
2
1
.
3
5
1
.
5
Depth of soil/Shear wave Velocity
N
u
m
b
e
r

o
f

S
t
o
r
i
e
s
n for RCC frame
n for steel frame
Figure 1.2.8 Limiting value of storeys for frames.
where, T = time period of the free field soil (i.e. without the structure); H = depth of
soil over bedrock
13
; n = number of mode; and V
s
= shear wave velocity of the soil.
Thus based on the explanation above it can be argued that if the fixed base frequency
of structure is in the close proximity of the free field time period of the site the structure
may be subjected to significant excitation.
The above statement can be extended to a very interesting hypothesis.
If we equate the free field time period of the site to the fixed base time period of
the structure we can arrive at some limiting design parameters which can result in
significant dynamic amplification and which should be avoided at the very out set of
planning of the structure.
For instance as per IS-1893 RCC moment resisting frames with no infill brick work,
the fundamental time period is given by
T = 0.075h
0.75
(1.2.8)
Thus equating it to fundamental free field time period of the site we have
0.075h
0.75
=
4H
V
s
, which gives h =
_
160H
3V
s
_
4/3
(1.2.9)
Considering 1 floor is of height 3.3 m, we can further simplify the equation to
n = 0.303
_
160H
3V
s
_
4/3
(1.2.10)
13 Here bedrock is perceived as that level where the shear wave velocity of soil is greater or equal to
600m/sec.
The curves shown in Figure 1.2.8 give limiting stories for RCC and steel frames for
which resonance can occur in a structure during an earthquake as per IS-1893
14
for
various values of H/V
s
.
Let us now probe the problem a bit more based on a suitable numerical problem.
Example 1.2.4
A particular site has been found to consist of 100 m soil overlying bedrock
when the shear wave velocity of the soil is 222.22 m/sec. Find the limiting
number of stories of height 3.3 meter for an RCC frame for which resonance
can occur. What would be resonance story if the depth of the overlying soft soil is
only 30 m.
Solution:
Based on above data H/V
s
= 100/222.2 = 0.45 when H = 100 m.
As per the chart as shown above the limiting story for which resonance can
occur is 18.
Thus for a 18 storied building resonance can very well occur and the strategy
would be to build the building at least ()25% away i.e. either it should be 23
storied or more or 14 storied or less.
When the depth of soil is only 30 m, H/V
s
=
30
222.2
= 0.135.
Based on the above chart the limiting story height is roughly 4-storey only.
Thus to avoid resonance the building should be either more than 5-storey or less
than 3-storey.
The above problem well explains the phenomenon as to what happened in the
Mexico and Turkey earthquakes and perhaps challenges the myth quite prevalent in
many design offices that for one or two storied building earthquake is not important
and can well be ignored.
It is evident from the above problem that the response depends on the depth of soil
on which it is resting and depending on the free field time period the response can
either amplify or attenuate. It can well affect even a one storied building.
The chart in Figure 1.2.9 shows limiting story height of buildings with infill brick
panels and all other type of frames as per IS 1893 for different width of building
varying from 10 meter to 50 m
15
.
The above theory is though explained in terms of building, can very well be adapted
for any class of structure for which it is possible to establish the fundamental time
period expression.
14 In this case time period for steel frame is considered as T = 0.085(h)
0.75
as per IS-1893.
15 Time period of the fixed base structure considered as T =0.09h/(d)
0.5
as per, Indian Standards Institution
(1984, 2002). Indian Standard Criteria for Earthquake Resistant Design of Structures, IS: 1893
(Part 1), ISI, New Delhi, India.
0
5
10
15
20
25
30
35
40
0
0
.
0
8
0
.
1
5
0
.
2
3
0
.
3
0
.
3
8
0
.
4
5
0
.
5
3
0
.
6
0
.
6
8
0
.
7
5
Depth of soil/Shear wave velocity
N
u
m
b
e
r

o
f

s
t
o
r
y
n for d=10m
n for d=15m
n for d=20m
n for d=25m
n for d=30m
n for d=35m
n for d=40m
n for d=45m
n for d=50m
Figure 1.2.9 Limiting story for building with inll brick panel.
Having assessed the resonance criteria and making sure at planning stage that the
two periods do not match one would still like to quantify the combined time period
of the overall soil structure system and assess whether there is any amplification or
attenuation of the earthquake force.
Before plunging into detailed analysis based on FEMor otherwise it would be useful
to have a rough estimate as to howmuch the underlying soil affects the overall response.
Veletsos and Meek (1974) has given a very useful expression based on which it is
possible to estimate the modified time period of a structure, and is given by
T = T
_
1 +
k
K
x
_
1 +
Kx
h
2
K
_
(1.2.11)
where

T = modified time period of the structure due to the soil stiffness, T = time
period of the fixed base structure,

k = stiffnessof the fixed base structure @
4
2
W
gT
2
,
K
x
, K
= horizontal and rotational spring constant of the soil (IS-1893),

h = effective
height or inertial centroid of the system, and, W = total weight of the structure.
Based on the above expression one can immediately arrive at a rough estimate as
to how strong could be soil response at the very outset of a design. We elaborate the
above based on two suitable problems hereafter.
Example 1.2.5
An RCC Chimney 150 meter in height has a uniform cross section area of A
c
=
8.5 m
2
and moment of inertia I = 92.5m
4
. Evaluate the base moment and
shear under earthquake considering the problem as fixed base as well as the
soil effect. The structure is located in zone IV as per IS 1893. The structure
is supported on raft of diameter 18 meter. The soil has a dynamic shear wave
velocity of 120 m/sec and unit weight of 19 kN/m
3
. Consider 5% damping
for the analysis.
16
The grade of concrete used is M30 having dynamic E
conc
=
3.12 10
8
kN/m
2
.
Solution:
Height of the structure = 150 m; Area of shell = 8.5 m
2
Weight of chimney = 150 8.5 25 = 31875 kN (unit weight of conc. =
25 kN/m
3
)
Radius of gyration of the chimney =
_
I/A =
_
92.5/8.5 = 3.298 m
Thus slenderness ratio H/r =
150
3.298
= 45.4.
As per IS 1893 C
T
= 82.8.
As per IS 1893 time period of a fixed base chimney is given by, T =
C
T
3.13
_
WH
E
c
A
c
.
where, W =weight of chimney in N; E
c
=Dynamic Youngs modulus of conc.
@ 3 10
8
kN/m
2
Thus, T =
82.8
3.13
_
31875 150
3.12 10
8
8.5
= 1.13 sec
For 5% damping referring to chart in IS-1893 we have Sa/g = 0.10.
Thus the horizontal seismic coefficient is given by,
h
= IF
o
Sa
g
.
Here (Soil foundation factor) = 1.0 for chimney resting on raft, I = 1.5
Importance factor, F
o
= Zone factor @ 0.25 for zone IV.
This gives
h
= 1.0 1.5 0.25 0.10 = 0.0375
The Bending moment and shear force are given by,
M =
h
W

H
_
0.6
_
x
H
_1
2
+0.4
_
x
H
_
4
_
and V = C
v
h
W
__
5x
3H
_
2
3
_
x
H
_
2
_
16 In this case it is presumed that reader has some idea of how to use the code IS-1893 or is at least
familiar with it.
Here, C
v
= a coefficient depends on the slenderness ratio and as per the present
problem is 1.47 as per IS 1893;

H = height of c.g. of the structure above base
@ 75 meter for the problem; x = distance from the top.
Substituting the appropriate values, we have
M = 0.0375 31875 75 [0.6(1.0)
1
2
+0.4(1.0)
4
] = 89648 kN/m
V = 1.47 0.0375 31875 [(5/3) (2/3)] = 1757 kN.
Considering the soil effect we have the dynamic shear modulus of soil, G = v
2
s
Or G = (19/9.81) 120 120 = 27890 kN/m
2
.
With radius of raft = 9.0 m, K
x
=
8GR
2
, = Poissons ratio of the soil
considered as 0.35,
K
x
=
8 27890 9
2 0.35
= 1217018.2 kN/m
And K
=
8GR
3
3(1 )
which gives, K
=
8 27890 9
3
3(1 0.35)
= 83412554 kN/m
The xed base stiffness of chimney is given by
k =
4
2
W
gT
2
=
4
2
31875
9.81 1.13
2
= 100458 kN/m
Substituting the above numerical values in Veletsos equation we have
T = T
_
1 +
k
K
x
_
1 +
Kx
h
2
K

T = 1.13
_
1 +
100458
1217018.2
_
1 +
1217018.2 75
2
83412554
_
= 3.2 sec
As per IS 1893 for T = 3.2 sec, Sa/g = 0.05, which gives
h
=
0.0375
0.10
0.05 = 0.01875 (By proportion)
The base moment and shear are given by
M =
89648
0.0375
0.01875 = 44824 kN m;
and V =
1757
0.0375
0.01875 = 878.5 kN
The results are compared hereafter
17
Case Moment Shear Remarks
Without soil 89648 1757 Reduction in moment and shear by 34%
With soil 44824 878.5
The problem shows a clear attenuation of the response.
We show another example hereafter.
Example 1.2.6
Shown in Figure 1.2.10 is a horizontal vessel having empty weight of 340 kN
and operating weight of 850 kN is placed on two isolated footing of dimension
8.5 m 3 m. The center to center distance between the two foundations is 5.5
meter. The center line of vessel is at height (H
f
) of 4.5 meters fromthe bottomof
the foundation. Thickness of the foundation slab is 0.3 meter. The RCCpedestal
is of width 1.0 meter, length 6 meter having height of 3.45 meter. The shear wave
velocity of the soil is 200 m/sec having Poissons ratio of 0.3. Allowable bearing
capacity of the foundation is 150 kN/m
2
. Calculate the design seismic moment
considering the effect of soil and without it, if the site is in zone III as per IS-1893.
Consider soil density @ 18 kN/m
3
and unit weight of concrete as 25 kN/m
3
?
Solution:
Plan are of footing = 8.5 3 = 25.5 m
2
Equivalent circular radius =
_
Af
=
_
25.5
= 2.849 m
Moment of inertia of the foundation about X-axis
_
1
12
BL
3
_
=
1
12
38.5
3
=
153.5313 m
4
Moment of inertia of the foundation about Y-axis
_
1
12
LB
3
_
=
1
12
8.53
3
=
19.125 m
4
17 Without an elaborate analysis it could be an effective calculation to convince the boss that
you can save some money and the worth of a dynamic soil-structure interaction analysis.
y
p
y
p

H
f
H
p
D
s
W
p
L
p

Y
L
f
X
B
f
B
f
L
s
Figure 1.2.10 A horizontal vessel.
Equivalent circular radius about X axis =
1
2
_
64I
xx
_
0.25
= 3.739183 m
Equivalent circular radius about Y axis =
1
2
_
64I
yy
_
0.25
= 2.221 m
Mass density of soil () =
18
9.81
= 1.835 kN/m
3
Dynamic shear modulus = v
2
s
= 1.835 200
2
= 73400 kN/m
2
Lateral spring in X and Y direction =
32Gr
0
(1 )
7 8
= 1018306 kN/m
Rocking spring about X axis =
8Gr
3
x
3(1 )
= 14627886 kN/m
Rocking spring about Y axis =
8Gr
3
y
3(1 )
= 3063462 kN/m
Moment of Inertia of the pedestal about X axis =
1
12
1 6
3
= 18 m
4
Moment of Inertia of the pedestal about X axis =
1
12
6 1
3
= 0.5 m
4
Structural stiffness of pedestal about X axis =
3EI
x
L
3
=
3 3.2 10
8
18
3.45
3
=
3.95 10
8
kN/m
Structural stiffness of pedestal about Y axis =
3EI
y
L
3
=
3 3.2 10
8
0.5
3.45
3
=
1.10 10
7
kN/m
Contributing mass for the vessel empty case =
340
2 9.81
= 17.33 kN-sec
2
/m
Contributing mass for the vessel operating case =
850
2 9.81
=
43.323 kN-sec
2
/m
Contributing uniformly distribute load for the pedestal = 25.5 25/9.81 =
64.98 kN/m
The mathematical model for the pedestal thus constitute of a beam element
(pedestal) having a mass lumped at its tip (mass contribution from the vessel) is
shown in Figure 1.2.11.
The time period of such fixed base model is given by (Paz 1991)
T = 2
_
(M +0.25m
b
)
K
m
b
M
Figure 1.2.11 Mathematical model for the pedestal.
and the modified time period considering soil effect is given by
T = T
_
1 +
k
K
x
_
1 +
Kx
h
2
K
_
The time periods and the corresponding Sa/g values as per IS-1893 for 5%
damping are as show hereafter.
Time period Time period Time period Time period
(vessel empty) (vessel empty) (operating) (operating)
about X about Y about X about
Sl no Case direction direction direction Y direction
1 Without soil 0.0018 0.011 0.0024 0.015
2 With soil effect 0.0579 0.1057 0.0771 0.1408
Corresponding Sa/g value is given by
Sa/g Sa/g Sa/g Sa/g
about about about about
Sl no Case X direction Y direction X direction Y direction
1 Without soil 0.1000 0.120 0.1000 0.130
2 With soil effect 0.2 0.2 0.2 0.2
Base shear as per IS 1893 considering Importance factor as 1.0 for vessel empty
case and 1.25 for vessel in operation case we have
Shear Shear Shear Shear
1 Without soil 21.25 25.5 26.5625 34.53125
2 With soil effect 42.5 42.5 53.125 53.125
The moment at the foundation level is given by
Moment Moment
(vessel (vessel Moment Moment
empty) empty) (operating) (operating)
1 Without soil 100.9375 121.125 126.1719 164.0234
2 With soil effect 201.875 201.875 252.3438 252.3438
This case clearly shows an amplification of force considering the soil effect.
Having established a basis of how to evaluate the coupled soil-structure interaction
under dynamic loading we now extend the above theory to system with multi degree
of freedom where the theory can be very well be adapted as a powerful tool for a
detailed yet economic dynamic analysis.
1.3 A GENERALISED MODEL FOR DYNAMIC SOIL STRUCTURE
INTERACTION
In this section we present a generalised model for dynamic soil-structure interaction.
Though the model is developed based on 3D frames can also be adapted for a three
dimensional Finite Element analysis.
1.3.1 Dynamic response of a structure with multi degree
of freedom considering the underlying soil stiffness
We had shown earlier that for a single degree of freedom system the modified time
period of a structure considering the soil effect is given by
T = T
_
1 +
k
K
x
_
1 +
Kx
h
2
K
_
(1.3.1)
Both ATC(1982) and FEMAhas adapted this formula for practical design office usage
(Veletsos & Meek 1974, Jennings & Bielek 1973). The nomenclatures of the formula
are as explained earlier. Now squaring both sides of the above equation we have
T
2
= T
2
_
1 +
k
K
x
+
kh
2
K
_
(1.3.2)
Considering the expression T =
2
we have
4
2

2
=
4
2
2
_
1 +
4
2
m
T
2
K
x
+
4
2
m
h
2
T
2
K
_
or
4
2

2
=
4
2
2
_
1 +
m
2
K
x
+

2
m
h
2
K
_
(1.3.3)
Simplifying and expanding the above we have
1

2
=
1
2
+
m
K
x
+
m
h
2
K
which can be further modified to

1

2
=
1
2
+
1
2
x
+
1
(1.3.4)
which gives the modified natural frequency relation for a system with single degree of
freedom. This formulation has also been shown in, Kramer, S. (2004).
Now considering generically =
_
k/m we have
m
k
e
=
m
k
+
m
K
x
+
m
h
2
K
, or,
1
k
e
=
1
k
+
1
K
x
+
h
2
K
(1.3.5)
where k
e
= equivalent stiffness of the soil structure system having single degree of
freedom.
We shall extend the above basis to multi degree of freedom hereafter (Chowdhury
and Dasgupta 2002).
1.3.2 Extension of the above theory to system
with multi degree of freedom
A 3-D frame shown in Figure 1.3.1, is considered for the presentation of the proposed
method. The frame structure has n degrees-of freedom and subjected to soil reactions
in the form of translational and rotational springs.
For a system having n degrees of freedom the above equation can be written in the
form
[M]
nn
[K
e
]
nn
=
[M]
nn
[K]
nn
+
[M]
nn
K
x
+
[M]
nn
_
h
2
_
nn
K
(1.3.6)
Y
X
O
Z
K
x

K

-- mass points
Figure 1.3.1 A 3-D Frame having multi-degree-of freedom with representative foundation spring.
Here, [K
e
] = equivalent stiffness matrix of the soil structure systemof order n, [M] =
a diagonal mass matrix of order n having masses lumped at the element diagonals,
[
h
2
] = radius vectors of the lumped masses to the center of the foundation springs
of order n, K
x
, K
= translation and rotation spring stiffness of the total foundation

system represented by a unique value.
Taking out the common factor [M], we have
[I]
[K
e
]
=
[I]
[K]
+
[I]
K
x
+
_
h
2
_
K
(1.3.7)
where, [I] = identity matrix of order n having its diagonal element as 1.
or [I][K
e
]
1
= [I][K]
1
+ [I/K
x
] + [h
2
/K
]
[Fe] = [F] + [F
x
] + [F
] (1.3.8)
where [F] = Flexibility matrix of the system with suffixes as mentioned earlier for
stiffness matrices.
Once the flexibility matrix of the equivalent soil structure system is known the
stiffness matrix may be obtained from the expression
[K
e
] = [Fe]
1
(1.3.9)
Now knowing the modified stiffness matrix the eigen solution may be done based
on the usual procedure of
[K
e
] [] = [
e
] [M] []. (1.3.10)
1.3.3 Estimation of damping ratio for the soil
structure system
While calculating the damping ratio, the normal process is to guess a damping ratio
for the structure like 25%, and consider the same damping ratio for all the mode and
obtain the value of Sa/g value for the particular structure per mode corresponding to
the time period based on the curves given in IS-1893.
The basis of assuming this damping ratio is purely judgmental and is dependent
on either the experience of the engineer, recommendation of codes, or based on field
observations on the performance of similar structure under previous earthquakes.
When the effect of soil is neglected it is possible to obtain the material damping ratio
of the structure depending on what constitute the material like steel, RCC etc.
However when the whole system is resting on soil an analyst is usually faced with
the following stumbling blocks for which clear solution is still eluding us specially for
modal analysis in time domain.
The difficulties encountered can be summarised as follows
The damping matrix of the coupled soil-structure system becomes non-
proportional for which the damping matrix does not de-couple based on
orthogonal transformation.
As the damping ratio of the structure and the soil could be widely varying it
becomes difficult to assess a common damping ratio which would affect the soil
as well as structure.
Even after elaborate FEM modelling of the soil, the damping ratio contribution
per mode still remains guess estimation at the best.
We present hereafter a method by which one can estimate approximately the contri-
bution of combined soil structure systemunder earthquake for various modes, without
resorting to an elaborate modelling of the soil itself.
We only estimate the contribution of the soil damping to the structural systemwhose
response we are interested in. The estimation is surely approximate but at least gives
a rational mathematical basis to arrive at some realistic damping value rather than
guessing a damping value at the outset and presuming that it remain same for each
mode, specially for coupled soil structure systemwhere widely varying damping for the
foundation and structure makes it difficult for the analyst to arrive at unified rational
value applicable to the system.
1.3.4 Formulation of damping ratio for single degree
of freedom
Neglecting the higher order, the material damping ratio for a soil structure system
having single degree of freedom is given (Kramer 2004) by

2
=

2
+

x
2
x
+

(1.3.11)
where,

= damping ratio of the equivalent soil structure system; = damping ratio
of the fixed base structure;
x
= horizontal damping ratio of the soil, where
x
=
0.288
B
x
and B
x
=
(78)mg
32(1)
s
r
3
x
, where m = total mass of the structure and foundation; g =
acceleration due to gravity; = Poissons ratio of the soil;
s
= mass density of the
soil; r
x
= Equivalent circular radius in horizontal mode;
= damping ratio of the

soil in rocking mode
x
=
0.15
(1+B
and B
=
0.375(1)J
s
r
5
; and J
= mass moment
of inertia of the foundation and the structure.
Converting the damping ratio equation to stiffness-mass basis we have
m
k
e
=
m
k
+
m
x
K
x
+
mh
2
or

k
e
=

k
+

x
K
x
+
h
2

= k
e
_
k
+

x
K
x
+

_
(1.3.12)
For very high value of K
x
and K
k
e
k when

.
1.3.5 Extension of the above theory to systems
with multi-degree freedom
On extending the above to multi degree of freedom of order n, we have
[
][M]
nn
[K
e
]
nn
=
[ ][M]
nn
[K]
nn
+
[
x
][M]
nn
K
x
+
[
][M]
nn
[h
2
]
nn
K
] = [K
e
]{[ ][F] + [
x
][F
x
] + [
][F
]} (1.3.13)
[
] = Damping ratio matrix of the combined soil structure system having n number
of modes.
It is to be noted that [
] is non-proportional and not a diagonal matrix, and based

on the matrix operation as shown above has off-diagonal terms.
A study on the parametric effect shows that [
] becomes nearly a diagonal matrix

(i.e. the off diagonal terms vanishes or approaches zero) when damping ratio of the
structure and the soil foundation system are nearly equal.
However, when the damping ratio are widely varying the off diagonal terms do not
vanish however there magnitudes are relatively smaller than the diagonal terms (
ii
)
which has the most dominant effect on the system.
Thus if it is possible to arrive at a foundation layout where the damping ratio of the
structure and foundation are closely spaced considering the diagonal terms as modal
damping ratio per mode is quite correct.
Even when the off diagonal term exists due to widely varying values for practical
design engineering purpose considering the
ii
term of damping ratio matrix is realistic
for it gives a reasonably rational basis of estimation of the damping ratio per mode
rather than guessing a value based on gut feeling.
We explain the above theory based on suitable example hereafter
Example 1.3.1
Shown in Figure 1.3.1 is a three storied steel frame subjected to dynamic forces.
The damping ratio for steel is found to vary between 2 to 5%. Determine
The fixed base natural frequencies of the structure.
The fixed base eigen-vectors.
Modified natural frequency with foundation stiffness.
Modified eigen.
Take K
x
= 35000 kN/m and K
= 50000 kN/m for the soil-foundation.

Analyse the floor shears for earthquake based on IS-1893 Zone III for
Fixed base.
Considering the soil effect.
G H X
3
3000
E F X
2
3000
C D X
1
3000
A B
Figure 1.3.2
Here,
1 K
AC
= K
DB
= 1.5 10
3
kN/m M
GH
= 200 kN sec
2
/m
2 K
CE
= K
DF
= 1.0 10
3
kN/m M
EF
= 400 kN sec
2
/m
3 K
EG
= K
FH
= 0.75 10
3
kN/m M
CD
= 400 kN sec
2
/m
Solution:
The stiffness and mass matrix is given by
[K] =
_
_
5000 2000 0
2000 3500 1500
0 1500 1500
_
_
and [M] =
_
_
400
400
200
_
_
Based on Figure 1.3.2, we have found earlier that
1
=

1.6426 = 1.281 rad/sec;
2
=

10.00 = 3.162 rad/sec;
3
=
17.104 = 4.135 rad/sec.

Thus the time periods for the fixed base structure is given by
18
T
1
= 4.97 sec, T
2
= 1.987 sec, T
3
= 1.52sec
The mode shapes or the eigen-vectors are
[] =
_
_
1.00 1.0 1.0
2.1715 0.5 0.9208
2.7816 1.50 0.719
_
_
Normalised eigen vectors
[
i
] =
_
_
0.01615 0.03244 0.0344512
0.0350718 0.01622 0.03172
0.04493 0.02433 0.02477
_
_
Calculation for the combined soil-structure system
Here stiffness matrix of the fixed base structure
[K] =
_
_
5000 2000 0
2000 3500 1500
0 1500 1500
_
_
which on inversion gives
[F] =
_
_
0.000333 0.000333 0.000333
0.000333 0.000833 0.000833
0.000333 0.000833 0.003145
_
_
[F
x
] =
_
_
1/35000 0 0
0 1/35000 0
0 0 1/3500
_
_
=
_
_
2.85714 0 0
0 2.85714 0
0 0 2.85714
_
_
10
5
[h
2
] =
_
_
9 0 0
0 36 0
0 0 81
_
_
18 You can check the value by any of the method as explained in Chapter 5 (Vol. 1) for eigen value
analysis.
Thus
[F
] =
_
_
9/50000 0 0
0 36/50000 0
0 0 81/50000
_
_
=
_
_
0.00018 0 0
0 0.00072 0
0 0 0.00162
_
_
As [Fe] = [F] + [F
x
] + [F
] we have
[Fe] =
_
_
0.000542 0.000333 0.000333
0.000333 0.001581905 0.000833
0.000333 0.0008333 0.001982
_
_
which is combined flexibility matrix of the soil structure system.
Inversion of the above flexibility matrix gives
[K
e
] =
_
_
2195.19 344.34 224.4212
344.34 804.662 306.223
224.42 306.223 671.0682
_
_
The above gives the combined stiffness matrix for structural system consider-
ing the soil compliance
19
.
Thus based on the above modified stiffness matrix and mass matrix as
[M] =
_
_
400
400
200
_
_
We have, based on eigen solution,
1
=

1.2163 = 1.10286 rad/sec;
2
=
3.9666 = 1.9916 rad/sec;

3
=

5.8255 = 2.4136 rad/sec.
Thus the time periods for the combined soil-structure system is given by
T
1
= 5.697 sec, T
2
= 3.154sec, T
3
= 2.603sec
19 Watch the numbers. . . . . it is symmetric and is completely different than when you add the
springs directly to the diagonal. This matrix has no rigid body mode and can be used directly
for static analysis too.
Moreover if we take K
x
and K
very high the K

e
converges to the fixed base matrix K.
Normalised modified eigen vectors considering soil stiffness is given by
[
i
] =
_
_
0.013 0.0479 0.00589
0.0409 0.00772 0.0276
0.0360 0.0169 0.05835
_
_
Calculation of modal damping
Considering, = 5% for the structure,
x
= 10% for the soil in translation
mode,
= 15% for the soil in rocking mode

We have, [
] = [K
e
]{[ ][F] + [
x
][F
x
] + [
][F
]}
Substituting the values as mentioned and calculated above we have
[
] =
_
_
0.092 0.028 0.020
0.007 0.10908 0.028
0.002 0.0126 0.1115
_
_
It will be seen that that the main diagonal terms are dominant and
can be considered as the modal damping ratio contribution for each
mode.
Suppose we had closely spaced damping data like = 5% for the structure;
x
= 6% for the soil in translation mode;
= 5.5% for the soil in rocking

mode, the modal damping matrix reduces to
[
] =
_
_
0.0525 0.0015 0.001016
0.0004 0.05312 0.00144
0.00014 0.00066 0.05315
_
_
When the matrix become practically diagonal dominant with off diagonal
terms having very low values.
Thus for the present problem may be considered as
1
= 9.2% for
first mode,
2
= 10.9% for second mode; and
3
= 11.1% for the third
mode.
Calculation of earthquake force fixed base structure
m
1
m
1
m
2
1

2
m
2
m
2
2

3
m
3
m
2
3
400 0.01615 6.46 0.104329 0.03244 12.976 0.420941 0.03445 13.7804 0.47475407
400 0.03507 14.028 0.491962 0.01622 6.488 0.105235 0.01372 5.488 0.07529536
200 0.04493 8.986 0.403741 0.02433 4.866 0.118389 0.02477 4.954 0.12271058
29.474 1.000032 14.598 0.644565 13.2464 0.67276001
Modal mass participation factor
1
=
29.474
1.000032
= 29.47306 for the first mode,
2
=
14.598
0.644565
= 22.64777 for the second mode,
3
=
13.2464
0.67276
= 19.689 for the third mode.
Assuming 5% damping for the structure we have,
Mode Time period (secs) Sa (m/sec
2
) Remarks
1 4.9 0.4905 Sa value obtained from the chart
given in IS-1893 for 5% damping
2 1.98 0.6867 Do
3 1.52 0.7848 Do
For zone III:
K = 1.0, = 1.0, I = 1.2F
0
= 0.2 as per the code
Thus base shear is given by; V =
3
i=1
K I, F
0

i
S
a
m
i
i
Substituting data on the above formula we have
Mode Base shear V Remarks
1 102 Fixed base case
2 5.45
3 4.91
Calculation for coupled soil-structure interaction.
m
1
m
1
m
2
1

2
m
2
m
2
2

3
m
3
m
2
3
400 0.013 5.2 0.0676 0.0479 19.16 0.917764 0.006 2.4 0.0144
400 0.041 16.4 0.6724 0.0077 3.08 0.023716 0.0276 11.04 0.304704
200 0.036 7.2 0.2592 0.0169 3.38 0.057122 0.0583 11.66 0.679778
28.8 0.9992 12.7 0.998602 3.02 0.998882
Modal mass participation factor
1
=
28.8
0.9992
= 28.8231 for the first mode,
2
=
12.7
0.998602
= 12.717 for the second mode, and
3
=
3.02
0.9988
= 3.0233 for the third mode.
Modal damping for each mode, as calculated earlier.
Mode Damping Time (sec) Sa (m/sec
2
) Remarks
1 9.2% 5.7 0.343 Calculated from curve based on interpo-
lation corresponding to 9.2% damping
2 10.9% 3.2 0.294 Do- with 10.9% damping
3 11.15% 2.6 0.245 Do- with 11.15% damping
Calculation for Base shear
Base shear for the frame with coupled soil-structure interaction is given by
Mode Base shear V Remarks
1 68.4 Couple soil-foundation system
2 11.4
3 0.537
Calculation of storey forces
The storey forces for the two cases are calculated hereafter
Coupled soil structure system Fixed base
Base Base Base Base Base Base
shear shear shear shear shear shear
Storey m h mh
2 mh
2
3
i=1
mh
2
mode 1 mode 2 mode 3 mode 1 mode 2 mode 3
1st 400 3 3600 0.10526 7.20 10
+00
1.20 10
+00
5.66 10
02
1.08 10
+01
5.74 10
+00
5.17 10
+00
2nd 400 6 14400 0.42105 2.88 10
+01
4.80 10
+00
2.26 10
01
4.31 10
+01
2.29 10
+01
2.07 10
+01
Top 200 9 16200 0.47368 3.24 10
+01
5.40 10
+00
2.55 10
01
4.84 10
+01
2.58 10
+01
2.33 10
+01
Comparison of results
Time period
Structure type T1 T2 T3
Fixed base structure 4.9 1.987 1.52
Soil-structure interaction 5.697 3.154 2.603
The time periods are increasing with introduction of soil springs as predicted
at the outset.
Acceleration
Structure type Mode 1 Mode 2 Mode 3
Fixed base structure 0.4905 0.6867 0.7848
The acceleration decreases with soil-structure effect in this case
Damping
Fixed base structure 5% 5% 5%
Soil-structure interaction 9.2% 10.9% 11.15%
Damping constant for all mode for fixed base case varies with mode for
coupled analysis but is neither 5% min. nor 15% maximum but somewhere
in-between which is quite logical.
Base Shear (kN)
Fixed base structure 102 54.5 49.1
A signicant reduction in base shear, considering the soil effect, though
conceptually it can be predicted that amplitude of vibration will increase.
Shear Force per floor
Fixed Coupled Fixed Coupled Fixed Coupled
base with soil base with soil base with soil
Modes 1 1 2 2 3 3
Storey
1 10.8 7.2 5.74 1.2 5.17 0.0056
2 43.1 28.8 22.9 4.8 20.7 0.226
Top 48.4 32.4 25.8 5.4 23.3 0.255
Significant variation in floor shears per mode.
Based on the above example it can be concluded that
The major advantage with this technique is the calculation of the time period
without resorting to an elaborate modelling of the soil. Two representative spring
value for the foundation is capable of modifying the stiffness of the super-structure
having any conceivable degree of freedom.
This cuts down significantly the modelling as well as the cost of computation.
No rigid body motion exists.
Stiffness matrix of the soil structure system is symmetric and real.
The structure can be discretized to as many degrees of freedoms one choose to
select.
Beam, plates, shell, bricks anything can be used to model the super structure
system thus do not generically violate the procedures followed for FEM analysis
of the superstructure.
Since the matrix has no rigid body mode may be also be used directly for calculating
the static response too. No additional computational effort is required.
Though approximate, furnishes a rational basis of estimating the modal damping
ratio per mode for the coupled soil structure-system.
The results are logical and in general satisfies the trend as observed based on more
rigorous analysis based on complex damping and eigen value problem (where a
matrix of order n n gets inflated to the order 2n 2n thus adding to the cost of
computation).
1.3.6 Some fallacies in coupling of soil and structure
(Chowdhury 2008)
You will observe here that we had advocated two types of coupling of soil spring, one
vide Equation (1.1.3) where the soil spring is directly added to the diagonal stiffness
element of the structural matrix, meaning thereby that it is a parallel connection and
the other by Equation (1.3.7) which shows that the spring are in series.
The first method has developed from the theory of nodal compatibility and is a very
popular technique in practice for the root of its development is in the realms of matrix
analysis of structure and can very well be adapted in commercially available software.
While the second formulation is developed in the frequency domain analysis as
suggested by Veletsos for a harmonic oscillator having single degree of freedomcoupled
to a translational and rocking spring.
The question that remains as to which one is more realistic and gives the true
interaction of the soil with structure especially when we model the soil as boundary
springs.
One of the major flaws in parallel spring model is, as the boundary elements are
discrete and not a continuum it only gives a local effect and also affects the structural
node only locally.
The intention here is not to challenge or shock the structural engineers who have
been doing this for ages. But putting on the hat of a theoretical physicist and probing
this formulation a bit more it comes up with some very interesting result.
Let us imagine that the beam in Figure 1.1.1 is made of RCC of say dimension
450900 supported on a compliant foundation where the soil is modeled as a spring.
Now we put a motor on the beam which gives some dynamic force Psin w
m
t-say. We
want to find out the dynamic response of the beam. The problem shows no ambiguity
for the beam along with the soil spring vibrates with natural frequencies that can be
obtained based on the lumped mass matrix at node i &j and the stiffness matrix derived
vide Equation (1.1.3) and then subsequent amplitude and stresses can be calculated
by the usual procedure.
Now let us presume 100 years down the road scientists have developed a material
whose Youngs Modulus is say 9 10
20
kN/m
2
and we build this beam (of same
dimension) with this material (not the spring which represents the soil) and pose the
question as to does the system vibrate? Looking at Figure 1.1.1 one can intuitively say
that yes it does vibrate, but the beamhere being very stiff (limk ) undergoes rigid
body mode and the vibration is now guided by the stiffness of the spring only. Now
the question is does Equation (1.1.3) reflects this phenomenon amazingly not! For
putting this value of E = 910
20
kN/m
2
we find that [K]
g
becomes an infinitely stiff
matrix where the poor K
ii
and K
jj
(whose order would be of 10
5
to 10
6
) is completely
gobbled up by the stiffness values of the beamthat are exponentially higher and would
start giving time periods that are zero.
Like patch test in FEM, it is a test we can use to check the sanctity of a stiffness
formulation. We call this an RB (short of Rigid Body) test and we see it fails this
test with parallel spring connection, especially when the structure has got signicant
stiffness compared to soil.
Now if we put Equation (1.3.7) which is the series connection, to RB test, we find
that it passes the test with flying colors for as Limit of K the first termin the right
hand side of Equation (1.3.7) approaches zero and we are left with the soil springs
values only based on which the body vibrates and satisfies RB test conditions posed
earlier.
In Equations (1.3.8) and (1.3.9) it is clearly seen that the soil flexibility gets directly
added to the diagonal and then on inversion affects all the terms of the [K
e
] and gives the
true interaction unlike parallel spring which affects only locally the interaction effects
and does not possibly gives a true picture when the stiffness of the superstructure
becomes quit high compared to that of the soil.
1.3.7 What makes the structural response attenuate
or amplify?
In Example 1.2.5 and 1.2.6 we had shown two opposite cases of dynamic soil structure
response. While in the case of the chimney the response is attenuated, in case of the
horizontal vessel the response is however significantly amplified. One would obviously
be curious and wonder why does it happen? The riddle is surely not difficult to answer.
Shown in Figure 1.3.3 is the generic nature of the acceleration curve used for design
of structures under earthquake. The nature of the curve is almost common/similar for
all the earthquake codes around the world.
Based on the curve (Figure 1.3.3) it is evident that when the structure is very stiff
or massive and its fixed base time period hovers around the vicinity of point A, the
dynamic soil structure effect can show signicant amplication so long as the cou-
pled time period of the soil-structure system is within the zone C. Thus structures
like massive gravity dams, nuclear reactor buildings, Massive turbine foundations
20
,
large vessels supported on short pedestals (which are stiff) could show signicant
amplication in response when the effect of soil is considered in the analysis.
While for any structure whose fixed base time period is somewhere between point
B and C, if exceeds the point C consideration of the soil effect can undergo a major
attenuation. Normal buildings, RCC, Steel Chimneys, elevated water tanks etc would
possibly fall in this category. Thus depending on the stiffness of the structure, its mass
distribution, dynamic property of the soil one can either save some money (if there is
20 For instance the structural congurations used for old LMW type Russian turbo-generators used
commonly in India for 210 MW plant.
B C
g
Sa
A
D
Time period (sec)
Figure 1.3.3 Generic response spectra curve for earthquake.
attenuation) or could result in more costly design (for amplified response) which may
vary from case to case.
1.4 THE ART OF MODELLING
Computer Modelling of soil & structure optimally to arrive at a meaningful solution
is an art by itself, and can well be a topic of a complete book. We present hereafter
some major techniques that has been found to effective & reasonable.
1.4.1 Some modelling techniques
Experience shows that in many cases young engineers eager on get-going mode would
start fromthe very outset with an elaborate model of the whole soil-structure system
21
.
They spend significant amount of time on data input and checking of such massive
model and come up with a result whose qualitative difference with a much simpler
model is only marginal. Moreover trying to handle a big data-base, an inadvertent
modeling or input data error passing the scrutiny is not at all uncommon.
So at the very outset our suggestion would be, start with a simple model without
trying to over sophisticate the issue from the very out set
22
.
Start with a test case or a simplified model to check the results. For instance a simple
model given in a book or the user manual of the software in use is a very good starting
to have some idea what types of element to choose, what order of refinement suffice
and what type of simplified idealization is acceptable.
21 Problem modeled with a minimum 1000 degrees of freedom!
22 Mentioning the fact that you have used eight nodded brick elements, or 9-noded plate elements based
on iso-parametric formulation may look impressive as technical jargons in a design basis report but may
not always be cost-effective solution.
Super structures above ground
Ground level
Railway Carriage
Underground Tunnel
Figure 1.4.1 An underground tunnel for movement of train in a metro city.
You will be amazed to find that in most of the cases, modeling the soil intelligently
as linear springs (whose values are judiciously chosen) can be good enough for many
major soil structure interaction analyses. Specially, when the structure is modeled in
3D, avoid using Finite elements to model soil and coupling it to the structure.
Firstly, the model becomes huge resulting in more engineering time plus gives results
which become difcult to decipher and does not necessarily always gives a more
accurate or better result compared to a relatively simplied model.
Start with a simple model (preferably a stick model) and add the soil spring to get
a first order feel of how much the soil affects its response
23
.
Get a basic feel as to how much the results vary in terms of fixed base problem-
if found significant one should then and only then resort to a much more detailed
analysis.
If the variation is say within 15%, one can well ignore the soil effect and consider
the problem as a standard fixed base problem and proceed with the analysis.
Keep your eyes open but do not be biased on the issue. Optimize your engineering
effort to the best possible way.
There are certain types of problem where resorting to FEM however would become
almost essential. For instance for the problem considered in Figure 1.4.1, it would be
impossible to arrive at reasonable solution without an application of FEM.
Shown in Figure 1.4.1 is a sketch of an underground tunnel catering to movement
of high speed trains. The movement of train generates dynamic forces which travels
through the soil to the surface and could adversely affect the structures built on the
surface like buildings, water tanks etc and becomes an important study for engineers
undertaking such kind of projects.
23 A computer analysis is not mandatory at this stage, a simple hand calculation or an analysis in spread
sheet or MATHCAD would suffice.
It is but evident that for these cases of modeling, the soil as spring element will not
work and a comprehensive finite element modeling of the soil based on plane strain
element is required. Here also, while doing the modeling, our suggestion would be
start with a crude model (say 20 to 30 elements) to get a fill of the first order effects
and then progressively refine the model to get a more accurate result.
In static loading case in Chapter 4 (Vol. 1) we had explained the principles of
meshing of such plane strain problem. Under dynamic loading the principles meshing
are generally done based on the following
1 Find the time period of the exciting frequency (T
s
) of the soil medium as 4H/v
s
.
2 If v
s
is the shear wave velocity of the soil mediumthen for being the wavelength of
the propagating waves they are related by v
s
= f . Here f is the natural frequency
of the medium and f = 1/T
s
.
3 Thus obtain = v
s
T
s
.
4 The mesh size should preferably be /10 to /4 for linear or bilinear/quadratic
elements chosen.
One of the major limitations in FEM for wave propagation problem is that the
boundary has to be taken to a signicant distance away from the source to ensure
no waves are reflected back which would otherwise generate spurious modes. This
often makes the problem expensive in terms of data input, checking and run time.
Moreover, it is difcult to gauge at the outset as to where can the boundary be
terminated.
Infinite finite element as discussed in Chapter 4 (Vol. 1) is one alternative which has
been found to have a strong potential for catering to such problem.
Other than this, paraxial boundaries or providing viscous dampers at the boundary
of soil domain capable of absorbing the propagating waves are often used for this type
of problems
24
.
Else boundary elements have also been used to model such infinite domains and are
coupled to the superstructure (modeled by FEM) and an effective solution has been
sought.
Unfortunately most of the commercially available software do not have the provi-
sion of adding matrix which can be assembled to the FEM matrix and an engineer has
to write his own special purpose software to cater to such problems.
Finally a word on the soil. . . . . .
Irrespective of whether we use springs or finite element to model the soil, the fun-
damental property on which the stiffness depends, are the value G (Dynamic shear
modulus) and , the Poissons ratio. We had discussed in detail as to how to arrive at
the appropriate design values of these two parameters in the next section.
In spite of all the techniques used it should be clearly mentioned that the parameters
are still marred by uncertainties and the results thus obtained should be mellowed with
some judgment which comes out only of experience and sustained practice.
24 Refer Chapter 5 (Vol. 1) for detailed discussion on this issue.
It is always preferable to do some parametric study by varying the design soil values
by ()15 to 20%(depending on howreliable and exhaustive has been the geotechnical
investigation) and check howmuch these results affect the design values and preferably
a conservative and safe value should be chosen (based on this variance).
We mention in Table 1.4.1, some suggestive models for different classes of structures
where we start with a primary model (i.e. to get a basic feel of the response) and a
secondary model which is a further improvement to the primary model.
Table 1.4.1 Some suggestive models.
Sl. Structure
No. type Primary model Secondary model Remarks
1 Framed Stick model with soil 2D or 3D frame system
building considered as two uni- with masses lumped at
que spring (rotational nodes. Soil modeled as
and translational) consti- springs under each
tuting all the foundations individual foundation
2 High-rise 2D frame for the beam 3D frame for the beam The horizontal slab
building column system while the column system with need not be too
with shear shear wall modeled as shear wall modeled as rened and should
walls an equivalent cantile- plain stress elements. be good enough to
ver with soil springs The horizontal slabs generate requisite
under each column and modeled as plane stress stiffness in its own
the shear wall elements. Soil modeled plane
as springs below each
foundation
3 Chimneys 2D stick model with soil No further renement is For local effect model
and eleva- idealized as springs usually warranted unless the shell or super-
ted water some local effect of soil structure as a stick
tanks is required to be studied and the soil a axis
on surrounding structure. symmetric plain
strain element
4 Frames 2D frames with soil A detailed 3D model Refer Chapter 2 (Vol. 2)
support- modeled as springs. constituting of beam on detailed modeling
ing rotary Bottom raft considered elements with master and technique for these
machines innitely stiff hence only slave node option. The type of foundation.
lumped mass contribu- bottom raft discretised
tion is taken. Soil mode- into beam or plate elem-
led as springs. 3 to 4 ents with soil modeled
degrees of freedom as springs and connected
usually sufce. at each node of the raft
elements.
5 Dams and A simple stick model with A comprehensive 2D
embank- soil modeled as springs model with the dam
ments else time period may be broken up into plain
found from formula strain element and soil
suggested in code and modeled as springs or
modied by Veletsoss further rened into 2D
formula. plain strain element
depending on the comp-
lexity of the soil or the
importance of the dam
1.4.2 To sum it up
Dynamic soil structure interaction is still in its early days and investigators are still
looking for answers to many problems which are encountered in practice.
For instance soil are modeled as linear springs based on elastic half space theory,
considering it as a linear isotropic medium, but in reality it is not so. Layered soil
phenomenon, pore pressure dissipation under dynamic loading, liquefaction potential
and its effect, infinite domain problem, non linear and inelastic behaviour, radiation
and geometric damping are some of the important factors on which research is still in
progress to arrive at a more realistic model amenable to design office practice.
What has been presented in this chapter is only an introductory concept and what
is in vogue in practice at the present.
Hopefully in days to come our understanding in some of the issues mentioned above
will be more profound and engineers and researchers would come up with results which
would be more realistic and reliable.
However a word of caution should be pertinent at this juncture.
As stated earlier as the uncertainty plaguing the problem is many, one should not
loose the final outcome of what we are trying to achieve i.e. a safe and sound structure
which can stand the vagaries of nature.
So one should not get lost in the maze of sophisticated mathematics and try to always
economize on the structure based on what the computer out put reflects
25
.
For facilities important to society the results should always be mellowed with sound
engineering practice like good detailing, robust geometric configuration, and good
quality of time tested construction practice.
All these aspects are equally important for a structure to survive the wrath of Mother
Nature whose ways are still not very clearly known to us.
1.5 GEOTECHNICAL CONSIDERATIONS FOR DYNAMIC
SOIL STRUCTURE INTERACTION
In this section we deal with the geotechnical considerations which go into the process
of a successful dynamic soil-structure interaction analysis.
At the very outset we would request readers specially with a strong structural leaning
not to ignore this section. For our experience shows that nemesis of many mistakes
lies in misinterpretation of this particular topic. As such before launching yourself
into linear or non-linear finite element analysis of soil-structure system, the conceptual
aspect of the influencing soil parameters, its limitations and its effects should be clearly
understood.
As a pre-requisite, we expect that you have some background on. . .
Some fundamental concepts of Soil Mechanics
Basic concepts in Soil Dynamics
25 The output is nothing but a reflection of mans limited knowledge of nature and only an approximate
quantification of an idealized mathematical model which could be in significant variance to reality in
spite of our best effort.
1.5.1 What parameters do I look for in the soil
report?
To start with we pose the above fundamental question. To readers having some back-
ground on this issue may find it intriguing for his obvious answer would be the dynamic
shear modulus (G) and Poissons ratio ().
The obvious query that subsequently comes to mind is, does it require a full section
to be devoted to this issue?
The answer would surely be an emphatic yes, for in our opinion the values
adapted are often misunderstood/abused in many a case, and often makes the analysis
questionable or unrealistic.
The reasons that could be attributed to it are as follows:
Geotechnical test (lab or field) based on which data evaluated are not understood
properly. As the limitations of such data are not clearly made; often results in
incorrect interpretation.
Data considered are often not relevant or correct in terms of real situation in the
field, specially for layered soil.
Insufficient data and or lack of knowledge on the strain level to which the
foundation-structure system will be subjected to specially during earthquake.
Lack of dynamic test data and improperly co-related value from static soil
parameter which could be widely varying with the reality.
Finally, often forgetting the bottom line that unlike man made material like con-
crete and steel, soil is far more heterogeneous and unpredictable; thus for a real
soil structure interaction it is unfair to have an analysis on an absolute scale. It
should preferably be done for a particular range of values and the best estimate is
to be made out of it and this is where engineering judgment would count to a
large extent.
Having made the above statements, let us evaluate various aspects of dynamic
property of soil which are important for an integrated soil-structure interaction
analysis.
Before even looking at soil report the analyst should be clear with himself on
The type of structure he is dealing with
Type of foundation that is anticipated like shallow foundation (could be isolated
or combined footing), raft or piles etc.
What analysis he is looking for like is it an analysis for machine induced load,
earthquake, blast force etc.
Understanding of the above criteria will not only help him in understanding the
data obtained from different tests but could also possibly make him realize their
interpretation in a more realistic perspective.
The engineering parameters we look for in the soil report for develop-
ing the soil model either for finite element or linear/non-linear spring dashpot
model are
(shear stress)
2
1
1 2
G
1
G
2
Figure 1.5.1 Shear stress-strain curve of soil under cyclic loading.
1 Dynamic shear modulus (G) or shear wave velocity (v
s
)
26
2 Poissons ratio (v)
3 Damping value of soil both radiative and material.
The values are usually obtained either from field test, laboratory test or from theo-
retical co-relation with other engineering soil parameters. Before we step further into
the topic it would possibly be worthwhile to understand how soil behaves under cyclic
loading and what its characteristics are.
It should be remembered that even under low strain, soil behavior is essentially
non-linear though at low strain it does show some kind of linearity.
Shown in Figure 1.5.1, is the shear stress-strain curve of soil under cyclic loading.
It is evident from the above figure that shear strain varies with stress, and goes on
increasing with number of cycles of loading.
Thus before an analysis is being carried out one has to have an idea about the average
strain range to which the soil will be subjected to under the induced dynamic loading.
The characteristic curve which shows the variation of shear modulus with respect
to shear strain is shown in Figure 1.5.1a.
The curve shown above is otherwise known as Seed and Idrisss (1970) curve which
shows the variation of dynamic shear modulus of soil with shear strain.
26 Relationship being G = v
2
s
.
0
0.2
0.4
0.6
0.8
1
1.2
0 0.01 0.1 1 10 100 1000
Strain Ratio
G
/
G
0
Figure 1.5.1a Variation of Shear Modulus with strain under cyclic loading. (Seed & Idriss 1970).
Soil subjected to stress by machine foundation are usually low strain and varies
between to 10
4
to 10
3
%.
However for an earthquake of even moderate magnitude this will be much higher-
having strain range varying to 10
2
to even 10
1
% for very severe earthquake.
Since it is difficult to gauge at the outset of an analysis how much strain the soil will
be subjected to, the correction factor to be used to modify the data as obtained in the
soil report becomes difficult to quantify.
On the contrary rendering no correction would result in assuming a more stiff soil
and the result obtained based on this could be significantly varying from the reality.
Fortunately or unfortunately most of the tests carried out in the field or in the
laboratory for determination of the dynamic shear modulus is based on low strain
range having values restricted to 10
4
%.
Thus it should be clearly understood that the dynamic shear modulus data furnished
in the soil report is only valid for LOW strain range and can be only used directly
for analysis where the strain induced in the soil is significantly low like in design of
machine foundations only. For earthquake analysis where the site is situated in an
area of moderate to severe earthquake zone, direct use of such soil dynamic data may
not be valid for design of normal structures, for the strain induced in soil is much
higher.
1.6 FIELD TESTS
The most common field tests that are carried out at site for evaluation of dynamic
shear modulus or shear wave velocity are
1 Block Vibration Test
2 Seismic cross hole
1.6.1 Block vibration test
In block vibration test as shown in Figure 1.6.1, an oscillator is placed on a concrete
block of size 1.5 m 0.75 m 0.7 m resting at foundation level and induces dynamic
Oscillator
L
x
Fdn Level
Propagating waves
H=0.6 to 1.2 m
Figure 1.6.1 Schematic diagram for block vibration test.
loading on the soil. Two geo-phones are placed at a distance to pick up the signal from
the oscillator.
Once the oscillator induces dynamic force on the soil the geo-phones pick up this
signal and transfer themto an oscilloscope which shows an elliptical figure of Lissajous.
The operating speed of the oscillator is varied till the time the natural frequency of the
soil and the operating frequency of the oscillator matches (the Lissajous figure in the
oscilloscope becomes a perfect circle).
The shear wave velocity of the site is then given by
v
s
= 4fL
x
(1.6.1)
where v
s
= shear wave velocity of the soil; f = operating frequency of the oscillator
in cps; L
x
= distance between the two geo-phones.
For arriving at meaningful results usually high frequency oscillators (>100 cps) are
put to use for which the waves generated are of the order of 0.6 to 1.2 m.
Thus results obtained from this test only influence soil of depth 0.6 to 1.2 m below
the depth of foundation and should not be used where piles or other types of deep
foundations having influence area propagating much deeper is used.
Trying to induce lower frequency calls for much heavier oscillators which make the
test uneconomical compared to other types of tests.
1.6.2 Seismic cross hole test
As shown in the schematic sketch in Figure 1.6.2, a probe is placed in a bore hole
to the desired depth and shear wave is generated in the soil by hitting it hard with a
hammer.
The waves are picked up by a geo-phone entrenched firmly to the casing of another
bore hole located at a known distance (L
x
) from the first hole.
The time taken to pick up the signal is measured by the oscilloscope.
Oscilloscope
Hammer
Ground level
Bore Hole casing
Probe
Geo-Phone
L
x
Figure 1.6.2 Schematic diagram for seismic cross bore-hole test.
The dynamic shear modulus G
dyn
is then obtained from the expression
G
dyn
=

g
_
L
x
t
_
2
(1.6.2)
where, G
dyn
= dynamic shear modulus of the soil; = weight density of soil; g =
acceleration due to gravity; L
x
= distance between the two bore holes, and t = elapsed
time.
One of the major advantage with this test is that dynamic shear modulus can be mea-
sured to any desired depth and can very well be an integrated part of a SPT program.
The test is very effective in case the soil is layered in nature where visual inspection
of each layer is possible based on SPT test.
However, the strain range for test is again restricted to 10
4
% which is normally
less than the strain range experienced by machine foundations and earthquake analysis
and needs to be corrected to arrive at the design value of G.
1.6.3 How do I co-relate dynamic shear modulus when
I do not have data from the dynamic soil tests?
A not so uncommon phenomenon, that even puts an experienced engineer under dif-
cult situation at times. In many cases it has been observed that no dynamic test has
been carried out during the geo-technical investigation especially if it is a building
project. Though not unusual, but should not happen as a rule, for this shows the lack
of foresight on the part of the engineer while submitting the technical and commercial
proposal for a project. Even at the proposal stage the process involved in a plant is well
known to the bidder and all the concerned civil engineer has to do is to check with his
process department and find out if rotating machines are part of the process or not.
On the other hand knowing the location of a particular site one can easily find out
from the codes how active this zone is seismically and if felt reasonable all he has
to do is to include this additional cost of dynamic geotechnical investigation in his
commercial bid. People suffer from misnomer that dynamic tests are expensive-which
is actually not true, for an average dynamic test in international market takes roughly
US$ 20,00025,000 which would however be 0.25%of a small petrochemical refinery
and possibly 0.1% of a combined cycle 350 MW power plant.
Lack of these tests can land up some of the equipments operating in such projects
into serious problem whose cost itself would constitute 3040% of the whole
project cost!
So one has to decide on the risk involved and come to a conclusion of its worth.
Though theoretical co-relation exist for evaluation of dynamic shear modulus of soil
from static soil test (which has been successfully used in project works), it is always
preferable to have these dynamic tests carried out at site, for not only does it imbibe
more condence in the design process but engineer should also be aware that theo-
retically co-related values have also varied widely with respect to actual field data, and
should be mellowed with judgment. Considering the uncertainty prevalent in soil, is
surely not an easy task to accomplish.
1.7 THEORETICAL CO-RELATION FROM OTHER SOIL
PARAMETERS
The most outstanding work in establishing theoretical co-relation for evaluating the
dynamic property of soil has been done by Hardin, Drnevich, Richart, Seed, Idriss
to name a few
27
. The expressions suggested by them have been successfully used for
many real projects by the engineers in the past. We are going to have a look at some
of them hereafter and understand their limitations if any.
1.7.1 Co-relation for sandy and gravelly soil
1.7.1.1 Hardin and Richarts Formula
For rounded grained soil having void ratio less or equal to 0.8 the dynamic shear
modulus is given by (Hardin and Richart 1963)
G =
2630(2.17 e)
2
1 +e
0
in psi (1.7.1)
27 This is by no mean to ignore other researchers who have contributed significantly to this difficult study.
We only name a few, which are popular in practice.
For angular grained soil the dynamic shear modulus is given by
G =
1230(2.97 e)
2
1 +e
0
in psi. (1.7.2)
where, G = dynamic shear modulus of the soil in psi, e = in-situ void ratio of the soil
sample,
0
= mean effective stress in psi = 0.333
v
(1 + 2K
0
),
v
= vertical effective
stress in psi,
h
= horizontal effective in psi = K
0
v
, K
0
= earth pressure at rest, and
is a function of the plasticity index and the over-consolidation ratio.
The relationship between plasticity index, over-consolidation ratio and K
0
is as
shown in the following figure.
Figure 1.7.1 Value of the K
0
after Brooker & Ireland (1965) Reproduced by permission of the National
Research Council of Canada from the candian geotechnical Journal Vol-2 (1965).
1.7.1.2 Seed and Idriss Formula
The formula for dynamic modulus in this case, Seed and Idriss (1970) have been related
to relative density of sand which can usually be quantified fromSPTtest and is given by
G = 83.3K
2
0
in psi (1.7.3)
Here K
2
is a function of the relative density of the sand which can again be estimated
from the SPT value.
The relationship between SPT value and the relative density is as given Table 1.7.1.
Table 1.7.1 Soil properties with SPT values.
SPT value Compactness Relative density Angle of friction
04 Very loose 015 <28
410 Loose 1535 2830
1030 Medium 3565 3036
3050 Dense 6580 3641
>50 Very dense >85 >41
Table 1.7.2 Values of K
2
versus relative density at
strain of 10
3
%(Seed and Idriss 1970).
Relative density (%) K
2
90 70
75 61
60 52
45 43
40 40
30 34
For case of computer programming K
2
can also be represented by the expression
K
2
= 0.6Dr +16 (1.7.3a)
It is to be noted that in this case to determine the relative density, the observed SPT
value has to be corrected for the overburden pressure and dilatancy to arrive at the
design SPT value before it is co-related with the above table.
1.7.1.3 Corrections to SPT value
Though available in standard textbooks of Soil Mechanics and Foundation Engineer-
ing, for brevity we present the correction expressions as mentioned hereafter.
For dilatancy correction if the observed SPT value (N
0
) is greater than 15 then the
corrected SPT value N
is given by (Terzaghi and Peck 1967).

N
= 15 +
1
2
(N
0
15) (1.7.4)
The overburden correction as per Peck et al. 1980 is given by
N
= 0.77N
log
10
2000
p
for p
25 kPa (1.7.5)
For p 25 kPa, as per Murthy (1991)
N
=
4N
2 +0.034p
(1.7.6)
in which, N
= corrected SPT value for overburden, N
= corrected SPT value for

dilatancy, p
= gross overburden pressure in kN/m

2
.
1.7.1.4 Ohsaki and Iwasakis formula
Ohsaki and Iwasaki (1973) have given co-relation for dynamic shear modulus directly
co-related to SPT value and is expressed as
G = 12000 N
0.8
in kPa (1.7.7)
Here N = design SPT value at the site after relevant corrections.
Example 1.7.1
As shown in Figure 1.7.2 is a small site having dimensions 18 m 6 m which
would be supporting a Compressor unit and a few pumps, for which four
boreholes were dug at four corners as shown. The soil was found to be cohe-
sionless in nature and SPT values observed at the four bore holes are as tabled
hereafter
Depth BH1 BH2 BH3 BH4
(meter) (SPT value) (SPT value) (SPT value) (SPT value)
2 4 6 4 3
4 8 6 6 5
6 12 9 11 8
8 15 12 16 11
10 20 18 24 16
14 22 24 28 20
18.0
BH2
6.0
BH4
BH1
BH3
Figure 1.7.2
Based on Laboratory and eld analyses following parameters were further
established:
Ground water table = 1.6 m, below grade level
Saturated density of soil = 22 kN/m
3
Void ratio e
0
= 0.58; Plasticity Index = 0.0; Poissons ratio = 0.32
Determine the best estimate of dynamic shear modulus (G) of soil at 10.6
meter below ground level presuming no dynamic soil test was done during
geo-technical investigation.
Solution:
Average observed SPT value at a depth of 10.0 meter
=
20 +18 +24 +16
4
= 19.5 = 20 (say)
Average observed SPT value at a depth of 14.0 meter
=
22 +24 +28 +20
4
= 23.5 = 24 (say)
At a depth of 10.6 meter below ground level based on linear inter-polation
average observed SPT Value
=
24 20
4
0.6 +20 = 20.6

= 21 (say)
The above observed SPT value has now to be corrected for dilatancy and
overburden pressure
1 Correction for dilatancy
As per Terzaghi, corrected SPT (N
) value is given by
N
= 15 +
1
2
(N
0
15) for N > 15; (1.7.8)
or N
= 15 +
1
2
(21 15) = 18 (1.7.9)
2 Correction for overburden pressure
As per Peck
N
= 0.77N
log
10
2000
p
for p
25 kPa (1.7.10)
N
= Corrected SPT value for overburden; N
= Corrected SPT value for

dilatancy; p
= Gross overburden pressure in kN/m

2
.
Here p
= 22 10.6 = 233.2 kN/m

2
Substituting above in Pecks formula we have, N
= 0.77 N
log
10
2000
233.2
= 13
Thus, corrected design SPT value = 13
Referring to table 1.7.1 for N = 13, Dr (Relative density) = 39.5%
Net overburden pressure at 10.6 meter level is expressed as
v
= (22 10) 9 +22 1.6 = 131.6 kN/m
2
(18.718 p.s.i)
As there is no previous history of loading on the site O.C.R. = 1.
Thus for P.I. = 0.0 and O.C.R = 1 as per Brooker and Irelands curve we
have K
0
= 0.48
Considering confining pressure
o
= 0.333
v
(1 +2K
0
); we have
0
=
18.718
3
(1 +2 0.48) = 12.22 p.s.i.
As per Hardin and Richarts formula
G =
2630(2.17 e)
2
1 +e
0
G =
2630(2.17 0.58)
2
1 +0.58
12.22 = 14710.5 p.s.i. (101426 kN/m

2
)
(1.7.11)
As per Seed and Idriss formula
Referring to the chart given above for Dr = 39.5% and strain in the range of
10
3
% (usually valid for machine foundation) K
2
= 40.
And as G = 83.3 K
2
0
we have
G = 83.3 40
12.22 = 11647 p.s.i. (80308 kN/m

2
)
Thus taking average value of G based on Hardin and Seeds method
Average G =
101426 +80308
2
= 90867 kN/m
2
As per Ohsaka and Iwaskis formula
G = 12000 N
0.8
in kPa G = 12000 (13)
0.8
= 93397.6 kpa (93398 kN/m
2
)
Thus it will be observed that variation with average G obtained based on
Hardin, Seeds and Ohsakas formula is not signicant and is of the order of
2.7%
28
.
1.7.2 Co-relation for saturated clay
1.7.2.1 Hardin and Drnevich formula
Hardin and Drnevich (1973) have given the following formula applicable to clayey
soil as
G
max
= 1230
(2.973 e)
2
(1 +e)
(OCR)
k
(
0
)
0.5
in psi (1.7.12)
where, e = void ratio; OCR = over consolidation ratio;
0
= mean effective stress in
psi = 0.333 (
v
+ 2
h
);
v
= vertical effective stress in psi;
h
= horizontal effective
stress in psi = K
0
v
, K
0
= earth pressure at rest, and is a function of the plasticity index
and the over-consolidation ratio.
k = is a function of the plasticity index (PI) of the soil and is given as
k = 5 10
8
(PI)
3
4 10
5
(PI)
2
+0.0092(PI) +0.0025 (1.7.12a)
It is to be noted that G
max
as obtained above corresponds to a shear strain range of
0.2510
4
%and needs to be modified for the appropriate strain range as appropriate
for a problem in hand based on the expression
G =
G
max
(1 + /
r
)
(1.7.13)
Here = desired strain range;
r
= reference strain range and is expressed as
r
=

max
G
max
100 and
28 The point we are trying to make here is not to go by one formula, but check with possibly all of them
and comparing them to arrive at result which would possibly be best fit and hopefully be most realistic.
Here again it is to be noted that we had not used the angular sand formula of Hardin, if the soil
description does not reflect it or the soil has both rounded and angular grains an intermediate value has
tobe chosen judiciously.
max
=
_
_
1 +K
0
2
(
v
u) sin +c cos
_
2
_
1 K
0
2
(
v
u)
_
2
_
0.5
(1.7.14)
in which,
v
= total vertical stressing in soil; u = pore pressure; c = cohesion of soil;
= angle of friction of soil, and K
0
= coefficient of earth pressure at rest.
Example 1.7.2
It has been decided to place foundation of an industrial structure at 4.0 meter
below the existing ground level.
Based on laboratory and field tests it has been found that the Ground water
table is at a depth of 1.0 meter below GL.
Unconsolidated undrained triaxial tests reveal the sample to have the following
values:
Cohesion value c = 0.21 kg/cm
2
Angle of resistance = 18 degrees
Pore pressure = 0.0 kg/cm
2
Consolidation tests reveal that it had a history of pre-consolidation pressure
of 200 kN/m
2
:
Initial void ratio = 0.61
Plasticity limit PI = 35
Saturated unit weight of soil = 19 kN/m
3
The site has a history of moderate to severe earthquake when from previous
record it is observed to generate a strain range up to 0.1%.
Calculate the dynamic shear modulus of soil for this predicted strain
range.
Solution:
For foundation located at 4.0 meter below the ground level net vertical pressure
v
= 19 1.0 + (19 10) 3.0 = 46 kN/m
2
(6.54 psi)
OCR =
200
46
= 4.34, for plasticity index of 35 from Brooker and Irelands,
chart K
0
= 1.1
Thus considering
o
= 0.333
v
(1 +2K
0
), we have
o
= 0.333 6.54(1 +2 1.1) = 6.976 psi
G
max
= 1230
(2.973 e)
2
(1 +e)
(OCR)
k
(
0
)
0.5
in psi
Here k = 0.27 for PI = 35 as per Equation 1.7.12a
Thus substituting the values we have
G
max
= 1230
(2.973 0.61)
2
(1 +0.61)
(4.34)
0.27
(6.976)
0.5
= 16746 psi (115465 kN/m
2
)
Calculation for Shear stress
max
=
_
_
1 +K
0
2
(
v
u) sin +c cos
_
2
_
1 K
0
2
(
v
u)
_
2
_
0.5
or,
max
=
_
_
1 +1.1
2
(6.54) sin 18 +3.0457 cos 18
_
2
_
1 1.1
2
(6.54)
_
2
_
0.5
= 5.00 psi.
r
= Reference strain range and is expressed as
r
=

max
G
max
100 =
5.00
16746
100 = 0.0299%
Thus for 0.1% strain. . .
G =
G
max
_
1 +

r
_ G =
115465
_
1 +
0.1
0.0299
_ = 26577 kN/m
2
It is thus observed that dynamic shear modulus is 23% of the theoretically calc-
ulated data.
Based on the above example it would perhaps be not difcult to realize that how
important role does the strain range plays on the design value of dynamic shear
modulus of soil.
1.8 ESTIMATION OF MATERIAL DAMPING OF SOIL
Damping plays a significant part in the overall response of soil structure system. While
for structural members material damping plays a significant part (mostly considered
as Rayleigh damping), for soil, two types of damping are basically involved.
Radiation damping
Material damping
Radiation or geometric damping of a soil foundation system is a mean by which
the energy is dissipated by means of radiation from the source and is a function of
mass and inertia of the system
29
. Material damping of the soil foundation system is
a mean by which the energy is dissipated by hysteresis and is an inherent property of
the constituting material of the soil.
This can very well be found from resonant column test in the laboratory when
after the soil has been vibrated the exciter is stopped and successive amplitudes are
measured. If a
1
and a
2
are two successive amplitudes then
D
m
=
_
ln
a
2
a
1
__
_
4
2
+
_
ln
a
2
a
1
_
2
(1.8.1)
The total damping ratio of a soil foundation systemis sumof radiation and material
damping. It is generally observed that material damping has a significant magnitude
relative to radiation damping specially in rotational modes. In such cases total damping
rather than geometric damping should be used to obtain the response of the structure
foundation system.
For translatory mode, on the contrary material damping plays an insignificant role
and may be neglected in the analysis. Thus for tall narrow structures like chimney,
Boiler structures, tall buildings where the coupled horizontal and rocking mode could
play significant role it would perhaps be realistic to also consider the material damping
of soil in order to have a meaningful response.
1.8.1 Whitmans formula
Whitman (1973) has suggested that total damping (geometric + material) for a
machine foundation can be obtained from the expressions
Horizontal Mode
D
h
=
0.31
_
M
r
3
0
(1.8.2)
29 We have dealt this detail in Chapter 2 (Vol. 2) Design of Machine foundations.
For vertical mode
D
v
=
0.49
_
M
r
3
0
(1.8.3)
For rocking mode
D
= 0.05 +0.1
_
_
_
I
r
5
0
_
0.5
_
1 +
_
I
4r
5
0
__
_
_
1
(1.8.4)
Here M = mass of foundation plus structure or machine vibrating; I
= mass
moment of inertia of foundation plus machine/structure about a horizontal axis
through the base of the foundation perpendicular to the plane of rocking; r
0
=
equivalent radius of footing, and = mass density of soil.
1.8.2 Hardin formula
Hardin (1965) has expressed material damping of sandy soil by the expression
D
m
=
0.985
0.2
r
0
(1.8.5)
Here notations are same as expressed earlier except the fact that the conning pres-
sure
0
is expressed in kPa. The equation is valid for shear strain amplitude of 10
6
to 10
4
with a conning pressure of 24 kPa to 144 kPa.
For a particular strain range the value obtained above can be corrected based on the
expression
D
c
D
m
=
/
r
1 + /
r
(1.8.6)
Example 1.8.1
For the example as shown in Example 1.7.2, estimate the damping ratio of the
soil as per Hardins formula. The soil properties remain same as given in Example
1.7.2.
Solution:
Based on the solution furnished in Example 1.7.1 value of dynamic shear
modulus is given by
G = 90867 kN/m
2
;
0
= 12.22 p.s.i. (85.9 kN/m
2
); K
0
= 0.48
For design SPT value N = 13, = 31
, and
v
= 131.6 kN/m
2
Considering,
max
=
__
1 +K
0
2
(
v
u) sin +c cos
_
2
_
1 K
0
2
(
v
u)
_
2
_
0.5
, we have
max
=
_
_
1 +0.48
2
131.6 sin 31
_
2
_
1 0.48
2
131.6
_
2
_
0.5
= 36.67 kN/m
2
As =

G
100, we have
r
=
36.67
90867
100 = 0.0404%
Considering D
m
=
0.985
0.2
r
0
, we have
D
m
=
0.985(0.0404)
0.2
85.9
= 0.056
Thus material damping ratio is estimated as 5.6%.
1.8.3 Ishibashi and Zhangs formula
Ishibashi and Zhang (1993) has proposed an expression for the damping ratio of
plastic and non-plastic soil and is given by
= 0.333
1 +exp(0.0145PI
1.3
)
2
_
0.586
_
G
G
max
_
2
1.547
G
G
max
+1
_
(1.8.7)
The notations for the above expression are already explained in earlier formulas.
We show below variation of damping ratio with G/G
max
for different Plasticity
Index based on the above formula.
It will observed (Figure 1.8.1) that as G/G
max
reduces, as damping ratio goes
on increasing meaning thereby that as strain increases damping ratio goes on
increasing. Variation of Damping with strain vide Equation (1.8.6) is shown in
Figure 1.8.1a.
Variation of damping ratio
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1
0
.
9
0
.
8
0
.
7
0
.
6
0
.
5
0
.
4
0
.
3
0
.
2
0
.
1 0
G/Gmax
D
a
m
p
i
n
g

r
a
t
i
o
PI = 10
PI = 20
PI = 30
PI = 40
PI = 50
PI = 60
PI = 70
PI = 80
PI = 90
PI = 100
Figure 1.8.1 Variation of damping with plasticity index as per Ishibashi and Zhang (1993).
0
0.2
0.4
0.6
0.8
1
1.2
0 0.01 0.1 1 10 100 1000
Strain Ratio
D
/
D
r
Figure 1.8.1a Variation of damping ratio with strain under cyclic loading.
Example 1.8.2
For the clayey soil sample as shown in Example 1.7.2, determine the damping
ratio for the strain range level of 0.1% based on Zhangs formula. Consider all
soil properties same as Example 1.7.2?
Solution:
Based on earlier example we have seen that plasticity index PI = 35.
In Example 1.7.2 we have already calculated that for 0.1% strain G/G
max
=
0.230 Substituting the above in Ishibashi and Zhangs formula we have
= 0.333
1 +exp(1.47)
2
[0.586(0.23)
2
1.547 0.23 +1] = 0.1382
Thus estimated damping ratio is 13.82%.
1.9 ALL THINGS SAID AND DONE HOWDO WE ESTIMATE
THE STRAIN IN SOIL, SPECIALLY IF THE STRAIN IS LARGE?
We acknowledge at the very outset that posing the question, though easy, is not very
easy to answer. The uncertainties involved are so widely varying that it would be
difcult to give a precise answer to this issue. To the best of our knowledge there is
no straight forward answer to this problem and the best one can achieve is a reason-
able estimate or can possibly study a range of values and try to predict the overall
behavior.
For high speed centrifugal machine foundation it does not pose a serious problem
for at the lowstrain range a fewpercent here and there does not contribute a significant
variation to these values.
But for impact type of machines (hammer foundations) and slow speed machines
(coal mill foundations, reciprocating compressors) induced strain could be larger than
strain developed during field test, for which the correct estimation of G
dyn
and damping
becomes important.
For earthquake of course the strain would invariably be larger than measured during
test, even for a moderate earthquake when as the strain range increases, degrada-
tion in soil stiffness becomes signicant and has a major contribution to the overall
response.
It is obvious that strain induced in soil will depend upon the strength of dynamic
loading, the geological condition of the site, stress history of soil and a number of
other factors. So the point remains that if there exists no previous records of strain
from similar machine in same site or from previously occurring earthquake data how
does one rationalize the strain?
We discuss below some of the techniques which could be used for evaluation of the
strain induced in the soil.
1.9.1 Estimation of strain in soil for machine
foundation
For machine foundations the present practice of arriving at strain dependent dynamic
shear modulus and damping can be structured as follows:
Start with the field observed/lab obtained data for G
dyn
and damping as furnished
in the soil report which would usually correspond to the strain range of 10
4
to
10
3
%.
Calculate natural frequency of the soil-foundation system based on free vibration
analysis.
For rotating mass type calculate the transmissibility factor based on the expression
T
r
=
r
2
_
1 + (2 r)
2
_
(1 r
2
)
2
+ (2 r)
2
(1.9.1)
For constant force excitation (like in hammer foundation) calculate transmissibil-
ity factor based on expression
T
r
=
_
1 + (2 r)
2
_
(1 r
2
)
2
+ (2 r)
2
(1.9.2)
where r =

m
n
,
m
= operating frequency of the machine;
n
= natural frequency
of the foundation; = damping ratio of the soil.
Find out the pseudo-static force by multiplying the vertical unbalanced force of
the machine by the transmissibility factor as mentioned above.
Find out the dynamic stress induced in the soil by dividing the above force by the
foundation plan area.
The approximate shear strain in the soil is given by the expression.
(%) =
12q
dyn
G
(1.9.3)
Verify the strain obtained against the initial value. If they vary significantly find
out the new G value based on the calculated strain and repeat the process as
mentioned above till it converges.
The above method is surely non-rigorous but generates an answer which will give
reasonably accurate results for practical analysis of machine foundations. For more
complicated soil with varying properties a more rigorous analysis based on Finite
element analysis is possible. This will be discussed later on.
The above technique is now explained based on a suitable numerical example.
Example 1.9.1
A centrifugal turbine driven compressor has foundation dimension of 6 m
3.2 m 2.5 m. The weight of the compressor is 300 kN. The unbalanced mass
on the shaft is 3.5 kN sec
2
/m rotating at an eccentricity of 0.4 mm having
operating frequency of 1800 rpm.
The soil investigation has revealed the soil data as follows
SPT = 13 (After dilatancy and overburden correction)
Plasticity Index (PI) = 0
Poissons Ratio = 0.3
Calculate the correct value of dynamic shear modulus and damping.
Solution:
Based on Ohsakas formula G = 12000 (13)
0.8
= 93397.65 kN/m
2
Assumed strain level = 1 10
4
%
Weight of foundation = 63.22.525 = 1200 kN; Weight of machine =
300 kN, Total weight = 1500 kN
Mass of foundation + machine = 1500/9.81 = 152.9052 kN sec
2
/m
Equivalent radius of the foundation (r
0
) =
_
6 3.2
= 2.472155 m
Equivalent vertical spring stiffness of soil
30
.
Kz =
4Gr
0
1
=
4 93397.65 2.472
0.7
= 1.32 10
6
kN/m
n
=
_
K
z
m
=
_
1.32 10
6
152.9
= 92.89 rad/sec;
m
=
1800 2
60
= 188 radian/sec r =
m
/
n
= 2.029.
P
dyn
= m e
2
m
= 3.5
0.4
1000
(188)
2
= 49.7428 kN.
Considering transmissibility as
T
r
=
r
2
_
1 + (2 r)
2
_
(1 r
2
)
2
+ (2 r)
2
we have, T
r
= 0.65078.
30 Refer to Chapter 5 (Vol. 1) Basic Concepts of Soil Dynamics, for details of this formula.
Equivalent static force on foundation =
0.65078 49.7428
6 3.2
= 1.686 kN/m
2
Considering (%) =
12q
dyn
G
=
12 1.686
93397.65
= 2.17 10
4
%
Considering G =
G
max
(1 +

r
)
we have, New G =
93397.65
_
1 +
2.1710
4
110
4
_ = 29497.88 kN/m.
We proceed with second cycle of iteration with this new value of G.
Shown below is such iteration for 14 cycles
Cycles 1 2 3 4 5 6 7
G
dyn
93397.65 29497.88 22553.77 20646.63 20025.37 19811.83 19737.08
Damping 0.012987 0.189764 0.21998 0.228656 0.231517 0.232505 0.232851
K
z
1.32 4.17 3.19 2.92 2.83 2.80 2.79
10
+06
10
+05
10
+05
10
+05
10
+05
10
+05
10
+05
n
92.89143 52.20392 45.64753 43.67493 43.01282 42.78288 42.7021
r(w
m
/w
n
) 2.029203 3.610755 4.129371 4.315875 4.382311 4.405864 4.414199
T
r
0.65078 0.298033 0.255622 0.243327 0.239243 0.23783 0.237334
P
dyn
49.74281 49.74281 49.74281 49.74281 49.74281 49.74281 49.74281
P
eq
32.37161 14.82497 12.71537 12.10377 11.90063 11.83032 11.80565
q
dyn
1.686022 0.772134 0.662259 0.630405 0.619824 0.616163 0.614878
Strain 2.17 3.14 3.52 3.66 3.71 3.73 3.74
10
04
10
04
10
04
10
04
10
04
10
04
10
04
Cycles 8 9 10 11 12 13 14
G
dyn
19710.75 19701.45 19698.17 19697.01 19696.60 19696.45 19696.40
Damping 0.232973 0.233016 0.233031 0.233037 0.233039 0.233039 0.23304
K
z
2.78 2.78 2.78 2.78 2.78 2.78 2.78
10
+05
10
+05
10
+05
10
+05
10
+05
10
+05
10
+05
n
42.6736 42.66353 42.65998 42.65872 42.65827 42.65812 42.65806
r(w
m
/w
n
) 4.417147 4.418189 4.418558 4.418688 4.418734 4.41875 4.418756
T
r
0.237159 0.237097 0.237075 0.237068 0.237065 0.237064 0.237064
P
dyn
49.74281 49.74281 49.74281 49.74281 49.74281 49.74281 49.74281
P
eq
11.79696 11.79388 11.7928 11.79241 11.79228 11.79223 11.79221
q
dyn
0.614425 0.614265 0.614208 0.614188 0.614181 0.614179 0.614178
Strain 3.74 3.74 3.74 3.74 3.74 3.74 3.74
10
04
10
04
10
04
10
04
10
04
10
04
10
04
We show hereafter the variation of strain, damping and G
dyn
per cycle in Figs. 1.9.1 to 1.9.3.
Variation of strain per cycle
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
4.00E-04
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number of cycles
S
t
r
a
i
n
(
%
)
Figure 1.9.1 Variation of strain (%) per cycle.
Variation of damping ratio with strain
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number of cycles
D
a
m
p
i
n
g

r
a
t
i
o
Figure 1.9.2 Variation of material damping per cycle.
Variation of Gdyn with strain
0
20000
40000
60000
80000
100000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Number of cycles
G
d
y
n
(
k
N
/
m
2
)
Figure 1.9.3 Variation of dynamic shear modulus per cycle.
From the tables and the above plots it is observed that the value becomes
constant after 7th cycle of iteration based on which we conclude that design
values are as follows
G
dyn
= 19700 kN/m; Material Damping ratio = 0.23, and Estimated strain
range = 3.74 10
4
%.
Thus actual design of the foundation shall be carried out based on this corrected
value instead of the initial values as mentioned in the soil report.
1.9.2 Estimation of soil strain for earthquake
analysis
For earthquake analysis things are surely more complicated for not only the forces
induced in the soil is much more complex, the behavior itself is different frommachine
foundations.
While in machine foundation the force is induced in the soil from the structure
in earthquake the force is induced within the soil where the soil first start vibrat-
ing based on the waves propagating through it. Thus acceleration, it is excited to,
depends on the free field vibration of the site. This acceleration induced in the soil
generates shear strain on which the stiffness degradation and damping ratio would
depend.
Though the analysis shown hereafter is based for isotropic homogenous medium
it can well be extended to layered soil having variable property based on weighted
average.
Propagation of Earthquake
H
Figure 1.9.4 Schematic diagram of an industrial site with propagating earthquake waves.
Shown in Figure 1.9.4 is a schematic diagram of an industrial site with propagating
earthquake waves. The depth of the site (H) is considered to the bedrock level from
where the waves are presumed to be propagating
31
.
The waves propagating at bedrock level travels upward and hits the surface (z = 0)
when the site surface undergoes a motion. However as surface is free, it is free to shake
as such no strain energy develops at the surface.
The motion of such elastic waves propagating through an elastic medium can be
defined by the partial differential equation
2
u
t
2
= v
2
s
2
u
z
2
(1.9.4)
Here u = displacement of the soil and is a function of time t and depth z, v
s
= Shear
wave velocity of the soil.
Considering u(z, t) = (z)(t), (1.9.5)
31 For site having no bedrock this level is usually considered at the depth where shear wave velocity of
the site is greater or equal to 600 m/sec. Based on SPT value this can be considered as the depth where
design SPT value is greater than 50.
we have substituting in the equation above
(z)

(t) = v
2
s

(z)(t)
or

(t)
v
2
s
(t)
=
(z)
(z)
= p
2
(say)
The above on separation gives two homogenous equations
(z) +p
2
(z) = 0 and

(t) +p
2
v
2
s
(t) = 0 (1.9.6)
The above gives solution
(z) = Acos pz +Bsin pz (1.9.7)
at z = 0 as there will be no shear strain hence du/dz = 0
(z) = Apsin pz +Bpcos pz = 0, at z = 0

The above gives the constant B = 0, from which we deduce, (z) = Acos pz
At z = H as the soil is confined, hence we have, u(z, t) = 0 Acos pH = 0
p =
(2n 1)
2H
(1.9.8)
Considering, (t) = Ccos t +Dsin t (1.9.9)
we have, = p
2
v
2
s
, where is the eigen-value of the problem.
Knowing from our fundamental knowledge of vibration that
2
n
= =
(2n 1)
2
2
4H
2
v
2
s
or
n
=
(2n 1)
2H
v
s
rad/sec (1.9.10)
Considering T =
2
n
, we have
T
n
=
4H
(2n 1) v
s
secs (1.9.11)
Here T
n
is known as the free field time period of the site for n numbers of mode.
The corresponding eigenvectors are given by
(z) = cos
(2n 1)z
2H
(1.9.12)
The displacement vector is given by
u(z) =
i
i
(z)S
d
(1.9.13)
where S
d
= Displacement spectrum of the site which can again be represented as
u(z) =
i
i
(z)
Sa
2
n
(1.9.14)
where Sa = acceleration spectrum of the site and is a function of the free field time
period of the site
32
.
In which,
i
= Modal mass participation factor =
m
i
m
i
2
i
(1.9.15)
The modal participation factor can thus be considered as
i
=
m
i
m
i
2
i
=
H
_
0
z cos
z
2H
_
H
_
0
z cos
2
z
2H
(1.9.16)
The above on integration by parts gives,
i
=
8
+2
(1.9.17)
Thus for the present problem
u(z) =
32S
a
H
2
(2n 1)
2
2
( +2)v
2
s
cos
(2n 1)z
2H
(1.9.18)
Here =
ZI
2R
the IS code factor
33
32 This response spectrum is usually available as site response spectra in absence of which charts furnished
in National codes are usually followed.
33 Presently code does not have any guidline for R for soil. It has been observed that a value between R = 2
to 3 usually gives realistic results.
Considering shear strain
z
=
u
z
we have
z
=
16S
a
H
(2n 1)( +2)v
2
s
sin
(2n 1)z
2H
(1.9.19)
Considering G = v
2
s
we have
z
=
16S
a
H
(2n 1)( +2)G
sin
(2n 1)z
2H
(1.9.20)
Here G = dynamic shear modulus of the soil, = mass density of the soil.
For foundation at a particular depth below the free surface for which we have
obtained the dynamic shear modulus based on field or lab test
34
. We start initially to
find out the shear strain in the soil based on this value considering a strain range of
10
3
/10
4
%.
The steps that are followed subsequently to arrive at the corrected G and damping
value are furnished hereafter (Chowdhury 2008).
1 Identify the bedrock level (H) of the site
2 Find out the shear wave velocity from the expression G = v
2
s
3 Find out the free field time period of the site from the expression T
n
=
4H
(2n1)v
s
4 Based on the site response spectra/spectra given in code and damping value as
obtained in soil report obtain the acceleration Sa
5 Obtain shear strain for the soil prole based on the expression
z
=
16S
a
H
(2n1)(+2)G
sin
(2n1)z
2H
6 Check if this strain is near or equal to the initial strain(10
3
to 10
4
)%
35
.
7 If there exists a significant variation correct Gbased on the equation G =
G
max
(1+/
r
)
8 Find out the ratio G/G
max
9 Obtain new damping ratio based on Zhangs expression
= 0.333
1 +exp(0.0145PI
1.3
)
2
_
0.586
_
G
G
max
_
2
1.547
G
G
max
+1
_
10 Repeat the steps as mentioned from 2 to 7 till the strain is same as previous cycle.
The value for which the strain becomes constant is the corrected Dynamic shear
modulus of the soil.
The above steps will now be further elaborated by a suitable problem.
34 Or from theoretical co-relation.
35 In absence of this input in soil report consider 10
3
% for soft soil and 10
4
% for medium stiff soil.
Example 1.9.2
For a particular site susceptible to earthquake it was observed based on soil
investigation that bed rock exists at 20 meters belowground level. Seismic cross-
hole test reveals average dynamic shear modulus of the soil to be 154897 kN/m
2
at a reference strain of 1 10
5
.
Considering density of soil as 19 kN/m
3
, and plasticity index as 35. Calculate
the corrected dynamic shear modulus of soil and damping at 2.5 meter below
GL where foundation of a particular structure will be placed. Consider IS 1893
curves to evaluate the acceleration pertaining to a particular time period.
Solution:
Depth of soil over bedrock = 20 m, Density of soil = 19 kN/m
3
, Thus mass
density of soil () =
19
9.81
= 1.936 kN sec
2
/m, Dynamic shear Modulus of soil
(G) = 154897 kN/m
2
.
Shear wave velocity of soil (v
s
) =
_
G
= 282.8 m/sec.
Considering T
n
=
4H
(2n 1)v
s
we have for fundamental mode
T
1
=
4 20
282.8
= 0.283sec
Considering Zhangs formula
= 0.333
1 +exp(0.0145PI
1.3
)
2
_
0.586
_
G
G
max
_
2
1.547
G
G
max
+1
_
Taking G/G
max
= 1 for first cycle and PI = 35 we have, = 0.798%
For damping @ 0.798% and time period of 0.283sec Sa = 5.75 m/sec
2
from
IS code.
For this case the code factor Z is considered as 0.24, I = 1.2, R = 3. The value
of R is chosen as 3 in this case because for PI = 35 it is assumed that the soil has
high plasticity and thus has reasonable ductility. As a matter of fact there is no
guideline at moment prevalent in the code and an engineer has to use his own
judgment here.
Substituting the above data in the expression
z
=
2S
a
H
(2n 1)G
sin
(2n 1)z
2H
We have at depth of 2.5 meter below ground level
z
= 2.5145 10
5
Considering G =
G
max
(1 +

r
)
we have G = 44073.2 kN/m
2
We proceed with next cycle of iteration with this new value of G.
In table below we show how the data converges for 10 successive cycles
Cycles 1 2 3 4 5
GkN/m
2
154897 44073.21001 39512.54 37131.85 35405.98
V
s
m/sec 282.7998707 150.8499063 142.8319 138.4621 135.206
T sec 0.282885561 0.53032847 0.560099 0.577775 0.59169
Sa m/sec
2
5.75 1.9 1.85 1.85 1.8
Strain 2.51454 10
05
2.9202 10
05
3.17 10
05
3.37 10
05
3.44 10
05
G/G
max
1 0.284532367 0.255089 0.23972 0.228578
Damping
ratio(%) 0.798 12.425 13.167 13.562 13.852
Cycles 6 7 8 9 10
GkN/m
2
34857.41 34437.41 34113.86 33863.42 33668.85
V
s
m/sec6 134.1545 133.3438 132.716 132.2279 131.8475
T sec 0.596327 0.599953 0.602791 0.605016 0.606762
Sa m/sec
2
1.8 1.8 1.8 1.8 1.8
Strain 3.5 10
05
3.54 10
05
3.57 10
05
3.6 10
05
3.62E
05
G/G
max
0.225036 0.222325 0.220236 0.218619 0.217363
Damping
ratio(%) 13.945 14.016 14.071 14.114 14.147
Thus based on above calculation we may take, corrected G value =
33600 kN/m
2
; Damping ratio = 0.14. Variation of shear modulus and damping
ratio (%) with number of cycles are shown in Figs. 1.9.5 and 1.9.6.
Varaition of Gdyn at foundation level
0
50000
100000
150000
200000
1 2 3 4 5 6 7 8 9 10
Iteration number
G
Figure 1.9.5 Variation of shear modulus with number of cycles.
Varaition Damping ratio(%) at foundation level
0.000
5.000
10.000
15.000
1 2 3 4 5 6 7 8 9 10
Number of iterations
D
a
m
p
i
n
g

R
a
t
i
o
(
%
)
Figure 1.9.6 Variation of damping ratio (%) with number of cycles.
1.9.3 What do we do if the soil is layered with varying
soil property?
Till now the theories we have presented assumes soil as a homogenous isotropic
medium but in reality in all possibility the soil encountered at a particular site will
be layered in nature.
Shown in Figure 1.9.7, is a typical stratified soil profile where the shear modulus,
density of soil and Poissons ratio vary with depth.
For most of the cases taking a weighted average is the normal practice where the
average dynamic property may be taken as
G
av
=
G
1
H
1
+G
2
H
2
+G
3
H
3
+G
4
H
4
H
1
+H
2
+H
3
+H
4
(1.9.21)
and same principle be applied for mass density and Poissons ratio.
However for very important structures or site susceptible to major earthquakes
methods based on finite element analysis may be applied to arrive at a design dynamic
modulus and damping value
36
.
Shown in Figure 1.9.8, is the finite element model of a site having layered soil
property. In this case the soil is modelled as plane strain element to the bedrock
boundary and each individual layers having different properties can very easily be
catered to.
To start with we assume G value as obtained from soil report and consider the
damping ratio based on Zhangs formula considering G/G
max
= 1 at the strain level
of 10
3
/10
4
% say.
Suppose the previous earthquake history shows that shaking has taken place for
duration of 3 sec maximum, we select duration of 6 sec for analysis.
36 In such cases preferably site response spectra of the particular should be used. Moreover some previous
history of shaking and its duration should be available for analysis.
Figure 1.9.7 Layered soil strata with varying soil property.
Super Structure
Layered Soil
Bed Rock
Figure 1.9.8 Finite element model of a site having layered soil property.
Next for duration of 6 sec we input the Sa/g curve for the particular damping and
perform a time history analysis of the system for 6 sec.
Fromthe output for each layer we find the average shear strain. Based on the output
average strain, we correct the value of G
dyn
for the next cycle and also the damping
ratio and do a second cycle of time history for 6 sec. We repeat this process for a
couple of times till the values have stabilised with respect to the previous cycle.
The value of G and damping considered in last cycle where the strain has stabilised
are the dynamic shear modulus and damping of the soil.
Calculation of shear modulus of soil based on the free field time period is an effective
tool for assessing the dynamic shear modulus of soil. However, there is a possibility
that the time period obtained by this method could be higher than the reality unless
proper consideration are given for the confining effect of the surrounding soil and
proper judgement of the depth is made. ATC (1982) has defined H
max
as the depth
limited to 183 m having low strain shear wave velocity of 760 m/sec.
1.9.4 Checklist of parameters to be looked in the
soil report
Based on above discussion, the parameters which require particular attention in a soil
report from the engineer are summarised as follows:
1.9.4.1 Field test
Has SPT test been carried out?
The obvious intention is to find out N on which G value depends. This can also be
utilised to cross check the field observed dynamic data.
If SPT values are furnished are the observed data or corrections need to be done?
A point to be checked for field observed data as shown earlier needs to be corrected.
While the soil consultant will do this correction during his own calculation of bearing
capacity of soil for foundation recommendations, usually furnishes observed field data
while furnishing the bore log detail in the report. So for your calculation this data needs
to be corrected. If you are not too sure you can back calculate it from recommended
value.
Has Ground Water Table been established during boring?
Usually provided in a soil report but better to check for this has significant effect on
the net vertical stress.
Has any dynamic field test carried out?
Block Vibration test
Seismic cross hole test
One of the tests should be a part of the soil report. But do not take the values
furnished sacrosanct. Back check with theoretical co-relation to establish if the order
is close, if not you do have the right to ask your soil consultant why there is this
discrepancy. There could be special geological condition which could result in such
discrepancy and you should be clear about it.
If the above tests are carried out, what is the strain range induced in the soil during
the test?
This is something usually not supplied by the soil consultant who usually would
recommend a unique G value. This should not be acceptable to you.
You should clear it at the very outset when providing him the specification for
Geotechnical investigation that this is an input you are looking for and it should be
a part of his report. It is more realistic to start with this value rather than guessing a
theoretical value of 10
3
/10
4
%.
1.9.4.2 Laboratory tests
Check Atterbergs limit gives values of liquid limit, plastic limit, plasticity
index etc.
Generally speaking
37
as a ritual, structural engineers/analysts ignore this topic. Our
suggestion would be, do not disregard this for this is the basic data which gives you
the first insight into the fact as to how the soil behaves. Moreover plasticity index
being an important property it is all the more important that you should pay attention
to this.
Triaxial test gives values of c, and pre-consolidation history
Again given a back handed treatment by the analyst who without going through
the test data would prefer to pick up the numbers which concerns him (c
a
and ). We
suggest go through the test and develop enough skill to interpret the pre-consolidation
stress. Make sure to ask during enquiry to the consultant to supply this data. For Over
Consolidation Ratio (OCR) plays a very important role in arriving at the correct value
of G
dyn
.
During interpretation if need be, seek help of a geotechnical specialist to make sure
what has been understood is correct this will save a lot of headache in the long run.
Bulk density and void ratio of soil
Grain distribution to check if the soil is gap graded, uniformly graded, or poorly
graded. Relative density of cohesion less soil is highly dependent on this.
1.10 EPILOGUE
The technology described in this chapter to our perception is still in its infancy and we
are optimistic that with time and research that is being carried out all over the world,
we shall be in a better position in future to predict more realistically the dynamic
properties of soil which affect the response of structure.
Whatever we have presented here is what we believe is simple to apply, provides
reasonably realistic results and practical for day to day design office practice.
There is hardly any comprehensive text which gives a defined picture on this
issue. Most of the techniques developed herein are based on research papers
(names furnished in the reference) and typical practices followed in some design
offices
38
.
We urge the readers to go through these papers which we believe will give them
further insight to the problem.
37 Exceptions are always there. . ..
38 Even consultants who require to use these type of technology is very limited.
The ideas presented in this chapter is to make the reader aware of the limitations
prevalent with soil and also to caution him on the fact that without these values
realistically estimated, the whole analysis related to dynamic soil structure interaction
could become a questionable exercise.
So be aware and use your judicious best to furnish a meaningful design.
SUGGESTED READING
39
1 Cohen, M. & Jennings, P., Silent Boundary Methods For Transient Analysis, Computa-
tional Method in Transient Analysis Computational Method in Mechanics, Vol. 1, North
Holland.
2 Dasgupta, S.P. & Kameswara Rao, N.S.V.K. 1976, Some nite element solutions in the
dynamics of circular footings, Proc 2nd International Conference on Numerical Methods
in Geomechanics, Blacksburg USA.
3 Dasgupta S.P. & Kameswara Rao, N.S.V.K. 1978, Dynamics of rectangular footings by
Finite elements, Journal of GT Division ASCE, Vol. 104, No. 5.
4 Gazetas, G & Tassoulas, A.L. 1987, Horizontal Stiffness of Arbitrarily shaped embedded
foundation, Journal of GT Division, ASCE, Vol. 113, No. 5.
5 Kameswar Rao, N.S.V. 1977, Dynamic soil structure system A Brief Review, J. Struct.
Engg., India, Vol. 4.
6 Lysmer, J. &Kuhlemeyer, R.L. 1969, Finite dynamic model of innite media, J.EM.Divn,
ASCE, EM4.
7 Segol, G., Abel, J.F. &Lee, P.C.Y. 1975, Finite element Mesh Gradation of surface waves,
J. GT Division, ASCE, Vol. 101, GT 11.
8 Wolf, J.P. 1985, Dynamic Soil Structure Interaction, Prenctice-Hall Englewood Cliffs, NJ.
9 Wolf, J.P. 1988, Dynamic Soil Structure Interaction in Time Domain, Prenctice-Hall,
Englewood Cliffs, NJ.
10 Wolf, J.P. 1994, Foundation Vibration Analysis: Using Simple Physical Model, Prentice-
Hall, Englewood-Cliffs, NJ.
11 Whitman, R.V. 1970, Soil Structure Interaction Seismic design for Nuclear power plants,
The MIT press, Cambridge, Massachusets.
39 This topic being relatively new, there are not much reference books (other than reference 8, 9 & 10)
which deal this topic comprehensively. Many literatures though have mentioned the interaction effect
in their work. The references suggested are thus mostly restricted to research papers, which we would
request you to get hold of and rummage through patiently.
Chapter 2
Analysis and design of machine
foundations
2.1 INTRODUCTION
This chapter deals with vibration analysis and design of machine foundations subjected
to dynamic load. As a pre-requisite to this chapter, you should be thoroughly familiar
with concepts that are put in chapter 5 (Vol. 1) on
Basic concepts in structural dynamics.
Basic concepts in soil dynamics.
Armed with these basics, we believe you will find this chapter interesting and find
design of machine foundation a challenging and intellectually stimulating task.
Machine foundations are one of the most important features of industrial devel-
opment. In both developed and developing countries, growth of economy is largely
attributed to development of industry and infra-structure facilities.
In industrial facilities like Power Plants, Steel Plants, Petrochemical Complexes,
Fertiliser Plants etc., consist of a number of centrifugal and reciprocating machines
and these play an important role to ensure smooth operation of the process and that
the output product is of right quality. If any of these equipments starts malfunc-
tioning or breaks down due to excessive vibration or settlement of the foundations,
cascading effect on the overall performance on engineering could be catastrophic at
times.
We give two case histories below making you aware of how far reaching could be
the consequences.
2.1.1 Case history #1
In Middle-East, in one of the oil producing nation there was a plant which had been
operating for last 25 years smoothly, sweetening the sour gas that was being pumped
into the complex from a nearby gas-field. The authorities hit a new source of natural
gas nearby this complex and the obvious choice was to pump gas from this new gas
field to the existing plant for further processing. This called for upgrading the plant
capacity. On engineering evaluation it was found that it necessitated certain changes in
diameter of the pipes, re-routing some of the existing pipes with newsupports and also
changing the rating of the two-stroke reciprocating compressor which was existing at
the plant. The company management wanted to expedite the issue for they perceived
that each day lost in production, the company stood to loose about 100,000 thousand
US dollars in profit.
In haste nobody thought to re-check performance of the foundation of the
compressor under dynamic load, now that its rating was changed!
When the plant started after this modification with additional gas being pumped
from the new gas field, whole pipe rack started shacking violently and the compressor
foundation started showing vibration amplitude that was well beyond acceptable limit.
The vibration became so high at 80% production level (at which the plant would
operate at most of the time) that the operation manager had no alternative but to stop
the plant completely.
Subsequent investigation revealed that with new rating of the compressor, the oper-
ating frequency nowhovered very near to natural frequency of the foundation resulting
in a resonant condition and also induced additional excitation to the fluid owing
through the pipe resulting in a force which the piping system was not capable of
taking care off, without undue distress.
This resulted in complete overhauling of the compressor foundation and stiffening
the pipe racks by additional bracings and all these re-engineering resulted in a delay
of about 5 months for full scale production and also a total revenue loss to company
in the tune of 300 million dollars
1
.
2.1.2 Case history #2
In another case a medium scale factory requiring heavy duty power for its production
opened a new unit in an industrial area (in India), where during peak summer season
power supply was reported to be sporadic. To maintain optimum production level,
the owner procured 3 numbers of standby generator sets to supply power during
periods of power cuts. During soil investigation nobody thought of doing a test for
dynamic property of the soil and the foundation was designed based on obtaining
the dynamic properties by theoretical co-relation with other static engineering soil
parameters. After generators were installed and started operating it was found that
amplitude of vibration was well beyond tolerable limits resulting in tripping of the
machines quite often. As the generator failed to meet the optimal power demand,
production output nose dived quite substantially.
On investigation of the problemby a consultant hired by the owner it was found that
field observed dynamic properties varied widely with those considered fromtheoretical
co-relation for the amplitude and resonance check.
The consultant suggested that the generator foundation be modified by providing
additional mass of concrete, adjusting the height of foundation and partially re-routing
the cables adding to an additional cost of only 20% of original installation cost. The
suggestion was vetoed by the owner arguing that company was going through difficult
times financially and it was not possible to generate further funds for such additional
capital expenditures.
The owner hired a couple of mechanics and with some adjustments coerced the
machines into operation but still the performance did not improve significantly. Within
1 This was time when oil was priced at 25 dollars per barrel. In todays index the loss would be 4 to 5
times the actual loss incurred.
Analysis and design of machine foundations 85
a period of six months the machines broke down completely. The equipment supplier
refused to replace the machines (though they failed within the warranty period) arguing
that conditions put in the contract in terms of amplitude and frequency restrictions
were violated from the very outset and as such they were not responsible for bad
performance of the machine.
By this time the company was in such a poor condition financially due to failure
of production target, that it could not generate fund to replace and overhaul the
equipment and its foundations and had no other option but to declare it sick and close
the unit.
So lesson learnt from the above two cases are that if proper attention is not paid to
the design of these type of foundations, consequences could be quite far reaching and
serious in nature.
2.2 DIFFERENT TYPES OF FOUNDATIONS
Foundations supporting machines can be classified into the following categories:
Block foundations resting on soil or piles
Frame foundations
Wall foundations
Spring mounted machines resting on rafts/grade slabs
2.2.1 Block foundations resting on soil/piles
These types of foundations usually consist of massive RCC blocks resting on soil or
on piles and are as shown in Figures 2.2.1 and 2.
Block type of foundation are usually used to support machines like
1 Pumps
2 Motors
3 Generators
Figure 2.2.1 Block foundation resting on soil.
Figure 2.2.2 Block foundation on piles.
4 Coal mill foundations
5 Gas Turbines
6 Motor Driven Boiler feed pumps
7 Centrifugal/Reciprocating type Compressors etc.
2.2.2 How does a block foundation supporting rotating
machines differ from a normal foundation?
The function of a foundation is basically to transfer the load coming on it from super-
structure or items like vessels, tanks, skids etc. to the underlying soil. For a normal
foundation supporting systems like building structure, vessels, tanks etc. the major
load coming on it is static in nature. For geometric sizing of the foundation the stress
induced on soil being less than the allowable bearing capacity of the soil sufce. The
dynamic force coming if any are quite rare and are mostly those due to earthquake
forces and in case of very tall structures may be a bit more often due to wind induced
vibration. In majority of the time for a normal foundation static load pre-dominates.
While for machine foundation it is just the reverse. In most of the industrial facilities
production being round the clock, the major load coming on the foundation is dynamic
in nature and the foundation should be so designed that it is capable enough to sustain
this dynamic loads over and above the static loads without causing any distress to
underlying soil or to the machine it is supporting.
So the question boils down to what are these conditions for which the foundation
can safely sustain the dynamic load coming from the machines in operation?
There are usually two conditions that are checked for while designing a machine
foundation:
Resonance check
Amplitude check
2.2.2.1 Resonance check
All machines under operation usually induce a periodic dynamic load on the founda-
tion and in most of the case can be represented by a function like P
0
sin
m
t, where
P
0
= magnitude of the unbalanced force from the machine during its operation and
m
= operating frequency of the machine usually expressed in
radians/sec.
Hertz or revolution per minute (rpm).
Due to this induced dynamic load from the machine the block foundation including
some portion of the soil underlying the foundation is subjected to vibration and it is
essential that the natural frequency (
n
) of this vibration should be well away from
the operating frequency of the machine i.e. resonance condition should not prevail.
Irrespective of any code the normal practise is to design the foundation in such a
way that its operating frequency is at least 20% away from the natural frequency of
the foundation.
2.2.2.2 Amplitude check
Under this condition, it is usually checked that the amplitude of vibration of the block
foundation is well within the acceptable limits of engineering practise.
The acceptable or the tolerable limits are usually suggested by the vendor supply-
ing the equipment or in absence of such data are usually obtained from the codal
stipulations.
If the amplitude of vibration is more than this acceptable limit can mar the
performance of the equipment in the following way
Rapid deterioration of the machine due to heavy wear and tear.
Excess amplitude of vibration inducing fatigue in the coupling and the connecting
shafts leading to repeated breakdowns.
Damage to the piping system connected rigidly to the equipment.
Increase in decibel level during operation causing discomfort to the operators.
Accumulation of explosive gases which at times could be dangerous to human life
and property.
Based on the above discussion it is imperative that for a foundation designed for
dynamic load the above two conditions are met.
Now let us see how we mathematically model the soil-foundation system to
theoretically check the two conditions as mentioned above.
A block foundation as shown earlier constitute of a massive RCC block resting on
the ground supporting the machine aligned over it.
For all practical purpose the block and the machine is considered as a rigid lumped
mass supported on an elastic base constituting the underlying soil/pile.
K
h
K
K
v
Figure 2.2.3 Mathematical model of foundation in 2D.
2
2
1
1
3
3
Figure 2.2.4 Degrees of freedom in space.
Where, for analysis purpose the soil is modelled as equivalent linear springs.
Shown in Figure 2.2.3 is the mathematical model of a machine foundation with soil
modelled as linear springs based on mechanical analog of elastic half space theory in
2D, and 6 degrees of freedom it has on space (Figure 2.2.4).
Before we go into further details of the state of the art theory for design of such
foundations, it would possibly be worthwhile to look back at its evolution and study
its subsequent metamorphosis to the various techniques used in present day design
office practices.
2.2.3 Foundation for centrifugal or rotary type
of machine: Different theoretical methods
for analysis of block foundation
2.2.3.1 Tschebotarioffs (1953) method
This is one of the early methods used for calculating the natural frequency of a founda-
tion. If f
n
is the natural frequency of the machine plus foundation in terms of contact
area then Tschebotarioff defined a term as reduced natural frequency f
nr
given by
f
nr
= f
n
, where =
W
A
f
t/ft
2
(2.2.1)
where W is the weight of the machine plus foundation and A
f
is the base area.
2.2.3.2 Alpans (1961) method
Alpan made use of Tschebotarioffs theory and developed an expression of natural
frequency as
f
n
=

W
(A
1/4
f
) (2.2.2)
where, f
n
= natural frequency of the foundation in cycles per minute; W = weight
of the machine plus foundation in Kilogram; A
f
= contact area of the foundation in
square meter, and = a constant whose value depends on the nature of the soil as
given in Table 2.2.1.
2.2.3.3 Newcombs (1951) method
Newcomb developed an empirical equation for natural frequency based on the static
deflection of the soil data and is expressed as
f
n
= 188
_
1
st
(2.2.3)
where, f
n
= natural frequency in cycles per minute and,
st
= static deflection in
inches.
The displacement parameter
st
can be obtained from plate load test for any design
bearing pressure.
The above theories were mostly based on observation and experience and as such
are empirical in nature. The theories can be put to use to check the resonant condition
only no check for amplitude is possible by these methods. As such they shall only be
used for preliminary design or sizing of the foundation only.
Table 2.2.1 Value of constant, .
Sl. No. Soil type Value of
1 Peat 3900
2 Plastic clay 69000
3 Sand 82000
4 Sand stone 111000
2.2.4 Analytical methods
2.2.4.1 Hsiehs (1962) method
Considering the soil as semi-infinite elastic medium, Hsieh put forward an analytical
treatment for vibration of circular foundation as given hereunder
For Translation: m x +r
2
0
F
2
_
G x +r
0
GF
1
x = P
0
sin
m
t (2.2.4)
For rocking:
+r
4
F
2
_
G x +Gr
3
F
1
= M
0
sin
m
t (2.2.5)
where, m = mass of the foundation plus the machine; x, = displacement vectors in
translational and rotational mode respectively; = mass moment of inertia; = mass
density of soil; G=dynamic shear modulus of the soil; r
0
= radius of the circular foun-
dation in translational mode which for rectangular foundations is
_
LB/; r
= radius
of the circular foundation in rocking mode which for rectangular foundations is
4
_
LB
3
3
or
4
_
L
3
B
3
as the case may be; a
0
= frequency factor and is given by
m
r
0
_
/G;
m
=
the operating frequency of the machine; P
0
, M
0
= amplitude of exciting force in trans-
lation and rocking mode, and F
1
, F
2
= are functions whose values are given below:
F
1
is usually represented in the form,
1

2
a
2
0
.
Now substituting this value of F
1
in the above differential equation we have,
For translation mode,
(m+
2
r
3
0
) x +r
2
0
F
2
_
G x +r
0
G
1
x = P
0
sin
m
t (2.2.6)
For rocking mode,
( +
2
r
5
+r
4
F
2
_
G x +
1
Gr
3
= M
0
sin
m
t (2.2.7)
Representing the above equations as
m
x
x +c
x
x +k
x
x = P
0
sin
m
t and I
+c

+k
= M
0
sin
m
t (2.2.8)
where, m
x
= (m +
2
r
3
0
); c
x
= r
2
0
F
2
G; k
x
= r
0
G
1
; I = ( +
2
r
5
); c
=
r
4
F
2
G; k
=
1
Gr
3
.
We had already seen earlier
2
that the solution of such equation can be repre-
sented as
2 Refer Chapter 3 (Vol. 1) for solution of such equations having damped single degree of freedom.
x
max
=
P
0
k
x
sin
m
t
_
[(1 r
2
)
2
+(2Dr)
2
]
(2.2.9)
in which, r =
m
/
nx
and D = c/c
c
where,
nx
=
_
k
x
/m
x
; and c
c
= critical damping
of the system and is 2
_
m
x
k
x
.
And for rocking mode,
max
=
M
0
k
sin
m
t
_
[(1 r
2
)
2
+(2Dr)
2
]
(2.2.10)
in which, r =
m
/
n
and D = c/c
c
where,
n
=
_
k
/I; and c
c
= critical damping
of the system and is 2
_
Ik
.
Here one point needs to be noticed is the additional term in the inertial coefficient
m
x
= (m+
2
r
3
0
) and I = ( +
2
r
5
) (2.2.11)
Here the original mass and mass moment of inertia terms get added up with an
additional term of
2
r
3
0
and
2
r
5
respectively.
This can be attributed as added mass of soil which starts vibrating with the
foundation in same phase.
This looks quite logical for it has indeed been observed during field observation
that a part of soil below foundation do indeed participates in the vibration of the
foundation system.
Table 2.2.2 gives values of F
1
and F
2
for various modes given by Hsieh.
For uniformand parabolic distribution of pressure, Hsieh suggest to use an effective
radius r
0
where is 0.78 and 0.59 respectively.
Table 2.2.2 Values of F
1
and F
2
.
Mode Poissons ratio F
1
F
2
Vertical (0 < a
0
< 1.5) 0.0 4.0 0.5a
2
0
3.3 + 0.4a
0
0.25 5.3 1.0a
2
0
4.4 + 0.8a
0
0.50 8.0 2.0a
2
0
6.9
Horizontal (0 < a
0
< 2.0) 0.0 4.5 0.2a
2
0
2.4 + 0.3a
0
0.25 4.8 0.2a
2
0
2.5 + 0.3a
0
0.50 5.3 0.1a
2
0
2.8 + 0.4a
0
Rocking (0 < a
0
< 1.5) 0.0 2.5 0.4a
2
0
0.4a
0
Torsion (0 < a
0
< 2.0) All 5.1 0.3a
2
0
0.5a
0
2.2.4.2 Barkan or IS-2974 method
This is by far the most popular method in the design office. Barkan (1962) developed
this method way back in 60s and is still in vogue for design of machine foundations
under rotating loads.
In this method Barkan assumed the block foundation, shown in Figure 2.2.5, as
a rigid lumped mass (i.e. he assumed the concrete block to have infinite stiffness in
comparison to the soil and neglected any internal deformation of the concrete block
itself) having three degrees of freedom.
The soil medium he idealised as linear springs which he defined in terms of soil
parameter c
z
, c
x
& c
which are otherwise known as coefficient of elastic uniform

compression, coefficient of elastic uniform shear, coefficient of elastic non-uniform
compression respectively.
2.2.4.2.1 Vertical direction
Here in the vertical direction the spring constant is considered as
k
z
= c
z
A
f
(2.2.12)
where, k
z
= equivalent spring in vertical direction; c
z
= coefficient of elastic uniform
compression, and A
f
= plan area of the foundation.
C/L of machine shaft
h
mx
x
H
Z
c
mz
H
x
0
V
P
z
sin
m
t
M
x
sin
m
t
P
x
sin
m
t
Figure 2.2.5 Mathematical model of Barkan for vertical and coupled sliding and rocking motion about
the minor axes of the foundation.
The natural frequency of the foundation is given by
z
=
_
k
z
m
and, (2.2.13)
the amplitude of vertical vibration is given by
z
=
P
z
sin
m
t
k
z
1 r
2
(2.2.14)
where, r =
m
/
n
;
m
= operating frequency of the machine.
2.2.4.2.2 For coupled horizontal and rocking mode
It has been observed by Barkan that when a foundation has horizontal force along its
minor axis the foundation undergoes sliding and rocking simultaneously. When the
foundation starts vibrating resistance is mobilised in the soil in terms of forces V and
H as shown in Figure 2.2.5.
The resistive force may thus be expressed as
V
R
=
_
C
dA (2.2.15)
and the resistive moment is expressed
M
R
=
_
C
2
dA = C
I
A
(2.2.16)
where = distance between rotation axis and the element of area dA; = angular
rotation of the machine foundation; I
A
= second moment of area of the foundation
contact surface with respect to the axis passing through the centroid of the area and
perpendicular to the plane of the vibration.
2.2.4.2.3 For horizontal force, H
H = c
A
f
x
0
= c
A
f
(x Z
c
) (2.2.17)
where, A = area of base contact; Z
c
, x, x
0
etc. are shown in Figure 2.2.5.
Now applying DAlemberts equation for dynamic equilibrium
3
, we have
m x +H = P
x
sin
m
t or m x +c
A
f
(x Z
c
) = P
x
sin
m
t (2.2.18)
where m is the mass of the machine foundation.
3 Refer Chapter 5 (Vol. 1) for definition of DAlemberts equation.
Similarly for the moment equation about the minor axis of the foundation we have
J
x

c
A
f
Z
c
x +(c
I
A
WZ
c
+c
AZ
2
c
) = M
x
sin
m
t (2.2.19)
where J
x
= mass moment of inertia of the machine-foundation block about the minor
axis of rotation.
From Equations (2.2.18) and (2.2.19), we see that they contain both x and , so a
coupled sliding and rocking motion will develop along this direction.
Using the above equations and considering free vibrations, Barkan developed the
following equation for calculation of the frequencies.
J
0
(
2
+
2
x
)
2
J
x
+

2
2
x
J
0
J
x
= 0 (2.2.20)
where J
0
= J
x
+mZ
2
c
;
2
=
c
I
A
WZ
c
J
0
and
2
x
=
c
A
f
m
.
Based on the above, the two principal frequencies for the coupled vibration is
given by
2
1,2
=
J
0
2J
x
_
_
+
2
x

_
(
2
x
+
2
)
2
4J
x
2
x
J
0
_
_
(2.2.21)
Considering the forced vibration, the amplitudes A
x
, A
may be expressed as
A
x
=
(c
I
A
WZ
c
+c
A
f
Z
2
c
J
x
2
m
)P
x
c
A
f
Z
c
M
x
mJ
x
(
2
1

2
m
)(
2
2

2
m
)
sin
m
t
A
=
c
A
f
Z
c
P
x
(c
A
f
m
2
m
)M
x
mJ
x
(
2
1

2
m
)(
2
2

2
m
)
sin
m
t
(2.2.22)
2.2.4.2.4 Torsional mode
For this mode, again the foundation considered is a lumped mass having single degree
of freedom when the frequency and amplitude are given by
=
_
K
and =
T sin
m
t
k
1 r
2
(2.2.23)
where, K
= c
; r =
m
/
n
.
This method is also recommended by IS 2974 Code and design practices for
machine foundation and still remains the most popular method for vibration analysis
of block foundations.
But let us see the limitations of Barkans method with respect to the reality under
field conditions.
2.2.4.2.5 The Limitations
The major limitations that can be attributed to Barkans method are as follows:
Barkans model does not take damping into consideration. It has been observed
fromfield instrumentation data that damping plays a significant role in the overall
response of the foundation especially when the operating frequency of the machine
is low.
It does not account for the embedment effect of the surrounding soil which could
play a significant role on the magnitude of soil stiffness and damping.
It does not take into cognisance the virtual mass of soil which vibrates in same
phase with the machine and the foundation.
Barkan suggested spring value (usually the coefcient of uniform elastic compres-
sion) of the soil to be obtained from dynamic plate load test
4
and may only give
correct values for a shallow depth below the surface while this may not be valid
for layered soil and also when the contact area of the foundation is large.
So based on the above limitations it was felt to upgrade the mathematical process
in the design of machine foundation.
Before we study the further enhancements it would be worth to write Barkans
equation in a more generic form.
The soil spring stiffness is described by the terms
K
x
= c
A
f
and K
x
= c
I
A
, (2.2.24)
while, the equations of equilibrium are defined as
m x +c
A
f
(x Z
c
) = P
x
sin
m
t, and
J
x

c
A
f
Z
c
x +(c
I
A
WZ
c
+c
A
f
Z
2
c
) = M
x
sin
m
t (2.2.25)
Substituting the values of K
x
and K
, we have
m x +K
x
(x Z
c
) = P
x
sin
m
t and
J
x

K
x
Z
c
x +(K
WZ
c
+K
x
Z
2
c
) = M
x
sin
m
t (2.2.26)
The above on writing in matrix form can be represented as
_
m 0
0 J
x
_ _
x
_
+
_
K
x
K
x
Z
c
K
x
Z
c
K
+K
x
Z
2
c
WZ
c
_ _
x
_
=
_
P
x
M
x
_
sin
m
t (2.2.27)
Based on the above equation we will see later howwe develop further realistic model
of the coupled horizontal and rocking mode
5
.
4 Usually carried out with a plate of size 300 mm 300 mm.
5 Structural Engineers be alert from this point.What we are going to apply herein subsequently are the
theories of structural dynamics for system with two degrees of freedom.
2.2.4.3 Richart and Lysmers model
Richart et al. (1970) idealised the foundation as a lumped mass supported on soil
which is idealised as frequency independent springs which he described in terms of
soil parameter dynamic shear modulus or shear wave velocity of the soil for circular
footing when footings having rectangular shape in plan can be converted into a footing
having equivalent circular radius.
Tables 2.2.3 and 4 along with Figure 2.2.6 show the different values of spring and
damping value as per Richart and Lysmer.
In which, G = dynamic shear modulus of the soil and is given by, G =
s
V
2
s
;
= Poissons ratio of the soil;
s
= mass density of the soil; V
s
= shear wave
velocity of the soil obtained from soil testing; g = acceleration due to gravity;
m = mass of the machine and foundation; J = mass moment of inertia of the
machine and the foundation about the appropriate axes; K = equivalent spring
stiffness of the soil; C = damping value of the soil; B = inertial factor contribut-
ing to the damping factor; D = damping ratio of the soil; r = equivalent radius
of a circular foundation; L = length of the foundation, and, B = width of the
foundation.
Many engineers in design offices prefer to use Richarts springs neglecting the damp-
ing and use Barkans formulation in matrix form as shown earlier to find out the
natural frequency and amplitude of the foundation. But, by neglecting the damp-
ing, he could signicantly over-estimate the amplitude of vibration (specially for low
tuned machines) thus adding to the cost by trying to restrict it within the acceptable
limits.
Let us now see as to what form the equations take when damping is introduced in
the system.
Table 2.2.3 Values of soil springs as per Richart and Lysmer (1970) model.
Sl. No. Direction Spring value Equivalent radius Remarks
1 Vertical K
z
=
4Grz
(1 )
r
z
=
_
LB
This is in vertical Z direction

2 Horizontal K
x
=
32(1 )Gr
x
(7 8)
r
x
=
_
LB
This induce sliding in horizontal

X or Y direction
3 Rocking K
x =
8Gr
3
x
3(1 )
r
x
=
4
_
LB
3
3
This produces rocking about
Y axis
3.1 Rocking K
y
=
8Gr
3
y
3(1 )
r
y
=
4
_
L
3
B
3
X axes
4 Twisting K
=
16Gr
3
3
r
=
4
_
L
3
B +BL
3
6
This produces twisting about
vertical Z axis
Table 2.2.4 Values of soil damping as per Richart and Lysmer (1970) model.
Damping ratio and
Sl. No. Direction Mass ratio (B) Damping value Remarks
1 Vertical B
z
=
0.25 m(1 )g
s
r
3
z
z
=
0.425
B
z
, This is damping value is in
vertical Z direction.
C
z
= 2
z
K
z
m
2 Horizontal B
x
=
(7 8)mg
32(1 )
s
r
3
x
x
=
0.288
B
x
, This damping value is in
lateral X or Y direction
C
x
= 2
x
K
x
m
3 Rocking B
x
=
0.375(1 )J
x
g
s
r
5
x
x
=
0.15
(1 +B
x
)
_
B
x
, This damping value is for
rocking about Y direc-
tion
C
x
= 2
x
_
K
x
J
x
3.1 Rocking B
y
=
0.375(1 )J
y
g
s
r
5
y
y
=
0.15
(1 +B
y
)
_
B
y
, This damping value is for
rocking about Y axes
C
y
= 2
y
_
K
y
J
y
4 Twisting B
=
J
s
r
5
=
0.5
1 +2B
, This damping value is valid

for twisting about verti-
cal Z axis.
C
= 2
_
K
Z
B
Y
L
X
Figure 2.2.6 3D View of the block foundation.
2.2.4.3.1 Vertical motion considering damping of the soil
For vertical direction the equation becomes that of a lumped mass having single degree
of freedom when
m z +C
z
z +K
z
z = P
0
sin
m
t (2.2.28)
solution is,
z
=
_
K
z
m
and
z
=
(P
0
/K
z
) sin
m
t
_
(1 r
2
)
2
+(2D
z
r)
2
(2.2.29)
where r =

m
n
and D
z
= damping ratio.
2.2.4.3.2 Coupled horizontal and rocking motion considering damping soil
We have seen that based on Barkans formulation the equation of motion in matrix
form is
_
m 0
0 J
x
_ _
x
_
+
_
K
x
K
x
Z
c
K
x
Z
c
K
+K
x
Z
2
c
WZ
c
_ _
x
_
=
_
P
0
M
0
_
sin
m
t (2.2.30)
Since the above equation is based on DAlemberts equation, the equation are said
to be statically coupled when the stiffness matrix and damping matrix have the same
matrix form (Meirovitch 1975). Thus, based on the above argument the damped
equation of motion in coupled rocking and sliding mode becomes
_
m 0
0 J
x
_ _
x
_
+
_
C
x
C
x
Z
c
C
x
Z
c
C
x
+C
x
Z
2
c
WZ
c
_ _
x
_
+
_
K
x
K
x
Z
c
K
x
Z
c
K
x
+K
x
Z
2
c
WZ
c
_ _
x
_
=
_
P
0
M
0
_
sin
m
t (2.2.31)
The above equations constitute the complete equation of motion for coupled sliding
and rocking mode considering the damping effect of the soil.
Actually for all practical calculations for finding out the dynamic response of the
foundation the term WZ
c
is usually neglected, for it has been observed that unless
the foundation is very massive and deep the term WZ
c
has no significant effect on the
overall response of the system.
Based on the above argument the above equation reduces to
_
m 0
0 J
x
_ _
x
_
+
_
C
x
C
x
Z
c
C
x
Z
c
C
x
+C
x
Z
2
c
_ _
x
_
+
_
K
x
K
x
Z
c
K
x
Z
c
K
x
+K
x
Z
2
c
_ _
x
_
=
_
P
0
M
0
_
sin
m
t (2.2.32)
The equation above surely looks elegant, but now comes the catch . . . , for this
damping matrix of soil is not proportional to either the mass or the stiffness of the soil,
moreover they are coupled by the termof Z
c
and W (the weight of the foundation) and
as such do not de-couple on orthogonal transformation
6
. This forms a major headache
to the designer as he is not in a position to guess the damping ratio.
As we had already discussed in Chapter 5 (Vol. 1) that the most appropriate tech-
nique in such case is to resort to Time history analysis for the correct answer, many
engineers find time history too intensive in terms of calculation
7
and prefer to use
modal response technique as a tool for analysis of the same. Of course the easiest way
out is to neglect the damping and argue that the design is conservative!
But this need not be done for we have already stated in Chapter 5 (Vol. 1) that it is
possible to by pass this problem of orthogonal de-coupling, even when the damping
matrix is non-proportional which though not exact would give still give a designer
a reasonable value to estimate a more realistic amplitude of vibration (it is surely a
better value than no damping considered).
We will consider application of this technique in subsequent section (section 2.2.5).
2.2.4.3.3 Torsional mode
In this mode again the block foundation is again considered as a lumped mass having
single degree of freedom, natural frequency and the torsional rotation, is given by
=
_
K
; and, =
T sin
m
t
K
_
(2Dr)
2
+(1 r
2
)
2
(2.2.33)
where K
= 16Gr
3
/3, D is the damping ratio in the torsion mode (Table 2.2.4) and
r is the ratio between the natural frequency of the foundation in torsion mode and the
operating frequency of the machine.
2.2.5 Approximate analysis to de-couple equations
with non-proportional damping
We have seen that the equation of motion is given by
_
m 0
0 J
x
_ _
x
_
+
_
C
x
C
x
Z
c
C
x
Z
c
C
x
+C
x
Z
2
c
WZ
c
_ _
x
_
+
_
K
x
K
x
Z
c
K
x
Z
c
K
x
+K
x
Z
2
c
WZ
c
_ _
x
_
=
_
P
0
M
0
_
sin
m
t
(2.2.34)
For finding the natural frequencies we perform the eigen value analysis when the
un-damped equation becomes (neglecting-WZ
c
for reasons as cited earlier)
_
K
x
m K
x
Z
c
K
x
Z
c
K
+K
x
Z
2
c
J
x
_
= 0 (2.2.35)
6 For further explanation on this property refer to Chapter 5 (Vol. 1), the topic of orthogonal trans-
formation for modal response technique.
7 And the theory underlying time history is eluding many is not too an uncommon a fact. . ..
Solving the above equation we find out the eigen value vis-a vis the natural
frequency of the foundation system. Let the eigen values be
1
and
2
respectively. Let
corresponding eigen vectors be
xx

x
T
and
x

T
respectively, when the
complete eigen vector matrix is expressed as,
_
xx

x
_
.
Since the eigen vector is known separately for each mode we find out the damping
ratio separately for each mode as follows.
As a first step we perform the operation {}
T
[C]{} for each mode.
For the first mode, we have
xx

x
_
C
x
C
x
Z
c
C
x
Z
c
C
x
+C
x
Z
2
c
_ _
xx
x
_
which gives,
xx

x
_
C
x
xx
C
x
Z
c
x
C
x
Z
c
xx
+(C
x
+C
x
Z
2
c
)
x
_
= C
x
2
xx
2C
x
Z
c
xx
+(C
x
+C
x
Z
2
c
)
2
x
(2.2.36)
It should be realised that the above is a unique value and we also know that the
operation {}
T
[C]{} breaks up the equation to form 2D
i
i
where i is the degrees of
freedom of the system.
Now considering,
2D
i
i
= C
x
2
xx
2C
x
Z
c
xx
+(C
x
+C
x
Z
2
c
)
2
x
,
for the first mode,
D
1
=
C
x
2
xx
2C
x
Z
c
xx
+(C
x
+C
x
Z
2
c
)
2
x
2
1
(2.2.37)
where D
1
= damping ratio for the first mode and;
1
= first natural frequency of the
foundation.
Similarly, for the second mode proceeding in same manner it can be proved that
D
2
=
C
x
2
x
2C
x
Z
c
+(C
x
+C
x
Z
2
c
)
2
2
2
(2.2.38)
Once the damping ratios are identied we assume, [C] = [M] + [K] and
performing the operation
{}
T
[C]{} = {}
T
[M]{} +{}
T
[K]{}, (2.2.39)
We have, for two degrees of freedom
2D
1
1
= +
2
1
and 2D
2
2
= +
2
2
(2.2.40)
Thus, we have two equation with two unknowns, and , and solving the above
two equations we get the value of and .
Once these values are known one can obtain an equivalent proportional soil damping
from the operation [C] = [M] +[K] which is now quite suitable for modal response
technique.
We now further explain the above method based on a suitable numerical problem.
Example 2.2.1
For a block foundation supporting a centrifugal pump was observed to have the
following design data M = 50 kN sec
2
/m, J = 100 kN sec
2
m, Z
c
= 1.5 m,
K
x
= 3000 kN/m, K
= 5000 kN/m, C
x
= 200 kN/m, C
350 kN/m.
Find out
The natural frequencies in coupled horizontal and rocking mode.
The normalized eigen vectors.
Find out the approximate damping ratios for each mode.
Correct the damping matrix based on equivalent Rayleigh damping.
Solution:
The complete equation of motion for the foundation under coupled rocking and
sliding mode is given by
_
m 0
0 J
x
_ _
x
_
+
_
C
x
C
x
Z
c
C
x
Z
c
C
x
+C
x
Z
2
c
_ _
x
_
+
_
K
x
K
x
Z
c
K
x
Z
c
K
x
+K
x
Z
2
c
_ _
x
_
=
_
P
0
M
0
_
sin
m
t
Based on the above the equation of motion becomes
_
50 0
0 100
_ _
x
_
+
_
200 300
300 988
_ _
x
_
+
_
3000 4500
4500 11750
_ _
x
_
=
_
P
0
M
0
_
sin
m
t
Calculation of the un-damped natural frequency
To find out the natural frequencies we have
_
3000 50 4500
4500 11750 100
_
= 0
35.2510
6
310
5
5.87510
5
+5000
2
20.2510
6
=0
The above equation on simplification reduces to
2
177.5 +3000 = 0
1
=
177.5 +
_
(177.5)
2
413000
2
= 158.5 = 12.58 rad/sec
and
2
=
177.5
_
(177.5)
2
413000
2
= 18.92 = 4.35 rad/sec
For the rst mode we have
_
3000 5018.92 4500
4500 11750 10018.92
_ _
x
_
= 0
Solving the above two homogeneous equations, considering x = 1.00, we
have, x : = 1.00 : 0.45644.
Similarly for the second mode we have
_
3000 50158.5 4500
4500 11750 100158.5
_ _
x
_
= 0
Solving the above two homogeneous equations, considering x = 1.00
We have, x : = 1.00 : 1.094
Thus, the complete eigen vector matrix becomes
_
x
_
i=1,2
=
_
1.00 1.00
0.45644 1.094
_
Calculation of normalized eigen vectors
On operation {}
T
[M]{} we have
For the first mode
_
1.00 0.45644
_
_
50 0
0 100
_ _
1.00
0.45644
_
= 70.83
_
M
r
=
70.83 = 8.416
Thus {}
i=1
=
_
0.1188173
0.0542329
_
For the second mode, on operation {}
T
[M]{}, we have
1.00 1.094
_
50 0
0 100
_ _
1.00
1.094
_
= 130.26
_
M
r
=
169.6836 = 13.026
Thus {}
i=2
=
_
0.07676
0.08398
_
The normalized eigen vectors for the two modes are,
_
x
_
i=1,2
=
_
0.1188173 0.07676
0.0542329 0.08398
_
Calculation of the modal damping ratios
Nowif we performthe orthogonal operation with the complete normalized eigen
vector matrix it will be observed that while {}
T
[M]{} diagonalise to [I] and
{}
T
[K]{} [].
But as will be seen nowthat {}
T
[C]{} will NOTde-couple to the form2D
i
i
.
_
0.1188173 0.0542329
0.07676 0.08398
_ _
200 300
300 988
_ _
0.1188173 0.07676
0.0542329 0.08398
_
=
_
0.1188173 0.0542329
0.07676 0.08398
_ _
7.49359 40.546
18.130205 106.00024
_
=
_
1.8736 0.9311626
0.9473666 12.014236
_
Since the above matrix has off-diagonal terms will NOT be equal to zero,
hence we conclude that the matrix has not de-coupled due to orthogonal
transformation.
As such we treat the eigen vectors separately for each individual modes,
Thus for the first mode we have
0.1188173 0.0542329
_
200 300
300 988
_ _
0.1188173
0.0542329
_
= 1.86308
2D
1
1
= 1.86308.
or, D
1
=
1.86308
24.35
= 0.214
0.07676 0.08398
_
200 300
300 988
_ _
0.07676
0.08398
_
= 12.0142
D
2
=
12.0142
212.58
= 0.4775.
Thus damping ratio is of the order of 21.4% for the first mode and 47.75%
for the second mode.
Correction to the damping matrix based on Rayleigh damping
Let us assume the damping matrix is of the form [C] = [M] + [K], then on
orthogonal transformation we have {}
T
[C]{} = {}
T
[M]{} +{}
T
[K]{},
which gives 2D
1
1
= +
2
1
and 2D
2
2
= +
2
2
.
On substituting the respective values, we have
For the first mode, +18.92 = 20.2144.35 +18.92 = 1.8618
For the second mode, + 158.5 = 20.477512.58 + 158.5 =
12.0139
Solving the above two equations we have = 0.48568 and = 0.0727
As [C] = [M] +[K] we have
[C] = 0.48568
_
50 0
0 100
_
+0.0727
_
3000 4500
4500 11750
_
=
_
242 327.15
327.15 902.795
_
is the modified damping matrix which satisfies the orthogonal property.
If we compare the value with the original damping matrix obtained from the
soil property we see each of the term has got slightly modified.
For practical design office calculation this is usually deemed sufficient. It at least
depicts a better result than no damping considered at all.
We will subsequently see howdata based on modified damping matrix compare with
time history response which we had stated would be the most appropriate accurate
method that could be adopted with non-proportional damping.
But prior to that let us evaluate another form in which equations of motion for
coupled rocking and sliding motion can be formulated too.
2.2.6 Alternative formulation of coupled equation
of motion for sliding and rocking mode
We had already shown in Chapter 5 (Vol. 1) that other then formulating equation of
motion based on DAlemberts equation we can also write themdown based on energy
concepts as put forward by Lagrange.
We use here Lagranges formulation to derive the equation of motion for coupled
rocking and sliding motion.
We had shown previously in chapter 5 (Vol. 1) that considering a conservative
system having kinetic energy as T and potential energy as U we have
d(T +U) = 0 and d(T +U) =
n
i=1
_
d
dt
_
T
q
i
_
T
q
i
+
U
q
i
_
dq
i
= 0 (2.2.41)
For the machine foundation subjected to coupled rocking and sliding motion
Kinetic energy (T) =
1
2
m( x +Z
c

)
2
+
1
2
J
2
and
The Potential Energy (U) =
1
2
K
x
x
2
+
1
2
K
2
Based on the above equation we have
T =
1
2
m( x +Z
c

)
2
+
1
2
J
2
T
x
= m( x +Z
c

) and
d
dt
_
T
x
_
= m( x +Z
c

) = m x +mZ
c

We have, U =
1
2
K
x
x
2
+
1
2
K
2
;
U
x
= K
x
x
Also, T =
1
2
m( x +Z
c

)
2
+
1
2
J
2
;
T
= mZ
c
( x +Z
c

) +J
d
dt
_
T
_
= mZ
c
x +mZ
2
c

+J
For U =
1
2
K
x
x
2
+
1
2
K
2
;
U
= K
Substituting the above values in the Lagrangian equation, we have

m x +mZ
c

+K
x
x = 0; and mZ
c
x +(mZ
2
c
+J)
+K
= 0 (2.2.42)
Thus writing in matrix form the free vibration equation becomes
_
m mZ
c
mZ
c
mZ
2
c
+J
_ _
x
_
+
_
K
x
0
0 K
_ _
x
_
= 0 (2.2.43)
Equation (2.2.42) is known as dynamically or inertially coupled equation. You will
observe in contrary to the formulation based on DAlemberts equation, where the
mass matrix is diagonal here the stiffness matrix is diagonal while the mass matrix is
non-diagonal but symmetric.
For equation coupled by inertia it has been observed that both the damping and
stiffness matrix remains diagonal (Meirovitch 1975) and the complete equation of
motion thus becomes
_
m mZ
c
mZ
c
mZ
2
c
+J
_ _
x
_
+
_
C
x
0
0 C
_ _
x
_
+
_
K
x
0
0 K
_ _
x
_
=
_
P
0
M
0
_
sin
m
t
(2.2.44)
Now that we have established the equation the question that obviously crops up
in mind is what is the advantage of this equation over the normal equation that was
derived based on static coupling/Barkans equation.
The first thing we will see subsequently that the eigen-values remain invariant with
this formulation.
Moreover it has been observed that damping ratio derived by this method are quanti-
tatively closer to the values derived from classical analysis based on frequency domain
analysis in complex domain. (Wolf 1988).
The reason for the better prediction of damping ratio could be that the damping
matrix derived by this formulation is in uncoupled form.
We now further explain the above based on suitable numerical example.
Example 2.2.2
For a block foundation as described in Example 2.2.1, calculate the following
based on Lagranges Formulation.
The natural frequencies in coupled horizontal and rocking mode
The normalized eigen vectors
Approximate damping ratios for each mode.
All design parameters pertaining to design remains same as in Example 2.2.1.
Solution:
The complete equation of motion for the foundation under coupled rocking and
sliding mode based on Lagranges formulation is given by
_
m mZ
c
mZ
c
mZ
2
c
+J
_ _
x
_
+
_
C
x
0
0 C
_ _
x
_
+
_
K
x
0
0 K
_ _
x
_
=
_
P
0
M
0
_
sin
m
t
Based on the above the equation of motion becomes
_
50 75
75 212.5
_ _
x
_
+
_
200 0
0 350
_ _
x
_
+
_
3000 0
0 5000
_ _
x
_
=
_
P
0
M
0
_
sin
m
t
To find out the natural frequencies we have
_
3000 50 75
75 5000 212.5
_
= 0 which on expansion gives,
5000
2
8.87510
5
+1510
6
= 0 =
177.5 139.66
2
::
1
=
18.92,
1
= 4.35 rad/sec;
2
= 158.5,
2
= 12.58 rad/sec.
It should be observed that the natural frequencies are identical to the one
obtained in Example 2.2.1 based on static coupling.
Calculation of the eigen vectors
For the first mode we have
_
3000 5018.92 7518.92
7518.92 5000 212.518.92
_ _
x
_
= 0
_
2054 1419
1419 979.5
_ _
x
_
= 0
Solving the above homogenous equations considering, x = 1.00 we have
x : = 1.00 : 1.447
Similarly for the second mode
_
3000 50158.5 75158.5
75158.5 5000 212.5158.5
_ _
x
_
= 0
_
4925 11887.5
11887.5 28681.25
_ _
x
_
= 0
Solving the above homogenous equations considering, x = 1.00 we have
x : = 1.00 : 0.4143.
Calculation of the normalized eigen-vectors
For the rst mode
{
1
}
T
[M]{
1
} = 1.00 1.447
_
50 75
75 212.5
_ _
1.00
1.447
_
= 711.878
_
M
r
= 26.68
Thus, the normalized eigen vector is {
1N
} =
_
0.03748
0.05423
_
.
For the second mode we have
{
2
}
T
[M]{
2
} = 1.00 0.4143
_
50 75
75 212.5
_ _
1.00
0.4143
_
= 24.32945
_
M
r
= 4.93248
Thus, the normalized eigen vector is {
2N
} =
_
0.20273
0.08399
_
.
Calculation of the damping ratios
For the first mode
{
1
}
T
[C]{
1
} = 0.03748 0.05423
_
200 0
0 350
_ _
0.03748
0.05423
_
= 1.1632 2D
1
1
= 1.1632
i.e. D
1
=
1.1632
2 4.35
= 0.133.
For the second mode
{
1
}
T
[C]{
1
} = 0.20273 0.08399
_
200 0
0 350
_ _
0.20273
0.08399
_
= 10.688 2D
2
2
= 10.688
i.e. D
2
=
10.688
2 12.58
= 0.4248.
Thus
for the first mode, the damping ratio is 13.3% and
for the second mode, the damping ratio is 42.48%.
The table below gives comparative results based static and dynamic coupling
formulations for the examples solved above.
Natural Natural Normalized Normalized Damping Damping
Formulation frequency frequency mode shapes mode shapes ratio 1st ratio 2nd
basis 1st mode 2nd mode 1st mode 2nd mode mode mode
Static
coupling 4.35 12.58 0.1181 : 0.0542 0.0767 : 0.0839 21.4% 47.75%
Dynamic
coupling 4.35 12.58 0.0375 : 0.0542 0.2027 : 0.0839 13.3% 42.48%
Based on the above it would possibly be worthwhile to know how the amplitudes
vary based on the above two methods vis--vis the time history response which we
advocated as the most appropriate and correct method for handling responses having
non-proportional damping.
This is what we are going to establish based on suitable numerical example hereafter.
Example 2.2.3
For a block foundation as described in Example 2.2.1. Calculate the following
amplitude of vibration based on
Static Coupling
Dynamic coupling
Time history response
Discuss the results based on the three answers.
The unbalanced dynamic force is 200 KN having operating frequency of
750 r.p.m acting at a height 600 mm from the top of foundation.
The total height of the concrete block (L) is 1.8 m.
All other design parameters pertaining to design remains same as Example
2.2.1.
Solution:
The operating frequency of the machine is 750 r.p.m. = (7502)/60 =
78.53 rad/sec.
Solution based on static coupling
The equation of motion in matrix form is given by
[M]{
X} +[C]{
X} +[K]{X} = {P}
Here, the force matrix is given by
_
P
0
M
0
_
=
_
200
180
_
sin 78.53t
For static coupling,
The normalized eigen-vector obtained earlier was
[] =
_
0.1188173 0.076746
0.0542333 0.0839942
_
Considering the operation, []
T
{P} we have
[]
T
{P} =
_
0.1188173 0.0542333
0.076746 0.0839942
_ _
200
180
_
sin 78.53t
=
_
33.52
0.2293
_
sin 78.53t
On orthogonal transformation we have
{}
T
[M]{}{
} +{}
T
[C]{}{
} +{}
T
[K]{}{} = {}
T
{P}
This in decoupled form reduces to
{
1
} +2D
1
1
{
1
} +[
2
1
]{
1
} = p
0
sin
m
t;
{
2
} +2D
2
2
{
2
} +[
2
2
]{
2
} = m
0
sin
m
t
Substituting the different values, we have
{
1
} +20.2144.35{
1
} +18.92{
1
} = 33.52sin 78.53t and
{
2
} +20.477612.58{
2
} +158.5{
2
} = 0.230sin 78.53t

1
=
33.53sin 78.53t
_
(18.92 78.53
2
)
2
(1.86278.53)
2
= 5.4506191810
3
sin 78.53t
2
=
0.230sin78.53t
_
(158.5 78.53
2
)
2
(12.01678.53)
2
= 3.81623525910
5
sin 78.53t
Thus back transferring on global co-ordinate we have
_
x
_
=
_
0.1188173 0.076746
0.0542333 0.08399942
_ _
5.45061918
0.03816235259
_
10
3
sin 78.53t
=
_
0.6505566
0.2923994
_
10
3
sin 78.53t
Solution based on dynamic coupling
[] =
_
0.0374669 0.2027373
0.0542333 0.0839942
_
Considering the operation []
T
{P} we have
[]
T
{P} =
_
0.0374669 0.0542333
0.2027373 0.0839942
_ _
200
180
_
sin 75.53t
=
_
17.255374
25.428504
_
sin 78.53t
On orthogonal transformation we have
[]
T
[M][]{
} +[]
T
[C][]{
} +[]
T
[K][]{} = []
T
{P}
{
1
} +2D
1
1
{
1
} +[
2
1
]{
1
} = p
0
sin
m
t;
{
2
} +2D
2
2
{
2
} +[
2
2
]{
2
} = m
0
sin
m
t
Substituting the different values we have
{
1
} +20.1505967 4.35{
1
} +18.92{
1
} = 17.25537 sin 78.53t and
{
2
} +20.424812.58{
2
} +158.5{
2
} = 25.429054 sin 78.53t

1
=
17.25537 sin 78.53t
_
(18.92 78.53
2
)
2
(1.310278.53)
2
= 2.80625310
3
sin 78.53t
2
=
25.428504 sin 78.53t
_
(158.5 78.53
2
)
2
(10.86978.53)
2
= 4.190049410
3
sin 78.53t
_
x
_
=
_
0.0374669 0.2027373
0.0542333 0.08399942
_ _
2.806253
4.1900494
_
10
3
sin 78.53t
=
_
0.9546209
0.1997474
_
10
3
sin 78.53t
Figure 2.2.7 gives a comparison of the amplitude value for the static and
dynamic coupling case.
It will be observed that
Dynamically coupled equation gives slightly higher amplitude than statically
coupled equations.
This is applicable for both translation and rotation.
-1.50E-03
-1.00E-03
-5.00E-04
0.00E+00
5.00E-04
1.00E-03
1.50E-03
1 20 39 58 77 96 115 134
Time steps
A
m
p
l
i
t
u
d
e

o
f

v
i
b
r
a
t
i
o
n
Translation Based
on Static coupling
Rotation based on
static coupling
Translation Based
on Dynamic
coupling
Rotation based on
dynamic coupling
Figure 2.2.7 Comparison of response values static versus dynamic coupling.
Next we compare these results with time history response where we treat the
complete mass, damping and stiffness matrix in uncoupled form and integrate
directly the equation
[M]{
X} +[C]{
X} +[K]{X} = {Psin
M
t}
8
Figure 2.2.8 shows a very interesting result pertaining to time history vis--vis
modal response technique based on dynamically coupled equation.
The time history technique used has been Wilson- method having a time step
of 0.0075 seconds and response has been calculated to 500 steps.
It will be observed that in modal response technique we have ignored the
transient response part and have only found out the response based on the steady
state part, while, the step by step integration considers both the transient and
the steady state responses.
In comparison to a steady state response of 1 mm the time history starts with
peak amplitude of approximately 5 mm in step 18 and slowly converges to a
value near to 1 mm at about 295th step and becomes steady after that
9
.
8 Refer Chapter 5 (Vol. 1) the topic of Time history/step by step integration.
9 Looking at the displacement value of 1 and 5 mmhardend professionals could well frown or smirk.
But herein the example is deliberately designed like this to give beginners and especially students
a quantitative feel which eludes many when working in micron level. Real life problems would
come surely latter in the chapter.
It will be observed that the values are quite closely matching at the steady
state position with step by step integration giving slightly higher values than
dynamically coupled modal response.
The initial response due to the transient part of the time history analysis is
significant (about 5 mm), as this decays down quickly after some time (here
about 2.0 seconds after the start) will really not have much effect on the over all
behavior of the foundation as such, but for pipes and nozzles rigidly connected
to the machine this initial high amplitude of 5 mm can have significant effect
and if proper care is not taken may induce severe reversal of stress and may even
induce failure.
-3.00E-03
-2.00E-03
-1.00E-03
0.00E+00
1.00E-03
2.00E-03
3.00E-03
4.00E-03
5.00E-03
6.00E-03
Time steps
A
m
p
l
i
t
u
d
e
Translation Modal response
Translation based on time history
1 52 103 154 205 256 307 358 409 460
Figure 2.2.8 Comparison of response time history versus modal response.
2.3 TRICK TO BY PASS DAMPING MAGNIFICATION
FACTOR, THE KEY TO THE PROBLEM. . .
A perfectionist may not like the methodology proposed regarding approximate esti-
mate of the damping ratio based on individual mode. At the same time he might argue
that for secondary equipment like mediumor small capacity pumps doing time history
analysis is too intense and not really called for.
Fair enough, for the argument is not without some sanctity so how do we tackle
this riddle?
The most logical solution to the above problem could be that if we can create a
condition where damping plays a negligible effect compared to un-damped situation
then we can surely ignore damping from our basic equation and arrive at a result
which is as good an answer with damping and the problem is solved.
So the next obvious query will be, what is this condition which will give an invariable
answer irrespective of damping taken or not?
Before we describe this condition it would be worthwhile to understand what the
magnification factor is.
Consider Figure 2.3.1. We had seen earlier that equation of motion for a body
having damped single degree of freedom is given by
z
=
P
0
K
z
sin
m
t
_
(1 r
2
)
2
+(2D
z
r)
2
(2.3.1)
where r =
m
/
n
(frequency ratio) and D
z
= damping ratio
In the above equation the denominator is known as the magnification factor.
i.e. M.F. =
1
_
(1 r
2
)
2
+(2D
z
r)
2
(2.3.2)
Based on the above expression we define magnification factor as a factor which
gives us a measure of how many times the dynamic amplitude gets magnified over the
equivalent static deflection of P
0
/K.
Nowthe above is a very interesting equation for if we plot the above equation for var-
ious values of damping against the frequency ratio we get curve as shown in Fig. 2.3.1.
In the below curve it will be observed that when the frequency ratio (
m
/
n
) is
about 3.0, all the lines irrespective of whatever is the damping converges nearly to
a single line.
Now what does this signify?
It implies that when the frequency ratio is more than 3.0 irrespective of the damping
ratio the magnication factor do not change and thus the damping ratio plays practi-
cally no part in the overall response of the system. The usual practice is that when the
frequency ratio is more than 3.5 the damping effect of the soil is neglected.
0
1
2
3
4
5
6
7
8
0
.
1
5
0
.
5
5
0
.
9
5
1
.
3
5
1
.
7
5
2
.
1
5
2
.
5
5
2
.
9
5
3
.
3
5
3
.
7
5
4
.
1
5
4
.
5
5
4
.
9
5
5
.
3
5
5
.
7
5
Frequency Ratio
M
a
g
n
i
f
i
c
a
t
i
o
n

f
a
c
t
o
r
Mag. Factor for
5% damping
Mag. Factor for
10% damping
Mag. Factor for
15% damping
Mag. Factor for
25% damping
Mag. Factor for
37.5% damping
Figure 2.3.1 Magnication factor for various damping ratio.
One can immediately draw conclusion from the above fact that when a block
foundation is resting on ground having soft to medium soil supporting machines hav-
ing high operating speed it would possibly be quite justied to neglect the damping
effect of soil.
However, for low tuned machine it has been observed that it is difcult to achieve
this frequency separation for block foundation and for such cases damping cannot be
ignored.
We now further explain the above with a suitable numerical example.
Example 2.3.1
For the block foundation as described in Example 2.2.1. Calculate the following
amplitude of vibration based on undamped equation of motion based on
Static Coupling
Dynamic coupling
Discuss the results based on the answers.
The unbalanced dynamic force is 200 kN having operating frequency of
750 r.p.m acting at a height 600 mm from the top of foundation.
The total height of the concrete block (L) is 1.8 m
All other design parameters pertaining to design remains same as
Example 2.2.1
Solution:
The operating frequency of the machine is 750 r.p.m. =
750
60
2 =
78.53 rad/sec
Solution based on static coupling
The equation of motion in matrix form is given by [M]{
X} +[C]{
X} +[K]{X} =
{P}; the force matrix is
_
P
0
M
0
_
=
_
200
180
_
sin 78.53t
For Static coupling
[] =
_
0.1188173 0.076746
0.0542333 0.0839942
_
T
{P}, we have
[]
T
{P} =
_
0.1188173 0.0542333
0.076746 0.0839942
_ _
200
180
_
sin 78.53t
=
_
33.52
0.2293
_
sin 78.53t
We had seen earlier that the two natural frequencies are
1
= 4.35 rad/sec;
2
= 12.58 rad/sec
Thus the frequency separation is r =
78.53
12.58
= 6.24 > 3.5
Thus we can say that as the frequency ratio is greater than 3.5 damping effect
of the soil can be neglected.
Thus, with orthogonal transformation and neglecting the damping, we have
[]
T
[M][]{
} +[]
T
[K][]{} = []
T
{P}
{
1
} +[
2
1
]{
1
} = p
0
sin
m
t; {
2
} +[
2
2
]{
2
} = m
0
sin
m
t
or, {
1
} +18.92{
1
} = 33.52sin 78.53t and
{
2
} +158.5{
2
} = 0.230sin 78.53t

1
=
33.53sin 78.53t
_
(18.92 78.53
2
)
2
= 5.453769810
3
sin 78.53t
2
=
0.230sin 78.53t
_
(158.5 78.53
2
)
2
= 3.827935410
5
sin 78.53t
_
x
_
=
_
0.1188173 0.076746
0.0542333 0.08399942
_ _
5.4537698
0.038273954
_
10
3
sin 78.53t
=
_
0.6509395
0.2925838
_
10
3
sin 78.53t
Solution based on dynamic coupling
[] =
_
0.0374669 0.2027373
0.0542333 0.0839942
_
T
{P} we have
[]
T
{P} =
_
0.0374669 0.0542333
0.2027373 0.0839942
_ _
200
180
_
sin 75.53t
=
_
17.255374
25.428504
_
sin 78.53t
As we have seen earlier that the frequency separation is greater than 3.5, hence
on orthogonal transformation without damping we have
[]
T
[M][]{
} +[]
T
[K][]{} = []
T
{P}
{
1
} +[
2
1
]{
1
} = p
0
sin
m
t; {
2
} +[
2
2
]{
2
} = m
0
sin
m
t
{
1
} +18.92{
1
} = 17.25537 sin 78.53t and
{
2
} +158.5{
2
} = 25.429054sin 78.53t

1
=
17.25537 sin 78.53t
_
(18.92 78.53
2
)
2
= 2.806645310
3
sin 78.53t
2
=
25.428504sin 78.53 t
_
(158.5 78.53
2
)
2
= 4.232116110
3
sin 78.53t
_
x
_
=
_
0.0374669 0.2027373
0.0542333 0.08399942
_ _
2.8066453
4.2321161
_
10
3
sin 78.53t
=
_
0.963164
0.2007444
_
10
3
sin 78.53t
On comparing the amplitudes for the damped and undamped case we have
the following results.
Damped amplitude Undamped amplitude Difference in (%)
Sl. No. Formulation basis X (10
3
) (10
3
) X (10
3
) (10
3
) X
1 Static coupling 0.650525 0.2924 0.6509 0.29258 0.06 0.06
2 Dynamic coupling 0.9546 0.1997 0.9631 0.20074 0.88 0.518
2.4 EFFECT OF EMBEDMENT ON FOUNDATION
The theories described above dominated the scenario of design of machine founda-
tion for quite sometime. But as the machines designed were progressively becoming
heavier and having higher and higher operating frequencies the foundations in turn
were becoming more and more massive in nature
10
, and it was realized that when
a foundation is constructed below ground level the surrounding soil in which it is
embedded plays a signicant role on the overall response of the foundation and needs
to be carefully evaluated too.
Number of theoretical formulations have been derived and field experiments (Gupta
1972, Erden & Stokoe 1975) have been conducted to study the embedment effect
of soil on the overall response of the foundations, though there exists disagreements
between the theories put forward however the general consensus about the embedment
of soil on the foundation both fromtheoretical and field observations are as follows:
The embedment effect increases the natural frequency of the foundation
It reduces overall amplitude of the foundation.
It is not difficult conceive from the above statements that:
Embedment effect increases the soil stiffness and
Also has an incremental effect on the damping of the soil.
The most popular theory which is in practice in design office is given
Tables 2.4.1 and 2 (Whitman 1972).
Here, h=depth of embedment of the foundation in the surrounding soil; =
Poissons ratio of the soil.
Table 2.4.1 Embedment coefcients for spring constants.
Equivalent
Sl. No. Direction Coefcient radius Remarks
1 Vertical
z
= 1 +0.6(1 )
h
r
z
r
z
=
_
LB
This is in vertical Z direction

2 Horizontal
x
= 1 +0.55(2 )
h
r
x
r
x
=
_
LB
This induce sliding in horizon-

tal x or y direction
3 Rocking
x
= 1 +1.2(1 )
h
r
x
r
x
=
4
_
LB
3
3
Y axis
+ 0.2(2 )
_
h
r
x
_
3
3.1 Rocking
y
= 1 +1.2(1 )
h
r
y
r
y
=
4
_
L
3
B
3
X axis
+ 0.2(2 )
_
h
r
y
_
3
4 Twisting None available This produces twisting about
vertical Z axis
10 In most of the cases as the plan area of the foundation is dependent on the equipment general arrangement
to increase the mass it was getting deeper and deeper.
Table 2.4.2 Embedment coefcients for soil damping ratio.
Sl. No. Direction Coefcient Equivalent radius Remarks
1 Vertical
z
=
1 +1.9(1 )
h
r
z
z
r
z
=
_
LB
z
is value as obtained
as coefcient for
soil spring constant
2 Horizontal
z
=
1 +1.9(2 )
h
r
x
x
r
x
=
_
LB
x
as coefcient for
3 Rocking
x
=
1 +0.7(1 )
h
r
x
+0.6(2 )
_
h
r
x
_
3
x
r
x
=
4
_
LB
3
3
x
as coefcient for
3.1 Rocking
y
=
1 +0.7(1 )
h
r
y
+0.6(2 )
_
h
r
y
_
3
y
r
y
=
4
_
L
3
B
3
y
as coefcient for
4 Twisting None available
It is suggested that if we multiply the spring constants available from Richart and
Lysmer formulation vide Tables 2.2.3 and 4 by the above factors we get the modified
spring constants valid for the embedded foundations.
Damping ratio as obtained from Richart and Lysmers model when multiplied by
the coefcients as furnished in Table 2.4.2 gives the damping ratio considering the
embedment effect of the soil.
2.4.1 Novak and Beredugos model
Novak &Beredugo (1972a) model for embedded foundation has already been worked
out in detail in the chapter 5 (Vol. 1) under soil dynamics and you may refer to
the same for further details. Both vertical and lateral mode coupled with rocking
(Novak & Beredugo 1972b) has been treated therein.
2.4.2 Wolfs model
Wolf (1985) has devised springs for dynamic foundation where he has con-
sidered additional soil mass vibrating with the foundation effect as shown in
Table 2.4.3.
2.5 FOUNDATION SUPPORTED ON PILES
In many cases due to poor soil condition machine foundations are loaded on piles and
obviously other than static loads they are also subjected to vibrations and dynamic
loading. Dynamic behaviour of piles is still to certain extent not very clearly understood
though theoretical formulas exist to predict their behaviour under time dependent
Table 2.4.3 Soil spring constants as per Wolf (1985).
Mode Spring stiffness
0

0
Vertical
4Gr
0
1
0.58 0.095
Horizontal
8Gr
0
1
0.85 0.27
Rocking
8Gr
3
3(1 )
0.3
1 +
3(1)m
8r
5
0.24
Torsion
16Gr
3
3
0.433
1 +
2m
r
5
_
m
r
5
_
0.045
In which, C =
r
Vs
k
0
and m =
_
r
Vs
_
2
k
0
, where r =equivalent radius and
shall be r
0
, r
, r
as the case may be; G = Dynamic shear modulus of the soil;

=mass density of the soil; v
s
= shear wave velocity of the soil; M=mass of
the soil participating in the vibration with the machine and the block foundation,
and C =damping of the soil.
loading; they have been co-related with field observations for only a few simplified
cases.
As such the decision of using piles below machine foundations should be taken
cautiously and not without some understanding of howit would behave under the load
induced from the machine. Though there are very few reports on the field observation
data on dynamic behaviour of piles under machine foundations it is however generally
accepted that under time dependent loads piles,
Have significant influence on the amplitude, especially near resonance,
Increases the natural frequency of the system,
Decrease the geometric damping of the soil foundation system.
Since in some cases particularly in lateral mode, the effect of piles could be adverse
we repeat that it should be used with caution.
For machine foundation on piles three mathematical models are usually in vogue.
1 Piles considered as frequency independent equivalent springs based on Novaks
(1974) formulation.
2 Piles considered as beam elements connected to soil springs based on Richarts
formula as mentioned earlier.
3 Considering the underlying soil and the pile as finite elements and executing a
detailed analysis based on appropriate boundary conditions.
We will discuss all the above methods nowin some detail but would like to emphasise
at this point is that each one of them has its pros and cons and are not self-sufficient.
As such which would be the most appropriate model for analysis varies from case to
case and one method of analysis may have to be complimented by another model.
While describing the model we start with the most exhaustive one and go the reverse
Figure 2.5.1 Machine foundation supported on piles.
way for we feel this will give you a better insight to the various problems that exists
with dynamic behaviour of piles.
Shown in Figure 2.5.1 is a machine foundation supported by piles.
2.5.1 Pile and soil modelled as f inite element
This would obviously the most exhaustive model one could perceive and is shown in
Figure 2.5.2.
A representative and conceptual 2D model of the pile and soil are shown in
Figure 2.5.2.
Actually, the most appropriate model would be in 3D, where the piles are modelled
as beam elements while the soil can be modelled as eight nodded brick element and a
comprehensive dynamic analysis of the whole system could be performed.
It can however be perceived that the analysis would time consuming and expensive
(both in terms of man-hours and input data generation) and is usually not warranted
except for large multi-shaft gas/steam turbines of high output resting on poor soil
where such analysis could become essential in order to ensure the performance of such
expensive machines under operating conditions.
Rigid body lumped as mass
Pile Cap modeled as beam element
Master Node
Slave Node
Figure 2.5.2 Finite element model of machine foundation with pile and soil.
One of the major disadvantages with this type of model is that the boundary of the
soil has to be extended to substantial distance away both at the sides and from the
pile tip in vertical direction enabling the model to predict correctly the response of the
system. If this boundary limit is inadequate from the pile tip then waves transmitted
to the soil due to the vibration of the machine will get reflected back and result in
spurious responses which could make the analysis completely wrong.
The question as to how far this boundary should extend, no rational basis has been
derived yet and is completely up to the engineers judgement
11
.
11 One thumb rule is to extend the boundary in vertical direction to 2.5 times the length of the pile.
Other than this there are certain practical problems encountered especially when
the piles are long (say 20/30 m), geotechnical data may not be available to the depth
to which an engineer might like to extend the boundary of the problem and as such
if comprehensive soil data to the desired level is not available it may be difficult to
model the system without the adequate data.
In spite of the above problems the model is not without its advantage and may be
summarised as follows:
It comprehensively caters to the 3D effect of the pile soil and the foundation
It can effectively model the soil if layered in nature where each of the layers has
different material property.
The group interaction effect of soil and pile is automatically catered for.
Piles having variable geometry (tapered piles) can also be modelled without any
problems.
If battered piles are provided to counter any lateral thrust can also be modelled
without any difficulty.
2.5.2 Piles modelled as beams supported on elastic springs
In this method the piles are modelled as beam elements connected to springs in hor-
izontal and vertical direction calculated out of the soil material property as shown
Figs. 2.5.3 and 4. Again for a pile cap supporting machines a 3D model will be
developed and dynamic analysis will be carried out.
Shown in Figure 2.5.5 is a 3D model of pile cap, pile with soil springs. In this case
the soil springs may be calculated based on Richarts formulation or by multiplying
the influence area of each node by the coefficient of uniform compression.
The piles are modelled as beam elements having 6 degrees of freedom at each node.
The pile cap is mathematically modelled as either beam or plate bending element
depending upon the overall aspect ratio of the cap and other design considerations.
Figure 2.5.3 Pile in soil. Figure 2.5.4 Mathematical model.
Figure 2.5.5 Mathematical model of pile cap, piles with soil springs.
It is obvious that with respect to the previous model one of the major advantages is
that it is a relatively less laborious model in terms of input generation and complexity
and many engineers prefer to use this in lieu of a detailed finite element 3D model as
shown previously.
However the above model suffers from one serious lacuna for which it should be
used with caution.
The model in Figure 2.5.5 does not take into cognisance the effect of the soil which
lies between the two piles and treats the soil as only discrete element based on springs.
This could significantly under rate or even over rate the dynamic response which
depends on the nature of the soil
It does not take in to cognisance the pile group interaction factor which has been
observed to have significant effect on the dynamic response on the systemspecially
when the pile spacing is between 2.5D to 3D, where D is the overall diameter of
the pile.
It is recommended that this model may be used when the centre to centre distance
between the piles are at least more than 5D.
2.5.3 Novaks (1974) model for equivalent spring
stiffness for piles
This is possibly the most popular method used in the design offices to evaluate springs
for piles subjected to dynamic loads and will be discussed in some detail.
Though not without limitations the major advantage with this method is that
It is simple to use.
The spring stiffness and damping values are frequency independent.
The group interaction effect of the piles can be to certain extent taken care of.
The spring and damping values thus obtained can be very easily implemented as
linear springs in commercially available finite element software.
Standard Chart and coefficients exists for piles that are quite easy to use.
Novaks method has found to be in excellent agreement with eld observations
specially when the pile group arrangement is not complicated.
2.5.4 Equivalent pile springs in vertical direction
For single end bearing piles undergoing vertical motion, the spring constants are given
by the expression
k
bz
=
_
E
p
A
p
r
0
_
f
18,1
(2.5.1)
where, k
b
z = equivalent spring constant for end bearing piles; E
p
= Youngs modulus
of pile material; A
p
= cross sectional area of the pile; r
0
= equivalent radius of the
pile, and, f
18,1
= a factor which depends on pile material (concrete, steel, timber etc.),
ratio of embedded length l to radius (r
0
) and V
s
/V
c
(shear wave velocity of the soil
above the tip to compression wave velocity in pile).
The damping value in vertical direction is given by
c
bz
=
_
E
p
A
p
v
s
_
f
18,2
(2.5.2)
where, c
bz
= damping value of the end bearing piles; v
s
= shear wave velocity of
the pile through which the soil is driven (
_
G
s
g/
s
); f
18,2
= is a factor as given in
Table 2.5.1.
Table 2.5.1 Values of factor f -as per Novak (1974) for stiffness
and damping factor for single pile. For concrete piles
(
s
/
p
= 0.7) having /r
0
> 25.
Slenderness ratio Stiffness and damping function f for vertical bearing pile
20 f
18,1
= 3.75(V
s
/V
c
)
2
0.05(V
s
/V
c
) + 0.0501
f
18,2
= 15.345(V
s
/V
c
)
2.0928
50 f
18,1
= 6.25(V
s
/V
c
)
2
+ 0.05(V
s
/V
c
) + 0.0199
f
18,2
= 10(V
s
/V
c
)
2
+ 1.5(V
s
/V
c
) 0.012
100 f
18,1
= 3.75(V
s
/V
c
)
2
+ 0.45(V
s
/V
c
) + 0.0061
f
18,2
= 1.4(V
s
/V
c
) 0.0083
The values of f
18,1
and f
18,2
are meant for end bearing piles. However it has been
observed that for friction piles having l/r
0
greater than 60 or V
s
/V
c
greater than 0.03
these values are in small error pertaining to timber and concrete piles.
For steel piles Novak has given a value of f
18,1
= 0.030 and f
18,2
= 0.045 where
V
s
/V
c
= 0.033 and l/r
0
greater than 80.
For relatively short friction piles the following expression has been suggested by
Novak for calculation of the stiffness and damping
k
z1
=
_
E
p
A
p
r
0
_
f
18,1
and c
z1
=
_
E
p
A
p
v
s
_
f
18,2
(2.5.3)
Table 2.5.2 Vertical stiffness coefcients for oating piles as per Novak (1983).
f

18,1
L/R E
p
/G = 10,000 E
p
/G =2500 E
p
/G =1000 E
p
/G =500 E
p
/G =250
10.8696 0.0021 0.0052 0.0104 0.0187 0.0332
21.7391 0.0031 0.0083 0.0166 0.0301 0.0509
32.6087 0.0042 0.0104 0.0218 0.0364 0.0571
43.4783 0.0042 0.0125 0.0260 0.0405 0.0582
46.7391 0.0052 0.0135 0.0270 0.0416 0.0582
54.3478 0.0052 0.0145 0.0281 0.0416 0.0582
65.2174 0.0062 0.0166 0.0291 0.0416 0.0582
76.0870 0.0062 0.0177 0.0301 0.0416 0.0582
86.9565 0.0073 0.0187 0.0301 0.0416 0.0582
100.0000 0.0083 0.0197 0.0301 0.0416 0.0582
Table 2.5.3 Vertical damping coefcients for oating piles as per Novak (1983).
f

18,2
L/R E
p
/G =10,000 E
p
/G =2500 E
p
/G =1000 E
p
/G =500 E
p
/G =250
10.8696 0.0032 0.0126 0.0295 0.0558 0.1032
16.3043 0.0053 0.0179 0.0421 0.0695 0.1137
21.7391 0.0074 0.0232 0.0495 0.0811 0.1126
27.1739 0.0084 0.0263 0.0537 0.0832 0.1095
32.6087 0.0105 0.0305 0.0568 0.0811 0.1053
38.0435 0.0116 0.0326 0.0589 0.0789 0.1021
43.4783 0.0137 0.0347 0.0579 0.0758 0.0989
48.9130 0.0147 0.0368 0.0568 0.0737 0.0979
54.3478 0.0147 0.0379 0.0558 0.0726 0.0979
59.7826 0.0168 0.0379 0.0537 0.0716 0.0979
65.2174 0.0168 0.0379 0.0526 0.0705 0.0979
70.6522 0.0179 0.0379 0.0516 0.0695 0.0979
76.0870 0.0189 0.0379 0.0516 0.0695 0.0979
81.5217 0.0189 0.0368 0.0505 0.0695 0.0979
86.9565 0.0200 0.0358 0.0505 0.0695 0.0979
92.3913 0.0211 0.0358 0.0505 0.0705 0.0989
100.0000 0.0211 0.0337 0.0495 0.0705 0.0989
where, f
18,1
and f
18,2
are stiffness and damping factors respectively as given
Tables 2.5.2 and 2.5.3 respectively.
It has been suggested that the values given in these charts are most appropriate
when used for the range of a
0
= 0.1 to 0.8, where a
0
=
2fr
0
v
s
here f = the operating
frequency of the machine. It has also been suggested (Steven 1978) that these values
are even valid for a
0
as low as 0.05 which means that reasonably good results may be
expected for even slender piles and low frequencies.
The damping ratio for the pile may be calculated from the expression
D
z1
= c
z1
/2
_
k
z1
m
c
(2.5.4)
where m
c
is the mass of the cap plus machinery or the portion of structure which
is vibrating in the same phase as the cap. Part of the mass of the pile may also
be included in the above equation but it has been generally found that this ratio
of the pile mass to the mass of the supported weight is very small and is usually
ignored.
2.5.5 The group effect on the vertical spring
and damping value of the piles
Piles in practise usually do not occur as a single pile and usually under a foundation
block there will be a number of piles to forma foundation. For instance under a normal
block foundation supporting some machinery the pile system could be something like
as shown in Figure 2.5.6.
We had shown in Section 2.5.4 as to how to calculate the spring stiffness for indi-
vidual single piles. When we try to find out the equivalent spring stiffness for the
pile group as shown below this does not constitute of the sum of the individual pile
Figure 2.5.6 Plan view of a block foundation supported over piles.
stiffness. It also depends upon the relative distance between the piles itself and the
slenderness ratio of the piles carrying the loads and is expressed as
k
g
z
=
N
1
k
z1
_
N
A
(2.5.5)
The equivalent damping for the pile group is given by
c
g
z
=
N
1
c
z1
_
N
A
(2.5.6)
where, N = number of piles in a group;
A
= displacement interaction factor (axial)
for a typical reference pile in the group relative to itself and to all other piles in the
group assuming the reference pile and all other pile are loaded to same magnitude,
and the factor
A
can be evaluated from the expression (Randolph and Poulos 1982);
and also recommended by API 351R.
A
=
0.5ln(l
p
/s)
ln(l
p
d
A
)
for s l
p
(2.5.6a)
Here l
p
= Pile length, s = spacing of piles, d = diameter of pile
A
= G
av
/G
b
;
G
av
= Average Shear modulus along pile depth and G
b
= Shear modulus at pile base.
Alternatively the value can also be deduced fromPouloss interaction curve for static
interaction (Poulos and Davis 1980).
2.5.6 Effect of pile cap on the spring
and damping stiffness
If the pile cap is not in contact with the ground the above equations can be directly
used in for the analysis. Pile caps embedded usually have a favourable effect on the
response of the group and should be adapted wherever possible.
It would be realistic to assume that the embedment effect generates only the side
friction between the cap and the soil and that to only when dense granular backfill is
used. For the soil below the pile cap which is likely to be of inferior quality can settle
away from the cap for non-cohesive soil, similarly for cohesive soil this can shrink
away from the sides of the pile cap and can become ineffective.
Table 2.5.4 Values of

S
1
and

S
2
for various Poissons ratio.
Poissons ratio

S
1

S
2
0.0 2.7 6.7
0.25 2.7 6.7
0.40 2.7 6.7
The expressions for calculating the stiffness and the damping constant for an
embedded cap of embedd depth h is given by
k
f
z
= G
s
h
S
1
and c
f
z
= hr
0
_
_
G
s
s
/g
_
S
2
(2.5.7)
Equation (2.5.7) should be added to the pile stiffness and damping of the pile group
as presented in the previous section to arrive at the complete spring and damping
constant of a pile group in vertical direction. Values of

S
1
and

S
2
for various Poissons
ratio are given in Table 2.5.4.
in the horizontal direction
For vibration in horizontal direction the expression for stiffness and damping is as
shown below:
k
x1
=
_
E
p
I
p
r
3
0
_
f
11,1
for /r
0
25, and c
x1
=
_
E
p
I
p
r
2
0
v
s
_
f
11,2
for /r
0
25.
(2.5.8)
Here I
p
is the moment of inertia of the pile cross section about the centroidal axis
perpendicular to the direction of the motion. Here x direction depicts the horizontal
motion and f
11,1
and f
11,2
are factors for fixed headed piles. The values of f
11,1
and
f
11,2
are furnished in Table 2.5.5.
The group effect is expressed as
k
g
x
=
N
1
k
x1
N
1

L
and c
g
x
=
N
1
c
x1
N
1

L
(2.5.9)
Table 2.5.5 Values of factor f -as per Novak (1974) stiffness and damping factors for horizontal and
rocking mode.
Poissons ratio Function f
0.25 f
11,1
= 7.25(V
s
/V
c
)
2
+ 0.38(V
s
/V
c
) 0.0013
f
11,2
= 17(V
s
/V
c
)
2
+ 0.915(V
s
/V
c
) 0.0032
f
7,1
= 55(V
s
/V
c
)
2
+ 9.3(V
s
/V
c
) + 0.1075
f
7,2
= 38.75(V
s
/V
c
)
2
+ 6.55(V
s
/V
c
) + 0.0734
f
9,1
= 1.81(V
s
/V
c
)
f
9,2
= 0.375(V
s
/V
c
)
2
2.67(V
s
/V
c
) + 0.0005
0.4 f
11,1
= 7.875(V
s
/V
c
)
2
+ 0.43(V
s
/V
c
) 0.0015
f
11,2
= 18.75(V
s
/V
c
)
2
+ 1.02(V
s
/V
c
) 0.0037
f
7,1
= 57.5(V
s
/V
c
)
2
+ 9.65(V
s
/V
c
) + 0.1113
f
7,2
= 41.25(V
s
/V
c
)
2
+ 6.85(V
s
/V
c
) + 0.0746
f
9,1
= 1.94(V
s
/V
c
)
f
9,2
= 0.75(V
s
/V
c
)
2
2.87(V
s
/V
c
) + 0.0006
Table 2.5.6 Values of Su
1
and Su
2
Poissons ratio

S
u1

S
u2
0.0 3.6 8.2
0.25 4.0 9.1
0.40 4.1 10.6
where,
L
= a displacement factor for lateral motion defined in similar way to
A
and
is given by
12
Lf
= 0.6
c
[E
p
G
c
]
1
7
_
r
0
s
_
(1 +cos
2
p
) (2.5.9a)
LH
= 0.4
c
[E
p
G
c
]
1
7
_
r
0
s
_
(1 +cos
2
p
);
H
=
2
LH
,
M
=
3
LH
(2.5.9b)
Here
c
= G
z
/G
av
where G
z
= Shear Modulus at depth l
c
/4
l
c
= 2r
0
[E
p
/G
c
]
2
7
and is known as the critical length of the pile where G
c
= Average
shear modulus over the critical length of the pile.
Lf
= The horizontal interaction factor for fixed headed piles (no head rotation).
LH
= The horizontal interaction factor due to horizontal force (rotation allowed).
H
= Interaction factor due to horizontal force for rotation.
M
= Interaction factor due to moment for rotation.
p
= Angle subtended by a pile in pile group with respect to the reference pile.
When the calculated interaction factor exceeds 1/3, its value needs to be replaced
by
= 1 2/
27, a correction made to avoid approaching infinity as s tends to

zero.
Alternatively Pouloss interaction curve for static load case under horizontal load
may also be used.
The stiffness and damping characteristics of the pile cap is expressed as
k
f
x
= G
s
h
S
u1
and c
f
x
= hr
0
S
u2
_
G
s
s
/g (2.5.10)
The factors

S
u1
and

S
u2
are as given in Table 2.5.6.
2.5.8 Equivalent pile springs and damping in rocking motion
The expression for spring stiffness and damping for simple pile has been expressed as
k
1
=
_
E
p
I
p
r
0
_
f
7,1
; c
1
=
_
E
p
I
p
v
s
_
f
7,2
(2.5.11)
12 Example 2.7.3 is a very good conceptual case study for the same.
Here I is the moment of inertia of the pile cross section about the axis of rotation
and f
7,1
and f
7,2
are factors for rotational direction for fixed head piles, as furnished
in Table 2.5.5.
2.5.9 Group effect for rotational motion
For a pile group the group stiffness, shown in Figure 2.5.7, is expressed as
k
g
=
N
1
[k
1
+k
z1
X
2
r
+k
x1
Z
2
c
2Z
c
k
x1
] +k
f
(2.5.12)
Here X
r
and Z
c
are shown in the Figure 2.5.7 and k
z1
and k
x1
are stiffness constant
of single piles as described earlier. In addition
k
x1
=
_
E
p
I
p
r
2
0
_
f
9,1
and
k
f
= G
s
r
0
h
S
1
+G
s
r
2
0
h
__
2
3
_
+
_
Zc
r
0
_
2
_
Zc
r
0
_
_
S
u1
(2.5.13)
where = h/r
0
, here h = embedment depth of pile cap, and S
1
is as given in
Table 2.5.7.
The damping matrix for the pile group is expressed by
c
g
=
N
1
[c
1
+c
z1
X
2
r
+c
x1
Z
2
c
2Z
c
c
x1
] +c
f
(2.5.14)
where c
z1
and c
x1
are damping constant of single piles as described earlier. In addition
c
x1
=
_
E
p
I
p
r
0
v
s
_
f
9,2
and
Table 2.5.7 Values of S
1
and S
2
Poissons ratio

S
1

S
2
0.0 2.5 1.8
0.25 2.5 1.8
0.40 2.5 1.8
Machine block
Soil line
X
r
Z
c
h
Figure 2.5.7 Geometric description of the pile group.
c
f
= r
4
0
_
G
s
s
/g
_
S
2
+
__
2
3
_
+
_
Z
c
r
0
_
2
_
Z
c
r
0
_
_
S
u2
_
(2.5.15)
We now explain the above theory further by a suitable numerical problem.
Example 2.5.1
Find the vertical, horizontal and rocking stiffness of the pile group based on
Novaks formulation as shown in Fig. 2.5.8 and with the following soil
properties:
Length of the pile = 45.0 m; Diameter of the pile = 950 mm; Grade of concrete
M20 having a dynamic modulus as 300 10
6
kN/m
2
. Consider Poissons ratio
of soil = 0.4.
Solution:
Since each of the layers has different velocity and thickness we take a weighted
average of the shear wave velocity of the three soil layers as follows
v
s
=
6010 +11020 +21515
45
= 134 m/sec
600
1400
10 m
v
s
= 60 m/sec
Layer#1
= 18 kN/m
3
v
s
= 110 m/sec
= 20 kN/m
3
20 m
Layer#2
30 m
Layer#3
2000
2000
1
4
7
2
5
8
3000 3000
500
(Typ)
2000
2000
3
6 9
Arrangement of the pile group in Plan
Geotechnical profile of soil in which the piles are located
v
s
= 215 m/sec
= 22 kN/m
3
Figure 2.5.8 Machine foundation supported on piles.
Average weight density of soil is = s =
1810+2020+2215
45
=
20.22 kN/m
3
Thus dynamic shear modulus of the soil is taken as
G
avg
=
20.22
9.81
(134)
2
= 37010 kN/m
2
; g, acceleration due to gravity is
taken as 9.81 m/sec
2
.
Equivalent Radius of pile cap =
_
LB
=
_
7 5
= 3.33 m
Vertical stiffness of piles
Cross-sectional area of each pile =
(0.95)
2
4
= 0.7088 m
2
Dynamic modulus of concrete = 30010
6
kN/m
2
Density of concrete (
c
) = 25/kN/m
3
v
c
=
_
E
p
g
p
=
_
30010
6
9.81
25
= 10849 m/sec;
v
s
v
c
=
134
10849
= 1.23510
2
0.01
L
r
0
=
45
0.475
= 94.7 > 25.
For L/r
0
= 95 and v
s
/v
c
= 0.01 we have from Novaks Chart
f
18,1
= 0.011 and f
18,2
= 0.005.
Thus for each pile we have,
K
zi
=
E
p
A
r
0
f
18,1
=
30010
6
0.7088
0.475
0.011
= 4.92410
6
kN/m
and C
zi
=
E
p
A
v
s
f
18,2

30010
6
0.7088
134
0.005
= 7934 kN-sec/m.
Calculation of group interaction factor for, (L/r
0
) = 95
Thus, for group effect
9
i=1
K
zg
=
94.92410
6
3.86
= 11480829 kN/m;
9
i=1
C
zg
=
97934
3.86
= 18498 kN-sec/m.
Calculation of group interaction factor (vertical)
Pile No. Spacing (S) L/S Remarks
1 0.0 0.0 1.0 Pile # 1 is the reference
pile
2 2.0 22.5 0.474
3 4.0 11.25 0.368 obtained vide
4 3.0 15.0 0.413 Eqn (7.5.6a)
5 3.6 12.5 0.384
6 5.0 9.0 0.334
7 6.0 7.5 0.307
8 6.32 7.12 0.299
9 7.21 6.24 0.278
Total sum 3.86
Effect of the pile cap on overall pile stiffness
Here the effect of layer#1 will be more dominant on the pile cap as such for
finding out the stiffness properties in context of the pile cap, we have
G
s
=
18
9.81
(60)
2
= 6605 kN/m
2
For embedded depth, h = 1.4 m and S
1
= 2.7, we have,
K
f
z
= G
s
hS
1
K
f
z
= 66051.42.7 = 24969 kN/m.
C
f
z
=

hr
0
_
G
s
s
/gS
2
= 1.43.33
_
660518
9.81
6.7 = 3438.6 kN/m.
Thus, total vertical stiffness = 11480829 + 24969 = 11505798 kN/m
Total damping for the pile and pile cap = 18498 + 3439 = 21937 kN-sec/m.
Calculation of horizontal stiffness
I =
(0.95)
4
64
= 0.04 m
4
; with l/r
0
= 95,
s
= 0.4 and v
s
/v
c
= 0.01, from
Novaks chart f
11,1
= 0.0036 and f
11,2
= 0.0084
K
x
=
E
p
I
r
3
0
f
11,1
=
30010
6
0.04
(0.475)
3
0.0036 = 403091 kN/m;
C
x
=
E
p
I
r
2
0
v
s
f
11,2
=
30010
6
0.04
(0.475)
2
134
0.0084 = 3334 kN-sec/m.
Calculation of interaction factor (lateral)
Pile number Spacing S/r
0
factor in degree
l
Remarks
1 0.0 0.0 0 0 Pile#1 is the reference pile
2 2.0 4.21 90 0.246
3 4.0 8.42 90 0.123
l
calculated
4 3.0 6.316 0 0.328 as per eqn (7.5.9a)
5 3.6 7.578 33.6 0.231
6 5.0 10.526 53.13 0.134
7 6.0 12.632 0 0.164
8 6.32 13.306 18.43 0.148
9 7.21 15.178 33.6 0.115
Total sum 1.489

=1.50
For the pile group we have,
9
i=1
Kx =
4030919
1.5
= 2418546 kN/m;
9
i=1
Cx =
33349
1.50
= 20004 kN-sec/m.
Effect of the pile cap on overall pile stiffness
Here the effect of layer#1 will be more dominant on the pile cap as such for
finding out the stiffness properties in context of the pile cap we have
G
s
=
18
9.81
(60)
2
= 6605 kN/m
2
For embedded depth, h = 1.4 m and S
u1
= 4.1 and S
u2
= 10.6,
K
f
x
= G
s
hS
u1
= 66051.44.1 = 37913 kN/m;
C
f
x
=

hr
0
_
G
s
s
/gS
u2
= 1.43.33
_
660518
9.81
10.6 = 5440 kN/m.
Thus, the total lateral stiffness = 2418546 +37913 = 2.45610
6
kN/m.
Total damping for the pile and pile cap in lateral direction = 20004 + 5440 =
25444 kN-sec/m.
Calculation of rocking stiffness and damping
The rocking stiffness of individual pile, with f
7,1
= 0.202, is given by
k
1
=
_
E
p
I
p
r
0
_
f
7,1
=
_
30010
6
0.04
0.475
_
0.202 = 5103158 kN/m
The coupled sliding and rocking stiffness with, f
9,1
k
x1
=
_
E
p
I
p
r
2
0
_
f
9,1
=
_
30010
6
0.04
(0.475)
2
_
(0.0194)
= 1031801 kN/m
The pile cap stiffness is given by
k
f
= G
s
r
0
h
S
1
+G
s
r
2
0
h
__
2
3
_
+
_
Zc
r
0
_
2
_
Zc
r
0
_
_
S
u1
Substituting the values, G
s
= 6605 kN/m
2
; h = 1.4 m; r
0
= 3.33 m;

S
1
=
2.5; =
h
r
0
= 0.42; Z
c
= 1.5, X
r
= 2.0 m and

S
u1
= 4.1, we have, k
f
=
84417 kN/m.
Thus the total stiffness of the pile group is given by
k
g
N
1
[k
1
+k
z1
X
2
r
+k
x1
Z
2
c
2Z
c
k
x1
] +k
f
k
g
= 9[5103158 +4.92410
6
4 +4030912.25
+21.5103801] +84417
k
g
= 2.34210
8
kN/m
Calculation of damping value in rocking mode
The damping value with, f
7,2
c
1
=
_
E
p
I
p
v
s
_
f
7,2
=
30010
6
0.04
134
0.139 = 12448 kN-sec/m.
The coupled sliding and rocking mode damping, with f
9,2
= 0.0280, is
given by
c
x1
=
_
E
p
I
p
r
0
v
s
_
f
9,2
=
30010
6
0.04
0.475134
(0.0280) = 5273 kN-sec/m
The damping contributed by pile cap is given by
c
f
= r
4
0
_
G
s
s
/g
_
S
2
+
__
2
3
_
+
_
Z
c
r
0
_
2
_
Z
c
r
0
_
_
S
u2
_
where, G
s
= 6605 kN/m
2
; h = 1.4 m; r
0
= 3.33 m;

S
2
= 1.8; = h/r
0
= 0.42;
Z
c
= 1.5 m;

S
u2
= 10.6;
s
= 18 kN/m
3
; g = 9.81 m/sec
2
,
Substituting the above values we have, c
f
= 14604 kN-sec/m. The damping
value of the pile group is given by
c
g
=
N
1
[c
1
+c
z1
X
2
r
+c
x1
Z
2
c
2Z
c
c
x1
] +c
f
c
g
= 9[12448 +79344 +33342.25 +21.55273] +14604

= 622145 kN-sec/m.
2.5.10 Model for dynamic response of pile
In previous section we had presented the dynamic stiffness of piles as proposed by
Novak. This is a very popular model for dynamic response of machine foundations in
the design offices. Based on the analytical solution of Baranov (1967), Novak (1974)
proposed the method for evaluating the vertical response of piles under dynamic load-
ing. Many researchers like, Wolf and Von Arx (1978), Waas (1981), Kaynia and
Kausel (1982), Banerjee and Sen (1987) have advanced solutions to this problem,
yet Novaks method remains the most popular due to its sheer simplicity in appli-
cation. Solution of pile and pile-group based on the method proposed by Banerjee
and Sen, which is based on Boundary Element Method gives quite accurate results
but it is computationally too exhaustive to find applications in a day-to-day design
office work.
Applying Finite Element Method, where the pile is modeled as beam elements
and the soil as Winkler springs, has yielded good results. But, they are found to
be valid only when piles are single or when the distance between piles is signi-
cant (5d, d being the diameter of the pile), when the pile-soil interaction can be
neglected.
Novaks solution is mostly based on charts, and it furnishes stiffness and damping
of a pile and the solution is addressed to the fundamental degree of freedom.
In spite of its immense popularity, the model do have a few limitations as
summarized below:
a The solution does not take into account inertial effect of the pile;
b Extrapolation is required when design data are out of range of the chart;
c Charts are available only for RCC or timber piles, whether these charts are
applicable to cases of steel piles
13
, there is no clear-cut guideline;
d The charts do not address to the case where a pile is partially embedded;
e The formulation do not cater to dynamic axial load, moments or shears induced in
pile due to dynamic loads.
In certain cases when piles are supporting reciprocating compressors, it becomes
essential to check the design for higher frequencies of the foundations to ensure that
they are not matching with the second or third harmonics, when higher forces may be
induced by the machine at harmonic other than the first. In such cases one has no other
options but to resort to an elaborate and expensive three-dimensional Finite Element
based soil-pile foundation model to arrive at an answer to this problem and in number
of cases uncertainties present in such results are many. However, it was shown by
Novak that results obtained for higher harmonics are not significantly different for
the type of problem that is normally encountered in pile vibration studies.
We present nowa model (Chowdhury and Dasgupta 2006) that overcomes many of
the limitations cited above. The solution is simple (yet realistic) and does not require
elaborate software to be developed for the analysis. Asimple spreadsheet would suffice
for the problem considering the solution is basically analytic.
2.5.10.1 Proposed method
We had stated at the outset that most of the work relating to dynamic stiffness of pile is
based on Baranovs (1967) theory on the response of a soil embedded foundation. The
present formulation is based on Novak and Beredugos (1972) approach on embedded
foundation.
2.5.10.2 Vibration of friction piles
Let us consider a pile as shown in Figure 2.5.9. The pile is assumed to provide resistance
both through bearing as well as friction.
Let K
f
represents the frictional stiffness of the pile and the pile tip bearing stiffness
is taken as K
b
. The longitudinal vibration of such beams having only the frictional
stiffness may be represented by the expression
EA
2
u
z
2
+K
f
u = m(z)
2
u
t
2
(2.5.16)
13 This is an important issue for many real life projects specially in Arctic condition (like North Siberia)
or very arid region (like Sudan, Algeria) due to extreme low temperature or absence of water makes
concreting hazardous and almost all the structures and foundations are built on steel piles.
Z
L
K
f
dz
K
v
K
b
Figure 2.5.9 Pile embedded in ground up to a depth L and its mathematical model.
in which, E=Youngs modulus of pile; A=area of pile; K
f
= dynamic frictional
stiffness of soil having dimension (F/L) and u(z, t) =dynamic amplitude of pile =
(z) q(t); and m(z) =mass of element dz.
One of the solutions of equation (2.5.16) is given by
q(t) = C
3
sin t +C
4
cos t (2.5.17)
With the denition of u and using Equation (2.5.17), Equation (2.5.16) may be
written as
EA
d
2
(z)
dz
2
+K
f
(z) = m(z)
2
(z) (2.5.18)
The above equation can further be simplified to
d
2
(z)
dz
2
+p
2
(z) = 0 (2.5.19)
where p
2
= (m
2
+K
f
).
If you observe Equation (2.5.19) carefully, you should realize that it suggests that
the presence of frictional stiffness K
f
does not affect the basic shape function of the
pile and would remain same for the case had the pile would not have been embed-
ded. However, the bearing stiffness K
b
connected at the end of pile would affect the
shape function depending on the appropriate boundary condition. For computing the
correct shape function of the system, one has to start with the model as shown in the
Figure 2.5.9. The general solution for Equation (2.5.19) is given by Humar (1990)
(z) = (C
1
cos pz +C
2
sin pz)(C
3
sin t +C
4
cos t) (2.5.20)
in which, C
1
, C
2
, C
3
and C
4
are the integration constants to be determined from the
appropriate boundary conditions.
The pile has at the free head, z = 0, EA
du
dz
= 0, which gives
EAp[C
1
sin pz +C
2
cos pz][C
3
sin t +C
4
cos t] = 0 C
2
= 0. (2.5.21)
and at the tip, z = L, EA
du
dz
= K
b
u(z)
z=L
, which gives
EAp[C
1
sin pL] = K
b
C
1
cos pL pLtan pL = K
b
L/(EA) (2.5.22)
in which K
b
= G
b
r
0
C
b
(2.5.23)
where, G
b
= dynamic shear modulus of the soil at pile tip; r
0
= radius of the pile;
C
b
= a frequency independent dimensionless constant as suggested by Novak and
Beredugo (1972) and is given in Table 2.5.8.
Combining Eqns. (7.5.22) and (7.5.23), one can have
pLtan pL =
G
b
C
b
L
Er
0
(2.5.24)
It will be observed that the right hand side of Equation (2.5.24) is a dimensionless
quantity.
If =
G
b
C
b
L
Er
0
=
_
G
b
E
_ _
C
b
_
; where =slenderness ratio (L/r
0
) of the pile, Equation
(2.5.24) can be represented as
pLtan pL = 0. (2.5.25)
Equation (2.5.25) is a transcendental equation in pL and can be solved numerically.
The values of pL for various values of for the first mode are shown in Table 2.5.9.
Writing, pL = , the arbitrary shape function of the problem is given by
(z) = cos
z
L
(2.5.26)
The potential energy d of an element of depth dz, shown in Figure 2.5.9, is given
by Shames and Dym (1995)
d =
EA
2
_
du
dz
_
2
+
K
v
2
u
2
(2.5.27)
where, E=Youngs modulus of pile; A=area of pile; K
v
= dynamic stiffness of soil
having dimension kN/m; w= displacement of pile in the z direction and may be written
as (z)q(t).
In Equation (2.5.27), K
v
consists of two parts, namely,
1 the bearing stiffness at pile tip, and
2 the friction stiffness along the shaft.
For a rigid circular embedded footing with embedment D
f
, the stiffness of the footing
may be expressed as per Novak & Beredugo as
K
v
= G
b
r
0
C
b
+GD
f
S
1
(2.5.28)
where, K
v
= foundation stiffness in the vertical direction; G=dynamic shear modulus
of the soil along the embedment length; G
b
= dynamic shear modulus of the soil at
the base; r
0
= radius of the foundation; C
b
and S
1
= dimensionless constant which
are basically frequency dependent.
However, it has been shown by Novak and Beredugo that considering C
b
and S
1
as frequency independent, no accuracy is lost for practical design problems and the
analysis becomes quite simplified for rigid circular embedded footing.
The frequency independent values of C
b
and S
1
are as given below in Table 2.5.8.
However, it should be remembered that an embedded circular footing is usually
considered to be rigid having infinite structural stiffness. On the contrary, a pile will
be far more flexible member whose structural stiffness will be much lower, thus the
above recommended value may be valid for certain pile geometry but may not be
valid for others. Comparing the stiffness data of piles obtained by Novak, Dobry and
Gazetas (1988) it is proposed that following value of S
1
be used for dynamic analysis
of piles in vertical direction.
S
1
=
9.553(1 +)
0.333
(2.5.29)
where = Poissons ratio of the soil; and = slenderness ratio of the pile.
This value of S
1
is derived based on similar technique used earlier by Lysmer and
Richart (1966) for deriving equivalent stiffness and damping of circular footings for
Lysmers analog from the solutions of a similar elasto-dynamic analysis as proposed
by Bycroft (1956). The value C
b
may be taken as suggested in Table 2.5.8 for it has no
bearing on the flexibility of pile and is a function of the base area only. Considering pile
base area is much smaller in comparison to a footing, its contribution is only marginal.
Table 2.5.8 Suggested frequency independent values suggested
by Novak and Beredugo (1972) for embedded
footing.
Poissons ratio C
b
S
1
0.0 3.9 2.7
0.25 5.2 2.7
0.5 7.5 2.7
Table 2.5.9 Roots of equation pL tan(pL)- = 0 for the rst or fundamental mode.
pL pL pL pL
0 0.02 0.7 0.75 2 1.077 15 1.473
0.1 0.322 0.8 0.791 2.5 1.142 20 1.496
0.2 0.433 0.9 0.828 3 1.192 25 1.51
0.3 0.522 1 0.86 3.5 1.232 30 1.52
0.4 0.593 1.25 0.931 4 1.265 35 1.527
0.5 0.653 1.5 0.988 5 1.314 40 1.533
0.6 0.705 1.75 1.036 10 1.429 50 1.54
Moreover in most of the practical cases its effect does not come into consideration (as
will be shown subsequently) for analysis of such piles are either considered as bearing
pile i.e. having infinite base stiffness or floating having no base effects. The first termin
Equation (2.5.28) represents the contribution of base resistance, while the second term,
the embedment effect of the foundation. Substituting Equation (2.5.28) in Equation
(2.5.27) for an element dz, d may be written as
d =
EA
2
_
du
dz
_
2
+
G
b
r
0
C
b
2
u
2
+
GS
1
dz
2
u
2
(2.5.30)
and the total potential energy over the total length of the pile (L) is given by
=
EA
2
L
_
0
_
du
dz
_
2
dz +
GS
1
2
L
_
0
u
2
dz +
G
b
r
0
C
b
2
u
2
(2.5.31)
Considering u(z, t) = (z)q(t), it can be proved (Hurty and Rubenstein 1967), that
K
ij
= EA
L
_
0
i
(z)
j
(z)dz +GS
1
L
_
0
i
(z)
j
(z)dz +G
b
r
0
C
b
i
(L)
j
(L) (2.5.32)
where the shape function of the problem is given by Equation (2.5.26).
The first derivative of the above with respect to z is given by
(z) =
L
_
sin
z
L
_
(2.5.33)
Using z/L = implying dz = Ld, and converting the shape function as furnished
in Equation (2.5.26) from local to generalized co-ordinates, the limits of the problem
get converted to 1 to zero.
Now, if one assumes
F() = cos (2.5.34)
F
i
() =
L
(sin ), and (2.5.35)
K
ij
=
EA
i
j
L
+1
_
0
F
i
()F
j
()d +GS
1
L
1
_
0
F
i
()F
j
()d +G
b
r
0
C
b
F
i
(1)F
j
(1).
(2.5.36)
For the fundamental mode i = j = 1 and Equation (2.5.36) reduces to
K
1
=
EA
2
L
+1
_
0
F
1
()
2
d +GS
1
L
1
_
0
F
1
()
2
d +G
b
r
0
C
b
F
1
(1)
2
(2.5.37)
Equation (2.5.37) can be rewritten as
K
1
=
EA
2
L
+1
_
0
(sin )
2
d +GS
1
L
1
_
0
(cos )
2
d +G
b
r
0
C
b
(cos )
2
(2.5.38)
Equation (2.5.38) on integration and after some simplification may be expressed as
K
1
= I
1
+I
2
+I
3
in which,
I
1
=
EA
2
L
_
1
2

sin 2
4
_
; I
2
= GS
1
L
_
1
2
+
sin 2
4
_
;
I
3
=
G
b
r
0
C
b
2
(1 +cos 2) (2.5.39)
Finally, K
1
can be written as
K
1
=
_
EA
2
2L
+
GS
1
L
2
+
G
b
r
0
C
b
2
_
+
_
GS
1
L
4
EA
4L
_
sin 2 +
G
b
r
0
C
b
2
cos 2
(2.5.40)
which may be further simplified to
K
1
= X
1
+X
2
sin 2 +X
3
cos 2 (2.5.41)
in which
X
1
=
_
EA
2
2L
+
GS
1
L
2
+
G
b
r
0
C
b
2
_
; X
2
=
_
GS
1
L
4
EA
4L
_
; X
3
=
G
b
r
0
C
b
2
(2.5.42)
Equation (2.5.42) gives the stiffness of the pile for the vertical mode, without any
limitation to slenderness ratio, E/G or the material type.
2.5.10.2.1 Mass of the pile
For a conservative system, if T is the kinetic energy of the system then at any time t,
the energy equations may be written as
T(t) =
1
2
H
_
0
m(z)
_
u(z, t)
t
_
2
dz (2.5.43)
Using, u(z, t) =
n
i=1
i
(z)q
i
(t) (2.5.44)
where u(z, t) = displacement function;
i
(z) = shape function; q
i
(t) = generalized
co-ordinate; m(z) =mass of element dz and substituting Equation (2.5.44) in Equation
(2.5.43), the energy equation may be written as
T (t) =
1
2
H
_
0
m(z)
_
_
n
j=1
i
(z) q
i
(t)
_
_
_
_
n
j=1
j
(z) q
j
(t)
_
_
dz
(2.5.45)
=
1
2
n
i=1
n
j=1
q
i
(t) q
j
(t)
_
_
H
_
0
m(z)
i
(z)
j
(z)dz
_
_
from which the mass matrix may be written as
m
ij
=
_
_
H
_
0
m(z)
i
(z)
j
(z)dz
_
_
for i, j = 1, 2, 3. . . n (2.5.46)
Similarly the stiffness value with transformation from local to natural co-ordinate,
the mass contribution of the pile may be obtained as
m
ij
=

p
AL
g
1
_
0
F
i
()F
j
()d (2.5.47)
were
p
=bulk density of pile material; A=area of pile cross section; L=pile length
embedded in soil, and g =acceleration due to gravity.
For the fundamental mode, i, j = 1, and one can have
m
1
=

p
AL
g
1
_
0
F
1
()
2
d (2.5.48)
The above on expansion results in
m
1
=

p
AL
g
1
_
0
(cos )
2
d (2.5.49)
Integration of Equation (2.5.49) gives
m
1
=

p
AL
2g
_
1 +
sin 2
2
_
(2.5.50)
which is the contributory mass of the pile for the fundamental mode in the vertical
direction.
2.5.10.2.2 Damping of the pile
The damping of the pile embedded in soil will constitute of two parts:
Material damping of the pile itself;
Radiation damping of the soil-pile system.
It is obvious that the material damping of the pile will be much lower than that of
the soil radiation damping. As the first step for calculating the soil damping one may
ignore the material damping of the pile for the time being. Material damping of soil
also is part of the system vibration. However, it has been found that for translational
vibration their effect is insignicant and may be neglected without any signicant effect.
Else, if one wishes, their values may be obtained from resonant column test from the
laboratory when damping may be obtained from ratio of successive amplitudes.
For a rigid footing embedded in soil for a depth D
f
, Novak and Beredugo has
proposed an expression
C
z
= r
0
_
b
G
b

C
b
+r
0
_
G

S
2
D
f
(2.5.51)
where, r
0
=radius of the foundation; G
b
= dynamic shear modulus at foundation
base; G=dynamic shear modulus of soil in which the foundation is embedded; D
f
=
depth of embedment;

C
b
and

S
2
= frequency independent constants as dened by
Novak and furnished in Table 2.5.10.
Table 2.5.10 Values of damping coefcients based on Novak and Beredugo (1972).
Poissons ratio

C
b

S
2
0.0 3.438a
0
+0.5742a
2
0
1.154a
3
0
+0.7433a
4
0
6.059a
0
+
0.7022a
0
a
0
+0.01616
0.25 5.06a
0
Do
0.5 7.414a
0
2.986a
2
0
+4.324a
3
0
1.782a
4
0
Do
where a
0
= r/v
s
in which, operating frequency of the system in rad/sec; r = radius of the pile; v
s
=shear wave
velocity of the soil.
With reference to Figure 2.5.1 for a pile element of length dz, embedded in the soil,
the above equation may expressed as
C
z
= r
0
_
b
G
b

C
b
+r
0
_
G

S
2
dz (2.5.52)
For systems having continuous function, the damping is usually expressed as
(Mario 1987):
C
z
= c(z)
_

i
(z)
j
(z)dz (2.5.53)
For the present case, Equation (2.5.53) can be expressed as
C
z
= r
0
_
G

S
2
L
_
0
i
(z)
j
(z)dz +r
0
_
b
G
b

C
b
i
(L)
j
(L) (2.5.54)
Considering (z) = cos
z
L
, for the fundamental mode, one can have
C
z
= r
0
_
G
S
2
L
_
0
cos
2
z
L
dz +r
0
_
b
G
b

C
b
cos
2
(2.5.55)
and hence
C
z
= r
0
_
G
S
2
L
1
_
0
cos
2
d +r
0
_
b
G
b

C
b
cos
2
(2.5.56)
Equation (2.5.56), on integration simplifies to
C
z
=
1
2
r
0
_
G

S
2
L +
1
2
r
0
_
b
G
b

C
b
+
r
0
S
2
L
4
sin 2 +
r
0
_
b
G
b

C
b
2
cos 2
(2.5.57)
Equation (2.5.57) expresses the soil damping for a single pile under vertical mode
of vibration. Here the Factor

S
2
and

C
b
are damping coefficients which are frequency
dependent. Fortunately the damping factor is required for calculation of the amplitude
when the eigen solution of the problem is already done vis-a-vis, the dimensionless
frequency number a
0
= r/v
s
term is known
14
. Polynomial t curve for

S
2
and

C
b
are
available in terms of a
0
which can be used to arrive at these parameters.
The damping constants are given in Table 2.5.10.
2.5.10.2.3 Consideration of material damping of pile
The structural stiffness contribution of the pile is given by Equation (2.5.40), while
that of the mass is given in Equation (2.5.50). Thus, if C
c
be the critical damping of the
pile then it can be expressed as C
c
= 2
_
Km
p
, where K (equals I
1
in Equation 2.5.39)
and m
p
are the stiffness and mass matrices of the pile. Depending on the material
used for pile like (RCC, steel etc.) a suitable damping ratio (D) can be assumed. The
damping (C) for the pile can expressed as
C
p
= DC
c
(2.5.58)
This, when added to the radiation damping, calculated earlier, gives the complete
damping quantity for the soil-pile system. It should be noted that for perfectly floating
piles structural contribution of pile vanishes, and the material damping of the pile
mentioned in the preceding need not be considered.
2.5.10.3 Vibration of bearing piles
The expressions derived so far give a general case when the load is transferred from
the pile to the soil, through both friction and bearing. There will be cases when the
pile is pre-dominantly bearing in load transfer. Using the above formulation when
lim (i.e. G
b
is very large compared to E), pL(tan pL) = , when /2,
the pile reduces to a perfectly bearing pile (i.e. xed at the base), however for practical
case when 50, it will not be too erroneous to assume /2, when the stiffness
of the pile reduces to
K
1
=
_
EA
2
8L
+
GS
1
L
2
_
and (2.5.59)
the damping may be expressed as
C
1
=
1
2
r
0
_
G
S
2
L (2.5.60)
and the mass is
m
1
=

p
AL
2g
. (2.5.61)
14 For calculation of damping it is considered =
n
for it is most critical at resonance.
2.5.10.4 Vibration of friction piles
When G
b
is very small the load is transferred mainly through pile friction. In the
above formulation when lim 0, pLtan pL = 0, when, 0, the pile becomes a
perfectly friction pile.
Thus, for 0, the stiffness of pile is given by
K
1
= (GS
1
L)/2 (2.5.62)
The damping matrix may be expressed as
C
1
=
1
2
r
0
_
GS
2
L +r
0
_
b
G
b

C
b
(2.5.63)
FromEquation (2.5.63) it should be noted that for a friction pile, the damping factor
increases, while the stiffness term in Equation (2.5.62) is less than the bearing case in
Equation (2.5.59). A similar observation has also been made by Novak (1974) in his
investigation. For very poor soil, the term G
b
in Equation (2.5.63) may be ignored.
However for cases when piles located in mediumto stiff homogenous clayey soil where
G = G
b
and yet the load is basically transferred through friction, the last term cannot
be ignored and would further enhance the radiation damping. The mass matrix shall
be same as stated in Equation (2.5.50).
2.5.10.5 Vertical vibration of partially embedded piles
In many instances, especially in the arctic condition, due to environmental reasons,
the steel piles are driven into the ground when they protrude about 2 to 3 m above the
ground over which the pile cap and vibrating equipments as placed (Figure 2.5.10). In
such cases Novaks (1976, 1983) chart cannot be used readily.
Rotating Machine
Pile Cap
Partially embedded piles
G.L.
L
L
1
Figure 2.5.10 Schematic diagram of partially embedded piles.
To evaluate the pile stiffness for such cases, the stiffness Equation (2.5.36) is to be
modified as
K
ij
= EA
L
_
0
i
(z)
j
(z)dz +GS
1
L
1
_
0
i
(z)
j
(z)dz +G
b
r
0
C
b
i
(L)
j
(L) (2.5.64)
where L
1
=partial depth of embedment of pile and L=total length of pile.
It is apparent from Equation (2.5.64) that the first and last term remains unchanged
and the second term based on depth of embedment gets modied, where the integra-
tion limits changes to (L
1
0) and the stiffness expression for the fundamental mode
reduces to
K
1
=
_
EA
2
2L
+
GS
1
L
1
2
+
G
b
r
0
C
b
2
_
+
_
GS
1
L
1
4
EA
4L
_
sin 2+
G
b
r
0
C
b
2
cos 2
(2.5.65)
The damping of the pile-soil system is given by
C
z
=
1
2
r
0
_
G
S
2
L
1
+
1
2
r
0
_
b
G
b

C
b
+
r
0
GS
2
L
1
4
sin 2 +
r
0
_
b
G
b

C
b
2
cos 2
(2.5.66)
The mass matrix remains the same as stated in Equation (2.5.50).
It should be noted that for this case while calculating the value of S
1
[Equation
(2.5.29)], the slenderness ratio is to be calculated based on the embedded length
of the pile.
2.5.10.5.1 Stiffness of the pile for soils with varying elastic property
In the previous section, the calculation of stiffness as well as the damping of soil was
based on the dynamic shear modulus of soil invariant with depth. While this could be
possible for clayey soils, there are many cases when the dynamic shear modulus of the
soil has been found to vary with depth. Generically this can be expressed as
G
= G(z/H)
(2.5.67)
where =a number varying from 02 [considered 0 when G is constant with depth,
assumed 1 for linear variation and 2 for parabolic distribution].
For instance for the soil with variable elastic property, Equation (2.5.67) may be
modified to
G
= G
(2.5.68)
where = z/H.
For the cases mentioned above, Novaks (1976) chart is possibly not valid. To
accommodate the above variation, the stiffness equation can be modified to
K
ij
=
EA
2
L
2
+1
_
0
F
i
()F
j
()d +GS
1
L
1
_
0
F
i
()F
j
()d +G
b
r
0
C
b
F
i
(L)F
j
(L)
(2.5.69)
2.5.10.5.2 Shear modulus having a linear variation
When the soil has linear distribution with depth, the stiffness Equation (2.5.69) may
be expressed as
K
1
=
EA
2
L
+1
_
0
(sin )
2
d +GS
1
L
1
_
0
(cos )
2
d +G
b
r
0
C
b
(cos )
2
(2.5.70)
which, on integration and subsequent simplification, gives rise to
K
1
=
1
2
_
EA
2
L
+
GS
1
L
4
_
1
1
2
_
+
G
b
r
0
C
b
2
_
+
1
2
_
GS
1
L
EA
L
_
sin 2
+
_
GS
1
L
4
2
+
G
b
r
0
C
b
2
_
cos 2 (2.5.71)
It may be noted that while for bearing pile = /2, for friction pile (unlike constant
G case), = 0 is an inadmissable function in this case. For the fundamental mode the
admissible function is = , which is the next higher mode. This is logical also for
the soil having stiffness increasing with depth and the pile will have a natural tendency
to wobble about its centre rather than moving en-mass.
The damping matrix in this case can be expressed as
C
z
= r
0
_
G
S
2
L
1
_
0
_
cos
2
d +r
0
_
b
G
b

C
b
cos
2
(2.5.72)
The integration of the first term in Equation (2.5.72) being cyclic in nature and can
be solved approximately by expanding the cosine function in series. On integration,
Equation (2.5.72) reduces to
C
z
= r
0
_
G
S
2
L
_
2
3
2
2
_
1
7

2
33
+
2
675
4
__
+r
0
_
b
G
b

C
b
cos
2
(2.5.73)
2.5.10.5.3 Shear modulus having a parabolic variation
When the soil modulus has a parabolic distribution with depth, the stiffness equation
may be expressed as
K
1
=
EA
2
L
+1
_
0
(sin )
2
d +GS
1
L
1
_
0
2
(cos )
2
d +G
b
r
0
C
b
(cos )
2
(2.5.74)
which on integration and subsequent simplification reduces to
K
1
=
_
EA
2
2L
+
GS
1
L
6
+
G
b
r
0
C
b
2
_
+
_
GS
1
L
2
_
1
2
+
1
3
3
_
EA
4L
_
sin 2
+
_
GS
1
L
+
G
b
r
0
C
b
2
_
cos 2 (2.5.75)
In this case, the first admissible function will be = for a friction pile and = /2
for a bearing pile.
The mass matrix for both the cases remains same as stated in Equation (2.5.50)
while the damping matrix can be obtained from the expression
C
z
= r
0
_
G
S
2
L
1
_
0
cos
2
d +r
0
_
b
G
b

C
b
cos
2
(2.5.76)
which on integration and simplification reduces to
C
z
=
_
r
0
_
b
G
b
C
b
2

r
0
GS
2
L
4
_
+
r
0
GS
2
L
8
sin 2
+
_
r
0
GS
2
L
4
+
r
0
_
b
G
b
C
b
2
_
cos 2. (2.5.77)
2.5.10.6 Group effect of pile
This has already been explained in detail in section 2.5.5 and may be used for the
present case also.
2.5.10.7 Effect of pile cap on pile stiffness
The pile cap has been found to affect the response of footing signicantly. Before
considering its effect within the proposed framework, it would be worthwhile to
recapitulate the practice in vogue.
D
f
Figure 2.5.11 Schematic diagram of pile and pile-cap with embedment.
The sketch given in Figure 2.5.11 can represent the pile group with pile cap.
In such case usually the embedment stiffness GS
f
D
f
is directly added to the pile
group stiffness and the system is considered as a lumped mass single degree freedom
system where
=
_
K
group
+G
f
S
f
D
f
M
(2.5.78)
where G
f
= dynamic shear modulus of the soil surrounding the pile cap; D
f
= depth
of embedment; S
f
= constant as suggested by Novak furnished in Table 2.5.4 (as S
1
);
M=mass of pile cap and machine placed on it.
It may be noted that contributing effect of the pile mass is ignored in the above which
could be significant for a pile group having large number of piles. To overcome the
above limitation and also to derive a better response we propose a two mass lumped
model has been proposed and shown in Figure 2.5.12.
The mass and stiffness matrices for the above model may be written as
[K] =
_
K
group
+G
f
S
f
D
f
G
f
S
f
D
f
G
f
S
f
D
f
G
f
S
f
D
f
_
(2.5.79)
M = Mass of pile cap + machine
K
2
= G
f
S
f
D
f
m
p
= Mass of piles in group
K
1
= K
group
Figure 2.5.12 Proposed two mass lumped model for the pile and pile cap.
and [M] =
_
_
n
p
AL
2g
_
1 +
sin 2
2
_
0
0 M
_
_
(2.5.80)
where n =number of piles in the pile group.
Since Equation (2.5.79) is statically coupled, the damping matrix is given by
C =
_
C
group
+C
f
C
f
C
f
C
f
_
(2.5.81)
where C
z
= r
0
_
G
S
2
D
f
and D
f
is the embedment depth of pile cap. (2.5.82)
Once the stiffness, mass and damping matrices are established, the natural frequency
of the system may be obtained from the standard expression
[K] [M]
2
= 0 (2.5.83)
leading to
1,2
=
(m
p
B +MA)
_
[(m
p
B +MA)
2
4m
p
MAB]
2m
p
M
(2.5.84)
in which m
p
=
n
p
AL
2g
_
1 +
sin 2
2
_
and A = K
group
+G
f
S
f
D
f
; B = G
f
S
f
D
f
.
where,
1
=

1
and
2
=

2
here
1
and
2
are the natural frequency of the
structure.
The damping matrix generated here is non-classical in nature and will not be
de-coupled on orthogonal transformation. However, since the degrees of freedom
considered here is two, the same can also be converted into an equivalent Rayleigh
damping (refer section 2.2.5 in this chapter where we have solved this) when the matrix
will decouple and standard modal solution can be applied.
2.5.10.8 Solutions for higher modes
This case is usually not considered in design office practices and neither any guidelines
presently exists for the same except treating the pile as a beam and the soil as Winkler
springs and solving the same based on finite element method. Using the proposed
methodology, the stiffness, damping and mass matrices can be computed for the higher
modes.
Referring to Equation (2.5.37), the stiffness matrix can be stated as
[Kij] =
EA
L
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
2
1
1
_
0
F
1
()
2
2
1
_
0
F
1
()F
2
()
1
3
1
_
0
F
1
()F
3
() .........
1
n
1
_
0
F
1
()F
n
()
1
1
_
0
F
2
()F
1
()
2
2
1
_
0
F
2
()
2
......... ........
2
n
1
_
0
F
2
()F
n
()
1
1
_
0
F
3
()F
1
()
3
2
1
_
0
F
3
()F
2
()
2
3
1
_
0
F
3
()
2
..........
3
n
1
_
0
F
3
()F
n
()
............. ........ ........ ....... ...........
1
1
_
0
F
n
()F
1
()
2
n
1
_
0
F
n
()
2
_
_
d +
GS
1
L
2
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
1
_
0
F
1
()
2
1
_
0
F
1
()F
2
()
1
_
0
F
1
()F
3
() .........
1
_
0
F
1
()F
n
()
1
_
0
F
2
()F
1
()
1
_
0
F
2
()
2
......... ........
1
_
0
F
2
()F
n
()
1
_
0
F
3
()F
1
()
1
_
0
F
3
()F
2
()
1
_
0
F
3
()
2
..........
1
_
0
F
3
()F
n
()
.............. ........ ........ ....... .............
1
_
0
F
n
()F
1
()
1
_
0
F
n
()
2
_
_
d +Gbr
0
C
b
F
1
(0)F
j
(0) etc. (2.5.85)
For first three modes this can simply be presented as
[K]
i=1,3 j=1,3
=
_
_
K
11
K
12
K
13
K
21
K
22
K
23
K
31
K
32
K
33
_
_
(2.5.86)
where for i = j
K
ii
=
_
EA
2
i
2L
+
GS
1
L
2
+G
b
r
0
C
b
_
+
_
GS
1
L
4
i
EA
i
4L
_
sin 2
i
+
G
b
r
0
C
b
2
cos 2
i
(2.5.87)
For i = j we have
K
ij
=
_
EA
i
j
2L

GS
1
L
2
_
sin(
i

j
)
i

j
+
_
EA
i
j
2L

GS
1
L
2
_
sin(
i
+
j
)
i
+
j
+G
b
r
0
C
b
cos
i
j
(2.5.88)
It should be noted at this point that there are no suggestive values available for S
1
and C
b
for higher modes either by Novak or any other research. However, it may
be reasonably stated that for higher modes the dimensionless frequency a
0
would
be 1.0 (or near 1.0 at worse) when the curve for S
1
becomes almost constant
(Novak 1974) and the values furnished in Table 2.5.6 may be used without much
error.
The value of for the fundamental mode is already furnished in Table 2.5.9 for the
next two modes the values of are furnished in Table 2.5.11 and Table 2.5.12.
Table 2.5.11 Roots of equation pL tan(pL) = 0 for second mode.
pL pL pL pL
0 3.141 0.7 3.348 2 3.644 10 4.425
0.1 3.173 0.8 3.374 2.25 3.689 20 4.491
0.2 3.204 0.9 3.4 2.5 3.732 25 4.533
0.3 3.234 1 3.426 3.0 3.809 30 4.561
0.4 3.264 1.25 3.486 3.5 3.876 35 4.582
0.5 3.292 1.5 3.542 4 3.935 40 4.598
0.6 3.320 1.75 3.595 5 4.034 50

=3/2
Table 2.5.12 Roots of equation pL tan(pL) = 0 for third mode.
pL pL pL pL
0 6.28 0.7 6.392 2 6.578 15 7.316
0.1 6.299 0.8 6.407 2.25 6.611 20 7.495
0.2 6.315 0.9 6.422 2.5 6.643 25 7.56
0.3 6.331 1 6.437 3.0 6.704 30 7.606
0.4 6.346 1.25 6.474 3.5 6.761 35 7.639
0.5 6.362 1.5 6.510 4 6.814 40 7.665
0.6 6.377 1.75 6.544 5 6.910 50

=5/2
Mass matrix is similarly given by
For i = j
m
ii
=

p
AL
2g
_
1 +
sin 2
i
2
i
_
+Mcos
2
i
(2.5.89)
where M=Mass of machine plus pile cap
For i = j
m
ij
=

p
AL
2g
_
sin(
i
+
j
)
i
+
j
sin(
i

j
)
i

j
_
+
M
2
_
cos(
i
+
j
) cos(
i

j
)
_
(2.5.90)
The damping matrix can be obtained as
For i = j
C
ii
=
1
2
r
0
_
G
S
2
L +
1
2
r
0
_
b
G
b

C
b
+
r
0
GS
2
L
4
sin 2
i
+
r
0
_
b
G
b

C
b
2
cos 2
i
(2.5.91)
For i = j
C
ij
= r
0
_
G
S
2
L
_
sin(
i
+
j
)
i
+
j
sin(
i

j
)
i

j
_
(2.5.92)
+
r
2
0
_
G
b

C
b
2
_
cos(
i
+
j
) cos(
i

j
)
_
It is apparent that the dynamic analyses of piles with pile cap are standard and
the validity of the same would depend on how correctly the pile stiffness values have
been obtained. For this, it would worthwhile to evaluate how the present formulation
matches with other established methods. To this end, the pile stiffness as obtained by
Equations (2.5.59) and (2.5.62) has been compared with Novaks chart (1974) and
equation based on rigorous analysis as proposed by Dobry and Gazetas (1988).
It should be noted that their expression is valid for floating piles of length say, L and
embedded in an elastic half space of length 2L. The results have been compared for a
single pile of various slenderness ratio varying from20 to 100 and E
p
/G
s
value of soil
varying from 250 to 10,000 for an RCC pile of diameter of 600 mm and having E
p
=
30 GPa. Poissons ratio value for soil considered is 0.4. Here E
p
=Youngs modulus
of pile material; G
s
=dynamic shear modulus of soil. The results for K
pile(bearing)
and
K
pile(friction)
are shown in Figure 2.5.13 through 20 for various slenderness ratios.
Finally, the natural frequency of a real life centrifugal compressor foundation
supported on 9 RCC piles, 45 meter long having diameter of 950 mm, have been
compared. The piles are spaced at 3.0 m c/c. The size of pile cap is 7 m5m2.0 m,
embedded to depth of 1.4 meter. The weight of the generator supported on it weighs
400 kN. The frequencies are again compared for a range of E
p
/G
s
varying from 250
to 10000.
Pile stiffness L/r=20
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
250 500 1000 2500 10000
E
p
/G
s
P
i
l
e

s
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(bearing)
Novak(bearing)
Gazetas
Figure 2.5.13 Comparison of bearing pile stiffness for slenderness ratio =20.
Stiffness of pile for L/R=20
E
p
/G
s
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
250 500 1000 2500 10000
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(friction)
Novak(friction)
Gazetas
Figure 2.5.14 Comparison of friction pile stiffness for slenderness ratio =20.
Vertical stiffness for L/r=40
E
p
/G
s
0.0000E+00
1.0000E+06
2.0000E+06
3.0000E+06
4.0000E+06
5.0000E+06
2
5
0
5
0
0
1
0
0
0
5
0
0
0
1
0
0
0
0
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(bearing)
Novak(bearing)
Gazetas
The results based on K
p(bearing)
and K
p(friction)
has been compared to Dobry and
Gazetas results and presented in Table 2.5.13. The results have not been compared
with Novak in this case for the charts are too crude especially in the range when
the ratio of E
p
/G
s
= 250010000 and significant variation can occur based on eye
estimate of stiffness function. Results have been found to be excellently matching
particularly for friction piles.
Stiffness of pile for L/r=40
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
250 500 1000 5000 10000
E
p
/G
s
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(friction)
Novak(friction)
Gazetas
0.0000E+00
1.0000E+06
2.0000E+06
3.0000E+06
4.0000E+06
5.0000E+06
6.0000E+06
250 500 1000 5000 10000
E
p
/G
s
P
i
l
e

s
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(bearing)
Novak(bearing)
Gazetas
Pile stiffness for L/r=80
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
250 500 1000 5000 10000
E
p
/G
s
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(friction)
Novak(friction)
Gazetas
As stated earlier, the results fromEquation (2.5.41) (with appropriate boundary con-
dition for bearing and friction) have been compared with Novaks chart and Dobry and
Gazetas expression. The results have been studied against both the bearing and friction
pile coefcients as suggested by Novak and El-Sharnouby (1983). It will be observed
in Figures 2.5.13 through 2.5.18 that the frictional stiffness values obtained are very
0.0000E+00
1.0000E+06
2.0000E+06
3.0000E+06
4.0000E+06
5.0000E+06
6.0000E+06
250 500 1000 5000 10000
E
p
/G
s
P
i
l
e

s
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(bearing)
Novak(bearing)
Gazetas
Pile stiffness for L/r=100
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
250 500 1000 5000 10000
E
p
/G
s
P
i
l
e

s
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kpile(friction)
Novak(friction)
Gazetas
Table 2.5.13 Variation of vertical frequency for compressor foundation.
Freq (rad/sec) Freq (rad/sec) As per Dobry and
Sl. No. E
p
/G
s
for K
pile
(bearing) for K
pile
(friction) Gazetas (rad/sec)
1 250 196 195 197
2 500 139 138 139
3 1000 99 98 98
4 2500 64 62 62
5 5000 47 44 44
6 7500 39 36 36
7 10000 35 31 31
close to Dobry and Gazetas results in all the cases for various L/r and E/G
s
values.
For the bearing piles, the values obtained are slightly higher than Dobry and Gazetas
values but matching very closely to Novaks data from E
p
/G
s
= 500 onwards. This
is expected. It was pointed out by Novak and others that bearing stiffness for a pile is
slightly more than that of friction stiffness.
At L/r = 20 the bearing values obtained are higher than that of Dobry and Gazetas
(which is logical considering his case is that of a floating pile) as well as from Novak
but the difference reduces considerably from E
p
/G
s
= 1000 onwards, and this is
the range in which piles are commonly used in practice. The values, where E
p
/G
s
is 1000 are actually far too stiff for any piles to be bored or driven. Moreover,
a pile with L/r = 20 is actually a fictitious values. For instance a standard pile of
length 30 meter, the diameter becomes 3.0 m, which is actually a cassion and not a
pile. It is possibly in such cases, the axial stiffness is far too high and this shows a
significant higher stiffness in bearing compared to friction piles for such an unrealistic
L/r ratio. For real life problems, the values of L/r is around 50100 and E
p
/G
s
>
1000. It will be observed that the values obtained by the proposed method are quite
close to the reported results useful for practical ranges of application. As for the
frequencies obtained for various E
p
/G
s
values the results in Table 2.5.10 are extremely
encouraging.
2.5.10.9 What is the major advantage of this model?
The major advantage with the proposed method is that instead of solving the differ-
ential equation (especially when the boundary condition gets complicated with cases
like partial embedment or variable soil) the stiffness, damping and mass matrices
are directly derived from energy principles and the subsequent derivation gets quite
simplied.
Finally, the formulation have been derived for a general case when pile can act
both as bearing and friction pile for which no direct solutions are available-and
this could be the reality in many cases when the pile is neither in full bearing or
full oating. Comparing the results it can be well inferred that the method can
be used for practical design office work without the limitations as stated at the
outset.
2.5.10.10 Design steps
Based on the derivations presented, the design steps may be summarized as follows:
Determine the soil properties like G, G
b
, G
f
and (Poissons ratio of the soil);
Determine the pile properties like Length of pile L and diameter of pile (2r
0
) and
also the Youngs Modulus E of the pile material;
Determine the pile cap property like its mass and depth of embedment D
f
;
Determine the weight of machine supported on the pile cap;
Obtain Novaks stiffness and damping coefficients C
b
, S
1
, C
b
, S
2
fromTable 2.5.9
and Table 2.5.10, Equation (2.5.29) etc.;
Establish the dimensionless parameter = (G
b
/E) (C
b
/) ;
For the given value determine the value of pL from Table 2.5.9;
If the pile is bearing (known priori) = /2;
Consider = pL;
Determine K
1
and m
p
from Equations (2.5.40) and (2.5.50) respectively;
Determine the embedment stiffness matrix from the Equation (2.5.79);
Form the mass, stiffness;
Perform eigen solution;
Find the value of the frequency and obtain the dimensionless frequency number a
0
;
Find the value of S
2
from Beredugos expression as given in Table 2.5.10;
Determine the damping of the system based on Equations (2.5.57), (2.5.60) &
(2.5.63);
Perform Modal or time history analysis to obtain the amplitude of vibration.
2.5.11 Dynamic analysis of laterally loaded piles
2.5.11.1 Piles under dynamic lateral loading
Having presented the mathematical model of vibration in vertical direction, we now
present a model of vibration of piles under lateral or horizontal load. This is an impor-
tant study for the pile supporting rotating machines under centrifugal and reciprocating
loads.
In majority of cases it has been found that of all modes (like vertical, rocking,
yawning, twisting etc.); lateral vibration (coupled with rocking) is most critical and
often governs the design. Thus a study of such motion is of paramount importance
for piles supporting important installations and also for facilities, which are valuable
to the community under earthquake threats. Recall Novaks method described earlier
for lateral pile you will realize on retrospection that the model has got the following
limitations
The method is coefficient based [function of the ratio of Youngs modulus of pile
(E
p
), and dynamic shear modulus of soil (G
s
), as such for intermediate values one
has to interpolate which may not be always very accurate.
The values are given for Poissons ratio of 0.25 and 0.40 only. Thus for any
value between 0.25 and 0.4, or beyond 0.4 another set of linear interpolation/
extrapolation is necessary.
Novak and El Sharnouby (1983) has given stiffness and damping coefficients for
soil having parabolic profile but in many cases the variation is linear and no
coefficients are available for this case.
The method does not have a solution for partially embedded piles, which is of great
practical importance for piles driven in arctic condition (especially in Northern
Siberia which constitute of a large number of Oil and Gas facilities).
The dynamic bending moment and shear force induced on pile cannot be evaluated.
Finally the formulation is valid for long piles (i.e. the failure takes place in the pile
body before soil yields) and do not cater to piles, which are short.
The simplified formulas given by Dobry and Gazetas (1988) is based on more rigor-
ous analysis, however it also does not address the issues of partial embedment, dynamic
bending moment and shear, or the issue- if the pile is short (i.e. L/r < 25) etc.
We nowpresent herein (Chowdhury and Dasgupta 2008) a mathematical model for
analysis of such piles under lateral load that overcomes many of the bottle necks cited
above.
Similar to the vertical vibration model presented earlier the present formulation is
based Novak and Beredugos (1972) formulation for a rigid cylinder embedded in
elastic half space. Shown in Figure 2.5.21 is a pile embedded in homogeneous elastic
medium and considered under plane strain condition. The pile is considered long and
slender, to start with. Under static conditions, the equation of equilibrium in the
M
P
X
Z
dz
Soil Stiffness = GSx
1
Figure 2.5.21 Conceptual model of pile under lateral load.
x-direction [similar to beams on elastic foundation] is given by Timoshenko (1956) as
E
p
I
p
d
4
x
dz
4
= k
s
x (2.5.93)
where, E
p
=Youngs modulus of the pile; I
p
=moment of inertia of the pile cross
section; k
s
=elastic stiffness of the soil and is expressed as GS
x1
; G=dynamic
shear modulus of the soil; S
x1
= Beredugos constant which are basically frequency
dependent.
However, it has been shown by Novak and Beredugo (1972) that considering this
term frequency independent, no accuracy is lost for practical design problems and
the analysis becomes quite simplified for rigid circular embedded footing. Elaboration
about this parameter, in terms of piles, will be made later.
The general solution of Equation (2.5.93) may be written as
x = e
pz
(C
0
cos pz +C
1
sin pz) +e
pz
(C
2
cos pz +C
3
sin pz) (2.5.94)
where p =
4
_
GS
x1
E
p
I
p
.
For long piles under load or moment at its head, it is reasonable to assume that
at significant distance from the pile head (where the load is applied), the curvature
vanishes. This condition can only be satisfied when C
2
and C
3
in Equation (2.5.94) is
considered insignificant. Hence, the deflection equation can be taken as
x = e
pz
(C
0
cos pz +C
1
sin pz) (2.5.95)
Considering the pile head undergoing specified deflection and rotation as well as
its head is fixed to the pile cap (same boundary condition as considered by Novak
(1974)), one can have [Figure 2.5.21],
At z = 0, let x = x
0
C
0
= x
0
, which gives
x = e
pz
(x
0
cos pz +C
1
sin pz) (2.5.96)
Again, at z = 0,
dx
dz
=
0
one can have
C
1
= x
0
+

0
p
(2.5.97)
Thus Equation (2.5.98) can now be represented as
x = e
pz
_
x
0
cos pz +
_
x
0
+

0
p
_
sin pz
_
(2.5.98)
For magnitude of rotation being small
0

= x
0
/L, x may be written as
x = e
pz
_
x
0
cos pz +
_
x
0
+
x
0
pL
_
sin pz
_
(2.5.99)
x
x
0
= e
pz
_
cos pz +
_
1 +
1
pL
_
sin pz
_
(2.5.100)
Now considering = pl and using Equation (2.5.100), for any arbitrary loading,
the generic shape function in dimensionless form can be represented as
(z) = e
z
L
_
cos
z
L
+
_
1 +
1
_
sin
z
L
_
(2.5.101)
in which
=
4
_
GS
x1
L
4
E
p
I
p
, L being the length of the pile. (2.5.102)
Equation (2.5.101) can be further reduced to
(z) = e
z
L
_
cos
z
L
+ sin
z
L
_
(2.5.103)
where = 1 +
1
(2.5.104)
The generic shape function of the pile for the fundamental mode as in Equation
(2.5.103) is shown in Figure 2.5.22 for E
p
/G = 5000.
The potential energy d of an element of depth dz, shown in Figure 2.5.21 is then
given (Shames and Dym 1995) by
d =
E
p
I
p
2
_
d
2
v
dz
2
_
2
+
K
h
2
v
2
(2.5.105)
where, E
p
=Youngs modulus of pile; I
p
=moment of inertia of pile; K
h
=lateral
dynamic stiffness of soil; v =displacement of the pile in the x direction and may be
written as [(z)q(t)].
For a rigid circular embedded footing of embedment D
f
may be expressed (Beredugo and Novak (1972)) as
K
h
= G
b
r
0
C
b
+GD
f
S
x
1
(2.5.106)
where, K
h
= foundation stiffness in horizontal direction; G=dynamic shear modu-
lus of the soil along foundation surface; G
b
=dynamic shear modulus of soil at the
foundation base; r
0
= radius of the foundation; C
b
and S
x1
= constants which are
basically frequency dependent.
Ignoring the first term in Equation (2.5.106), which represents the contribution
of base resistance, and substituting the same in Equation (2.5.105), for a cylindrical
element of depth dz, embedded in soil, the potential energy d may be expressed as
d =
E
p
I
p
2
_
d
2
v
dz
2
_
2
+
GS
x1
dz
2
v
2
. (2.5.107)
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
0
0
.
1
5
0
.
3
0
.
4
5
0
.
6
0
.
7
5
0
.
9
z/L
S
h
a
p
e

f
u
n
c
t
i
o
n

Figure 2.5.22 Generic shape function long pile in the horizontal mode for E
p
/G = 5000.
The total potential energy over the length of the pile (L) is then given by
=
E
p
I
p
2
L
_
0
_
d
2
v
dz
2
_
2
dz +
GSx
1
2
L
_
0
v
2
dz (2.5.108)
Considering v (z, t) = (z)q(t), it can be proved (Hurty and Rubenstein 1967) that
K
ij
= E
p
I
p
L
_
0
i
(z)
j
(z)dz +GS
x1
L
_
0
i
(z)
j
(z)dz (2.5.109)
For the fundamental mode, stiffness of the pile is then given by
K = E
p
I
p
L
_
0
i
(z)
2
dz +GSx
1
L
_
0
i
(z)
2
dz (2.5.110)
On double differentiation, Equation (2.5.103) reduces to
(z) =
2
2
L
2
e
z
L
_
sin
z
L
cos
z
L
_
and (2.5.111)
(z)
2
=
4
4
L
4
e
2z
L
_
X
2

Y
2
cos
2z
L
sin
2z
L
_
(2.5.112)
where, X = 1 +
2
; Y = 1
2
and is given in Equation (2.5.104).
Again from Equation (2.5.103)
(z)
2
= e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
(2.5.113)
Substituting Equations (2.5.112) and (2.5.113) in Equation (2.5.110), the stiffness
reduces to
K =
4E
p
I
p
4
L
4
L
_
0
e
2z
L
_
X
2

Y
2
cos
2z
L
sin
2z
L
_
dz
+GS
x1
L
_
0
e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
dz (2.5.114)
Equation (2.5.114) on integration by parts and on simplification may be expressed as
K =
4E
p
I
p
4
L
4
_
X
2
L
2
(1 e
2
)
Y
2
L
4
_
e
2
(sin 2 cos 2) +1
_
_
4E
p
I
p
4
L
4
_
L
4
(1 e
2
(sin 2 +cos 2))
_
+GS
x1
_
X
2
L
2
(1 e
2
) +
Y
2
L
4
(e
2
(sin 2 cos 2) +1)
_
+GS
x1
_
L
4
(1 e
2
(sin 2 +cos 2))
_
(2.5.115)
In Equation (2.5.115), e
2
(sin 2 + cos 2) and e
2
(sin 2 cos 2) may be
ignored as their values are exceedingly small (highest is of the order 10
3
and the
lowest is 10
30
for E
p
/G value varying from 250 to 10,000) and has practically no
effect on the stiffness value and this also considerably simplifies the expression.
Based on the above simplication, Equation (2.5.115) may be rewritten as
K =
4E
p
I
p
4
L
4
_
X
2
L
2
(1 e
2
)
Y
2
L
4
L
4
_
+GS
x1
_
X
2
L
2
(1 e
2
) +
Y
2
L
4
+
L
4
_
(2.5.116)
K =
E
p
I
p
3
L
3
_
X(1 e
2
)
Y
2

_
+
GS
x1
L
4
_
X(1 e
2
) +
Y
2
+
_
(2.5.117)
Taking E
p
I
p
3
/L
3
as common in Equation (2.5.117) and substituting the value of
from Equation (2.5.102a), Equation (2.5.117) reduces to
K =
E
p
I
p
3
L
3
_
5X
4
(1 e
2
)
3Y
8

3
4
_
which can be further simplified to
K =
E
p
I
p
L
3
_
5X
4
(1 e
2
)
3Y
8

3
4
( 1)
3
_
(2.5.118)
The accuracy of Equation (2.5.118) will be dependent on the correct selection of
S
x1
. For instance for rigid circular footing Beredugo and Novak (1972) has furnished
a frequency independent value of S
x1
=4.0 to 4.1 (depending on Poissons ratio). This
has been found to give adequate accuracy for practical engineering design. Comparing
the stiffness data with Novak (1974), Dobry and Gazetas (1988), it is proposed that the
following values of S
x1
as furnished in Tables 2.5.14 to 16 be used for the calculation
of dynamic response of the pile in the lateral direction.
For a particular pile having specific slenderness ratio and Poissons ratio of the
soil, the value of S
x1
can be selected from Tables 2.5.14, 2.5.15 and 2.5.16 and on
substitution of the same in Equation (2.5.102), Equation (2.5.118), gives the solution
of pile stiffness in the lateral direction.
Table 2.5.14 Suggested value of S
x1
for Poissons ratio of soil = 0.25.
L/r
0
Poissons (Slenderness
ratio ratio) S
x1
(250) S
x1
(500) S
x1
(1000) S
x1
(2500) S
x1
(5000) S
x1
(10000)
0.25 25 2.00 1.83 1.66 1.43 1.25 1.07
40 2.19 2.05 1.90 1.70 1.55 1.39
60 2.30 2.17 2.05 1.87 1.74 1.60
80 2.36 2.24 2.12 1.96 1.84 1.71
100 2.39 2.28 2.17 2.01 1.90 1.78
Note: The value in parenthesis after S
x1
indicates E
p
/G
s
value of the soil.
x1
for Poissons ratio of soil = 0.40.
L/r
0
ratio ratio) S
x1
(250) S
x1
(500) S
x1
(1000) S
x1
(2500) S
x1
(5000) S
x1
(10000)
0.40 25 2.27 2.08 1.89 1.63 1.43 1.23
40 2.48 2.32 2.16 1.94 1.76 1.59
60 2.60 2.46 2.31 2.12 1.97 1.82
80 2.66 2.53 2.40 2.22 2.08 1.94
100 2.70 2.57 2.45 2.28 2.15 2.02
x1
indicates E
p
/G
s
value of the soil.
x1
for Poissons ratio of soil =0.50.
L/r
0
ratio ratio) S
x1
(250) S
x1
(500) S
x1
(1000) S
x1
(2500) S
x1
(5000) S
x1
(10000)
0.50 25 2.45 2.25 2.05 1.77 1.55 1.34
40 2.67 2.50 2.33 2.09 1.91 1.72
60 2.80 2.65 2.50 2.29 2.13 1.96
80 2.87 2.72 2.58 2.39 2.24 2.10
100 2.91 2.77 2.63 2.45 2.32 2.18
x1
indicates E
p
/G
s
value of the soil.
2.5.11.1.1 Estimation of contribution of pile mass
The mass matrix of the pile may be expressed as (Meirovitch 1967)
M
x
= m(x)
_

i
(z)
j
(z)dz (2.5.119)
For the present case of the pile of length L, Equation (2.5.118), can be expressed as
M
x
=

p
A
p
g
L
_
0
(z)
2
dz (2.5.120)
where,
p
= unit weight of the pile material; A
p
= cross sectional area of the pile;
g =acceleration due to gravity.
or, M
x
=

p
A
p
g
L
_
0
e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
(2.5.121)
Equation (2.5.121) on integration and after simplification gives
M
x
=

p
A
p
L
4g
_
X(1 e
2
) +
Y
2
+
_
(2.5.122)
Equation (2.5.122) is the inertial contribution of the pile material for the fundamen-
tal mode. Incidentally, the inertial effect is usually ignored in design but could have
signicant effect if the number of piles is large in a pile group.
2.5.11.2 Radiation damping for pile under lateral load
f
, Beredugo and Novak (1972) have
proposed the expression
C
z
= r
0
_
b
G
b

C
b
+r
0
_
G

S
2
D
f
(2.5.123)
where, r
0
=radius of the foundation; G
b
=dynamic shear modulus at the foundation
base; G=dynamic shear modulus of the soil in which the foundation is embedded;
D
f
=depth of embedment;

C
b
and

S
2
=frequency independent constants as defined
by Novak and Beredugo (1972).
With reference to Figure 2.5.21 for a pile element of length dz embedded in the soil,
and ignoring the bearing effect, Equation (2.5.123) may be expressed as
c(x) = r
0
_
GS
x2
dz (2.5.124)
For systems having continuous response function, the damping may be expressed as
C
x
= c(x)
_

i
(z)
j
(z)dz (2.5.125)
For the pile of length L, Equation (2.5.125) may be expressed as
C
x
= r
0
_
GS
x2
L
_
0
(z)
2
dz (2.5.126)
or, C
x
= r
0
_
GS
x2
L
_
0
e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
(2.5.127)
On integration and after simplification Equation (2.5.127) reduces to
C
x
= r
0
_
GS
x2
L
_
X(1 e
2
) +
Y
2
+
4
_
(2.5.128)
Equation (2.5.128) expresses the soil damping for a single pile under horizontal
mode of vibration. The factor S
x2
is a frequency dependent damping coefficient. The
damping factor is required for calculating the amplitude only after the eigen solution of
the problemis already done vis-a-vis, the dimensionless frequency number a
0
= r
0
/v
s
term is known a priori. Polynomial fit curve for S
x2
are available in terms of a
0
which
can be used directly to obtain these parameters. S
x2
for different Poissons ratios are
Table 2.5.17 Values of S
x2
(Beredugo & Novak 1972).
Poissons ratio S
x2
0.0 7.334a
0
+
0.8652a
0
a
0
+0.00874
0.25 0.83a
0
+
41.59a
0
a
0
+3.90
0.5 0.96a
0
+
56.559a
0
a
0
+4.68
2.5.11.3 Material damping of pile
The structural stiffness contribution of the pile is given in the first part of Equation
(2.5.117), while that of the mass is given in Equation (2.5.122). Thus, if C
c
be the
critical damping of the pile then it can be expressed as C
c
= 2
_
Km
p
; K (the rst term
in Equation (2.5.117)) and m
p
being the stiffness and mass of the pile.
Depending on the material used for pile (like RCC, steel etc.) a suitable damping
ratio (D) can be assumed. The damping (C
p
) for the pile can then be expressed as
C
p
= DC
c
(2.5.129)
This, when added to the radiation damping, calculated through Equation (2.5.128)
gives the complete damping quantity for the soil-pile system.
2.5.11.4 Piles with other boundary conditions
Having established the stiffness, mass and damping of the pile in lateral direction
based on minimization of the potential energy of the system, the above method can be
extended for the piles with other boundary conditions for which there are no standard
solutions available.
2.5.11.5 Partially embedded piles
In Arctic and North Siberian condition, due to environmental reasons, the steel piles are
driven into the ground when they protrude about 23 m above the ground over which
the pile cap and vibrating equipments are placed. Piling configuration has already been
shown earlier while explaining the vertical vibration of pile. In such cases the existing
solutions cannot be used. However, a solution of the same is proposed hereunder. Let
L be the full length of the pile and the length of the embedment in soil be L
1
(refer
Figure 2.5.10).
For this case, one may write
e
=
4
_
GS
x1
L
4
1
E
p
I
p
(2.5.130)
Here subscript e represents embedment of the pile.
The shape function can thus be represented by
(z) = e
e
z
L
1
_
cos

e
z
L
1
+ sin

e
z
L
1
_
(2.5.131)
and
(z) =
2
2
e
L
2
1
e
e
z
L
1
_
sin

e
z
L
1
e
cos

e
z
L
1
_
(2.5.132)
and hence
(z)
2
=
4
4
e
L
4
1
e
2
e
z
L
1
_
X
e
2

Y
e
2
cos
2
e
z
L
1
e
sin
2
e
z
L
1
_
(2.5.133)
where, X
e
= 1 +
2
e
; Y
e
= 1
2
e
and
e
= 1 +
1
e
.
Now, considering the fact that the embedment of a pile does not have any effect on
the shape function of the system, the stiffness of the pile for the fundamental mode
may be written as
K = E
p
I
p
L
_
0
i
(z)
2
dz +GS
x1
L
1
_
0
i
(z)
2
dz (2.5.134)
Considering, = L
_
L
1
, Equation (2.5.134) may be rewritten as
K =
4E
p
I
p
4
e
L
4
1
L
1
_
0
e
2
e
z
L
1
_
X
e
2

Y
e
2
cos
2
e
z
L
1
e
sin
2
e
z
L
1
_
dz
+GS
x1
L
1
_
0
e
2
e
z
L
_
X
e
2
+
Y
e
2
cos
2
e
z
L
1
+
e
sin
2
e
z
L
1
_
dz (2.5.135)
Equation (2.5.135) on integration by parts and after simplification, may be
expressed
K=
E
p
I
p
3
e
L
3
1
_
X
e
_
1
4
+
_
+Y
e
_
1
8

2
_
+
e
_
1
4

_
X
e
e
2
e
_
4
e
2
e
(1)
+1
_
_
(2.5.136)
this can be further simplified to
K =
E
p
I
p
L
3
1
_
X
e
_
1
4
+
_
+Y
e
_
1
8

2
_
+
e
_
1
4

_
X
e
e
2
e
_
4
e
2
e
(1)
+1
_
_
(
e
1)
3
(2.5.137)
Equation (2.5.137) gives the solution for stiffness of a partially embedded pile in the
ground. The correctness of the equation can be back checked by the fact that when
the pile becomes fully embedded i.e. for L
1
= L 1,
e
= , X
e
= X etc., when
Equation (2.5.137) degenerates to Equation (2.5.118).
Proceeding in an identical manner as done before, the mass and damping terms may
be computed as
M
x
=

p
A
p
L
1
4g
_
X
e
(1 e
2
e
) +
Y
e
2
+
e
1/(
e
1)
_
(2.5.138)
C
x
= r
0
_
GS
x2
L
1
_
X
e
(1 e
2
e
) +
Y
e
2
+
e
4/(
e
1)
_
(2.5.139)
2.5.11.6 Pile embedded in soils with varying elastic property
We present now the effect of variation of shear modulus with respect to depth. In the
previous section, the calculation of stiffness as well as the damping of soil was based on
constant dynamic shear modulus of the soil. For varying shear modulus, the variation
with depth can be expressed as
G
= G(z/L)
m
(2.5.140)
where m=a number varying from 02 [considered 0 when G is constant with depth,
assumed 1 for linear variation and 2 for parabolic distribution].
For a linearly varying soil the stiffness matrix can be written as
K =
4E
p
I
p
4
L
4
L
_
0
e
2z
L
_
X
2

Y
2
cos
2z
L
sin
2z
L
_
dz
+GS
x1
L
_
0
_
z
L
_
e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
dz (2.5.141)
Integration of above and ignoring the terms containing the factor, e
2
cos 2,
e
2
sin 2 etc., having extremely small contributions, Equation (2.5.141) reduces to
K =
E
p
I
p
3
L
3
_
X(1e
2
)
Y
2

_
+
GS
x1
L
4
2
_
X[1 e
2
(1 +)] +
3Y
4
+

2
_
(2.5.142)
and can be further simplied to
K =
E
p
I
p
3
L
3
_
X
_
1 e
2
_
1 +
1
4
+
1
4
__
Y
_
1
2

3
16
_
1
1
8
__
(2.5.143)
The damping matrix for this case, proceeding in same manner as outlined earlier,
can be represented by
C
x
=
r
0
_
G S
x2
L
4
2
_
X[1 e
2
(1 +)] +
3Y
4
+

2
_
(2.5.144)
The mass coefficient remains the same as expressed in Equation (2.5.122).
When the dynamic shear modulus variation is parabolic with depth, the stiffness
equation of the pile can be expressed as
K =
4E
p
I
p
4
L
4
L
_
0
e
2z
L
_
X
2

Y
2
cos
2z
L
sin
2z
L
_
dz
+ GS
x1
L
_
0
_
z
L
_
2
e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
dz (2.5.145)
Equation (2.5.145) on integration and on subsequent simplification reduces to
K =
E
p
I
p
3
L
3
_
X(1 e
2
)
Y
2

_
+
GS
x1
L
4
_
X
_
1
4
2
e
2
_
2 +
1
2
___
(2.5.146)
which can be further simplified to
K =
E
p
I
p
3
L
3
_
X
__
1 +
1
16
2
_
e
2
_
3
2
+
1
4
1
8
2
__
Y
2

_
(2.5.147)
Equation (2.5.147) gives the stiffness expression of pile under parabolic variation
of G along the length of pile.
Proceeding in same manner as stated above the damping matrix may be expressed as
C
x
=
r
0
GS
x2
L
4
_
X
_
1
4
2
e
2
_
2 +
1
2
___
(2.5.148)
The mass coefficient remains the same as expressed in Equation (2.5.122).
2.5.11.7 Computation of bending moment and shear force
For machine foundation subjected to a lateral load of P
0
sin
m
t, the amplitude of
vibration is given by
v(t) =
P
0
K
sin
m
t
_
(1 r
2
)
2
+(2Dr)
2
(2.5.149)
where,
m
= operating frequency of the machine; P
0
=unbalanced dynamic force;
r =
m
/
n
=the ratio of operating and natural frequency; D=damping ratio of the
system.
Thus the peak amplitude is given by
v(t) =
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
(2.5.150)
The complete displacement function is then given by
v(z, t) =
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
(z) (2.5.151)
or v(z, t) =
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
e
z
L
_
cos
z
L
+ sin
z
L
_
(2.5.152)
The bending moment is given by
E
p
I
p
v
= M(z) =
E
p
I
p
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
2
2
L
2
e
z
L
_
sin
z
L
cos
z
L
_
(2.5.153)
M
z
=
E
p
I
p
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
2
2
L
2
e
z
L
_
sin
z
L
cos
z
L
_
(2.5.154)
The maximum moment will be at the head i.e. at z = 0, and it can be expressed as
M
max
=
2E
p
I
p
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
_
( +1)
L
2
_
(2.5.155)
The shear force is given by
E
p
I
p
v
= V(z)=
E
p
I
p
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
2
3
L
3
_
( 1) sin
z
L
+( +1) cos
z
L
_
or
V(z) =
E
p
I
p
P
0
K
_
(1 r
2
)
2
+(2Dr)
2
2
3
L
3
_
( 1) sin
z
L
+( +1) cos
z
L
_
(2.5.156)
2.5.11.8 Dynamic response of short piles
in the horizontal mode
There are no solutions till date for this type of piles. Existing solutions are based on
long piles with the implicit assumption that under ultimate load piles fail before the
soil. However there are number of areas (e.g. Bonny River Delta in Nigeria, where the
topsoil constitute of very weak clay underlain by dense sand) where the soil will yield
much before the pile. Broms (1965) has shown that the displacement curvatures for
such piles are completely different than that of long piles.
While a long pile embedded in soil behaves as a semi-infinite beam on elastic foun-
dation, a short pile behaves as a beam of finite length on elastic foundation. Bojtsov
et al. (1982) has given solution to the generic displacement curves of such short beams
on elastic foundation that is given by
x = C
0
cos hpz cos pz +C
1
cos hpz sin pz +C
2
sin hpz sin pz +C
3
sin hpz cos pz
(2.5.157)
where p is same as expressed in Equation (2.5.96).
Expressing in terms of Puzrevsky function (Karnovsky and Lebed 2001), Equation
(2.5.157) can be expressed as
x = C
0
V
0
(pz) +C
1
V
1
(pz) +C
2
V
2
(pz) +C
3
V
3
(pz) (2.5.158)
where, V
0
(pz) = cosh pz cos pz (2.5.159)
V
1
(pz) =
1
2
(cosh pz sin pz +sinh pz cos pz) (2.5.160)
V
2
(pz) = sinh pz sin pz (2.5.161)
V
3
(pz) =
1
2
(cosh pz sin pz sinh pz cos pz) (2.5.162)
Puzrevskys functions, defined below, have some unique functional properties,
which will be used for subsequent analysis for derivation of the stiffness, damping
and mass of the piles.
V
0
(0) = 1; V
0
(0) = 0; V
0
(0) = 0; V
0
(0) = 0 (2.5.163)
V
1
(0) = 0; V
1
(0) = p
2, V
1
(0) = 0; V
1
(0) = 0 (2.5.164)
V
2
(0) = 0; V
2
(0) = 0; V
2
(0) = 2p
2
, V
2
(0) = 0 (2.5.165)
V
3
(0) = 0; V
3
(0) = 0; V
3
(0) = 0; V
3
(0) = 2
2p
3
(2.5.166)
V
3
(pz) = p
2V
2
(pz); V
2
(pz) = p
2V
1
(pz) (2.5.167)
V
1
(pz) = p
2V
0
(pz); V
0
(pz) = p
2V
3
(pz) (2.5.168)
For a solution of the short pile one may use the model shown in Figure 2.5.23.
For the analysis (similar to long piles) the pile may be assumed as fixed at base and
can undergo deflection and rotation at the pile head.
M
P
X
Z
Soil Stiffness=GSx
1
dz
Figure 2.5.23 Conceptual model of short pile under lateral load.
Considering base of pile at z = 0, shown in Figure 2.5.23, one may write
At z = 0, x = 0 C
0
= 0
At z = 0, x
= 0 C
1
= 0
which gives, x = C
2
V
2
(pz) +C
3
V
3
(pz) (2.5.169)
At the pile head, i.e. at z = L x = 1 yielding,
C
2
V
2
(pL) +C
3
V
3
(pL) = 1 (2.5.170)
Again at z = L x
= 1/L which gives,

C
2
V
2
(pL) +C
3
V
3
(pL) = 1/L. (2.5.171)
Using Equations (2.5.167) and (2.5.170), one may write
C
2
V
1
(pL) +C
3
V
2
(pL) =
1
pL
2
(2.5.172)
The above may be expressed in matrix form as
[V] {C} =
_
p
_
(2.5.173)
which can be further reduced to
{C} = [V]
1
{p} (2.5.174)
Performing the above operation gives
_
C
2
C
3
_
=
1
_
V
2
(pL) V
3
(pL)
V
1
(pL) V
2
(pL)
_ _
1
1/pL
2
_
(2.5.175)
where = V
2
2
(pL) V
1
(pL)V
3
(pL) which implies
C
2
=
1
_
V
2
(pL)
V
3
(pL)
pL
2
_
and C
3
=
1
_
V
2
(pL)
pL
2
V
1
(pL)
_
(2.5.176)
Thus, the displacement for the given boundary condition is then expressed as
x =
1
_
V
2
(pL)
V
3
(pL)
pL
2
_
V
2
(pz) +
1
_
V
2
(pL)
pL
2
V
1
(pL)
_
V
3
(pz) (2.5.177)
Based on above, the generic shape function in dimensionless form is given by
(z) =
1
_
V
2
()
V
3
()
2
_
V
2
_
z
L
_
+
1
_
V
2
()
2
V
1
()
_
V
3
_
z
L
_
(2.5.178)
where the determinant gets modified to = V
2
2
() V
1
()V
3
().
Considering A = C
2
/ and B = C
3
/ the shape function can now be expressed as
(z) = AV
2
_
z
L
_
+BV
3
_
z
L
_
(2.5.179)
Atypical shape function for the short piles E
p
/G
s
= 2500 is shown in Figure 2.5.24.
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0
0
.
1
0
.
2
0
.
3
0
.
4
0
.
5
0
.
6
0
.
7
0
.
8
0
.
9 1
z/L
S
h
a
p
e

F
u
n
c
t
i
o
n
Figure 2.5.24 Generic shape function of short pile for E
p
/G = 2500.
Differentiating Equation (2.5.179) and using properties mentioned earlier one
could have
(z) =
2
2
L
2
_
AV
0
_
z
L
_
+BV
1
_
z
L
__
(2.5.180)
Substituting the above functions in Equation (2.5.110), the stiffness can be
expressed as
K =
4E
p
I
p
4
L
4
L
_
0
_
AV
0
_
z
L
_
+BV
1
_
z
L
__
2
dz
+GS
x1
L
_
0
_
AV
2
_
z
L
_
+BV
3
_
z
L
__
2
dz (2.5.181)
Equation (2.5.181) is too complicated to solve in closed form and a numerical
quadrature scheme may be used to obtain K.
Considering = z/L we have L d = dz and as z L; 1; as z 0 0;
which gives
K=
4E
p
I
p
4
L
4
1
_
0
[AV
0
() +BV
1
()]
2
Ld +GS
x1
L
_
0
[AV
2
() +BV
3
()]
2
Ld
(2.5.182)
Substituting the value of [Equation (2.5.102)] in Equation (2.5.182), the stiffness
may be written as
K = GS
x1
L
_
_
4
1
_
0
[AV
0
() +BV
1
()]
2
d +
1
_
0
[AV
2
() +BV
3
()]
2
d
_
_
(2.5.183)
K = GS
x1
L[4I
1
+I
2
] (2.5.184)
where I
1
=
1
_
0
[AV
0
() +BV
1
()]
2
d and I
2
=
1
_
0
[AV
2
() +BV
3
()]
2
d
(2.5.185)
The integrals I
1
and I
2
can very easily be solved by using Simpsons 1/3rd rule
between limits 01 and can be back substituted in Equation (2.5.184) to compute the
stiffness for the short pile.
However, one should note that there is no theoretical or experimental benchmarking
against which the stiffness values can be checked or compared. So, use of the expression
must always be backed up by dynamic field test of the piles to adjust the data (especially
S
x1
or E
p
/G) to match with the field observed values. In absence of comparative
benchmarks the design may be initiated with the suggestive values of S
x1
for various
E
p
/G
s
The values mentioned in Table 2.5.18, are based on the formulation for long pile
(with L/r < 25) but may be used as a starting point for the iteration based on field
observed data.
The mass of pile for the fundamental mode is given by
M
x
=

p
A
p
g
L
_
0
(z)
2
dz
or M
x
=

p
A
p
L
g
1
_
0
[AV
2
() +BV
3
()]
2
d (2.5.186)
M
x
=

p
A
p
L
g
I
2
(2.5.187)
To start the design a value of S
x1
is selected for specic E
p
/G
s
from Table 2.5.18
and find out the value of the frequency (
_
K/M) based on Equations (2.5.184)
and (2.5.187). Let this be defined as
c
where the subscript c stands for the word
computed. Let the field-tested natural frequency of the pile be
f
, where,
f
=
c
.
In most of cases it has been seen (Jadi 1999) that the field observed frequency value
deviates from the computed ones and usually varies by about 3040%. This is logical,
for when the pile is bored or driven the soil gets displaced and clayey soil may loose a
part of its shear strength thus resulting in reduced dynamic shear modulus compared
to the value observed during geo-technical investigation. There could be cases where
the field observed values might be more than the computed ones, especially in sandy
soil where the soil gets densified due to pile driving. The bottom line is that in rare
cases the computed and observed values would match.
Table 2.5.18 Suggested for S
x1
for short piles (L/r 20) for eld data
iteration.
E
p
/G
s
S
x1
( = 0.25) S
x1
( = 0.4) S
x1
( = 0.5)
250 1.53 1.75 1.89
500 1.35 1.54 1.68
1000 1.17 1.34 1.46
2500 0.95 1.09 1.46
5000 0.95 1.09 1.46
10000 0.95 1.09 1.46
Based on the above argument the error () in the analysis is then given by
=
c

f
For 0 we have
c
=
f

2
c
=
2
f
.
Considering
2
c
=
K
M
x
and using Equations (2.5.184) and (2.5.187), one can have
GS
x1
g
p
A
p
_
4
I
1
I
2
+1
_
2
f
= 0 (2.5.188)
It will be observed that all the factors , I
1
, I
2
in Equation (2.5.188) is a function
of E
p
/G. The difference (=the error ) can now be set to zero or minimum by varying
the value of E
p
/G for which lim 0.
This can very easily be done by using the standard solver or goal-seek in a spreadsheet
with boundary constraint that S
x1
> 0.
The solver basically uses an algorithmcalled generalized reduced gradient technique
(GRG2) used for constrained optimization (Lasdon et al. 1978). The procedure begins
with the nonlinear optimization technique with equality constraints. The necessary
slack and surplus variables are added as x
s
or x
2
s
to any inequality constraints, and the
problem is to
Optimize: y(x)
Subject to: f
i
(x) = 0: for i = 1, 2, . . . , j
where j is the number of constrained equations and n is the number of independent
variables where n > m.
This is a very standard technique used in all nonlinear programming and is used
routinely as a mathematical tool in many standard commercially available software
like MS excel, MATLAB etc. having varied applications in engineering, science and
economics modeling.
Use of the above will automatically revise the value of E
p
/G and upgrade the values
of I
2
and I
1
(dimensionless but a function of E
p
/G), which may then be used to calculate
the revised and exact stiffness and mass contribution of the pile which would closely
simulate the field condition.
The steps are furnished in detail in Figure 2.5.27 as to how the data are updated
and corrected for the example cited in example mentioned below.
Having established the mass and stiffness coefficients of the pile correctly based on
field data the damping may now be established as
C
x
= r
0
_
GS
x2
LI
2
(2.5.189)
where I
2
is the corrected upgraded value and S
x2
is as obtained from Table 2.5.17.
2.5.11.8.1 Comparison of results
A comparison of results against established methods to ensure that the method is not
an utopian exercise with differential equations and it does have applications.
0.00E+00
1.00E+05
2.00E+05
3.00E+05
4.00E+05
5.00E+05
6.00E+05
2
5
0
5
0
0
1
0
0
0
2
5
0
0
5
0
0
0
1
0
0
0
0
E
p
/G
s
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kxx
Novak
Gazetas
Figure 2.5.25 Comparison of stiffness values for, r = 0.3 m and length = 30 m.
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
2
5
0
5
0
0
1
0
0
0
2
5
0
0
5
0
0
0
1
0
0
0
0
E
p
/G
s
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kxx
Novak
Gazetas
Figure 2.5.26 Comparison of stiffness values for, r = 0.6 m and length = 30 m.
For this two RCC piles of radius 0.3 m, 0.6 m of length 30 m has been has been
checked with the reported results for comparison. The values K
xx
[Equation (2.5.106)]
is shown in Figures 2.5.25 and 2.5.26 for comparison.
Next, the results of uncoupled horizontal frequency of a real time compressor foun-
dation weighing 400 kNsupported on 9 RCCpiles of length 36 mand diameter 1.8 m.
The pile cap size is 7 m 5 m 2 m. The piles are spaced at distance of 3.0 m.
The natural frequencies of the foundation are compared for E
p
/G value varying from
25010,000. Weight of the compressor is 400 kN.
Table 2.5.19 clearly shows that the values are in very good agreement for the base
case and thus can well be used for other cases as mentioned above for which there are
no direct solutions.
Finally, the stiffness of a short pile has been computed. This is based on the field
observed data having the following properties:
Length of pile = 10 m, diameter of pile = 1.2 m. Material of pile = RCC.
method of installation-bored pile.
based on soil test, observed E
p
/G = 5000.
E
p
considered = 3 10
7
kN/m
2
.
unit weight of pile material = 25 kN/m
3
.
field observed natural frequency of the pile is = 58 rad/sec (9 Hz).
Poissons ratio of soil considered = 0.4.
Table 2.5.19 Comparison of frequency for a compressor foundation proposed versus Novak and
Gazzetas.
Frequency
Frequency (rad/sec) Frequency (rad/sec) (rad/sec) with
Ep/G with K
proposed
with K
Novak
K
Gazzetas
250 252.64 251.44 252.57
500 192.10 194.85 192.07
1000 146.14 150.76 146.07
2500 101.79 107.21 101.71
5000 77.30 94.66
77.35
10000 58.87 63.98 58.82
The stiffness value was linearly interpolated from Novak (1983) Table for Ep/G = 500.
For the above conditions:
Selected value of S
x1
from Table 6 = 1.09
E
p
/G
s
= 5000 (given),
= 2.1512 Equation (2.5.102)
A = 0.50135; B = 0.02705 Equation (2.5.179)
I
1
= 0.23802, I
2
= 0.9035 Equation (2.5.185)
Computed natural frequency
__
K
p
/M
p
_
= 68.26 rad/sec (11 Hz)
Error () = 10.26
Setting the error () = 0 and running the solver function in a spread sheet for
changing E
p
/G
s
for boundary constraint S
x1
> 0, the following upgraded data have
been obtained:
S
x1
= 1.09; E
p
/G
s
= 7246; =1.96064; A = 0.65984; B = 0.04832; I
1
=
0.27266 and I
2
= 0.949504.
Computed natural frequency based on above data = 58 rad/sec (9 Hz).
Revised error () = 2.79 10
7
.
Thus based on the above data as per Equation (2.5.184), the correct stiffness of the
pile is deduced as K
pile
= 9.206 10
4
kN/m.
It is to be noted here that the E
p
/G value has increased from 5000 to 7246 meaning
thereby that the soil had lost some of its initial strength due to boring of the pile-which
is quite logical.
2.5.11.8.2 Computer run steps for short pile based on f ield observed data
The following section shows the computer run for evaluation of the stiffness of the
pile in lateral direction in three steps.
1 Stiffness and frequency calculation of pile based on theoretical data and calculating
the error based on field observed data.
2 The data screen just prior to run of the solver with command to change E
p
/G
value keeping the Sx value > 0.
3 Final value of the stiffness and frequency of the pile after solver has optimized the
data.
Step-1: Shows the initial calculation of frequency and stiffness of pile including the error with
respect to field observed frequency.
Figure 2.5.27 Steps of computation.
Steps of calculations are given in Figure 2.5.27.
A comprehensive analytical solution for dynamic analysis of long piles has been
presented and is in good agreement with the existing solution. Based on this, piles
with boundary conditions like partial embedment and soils with varying G can also
be analyzed.
Considering the fact that the dynamic bending moment and shear force can also be
obtained by this method, the standard practice of restricting the pile capacity to 50%
of its capacity will not be necessary. It will be observed from Equations (2.5.153) and
(2.5.156) that the moment and shear takes care of the dynamic magnification factor
of the load at the same time gives a complete distribution of its magnitudes along the
depth of the pile. This when combined with static load would give the design moment
for the pile. Considering that there is no uncertainty with this formulation, one can
perhaps restricts the pile load limit to 80% of its capacity in lieu of 50% as in vogue
presently and this would bring significant economy in design and for large project
savings could quite significant.
Short piles, for which no established method exists, also can be solved by the present
method.
Step-2: Showing solver on the verge of optimizing by changing Ep/G value by setting the error
to zero.
Figure 2.5.27 (continued).
2.5.11.9 Dynamic analysis of piles under rocking or rotational mode
We present herein the mathematical model for rocking or rotational mode. This mode
generally comes coupled with translational mode.
Shown in Figure 2.5.21 is a pile embedded in ground considered in a homogeneous
elastic medium under plane strain condition. The pile is considered to be long and
slender.
Under static loading, the equation of equilibrium in the x-direction for such beam
on elastic foundation is given by Timoshenko (1956)
E
p
I
p
d
4
x
dz
4
= k
s
x (2.5.190)
where E
p
= Youngs modulus of the pile; I
p
= moment of inertia of the pile cross
section; k
s
= elastic stiffness of the soil and is expressed as GS
1
; G
s
= dynamic shear
modulus of the soil; S
1
=Berdugos rotational constant which are basically frequency
dependent,
Step-3: Final value of stiffness of piles after the solver has optimized the error.
Figure 2.5.27 (continued).
The general solution of Equation (2.5.190) is given by
x = e
qz
(C
0
cos qz +C
1
sin qz) +e
qz
(C
2
cos qz +C
3
sin qz) (2.5.191)
where q =
4
_
GS
1
E
p
I
p
(2.5.192)
Similar to lateral load case the deflection equation can be considered as
x = e
qz
(C
0
cos qz +C
1
sin qz) (2.5.193)
Considering the pile head undergoing specified deflection and rotation as well as
its head is fixed to the pile cap (same boundary condition as considered by Novak
(1974)), we have
At z = 0, let x = x
0
C
0
= x
0
, which gives
x = e
qz
(x
0
cos qz +C
1
sin qz) (2.5.194)
Again considering at z = 0,
dx
dz
=
0
, we have
C
1
= x
0
+

0
q
(2.5.195)
Thus Equation (2.5.194) can now be represented as
x = e
qz
_
x
0
cos qz +
_
x
0
+

0
q
_
sin qz
_
(2.5.196)
Dividing each of the above term by L we have
x
L
= e
qz
_
x
0
L
cos qz +
_
x
0
L
+

0
qL
_
sin qz
_
(2.5.197)
For magnitude of rotation being small
0

=
x
0
L
and
z

=
x
L
when we have
z
=
0
e
qz
_
cos qz +
_
1 +
1
qL
_
sin qz
_
(2.5.197a)
Now considering = qL and looking at Equation (2.5.197) we can say that
for any arbitrary loading, the generic shape function in dimensionless form can be
represented as
(z) = e
z
L
_
cos
z
L
+
_
1 +
1
_
sin
z
L
_
(2.5.198)
where
=
4
_
GS
1
L
4
E
p
I
p
; L = Length of the pile. (2.5.199)
Equation (2.5.198) can thus be written as
(z) = e
z
L
_
cos
z
L
+ sin
z
L
_
(2.5.200)
Thus it is observed that shape function for rotational mode remains invariant with
respect to the lateral motion of pile for the given boundary condition.
Generic Shape Function of Pile in Rotational Mode
-0.5
0
0.5
1
1.5
0
0
.
1
5
0
.
3
0
.
4
5
0
.
6
0
.
7
5
0
.
9
z/L
S
h
a
p
e

f
u
n
c
t
i
o
n
F(z)
Figure 2.5.28 Generic shape function of pile for E
p
/G = 5000.
Differentiating Equation (2.5.200) we have
(z) =

L
e
z
L
_
( 1) cos
z
L
(1 +) sin
z
L
_
(2.5.201)
when, = 1 +
1
(2.5.202)
The generic shape function of the pile in fundamental mode as per Equation
(2.5.200) is as shown in Figure 2.5.28 for E
p
/G = 5000.
The potential energy d of an element of depth dz, shown in Figure 2.5.25, under
rotational mode is then given by (Craig 1981)
d =
E
p
I
p
2
_
d
dz
_
2
+
K
2

2
(2.5.203)
where, E
p
=Youngs modulus of pile; I
p
=moment of inertia of pile; K
=rotational
stiffness of soil having dimension kN/m; =rotational displacement of pile in x
direction and may be written as (z)q(t).
For a rigid circular embedded footing with embedment D
f
in rotational mode may be expressed as
K
= G
b
r
3
0
_
C
1
+
G
s
G
b
D
f
r
0
_
S
1
+
D
2
f
3r
2
0
S
x1
__
(2.5.204)
where, K
= foundation stiffness in horizontal direction; G

s
= dynamic shear mod-
ulus of the soil along foundation surface; G
b
= dynamic shear modulus of the soil
at foundation base; r
0
= Radius of the foundation; C
1
and S
1
S
x1
= Beredugos
Constant which are basically frequency dependent.
Ignoring the first term within bracket in Equation (2.5.204) which represents the
contribution of base resistance, and substituting the same in Equations (2.5.203) for a
cylindrical element of depth dz, embedded in soil, and also ignoring the termcontaining
dz
2
which is exceedingly small the potential energy d may be written as
d =
E
p
I
p
2
_
d
dz
_
2
+
Gr
2
0
S
1
dz
2

2
(2.5.205)
the total potential energy over the whole length of the pile (L) is then given by
=
E
p
I
p
2
L
_
0
_
d
dz
_
2
dz +
Gr
2
0
S
1
2
L
_
0
2
dz (2.5.206)
Considering v(z, t) = (z)q(t), it can be proved that
K
ij
= E
p
I
p
L
_
0
i
(z)
j
(z)dz +Gr
2
0
S
1
L
_
0
i
(z)
j
(z)dz (2.5.207)
Thus for fundamental mode the rotational stiffness of the pile is then given by
K
= E
p
I
p
L
_
0
(z)
2
dz +Gr
2
0
S
1
L
_
0
(z)
2
dz (2.5.208)
where,
(z) is as expressed in Equation (2.5.201).

Substituting the value of (z) and
(z) in Equation 2.5.208, and carrying out

integration by parts and some simplication, we nally get the rotational stiffness as
K
=
E
p
I
p
L
_
_
X(1 +) (1 e
2
) +Y
_
1
2
+

4
_
_
1

2
_
2( 1)
_
_
(2.5.209)
where =
4G
2
S
1
E
p
2
and = L/r
0
the slenderness ratio of the pile. It is to be noted that
is a dimensionless quantity, X, Y, etc. are same as derived for lateral stiffness case.
The accuracy of Equation (2.5.209) will be dependent on the correct selection of S
1
.
For instance for rigid circular footing Novak and Beredugo (1972) has furnished
a frequency independent value of S
1
=2.5 (for any value Poissons ratio) which has
been found to give adequate accuracy for practical engineering design.
Comparing the stiffness data with Novak (1974) and Gazetas (1988) data it is pro-
posed that the following values [Tables 2.5.20 to 22] of S
1
be used for the calculation
of dynamic response of pile under rocking mode.
1
L/r
0
Poissons (slenderness
ratio ratio) S
1
(250) S
1
(500) S
1
(1000) S
1
(2500) S
1
(5000) S
1
(10000)
0.25 25 16.968 23.089 30.776 43.412 54.647 66.877
40 17.358 23.656 31.586 44.678 56.390 69.253
60 17.567 23.961 32.016 45.333 57.272 70.418
80 17.674 24.110 32.225 45.648 57.688 70.958
100 17.736 24.199 32.348 45.833 57.930 71.267
Note: The value in Parenthesis after S
1
depicts the value of E
p
/G
s
value of the soil.
1
L/r
0
ratio ratio) S
1
(250) S
1
(500) S
1
(1000) S
1
(2500) S
1
(5000) S
1
(10000)
0.40 25 18.037 24.623 32.937 46.707 59.054 72.614
40 18.448 25.221 33.794 48.05 60.909 75.145
60 18.671 25.543 34.249 48.748 61.851 76.393
80 18.781 25.702 34.471 49.084 62.298 76.974
100 18.847 25.795 34.603 49.281 62.557 77.307
1
p
/G
s
value of the soil.
1
L/r
0
ratio ratio) S
1
(250) S
1
(500) S
1
(1000) S
1
(2500) S
1
(5000) S
1
(10000)
0.50 25 18.717 25.599 34.316 48.817 61.888 76.316
40 19.141 26.217 35.202 50.21 63.813 78.946
60 19.37 26.55 35.674 50.936 64.794 80.247
80 19.484 26.714 35.905 51.285 65.259 80.853
100 19.552 26.811 36.041 51.49 65.531 81.203
1
p
/G
s
value of the soil.
For a particular pile having specific slenderness ratio and Poissons ratio of the
soil we select the value of S
1
from the above table and on substitution of the same
in Equation (2.5.199) and Equation (2.5.209) gives the solution of pile stiffness in
rocking mode.
2.5.11.9.1 Estimation of mass contribution of pile
The mass matrix of the pile may be expressed as
M
x
= m(z)
_

i
(z)
j
(z)dz (2.5.210)
For the present case of pile of length L, mass moment of inertia J
x
is represented by
J
x
=
M
x
L
L
_
0
_
r
2
0
4
dz +z
2
dz
_
(2.5.211)
Substituting Equation (2.5.210), we may now write
J
x
=

p
A
p
r
2
0
4g
L
_
0
(z)
2
dz +

p
A
p
L
2
g
L
_
0
_
z
L
_
2
(z)
2
dz (2.5.212)
where
p
= weight density of the pile material; A
p
= cross sectional area of pile;
g = acceleration due to gravity.
Equation (2.5.212) on integration by parts and simplification nally gives
J
x
=

p
A
p
r
2
0
L
16g
_
XF() +
Y
2
+
_
(2.5.213)
where F()
__
1 e
2
+

2
2
4
2
e
2
_
2 +
1
2
_
__
and = L/r
0
the slenderness ratio of the pile.
Equation (2.5.213) gives the inertial contribution of pile in the fundamental mode.
Incidentally the effect of this is usually ignored in design but could have significant
effect if the number of piles is large in a pile group.
2.5.11.9.2 Radiation damping factor for pile under rocking mode
f
, Novak and Beredugo (1972) has
proposed an expression
C
= r
4
0
_
G
_
C
2
+
G
s
G
D
f
r
0
_
S
2
+
D
2
f
3r
2
0
S
x2
__
(2.5.214)
where, r
0
= radius of the foundation; G = dynamic shear modulus at foundation
base; G
s
=dynamic shear modulus of soil in which the foundation is embedded; D
f
=
depth of embedment; C
2
, S
2
and S
x2
= frequency independent constants as defined
by Novak and Beredugo (1972).
Ignoring the first term in Equation (2.5.214) which represents the contribution of
base damping for a cylindrical element of depth dz, embedded in soil, and ignoring
the term, containing dz
2
which is again exceedingly small, we have
c() = r
3
0
_
G
s
S
2
dz (2.5.215)
For systems having continuous function, the damping is usually expressed as
C
= c()
_

i
(z)
j
(z)dz (2.5.216)
For the present case of pile of length L, Equation (2.5.216) can be expressed as
C
= r
3
0
_
G
s
S
2
L
_
0
(z)
2
dz (2.5.217)
Equation (2.5.217) on integration by parts and simplification, we have
C
= r
3
0
_
GS
2
L
_
X(1 e
2
) +
Y
2
+
4
_
(2.5.218)
Equation (2.5.218) expresses the soil damping for a single pile under horizontal
mode of vibration. Here the Factor S
2
is damping coefcient which is frequency
dependent. Fortunately the damping factor is required for calculation of the amplitude
when the eigen solution of the problem is already done vis a vis, the dimensionless fre-
quency number a
0
= r
0
/v
s
term is known. Polynomial fit curve for S
2
are available
in terms of a
0
which can be used directly to arrive at these parameters.
The value of S
2
is as given hereafter as per Novak and Beredugo (1972)
S
2
= 0.0144a
0
+5.263a
2
0
4.177a
3
0
+1.643a
4
0
0.2542a
5
0
(2.5.218a)
This value unlike other Beredugos constant is independent of Poissons ratio.
2.5.11.9.3 Consideration of material damping of pile
The structural stiffness contribution of the pile is given in the first part of Equation
(2.5.208), while that of the mass moment of inertia is given in Equation (2.5.212).
Thus, if C
c
is the critical damping of the pile then it can be expressed as C
c
= 2
_
KJ
x
,
where K and J
x
are the stiffness and mass moment of inertia of the pile.
Depending on the material used for pile like (RCC, steel etc.) a suitable damping
ratio (D) can be assumed. The damping (C
p
) for the pile can be expressed as
C
p
= DC
c
(2.5.219)
This, when added to the radiation damping, calculated in Equation (2.5.218) gives
the complete damping quantity for the soil-pile system.
2.5.11.9.4 Piles with other boundary conditions
Having established the stiffness, inertial and damping contribution of the pile in rock-
ing mode based on minimization of the potential energy of the system we extend the
above method for piles with other boundary conditions for which there are no standard
solutions.
2.5.12 Partially embedded piles under rocking mode
As stated earlier, this is a very common practice in Arctic and North Siberian condition,
where due to environmental reasons; the steel piles are driven into the ground when
they protrude about 23 m above the ground over which the pile cap and vibrating
equipments are placed. In such cases Novaks (1974, 1983) chart cannot be used, nor
is Gazetas formulation valid.
We provide the solution of the same as hereafter.
Let L be the full length of the pile and let the length of the embedment is soil be L
1
.
For this case we have
e
=
4
_
GS
1
L
4
1
E
p
I
p
(2.5.220)
Here subscript e represents embedment of the pile.
The shape function can thus be represented by
(z) = e
e
z
L
1
_
cos

e
z
L
1
+ sin

e
z
L
1
_
(2.5.221)
The stiffness function can thus be represented as
(z) =

e
L
1
e
e
z
L
1
_
(
e
1) cos

e
z
L
1
(1 +
e
) sin

e
z
L
1
_
(2.5.222)
Square of the above is given by
(z)
2
=

2
e
L
2
1
e
2
e
z
L
_
X
e
2
2
e
cos
2
e
z
L
1
+Y
e
sin
2
e
z
L
1
_
(2.5.223)
Here X = 1 +
2
e
; Y = 1
2
e
and
e
= 1 +
1
e
.
Now considering the fact that embedment of a beam does not have any effect on
the shape function of the system, the stiffness of the pile is expressed as
K
= E
p
I
p
L
_
0
i
(z)
2
dz +GS
1
L
1
_
0
[
i
(z)]
2
dz (2.5.224)
Equation (2.5.224) on integration by parts and simplification may be expressed as
K
=
E
p
I
p
e
2L
1
_
X
e
( +) X
e
{e
2
e
+e
2
e
}
+ Y
e
_
2
+

4
_
e
_

2
__
(2.5.225)
which can further be expressed as (here = L/L
1
)
K
=
E
p
I
p
L
1
_
X
e
( + e
2
e
e
2
e
) +Y
e
_
2
+

4
_
e
_

2
__
2(
e
1)
(2.5.226)
Equation (2.5.226) gives the solution for stiffness of partially embedded piles in the
ground. The correctness of the equation can be back checked by the fact that when the
pile becomes fully embedded i.e. L
1
= L we have 1,
e
= , X
e
= X etc. when
Equation (2.5.226) degenerates to Equation (2.5.209), the stiffness for fully embedded
pile.
Proceeding in identical manner as done before, the mass and damping terms can be
obtained as given earlier.
The mass moment of inertia of pile remains same as stated in Equation (2.5.213).
The damping matrix is given by the expression
C
= r
3
0
_
GS
2
L
1
_
X
e
(1 e
2
e
) +
Y
e
2
+
e
4/(
e
1)
_
(2.5.227)
2.5.12.1 Stiffness of the pile for soils with varying elastic property
Considering the variation of shear modulus with depth as
G
= G(z/L)
m
(2.5.228)
where m = a number varying from 02 [considered 0 when G is constant with depth,
assumed 1 for linear variation and 2 for parabolic distribution] we derive the pile
stiffness and other parameters as hereafter.
Thus for linearly varying soil the stiffness matrix can be written as
K
=
E
p
I
p
2
L
2
L
_
0
e
2z
L
_
X 2 cos
2z
L
Y sin
2z
L
_
dz
+Gr
2
0
S
1
L
_
0
_
z
L
_
e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
dz (2.5.229)
Equation (2.5.229) can be further simplified to
K
=
E
p
I
p
2L
_
X
_
1 +

2
e
2
_
1 +

2
(1 +)
__
+Y
_
3
8
+
1
2
_
_
1

4
__
(2.5.230)
The damping matrix for this case can thus be represented by
C
x
=
r
3
0
GS
2
L
4
2
_
X[1 e
2
(1 +)] +
3Y
4
+

2
_
(2.5.231)
The mass coefficient remains same as expressed in Equation (2.5.213).
When the dynamic shear modulus variation is parabolic with depth the stiffness
equation of the pile is expressed as
K
=
E
p
I
p
2
L
2
L
_
0
e
2z
L
_
X 2 cos
2z
L
+Y sin
2z
L
_
dz
+Gr
2
0
S
1
L
_
0
_
z
L
_
2
e
2z
L
_
X
2
+
Y
2
cos
2z
L
+ sin
2z
L
_
dz (2.5.232)
which can be further simplified and expressed as
K
=
E
p
I
p
2L
_
X
__
1 +

8
2
_
e
2
_
1 +

2
_
2 +
1
2
___
+
Y
2

_
(2.5.233)
Equation (2.5.233) gives the stiffness expression of pile under parabolic variation
of G along the length of pile.
Proceeding in same manner as stated above the damping matrix is expressed as
C
=
r
3
0
GS
2
L
4
_
X
_
1
4
2
e
2
_
2 +
1
2
___
(2.5.234)
The mass coefficient remains same as expressed in Equation (2.5.213).
2.5.12.1.1 Calculation of dynamic bending moment and shear force in pile
Neither Novak nor Gazetas method can be used for this purpose. For machine foun-
dation subjected to a dynamic moment of M
0
sin
m
t, the amplitude of vibration is
given by
(t) =
(M
0
/K
) sin
m
t
_
(1 r
2
)
2
+(2Dr)
2
(2.5.235)
where,
m
=operating frequency of the machine; M
0
=unbalanced dynamic moment;
=
m
/
n,
the ratio of operating and natural frequency; D = damping ratio of the
system.
Thus the peak amplitude is given by
(t) =
(M
0
/K
)
_
(1 r
2
)
2
+(2Dr)
2
(2.5.236)
The complete displacement function is then given by
(z, t) =
(M
0
/K
)
_
(1 r
2
)
2
+(2Dr)
2
e
z
L
_
cos
z
L
+ sin
z
L
_
(2.5.237)
Thus bending moment is given by
E
p
I
p
= M(x)
or, M(x) =
(E
p
I
p
M
0
)/K
_
(1 r
2
)
2
+(2Dr)
2
L
e
z
L
_
(1 +) sin
z
L
( 1) cos
z
L
_
(2.5.238)
The dynamic shear force is given by
E
p
I
p
= V(z) =
(E
p
I
p
M
0
)/K
_
(1 r
2
)
2
+(2Dr)
2
2
2
L
2
e
z
L
_
sin
z
L
cos
z
L
_
(2.5.239)
V(z) =
(E
p
I
p
M
0
)/K
_
(1 r
2
)
2
+(2Dr)
2
2
2
L
2
e
z
L
_
sin
z
L
cos
z
L
_
. (2.5.240)
2.5.12.2 Dynamic response of short piles under rotational mode
As mentioned earlier no solution exists till date for this type of piles. Bojtsov (1982)
has given solution to the generic displacement curvature of such short beams on elastic
foundation which is given by
x = C
0
cos hpz cos pz +C
1
cos hpz sin pz +C
2
sin hpz sin pz +C
3
sin hpz cos pz
(2.5.241)
where p = q is same as expressed in Equation (2.5.192).
Expressing the above in terms of Puzrevsky function (Karnovsky 2001) Equation
(2.5.257) can be expressed as
x = C
0
V
0
(pz) +C
1
V
1
(pz) +C
2
V
2
(pz) +C
3
V
3
(pz) (2.5.242)
where,
V
0
(pz) = cos hpz cos pz (2.5.243)
V
1
(pz) =
1
2
(cos hpz sin pz +sin hpz cos pz) (2.5.244)
V
2
(pz) = sin hpz sin pz (2.5.245)
V
3
(pz) =
1
2
(cos hpz sin pz sin hpz cos pz) (2.5.246)
For analysis similar to previous case we assume the pile as fixed at base and is fixed
also at pile cap level and can undergo deflection and rotation at pile head. Considering
base of pile as z = 0 and applying the Puzrevskys functional properties as elaborated
in case of piles under lateral load we have
At z = 0, x = 0 C
0
= 0
At z = 0, x
= 0 C
1
= 0 which gives
x = C
2
V
2
(pz) +C
3
V
3
(pz) (2.5.247)
At the pile head we have at z = L x = 1 which gives
C
2
V
2
( pL) +C
3
V
3
( pL) = 1 (2.5.248)
Again at z = L x
= 1/L which gives

C
2
V
2
(pL) +C
3
V
3
(pL) = 1/L (2.5.249)
Using the derivative properties as shown above we have
C
2
V
1
(pL) +C
3
V
2
(pL) =
1
pL
2
(2.5.250)
Expressing the above in matrix form we have
[V] {C} = {p} (2.5.251)
which can be further expressed as
{C} = [V]
1
{p} (2.5.252)
Performing the above operation gives
_
C
2
C
3
_
=
1
_
V
2
(pL) V
3
(pL)
V
1
(pL) V
2
(pL)
_
_
_
_
1
1
pL
2
_
_
_
(2.5.253)
where = V
2
2
(pL) V
1
(pL)V
3
(pL) which gives
C
2
=
1
_
V
2
(pL)
V
3
(pL)
pL
2
_
and C
3
=
1
_
V
2
(pL)
pL
2
V
1
(pL)
_
(2.5.254)
Thus the displacement for the given boundary condition is then expressed as
x =
1
_
V
2
(pL)
V
3
(pL)
pL
2
_
V
2
(pz) +
1
_
V
2
(pL)
pL
2
V
1
(pL)
_
V
3
(pz) (2.5.255)
Considering the fact that for long piles the shape function remains invariant for
rocking mode with respect to lateral motion, for same boundary condition it may
be concluded that for short piles also the same condition would hold good thus the
generic shape function in dimensionless form in rocking mode is given by
(z) =
1
_
V
2
()
V
3
()
2
_
V
2
_
z
L
_
+
1
_
V
2
()
2
V
1
()
_
V
3
_
z
L
_
(2.5.256)
where the determinant gets modified to = V
2
2
() V
1
()V
3
().
Considering A = C
2
/ and B = C
3
/ the shape can now be expressed as
(z) = AV
2
_
z
L
_
+BV
3
_
z
L
_
(2.5.257)
Typical generic shape function for the short piles E
p
/G
s
= 2500 is as shown in
Figure 2.5.29.
Differentiation of above and using the differential properties as mentioned earlier
we have
(z) =

2
L
_
AV
1
_
z
L
_
+BV
2
_
z
L
__
(2.5.258)
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0
0
.
1
0
.
2
0
.
3
0
.
4
0
.
5
0
.
6
0
.
7
0
.
8
0
.
9 1
z/L
S
h
a
p
e

F
u
n
c
t
i
o
n
Figure 2.5.29 Generic shape function of short pile for E
p
/G = 2500.
Substituting the above functions, we have
K =
2E
p
I
p
2
L
2
L
_
0
_
AV
1
_
z
L
_
+BV
2
_
z
L
__
2
+Gr
2
0
S
1
L
_
0
_
AV
2
_
z
L
_
+BV
3
_
z
L
__
2
(2.5.259)
The above is too complicated to solve in closed form as such numerical integration
may be used to arrive at the stiffness value.
Considering =
z
L
we have L d = dz and as z L; 1; as z 0 0;
which gives
K =
2E
p
I
p
2
L
2
1
_
0
[AV
1
() +BV
2
()]
2
Ld
+Gr
2
0
S
1
L
_
0
[AV
2
() +BV
3
()]
2
Ld (2.5.260)
Substituting the value of (from Equation 2.5.209) in Equation (2.5.260), we have
K = Gr
2
0
S
1
L
_
_
2
1
_
0
[AV
1
() +BV
2
()]
2
d +
1
_
0
[AV
2
() +BV
3
()]
2
d
_
_
(2.5.261)
K = Gr
2
0
S
1
L
_
2
I
1
+I
2
_
(2.5.262)
Table 2.5.23 Suggested for S
1
for short piles (L/r 20) for eld data iteration.
E
p
/G
s
S
1
( = 0.25) S
1
( = 0.4) S
1
( = 0.5)
250 15.563 16.561 17.197
500 21.046 22.468 23.372
1000 27.873 29.860 31.135
2500 39.05 42.041 43.976
5000 49.07 53.014 55.576
10000 60.187 65.311 68.598
Here
I
1
=
1
_
0
[AV
1
() +BV
2
()]
2
d and I
2
=
1
_
0
[AV
2
() +BV
3
()]
2
d
(2.5.263)
The integrals I
1
and I
2
can very easily be solved by using Simpsons 1/3rd rule
between limits 01 and can be back substituted in Equation (2.5.261) to arrive at the
stiffness for the short pile.
As there is no theoretical or experimental benchmarking against which the stiffness
values can be checked or compared. So use of this expression must always be backed
up by dynamic field test of the piles to adjust the data (especially S
1
or E
p
/G) to match
the field observed value.
In the absence of comparative benchmarks we may start the design with the fol-
lowing suggestive values of S
1
for various E
p
/G
s
values given in Table 2.5.23. These
values as mentioned above, is based on formulation for long pile (with L/r <25) but
may be used as a starting point for the iteration based on field observed data.
The mass moment of inertia of the pile for fundamental mode is given by
J
x
=

p
A
p
r
2
0
4g
L
_
0
(z)
2
dz +

p
A
p
L
2
g
L
_
0
_
z
L
_
2
(z)
2
dz (2.5.264)
J
x
=

p
A
p
r
2
0
L
4g
I
1
+

p
A
p
L
3
g
I
3
(2.5.265)
Here I
3
=
1
_
0
2
[AV
2
() +BV
3
()]d (2.5.266)
or J
x
=
M
p
r
2
0
4
I
1
+M
p
L
2
I
3
(2.5.267)
where, M
p
=

p
A
p
L
g
J
x
=
M
p
r
2
0
4
[I
1
+4
2
I
2
] (2.5.268)
To start the design we select a value of S
1
for a specified E
p
/G
s
from Table 2.5.23
and find out the value of the frequency based on Equation (2.5.262) and (2.5.267).
Let this be defined as
c
where the subscript c stands for the word computed.
Let the field tested natural frequency of the pile be
f
where
f
=
c
.
Based on the above argument the error() in the analysis is then given by
=
c

f
For 0 we have,
c
=
f
or
2
c
=
2
f
Considering
2
c
=
K
M
x
, we have
4GS
1
L
M
p
_
2
I
1
+I
2
I
1
+4
2
I
2
_
2
f
= 0 (2.5.269)
It will be observed that all the factors , I
1
, I
2
in Equation (2.5.269) is a function
of E
p
/G
s
. The difference (which is the error ) can now be set to zero or minimum by
varying the value of E
p
/G
s
for which, lim 0.
This can very easily be done by using the standard solver or goal seek in a spread
sheet with boundary constraint that S
1
> 0.
The above will automatically revise the value of E/G
s
and upgrade the values of I
3
,
I
2
and I
1
(which are dimensionless functions), which may then be used to calculate
the revised and exact stiffness and mass contribution of the pile which would closely
simulate the field condition.
Having established the mass and stiffness coefficients of the pile correctly based on
field data the damping may now be established as
C
= r
3
0
_
GS
2
LI
2
(2.5.270)
2.5.13 Group effect of pile
Refer Section 2.5.7 where this has been dealt in detail and may well be used for this
case too.
2.5.13.1 Effect of pile cap on pile stiffness
The sketch given in Figure 2.5.30 represent the pile group with pile cap. In such case
usually the embedment stiffness GS
f
D
f
is added to the pile group stiffness and the
system is considered as a lumped mass single degree freedom system, the details of
which are furnished in Novak (1974) and Prakash and Puri (1988).
In conventional formulation as the stiffness matrix is statically coupled another set
of stiffness K
x
needs to be derived in addition to what has been derived above. To
circumvent this issue we propose to use the following model as shown in Figure 2.5.31.
To derive the equations we use the Lagranges equation from the energy principle
as derived earlier when we finally get the stiffness and mass matrix as
_
_
M
f
M
f
M
f
Z
c
M
f
M
f
+M
x
M
f
Z
c
M
f
Z
c
M
f
Z
c
J
x
+M
f
Z
2
c
_
_
_
_
_
x
u
_
_
_
+
_
_
K
f
0 0
0 K
x
0
0 0 K
_
_
_
_
_
x
u
_
_
_
= 0 (2.5.271)
The above gives the complete free vibration equation of motion for pile plus pile
cap with machine considering pile springs in translation and rocking mode.
Considering the equation to be dynamically coupled the damping matrix can now
be expressed as
[C] =
_
_
C
f
0 0
0 C
x
0
0 0 C
_
_
(2.5.272)
Z
c
D
f
Figure 2.5.30 Schematic diagram of pile and pile-cap with embedment.
x
M = Mass of (Pile cap + Machine)
Z
c
K
f
= Embedded stiffness of soil @ GS
f
D
f
mp = Mass of pile group
K
x
u
J
x
= Moment of inertia
of Pile group
K
Figure 2.5.31 Mathematical model of pile group and pile cap under coupled sliding and rocking mode.
where, M
f
= mass of pile cap plus mass of machine; M
x
= mass of pile group; J
x
=
mass moment of inertia of piles; Z
c
= center of mass of foundation plus machine
along vertical axes; K
f
= lateral embedded stiffness of pile cap @ G S
fx
D
f
; G =
dynamic shear modulus of soil; S
fx
= Berdugos constant @ 3.6, 4, 4.1 for = 0.0,
0.25, 0.4 respectively; D
f
= depth of embedment; K
x
= Group lateral stiffness of pile
group based on Equation (2.5.118) where K
=rotational stiffness of pile group; U =

potential energy of the system, and T = kinetic energy of the system.
It is to be noted that for pile group for calculation of mass and mass moment of
inertia the mass and inertia of single pile has to be multiplied by the number of piles
in the group. While for stiffness and damping the group stiffness and damping has to
be derived according to Equation in section 2.5.7.
2.5.14 Comparison of results
The method proposed herein is now compared withy Novak and Gazetas values to
check their accuracy.
For this two RCC piles of radius 0.4 m, 1.0 m of length 40 m has been has been
checked with the reported results for comparison. The values K
[Equation (2.5.209)]
is shown in Figures 2.5.32 and 33 for comparison.
The results clearly shows that the values are in very good agreement for the base
case and thus can well be used for other cases as mentioned above for which there are
no direct solution.
We finally calculate the stiffness of a short pile based on field observed data having
the following properties.
Length of pile = 10 m, Diameter of pile = 1.2 meter. Material of pile RCC.
Method of installation-bored pile.
Based on soil test, observed E
p
/G =5000.
E
p
considered @ 3 10
7
kN/m
2
.
Density of pile material = 25 kN/m
3
.
Field observed natural frequency of the pile is = 28 rad/sec (4 Hz).
Poissons ratio of soil considered = 0.4.
For the above conditions
Selected value of S
1
from Table 2.5.23 = 53.014.
E
p
/G
s
= 5000 (given); = 5.681 : vide Equation (2.5.198); A = 0.000912 : vide
Equation (2.5.271); B = 0.003447 do; I
1
= 0.0277902, I
2
= 0.201259,I
3
=
0.16886 : vide Equation (2.5.263).
Computed Stiffness = 7.7810
5
kN/m.
Computed natural frequency (
_
K
/J
x
) = 39.98 rad/sec (6 Hz).
Error () = 11.98
Setting the error () = 0 and running the goal seek function in a spread sheet for
changing E
p
/G
s
for boundary constraint S
1
> 0, we have the following upgraded
data:
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
2
5
0
5
0
0
1
0
0
0
2
5
0
0
5
0
0
0
1
0
0
0
0
E
p
/G
s
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kxx
Novak
Gazetas
Figure 2.5.32 Comparison of stiffness values for r = 0.4 m and length = 40 m.
Comparison of Rocking Stiffness for piles
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2
5
0
5
0
0
1
0
0
0
2
5
0
0
5
0
0
0
1
0
0
0
0
E
p
/G
s
S
t
i
f
f
n
e
s
s

(
k
N
/
m
)
Kxx
Novak
Gazetas
Figure 2.5.33 Comparison of stiffness values for r = 1.0 m and length = 40 m.
S
x1
= 53.014; E
p
/G
s
= 3354; = 6.2777; A = 0.000889; B = 0.00528; I
1
=
0.00111, I
2
= 0.138856, I
3
= 0.11672.
Computed natural frequency based on above data = 28 rad/sec (4 Hz).
Revised error () = 0.00092
Thus based on the above data as per Equation (2.5.261) the correct stiffness of the
pile is given by K
pile
= 2.6410
5
kN/m.
In case the above correction is already done for lateral pile stiffness and E/G value
has been already modified to suite the field observed data, the same can directly be
used without carrying out the above mentioned modification again.
Referring to Figures 2.5.32 and 2.5.33, it is observed that the results are in excellent
agreement with both Novak (1983) and Gazetas (1988) stiffness. Considering the base
case being in such agreement formulations for other cases like partial embedment,
varying shear modulus etc., can now be very easily adapted for which there are no
standard solutions.
The short pile case is basically a theoretical solution and needs significant field test
and lab testing to arrive at a predefined S
1
values which would make the method
more powerful.
However in absence of such data the present algorithm as mentioned herein could
become a very powerful tool for dynamic analysis of such piles for which no solution
is available till date and yet remains a serious practical problem.
Y
B
X Z X
L
Y
Foundation resting on ground Foundation on piles/springs
Figure 2.5.34
2.5.15 Practical aspects of design of machine
foundations
Enough gazing at the moon and theoretical contemplation, from the hallowed domain
of academics let us now digress into the real world of a professional engineer and see
based on the above theories how he goes about to design the foundation
What are the input data he looks for?
What are the assumptions he considers in his process of design.
There are three aspects to be considered at the start of the design
1 Environmental and economic impact.
2 Machine data.
3 Soil data.
2.5.15.1 Environmental and economic impact
This is the first point that a designer should assess, but unfortunately the effect of
environmental impact on the machine foundation is often overlooked.
There could be a situation, where other than the vibration of the machine itself
there are external source of vibration affecting the foundation and this could be in the
form of
Blasting in the vicinity of the foundation
Pile driving
Waves transmitted by other machines operating in the vicinity of the machine
foundation in question.
Our experience shows that young engineers while doing their design of machine
foundations are more focussed on the quantitative magnitude of the natural frequency
and the amplitude and often overlooks this point.
To asses the effect of environmental impact on the foundation, if need be, seek help
of a vibration specialist and try to assess what could be the cascading effect of this
secondary source of disturbance.
If it is felt that this may possibly have some effect on the foundation isolate the
foundation by providing pockets/cut outs all-round the foundation and leaving this
space void or feeling it up with suitable dampers like cork boards, felt sheets etc.
Next try to assess how important role does the machine play in the overall process
system.
In other words, What would be the economic impact of the machine on the overall
process vis--vis its performance?
For instance if a minor chemical pump stops during an engineering process the over-
all cost impact on the process could vary from a few hundreds of dollars to thousand
dollar.
While for a major generator or a compressor foundation if the performance is not
up to the acceptable standard the client could stand to loose millions of dollars in
terms of production output and man-hours lost.
If required talk to your process engineering or mechanical engineering colleagues to
asses the criticality of the machine.
More important is the machine be more conservative in your design approach.
Do not try to economise on the material. The money that could be saved by cutting
down on a few cubic meter of concrete or hundred Kilogram of reinforcement, could
be well be offset by manifolds if your company stands to pay liquidated damages due
to malfunctioning of the foundation
15
.
For machine foundations economy lies more on the smooth performance of the
machine rather than any other factors.
2.5.15.2 Machine data
Once you have assessed the above aspect, as a next step, you should have the
machine data at your disposal in the form of a General Arrangement Drawing of
the machine.
On study of the drawing see if the following check list is satisfied as a minimum
1 Do the drawing furnishes the overall dimension of the machine/skid on which it
is mounted?
2 Are the anchor bolt locations, size of the bolts (both diameter and length) and
details of how it should be anchored to the foundation furnished by the vendor?
3 Do the drawing supply the height at which the centre line of the shaft of the
machine is located from the bottom of the machine frame (which will be the top
of concrete or top of grout for you)?
4 Is it clear to you what type of machine it is i.e. if it is centrifugal or reciprocating
in nature?
15 And this we are sure will not have a very positive outcome on your annual performance
appraisal. . . . . .
5 Does the drawing supply you with the operating speed of the machine or the range
which should be cleared during the design of the machine foundation?
6 Do the foundations need to support any pipes or valves on it other than the
machine itself?
7 If so, are all the loads and locations of these valves and pipes are mentioned in
the drawing?
8 Does the drawing clearly mention the unbalanced mass, eccentricity or the
dynamic loads generated during the operation of the machine including any
specic direction?
9 Is it clear to you what would be the level of the top of concrete of the foundation?
This is very important for the top of foundation usually fixed from the process
engineering group and if there is any mismatch in the level in the field could create
problems in terms of alignment of pipe flanges or variation in the net positive
suction head (NPSH) for the pump.
10 Is the location of the equipment in terms of co-ordinates with respect to the overall
plant available with you?
11 Finally has the equipment supplier defined any performance criterion which needs
to be met in terms of amplitude, frequency etc.
The above are very vital points both from performance and contractual point of
view. For if the equipment supplier has furnished this information then it should be
strictly adhered to, for once this is complied with the supplier alone stands guarantee
for the performance of the machine.
On the contrary if this is violated, even if the equipment supplied is faulty, the
vendor can always wriggle out of the situation by saying that his specifications were
violated and as such he cannot stand guarantee for the performance of the machine
16
.
If the vendor has not specified such conditions the usual de-fault is the local code
stipulation.
But do not presume this, ask him specifically to define his performance criteria and
if he is unable to do so, make it clear to him (in writing) as to what performance
criterion you are using based on which code (could be IS, DIN, BS, ASTM special
publications etc). If possible seek his written compliance that the code-norms that is
being followed by you is acceptable to him.
Remember for important machines you are fiddling with millions of dollars so play
safe. Guard yourself both technically as well as contractually.
2.5.15.3 My Machine is perfectly balanced. . . you dont need
to worry about the dynamic force!
A standard sales talk you will hear time and again from the equipment sales
engineer.
Rookie engineers often get carried away by this and fall prey to this over sales
strategy.
16 Refer to case history 2 at the outset of this chapter and retrospect a bit.
Many equipment suppliers do not supply any unbalanced dynamic load claiming
their machines to be perfectly balanced!
This often leaves an inexperienced engineer with the option of doing only a resonance
check and leaves it at that for he has no other data as a guideline to performany further
check.
What should be realised at this point is that it is possible perhaps to achieve
a perfect balance in the manufacturing unit under a controlled condition at the
outset.
But when such machines are performing under a much gruelling conditions of oper-
ating day in day out and often left exposed to the vagaries of nature, due to normal
wear and tear some imbalance will invariably be generated in the system which will
induce dynamic loads on the foundation.
So do not get carried away by the claims of the vendor, for you as designer alone
remain responsible for the performance of the foundation.
In absence of such data from the vendor you may use the following guidelines
(Arya et al. 1979).
2.5.15.3.1 Design eccentricities of centrifugal machines
Ecentricities of machines under varying speeds are given in the table below.
Eccentricity in double
Sl. No. Operating speed amplitude(inch)
1 750 0.0140.032
2 1500 0.008
3 3000 0.002
Here unbalanced dynamic force for centrifugal machine is given by
F
dyn
= me
2
m
where, m = mass of the rotating shaft; e = eccentricity developed in the shaft, and
m
= operating speed of the machine.
2.5.15.3.2 For Centrifugal compressors
e(mil) =
_
12,000
r.p.m.
1.0(mil)
where, = 0.5 at installation time = 1.0 after several years of operation
r.p.m. = Operating frequency of the machine.
1 mil = 0.001 inch.
2.5.15.3.3 For electrical motors
Peak to peak
displacement
Sl. No. Motor type Speed (RPM) amplitude (inch)
1 Integral horsepower electric motor 30004000 0.0010
15002999 0.0015
10001499 0.0020
999 and below 0.0025
2 Large induction motor 3000 and above 0.0010
15002999 0.0020
10001499 0.0025
999 and below 0.0030
2.5.15.4 Soil data
These data are furnished in the geo-technical report and should supply you with the
following parameters
17
Ground water table prevalent at the site
Atterbergs limits
Poissons ratio of the soil,
Unit weight of the soil,
Dynamic shear modulus of the soil, G
The foundation depth and bearing capacity of the soil at which the above
parameters are valid
All other information, regarding the static design of the foundation.
The knowledge of ground water table is essential for all block foundations and
should preferably have the bottom of foundation above the ground water table for
waves passing through water attenuates the dynamic response.
A check on the Atterbergs limit can give a very good indication qualitatively about
the fundamental property of the soil as to how it will behave. But unfortunately very
little attention is paid to this aspect in design offices.
The various Atterbergs limits like liquid limit, plastic limit etc not only give a clear
indication of how the soil would behave but also holds key to the fact that if the soil
is sensitive to shocks induced by vibration or not.
We do not discuss the details of Atterbergs limit and its interpretations but make
you aware of one criterion which is quite important in context of machine foundation
design.
Generically when the natural moisture content of the soil is closer to the liquid limit
the soil is deemed soft and when the natural moisture content is close to the plastic
17 Here we assume the reader has some knowledge about the static design procedure of a foundation.
limit it is considered as stiff. However there are certain types of soils whose natural
moisture content is greater than the liquid limit. If you ever encounter such case you
should immediately be on the alert.
For such soils generally belong to the montmorillonite group and constitutes a brittle
structure. This type of soil, when disturbed by vibration, flows like a liquid. If this soil
is allowed to remain in place it can be very dangerous for the foundation which may
undergo sudden settlement without any notice.
The liquidity index values of such soils are greater than unity. If such of soils are
encountered at a level where foundation would be resting, the complete layer should
be replaced by PCC or removed and back-filled with hydraulically compacted sand
fill compacted to a Procter Density as specified by the soil consultant.
If this strata is quite deep possibilities to be investigated to provide piles (driven/
bored) to a substantial depth below this strata and ignoring the stiffness effect of this
montmorillonite clay strata while calculating the equivalent springs for the piles.
The Poissons ratio of the soil is usually supplied in the soil report. This is required
for calculation of the soil springs used for dynamic analysis of the foundation. In
absence of such data = 0.4 would suffice for most of the cases.
The weight density of soil is usually furnished in the soil report this needs to be
divided by acceleration due to gravity (g) to arrive at the mass density.
or, = /g here, = mass density of soil; =unit weight of the soil, and g =
acceleration due to gravity @ 9.81m/sec
2
or 32.2 ft/sec
2
.
The Dynamic shear modulus plays a key role in evaluation of the spring data.
Though co-relation exists for theoretical evaluation of G from other engineering
parameters of the soil
18
for important foundations we still advocate that you insist on
field test to get the field observed value of G.
Try to convince the client
19
, it is worth spending a few thousand dollars now rather
than to pay through your nose in terms of performance compensations and could lead
to a classic case of being penny wise and pound foolish.
Designing a foundation with improper G value will completely waste the design
effort for the said foundation.
2.5.15.5 Trial sizing of the block foundation
Based on the above input data the next step for the designer is to do a trial sizing of
the block foundation with which he starts his first check for resonance and amplitude.
The basic guideline for the same could be summarised as follows:
The rigid type block foundation should be so proportioned that it should have
following mass ratio with respect to the machine
For centrifugal machine it should be 2 to 3 times the weight of the machine.
For Reciprocating type it should be 3 to 5 times the weight of the machine.
18 Refer Chapter 1 (Vol. 2) for these theoretical co-relations.
19 Even your boss at times. . . . . .
The top of foundation is usually kept about 300 mm above the finished grade
elevation to prevent damage due to surface water run-off.
However this should be back checked with process department to ensure that
NPSH of the pump or piping connections will not be affected adversely.
The vertical thickness of the foundation should be selected based on maximum
value of the following:
Maximum embedded length of the anchor bolts plus 250 mm
One fifth the width (least dimension) of the foundation in plan
One tenth of largest dimension in plan
A depth of 600 mm.
The width of the foundation is selected based on the maximum value of the
following:
Centre to centre distance of the anchor bolts plus 150 mm on both the side of
the foundation
Length to the edge of the machine plus 300 mm at the both the ends of the
foundation
1 to 1.5 times the vertical distance from the bottom of foundation to the
machine centre line
Once the width and height of foundation is selected the length can be
calculated based on the mass criteria as stated above.
The plan dimension of the machine should be so adjusted that c.g. of the machine
assembly matches with c.g. of the foundation.
For foundation resting on soil, eccentricity in c.g. of the machine and the
foundation shall not be more than 5%.
For large reciprocating machines the embedded depth to be so adjusted that at
least 60 to 80% of the depth of the foundation is embedded in the soil. This will
increase the lateral restraint and damping ratio for modes of vibration.
We now give below some useful data and mathematical expressions which could
effective in day to day design office practise for design of block foundations.
2.5.15.6 Centre of gravity of the machine foundation
Here the whole machine foundation is broken into different segments having mass
as m
i
having co-ordinates as x
i
, y
i
, z
i
respectively then the c.g. of the foundation is
given by
x =
i
m
i
x
i
i
m
i
, y =
i
m
i
y
i
i
m
i
,
and z =
i
m
i
z
i
i
m
i
Second moment of inertia of standard geometric shapes (Refer to Figures 2.5.35
and 36).
For foundations resting on soil the moment of inertia is calculated by
I
xx
=
1
12
LB
3
; I
yy
=
1
12
BL
3
, and I
zz
= I
xx
+I
yy
For foundations resting on piles or springs moment of inertia is calculated by
I
xx
=
i
y
2
i
; I
yy
=
i
x
2
i
, and I
zz
=
i
_
x
2
i
+y
2
i
_
e
Z
lx
ly
lz
X
Y
Figure 2.5.35 A solid rectangular prism.
Z
X
Y
Figure 2.5.36 A solid circular cylinders.
2.5.15.7 Mass of inertia of geometrical shapes
For solid rectangular prisms (Refer to Figure 2.5.36)
Jx = m/12(ly
2
+lz
2
)
Jy = m/12(lx
2
+lz
2
)
Jz = m/12(lx
2
+ly
2
)
For solid circular cylinders
The second moment of inertia is given by
Jx =
m
12
_
3
4
D
2
+l
2
_
; Jy =
m
8
D
2
; Jz =
m
12
_
3
4
D
2
+l
2
_
Here D = diameter of the cylinder; L = length of the cylinder.
2.6 SPECIAL PROVISIONS OF IS-CODE
We now give below some salient provisions and recommendations of IS-2974 for
rotary and reciprocating types of machines that constitute the normal design office
practice in India.
2.6.1 Recommendations on vibration isolation
To avoid transmission of vibration to adjoining parts of the buildings or other founda-
tions, it is necessary to provide a suitable isolation between the equipment foundation
and the adjoining structures.
This may be achieved by providing sand trench around the foundation block, the
thickness and depth of which shall be determined for each individual case.
As a rule the equipment foundation shall not be allowed to serve as a support for
other structures or for machines not related to the particular equipment.
In case it becomes necessary to support unimportant parts of other structures on the
machine foundation itself, measure shall be taken to make the connections resilient by
introducing gaskets made of rubber, cork, felt or other resilient materials.
2.6.2 Frequency separation
The natural frequency of the foundation systemshall be such as to avoid resonance with
operating frequency of the machine and the amplitudes be kept below the permissible
limit.
Foundations for low frequency machine shall preferably be designed as such that
the natural frequency of the foundation is higher then the operating frequency of the
machine.
The natural frequency of any foundation should not preferably be within 20% of
the operating speed of the machine.
2.6.3 Permissible amplitudes
Normally the recommendation of the vendor supplying the equipment shall guide the
design, however in absence of such data the code recommends that if no resonance
is to occur in adjoining structure the amplitude of vibrations of a foundation at the
upper edge shall not exceed 0.20 mm in both directions.
When several foundations for similar machines are erected on a common mat the
computation for vibration shall proceed assuming that each machine foundation is
independent of others by breaking up the raft into sections corresponding to separate
foundations.
The design value for the permissible amplitude of vibrations may be increased
by 30%.
2.6.4 Permissible stresses
Concrete of grade M15 or higher shall be used for foundations. Concrete and steel
stresses are as specified in IS: 456-2000 shall be used for considering the dynamic loads
separately in detailed design. The following elastic moduli of concrete may be used in
design.
Grade of concrete E
dyn
(kN/m
2
)
M15 250 10
6
M20 300 10
6
M25 340 10
6
M30 370 10
6
2.6.5 Concrete and its placing
The concrete used shall be controlled concrete conforming to design requirements.
The grade of concrete should generally be M15 to M20 for block foundation and M20
for frame foundation. The concrete shall be placed and designed in accordance with
IS: 456-2000.
The concrete used shall be of plastic consistency having an allowable slump, which
may vary between 50 to 80 mm. The water cement ratio shall not exceed 0.45. The
same consistency shall be maintained throughout the foundation.
2.6.6 Reinforcements
All foundation units of foundation shall be provided with top and bottom reinforce-
ment in two directions. Reinforcement shall be provided along the surface only in case
of block foundation.
The reinforcement in block foundation shall not be less than 25 kg/m
3
.
The minimumdiameter of bars shall be 12 mmwith a maximumspacing of 200 mm
in order to care of the shrinkage.
500
500
1442.5
1442.5
1442.5
1442.5
1 4 6 15 17
9 12
7 2
10
13
6770
A B C
E
D
11 14
3 5 8 16 18
O
800 4000 800 800 2400 800 425
PLAN VIEW OF THE BLOCK FOUNDATION
16100
C/L of Shaft of the machine
2000
1600
(typ.)
600
3600
800 800
ELEVATION OF THE BLOCK FOUNDATION
2405 3285 385
Figure 2.6.1 Plan and elevation of a gas turbine foundation.
2.6.7 Cover to concrete
For block foundation the concrete cover for protection of reinforcement shall be
75 mm at the bottom, 50 mm on both sides and 40 mm at top.
We now solve a practical design problem for a Gas Turbine resting on a block
foundation for your perusal and we hope that this will give you a better insight to
the aspect of how to apply the previously mentioned theories to the day-to-day design
office work of design of machine foundation (Figure 2.6.1).
Example 2.6.1
Design the gas turbine foundation shown in Fig. 2.6.1.
Design data
1 Bearing capacity of soil = 200 kN/m
2
2 Shear wave velocity of soil = 125 m/sec
3 Density of soil = 20 kN/m
3
4 Poissons ratio of soil = 0.25
5 C/L of shaft of the machine = 2.0 above T.O.C.
6 Operating frequency of turbine = 2250 r.p.m
7 Grade of concrete = M25
8 Grade of steel = Fe415
9 Allowable amplitude = 0.2 mm
10 Load at various anchor locations are as shown in the table hereafter.
Equipment load data at various Anchor Bolt locations
Static load Dynamic load
Vertical load Vertical Horizontal
Anchor bolt # (kN) load (kN) load (kN) Remarks
1 311
2 42 7 7
3 311
4 517
5 517
6 311
7 50 6.76 6.76
8 311
9 200
10 200 51 51 All horizontal force
is along global Y axes
11 200
12 350
13 350 23 23
14 350
15 185 Anchor Bolt for generator
Total 4760
Calculate the natural frequency and amplitude based on Figure 2.6.1 and
using
Richart and Lysmer model
Richart and Lysmer model with embedment
Wolfs model
Time history analysis based on Newmark-Beta method.
Solution:
Geometric property of the foundation
Area of foundation = 16.16.77 = 108.9 m
2
Second Moment of Inertia =
1
12
LB
3
=
1
12
16.16.77
3
= 416.30 m
4
Equivalent radius in vertical and horizontal mode:
r
0
=
_
108.9
= 5.89 m
Equivalent radius in rocking mode about the X-axis
r
=
4
_
4416.30
= 4.80 m
Calculation of soil springs
Shear Wave velocity = 125 m/sec; Poissons ratio = 0.25.
Unit weight of soil = 20 kN/m
3
Dynamic shear modulus (G) = Vs
2
=
20
9.81
(125)
2
= 31855.25 kN/m
2
Springs based on Richarts model
K
z
=
4Gr
0
(1 )
=
431855.25 5.89
(1 0.25)
= 1000679.6 kN/m
K
y
=
32Gr
0
(1 )
(7 8)
=
32(1 0.25) 31855.25 5.89
(7 80.25)
= 900611.63 kN/m
K
y
=
8Gr
3
3(1 )
=
831855.25 (4.8)
3
3(1 0.25)
= 12511907.52 kN/m.
Table for calculation of c.g. and second moment of inertia of m/c & fdn.
(explained in next page)
Centre of gravity
x =
11291.04
1476.4
= 7.65 m from the point O; y =
4997.63
1476.4
= 3.39 m from the
point O; and z =
3722.95
1476.4
= 2.52 m from the bottom of the foundation.
Eccentricity in x direction =
8.05 7.65
16.1
100 = 2.5% < 5% hence OK.
Eccentricity in y direction =
3.399 3.385
6.77
100 = 0.07% < 5%hence OK.
Total mass moment of inertia =
m
12
_
l
2
y
+l
2
z
_
+m(y
2
oi
+z
2
oi
) +m z
2
= 4703.00 +4509.65 +1476.4 (2.52)
2
= 18588.4 kN-m-sec
2
.
2
1
8
D
y
n
a
m
i
c
s
o
f
S
t
r
u
c
t
u
r
e
a
n
d
F
o
u
n
d
a
t
i
o
n
:
2
.
A
p
p
l
i
c
a
t
i
o
n
s
Loading L
x
L
y
L
x
Weigt m/l2(Ly
2
+ m(y
2
oi
+
location (m) (m) (m) ht(kN) Mass X
i
Y
i
Z
i
m
i
x
i
m
i
y
i
m
i
z
i
Lz
2
) y
oi
z
oi
z
oi
)
1 311 31.70 0.385 6.27 3.6 12.21 198.77 114.13 0.00 2.89 1.078 300.73
2 42 4.28 0.385 3.385 3.6 1.65 14.49 15.41 0.00 0 1.078 4.98
3 311 31.70 0.385 0.5 3.6 12.21 15.85 114.13 0.00 2.885 1.078 300.73
4 517 52.70 3.67 6.27 3.6 193.41 330.44 189.72 0.00 2.89 1.078 499.93
5 517 52.70 3.65 0.5 3.6 192.36 26.35 189.72 0.00 2.885 1.078 499.93
6 311 31.70 6.075 6.27 3.6 192.59 198.77 114.13 0.00 2.89 1.078 300.73
7 50 5.10 6.075 3.385 3.6 30.96 17.25 18.35 0.00 0 1.078 5.93
8 311 31.70 6.075 0.5 3.6 192.59 15.85 114.13 0.00 2.885 1.078 300.73
9 200 20.39 7.3 4.828 4.6 148.83 98.42 93.78 0.00 1.44 2.078 130.49
10 200 20.39 7.3 3.385 4.6 148.83 69.01 93.78 0.00 0 2.078 88.07
11 200 20.39 7.3 1.943 4.6 148.83 39.60 93.78 0.00 1.443 2.078 130.49
12 350 35.68 9.7 4.828 4.6 346.08 172.23 164.12 0.00 1.44 2.078 228.35
13 350 35.68 9.7 3.385 4.6 346.08 120.77 164.12 0.00 0 2.078 154.11
14 350 35.68 9.7 1.943 4.6 346.08 69.30 164.12 0.00 1.443 2.078 228.35
15 185 18.86 11.3 6.27 3.6 213.10 118.24 67.89 0.00 2.89 1.078 178.89
16 185 18.86 11.3 0.5 3.6 213.10 9.43 67.89 0.00 2.885 1.078 178.89
17 185 18.86 15.3 6.27 3.6 288.53 118.24 67.89 0.00 2.89 1.078 178.89
18 185 18.86 15.3 0.5 3.6 288.53 9.43 67.89 0.00 2.885 1.078 178.89
A 6.5 6.77 3.6 3960.5 403.72 3.25 3.385 1.8 1312.08 1366.58 726.69 1977.97 0 0.722 210.24
B 4 6.77 3 2031 207.03 8.5 3.385 1.6 1759.79 700.81 331.25 946.02 0 0.922 175.86
C 5.6 6.77 3.6 3412.1 347.82 13.3 3.385 1.8 4625.96 1177.36 626.07 1704.10 0 0.722 181.13
D 0.8 5 1.6 160 16.31 7.3 3.385 3.8 119.06 55.21 61.98 37.46 0 1.278 26.65
E 0.8 5 1.6 160 16.31 9.7 3.385 3.8 158.21 55.21 61.98 37.46 0 1.278 26.65
14484 1476.40 11291.04 4997.63 3722.95 4703 4509.65
Calculation for damping
Based on Richarts Formula we have
In the vertical direction
B
z
=
0.25 m(1 )g
s
r
3
z
=
0.251476.40.75
2.03 (5.89)
3
= 0.667
D
z
=
0.425
_
B
z
=
0.425
0.667
= 0.52
C
z
= 2D
z
_
K
z
m = 20.52
1000679.61476.4 = 39974 kN sec/m

In the horizontal direction
B
y
=
(7 8) mg
32(1 )
s
r
3
y
=
5
320.75

1476.4
2.04 (5.89)
3
= 0.738
D
y
=
0.288
_
B
y
=
0.288
0.738
= 0.3352 and
C
y
= 2D
y
_
K
y
m = 20.3352
900611.631476.4 = 24446 kN sec/m

For the rocking mode
B
y
=
0.375(1 )J
y
g
s
r
5
y
=
0.3750.7518588.4
2.03 (4.798)
5
= 1.013
D
y
=
0.15
(1 +B
y
)
_
B
y
=
0.15
2.013
1.013
= 0.074
C
y
= 2D
y
_
K
y
m = 20.074
12511907.5218588.4 = 71375 kN sec/m

Calculation of natural frequencies
In the vertical direction
z
=
_
K
z
m
=
_
100679.6
1476.4
= 26 rad/sec (249 r.p.m)
In the horizontal direction the equation of motion for free vibration is given by
_
m 0
0 J
y
_ _
y
_
+
_
C
x
C
y
Z
c
C
y
Z
c
C
y
+C
y
Z
2
c
WZ
c
_
_
y
_
+
_
K
y
K
y
Z
c
K
y
Z
c
K
y
+K
y
Z
2
c
WZ
c
_ _
y
_
=
_
0
0
_
For eigen value analysis we have
_
K
y
m K
y
Z
c
K
y
Z
c
K
y
+K
y
Z
2
c
WZ
c
J
_
= 0
Here m = 1476.4; J
y
= 18588.4; Z
c
= 2.52 m; W = 14484 kN; K
y
=
900611.63 kN/m; K
y
= 12511907.52 kN/m
or,
_
900611 1476.4 2269540
2269540 18194648 18588.4
_
= 0
The above matrix on expansion and simplification reduces to
2
1589 +409398 = 0;
=
1589
_
(1589)
2
41409398
2
= 323.5, 1265

2
= 17.98 rad/sec (172 r.p.m.); and
3
= 35.56 rad/sec (340 r.p.m.)
Calculation of eigen vectors
For first mode, for = 17.98 rad/sec, we have
_
900611 1476.4323.5 2269540
2269540 18194648 18588.4323.5
_ _
11
12
_
= 0
The above on expansion gives the following two equations
422995.6
11
2269540
12
= 0 and 2269540
11
12181301
12
= 0
Considering
11
= 1.00 and solving the above homogenous equation we have,
12
= 0.186379442
Thus,
_
11
12
_
=
_
1.00
0.186379442
_
For the Second mode we have
_
900611 1476.41265 2269540
2269540 18194648 18588.41265
_ _
21
22
_
= 0
The above on expansion gives the following two equations:
967035
21
2269540
22
= 0 and 2269540
21
5319678
22
= 0
Considering
21
= 1.00 and solving the above homogenous equation we have
22
= 0.426092952
Thus the complete eigen vector matrix is given by
[] =
_
1.00 1.00
0.186379442 0.426092952
_
Calculation of normalised eigen vectors
For the first mode
{}
T
[M] {} = 1.00 0.186379442
_
1476.4 0.0
0.0 18588.4
_ _
1.00
0.186379442
_
The above on simplification gives,

M
r
=

2122.110 = 46.06
Thus dividing each term of the eigen vector by above we have,
_
11
12
_
N
=
_
0.021707808
4.04588918410
3
_
.
{}
T
[M] {} = 1.00 0.426092952
_
1476.4 0.0
0.0 18588.4
_ _
1.00
0.426092952
_
The above on simplification gives,

M
r
=

4851.22 = 69.65.
Thus dividing each term of the eigen vector by above we have,
_
21
22
_
N
=
_
0.014357356
6.11756837 10
3
_
Thus the complete normalised eigen vector matrix is
[] =
_
21.707808 14.357356
4.045889184 6.11756837
_
10
3
Correction of damping matrix based on Rayleigh coeff icient
for modal analysis
We had already stated that damping matrix obtained from soil property is
non proportional and when considered in the analysis will not de-couple under
orthogonal transformation as such we correct the matrix enabling us to de-couple
the same
20
.
20 Based on the theory of magnification factor damping may be ignored for this case for the ratio
of the fundamental frequencies of the foundation to the operating frequency of the machine is
more than 3.5. However for sake of clarification of the problem we continue to consider it in our
analysis.
Here,
[C] =
_
C
x
C
y
Z
c
C
y
Z
c
C
y
+C
y
Z
2
c
WZ
c
_
=
_
24446 61604
61604 190117.2
_
For the First mode we have
{}
T
[C] {} = 21.707808 4.045889184 10
3
_
24446 61604
61604 190117.2
_
_
21.707808
4.045889184
_
10
3
The above on simplication gives, {}
T
[C] {} = 3.8106 2
1
1
=3.8106 or
1
= 0.105.
For the Second mode, we have
{}
T
[C] {} = 14.357356 6.11756837 10
3
_
24446 61604
61604 190117.2
_
_
14.357356
6.11756837
_
10
3
The above on simplification gives
{}
T
[C] {} = 22.97 2D
2
2
= 22.97 or D
2
= 0.323.
Now considering the design damping as proportional Rayleigh damping,
we have
[C] = [M] + [K] or []
T
[C] [] = []
T
[M] [] + []
T
[K] []
and we have, 2D
1
1
= +
2
1
and 2D
2
2
= +
2
2
.
i.e. +323 = 3.8106 and +1265 = 22.97
Solving the above two simultaneous equations, we have: = 2.7589 and
= 0.0203
Substituting the above value of and we have,
[C] = 2.7589
_
1476.4 0
0 18588.4
_
+0.0203
_
900611 2269540
2269540 18194648
_
[C] =
_
14222 46071
46071 318233
_
.
Calculation of amplitude in the vertical mode
Vertical force = (7 +6.76 +51 +23) = 87.76sin 236t
We had calculated earlier that D
z
= 0.52 and hence,
z
=
P
0
sin
m
t/K
z
_
(1 r
2
)
2
+(2 r)
2
, with, r =
236
26
= 9.0.

z
=
87.76sin 236t
1000679.6
_
(1 81)
2
+(2 0.52 9)
2
= 1.08822 10
6
sin 236t m.
For, coupled sliding and rocking mode, we have
Lever arm = 3.6 + 2.0 2.52 = 3.08 m; Horizontal force = 87.76 sin 236t
Thus moment about the vertical centroid, Z
c
= 270sin 236t
The force matrix can be represented as, {P} =
_
87.76
270
_
sin 236t.
The equation of motion can be written as
[M]
_
Y
_
+[C]
_
Y
_
+[K] {Y} = {P}
With orthogonal transformation, we can write
[]
T
[M] []
_
y
_
+[]
T
[C] []
_
y
_
+[]
T
[K] [] {Y} = []
T
{P}
which gives the following two equations
y +2D
1
1
y +
2
1
y = psin
m
t and

+2D
2
+
2
2
= msin
m
t
Here []
T
{P} =
_
21.707808 4.045889184
14.357356 6.11756837
_ _
88
270
_
sin 236t
_
p
m
_
=
_
3.0
0.388
_
sin 236t
i.e. y + 20.10517.98 y + 324y = 3.0sin 236t y + 3.776 y + 324y =
3.0sin 236t
and

+ 20.32335.56
+ 1265 = 0.388sin 236t

+ 22.97
+
1265 = 0.388sin 236t
Thus for the horizontal translation, we have y+3.776 y+324y = 3.0sin 236t
where r =
236
324
= 13.11 and =
3.776
2
324
= 0.104.
Hence,
y
=
3 sin 236t
324
_
(1171.8)
2
+(2 0.104 13.11)
2
= 5.4204 10
5
sin 236t
For the rocking mode
+22.97
+1265 = 0.388 sin 236t

where r =
236
1265
= 6.63 and =
22.97
2
1265
= 0.322
=
0.388 sin 236t
1265
_
(1 43.95)
2
+(20.3226.63)
2
= 7.106310
6
sin 236t
Thus in global co-ordinate, we have
_
Y
_
=
_
21.707808 14.357356
4.045889184 6.11756837
_ _
5.4204
0.7106
_
10
8
=
_
107.462
26.27
_
10
8
sin 236t
Net horizontal amplitude at top of foundation
Y=107.46310
8
+(3.6 2.52) 26.27 10
8
=1.35810
6
m < 0.2 mm OK
Net horizontal amplitude at base of the foundation
Y = 107.46310
8
2.5226.27 10
8
= 41.26310
8
m < 0.2 mm OK
We make here a very interesting comparison, shown in Figure 2.6.2, is the time
history response of the block foundation with non-proportional soil damping
and corrected proportional Rayleigh damping, we have obtained earlier.
It will be observed that values are quite closely matching and for practical
engineering work this is deemed sufficient.
Comparison of amplitude based on time history
-0.000015
-0.00001
-0.000005
0
0.000005
0.00001
0.000015
1 23 45 67 89 111 133 155 177 199 221 243 265
Time steps
A
m
p
l
i
t
u
d
e
Displacement with non
proportional damping
Displacement with corrected
proportional damping
Figure 2.6.2
Check of local vibration of the pedestals
Width of the pedestal = 800 mm; Depth of the pedestal = 5000 mm; Height
of pedestal = 1600 mm.
I =
BD
3
12
=
0.8125
12
= 8.33 m
4
Considering the pedestal as cantilever beam,
K
h
=
3EI
L
3
=
330010
6
8.33
(1.6)
3
= 1.8310
9
kN/m
Self weight of the pedestal = 0.8 5.0 1.6 25 = 160 kN; Weight from
machine = 3 200 +3 350 = 1650 kN
Total weight = 1810 kN
Thus total mass(m) = W/g = 184.5 kN-sec
2
/m and hence =
_
K
h
m
=
_
1.8310
9
184.5
= 3149 rad/sec
And, r =

m
n
=
236
3149
= 0.07.
As the frequency ratio is very low we neglect the damping we have,
y =
P
0
/k
(1 r
2
)
=
51sin 236t
1.8310
9
(1 0.07
2
)
= 2.8010
8
m < 0.2 mm.
Calculation based on embedment effect
We had already calculated before that based on property of soil individual soil
stiffness in various modes as
K
z
= 1000679.6 kN/m,C
z
= 39974kN.sec/m; K
y
= 900611.63kN/m,
and C
y
= 24446kN/m, K
y
= 12511907.52 kN/m, C
y
= 71375 kN/m.
Embedment factor based on Richarts table for soil stiffness
z
= 1 + 0.6(1 )
h
r
z
, here h = 3.0 m as per the problem, and r
z
=5.89 m
which gives,
z
= 1.23 similarly,
y
= 1 +0.55(2 )
h
r
y
and this gives
y
= 1.497 and
y
= 1 +1.2(1 )
h
r
y
+0.2(2 )
_
h
r
y
_
3
and this gives
y
= 1.5625.
Embedment factor based on Richarts table for soil damping
z
=
1 +1.9(1 )
h
r
z
z
= 1.556;
y
=
1 +1.9(2 )
h
r
y
y
= 2.20 and
y
=
1 +0.7(1 )
h
r
y
+0.6(2 )
_
h
r
y
_
3
y
= 1.267.
Thus considering the embedment factor the stiffness and damping value gets
modified to:
K
e
z
= 1230836 kN/m, C
e
z
= 62204 kN sec/m
K
e
y
= 1348214 kN/m, C
e
y
= 53812 kN/m
and K
e
y
= 19549856 kN/m, C
e
y
= 90432 kN/m
For the vertical direction, we have
z
=
_
K
e
z
m
=
_
1230836
1476.4
= 29 rad/sec (277 r.p.m);
We had calculated earlier that D
z
= 0.73;
Thus
z
=
P
0
sin
m
t/K
z
_
(1 r
2
)
2
+(2 r)
2
, where r =
236
29
= 8.13
z
=
87.76sin 236t
1230836
_
(1 66)
2
+(20.528.13)
2
= 1.0775410
6
sin 236t m
In horizontal direction the equation of motion for free vibration is given by:
_
m 0
0 J
y
_ _
y
_
+
_
C
x
C
y
Z
c
C
y
Z
c
C
y
+C
y
Z
2
c
WZ
c
_ _
y
_
+
_
K
y
K
y
Z
c
K
y
Z
c
K
y
+K
y
Z
2
c
WZ
c
_ _
y
_
=
_
0
0
_
_
K
e
y
m K
e
y
Z
c
K
e
y
Z
c
K
y
+K
e
y
Z
2
c
WZ
c
J
_
= 0
Here m = 1476.4; J
y
= 18588.4; Z
c
= 2.52 m; W = 14484 kN; K
e
y
=
1348214 kN/m; K
e
y
= 19549856 kN/m.
_
1348214 1476.4 3397499
3397499 28075055 18588.4
_
= 0
2
2423 +958897 = 0 = 498, 1925

2
= 22.3 rad/sec (213 r.p.m.), and
3
= 43.87 rad/sec (419 r.p.m.).
Now proceeding in the exact manner as explained in the previous case, we
arrive at the result:
Net horizontal amplitude at top of foundation
Y = 107.1410
8
+(3.6 2.52) 26.0910
8
= 1.35310
6
m < 0.2 mm OK
Net horizontal amplitude at the base of foundation
Y = 107.1410
8
2.5226.0910
8
= 41.410
8
m < 0.2 mm OK
Calculation based on Wolfs Model
Based on formulation proposed by Wolf
K
z
=
4Gr
0
(1 )
=
431855.255.89
(1 0.25)
= 1000679.6
C
z
=
r
z
Vs
K
z
0
=
5.891000679.60.58
125
= 27348 k sec/m
m
z
=
_
r
z
Vs
_
2
K
z
0
=
_
5.89
125
_
2
1000679.60.095 = 211 k-sec
2
/m
In the horizontal direction we have
K
y
=
8Gr
y
1
=
831855.255.89
0.75
= 2001359 k/m
C
y
=
r
y
Vs
K
y
0
=
5.8920013590.85
125
= 80158 k sec/m
m
y
=
_
r
y
Vs
_
2
K
y
0
=
_
5.89
125
_
2
20013590.27 = 1200 k-sec
2
/m
In rocking mode, we have
K
y
=
8Gr
3
3(1 )
=
831855.25 (4.8)
3
3(1 0.25)
= 12511907.52 k/m
J
y
=
_
r
y
Vs
_
2
K
y
0
=
_
4.8
125
_
2
12511907 0.24 = 4428 k-sec
2
/m
0
=
0.3
1 +
3(1)m
8r
5
=
0.3
1 +
30.754428
8(4.8)
5
2.04
= 0.242
Thus, C
y
=
r
y
Vs
K
y
0
=
4.812511907 0.242
125
= 116271 k sec/m
For vertical direction we have
z
=
_
K
z
m
=
_
1000679
1687
= 24.35 rad/sec (277 r.p.m)
Damping Ratio D
z
=
27348
2
10006791687
= 0.332
Thus
z
=
P
0
sin
m
t/K
z
_
(1 r
2
)
2
+(2Dr)
2
where r =
236
24.34
= 9.7

z
=
87.76sin 236t
1000679
_
(1 94)
2
+(20.3329.7)
2
= 0.9407610
6
sin 236t m
For coupled sliding and rocking mode
_
m 0
0 J
y
_ _
y
_
+
_
C
x
C
y
Z
c
C
y
Z
c
C
y
+C
y
Z
2
c
WZ
c
_ _
y
_
+
_
K
y
K
y
Z
c
K
y
Z
c
K
y
+K
y
Z
2
c
WZ
c
_ _
y
_
=
_
0
0
_
For eigen value analysis we
_
K
e
y
m K
e
y
Z
c
K
e
y
Z
c
K
y
+K
e
y
Z
2
c
WZ
c
J
_
= 0
Here m = 2676; J
y
= 23016; Z
c
= 2.52 m; W = 14484 kN; K
e
y
= 2001359
kN/m; K
e
y
= 12511907 kN/m.
It is to be noted that here mass and moment of inertia is the mass/inertia
of machine and foundation plus the mass/inertia of soil participating in the
vibration.
Thus
_
2001359 2676 5043425
5043425 25184838 23016
_
= 0
2
1842 +405381 = 0 = 255, 1587
Hence,
2
= 15.96 rad/sec (152 r.p.m.) and
3
= 39.83 rad/sec (380 r.p.m.)
Now proceeding in the manner as explained in the case of Richarts model we
arrive at the result;
Net horizontal amplitude at top of the foundation
Y = 6310
8
+(3.6 2.52) 37.64610
8
= 1.036510
6
m < 0.2 mm OK
Net horizontal amplitude at the base of foundation
Y = 6310
8
2.5237.64610
8
= 31.8610
8
m < 0.2 mm OK
The values obtained by the three methods are summarised hereafter:
Comparison of natural frequencies
Based on the method
z
(r.p.m)
y
(r.p.m)
y
(r.p.m)
Richarts formula 249 172 340
Richart with embeddment 277 213 419
Wolfs method 277 152 380
Comparison of amplitude
Based on the method
z
(mm)
y
(mm)
y
Richarts formula 1.0882 10
3
1.358 10
3
41.263 10
8
Richart with embeddment 1.077 10
3
1.353 10
3
41.4 10
8
Wolfs method 0.94076 10
3
1.0365 10
3
31.86 10
8
Based on Time History Analysis
We perform time history analysis for springs based on Richarts method and
Wolfs Method.
Here time history response has been done for 215 steps with complete soil
damping into consideration and shown in Figures 2.6.3 and 4.
Time History based on Newmark Beta Method with Richarts Spring
-0.000006
-0.000004
-0.000002
0
0.000002
0.000004
0.000006
0.000008
1 17 33 49 65 81 97 113 129 145 161 177 193 209
Time Steps
A
M
P
L
I
T
U
D
E
Displacement in Y direction
Angular Rotation
Figure 2.6.3
Time History based on Newmark Beta Method with Wolf's Spring
Spring
-0.000003
-0.000002
-0.000001
0
0.000001
0.000002
0.000003
1 20 39 58 77 96 115 134 153 172 191 210
Time Steps
A
M
P
L
I
T
U
D
E
Displacement in Y direction
Angular Rotation
Figure 2.6.4
2.7 ANALYSIS AND DESIGN OF MACHINE FOUNDATION
UNDER IMPACT LOADING
2.7.1 Introduction
In this section we will deal with foundations subjected to impact loading. These type
of foundations usually constitute of hammer foundations used for forging or hydraulic
stamps used to flatten steel billets to make plates out of them.
The arrangement of the hammer foundation is usually as shown Figure 2.7.1.
The hammer foundation, consists of a hammer or a tup which falls repeatedly on
an anvil. The anvil in turn is placed on an elastic pad resting on a massive RCC block.
The elastic pad is used to isolate the foundation from the surrounding and minimize
the harmful effect of the vibration induced by the hammer dropping on the anvil. The
elastic pad also acts as damper to reduce the net amplitude of vibration of the anvil
and the foundation.
Depending upon the functionality, the frame of the hammer may either rest on the
foundation block as shown above or may even rest on a separate foundation.
While planning the foundation it is usually ensured that the center line of the anvil
is concentric with the center of gravity of the base of the foundation. This ensures that
the amplitude of vibration is restricted to vertical translation only and does not give
rise to any coupled motion including rocking
21
.
At times when the hammer is very heavy the foundation is further isolated by
providing elastic pad/springs along with dampers below the RCC block too.
Shown in Figure 2.7.2 is a hammer foundation where other than the anvil the RCC
block is also mounted on springs and dampers to isolate the transmittal of vibration
to the surrounding.
The springs or the elastic pad which are placed below the RCC block is usually
an expensive item and care should be taken to protect them from exposure to water,
chemicals, oils etc which could otherwise damage their properties. This is usually done
by providing a protective RCC trough all round the foundation and sealing the same
at the top of the foundation level.
The elastic material used under the anvil or the RCC block could be of cork, tim-
ber or even specialized mechanical springs and dampers supplied by vendors having
technological expertise in isolation techniques of these type of foundations.
2.7.1.1 How does the behavior of a mechanical system
under impact differ from externally applied
harmonic loads?
We do not tender any apologies for posing so fundamental a question, for in our
experience in teaching this subject, as well as interacting with professionals in the
industry for over two decades, we have been somewhat startled to find that though
people can arrive at the design values for various type of machine foundation quite
accurately by following the code stipulation blindly, but how does the characteristics
21 Hammer foundation having eccentric anvil is though uncommon but surely not rare. We will deal with
this particular case separately later.
Frame
Hammer/Tup
RCC Fdn.
Anvil
Elastic Pad
F.G.L.
Figure 2.7.1 General arrangement of a Typical Hammer Foundation.
Frame
Hammer/Tup
RCC Fdn.
RCC Trough
Anvil
Elastic Pads/Springs
Figure 2.7.2 General arrangement of a Typical Hammer Foundation mounted on spring with R.C.C.
Trough.
of the behavior differs in the above two cases. . . the picture has remained gray to
many. So before we take a plunge into the mathematical aspect of the design it would
be worthwhile to understand the conceptual aspect of the problem and reflect a bit on
how the two cases differ in transmitting the vibration to the system.
We hypothesize two pictures for this from our day to day life.
1 Imagine a boy continuously jumping on a plank supported at two ends for some
time.
2 A Karatika giving a vicious chop to the same plank at some point on it
22
.
2.7.1.2 How do the planks behave under these two conditions?
For the first case if we have a stop watch we can measure the time taken by the boy
when he is at rest on the plank to the time he jumps (presuming with same monotony)
and again comes to rest on the plank. If we take this as his time period of vibration
T it is possible to find out the frequency of his motion from the relation, = 1/T.
If we now measure the weight of the boy we can say that the plank is subjected to
a continuous external force of W sin t, where W is the weight of the boy and
is the frequency with which he is jumping on the plank. Thus the plank is under a
forced harmonic load and will also produce amplitudes which will be a function of
the external force expressed as W sin t. This is called harmonic force
23
.
While in the second case when the Karatika executes the chop he is transferring his
potential energy into a kinetic energy and is transferring this energy to the plank in a
very short period of time (may be some small fraction of a second) and then it ceases
to exist. This is quite unlike the earlier case when the external force continues to excite
the plank till the boy continues to jump on it.
These type of forces when induced on a body where it is subjected to force for a
very short time is known as an impact load or in technical term we call it a transient.
We consciously or otherwise often observe this phenomenon quit often in our day
to day life like
A hammer used to put a nail in place
A mallet used to hit a golf ball
A ship hitting the jetty fenders when coming to rest on a port
These are all cases of impact forces acting on a system.
For instance, if we take the case of a mallet hitting the golf ball what we do is take
a swing up when we concentrate our potential energy and with the down swing of the
club we transfer the potential energy to kinetic energy which is then transferred to the
ball at the instant of impact.
Now suppose we connect the ball to a spring, we will observe that the ball starts
vibrating to and fro with respect to its mean position.
22 The two cases are mutually exclusive.
23 Block Foundation dealt earlier is a typical example of this when the unbalanced mass induces an external
force that is harmonic in nature.
To understand the effect of impact further we a take a step backward and formulate
a problem from our days of engineering mechanics/School Physics as hereafter.
Example 2.7.1
Shown in Figure 2.7.3 is a metal block of weight 100 kN suspended from a
point O by a mass less inextensible string having a length of 2.5m. It is released
from rest from a position 90 degree to vertical position of rest as shown below.
The block of 100 kN hits another metal block of weight 500 KN connected to a
spring of stiffness 2500 kN/mat point X. Considering the collision to be perfectly
elastic find out the amplitude of vibration of the body considering friction less
surface having
O
W
1
= 100 kN
2500 mm
X W
2
= 500 kN
K = 2500 kN/m
Figure 2.7.3 Conceptual diagram of the system.
Un-damped motion.
Damped Motion having a damper connected to W
2
of magnitude 125 kN
sec/m
Solution:
When the body is released form its position of rest it takes a swing and hits the
500 kN body at point X.
The potential energy of the 100 kN body at its initial position = W
1
h
Kinetic energy of the 100 kN body at the point of impact =
1
2
_
W
1
g
_
u
2
1
Applying the law of conservation of energy i.e. KE = PE,
we have, u
1
=
_
2gh m/sec: For h = 2.5 m u
1
= 7m/ sec
Now let,
u
1
= Initial velocity of the body W
1
@ 7.0 m/sec before collision;
u
2
= Initial velocity of the body W
2
@ 0.0 m/sec before collision;
v
1
= Final velocity of the body W
1
after collision, and
v
2
= Final velocity of the body W
2
after collision.
Then based on conservation of momentum and the collision being elastic we
have,
W
1
g
u
1
+
W
2
g
u
2
=
W
1
g
v
1
+
W
2
g
v
2
or, m
1
u
1
+m
2
u
2
= m
1
v
1
+m
2
v
2
and
1
2
m
1
u
2
1
+
1
2
m
2
u
2
2
=
1
2
m
1
v
2
1
+
1
2
m
2
v
2
2
Based on the above boundary conditions, we have
m
1
u
1
= m
1
v
1
+m
2
v
2
and m
1
u
2
1
= m
1
v
2
1
+m
2
v
2
2
Substituting the numerical data mentioned in the problem, one can have
100v
1
+500v
2
= 700 and 100v
2
1
+500v
2
2
= 4900
Solving the above two equations we have ; v
1
= 3.90 m/ sec and v
2
=
2.18 m / sec.
Here the negative value for v
1
means that the 100 kN body will rebound back
with a velocity of 3.9 m/sec.
Now applying DAlemberts equation to the body connected to the spring we
have
m x +Kx = 0 where K = spring stiffness.
m
dv
dt
+Kx = 0, where v = velocity vector of the body
i.e. m
dv
dx
dx
dt
+Kx = 0, as
dx
dt
= v we have, mvdv +Kxdx = 0
The above differential equation has boundary condition as at v = v
0
, x = 0
and at v = 0, x =
m
0
_
v
0
vdv +
_
0
Kxdx = 0, this on simplification gives, = v
_
m
K
.
Also, =
v
where = the natural frequency of the structure.

Thus, =
_
K
m
=
_
25009.81
500
= 7 rad/ sec
The amplitude, =
v
=
2.18
7
= 0.3114 m.
It is to be noted that from the above calculation we have managed to find out
only the magnitude of the maximum amplitude.
It does not tell us how the body will vibrate under this impact force.
To get this history let us consider the differential equation
m x +Kx = 0; and let, x = C
1
sin t +C
2
cos t
be the solution to the above.
Applying the boundary condition at t = 0 x = 0 and v = 2.18 m/sec, we have
At t = 0 x = 0 C
2
= 0 or x = C
1
sin t and x = C
1
cos t at t =
0; C
1
=
x
=
v
0
from which we deduce,

x =
v
0
sin t x = 0.3114sin 7t
The above when plotted at time step of 0.05 seconds shows a curve as furnished
in Figure 2.7.4.
With damped vibration for single degree of freedom, we have seen earlier in
Chapter 3 (Vol. 1) that amplitude of vibration is given by
x = e
D
n
t
[C
1
cos
d
t +C
2
sin
d
t]
where,
d
=
n
_
(1 D
2
) and D = c/c
c
and c
c
= 2
km.
Displacement history under initial velocity v
0
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
16
1
1
1
6
2
1
2
6
3
1
3
6
4
1
4
6
5
1
5
6
6
1
6
6
Time Steps
A
m
p
l
i
t
u
d
e

(
m
e
t
e
r
)
Amplitude (meter)
Figure 2.7.4 Displacement history undamped case.
Now, for t = 0 when x = 0 C
1
= 0 which reduces the above equation to
x = C
2
e
D
n
t
sin
d
t
Again for t = 0, x = v
0
we have, C
2
= v
0
/
d
which results in the equation,
x =
v
0
n
_
(1 D
2
)
e
D
n
t
sin
d
t
Here we have, v
0
= 2.18 m/ sec,
n
= 7 rad/ sec, D = 0.175.
Substituting the above values, we have x = 0.316e
1.225t
sin 6.892t.
Plotting the above values at time step of 0.05 sec we see the time history curve
is as given in Figure 2.7.5.
Damped displacement under initial velocity v
0
-0.2
-0.1
0
0.1
0.2
0.3
159
1
3
1
7
2
1
2
5
2
9
3
3
3
7
4
1
4
5
4
9
5
3
5
7
6
1
6
5
Time Steps
A
m
p
l
i
t
u
d
e

(
m
e
t
e
r
)
Amplitude (m)
Figure 2.7.5 Displacement history damped case.
The above response shows some very interesting results. While for un-damped
motion the curve follows a sinusoidal pattern, for damped case it initially starts with
peak amplitude and quickly dies down due to the inherent damping in the system
in contrary to the harmonic loading, where the body continues to vibrate under the
application of the externally applied force.
So far so good, we have managed to arrive at the behavior pattern of a systemhaving
a single degree of freedom subjected to impact load albeit some idealization such as
The string is inextensible and mass less
The collision is perfectly elastic
24
The spring and damper is mass less having identified definite values.
24 There is no collision in nature that is perfectly elastic for some energy is always dissipated out in form
of heat or sound thus we usually use a term co-efficient of restitution. We will learn more about it
subsequently.
Now the question boils down to how does the above problem relates to a hammer
foundation which we are supposed to discuss herein?
To explain this, we need to clarify how does a hammer foundation work?
Based on the General arrangement of hammer foundation shown earlier, the ham-
mer or the tup either undergoes a free fall on the anvil or falls under a certain pressure
(for double acting hammers). It either attens or forge the metal on the anvil to a desired
shape or may even crush it to lower particle size (in case of a crusher) depending upon
for what purpose the machine is being put to use.
Irrespective of its function, the basic point that remains unaltered is the following:
In contrary to the foundation supporting centrifugal or reciprocating type of
machines where the foundation is subjected to a constant external harmonic force
the hammer foundation induces a transient force at the point of collision and then
ceases to exist till the next blow is induced
25
. Thus based on the above statement we
can postulate that for design of machine foundations of this type we need to analyze
the system subject to transient shocks. Hence as a first step let us see what type of
mathematical model is in vogue for analysis of these types of foundations.
2.7.2 Mathematical model of a hammer foundation
For foundations resting on ground supporting anvils mounted on elastic base we
usually consider a system having two degrees of freedom as shown hereafter.
Shown in Figure 2.7.6 is the mathematical model of a hammer foundation resting
on soil where,
m = mass of the hammer or the tup;
m
1
= mass of the foundation block plus frame resting on it if any;
m
2
= mass of the anvil resting on elastic pad;
k
1
= soil spring to be calculated from Barkan or Richarts formula
26
;
k
2
= spring value for the elastic pad, this value is normally furnished by the vendor
supplying these pads;
c
1
= damping of the soil to be obtained from Richarts formula;
c
2
= damping of the elastic pad again furnished by the vendor;
H = height of the free fall of the hammer;
x
2
= amplitude vector of the anvil; and
x
1
= amplitude vector of the Foundation.
We had already seen in earlier
27
that for bodies having two degrees of freedom the
free equation of vibration is given by
_
m
1
0
0 m
2
_ _
x
1
x
2
_
+
_
k
1
+k
2
k
2
k
2
k
2
_ _
x
1
x
2
_
= 0, (2.7.1)
25 We hope by now the reader can smell the congruence with the worked out example 2.7.1.
26 Refer to section of block foundation for the formula of the springs and dampers.
27 Refer Chapter 5 (Vol. 1) on basic concepts in Structural Dynamics.
m
H
m
2
x
2
k
2
c
2
c
1
x
1
k
1
m
1
Figure 2.7.6 Mathematical model of hammer foundation resting on soil directly.
and since this is a statically coupled equation the damped free vibration of motion is
given by (Meirovitch 1975)
_
m
1
0
0 m
2
_ _
x
1
x
2
_
+
_
c
1
+c
2
c
2
c
2
c
2
_ _
x
1
x
2
_
+
_
k
1
+k
2
k
2
k
2
k
2
_ _
x
1
x
2
_
= 0 (2.7.2)
2.7.2.1 Mathematical model of foundation
resting inside a trough
In this case the mathematical model for analysis of the system is as shown in
Figure 2.7.7.
In the above mathematical model,
m = mass of the hammer or the tup;
m
1
= mass of the trough resting on soil;
m
2
= mass of the foundation plus hammer frame resting on it; if any,
m
H
m
3
x
3
k
3
c
3
m
2
x
2
k
2
c
2
x
1
m
1
k
1
c
1
Figure 2.7.7 Mathematical model of hammer foundation resting inside a trough.
m
3
= mass of the anvil resting on elastic pad supplied by vendor;
k
1
= soil spring to deduced from either Barkan, Richarts formula;
k
2
= spring value for the elastic pad/spring on which the foundation is resting whose
value is normally furnished by the vendor supplying these pads;
k
3
= spring value for the elastic pad supporting the anvil whose value is normally
furnished by the vendor supplying these pads;
c
1
c
2
= damping of the elastic pad /spring supporting the foundation on the trough;
c
3
= damping of the elastic pad supporting the anvil;
H = height of the free fall of the hammer;
x
1
= amplitude of the trough resting on soil;
x
2
= amplitude of the RCC block resting on springs; and
x
3
= amplitude of the anvil resting on the elastic pad.
The free vibration of motion for the above is given by
_
_
m
1
0 0
0 m
2
0
0 0 m
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
+
_
_
k
1
+k
2
k
2
0
k
2
k
2
+k
3
k
3
0 k
3
k
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
= 0 and
the damped equation of motion is given by
_
_
m
1
0 0
0 m
2
0
0 0 m
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
+
_
_
c
1
+c
2
c
2
0
c
2
c
2
+c
3
c
3
0 c
3
c
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
+
_
_
k
1
+k
2
k
2
0
k
2
k
+
2
k
3
k
3
0 k
3
k
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
= 0 (2.7.3)
Based on Equation (2.7.3), we have managed to formulate the equation of motion
for the hammer foundation, but how it will behave under transient load is yet to be
ascertained, though. Based on Example 2.7.1, we have shown the vibration charac-
teristics of a body having single degree of freedom under impact loading, we will now
extend this theory to multi-degrees of freedom.
2.7.2.2 Un-damped Response of a system under
impact loading having multi-degree of freedom
We write equation of motion in matrix notation as
[M]
_
X
_
+[K] {X} = 0 (2.7.4)
Here [M] = A square mass matrix of the order n n;
[K] = A square stiffness matrix of the order n n;
{X} = A Column deflection matrix of order n 1 (which means n rows and 1
column).
Based on the orthogonal property of the matrix
28
we have
[]
T
[M][]{
X} +[]
T
[K][]{X} = 0 which de-couples to (2.7.5)
28 Refer Chapter 5 (Vol. 1) for the orthogonal property of the Matrix.
N
i=1
{{
i
} +
2
i
{
i
}} = 0 where {X} =
N
i=1
[]{
i
} (2.7.6)
Now let {
i
} = {A
i
sin
i
t +B
i
cos
i
t}. (2.7.7)
Since, {X} =
N
i=1
[] {
i
} , we have, {X} =
N
i=1
[] {A
i
sin
i
t +B
i
sin
i
t}
Multiplying both sides of the above expression by the term []
T
[M] we have
[]
T
[M] {X} =
N
i=1
[]
T
[M] [] {A
i
sin
i
t +B
i
sin
i
t} which reduces to
[]
T
[M] {X} =
N
i=1
{A
i
sin
i
t +B
i
sin
i
t}
Now imposing the boundary condition at t = 0 {X} = {0}, we have
{0} =
N
i=1
{A
i
sin
i
t +B
i
cos
i
t}, which implies
{B
i
} = {0}, thus we have, []
T
[M]{X} =
N
i=1
{A
i
sin
i
t}
Again imposing the boundary condition at t = 0, {
X} = {V
0
} we have
[]
T
[M] {V
0
} =
N
i=1
{A
i
i
} {A
i
} =
[] [M] {V
0
}
i
, which gives
{X} =
N
i=1
_
[]
T
[M] {V
0
}
i
_
[] {sin
i
t} (2.7.8)
Based on the above we can clearly infer that for {sin
i
t} = {1}, we have the
maximum value of the amplitude vector.
We now explain further, the phenomenon based on a suitable numerical example.
For the numerical worked out problem below, we have deliberately used a
theoretical data with an objective that you can follow the process clearly.
Example 2.7.2
For a system having the following data find out the amplitude of vibration when
the mass m
2
is subjected to an initial velocity of 0.5 m/sec.
[K] =
_
30 10
10 30
_
, [M] =
_
1 0
0 1
_
, and {V} =
_
0.0
0.5
_
Solution:
The free vibration of motion for the body is given by
_
m
1
0
0 m
2
_ _
x
1
x
2
_
+
_
k
11
k
12
k
21
k
22
_ _
x
1
x
2
_
= 0
The eigen value solution of the problemis expressed as
_
30 10
10 30
_
= 0
and this on simplification reduces to
( 20) ( 40) = 0 which gives
= 20 or
1
= 4.35 rad/sec; and = 40 :
2
= 6.32 rad/sec.
Calculation of the eigen-vectors
For the first mode having = 20
10
1
10
2
= 0 and 10
1
+10
2
= 0
Thus for
1
= 1, we have
2
= 1
For the second mode having = 40
10
1
10
2
= 0 and 10
1
10
2
= 0
Thus for
1
= 1 we have
2
= 1
This the complete eigen vector matrix as [] =
_
1 1
1 1
_
.
The normalized eigen vector, based on orthogonal transformation theory is
given by
[]
n
=
_
1/
2 1/
2
1/
2 1/
2
_
We had deduced earlier that {X} =

N
i=1
_
[]
T
[M]{V
0
}
i
_
[] {sin
i
t} ,
where {X} = [] {
i
} , thus based on mathematical symmetry we have {
i
} =
N
i=1
_
[]
T
[M]{V
0
}
i
_
{sin
i
t} .
Hence for the first mode
{
1
} =
_
1
2
1
2
_ _
1 0
0 1
_ _
0
0.5
_
1
4.5
sin 4.5t
the above on simplification reduces to [
1
] = 0.07856sin 4.5t
{
2
} =
_
1
2

1
2
_ _
1 0
0 1
_ _
0
0.5
_

1
6.32
sin 6.32t
the above on simplification reduces to {
2
} = 0.056sin 6.32t
As {X} = [] {
i
} , we have, {X} =
_
1/
2 1/
2
1/
2 1/
2
_ _
0.07856sin 4.5t
0.056sin 6.32t
_
{X} =
_
0.0555sin 4.5t 0.0396sin 6.32t
0.0555sin 4.5t +0.03965sin 6.32t
_
The above when plotted at a time step of 0.05 sec gives plots as shown in
Figure 2.7.8.
Undamped motion of the system
-0.1
-0.08
-0.06
-0.04
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65
Time Steps
A
m
p
l
i
t
u
d
e
x1
x2
Figure 2.7.8 Undamped response of a system under impact loading having multi-degree of
freedom.
The above plot shows how the two degree system body vibrates under a
transient initial velocity of 0.5 m/sec.
2.7.2.3 Damped response of a system under impact
loading having multi-degree of freedom
We had already discussed in our earlier chapters that undamped motion of vibra-
tion is an idealization which is practically rare in nature and as such it would now
be worthwhile to asses (Chowdhury et al. 2002) how damping affect the above
phenomenon.
We write equation of motion in matrix notation as
[M]{
X} +[C]{
X} +[K]{X} = 0 (2.7.9)
where
[M] = a square mass matrix of the order n n;
[K] = a square stiffness matrix of the order n n;
[C] = a square damping matrix of order n n;
{X} = a column deflection matrix of order n 1
(which means n rows and 1 column)
Based on the orthogonal property of the matrix we have
[]
T
[M][]{
X} +[]T[C][]{
X} +[]
T
[K][]{X} = 0 (2.7.10)
which de-couples to
N
i=1
{{
i
} +2D
i
i
{
i
} +
2
i
{
i
}} = 0 where {X} =
N
i=1
[]{
i
} (2.7.11)
We have already proved earlier that for body having single degree of freedom the
free damped equation of motion is given by
x = e
D
n
t
[C
1
cos
d
t +C
2
sin
d
t] (2.7.12)
where,
d
=
n
_
[1 D
2
].
Thus in transformed co ordinate when the equations get de-coupled we can write
Let {
i
} = e
D
i
ni
t
{A
i
sin
di
t +B
i
cos
di
t} , and (2.7.13)
since {X} =
N
i=1
[] {
i
} , we have,
{X} =
N
i=1
[] e
D
i
ni
t
{A
i
sin
di
t +B
i
cos
di
t} (2.7.14)
Multiplying both sides of the above expression by the term []
T
[M] we have
[]
T
[M]{X}=
N
i=1
[]
T
[M][]e
D
i
ni
t
{A
i
sin
di
t +B
i
cos
di
t} which reduces to
[]
T
[M] {X} =
N
i=1
e
D
i
ni
t
{A
i
sin
di
t +B
i
cos
di
t}
Now imposing the boundary condition at t = 0 {X} = {0}, we have,
{0} =
N
i=1
e
D
i
ni
t
{A
i
sin
di
t +B
i
cos
di
t} which implies,
[B
i
] = [0] , thus we have []
T
[M] [X] =
N
i=1
e
D
i
ni
t
{A
i
sin
di
t}
Again imposing the boundary condition at t = 0 {
X} = {V
0
} we have,
[]
T
[M] {V
0
} =
N
i=1
{A
i
di
} {A
i
} =
[]
T
[M] {V
0
}
i
, which gives
{X} =
N
i=1
_
[]
T
[M]{V
0
}
di
_
[]e
D
i
ni
t
{sin
i
t} (2.7.15)
It will be interesting to note that in this case the maximumamplitude does not occur
at
i
t = /2 as in the case of undamped vibration.
To get the maximum amplitude we need to plot the complete time history which
was not required for the undamped case.
We now further explain the phenomenon based on a suitable numerical example.
Example 2.7.3
Repeat the problem worked out in Example 2.7.1 with following damping
ratio D
1
= 0.15 and D
2
= 0.20. All other parameters remain the same as
in Example 2.7.1.
Solution:
The free damped equation of motion for the problem is given by
_
m
1
0
0 m
2
_ _
x
1
x
2
_
+
_
c
11
c
12
c
21
c
22
_ _
x
1
x
2
_
+
_
k
11
k
12
k
21
k
22
_ _
x
1
x
2
_
= 0
We had already proved earlier that based on eigen solution

1
= 4.35 rad/sec and
2
= 6.32 rad/sec.
The normalised eigen vector matrix for this case was calculated in the previous
example as
[]
n
=
_
1/
2 1/
2
1/
2 1/
2
_
Since {X} =
N
i=1
_
[]
T
[M] {V
0
}
di
_
[] e
D
i
ni
t
{sin
i
t} and
{X} =
N
i=1
[] {
i
} we have
{
i
} =
N
i=1
_
[]
T
[M] {V
0
}
di
_
e
D
i
ni
t
{sin
i
t}
Thus for first mode we have
{
1
} =
_
1
2
1
2
_ _
1 0
0 1
_ _
0
0.5
_
1
4.5
(10.0225)
e
0.675t
sin 4.449t
or {
1
} = 0.07946 e
0.675t
sin 4.449t
For second mode we have
{
2
} =
_
1
2

1
2
_ _
1 0
0 1
_ _
0
0.5
_
e
1.264t
sin 6.19t
or {
2
} = 0.0571168e
1.264t
sin 6.19t
Now since
{X} =
N
i=1
[] {
i
} we have, {X} =
_
1/
2 1/
2
1/
2 1/
2
_
_
0.07946e
0.675t
sin 4.49t
0.0571168e
1264t
sin 6.19t
_
or {X} =
_
0.0562e
0.675t
sin 4.49t 0.0404e
1.264t
sin 6.19t
0.0562e
0.562t
sin 4.49t +0.0404e
1.264t
sin 6.19t
_
The above equations when plotted at a time step of 0.05 sec, shows the history
as given in Figure 2.7.9.
Damped response of motion
-0.04
-0.02
0
0.02
0.04
0.06
0.08
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67
Time steps
A
m
p
l
i
t
u
d
e
x1
x2
Figure 2.7.9 Damped response for the two man system.
Observations:
On studying the above mentioned plot you will see that the characteristic of the
curve is very similar to that plotted for the single degree of freedom.
Also observe that the response is largely reduced compared to Example 2.7.2 due
to the consideration of damping in this case.
2.8 DESIGN OF HAMMER FOUNDATION
2.8.1 Design criteria for hammer foundation
Its time we emerge form the exotic world of Newtonian Mechanics and start looking
into the design parameters pertaining to hammer foundations.
In the design criteria we use the following nomenclatures
W
h
= weight of Hammer in kN;
W
a
= weight of the anvil in kN;
W
fr
= weight of hammer resting on foundation in kN;
W
f
= weight of the RCC foundation in kN;
L
a
B
a
= contact Area of the anvil in m
2
;
L B = contact are of the foundation in m
2;
E
a
= Youngs Modulus of the elastic pad below the anvil in kN/m
2
;
t
a
= thickness of the elastic pad in m;
G = Dynamic shear modulus of the soil in kN/m
2
;
g = acceleration due to gravity @ 9.81 m/sec
2
;
D
a
= critical damping ratio of the elastic pad;
D
s
= critical damping ratio of the soil;
p = pressure on piston of double acting hammer kN/m
2
;
A
p
= area of piston in m
2
;
L = length of the stroke in m;
= correction factor @ 1.0 for well adjusted hammer and varies between 0.5 to
0.8 for double acting hammers. Usual design value taken is 0.65;
k = coefficient of impact @ 0.5 for stamping hammers and 0.25 for forging
hammers;
v = velocity of the hammer at the point of impact in m/sec;
V
a
= velocity of anvil after the impact in m/sec;

s
= allowable static bearing capacity of the soil in kN/m
2
; and

p
= allowable stress of the elastic pad in kN/m
2
.
2.8.1.1 Maximum permissible amplitudes for foundation
The maximum permissible amplitude of vibration for a hammer foundation shall not
exceed 1.2 mm.
2.8.1.2 Maximum permissible amplitudes for the anvil
The maximumpermissible amplitude is usually dependent on the weight of the hammer
and is also at times prescribed by the vendor supplying the elastic pad who limits the
deflection based on the allowable stress of the elastic pad. Following Table 2.8.1
furnishes the allowable amplitude of the anvil based on hammer weight.
2.8.1.3 Minimum weight of foundation
The minimum weight of foundation is given by
W
min
= W
h
_
8(1 +k)v
W
a
+W
fr
W
h
_
Table 2.8.1
Weight of Max. permissible
Sl. No. hammer (KN) amplitude (mm)
1 30 34
2 20 2
3 Upto 10 1
2.8.1.4 Minimum base area of the foundation
The minimum base area of the foundation may be obtained from the following
expression
A
min
=
20(1 +k)
s
v W
h
2.8.1.5 Minimum thickness of foundation (Major 1980)
This is given in Table 2.8.2.
2.8.1.6 Velocity, V, of hammer at point of impact
For free falling hammer
V =
_
2gH
If energy of impact E
i
is given by the manufacturer then
H =
E
i
W
h
and V =
_
2gH
For double acting hammer
V = 0.65
_
2g(W
h
+pA
p
)l
W
h
2.8.1.7 Velocity of anvil after of impact
V =
(1 +k)
_
1 +
W
a
W
h
_v
Table 2.8.2
Weight of Thickness of
Sl. No. hammer (kN) foundation (mm)
1 >60 >2250
2 60 2250
3 40 1750
4 20 1250
5 10 1000
2.8.1.8 Calculation of natural frequency based on IS-2974
Calculate
a
=
_
k
a
m
a
Calculate
z
=
_
k
z
m
a
+m
f
+m
fr
Calculate
n1
and
n2
from the equation
4
n
(
2
a
+
2
z
) (1 +)
2
n
+(1 +)
2
a
2
z
= 0; =
m
a
m
f
+m
fr
where, k
a
=
E
a
A
a
t
a
and k
z
=
4Gr
0
(1)
in which, r
0
=
_
LB
and, G = dynamic shear

modulus of the soil.
When supported on short bearing piles
k
z
=
k
p
k
z
k
p
+k
z
where, k
p
= vertical stiffness of pile which may be obtained from formulas derived
earlier and m
a
=
W
a
g
, m
f
=
W
f
g
and m
fr
=
W
fr
g
.
2.8.1.9 Amplitude of vibration based on IS-2974
The amplitude of the RCC foundation (x
1
) is given by
x
1
=
(
2
a

2
n2
)(
2
a

2
n1
)
2
a
(
2
n1

2
n2
)f
n2
V
_
sin
n1
t
n1
sin
n2
t
n2
_
The amplitude of vibration of the anvil (x
2
) is given by
x
2
=
V
(
2
n1

2
n2
)
_
(
2
a

2
n2
) sin
n1
t
n1
(
2
a

2
n1
) sin
n2
t
n2
_
2.8.1.10 Stability of pad between anvil and block
Total deflection of the pad under impact is given by
tot
=
st
+
dyn
where
st
=
W
a
+W
fr
K
a
and
dyn
=
V
na
2.8.1.11 The loading intensity on the pad
The loading intensity on the pad is given by
p
=
K
a
tot
A
a
2.8.1.12 Stress in soil below foundation
The stress in soil below foundation is given by
s
=
W
a
+W
f
+W
fr
+K
z
A
f
in which, =
V
2 f
z
, where

V =
1 +k
1 +
W
f
W
a
V.
2.8.2 Discussion on the IS-code method of analysis
The discussion should not be deemed as a criticism of the method, for normal
engineering design of standard hammer foundations the method advocated by the
code is adequate. But when a foundation subjected to loading of large magnitude
(heavy hammer >40 kN), the method proposed by the code may lead to conservative
and expensive design. Moreover if the environmental criteria calls for more stringent
restriction of amplitude or transmittal of waves, a designer may find it difficult to meet
the requirements based on IS-code.
The reasons attributable for the same may be summarized as follows:
IS-code method does not take damping of the pad or that of the soil into considera-
tion. It has been observed that damping plays a very signicant role in minimizing
the amplitude of vibration for such hammer foundation (Novak and El Hifnawy
1983).
It also does not take into consideration the embedment effect which could play a
very significant role for heavy hammer foundation when the depth of the block
could be quite large.
The IS-code formula of k
z
= 7.6Gr
0
apparently looks overestimated
29
.
The dynamic displacement (
dyn
) is based on uncoupled form when the actual
response should based be coupled response. This could either under-estimate or
could also over estimate the stress induced in the foundation.
Based on the above the design procedure suggested herein may be structured as
follows.
29 For if we equate 4Gr
0
/(1 v) = 7.6Gr
0
. We get v = 0.473 => 0.5. Poissons ratio @ 0.5 depicts
perfectly plastic clay which is rarely obtained. Value of v is usually taken as 0.4.
2.8.3 Check list for analysis of hammer foundation
Check if the following data regarding the machine and the foundation are available
with you
Weight of the hammer.
Type of hammer (free falling or double acting etc.).
Sufficient data to calculate the velocity of hammer at point of collision.
(like energy of impact, height of free fall etc.).
Geometric dimension and weight of anvil.
Geometric dimension, stiffness and damping property of the elastic pad supporting
the anvil.
Anchoring detail of the elastic pad.
Stiffness and damping data if mechanical springs and dampers are used in lieu of
elastic pads.
Mechanical detail of the springs and dampers including their xing detail as
suggested by the supplier.
Anchoring detail of the frame supporting the hammer on the foundation.
Top elevation of the anvil based on the mechanical process.
Allowable bearing capacity of the soil.
Dynamic shear modulus of the soil.
Grade of concrete.
2.8.4 Other techniques of analysis of Hammer foundation
For most of the cases analysis as mentioned above suffice. However there are cases
where due to the massiveness of the foundation and hammer more detailed analysis is
envisaged where the effect of generated shock to its surrounding could be significant
30
.
In such cases the best way to analyze such problem would be to resort to FEM
analysis of such foundations.
Shown in Figure 2.8.1 is a conceptual Finite element mathematical model of hammer
foundation housed inside a structural building. If the hammer foundation is very heavy,
though the hammer foundation itself may be within the acceptable limit of codal
stipulation can yet have adverse effect on the building in which it is located.
In such cases studying the problem based on FEM could be quit advantageous.
2.8.4.1 Selection of elements
Here the anvil is modeled as a beam element having thee degrees of freedom
(horizontal, vertical and rotational degrees).
The elastic pad below the anvil which is usually modeled as a spring can be
converted to equivalent truss element having stiffness @ AE/L for computer
implementation.
30 It has not been uncommon that the shocks generated by hammer foundation has done secondary damages
to the building in which it is placed or have rendered crane girders unserviceable at the Gantry level due
to distortion.
Anvil modeled as beam element
Lumped mass(Typ)
Pad modeled
as truss
element
Figure 2.8.1 Finite element model of hammer foundation.
The block foundation itself can be modeled as a 2D plane stress element having
incompatible modes.
The soil medium can be modeled as 2D plane strain element again having
incompatible modes.
Finally the building and the foundation may be modeled as a plane frame
constituting of beam elements connected to the soil elements.
2.8.4.2 Boundary conditions
This is surely a major problem for failing to take a correct judgment can render the
analysis useless for waves generated due to the transient shock from the hammer may
get reflected from the boundary and generate spurious response of the foundation as
well as the building rendering the analysis useless.
We had already discussed earlier that for this type of infinite domain problem the
boundary should be extended far enough ensuring that no reflection of the waves take
place.
This on the other hand makes the analysis voluminous and also expensive in terms
of man-hours.
Moreover from practical point of view geo-technical data may not be available to
the depth required to ensure that no reflection takes place.
This can however be done by following techniques:
To provide spring and dash-pots having high damping value at the boundaries which
ensures that it absorbs all energy transmitted to it
31
.
Provide paraxial or quite boundaries to suppress the spurious modes
32
.
Providing boundaries at a distance at least 2.5 times the length of the Rayleigh waves
to ensure that radiation damping is good enough to dissipate away the energy.
2.8.4.3 Material input
Following, material input shall be provided for the analysis:
Dynamic Modulus of concrete for the RCC block.
Dynamic modulus of the elastic pad.
Dynamic shear modulus of the soil
33
.
Poissons ratio of the soil.
Damping property of elastic pad below anvil (in this case the truss elements).
Dynamic elastic modulus of the structure enabling the computer to generate the
stiffness matrix.
2.8.4.4 Input loading
The velocity of the anvil after impact shall be directly provided as input velocity for
the nodes of the beam element where masses are lumped.
2.8.4.5 Method of analysis
Due to the heterogeneous properties of the material we advocate here a time history
analysis having a time step of 0.1T
n
where T
n
is the least period of the system.
31 Unfortunately most of the commercially available FEM software does not have this feature of directly
inputting dash pots except ANSYS.
32 Refer to the Chapter 5 (Vol. 1) for further explanation.
33 If the soil is layered then for each layer a separate value of shear modulus has to be provided enabling
the computer to develop the material stiffness matrix for the plane strain element.
If the building system is found to have unacceptable amplitude ways and means
have to be sought to reduce the vibration transmitted to it by either providing air gaps
around the foundation or by providing suitable dampers around the foundation to
absorb this energy.
Example 2.8.1
Ahammer foundation (Figure 2.8.2) having the following data has to be designed
for a particular site.
Frame Column(Typ.)
1400
2000
1400
1150 1250 1250 1150
1290
2150
Figure 2.8.2 General arrangement of the hammer foundation.
Calculate the natural frequency of vibration
The damped amplitude of vibration
Compare the same with time history response
Re-analyse the foundation based on IS-Code or Barkans mathod and
compare the results.
Design input data
1 Weight of hammer = 20 kN
2 Weight of frame = 85 kN
3 Plan area of anvil = 2.0 m 2.5 m
4 Thickness of oak pad = 0.4 m
5 Height of free fall = 1.8 m
6 Unit weight of soil = 10 kN/m
3
7 Dynamic shear modulus of soil = 120,0000 kN/m
2
8 Poissons ratio = 0.40
9 Dynamic modulus of oak wood = 500,0000 kN/m
2
10 Damping ratio for oak = 0.10
11 Plan area of foundation = 5 m 5 m
12 Co-efficient of restitution = 0.65
13 Allowable bearing capacity of soil = 180 kN/m
2
14 Air gap between anvil and foundation = 100 mm
Let height of the R.C.C. anvil be = 1250 mm
Weight of anvil (W
a
) = 2.52.01.2525 = 156.25 kN
Weight of hammer = 20 kN
Velocity of hammer at point of collision =
_
2gH = v =

29.811.80 =
5.94 m/sec
Velocity of Anvil after the impact
V =
(1 +k)
_
1 +
W
a
W
h
_v =
1.65
1 +
156.25
20
5.94 = 1.112 m/sec
Minimum weight of foundation
W
min
= W
h
_
8(1 +k)v
W
a
+W
fr
W
h
_
= 20
_
81.655.94
156.25 +85
20
_
= 1327.16 kN
Thus minimum depth required =
1327.16
5 5 25
= 2.12 m

= 2.15 m(say)
Thus based on the sketch furnished earlier,
Weight of foundation = (5 5 3.44 2.7 2.2 1.29) 25 =1958 kN.
Weight of the oak pad = 2.5 2.0 0.4 20 = 40 kN
Minimum base area required = A
min
=
20(1 +k)
s
v W
h
=
20 1.65
180
5.94 20 = 21.78 > 25 m
2
OK
Stiffness properties
For Pad stiffness (k
2
) =
EA
p
t
=
500 10
4
5
0.4
= 62510
4
kN/m
For Soil
Equivalent Radius r
0
=
_
25
= 2.82 m
k
z
=
4Gr
0
(1 )
=
4 120 10
4
2.82
0.6
= 22560000 kN/m
Damping properties
Mass of oak pad =
40
9.81
= 4.0 kN sec
2
/m
C
c
= 2
km = 2
_
625 10
5
4 = 31623 kN sec/m
C = D2
km = 0.1 31623 = 3162 kN sec/m

Mass of foundation and machine =
1958 +156 +85
9.81
= 224 kN sec
2
/m
B
z
=
0.25(1 ) mg
s
r
3
0
=
0.25 0.6 224 9.81
19 (2.82)
3
= 0.773
D =
0.425
_
B
z
= 0.4832
Thus damping of the soil is given by
C
s
= D2
km = 20.4832
22560000224 = 68699 kN sec/m

The mathematical model
The mathematical model perceived is given in Figure 2.8.3.
m
2
x
2
k
2
c
2
m
1
x
1
k
1
c
1
Figure 2.8.3 Mathematical model of the foundation.
The equation of motion for the above is given by
_
m
1
0
0 m
2
_ _
x
1
x
2
_
+
_
C
1
+C
2
C
2
C
2
C
2
_ _
x
1
x
2
_
+
_
k
1
+k
2
k
2
k
2
k
2
_ _
x
1
x
2
_
= 0
For free vibration we have
_
m
1
0
0 m
2
_ _
x
1
x
2
_
+
_
k
1
+k
2
k
2
k
2
k
2
_ _
x
1
x
2
_
= 0
_
208 0
0 16
_ _
x
1
x
2
_
+
_
850.6 625
625 625
_ _
x
1
x
2
_
10
5
= 0
For eigen solution we have
_
85060000 208 62500000
62500000 62500000 16
_ _
x
1
x
2
_
= 0
The above on expansion and simplification reduces to
2
4315192 +4.2367788 10
11
= 0;
1
= 100525 and
2
= 4214667
1
= 317 rad/sec and
2
= 2052 rad/sec
Calculation of eigen vectors
For the first mode
_
85060000 20909200 62500000
62500000 62496672
_ _
11
12
_
= 0;
for
11
= 1.0 we have
12
= 1.0264
For the second mode
_
85060000 8.766507410
8
62500000
62500000 4934672
_ _
21
22
_
= 0; for
21
= 1.0
we have
22
= 12.665
Thus the complete eigen vector is
[] =
_
1.0 1.0
1.0264 12.665
_
Normalization of the eigen vectors
For the first mode
{}
T
[M] {} = 1.00 1.0264
_
208 0
0 16
_ _
1.00
1.0264
_
= 224.856 and
_
M
r
=
224.856 = 14.495 {
1
}
N
=
_
0.066688
0.06844
_
For the second mode
{}
T
[M] {} = 1.00 12.665
_
208 0
0 16
_ _
1.00
12.665
_
= 2774 and
_
M
r
=
2774 = 52.67 {
1
}
N
=
_
0.01898
0.240
_
Thus the complete normalized eigen vector is []
N
=
_
0.0667 0.01898
0.0684 0.240
_
.
Damping matrix and its correction for de-coupling
The damping matrix is given by [C] =
_
c
1
+c
2
c
2
c
2
c
2
_
=
_
71861 3162
3162 3162
_
The above matrix on orthogonal transformation does not de-couple as such
we need to correct it.
For the first mode
{}
T
[C] {} = 0.0667 0.0684
_
71861 3162
3162 3162
_ _
0.0667
0.0684
_
= 305
2D
1
1
= 305 D
1
= 0.480.
For the second mode
{}
T
[C] {} =0.01898 0.0240
_
71861 3162
3162 3162
_ _
0.01898
0.240
_
= 208
2D
2
2
= 208 D
2
= 0.050.
Based on Rayleigh damping we know that on orthogonal transformation the
expression should reduce to
2D
1
1
= +
2
1
and 2D
2
2
= +
2
2
or +100525 = 305 and +4214667 = 208
the equation on solving gives = 302.63 and = 2.3577 10
5
Thus, based on Rayleigh damping
[C] = [M] + [K] [C] = 302.63
_
208 0
0 16
_
+2.3577 10
5
_
850.6 625
625 625
_
10
5
=
_
64952 1474
1474 6316
_
Calculation of amplitude
The equation for damped amplitude is given by
{X} =
N
i=1
_
[]
T
[M]{V
0
}
di
_
[]e
D
i
ni
t
{sin
i
t}
{
i
} =
N
i=1
_
[]
T
[M]{V
0
}
di
_
[]e
D
i
t
ni
{sin
i
t}
or
1
= 0.0667 0.0684
_
208 0
0 16
_ _
0
1.112
_
1
317
(1 0.2304)
e
0.48317t
sin(317
1 0.2304)t

1
= 4.3776e
152.16t
sin 278t 10
3
2
= 0.01898 0.240
_
208 0
0 16
_ _
0
1.112
_
1
2052
(1 0.0025)
e
102.6t
sin(2052
1 0.0025)t

2
= 2.08398e
102.6t
sin 2049t 10
3
Since {X} = [] {}, hence
{X} =
_
0.0667 0.01898
0.0684 0.240
_ _
4.3776e
152.16t
sin 278t
2.08398e
102.6t
sin 2049t
_
10
3
or {X} =
_
0.29198e
152.16t
sin 278t 0.03955e
102.6t
sin 2049t
0.299428e
152.16t
sin 2049t +0.50015e
102.6t
sin 2049t
_
10
3
m.
The above when plotted at a time step of 0.0005 seconds shows displacement
plots as depicted in Figure 2.8.4.
Displacement History of Hammer Foundation
with time step of 0.0005 seconds
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
1 14 27 40 53 66 79 92 105 118 131 144
Time steps
A
m
p
l
i
t
u
d
e

(
m
m
)
Displacement amplitude
of foundation (mm)
of anvil (mm)
Figure 2.8.4 Displacement history of foundation and the anvil.
Time History analysis based on Newmark-beta method
Here we do time history analysis of the equation
_
208 0
0 16
_ _
x
1
x
2
_
+
_
71861 3162
3162 3162
_ _
x
1
x
2
_
+
_
85060000 62500000
62500000 62600000
_ _
x
1
x
2
_
= 0
having boundary condition at t = 0 v = 1.112 m/sec. Solution is given in
Figure 2.8.5.
Time history plot of Hammer Foundation
(Newmark-beta Method)
-0.0004
-0.0002
0
0.0002
0.0004
0.0006
1 14 27 40 53 66 79 92 105 118 131 144
Time steps
A
m
p
l
i
t
u
d
e
(
m
e
t
e
r
)
Amplitude of fdn.
Amplitude of anvil
Figure 2.8.5 Time history response of amplitude for foundation and the anvil.
It is to be noted that we use here the original non proportional damping matrix
and not the corrected one used above. we give the following results for 98 steps
(explanation in next page).
Next we compare the response of the foundation and the anvil separately to see
what is the variation in the results. The results are as plotted in Figures 2.8.6 to 7.
Comparison of amplitude of foundation
based on Time History and Closed form
-0.05
0
0.05
0.1
0.15
0.2
1 15 29 43 57 71 85 99 113 127 141
Time steps
A
m
p
l
i
t
u
d
e
(
m
m
)
Amplitude of foundation
based on Newmark Method
of foundation(mm)
Figure 2.8.6 Comparison of response, time history versus approximate damping.
Sl. Time
No. step x
1
(disp) x
1
(vel) x
1
(acc) x
2
(disp) x
2
(vel) x
2
(acc)
0 0 0 0 0 0.00 10
+00
1.112 0
1 0.0005 8.0310
06
3.21 10
02
1.28 10
+02
4.42 10
04
0.65707 1819.7
2 0.001 3.90 10
05
9.18 10
02
1.10 10
+02
5.39 10
04
0.2687 1883.2
3 0.0015 9.10 10
05
1.16 10
01
1.27 10
+01
2.58 10
04
0.855 462.04
4 0.002 1.40 10
04
7.99 10
02
1.32 10
+02
1.24 10
04
0.6752 1181.16
5 0.0025 1.62 10
04
8.62 10
03
1.53 10
+02
2.80 10
04
0.05021 1720.53
6 0.003 1.53 10
04
4.58 10
02
6.51 10
+01
9.60 10
05
0.68719 827.405
7 0.0035 1.29 10
04
4.89 10
02
5.27 10
+01
2.58 10
04
0.72961 657.74
8 0.004 1.15 10
04
9.27 10
03
1.06 10
+02
4.88 10
04
0.19049 1498.7
9 0.0045 1.20 10
04
3.21 10
02
5.96 10
+01
4.22 10
04
0.4548 1082.6
10 0.005 1.38 10
04
3.73 10
02
3.87 10
+01
1.36 10
04
0.6883 148.758
11 0.0055 1.47 10
04
1.53 10
03
1.05 10
+02
1.27 10
04
0.365 1144.36
12 0.006 1.36 10
04
4.63 10
02
8.6810
+01
1.6710
04
0.20483 1135
13 0.0065 1.07 10
04
6.94 10
02
5.62 10
+00
2.01 10
05
0.54351 219.711
14 0.007 7.67 10
05
5.33 10
02
7.01 10
+01
2.56 10
04
0.40073 790.8
15 0.0075 5.95 10
05
1.53 10
02
8.19 10
+01
3.39 10
04
0.0677 1082.9
16 0.008 5.87 10
05
1.20 10
02
2.72 10
+01
2.07 10
04
0.4603 487.52
17 0.0085 6.37 10
05
7.88 10
03
4.36 10
+01
2.55 10
05
0.4714 443.13
18 0.009 6.04 10
05
2.10 10
02
7.19 10
+01
1.75 10
04
0.1257 939.469
19 0.0095 4.30 10
05
4.85 10
02
3.83 10
+01
1.39 10
04
0.27068 646.244
20 0.01 1.79 10
05
5.18 10
02
2.53 10
+01
2.89 10
05
0.39924 132.01
21 0.0105 2.27 10
06
2.89 10
02
6.61 10
+01
1.74 10
04
0.18314 732.36
22 0.011 9.23 10
06
1.09 10
03
5.39 10
+01
1.77 10
04
0.1732 692.94
23 0.0115 5.11 10
06
1.54 10
02
3.34 10
+00
4.07 10
05
0.372 102.31
24 0.012 1.84 10
07
5.76 10
03
4.19 10
+01
1.19 10
04
0.2675 520.295
25 0.0125 2.46 10
06
1.63 10
02
4.64 10
+01
1.78 10
04
0.03175 676.689
26 0.013 1.42 10
05
3.05 10
02
1.02 10
+01
1.02 10
04
0.27195 284.117
27 0.0135 2.79 10
05
2.46 10
02
3.39 10
+01
3.30 10
05
0.26895 296.12
28 0.014 3.50 10
05
3.58 10
03
5.00 10
+01
1.12 10
04
0.04846 585.85
29 0.0145 3.19 10
05
1.58 10
02
2.75 10
+01
7.62 10
05
0.1932 380.92
30 0.015 2.31 10
05
1.96 10
02
1.25 10
+01
3.73 10
05
0.2607 111.02
31 0.0155 1.64 10
05
7.21 10
03
3.70 10
+01
1.31 10
04
0.1148 472.61
32 0.016 1.68 10
05
9.05 10
03
2.81 10
+01
1.32 10
04
0.11024 427.577
33 0.0165 2.29 10
05
1.51 10
02
3.79 10
+00
4.75 10
05
0.22911 47.8826
34 0.017 2.83 10
05
6.44 10
03
3.09 10
+01
4.91 10
05
0.1574 334.73
35 0.0175 2.76 10
05
9.28 10
03
3.19 10
+01
8.09 10
05
0.0303 415.92
36 0.018 2.04 10
05
1.93 10
02
8.03 10
+00
2.99 10
05
0.1738 158.38
37 0.0185 1.15 10
05
1.63 10
02
1.97 10
+01
5.43 10
05
0.1628 202.42
38 0.019 6.40 10
06
4.13 10
03
2.91 10
+01
9.99 10
05
0.0197 369.994
39 0.0195 7.06 10
06
6.79 10
03
1.45 10
+01
7.24 10
05
0.12982 228.176
40 0.02 1.07 10
05
7.89 10
03
1.01 10
+01
1.62 10
06
0.1662 82.648
41 0.0205 1.25 10
05
6.95 10
04
2.42 10
+01
6.08 10
05
0.07049 300.2
42 0.021 9.58 10
06
1.11 10
02
1.74 10
+01
6.10 10
05
0.0695 259.89
43 0.0215 3.15 10
06
1.46 10
02
3.19 10
+00
8.95 10
06
0.1388 16.976
44 0.022 2.72 10
06
8.86 10
03
1.99 10
+01
4.79 10
05
0.0887 217.091
45 0.0225 4.66 10
06
1.10 10
03
1.99 10
+01
6.27 10
05
0.02976 256.868
46 0.023 2.58 10
06
7.24 10
03
4.67 10
+00
2.62 10
05
0.11601 88.1138
47 0.0235 5.62 10
07
5.33 10
03
1.23 10
+01
2.88 10
05
0.10398 136.24
(Continued)
Sl. Time
No. step x
1
(disp) x
1
(vel) x
1
(acc) x
2
(disp) x
2
(vel) x
2
(acc)
48 0.024 1.37 10
06
2.09 10
03
1.74 10
+01
5.77 10
05
0.01174 232.72
49 0.0245 1.23 10
06
8.32 10
03
7.58 10
+00
4.05 10
05
0.0804 136.01
50 0.025 5.37 10
06
8.24 10
03
7.90 10
+00
4.45 10
06
0.0995 59.672
51 0.0255 7.99 10
06
2.22 10
03
1.62 10
+01
3.86 10
05
0.0371 189.843
52 0.026 7.37 10
06
4.66 10
03
1.13 10
+01
3.55 10
05
0.04953 156.874
53 0.0265 4.44 10
06
7.06 10
03
1.72 10
+00
8.53 10
07
0.08917 1.66764
54 0.027 1.76 10
06
3.66 10
03
1.19 10
+01
3.50 10
05
0.05426 141.3
55 0.0275 1.38 10
06
2.16 10
03
1.14 10
+01
4.33 10
05
0.0209 159.34
56 0.028 3.28 10
06
5.42 10
03
1.64 10
+00
1.99 10
05
0.073 49.244
57 0.0285 5.54 10
06
3.65 10
03
8.73 10
+00
1.41 10
05
0.0629 89.9718
58 0.029 6.11 10
06
1.38 10
03
1.14 10
+01
3.09 10
05
0.0041 145.004
59 0.0295 4.41 10
06
5.42 10
03
4.82 10
+00
1.89 10
05
0.05205 79.6655
60 0.03 1.71 10
06
5.39 10
03
4.92 10
+00
9.41 10
06
0.06115 43.237
61 0.0305 6.34 10
08
1.70 10
03
9.86 10
+00
2.97 10
05
0.02016 120.74
62 0.031 1.10 10
07
2.39 10
03
6.52 10
+00
2.63 10
05
0.0338 95.228
63 0.0315 1.61 10
06
3.61 10
03
1.66 10
+00
3.69 10
06
0.0567 3.79594
64 0.032 2.83 10
06
1.26 10
03
7.72 10
+00
1.87 10
05
0.033 90.9929
65 0.0325 2.54 10
06
2.42 10
03
7.02 10
+00
2.34 10
05
0.01427 98.0753
66 0.033 8.48 10
07
4.35 10
03
7.01 10
01
8.48 10
06
0.04544 26.603
67 0.0335 1.01 10
06
3.10 10
03
5.71 10
+00
1.22 10
05
0.03721 59.519
68 0.034 1.76 10
06
1.08 10
04
7.11 10
+00
2.14 10
05
0.0003 90.463
69 0.0345 1.09 10
06
2.60 10
03
2.84 10
+00
1.27 10
05
0.0345 46.558
70 0.035 1.90 10
07
2.51 10
03
3.20 10
+00
5.56 10
06
0.0385 30.5594
71 0.0355 8.65 10
07
1.95 10
04
6.05 10
+00
1.81 10
05
0.0118 76.5631
72 0.036 3.52 10
07
2.25 10
03
3.73 10
+00
1.56 10
05
0.02179 57.6405
73 0.0365 9.15 10
07
2.82 10
03
1.44 10
+00
1.45 10
06
0.03487 5.3432
74 0.037 1.92 10
06
1.19 10
03
5.09 10
+00
1.20 10
05
0.01895 58.344
75 0.0375 1.92 10
06
1.19 10
03
4.44 10
+00
1.41 10
05
0.0107 60.105
76 0.038 1.02 10
06
2.40 10
03
3.82 10
01
4.12 10
06
0.0291 13.827
77 0.0385 1.79 10
08
1.60 10
03
3.56 10
+00
8.86 10
06
0.0228 39.3595
78 0.039 2.96 10
07
3.50 10
04
4.26 10
+00
1.43 10
05
0.0012 56.5166
79 0.0395 2.38 10
07
1.79 10
03
1.48 10
+00
8.42 10
06
0.02214 27.2158
80 0.04 1.08 10
06
1.59 10
03
2.26 10
+00
3.04 10
06
0.02367 21.079
81 0.0405 1.49 10
06
5.06 10
05
3.89 10
+00
1.05 10
05
0.00634 48.257
82 0.041 1.13 10
06
1.49 10
03
2.29 10
+00
8.53 10
06
0.0144 34.597
83 0.0415 3.02 10
07
1.82 10
03
9.82 10
01
4.86 10
07
0.0217 5.36525
84 0.042 3.49 10
07
7.82 10
04
3.18 10
+00
8.65 10
06
0.011 37.4775
85 0.0425 3.75 10
07
6.77 10
04
2.66 10
+00
9.49 10
06
0.00762 36.8928
86 0.043 1.34 10
07
1.36 10
03
7.52 10
02
2.94 10
06
0.01859 7.00279
87 0.0435 6.73 10
07
7.98 10
04
2.32 10
+00
5.19 10
06
0.0139 25.792
88 0.044 7.62 10
07
4.42 10
04
2.64 10
+00
8.33 10
06
0.0013 35.146
89 0.0445 3.25 10
07
1.31 10
03
8.13 10
01
4.48 10
06
0.0141 15.722
90 0.045 2.84 10
07
1.13 10
03
1.52 10
+00
2.63 10
06
0.0144 14.4488
91 0.0455 6.01 10
07
1.39 10
04
2.45 10
+00
7.01 10
06
0.0032 30.4011
92 0.046 4.31 10
07
8.16 10
04
1.37 10
+00
5.40 10
06
0.00962 20.7196
93 0.0465 1.97 10
08
9.89 10
04
6.82 10
01
4.16 10
07
0.01365 4.5961
94 0.047 3.47 10
07
3.20 10
04
1.99 10
+00
5.45 10
06
0.0065 24.005
95 0.0475 2.84 10
07
5.72 10
04
1.58 10
+00
5.79 10
06
0.0052 22.597
96 0.048 9.65 10
08
9.51 10
04
6.63 10
02
1.59 10
06
0.0116 3.3553
97 0.0485 4.72 10
07
5.52 10
04
1.53 10
+00
3.39 10
06
0.0083 16.8157
98 0.049 5.49 10
07
2.43 10
04
1.66 10
+00
5.11 10
06
0.00138 21.7888
Comparison of amplitude of the anvil
based on Time History and Closed form
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
1 16 31 46 61 76 91 106 121 136
Time steps
A
m
p
l
i
t
u
d
e
(
m
m
)
Amplitude of anvil based
on Newmark Method
of anvil(mm)
Figure 2.8.7 Comparison of response time history versus approximate damping.
The above displacements plot show some very interesting results
The closed form solution and the time history results are very closely
matching
Since the time history is done with the original soil damping matrix and com-
pared with modal response having damping matrix corrected to Rayleigh
format (and yet it gives reasonably good results), it may be concluded that
the technique of separating the damping for each mode and correcting the
damping matrix based on Rayleigh damping may well be adapted without
any signicant error in cases where the damping matrix is non-proportional.
Analysis based on Code
a
=
_
k
a
m
a
=
_
62500000
16
= 1979 rad/sec
z
=
_
k
z
m
a
+m
f
+m
fr
=
_
25600009.81
156 +1958 +85
= 317 rad/sec
=
m
a
m
f
+m
fr
=
156
1958 +85
= 0.076
Equation for natural frequency is given by
4
n
(
2
a
+
2
z
)(1 +)
2
n
+(1 +)
2
a
2
z
= 0
Substituting the numerical values calculated above, we have
4
n
4322216.7
2
n
+4.234697410
11
= 0.
The above on solving gives,
2
n1
= 100302.9
n1
= 316.7 rad/sec and
2
n2
= 4221913.9
n2
= 2054 rad/sec.
The amplitude of vibration is given by the formula
x
1
=
(
2
a

2
n2
)(
2
a

2
n1
)
2
a
(
2
n1

2
n2
)f
n2
V
_
sin
n1
t
n1
sin
n2
t
n2
_
for the foundation and,
x
2
=
V
(
2
n1

2
n2
)
_
(
2
a

2
n2
) sin
n1
t
n1
(
2
a

2
n1
) sin
n2
t
n2
_
for the anvil.
Substituting the numerical values as calculated above and plotting @ 0.0005
sec time steps we time history plot for this case as shown in Figures 2.8.8 and 9.
Undamped response of hammer foundation
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
1
1
3
2
5
3
7
4
9
6
1
7
3
8
5
9
7
1
0
9
1
2
1
1
3
3
1
4
5
Time Steps
A
m
p
l
i
t
u
d
e

o
f

f
o
u
n
d
a
t
i
o
n

b
l
o
c
k
Amplitude of foundation
Figure 2.8.8 Amplitude of foundation as per IS-code.
Undamped response of anvil
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
9
1
1
0
1
1
1
1
1
2
1
1
3
1
1
4
1
Time steps
A
m
p
l
i
t
u
d
e

o
f

a
n
v
i
l
(
m
m
)
Amplitude of anvil
Figure 2.8.9 Amplitude of anvil as per IS-code.
The maximum amplitude of vibration is 0.284 mm in lieu of 0.156 mm for
damped case. This proves that without considering the effect of damping the
designer may have to make the foundation more heavy (thus expensive) to reduce
the amplitude if the value exceeds the acceptable limit as shown in Figure 2.8.9.
In this case the undamped amplitude of the anvil is 0.754 mm in lieu of
0.471 mm when damping is considered.
We give below a compare the various data as calculated above.
Comparison of results of the hammer foundation by various techniques:
Close Close Time history
formed formed response of
Engineering (undamped (damped the damped
parameter solution) Solution) equation Remarks
1st Natural 316.7 317 317 Practically no difference
frequency
2nd Natural 2054 2052 2052 Practically no difference
frequency
Amplitude of 0.754 0.471 0.539 Significant difference between un-
anvil (mm) damped and damped amplitude but
marginal difference with time history
Amplitude of 0.286 0.156 0.1622 Comments same as above
foundation (mm)
2.9 DESIGN OF ECCENTRICALLY LOADED HAMMER FOUNDATION
2.9.1 Mathematical formulation of anvil placed
eccentrically on a foundation
Sometimes due to the functional requirements anvils may be placed eccentric to the
RCC block as shown in Figure 2.9.1.
In this case the hammer hits the anvil concentrically, but as the anvil is placed at
an eccentricity e mm, say, the foundation block other than vertical mode also gets
subjected to a coupled horizontal and rocking mode.
The mathematical model used for such case is as shown in Figure 2.9.2.
Here additional sliding (x) and rocking mode () is also simulated due to the eccentric
impact.
Hammer(Wh)
e
Anvil(Wa)
Foundation(Wf)
Figure 2.9.1 Eccentrically loaded anvil.
m
2
z
2
k
2
c
2
C
x
X
m
C
k
1
z
1
k
x
k
1
c
1
Figure 2.9.2 Mathematical model of eccentrically loaded anvil.
The nomenclatures used here are:
m
1
= mass of the foundation block plus frame resting on it if any;
m
2
= mass of the anvil resting on elastic pad;
k
1
= soil spring to be calculated from Barkan or Richarts formula
34
;
k
2
= spring value for the elastic pad, this value is normally furnished by the vendor
supplying these pads;
k
x
= soil spring to be calculated from Barkan or Richarts formula in x direction;
k
= soil spring to be calculated from Barkan or Richarts formula in direction;

c
1
c
2
= damping of the elastic pad again furnished by the vendor;
c
x
= damping of the soil to be obtained from Richarts formula in x direction;
c
= damping of the soil to be obtained from Richarts formula in direction;

z
2
= amplitude vector of the anvil in vertical direction;
z
1
= amplitude vector of the Foundation in vertical direction;
x = amplitude vector of the Foundation in horizontal direction;
= rotational amplitude of the foundation, and
e = eccentricity of the anvil with respect to the c.g. of the foundation.
34 Refer to section on block foundation for the formula of the springs and dampers.
Based on dAlemberts equation and the free body diagram the equation of motion
for vibration are as follows:
m
2
z
2
+k
2
(z
2
z
1
) = 0
m
1
z
2
+k
1
(z
1
e) +k
2
(z
1
z
2
) = 0
m
1
x +k
x
(x Z
c
) = 0
J
k
x
Z
c
x +[k
W
f
Z
c
+k
1
e
2
+k
x
Z
2
c
] k
1
ez
1
= 0 (2.9.1)
The above when written in matrix form gives the equation
_
_
_
_
m
1
0 0 0
0 m
1
0 0
0 0 m
2
0
0 0 0 J
_
_
_
_
z
1
x
z
2
_
+
_
_
_
_
k
1
+k
2
0 k
2
k
1
e
0 k
x
0 k
x
Z
c
k
2
0 k
2
0
k
1
e k
x
Z
c
0 (k
W
f
Z
c
+k
1
e
2
+k
x
Z
2
c
)
_
_
_
_
z
1
x
z
2
_
= {0}
(2.9.2)
2.9.2 Damped equation of motion with eccentric anvil
The damped equation of motion is given by
[M]{
X} +[C]{
X} +[K]{X} = 0 (2.9.3)
Here,
[M] = a square mass matrix of the order 4 4 as shown above
[K] = a square stiffness matrix of the order 4 4 as shown above
{X} = a Column deflection matrix of order 4 1 (which means 4 rows and 1 column)
[C] = a square damping matrix of order 4 4.
The equation being statically coupled, the damping matrix is given by
[C] =
_
_
_
_
c
1
+c
2
0 c
2
c
1
e
0 c
x
0 c
x
Z
c
c
2
0 c
2
0
c
1
e c
x
Z
c
0 (c
W
f
Z
c
+c
1
e
2
+c
x
Z
2
c
)
_
_
(2.9.4)
Based on the above matrices one can now do the analysis in identical fashion as
shown earlier
35
once the initial velocity of the anvil after the impact is known.
35 Here the order of matrices being 4 4 eigen solution may be done by Bairstows method or one can
directly solve for them in solution tools in computer like MATHCAD or MATLAB etc.
Figure 2.10.1 Typical reinforcement detail for hammer foundation.
2.10 DETAILS OF DESIGN
2.10.1 Reinforcement detailing
Usually for this type of foundation moments induced and reinforcement required for
the same are nominal and nominal reinforcement is only provided.
Usually 16 mm diameter bar is placed @ 150/200 mm c/c subject to the minimum
condition that the volume of reinforcement shall not be less than 25 Kg/m
3
.
In this type of foundations, the reinforcement shall be placed along the three axis
and also diagonally at the corner to prevent shear failure.
Additional reinforcement shall be provided at the top side of the foundation block
than at other sides.
Topmost layer of reinforcement shall be provided with a cover not less than 50 mm.
The sketch in Figure 2.10.1 shows a typical detail of re-bars for hammer foundations.
2.10.2 Construction procedure
The foundation block should preferably be cast in one go. If this is not possible and
a construction joint is unavoidable then such joints shall be horizontal in orientation
and measures shall be taken to provide a proper joint by providing dowels of 12 mm
or 16 mm dia bars embedded at 60 mm center to center to depth of at least 300 mm
at both sides of the joint.
Before placing the next layer of concrete the previously laid layer of concrete should
be roughened, cleaned thoroughly and washed by water jet and then covered by a layer
of rich 1:2 cement grout (1 cement, 2 sand) at least 20 mm thick. Concrete should be
placed not later than 2 hours after the grout is laid.
2.11 VIBRATION MEASURING INSTRUMENTS
2.11.1 Some background on vibration measuring
instruments and their application
Design of machine foundation is not only a case of design of foundation and restricting
the amplitude of vibration within the acceptable limit. In many cases if there are
sensitive instruments around the foundation a foundation specialist is required to
measure the vibration amplitude of the surrounding instruments (even though the
foundation itself is safe) to ensure their safe operation. There could be properties
or heritage buildings around the foundations that would require protection from the
foundation vibration. In such cases an important function is to measure the vibration
induced by the machine and its effect on the surrounding. This is usually done by
instruments which are generally termed as vibration pick up instruments.
In this section we will see what the theoretical background for development of these
machines is and how they are used to measure such vibrations.
2.11.2 Response due to motion of the support
In many situations vibration of a system is not due to forces acting directly on the
mass but resulting from the motion of the base. Consider the situations shown in
Figure 2.11.1.
Figure 2.11.1(a) shows the basics of a ground-measuring device. A vibration meter
shown therein measures only a relative displacement of the ground. The relative motion
is usually converted to an electric voltage by making the seismic mass a magnet mov-
ing relative to coils fixed in the case shown in Figures 2.11.1(b) and (c). The electric
voltage produced is a measure of the ground displacement. These types of instruments
are called velocity meters. The voltage generated is proportional to the rate of cutting
of the magnetic field and the output of the system will be proportional to the velocity
of the vibrating body. If a rotating drum is fixed and a needle is moving on a drum,
the relative motion of the instrument-soil system will be recorded on the drum [rela-
tive displacement = [z
2
(t) z
1
(t)], z
2
(t) being the displacement of the mass]. A basic
description of a seismic pick up is shown in Figure 2.11.1(c). The relative displacement
is the e. m. f. produced in the coil is the electrical signal from it mechanical counter-
part. A hypothetical pick up is shown in Figure 2.11.1(d) wherein a magnetic material
moves up and down in a electro-magnetic field and the electrical signal produced is a
measure of the actual ground displacement.
2.11.3 Vibration pick-ups
The vibration-displacement amplitudes are most often measured in soils and founda-
tions ranging from millionths to thousandths of a centimetre and occur at frequencies
ranging from less than 10 Hz to more than 100 Hz. The instruments required to mea-
sure motions of this magnitude are designed on the basis of a single-degree-of-freedom
system. Instruments based on this design have two distinct advantages. First, in the
S.D.O.F. system, the suspended mass is used as a reference from which vibrations
are measured because in cases such as ground-motion measurements no reference is
available. The second, some electrical phenomena are readily adapted to measuring
(a). General set up
(b). General set up of vibration pick up
(c). Seismic pick up
(d). Hypothetical ground displacement measuring device
Electromagnetic field
z
2
-z
1
z
1(t)
z
2
(t)
m
Rotating drum
Displacement here is same as that of the ground;
i.e. the actual displacement is measured
Vibration meter
z
2
-z
1
z
1(t)
m
Rotating drum
z
2
-z
1
z
1(t)
z
2
(t)
m
z
2
(t)
Figure 2.11.1 Seismic pick up.
z
1(t)
z
2
(t)
m
k
c
Figure 2.11.2 Displacement pick up.
the response of the system by producing an electrical signal that can be observed with
an oscilloscope or recorded for subsequent analysis.
An instrument that converts mechanical motion into an electrical signal is called a
transducer. For vibration measurements there are three general types of transducers,
namely, velocity, acceleration and displacement transducers.
2.11.3.1 Displacement transducer
In Figure 2.11.2 a schematic sketch of a displacement transducer is shown. Here we
record Z
r
(t) = z
2
(t) z
1
(t).
Now if it so happens that
Z
r
(t)
z
1
(t)
1 and
z
2
(t)
z
1
(t)
1
then whatever Z
r
we measure will be the z
1
(t).
If z
2
(t) and z
2
(t) are velocities and accelerations of the mass, one can write the
equation of motion of the system as
m z
2
+c( z
2
z
1
) +k(z
2
z
1
) = 0 (2.11.1)
Again, let
Z(t) = z
2
(t) z
1
(t), so

Z(t) = z
2
(t) z
1
(t),
and

Z(t) = z
2
(t) z
1
(t), also, z
2
(t) =

Z(t) + z
1
(t).
Substituting above, one can write:
m
Z(t) +c
Z(t) +kZ(t) = m z
1
(t) (2.11.2)
If we assume a ground motion of the type: z
1
(t) = A sin t
m z
1
(t) = mA
2
sin t: Substituting the above, Equation (2.11.2) reduces to:
m
Z(t) +c
Z(t) +kZ(t) = mA
2
sin t (2.11.3)
Solution of Equation (2.11.3) is
Z =
mA
2
/k
_
(1 r
2
)
2
+(2Dr)
2
sin (t ) = r
2
Asin(t ) = X sin(t )
and tan =
2Dr
1 r
2
where =
1
_
(1 r
2
)
2
+(2Dr)
2
(2.11.4)
Solution of Equation (2.10.4) is shown in Figure 2.11.3. It shows that when r is
very high, Z
max
/A approaches one. That is if we choose pick-ups having a very low
n
(natural frequency of m-k system), r will be very high. This will result in Z A.
Also the phase angle , between the exciting force (ground motion) and the instrument
should be nearly zero or 180
. Figure 2.11.3 shows that is nearly 180
for large values

of r. Regarding D of the transducer, we see that the curve with D =0.6 is better suited
as the amplitude is not amplified near the natural frequency and secondly Z
max
/A
reaches unity faster.
2.11.3.2 Instrument with low natural frequency
Instruments with low natural frequency, the r-values approach a large value and the
relative displacement Z approaches A, regardless of the damping, D [Figure 2.11.3].
The mass here remains stationary while the supporting case moves with the vibrating
body. The instrument just described is the basis of what is known as seismometer.
These instruments are of large size as the relative motion of the seismic mass must
be of the same order of magnitude as that of the vibration to be measured. Since the
1
a
c
c
e
l
e
r
o
m
e
t
e
r
seismometer
0.2
0
2 4 6 8
0.707
D = 0
0
40
90
140
180
0 0.5
1 1.5 2 2.5 3
Frequency ratio, r
P
h
a
s
e

a
n
g
l
e
,
D = 0.0
D = 0.6
1
D = 0.707
D = 0.4
D = 0.05
4
6
8
Z
m
a
x
/
A
Frequency ratio, r
Figure 2.11.3 Response curves.
seismic mass (m as shown in Figure 2.11.2) is a magnet moving relative to coils fixed
in the case, the voltage generated is proportional to the velocity of the vibrating body.
The displacement and acceleration can be obtained from this velocity type transducer
through integrator and differentiator provided in most signal-recording units.
2.11.3.3 Instrument with high natural frequency
Instruments with high natural frequency, the r-values approach a very low value
and the relative displacement Z approaches A, regardless of the damping, D [Figure
2.11.3]. Again, the denominator of Equation (2.11.4),
i.e.
_
(1 r
2
)
2
+(2Dr)
2
approaches unity as r 0.
Under this condition, Equation (2.10.4)
Z
max
= Ar
2
=

2
A
2
n
=
Acceleration
2
n
,
implying that Z
max
is now proportional to the acceleration of the motion, 1/
2
n
being
the constant of proportionality. Range of accelerometer can be seen from a magni-
ed plot of
1
(1r
2
)
2
+(2Dr)
2
for various values of D. For D = 0.7, the useful range is
D = 0.65
D = 0.7
D = 0.75
D = 0.6
D = 0
(1- r
2
)
2
+ (2Dr)
2
1
0.96
0.98
1
1.02
1.04
0 0.2 0.4 0.6 0.8 1
Frequency ratio, r
Figure 2.11.4 Acceleration error with varying frequency for various D.
0 D 0.2 with a maximum error of 0.01 percent (Figure 2.11.4). Electromagnetic
type of accelerometers generally utilizes a damping around D = 0.7, which not only
extends the useful frequency range but also prevents phase distortion for complex
waves.
2.11.3.4 Velocity transducers
The displacement transducer described in the preceding may also be used as a
velocity transducer. The relation between the relative velocity amplitude and the
ground-velocity amplitude is identical to Z
max
/A = r
2
/
_
(1 r
2
)
2
+(2Dr)
2
, since
multiplication of Z
max
and A by leaves the equation unchanged.
Output from a velocity transducer is generated by coil moving through a magnetic
field as mentioned earlier (Figure 2.11.5). Since voltage induced in the coil is directly
proportional to the relative velocity between the coil and the magnetic field, either the
coil or the magnet is made part of the mass and the other component is attached to
the frame.
Phase angles : tan
1
=
2D
i
n
i
1
_

n
i
_
2
: tan =
2D
n
1
_
n
_
2
(2.11.5)
in which = operating frequency, frequency of the forcing function (say 50 Hz);
ni
= natural frequency of the instrument (say 4.75 Hz);
n
= natural frequency of
(soil mass + footing); D
i
= Damping coefficient of the instrument; D = damping
coefficient of (soil mass + footing).
For a soil-footing vibration system:
Phase angles for the above system (Figure 2.11.6) may be written as
tan
1
=
2D
i
/
n
i
1
_

n
i
_
2
; tan =
2D/
n
1
_
n
_
2
(2.11.6)
Z(t) = A sin ( t-
1
-
= relative displacement
(as Z A)

F
0
sin t
z
1
(t)
= A sin( t-
z
2
(t)
m
k c
Figure 2.11.5 Ground displacement measuring instrument.
Pick up
F = F
0
sin t
soil
footing
+
F(t), A(t)
t
A
0
sin t
F
0
sin t
Figure 2.11.6 Footing-soil system.
where
n
is the natural frequency of soil + footing system; other terms are as defined
in the preceding.
To obtain D one may use (
1
+) and (
1
), if possible.
2.11.3.5 Acceleration transducers
There is no magnetic field here. The Figure 2.11.7 shows a schematic sketch of an
accelerometer. The principle is when there is a pressure difference between the faces
it produces a voltage difference, which is a measure of the force acting and hence the
acceleration of the mass of the crystal.
For the mass m of the crystal:
m z
2
+c( z
2
z
1
) +k(z
2
z
1
) = 0; or m z
2
+c z
2
+kz
2
= c z
1
+kz
1
(2.11.7)
Assuming,
z
1
= A
1
sin t, z
1
= A
1
cos t, hence c z
1
+kz
1
= cA
1
cos t +kA
1
sin t
Thus the right hand side of Equation (2.11.7) reduces to F sin ( +
1
): in which
F = A
1
c
2
2
+k
2
and tan
1
=
c
k
= 2D

n
. The angle
1
is the angle between force
(F) and the displacement of the ground z
1
.
z
1
(t)
Piezo-electric crystal
z
2
(t)
Accln. measured
O
O
k
c
m
Figure 2.11.7 Accelerometers.
Solution of Equation (2.11.7) is given by
z
2
=
(A
1
c
2
2
+k
2
)/k
_
(1 r
2
)
2
+(2Dr)
2
sin (t +
1

2
) = A
2
sin (t +
1

2
) (2.11.8)
A
2
A
1
=
z
2 max
A
1
=
_
1 +(2Dr)
2
_
(1 r
2
)
2
+(2Dr)
2
(2.11.9)
To have A
2
/A
1
1, r-value should very low. For a typical instrument where
f
n
=3200 Hz (say) which much higher than the operating frequency normally encoun-
tered in practice. Within a range of r 0.2, such a situation is encountered. These are
shown in Figures 2.11.8 to 10.
2.11.3.5.1 Phase angles
Between z
1
and generating force:
1
tan
1
=
2D
n
;
Between generating force and z
2
:
2
tan
2
=
2D/
n
1
_
n
_
2
(2.11.10)
Between z
2
and z
1
= (
2

1
) = tan
1
_
2Dr
1 r
2
_
tan
1
(2Dr) = tan
1
_
2Dr
3
1 r
2
(1 4D
2
)
_
(2.11.11)
An ideal accelerometer is the one in which instrument mass is servo-controlled to
have zero relative displacement; the force necessary to accomplish this becomes a
measure of the acceleration.
D = 0
0.707
0
40
90
140
180
0 0.5
1 1.5 2 2.5 3
Frequency ratio, r
P
h
a
s
e
a
n
g
l
e
,

2
6
r
2 1
0
1
A
2
/
A
1
4
Range for making A
2
/A
1
= 1
Figure 2.11.8 Accelerometer response.
0.01
0.15
1
D=.05
r
0
2
4
6
0
2
4

i
n

r
a
d
i
a
n

0.5
Figure 2.11.9 Phase angles.
(
2
-
1
)
cz
1
kz
1
kz
2
z
1
z
2
F
Reference
2
cz
1
m
2
z
2
t
Figure 2.11.10 Vector diagram for phase angles.
The piezoelectric crystals are mounted in such a manner that under acceleration
they are either compressed or bent to produce an e.m.f. which is ultimately converted
to electrical signals. The natural frequencies of such accelerometers can be made very
large, say in the range of 50 000 Hz. The piezoelectric crystal mounted accelerometers
can be made very small in size, may be of the range of 10 mm in diameter and height
and are very rugged to withstand a shock as high as 10 000 g acceleration. A typical
instrument may have fn
i
1 Hz to 5 Hz useful frequency, f
operating
10 Hz to
2000 Hz which means r is more than 10. Sensitivity of such instruments may be in the
range 20 mV/(cm/sec) to 350 mV/(cm/sec) with maximum displacement = 0.5 cm
(double amplitude) [Note A
rms
= 0.707 A].
Sensitivity of crystal-type accelerometer is denoted either in terms of charge, i.e. pico-
coulombs = pC =10
12
coulombs per g or in terms of voltage, i.e. millivolts = mV =
10
3
V per g. Sensitivity of a crystal-type accelerometer can be established from: say a
typical crystal accelerometer is 25 pC/g with crystal capacitance equal to 500 pF (pico-
farads). Voltage fromthe classical equation E = Q/C, gives the sensitivity =25/500 =
0.05 V/g or 50 mV/g as sensitivity in terms of voltage. Again, if the accelerometer is
connected to a vacuum tube voltmeter through a 3 m long cable of capacitance 300
pF, the open circuit output voltage of the accelerometer is reduced to (50)(500)/(500
+ 300) = 31.3 mV/g. This loss can be avoided by using a charge amplier, in which
case the capacitance of the cable has no effect.
2.11.3.6 Amplitude distortion
Normally the measured vibrations consist of a number of harmonic motions of various
frequencies. Amplitude distortion occurs in an accelerometer if the acceleration of one
harmonic is amplified more than another. From a harmonic solution, the amplitude
of acceleration can be written as
2
A. For an equal amplification to acceleration, it
is desirable to have /
2
n
nearly same for all frequencies. For r = 0, = 1. Thus, the
amplitude distortion can be defined as the change in /
2
n
with respect to r = 0. The
percent amplitude distortion is
100
2
n
2
n
1
2
n
= 100 ( 1)
It can be observed from Figure 2.11.11 that accelerometer should be built with D
lying between 0.6 and 0.7 to minimize the amplitude distortion.
2.11.3.7 Phase distortion
This distortion occurs if the relative phase of the harmonics recorded is different from
that of the vibration to be measured. For zero distortion, the shift should increase
linearly with frequency of the harmonic motion. The phase shift at r =1 is always /2.
For zero distortion, the phase shift for 0 < r < 1 should be 90r degree. Hence phase
distortion in an accelerometer can be defined as: Phase distortion = ( 90r) degree.
It can be noticed in Figure 2.11.11 that appropriate damping in an accelerometer is
necessary for minimizing the phase distortion.
To record any complex wave without changing its shape, the phase of all harmonics
must remain unchanged with respect to the fundamental. This requires that the phase
angle be zero or that all the harmonic components must be shifted equally. The first
case of zero phase shift corresponds to D =0 for r < 1. The second case of equal time-
wise shift of all harmonics is nearly satisfied for d = 0.7 for r < 1 (Figure 2.11.12)
and when D =0.70, the phase for r < 1 can be expressed by r/2. Thus for d = 0
or 0.70, the phase distortion is completely eliminated.
A
m
p
l
i
t
u
d
e

d
i
s
t
o
r
t
i
o
n

i
n

%
D = 0.65
D = 0.7
D = 0.75
D = 0.6
D = 0
10
0
+
10
0 0.2 0.4 0.6 0.8 1
Frequency ratio, r
Figure 2.11.11 Amplitude distortion in accelerometer.
P
h
a
s
e

d
i
s
t
o
r
t
i
o
n
Frequency ratio, r
D = 0.75
0.5
0.6
0.70
-12
-10
-8
-6
-4
-2
0
2
0
0.2 0.4 0.6 0.8 1
1.2
Figure 2.11.12 Phase distortion in accelerometer.
Example 2.11.1
1 A manufacturer of vibration measuring instruments gives the following
specif ication for one of its vibration pick ups;
Frequency range: velocity response flat from 10 Hz to 1000 Hz.
Sensitivity: 0.096 V/cm/sec, both volts and velocity in rms values.
Amplitude range: Almost no lower limit to maximum stroke between stops
of 0.60 in.
a This instrument was used to measure the vibration of a machine with
a known frequency of 30 Hz. If a reading of 0.024 V is indicated,
determine the rms amplitude.
b Could this instrument be used to measure the vibration of a machine
with known frequency of 12 Hz and double amplitude of 0.80 cm?
Give reasons.
Solution:
a Voltage = 0.024 V: Sensitivity = 0.096 V/(cm/sec) : Velocity = 0.024/0.096
cm/sec = 0.25 cm/sec.
f = 30 Hz : = 2 30 rad/sec : Amplitude = velocity/ = 0.132 10
2
cm = 0.0133 mm.
f = 12 Hz : = 2 12 rad/sec : Amplitude = 0.40 cm.
Velocity = 2 12 0.40 = 30.159 cm/sec
b Now for f = 10 Hz : = 2 10, amplitude = 0.3 2.54 = 0.762 cm:
velocity = 0.762 20 = 47.88 cm/sec.
For f = 1000 Hz, velocity = 0.762 2000 = 478877.8 cm/sec
Velocity required is 30.159 cm/sec.
So the instrument cannot be used.
2 The sensitivity of a certain crystal accelerometer is given as 18 pC/g, with
its capacitance equal to 450 pF. It is used with a vacuum tube voltmeter and
its cable is 5 m long with a capacitance of 50 pF/m. Determine its voltage
output per g. Ans. E = 25.7 mV/g.
Solution:
Sensitivity = 18 pC/g, crystal capacitance = 450 pF.
Sensitivity in terms of voltage = 18/450 0.04 V/g [E = Q/C].
Total cable capacitance = 50 5 = 250 pF
Output voltage = 40 450/(450 +250) = 25.7 mV/g
3 A vibration pickup has a sensitivity of 40 mV/(cm/sec) between f = 10 Hz
to 2000 Hz. If 1 g acceleration is maintained over this frequency range,
what will be the output voltage at (a) 10 Hz and (b) at 2000 Hz. Ans. (a)
624.5 mV, (b) 3.123 mV.
Solution:
Sensitivity = 40 mV/(cm/sec) = 0.04 V/(cm/sec)
a F = 10 Hz = 2 10 = 62.83 rad/sec
Acceleration: 1 g =981 cm/sec
2
velocity =981/62.83 =15.61356 cm/sec.
Voltage = (40 981)/62.83 mV = 624.5 mV
b f = 2000 Hz = 2 2000 rad/sec
Velocity = 981/(2 2000)
Voltage = (40 981)/(2 2000) = 3.123 mV.
2.12 EVALUATION OF FRICTION DAMPING FROM ENERGY
CONSIDERATION
E = F
0
A sin , phase angle at resonance is 90
, and the energy input is F

0
A.
Energy dissipation, using friction concept is 4fA. Now, if 4f /F
0
< 1, energy input
exceeds the energy dissipation, and the excess energy accumulated over the cycles
builds up the amplitude of oscillation.
Consider an embedded footing, Figure 2.12.1.
Governing equation is
m z +c z +C
eq
z +kz = F
0
sin t (2.12.1)
Steady state solution is
A =
F
0
/k
_
(1 r
2
)
2
+
_
C
eq

k
_
2
+(2Dr)
2
sin (t ) = A
0
sin (t ) (2.12.2)
A
0
can be obtained from
A
0
=
_
F
0
k
_
_
1
_
4f
F
0
_
2
_
(1 r
2
)
2
+(2Dr)
2
. (2.12.3)
If we have two observations with forcing functions: F
1
(t) = F
01
sin t and F
2
(t) =
F
02
sin t, there will be two responses namely, A
01
sin (t ) and A
02
sin (t ).
f/2
f/2
k c
m
F = F
0
sin t
Figure 2.12.1 Vibration of a footing with side friction.
A
0
r
r = 1
Figure 2.12.2
Now using Equation (2.12.3) we can write
A
01
A
02
=
F
01
F
02
_
1
_
4f
F
01
_
2
_
1
_
4f
F
02
_
2
(2.12.4)
A
01
, A
02
, F
01
and F
02
are known and hence we can obtain a representative value of
f , the friction force.
Now, for D = 0 Equation (2.12.2) gives
A
0
=
_
F
0
k
_
_
1
_
4f
F
_
2
1 r
2
and the response is given in Figure 2.12.2.
But 4f < F
0
has to be satisfied for a real system, i.e. 4fA
0
sin < F
0
A
0
sin .
Hence, work done by the friction force is less than the work done by exciting force.
This implies building up of energy and hence a resonant situation will arise.
If f is large we have to use more exact analysis for solution as the motion cannot be
assumed to be harmonic.
2.13 VIBRATION ISOLATION
Vibratory forces generated by machines and engines are often unavoidable; however,
their effect on a dynamical system can be reduced substantially by properly designed
springs, which are referred to as isolators. Protection of the base against the action of
driving forces is called active isolation and protection against kinematic disturbances
is called passive isolation. Thus, when the noise-making source itself is isolated from
other structures, the isolation is an active isolation whereas when other structures
are isolated from the noise making sources, the isolation is a passive one. In active
isolation the basic problem is that of determining the force transmitted to the base;
in the theory of passive isolation, it is the problem of finding the amplitude of the
vibration the object is to be protected is forced into.
k c
m
F = F
0
sin t
Figure 2.13.1 Vibration isolation (active).
F
x
kx
m
2
x
F
Tr
cx
Figure 2.13.2 Vector diagram for the system shown in Figure 2.13.1.
2.13.1 Active isolation
Let us consider a system shown in Figure 2.13.1. Let F = F
0
sin t acting on a SDOF
system.
The force transmitted to the ground through springs and damper:
F
Tr
=
_
(kx)
2
+(cx)
2
= kx
_
1 +
_
c
k
_
2
(2.13.1)
The vector diagram can be shown as in Figure 2.13.2.
Solution of the problem can be written as
x =
F
0
k
_
(1 r
2
)
2
+(2Dr)
2
sin (t ) =
F
0
k
sin (t ) (2.13.2)
Thus from Eqns. (7.13.1) and (7.13.2) one can write
F
Tr
F
0
=
_
1 +(2Dr)
2
_
(1 r
2
)
2
+(2Dr)
2
(2.13.3)
This is identical to the one developed for accelerometers.
Thus the problem of isolating a mass from motion of the support point is identical
to that of isolating the disturbing forces. The ratio, F
Tr
/F
0
is called the transmissibility.
Without the loss of generality, setting D = 0, Equation (2.13.3) can be written as
F
Tr
= F
0
(2.13.4)
Hence, the efciency of active isolation depends on the magnitude of the magnica-
tion factor to be used. This requires accordingly a low value of the natural frequency
n
, which may be achieved by reducing the stiffness of the mounting of the machine
or by increasing the vibrating mass.
Transmissibility is less than unity only for r >
2. Isolation is possible only for

/
n
>
2 (refer to Figure 2.11.3). An un-damped spring is superior to a damped

system in reducing transmissibility. However, to reduce amplitude near resonance
some damping is desirable.
It should be noted that vibration isolation of slow-speed machines (when the fre-
quency of the disturbance is not high) may require a very low natural frequency and
accordingly impractically great exibility of vibration absorbers. To overcome this dif-
culty the vibrating mass is articially increased in such cases. This serves a twofold
objective; first, the natural frequency is reduced and, second, sufcient stiffness of the
system is preserved.
It is possible to reduce the amplitude of vibration by supporting the machine on a
large mass or by other means is shown in Figure 2.13.3.
Again a set of elastic constraints (vibration isolators), in the form of steel springs or
rubber elements are introduced under the frame of the machine to be isolated.
(a) (b)
(c) (d)
Machine
m
Machine
M
Figure 2.13.3 Active isolation.
Two equally efficient types of mounting are in use. These are supporting type when
vibration isolators are placed under the base of the machine (Figure 2.13.3a) and
suspension type when vibration isolators are placed above the bottom of the base in
the latter case the vibration isolators may be either in compression (Figure 2.13.3b)
or in tension (Figure 2.13.3c). If horizontal vibration prevails in the machine to be
isolated, a pendulum type suspension may be used to advantage (Figure 2.13.3d).
To keep transmissibility same, k must be increased in the same ratio so that
(m + M)/k remains the same.
Say for some transmissibility, if we increase the mass m to m+M
F
Tr
F
0
=
_
1 +(2Dr)
2
_
(1 r
2
)
2
+(2Dr)
2
(2.13.5)
k must be increased in the same proportion so that (m+M)/k remains the same.
Thus as k increases
x =
F
0
k
_
(1 r
2
)
2
+(2Dr)
2
sin (t ) (2.13.6)
x will be reduced.
When damping is negligible
Transmissibility =
F
Tr
F
0
=
1
_
n
_
2
1
(2.13.7)
where it is understood that /
n
to be used is always greater than

2.
2.13.2 Passive isolation
Consider the case in which, there is a vibration of the ground in Figure 2.11.7 instead
of the force, F.
Following Equation (2.11.9), we can write
m x
2
+c x
2
+kx
2
= c x
1
+kx
1
(2.13.8)
If x
1
(t) = X
1
sin t, Equation (2.13.8) reduces to
m x
2
+c x
2
+kx
2
= X
1
[ksin t +c cos t]
= X
1
_
k
2
+c
2
2
sin (t +) = F
x
sin (t +) (2.13.9)
where tan =
c
k
= 2Dr. (2.13.10)
(a) (b)
Figure 2.13.4 Passive isolation.
The steady state solution is
x
2
= X
1
_
1 +(2Dr)
2
sin (t +) (2.13.11)
in which
tan =
2Dr
1 r
2
(2.13.12)
The passive isolation is used to protect instruments and precision machines against
vibrations transmitted from the supporting structure. The amplitude of vibration of
an isolated object is expressed in terms of the amplitude of vibration of the base by
Equation (2.13.10). Thus a passive isolation should use the same idea of making the
mounting soft, as in the case of an active isolation. It is generally required that the
natural frequency of the isolated object shall not exceed one-fourth of the frequency of
vibration of the base. If the frequency of the disturbance is not known, it is necessary
to introduce in elastic pads in the mounting system. Thus the irregularities of a road
may have the shape of a sine curve with the wave length varying over a wide range.
Therefore, there is a real danger that the body of a moving vehicle may be in a state
of resonance; to limit resonant amplitudes the vehicle suspension is always provided
with hydraulic shock absorber which dissipate a considerable amount of energy during
vibration [Figure 2.13.4(a)]. This absorber system has a disadvantage: it does not
afford sufficient comfort of passengers when subjected to shocks which are transmitted
to the automobile body with almost no relief. To obtain the necessary softness of
the suspension it may be provided with additional flexible elastic damper shown in
Figure 2.13.4(b).
2.13.3 Isolation by trench
An exhaustive field study was carried out by Woods to examine the effectiveness
of open trenches as barrier for vibration isolation. Lamb analysed the problem of
the propagation of tremors over the surface of an elastic solid. He solved the problem
of spreading out of a symmetrical annular wave disturbance around a point source.
These waves consist of body waves, namely longitudinal wave (Primary wave or
P-wave), transverse wave (Shear wave or S-wave) and surface waves, namely, Rayleigh
surface waves (R-wave). At any point on the surface P-wave arrives rst and it under-
goes an oscillatory displacement. It is followed by a relatively quiet period till another
oscillatory displacement owing to the arrival of the S-wave. Lamb termed this phase
of motion as minor tremor. A much larger oscillatory movement is followed subse-
quently due to the arrival of Rayleigh waves termed as major tremor. P-wave travels
faster then S-wave and R-wave is slightly slower than the S-wave. As to the nature of
wave propagation, a compressional wave (i.e. P-wave) propagates radially outward
from the source as hemispherical wave front whereas Rayleigh waves propagate radi-
ally outward in a cylindrical wave front. As waves travel outward the energy density
decreases with the distance from the source of disturbance. This is known as radiation
or geometric damping. The amplitudes of compressional wave attenuates approxi-
mately in proportion to 1/r, r is the distance of the source of disturbance. Along the
surface of an elastic halfspace, the attenuation is proportional to 1/r
2
. For the Rayleigh
waves, the decrease in amplitude is proportional to 1/
r. Again, about two-third of

the total energy of vibration is normally carried through the Rayleigh wave and its
smaller decay with the distance in comparison to other waves. Thus, the Rayleigh
wave is more important for structures near the surface, particularly in the event of
earthquakes, blasts and other dynamic operations.
2.14 MACHINE FOUNDATION SUPPORTED ON FRAMES
2.14.1 Introduction
In this section we will deal with machines supported on frames. These are also
sometimes termed as frame foundations. These type of foundations usually support
equipment like steam turbines (ST), boiler feed pumps (BFP), in power plants, com-
pressors in petroleum reneries, air blowers in automobile industry etc. Though the
basic analytical principle remains the same, for the present chapter however, we will
restrict our discussion mainly to foundations related to steam turbines and Boiler feed
pump only.
Apre-requisite to this section is again Chapter 5 (Vol. 1) (basic concepts in structural
and soil dynamics) and you should have gone through the previous section on analysis
and design of block foundation.
We also envisage that you have some basic concepts on Matrix Analysis of Structure
whose concepts we are going to use quite in detail.
Turbines and Boiler feed pumps form the heart of any power plant. Thus for any
developed and developing nation, capacity of supplying unhindered energy not only
ensures a steady industrial growth, but also goes on to improve the quality of life in
a long way. The main source of this energy is obviously electricity and this is what
a turbo-generator generates, based on the electro-mechanical process. Thus if the
foundation which supports these critical machines misbehave and the machine trips
during operation, the cascading effect on the end users and the industry dependent on
the power generated could suffer severe losses. If the shortage is severe in nature, this
could even have a very adverse effect on the economic growth to a complete part of a
country.
Thus for successful operation two aspects become critical for these machines
The machine itself should run smoothly (round the clock).
The foundation supporting the equipment is capable of sustaining the various
loads coming fromthe turbine under operation as well as those that could develop
due to the vagaries of nature or otherwise like earthquake, thermal, electrical
faults, short circuits etc.
In factory, quality and performance of the machine is controllable since materials
used are all man made and all requisite appurtenances are manufactured under a
careful controlled condition.
Thus, it is not difcult to arrive at a condition in the manufacturing process where
quality of output for two machines coming out of a factory could be stated to have
identical mechanical characteristics. However, for a civil engineer designing its foun-
dation the situation is completely different. He neither has control on the subsoil on
which it is being built nor he has any control on the vagaries of nature like earth-
quake, wind etc. In addition to this he has to cater to a number of uncertain loads at
the start of his design like piping loads, stator loads, and electrical fault loads etc and
still make sure that the foundation functions within acceptable limits of engineering
norm. Considering the difcult natural parameters, enormity of the machines and risk
involved in terms of public outcry, turbine foundations still remain one of the most
difcult and challenging task in civil engineering profession.
The engineer not only needs a very specialized knowledge in various aspects of
civil engineering like structural mechanics, dynamic theories related to structures and
soil, he should also have some interdisciplinary appreciation about mechanical and
electrical aspect of the machine itself. Though advent of digital computer has made
the life much simpler in terms of accurate calculations and analyzing the output results
visually, for turbine foundations this should be supplemented with some engineering
judgment and experience. For this is a case where the computer output numbers only,
do not reflect the actual picture. The engineer has to carefully weigh the effect of the
idealization in his mathematical modeling that has created these numbers and take
design decision using his engineering judgment. So before we get into the main topic
itself. . .
We plead with our readers to be cautious with this type of foundations and not hesi-
tate to take help of engineers who are experienced and also the construction people who
has constructed and commissioned such turbines and monitored their performance
36
.
2.14.2 Different types of turbines
and the generation process. . .
Before we go into the analysis and design aspect of such a turbine foundations it would
be useful to know something about the machine itself, how it behaves and why we
36 He may be an old man not so expert with computers as our modern day engineers but remember that
his experience is worth more than a million dollar software you may write for he has a feel of this giant
who if starts misbehaving can have a very serious consequence.
take the trouble of mounting it on a frame when putting it on a massive block resting
on ground would have made our life much easier.
Irrespective of the nature of fuel like fossil fuel (coal red plant), LPG/ Naptha (like
in Combined cycle or open cycle plant), processed uranium rods (in nuclear power
plant) basic principle of operation of turbine remains the same.
The fuel is used to generate steam to a pre-designed temperature in boiler and is
allowed to expand within a turbine under pressure. This generates a mechanical energy
which makes the turbine rotate.
The turbine shaft in turn is connected through a coupling or a synchronous clutch
to a generator rotor, that is rotated by the turbine and generates electrical forces due
to mechanical movement of the generator shaft in a magnetic eld. The electric power
thus generated is transferred through bus duct connections to a primary transformer
where after stepping up the voltage it is supplied to power grid through a switch yard.
This in essence is the simplied process of electricity generation.
The machine itself is a centrifugal machine and are usually of two types
Gas driven
Steam driven
The gas driven one basically uses Naptha or natural gas as the base fuel and even at
exhaust, it contains substantial thermal energy. This is usually recycled through a heat
recovery system to further heat water into steam and is passed off to a steam driven
turbine to generate further electricity.
While gas driven turbine does not require any condenser at the gas exhaust, steam
turbines will always have a condenser connected to the steam exhaust to condense
off steam coming out of the turbine. This is collected in a hot-well from where it is
Generator on bearing Coupling Turbine on bearing
Shaft
Condenser
(spring mounted)
Figure 2.14.1 Longitudinal prole of a turbine foundation with the equipment.
further recycled back to boiler through a condensate extraction pump. For steam to
condense, usually advantage of gravity force is taken and also from plant layout and
pipe routing consideration, the condenser should preferably be positioned at a level
which is lower than the turbine operating floor level. Based on this concept the best
location for placing the condenser is usually at a location directly below the turbine.
It is for this steam driven turbines are usually mounted on frames to take advantage
of the space beneath it, while for gas turbines, as no such requirements are essential,
are usually mounted on block foundations.
Besides this, the frame mounted machines also provides easy access to electrical
connections to generator and main steam pipes. Connecting the steam pipe from the
bottom is preferable for it avoids dismantling of pipe work during maintenance; this
also prevents pipe work draining into the turbine.
A typical schematic sketch of a turbine foundation is as shown in Figure 2.14.1.
2.14.3 Layout planning
For turbines placed in a power house typical layout which is most common is as shown
in Figure 2.14.2.
Spring mounted
Boiler feed pump
(Turbine driven)
Condenser
Spring mounted
Figure 2.14.2 Typical cross section of turbine pedestal and power house.
For a typical 900 MWpower plant this is usually about 16/17.00 mabove the power
house floor level with condenser mounted on springs.
The 17.00 m level is also known as the operating floor level of the power house.
In such case, the boiler feed pump (BFP) which feeds the water to the boiler is driven
by turbine itself and usually rest on RCC block foundation mounted on springs over
steel structure.
For plants of lower capacity when the top deck height is much lower, BFP is also
sometimes mounted on RCC frames similar to the turbine.
2.14.4 Vibration analysis of turbine foundations
We stated earlier that analysis of turbine foundation is a complicated job and requires
a lot of ingenuity and deep insight to the problem of dynamics for these are machines
which are massive and rotates at a very high speed
37
.
We present here the following methods of analysis of the framed foundation:
Resonance or Rauschs (1959) method
Amplitude or Barkans (1962) method
Combined or Majors (1980) method
A 2D soil structure interaction model of framed foundation that can take into
cognizance the effect of underlying soil/pile as frequency independent springs
(Chowdhury 1984).
The 3Dnite element model of the foundation considering the underlying soil into
cognizance. This is analyzed through a computer.
2.14.4.1 Rauschs method
Rausch proposed a method where the basic criteria that needs to be satised is that the
fundamental natural frequency of the foundation should be out of tune to the operating
frequency of machine by about () 20%. He suggested a mathematical model where
for natural frequency in vertical direction for the individual cross frames self weight
and superimposed load on longitudinal girders and the load coming from the machine
is considered as lumped mass over the columns having single degree of freedom(Figure
2.14.3)
38
.
For horizontal direction he assumed the bottom raft to be innitely rigid and again
proposed a mathematical model having single degree of freedom. He also assumed
that in vertical direction the average of natural frequency of the frames is the natural
frequency of the system in that direction.
37 For 50 Hz power grids the typical RPM of machines are 3000 RPM. For 60 Hz grids the speed is about
3600 RPM. For Nuclear power plants these are about 15001800 RPM. 50 and 60 Hz are standard
Power grid cycles available globally.
38 This is surely an over simplication of the problem.
N
Typical cross frame Equivalent mathematical model
P N
Q
m=(P+Q+2N)/g
K
V
Figure 2.14.3 Idealization in Rausch method.
1 Frequency in vertical direction
As per Rausch if there is n number of frames in the foundation and if f
v
is the natural
frequency of the structure in the vertical direction, then
f
v
=
n
i=1
f
i
/n (2.14.1)
Knowing,
n
=
_
K
v
/m rad/sec
we have,
n
=
_
(K
v
g)/W
where, g = acceleration due to gravity; W = weight acting in the vertical direction.
If
st
is the static deflection of the frame then,
st
= W/K
v
i.e.
n
=
_
g/
s
t
rad/sec.
Using T = 2/
n
,
we have, T=2
_
st
g
secs; f =
1
2
_
g
st
cps, f =
60
9.81
2
st
=
30
st
cycles/min
(2.14.2)
The vertical frequency of the of individual frame in vertical direction is thus given by
f
v
= 30/
_
v
cpm (2.14.3)
where
v
= the total vertical deflection at mid-point of the cross beam in meters.
Hence, for different types of loading as shown above,
v
=
1
+
2
+
3
+
4
(2.14.4)
where,
1
=
PL
3
96EI
b
_
2 +1
+2
_
is deflection due to concentrated load;
2
=
QL
3
384EI
b
_
5 +2
+2
_
is the deflection due to uniformly distributed load;
3
=
3
5
L
EA
b
_
P +
Q
2
_
is the deflection due to shear; (2.14.5)
4
=
h
EA
c
_
N +
P +Q
2
_
is the axial deflection of column due to the
concentrated load transferred from the longitudinal girder (N)
in which, P = concentrated load from the machine; Q = UDL of the cross beam (qL);
q = self weight per unit length of the cross beam; N = concentrated load on the
column; A
b
= area of cross section of the beam; A
c
= area of cross section of the
column; I
b
= moment of inertia of the beam; I
c
= moment of Inertia for the column;
E = dynamic modulus of elasticity of the frame; h = effective height of the column;
L = effective length of the cross beam, and = (I
b
h)/(I
c
L).
2 Frequency in horizontal direction
Again considering single degree of freedom the natural frequency f
h
is given by
f
h
= 30
_
K
h1
+K
h2
+ +K
hn
W
(2.14.6)
where, W = total load of machine plus the top deck and K
hi
=
12EI
c
h
3
_
6 +1
3 +2
_
.
This method does not have any provision of calculation of amplitude and suffers
from following drawbacks:
Over simplication of the mathematical model based on single degree of freedom.
Aresonance check does not necessarily ensure that the design is safe and the ampli-
tudes are within acceptable limits, especially for low tuned foundation which has
been observed to undergo signicant displacement when the machine speed passes
through the natural frequency value during start and stopping of the machine.
It considers the bottomraft as stiff and nds frequency in translational mode only,
no rocking mode frequency has been calculated, and this could have signicant
contribution to the overall dynamic response (which of course depends on the
geometry of the foundation system).
2.14.4.2 Amplitude or Barkans method
Barkan improved upon Rauschs method by taking into consideration the following
steps for the analysis:
In vertical direction he considered a two mass lumped model for analysis.
He derived the translational amplitude by taking coupled rotation of the top deck
plain considering the top deck as rigid mass supported on a series of leaf springs
which represented lateral stiffness of the columns.
However, like Rausch he also assumed the frames to be supported on slab that
is innitely stiff and thus ignoring the effect of elastic base (soil) supporting the
bottom raft.
1 Calculation in the vertical mode
Barkan argued that under vertical mode, the transverse frame will take the deformed
shape as shown in Figures 2.14.4 and the mathematical idealization may be showed
as given in Figure 2.14.5.
m
1
/2
m
1
/2
m
2
, k
2
k
1
/2 k
1
/2
Figure 2.14.4 Transverse frame of the foundation.
m
2
z
2
k
2
m
1
z
1
k
1
Fixed at base
Figure 2.14.5 Mathematical model in vertical direction.
Since the columns are stiff and have similar inertia they would deform uniformly
under axial compression while the longitudinal girders will try to resist the flexural
deformation of the transverse beam based on their torsional stiffness.
As torsional stiffness of the longitudinal girder is much less than axial stiffness of
the columns or flexural stiffness of the transverse girder, its effect on overall dynamic
response of the system is marginal and can be neglected.
Similar to the proposition of Rausch he also suggested that the transverse frames
can be treated independent of each other in the vertical direction
39
.
Based on the above he dened the various analytical parameters for each transverse
frame as follows:
k
1
=
2EA
c
h
;
v
=
L
3
(1 +2)
96EI
b
(2 +)
+
3L
8GA
b
; k
2
=
1
v
(2.14.7)
where, G = dynamic shear modulus of concrete @ 0.5E.
Calculation of mass m
m
2
= m
0
+0.45m
b
where, m
0
= P/g is the concentrated mass of the machine carried by the beam and
m
b
= the mass of the transverse girder and
m
1
= m
L
+0.255m
b
+0.35m
c
.
in which, m
L
= mass from longitudinal girder transferred to the frame; m
c
= mass of
the column.
The natural frequency of each frame is then obtained from the equation
_
m
1
0
0 m
2
_ _
z
1
z
2
_
+
_
k
1
+k
2
k
2
k
2
k
2
_ _
z
1
z
2
_
= 0 (2.14.8)
Similarly the amplitude of each frame can be obtained based on the method we have
explained earlier
40
.
For amplitude calculation, the vertical dynamic load was assumed as
P
v
= C
i
sin
m
t (2.14.9)
where, C
i
= (R/g)e
2
m
, in which, R = weight of the rotor; e = eccentricity of the
rotor, and
m
39 For a modern day engineer this might appear as Barkan was trying to simplify the case but what we
should realize was that he did not had a desk top computer readily available on his desk nor were
computers so easily available. It was an era when most of the calculations were done manually. What
is most appreciable was that he idealized and modeled an extremely complex problem to a level which
was amenable to manual calculation and in-spite of the simplication gave results which were very
reasonable.
40 We have explained the method of calculation of natural frequency and amplitude of vibration for
harmonic load for system with two degrees of freedom quite in detail in Chapter 5 (Vol. 1).
In this case it was assumed that the top deck and bottom slab as innitely rigid in
its own plane and the columns act as leaf springs where the stiffness of the springs
tantamount to transverse stiffness of the individual columns. Figure 2.14.6 shows the
mathematical model perceived by Barkan.
Mass on each horizontal spring is given by
m
i
= m
0i
+m
bi
+0.3m
ci
+m
Li
(2.14.10)
Here the horizontal displacement is given by
hi
=
h
3
(2 +3)
12EI
c
(1 +6)
and K
hi
=
1
hi
(2.14.11)
Here the term i represents the ith cross frame of the system.
W
a
W
b
W
c
W
d
X
gb
X
gc
X
ga
X
gd
H G
C/L axis
K
a
K
b
X
hb
K
c
K
d
A
X
ha
B
X
hc C
D
W
a
W
b
W
c
W
d
K
a
K
b
K
c
K
d
H
G
Figure 2.14.6 Mathematical model in horizontal vibration.
It was argued that due to difference in geometry and shape, there will be some
difference between the center of gravity (G) and center of stiffness (H).
While the resultant of all the masses
_
N
i=1
m
i
_
will pass through the point G, there
exists another point known as the center of elasticity (H) through which resultant of
all the column stiffness
_
N
i=1
K
hi
_
will pass.
As these two points do not coincide, as such other than translation the top deck
will also undergo a rotation in the horizontal plane () which will be coupled with the
translation (x).
Taking the center of gravity as the reference co-ordinate point he obtained the
following differential equation of motion
M x +K
h
x +K
h
e = P
h
cos
m
t and J

+K
h
ex +(K
h
e
2
+ ) = M
h
cos
m
t
(2.14.12)
in which, M =
_
N
i=1
m
i
_
for N number of frames; K
h
=
_
N
i=1
K
hi
_
for N number of
frames; e =distance between the points H and G; J
=

N
i=1
m
i
X
2
gi
; =

N
i=1
K
hi
X
2
hi
;
P
hi
= C
i
cos
m
t; P
h
=

N
i=1
P
hi
and M
h
=

N
i=1
P
hi
X
gi
.
Writing the above equation in the matrix form, we have
_
M 0
0 J
_ _
x
_
+
_
K
h
K
h
e
K
h
e K
h
e
2
+
_ _
x
_
=
_
P
h
cos
m
t
M
h
cos
m
t
_
(2.14.13)
The coupled natural frequency of the system can be obtained from the equations
f (
2
) =
4
n
(
2
x
+
2
)
2
n
+
2
x
= 0 (2.14.14)
Here
x
=
_
K
h
/M,
=
_
N
i=1
K
h
i
X
2
h
i
/J
and = (1+e
2
)/r
2
where r
2
= J
/M
41
3 Amplitude of vibration
The amplitude of vibration is obtained from the expression
x =
_
e
2
r
2
+
2
x
+
2

2
m
_
P
h
M
e
2
x
M
h
J
f (
2
)
; =
_
e
2
r
2
2
x
_
P
h
M
(
2
x

2
m
)
M
h
J
f (
2
)
(2.14.15)
The net amplitude of horizontal vibration is given by
x
net
= x +X
(2.14.16)
41 Alternatively this can also be calculated based on the eigen value technique for two degree of freedom
showed in Chapter 5 (Vol. 1).
where, X
= is the farthest point from the center of gravity point G.

We can also apply the modal technique shown earlier in the chapter 5 (Vol. 1) and
directly nd out the amplitude of vibration in matrix notation too.
2.14.4.3 Combined or Majors method of analysis
Major actually combined the above two methods (Barkan & Rausch) to arrive
at a method which is usually known as the combined method of analysis.
He realized that the resonance method of Rausch and amplitude method of
Barkan are actually mutually complementary and blended the better of the two
approaches
42
.
Thus it would be worth knowing as to what improvements Major did with respect
to the previous two models.
The improvements may be summarized as follows:
Both Rausch and Barkan neglected the effect of underlying soil from their cal-
culation
43
, Major did try to cater for the effect of soil at least in vertical
mode of vibration by adding the soil deformation to elastic deformation of the
frame.
As stated earlier that resonance check does not always prove to be an ade-
quate design especially for under-tuned foundation which are found to show
signicant vibration during start and stop of the machine, Major did devise
a model where the foundation behavior under this transient can also be
checked.
These, in essence, are the two signicant contribution of Major in his combined
method.
The methodology applied in this method is explained hereafter
44
.
For vertical frequency analysis Major followed in essence the method proposed by
Rausch except that he took Barkans two-mass model as shown in Figure 2.14.5.
Here, m
2
=mass of the (upper slab +machine) +0.5 times the mass of the column;
m
1
= mass of the bottom slab + mass of the condenser +0.5 times the mass of the
column; k
2
=equivalent spring constants for the columns, and k
1
=equivalent spring
constants of the soil.
42 IS 2974 also recommends Majors method for design of the Turbo-generator foundations.
43 Though Barkan acknowledged that this might affect the response but conceded that the analy-
sis was too complex to be done manually and for very thick bottom raft, the effect of soil was
negligible.
44 We apologize, for there would be some repetition with respect to earlier method of Rausch and Barkan.
But we would still like to repeat it for rstly- the clarity and secondly to highlight what is the difference
in approach with respect to the previous two methods.
He proposed that total vertical deflection is given by
45
v
=
1
+
2
+
3
+
4
+
s
(2.14.17)
Here,
1
=
PL
3
96EI
b
_
2 +1
+2
_
is deflection due to concentrated load;
2
=
QL
3
384EI
b
_
5 +2
+2
_
is the deflection due to uniformly distributed load;
3
=
3
5
L
EA
b
_
P +
Q
2
_
is the deflection due to shear; (2.14.18)
4
=
h
EA
c
_
N +
P+Q
2
_
is the axial deflection of column due to the concentrated
load transferred from the longitudinal girder (N).
s
=
(P +Q+2N) +W
f
L
f
B
f
c
u
(2.14.19)
is the elastic deformation of soil in vertical mode.
Here, P = concentrated load from the machine; Q = UDL of the cross beam (qL);
q = self weight per unit length of the cross beam; N = concentrated load on the
column; W
f
= weight of bottom slab + half the weight of the columns; A
b
= area of
cross section of the beam; A
c
= area of cross section of the column; I
b
= moment of
inertia of the beam; I
c
= moment of Inertia for the column; E = dynamic modulus of
elasticity of the frame; h = effective height of the column; L = effective length of the
cross beam; L
f
= length of the foundation; B
f
= width of the foundation and c
u
=
co-efcient of elastic compression of the soil = I
b
h/(I
c
L).
The fundamental frequency in vertical direction is then given by
f
v
= 30/
_
v
cpm. (2.14.20)
Considering n number of cross frames, in horizontal direction, Major followed the
same procedure of Barkan, as shown in Figure 2.14.6 like idealizing the top deck as
45 This is a very interesting proposition of adding elastic deformation of the soil directly to the structure
just note it for the time being we will discuss more about it later at appropriate time.
rigid in its own plane and considering an eccentricity e between the center of mass and
center of stiffness he arrived at an expression
_
f
n
_
h
= 30
0

_
2
0

n
i=1
K
hi
n
i=1
W
i
I
h
J
cpm (2.14.21)
where, K
hi
= lateral stiffness of the ith frame i; and K
hi
=
1
hi
where
hi
=
h
3
(2+3)
12EI
c
(1+6)
;
W
i
= total weight of the ith frame plus weight of the machine plus weight of
the transverse beam and the longitudinal beams; J
= mass moment of inertia

=
n
i=1
W
i
X
2
gi
; X
g
= distance of weight W from the resultant center of mass point G
46
;
I
h
=

n
i=1
K
hi
X
2
hi
; X
h
= distance of each frame from the center of rigidity H,
and
0
=
1
2
_
e
2
n
i=1
K
hi
J
n
i=1
K
hi
n
i=1
W
i
+
I
h
J
_
. (2.14.22)
3 Calculation of amplitude
We had seen earlier in Chapter 5 (Vol. 1) that under harmonic load the amplitude of
vibration is given by the expression
x
max
=
P
0
k
sin
m
t
_
_
(1 r
2
)
2
+(2Dr)
2
_
(2.14.23)
where r =
m
/
n
and D = c/c
c
with, c
c
= Critical damping of the system and is
2
mk.
For sin
m
t = 1, we have
x
max
=
P
0
k
_
_
(1 r
2
)
2
+(2Dr)
2
_
x
max
=

st
_
_
(1 r
2
)
2
+(2Dr)
2
_
(2.14.24)
Here, we introduce a term called logarithmic decrement given by = 2D, where
= Logarithmic decrement; D = damping ratio.
Major replaced the 2D by
and dened amplitude of vibration as

x
max
=

st
_
_
(1 r
2
)
2
+
_
_
2
(r)
2
_
. (2.14.25)
46 Refer to Figure 2.14.6.
4 Under-tuned foundation
For under tuned foundation when
n
<
m
during starting and stopping of machine,
there will be a case, when for a fleeting moment
n
=
m
and as such the frequency
ratio (r) will be equal to 1.0 for that instant.
During this point considering r = 1 the amplitude of vibration reduces to
x
max
=
st
/ (2.14.26)
Major has suggested that the logarithmic decrement () be taken for concrete as
0.4 when the maximum amplitude becomes, x
max
= 7.85
st
.
5 Over tuned foundation
For over tuned foundation when
n
>
m
the maximum amplitude can be found out
from the expression
x
max
=

st
_
_
(1 r
2
)
2
+
_
_
2
(r)
2
_
(2.14.27)
where
st
=
v
or
h
as the case may be.
6 Calculation of unbalanced centrifugal force
For under tuned foundation (
n
<
m
) the centrifugal force C
i
is given by
C
i
= R
_
m
_
2
(2.14.28)
For over tuned case (
n
>
m
) the centrifugal force C
i
is given by
C
i
= R (2.14.29)
where the value of is as given in Table 2.14.1.
While in the vertical direction, Major considered the deection of individual frame
which when multiplied by the above factors gives the dynamic amplitude under
transient condition.
In horizontal direction a stick model has been considered, where the stiffness of all
frames are clubbed together to arrive at a unique value of amplitude.
Table 2.14.1 Values of .
Sl. No. rpm rating of machine
1 0.2 3000
2 0.16 1500
3 0.1 750
Since individual amplitude of each frame is also necessary he approached the
problem in the following manner.
If C =

N
i=1
C
i
denotes the total centrifugal force in horizontal direction then the
centrifugal force on each individual frame is distributed in terms of their individual
stiffness
C
i
= C
K
hi
N
i=1
K
hi
+e
CK
hi
X
hi
N
i=1
K
hi
X
2
hi
(2.14.30)
Here e
is the distance between center of rigidity X

hi
and center of the resultant of
the horizontal dynamic forces, C =

N
i=1
C
i
.
Once C
i
is obtained the deflection of the ith frame is obtained from the expression
hi
=
C
i
K
hi
(2.14.31)
with the value of
hi
, the amplitude of vibration in horizontal direction is obtained
from the expression
a
hi
=

hi
_
_
(1 r
2
)
2
+
_
_
2
(r)
2
_
(2.14.32)
Major states that since the structure is usually more flexible in transverse direction
and considering the high speed of the machine is usually under tuned in this direction
and as such it is a common practice to consider for horizontal mode
a
hi
= 7.85
hi
. (2.14.33)
7 Dynamic forces
The dynamic forces to be accounted for in structural design of the frame have been
expressed by Major as follows:
To account for idealization made in calculation of natural frequency it is suggested
to correct the calculated natural frequency by a term, f
n
= f
n
(1 ), where is a
correction factor and may be considered as 0.2.
For under-tuned foundation (f
n
< f
m
) plus signed should be considered while for
over tuned foundation minus sign to be considered
47
.
When f
n
lies between
f
m
1+
and
f
m
1
, then f
n
= f
m
.
47 This actually means Major is assuming that the frequency calculation could be out from actual by ()
20% and based on the correction factor is actually trying to develop a conservative estimate of the
dynamic force.
Table 2.14.2 Values of dynamic force under various conditions.
Operating frequency
Case of machine (rpm) Dynamic force Remarks
f

n
< f
m
3000 F = 16R
_
f
n
f
m
_
2
F
max
= 16R
1500 F = 12R
_
f
n
f
m
_
2
F
max
= 12R
750 F = 8R
_
f
n
f
m
_
2
F
max
= 8R
f

n
> f
m
3000 F =
2F
max
_
_
1
f
2
m
f
2
n
_
2
+
_
_
2
_
f
2
m
f
2
n
_
F
max
= 1.0R
1500 Do F
max
= 0.8R
750 Do F
max
= 0.5R
f
m
1
< f

n
<
f
m
1+
3000 F = 16R
1500 F = 12R
750 F = 8R
R = rotating weight on the frame.
Based on the above, Major suggested Table 2.14.2 for calculating the dynamic
forces. For vertical dynamic force that acts on the center of the transverse beam the
rotating weight on the beam only should considered as the expression R.
For calculation of the horizontal dynamic force in transverse direction total rotat-
ing weight on the transverse beam plus rotating weight on the longitudinal girder
transferred to the column shall also be considered while calculating the term R.
2.15 DYNAMIC SOIL-STRUCTURE INTERACTION MODEL
FOR VIBRATION ANALYSIS OF TURBINE
FOUNDATION
We present here a further modication of Majors method considering the effect of
underlying soil on the vibration analysis of turbine foundation. The history behind
its evolution is quite interesting and would not be possibly out of context to share
the background with you. Till 1980s Indian power industry was mostly restricted to
Turbine units having capacity up to 210 MW. These turbines were all supplied by
BHEL
48
and were prototype of LMW models used in the USSR. While the machines
were quite massive and sturdy, the foundation system for these types of machines was
usually wall mounted and not frame type. In reality they were actually massive RCC
blocks having cutouts in it for laying the piping and xing other sundry xtures. They
were generically short in height and because of their massiveness and immense rigidity
these foundations were mostly over tuned. Thus conventional theories as proposed by
Barkan/Major justied their analysis quite well.
48 Bharat Heavy Electrical Limited, they are the premier Turbine manufacturing company in India.
The rst 500 MW turbines supplied in India were from Siemens KWU
49
and had
a complete different structural conguration from the erstwhile models that were in
vogue in the industry. The top deck level was much taller (about 16.0 m); they were
much slicker and called for much more space below the turbine and truly represented
a framed structure having under tuned characteristics.
In the meantime a number of turbines in operation in various parts of India were
monitored for vibration and it was observed that some of them which were designed
as over-tuned system still showed transient excitation during start and stop of the
machine (meaning thereby that they were showing under tuned characteristics).
The question was why it was happening so? It was realized that it was possibly the
soil below the bottom mat which was participating in the vibration and changing the
characteristic behavior of the foundation.
Wedpathak, Pandit and Guha (1977) conducted vibration monitoring on various
TG foundations at different power plant in India and showed that there existed
a considerable variation in amplitudes observed in the eld and those calculated
theoretically.
The above discrepancy suggested that there was denitely a necessity to arrive at a
more realistic mathematical model to predict the response of the turbine foundations.
It also proved that the assumption made in conventional analysis by Barkan and
Major, that making the bottom raft thick- nullies any participation of the under-
lying soil in the vibration may not be true in all cases. Especially for 500 MW
class of turbine where to suppress the vibration of the underlying soil the thickness
of the bottom, mat would have to be so thick that the foundation could become
prohibitively expensive.
Moreover, due to their height and slenderness in transverse direction it was realized
that translation in this direction will also induce a coupled rocking mode in the trans-
verse plane which was not accounted for in the conventional method. Considering the
inadequacy in the conventional method in the context of present day class of turbines,
we started our investigation into this problemto arrive at a more rational model where
the contribution of the soil in vibration of such frame foundations can be catered for.
While it was always possible to solve this problembased on FEM
50
, we realized that
prior to that one should have the feel as to how the system is behaving and moreover
considering the expense incurred for doing a major FEManalysis in terms of man hour
spent in data generation, data input, checking the output and result interpretations,
was there an alternative model which would give reasonable results if needed to be
done manually or use computer to a minimum?
That was the philosophy based on which we started our quest for a solution and
the outcome is what we would like to share with you.
Unlike Majors model we consider here a three-mass lumped system as shown in
Figure 2.15.1. We use here a judicious mixture of Barkan and Majors method and
couple the soil springs based on Richart or Wolfs formulation.
49 The rst Siemens machine of 500 MW was supplied to Trombay (Tata Electric) and the second to
Singrauli NTPC.
50 This we had tackled too and will be presented at a later stage.
m
3
z
3
k
3
m
2
z
2
k
2
m
1
k
z
1
Figure 2.15.1 2D-Mathematical model for soil-structure interaction (Vert. mode).
Here for n number of cross frames, we have
m
3
=

n
i=1
(Concentrated mass of the machine carried by the transverse girder
+0.45 times self weight of the transverse girder)
m
2
=

n
i=1
(0.25 times the Mass of the transverse girder +mass fromthe longitu-
dinal girder including machine weight if any transferred to the cross frame +0.30
times the mass of the column)
m
1
= mass of the bottom slab + mass of the condenser +
n
i=1
(0.3 times the
mass of the column) + mass of the soil participating in the vibration.
For spring k
3
, for the beams we have
n
i=1
v
=
L
3
(1 +2)
96EI
b
(2 +)
+
3L
8GA
b
and k
3
=
n
i=1
1
v
(2.15.1)
where, k
2
= equivalent spring constants for the columns @

n
i=1
2A
c
E
c
h
; k
1
=
equivalent spring for the soil obtained from Richart or Wolfs formulation
51
and G =
dynamic shear modulus of concrete @ 0.5E.
Applying DAlemberts equation free vibration of the system can deduced as
_
_
m
1
0 0
0 m
2
0
0 0 m
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
+
_
_
k
1
+k
2
k
2
0
k
2
k
2
+k
3
k
3
0 k
3
k
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
= 0 (2.15.2)
51 Refer section on block foundation for the values of the soil springs.
Calculation is now quite straight forward for frequency analysis based on eigen
value solution.
Instead of soil, if the foundation is resting on piles then we can straight use pile
springs based on Novaks formulation or other methods as cited previously and use
this spring as the spring k
1
.
The above matrix on expansion will give an equation of third degree whose
characteristics roots will give the eigen values of the above problem.
2 Calculation of horizontal frequency
Based on the discussion in the preceding page, it was highlighted that major lacunae
lies in this mode. While conventional analysis considers translation and rotation in
plan it does not consider the rotation in elevation which will also get coupled when
the height of turbine deck is high. We consider all these aspects in our formulation
and develop a matrix of order 4 4 that we feel takes into cognizance all the short
comings of the conventional method we discussed. We show hereafter an analytical
model conceived to cater to all the above aspects.
While the conventional analysis considers lateral translation x and rotation in
plan it considers the bottom raft to be completely rigid and the soil has no effect on
the vibration.
Since the major horizontal motion of the machine is in the transverse direction we
have added additional degrees of freedom
u which is the translational displacement of the foundation.
For turbines of capacity 500 MW and above as the height h of the column is quite
large this will also induce a rocking of the foundation (in transverse plane) and
assigned a value .
Thus while the conventional analysis has two degrees of freedom x and , in our
model shown in Figure 2.15.2, we have four degrees of freedom, namely x, , u, .
Here, K
x
= translation spring value of soil; K
= rocking spring value of soil; m

0
,
J
= mass and mass moment of inertia of top deck + Machine; mf and J
= mass and
mass moment of inertia of the bottom raft.
To arrive at the equation of motion based on DAlemberts principle will be quite
difcult as the coupled motion is quite complicated.
So to derive the equations we use the famous Lagranges equation from the energy
principle when
d(T +U) =
n
i=1
_
d
dt
_
T
q
i
_
T
q
i
+
U
q
i
_
dq
i
= 0 (2.15.3)
T = f (q
1
, q
2
, q
3
. . . . . . . q
n
; q
1
, q
2,
q
3
, . . . . . . . . . . q
n
) and
U = f (q
1
, q
2
, q
3
, . . . . . . . . . , q
n
) (2.15.4)
The kinetic energy, T for the system is given by
T =
1
2
m
f
u
2
+
1
2
J
2
+
1
2
m
0
( u + x +h
+e
)
2
+
1
2
J
2
(2.15.5)
K
h
h
Kx
m
0
, J
m
f
, J
K
u
Figure 2.15.2 2D-Mathematical model for soil-structure interacton (Horz. mode).
The potential energy, U is given by
U =
1
2
K
x
u
2
+
1
2
K
2
+
1
2
K
h
x
2
+
1
2
I
2
(2.15.6)
Differentiating,
T
u
= m
f
u +m
0
( u + x +h
+e
) and
d
dt
_
T
u
_
= m
f
u +m
0
( u + x +h
+e
)
T
= J

+m
0
h( u + x +h
+e
) and
d
dt
_
T
_
= J

+m
0
h( u + x +h
+e
)
(2.15.7)
Similarly
d
dt
_
T
x
_
= m
0
( u + x +h
+e
) and
d
dt
_
T
_
= J

+m
0
e( u + x +h
+e
)
(2.15.8)
For potential energy, we have
U
u
= K
x
U;
U
= K
;
U
x
= K
h
x and
U
= K
h
e
2
+I
Substituting the above values in the equation

d(T +U) =
n
i=1
_
d
dt
_
T
q
i
_
T
q
i
+

q
i
_
dq
i
= 0
and writing in matrix form we have
_
_
_
_
m
0
m
0
e m
0
m
0
h
m
0
e J
+m
0
e
2
m
0
e m
0
eh
m
0
m
0
e m
0
+m
f
m
0
h
m
0
h m
0
eh m
0
h J
+m
0
h
2
_
_
_
_
x
_
+
_
_
_
_
K
h
0 0 0
0 K
h
e
2
+I
0 0
0 0 K
x
0
0 0 0 K
_
_
_
x
_
= 0 (2.15.9)
Equation (2.15.9) gives the complete free vibration equation of motion for the
turbine foundation system considering the soil springs the translation and rocking
modes.
Here, K
h
=
N
i=1
K
hi
=
N
i=1
1
hi
where,
hi
=
h
3
(2 +3)
12EI
c
(1 +6)
;
J
=
N
i=1
m
i
X
2
gi
and I
=
N
i=1
K
i
X
2
hi
(2.15.10)
Before we go further a few things needs to be noticed
The matrix is real and symmetric.
The equations are dynamically coupled thus the reference co-ordinate is the center
of rigidity and not center of mass as is the case with DAlemberts equation.
Due to dynamic coupling, the mass matrix is a full matrix while stiffness and the
damping matrix would remain in uncoupled form.
Expansion of the eigen value matrix will give a fourth order polynomial whose roots
can be found based on Bairstows method or else can be very easily solved based on
software tools like MATH CAD/ MATLAB etc.
3 Calculation of amplitude of vibration
We will use here the generic modal response technique to obtain the amplitude of
vibration using the orthogonal property of the matrix.
Moreover since the Turbine once started will continue to operate for a long time as
such the steady state response is critical and we shall ignore the transient part. Thus
in the vertical /horizontal direction, we have
[M]{
X} +[C]{
X} +[K]{X} = {P(t)} (2.15.11)

where, [M] = mass matrix of the system; [C] = damping matrix of the system; [K] =
stiffness matrix of the system, and {P(t)} = Psin
m
t/Pcos
m
t the dynamic force
with sine or cosine function for the vertical or horizontal case respectively.
Now considering the operation,
[]
T
[M][]{
X} +[]
T
[C][]{
X} +[]
T
[K][]{X} = []
T
{P(t)}[] (2.15.12)
If the total numbers of degrees of freedom is j say then we have j numbers of
uncoupled equation depicted by
j
i=1
i
+2D
i
i
+
i
2
i
= p
0
i
(t) (2.15.13)
when
j=3
i=1
i
=
p
0i
sin
m
t
_
(1 r
2
)
2
+(2Dr)
2
in the vertical direction. And
j=4
i=1
i
=
p
0i
cos
m
t
_
(1 r
2
)
2
+(2Dr)
2
in the horizontal direction.
(2.15.14)
Once the displacement in uncoupled form are known the global amplitude is found
out based on the expression, {X} = []{}.
The net amplitude at the top deck, is given by the expression
x
i
net
= x
i
+U
i
+X
hi
+h
52
(2.15.15)
It would be worth now to objectively evaluate the advantage of this method.
Some of the advantages that can be attributed to this model are:
It takes all the fundamental degrees of freedom considered by the conventional
method and also takes into consideration the effect of the soil in vibration analysis.
52 We are not trying to take a short cut. We will further elaborate the whole technique including the
complete design based on a suitable problem hereafter.
It can take both soil or pile springs (with and without embedded effect) as an input
to the overall matrix.
The calculation though more intense than conventional method it is yet amenable
to manual computation and gives the engineer a rst order feel as to how the
coupled soil-structure is behaving under dynamic loading.
It will surely give quantitatively a clear idea as to how much is the effect of soil
on the overall vibration vis-a-vis xed base frequency when the effect of soil is
neglected
53
.
It will also help in taking a better decision if further elaborate analysis based on
3D space frame model is envisaged or not.
To people of orthodox school as well as the computer buffs
54
we can assure that
this technique works quit ne.
This technique has been put into practice for a boiler feed pump framed foundation
for a power plant in India and we are happy to inform that it has been operating
smoothly without any problem for more than 15 years (Chowdhury and Som 1993).
2.16 COMPUTER ANALYSIS OF TURBINE FOUNDATION
BASED ON MULTI DEGREE OF FREEDOM
In this section we discuss the method of analysis and design of turbine foundations
considering it as a frame having multi degree of freedom through computer
55
.
In this case the steps followed for analysis of the frame foundation is as follows:
The system is broken up into three parts as shown
The super structure
The raft
The soil
We basically use here the concept of nite element to solve the above problem.
Though application of nite element is more appropriate for continuum, however
basic principle of its application is well valid for this case also.
Shown in Figure 2.16.1 is a typical conceptual model of a turbine foundation resting
on a bottom raft which in turn is resting on soil.
53 If you are solving the problem in MATHCAD/MATLAB just put Kx = 10
20
and K
= 10
20
this
will effectively make the problem a xed base one. Else delete the rows and column in the matrix
pertaining to the soil degrees of freedom and reduce it to a 2X2 matrix having x and as the active
degrees of freedom.
54 Whose staple diet is a problem having 1000 degrees of freedom. Anything less than that is
surely crude!
55 It is not that we would like to continue our designs based on a paper, pencil and a calculator at best. At
the door of the 21st century we do not want to carry the stigma of being Rip Van Winkle though we
conrm that we discourage the use of software as a black box.
Figure 2.16.1 3D computor model of a turbine frame with bottom raft and soil spring.
We discuss below step by step the concepts underlying the development of its
mathematical model for analysis in computer.
1 The super structure
What element do I choose and how many nodes do I consider?
The intuitive choice for the super-structure is obviously to model it as a space frame
where the beams and columns are idealized as beam elements having six degrees of
freedomat each node. But for modeling a turbine foundation frame there is a difference
with normal building frames.
Mathematical model for the beam and column are usually taken at the center line
of the element as shown in Figure 2.16.2.
Based on Figure 2.16.2, during computer analysis, the moment and shear output
will be given at the chosen nodes. For normal building frame this does not digress
fromthe reality much for the dimension of the columns are small. However for turbine
foundation the columns are of large dimension (usually they are about 1500/2000 mm).
During design of beams since we know that the design bending moment at support
is to be taken at the face of the column, the large dimension of the column makes
a signicant reduction in the design moment of the beam at the support. The major
advantage is that it helps in reduction of congestion of reinforcement at the beam
column junction.
As such to correctly predict this phenomenon the model should consist of three
nodes instead of one connected by rigid links as shown Figure 2.16.3.
Idealized Model
Node (Typ.)
Actual Frame
Figure 2.16.2 Idealized model of a normal frame.
Moment based on one node at C/L of beam column
Design moment at column face
Based on three node concept
Beam Node
Rigid Link
Bending Moment profile
Column Node
Figure 2.16.3 Typical connection of beam column junction with rigid link.
Note: In some software packages this may also be input as master and slave option where
the beam node is usually taken as the master and the column node as the slave node.
For the beam elements as the span by depth ratio is signicant it is preferable to
consider the shear deformation of the girder during the analysis.
The loads that are induced by the machine to the deck are mostly transferred through
the bearing/sole plate. The sole plates are not necessarily always co-aligned with beam
center line. Thus to simulate this situation two of the following techniques could
be used.
Sole plate
T=FxD d
F
F
Link
C/L
Beam
Figure 2.16.4 Vertical load acting on sole plate eccentric to the center line of the beam.
By dening the load with additional torsion about the center line of the beam
based on the eccentricity between the bearing plate and center line axis of the
beam
Providing node at the point of incident of the load and connecting this point to
the mathematical model by a rigid link as shown in Figure 2.16.4.
2 How many nodes do I consider?
Intuitively the primary choice of nodes will be the beamcolumn junctions. On identify-
ing these nodes we further break it up into two additional nodes based on the concepts
as mentioned above. Other than this points at which direct load is transferred to the
girders nodal points are to be considered also.
For members under complex loading in span the number of nodes to be provided
for each beam member should be sufcient to plot the bending moment and the shear
force diagram.
For dynamic analysis enough nodes should be considered along the length of the
beam and column so that all the modes having a natural frequency less than or equal
to the operating frequency of the machine are simulated. The lower rigid body mode
of the top deck as a unit is not affected signicantly by the number of nodes along the
length of the beams.
However higher modes simulating the differential deflection of the top deck are
affected by the distribution of nodes. If not modeled with enough nodes these modes
may be entirely missed leading to an incorrect result.
The suggested number of nodes n, to be placed along the length of the span is given
by the larger of the following two values
56
:
n
L
m
_
m
EI
_1
4
: n
1
2
+
L
m
_
m
EA
_1
2
(2.16.1)
56 The expressions are derived from frequencies of a simple supported beam in flexural and axial mode.
The basis of this expression is that if the nth natural frequency of the beam is at or below the operating
frequency of the machine then at least n mid-span nodes will be required to calculate the n modes using
the discrete model.
where, L = span of the member; E = modulus of elasticity; m = mass per unit length;
m
= operating frequency of the machine in rad/sec; I = moment of inertia about the
beam about its weaker axis of bending, and A = cross sectional area of the beam.
The nodal mass may be calculated either based on lumped mass approach or con-
sistent mass approach (Archer 1963). The consistent mass approach accounts for the
distributed mass and variation of deection along the length of the beam.
However, one major disadvantage with the consistent mass matrix is that it is a full
matrix in contrary to lumped mass which is a diagonal matrix and thus calls for more
computational effort.
It has been observed that the natural frequency obtained by consistent mass
approach is more accurate than lumped mass approach though the difference may be
small for most of the practical problems. For practical analysis of Turbine foundation
considering masses lumped at the nodes is the common industrial practice.
Once the beam elements and the nodes are chosen and their properties like moment
of inertia and sectional area etc are provided as input, the computer generates the local
stiffness matrix of each beam (of size 1212) and then based on their direction cosine
transfers the local stiffness matrix into the global axis and assembles them to form the
global stiffness matrix of the superstructure.
3 The foundation raft
The foundation raft usually consists of a slab resting on soil or pile which is about
2000/2500 mm thick.
What element to use which would be optimal as well as provide the best result is
still a debate among the nite element analysts.
Some literature (Design Criteria for Turbine Generator Pedestal, 1970) recommend
to model the raft as plate bending elements while the others (Arya et al. 1979) insist
on to model it as beam elements supported on soil springs. While some advocate to
use even 8 nodded brick element to model the raft. With such controversies prevailing
on this issue it would possibly be worthwhile to evaluate the pros and cons of each of
these elements.
Plate elements
Plate elements apparently look to be a good choice for physically, it best reflects the
continuum. But as far as mathematical formulation of plates based on Finite Element
formulation is concerned the best available element for plate bending considering its
numerical convergence is the Discrete Kirchoff Triangular (DKT) plate element. The
stiffness matrix formulation of DKT plate element is based on the thin plate theory
having three (two translation and one rotation) degrees of freedom per node. The
basic idealization is that the thickness of the plate is negligible in comparison to its
plan dimension and as such the effect of transverse shear acting along the edge of the
plate is neglected.
For the turbine raft having thickness of 2000/2500 mmit is evident that the thickness
of the raft is quite large and as such it would not be perhaps prudent to neglect the
thickness vis--vis the effect of shear strain energy contribution of the overall system.
Which catapults the problem from Kirchoff-type of thin plate to Mindlin-Reissner
type of thick plate where solution is sought taking into consideration the shear
deformation along the edge of the plate.
Though many researcher have tried to formulate these type of plate based on FEM
most of them suffers from one technical snag or other, namely
Failing to pass important patch test
Spurious zero energy mode
Being sensitive to geometric distortion and meshing.
These can lead to poor solutions and even results which are unacceptable at times
(Kardestuncer 1987). Hence, without a proper mathematical formulation of the thick
plate in hand, specially the numerical problems it can create while seeking solution to
the problem, we would suggest not to use such elements in modeling this problem.
Brick elements
Brick elements could also become a plausible choice for modeling the turbine raft.
From convergence point of view brick elements are stable and have been successfully
adapted to solve different class of problems in fracture, rock and fluid mechanics.
However it has been observed that the eight nodded brick element usually have poor
approximation capability and higher order elements having 16 or 24 nodes are usually
used for efcient solution.
But use of such higher order elements calls for a much more expensive analysis in
terms of computer time, data preparation, input, output etc and is usually not essential.
Besides this brick element suffers from one serious lacunae in terms of design. Brick
elements in most of the commercially available software give output in terms of normal
and shear stress parameters. While this is ne in terms rock or fracture mechanics
problem where design check is done against allowable stresses, for the turbine raft
design we are basically looking for output in terms of moment, shear and torsion.
To back calculate these parameters from the computer out put and subsequent
interpolation to arrive at the design moments, shears etc can be extremely tedious and
chances are very high that the engineer assigned to performthis task gets lost in a maze
of numbers and gets totally confused.
For eigen-solution though use of brick element is OK we would however suggest
users the use of brick elements for design purpose with caution for the enormous
difculty one could face in back calculating the stress output in terms of moment,
shear and torsion.
Beam elements
This brings us to the last of element in use, the beam element to model the turbine raft.
Fromconvergence and correctness of results we had already discussed in quite detail
in Chapter 2 (Vol. 1) that if properly modeled beam elements gives results which is
very close to plate elements in simulating a raft problem
57
.
Moreover for derivation of stiffness matrix irrespective of the methodology used like
moment area theorem, strain energy method or numerical methods like nite element,
the results converge to an exact solution.
57 Refer Chapter 4 (Vol. 1) on Static soil structure Interaction where we have discussed in detail the use of
beam vis-a-vis plate bending elements for simulation of rafts resting on soil.
Even when the stiffness formulation takes into consideration transverse shear defor-
mation unlike thick plate element the formulation is consistent and conforming. Finally
computer output is in terms of moment shear and torsion directly and may be directly
used for re-bar calculation without having recourse to deriving themfromstress output
unlike brick elements.
Moreover if we take the elements with reasonable mesh renements including the
transverse shear deformation into cognizance
58
, we can approach a state where the
energy compatibility in terms of external work done and consequent strain energy
induced can be well satised.
Thus in terms of ease of use as well as convergence of results beam elements do
make a very attractive choice
59
.
For the raft, as the thickness is signicant considering the shear deformation
characteristics is a must for maintaining the strain energy compatibility.
4 The soil
The basic soil parameter which needs to be known to mathematically model the soil is
dynamic shear modulus (G)
60
. The soil being a continuumitself can either be modeled
based on FEM as 3D brick elements
61
, 2D plane strain elements or discrete springs.
For modeling the soil, the choice is again multiple. However as soil itself is an
innite domain successful application of FEM has been mostly in cases where the
problem could be simulated by a two dimensional model where the soil itself has been
modeled as plane strain elements or innite nite elements to arrive at a meaningful
result.
Rarely, we have come across cases where in practical problems pertaining to soil
has been modeled in 3D elements for the effort and cost in terms of man-hour and
output interpretation can make the analysis prohibitively expensive.
For the particular case of turbine foundation analysis as we are interested to know
more about the behavior of the frame and the bottom raft rather then the intricate
behavior of the soil itself, the common practice is to model the soil as frequency
independent linear springs based on Richart or Wolfs springs as described in section
of block foundation.
For practical application this has been found to be quite adequate. More sophisti-
cated model based on frequency dependent complex stiffness is usually not warranted
in this case.
Depending on the soil stiffness and the stiffness of the raft a correction to the spring
needs to be done for correct evaluation of the response
62
.
Once the spring values are evaluated they are connected to the node of the raft ele-
ment based on usual nite element procedure to arrive at the complete stiffness matrix
58 Whose contribution becomes signicant as the ratio of span by depth reduces.
59 For protagonists of classical school this is to inform that many Turbines raft has been modeled as beam
elements which have been analyzed, designed and constructed and has successfully stood the vagaries
of nature and the test of time.
60 We have dealt this topic in detail in the Chapter 1 (Vol. 2).
61 Refer Chapter 4 (Vol. 1) where we have discussed such problems in detail in terms of static loading.
62 This we have dealt in detail in the chapter 1 (Vol. 2).
of soil foundation system. A typical example is shown in Chapter 2 (Vol. 1) where we
have solved a table top centrifugal compressor foundation based on this method.
5 The machine
Do we model the machine resting on the top deck also in our analysis?
A debate which has been in the profession for quite some time and we do not want to
pass a sacrosanct sermon on this issue.
However our objective analysis of this Shakespearean dilemma
63
is as follows:
For Turbines of low capacity (<350 MW) the foundations are usually designed
having over tuned characteristics. Moreover as the overall dimension of the machine
is also relatively smaller, as such it is reasonable to consider the whole turbine and the
generator as a rigid mass whose inertial contribution as a lumped mass is taken into
cognizance in the analysis only.
However with increasing demand for energy, power manufactures are coming out
with Turbines having higher and higher capacities.
This has made the overall dimension of the turbine larger and the foundation size
have also increased and have made it exible and more susceptible to dynamic excita-
tion. For the equipment, the main shaft which connects the turbine and the generator
has become longer, thus exible, and with increase in the operating speed a slight
imbalance in the rotating mass can induce signicant dynamic load on the shaft and
also the over all deformation of the soil, raft and the frame (specially in the exural
mode) can generate a phenomenon which is know as the bowing of the turbine shaft.
Bowing or bending of the shaft about its center line axis can create damage to
the machine components, induce large forces at the bearing and can also reduce the
operating efciency of the turbine.
Thus for larger turbines (>500 MW) it would be possibly justied to consider the
machine as an integral part of the analysis too.
For such consideration an elaborate Finite Element modeling of the turbine and the
generator is usually not warranted a simple mathematical model consisting of masses
lumped at strategic nodes connected by beams, springs, rigid links etc would usually
sufce
64
.
2.17 ANALYSIS OF TURBINE FOUNDATION
2.17.1 The analysis
The analysis is usually done in the computer in four steps:
Dynamic analysis to calculate the natural frequencies of the system to ensure that
it is out of tune to the operating frequency of the machine by 20%.
63 To be or not to be . . . .
64 At this point we would strongly recommend you to take help of your equipment specialist while modeling
the equipment connected to the super-structure.
Calculation of the dynamic amplitude to check that the same are within the
acceptable limits as prescribed in the code or as pre-dened by the equipment
supplier.
Earthquake analysis if the same is perceived critical for the foundation.
Apseudostatic analysis to obtain the design Moment, Shear and Torsion induced
in the members check the stresses induced in the different structural elements like
beam column and slabs.
2.17.2 Calculation of the eigen values
For calculation of the natural frequencies or the eigen values the rst choice the user
has to make as to how many modes do I consider for the analysis?
First three modes, six modes or twenty modes . . . We have heard variety of such
numbers
65
. Unfortunately, none of the answers are universally correct, for how many
modes are signicant for the analysis varies from case to case and it also depends on
what we are looking for in terms of cases like checking the resonance, checking the
transient response or checking the response against earthquake.
The most rational basis of choice of modes would be based on modal mass participa-
tion factor
66
which should always be the basis of arriving at the number of signicant
modes to be considered for dynamic analysis when we are doing a resonance check.
As a rst step start with say ve or six signicant modes check the frequency with
the operating speed of the machine and also at the same time check the modal mass
participation factor for these modes
67
. If the mass participation is of the order of say
50 or 60% it is evident more number of modes need to considered.
Number of modes that excite at least 95 to 99% of the mass should be the basis of
number of signicant modes to be considered in the analysis. The reason is as explained
hereafter.
Suppose for the rst six modes we nd the natural frequency of the system is below
the operating speed of the machine by 20%but it has only excited say 60%of the mass
while higher modes which are in the vicinity of the operating frequency has excited
say 89% of the mass (say the 9th or the 10th mode) it is obvious that these modes
will excite the structure much more and this we will completely miss if we restrict our
analysis to a preconceived six-mode analysis only.
The other advantage is that as the eigen values go on increasing with each mode
there will always be some value which would match or be very near to the operating
frequency of the machine. But, if nearly 100% of the mass has already participated
in the vibration in the earlier modes this will have no effect on the response of the
structure even though the frequency is in the vicinity of the resonance range.
However, this can only be predicted condently provided you know exactly how
much mass has already participated in the vibration.
65 With comments such as From my experience, Normal engineering practice, and nally From
previous experience from an engineer with 2 years of experience(!!!) etc. to name a few.
66 For details of modal mass participation refer to Chapter 3 (Vol. 2).
67 Most of the commercially available FEM and dynamic analysis software have this option as an output
for the user to check the mass participation in the X,Y and Z direction.
2.17.3 So the ground rule is. . .
Do not guess or start with the pre-conceived notion that n number of modes would
sufce check the modal mass participation factor and then decide.
To arrive at the eigen value vis a vis the natural frequency, though the basic equation
for the solution remains same i.e.
[K]{x} =
2
[M]{x} (2.17.1)
computation of
2
is surely not done in the way we have described in our ear-
lier chapter
68
. For solution of eigen-values having large degrees of freedom special
numerical techniques are usually used.
When earthquake analysis is also critical, number of modes signicant enough to
simulate the natural frequency to 33Hz should be considered for the analysis.
2.17.4 Calculation of amplitude
Once the resonance condition is checked the next step is to ensure that the amplitude
of vibrations is restricted within the acceptable limits. The techniques explained earlier
based on modal analysis and orthogonal transformation
69
is usually used to obtain
the amplitude of vibration under operating conditions.
It has been mostly seen that the response of the turbine foundation, especially con-
sidering the soil effect is usually not critical under the normal operating condition. It
is only during the start and stop of the machine when the system goes on transient
resonance that it shows signicant excitation. As explained and shown earlier, in the
previous example of the 2D soil-structure interaction model, the best technique to nd
such responses would be based on time history analysis where both the transient and
steady state response needs to be checked, to ensure that such eeting response are also
within the acceptable limits as prescribed by the manufacturer or the code of practice.
2.17.5 Calculation of moments, shears and torsion
If earthquake load is not a governing case usually an equivalent static analysis will
sufce where an equivalent static load for the induced dynamic loads is obtained,
based on magnication factors as suggested in the code.
The table suggesting such factors has already been shown earlier while describing
Majors combined method. IS 2974 usually recommends the use of this table to obtain
an equivalent static force for the rotating mass and advocates to add these loads to
other loads for an equivalent static analysis and structural design of the members.
68 Different techniques used for calculation of eigen values of the system having large degrees of freedom
has been dealt in detail in Chapter 5 (Vol. 1) and may please be referred to.
69 Refer the calculations for 2D model we have derived earlier or Chapter 5 (Vol. 1) for the details of such
analysis.
2.17.6 Practical aspects of design of Turbine foundation
Here we digress from the theoretical contemplation and enter the real world of prac-
ticing engineer to evaluate further what other parameters and decisions form the back
bone of a successful design of a Turbine foundation. As a rst step we start with a
check list to see what inputs we require to start a design.
2.18 DESIGN OF TURBINE FOUNDATION
2.18.1 Check list for turbine foundation design
1 Does the drawing furnish the overall dimension of the machine?
2 Are the anchor bolt locations, size of the bolts (both diameter and length) and
details of how it should be anchored to the foundation furnished by the vendor?
3 Does the drawing supply the height at which the centre line of the shaft of the
machine is located from the bottom of the machine frame (which will be the top
of concrete or top of grout for you).
4 Does the drawing supply you with the operating speed of the machine or the range
which should be cleared during the design of the foundation?
5 Does the top deck need to support any pipes or valves on it other than the machine
itself?
6 If so are all the loads and locations of these valves and pipes are mentioned in the
drawing?
7 Does the drawing clearly mention the unbalanced mass, eccentricity or the
dynamic loads generated during the operation of the machine?
8 Are all the cut outs in the top deck including its size and location has been made
clear in the drawing?
9 Is the location of all embedded part on the top deck including their size, location
and thickness has been made clear in the drawing?
10 Is the location of the condenser support including the load coming from it is
available to you?
11 Is the Plan area of the working platform for accessing valves and for maintenance
is made clear?
12 Different load combinations for which the turbine foundation has to be designed
specially from mechanical considerations like short circuit moments, breaking of
impeller, Thermal differential etc has been furnished?
13 Finally has the equipment supplier dened any performance criterion which needs
to be met in terms of amplitude, frequency etc.
The importance of this has already been made clear previously in the chapter of
block foundations.
14 Allowable bearing capacity of the soil.
15 Dynamic shear modulus of the soil.
16 Grade of concrete to be used.
Once the above check list is satised the engineer starts his analysis with the tentative
sizing of the geometry of the super-structure.
The guideline furnished below, are suggestive as a rst trial and the adequacy of the
same shall be checked against a thorough dynamic analysis.
The designer should give enough thought to the sizing of the equipment, its size
and clearance requirements in terms of maintenance and access during operation.
The size of such access corridor should be clearly discussed with the equipment
vendor and also with the plant operation people to nalize the overall dimension
of the top deck.
All columns should be sized in such a way that they are almost equally stressed
under vertical loads (i.e. = P/A shall be constant for all the columns as far as
possible). As a rule of thumb, the columns shall have load carrying capacity of
about six times the vertical load and shall be placed not less than 3.6 meters center
to center.
The depth of the longitudinal and the transverse beam shall be one fth the clear
span with the width equal to the width of the column. Care should also be taken
that if some anchor bolts are embedded in the beam the depth of the beam is
adequate for generating the full strength of the anchor bolts. The deflection of the
beam under static load shall be restricted to 0.5 mm.
The turbine frame should in principle act as a rigid shear frame as such the flexural
stiffness of the top deck beams shall be two times the flexural stiffness of the
columns.
The bottom of the raft shall not be placed above the level as suggested by the
geo-technical consultant where the thickness (t) of the slab shall not be less than,
t = 0.07L
4/3
, where L is the average distance between columns.
The mass of the top deck plus mass of half the length of the column shall not be
less than the mass of the supported turbine and its auxiliaries on the top deck.
The total mass of the frame plus the raft shall not be less than three times the mass
of the machine.
The stress induced in soil shall not exceed 50% of the allowable bearing capacity
of the soil.
For foundations supported on piles the most heavily loaded pile shall not carry
50% of its allowable load.
The center of resistance for the pile group or the soil shall not be more than 300mm
from c.g. of the superimposed loads.
The center of rigidity of the columns shall coincide with the c.g. of the equipment
plus the top half of the structural loads both in the transverse and longitudinal
direction. This shall be done based on the equations:
x =
n
i=1
x
i
I
x
i
_
n
i=1
I
x
i
, z =
n
i=1
z
i
I
z
i
_
n
i=1
I
z
i
(2.18.1)
where, x = co-ordinate of the center of rigidity in longitudinal direction; z =
co-ordinate of the center of rigidity in transverse direction; I
x
and I
z
= moment
of inertia of the columns, and n = number of columns.
All columns should deflect equally in vertical, transverse and longitudinal direc-
tions as far as possible when subjected to equivalent static load with a limit on
deection for all cases as 0.5 mm.
Intermediate platforms are some times provided below the turbine deck for access
frombottomand maintenance. These platforms should preferably be placed below
the high pressure turbine and should be of RCC. The beams are usually of depth
varying from 0.9 to 1.2 meter with a slab thickness not less than 300 mm. During
computer analysis stiffening effect of such platforms on the superstructure shall
be considered in the analysis and it should also be ensured that the platform itself
is not in resonance with the operating speed of the machine.
1 Loads and load combinations for analysis
This we are going to deal in some detail. For unlike normal civil engineering structure
the turbine foundation is a very specialized structure where different types of loading
arise from the mechanical and electrical aspects of the machine.
If the engineer analyzing the foundation does not have a clear idea about these loads
he may land up with an analysis which could be deemed useless.
Irrespective of how sophisticated FEM package you use or use the most comprehen-
sive mathematical model if the loading input is not correct the result output is always
useless.
The different loads which come on the turbine are as discussed hereafter. While civil
engineers are quite comfortable with loading like Dead Load (DL), Live Load (LL),
Seismic load (SL) etc., our observation is that many of them are not very clear about
the typical loads which come on a turbine foundation like condenser vacuum loading
(CVL), normal torque loading (NTL) etc and how they could effect the behavior of
the foundation.
We break up the loading in three different categories: Civil Loads; Mechanical loads,
and, Electrical loads.
a Civil Loads
This is constitutes of the following:
1 Dead Load (DL)
As the name suggests this combines the self weight of all the frame members and weight
of the foundation.
2 Live Load (LL)
The live load includes those loads that vary in its magnitude and occurrence. The
normal practice is to consider a Live Load @10 kN/m
2
, on the top deck for the analysis
and design.
If based on the maintenance concept it is expected that maintenance load and lay
down load shall also come on the top deck then they shall be considered as live load
in the design.
3 Wind Load (WL)
This is usually not considered in the analysis of Turbine foundations for in most of
the cases the TG foundation is placed inside a building (the power house) where all
the wind load is transferred to the ground through the power house structure itself.
There are exceptional cases only when it needs to be taken into cognizance
70
.
4 Earthquake/Seismic Load (SL)
This could be the major design load if the power plant is being built in area prone to
major earthquakes.
Earthquake is itself a major topic of study and we will not go in detail here on this
issue for we have dealt this issue separately
71
.
5 Loading due to Creep and Shrinkage (SCL)
This usually applies to RCC frames where after initial deection the structure
undergoes deformation under sustained loading.
This time dependent deflection at the bearing location can be two or three times more
than the short term elastic deflection. However this phenomenon was not considered
earlier for design due to the following reason. For a typical coal red power plant the
initial machine alignment use to occur at about 24 to 36 month after the foundation
has been constructed. By this time most of the deformation due to shrinkage and creep
would have taken place thus further deformations were negligible and had practically
no effect on the shaft alignment.
However under present scenario with demand in power on the rise globally the
turnkey contractors are expected to nish and hand over one whole plant in 20 to 29
months only. As such it is obvious now the loading on the turbine frame would come
much earlier when the secondary deformation effect of creep and shrinkage could be
signicant at the bearing level and should be carefully evaluated.
b Mechanical Loads
1 Machine Dead Load (MDL)
This constitutes of the weight of the various turbine components and is usually termed
as the machine dead load. The turbine manufacturer in their equipment layout drawing
usually supplies these loadings and their locations.
2 Condenser Dead Load (CDL)
We had already explained earlier that the condenser is normally mounted below
the turbine top deck. Depending upon the supporting system used for installing the
condenser the loading induced on the foundation varies. The common practice for
installing the condenser is either of the two systems as discussed hereafter:
The condenser is spring mounted on the bottom raft while the top neck is rigidly
connected to the turbine exhaust nozzle. The springs are of adjustable type enabling
themto transfer specied loads to the turbine exhaust nozzle. They are also sometimes
70 There are cases where the turbine deck is spring mounted and rests on steel columns, which in turn is
connected to the power house structure. In such cases WL load has to be taken in consideration in the
analysis specially the load combinations. In such case usually a combined power house and TG frame
analysis is carried out.
71 For more detail on this issue refer to Chapter 3 (Vol. 2).
used to balance the loading eccentricity that can develop due to circulating water
pressure loads.
The condenser bottom is mounted on a rigid frame and an expansion joint is pro-
vided between the condenser and the turbine exhaust nozzle to relieve the thermal
expansion and variations in the condenser loads.
For condenser mounted on rigid frame the total weight of the condenser is trans-
ferred to the bottom raft. For spring-mounted condensers, it is mostly welded to the
turbine exhaust nozzle when the proportion of load that will be distributed between
the top deck and the bottom raft depends upon the stiffness of the spring and their
alignment. The equipment supplier usually supplies this loading.
3 Condenser Vacuum Load (CVL)
For condensers mounted on rigid frame we had already mentioned that an expansion
joint is provided between it and the turbine exhaust nozzle, for this the difference
between the atmospheric pressure on the casing of the turbine and the vacuumpressure
inside the condenser develops a force on the turbine. This load can be several times in
magnitude to the weight of the condenser itself and is transmitted to the foundation
through the turbine soleplates. The turbine manufacturer provides the distribution of
this loading.
For spring mounted condensers when the condenser is rigidly connected to the
turbine exhaust nozzle no vacuum load is transmitted to the turbine top deck.
4 Normal Torque Load (NTL)
The steam expanding within the turbine imposes a torque on the stationary casing
in the opposite direction of the rotation of the rotor. The magnitude of the torque
depends on the angular speed and the power output of the turbine. The equipment
vendor usually supplies this load in the vendor drawing as equivalent vertical loads on
the sole plate.
5 Other Equipment Loads (OEL)
Other than the turbine itself the foundation may support other equipment such as
turbine stops, control valves, interceptor valves, main steam pipeline hangers etc.
Thus additional dead loads from these, which are not included under the heading
MDL, shall also be considered in the design.
6 Thermal Load (ThL)
During operation of the turbine, temperature change of the turbine and the generator
causes expansion and contraction to take place resulting in various parts to slide. As
the progressive heating of the machine take place the turbine shaft expands, however
the expansion does not induce any loading on the foundation for the shaft is xed
longitudinally by single thrust bearing when the shaft slides freely across the journal
bearings which are adequately lubricated.
Unlike the shaft during the heat build up in the system during operation the turbine
casing also gets heated and imposes thermal loading on the foundation. The transverse
beams usually support the sole plates supporting the high pressure and the intermediate
pressure turbine casing. The low-pressure turbine casings, the generator casing and the
exciter are supported on the sole plates of the longitudinal and the transverse beams.
During the heat built up the casing expands from their anchor points thus producing a
friction load. Though exact calculation of such forces is very difcult for it depends on
a number of factors however the common practice is to use the following simplifying
analysis in lieu of an exact analysis has been found to be adequate.
The total thermal loading in longitudinal or transverse direction is considered as the
vector sum of the forces acting on that direction.
The magnitude of the force on any sole plate is calculated as:
Force = x
(Here x is sum of machine dead load, condenser load, normal torque load and piping
load if any), where, = coefcient of friction which varies from 0.2 to 0.5. The value
has to be conrmed with the turbine manufacturer.
At the startup condition the expansive load shall be taken as acting away from the
center line of the turbine while during shutdown it will considered acting towards the
center line of the turbine.
In case of the longitudinal expansion an approximation is made to the direction of
the force and the unbalanced force between two anchorage points, which prevent the
movement of the turbine, is applied as the concentrated load at the anchorage points.
7 Turbine Casing Pipe Load (TCPL)
The pipes connected to the turbine casing also induce loads to the foundation. The
turbine generator manufacturer to prevent distortion or overturning of the turbine
components species maximum loads. The turbine casing may be assumed to be rigid
and the forces are then calculated at the support points on the foundation. The types
of piping that generate most of the loads are: main steam inlet piping; reheat steam
piping, and extraction steam piping.
8 Piping Load from Equipment Attached to the Foundation (PEL)
As we had stated earlier that various auxiliary equipment are also supported on the
turbine deck. Positioning and aligning piping for this equipment creates erection forces.
Turbine piping is assembled and welded to these equipment and is anchored to the
foundation. The remainder of the steaminlet pipes is then welded to the assembly inlet
connections. Different forces are created due to thermal expansion during operation.
Erection forces, static and dynamic forces should be evaluated to check if they have
any signicant contribution or not. For instance a rapid closing of the steamstop valve
attached to the foundation can induce a major loading.
9 Load due to Machine Unbalance (MUL)
Irrespective of however care is taken in balancing the turbine generator rotor it practi-
cally impossible to do away with some imbalance in force which it will generate during
its rotation.
The magnitude of this imbalance depends on a number of factors like design
considerations, installation and maintenance procedures. The factors which usually
contribute to such imbalanced dynamic load can be summarised as follows:
i Axis of rotation eccentric to the center of mass of the rotor;
ii Deflection of the shaft due to gravity load;
iii Uneven thermal expansion;
iv Misalignment during installation;
v Normal wear and tear during operation and,
vi Corrosion.
The combined or few of the reasons as mentioned above contribute to the dynamic
imbalance in the in the rotating shaft which is synchronous with the shaft rotational
speed. These forces are transmitted through the bearing shaft to the foundation.
The dynamic load is dened by
P
dyn
= m e
2
(2.18.2)
Here, m = unbalanced mass of the rotor; e = eccentricity of the rotor shaft, and
10 Load due to Bowed Rotor (BRL)
A bowed rotor can impose large dynamic loads on turbine generators foundation.
The bowed condition of the rotor will create unbalance force which are transmitted
through the machine bearings to the sole plates. The magnitude of the force will vary
with the unbalanced dynamic force as mentioned above.
The phenomenon can happen due to:
i Failure to put the rotor on turning gear when the machine is shut down;
ii Deflection of the raft, soil and the frame in flexural mode;
iii Water Induction and
iv Very severe packing rub.
The largest bowed rotor response occurs at the rst critical speed for the rotor. The
time taken by the turbine rotor to pass through the critical speed is shorter when going
on-line.
However it takes much longer time when it goes off-line and the rotor coasts through
the resonant speed. Since this is a condition that usually requires turbine generator shut
down it will exist only for the time required for the rotor to coast down to rest. Thus
it is sufcient to ensure that the foundation stresses are low enough to eliminate the
chance of any permanent damage to the structure during the shut down period.
The magnitude and the location of the bowed rotor is usually supplied by the man-
ufacturer of the turbine in question and is dependent on the specic assumption made
by the vendor.
The force due to bowed rotor is function of the unbalanced dynamic force
P
dyn
= m e
2
(2.18.3)
The loading is normally provided in the formof a sinusoidal function for the dynamic
analysis or an equivalent static load for simplied analysis.
It is to be noted that, some turbine manufacturer may not supply this load for
depending on their own design some consider bowed rotor as worst case of accidental
loading for the high pressure and intermediate pressure turbine while others consider
loss of turbine blade as worst case of accidental load that can come on the foundation
as an emergency load.
11 Load due to Missing Rotor Blade (MRBL)
A turbine rotor must be balanced dynamically to ensure satisfactory operation and
cause no adverse effects on the turbine equipment or the foundation.
However it has been observed from previous operational experience that in some
cases (though very rare) that the last row of blade in a low pressure rotor breaks loose
from the rotor and causes a severe imbalance in the system.
The lose of blade which can vary in size from 500 mm to 1000 mm in length
can cause substantial force on the rotor, the bearing and the foundation system. The
magnitude of this unbalance is a function of the rotor blade weight, its center of gravity
with respect to the rotor and rotational speed of the rotor. As this can happen with
any of the several rows of last row blades a separate analysis should be made with a
single unbalance equivalent to the loss of one last row blade applied to the mass point
corresponding to each of the last row blade in each low pressure turbine.
Since this is an emergency situation and will require the turbine to be shut down it
will only exist only for the time period required for turbine to come down to a stop.
Thus it is sufcient to ensure that the stresses in the foundation are low enough to
preclude any permanent damage during the coast down period.
The loading is normally provided by the vendor in the form of a sinusoidal function
for the dynamic analysis or an equivalent static load for simplied analysis.
12 Electrical Loads
1 Generator Emergency Torque (GET)
Of all the loads that can occur a line-to-line short circuit at the generator terminal
causes the most severe loading of the turbine generator loading. Such a fault occurs
when any two of the three generators phase are shorted. The calculation of the maxi-
mum generator air gap torque during symmetrical three phase and unsymmetrical line
to line or line to ground terminal short circuits is normally performed assuming no
electrical damping in order to obtain greatest possible forces that can be transmitted
to the foundation under different fault condition.
Experience and previous data shows that the maximum torque resulting from a line
to line short circuit is about 25% greater than that caused by a single terminal to
ground fault and roughly 30% more than that with a symmetrical 3 phase fault at the
terminal of the generator.
The vendor in the form of a forcing function or an equivalent static force normally
provides the loading due to generator short circuit.
The use of equivalent static force for the maximum short circuit torque assume that
the foundation is innitely rigid and thus must directly absorb the full impact of the
severe shock forces.
Since this assumption may result in over designing the foundation the more realistic
approach of a dynamic analysis is on the basis of the short circuit moment as a time
dependent function is usually preferred.
In view of the very severe transient nature of the maximum short circuit loading the
foundation in the vicinity of the generator, the designer should performan appropriate
dynamic analysis of this abnormal load case.
2 Load combination for design
The following load combination is generally considered for the design as per American
practice
72
.
i Operating conditions
The loading condition for which the foundation has to checked for and designed is
= 1.4(DL+MDL+OEL+PEL +CDL+CVL)+1.7(LL+NTL+ThL+TCPL+MUL)
ii Accident Conditions
Generator Emergency
= DL+MDL+OEL+PEL+CDL+CVL+LL+GET+ThL+TCPL+MUL
Bowed Rotor case
= DL+MDL+OEL+PEL+CDL+CVL+LL+NTL+BRL+ThL+TCPL
Missing Rotor Blade
= DL+MDL+OEL+PEL+CDL+CVL+MRBL+NTL+ThL+TCPL
Seismic load
= 0.75 [1.4(DL+MDL+OEL+PEL+CDL+CVL)+1.7(LL+NTL+ThL+TCPL
+MUL+1.1 SL)
It is to be noted that 1.4 and 1.7 are load factors for design of concrete section based
on ACI-318. For design of sections based on other codes like IS or BS appropriate load
factors in place 1.4 and 1.7 has to be taken.
2.18.2 Spring mounted turbine foundation
In this section we discuss the method of analysis and design of turbine foundations
mounted on springs.
This is a practice which is quite common in European countries and is being put to
increasing use in this part of the world nowa days specially for foundations supporting
Turbines of high capacity.
In this case the top deck is usually mounted on springs of pre-designed specication
and is supported in turn on a frame as shown in Figure 2.18.1.
From the conceptual Figure 2.18.1, the obvious question that comes to mind is why
do we do such a thing and what advantage we gain from it?
To understand this we take up hereafter a concept which is otherwise known as
vibration isolation.
72 IS 2974 Part III though discusses the vibration analysis in detail it is silent on how and what load
combinations should be considered for design.
Top Deck
Mechanical Springs
Frame supporting
Top Deck
Bottom Raft
C.G of top deck
Figure 2.18.1 Spring mounted turbine foundation.
2.18.2.1 Theory of vibration isolation
We had seen in the section for analysis of block foundation that under dynamic load
the amplitude of vibration is expressed by the formula
z
=
P
0
K
z
sin
m
t
_
(1 r
2
)
2
+(2D
z
r)
2
(2.18.4)
where r =

m
n
; D
z
= damping ratio, and K
z
= equivalent spring of the soil.
Now instead of soil spring if we support the block on some mechanical springs (K
s
)
only the amplitude of vibration of the spring can be expressed as
s
=
P
0
K
s
sin
m
t
1 r
2
where r =

m
n
. (2.18.5)
Now, K
s
s
=
P
0
sin
m
t
1 r
2
= P
s
=
P
0
sin
m
t
1 r
2
where P
s
= K
s
s
(2.18.6)
Considering, T
F
= P
s
/P
0
= transmissibility factor, we have
T
F
=
1
1 r
2
(2.18.7)
The transmissibility factor is thus a measure of how much of the dynamic force is
transmitted to the supporting springs.
For transmissibility in the range less than unity the above equation is written in
the form
T
F
=
1
r
2
1
(2.18.8)
Considering the limiting case of T
F
= 1 we have, r
2
1 = 1 i.e. r =

2.
Thus it is seen that the transmissibility factor T
F
shall have a value less than unity
for all values of r

2.
For damping prevalent in the system the transmissibility factor is given by
expression
T
F
=
_
1 +(2D
z
r)
2
_
(1 r
2
)
2
+(2D
z
r)
2
(2.18.9)
where r =

m
n
and D
z
= damping ratio.
If we plot the above equations for different values of frequency ratio and T
F
we
have curves as shown in Figure 2.18.2.
Observing the curves, it will be seen that even with damping existing in the system
T
F
value is less than 1 when the frequency ratio r

2 i.e. the force transmitted to
the support is less than the induced dynamic force.
To get a further insight into how the frequency ratio affects transmissibility factor
we study an expression called isolation efciency expressed as
I =
r
2
2
r
2
1
100 in % where r =

m
n
and is the measure of the reduction of Trans-
missibility factor of the system (Crede 1951).
We plot a curve, shown in Figure 2.18.3, based on the above expression.
Based on this gure we nd that when frequency ratio is

2 the isolation efciency
is 0% i.e. 100% of the dynamic load gets transmitted to the support.
However when r = 2.45 the reduction efciency increases to 80% i.e. a signicant
amount of reduction of force transmittal to the support system is obtained.
0
2
4
6
8
10
12
0
0
.
2
5
0
.
5
0
.
7
5 1
1
.
2
5
1
.
5
1
.
7
5 2
2
.
2
5
2
.
5
2
.
7
5 3
3
.
2
5
Frequency Ratio
T
r
a
n
s
m
i
s
s
i
b
i
l
i
t
y

F
a
c
t
o
r
Damping ratio @ 5%
Damping ratio @ 10%
Damping ratio @ 15%
Damping ratio @ 20%
Damping ratio @ 25%
Damping ratio @ 30%
Figure 2.18.2 Variation in transmissibility factor.
Isolation Efficiency(%)
0
20
40
60
80
100
120
1
.
4
1
1
.
7
5
2
.
1
2
.
4
5
2
.
8
3
.
1
5
3
.
5
3
.
8
5
4
.
2
4
.
5
5
4
.
9
Frequency Ratio
I
s
o
l
a
t
i
o
n

e
f
f
i
c
i
e
n
c
y
(
%
)
Isolation
Efficiency(%)
Figure 2.18.3 Isolation efciency (%).
We give below some data showing variation of Isolation efciency with respect to
the frequency ratio
Frequency ratio 1.414 1.5 2.0 2.5 3.0 3.5 4.0 5.0
Isolation efciency(%) 0.0 20 66.66 80.95 87.5 91.11 93.33 95.8
It will be observed both from the above gures as well as from Figure 2.18.3 that
up to a frequency ratio of 3.0, the reduction in transmitted force to the support is
signicant but beyond that as the curve flattens asymptotically not much reduction in
transmissibility is obtained.
For instance if we increase the frequency ratio from 3 to 5 say the variation in
isolation frequency is only 8.6% however the manufacturing cost for such mechanical
springs as per some vendors nearly gets doubled.
Thus the common practise is to restrict the frequency ratio to maximum between 3
and 4 in practical engineering design.
Hence it is seen that if we can provide elastic supports like springs belowa foundation
and can maintain a separation ratio of 3 to 4 with respect to the operating frequency
of the machine following advantages may be obtained
The dynamic force transmitted to the supporting system for the springs could be
signicantly reduced.
Based on the reduced dynamic force it is possible to restrict the amplitude of
vibration to manageable limits.
The foundation remains isolated/de-coupled to the surrounding and does not
transfer any dynamic load.
The above points are in a nutshell major advantage gained by providing springs for
vibration isolation.
Moreover as the springs are man made (unlike soil where we do not have any
control on its property) under a careful controlled condition, it is possible to design
these springs in such a way that they do have a frequency ratio between 3 to 4 with
the operating frequency of the machine.
2.18.2.2 Effect of damping on the transmissibility factor
Since any physical system in this world has some amount of damping (even air) it
would be worthwhile to evaluate how damping affects the transmissibility coefcient.
Freq. ratio Damping ratio 0.0 0.05 0.10 0.15 0.20 0.25 0.3
2.0 T
F
0.333 0.339 0.356 0.381 0.412 0.447 0.483
3.0 T
F
0.125 0.130 0.145 0.167 0.193 0.221 0.251
4.0 T
F
0.066 0.072 0.085 0.103 0.125 0.147 0.171
Studying the above table it will be observed that having high damping value in the
system is counter productive to transmissibility. On the contrary a little amount of
damping in the system is advantageous in terms of transmittal of dynamic forces to
the foundation.
On the other hand, as we know that amplitude gets reduced due to the effect of
damping in the resonant zone the most ideal damper that can be introduced in a
system should thus have the following properties:
High damping value when the frequency of the machine is passing through the
resonant range.
Nominal damping value when the machine is operating at its normal speed.
Vendors specialising in supplying such viscous-dampers have their patented products
which exhibits such property as discussed above thus suppressing the dynamic effect
of the machine on the foundation and to its surrounding considerably.
2.18.2.3 How springs affect turbine foundation?
Based on the above theoretical discussion the query is but obvious.
For Turbine foundation as shown in the Figure 2.18.1, the top deck is usually
mounted on the spring and whole spring mounted assembly is then supported on a
frame which could be made of either RCC or steel.
The springs are supported in such a manner that the support points match with the
c.g. of the top deck and machine. This helps in suppressing the coupled translation and
rocking mode of the top deck. The technique of providing spring mounted Turbine
top deck is mostly in vogue in Europe where the Germans pioneered this technique
about 40 years ago.
Surprisingly in spite of certain advantage it provides (specially for turbines oper-
ating in Nuclear power plants) in terms of cost, plant layout etc it has not
been a popular concept in USA where engineers still opt for conventional framed
foundations.
Conventional turbine frame foundations, usually calls for columns of heavy section
and also a huge base mat to suppress the dynamic effect.
When the top deck is mounted on springs the major advantage is that the dynamic
effect of the machine is restricted up to the spring part only and the rest of the
foundation needs to be only designed for static loading.
The obvious advantage is that it calls for much slicker frame resulting in considerable
saving in material cost and as far as analysis is concerned, uncertainties prevalent with
a comprehensive dynamic soil structure interaction analysis (specially if resting on
piles) for such complex system is not required.
For nuclear power plants, the operating frequency of turbine is usually around
1500 rpm unlike conventional power plants (where it is about 3000 to 3600 rpm)
thus while designing the pedestals for these foundations engineers faced difculties
to keep them signicantly away from the operating frequency of the machine as they
were becoming far too flexible to their discomfort. The obvious choice was then to
mount them on springs and isolate the rest of the foundation from the dynamic effect.
Though the above was a starting point of such concepts, spring mounted
turbine foundations are now quite common in conventional fossil fuel power
plant also.
The major advantages gained in this case can be summarised as given hereunder:
The top deck remains dynamically uncoupled with respect to the supporting frame,
thus the supporting frame is only subjected to static load and needs to be designed
accordingly.
This makes the supporting framed structure slicker and also does away with
the necessity of providing a heavy bottom mat which is otherwise essential for
a conventional frame foundation.
The springs are capable to certain extent adjust themselves to cater to the
differential settlement, if any.
Even due to the overall settlement of the foundation which can cause additional
stress to the critical pipe connection, adjustment can be directly made using the
springs to adjust the levels and that too without interrupting the operation of the
machine.
More space is usually available belowthe foundation thus maintenance and laying
of piping and cables become more accessible and easy.
Substantial gain in material and cost is evident. Some vendors claim that with
spring mounted turbine foundations saving in cost could be as high as 45% when
compared to conventional frame foundation.
Cost of piling is reduced as there is a signicant reduction in weight.
No dynamic loads need to be considered for the piles.
The structural uncoupling of the top deck allows for the use of even steel structures
for the supporting frames.
Use of steel structures gives additional advantage in terms of construction sequence
for they can be installed parallel to the power house structure which gives a
signicant saving in construction time.
Differential settlements can be easily measured based on the variation of spring
heights. Instrumentation techniques are available which monitors these spring
heights and when it exceeds preset-values automatically give visual signals or sends
alarms.
The advantages as mentioned above are making this concept progressively pop-
ular in the Industry. In many projects in India also this concept has been put to
practice and the turbines are found to be operating quite smoothly without any
hindrance.
2.18.2.4 Mathematical modelling of spring mounted turbines
The intuitive computer model that could be conceived for this case is to conceive
spring elements connected to the top deck directly supported in the bottom frame
[Figure 2.18.4].
However other than ANSYS most of the normally available structural engineering
package do not have the provision of adding springs directly between two members
(the basic pre-condition is one end of the spring should be xed and not an active
node).
Top deck
Spring Elements
Supporting frame
Figure 2.18.4 Actual model of the top deck mounted on spring.
In such cases the most effective way to model the spring would be to represent
it by equivalent truss element having stiffness as AE/L, where AE/L shall have a
magnitude equal to the individual spring stiffness as considered in the vendors
catalogue.
2.18.2.5 Turbine foundations concrete versus steel structure
In USA as well as India conventional Turbine Foundation design is still dominated by
RCC structure.
However in many countries in Europe (especially Germany, France, Hungary etc.)
and Canada, Turbine foundation made of steel has been successfully implemented.
One of the major advantages with the RCC is its high damping property and not
requiring a very sophisticated construction technology to construct it. At the early
stage of advent of turbine foundations thus RCC made a very attractive choice.
However with turbine capacities increasing progressively the size of the turbine
foundations are also getting bigger and the construction technology is getting more
and more complex.
One of the major requirements of casting of RCC Turbine Foundation is that it
should be preferably cast in one go. Else additional steel has to be provided at such cold
joints and additional care has to be taken during construction to ensure its monolithic
property.
While for a foundation having concrete volume of 500 m
3
this was not a very
difcult task, but when foundation capacity gets increased to 1500 m
3
or more it
surely becomes a different ball game.
Firstly one needs a complete batching plant to be erected at the turbine foundation
site capable of supplying continuously concrete of same quality. This requires a very
elaborate arrangement to be made by the contractor at site including a building of a
make shift testing laboratory, where samples are collected and tested from different
batches continuously to ensure that the concrete is of the desired specied quality.
The next major difculty encountered while casting is the heat of hydration which
RCC generates during hardening.
When the volume of concrete is large the heat of hydration can be substantial to cre-
ate cracks at the surfaces and needs to be carefully controlled at site (usually controlled
by using pre xed quantity of Ice blocks in lieu of water) to nullify its effects.
The time consumed for laying reinforcement is substantial and needs to be
thoroughly checked with respect to drawing.
For large turbine foundations (>500 MW) at times client also insists in his contract
that the contractor to ensure based on non-destructive test that there are no voids or
honeycombing within the concrete. This calls for expensive ultra-sonic taste of the
foundation which is not only an expensive exercise but time consuming too.
As far as design aspect is concerned one of the major difculties encountered is the
laying of the embedded plates and hangers in top deck for pipe supports which are
usually large in numbers.
Firstly when the turbine foundation is getting designed the critical steam piping
design is yet not nalised and thus the location of embedded plates and supports
furnished by the piping engineer is only tentative and could be subjected to change.
This surely makes the foundation designers task a difcult one.
For once the turbine deck is cast no perforation or anchoring to the deck is allowed
which could be detrimental to the concrete strength
73
. Thus the engineer has to be
doubly careful with this drawing and check it very thoroughly to ensure that not a
single plate is missed, thus time taken for engineering is more.
A general tendency is thus to provide more number of embedded plates or plates
of bigger size to cover the eventualities. Since this is completely dependent on the
engineers personal judgement it has not been uncommon that at times the tonnage
has become as high as 30% extra then the estimated value and incurred unnecessary
wastage.
Similarly for any valves or other sundry equipment resting on the turbine top deck
their location and anchor bolt details etc needs to be nalised during drawing prepara-
tion stage. If the procurement department has not nalised with these equipment or the
vendor drawings are not available the design engineer could be in a lot of difculty.
With steel structure on the contrary most of the difculties as discussed above is not
encountered. In spite of the fact that steel structure provides low damping, for large
turbine foundations steel as a construction material do have some distinct advantage
over RCC.
Firstly every thing need not be erected at the site; the top deck, which generally
consists of a rigid grillage, can be constructed at the shop under a careful controlled
condition and be carried to the site and erected over the columns.
As welds are susceptible to rapid fatigue failure under dynamic loads due to
reversible of stress the connections are usually bolted (bolted connections also provide
good damping and is more advantageous in such cases) and providing site connected
bolting is not a problem at all.
The major advantage in terms of construction is that the elaborate arrangement
one requires for RCC structures in terms of inspection and checking of laying of re-
bars, controlling the concrete quality and large amount of human resources one has
to deploy at the site is not required at all.
In fact the fabrication of the top deck at shop can start much ahead of the erection
of powerhouse it self and can be erected at site simultaneously.
This signicantly saves construction cost as well as time too.
From design engineering point of view one need not worry about the location of
embedded plates and hangars, even with very late information welding locally steel to
steel is never a problem unlike anchoring plates on concrete top deck.
It can be logically perceived that steel foundation would be relatively high tuned one
compared to RCC foundations due to its lower mass. However they can be suitably
designed and adjusted to have the requisite frequency separation of 20%.
As we had stated earlier that turbine foundation usually does not become critical
during its normal operation but shows signicant excitation during the start and stop
of the machine (mostly due to the soil participation) if the amplitude of vibration can
limited within the acceptable limit steel structures do have a very high potential as a
construction material for such type of structures.
73 Though technology exists where embedded plates can be anchored to concrete slabs after it is cast but
considering the critical nature of the turbine foundation such processes are usually not allowed for
Turbine top deck by the client and is not a good engineering practice too.
One of the major constraints encountered for turbines mounted on steel structures
is however the limitation in the available of ready made rolled sections.
For large turbines composite columns made out of industrially available rolled
section could become inadequate in terms of strength. This calls for usage of plate
girders in lieu of composite section. Due to inherent weakness of welds under dynamic
loading continuous butt welds are usually preferred instead of llet welds. Continuous
butt welds specially at the flange and web junction calls for rigorous quality assurance
in terms of Radiographic test or dye test to ensure 100 % weld penetration and could
make the fabrication expensive.
However if the steel industry in India agrees to manufacture rolled sections of higher
sizes (beyond ISMB 600) steel structure can become a very strong competitor to RCC
foundations.
In Europe since rolled steel sections having much higher moment of inertia are
available, use of steel structure as an alternate to RCC has become a viable solution
there.
2.18.2.6 Design of RCC sections
The structural members are usually designed by using IS-456(2000) or the local code
of the country in which it is being constructed or as specied in the contract like
ACI318, BS8110, DIN etc.
In most of the case the geometric sizing is decided by the equipment supplier, based
on which the stress induced in the members itself are normally low and to our knowl-
edge there has been no such cases where Turbine foundation members have misbehaved
or failed due to strength failure.
Most of the cases where members have misbehaved can be attributed to
improper detailing or faulty construction for which cracks have been observed to
develop.
Based on above, proper detailing of the members are of primary importance.
Some good detailing practices are mentioned hereafter which could be followed
while detailing such foundations.
The vertical reinforcing bars of the column shall have sufcient embedment in the
base slab to develop the required stresses.
Reinforcement in beams and columns shall be provided in all four sides irrespective
of they are required or not.
If design requirements do not guide the percentage of steel, the re-bars shall be
placed symmetrically on all four sides.
The minimum Steel provided in different parts of the members are mentioned
hereunder
Sl. No. Structural member type Steel quantity
1 Base slab 40 kg/m
3
2 Columns 70 kg/m
3
3 Top deck (beam and slab) 90 kg/m
3
Shear stirrups to be provided to account for the total shear in the foundation
element.
Splicing in columns if any shall always be done at the mid-height.
The diameter of bar in beams and columns should be so chosen that the maximum
spacing of the bars are not more than 150 mm.
Try to use lower diameter bar as far as practicable. For with lower diameter bars,
number of bars is more and distribution of stress and transfer of load between
concrete and steel is more uniform.
Unless specied by the contract the cover to reinforcement is usually taken as
follows:
Base Slab 100 mm on top, bottom and sides.
Columns and Pedestals 50 mm on sides
Beams 40 mm on all sides
Minimum development length for all bars irrespective of requirement shall not be
less than 50 times the diameter of the bar.
Beam column junction should be provided with additional steel to ensure that
cracks do not develop due to continuous reversal of stresses due to the application
of cyclic loads.
Example 2.18.1
Shown in Figure 2.18.5 is the layout plan of a Boiler feed pump framed
foundation with location of equipment loads as shown.
The dynamic loads under various operating conditions are as shown hereafter.
1315
Y 16.65
350
195kN
22.1kN 35 5 508 54
775
16.65 1200
1580
= =
4590 Pump Side 4590 C/L Coupling Motor Side
3.0m(TOC)
+3.5m(T.O.C.)
X + 4.0m(T.O.C.) 100kN
1765 4192 1585 2238
+3.5m(TOC)
Figure 2.18.5 Plan view of top deck with location of equipment load.
Table for dynamic loads on boiler feed pump top deck
Load condition Remarks P
x
P
y
P
z
M
x
M
y
M
z
Short circuit moment Total force at top deck 0.0 226 268 0 0 0
Operating load (1) End frame (pump side) 58 44 0 0 0
Middle frame 80 0 0 0
End frame (motor side) 100 8.3 0 0 0
Operating load (2) Load per long beam (pump side) 37 0.0 130 0 0 0
Operating load (2) Load per long beam (at coupling) 0 0 25 0 0 0
Operating Frequency of the machine = 5100 rpm
Center line axis of shaft = 1.1 m above the top deck
Bearing capacity of soil 150 kN/m
2
Shear wave velocity of the soil = 115 m/sec
Poissons Ratio of soil = 0.30
Live load of top deck slab during operation = 5 kN/m
2
Unit Weight of soil = 19 kN/m
3
All columns = 600 mm 600 mm
All beams = 600 mm 900 mm
Grade of concrete M25
Analyze the frame by
Rauschs method
By Barkans method
By Majors method
By 2D soil-structure interaction model.
Compare the results of the analysis based on the above methods with time
history
Do detailed design of the frame.
Solution:
We start the problem sequentially.
Here the top deck consists of a flat slab 900 mm thick supported on columns
(600 mm600 mm) and practically does not have a framing system.
Here for analysis and design we perceive a frame having edge beams in both
transverse and longitudinal direction having depth of 600 mm 900 mm as
shown in Figure 2.18.5 and 6.
The load from the slab is transferred to the idealized frame (as shown by the
dotted lines, Figure 2.18.7) and the frame is analyzed for vibration in vertical
and horizontal mode.
900
3600
(+/-)0.0
1500
3200
Figure 2.18.6 Elevation of the frame in transverse direction.
A
2
54
B
4590 4590
2600 22.5 195 16.7 16.7
690 2571 1329 1417 965 1308 900
1 2 3
508
1
Figure 2.18.7
Calculation of UDL load transferred to frame
Thickness of slab = 900 mm
Self weight of slab = 0.9 25 = 22.5 kN/m
2
Live load = 5.0 kN/m
2
: Thus, DL +LL = 27.5 kN/m
2
Equivalent UDL transferred to frame 1 and 3 =
wl
x
3
=
27.5 2.6
3
=
23.83 kN/m
Equivalent UDL on frame 2 = 23.83 2 = 47.66 kN/m
From Figure 2.18.7,
Equivalent UDL on frame A and B =
wl
x
6
_
3
_
l
x
l
y
_
2
_
=
27.5 2.6
6

_
3
_
2.6
4.59
_
2
_
= 31.92 kN/m
Load on longitudinal beam from the area of hydraulic coupling having local
projection of 1.4 m, of width 1585 mm
w =
0.5 2.6 25
2
= 16.21 kN/m
Thus the total superimposed UDL coming on the frame is as shown in
Figure 2.18.8
48 KN/m
24 KN/m
32 KN/m
48 KN/m
48 KN/m
24 KN/m
32 KN/m
Figure 2.18.8 Frame with uniformly distributed load from top deck slab.
Calculation of Concentrated load
For slab panel 1 along center line axis the distribution of load is as shown in
Figure 2.18.9.
R=217.5
R=217.5
22.5 195
Lx
690 2571 1329
2995
Ly
Figure 2.18.9 Load distribution slab panel-1.
Here we first out the point through which the resultant of this two concen-
trated force acts
y
c
=
22.5 690 +195 3261
22.5 +195
= 2995 mm from frame 1
Now the slab being restrained at all sides subjected to a load of 217.5 at
distance of 2995 mm from frame 1 it is evident that displacement at point O
shall be same for long and short span.
Thus considering the middle strip as a beamfixed at both ends in long direction
l
=
P
y
a
3
b
3
3EIL
3
y
Here P
y
= the net concentrated load acting in long direction; a = 2995 mm;
b = 1595 mm; L
y
= 4590 mm, and L
x
= 2600 mm
Displacement in short span is given by,
s
=
P
x
L
3
x
192EI
Here P
x
= The load transferred to short span.
Since here due to displacement compatibility,
l
=
s
, we have
P
x
L
3
x
192EI
=
P
y
a
3
b
3
3EIL
3
y
P
x
= 64P
y
a
3
b
3
_
L
y
L
x
_
3
Since by law of static

V = 0 we have, P
x
+P
y
= 217.5
P
y
= 218
_
_
64
_
ab
(L
y
/L
x
)
_
3
+1
_
Now substituting the values a, b, L
y
and L
x
, we have P
y
= 19.27 kN and
P
x
= 198.73 kN.
Since in short direction the load is symmetrical load on long beams along
Row A and B =
198.73
2
99.4 kN
In long direction, the position of the load is as shown in Figure 2.18.10.
19.3
2995 1595 R
2
R
1
Figure 2.18.10
Thus R
1
=
19.3 1595
4590
= 6.7 kN and R
2
= 12.6 kN.
Concentrated load transferred to edge beams from the slab panel 1 is shown
in Figure 2.18.11.
99.4
12.6
99.4
6.7
Figure 2.18.11 Concentrated load transferred to edge beams from the slab panel-1.
For slab panel 2 the loading arrangement is as shown Figure 2.18.12.
L
y
= 4500
508
54
1417
16.7 16.7
L
x
= 2600
965 1308 900
Figure 2.18.12 Loading on slab panel-2.
L
y
= 4500
L
x
= 2600
194
87.4
1584
3006
Figure 2.18.13 Resultant load on slab panel-2.
Total concentrated load acting on the slab = 16.7 2 +54 = 87.4 kN
The c.g. of the loads in y direction (Figure 2.18.13) is given by
y
c
=
16.7 1308 +16.7 2273
87.4
= 684 mm and
x
c
=
33.4 508
87.4
= 194 mm
Here
l
=
P
y
a
3
y
b
3
y
3EIL
3
y
and
s
=
P
x
a
3
x
b
3
x
3EIL
3
x
, when based on displacement compatibility
l
=
s
when we have
P
y
a
3
y
b
3
y
3EIL
3
y
=
P
x
a
3
x
b
3
x
3EIL
3
x
P
y
= P
x
_
a
x
b
x
a
y
b
y
_
3
_
L
y
L
x
_
3
Here a
x
= 1494 mm; b
x
= 1106 mm; a
y
= 3006 mm; b
y
= 1584 mm;
L
y
= 4590 mm and L
x
= 2600 mm.
Substituting the values in the above equation we have, P
y
= 0.2299P
x
For

V = 0 we have, P
x
+P
y
= 87.4 kN
Thus, we have P
x
= 71.06 kN and P
y
= 16.34 kN. Now, proceeding in the
similar manner as explained for slab panel 1 we find out the load transferred on
the frame beams shown in Figure 2.18.14.
Concentrated load transferred to edge beams from the slab panel 2 in
Figure 2.18.14.
30.22
5.64
3006
1494
10.7
140.83
Figure 2.18.14 Concentrated load transferred to edge beams from slab panel-2.
The net distribution of concentrated force on the frame is as shown in
Figure 2.18.15.
30.2
10.7
18.2
141
1584
99.4
6.7
99.4
Figure 2.18.15 Frame with concentrated load from top deck slab.
1 Rauschs method
Calculation of geometric properties of the frame
Area of transverse beam A
b
= (600 900) = 0.6 0.9 = 0.54 m
2
Area of column A
c
= (600 600) = 0.6 0.6 = 0.36 m
2
I
b
=
1
12
0.6 0.9
3
= 0.03645 m
4
;
I
c
=
1
12
0.6 0.6
3
= 0.0108 m
4
; and
=
I
b
h
I
c
L
=
0.03645 4.05
0.0108 2.6
= 5.257
Calculation of load transverse frame 1
Self weight of beam = 0.60.925 = 13.5 kN/m; UDL fromslab = 24 kN/m
Total, UDL(q) = 24 +13.5 = 37.5 kN/m
Self weight of long beam = 13.5 kN/m; UDL on long beam = 32 kN/m
Total UDL on long beam = 32 +13.5 = 45.5 kN/m
Load from long beam = 45.5
4.59
2
+0.5 99.4 = 154.15 kN
Load from column = 0.6 0.6 1.8 25 = 16.2 kN
Total load transferred to column (N) = 16.2 +154.15 = 170.4 kN
Load on beam from machine (P) = 6.7 kN; Modulus of concrete E
c
= 3
10
8
kN/m
2
E
c
I
b
= 3 10
8
0.03645 = 1.0935 10
7
kN m
2
;
E
c
A
b
= 3 10
8
0.54 = 1.62 10
8
kN
E
c
A
c
= 3 10
8
0.36 = 1.08 10
8
kN
Calculation of displacements transverse frame 1
1
=
PL
3
96EI
b
_
2 +1
+2
_
=
6.7 (2.6)
3
96 1.0935 10
7
_
2 5.257 +1
5.257 +2
_
= 1.769 10
7
m
2
=
QL
3
384EI
b
_
5 +2
+2
_
=
37.5 2.6 (2.6)
3
384 1.0935 10
7
_
5 5.257 +2
5.257 +2
_
= 1.5893 10
6
m
3
=
3
5
L
EA
b
_
P +
Q
2
_
=
0.6 2.6
1.62 10
8
_
6.7 +
98
2
_
= 5.364 10
7
m
4
=
h
EA
c
_
N +
P +Q
2
_
=
4.05
1.08 10
8
_
170.4 +
6.7 +98
2
_
= 8.353 10
6
m
4
i=1
i
= 1.06557 10
5
m
Thus, f
v
=
30
v
=
30
1.06557 10
5
= 9190 rpm
Total UDL(q) = 48 +13.5 = 61.5 kN/m
Self weight of long beam = 13.5 kN/m; UDL on long beam = 32 kN/m
Load from machine (P) = 18.2 kN
Total UDL on long beam = 32 +13.5 = 45.5 kN/m
Load from long beam = 45.5 4.59 = 208.85 kN; Load from column =
0.6 0.6 1.8 25 = 16.2 kN
Shown in Figure 2.18.16.
141 for Row B
30.2 for Row A
99.4
4590
1585
1584
2730
1 2
3
16 kN/m
Figure 2.18.16 Load distribution on longitudinal girder.
Load on beam from machine on row A.
Load on row A =
99.4
2
+30.2
1584
4590
+16 1.585
2730
4590
= 75.2 kN
Load on row B =
99.4
2
+141
1584
4590
+16 1.585
2730
4590
= 113.4 kN
Load on column along row A = 208.85 +75.2 +16.2 = 300 kN
Load on column along row B = 208.85 +113.4 +16.2 = 338.45 kN
Average load = 319.225 kN
Thus loading on frame 2 is as shown in Figure 2.18.17.
319 (kN) 18.2 (kN) 319 (kN)
62 kN/m (q)
Figure 2.18.17 Load on transverse frame-2.
Calculation of displacements for transverse frame 2
1
=
PL
3
96EI
b
_
2 +1
+2
_
=
18.2 (2.6)
3
96 1.0935 10
7
_
2 5.257 +1
5.257 +2
_
= 4.832 10
7
m
2
=
QL
3
384EI
b
_
5 +2
+2
_
=
62 2.6 (2.6)
3
384 1.0935 10
7
_
5 5.257 +2
5.257 +2
_
= 2.614 10
6
m
3
=
3
5
L
EA
b
_
P +
Q
2
_
=
0.6 2.6
1.62 10
8
_
18.2 +
161.2
2
_
= 9.514 10
7
m
4
=
h
EA
c
_
N +
P +Q
2
_
=
4.05
1.08 10
8
_
319 +
18.2 +161.2
2
_
= 1.53263 10
5
m
4
i=1
i
= 1.9375 10
5
m f
v
=
30
v
=
30
1.9375 10
5
= 6815 rpm.
Total UDL(q) = 24 +13.5 = 37.5 kN/m
Self weight of long beam = 13.5 kN/m; UDL on long beam = 32 kN/m;
Load from machine (P) = 10.7 kN
Total UDL on long beam= 32+13.5 = 45.5 kN/m; Load from long beam=
45.5
4.59
2
= 113 kN
Load from column = 0.6 0.6 1.8 25 = 16.2 kN
As shown in Figure 2.18.18, Load due to concentrated load on long beam on
Row A and B
Load on row A = 30.2
3006
4590
+16 1.585
1860
4590
= 30.6 kN
Load on row B = 141
3006
4590
+16 1.585
1860
4590
= 102.62 kN
Average load = 66.61 kN
Total load on column (N) = 66.61 +113

= 180 kN
30.1 for Row A
141 for Row B
16 kN/m
1584
2
3
2730
Figure 2.18.18 Load on longitudinal girder.
Calculation of displacements for transverse frame 3
1
=
PL
3
96EI
b
_
2 +1
+2
_
=
10.7 (2.6)
3
96 1.0935 10
7
_
2 5.257 +1
5.257 +2
_
= 2.825 10
7
m
2
=
QL
3
384EI
b
_
5 +2
+2
_
=
37.5 2.6 (2.6)
3
384 1.0935 10
7
_
5 5.257 +2
5.257 +2
_
= 1.5893 10
6
m
3
=
3
5
L
EA
b
_
P +
Q
2
_
=
0.6 2.6
1.62 10
8
_
10.7 +
98
2
_
= 5.75 10
7
m
4
=
h
EA
c
_
N +
P +Q
2
_
=
4.05
1.08 10
8
_
180 +
10.7 +98
2
_
= 8.788 10
6
m
4
i=1
i
= 1.1234 10
5
m; Thus, f
v
=
30
v
=
30
1.1234 10
5
= 8950 rpm.
Thus average vertical frequency of the frame =
9190 +6815 +8950
3
=
8318 > 5100 rpm
Frequency in horizontal direction
Weight of top deck = 9.78 3.2 0.9 25 = 704 kN
Weight from machine = 22.1 +195 +16.65 2 +54 +100 = 404
Total weight = 704 +404 = 1108 kN
K
hi
=
12EI
c
h
3
_
6 +1
3 +2
_
=
12 3 10
8
0.0108
(4.05)
3

_
6 5.257 +1
3 5.257 +2
_
= 1071751 kN/m
f
h
= 30
_
K
h1
+K
h2
+ +K
hN
W
= 30
_
3 1071751
1108
= 1616 < 5100 rpm.
The method does not have any provision of amplitude check and only check
for resonance with the operating frequency of the machine.
2 Barkans method
Calculation of stiffness for transverse frame 1 in vertical direction
k
1
=
2EA
c
h
is the stiffness of the columns =
2 3 10
8
0.6 0.6
4.05
=
5.333 10
7
kN/m
For transverse beam
v
=
L
3
(1 +2)
96EI
b
(2 +)
+
3L
8GA
b
=
(2.6)
3
96 1.0935 10
7

(1 +2 5.257)
7.257
+
3 2.6
8 1.5 10
8
0.6 0.9
= 3.86014 10
8
m
k
2
=
1
v
k
2
= 2.6 10
7
kN/m
The stiffness matrix thus becomes
[K] =
_
k
1
+k
2
k
2
k
2
k
2
_
=
_
7.93 2.6
2.6 2.6
_
10
7
kN/m
m
1
= m
L
+0.255m
b
+0.35m
c
m
b
= Mass of cross girder =
(24 +13.5) 2.6
9.81
= 9.938 kN sec
2
/m
m
L
= Mass transferred from long girder
=
104.45 2 +99.4
9.81
= 31.43 kN sec
2
/m
m
c
= Mass of column =
0.6 0.6 3.6 25 2
9.81
= 6.605 KN/sec
2
/m
m
1
= 0.255 9.938 +0.35 6.605 +31.43 = 36.27 kN sec
2
/m
m
2
= m
0
+0.45m
b
m
0
= Load from machine on transverse girder;
m
2
=
6.7
9.81
+0.45 9.938 = 5.15kN sec
2
/m
Thus, [M] =
_
36.27 0
0 5.15
_
Thus based on eigen value solution
=
_
7.93 10
7
36.27 2.6 10
7
2.6 10
7
2.6 10
7
5.15
_
= 0
The above on solution gives
74
1
= 1.2369 10
6

1
= 1112 rad/sec(10168 rpm) > 5100 rpm
2
= 5.998 10
6

2
= 2449 rad/sec(23386 rpm) > 5100 rpm
The normalized eigen vector is given by
[]
n
=
_
0.14837 0.07413
0.19654 0.39428
_
Based on loading table for dynamic load, P
v
= 44 +130 = 174 KN.
Thus equation of motion becomes
_
36.27 0
0 5.15
_ _
z
1
z
2
_
+
_
7.93 2.6
2.6 2.6
_
10
7
_
z
1
z
2
_
=
_
0.0
174
_
sin 534t
74 We will not solve this equation directly. Please refer to Chapter 5 (Vol. 1) on structural dynamics
where we have solved in detail such eigen value problem.
Calculation of amplitude for transverse frame 1 in vertical direction
Thus []
T
n
{P} =
_
0.14837 0.19654
0.07413 0.39428
_
_
0
174
_
=
_
34.19
68.60
_
Considering 5%
damping for concrete on orthogonal transformation of the above equation of
motion we have
1
+111
1
+1.2369 10
6
1
= 34.19sin 534t;
2
+245
2
+5.998 10
6
2
= 68.60sin 534t
Here, r =

m
n
= 0.48 for mode 1 and r =
534
2449
= 0.21,

1
=
34.19sin 534t
1.236910
6
_
(1(0.48)
2
)
2
(20.050.48)
2
= 3.58410
5
sin 534t m
and
2
=
68.60sin 534t
5.99810
6
_
(1(0.21)
2
)
2
(20.050.21)
2
= 11.9210
6
sin 534t m
We have then, {Z} = []{} =
_
z
1
z
2
_
=
_
0.14837 0.07413
0.19654 0.39428
_ _
35.48
11.92
_
10
6
sin 534t =
_
4.38
11.657
_
sin 534t 10
6
m
k
1
=
2EA
c
h
is the stiffness of the columns =
2 3 10
8
0.6 0.6
4.05
=
5.333 10
7
kN/m
For transverse beam
v
=
L
3
(1 +2)
96EI
b
(2 +)
+
3L
8GA
b
=
(2.6)
3
96 1.0935 10
7

(1 +2 5.257)
7.257
+
3 2.6
8 1.5 10
8
0.6 0.9
= 3.86014 10
8
m
k
2
=
1
v
= 2.6 10
7
kN/m
The stiffness matrix thus becomes
[K] =
_
k
1
+k
2
k
2
k
2
k
2
_
=
_
7.93 2.6
2.6 2.6
_
10
7
kN/m
m
1
= m
L
+0.255m
b
+0.35m
c
m
b
(48 +13.5) 2.6
9.81
= 16.29 kN sec
2
/m
m
L
=
(32 +13.5) 4.95
9.81
+
75.2
9.81
+
113.4
9.81
= 42.20 kN sec
2
/m.
m
c
= Mass of column =
0.6 0.6 3.6 25 2
9.81
= 6.605 kNsec
2
/m
m
1
= 0.255 16.29 +0.35 6.605 +42.20 = 48.65 kN sec
2
/m
m
2
= m
0
+0.45m
b
m
0
= Load from machine on transverse girder
m
2
=
18.2
9.81
+0.45 16.29 = 9.185 kN sec
2
/m.
Thus [M] =
_
48.65 0
0 9.185
_
Thus the eigen value solution =
_
7.93 10
7
48.65 2.6 10
7
2.6 10
7
2.6 10
7
9.185
_
= 0
75
1
= 861700
1
= 928 rad/sec(8861 rpm) > 5100 rpm
2
= 3.599 10
6

2
= 1897 rad/sec(18115 rpm) > 5100 rpm
The normalized eigen vector is given by, []
n
=
_
0.1215 0.0754
0.1748 0.27985
_
Based on loading table for dynamic load
P
v
= 25 +130 = 155 kN.
75 We will not solve this equation directly. Please refer to Chapter 5 (Vol. 1) where we have solved in
detail such eigen value problem.
_
48.65 0
0 9.185
_ _
z
1
z
2
_
+
_
7.93 2.6
2.6 2.6
_
10
7
_
z
1
z
2
_
=
_
0.0
155
_
sin 534t
[]
T
n
{P} =
_
0.1215 0.1748
0.0759 0.27985
_ _
0
155
_
=
_
27.094
43.3768
_
Considering 5% damping for concrete on orthogonal transformation of the
above equation of motion we have
1
+92.8
1
+861700
1
= 27.094sin 534t;
2
+189.7
2
+3.599 10
6
2
= 43.3768sin 534t
Here r =

m
n
= 0.575 for mode 1 and r =
5100
18115
= 0.28

1
=
27.094sin 534t
861700
_
(1 (0.575)
2
)
2
(2 0.05 0.575)
2
= 4.68 10
5
sin 534t m
and
2
=
43.3768sin 534t
3.59910
6
_
(1 (0.28)
2
)
2
(2 0.05 0.28)
2
= 13.07 10
6
sin 534t m
Since {Z} = [] {},
_
z
1
z
2
_
=
_
0.1215 0.0754
0.1748 0.27985
_ _
46.8
13.07
_
sin 534t 10
6
=
_
4.70
11.838
_
sin 534t 10
6
m.
Referring to the previous calculation the stiffness matrix is
[K] =
_
k
1
+k
2
k
2
k
2
k
2
_
=
_
7.93 2.6
2.6 2.6
_
10
7
kN/m
m
1
= m
L
+0.255m
b
+0.35m
c
m
b
(24 +13.5) 2.6
9.81
= 9.93 kN sec
2
/m
m
L
=
(32 +13.5) 4.95
9.81 2
+
30.6
9.81
+
102.62
9.81
= 25.06 kN sec
2
/m.
m
c
= Mass of column =
0.6 0.6 3.6 25 2
9.81
= 6.605 kN sec
2
/m
m
1
= 0.255 9.93 +0.35 6.605 +25.06 = 29.90 kN sec
2
/m
m
2
= m
0
+0.45m
b
m
0
= Load from machine on transverse girder
m
2
=
10.7
9.81
+0.45 9.93 = 5.56 kN sec
2
/m
Thus [M] =
_
30 0
0 5.56
_
Thus based on eigen value solution
=
_
7.93 10
7
30 2.6 10
7
2.6 10
7
2.6 10
7
5.56
_
= 0
76
1
= 1404600
1
= 1185 rad/sec(11316 rpm) > 5100 rpm
2
= 5915000
2
= 2432 rad/sec(23224 rpm) > 5100 rpm
The normalized eigen vector is given by
[]
n
=
_
0.1555 0.0956
0.2194 0.3612
_
Based on loading table for dynamic load
P
v
= 8.3 +25 = 33.3 kN.
76 We will not solve this equation directly. Please refer to Chapter 5 (Vol. 1) where we have solved in
detail such eigen value problem.
_
30 0
0 5.56
__
z
1
z
2
_
+
_
7.93 2.6
2.6 2.6
_
10
7
_
z
1
z
2
_
=
_
0.0
33.3
_
sin 534t
Thus []
T
n
{P} =
_
0.1555 0.2194
0.0956 0.3612
__
0
33.3
_
=
_
7.30
12.03
_
1
+118.5
1
+1404600
1
= 7.30sin 534t;
2
+243.2
2
+5915000
2
= 12.03sin 534t.
Here r =

m
n
= 0.45 for mode 1 and r = 0.22
1
=
7.30sin 534t
1404600
_
(1 (0.45)
2
)
2
(2 0.05 0.45)
2
= 6.5 10
6
sin 534t and
2
=
12.03sin 534t
5.91510
6
_
(1 (0.22)
2
)
2
(2 0.05 0.22)
2
= 2.136 10
6
sin 534t
Since {Z} = []{}
Hence
_
z
1
z
2
_
=
_
0.1555 0.0956
0.2194 0.3612
_ _
6.5
2.136
_
sin 534t 10
6
=
_
0.822
2.219
_
sin 534t 10
6
m
Calculation of horizontal frequency
Table for calculation of mass and stiffness
Frame m
0
m
b
m
c
m
l
m
i

hi
K
hi
1 0.683 9.938 6.605 31.43 44.030 9.333E-07 1071467
2 1.855 16.29 6.605 42.2 62.326 9.333E-07 1071467
3 1.0907 9.93 6.605 25.06 38.060 9.333E-07 1071467
144.416 3214401
Here mi = m
0i
+m
bi
+0.3m
ci
+m
Li
and
hi
=
h
3
(2 +3)
12EI
c
(1 +6)
Table for calculation of second, mass moment and stiffness inertia
Frame m
i
K
hi
d
i
mi
di
K
hi
d
I
x
gi
(X
g
d
i
) X
hi
m
i
x
2
gi
K
hi
X
2
hi
1 44.030 1071467 0 0 0 4.40 4.59 852.5194 22573770
2 62.326 1071467 4.59 286.08 4918032.8 0.190 0 2.243947 0
3 38.060 1071467 9.18 349.39 9836065.6 4.78 4.59 869.5175 22573770
635.47 14754098 1724.281 45147541
X
g
=
635.476
144.416
= 4.40 m; X
h
=
14754098
3214401
= 4.59 m
Thus e = 4.59 4.40 = 0.190 m
Dynamic loads
P
hi
= 58 +80 +100 = 238 kN and
M
h
=
N
i=1
P
hi
X
gi
= 238 4.40 = 1047.2
_
M 0
0 J
_ _
x

_
+
_
K
h
K
h
e
K
h
e K
h
e
2
+
_ _
x
_
=
_
P
h
cos
m
t
M
h
cos
m
t
_
_
144.416 0.0
0.0 1724.3
_ _
x

_
+
_
3.215 10
6
0.6095 10
6
0.6095 10
6
4.53 10
6
_ _
x
_
=
_
238cos 534t
1047.2cos 534t
_
Thus for natural frequency we have
_
3.215 10
6
144.416 0.6095 10
6
0.6095 10
6
4.53 10
6
1724.3
_
= 0
The above on expansion and solution gives
1
= 21919
1
= 148 rad/sec(1413 rpm) < 5100 rpm
2
= 26614
2
= 163 rad/sec(1557 rpm) < 5100 rpm
The normalized eigen vector is given by,
[]
n
=
_
0.08011 0.022485
6.5109 10
3
0.023186
_
[]
T
n
{P} =
_
0.08011 6.5109 10
3
0.022485 0.023186
_ _
238cos 534t
1047.2cos 534t
_
=
_
12.25cos 534t
248.15cos 534t
_
1
+14.8
1
+21919
1
= 12.25cos 534t;
2
+16.3
2
+26614
2
= 248.15cos 534t.
Here r =

m
n
= 3.609 for mode 1 and r = 3.27
i.e.
1
=
12.25cos 534t
21919
_
(1(3.6)
2
)
2
(2 0.05 0.3.6)
2
= 4.6707 10
5
cos 534t
and
2
=
248.1552cos 534t
26614
_
(1(3.27)
2
)
2
(20.053.27)
2
= 9.61416 10
4
cos 534t
Since {X} = []{}, hence
_
x
_
=
_
0.08011 0.022485
6.5109 10
3
0.023186
_ _
4.6707
96.1416
_
cos 534t 10
5
=
_
2.5359
2.1987
_
cos 534t 10
5
m
Thus, displacement of frame 3
This is generically given by x
net
= x +X
where
X = is the farthest point form the center of gravity point G
x
3
= 2.5359 10
5
+4.78 2.1987 10
5
= 1.30456 10
4
Displacement of frame 2
x
2
= 2.5359 10
5
+0.190 2.1987 10
5
= 2.953 10
5
Displacement of frame 3
x
1
= 2.5359 10
5
4.40 2.1987 10
5
= 7.13828 10
5
3 Majors combined method of analysis
In this case the natural frequency in vertical direction is same as shown in
Rauschs method earlier except for the case of soil when Major considers the
total frame as stick model and combines with soil displacement.
Thus for transverse frame 1, we have
4
i=1
i
= 1.06557 10
5
m
For calculation of

4
i=1
i
refer to previous calculation by Rauschs method
and f
v
=
30
v
=
30
1.06557 10
5
= 9190 r.p.m
While Major calculated the dynamic force based on Rotor weights, here
dynamic force coming on the frame has directly been given, thus
Load on transverse beam (C
b
) = 44 kN
Load on column from Long beams(C
c
) = 130 kN
Thus
1
=
C
b
L
3
96EI
b
_
2 +1
+2
_
+
3
5
L
EA
b
C
b
+
h
EA
c
_
C
b
+C
c
2
_
=
44 (2.6)
3
96 1.0935 10
7

11.514
7.257
+
0.6 2.6 44
1.62 10
8
+
4.05
1.08 10
8
_
44 +130
2
_
= 4.855 10
6
m
As
n
>
m
hence corrected value of operating frequency is
n
=
0.8 9190 = 7352 r.p.m
M.F =
1
_
_
(1 r
2
)
2
+(
)
2
(r)
2
_
, here r =
5100
7352
= 0.69 and = 0.4
Substituting the above values, we have
M.F. = 1.900
Thus
v1
= 4.855 10
6
1.9 = 9.2245 10
6
m
For transverse frame 2, we have
4
i=1
i
= 1.9375 10
5
m
For calculation of

4
i=1
i
and f
v
=
30
v
=
30
1.9375 10
5
= 6815 rpm
b
) = 0.0 kN
c
) = 130 +25 = 155 kN
Thus
1
=
C
b
L
3
96EI
b
_
2 +1
+2
_
+
3
5
L
EA
b
C
b
+
h
EA
c
_
C
b
+C
c
2
_
= 0.0 +0.0 +
4.05
1.08 10
8
_
155
2
_
= 2.9063 10
6
m
As
n
>
m
n
= 0.86815 =
4948 rpm which is less than the operating speed of 5100 r.p.m.
At transient resonant condition as per Major M.F = 7.85, here r =
5100
7352
= 0.69
Substituting the above values we have
Thus,
v2
= 2.9063 10
6
7.85 = 2.281 10
5
m
For transverse frame 3 we have
4
i=1
i
= 1.1234 10
5
m
For calculation of

4
i=1
i
and f
v
=
30
v
=
30
1.1234 10
5
= 8950 rpm
b
) = 8.30 kN
c
) = 25 kN
Thus
1
=
C
b
L
3
96EI
b
_
2 +1
+2
_
+
3
5
L
EA
b
C
b
+
h
EA
c
_
C
b
+C
c
2
_
=
8.3 (2.6)
3
96 1.0935 10
7

11.514
7.257
+
0.6 2.6 8.3
1.62 10
8
+
4.05
1.08 10
8
_
8.3 +25
2
_
= 9.2478 10
7
m
As
n
>
m
n
= 0.88950 =
7160 rpm which is greater than the operating speed of 5100 rpm.
Here
M.F =
1
_
[(1 r
2
)
2
+(
)
2
(r)
2
]
, r =
5100
7352
= 0.69 and = 0.4
which gives M.F. = 1.995
Substituting the above values we have
v3
= 9.2478 10
7
1.995 = 1.8446 10
6
m
Calculation of vertical frequency including the soil effect
In this case we consider the total top deck including the column as a stick
model
Thus, for frame 1,

4
i=1
i
= 1.06557 10
5
; For frame 2,
4
i=1
i
= 1.9375 10
5
m;
For frame 3,

4
i=1
i
= 1.1234 10
5
m; Thus,
av
= 1.37549 10
5
m.
Shear wave velocity =115 m/sec; unit weight of soil =19 kN/m
3
; mass density
of soil = 1.936 kN sec
2
/m
4
.
Dynamic shear modulus, G = 1.936 115 115 = 25614 kN/m
2
.
Base Area = 9.78 m 3.2 m; equivalent radius r
0
=
_
9.78 3.2
= 3.156 m
Thus vertical spring stiffness of soil =
4Gr
0
(1 )
=
4 25614 3.156
0.7
=
461930 kN/m
Weight of top deck = 9.78 3.2 0.9 25 = 704.16 kN
Weight of column = 6 0.36 3.625 = 194.4 kN
Weight of machine = 22.1 +195 +2 16.65 +54 +100 = 404 kN
Weight of base mat = 9.78 3.2 1.5 25 = 1173.6 kN
Total vertical Load = 704.16 +194.4 +404 +1173.6 = 2476.16 kN
s
=
P
v
K
v
=
2476.16
461930
= 5.36046 10
3
m
Total displacement
v
=
1
+
2
+
3
+
4
+
s
= 1.3754910
5
+5.36046
10
3
= 5.37422 10
3
m
Knowing f
v
=
30
v
cpm; we have f
s
=
30
5.37422 10
3
= 409 rpm
Calculation of horizontal frequency
For horizontal frequency we know that
(f
n
)
h
= 30
2
0
s
N
i=1
K
hi
N
i=1
W
i
I
h
J
c.p.m
Here, K
hi
= Lateral stiffness of the ith frame i and K
hi
=
1
hi
where
hi
=
h
3
(2 +3)
12EI
c
(1 +6)
W
i
= total weight of the ith frame plus weight of the machine plus weight of
the transverse beam and the longitudinal beams, J
=Mass moment of inertia
N
i=1
W
i
X
2
gi
; Xg =distance of weight W from the resultant center of mass point
G; I
h
=

N
i=1
K
hi
X
2
hi
; X
h
= distance of each frame from the centre of rigidity
H,
and
0
=
1
2
_
e
2
N
i=1
K
hi
J
N
i=1
K
hi
N
i=1
W
i
+
I
h
J
_
Here,
0
=
1
2
_
0.190
2
3 1.0717 10
6
1724.3 9.81
+
3 1.0717 10
6
144.416 9.81
+
2 2.257 10
7
1724.3 9.81
_
= 2472.42
Here all the data within the parenthesis were calculated while doing the
calculation based on Barkans method.
Thus (f
n
)
1
= 30
0

_
2
0

N
i=1
K
hi
N
i=1
W
i
I
h
J

or, (f
n
)
1
= 30
_
2472.4
_
6112853
3 1.0717 10
6
144.416 9.81

2 2.257 10
7
1724.3 9.81
= 30
_
2472.4
56797
(f
n
)
1
= 30
2472.4 238 = 1418 rpm(148 rad/sec);

(f
n
)
2
= 30
2472.4 +238 = 1562 rpm(164 rpm/sec).

Calculation of Horizontal amplitude as per Figure 2.18.19
4590
4590
58 kN 80 kN 100 kN
Figure 2.18.19 Top deck slab with transverse load.
The resultant of the Horizontal dynamic load acts at
x =
80 4.59 +58 4.59 2
100 +58 +80
= 3.78 m
Thus eccentricity between center of rigidity and x is 0.81 m is
C
i
= C
K
hi
N
i=1
K
hi
+e
CK
hi
X
hi
N
i=1
K
hi
X
2
hi
hence for frame 1 we have
C
1
= 238
1.071610
6
31.071610
6
+0.81
2381.071610
6
4.59
22.25710
7
= 100.33 kN
Knowing
hi
=
C
i
K
hi
we have,
h1
=
100.33
1.0717 10
6
= 9.36264 10
5
As the foundation is under tuned thus maximum amplitude at transient
condition is given by
h1
= 9.36264 10
5
7.85 = 7.34967 10
4
m.
For frame 2, we have
C
2
= 238
1.0716 10
6
3 1.0716 10
6
+0.81
238 1.0716 10
6
0.0
2 2.257 10
7
= 79.33 kN
Knowing
hi
=
C
i
K
hi
we have,
h2
=
79.33
1.0717 10
6
= 7.4277 10
5
h2
= 7.4277 10
5
7.85 = 5.8307 10
4
.
For frame 3 we have
C
1
= 238
1.0716 10
6
3 1.0716 10
6
+0.81
238 1.0716 10
6
4.59
2 2.257 10
7
= 58.33 kN
Knowing
hi
=
C
i
K
hi
we have,
h3
=
58.33
1.0717 10
6
= 5.44306 10
5
h3
= 5.44306 10
5
7.85 = 4.273 10
4
m
4 Based on 2D soil structure interaction model
Calculation in vertical direction
The mathematical model for this case is as shown in Figure 2.18.20. Mathe-
matical model of the turbine foundation with soil spring in vertical direction.
The equation of motion is given by
_
_
_
m
1
0 0
0 m
2
0
0 0 m
3
_
_
_
_
z
1
z
2
z
3
_
_
+
_
_
_
c
1
+c
2
c
2
0
c
2
c
2
+c
3
c
3
0 c
3
c
3
_
_
_
_
z
1
z
2
z
3
_
_
+
_
_
k
1
+k
2
k
2
0
k
2
k
2
+k
3
k
3
0 k
3
k
3
_
_
_
_
_
z
1
z
2
z
3
_
_
_
=
_
_
_
0
0
P
v
_
_
_
sin
m
t
k
3
m
3
m
2
k
2
k
1
m
1
Figure 2.18.20 Mathematical model of the turbine foundation with soil spring in vertical
direction.
Here, mass of column
m
c
=
0.36 3.6 6 25
9.81
= 19.816
Here m
1
= Mass of the bottom raft
m
1
=
9.78 3.2 1.5 25
9.81
+5.944 = 125.57 126 kN sec
2
/m
m
2
= 0.25 times the weight of the transverse girder + weight of machine from
longitudinal girder +self weight fromlongitudinal girder +0.3 times the weight
of column.
m
2
=
_
24 +13.5
9.81
2.6 2 +
48 +13.5
9.81
2.6
_
0.25
+
4.59 2 (32 +13.5)
9.81
+
16 1.585 2
9.81
+
99.4 2 +30.2 +141
9.81
+
0.3 6 0.36 3.6 25
9.81
= 100.451 100 kN sec
2
/m
m
3
= Concentrated mass on transverse girder + 0.45 times the self weight
For frame 1 we have
m
31
=
6.7
9.81
+
(24 +13.5)
9.81
2.6 0.45 = 5.15 kN sec
2
/m
For frame 2 we have
m
32
=
18.2
9.81
+
(48 +13.5)
9.81
2.6 0.45 = 9.19 kN sec
2
/m
For frame 3 we have
m
33
=
10.7
9.81
+
(24 +13.5)
9.81
2.6 0.45 = 5.56 kN sec
2
/m
Thus for the complete frame we have, m
3
= 5.15+9.19+5.56 = 19.9 20
The mass matrix is thus given by
[M] =
_
_
126 0 0
0 100 0
0 0 20
_
_
Calculation of stiffness matrix
In vertical direction the displacement of the transverse girder is given by
v
=
L
3
(1 +2)
96EI
b
(2 +)
+
3L
8GA
b
v
=
2.6
3
(1 +2 5.257)
96 1.0935 10
7
(2 +5.257)
+
3 2.6
8 1.5 10
8
0.54
= 3.86014 10
8
Thus for three frames we have, k
3
=
3
3.86014 10
8
= 7.772 10
7
For columns we have, k
2
=
2EA
c
h
=
2 3 10
8
0.36
3.6
3 = 18.0 10
7
For the soil the equivalent spring stiffness is given by
k
1
=
4Gr
0
1
= 461930
[K] =
_
_
k
1
+k
2
k
2
0
k
2
k
2
+k
3
k
3
0 k
3
k
3
_
_
=
_
_
18.0461930 18.0 0
18 25.772 7.772
0 7.772 7.772
_
_
10
7
Thus for eigen value solution we have
77
_
_
180461930 126 180000000 0
180000000 257720000 100 77720000
0 77720000 77720000 20
_
_
= 0
This gives the eigen values and the corresponding three natural frequencies as
[] =
_
_
0.002 10
6
0 0
0 2.5641 10
6
0
0 0 5.3294 10
6
_
_
The corresponding eigen vector are given by
[] =
_
_
0.5768 0.3675 0.1266
0.5775 0.2983 0.3454
0.5778 0.8771 0.9299
_
_
;
[] =
_
_
44.721 0 0
0 1601 0
0 0 2308
_
_
rad/sec
77 We have solved the eigen problem in Math-Cad directly.
Based on orthogonal transformation []
T
[M][]
[]
T
[M][] =
_
_
81.9436
42.1484
31.2429
_
_
Thus the scale factors are given by
M
r1
=
81.9436 = 9.052; M
r2
=
42.1484 = 6.492;
M
r3
=
31.2429 = 5.5895.
Thus the normalised eigen vector is given by
[]
n
=
_
_
0.06372 0.05661 0.02265
0.06379 0.04595 0.06179
0.06383 0.13510 0.16636
_
_
Calculation of damping matrix
Critical damping, C
c
= 2
km
Let the damping ratio for RCC structure be 5%.
Thus critical damping C
c
is given by
c
3
= 0.05 (2
_
7.772 10
7
20) = 3942.6;
c
2
= 0.05 (2
_
18.00 10
7
100) = 13416
The soil spring is calculated based on Richarts formula as shown hereafter
78
B
z
=
(1 )W
f
4
s
r
3
0
=
0.7 1173.6
4 19 (3.156)
3
= 0.3438; D
z
=
0.425
_
B
z
= 0.7247
c
1
= 0.7247 (2
461930 126) = 11057.6

Thus the damping matrix is given by
[C] =
_
_
c
1
+c
2
c
2
0
c
2
c
2
+c
3
c
3
0 c
3
c
3
_
_
=
_
_
24474 13416 0
13416 17359 3943
0 3943 3943
_
_
78 For further details refer to the section of Design and Analysis of Block Foundation.
Thus on orthogonal transformation for each individual mode we have
[]
T
[C][] =
_
_
44.8982
207.9
306.57
_
_
2D
1
1
= 44.8902 D
1
= 0.502
Calculation for load vectors
The total vertical dynamic force is given by, P
v
= 44+8.3+1302+252 =
362.3 kN
Performing the operation
[]
T
n
{P} =
_
_
0.06372 0.06379 0.06383
0.05661 0.04595 0.13510
0.02265 0.06179 0.16636
_
_
_
_
_
0
0
362.3
_
_
_
sin 534t
=
_
_
_
23.1256
48.9467
60.2772
_
_
_
sin 534t
Thus the three uncoupled equation of motion is given by
1
+44.89
1
+2000
1
= 23.1256sin 534t
2
+207.89
2
+2564100
2
= 48.9467 sin 534t
3
+207.89
3
+2564100
3
= 60.2772sin 534t
1
= 0.502,
2
= 0.0649 and
3
= 0.0664
Based on the above the displacement vector is given by
i =
3
i=1
p
i
sin
m
t
_
(1 r
2
i
)
2
+(2D
i
r
i
)
2
Once we know the displacement vectors in un-coupled state the displacement
in the global structural co-ordinate is given by
{Z} = []{}
The results are shown hereafter in tabular form
C
a
s
e
1
O
p
e
r
a
t
i
n
g
f
r
e
q
u
e
n
c
y
o
f
t
h
e
m
a
c
h
i
n
e
=
5
3
4
r
a
d
/
s
e
c
E
i
g
e
n
E
i
g
e
n
E
i
g
e
n
N
a
t
u
r
a
l
C
o
r
r
e
c
t
e
d
v
e
c
t
o
r
1
s
t
v
e
c
t
o
r
2
n
d
v
e
c
t
o
r
3
r
d
D
a
m
p
i
n
g
D
i
s
p
A
m
p
l
i
t
u
d
e
f
r
e
q
u
e
n
c
y
f
r
e
q
u
e
n
c
y
m
o
d
e
m
o
d
e
m
o
d
e
P
r
a
t
i
o
r
M
.
F
.
(
u
n
c
o
u
p
l
e
d
)
Z
z
1
4
4
.
7
2
5
3
.
6
6
4
0
.
0
6
3
7
2
0
.
0
5
6
6
1
0
.
2
2
6
5
2
3
.
1
3
0
.
5
0
2
9
.
9
5
0
0
.
0
1
0
1
4
1
.
1
7
1
0
0
4
8
.
9
6
1
0
0
6
z
2
1
6
0
1
1
2
8
0
.
8
0
.
0
6
3
7
9
0
.
0
4
5
9
5
0
.
0
6
1
7
9
4
8
.
9
5
0
.
0
6
5
0
.
4
1
6
1
.
2
0
7
8
0
2
.
3
0
4
1
0
0
5
9
.
3
0
1
0
0
6
z
3
2
3
0
8
1
8
4
6
.
4
0
.
0
6
3
8
3
0
.
1
3
5
1
0
.
1
6
6
3
6
6
0
.
2
8
0
.
0
6
6
0
.
2
8
9
1
.
0
9
0
3
3
1
.
2
3
5
1
0
0
5
1
.
2
6
1
0
0
5
C
a
s
e
2
T
r
a
n
s
i
e
n
t
f
r
e
q
u
e
n
c
y
d
u
r
i
n
g
s
t
a
r
t
a
n
d
s
t
o
p
o
f
m
a
c
h
i
n
e
=
4
4
.
7
2
r
a
d
/
s
e
c
E
i
g
e
n
E
i
g
e
n
E
i
g
e
n
N
a
t
u
r
a
l
v
e
c
t
o
r
1
s
t
v
e
c
t
o
r
2
n
d
v
e
c
t
o
r
3
r
d
D
a
m
p
i
n
g
Z
f
r
e
q
u
e
n
c
y
m
o
d
e
m
o
d
e
m
o
d
e
P
r
a
t
i
o
r
M
.
F
.
D
i
s
p
(
u
n
c
o
u
p
l
e
d
)
Z
Z
1
4
4
.
7
2
0
.
0
6
3
7
2
0
.
0
5
6
6
1
0
.
2
2
6
5
2
3
.
1
2
5
6
0
.
5
0
2
1
0
.
9
9
6
0
1
6
1
.
1
5
1
7
1
0
0
2
7
.
3
5
3
7
1
0
0
4
Z
2
1
6
0
1
0
.
0
6
3
7
9
0
.
0
4
5
9
5
0
.
0
6
1
7
9
4
8
.
9
4
6
7
0
.
0
6
5
0
.
0
2
7
9
3
3
1
.
0
0
0
7
7
4
1
.
9
1
1
1
1
0
0
5
7
.
3
6
2
7
1
0
0
4
Z
3
2
3
0
8
0
.
0
6
3
8
3
0
.
1
3
5
1
0
.
1
6
6
3
6
6
0
.
2
7
7
2
0
.
0
6
6
0
.
0
1
9
3
7
6
1
.
0
0
0
3
7
2
1
.
1
3
2
0
1
0
0
5
7
.
3
7
7
4
1
0
0
4
The results are compared hereafter by bar chart shown in Figure 2.18.21.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Operating Ist Transient
Operating Case
A
m
p
l
i
t
u
d
e

(
m
m
)
Vertical amplitude
columns
Vertical amplitude columns
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Operating Case
A
m
p
l
i
t
u
d
e

(
m
m
)
Vertical amplitude Transverse girder
Vertical amplitude
Transverse girder
Vertical amplitude Bottom Raft
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Operating Case
A
m
p
l
i
t
u
d
e

(
m
m
)
Vertical amplitude
Bottom Raft
Figure 2.18.21 Comparison of transient and operating response.
Analysis in coupled horizontal and rocking mode
The mathematical model for the turbine foundation for this mode is perceived
as Figure 2.18.22.
K
h
h
m
0
, J
Kx
u
y
Figure 2.18.22 2Dmathematical model for coupled translation and rocking including soil springs.
The un-damped equation of motion for free vibration in coupled horizontal
and rocking motion is given by
_
_
_
_
m
0
m
0
e m
0
m
0
h
m
0
e J
+m
0
e
2
m
0
e m
0
eh
m
0
m
0
e m
0
+m
f
m
0
h
m
0
h m
0
eh m
0
h J
+m
0
h
2
_
_
_
_
y

_
+
_
_
_
_
K
h
0 0 0
0 K
h
e
2
+I
0 0
0 0 K
y
0
0 0 0 K
_
_
_
y
_
= 0
Calculation of mass matrix
Here
m
0
= mass of the top deck + weight of the machine + 0.3 times the weight
of the column
e = distance between the centre of gravity and centre of rigidity
h = height between centre of the top deck mass to the centre of the bottom
raft mass
m
f
= mass of the bottom raft.
J
= mass moment of inertia of top deck in plan for the lumped masses @
J
=

N
i=1
m
i
X
2
gi
J
= mass moment of inertia of bottom raft in transverse plane

Referring to calculation done in Barkans method earlier we have
m
0
= 144.416 kN sec
2
/m; J
=
N
i=1
m
i
X
2
gi
= 1724.3;
e = 4.59 4.40 = 0.19 m
z =
44.03 0.45 +62.326 0.45 +38.06 0.45
144.416
= 0.45 (refer to calcu-
lations based on Barkans method for individual mass data)
Thus, h = 3.6 +(0.9 0.45) +1.5/2 = 4.8 m
m
f
=
9.78 3.2 1.5 25
9.81
= 120 kN sec
2
/m;
J
=
m
12
(l
2
y
+l
2
z
) =
120
12
(3.2
2
+1.5
2
) = 125
Thus
[M] =
_
_
_
_
m
0
m
0
e m
0
m
0
h
m
0
e J
+m
0
e
2
m
0
e m
0
eh
m
0
m
0
e m
0
+m
f
m
0
h
m
0
h m
0
eh m
0
h J
+m
0
h
2
_
_
=
_
_
_
_
144.416 27.44 144.416 693
27.44 1730 27.44 132
144.416 27.44 264 693
693 132 693 3452
_
_
kN sec
2
/m
Calculation of stiffness matrix
The stiffness matrix of the frame including the soil spring is given by
[K] =
_
_
_
_
K
h
0 0 0
0 K
h
e
2
+I
0 0
0 0 K
y
0
0 0 0 K
_
Here shear wave velocity of the soil = 115 m/sec
Unit weight of soil
s
= 19 kN/m
3
; Poissons ratio of soil
s
= 0.3
Thus dynamic shear modulus of soil G =
19
9.81
(115)
2
= 25614 kN/m
2
r
y
=
_
9.78 3.2
= 3.156 m; r
=
4
_
9.78 (3.2)
3
3
= 2.414 m
Thus, K
y
=
32(1
s
) Gr
y
7 8
s
=
32 0.7 25614 3.186
7 2.4
= 3.93645
10
5
kN/m
K
=
8Gr
3
3(1
s
)
=
8 25614 (2.414)
3
2.1
= 13.73 10
5
kN/m
Referring to the table for calculation of stiffness in Barkans method, done
earlier
3
i=1
K
hi
= 3 1.072 10
6
= 3.216 10
6
kN/m;
I
=
N
i=1
K
i
X
2
hi
= 2 2.257 10
7
= 4.514 10
7
We had calculated earlier
Thus K
h
e
2
+I
= 3.216 10
7
0.19
2
+4.514 10
7
= 4.5256 10
7
kN/m
Thus the stiffness matrix can be represented by
[K] =
_
_
_
_
3.216 10
6
0 0 0
0 4.5256 10
7
0 0
0 0 3.94 10
5
0
0 0 0 1.373 10
6
_
_
Calculation for eigen value
_
_
_
_
3.216 10
6
144.416 27.44 |
27.44 4.5256 10
7
1730 |
144.416 27.44 |
693 132 |
| 144.416 693
| 27.44 132
| 3.94 10
5
264 693
| 693 1.373 10
6
3452
_
_
= 0
Solving the above we have
79
[] =
_
_
_
_
644590
340
3650
26260
_
_
which gives the natural frequency as
[] =
_
_
_
_
803
18.43
60.41
157
_
_
rad/ sec
80
The corresponding eigen vectors are given by
[] =
_
_
_
_
0.9812 0.0675 0.0128 0.0043
0.0005 0.0009 0.0002 0.9992
0.0411 0.6156 0.9761 0.0051
0.1888 0.7851 0.2170 0.0386
_
_
Now performing the operation []
T
[M][] we have,
[]
T
[M][] =
_
_
_
_
4.9
2984.4
120.8
1722.3
_
_
Thus the scaled factors are given by
M
r1
=
4.9 = 2.213; M
r2
=
2984.4 = 54.63;
M
r3
=
120.8 = 10.99, and M

r4
=
1722.3 = 41.5.
Thus the normalized eigen vector is given by
[]
n
=
_
_
_
_
0.4438 1.2392 10
3
|
2.2593764 10
4
1.6474 10
5
|
0.018798 0.0112685 |
0.085314 0.0142961 |
79 We have solved the eigen value problem directly in Math CAD.
80 Reader to check that in case we ignore the effect of soil, the mass matrix reduces to M =
_
144.416 27.44
27.44 1730
_
and the stiffness matrix K =
_
3.216 0
0 45.256
_
10
6
and the above on
eigen value solution gives natural frequencies as 148 rad/sec and 163 rad/sec which is exactly
same as what we have obtained based on Barkan or Majors method.
| 1.1646952 10
3
1.0361446 10
4
| 1.8198362 10
5
0.0240771
| 0.08882 1.2289157 10
4
| 0.0197452 9.30212048 10
4
_
_
Calculation of damping matrix
Let the damping ratio for the RCC frame be 5% then we have
Frame No. Mass Kh Xh C
c
= 2
km C
h
(5% of C
c
) C
h
xh
2
1 44.03 1.0717 10
6
4.59 13738 687 14474
2 62.326 1.0717 10
6
0.0 16346 817 0.0
3 38.06 1.0717 10
6
4.59 12773 639 13463
144.416 2143.05 27937
Calculation for soil damping
B
y
=
(7 8) W
f
32(1 )
s
r
3
x
=
(7 2.4) 1177.2
32 19 0.7 (3.156)
3
= 0.404 and
D
y
=
0.288
B
x
=
0.288
0.404
= 0.452
C
y
= 2D
y
_
K
y
m = 2 0.452
_
3.94 10
5
120 = 6216 kN sec /m
B
=
0.375(1 )J
s
r
5
=
0.375 0.7 125 9.81
19 (2.414)
5
= 0.2066
D
=
0.15
(1 +B
)
_
B
=
0.15
1.2066 0.454
= 0.274
Thus C
= 2D
_
K
= 2 0.274

1.373 10
6
125 = 7179
KN sec/m.
The equation being dynamically coupled, the damping matrix is given by
[C] =
_
_
_
_
C
h
0 0 0
0 C
h
e
2
+C
h
X
2
h
0 0
0 0 C
y
0
0 0 0 C
_
Substituting the values calculated above we have
[C] =
_
_
_
_
2143 0 0 0
0 28014 0 0
0 0 6216 0
0 0 0 7179
_
_
Now performing the operation []
T
[C][] and for each mode separately we
have
[2D
i
i
] =
_
_
_
_
476.532 0 0 0
0 2.26 0 0
0 0 51.84 0
0 0 0 16.2462
_
_
Thus dividing each term of the above matrix by 2
i
, we have
[D
i
] =
_
_
_
_
0.296 0 0 0
0 0.0605 0 0
0 0 0.429 0
0 0 0 0.052
_
_
Calculation of load vector
Here the load vector is given by {P
h
} =
_
_
238
1092.42
0.0
1511.3
_
_
, here M
=

N
i=1
P
hi
X
hi
and M
=

N
i=1
P
hi
h and h = 3.6 +0.9 +1.1 +0.75 = 6.35 m, this is from the
center line of shaft to the c.g of the bottom raft.
Performing the operation []
T
{P} we have, []
T
{P} =
_
_
23.5575
21.9186
30.138
24.9211
_
_
Thus the four uncoupled equation of motion is given by
1
+476.532
1
+644590
1
= 23.5575sin 534t
2
+2.26
2
+340
2
= 21.9186sin 534t
3
+51.849
3
+3650
3
= 30.138sin 534t
4
+16.2462
4
+24670
4
= 24.9211sin 534t
Once we know the displacement vectors in un-coupled state the displacement
in the global structural co-ordinate is given by
{Y} = []{}
i
=
4
i=1
p
i
sin
m
t
_
(1 r
2
i
)
2
+(2D
i
r
i
)
2
The results are tabulated as follows: Here we have calculated the response for
four cases
1 One during normal operation
2 For three transient case during start and stop of the machine
3
8
0
D
y
n
a
m
i
c
s
o
f
S
t
r
u
c
t
u
r
e
a
n
d
F
o
u
n
d
a
t
i
o
n
:
2
.
A
p
p
l
i
c
a
t
i
o
n
s
Displacement under normal operation
Operating frequency = 534 rad/sec
Eigen Eigen Eigen Eigen Uncoupled
Corrected Vectors 1st Vectors 2nd Vectors 3rd Vectors 4th force Damping Frequency M.F. Uncoupled Net
Displacement Frequency frequency mode mode mode mode vector Ratio Ratio - amplitude displacement
Y 802 642 0.443 1.24 10
03
1.16 10
03
1.04 10
04
23.6 0.30 0.83 1.72 6.32 10
05
2.77 10
05
18 22 2.26 10
04
1.65 10
05
1.82 10
05
2.4110
02
21.9 0.06 24.15 0.00 1.11 10
04
3.48 10
06
U 60 72 0.0188 0.11268 0.08882 1.23 10
04
30.13 0.43 7.37 0.02 0.000154 1.17 10
08
157 188 0.0853 0.01429 0.01975 9.30 10
04
24.9 0.05 2.83 0.14 0.00014 9.88 10
06
Displacement under 1st transient
Transient frequency = 18.43 rad/sec
Vectors 1st Vectors 2nd Vectors 3rd Vectors 4th Damping Ratio Uncoupled Net
Displacement Frequency mode mode mode mode vector force Frequency M.F. amplitude displacement
Y 802.8 0.44338 1.24 1.1610
03
1.0410
04
23.55 0.296 0.02 1.00 3.67210
05
6.5510
04
18.43 2.2610
04
1.6510
05
1.8210
05
2.4110
05
21.91 0.061 1.00 8.26 0.5333 3.3610
05
U 60.41 0.018798 0.11268 0.08882 1.2310
04
30.13 0.429 0.31 1.06 0.00875 5.9310
02
157 0.085314 0.01429 0.019745 9.3010
04
24.9 0.052 0.12 1.01 0.001025 7.8010
03
A
n
a
l
y
s
i
s
a
n
d
d
e
s
i
g
n
o
f
m
a
c
h
i
n
e
f
o
u
n
d
a
t
i
o
n
s
3
8
1
Displacement under 2nd transient
Transient frequency = 60.41 rad/sec
Vectors 1st Vectors 2nd Vectors 3rd Vectors 4th force Damping Frequency M.F. Uncoupled Net
Displacement Frequency mode mode mode mode vector Ratio Ratio - amplitude displacement
Y 802.8 0.44338 1.2410
03
1.1610
03
1.0410
04
23.55 0.296 0.08 1.00 3.68710
05
3.1810
05
18.43 2.2610
04
1.6510
05
1.82 10
05
2.4110
02
21.91 0.0605 3.28 0.10 0.006617 2.8810
05
60.41 0.018798 0.112685 0.08882 1.2310
04
30.13 0.429 1.00 1.17 0.009625 1.0910
03
157 0.085314 0.014296 0.019745 9.3010
04
24.92 0.052 0.38 1.17 0.00118543 2.87 10
04
Displacement under 3rd transient
Transient frequency = 157 rad/sec
Vectors 1st Vectors 2nd Vectors 3rd Vectors 4th force Damping Frequency M.F. Uncoupled Net
Displacement Frequency mode mode mode mode vector Ratio Ratio - amplitude displacement
Y 802.8 0.44338 1.24 10
03
1.16 10
03
1.04 10
04
23.55 0.296 0.20 1.03 3.789 10 05 1.31 10
05
18.43 2.2610
04
1.65 10
05
1.8210
05
2.41 10
02
21.91 0.0605 8.52 0.01 0.00090156 2.34 10
04
U 60.41 0.018798 0.112685 0.08882 1.2310
04
30.13 0.429 2.60 0.16 0.001334 1.7710
05
157 0.085314 0.014296 0.019745 9.30 10
04
24.92 0.052 1.00 9.62 0.009722 3.3510
05
Net amplitude of the frames
The net amplitude of the frame is given by the expression
yi
net
= y
i
+U
i
+X
hi
+h
where h = 4.8 m and X
hi
= +4.59 m, 0.0, 4.59 m respectively for
Frame 1, 2 and 3.
The results are as shown hereafter
Operation condition Frame 1 Frame 2 Frame 3
Normal operating case 3.57 10
05
1.97 10
05
3.78 10
05
1st Transient (18.43 rad/sec) 9.76 10
02
9.74 10
02
9.73 10
03
2nd Transient (60 rad/sec) 1.4 10
03
1.27 10
03
1.14 10
03
3rd Transient (157 rad/sec) 1.20 10
03
1.30 10
04
9.45 10
04
The results are shown based on bar chart as given in Figure 2.18.23.
Amplitude of the frames
-2.00E-02
0.00E+00
2.00E-02
4.00E-02
6.00E-02
8.00E-02
1.00E-01
1.20E-01
Frame1 Frame2 Frame3
Frame Numbers
D
i
s
p
l
a
c
e
m
e
n
t

o
f

t
o
p

d
e
c
k
Normal operating force
1st transient
2nd transient
3rd transient
Figure 2.18.23 Comparison of displacement operating vs transient.
The results of natural frequencies are compared based on various methods.
Frequency in vertical direction
Mode number Rausch Barkan Major 2D soil structure interaction
1st Mode 409 427
2nd Mode 8318 10265 8318 15288
3rd Mode 21575 22040
Frequency in horizontal direction
Mode number Rausch Barkan Major 2D soil structure interaction
1st Mode 176
2nd Mode 1616 1413 1418 577
3rd Mode 1557 1562 1547
4th Mode 7667
Comparison of results with time history response
Here we compare the results obtained above with time history response which we
have stated earlier as the most appropriate method of analysis for such coupled
soil-structure interaction analysis
We show below the plots (Figures 2.18.24 and 25) for the amplitude
of vibration and displacement of the frames under operating frequency of
534 radians/sec.
Amplitude based on Newmark - method
-0.00004
-0.00003
-0.00002
-0.00001
0
0.00001
0.00002
0.00003
0.00004
0
0
.
0
4
7
5
0
.
0
9
5
0
.
1
4
2
5
0
.
1
9
0
.
2
3
7
5
0
.
2
8
5
0
.
3
3
2
5
0
.
3
8
0
.
4
2
7
5
0
.
4
7
5
0
.
5
2
2
5
0
.
5
7
0
.
6
1
7
5
Time steps in seconds
A
m
p
l
i
t
u
d
e
Y
Figure 2.18.24 Amplitudes under operating condition having frequency @ 534 rad/sec.
Amplitude of frames at top deck
-1.50E-04
-1.00E-04
-5.00E-05
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
0
0
.
0
3
0
.
0
7
0
.
1
0
.
1
3
0
.
1
6
0
.
2
0
.
2
3
0
.
2
6
0
.
2
9
0
.
3
3
0
.
3
6
0
.
3
9
0
.
4
2
0
.
4
6
0
.
4
9
0
.
5
2
0
.
5
5
0
.
5
8
0
.
6
2
Time steps (sec)
A
m
p
l
i
t
u
d
e

(
m
e
t
e
r
)
Y1
Y2
Y3
Figure 2.18.25 Amplitude of displacement of the top deck under operating frequency of
534 rad/sec.
Under transient load like Major we assumed the operating frequency in reso-
nance with the natural frequency instantaneously and considering the function,
sin
m
t = 1.
For time history response we consider the operating frequency equal to first,
second and third transient frequency respectively and find out the transient
response (peak amplitude). The results are plotted graphically in Figures 2.18.26
through 31.
-0.01
-0.008
-0.006
-0.004
-0.002
0
0.002
0.004
0.006
0.008
0.01
0
0
.
4
5
0
.
9
1
.
3
5
1
.
8
2
.
2
5
2
.
7
3
.
1
5
3
.
6
4
.
0
5
4
.
5
4
.
9
5
5
.
4
5
.
8
5
A
m
p
l
i
t
u
d
e
Y
U
Amplitude based on Newmark- method
Figure 2.18.26 Amplitudes at the rst transient frequency @ 18.43 rad/sec.
-5.00E-02
-4.00E-02
-3.00E-02
-2.00E-02
-1.00E-02
0.00E+00
1.00E-02
2.00E-02
3.00E-02
4.00E-02
5.00E-02
0
0
.
3
0
.
6
0
.
9
1
.
2
1
.
5
1
.
8
2
.
1
2
.
4
2
.
73
3
.
3
3
.
6
3
.
9
4
.
2
4
.
5
4
.
8
5
.
1
5
.
4
5
.
76
Time steps (sec)
A
m
p
l
i
t
u
d
e

(
m
e
t
e
r
)
Y1
Y2
Y3
Figure 2.18.27 Amplitude of top deck at the rst transient frequency @ 18.43 rad/sec.
-0.0015
-0.001
-0.0005
0
0.0005
0.001
0.0015
0
0
.
2
0
.
4
0
.
6
0
.
8
1
1
.
2
1
.
4
1
.
6
1
.
8
2
2
.
2
2
.
4
2
.
6
2
.
8
3
A
m
p
l
i
t
u
d
e
Y
U
Figure 2.18.28 Amplitudes at the second transient frequency @ 60 rad/sec.
-2.50E-03
-2.00E-03
-1.50E-03
-1.00E-03
-5.00E-04
0.00E+00
5.00E-04
1.00E-03
1.50E-03
2.00E-03
2.50E-03
0
0
.
1
5
0
.
3
0
.
4
5
0
.
6
0
.
7
5
0
.
9
1
.
0
5
1
.
2
1
.
3
5
1
.
5
1
.
6
5
1
.
8
1
.
9
5
2
.
1
2
.
2
5
2
.
4
2
.
5
5
2
.
7
2
.
8
5
3
Time steps (sec)
A
m
p
l
i
t
u
d
e

(
m
e
t
e
r
)
Y1
Y2
Y3
Figure 2.18.29 Amplitudes of top deck at the second transient frequency @ 60 rad/sec.
-0.0003
-0.0002
-0.0001
0
0.0001
0.0002
0.0003
0
0
.
0
5
0
.
0
9
0
.
1
4
0
.
1
8
0
.
2
3
0
.
2
7
0
.
3
2
0
.
3
6
0
.
4
1
0
.
4
5
0
.
5
0
.
5
4
0
.
5
8
A
m
p
l
i
t
u
d
e
Y
U
Figure 2.18.30 Amplitudes at third transient frequency @ 157 rad/sec.
-2.00E-03
-1.50E-03
-1.00E-03
-5.00E-04
0.00E+00
5.00E-04
1.00E-03
1.50E-03
2.00E-03
0
0
.
0
3
0
.
0
5
0
.
0
8
0
.
1
0
.
1
3
0
.
1
5
0
.
1
8
0
.
2
0
.
2
3
0
.
2
5
0
.
2
8
0
.
3
0
.
3
3
0
.
3
5
0
.
3
8
0
.
4
0
.
4
3
0
.
4
5
0
.
4
8
0
.
5
0
.
5
3
0
.
5
5
0
.
5
7
0
.
6
Time steps (sec)
A
m
p
l
i
t
u
d
e

(
m
e
t
e
r
)
Y1
Y2
Y3
Figure 2.18.31 Amplitude of top deck at third transient frequency @ 157 rad/sec.
Discussions of the results
The basic amplitudes as obtained by the two methods are shown hereafter
Modal Time Critical
Sl. No. Displacements response history frequency Remarks
1 Y 2.77 10
05
2.617 10
05
534 This is under operating case
2 2.34 10
04
2.09 10
04
157 Operating frequency in
resonance with this mode
3 1.09 10
04
8.33 10
04
60.41 Operating frequency
in resonance with this mode
4 7.80 10
03
7.38 10
03
18.43 Operating frequency in
resonance with this mode
The displacement of the top deck is as shown hereafter
Modal Time Critical
Sl. No. Displacements response history frequency Remarks
1 Y 3.57 10
05
1.6 10
04
534 This is under operating case
2 Y 1.20 10
03
1.31 10
03
157 Amplitude at third transient
3 Y 1.4 10
03
2.21 10
03
60 Amplitude at second transient
4 Y 9.76 10
02
4.12 10
02
18.43 Amplitude at first transient
Discussion on the results
It is pretty obvious from the results that time history gives a clearer picture so far as
the amplitude is concerned.
While by modal technique, taking sin
m
t = 1, we calculate the maximumamplitude
at various resonance points but actually at the time when the operating frequency
passes this particular frequency sin
m
t = 1 but something lesser than that. Since we
do not know this data as a conservative value this is usually taken as unity. Thus it is
evident from the result that while by time history we get a value of deflection of top
deck as 41 mm this gives about 98 mm by modal technique. While in all other cases
though the order of displacement does remain same they do show discrepancy in top
deck amplitude.
Specially under operating condition while modal response gives amplitude of
deection of top deck as 3.57E-05 m time history gives an amplitude of 1.6E-04
(about 5 times). This can be attributed to the fact that while by modal response we
only find the steady state part, by time history the transient part of the vibration is
also considered in the response and this possibly gives an initial higher response at
operating frequency.
Thus for resonance check we can followthe modal response technique to find out the
eigen values, however for amplitude check carrying out time history response specially
to check the transient response is much more sensible.
Based on the above calculation following points can be concluded
Soil has significant influence on the natural frequency of the system.
Majors hypothesis of transient response is correct, and there too soil could
play a signicant role (Though Majors method of calculating the response is
conservative).
Increasing the thickness of bottom raft does not necessarily help. It could reduce
the peak operating frequency response but will not have much effect on the
transient response of the system during start and stop of the machine.
Transient response can be signicant, and all the signicant transients to be
checked for the frame.
Though this may not structurally effect the frame (for soil transients would
mostly be rigid body displacements) its effect on the stress level of the connected
piping and load induced at nozzles could be significant.
For critical structures amplitude should preferably be checked based on time his-
tory response to ensure that amplitudes are within acceptable limit. This should be
done specially under operating frequency where modal technique usually ignores
the transient part and only gives the maximum value for the steady state part.
SUGGESTED READING
1 Barkan, D.D. 1962, Dynamics of Bases and Foundations, McGraw-Hill Book Co. NY.
2 Srinivasalu, P. & Vadiyanathan, C.V. 1977, Handbook of Machine Foundations, Tata
Mcgraw-Hill, New Delhi.
3 Major, A. 1980, Dynamics in Civil Engineering Analysis and Design, Vols. IIV,
Akademia Kiado, Budaapest and Collets Holding London.
4 Arya, S.C., ONeill, M.W. & Pincus, G. 1979, Design of Structures and Foundations for
Vibrating Machines, Gulf Publishing Co., Houston, Texas.
5 Kameswara Rao, N.S.V. 1998, Vibration Analysis and Foundation Dynamics, Wheeler
Publishing, New Delhi.
6 Verma, C.V.J. &Lal, P.K. Ed., Treatise on the design, analysis and testing of High capacity
Turbo Generator foundation, Central Board of Irrigation and Power Publication #262.
We also furnish some selected papers which we feel would
further ameliorate your insight to the problem
1 Almuti, A.M. 1976, Large exible Turbine foundation, Methods of Structural Analysis,
ASCE, NY, pp. 707719.
2 Aneja, I. 1975, Dynamic Response of Systems Turbine generators on Various Founda-
tions, Proceedings of the American Power conference, Vol. 37, pp. 528540.
3 Arya, A.S. & Drewer, R. 1997, Mathematical modeling and computer simulation of ele-
vated foundations supporting vibrating Machinery, Transaction of IMACS, Vol. XIX,
No. 4, Dec.
4 Design Criteria for Turbine Generator Pedestal 1970, Journal of Power Division, ASCE,
Vol. 96, Jan, pp. 122.
5 Kasten, H.L. & Kirkland, W.D. 1970, Spring mounted Turbine Generator Spins Quitely,
Efciently, Electric Light and Power, E/D Edition, Nov., pp. 3840.
6 Shen, G.T. & Stone, N.E. 1975, Natural frequencies of turbine foundation, Structural
Design of Nuclear Plant Facilities, Vol. II, ASCE, NY, pp. 302330.
7 Srinivasulu, P. &Lakshmannan, N. 1978, Dynamic response of turbo-generator pedestal,
ASCE, Spring convention, Pittsburgh, Pensylvania, April, pp. 2428.
8 Chowdhury, I. & Som, P.K. 1993, Dynamic Pile structure interaction of Boiler Feed Pump
Frame Foundation, Indian Geotechnical Conference, Vol. 1, pp. 411414.
9 Task Committee on Turbine Foundations 1987, Design of large steam turbine-generator
foundations, ASCE, NY.
10 Rausch, E. 1959, Maschinen Fundamente und andere dynamisch beanspruchte Baukon-
structionen, VDI Verlag, Dusseldorf.
11 Wedpathak, A.V., Pandit, V.K. Guha, S.K. 1977, Soil-Foundation interaction under sinu-
soidal and impact type dynamic loads, Int. Symp. on Soil-Structure Interaction, University
of Roorkee, Roorkee.
Chapter 3
Analytical and design concepts
for earthquake engineering
3.1 INTRODUCTION
In this chapter we will deal with some of the fundamental concepts pertaining to
earthquake engineering.
On completion of this chapter you should have an understanding of
Why earthquake happens in nature.
Essential engineering parameters, which affect the geo-technical and structural
aspect of a system under earthquake.
Basic concepts of dynamic analysis as applied to Earthquake engineering pertain-
ing to buildings, and different types of industrial and infra-structural systems like
chimney, retaining wall, water tank, RCC and earth dams etc.
Have an understanding of different provisions of IS 1893 (2002) code.
Before reading this chapter we however feel that you should have following back-
ground as a pre-requisite.
1 Basic concepts in structural and soil dynamics as furnished in Chapter 5 (Vol. 1).
2 Also have some fundamental awareness of how earthquake can affect a structure-
foundation system.
Earthquake is perhaps the most complex natural phenomenon which human being
is trying to understand, combat and harness, from the early history of mankind. In
spite of scientific study of the subject for the last 100 years or so, it is felt that we
are still in the infancy of our knowledge on the subject. The parameters affecting
this phenomenon are so large and varying and also covering different branches of
science, we can at best arrive at a simplified model of the problem amenable to human
perception and try to arrive at a solution which would in all probability survive this
natures assault with some limited damage, if ever the structure faces such vagary.
The basic objective of an earthquake resistant design is not to make the structure
fool proof but to limit its damage to the extent of minimizing the loss of human life
and property.
Though earthquake is a global phenomenon, yet there are some countries in
the world like USA, Japan, Turkey, India, Iran, Newzealand etc that are severely
affected by earthquakes leading to signicant loss of human life and properties, while
there are others whose geological characteristics are considered seismically inert like
United Kingdom, Gulf countries like Oman, Kuwait, UAE, Qatar etc. which have no
signicant history of earthquakes.
Based on the above, it is evident that there are countries where significant research
and investigation have been carried out to develop procedures for earthquake resistant
design of structures. Countries like USA, Japan, India, Mexico etc have contributed
significantly on this issue.
3.1.1 Why do earthquakes happen in nature?
The topic itself can be subject matter of a complete book. A detailed discussion on
this is beyond the scope of this work, however as civil engineers to design structures,
which can withstand such calamity- some fundamental understanding on this issue is
essential.
As shown in Figure 3.1.1 the earth constitute of a central core, consisting of molten
magma which is undergoing continuous upheaval. While the outer core, which has
solidified in million of years forms the outer earth crust. The inner magma (the molten
core) is continuously creating a pressure on the outer core and trying to come out by
seeking some weaknesses in earth crust. Whenever it can come out it generates what
is known as a volcanic eruption. When it cannot, it tries to push the crust upward
thus creating folds and faults resulting in a source which stores a significant amount
of potential/strain energy.
As by law of nature, all systems in course of time try to achieve minimum state
of energy; these storehouses of potential energy keep on releasing their stored strain
energy as kinetic energy generating waves on surface of the outer crust which is com-
monly known as earthquake. It is said that Himalayan mountain range is one such
formation due to pressure of the inner magma. Deformation which the earth crust
underwent, due to formation of the mountain range, is still being adjusted naturally.
Earth Crust
Earth Core
Molten Magma
Figure 3.1.1 Earth with its central core.
Analytical and design concepts for earthquake engineering 391
It is for this reason, areas in its close proximity like Assam, Nepal and portions of
south China is often subjected to severe earthquakes.
There is also a phenomenon called seismotectonic movement otherwise known as
continental drift that generates earthquake at certain location of the earth. According
to this theory, the outer crust of earth is made up of undistorted plates of lithosphere.
These plates are in differential motion, and at places they move away from each other
where new plates are added from the interior of the earth while in places they collide
with each other.
All major earthquakes which mark the active zones of the earth closely follows the
plate boundaries and has been found to be a function of the movements of these plates
(Stevens 1980).
Human interference can also sometimes modify stresses on the earth surface to
trigger minor or even moderate earthquakes. In many mining areas tremors and shocks
results due to underground explosion in mines, causing damages to structures on
ground.
One of the classic cases of man made earthquake was Koyna Dam incident in 1967
in India, when pounding of large amount of water behind the dam resulted in an
earthquake causing extensive damage to surrounding (Chopra & Chakrabarti 1973).
3.1.2 Essential difference between systems subjected
to earthquake and vibration from machine
In Chapter 5 (Vol. 1) and Chapter 2 (Vol. 2) we had discussed in detail response
of machine foundation under dynamic loading. In machine foundations, unbalanced
force from the machine gets transmitted to ground via structure/foundation to the
soil media. In such cases normally a limited part of the soil is affected significantly.
Moreover the strain range induced in soil is usually limited to lowstrain range (usually
10
3
%). However in case of earthquake the phenomenon is quite different. In this
case when an earthquake shock is generated due to rupture of a fault within the earth
surface it generates waves within soil that induces a much larger strain (10
2
to 10
1
%)
for a major earthquake.
Shown in Figure 3.1.2, is a typical propagation of waves through the soil medium
and is usually a combination of four types of waves namely,
1 P-waves (body waves)
2 S-waves (body waves)
3 Rayleigh Waves (surface waves)
4 Love waves (surface waves)
The primary or P-waves are the fastest traveling of all waves and generally produce
longitudinal compression and extension within a soil medium. This wave can travel
both through soil and water and is the first one to arrive at a site. However soil being
relatively more resistant to compression and dilation, effect of its impact on ground
distortion is minimal.
The S-waves, also otherwise known as secondary or shear waves usually cause shear
deformation in the medium through which they propagate. The S-waves can usually
Time History response Velocity
-0.3
-0.2
-0.1
0
0.1
0.2
0
1
.
0
8
2
.
1
6
3
.
4
2
4
.
3
2
5
.
4
6
.
4
8
7
.
5
6
8
.
6
4
9
.
7
2
1
0
.
8
1
1
.
9
1
3
1
4
1
5
.
1
1
6
.
2
1
7
.
3
1
8
.
4
1
9
.
4
Time steps
V
e
l
o
c
i
t
y

v
(
m
/
s
e
c
)
Figure 3.1.2 Typical propagation of earthquake waves through surface.
propagate through soil only
1
. It travels at a much slower speed through the ground
than primary waves and soil being weak in resisting shear deformation; it is found to
cause maximum damage to ground surface.
Rayleigh waves are surface waves which are found to produce ripples on surface
of the ground
2
. These waves produce both horizontal and vertical movement of earth
surface as the waves travel away from the source.
Love waves are similar to S-waves and produces transverse shear deformation to
the ground.
These entire waves combine together to produce shock waves from which an engi-
neer extracts value of the maximum ground acceleration (a
max
) which is the major
parameter that governs his design.
Based on above it is apparent that mechanics of earthquake is opposite to dynamics
of machine foundation in the sense that here forces are transmitted from soil to the
structure. It is the shock within the ground which excites the structure and induces
inertial force in the system.
3.1.3 Some history of major earthquakes
around the world
A number of major earthquakes have been recorded that resulted in massive losses of
human lives and destructions of thousands of buildings and structures. Some of them
are cited in Table 3.1.1.
1 Since liquid have no shear resistance it cannot travel through water.
2 This is very much similar to ripples produced by a pebble dropped in a pond.
Table 3.1.1 Various Earthquakes & their casualities.
Name, Location & Year Casualty
Calcutta, 1737 Destroyed 300 000 lives.
Portugal, Spain & Northern Devastated Lisbon, loss of life was heavy. The disaster was
Morocco, subjected to three more detrimental since the first shock was followed by a
strong shocks in the afternoon colossal whirling wall of water sweeping everybody and
of Nov. 1, 1775 everything it met on its path.
The Alma-Ata earthquake, 1910 Demonstrated continuous vibrations, that lasted for 5 minutes.
Tokyo and Yokohama, Sept. 1, 1923 11 000 buildings were ruined, 59 000 houses burned
in Yokohama. Entire area of Tokyo was affected. Death
toll was 100 000, while 43 000 missing. 300 000 houses
were damaged.
Himalayan earthquake, 1950 One of the severest events recorded instrumentally.
Equivalent to the explosion of 100 000 A-bombs.
Mongolian earthquake, Vast devastation, A mountain peak was split into two parts.
December 4, 1956 Part of mountain, 400 m high, collapsed and fell down a
precipice. A depression, up to 18 km in length and 800 m
in width, originated. Broad fissures, up to 20 m in width
appeared on the ground surface. One of these broke for
length of 250 km. The intensity of the earthquake
approached a force 11.
Alaskan earthquake, 1964 The most severe of all known seismic events. Intensity was
over 11.
Chile, 1960 The most violent earthquake of the 20th century. It affected
an area over 200 000 km
2
and caused numerous landslides.
Latur and Osmanabad, Magnitude of earthquake was 6.3. Total of 7601 people lost
September 1993 their lives in Latur and Osmanabad. Number of houses
destroyed in the earthquake was about 30 000.
Bhuj earthquake, 26 January 2001 Some historical structures that survived the 1819 (M = 7.7)
earthquake have been destroyed in the 2001 earthquake.
The death toll was 19 727 and the number of injured
166 000. Indications are that 600 000 people were left
homeless, with 348 000 houses destroyed and an additional
844 000 damaged. Magnitude was 7.6.
Liquefaction On 4 February liquefaction phenomenon
were reported by hydrologists and by local villagers, with
an indication that the flow was sufficient in some cases
to activate desert rivers that have been dry for more than
a century. Widespread liquefaction was confirmed by
SPOT imagery and by field observation (5 Feb.). Many mud-
volcanoes in the Rann of Kachchh have dimensions of
hundreds of meters: one covers a 5 km diameter stretch of
the southern Rann with dark sand and mud. Numerous
ancient river channels have been illuminated by a pock
mark pattern of sand vents, and some have clearly flowed,
and breached their old channels.
3.1.4 Intensity
The severity of shaking of an earthquake as felt through damage is described as inten-
sity at a certain place on an arbitrary scale. One such scale is Modied Mercalli Scale
(MMS). This is shown in Table 3.1.2.
Table 3.1.2 Modified Mercalli Intensity Scale.
Class of
earthquakes Description
I Not felt except by a few under specially favourable circumstances.
II Felt by a few persons at rest, specially on upper floors of building; and delicately
suspended objects may swing.
III Felt quite noticeably indoors; specially on upper floors of buildings but many people
do not recognize it as an earthquake; standing motor cars may rock slightly; and
vibration may be felt like the passing of a truck.
IV During the day felt indoors by many, outdoors by a few; at night some awakened,
dishes, windows, doors disturbed, walls make cracking sound, sensation like
heavy truck striking the building; and standing motor car rocked noticeably.
V Felt by nearly everyone; many awakened; some dishes, windows etc. broken; a few
instances of cracked plasters; unstable objects overturned; disturbance of trees;
poles and other tall objects noticed sometimes and pendulum clocks may pop.
VI Felt by all; many frightened and run outdoors; some heavy furnitures moved; a
few instances of fallen plaster or damaged chimneys; damage slight.
VII Everybody runs outdoors, damage negligible in buildings of good design and
construction; slight to moderate in well built ordinary structures, considerable
in poorly built or badly designed structures; some chimneys broken; noticed
by persons driving motor cars.
VIII Damage slight in specially designed structures; considerable in ordinary substantial
buildings with partial collapse; very heavy in poorly built structures; panel walls
thrown out of framed structures; heavy furniture overturned; sand and mud
ejected in small amounts; changes in well water, and disturbs persons driving
motor cars.
IX Damage considerable in specially designed structures; well designed framed
structures thrown out of plumb; damage very heavy in substantial buildings with
partial collapse; buildings shifted off foundations; ground cracked conspicuously;
and underground pipes broken.
X Some well built wooden structures destroyed; most masonry and framed
structures with foundations destroyed; ground badly cracked; rails bent; land-slides
considerable from river banks and steep slopes; shifted sand and mud; and
water splashed over banks.
XI Few, if any, masonry structures remain standing; bridge destroyed; broad fissures
in ground, underground pipe lines completely out of service; earth slumps and
landslips in soft ground; and rails bent greatly.
XII Total damage; waves seen on ground surface; lines of sight and level distorted; and
objects thrown upward into the air.
3.1.5 Effect of earthquake on soil-foundation system
Having explained the primary source of disturbance is in the soil itself, it is important
to assess and know what could be the effects of an earthquake on the soil on which
a structure is built. For it should be understood that irrespective of how well an
earthquake resistant design is carried out for a structure if the ground supporting it
fails, the structure will invariably undergo significant damage and which at times could
even be catastrophic
3
.
The major effect on soil affected by an earthquake can be classified as follows:
1 Liquefaction of soil
2 Settlement of foundation due to deep seated liquefaction failure
3 Reduction of bearing capacity
4 Ground Subsidence
5 Land Slides
Of all the phenomena defined above, liquefaction is perhaps the most important
factor that has caused major damage in many previous earthquakes, and unfortu-
nately gets very little attention from structural engineers in a design office
4
. Thus it is
important to understand what the phenomenon is and what are the methods available
to assess and mitigate it?
3.1.6 Liquefaction analysis
3.1.6.1 What is liquefaction?
Conceptually speaking liquefaction is very much akin to giving a rapid squeeze to a
sponge ball saturated with water. When the squeeze is applied, we observe that water
stored inside the sponge comes out and the sponge feels lighter as the water comes out.
For soil sample (especially when it is cohesionless) shear strength is given by the
expression
s = ( u) tan (3.1.1)
where, s = shear strength of the soil; = overburden pressure of the soil sample; u =
in-situ pore pressure within the soil sample, and = angle of internal friction of the
soil sample.
When earthquake force acts on the soil sample it produces a rapid shock or a squeeze
on the soil body, by virtue of which there is a sudden increase in pore pressure. But
unlike the sponge ball the pore pressure cannot dissipate readily.
When force due to earthquake is significantly high (M 6.5) which also results
in ground shaking for a good amount of time the pore pressure increment becomes
sufficiently high such that it equals the overburden pressure and the soil looses its shear
3 Nigaata Earthquake (1964) in Japan was one of the primary example where a number of structures
underwent significant damages due to ground subsidence and liquefaction of soil.
4 Especially in India where in previous earthquakes a significant damage has been recorded due to this
phenomenon.
strength altogether (i.e. s = 0) and starts flowing like a liquid. This phenomenon is
otherwise known as liquefaction of soil.
When such phenomenon is observed during an earthquake soil collapses completely
and sand boils are observed in the ground. Even c- soils losses signicant part of
its strength resulting in bearing capacity failures of foundation and or signicant
settlement.
Liquefaction of soil has been observed in a number of earthquakes throughout the
world like Nigaata in Japan (1964), Kobe in Japan (1995), Dhubri and Koyna (1967)
earthquakes in India.
From the above discussion it is obvious that non-plastic cohesionless soils under
saturated condition are most susceptible to liquefaction during an earthquake.
As SPT value has been extensively used to define the static engineering strength
of cohesionless soil consistently it was but natural that researchers tried to co-relate
SPT values of cohesion less sandy soil to liquefaction potential of soil samples due to
earthquake shocks. Pioneering research in this area was done by Seed et al. (1984)
who correlated the observed SPT values to cyclic resistance ratio which is one of the
major parameters used to define the liquefaction potential of a soil sample. We will
talk more about this later; first let us see how liquefaction is measured for a particular
soil sample.
The susceptibility of a soil sample undergoing liquefaction is measured by a term
called liquefaction potential, which is measured as a Factor of Safety (FS) against
Cyclic Resistance Ratio (CRR) to Cyclic Stress Ratio (CSR). It is defined as
FS =
CRR
CSR
1.0 (3.1.2)
In other words (based on Equation (3.1.2)), if the factor of safety is less than or
equal to 1.0, the soil has very good possibility of undergoing liquefaction under an
earthquake, however if the value is greater than 1.0, the possibility of soil failure due
to liquefaction is remote.
Thus it is obvious that we need to first understand what does CSR and CRR stand
for. During earthquake soil under the influence of an earthquake will be subjected to
repetitive shear stress (known as cyclic shear stress) and it is estimated by the expression
CSR =

av
v
= 0.65
_
a
max
g
__
v
_
r
d
(3.1.3)
where, a
max
= maximum acceleration at the ground surface;
v
= total overburden
pressure at the design depth;
v
= effective overburden pressure at the design depth;
g = acceleration due to gravity, and, r
d
= stress reduction factor which varies with
depth and is given by
r
d
= 1.0 0.000765z for z 9.15 m (3.1.4a)
r
d
= 1.174 0.0267z for 9.15 m z 23 m (3.1.4b)
r
d
= 0.744 0.008z for 23 m z 30 m (3.1.4c)
r
d
= 0.5 for z 30 m (3.1.4d)
For ease of electronic computation r
d
may also be expressed by the expression (Blake
et al. 2002)
5
r
d
=
_
1.000 0.4113z
0.5
+0.04052z +0.001753z
1.5
_
_
1.000 0.4177z
0.5
+0.5729z 0.006205z
1.5
+0.001210z
2
_ (3.1.5)
The maximum acceleration of the ground (a
max
) is another factor, which needs
careful evaluation.
For practical design office purpose one of the expressions used to evaluate a
max
is
a
max
= 0.18410
0.320M
(D)
0.8
g (3.1.6)
where a
max
= maximum ground acceleration; M = expected moment magnitude of
earthquake, D = maximum epicenter distance in km, and g = acceleration due to
gravity @ 9.81 m/sec
2
.
It may be noted that if more reliable observed earthquake data is available for the
site (predicting ground acceleration more accurately) it may well be used in lieu of the
above formula.
Having calculated the cyclic stress ratio based on the above expressions it is essential
to evaluate the cyclic resistance ratio (CRR) of the in-situ soil.
It is evident that the CRR value of the soil sample will depend on its in-situ strength.
Since Laboratory testing can be carried out under a better controlled environment,
one of the plausible methods which have been tried is to collect in-situ undisturbed
soil sample for evaluation of the parameter CRR in the laboratory.
However, one of the major difculties encountered in this case is that generally the
in-situ stress state cannot be established in the laboratory, and specimens of granu-
lar soil retrieved with typical drilling techniques are far too disturbed to yield any
meaningful results.
Only through very specialized sampling techniques such as ground freezing, suf-
ciently undisturbed sample can be obtained, which again becomes prohibitively
expensive for all but most critical projects.
It is for this reason, co-relating the CRR value with field observed test data is still
the state of the art practice.
3.1.6.2 Co-Relation between CRR and SPT value
For calculation of CRRbased on observed SPT value (N
o
), as a first step, the observed
SPT value is subjected to certain corrections is as expressed by
(N
1
)
60
= N
o
(C
N
)(C
E
)(C
B
)(C
R
)(C
S
) (3.1.7)
in which, N
o
= measured SPT value at the site; C
N
= a correction factor for overbur-
den pressure; C
E
= a correction factor for hammer energy ratio; C
B
= a correction
5 This formula was proposed as guidelines for analyzing and mitigating landslide hazards in California,
Southern California Earthquake Center, Univ. of Southern California, Los Angeles.
Table 3.1.3 Correction factors to observed SPT values.
Factor Equipment parameter Term Correction factor
Overburden pressure Independent of Equipment C
N
_
P
a
v
Energy ratio Safety Hammer C
E
0.6 to 1.17
Doughnut Hammer 0.45 to 1.0
Rod length 3 to 4 m C
R
0.75
4 to 6 m 0.85
6 to 10 m 0.95
10 to 30 m 1.0
>30 m >1.0
Bore Hole Diameter 65 to 115 mm C
B
1.0
150 mm 1.05
200 mm 1.15
Sampling Method Standard Sampler C
S
1.0
Sampler without Liners 1.2
In the table, P
a
= atmospheric pressure or 100 kPa (100 kN/m
2
);
v
= effective overburden pressure at depth
of the standard penetration sample.
factor for borehole diameter; C
R
= a correction factor for rod length; C
S
= a cor-
rection factor for sampler with or without liners, and, (N
1
)
60
= corrected SPT value
with 60% hammer efciency.
The correction factors for various equipment parameters are as shown in
Table 3.1.3.
Having established the design SPT value (N
1
)
60
the cyclic resistance ratio (CRR) is
given by the expression for clean sands (i.e. <5% contents) as
CRR =
a +by +cy
2
+dy
3
1 +ey +fy
2
+gy
3
+hy
4
(3.1.8)
where, a = 0.048, b = 0.004721, c = 0.0006136, d = 1.67310
5
, e = 0.1248,
f = 0.009578, g = 0.0003285, h = 3.71410
6
, and y = (N
1
)
60
.
Equation (3.1.8) is valid for (N
1
)
60
less than 30. For clean granular soil having
N > 30 are far too dense to liquefy and are generally classed as non-liquefiable.
Another expression, which is used for clean sand base for computation of CRR is
6
CRR
7.5
=
1
34 (N
1
)
60
+
(N
1
)
60
135
+
50
[10(N
1
)
60
+45]
2

1
200
(3.1.9)
where, CRR
7.5
= the cyclic resistance ratio at earthquake magnitude of 7.5.
6 After Alan. F. Rauch at the University of Texas, 1998.
3.1.6.3 Inf luence of f ine contents on CRR value
While developing the original expression Seed et al. (1984) noted an apparent increase
of CRR value with an increase in fine contents. Whether this can be attributed to an
increase in resistance or decrease in penetration resistance is not clear.
However to cater to this, correction has been recommended to SPT values for the
influence of fine contents. Other grain characteristics like Plasticity index (PI) may
also affect the liquefaction resistance as well, however is not so well defined till date.
Hence, corrections based solely on fine contents are used and should be mellowed with
judgment and caution.
Seed et al. (1983) proposed corrections of (N
1
)
60
to an equivalent clean sand value
(N
1
)
60CS
given by
(N
1
)
60CS
= + (N
1
)
60
(3.1.10)
where and are determined from the following relationships as shown in
Table 3.1.4.
Table 3.1.4 Modification factor to SPT value based on fine contents.
Sl. No. Values of and Fine content
1 = 0 For FC 5%
2 = e
_
1.76
_
190
FC
2
__
5% FC 35%
3 = 5.0 FC 35%
4 = 1.0 For FC 5%
5 =
_
0.99 +
_
FC
1.5
1000
__
5% FC 35%
6 = 1.2 FC 35%
The above equations can now be used for routine liquefaction resistance calculation
for soil subjected to SPT at field.
3.1.6.4 Effect of earthquake magnitude on liquefaction resistance
The original study of the liquefaction potential was based on an earthquake magni-
tude of 7.5. To evaluate the potential of earthquake at other magnitudes, correction
factors were proposed that allows induced stress ratios for other magnitudes be
adjusted to a magnitude of 7.5 by dividing the stress ratios by the factors as shown in
Table 3.1.5. The magnitude scaling factor (MSF) as proposed in Table 3.1.5 based
on recent research is now believed to be very conservative for moderate size earth-
quake. A new set of MSF has now been proposed by Idriss where the MSF is defined
as function of Moment Magnitude and is given by
MSF =
10
2.24
M
2.56
(3.1.11)
Table 3.1.5 Magnitude scaling factor as proposed by Seed and Idriss (1970).
Sl. No. Earthquake magnitude Magnitude scaling factor
1 5.25 1.5
2 6 1.32
3 6.75 1.13
4 7.5 1.0
5 8.5 0.89
Table 3.1.6 Magnitude scaling factor as proposed by various investigators**.
Seed and
Idriss (original) Idriss Ambreseys Andrus &
Sl. No. Magnitude (1970) (1999) Arango (1995) Stokoe
1 5.5 1.43 2.2 3 2.2 2.86 2.8
2 6 1.32 1.76 2 1.65 2.2 2.1
3 6.5 1.19 1.44 1.6 1.4 1.69 1.6
4 7 1.08 1.19 1.25 1.1 1.3 1.25
5 7.5 1 1 1 1 1 1
6 8 0.94 0.84 0.75 0.85 0.67 0.8
7 8.5 0.89 0.72 0.44 0.65
** American Practice.
We furnish in Table 3.1.6, the data furnished by other researchers on the MSF value
varying with earthquake magnitude. The factor of safety against Liquefaction can now
be expressed as
FS =
_
CRR
7.5
CSR
_
MSF (3.1.12)
where CRR
7.5
= Cyclic resistance ratio for an earthquake magnitude 7.5.
Whatever has been discussed previously will now be further clarified by a suitable
problem, which covers the whole gamut of the above conditions.
Example 3.1.1
As shown in Figure 3.1.3, is a site soil profile which consists of 3.0 m of silty
clay underlain by 6 m of sand whose average SPT value is 13. The ground water
table is observed to be at a level of 1.0 meter belowground level. The dry density
of the silty clay is 18 kN/m
3
, while that in saturated condition is 20 KN/m
3
. The
saturated density of sand is 19.6 kN/m
3
. Sieve analysis shows the sand to have
Fines content as 15%. Find the liquefaction potential when the site is considered
to be 150 kmaway fromthe epicentre having an earthquake moment magnitude
of 6.5? The SPT test was carried out by standard sampler with safety hammer &
having rod length of 6.0 m. The diameter of the bore hole was 150 mm.
Density of soil =18/m3 1.0m
GWL
Density of soil =20kN/m3 2.0 m
Average SPT Value=13
Saturated Density of soil=19.6 kN/m3
6.0 m
Figure 3.1.3 Soil Profile of a site with typical soil properties.
Solution:
Considering a
max
= 0.18410
0.320M
(D)
0.8
g.
Here M = 6.5 and D = 150 km which gives, a
max
=
0.18410
0.3206.5
(150)
0.8
g = 0.4017 g.
Effective vertical stress at center of the sand layer is,
v
= 181.0+102+
9.63 = 66.8 kN/m
2
.
The gross vertical pressure at center of the sand layer,
v
= 181.0+202+
19.63 = 116.8 kN/m
2
.
The depth below ground, where liquefaction potential is calculated is 1 + 2 +
3 = 6 m.
Thus z = 6.0 m < 9.15 m, which gives, r
d
= 1.0 0.000765z r
d
=
1.0 0.0007656 = 0.9954.
Considering, CSR = 0.65
_
a
max
g
__

v
v
_
r
d
, we have, CSR = 0.65
_
0.4017g
g
__
116.8
66.8
_
0.9954 = 0.4544
The corrected SPT value is given by
(N
1
)
60
= N
o
(C
N
)(C
E
)(C
B
)(C
R
)(C
S
)
Here, N
o
= 13, C
N
=
_
100
66.8
, C
E
= 1.0, C
B
= 0.85, C
R
= 1.05, C
S
= 1.0
Thus (N
1
)
60
= 13
_
100
66.8
1.00.851.051.0 = 14.2
For FC = 15% we have
= e
_
1.76
_
190
FC
2
__
= e
_
1.76
_
190
15
2
__
= 2.498; and
=
_
0.99 +
_
FC
1.5
1000
_
_
=
_
0.99 +
_
15
1.5
1000
_
_
= 1.048.
Thus corrected SPT value is given by (N
1
)
60CS
= + (N
1
)
60
i.e. (N
1
)
60CS
= 2.498 +1.04814.2 = 17.3817 (say)
Considering CRR
7.5
=
1
34 (N
1
)
60
+
(N
1
)
60
135
+
50
[10(N
1
)
60
+45]
2

1
200
We have, CRR
7.5
=
1
34 17
+
17
135
+
50
[1017 +45]
2

1
200
= 0.1808
The Magnitude scaling factor is given by
MSF =
10
2.24
M
2.56
=
10
2.24
6.5
2.56
= 1.44;
Thus, FS =
_
CRR
7.5
CSR
_
MSF =
_
0.1808
0.4544
_
1.44 = 0.572 < 1.0.
Hence, as the factor of safety being less than 1.0, the soil has a high chance of
liquefaction during the earthquake.
3.1.6.5 Correlation between CRR and CPT value
Other than SPT, cone penetration test (CPT) is also used in field for evaluation of geo-
technical engineering parameters. As such investigators have also tried to co-relate the
CPT value with CRR for evaluation of liquefaction potential. One of the advantages
with CPT being that since it is a continuous process; thin layers of soil that one can
miss by SPT will not be missed in this case.
As stated earlier, Equation (3.1.3) is used to determine the CSR value. The CRR
value is indirectly co-related to CPT by developing relationship between CPT and SPT
value.
As per Seed and Idriss
q
c
= 4 to 5 N for clean sand, and
q
c
= 3.5 to 4.5 N for silty sand (3.1.13)
where, q
c
= the observed CPT value in MPa.
Once an equivalent SPT value is obtained fromthe observed q
c
, rest of the procedure
remains same as stated earlier. Murthy et al. (1991) has given the relationship which
can also be used to obtain equivalent SPT values from the observed cone penetration
values. This is given in Table 3.1.7.
Schmertmann (1978) presented a relationship between SPT and CPT values for
various types of soil [Table 3.1.8] which are also extensively used in the design offices
to determine equivalent SPT values from the observed CPT values.
Table 3.1.7 Relationship between relative density of fine sand, SPT, cone resistance and angle of
friction.
State of sand D
r
N q
c
(MPa)
Very loose <0.2 <4 <2.0 <30
Loose 0.20.4 410 24 3035
Medium dense 0.40.6 1030 412 3540
Dense 0.60.8 3050 1220 4045
Very dense 0.81.0 >50 >20 >45
Table 3.1.8 Relationship between SPT, CPT values for different
types of soil after Schmertmann (1978).
Type of soil q
c
/N
Sand and gravel mixture 6
Sand 4
Sandy silt 3
Clay-silt sand mixture 2
Insensitive clay 1.5
3.1.6.6 Liquefaction of clay
Normally clay is deemed non liquefiable. However based on experience of earthquake
in China it is now established that there are certain types of clay, which under shaking
do undergo liquefaction.
As a rule of thumb, a clay sample will be deemed liquefiable provided all of the
following criteria as mentioned below are complied with,
Weight of soil particles finer than 0.005 mm is less than 15% of the dry weight
of the soil.
The liquid limit (LL) of the soil is less than 35%.
The moisture content of the soil is less than 0.9 times the liquid limit of soil.
Clayey soil meeting not all of the above criteria are usually considered non
liqueable.
3.1.6.7 Settlement of foundation due to liquefaction failure
We had stated in our earlier section of liquefaction that during an earthquake, due to
shock there is a sudden increase in pore pressure that cannot dissipate immediately
resulting in lose of shear strength of soil. However, in course of time, this pore pressure
dissipates away towards the surface resulting in volumetric deformation of the ground.
Considering the above phenomenon and heterogeneous nature of soil, the soil may
undergo differential settlement which could be critical for building foundations and
underground lifelines.
A technique to estimate the ground settlement has been proposed by Ishihara and
Yoshimine (1992) wherein they developed a chart based on which the post liquefaction
volumetric strain is co-related to the FS value (CRR/CSR) and the SPT value as shown
in Fig. 3.1.4.
Figure 3.1.4 Curves for volumetric strain versus FS after Yoshimine (1992).
Based on above, once we know FS and SPT value, the volumetric strain is read off
from the curve and the settlement is obtained by multiplying this strain with the depth
of the soil.
The above is now further elaborated by a problem shown below.
Example 3.1.2
For the soil sample as described in Example 3.1.1 estimate the settlement of the
sandy layer considering all other boundary conditions remaining identical.
Solution:
From previous example we have seen FS =
_
0.1808
0.4544
_
1.44 = 0.572<1.0,
which shows that the soil can undergo liquefaction.
We has also seen that the corrected SPT value of the soil is N = 17.38 say 17.
Referring to Ishihara & Yoshimines chart we find volumetric strain = 2.0%.
Thus settlement of the sand layer of 6 m is =
2.0
100
6000 = 120 mm.
3.1.6.8 Reduction of bearing capacity of soil
Normally it is believed that earthquake has marginal effect on bearing capacity of soil.
As a matter of fact it is often a common practice and advised in many codes to increase
the allowable bearing capacity by 25%.
The reason for this can be attributed to the fact that normally when we find bearing
capacity of soil, we find out ultimate bearing capacity of soil. Dividing it by a factor
of safety we arrive at the allowable bearing capacity of soil. This is mostly as per the
general shear failure theory of soil, where Terzaghi, Meyerhof or Brinch Hansens
formula is used.
However in many cases and especially for cohesive soil, it is the settlement that
governs the design bearing capacity of soil.
Thus during an earthquake which is considered once in a lifetime phenomenon on
the structure, a lowering of the factor of safety on the bearing capacity is usually
deemed acceptable, and hence allowable bearing capacity is increased.
However it should be made clear that such increment is valid for a case when the
foundation is resting on
Crystalline rocks having no horizontal fragments or laminations;
Dense compacted sand having SPT value >30;
Stiff to very stiff clay with nominal plastic flow.
If the soil is otherwise made of fragmented rock, loose sand or soft plastic clay
sensitive to vibration this increased bearing capacity value should not be used
7
.
In such cases there could be significant reduction in strength when the foundation can
undergo either a local shear failure (when the foundation punches through overlying
soil due to liquefaction of bottom layer) or undergo a general shear failure when there
is a significant change in soil property for which bearing capacity factors N
c
, N
q
, and
N
undergo reduction resulting in a reduced bearing capacity.

3.1.6.9 Punching shear failure of soil
To understand how local shear failure can occur, let us consider the soil profile as
shown in Figure 3.1.5.
Let us consider the case of a foundation resting on the top layer of shallowclayey soil
which is non liquefiable, underlain by a layer of loose sand susceptible to liquefaction.
It is apparent from the figure that depth of the layer below the footing to the top of
liqueable sand layer is quite less and it might so happen that if bottom layer looses its
strength and the foundation is subjected to heavy load from superstructure, the foun-
dation may punch through this thin layer of soil and collapse, causing serious damage
to the super-structure. Similar to a column punching through a RCC footing here the
whole foundation punches through the soil along the vertical dotted line to collapse.
7 Unfortunately many design engineers hardly give consideration to this and believes that this increase of
bearing capacity of foundation almost a sacrosanct issue.
P
Figure 3.1.5 Soil Profile of a site with foundation resting on top layer on non-liquefiable soil.
To prevent this happening we calculate a factor of safety (FS) expressed as
FS =
2(B +L)Z
f
P
for isolated footing, and,
FS =
2Z
f
P
for strip footing. (3.1.14)
where, B = width of foundation in meter; L = length of foundation in meter; Z =
depth of soil layer from bottom of footing to the top of liquefiable soil, and
f
= shear
strength of un-liquefiable layer of soil in kN/m
2
.
If the top layer of non liquefiable soil is cohesive in nature (clay) then the shear
strength is given by
f
= S
u
, where S
u
= undrained shear strength of the soil. (3.1.15a)
For c soil (undrained shear strength parameters) the shear strength is given by
f
= c +
h
tan (3.1.15b)
where
h
= horizontal total stress in kN/m
2
; for cohesive soil this is often assumed
as 0.5
v
.
For a non-liquefiable soil layer of cohesionless soil, the shear strength is given by
f
=
h
tan
= k
0
v
tan (3.1.15c)
where
h
= effective horizontal stress in kN/m
2
and is equal to the coefficient of
passive pressure at rest times the vertical effective stress
v
, and,
= effective angle
of friction of the cohesionless soil.
We now show the application of the above based on a suitable problem as shown
in Example 3.1.3.
Example 3.1.3
Shown in Figure 3.1.5, is a footing of size 3 m2 m placed on a stiff clayeysilt
layer of undrained shear strength S
u
= 50 kN/m
2
and = 10
. The footing has

maximumload of 650 kNon it (including its own weight). The clay layer (3.0 m
deep) is underlain by a layer of loose sand 9.0 meter deep which is susceptible
to liquefaction. Find the factor of safety of the foundation under punching shear
failure. The foundation is resting at depth of 1.5 m below ground level. Unit
weight of soil of the top layer is 20 kN/m
3
.
Solution:
As per the problem Z = 3.0 1.5 = 1.5 m;
v
= 201.5 = 30 kN/m
2
.
Thus,
h
= 0.530 = 15 kN/m
2
and
f
= 50 +15 tan 10 = 52.64 kN/m
2
.
The resistive force = 2(B +L) Z
f
= 2(3 +2) 1.552.64 = 789.6 kN.
And, FS = 789.6/650 = 1.214. Considering the uncertainty in soil, FS = 1.2
could be a low value.
3.1.6.10 General shear failure capacity reduction
due to liquefaction
This phenomenon is generally observed in case of the soil supporting the foundation is
a stiff clay layer underlain by sandy layer susceptible to liquefaction. The ultimate bear-
ing capacity of foundation based on general shear failure theory is given by Terzaghis
equation as
q
ult
= cN
c
+qN
q
+
1
2
s
BN
(3.1.16)
The first term cN
c
gives the strength of the soil due to its cohesive property. The
second term depicts the effect of overburden soil which goes on to increase the bearing
capacity of the soil and the last term
1
2
s
BN
gives the frictional strength of the soil

where the term N
is a function of the friction angle .

For clayey soil, as = 0, it gives N
= 0 and N
q
= 1; For spread footing,
considering the aspect ratio (B/L) correction, we have
q
ult
= cN
c
_
1 +0.3
B
L
_
+ D
f
, further modified to
q
ult
= S
u
N
c
_
1 +0.3
B
L
_
+ D
f
. (3.1.17)
For shallow foundation near the ground as the second term has minimal effect, for
all practical purpose we can consider the equation to be
q
ult
= S
u
N
c
_
1 +0.3
B
L
_
(3.1.18)
For the bottom layer of liquefiable soil there is obviously a reduction in value of N
c
and this is usually function of the ratio of Z/B as given in Table 3.1.9.
Table 3.1.9 Reduction in value of N
c
for Z/B ratio.
Z/B N
c
0 0
0.25 0.7
0.5 1.3
1.0 2.5
1.5 3.8
5.5
where B = width of the foundation; Z = height of soil from
bottom of foundation to the top of liquefiable soil.
Example 3.1.4
For the example problem cited in Example 3.1.3, find the reduced bearing capa-
city of the foundation considering the top layer of soil as stiff clay of undrained
shear strength of 50 kN/m
2
. All other parameters remain the same as the earlier
problem.
Solution:
Under unliquefied state the ultimate bearing capacity is given by
q
ult
= S
u
N
c
_
1 +0.3
B
L
_
+ D
f
For = 0 N
c
= 5.5, q
ult
= 505.5
_
1 +0.3
2
3
_
+201.5 = 357.3 kN/m
2
.
Considering foundation size as 2 m3 m we have, Q
ult
= 357.323 =
2143.5 kN
FS =
2143.5
650
= 3.3
When the bottom soil is liqueed considering, Z/B = 1.5/2.0 = 0.75,
referring to Table 3.1.8, reduced N
c
value = 1.9.
Thus, q
ult
= 501.9
_
1 +0.3
2
3
_
+ 201.5 = 144 kN/m
2
Q
ult
=
14423 = 864 kN.
Thus, FS =
864
650
= 1.3, which is low and should preferably be about 1.5.
3.1.6.11 Ground subsidence due to earthquake
During an earthquake of major magnitude there are many cases of ground subsi-
dence and land slides which have wrecked havoc on many structures and especially
underground services which may get severely damaged due to this.
In the San Francisco Bay Earthquake (1906), the major source of damage was the
fire which broke out as an aftermath of the earthquake and could not be contained
as most of the underground water pipe lines were severely damaged due to ground
subsidence and became non-functional.
The major reason for this subsidence is again deep seated liquefaction for which the
soil starts to flowand due to differential or non uniformflowcan split apart a structure
built on it. Roads and pavements were observed to undergo extensive damage due
to subsidence and similar was a major observation in ChiChi earthquake in Taiwan
(1999). When the slope of the ground is less or equal to 6%the flowof soil is generally
defined as a lateral displacement of soil. When this slope is more than 6% the same is
know as a land slide.
A number of researches have been carried out to develop a mathematical model,
which would effectively predict the subsidence of the ground during a major earth-
quake. However, parametric functions being so many in numbers and uncertain that
there is yet a model which can be stated as unconditionally applicable.
The most used mathematical model for practical engineering purpose is one
empirical model by Bartlet &Youd (1992) developed based on historical data collected
from the six earthquakes in USA and two in Japan. They proposed two expressions
one for sites near steep banks with a free face, the other with sites having gently sloped
terrain.
For free faced condition
log D
H
= 16.3658 +1.1782M 0.9275log R 0.0133R +0.6572log W
+0.3483log T
15
+4.5270log(100 F
15
) 0.9224D50
15
(3.1.19)
For sloped terrain condition
log D
H
= 15.7870 +1.1782M 0.9275log R 0.0133R +0.4293log S
+0.3483log T
15
+4.5270log(100 F
15
) 0.9224D50
15
(3.1.20)
in which, D
H
= estimated average ground displacement in meters; D50
15
=
average mean grain size of the liqueable layers included in T
15
in mm; M = moment
magnitude of the earthquake; R = epicentral distance in kM; F
15
= average fine con-
tent (passing ASTM 200 sieve) for the liqueable layer in% included in T
15
; T
15
= the
cumulative thickness (in meter) of the saturated granular layer having blow count
<15; S = ground slope in percent, and, W = ratio of height (H) of the free face to the
distance (L) from the base of the free pace to point in question percent.
Example 3.1.5
Shown in Figure 3.1.6 is a site soil prole which consists of 3.0 m of clay under-
lain by 6 m of sand whose average SPT value is 13 which is susceptible to
earthquakes. The site consists of a canal owing across as shown in the figure
shown below.
The unit weight of the clay is 20 kN/m
3
. The saturated unit weight of sand
is 19.6 kN/m
3
. Sieve analysis shows the sand to have fines content as 15%.
The average grain size diameter of the sand layer is 0.032. A power house is
to be in built on this site located at distance of 30 meter from the canal bank.
The site is considered to be 50 kmaway fromthe epicentre having an earthquake
Moment magnitude of 6.75. Find the estimated movement of soil with this free
face condition.
30m
Unit weight of soil =20kN/m
3
3.0m
Average SPT Value=13
Saturated unit weight of soil=19.6 kN/m
3
6.0m
Figure 3.1.6 Soil Profile of a site with typical soil properties.
Solution:
Here, R = 50 km; M = 6.75; W = H/L = 3/30 = 0.1 = 10%; T = 6 m, and,
D
50
= 0.32.
Considering,
log D
H
= 16.3658 +1.1782M 0.9275log R 0.0133R +0.6572log W
+0.3483log T
15
+4.5270log(100 F
15
) 0.9224D50
15
,
we have
log D
H
= 16.3658 +1.17826.75 0.9275log 50 0.013350 +0.6572
log 10 +0.3483log 6 +4.5270log(100 10) 0.92240.032
log D
H
= 16.3658 +7.95285 1.576 0.665 +0.6572 +0.271 +8.846
0.295 = 1.1751
which gives D
H
= 0.068 m.
Considering uncertainties this value can vary from half to double thus
estimated value is 0.034 m to 0.136 m.
3.1.6.12 Effect of earthquake on structures
From above discussion it is obvious that earthquake has a profound influence on soil,
and since a structure is built on this soil it do also affects its response.
Potential energy stored in earth faults are released due to its rupture and generates
kinetic energy in form of stress waves in soil which propagates as P- and S-waves on
the surface of earth and induces acceleration on structures and foundations built on
the surface of the earth.
Thus, as per Newtons law of motion the structure is subjected to force as a result of
its inertial mass, which it has to resist depending on its stiffness and ensure that stresses
and deformations induced in the structure and foundation are within safe limits.
Above in essence is the basic philosophy of earthquake resistant design. The ana-
lytical methods adapted for earthquake analysis for different class of structures and
foundations may be classied into following category:
Seismic coefficient method or equivalent static method
Response spectrum method or psuedo static analysis
Dynamic analysis which is further subdivided into:
Modal analysis
Time history analysis.
We, as a first step, would study in general the basic principles underlying the above
methods and finally see their application to different class of structures and foundations
like buildings, tall chimneys, elevated water tank, retaining walls, earth dams etc.
3.2 EARTHQUAKE ANALYSIS
3.2.1 Seismic coeff icient method
This is an approach where the earthquake force is treated as an equivalent static force
based on the zonal classification of a country
8
.
Though earthquake force in essence is dynamic in nature based on the potential
occurrence of earthquakes in a particular zone, the soil condition, the type of foun-
dation, code recommends a certain percentage of weight of the structure which it is
expected to resist as lateral force.
It should be noted that this method is nowobsolete in terms of latest code IS-1893
2002 and may only be used with caution just to get an idea about the extent of force it
may generate in a particular zone for a particular type of structure, and that too only
for cases where large number of human life is not endangered either due to direct or
indirect effect of earthquake.
Based on the seismic zoning, soil foundation system, importance factor etc we derive
a factor,
h
, which is given by
h
= I
0
(3.2.1)
where, = a coefcient depending on the soil foundation system as given in
Table 3.2.1, I = importance factor as furnished in Table 3.2.2,
0
= basic horizontal
seismic coefcient as given in Table 3.2.3.
Based on above having derived, the value of
h
, the base shear acting at the soil
foundation level, is given by
V = KC
h
W, for multistoried frames or buildings and
V =
h
W, for all other type of structures (3.2.2)
where, V =base shear on the structure due to a given earthquake; K =a factor known
as the performance factor of the frame; C = a coefficient defining flexibility of the
structure with increase in number of storey, depending on fundamental time period.
The value of flexibility factor C versus time is as given in Figure 3.2.1.
The value of performance factor K for different type of framing is as given in
Table 3.2.4.
For calculation of time period (T), code has furnished some empirical formulas from
which T may be found out as follows:
For moment resisting frame without bracings or shear walls resisting lateral loads
T = 0.1n (3.2.3)
here n = number of storey including basement.
8 It is presumed the reader has a copy of the earthquake code like IS-1893 (1984 and 2002) for cross
reference.
Table 3.2.1 Soil foundation factor for various soil foundation system as per IS-1893, 1984.
Pile passing Combined
Type of soil through any or Isolated RCC Isolated Fdn Well
constituting soil but resting Piles on any Raft foundation without tie founda-
the foundation on rock other soil foundations with tie beams beams tions
Rock or Not
hard soil 1.0 applicable 1.0 1.0 1.0 1.0
Medium soil 1.0 1.0 1.0 1.0 1.2 1.2
Soft soil 1.0 1.2 1.0 1.2 1.5 1.5
Table 3.2.2 Value of importance factor I as per IS-1893, 1984.
Type of structure Importance factor (I)
Dams (all types) 3.0
Containers of inflammable or poisonous gases or liquids 2.0
Important service and community structures such as hospitals, water
towers and tanks, schools important bridges, important power houses,
monumental structures, emergency buildings like telephone exchange
fire bridge, large assembly buildings like structures like cinemas,
assembly halls and subway stations 1.5
All others 1.0
Table 3.2.3 Basic seismic coefficient
0
as per IS-1893, 1984.
Zone classification Seismic coefficient (
0
)
V 0.08
IV 0.05
III 0.04
II 0.02
I 0.01
For all others
T =
0.09H
d
(3.2.4)
where, H = total height of the main structure in meters and, d = maximum base
dimension of building in meters in direction parallel to the applied seismic force.
The above formulations are valid only for buildings which are regular in shape and
have regular distribution of mass or stiffness both in horizontal and/or vertical plane.
The value of
0
@ 0.08 (Table 3.2.3) has been obtained for zone V based on obser-
vations of earthquake occurrence in that zone however the values for other has been
reduced proportionally, the basis of this reduction has never been very explicit.
Though the above method has now been made obsolete in the recent code (IS-1893-
2002) but it still remains in practice in design offices to estimate preliminarily the
magnitude of earthquake force before a more detailed analysis is carried out.
Figure 3.2.1 Value of flexibility factor C as per IS-1893, 1984.
Table 3.2.4 Value of performance factor I as per IS-1893, 1984.
Structural framing system Value of performance factor K
Moment resistance (MR) frame with appropriate ductility
details as given in IS-4326 1.0
Frame as above with RC shear walls or steel bracing
members designed for ductility 1.0
Frame with either steel bracing members or plain or
nominally reinforced concrete infill panels 1.3
MR Frame as above in combination with masonry infill 1.6
Reinforced concrete framed buildings (Not covered by
1 or 2 above) 1.6
Example 3.2.1
An RCC building having frame layout is as shown in Figure 3.2.2. The trans-
verse cross section of the frame is also shown in the figure. Given the following
loading and geometric dimensions of the various structural members calculate
the base shear on the building as per seismic coefcient method IS-1893 (1984)
considering zone IV. Consider soil foundation system as of medium stiffness.
Loadings
Live load on roof = 2 kN/m
2
Live load on other floors = 4 kN/m
2
Parapet wall on roof = 1.5 m all round
Internal Partition walls = 1 kN/m
2
Floor finish = 1.5 kN/m
2
Cement plaster on ceiling = 50 mm.
4
4
6.0 6.0 6.0 6.0
Plan view of the frame
EL 116.4
EL 112.8
EL 109.2
EL 105.6
EL102.0
Tie beam all round
EL 100.0
4.0 4.0
Transverse elevation of the frame
Figure 3.2.2 A four storied RCC frame.
Geometric properties (Dimensions in mm)
Column size = 300600
Beam size in transverse direction = 300450
Beam size in longitudinal direction = 300600
Average thickness of water proofing on roof = 75 mm
All external walls 250 mm thick.
Material properties
Unit weight of concrete = 25 kN/m
3
Unit weight of brick = 20 kN/m
3
Unit weight of cement plaster = 24 kN/m
3
Grade of concrete = M25.
Seismic zone properties
Seismic zone = Zone IV
Soil type Medium stiff
Foundation type = Isolated footings with tie beams at 1.0 m below Ground
level.
Consider no live load on roof and 50% reduction in live load for other floors
during earthquake.
Solution:
Calculation of roof load (El 116.4)
Assume slab thickness = 125 mm;
Wt of slab =0.12524825 = 600 kN; Live Load on roof =2.0248 =
384 kN;
Parapet wall (1.5 m high) = 1.50.252 (24 +8) 20 = 480 kN;
Water proofing on roof = 0.07524824 = 345.6 kN;
Cement plaster on ceiling = 0.0524824 = 230.4 kN;
Wt of long beam = 0.3(0.6 0.125) 24325 = 256.5 kN;
Wt of short beam = 0.3(0.450 0.125) 8525 = 97.5 kN;
Wt of columns = 0.30.61.81525 = 121.5 kN, and,
Total load on roof = 600 + 384 + 480 + 346 + 230 + 257 + 98 + 122 =
2517 kN.
Calculation of load on other floors (El 112.8 109.2 and 105.6)
Wt of slab = 0.12524825 = 600 kN; Live Load on floor =
4.0248 = 768 kN
Wt of partition wall = 1.0248 = 192 kN;
Load from external brick wall = (3.6 0.475) 0.254820 +
(3.6 0.325) 0.251620 = 1012 kN;
Cement plaster on ceiling = 0.0524824 = 230.4 kN; Flooring on slab =
1.5248 = 288 kN;
Wt of long beam = 0.3(0.6 0.125) 24325 = 256.5 kN;
Wt of columns = 0.30.63.61525 = 243 kN, and,
Total load on each floor = 600 + 768 + 192 + 1012 + 230 + 288 + 257 +
98 + 243 = 3688 kN.
Calculation of load on ground floor (El 102.0)
Load from external brick wall = (3.6 0.475) 0.254820 +
(3.6 0.325) 0.251620 = 1012 kN;
Wt of long beam = 0.3(0.6 0.125) 24325 = 256.5 kN;
Wt of columns = 0.30.62.81525 = 189 kN, and,
Total load on ground floor = 1016 +258 +98 +189 = 1561 kN.
Total load to be considered for earthquake
Load at roof level = 2517 384 = 2133 kN (Considering no live load on roof
during earthquake);
Load at EL 112.8 = 3688768+0.5768 = 3304 kN (Considering 50% live
load on each floor during earthquake);
Load at El 109.2 = 3304 kN (Same as other floor);
Load at El 104.6 = 1561 kN, and,
Total Weight = 2133 +33304 +1561 = 13606 kN.
Calculation of seismic coefficient
As stated in the theory above, a
h
= I
0
For Seismic zone IV,
0
= 0.05
For medium stiff soil with isolated foundations connected by tie beam, = 1.0
For normal residential building importance factor I = 1.0
Thus, a
h
= 1.01.00.05 = 0.05
Considering T = 0.1n where n = number of storey, we have, T = 0.5 secs,
based on which as IS-1893 1984 flexibility factor C = 0.75.
Considering Moment resistant frame with ductile detailing, K = 1.0
And V
b
= KC
h
W = 1.00.750.0513606 = 510.225

= 510 kN
Thus, the total base shear acting on building for an earthquake force acting in
either transverse or longitudinal direction is = 510 KN
9
.
3.2.2 Response spectrum method
This method has undergone almost a radical change compared to what is furnished in
IS-1893 2002 and that what was furnished in IS-1893 1984.
In the previous code (1984 version) it was observed that base shear developed based
on seismic coefficient method and that by response spectrum method were almost
matching or were very close for 5% damping in the system. However with the present
version (2002) this force is almost double to the previous version. This we believe
would significantly enhance the project cost of all projects to come in near future.
9 This is strictly not correct for we will see later that time period will vary in both direction based on its
stiffness and mass thus earthquake force will also vary accordingly. Moreover the force calculated herein
is the total force acting on the building considered as stick model.
Figure 3.2.3 Response Spectrum Curve S
a
/g as per 1893 (1984).
3.2.2.1 Response spectrum method as per 1984 version
Though the 1984 version has been made obsolete however for historical reason and
also for comparison with the present code we present below the steps followed in this
method.
The 1984 version gave a set of curves representing the values S
a
/g versus different
time period in seconds for different level of damping. The sets of curves are as shown
in Figure 3.2.3.
The curve as shown in Figure 3.2.3 is actually based on the curves generated by Hous-
ner based on his observations and average spectrum obtained using four earthquake
time histories.
Based on the response spectra curve as furnished in Figure 3.2.3 for a particular time
period of a structure, the corresponding Sa/g is obtained for a particular damping ratio.
Based on the zonal demarcation like I, II, III, IV etc. code gives a values of response
spectrum factor F
0
10
based on which the coefcient of horizontal seismic force is
given by
h
= IF
0
S
a
g
(3.2.5)
Here and I are as already defined factors in the seismic coefficient method and
factor F
0
is as defined in Table 3.2.5.
10 This is exactly 5 times the value of
0
as given for seismic coefficient method.
Table 3.2.5 Value of Seismic zone factor F
0
as per IS-1893, 1984.
Zone classification Seismic zone factor (F
0
)
V 0.40
IV 0.25
III 0.20
II 0.10
I 0.05
Once the value of
h
is known the rest of the procedure remains same as that for
seismic coefficient method.
It may be noted that here that the time period may either be obtained based on
formulations as given in code or may be found out based on a detailed dynamic analysis
and forces are then obtained based on modal response technique
11
.
We now explain the above procedure based on a suitable numerical problem.
Example 3.2.2
For the building cited in Example 3.2.1, find the base shear as per response
spectrum technique based on IS-1893, 1984. Consider the site to be zone 4 with
medium stiff soil. Consider 5% damping ratio for the structure.
Solution:
Referring to Example 3.2.1 the time period of the building is given by
T = 0.1n = 0.5 sec.
For 0.51 sec and 5% damping the S
a
/g obtained from the curve as shown in
Fig. 8.2.4 is
S
a
/g = 0.16
As stated previously in Example 3.2.1, = 1.0 and I = 1.0 and F
0
= 0.25 as
per Table 3.1.6.
Thus
h
= IF
0
S
a
g
Or
h
= 1.0 1.0 0.25 0.16 = 0.04
As shown in Example 3.2.1, total weight of the structure W = 10310 kN
For T = 0.5 sec C = 0.75 K = 1.0.
11 This we are going to study in detail subsequently.
Thus considering
V = KC
h
W,
we have, V = 1.0 0.75 0.04 13606 = 408.18 kN = 408 kN.
3.2.2.2 Response spectrum method as per IS-1893 2002
As stated at the outset, the method has undergone a drastic modification with respect
to the present code. In lieu of the soil foundation factor () considered in the earlier
code, the latest version now defines the S
a
/g curve for different type of soil starting
with rock to soft soil. S
a
/g curve for various type of soil as per IS-1893 (2003) is
shown in Figure 3.2.4 for 5% damping.
Spectral Response as per IS -1893 2002
0
0.5
1
1.5
2
2.5
3
0
0
.
3
4
0
.
6
8
1
.
0
2
1
.
3
6
1
.
7
2
.
0
4
2
.
3
8
2
.
7
2
3
.
0
6
3
.
4
3
.
7
4
Time Period (secs)
S
p
e
c
t
r
a
l

A
c
c
e
l
e
r
a
t
i
o
n

C
o
e
f
f
i
c
i
e
n
t

(
S
a
/
g
)
Sa/g(Hard
soil/Rock)
Sa/g(Medium
soil)
Sa/g(Soft soil)
a
/g as per IS-1893 (2002).
Moreover as computer analysis has almost become a daily routine work in day to day
design office practice-where it is preferable to have digitised data of S
a
/g for computer
input, the code now defines the S
a
/g curve by direct formulas enabling one to furnish
numerical input for earthquake analysis by computer. The formulas suggested by code
for various types of soil as per Clause 6.4.4 of the code for 5% damping ratio are as
shown in Table 3.2.6:
The code has given factors based on which the values of S
a
/g obtained above may
be modified for different damping ratio.
Typical S
a
/g curve for soft soil with different damping ratio are shown in
Figure 3.2.5 while multiplication factors to be considered for different damping ratios
are furnished in Table 3.2.7.
Table 3.2.6 Expressions for S
a
/g for different types of soil as per IS-1893 2002.
Type of soil Value of S
a
/g Range
Rock or hard soil 1 + 15T 0.0 < T < 0.1
2.5 0.1 < T < 0.4
1.00/T 0.4 < T < 4.0
Medium soil 1 + 15T 0.0 < T < 0.1
2.5 0.1 < T < 0.55
1.36/T 0.55 < T < 4.0
Soft soil 1 + 15T 0.0 < T < 0.1
2.5 0.1 < T < 0.67
1.67/T 0.67 < T < 4.0
Spectral Acceleration Soft soil
0
0. 5
1
1. 5
2
2. 5
3
0
0
.
2
8
0
.
5
6
0
.
8
4
1
.
1
2
1
.
4
1
.
6
8
1
.
9
6
2
.
2
4
2
.
5
2
2
.
8
3
.
0
8
3
.
3
6
3
.
6
4
3
.
9
2
Time period(secs)
S
p
e
c
t
r
a
l

a
c
c
e
l
e
r
a
t
i
o
n
c
o
e
f
f
i
c
i
e
n
t
s

(
S
a
/
g
)
Sa/g(5%)
Sa/g(7%)
Sa/g(10%)
Sa/g(15%)
Sa/g(20%)
Sa/g(25%)
Sa/g(30%)
a
/g for soft soil as per IS-1893 (2002).
Table 3.2.7 Multiplying factors for obtaining values for other damping as per IS-1893 (2002).
Damping ratio (%) 0 2 5 7 10 15 20 25 30
Factors 3.2 1.4 1.0 0.9 0.8 0.7 0.6 0.55 0.5
Table 3.2.8 Seismic zone factor as per IS-1893 (2002).
Seismic zone II III IV V
Seismic intensity Low Moderate Severe Very severe
Z 0.1 0.16 0.24 0.36
The country unlike previously that was classified into 5 zones (zone I to V) in the
present code zone I has now been merged with zone II and the zones now constitute
of zone II to V only. The zone factors to be considered as per the present code are as
presented in Table 3.2.8.
Table 3.2.9 Response reduction factor R as per IS-1893 (2002).
Sl. No. Lateral load resistant system R
1 Ordinary moment resistant frame 3.0
2 Special Moment resisting frame specially detailed to
provide ductile behaviour 5.0
3 Steel Frame with
3a Concentric Bracing 4.0
3b Eccentric Bracing 5.0
4 Special moment resistant frame with ductile detailing 5.0
Buildings with shear walls
5 Load bearing Masonry wall buildings
5a Un-reinforced 1.5
5b Reinforced with horizontal RC band 2.5
Reinforced with horizontal RC band and vertical bars
5c at corners of rooms and jamb openings 3.0
6 Ordinary reinforced concrete shear walls 3.0
7 Ductile shear walls 4.0
Buildings with Dual systems
Ordinary shear wall with OMRF 3.0
Ordinary shear wall with SMRF 4.0
Ductile shear wall with OMRF 4.5
Ductile shear wall with SMRF 5.0
The importance factor, I has remained unchanged and as such the factors furnished
earlier in Table 3.2.2 still holds good.
To bring it in line with international practice followed by other countries
12
, the
code has now introduced a new factor R which is known as the response reduction
factor and also called the ductility factor in many literatures. This is the property of a
body to dissipate energy by means of its ductile behaviour and may be generated by
means of special detailing
13
.
The R factor for buildings is basically a function of the structural configuration of
the building like whether it is a Ordinary Moment Resistant frame (OMRF), special
moment resistant frame (SMRF) or has shear wall etc. The value of the response
reduction factor R for different types of structural system as defined in IS-1893 2002
is furnished in Table 3.2.9.
Based on the above data the design horizontal seismic coefficient A
h
for a structure
is determined by the expression
A
h
=
ZIS
a
2Rg
(3.2.6a)
12 Especially UBC 1997, and NEHRP as followed in USA.
13 We will discuss more about this later.
and the base shear is furnished by the expression
V = A
h
W (3.2.6b)
The empirical relation furnished by time period has also undergone some modica-
tions. As per the latest code the approximate fundamental time period in seconds for a
moment resistant frame without brick infill panels may be estimated by the empirical
expression
T
a
= 0.075h
0.75
for RCC frame building
T
a
= 0.085h
0.75
for steel buildings (3.2.7)
where, h = height of the building.
For all other buildings including moment resistant frame buildings with brick infill
panels is estimated from the formula as furnished in Equation (3.2.4).
Based on above we now solve a numerical problem to illustrate how base shear is
obtained as per latest IS-1893.
Example 3.2.3
For the building cited in Example 3.1.1, find the base shear as per response
spectrum technique based on IS-1893, 2002. Consider the site to be zone 4 with
medium stiff soil. Consider 5% damping ratio for the structure.
Solution:
Referring to Example 3.1.1, the time period of the building is given by
T =
0.09h
d
Here h = 16.4 meter and d = 8 m in transverse direction and d = 24 m in
long direction thus
T =
0.09 16.4
8
= 0.5218 sec in short direction and
T =
0.09 16.4
24
= 0.3012 sec in long direction.
Thus based on the response spectrum curve S
a
/g = 2.50 for both short and
long direction
As per IS-1893 2002 for Zone IV Z = 0.24
Considering SMRF with ductile detailing as per Table 3.2.9, R = 5.0.
Thus A
h
=
ZIS
a
2Rg
=
0.24 1 2.5
2 5
= 0.06
As shown in Example 10.1 total weight of the structure W = 13606 kN
Thus considering, V = A
h
W we have,
V = 0.06 13606 = 816.36 kN.
Thus based on the above three examples if we compare the base shear, for the
given building we have as follows:
Sl. No. Code Method Base shear (kN) Remarks
1 IS-1893-1984 Seismic 510
coefficient
method
2 IS-1893-1984 Response 408 Shear force value
spectrum Method very near to seismic
coefficient method
3 IS-1893-2002 do 816 Shear force almost
double the value
of base as calculated
by IS-1893-1984
Comparison of base shear as per IS-1893 (1984) and IS-1893 2002.
3.2.3 Dynamic analysis under earthquake loading
To understand the basic concept, we start with systemhaving single degree of freedom
and subsequently extend this to system having multi-degree of freedom.
As shown in Figure 3.2.6, a single bay portal is subjected to an earthquake force
for which the body moves through a distance u
g
at base and undergoes additional
deformation of u
t
at top.
We had shown earlier that under time dependent force the equation of motion is
given by
m u +c u +ku = 0 (3.2.8)
where m = mass of the system; c = damping of the system (usually represented by a
dash pot); k = stiffness of the system; u, u, u = acceleration, velocity and displacement
vectors, respectively.
As during the motion the body undergoes a rigid body motion in terms of u
g
, it does
not affect the stiffness and damping of the system, which are affected by u
t
only.
Y
u
t
X
u
g
Figure 3.2.6 Single bay portal subjected to earthquake force.
Thus the above equation may represented as
m( u
g
+u
t
) +c u
t
+ku
t
= 0 (3.2.9)
From which we arrive at the expression
m u
t
+c u
t
+ku
t
= m u
g
or m u
t
+c u
t
+ku
t
= F
e
(3.2.10)
where F
e
= the earthquake force induced on the system and is equal to the mass of
the body times ground acceleration due to earthquake.
3.2.4 How do we evaluate the earthquake force?
Before we proceed further to analyse the above equation of equilibrium, it is essen-
tial to understand the nature and characteristics of earthquake force and how do we
evaluate it.
The earthquake force in essence is a transient force and acts on a body for a small
instant of time. In terms of Newtonian mechanics this can also be termed as an
impulsive force acting on a body.
According to the basic law of physics an impulse force is expressed as,

F =
_
F(t)dt.
This expression means a force F which is function of time is acting upon a body for a
very small duration of time dt and is normally defined as an impulse.
As, F = m
dv
dt
we can write this as, Fdt = mdv.
Thus if an impulse force,

F, is acting on a body, it will result in a sudden change in
its velocity without significant change in its displacement.
For spring mass systemunder free vibration we had seen earlier that the displacement
is given by
x = Asin
n
t +Bcos
n
t, (3.2.11)
where Aand Bare integration constants and their magnitudes depend on the boundary
condition.
For boundary conditions at t = 0, velocity = v
0
and displacement x = x
0
, the above
expression can be written as
x =
v
0
n
sin
n
t +x
0
cos
n
t where
n
=
_
k
m
(3.2.12)
Thus for the spring mass initially at rest and acted upon by an impulse force is
given by
x =
F
m
n
sin
n
t (3.2.13)
When considering damping for the system the free vibration equation is written as
x = Ae
D
n
t
sin
_
_
1 D
2
n
t +
_
(3.2.14)
Considering the impulse load the above can modified to
x =
F
m
n
1 D
2
e
D
n
t
sin
_
1 D
2
n
t (3.2.15)
where D is the damping ratio of the system
While considering earthquake the above expression can be further reduced to the
expression
x =
x
1 D
2
e
D
n
t
sin
_
1 D
2
n
t (3.2.16)
Under earthquake the shock induced on the ground is generally represented by
response spectra or velocity spectra. Moreover, as we are interested in the peak value
(or maximum force in the system) the above integral can effectively used to obtain
the peak velocity from which maximum displacement and acceleration are obtained
subsequently a shown here after.
We had seen earlier that equation of motion for the portal structure under
earthquake is given by the expression
m u
t
+c u
t
+ku
t
= F
e
(3.2.17)
Dividing each term by m we have
u
t
+
c
m
u
t
+
k
m
u
t
=
F
e
m
or u
t
+2D
n
n
u
t
+
2
n
u
t
= u
g
(3.2.18)
Since the force is impulsive in nature acting for duration of time (say), the
displacement u
t
can be represented by
u
t
=
1
1 D
2
t
_
0
u
g
()e
D
n
(t)
sin
_
_
1 D
2
_
n
(t )d (3.2.19)
Differentiating the above we have
u
t
=
1
1 D
2
t
_
0
u
g
()e
D
n
(t)
_
D
n
sin
_
1 D
2
n
(t )d
+
n
_
1 D
2
cos
_
1 D
2
n
(t )
_
d (3.2.20)
Considering,
C
1
=
t
_
0
u
g
()
D
n
t
cos
_
1 D
2
n
d and
C
2
=
t
_
0
u
g
()e
D
n
t
sin
_
1 D
2
n
d, the velocity can be expressed as
u
t
=
e
D
n
t
_
1
2
__
C
1
DC
2
_
1 D
2
_
sin
_
1 D
2
n
t
+
_
C
1
_
1 D
2
+C
2
D
_
cos
_
1 D
2
n
t
_
u
t
=
e
D
n
t
1 D
2
_
C
2
1
+C
2
2
sin
_
_
_
1 D
2
_
n
t
_
(3.2.21)
The velocity spectrum or the peak velocity is given by the maximum value of the
above
i.e. S
v
=
u
g
e
D
n
t
1 D
2
_
C
2
1
+C
2
2
max
when the maximum displacement is given by
S
d
=
S
v
n
and S
d
=
S
a
2
n
(3.2.22)
where S
a
is the acceleration spectrum.
Thus the maximum force the system may experience is given by
F
max
= m
2
n
S
d
(3.2.23)
It is obvious that that for response spectrum analysis the value S
a
is function of the
time period or natural frequency of the system which is given by the expression
=
_
k
m
and T =
2
. (3.2.24)
Certain type of structures can very well be modelled as systems with single degree
of freedom and the base force can be found out as follows:
Example 3.2.4
Shown in Figure 3.2.7 is an air cooler of weight 450 kN is supported on a
structure as shown. Determine the force on the system calculating time period
based on dynamic analysis. Consider the soil is medium stiff and the site is
in zone III. Consider 5% damping for the structure. For beams and columns
section properties are as follows I
xx
= 1268.6 cm
4
, I
yy
= 568 cm
4
and A =
78 cm
2
, Area of the bracing members = 12 cm
2
, E
steel
= 2 10
8
kN/m
2
.
Unit weight of column material = 78.5 kN/m
3
What will be the force on the
frame based formulation as given in the code?
6500
6000 3000
Figure 3.2.7 Structure supporting an air cooler.
Solution:
For earthquake force in transverse direction
Stiffness of each column is given by K = 12EI/L
3
Here I = 1268.6 cm
4
= 1.2686 10
5
m
4
E = 2 10
8
kN/m
2
L = 6.5 m,
Thus, K =
12 2 10
8
1.2686 10
5
(4.5)
3
= 110.86 kN/m
Considering four columns
4
i=1
K
i
= 4 110.86 = 443.46 kN/m
Weight of the air cooler = 450 kN
Mass of the air cooler =
450
9.81
= 45.87 kN-sec
2
/m
Mass of each column =
7.8 10
3
78.5
9.81
= 0.0624 kN-sec
2
/m
Considering 1/3rd weight of column contributing to top mass of 4 columns
4
i=1
m
i
=
0.0624 4
3
= 0.0832 kN sec
2
/m
Weight of top beam = (6 +3) 2 78.5 7.8 10
3
= 11 kN
Mass of beam =
11
9.81
= 1.123 kN-sec
2
/m
Thus total mass = 45.87 +0.0832 +1.123 = 47.07 kN-sec
2
/m
Considering T = 2
_
m
K
we have, T = 2
_
47.07
443.46
= 2.04 sec.
for which as per IS-1893(2002), S
a
/g = 0.666
Considering A
h
=
ZIS
a
2Rg
,
where Z = 0.16 for zone III, I-1.0; S
a
/g = 0.666 and R = 3.0,
we have, A
h
=
0.16 1.0 0.666
2 3
= 0.0177
Thus V
h
= 0.0177 47.07 9.81 = 8.20 kN.
For earthquake in longitudinal direction (i.e. in the direction of the braced bay)
Stiffness of per column (considered hinged at base, Figure 3.2.8)
=
3EI
L
3
=
3 2.0 10
8
5.68 10
6
(6.5)
3
= 12.41 KN/m
= tan
1
6.5
3.0
= 65.22

6500
6000
Figure 3.2.8 Frame in longitudinal direction
Stiffness of each bracing
_
AE
L
cos
2
_
=
1.210
3
2.010
8
6.5
cos
2
65.22 =
6486kN/m.
Thus total stiffness of the frame in longitudinal direction = 412.4+6486
4 = 25993.6kN/m.
Considering T = 2
_
m
K
we have
T = 2
_
47.07
25993.6
= 0.267 for which as per IS-1893(2002), S
a
/g = 2.5.
Considering A
h
=
ZIS
a
2Rg
, here Z = 0.16 for zone III, I-1.0 and R = 4.0 (for
concentric bracing) we have, A
h
=
0.16 1.0 2.5
2 4
= 0.05.
Thus V
h
= 0.05 48.2 9.81 = 23.6kN in longitudinal direction.
As per code for steel frame (vide Equation 3.2.9), T
a
= 0.085 h
0.75
T
a
= 0.085 6.5
0.75
= 0.346 sec for which the value S
a
/g = 2.5.
Thus considering A
h
=
ZIS
a
2Rg
here Z = 0.16 for zone III, I-1.0 and R = 4.0
we have, A
h
=
0.16 1.0 2.5
2 4
= 0.05.
Thus, the maximumforce on the frame = 23.1 kN, this is same as we obtained
using the dynamic analysis.
3.2.5 Earthquake analysis of systems with
multi-degree of freedom
Before we delve into the detailed dynamic analysis of systems with multi-degree of
freedom under earthquake force (based on modal analysis or time history response),
we deal with a particular technique often used in practical engineering design where for
many buildings effect of fundamental time period is most pre-dominant. In such cases
higher mode participation vis-a-vis its effect being insignificant are ignored without
causing any significant errors.
3.2.5.1 Analysis based on assumed shape function
This is a technique in which a multi-degree freedom system is converted into an equiv-
alent system having mass and stiffness of that of a single degree of freedom based on
an assumed shape function to find out the time period of a system.
To start with let us consider a stick model of a systemhaving multi-degree of freedom
The kinetic energy of the system is given by
T(t) =
1
2
n
i=1
m
i
_
y(z, t)
t
_
2
(3.2.25)
We consider here,
y(z, t) = (z)(t) (3.2.26)
where, (z) = admissible shape function which satisfies the boundary condition of
the system; (t) = generalized co-ordinate.
M
n
M
3
M
2
M
1
K
n
K
3
K
2
K
1
Displaced Shape(1
st
Mode)
Figure 3.2.9 A stick model having multi-degree of freedom.
Thus,
T(t) =
1
2
n
i=1
m
i
_
_
n
j=1
j
(z)
j
(t)
n
k=1
k
(z)
k
(t)
_
_
T(t) =
1
2
n
j=1
n
k=1
j
(t)
k
(t)
_
m
i
n
i=1
j
(z)
k
(z)
_
(3.2.27)
from which we conclude that the generalized mass of the system is given by,
M
=
_
m
i
n
i=1
j
(z)
k
(z)
_
(3.2.28)
Thus for fundamental mode for j = k we have
M
=
_
n
i=1
m
i
2
i
(z)
_
(3.2.29)
Similarly potential energy is given by
V(t) =
1
2
n
i=1
k
i
[y(z, t)]
2
(3.2.30)
where, = difference in the displacement between two adjacent level.
Hence,
V (t) =
1
2
n
i=1
k
i
_
_
n
j=1
j
(z)
j
(t)
n
k=1
k
(z)
k
(t)
_
_
V(t) =
1
2
n
j=1
n
k=1
j
(t)
k
(t)
_
k
i
n
i=1
j
(z)
k
(z)
_
(3.2.31)
Thus for the fundamental mode, j = k, we have
K
=
_
n
i=1
k
i
2
i
(z)
_
(3.2.32)
Now knowing, T = 2
_
m
K
, we have for this generalized case
T
= 2
_
M
(3.2.33)
From the above mathematical derivation it is obvious that if we know what could
be the assumed shape function correctly it is possible to arrive at the fundamental time
period of the system.
Based on the aspect ratio (H/D), Naeem (1989) has proposed the following shape
functions which may be considered for buildings modeled as stick having multi-degrees
of freedom.
Sl. No. H/D Shape function
1 H/D < 1.5 sin
x
2H
2 1.5 < H/D < 3
x
H
3 H/D > 3.0 1 cos
x
2H
where, H =height of the building; D =width of building in direction of the earthquake
force considered.
We will now solve the previous building problem (vide Example 3.2.1) to see how
base shear results differ with what we have calculated earlier.
Example 3.2.5
Refer the problem as shown in Example 3.2.1 calculate the time period of the
building based on assumed shape function method and calculate the base shear
in both transverse and longitudinal direction and find out the base shear based
on IS-1893-2002. Consider all other boundary conditions remains same as was
defined in the previous problem (Figure 3.2.10).
EL 116.4
EL 112.8
EL 109.2
EL 105.6
EL102.0
Tie beamall round
EL 100.0
4.0 4.0
Figure 3.2.10 Transverse elevation of frame.
Solution:
Considering the frame as a stick model in transverse direction we have the model
El-116.4
K1
EL-112.8
K2
El-109.2
K3
El-105.6
K4
EL-102.00
K5
EL-100.0
Figure 3.2.11 Stick model of the frame.
Dimension of column = 300 600
Moment of inertia of the column =
1
12
300600
3
= 5400000000 mm
4
=
0.0054 m
4
Stiffness of column =
12EI
H
3
Here, E
conc
= 2.85 10
7
kN/m
2
K
i
=
12 2.85 10
7
0.0054
(3.6)
3
= 39583.33 KN/m
For fteen column per level total stiffness K
i
= 1539583.33 = 593750 kN/m
Thus, K
1
= K
2
= K
3
= K
4
= 593750 kN/m
And K
5
= 15
12 2.85 10
7
0.0054
(2)
3
= 3462750
Since H/D in transverse direction is =
16.4
8
= 2.05 < 3.0 thus shape function
considered is x/H
Level Weight Mass Stiffness
i

i
m
i
2
i
k
i
2
i
5 2133 217.4 1.00 217.4
593750 0.22 28737.5
4 3304 337.0 0.780 205.03
593750 0.219 28476.84
3 3304 337.0 0.561 106.06
593750 0.22 28737.5
2 3304 337.0 0.341 39.18
593750 0.22 28737.5
1 1561 159.12 0.121 2.33
3462750 0.121 50698.12
567.67 165387.46
Considering T
= 2
_
M
we have, T
= 2
_
567.67
165387.46
= 0.368 sec
Based on response spectrum curve, S
a
/g = 2.5
Considering all other parameters remaining constant, A
h
=
ZIS
a
2Rg
=
0.24 1.0 2.5
2 5
= 0.06.
Base shear = 0.06 13606 = 816 kN, which is same as we got earlier, based
on method as suggested by the code.
For longitudinal direction, we have
Dimension of column = 300 600
Moment of inertia of the column =
1
12
600300
3
= 1350000000 mm
4
=
0.00135 m
4
Stiffness of column =
12EI
H
3
Here, E
conc
= 2.85 10
7
kN/m
2
K
i
=
12 2.85 10
7
0.00135
(3.6)
3
= 9895.833 kN/m
For fteen column per level total stiffness K
i
= 15 9895.833 =
148437.5 kN/m
Thus K
1
= K
2
= K
3
= K
4
= 148437.5 kN/m
And K
5
= 15
12 2.85 10
7
0.00135
(2)
3
= 865687.5 kN/m
Since H/D in transverse direction is =
16.4
24
= 0.683 < 1.5 thus shape
function is sin
x
2h
Level Weight Mass Stiffness
i

i
m
i
2
i
k
i
2
i
5 2133 217.4 1.00 217.4
148438 0.059 516.71
4 3304 337 0.941 298.4
148438 0.179 4756.1
3 3304 337 0.771 200.32
148438 0.260 10034.4
2 3304 337 0.511 87.99
148438 0.321 15295.2
1 1561 159.12 0.190 5.744
865688 0.190 31251.33
809.85 61853.74
Considering T
= 2
_
M
we have, T
= 2
_
809.85
61853.74
= 0.7189 sec
Based on response spectrum curve, S
a
/g = 1.39
Considering all other parameters remaining constant
A
h
=
ZIS
a
2Rg
=
0.24 1.0 1.39
2 5
= 0.0334
Thus, the base shear = 0.0334 13606 = 454 kN.
We see that when actual stiffness and mass distribution of the system is con-
sidered for calculation of the time period the base shear is in signicant variation
to that as to what has been considered in the code.
3.2.5.2 Dynamic analysis of systems having multi-degree
of freedom under earthquake forces
In this section we discuss the time history and modal analysis technique as applied to
earthquake. We had already discussed in detail the basic concepts underlying the same
in Chapter 5 (Vol. 1) (basic concepts of structural dynamics) as applied to harmonic
forces.
The fundamental steps for earthquake analysis, essentially remains the same as that
of harmonic force except the fact that calculation of amplitude and interpretation of
forces in the system is different.
For a structural system having N degrees of freedom we have seen earlier in
Chapter 5 (Vol. 1) that the equation of motion is expressed as
[M]{

X} + [C]{

X} + [K]{X} = {P(t)} (3.2.34)
were [M] = mass matrix of the system of size N N
[C] = damping matrix of size N N
[K] = stiffness matrix of size N N
{

X}, {

X}, {X} = acceleration, velocity, displacement vector of the system
Considering the displacement vector as {X} = [
(x)
]{
(t)
} the eigen value of the
problem is given by
[K] [M]
2
][] = 0, (3.2.35)
from which we find out the time period of the system for m number of significant
modes.
The different techniques to find out the eigenvalues for the above equation have
already been discussed in Chapter 5 (Vol. 1).
The equation of motion can now be expressed as
[M][]{
} + [C][]{
} + [K][]{} = {P(t)} (3.2.36)

Pre-multiplying the above by []
T
we have
[]
T
[M][]{
} + []
T
[C][]{
} + []
T
[K][]{} = []
T
{P(t)} (3.2.37)
Based on orthogonal property we had seen earlier that the above de-couples into
N number of equations expressed by
{
n
} +2D{
n
} +
2
{
n
} =
[
n
]
T
{P(t)}
[
n
]
T
[M][]
(3.2.38)
For earthquake as the force induced in the system can be expressed as {P(t)} =
[M]{ u
g
}, the above general equation can be modified into
{
n
} +2D{
n
} +
2
{
n
} =
L
n
{ u
g
}
[
n
]
T
[M][]
(3.2.39)
where, L
n
= []
T
[M][I] here [I] is a unit column vector of dimension, N.
The solution of the above equation for nth mode at any time, t, is given by the
expression
14
n
(t) =
L
n
T
n
[M]
n
1
n
t
_
0
u
g
()e
D
n
n
(t)
sin
n
(t )d (3.2.40)
The displacement for each mass i at time t is then obtained by superimposition of
all modes calculated at this time t and is given by
x
i
=
N
n=1
in
n
(t) (3.2.41)
The earthquake force on the structure is then expressed in terms of the effective
acceleration
neff
(t) =
2
n
n
(t) (3.2.42)
Considering f = k, we have, the earthquake force at any floor i at time is t is
given by
f
in
(t) = k
i
n
x
in
or f
in
(t) = k
in
in
n
(t) (3.2.43)
Since based on the eigen value expression, we have
k = m
2
(3.2.44)
Substituting the value of k in terms of inertial force, we have
f
in
(t) = m
i
2
n
in
n
(t) (3.2.45)
Superimposing all modal contribution the earthquake force on the structure is
expressed as
f (t) = [M]
2
n
[](t) (3.2.46)
14 This, we explained in the case of systems having single degree of freedom.
where [M] = mass matrix of the system of order N N; [] = relative amplitude
distribution of order N N; [
2
] = diagonal matrix of order N N having eigen-
values in the diagonal term.
Based on the above theory the entire history of displacement and force response
can be defined for any multi-degree of freedom system having calculated the modal
response amplitudes.
When the above theory is applied to response spectrum, as discussed earlier with
single degree of freedom the maximum response for the each mode is considered.
If the maximumvalue of
n max
of the Duhamel integral is considered, the maximum
displacement in that mode is given by
x
n max
=
n
n max
=
n
L
n
T
n
[M]
n
S
vn
n
(3.2.47)
The maximum earthquake force in the structure is then given by
f
n max
= [M]
n
L
n
2
n
[M]
n
S
an
(3.2.48)
where S
an
, S
vn
are spectral velocity as furnished in the code.
The base shear which is the algebraic sum of all the force is given by
V
0
(t) =
N
i=1
f
n max
(t) =
N
i=1
L
2
n
n
M
n
S
an
(3.2.49)
The expression
L
2
n
M
n
is usually called the effective modal mass of the systemand when
divided by the total mass (represented in percentage), reflects the percentage of modal
mass responding to the earthquake force in each mode.
We now further illustrate the above theory by a suitable numerical problem.
Example 3.2.6
Shown in Figure 3.2.12 is a three storied RCC frame subjected to earthquake in
zone IV having medium soil condition. The damping ratio for RCC considered
is 5%. Determine
The natural frequencies of the structure.
The eigen-vectors.
The acceleration, velocity and displacement as per IS-1893 2002 based on
response spectrum method.
Effective Modal mass participation for each mode.
The nodal displacement per mode.
The nodal shear force per mode.
Base shear for the three modes.
G H X
3
E F X
2
C D X
1
A B
Figure 3.2.12 A three story RCC Frame.
Here
1 K
AC
= K
DB
= 15000 kN/m M
GH
= 20 kN sec
2
/m
2 K
CE
= K
DF
= 10000 kN/m M
EF
= 40 kN sec
2
/m
3 K
EG
= K
FH
= 5000 kN/m M
CD
= 40 kN sec
2
/m
Solution:
The free body diagram of the structure is as shown below in Figure 3.2.13:
k
2
(x
2
-x
1
) k
3
(x
3
-x
2
)
k
2
(x
2
-x
1
) k
3
(x
3
-x
2
)
m
3
x
3
m
2
x
2
m
1
x
1
k
1
x
1
Figure 3.2.13 Free body diagram at each floor.
Based on the free-body-diagram, for free vibration we have,
m
3
x
3
+k
3
(x
3
x
2
) = 0
m
2
x
2
+k
2
(x
2
x
1
) k
3
(x
3
x
2
) = 0
m
1
x
1
+k
1
x
1
k
2
(x
2
x
1
) = 0
The above on simplification while writing in the matrix form gives
_
_
m
1
0 0
0 m
2
0
0 0 m
3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
+
_
_
k1 +k2 k2 0
k2 k2 +k3 k3
0 k3 k3
_
_
_
_
_
x
1
x
2
x
3
_
_
_
= 0
The above on substituting the values gives the following matrices.
[K] =
_
_
50000 20000 0
20000 30000 10000
0 10000 10000
_
_
and [M] =
_
_
40
40
20
_
_
Natural frequencies and modal values of the system
Applying any one of the methods for determination of eigen-values and
eigenvectors as cited in Chapter 5 (Vol. 1) we have
1
= 156.93,
2
= 750,
3
= 1593
Thus,
1
=

156.93 = 12.527 rad/sec, which implies T
1
= 0.502 sec
2
=

750 = 27.386 rad/sec, which implies T
2
= 0.229 sec
3
=

1593 = 39.913 rad/sec, which implies T
3
= 0.157 sec
The corresponding eigenvectors are given as
For the first mode, {}
T
1
= 0.314 0.686 1
For second mode, {}
T
2
= 0.5 0.5 1
For third mode, {}
T
3
= 1 0.686 0.314
The Matrix M
n
and L
n
The identity matrix is given by
[I]
T
= 1 1 1
For first mode
M
n1
= {}
T
1
[M]{}
1
= 0.314 0.686 1
_
_
40
40
20
_
_
_
_
_
0.314
0.686
1
_
_
_
= 42.772
L
n1
= {}
T
1
[M]{I} = 0.314 0.686 1
_
_
40
40
20
_
_
_
_
_
1
1
1
_
_
_
= 60
For second mode
M
n2
= {}
T
2
[M]{}
2
= 0.5 0.5 1
_
_
40
40
20
_
_
_
_
_
0.5
0.5
1
_
_
_
= 40
L
n2
= {}
T
2
[M]{I} = 0.5 0.5 1
_
_
40
40
20
_
_
_
_
_
1
1
1
_
_
_
= 20
Similarly for third mode
M
n3
= {}
T
3
[M]{}
3
= 60.802; L
n3
= {}
T
3
[M]{I} = 18.832
Thus
L
2
n
M
n
for three modes are given
L
2
n
M
n
=
_
_
_
84.167
10
5.833
_
_
_
this, when divided by
the total mass of the system (i.e. 40 + 40 + 20 = 100 kN), and multiplied by
100 we have
=
_
_
_
84.167
10
5.833
_
_
_
%
which represents the percentage mass participating in each mode.
Calculation of Acceleration and velocity based on IS 1893 (2002)
For the first mode, considering T
1
= 0.502 sec
S
a
g
= 2.5 or S
a
= 2.5 9.81 = 24.525 m/sec
2
Considering zone IV severe earthquake condition Z = 0.24 and considering
ductility factor R = 3.0 (for ordinary moment resisting frame)
S
a
(design) =
0.24 24.525
2 3
= 0.981m/sec
2
;
S
v
(design) =
S
a
=
0.981
12.527
= 0.078m/sec.
For the second mode, considering T
2
= 0.229 sec
S
a
g
= 2.5 S
a
= 2.5 9.81 = 24.525m/sec
2
Considering Z = 0.24 as before ductility factor R = 3.0 (for ordinary moment
resisting frame)
S
a
(design) =
0.24 24.525
2 3
= 0.981 m/sec
2
;
S
v
(design) =
S
a
=
0.981
27.386
= 0.036 m/sec.
For third mode considering, T
3
= 0.157 sec
S
a
g
= 2.5 S
a
= 2.5 9.81 = 24.525 m/sec
2
.
Considering Z = 0.24 and R = 3.0, we have
S
a
(design) =
0.24 24.525
2 3
= 0.981 m/sec
2
;
S
v
(design) =
S
a
=
0.981
39.913
= 0.025 m/sec.
Thus based on above for the three modes, we have
{S
a
} =
_
_
_
0.981
0.981
0.981
_
_
_
m/sec
2
and {S
v
} =
_
_
_
0.078
0.036
0.025
_
_
_
m/sec
Calculation of displacement
The displacement in the first mode is given by (Equation 3.2.49)
= []
L
n
1
M
n1
S
v1
1
=
_
_
_
0.314
0.686
1
_
_
_
60
42.772

0.078
12.527
=
_
_
_
2.752 10
3
6.017 10
3
8.769 10
3
_
_
_
m
The displacement in the second mode is given by
= []
L
n2
M
n2
S
v2
2
=
_
_
_
0.5
0.5
1
_
_
_
20
40

0.036
27.386
=
_
_
_
3.27 10
4
3.27 10
4
6.54 10
4
_
_
_
m
The displacement in the third mode is given by
= []
L
n3
M
n3
S
v3
3
=
_
_
_
1.00
0.686
0.314
_
_
_
18.832
60.802

0.025
39.913
=
_
_
_
1.907 10
4
1.309 10
4
5.986 10
5
_
_
_
m
The shear force per floor is given by
[V]
i=n
= [M]
n
L
n
M
n
n
S
vn
[V]
i=1
=
_
_
40 0 0
0 40 0
0 0 20
_
_
_
_
_
0.314
0.686
1
_
_
_
n=1
60
42.772
12.527 0.078 =
_
_
_
220.618
482.302
351.46
_
_
_
kN
[V]
i=2
=
_
_
40 0 0
0 40 0
0 0 20
_
_
_
_
_
0.5
0.5
1.0
_
_
_
n=2
20
40
27.3860.036=
_
_
_
9.81
9.81
9.81
_
_
_
kN
Thus for the third mode we have
[V]
i=3
=
_
_
40 0 0
0 40 0
0 0 20
_
_
_
_
_
1.00
0.686
0.314
_
_
_
n=3
18.832
60.802
39.913 0.025 =
_
_
_
12.153
8.339
1.907
_
_
_
kN
Base shear per mode
The base shear per mode is given by
For the first mode, V
b
= 221 + 482 + 351 = 1054 kN; the second Mode,
V
b
= 9.8+9.89.8 = 9.8 kN; and for the third mode, V
b
= 128+2 = 6 kN.
Thus, V
b
=
_
_
_
1054
9.8
6
_
_
_
kN
3.2.6 Modal combination of forces
Once the maximum response for each mode is obtained as described above, it is
essential to obtain the combined response of all modes. As the modal maxima may or
may not occur at the same time and nor have the same sign they cannot be combined
to give accurate total maximum response. The most convenient way to represent this
is to combine them based on probability basis.
Three techniques often used for modal combination of forces are
Absolute Sum Method (ABSSUM)
Square root of Sum Square (SRSS)
Complete Quadratic Combination (CQC)
3.2.6.1 The absolute sum method (ABSSUM)
As the name suggests by this method the modal combination of all responses are
obtained by summing up the absolute values of the response without considering their
algebraic signs.
Thus, based on above
n
=
n
i=1
|
i
| (3.2.50)
where, |
i
| represents the absolute value of the responses, without consideration of
their algebraic sign.
This method though still in practice sometimes has been observed to give results
which are too conservative and is nowa days only used in case of non-critical structure.
Use of this method for important and critical structures has almost been abolished.
3.2.6.2 The square root of sum square method (SRSS)
In this method the modal response are obtained by summing up the square of the
responses and taking its root and has been found to give a much better result
(Rosenblueth 1951).
This method is however valid only when the frequencies of the structure are widely
spaced. For structures having repeated roots or closely spaced roots, CQC is found to
be superior, however when eigen values are widely spaced SRSS and CQC converges
to almost identical results.
Thus based on the above we have,
n
=
_
n
i=1
2
i
(3.2.51)
3.2.6.3 The complete quadratic combination method (CQC)
In this method (Der Kiureghian 1981) the response of the system is obtained by the
expression
n
=
_
n
i=1
n
j=1
ij
j
(3.2.52)
in which, n = number of modes being considered;
i
= response quantity in mode i;
j
= response quantity in mode j, and,
ij
= Cross modal coefficient and is given by
ij
=
8
_
D
i
D
j
(D
i
+
ij
D
j
)
3
2
ij
(1
2
ij
)
2
+4D
i
D
j
ij
(1 +
ij
)
2
(3.2.53)
where, D
i
= Modal damping ratio for mode i; D
j
= modal damping ratio for mode j,
and,
ij
= frequency ratio (
i
/
j
).
For normal seismic dynamic analysis the damping ratio is usually considered
constant for all modes when the above equation reduces to
ij
=
8D
2
(1 +
ij
)
3
2
ij
(1
2
ij
)
2
+4D
2
ij
(1 +
ij
)
2
(3.2.54)
Cross modal coefficient
0
0.2
0.4
0.6
0.8
1
1.2
0
.
5
0
.
6
8
0
.
8
6
1
.
0
4
1
.
2
2
1
.
4
1
.
5
8
1
.
7
6
1
.
9
4
Frequency Ratio
C
r
o
s
s

m
o
d
a
l

c
o
e
f
f
i
c
i
e
n
t

2% DR
5% DR
7% DR
10% DR
15% DR
20%DR
25% DR
Figure 3.2.14 Variation of cross modal frequency for different frequency ratios.
The variation of the cross modal response with frequency ratio for various damping
ratio is as shown in Figure 3.2.14.
From the curve we make a very interesting observation. The cross modal ratio plays
a significant part in the magnitude when the frequency ratio varies between 0.88 to
1.14. For other frequencies (which are widely apart) they diminish rapidly and their
contribution is insignificant. In other words for widely space frequencies the CQC
method in effect converges to the SRSS method.
We now further elaborate the above theories based on a suitable example
Example 3.2.7
For a typical three storied frame, the natural frequencies calculated are 4.257,
8.66 and 14.382 rad sec respectively. The corresponding base shear estimated
are V
b
=
_
_
_
330
75
33
_
_
_
kN, find out the combined maximum base shear based on
Absolute sum method.
Square root of sum square method.
Complete quadratic combination method.
Consider 5% damping in all modes.
Solution:
Absolute sum method
As per this the base shear is given by
V
b
= 330 +75 +33 = 438 kN.
Square root of sum square method
As per this the base shear is given by
V
b
=
_
330
2
+75
2
+33
2
= 340 kN
Complete quadratic combination method
As a first step we find out the values
ij
= (
i
/
j
) for the three modes which is as given below
Mode 1 2 3
1 1 0.49157 0.295995
2 2.034296 1 0.602142
3 3.378436 1.660739 1
Considering 5% damping as constant for all mode we have the cross modal
values as
ij
=
8D
2
_
1 +
ij
_
3
2
ij
(1
2
ij
)
2
+4D
2
ij
_
1 +
ij
_
2
Mode 1 2 3
1 1 0.0123 0.002712
2 0.025022 1 0.027546
3 0.009162 0.045746 1
Now considering V
b
=
_
_
_
330
75
33
_
_
_
kN. and applying the equation,
n
=
_
n
i=1
n
j=1

i
ij
j
, we have base shear based on CQC expression as
Mode 1 2 3
1 108900 304.432 29.53152
2 619.3049 5625 2475
3 99.77034 113.2222 1089
Adding all the nine terms in the above table and taking square root we have
V
b
=
119255.3 = 345.33 kN.

Thus it will be observed that based on CQC method base shear is 345 kN in
lieu of 340 kN based on SRSS method. Since the frequencies are widely spaced
the variation is only marginal about 1.56% only.
3.3 TIME HISTORY ANALYSIS UNDER EARTHQUAKE FORCE
Time history analysis under earthquake force is possibly the most comprehensive anal-
ysis one can undertake. However in spite of its rigorous mathematical basis
15
, modal
response technique has still remained a more popular method in day to day design
office practice.
The reason underlying the same can be attributed primarily to lack of site accelero-
grams which is the basic input for such an analysis. Previously site specic ground
acceleration data available was few and far for which engineers always preferred to
use the modal response technique using the response spectrumcurve which is available
in all codes of all countries having a specic earthquake code.
However in last thirty years there has been a significant technological advancement
based on which earthquake accelerograms are now almost globally available for all
major earthquakes. All major and minor tremors occurring around the World are now
being manned constantly.
This has signicantly enhanced our data base and in years to come for important
structures time history analysis would hopefully become a routine affair
16
. We show
hereafter a typical acceleration spectrum for the famous El-Centro Earthquake in
Figure 3.3.1.
When an earthquake occurs anywhere in the world the seismic monitoring station
picks up the tremor signals and based on such data ground acceleration/velocity at
different time steps are obtained. This data is further used as input ground acceleration
for time history analysis of structure to be build at that site or at its close proximity.
The theory underlying the method remains the same as shown in Chapter 5 (Vol. 1)
except the fact that we had earlier solved the problem with the forcing function as
harmonic force which in case of earthquake is the ground acceleration, { a
g
}
17
. Thus,
the basic equation of motion is
[M]{

X
t+t
} + [C]{

X
t+t
} + [K]{X
t+t
} = {R
t+t
} (3.3.1)
15 Refer Chapter 5 (Vol. 1) where we have discussed the various techniques of time history analysis.
16 For Nuclear power plants time history response analysis is now mandatory for all class 1 type structures
like turbine building, reactor building, spent fuel chamber etc.
17 This is usually obtained as an input from the site based on observed data like the one as shown for the
El-Centro Earthquake.
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.0 5.0 10.0 15 20 25 30 35
Time in seconds
A
c
c
e
l
e
r
a
t
i
o
n
,

g

Figure 3.3.1 Accelerogram of EL centro earthquake of May 18, 1940 NS component.
The termR
t+t
is obtained by multiplying the ground acceleration data by the mass,
[M]. In other words, here R
t+t
= [M]{ a
g
}
t+t
at every time step, t + t.
Thus once the force R
t+t
is known, rest of the procedure remains the same as what
has been described earlier in Chapter 5 (Vol. 1).
For instance the steps of Newmark- method gets slightly modified for earthquake
case as follows:
Steps for Newmark- method for earthquake analysis
Assemble the mass matrix, [M], damping matrix [C] and stiffness matrix [K].
Evaluate {

X
0
} (This will be obtained from the accelerometer data of the site).
Select time step size t and parameters and where 0.50 and =
0.25(0.5 +)
2
.
Calculate integration constant
0
=
1
t
2
,
1
=

t
,
2
=
1
t
,
3
=
1
2
1,
4
=

1,
5
=
t
2
_
2
_
,
6
= t(1 ),
7
= t . (3.3.2)
Form the modified stiffness matrix as
[

K] = [K] +
0
[M] +
1
[C] (3.3.3)
Calculate modified load at time t +t
{
R
t+t
} = [M]{ a
g
}
t+t
+ [M]{
0
X
t
+
2

X
t
+
3

X
t
}
+ [C]{
1
X
t
+
4

X
t
+
5

X
t
} (3.3.4)
Solve for displacement vector
[

K]{X
t+t
} = {
R
t+
t
} (3.3.5)
Calculate the acceleration and velocity at time t +t
{

X
t+t
} =
0
{X
t+t
X
t
}
2
{

X
t
}
3
{

X
t
} (3.3.6)
{

X
t+t
} = {

X
t
} +
6
{

X
t
} +
7
{

X
t+t
} (3.3.7)
For the sake of brevity we now explain the above through a suitable numerical
problem.
Example 3.3.1
A frame foundation supporting a compressor is subjected to El-Centro
accelerogram as shown in Figure 3.3.1. The stiffness, mass and damping
(non-proportional) matrix are given. Determine the response of the machine
foundation based on the time history response.
[M] =
_
200 0
0 1000
_
[C] =
_
7000 2800
2800 12300
_
and [K] =
_
3000 1200
1200 51000
_
Solution:
The displacement history is shown in tabular form for the first 10 steps at time
step of 0.02 seconds and the results of displacement and acceleration for node 2
and node 1 are finally shown graphically for 1566 steps in Figures 3.3.2 and 3.
Displacement Velocity at Acceleration Displacement Velocity at Acceleration
Sl. No. at node 1 node 1 at node 1 at node 2 node 2 at node 2
1 0 0 0 0.00 10
+00
0 0
2 2.98 10
06
2.98 10
04
2.98 10
02
3.24 10
06
0.000324 0.032396
3 1.18 10
05
5.82 10
04
1.36 10
03
1.32 10
05
0.000672 0.002417
4 2.59 10
05
8.35 10
04
2.66 10
02
2.99 10
05
0.001001 0.030511
5 4.98 10
05
1.55 10
03
4.53 10
02
5.83 10
05
0.001838 0.053184
6 9.13 10
05
2.59 10
03
5.83 10
02
1.07 10
04
0.003077 0.07067
7 1.50 10
04
3.25 10
03
7.55 10
03
1.78 10
04
0.003967 0.018316
8 2.13 10
04
3.06 10
03
2.68 10
02
2.57 10
04
0.003899 0.02506
9 2.66 10
04
2.27 10
03
5.15 10
02
3.26 10
04
0.003049 0.06
10 3.07 10
04
1.80 10
03
4.09 10
03
3.80 10
04
0.002377 0.00713
Displacement(d
2
)
-2.00E-02
-1.00E-02
0.00E+00
1.00E-02
2.00E-02
3.00E-02
0
1
.
4
2
.
8
4
.
2
5
.
6
7
8
.
4
9
.
8
1
1
.
2
1
2
.
6
1
4
1
5
.
4
1
6
.
8
1
8
.
2
1
9
.
6
Time step(sec)
D
i
s
p
l
a
c
e
m
e
n
t
(
m
)
Figure 3.3.2 Displacement History at node 2.
Acceleration(node 1)
-4
-2
0
2
4
0
1
.
3
2
2
.
6
4
3
.
9
6
5
.
2
8
6
.
6
7
.
9
2
9
.
2
4
1
0
.
6
1
1
.
9
1
3
.
2
1
4
.
5
1
5
.
8
1
7
.
2
1
8
.
5
1
9
.
8
Time steps
A
c
c
e
l
e
r
a
t
i
o
n

a
t

n
o
d
e

1

(
m
/
s
e
c
2
)
Figure 3.3.3 Acceleration response at node 1.
This method as discussed earlier in Chapter 5 (Vol. 1) is applicable on full matrix
when the mass, stiffness and damping matrix are all known. The technique is par-
ticularly suitable for cases, which has non-classical damping (where the matrix on
orthogonalization does not de-couple).
However for systems with large degree of freedom we rarely know the complete
damping matrix and we normally deal with the modal damping ratio usually defined
as a constant value for each mode for normal structural analysis.
For instance in case of analysis of 3Dframed building structure we do not (or cannot)
define the damping matrix and the usual input is the modal damping ratio assumed
constant for all modes.
In such cases we can either form the Rayliegh damping coefficient and adapting
the method as stated in Chapter 5 (Vol. 1) or proceed as mentioned hereafter.
As a first step we performthe usual eigen-value analysis and obtain the frequencies
and the eigen vectors.
Now knowing the modal damping ratio D (which is usually pre-defined) we de-
couple the equation into n number of equations (here n is the total numbers of
degree of freedom of the system) of the form
{
i=1,n
} +2D
i=1,n
i
= 1, n{
i=1,n
} +
_
2
i=1,n
_
{
i=1,n
} = { u
g
} (3.3.8)
For a given time history the above can be expressed as
{
i=1,n
} +2D
i=1,n
i=1,n
{
i=1,n
} +
_
2
i=1,n
_
{
i=1,n
}
=
_
1
1 D
2
t
_
0
u
g
()e
D
n
(t)
_
D
n
sin
_
1 D
2
n
(t )d
+
n
_
1 D
2
cos
_
1 D
2
n
(t )
_
d
_
(3.3.9)
For each of this equation we perform the time history response either by integra-
tion of the Duhamel Integral or by numerical integration based on any one of the
methods as explained in Chapter 5 (Vol. 1) and find out the values of the displace-
ment, velocity and acceleration and finally do a modal combination to obtain the
response for the different mode.
In such case the displacement {u
t
} =

n
1
[
n
]{
t
} and acceleration { u
t
} =
n
1
[
n
]{
t
}.
The corresponding effective earthquake force is given by
{V
n
(t)} = [M] { u
t
} . (3.3.10)
The total modal force for all modes are given by
{V
n
(t)} = [M] []
2
n
{
t
} (3.3.11)
For many large complex structures or finite element system with many degrees of
freedom even the above process could be time consuming and very laborious, fortu-
nately for many such systems, it is the first few modes which contribute signicantly
to the inertial forces when the subsequent higher modes can be neglected without any
appreciable error.
In such case if for a system NN if J number of modes (J << N) are deemed to be
significant (which can very well be estimated fromthe modal mass participation). Then
the mass matrix [M]
NN
, stiffness matrix [K]
NN
and the damping matrix [C]
NN
can well be crunched down to a matrix of order J J by the following operations
[

M]
JJ
= []
JN
[M]
NN
[]
NJ
similarly,
[

K]
JJ
= []
JN
[K]
NN
[]
NJ
and
[

C]
JJ
= []
JN
[C]
NN
[]
NJ
. (3.3.12)
were, [

M]
JJ
= modified mass matrix of size J J; [

K]
JJ
= modified stiffness matrix
of size J J; [C]
JJ
= modified damping matrix of size J J, and, []
JN
= the eigen
vector for the first J modes of the structure of size N N.
Once the modified matrix is known we can very well undertake a time history
analysis of this modified matrix and greatly reduce our computation time.
We now explain the above theory by a suitable numerical problem.
Example 3.3.2
Shown in Figure 3.3.4 is a three-storied frame subjected to dynamic forces based
on EL-Centro Earthquake as shown in Figure 3.3.1. The damping ratio for the
structure is considered as 5%. Determine
G H X3
3000
E F X2
3000
C D X1
3000
A B
( All dimensions are in mm )
Figure 3.3.4 Sketch diagram of three-storied space frame.
The fixed base natural frequencies of the structure.
The fixed base eigen vectors.
Displacement and shear force.
Let us take,
K
AC
= K
DB
= 1.5 10
3
kN/m; K
CE
= K
DF
= 1.0 10
3
kN/m;
K
EG
= K
FH
= 0.75 10
3
kN/m; M
GH
= 200kN sec
2
/m;
M
EF
= 400 KN sec
2
/m; M
CD
= 400 KN sec
2
/m.
Solution:
The stiffness and mass matrix is given by
[K] =
_
_
5000 2000 0
2000 3500 1500
0 1500 1500
_
_
and [M] =
_
_
400
400
200
_
_
Considering, [[K] [M][
2
]][] = [0]
1
= 1.281 rad/sec;
2
= 3.162 rad/s ec;
3
= 4.135 rad/sec.
Thus the time periods for the fixed base structure is given by
T
1
= 4.97 sec, T
2
= 1.987 sec, T
3
= 1.52sec .
The mode shapes or the eigen vectors and normalised eigen vectors are
[] =
_
_
1.00 1.0 1.0
2.1715 0.5 0.9208
2.7816 1.50 0.719
_
_
;
[
i
] =
_
_
0.01615 0.03244 0.0344512
0.0350718 0.01622 0.03172
0.04493 0.02433 0.02477
_
_
Considering the orthogonal equation
{
i
} +2D
i
i
{
i
} +
2
{
i
} = { a
g
}
we have the three equations as:
1
+0.281
1
+1.640961
1
= a
g
;

2
+0.3162
2
+9.99824
2
= 0;
3
+0.4135
3
+17.098
3
= a
g
.
Performing the time history analysis based on Wilson- method for input
accelerogram of El-Centro earthquake and combining the response based on the
equation, {X} = []{}. We plot below the displacement and force history in
Figures 3.3.5 and 6.
Displacement History
-0.04
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0.05
0
2
.
2
2
4
.
4
4
6
.
6
6
8
.
8
8
1
1
.
1
1
3
.
3
1
5
.
5
1
7
.
8
2
0
2
2
.
2
2
4
.
4
2
6
.
6
2
8
.
9
3
1
.
1
Time steps(sec)
D
i
s
p
l
a
c
e
m
e
n
t
(
m
e
t
e
r
)
Modal disp1
Modal disp2
Modal disp3
Figure 3.3.5 Displacement history of the frame for the three modes.
Modal shear response(kN)
-400
-200
0
200
400
0
2
.
2
2
4
.
4
4
6
.
6
6
8
.
8
8
1
1
.
1
1
3
.
3
1
5
.
5
1
7
.
8
2
0
2
2
.
2
2
4
.
4
2
6
.
6
2
8
.
9
3
1
.
1
Time step(sec)
F
o
r
c
e
(
k
N
)
Shear1
shear 2
shear 3
Figure 3.3.6 Modal shear history of the frame for the three modes.
It will be observed that the major contribution is from the fundamental mode,
the higher mode contribution is practically insignificant.
What has been explained above is the generic theory pertaining to earthquake
dynamic and pseudo-static analysis. Though the above has been explained with respect
to frames (or buildings) can be very easily be extended to a generic finite element
model with the underlying principle remaining the same be the analysis is done based
on response spectrum method or step by step integration.
We nowshowapplication of the above theories as applied to some special structures
which are important to society and industry, have got some unique features and require
some special analytical techniques.
3.3.1 Earthquake analysis of tall chimneys
and stack like structure
Tall chimneys and vertical self-supporting vessels are an important feature of power
and petrochemical industry. Damage to them during an earthquake can have a severe
consequence both in terms of economy and loss of human life.
While it is expected that a power plant remains functional after an earthquake,
which is essential to fight the aftermath of the disaster, leakage or damage of vertical
vessels in refinery or chemical plants containing flammable or toxic liquid can create
havoc to the environment and surrounding life
18
.
Though the analysis herein is discussed in terms of tall chimneys can well be applied
for vertical self-supporting vessels also.
With ever growing demand for power, engineering industry is churning out power
plants of progressively higher capacity. To maintain the ecological balance as well as
limit the environmental pollution, chimneys emitting the spent flue gas are also get-
ting higher and higher everyday. In India it is now mandatory that for all fossil fuel
power plants the height of chimneys be minimum 220 meter for 210 MW unit and
275 meter for 500 MW unit. While this though reduces the ground pollution concen-
tration signicantly has posed new challenges to the structural engineers to come up
with a safe design of these tall chimneys especially under wind and earthquake, which
affects its behaviour signicantly. Unlike other tall structures the most dangerous thing
about chimney is that these structures are basically a cantilever structure having one
line of defence (the structures itself) and has practically no redundancy built in it. Thus
during an earthquake if any portion of it develops a hinge would invariably make the
system a mechanism with collapse being imminent.
It is for this knowing the dynamic behaviour of the same under an earthquake
loading is of primary importance.
Fortunately as these structures have uniform distribution of mass and stiffness are
more amenable to classical mathematical treatment; however one of the major con-
troversies that remain with its behaviour is the level of damping to be considered in
the analysis.
While one school of thought prefers to use the standard damping ratio as used for
RCC (57%), the other school of thought is that since of its huge mass (due to its
self weight and lining) a major portion of the chimney remains under compression
even under wind and seismic loading and thus remains un-cracked. Since propagation
of cracking enhances the damping property of the system and does not occur in the
major portion of the chimney, a much lower damping ratio of say 2% should be a
18 Though the reason was different some of the readers may remember the Bhopal gas tragedy in 1980s in
India where huge number of people perished and got disabled for life due to leakage of toxic gas from
vessels in the plant of a multi-national Company.
Plan View
EI = constant
Figure 3.3.7 Typical multi- flue-chimney with its mathematical model.
more reasonable value. Unfortunately very little field observed instrumented data are
available to come to any decisive conclusion on this issue.
Chimneys, shown in Figure 3.3.7, are usually of two types
Multi-flue chimneys (used to cater to more than two power units at a time) having
uniform cross section.
Single flues (used to cater one or two units) usually having a tapered profile.
3.3.1.1 Analysis as proposed in IS-code
Before we start with the dynamic analysis of such tall structures we present herein the
method as proposed in IS Code:
As per IS code the time period of such chimneys considered as shown in Figure 3.3.4
are fixed at base and is given by
T = C
T
_
WH
EAg
(3.3.13)
where, W = weight of chimney plus lining and all other accessories; H = height of
chimney above the base; E = modulus of elasticity of the structural shell; A = area of
cross section of the base; g = acceleration due to gravity; C
T
= constant which is a
Table 3.3.1
Slenderness ratio C
T
Cv
5 14.4 1.02
10 21.2 1.12
15 29.6 1.19
20 38.4 1.25
25 47.2 1.30
30 56 1.35
35 65 1.39
40 73.8 1.43
45 82.8 1.47
50 or more 1.8 (H/r) 1.50
function of the slenderness ratio; For circular section A = 2rt; r = mean radius of
the shell and t = thickness of the shell.
The design base shear and moment for xed base is given by
V = C
v
A
h
W
_
1.1
_
z
H
_
0.5
+0.75
_
z
H
_
+0.9
_
z
H
_
4
_
;
M = A
h
W

H
_
0.4
_
z
H
_
0.5
+0.6
_
z
H
_
4
_
(3.3.14)
where, C
v
= a coefficient which is a function of slenderness ratio;

H = height of
centre of gravity of the structure above base, and A
h
=
ZI
2R
, the seismic coefficient as
per code.
The values of C
v
and C
T
are as furnished in Table 3.3.1.
IS code does not furnish any expression for the tip deflection.
3.3.1.2 Dynamic analysis of tall chimneys
We start with the analysis of a multi flue chimney of uniform cross section based on
Rayleigh Ritz technique to arrive at a closed form solution before extending the same
to a numerical solution for a tapered cantilever.
Since the outer core of a multi-flue chimney (usually termed as the wind shield) is
of uniform cross section we consider it as a cantilever beam fixed at base.
The free vibration equation of such beam is given by the expression (Hurty and
Rubenstein 1967),
EI
_
4
w
z
4
_
+ A
_
2
w
t
2
_
= 0 (3.3.15)
here, E = elastic modulus of the beam material; I = moment of inertia of the beam;
= mass density of the beam material; A = area of cross section of the beam, and,
w =displacement of the beamand is a function of time and geometry and is depicted as
w(z, t) = Y(z) q(t) (3.3.16)
Based on separation of variable technique the above partial differential equation
can be separated into two linear differential equation and one of which is
EI
_
d
4
Y
dz
4
_

4
Y = 0 where
4
=
A
2
EI
(3.3.17)
The generic solution to this equation is given by (Murray 1967)
Y = C
1
sin z +C
2
cos z +C
3
sin hz +C
4
cos hz (3.3.18)
Imposing the four boundary conditions:
1 Y = 0 at z = 0;
2
dY
dz
= 0 at z = 0;
3
d
3
Y
dz
3
= 0 at z = L;
4
d
2
Y
dz
2
= 0 at z = L.
(3.3.19)
We have the shape function solution as
Y
m
= sin
m
z
H
sin h
m
z
H

m
_
cos
m
z
H
cos h
m
z
H
_
(3.3.20)
Here m = number of modes 1, 2, 3, . . . . . . . . . .
m
= 1.875, 4.694, 7.855,
2m1
2
.
For m = 1, 2, 3. . . . . . m, etc.
m
=
sin
m
+sin h
m
cos
m
+cos h
m
We apply here the Rayliegh Ritz technique as described below.
For a conservative system if T is kinetic energy and V is the Potential energy of the
system then at any time t the energy equations may be written in the form
T(t) =
1
2
H
_
0
m(z)
_
y(z, t)
t
_
2
dz (3.3.21)
here y(z, t) =
n
i=1
i
(z)q
i
(t) (3.3.22)
where y (z, t) = displacement function;
i
(z) = admissible function, and, q
i
(t) =
generalized co-ordinate.
Substituting the above in the energy equation we have
T(t) =
1
2
H
_
0
m(z)
_
n
i=1
i
(z) q
i
(t)
_
_
_
n
j=1
j
(z) q
j
(t)
_
_
dz
=
1
2
n
i=1
n
j=1
q
i
(t) q
j
(t)
_
_
H
_
0
m(z)
i
(z)
j
(z)dz
_
_
(3.3.23)
from which we conclude that the mass coefficient has the form
m
ij
=
_
_
H
_
0
m(z)
i
(z)
j
(z)dz
_
_
for i, j = 1, 2, 3. . . . . . n (3.3.24)
For potential energy V we have
V(t) =
1
2
H
_
0
EI(z)
_
2
y(z, t)
z
2
_
2
dz
=
1
2
n
i=1
n
j=1
q
i
(t) q
j
(t)
_
_
H
_
0
EI(z)
d
2
i
(z)
dz
2
d
2
j
(z)
dz
2
_
_
dz, (3.3.25)
from which we conclude that stiffness has the form
k
ij
=
H
_
0
EI(z)
d
2
i
(z)
dz
2
d
2
j
(z)
dz
2
dz for i, j = 1, 2, 3. . . . . . n (3.3.26)
Since a multi-flue stack is considered to have a constant EI the stiffness and mass
expression is given as
k
ij
= EI
H
_
0
d
2
i
(z)
dz
2
d
2
j
(z)
dz
2
dz and m
ij
=
_
_
A
g
H
_
0
i
(z)
j
(z)dz
_
_
(3.3.27)
Considering the shape function as
i
= sin
i
z
H
sin h
i
z
H

i
_
cos
i
z
H
cos h
i
z
H
_
and
(3.3.28)
j
= sin
j
z
H
sin h
j
z
H

j
_
cos
j
z
H
cos h
j
z
H
_
The double derivative of the above is given by
i
=

2
i
H
2
_
sin
i
z
H
sin h
i
z
H
+
i
_
cos
i
z
H
+cos h
i
z
H
__
and
(3.3.29)
j
=
2
j
H
2
_
sin
j
z
H
sin h
j
z
H
+
j
_
cos
j
z
H
+cos h
j
z
H
__
Before performing the integration we change the above to generalized co-ordinate
by considering, =
z
H
when d =
dz
H
and as z 0, 0 and as z H, 1
based on above we can now express the double derivative as
f
()
i
=

2
i
H
2
_
sin
i
sin h
i
+
i
(cos
i
+cos h
i
)
_
, and
(3.3.30)
f
()
j
=
2
j
H
2
_
sin
j
sin h
j
+
j
(cos
j
+cos h
j
)
_
Thus stiffness of the system can now be expressed as
k
ij
=
EI
2
i

2
j
H
3
1
_
0
f
()
i
f
()
j
d (3.3.31)
and mass of the system is given by
m
ij
=
AH
g
1
_
0
f ()
i
f ()
j
d, where i = j = 1, 2, 3, . . . . . . m, (3.3.32)
For most of the chimneys it is found that first three modes are sufficient to predict
the dynamic response, as modal mass participation is almost 100% by this.
Thus for the first three modes the stiffness matrix
19
is given by
[K]
ij
=
EI
H
3
_
_
_
_
_
_
_
_
_
_
4
1
1
_
0
f
()
2
1
d
2
2
2
1
1
_
0
f
()
2
f
()
1
d
4
2
1
_
0
f
2
()
2
d
2
3
2
1
1
_
0
f
()
3
f
()
1
d
2
3
2
2
1
_
0
f
()
3
f
()
2
d
4
3
1
_
0
f
()
2
3
d
_
_
(3.3.33)
and the mass matrix is given by
[M]
ij
=
AH
g
_
_
_
_
_
_
_
_
_
_
1
_
0
f ()
2
1
d
1
_
0
f ()
2
f ()
1
d
1
_
0
f
2
()
2
d
1
_
0
f ()
3
f ()
1
d
1
_
0
f ()
3
f ()
2
d
1
_
0
f ()
2
3
d
_
_
(3.3.34)
The above integrals can very easily be solved based on Simpsons 1/3rd rule between
the limits 1 to 0 when we have
[K]
33
=
EI
H
3
_
_
22.936 0.002 0.006
0.002 468.044 0.11
0.006 0.11 3812.81
_
_
(3.3.35)
and the mass matrix is given by,
[M]
33
=
W
g
_
_
1.855 0 0
0 0.964 0
0 0 1.002
_
_
(3.3.36)
where, W = total weight of the shell + internal slabs + brick linings.
Converting the above into standard eigen-value form of A = and applying the
generalized Jacobi technique
20
we have
[] =
EIg
WH
3
_
_
12.364 0 0
0 485.523 0
0 0 3805
_
_
and (3.3.37)
19 The stiffness and mass matrix is symmetric about is diagonal.
20 The technique has been worked in detail in Chapter 5 (Vol. 1).
Eigen vectors for the first three modes
-3
-2
-1
0
1
2
3
0
0
.
1
0
.
2
0
.
3
0
.
4
0
.
5
0
.
6
0
.
7
0
.
8
0
.
9
1
Z/H
E
i
g
e
n
v
e
c
t
o
r
s
f1(x)
f2(x)
f3(x)
Figure 3.3.8 Eigen vectors for the first three modes.
the eigen vectors are given as
[] =
_
_
_
_
1.0 2.278 10
6
8.528 10
7
4.384 10
6
1 3.437 10
5
1.579 10
6
3.307 10
5
1
_
_
_
_
_
f
1
()
f
2
()
f
3
()
_
_
(3.3.38)
The eigen vector plots for the first three modes are as shown in Figure 3.3.8. since
[] =
2
and T =
2
we have
[T] =
_
_
1.787 0 0
0 0.285 0
0 0.102
_
_
_
WH
3
EIg
(3.3.39)
Thus for the first three modes we have
Mode number Time period (secs)
1 T
1
= 1.787
_
WH
3
EIg
2 T
2
= 0.285
_
WH
3
EIg
3 T
2
= 0.102
_
WH
3
EIg
3.3.1.3 Transformation to the format of IS-code
We have shown above that for fundamental mode the time period is given by
T
1
= 1.787
_
WH
3
EIg
, (3.3.40)
Now, considering I = Ar
2
, where A = area of the stack at the base and r = radius
of gyration, the equation can be written in the format of
T
1
= 1.787
_
WH
EAg
(3.3.41)
where = slenderness ratio of the stack @ H/r.
Considering, C
T
= 1.7873, we have
T
1
= C
T
_
WH
EAg
(3.3.42)
which is the same format as presented in the code. If we compare the values of C
T
as
furnished in code and as derived here it will be observed that code gives a higher value
of time period vis-a-vis what is presented here.
Since the accuracy of Rayleigh Ritz Method is dependent on the choice of the
assumed shape function it is evident that code had used a different shape function
then what has been presented herein
21
.
The various values of C
T
as proposed by the present method and what has been
proposed in the code are as mentioned hereunder
22
.
Slenderness ratio (H/r) C
T
(as per IS code) C
T
(1st Mode) C
T
(2nd Mode) C
T
(3rd Mode)
5 14.4 8.935 1.425 0.51
10 21.2 17.87 2.85 1.02
15 29.6 26.81 4.275 1.53
20 38.4 35.74 5.70 2.04
25 47.2 44.675 7.125 2.55
30 56 53.61 8.55 3.06
35 65 62.54 9.975 3.57
40 73.8 71.5 11.4 4.08
45 82.8 80.41 12.83 4.59
50 1.8 (H/r) 89.35 14.25 5.10
21 Present analysis would give slightly different (higher) values of moments and shears then what has been
proposed in the code.
22 IS-1893 does not propose any C
T
values for 2nd or 3rd mode.
In terms of response spectrum analysis displacement S
d
is given by, S
d
= S
a
/
2
.
Expressing it in terms of codal formulation, we may express it as
S
d
=
i
ZI
2R
S
a
2
(3.3.43)
where,
i
= modal participation factor and is given by
n
i=1
m
i
i
/
n
i=1
m
i
2
i
.
For an element of length dz the above can be expressed as
i
=
A
g
_
H
0

i
dz
A
g
_
H
0

2
i
dz
=
_
1
0
f
i
() d
_
1
0
f
i
()
2
d
(3.3.44)
in which, Z = zone coefficient; I = importance factor, and, R = ductility factor.
Integration of the mass participation factor within limits 1 to 0 for the first three
modes gives
Mode number Mass participation factor (
i
)
1 0.575
2 0.442
3 0.254
Now considering, =
ZI
2R
, an IS code factor, we can write the time dependent
function of displacement as
S
d
=
i
S
a
2
(3.3.45)
Thus for the first mode, we have
S
d
= 0.575
S
a1
WH
3
12.364EIg
= 0.0465
S
a1
WH
3
EIg
(3.3.46)
Let the complete function is given by, w(z, t) = (z) q(t), thus for this case
w(z, t) = 0.0465
S
a1
WH
3
EIg
[f
1
() +4.384 10
6
f
2
() +1.579 10
6
f
3
()],
neglecting the influence of the second and third mode whose influence are negligible
we have
w(z, t) = 0.0465
S
a1
WH
3
EIg
[f
1
()] (3.3.47)
For calculation of moment and shear we know that, EI
d
2
w
dz
2
= M
z
, and hence,
M
z
= 0.0465(WH
3
)
_
S
a1
g
_
1
H
2
[
2
1
f
()]
= 0.163 (WH)
_
S
a1
g
_
[f
1
()] (3.3.48)
Again considering, V
z
=
dM
z
dz
, we have
V
z
= 0.306W
_
S
a1
g
_
[f
1
()] (3.3.49)
Proceeding in identical manner for the second mode, we have
w(z, t) = 9.103 10
4
S
a2
WH
3
EIg
[f
2
()], (3.3.50)
ignoring the influence of mode one and three as their influence are very small.
M
z
= 9.103 10
4
WH
_
S
a2
g
_
[
2
2
f
2
()]
= 2.005 10
2
WH
_
S
a2
g
_
[f
2
()] and
V
z
= 9.415 10
2
W
_
S
a2
g
_
[f
2
()] (3.3.51)
Similarly for the third mode, we have
w(z, t) = 6.675 10
5
S
a3
WH
3
EIg
[f
3
()]
M
z
= 4.12 10
3
WH
_
S
a3
g
_
[f
3
()], and
V
z
= 0.0323W
_
S
a2
g
_
[f
3
()] (3.3.52)
The above can thus be generalized along the height of the chimney as
w
i
(, t) = (Coeff d)
S
ai
WH
3
EIg
,
M
= (Coeff m) W H
_
S
ai
g
_
, and V
= (Coeff v)W
_
S
ai
g
_
. (3.3.53)
Here = z/H, the height ratio and i = number of mode. It will be observed
that once we know the values of coefficients within parenthesis for i = 1, 2, 3, we
can immediately find out the dynamic amplitude, shear and moments without going
through the elaborate process of dynamic analysis.
The coefficients for dynamic amplitude, moment and shears are as stated hereafter
= z/H Coeff d
1
Coeff d
2
Coeff d
3
Coeff m
1
Coeff m
2
Coeff m
3
Coeff v
1
Coeff v
2
Coeff v
3
0 0.00000 0.00000 0.00000 0.44401 0.03938 0.00825 0.61200 0.18830 0.06460
0.05 0.00054 0.00005 0.00001 0.41345 0.02997 0.00502 0.61189 0.18793 0.06401
0.1 0.00212 0.00017 0.00003 0.38291 0.02063 0.00189 0.61112 0.18549 0.06040
0.15 0.00466 0.00034 0.00006 0.35244 0.01149 0.00095 0.60907 0.17947 0.05212
0.2 0.00809 0.00054 0.00008 0.32211 0.00276 0.00325 0.60518 0.16887 0.03896
0.25 0.01232 0.00075 0.00010 0.29203 0.00531 0.00479 0.59892 0.15328 0.02192
0.3 0.01729 0.00094 0.00010 0.26234 0.01248 0.00542 0.58982 0.13276 0.00290
0.35 0.02291 0.00110 0.00009 0.23318 0.01851 0.00509 0.57746 0.10786 0.01567
0.4 0.02911 0.00122 0.00007 0.20472 0.02320 0.00390 0.56144 0.07951 0.03129
0.45 0.03584 0.00128 0.00004 0.17717 0.02642 0.00204 0.54143 0.04899 0.04183
0.5 0.04300 0.00128 0.00000 0.15072 0.02809 0.00019 0.51713 0.01780 0.04589
0.55 0.05054 0.00120 0.00003 0.12559 0.02821 0.00244 0.48829 0.01241 0.04301
0.6 0.05840 0.00105 0.00006 0.10203 0.02689 0.00439 0.45470 0.03993 0.03382
0.65 0.06652 0.00084 0.00008 0.08026 0.02429 0.00575 0.41618 0.06308 0.01995
0.7 0.07483 0.00057 0.00009 0.06054 0.02068 0.00635 0.37259 0.08032 0.00380
0.75 0.08330 0.00024 0.00007 0.04313 0.01639 0.00614 0.32382 0.09027 0.01175
0.8 0.09188 0.00012 0.00005 0.02829 0.01180 0.00523 0.26979 0.09181 0.02373
0.85 0.10053 0.00052 0.00001 0.01628 0.00736 0.00387 0.21044 0.08409 0.02945
0.9 0.10922 0.00093 0.00004 0.00736 0.00356 0.00242 0.14574 0.06652 0.02674
0.95 0.11793 0.00136 0.00009 0.00181 0.00088 0.00136 0.07568 0.03877 0.01412
1 0.12664 0.00178 0.00015 0.00011 0.00015 0.00119 0.00024 0.00071 0.00933
We now elaborate the theory with a suitable numerical solution.
Example 3.3.3
A multi-Flue chimney has height of 220 m. Its estimated weight including lining
and internal slab is 175,000 kN. The diameter of the chimney at is 22.0 mhaving
average shell thickness of 650 mm. The chimney is situated in a place depicted
by zone IV as per IS-code resting on medium soil. Find the deflection, moment
and shear for first three modes and the maximum design moments and shears.
Consider grade of concrete used as M30 and damping ratio as 5% for the three
modes.
Solution:
Outside diameter of chimney =22 m; Shell thickness =650 mm, Inside diameter
of chimney = 20.7 m.
Youngs Modulus of concrete = 5700
30 10
3
= 31220186 kN/m
2
.
Moment of Inertia at base =

64
_
22
4
20.7
4
_
= 2486.39 m
4
Area of chimney at base =

4
_
22
2
20.7
2
_
= 43.59745 m
2
Radius of gyration =
_
I
A
= 7.552 m; Slenderness Ratio =
220
7.552
= 29.18139.
Considering Time period = T
1
= 1.787
_
WH
EAg
for first mode, we have
T
1
= 1.787 29.18139
_
175000 220
31220186 43.59745 9.81
= 2.8sec .
Similarly for the mode 2, we have
T
2
= 0.285 29.18139
_
175000 220
31220186 43.59745 9.81
= 0.446sec, givesS
a
/g = 2.5.
And for the mode 3, we have
T
2
= 0.102 29.18139
_
175000 220
31220186 43.59745 9.81
= 0.1594sec s, gives S
a
/g = 2.5.
For Zone IV medium stiff soil as per IS-1893
Z = 0.24 I = 1.5 R = 2.0 =
0.24 1.5
2 2
= 0.09
Substituting the values of W, H, E, I, S
a
/g and , and multiplying it by the
coefficients as furnished earlier we have
1

2

3
M
1
M
2
M
3
V
1
V
2
V
3
(m) (m) (m) (kN m) (kN m) (kN m) (kN) (kN) (kN)
0 0 0 0 327879 12543.67 865.5315 2823.571 1303.114 242.3699
0.05 4.9E-06 1.2E-05 7.97E-06 284303.7 7265.948 320.4488 2822.533 1297.935 237.9722
0.1 7.5E-05 0.00016 9.23E-05 243858.8 3441.049 45.18539 2815.416 1264.562 211.8813
0.15 0.00036 0.00068 0.00032 206595.1 1067.111 11.61869 2796.584 1183.756 157.8184
0.2 0.00109 0.00173 0.000648 172575.8 61.55624 134.9782 2760.962 1048.157 88.21899
0.25 0.00253 0.00333 0.000931 141858.9 228.489 292.6448 2704.16 863.554 27.95214
0.3 0.00498 0.00529 0.001015 114482.5 1260.757 374.0941 2622.622 647.893 0.499664
0.35 0.00875 0.00729 0.00084 90452.24 2773.24 330.6002 2513.781 427.7108 14.16074
0.4 0.01413 0.00893 0.000491 69730.34 4358.467 194.2332 2376.22 232.5303 56.56307
0.45 0.0214 0.00983 0.000151 52228.61 5651.426 53.75978 2209.821 88.35721 101.0252
0.5 0.03082 0.00974 6.82E-07 37803.75 6388.65 0.338409 2015.894 11.71511 121.2799
0.55 0.04258 0.00861 0.00011 26256.23 6447.994 73.71389 1797.282 5.593723 106.0343
0.6 0.05685 0.00664 0.000398 17332.5 5859.733 239.4717 1558.446 58.30672 64.83385
0.65 0.07374 0.00423 0.000678 10730.79 4785.996 409.8315 1305.506 145.6543 21.74064
0.7 0.09334 0.00192 0.000767 6110.51 3472.652 495.0058 1046.263 236.0995 0.493604
0.75 0.11566 0.00035 0.0006 3105.154 2184.266 454.2715 790.1897 298.0526 9.901529
0.8 0.14071 9.4E-05 0.000276 1338.709 1137.157 316.2084 548.3924 307.9525 38.26835
0.85 0.16845 0.00163 2.37E-05 445.3734 447.2201 156.1138 333.5583 257.7331 60.66253
0.9 0.19883 0.00524 9.28E-05 92.42883 107.6952 44.97968 159.8887 160.4891 55.77621
0.95 0.23182 0.01107 0.000657 6.071341 8.071409 3.881338 43.02797 53.6873 24.17968
1 0.26735 0.01911 0.001775 0 0 0 5.67E-05 1.04E-05 1.43E-05
The design values are obtained by the SRSS values of the three modes and are
as given here under
D(comb) m Mcomb kN m Vcomb kN
0 0 328120 3119.199
0.05 1.55E-05 284396.7 3115.759
0.1 0.000203 243883 3093.635
0.15 0.000836 206597.9 3040.899
0.2 0.002147 172575.9 2954.543
0.25 0.004283 141859.4 2838.836
0.3 0.007337 114490.1 2701.465
0.35 0.011419 90495.35 2549.947
0.4 0.01672 69866.69 2388.24
0.45 0.023551 52533.5 2213.893
0.5 0.032319 38339.78 2019.572
0.55 0.043441 27036.49 1800.416
0.6 0.057237 18297.8 1560.883
0.65 0.073869 11756.85 1313.786
0.7 0.093362 7045.756 1072.572
0.75 0.115663 3823.528 844.5906
0.8 0.140709 1784.729 630.1059
0.85 0.168457 650.1806 425.8726
0.9 0.198904 148.8775 233.3069
0.95 0.23208 10.82006 72.92729
1 0.268037 0 5.94E-05
Bending Moment under earthquake
-100000
0
100000
200000
300000
400000
0
0
.
1
5
0
.
3
0
.
4
5
0
.
6
0
.
7
5
0
.
9
Z/H
M
o
m
n
e
t
(
k
N
.
M
)
M1
M2
M3
Mcomb
Figure 3.3.9 Bending Moment diagram of the chimney.
Shear force diagram for three modes
0
1000
2000
3000
4000
0
0
.
1
5
0
.
3
0
.
4
5
0
.
6
0
.
7
5
0
.
9
Z/H
S
h
e
a
r

f
o
r
c
e
(
k
N
)
V1
V2
V3
Vcomb
Figure 3.3.10 Shear force diagram of the chimney.
We show in Figs. 3.3.9 and 10, the Modal moments shear and SRSS values.
3.3.1.4 Analysis of single f lue tapered chimney
For single flue chimney due to thermodynamic reason and to enhance the exit velocity
of the flue gas
23
, are usually provided with a tapered section as shown in Figure 3.3.11.
Besides the reason cited, the choice is also structurally reasonable for the moment and
shear increases from zero at top to maximum at base. Based on this the obvious
economic design be that which has a section minimum at top and maximum at base.
As the section has a varying profile (generally linear) the mathematical treatment
as shown for the multi-flue chimney with constant EI becomes complex for a closed
form solution except for the fundamental mode.
23 The flue gas needs sufficient exit velocity to reduce the ground level pollution concentration.
Variable EI
Figure 3.3.11 Single Flue Tapered Chimney and its mathematical model.
However this can very easily be solved by applying numerical techniques and arrive
at an accurate answer.
In case of tapered chimneys the numerical solution is preferable because though
in most of the cases the profile is linear however from stress point of view and also
to diminish the amplitude at the top, the profile usually has a number of transition
zones (i.e. the slope often changes at two or three positions thus have varying integral
functions with different limits).
Secondly the brick liner inside the chimney shell which reduces the temperature
differential across the chimney shell also undergoes change in thickness after a certain
level thus making the mass function discontinuous which surely makes the choice of
a numerical solution more attractive.
However one additional step on has to do in this case is to perform the eigen value
analysis which was already implicit in the calculation for chimneys with constant
sections.
The theory presented earlier can be modified for numerical analysis as follows:
As the moment of inertia of the section is varying the stiffness equation can be
expressed as
k
ij
=
E
2
i

2
j
H
4
1
_
0
I
z
(z)
i
(z)
j
dz (3.3.54)
In which, I
z
is moment of inertia considered to be varying at different height z.
where,
i
=

2
i
H
2
_
sin
i
z
H
sin h
i
z
H
+
i
_
cos
i
z
H
+cos h
i
z
H
__

i
=

2
i
H
2
F
(z) and F
(z) =

2
i
H
2
_
sin
i
z
H
sin h
i
z
H
+
i
_
cos
i
z
H
+cos h
i
z
H
__
(3.3.55)
Thus the stiffness matrix can now be written as
[K]
33
=
E
H
4
_
_
_
_
_
_
_
_
_
_
4
1
H
_
0
I
z
F
1
2
(z)dz
2
2
2
1
H
_
0
I
z
F
2
(z)F
1
(z)dz
4
2
H
_
0
I
z
F
2
2
(z)dz
2
3
2
1
H
_
0
I
z
F
3
(z)F
1
(z)dz
2
3
2
2
H
_
0
I
z
F
3
(z)F
2
(z)dz
4
3
H
_
0
I
z
F
3
2
(z)dz
_
_
(3.3.56)
Similarly Mass equation can now be written as
[M]
ij
=
1
g
_
_
_
_
_
_
_
_
_
(
c
A
c
+
b
A
b
)
H
_
0
2
1
(z)
(
c
A
c
+
b
A
b
)
H
_
0
2
(z)
1
(z)
1
(
c
A
c
+
b
A
b
)
H
_
0
2
2
(z)
(
c
A
c
+
b
A
b
)
H
_
0
3
(z)
1
(z) (
c
A
c
+
b
A
b
)
H
_
0
3
(z)
2
(z) (
c
A
c
+
b
A
b
)
H
_
0
2
3
(z)
_
_
(3.3.57)
where,
c
= unit weight of concrete;
b
= unit weight of brick lining; A
c
= area of
RCC shell at any height z, and A
b
= area of Brick lining at any height z.
Each of the above term of the stiffness and mass matrix are to be obtained by
numerical integration between the limits 0-H. For numerical solution for and
following values are to be adopted.
Mode 1 2 3
1.875 4.694 7.855
1.362221 0.981868 1.000776105
After the stiffness and mass matrix are formed, an eigen values analysis needs to
performed based on the equation
[K] [M]
2
= 0 (3.3.58)
Once the eigen values vis--vis time periods are known, the S
a
/g values can be
obtained from the response curve as furnished in the codes.
The displacement amplitude is thus furnished by the equation
S
d
=
i
ZI
2R
S
a
2
(3.3.59)
and the complete solution is given by the expression
24
24 Here
i
remains same as the case with constant EI as show earlier.
w
i
(z, t) =
i
ZI
2R
S
a
2
[
ii
] F
i
(z). (3.3.60)
Thus
w
1
(z, t) =
1
ZI
2R
S
a1
2
1
[
11
F
1
(z) +
12
F
2
(z) +
13
F
3
(z)]
w
2
(z, t) =
2
ZI
2R
S
a2
2
2
[
21
F
1
(z) +
22
F
2
(z) +
23
F
3
(z)] and (3.3.61)
w
3
(z, t) =
3
ZI
2R
S
a3
2
3
[
31
F
1
(z) +
32
F
2
(z) +
33
F
3
(z)]
where, [
ii
] = eigen vector value for the mode I.
The moment and shear are then obtained from the equation
EI
d
2
w
dz
2
= M
z
and EI
d
3
w
dz
3
= V
z
, which gives
M
1
(z, t) =
1
S
a1
EI
2
1
H
2
[
11
2
1
F
1
(z) +
12
2
2
F
2
(z) +
13
2
3
F
3
(z)] (3.3.62)
where, =
ZI
2R
.
M
2
(z, t) =
2
S
a2
EI
2
2
H
2
[
21
2
1
F
1
(z) +
22
2
2
F
2
(z) +
23
2
3
F
3
(z)] and
M
3
(z, t) =
3
S
a3
EI
2
3
H
2
[
31
2
1
F
1
(z) +
32
2
2
F
2
(z) +
33
2
3
F
3
(z)] (3.3.63)
Considering EI
d
3
w
dz
3
= V
z
,
V
1
(z, t) =
1
S
a1
EI
2
1
H
3
[
11
3
1
F
1
(z) +
12
3
2
F
2
(z) +
13
3
3
F
3
(z)]
V
2
(z, t) =
2
S
a2
EI
2
2
H
3
[
21
3
1
F
1
(z) +
22
3
2
F
2
(z) +
23
3
3
F
3
(z)] and
V
3
(z, t) =
3
S
a3
EI
2
3
H
3
[
31
3
1
F
1
(z) +
32
3
2
F
2
(z) +
33
3
3
F
3
(z)] (3.3.64)
where,
F
i
(z) =
_
cos
i
z
H
cos h
i
z
H
+
i
_
sin
i
z
H
+sin h
i
z
H
__
(3.3.65)
We explain the above with a suitable numerical example.
Example 3.3.4
A 220 m tall RCC chimney has properties as shown hereafter. Calculate the first
three fundamental time period and seismic response for seismic zone IV, with
site having medium soil.
The data for the chimney are given below:
Height of chimney = 220 m; Diameter of shell at bottom = 22 m; Shell
thickness at bottom=650 mm; Diameter of shell at top =5.0; Shell thickness at
top =250 mm; Air gap between shell and lining =100 mmalthrough; Thickness
of brick lining =150 mmfrom220 to 150 m; Thickness of brick lining =230 mm
from 150 to 25 m; unit weight of concrete = 25 kN/m
3
; Unit weight of brick =
22 kN/m
3
; Grade of concrete = M35; Zone coefficient = 0.24; Importance
factor = 1.5; R(Ductility factor) = 2.0.
Solution:
For the problem the earthquake factor =
ZI
2R
=
0.24 1.5
4
= 0.09
E
conc
= 31220185.78 kN/m
2
Outside Inside Outside dia Inside dia Area of Area of Moment of
z diameter diameter lining Lining lining concrete inertia
0 22 20.7 0 0 0 43.5974521 2486.389939
11 21.15 19.8925 0 0 0 40.5351404 2135.781491
22 20.3 19.085 0 0 0 37.5834816 1823.567498
33 19.45 18.2775 18.0775 17.6175 12.89600222 34.7424755 1546.84251
44 18.6 17.47 17.27 16.81 12.31252993 32.0121223 1302.815147
55 17.75 16.6625 16.4625 16.0025 11.72905763 29.3924218 1088.808108
66 16.9 15.855 15.655 15.195 11.14558534 26.8833741 902.2581633
77 16.05 15.0475 14.8475 14.3875 10.56211304 24.4849792 740.7161567
88 15.2 14.24 14.04 13.58 9.978640746 22.1972371 601.8470066
99 14.35 13.4325 13.2325 12.7725 9.39516845 20.0201477 483.4297052
110 13.5 12.625 12.425 11.965 8.811696154 17.9537111 383.3573185
121 12.65 11.8175 11.6175 11.1575 8.228223859 15.9979274 299.6369863
132 11.8 11.01 10.81 10.35 7.644751563 14.1527964 230.3899223
143 10.95 10.2025 10.0025 9.5425 7.061279268 12.4183182 173.851414
154 10.1 9.395 9.195 8.895 4.262355833 10.7944927 128.3708229
165 9.25 8.5875 8.3875 8.0875 3.881830423 9.28132008 92.41158428
176 8.4 7.78 7.58 7.28 3.501305012 7.87880022 64.55120714
187 7.55 6.9725 6.7725 6.4725 3.120779602 6.58693314 43.48127451
198 6.7 6.165 5.965 5.665 2.740254192 5.40571884 28.00744321
209 5.85 5.3575 5.1575 4.8575 2.359728782 4.33515733 17.04944393
220 5 4.55 4.35 4.05 1.979203372 3.37524861 9.641081217
Next we define the function f
i
() for the first three modes and then multiplying
and integrating the expression m
1
= (
c
A
c
+
b
A
b
)
_
H
0

2
1
(z)dz etc. we obtain
mass matrix as shown hereafter.
z f
1
() f
2
() f
3
() f
1
(x) f
1
(x) f
2
(x) f
1
(x) f
2
(x) f
2
(x) f
3
(x) f
1
(x) f
3
(x) f
2
(x) f
3
(x) f
3
(x)
0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
11 0.01 0.05 0.13 0.14 0.59 2.51 1.59 6.77 18.25
22 0.05 0.18 0.46 1.96 7.81 31.09 19.60 78.02 195.82
33 0.10 0.37 0.85 11.60 42.86 158.35 98.31 363.21 833.10
44 0.17 0.59 1.21 32.43 110.18 374.36 225.52 766.23 1568.28
55 0.27 0.82 1.45 69.73 215.60 666.56 381.57 1179.70 2087.90
66 0.37 1.03 1.51 126.80 352.36 979.13 516.23 1434.49 2101.62
77 0.49 1.21 1.38 205.03 504.71 1242.41 573.05 1410.63 1601.62
88 0.63 1.34 1.05 303.73 650.94 1395.06 510.50 1094.07 858.02
99 0.77 1.41 0.58 420.17 767.47 1401.84 318.23 581.27 241.03
110 0.92 1.40 0.04 549.82 833.09 1262.32 23.32 35.34 0.99
121 1.09 1.32 0.50 686.68 832.57 1009.45 315.27 382.25 144.75
132 1.26 1.16 0.95 823.77 759.11 699.53 621.92 573.11 469.53
143 1.43 0.92 1.24 953.56 615.44 397.22 824.51 532.15 712.92
154 1.61 0.62 1.32 942.19 364.48 141.00 770.21 297.95 629.62
165 1.79 0.27 1.16 1019.18 150.83 22.32 661.89 97.96 429.85
176 1.98 0.14 0.79 1070.23 74.41 5.17 427.87 29.75 171.06
187 2.16 0.57 0.23 1091.06 288.90 76.50 116.81 30.93 12.51
198 2.35 1.03 0.46 1078.68 472.15 206.67 210.18 92.00 40.95
209 2.54 1.49 1.22 1031.49 607.65 357.97 495.41 291.85 237.94
220 2.72 1.96 2.00 949.38 684.29 493.21 697.70 502.88 512.74
114105.17 43303.49 112513.56 4217.34 54348.47 136827.21
Integrating each of the above term by Simpsons 1/3rd rule
25
and dividing
each of the above terms by g = 9.81 we have the mass matrix as
[M] =
_
_
11631.5 4414.2 429.997
4414.2 11469.27 5540
429.997 5540 13948
_
_
kN-sec
2
/m
Again for stiffness matrix we show the functions f
i
() as hereafter and
applying the expression k
ij
=
E
2
i

2
j
H
4
_
1
0
Iz
(z)
i
(z)
j
dz we have
z f

1
(x) f

2
(x) f

3
(x) f

1
(x) f

1
(x) f

2
(x) f

1
(x) f

2
(x) f

2
(x) f

3
(x) f

1
(x) f

3
(x) f

2
(x) f

3
(x) f

3
(x)
0 2.72 1.96 2.00 3283.58 14833.25 67007.83 42337.64 191256.31 545891.08
11 2.54 1.49 1.22 2445.70 9030.09 33341.16 20605.66 76080.79 173607.84
22 2.35 1.03 0.46 1791.12 4913.99 13481.69 6118.56 16786.46 20901.32
33 2.16 0.57 0.23 1287.15 2136.53 3546.40 2422.39 4020.89 4558.87
44 1.98 0.14 0.79 905.58 395.01 172.30 6355.70 2772.32 44606.72
55 1.79 0.27 1.16 622.12 576.65 534.50 7091.03 6572.75 80825.06
(continued)
25 I = h/3[(y0 +4(y1 +y3 +y5) + +yn 1 +2)(y2 +y4 + +yn 2) +yn]
z f

1
(x) f

2
(x) f

3
(x) f

1
(x) f

1
(x) f

2
(x) f

1
(x) f

2
(x) f

2
(x) f

3
(x) f

1
(x) f

3
(x) f

2
(x) f

3
(x) f

3
(x)
66 1.61 0.62 1.32 416.04 1008.36 2443.96 5968.29 14465.37 85618.10
77 1.43 0.92 1.24 269.86 1091.32 4413.37 4094.23 16557.31 62116.76
88 1.26 1.16 0.95 169.03 976.03 5635.75 2238.81 12927.29 29652.64
99 1.09 1.32 0.50 101.70 772.62 5869.80 818.80 6220.62 6592.42
110 0.93 1.40 0.04 58.37 554.21 5261.92 43.83 416.12 32.91
121 0.77 1.41 0.58 31.69 362.68 4150.99 421.35 4822.53 5602.72
132 0.63 1.34 1.05 16.08 215.99 2900.50 474.44 6371.21 13994.95
143 0.49 1.21 1.38 7.51 115.90 1787.65 368.52 5684.07 18073.30
154 0.37 1.03 1.51 3.16 55.01 957.77 225.67 3929.12 16118.72
165 0.27 0.82 1.45 1.16 22.39 433.67 110.94 2148.97 10648.68
176 0.17 0.59 1.21 0.35 7.41 157.71 42.45 903.64 5177.64
187 0.10 0.37 0.85 0.08 1.81 41.78 11.59 268.21 1721.86
198 0.05 0.18 0.46 0.01 0.26 6.48 1.83 45.51 319.55
209 0.01 0.05 0.13 0.00 0.01 0.30 0.08 2.23 16.79
220 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
1.1961 2.3089 1.3445 2.4450 2.3870 8.9746
10
+05
10
+05
10
+06
10
+05
10
+06
10
+06
Integrating each of the above term numerically by Simpsons 1/3rd rule we
have stiffness matrix as
[K] =
_
_
_
_
_
1.1961 10
5
2.3089 10
5
2.445 10
5
2.3089 10
5
1.3445 10
6
2.387 10
6
2.445 10
5
2.387 10
6
8.9746 10
6
_
_
Performing the eigen value analysis by any of the methods as shown in
Chapter 5 (Vol. 1) we have
(rad/sec) T(sec)
6.382 2.53 2.48
76.771 8.76 0.717
664.385 25.77 0.244
Thus for T = 2.48 sec we have S
a
/g = 0.548, for T = 0.717; S
a
/g = 1.896
and for T = 0.244 S
a
/g = 2.5.
The corresponding eigen vectors are given by
0.967 0.165 0.061
0.25 0.958 0.227
0.052 0.236 0.972
Now applying the expression w
i
(z, t) =
i
ZI
2R
S
a
2
[
ii
]F
i
(z) for the first three
modes and performing an SRSS we have the deflection as
Z d
1
(meter) d
2
(meter) d
3
(meter) d
(comb)
0 0 0 0 0
11 0.000251703 0.00013216 0.00012267 0.000309628
22 0.000966707 0.0005533 0.00041479 0.001188575
33 0.002091173 0.00129389 0.00076611 0.002575668
44 0.003590126 0.00236493 0.00107735 0.004431993
55 0.005461638 0.00374318 0.00126895 0.006741745
66 0.007748398 0.00535884 0.00128912 0.00950877
77 0.010542958 0.00708952 0.0011191 0.012754123
88 0.013984374 0.00876281 0.00077447 0.016521178
99 0.018245647 0.01016819 0.00030182 0.020889875
110 0.023513101 0.01107722 0.00022885 0.02599275
121 0.029960377 0.01126932 0.00073501 0.03201815
132 0.037720885 0.01055956 0.00113413 0.039187443
143 0.046863104 0.00882398 0.00135609 0.047705891
154 0.057373044 0.00601826 0.00135336 0.057703702
165 0.069147478 0.00218613 0.00110751 0.069190892
176 0.082000282 0.00254487 0.00063079 0.082042187
187 0.095682647 0.00798104 3.7726E-05 0.096014933
198 0.109916343 0.01389279 0.00084109 0.110794041
209 0.124437733 0.02005545 0.00171779 0.126055232
220 0.139049403 0.02629749 0.00261821 0.141538509
Again applying the expressions
M
i
z
=
i
EI
z
S
ai
2
i
H
2
3
i=1
2
i
F
i
(z)
z M
1
(kN m) M
2
(kN m) M
3
(kN m) M
(comb)
0 3.6310
+05
1.7410
+05
1.9610
+05
448077.79
11 2.7510
+05
1.7210
+05
1.0110
+05
340170.59
22 2.0710
+05
1.6310
+05
2.9610
+04
265471.40
33 1.6010
+05
1.4410
+05
1.8010
+04
215984.91
44 1.3410
+05
1.1310
+05
4.4310
+04
180701.73
55 1.2510
+05
7.3410
+04
5.2910
+04
154065.57
66 1.2710
+05
3.0010
+04
4.8610
+04
138948.07
77 1.3310
+05
1.1210
+04
3.6610
+04
138323.60
88 1.3710
+05
4.4710
+04
2.1810
+04
146038.98
99 1.3610
+05
6.7010
+04
7.9610
+03
151530.17
110 1.2610
+05
7.6910
+04
2.6210
+03
147996.14
121 1.1010
+05
7.5710
+04
9.0310
+03
133959.24
132 8.9210
+04
6.6310
+04
1.1510
+04
111706.12
143 6.6610
+04
5.2210
+04
1.1010
+04
85366.09
154 4.5510
+04
3.7010
+04
8.7910
+03
59278.77
165 2.7910
+04
2.3310
+04
6.0110
+03
36847.91
176 1.5010
+04
1.2710
+04
3.4810
+03
19963.65
187 6.6710
+03
5.7610
+03
1.6410
+03
8961.31
198 2.2010
+03
1.9210
+03
5.6510
+02
2979.07
209 3.8110
+02
3.3510
+02
1.0110
+02
517.31
220 0.0010
+00
0.0010
+00
0.0010
+00
0.00
Considering
V
i
z
=
i
EI
z
S
ai
2
i
H
3
3
i=1
3
i
F
i
(z)
z V
1
(kN) V
2
(kN) V
3
(kN) V
(comb)
0 3878.895 2420.965 7221.983 8.547710
+03
11 3227.414 1972.271 6144.799 7.215610
+03
22 2217.082 1129.664 4938.797 5.530210
+03
33 859.165 97.013 3591.885 3.694510
+03
44 597.419 41453.868 2226.129 2.725110
+03
55 1856.457 2636.536 997.189 3.375210
+03
66 2693.050 3411.105 30.551 4.346210
+03
77 3008.120 3670.259 609.604 4.784510
+03
88 2831.386 3437.311 925.257 4.548410
+03
99 2289.459 2834.863 971.298 3.771110
+03
110 1557.132 2036.228 832.931 2.695310
+03
121 808.201 1216.009 601.699 1.579210
+03
132 177.805 511.899 356.071 6.484210
+02
143 257.491 4.045 149.873 2.979610
+02
154 481.345 287.389 9.173 5.606910
+02
165 525.028 390.325 63.902 6.573410
+02
176 447.041 360.932 83.191 5.805510
+02
187 311.860 262.670 69.421 4.136110
+02
198 173.245 149.620 42.549 2.328310
+02
209 65.076 57.113 16.980 8.823310
+01
220 0.008 0.011 0.007 1.492310
02
The Bending moment diagram is shown Figure 3.3.12.
Bending Moment Diagram
-3.00E+05
-2.00E+05
-1.00E+05
0.00E+00
1.00E+05
2.00E+05
3.00E+05
4.00E+05
5.00E+05
0 2
2
4
4
6
6
8
8
1
1
0
1
3
2
1
5
4
1
7
6
1
9
8
2
2
0
Height z(m)
B
e
n
d
i
n
g

m
o
m
e
n
t
(
k
N
-
m
)
M1
M2
M3
M(comb)
Figure 3.3.12 Bending Moment diagram of tapered chimney.
3.3.1.5 Computer analysis of tall chimneys
Many corporate houses have developed in-house computer program for analysis and
design of tall chimneys based on IS-4998 or ACI 307, CICINDetc. The seismic analysis
part however can also be done in generic finite element commercially available package
like GTSTRUDL, SAP 2000, STAAD PRO etc.
For dynamic analysis of tall chimneys normally a stick model with masses lumped
at convenient nodes sufce. The structural element constitutes of beam elements
with the mass of the shell and brick lining lumped at each end nodes i and j.
The computer assembles the stiffness matrix based on the principle of finite ele-
ment [K] =
_
[B]
T
D[B]dz and forms the lumped mass matrix [M] which is diagonal
in nature. On formation of these elements it performs the eigen value analysis based
on the expression [K][] = [M][]
2
= 0 and then perform the modal analysis based
on response spectrum method as explained earlier.
3.3.1.6 Discussion on factors affecting the dynamic analysis
of tall chimneys
The major factors which affect the dynamic response of tall chimneys under earthquake
are the code factors
Z (zone factor)
I (Importance factor)
R (Ductility factor)
While IS code recommends the value of Z for different zones, the importance factor
for chimney considering its slenderness requires special consideration while the usual
practice is to apply a factor of 1.5, however for zones which a more susceptible to earth-
quake (like zone IV and V) it is recommended (Wilson 2003) that importance factor
considered be 2.0. For structures of category 2(RCC Chimney) IS code recommend an
Importance factor of 1.75.
CICIND recommends ductility factor R = 1 for non ductile detailing and R = 2
when ductile detailing is to be adopted. IS-code recommends a value of R = 3.0 for
RCC Chimneys.
3.3.1.7 Do we consider soil structure interaction
for dynamic analysis of chimney?
This is a question which has plagued many engineers undertaking the task of design of
tall chimneys. While research papers (Ghosh and Batavyal 1985, Navarro 1992, Luco
1986) and code do recommend considering this, some assessments need to be made
whether it has any value addition in undertaking this complicated task.
Most of the tall chimneys are structurally flexible in nature and in all possibility have
its fixed base time period which would induce a base acceleration of the type k/T where
k varies with the nature of soil. Considering the soil as equivalent springs based on
Richart/Wolfs formulation and correcting the time for soil-structure interaction based
on the expression
T = T
_
1 +
k
K
x
_
1 +
Kx
h
2
K
_
(3.3.66)
where,

T = modied time period of the structure due to the soil stiffness; T =
time period of the fixed base structure;

k = stiffness of the fixed base structure @
4
2
W/(gT
2
); K
x
, K
= horizontal and rotational spring constant of the soil

26
;

h =
effective height or inertial centroid of the system, and W =total weight of the structure.
It will be observed that in most of the cases the time period will further prolong
and which would reduce the value of Sa/g as given in code. While one may feel happy
that it would give a more economic design considering the attenuation of response,
however is not true in all cases specially for chimney like structures.
Firstly for such flexible structures vibrating during earthquake the acceleration at
the top portion of the chimney will be subjected to much higher acceleration then the
ground acceleration input we furnish in the analysis. Only if we do a time history
analysis it will be observed that the acceleration at top is indeed much more than the
input base acceleration.
Thus forces in reality could be more at top portion then what we observe considering
soil-structure interaction and further reducing the design moments and shears may not
always be a safe decision, even with the soil damping attenuating the responses further
at the higher mode.
UBC 97 tries to cater to this phenomenon by a provision of a ctitious force
Ve = 0.07 V.T for time period greater than 0.7 seconds. As IS code does not have this
provision, considering soil compliance may under rate the response on the top portion
of the structure. Moreover as the ductility-design for these types of chimneys is still not
well defined, it would perhaps be preferable to design it as a fixed base structure and
render a conservative design. Unless the structure is itself so rigid that one is reason-
ably sure that considering soil structure interaction can amplify the response instead
of attenuation.
It has however been observed that soil-structure interaction analysis plays a
critical role in aerodynamic response of such chimneys especially the along wind
response which shows amplication while considering the foundation compliance
effect (Sadhegpour & Chowdhury 2008).
3.3.1.8 Thus to sum it up. . . .
A chimney should preferably be designed as a fixed base cantilever.
Minimum three modes should be considered for dynamic analysis since the higher
mode at times can give higher response in terms of shears and moments.
If by three modes at least 90% mass participation is not there higher mode
participation needs to be considered.
26 Refer Chapter 2 (Vol. 2) for the expressions of Kx and K
.
Consider damping ratio somewhere between 25%.
As the frequencies in most cases are widely spaced SRSS method of modal
combination would sufce.
If ground acceleration spectra for the particular site is available on may undertake
a time history analysis, when soil structure interaction can be considered.
In case codal spectrum is used for design, fixed base analysis is always preferable.
3.4 ANALYSIS OF CONCRETE GRAVITY DAMS
3.4.1 Earthquake analysis of concrete dam
The topic of engineering of concrete dam can itself be a subject of a book. For dams
only of very large magnitude are built with concrete and obviously requires a detailed
analysis as the failure risk is far too catastrophic in nature. However based on the state
of art as present till date no concrete dam in the world has yet undergone a failure
(except some few cracks) due to earthquake though many of them has faced seismic
force as high as 1.0 g.
While we are writing this section, IS-1893 (2002) code is yet to come up with
procedural practice for earthquake analysis of concrete dams. As such, our discussion
and benchmarking herein will be based on IS-1893 (1984) which has been in practice
for last 20 years and based on which a number of dams have been built.
3.4.1.1 The IS-code method
Shown in Figure 3.4.1 is a typical concrete dam with water upstream. During an
earthquake it is usually assumed that the dam vibrates on its own thus generating
a force (W/g) a due to its inertia. It is also subjected to a hydrodynamic force due
to propagation of waves through the liquid see figure 3.4.1 that produces additional
force on the dam face over and above the hydrostatic force it sustains.
Water level
z
a
g
W
Hydrodynamic
pressure
W
Figure 3.4.1 Typical concrete dam with water in upstream.
The earthquake dynamic analysis for a dam is based on certain simplifying
assumptions as stated hereunder
The water the dam supports is incompressible.
The dam is rigid and has the same motion throughout its body.
The dam and the reservoir motion is uncoupled and the interaction between them
is negligible for all practical purpose
27
.
The code has also recommended to adapt seismic coefficient method for damup to 100
meter high while response spectrum method with dynamic analysis (i.e. calculation of
time period and taking its effect) for dams that are greater than 100 m.
The seismic coefficient method is quite straightforward and does not require any
elaboration as the steps are same for building analysis except the fact that importance
factor to be considered should be as recommended by the code (for dams I = 3.0).
For response spectrum analysis IS-1893 (1984) code recommends to derive the
fundamental time period of concrete dam as
T = 5.55
H
2
B
_

c
gE
s
(3.4.1)
where, B = width of the dam base in m; H = height of the dam in m;
c
= unit weight
of material of dam in kg/m
3
; g = acceleration due to gravity @ 9.81 m/sec
2
, and, E
s
=
modulus of elasticity of the material of the dam in kg/m
2
.
However, the basis of derivation of this formula has not been elaborated either in
the code or in its explanatory manual
28
.
Looking at the formula it appears that Eqn. (3.4.1) is derived by applying the
Rayleigh Ritz method to some assumed shape functions and considering the dam as a
cantilever beam having varying cross section. The code also does not give any value
of time periods for higher modes whose effects are perhaps considered as non-critical.
The base shear and moment for the dam is given by the expression
V
B
= 0.6
h
W; M
B
= 0.9
h
W
h (3.4.2)
where
h
= IF
0
S
a
g
(3.4.3)
the notations in Eqn. (3.4.3) are as explained earlier in the section where we have
discussed on the code IS-1893 (1984).
27 This may not be true in all cases and could have significant effect. We will study this later on.
28 Except for note that this has been developed based on some research work carried out at University of
Roorkee India.
As per latest version of the code the seismic design coefficient can be taken as
h
=
ZI
2R
S
a
g
(3.4.4)
where, W = weight of the concrete dam;

h = height of the center of gravity of the
dam, above its base.
The value S
a
/g is read from the chart and is a function of the time period of the dam
as mentioned in Equation (3.4.1).
For any horizontal section at a depth z below the top of dam the shear force V
z
and
bending moment M
z
may be obtained from the expression
V
z
= C
v
V
B
and M
z
= C
M
M
B
(3.4.5)
where the values of Coefficients C
V
and C
M
are as furnished in figure below.
Figure 3.4.2 Values of C
v
and C
m
along the height of dam.
Considering a concrete dam as massive where the weight plays a significant part in
its stability, it is evident that unlike other structures, the vertical mode of earthquake
acceleration plays a significant part in its stability and cannot be ignored.
As per IS code based on response spectrum the force due to vertical acceleration is
considered as 0.75 times the value of
h
at the top of the dam and reducing linearly
to zero at base.
3.4.1.2 Hydrodynamic pressure on dam from the reservoir
As wave propagates through the ground there is an instantaneous hydrodynamic
pressure (or suction) exerted on the dam over and above the hydrostatic pressure
it sustains.
0
0.2
0.4
0.6
0.8
1
1.2
0
0
.
1
0
.
2
0
.
3
0
.
4
0
.
5
0
.
6
0
.
7
0
.
8
0
.
9 1
z/H
V
a
l
u
e

o
f

C
s
Figure 3.4.3 Value of C
s
with depth.
Based on the assumption that the water is incompressible the hydrodynamic pressure
at any depth z below the reservoir is determined by the expression
p = C
s
w
h (3.4.6)
where, p = hydrodynamic pressure of water; C
s
= a coefficient which varies with
shape and depth;
h
= seismic design coefficient as explained in Equation 3.4.4.
w
=
unit weight of water, and h = height of water in the reservoir.
The variation of C
s
with depth is given by the expression
C
s
=
C
m
2
_
z
h
_
2
z
h
_
+
_
z
h
_
2
z
h
_
_
(3.4.7)
where C
m
value varies (almost) linearly from a value of 0.735, when the vertical
upstream angle of the dam face varies from 0 degree (i.e. perfectly vertical), to a value
of 0.0, when this angle is 90 degree.
The variation of C
s
with depth is shown in Figure 3.4.3. The values in graph
multiplied by the value C
m
will give the value C
s
.
The approximate values of Shear and Moment at depth z below the free surface is
given by the expression
V
z
= 0.726 pz and M
z
= 0.299 pz
2
. (3.4.8)
where, p = hydrodynamic pressure is given in Equation (3.4.6).
3.4.1.3 Some comments and review of the IS-code method
The above method though has been in practice for quite some time, yet there are some
approximation and one assumption that is perhaps not conceptually correct.
Firstly, the value of time period as proposed in Equation (3.4.1) over estimates the
time period of a dam by about 25% and is not unique. The top width at crest has a
significant effect on the time period
29
.
Secondly, while calculating the hydrodynamic pressure vide Equation (3.4.6) code
uses the value of
h
as obtained based on Equation (3.4.1).
This is in violation to the basic assumption made at the outset that the dam face
is rigid and the vibration of the two systems (dam and the fluid in the reservoir) is
uncoupled/independent.
In this case as per the above mentioned assumption the free field time period of
the fluid in the reservoir (assumed tending to infinity) in horizontal direction should
govern the value of
h
g rather than the time period of the dam whose stiffness is
considered as infinite in comparison to the fluid. The present assumption could lead
to a significant variation in end results in some cases.
3.4.2 A method for dynamic analysis of concrete dam
We present (Chowdhury & Dasgupta 2008) here a method wherein we have tried to
overcome a number of deficiencies as mentioned in the above IS-code method. The
salient feature of the method are summarized as hereafter
1 Calculation of time period of the damhaving varying cross section for three modes
and studying the effect of the varying section on the time period.
2 Calculating the free field time period of the fluid in the reservoir and estimating
the hydrodynamic pressure assuming the dam wall to be perfectly rigid when the
two systems are not coupled.
3 A practical simplied approach considering fluid-structure interaction (for funda-
mental mode analysis) wherein we study the effect of hydrodynamic pressure on
the wall vis--vis the response of the dam when the dam is considered to have a
finite stiffness and the fluid is considered compressible.
4 How to model the fluid stiffness and mass when we carry out a finite element
analysis of the dam section considering fluid structure interaction.
3.4.2.1 Calculation of time period of the dam
having variable cross section
As shown in Figure 3.4.4, we show a dam of height H having base width B and top
width as B
t
supporting water of height H
w
. The dam is assumed to be resting on firm
ground (usually rock) and is considered to be fixed at base. Thus for mathematical
analysis the dam is considered as a cantilever flexural member (fixed at base free at
top) having varying width B
z
where this variation is considered as linear with respect
to H.
29 This we are going to study in some detail subsequently.
B
t
H Hw
Z
B X
a =
Figure 3.4.4 A dam supporting water in reservoir.
As derived in the case of the chimney, solving the fourth order differential equation
and differentiating the potential and kinetic energy of the system, the stiffness of the
dam can be expressed as
k
ij
=
H
_
0
EI(z)
d
2
i
(z)
dz
2
d
2
j
(z)
dz
2
dz (3.4.9)
for i, j = 1, 2, 3. . . . . . n and the mass matrix is expressed as
m
ij
=
_
_
c
A
z
g
H
_
0
i
(z)
j
(z)dz
_
_
. (3.4.10)
For a cantilever beam the fourth order differential equation is
EI
_
4
w
z
4
_
+ A
_
2
w
t
2
_
= 0 (3.4.11)
where, w(z) =
i
(z)q(t), and
i
= sin
i
z
H
sin h
i
z
H

i
_
cos
i
z
H
cos h
i
z
H
_
(3.4.12)
in which,
i
= 1.875, 4.694, 7.855,
2m1
2
for i = 1, 2, 3. . . . . . m, and
i
=
sin
i
+sin h
i
cos
i
+cos h
i
(3.4.13)
Since the Moment of inertia and the area of the dam vary with depth. At any height
z from the bottom the moment of inertia at any height z is expressed as
I
z
= I
0
_
1 +
z
H
_
3
(3.4.14)
where, I
0
= moment of inertia of the dam at base, and
=
B
t
B
B
. (3.4.15)
Similarly area at any height z is given by
A
z
= A
0
_
1 +
z
H
_
(3.4.16)
Based on above the stiffness and mass equation gets modified to
k
ij
= EI
0
H
_
0
_
1 +
z
H
_
3
d
2
i
(z)
dz
2
d
2
j
(z)
dz
2
dz (3.4.17)
and m
ij
=
_
_
c
A
0
g
H
_
0
_
1 +
z
H
_
i
(z)
j
(z)dz
_
_
(3.4.18)
The double derivative of Equation (3.4.12) is given by
i
=

2
i
H
2
_
sin
i
z
H
sin h
i
z
H
+
i
_
cos
i
z
H
+cos h
i
z
H
__
and (3.4.19)
To evaluate the stiffness and mass matrix in generic form by integration we change
the above to generalized co-ordinate by considering;
=
z
H
when d =
dz
H
and as z 0, 0 and as z H, 1
based on above we can now express the double derivative as
F
()
i
=

2
i
H
2
_
sin
i
sin h
i
+
i
(cos
i
+cos h
i
)
_
(3.4.20)
=

2
i
H
2
f
()
i
(say) (3.4.21)
k
ij
=
EI
0
2
i

2
j
H
3
1
_
0
(1 + )
3
f
()
i
f
()
j
d (3.4.22)
and mass of the system is given by
m
ij
=
A
0
H
g
1
_
0
(1 + ) f ()
i
f ()
j
d; where, i = j = 1, 2, 3, . . . . . . . . . , m
(3.4.23)
Thus, for the first three modes, the stiffness matrix is given by
[K]
ij
=
EI
H
3
_
_
_
_
_
_
_
_
_
4
1
1
_
0
(1 + )
3
f
()
2
1
d Symmetrical
2
2
2
1
1
_
0
(1 + )
3
f
()
2
f
()
1
d
4
2
1
_
0
(1 + )
3
f
2
()
2
d
2
3
2
1
1
_
0
(1 + )
3
f
()
3
f
()
1
d
2
3
2
2
1
_
0
(1 + )
3
f
()
3
f
()
2
d
4
3
1
_
0
(1 + )
3
f
()
2
3
d
_
_
(3.4.24)
and the mass matrix
30
is given by
[M]
ij
=
AH
g
_
_
_
_
_
_
_
_
_
1
_
0
(1 + ) f ()
2
1
d
1
_
0
(1 + ) f ()
2
f ()
1
d
1
_
0
(1 + ) f
2
()
2
d
1
_
0
(1 + ) f ()
3
f ()
1
d
1
_
0
(1 + ) f ()
3
f ()
2
d
1
_
0
(1 + ) f ()
2
3
d
_
_
(3.4.25)
It is apparent that the values of the stiffness and mass matrix are dependent on the
parameter , which would surely influence these values. As an example we solve the
above for B
t
/B = 0.1 (which is the most standard for concrete dams of about 100 m
high) i.e. the top width is 10% of base width for some space is normally kept at the
top of the dam for motor and pedestrian access, maintenance and inspection.
Thus, considering = 0.9, we have based on numerical integration between 1 to 0.
30 The matrix is symmetric about is diagonal.
The stiffness matrix is
[K]
33
=
EI
0
H
3
_
_
13.965 24.495 23.911
24.495 165.575 289.015
23.911 289.015 1.167 10
3
_
_
The mass matrix is given by
[M]
33
=

c
A
0
H
g
_
_
0.509 0.185 0.025
0.185 0.449 0.169
0.025 0.169 0.522
_
_
generalized Jacobi technique, we have
[] =
EI
0
g
c
A
0
H
4
_
_
21.62866 0 0
0 255.8406 0
0 0 2288.56
_
_
and the corresponding eigen-
vectors are given as
[] =
_
_
0.986 0.242 0.127
0.167 0.945 0.169
0.020 0.218 0.977
_
_
T
_
_
_
f
1
()
f
2
()
f
3
()
_
_
_
, since [] =
2
and T =
2
, we
have,
[T] =
_
_
1.3509 0 0
0 0.3928 0
0 0.13135
_
_
_
c
A
0
H
4
EI
0
g
Now, considering width of the dam as 1 m in the Y direction (perpendicular to the
plane of the paper),
A
0
= B 1m
2
and I
0
= 1
B
3
12
m
4
.
Substituting the above, we have for the first three modes:
Mode number Time period (secs)
1 T
1
= 4.68
H
2
B
_
c
Eg
2 T
2
= 1.36
H
2
B
_
c
Eg
3 T
3
= 0.455
H
2
B
_
c
Eg
The above can be generalized to an expression, T
1
= C
T
H
2
B
_
c
Eg
.
We present here the values of C
T
for the first three modes for various values of B
t
/B
Natural frequency Time period factor (C
T
) in
B
t
/B Eigen value (rad/sec) form as presented in code
24.306 4.93011156 4.41482756
0.05 261.863 16.1821816 1.34503449
2348 48.4561658 0.44918107
21.62866 4.65066232 4.68010595
0.1 255.84 15.9949992 1.36077483
2288 47.833043 0.45503257
19.807 4.45050559 4.89058871
0.15 257.77 16.0552172 1.355671
2268 47.6235236 0.45703448
18.485 4.29941857 5.0624502
0.2 264.047 16.2495231 1.33946038
2270 47.644517 0.4568331
17.477 4.1805502 5.20639421
0.25 273.01 16.5230143 1.31728945
2289 47.8434949 0.45493316
16.678 4.08387071 5.32964775
0.3 283.7957 16.846237 1.29201509
2322 48.1871352 0.45168886
To substantiate further, we have also compared the fundamental natural frequency
of a tapered cantilever beam of various width ratio (B
t
/B) derived by exact solu-
tion (Karnovsky & Lebed 2001) with the method as proposed above and compared
hereafter.
B
t
/B Proposed analysis Exact solution
0 5.348832 5.3151
0.1 4.650806 4.6307
0.2 4.299419 4.2925
0.3 4.083871 4.0817
3.4.2.1.1 Comparison of natural frequency (rad/sec) proposed
and exact analysis
It is evident from above that the value of time period as given in code is over esti-
mated
31
. For even with B
t
/B as 0.3 (which is very high) the time period coefcient is
5.33 in lieu of 5.55. While for a 10% crest width the error is of the order of (+)19%.
31 Though we agree that this may not have much effect on S
a
/g finally in some cases.
3.4.2.2 Fundamental uncoupled amplitudes of vibration
of the dam
The maximum amplitude of vibration as per modal response analysis is expressed as
S
d
= S
a
/
2
, thus based on provisions of code it may be further written as
S
d
=
i
S
a
2
(3.4.26)
where
i
=
n
i=1
m
i
i
/
n
i=1
m
i
2
i
for the present problem this is expressed as
i
=
1
_
0
(1 + )f
i
()d
_
1
_
0
(1 + ) f
i
()
2
d
For a typical value of B
t
/B = 0.1, the
i
values for the three modes are given as:
1
= 0.726,
2
= 0.771 and
2
= 0.432, =
ZI
2R
, and the code factors remain the
same as explained earlier.
The displacement along the height z is thus expressed as
w(z) =
i
S
a
2
[
11
f
1
(z) +
12
f
2
(z) +
13
f
3
(z)] (3.4.27)
where, f
i
(z) = sin

i
z
H
sin h
i
z
H

i
_
cos

i
z
H
cosh

i
z
H
_
.
Considering, T
1
= C
T
H
2
B
_
c
Eg
in which, C
T
= 4.68, 1.36, 0.455 etc. and also
considering, = 2/T, we have finally,
w(z) =

i
C
2
T
4
2
B
2
E
S
a
g

c
H
4
[
11
f
1
(z) +
12
f
2
(z) +
13
f
3
(z)] (3.4.28)
where for B
t
/B = 0.1, the displacement for fundamental mode may be expressed as
w(z) =
0.403
i
B
2
E
S
a
g

c
H
4
[0.986f
1
(z) 0.167f
2
(z) +0.020f
3
(z)]. (3.4.29)
3.4.2.3 Dynamic Bending Moment and Shear force for the
fundamental mode
For a beam element the bending moment is given by
M(z) = EI
z
d
2
w
dz
2
, (3.4.30)
for the present case, this may be expressed as
M(z) = EI
0
_
1 +
z
H
_
3

i
C
2
T
H
4
c
4
2
B
2
E
_
S
a
g
_
d
2
dz
2
(
11
f
1
(z) +
12
f
2
(z) +
13
f
3
(z))
=
i
C
2
T
48
2
_
S
a
g
_
c
BH
2
_
1 +
z
H
_
3
(
11
2
1
f
1
(z) +
12
2
2
f
2
(z) +
13
2
3
f
3
(z))
(3.4.31)
For B
t
/B = 0.1, Equation (3.4.30) is expressed as
M(z) = 0.0335
_
S
a
g
_
c
BH
2
_
1 +
z
H
_
3
(3.466f
1
(z) 3.679f
2
(z) +1.234f
3
(z)) (3.4.32)
Similarly, the shear force is written as
V(z) =
dM(z)
dz
=

i
C
2
T
48
2
_
S
a
g
_
c
BH
_
1 +
z
H
_
3
(
11
3
1
f
1
(z) +
12
3
2
f
2
(z) +
13
3
3
f
3
(z))
V(z) = 0.0335
i
C
2
T
48
2
_
S
a
g
_
c
BH
_
1 +
z
H
_
3
(6.5f
1
(z) 17.27
12
f
2
(z) +9.693f
3
(z)) (3.4.33)
The values of f
(z) and f
(z) are as given hereafter

Z/H f

1
(z) f

2
(z) f

3
(z) f

1
(Z) f

2
(Z) f

3
(Z)
0 2.72444111 1.96373508 2.001552 2 2 2
0.2 1.9765624 0.13756457 0.790419 1.977702 1.793709 1.20662
0.4 1.2564103 1.1575438 0.948172 1.834738 0.844848 0.966178
0.6 0.62639921 1.3421761 1.052816 1.485855 0.423056 1.034407
0.8 0.17408595 0.5912631 1.209796 0.881407 0.972256 0.79471
1 0 0 0 0.000284 0.000179 0.000486
3.4.2.4 Free f ield time period of the reservoir and the hydrodynamic
pressure from reservoir
In this section we determine the free field time period of the water extending to infinity
in horizontal direction and having a depth H
w
as shown in Figure 3.4.4. As assumed
by the code we presume the dam wall is acting rigidly with respect to the fluid and the
fluid vibration remains uncoupled with respect to the vibration of the dam.
Since the damprofile is considered in two dimensions we start with two dimensional
propagation of wave due to earthquake through the water.
This is expressed as
2
w(x, z, t)
x
2
+

2
w(x, z, t)
z
2
=
1
c
2
2
w(x, z, t)
t
2
(3.4.34)
in which, c =velocity of sound in water and is expressed c =
_
B
m
/
w
, normally taken
as 1439 m/sec; where, B
m
= bulk modulus of water (usually considered as 2.11 10
6
kN/m
2
), and
w
= mass density of water and w = displacement of fluid medium.
We will not solve this problem in detail here since we have already solved an iden-
tical problem having same boundary condition for estimation of dynamic pressure
due to earthquake on a rigid wall later under the topic of earth retaining structures
(Section 3.7.1).
Based on this analysis the fundamental frequency of the reservoir can be expressed
as
1
= c/(2H
w
)(H
w
= height of water in the reservoir).
Considering, T = 2/, we have, T = 4H
w
/c and the eigenvector is expressed as
(z) = cos
(2n1)z
2H
, where n = 1, 2, 3. . .; the number of modes.
Based on modal response technique, the maximum amplitude function can be
defined as
S
d
=
S
a
2
(3.4.35)
where S
d
= maximum displacement; S
a
= acceleration which is the function of time
period 4H
w
/c, and can be read off from normalized response given in the code.
Considering = ZI/2R a code factor depending on earthquake zone (low, moder-
ate, severe etc.), Importance factor, ductility factor etc. (here R may be considered as
1.5 for un-reinforced concrete), we can write
w(z) =
i
S
a
2
(z) (3.4.36)
Now substituting the value of and using B
m
=
w
c
2
, we have
w(z) =
4
S
a
H
2
w
c
2
cos
z
2H
=
4
S
a
w
H
2
w
B
m
g
cos
z
2H
(3.4.37)
where
w
= unit weight of water,
i
= modal mass participation factor =
m
i
m
i
2
i
, and,
g = acceleration due to gravity.
Thus, the modal participation factor can be considered as
i
=
m
i
i
/
m
i
2
i
=
_
_
H
_
0
w
z cos
z
2H
dz
_
_
_
_
_
H
_
0
w
z cos
2
z
2H
dz
_
_
(3.4.38)
The above on integration by parts gives,
i
=
8
+2
Thus, w(z) =
32
2
( +2)
S
a
w
H
2
B
m
g
cos
z
2H
(3.4.39)
The deformation in fluid in z direction is given by [as
z
=
w
z
]:
z
=
16
( +2)
S
a
w
H
w
B
m
g
cos
z
2H
(3.4.40)
Thus dynamic pressure is given by (considering 15% damping for fluid)
p
dyn
= B
m
u
z
=
12
( +2)
S
a
w
H
w
g
sin
z
2H
, (3.4.41)
the negative sign indicates that the pressure is acting on the dam. Here the factor 16
in Equation (3.4.40) is multiplied by a factor 0.75 as per IS-1893 (2002) to cater to
15% damping ratio.
The variation of pressure is as shown in Figure 3.4.5.
Based on above the pressure can be expressed as
p
dyn
= Coeff
S
a
w
H
w
g
(3.4.42)
where the coefficients can be read from the graph in Figure 3.4.5.
0
0.2
0.4
0.6
0.8
1
1.2
0
0
.
2
0
.
4
0
.
6
0
.
8 1
z/H
p
r
e
s
s
u
r
e

c
o
e
f
f
i
c
i
e
n
t
Figure 3.4.5 Variation of hydrodynamic pressure with depth.
Z
dy
dz
dx
Y
X
Figure 3.4.6 An elastic elemental body in space.
Considering V
z
=
_
Hw
0
p
z
dz, the shear force on the dam face is
V
dyn
= 0.645p
z
z, and the moment is given by (3.4.43)
M
dyn
= 0.4014pz z
2
(3.4.44)
where, p
z
= pressure at any point z from top of the dam, and z = distance from top
of the dam.
3.4.2.5 A f luid-structure interaction model for earthquake
analysis of dam
In previous section we had proposed a technique for dynamic analysis where the
vibration is uncoupled. By this we mean the two vibrations are independent of each
other.
The stiffness of one system does not affect the stiffness of the other vis-a-vis the
frequency of each of the system is also mutually exclusive. In reality this is only an
idealization and the fluid and structure do have finite stiffness and vibrate in coupled
mode.
We present a here two practical easy to apply model based on which such fluid
structure interaction can be carried out. There are off course sophisticated finite elem-
ent models available where the fluid and the structure could both be modeled in 2D
for a comprehensive analysis but is not without its lacunae.
Firstly such analysis is quite expensive and requires a significant large model for
problems such as dams. Moreover fluid elements having no shear strength and being
almost an incompressible medium often gives rise to numerical difficulty in arriving
at a meaningful solution. In the present model as proposed herein we use a simplified
lumped massspring based model.
3.4.2.6 Determination of the f luid spring and mass
We show in Figure 3.4.6, an elemental body in space of length dx, dy and dz.
For such a body under load, the strain energy equation as per theory of elasticity in
three dimensions is given by the expression
V =
e
2
2
+G
_
2
x
+
2
y
+
2
z
_
+
G
2
_
2
xy
+
2
yz
+
2
xz
_
(3.4.45)
If the body is a fluid medium then as G = 0, we have, V = (B
m
e
2
)/2 where,
B
m
= bulk modulus of water, and e =
x
+
y
+
z,
thus
V =
B
m
(
x
+
y
+
z
)
2
2
(3.4.46)
For two dimensional case for a reservoir as x in horizontal direction, we have
x
= 0, and hence,
V =
B
m
(
y
+
z
)
2
2
(3.4.47)
Similarly, since we are only considering unit width of the dam in the direction
perpendicular to the paper, the displacement is invariant in this direction which gives
y
= 0 and we are finally left with
V =
B
m
2
z
2
=
B
m
2
_
w
z
_
2
(3.4.48)
Considering, w(z) = (z), q(t) one can have
V
q
r
= B
m
w
z
q
r
_
w
z
_
V
q
r
= B
m
i
z
r
z
q
i
q
r
(3.4.49)
2g
m
f
H
w
2
=
8
K
f
=
Figure 3.4.7 Equivalent stiffness and lumped mass of the fluid element.
where, (z) = generalized shape function with respect to the x and z co-ordinates,
and q(t) = displacement function with respect to time in the generalized co-
ordinate.
From which it can be proved that the stiffness and mass matrix can be written as
K
ir
=
H
w
_
0
B
m
i
z
r
z
dz and M
ir
=

w
g
H
w
_
0
r
dz (3.4.50)
where K = stiffness matrix of the fluid medium; M = mass matrix of the fluid medium;
i and r are different modes 1, 2, 3. . .
K and M for the fundamental mode are given by
K
11
= B
m
Hw
_
0
_
_
z
_
2
_
dz and M
11
=

w
g
Hw
_
0
()
2
dz (3.4.51)
Considering the shape function as given in (z) = cos
(2n1)z
2H
w
and substituting it in
Equation (3.4.50) and by integrating, we have
K
11
=

2
B
m
8H
w
and M
11
=

w
H
w
2g
It may be observed that the unit of the stiffness derived here is kN/m
2
, which means
that the expression gives stiffness per unit area. Thus to determine the total stiffness
of the water in contact with dame face one has to multiply this by the contact area
which in this case is H
w
x1.
This gives K
11
=

2
B
m
8
kN/m per meter width.
Similarly the effective mass is given by M
11
=

s
H
2
w
2g
kN sec
2
/m per meter width.
Considering, T = 2
_
M/K and substituting it in the above expression, one can
arrive at the same expression, T = 4H
w
/c, as was derived earlier.
This shows that the stiffness and mass matrix formulation as represented here is
dimensionally correct.
Based on above as shown in Figure 3.4.7, we have managed to derive an equiva-
lent stiffness and lumped mass of the fluid contained by the dam whose fundamental
frequency matches with the free field time period of the fluid continuum.
3.4.2.7 A semi-analytic model for f luid structure interaction
The spring element derived above is obtained per unit area this means this spring is
distributed on the surface of the dam.
The potential energy d of an element of depth dz, of the dam section and the fluid
shown in Figure 3.4.3, is then given by
d =
E
c
I
z
2
_
d
2
w
dz
2
_
2
+
K
f
2
w
2
(3.4.52)
where, E
c
= Youngs modulus of the concrete dam; I
z
= moment of inertia of dam
expressed
B
3
12
(1 +
z
H
)
3
K
f
= fluid stiffness; w = displacement of the fluid dam
system and may be written as [(z)q(t)].
The total potential energy over the height (H) of the dam is then given by
=
E
c
I
0
2
H
_
0
_
1 +
z
H
_
3
_
d
2
v
dz
2
_
2
dz +

2
B
m
16H
w
H
w
_
0
v
2
dz (3.4.53)
Considering w(z, t) = (z)q(t), the stiffness expression becomes
K
ij
= E
c
I
o
H
_
0
_
1 +
z
H
_
3
i
(z)
j
(z)dz +

2
B
m
8H
w
H
w
_
0
i
(z)
j
(z)dz (3.4.54)
Thus, for the fundamental mode, Equation (3.4.54) may be written as
K = E
c
I
o
H
_
0
_
1 +
z
H
_
3
(z)
2
dz +

2
B
m
8H
w
H
w
_
0
(z)
2
dz (3.4.55)
It has been shown by Timoshenko that when a beam is supported by distributed
spring it only affects the natural frequency while the mode shape remains unaltered
with respect to that the original beam (i.e. in air), as such like in the case of the
uncoupled dynamic analysis carried out previously (where effect of water was ignored)
we consider
i
= sin
i
z
H
sin
i
z
H

i
_
cos
i
z
H
cos h
i
z
H
_
(3.4.56)
where,
i
= 1.875, 4.694, 7.855,
2m1
2
; for i = 1, 2, 3. . .. . . m.
Thus, in natural co-ordinates, the stiffness expression can be expressed as
K =
E
c
I
0
4
1
H
3
1
_
0
(1 + )
3
f
()
2
d +

2
B
m
8
1
_
0
f ()
2
d (3.4.57)
The above can be simplified to
K =
E
c
I
0
4
1
H
3
_
_
1
_
0
(1 + )
3
f
()
2
d +
s
f ()
2
d
_
_
(3.4.58)
where,
s
= 1.2
_
B
m
E
c
_ _
H
B
_
, a constant number.
Thus, K may be written as
K =
E
c
I
0
4
1
H
3
[I
1
+
s
I
2
] (3.4.59)
where, I
1
=
_
1
0
(1 + )f
()
2
d and I
2
=
_
1
0
f ()
2
d.
Similarly the mass coefficient based on kinetic energy principle can be as expressed as
M =

c
A
0
g
H
_
0
_
1 +
z
H
_
(z)
2
dz +

w
H
w
2g
H
w
_
0
(z)
2
dz (3.4.60)
The above in natural co-ordinate can now be expressed as
M =

c
A
0
H
g
1
_
0
(1 + ) f ()
2
d +

w
H
2
w
2g
1
_
0
f ()
2
d (3.4.61)
This can be further expressed as
M =

c
A
0
H
g
_
_
1
_
0
(1 + ) f ()
2
d +
m
1
_
0
f ()
2
d
_
_
(3.4.62)
where,
m
=
1
5
_
H
w
H
_
2
_
H
B
_
, a constant number.
Thus, M may be expressed as
M
=
c
A
0
H
g
[I
3
+
m
I
2
] ; where, I
3
=
1
_
0
(1 + )f ()
2
d (3.4.63)
Now considering the expression, T = 2
_
M
K
and substituting it in the above
expressions and after some simplication we have,
T = 6.19
H
2
B
_
c
C
FS
E
c
g
(3.4.64)
where, C
FS
is fluid-structure interaction coefficient given by C
FS
=
I
2
m
+I
3
I
2
s
+I
1
.
The amplitude of displacement is given by
w(z)
max
=
i
S
a
2
f
1
()
max
at z = H. (3.4.65)
where
i
=
i=1
m
i
i
/
i=1
m
i
2
i
, for this particular case, the expression becomes
i
=
c
A
0
H
g
_
1
0
(1 + )f ()d +

w
H
2
w
2g
_
1
0
f ()d
c
A
0
H
g
_
1
0
(1 + ) f ()
2
+

w
H
2
w
2g
_
1
0
f ()
2
=
I
4
+
m
I
5
I
3
+
m
I
2
(3.4.66)
Based on this the maximum displacement becomes
w(z)
max
= 1.01175
i
c
C
FS
B
2
E
c
_
S
a
g
_
H
4
(3.4.67)
The moment and shear are thus given by
M = EI
d
2
w
dz
2
and V = EI
d
3
w
dz
3
(3.4.68)
Hence, M
z
= 0.2964(1 + )
3
i
C
FS
BH
2
(S
a
/g)f
() (3.4.69)
V
z
= 0.557 (1 + )
3
i
C
FS
BH(S
a
/g)f
() (3.4.70)
The values of f
() and f
()are given Table 3.4.1.

The integral values I
1
to I
5
are as given hereafter for various B
t
/B ratio are given in
Table 3.4.2.
Table 3.4.1
Z/H Moment coefficients Shear coefficient
0 1 1
0.2 0.725493 0.988851
0.4 0.461163 0.917369
0.6 0.229918 0.742928
0.8 0.063898 0.440703
1 0 0
Table 3.4.2 Values of integral constants for various B
t
/B ratio.
Bt/B I
1
I
2
I
3
I
4
I
5
0.05 1.102 1.855 0.434 0.33 1.067
0.1 1.13 1.855 0.509 0.369 1.067
0.15 1.159 1.855 0.583 0.408 1.067
0.2 1.189 1.855 0.658 0.447 1.067
0.25 1.221 1.855 0.733 0.485 1.067
0.3 1.254 1.855 0.808 0.524 1.067
Table 3.4.3 Values stiffness (s) and mass (m) parameter for dam.
H/B 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5
s
0.300 0.244 0.195 0.154 0.118 0.089 0.065 0.045 0.030 0.019 0.011
m
0.293 0.273 0.254 0.234 0.215 0.195 0.176 0.156 0.137 0.117 0.098
Table 3.4.4 Values of fluid structure influence coefficient C
FS
.
B
t
/B 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5
0.05 0.589 0.605 0.618 0.626 0.630 0.628 0.621 0.610 0.593 0.572 0.548
0.1 0.624 0.642 0.656 0.667 0.672 0.672 0.668 0.657 0.642 0.623 0.600
0.15 0.656 0.676 0.692 0.704 0.712 0.714 0.710 0.702 0.688 0.670 0.648
0.2 0.688 0.709 0.727 0.741 0.750 0.753 0.751 0.744 0.732 0.715 0.694
0.25 0.718 0.741 0.760 0.775 0.785 0.790 0.789 0.783 0.772 0.756 0.736
0.3 0.746 0.770 0.791 0.807 0.818 0.824 0.825 0.820 0.810 0.795 0.776
Values of
s
and
m
are given in Table 3.4.3 for various H/B ratio. The coefficient
C
FS
the fluid structure coefficient is given in Table 3.4.4 for varying B
t
/B and H/B
ratio. The modal mass participation
i
factors are as given in Table 3.4.5 for various
H/B and B
t
/B ratio.
With the above tables now available it becomes quite simple to analyze the dam for
fluid structure interaction under earthquake force.
Intermediate values can be linearly interpolated without any significant error as the
variations as observed are not significant. The analysis can well be carried out by a
simple calculator or can very well be programmed in a spread sheet.
Table 3.4.5 Values of modal mass participation factor
i
.
B
t
/B 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5
0.05 0.657 0.661 0.664 0.668 0.672 0.676 0.681 0.686 0.692 0.699 0.706
0.1 0.648 0.650 0.653 0.656 0.659 0.663 0.667 0.671 0.675 0.680 0.686
0.15 0.640 0.642 0.644 0.647 0.649 0.652 0.655 0.658 0.662 0.666 0.670
0.2 0.632 0.634 0.636 0.638 0.640 0.642 0.645 0.648 0.650 0.654 0.657
0.25 0.625 0.626 0.628 0.630 0.631 0.633 0.635 0.637 0.639 0.642 0.645
0.3 0.619 0.620 0.622 0.623 0.624 0.626 0.627 0.629 0.631 0.633 0.635
A word regarding damping ratio of the system based on which S
a
/g is to be chosen.
As per ICOLD (International Committee Of Large Dams) the damping values of the
concrete in a dam is low and would be around 25%.While based on study by New-
mark and Roesenbleuth and also by Zangar (1953) water is usually found to have a
damping in the range of 15%. Thus as conservative estimate a value of 810% may
be chosen while selecting the value of S
a
/g for a coupled model.
3.4.2.8 Lumped mass and stiffness approach for Fluid-Structure
interaction
We nowpresent herein a lumped mass model for the same analysis where the fluid and
the dam are coupled together. The reasons for choosing this model are as mentioned
hereunder:
Engineers working with structural dynamics are quite conversant with this
approach.
The model is simple and easy to work with in practical design situation.
It is particularly advantageous to use this model when over and above the fluid the
foundation interaction is also to be considered (soil-fluid-structure interaction).
For large dams it has been particularly observed that considering fluid as an incom-
pressible medium and performing the analysis separately (as suggested by codes of
many countries including IS) the forces at times are in considerable variation to the
case when soil is considered in the analysis.
As such at least for economic reasons ICOLD recommends to use soil-fluid struc-
ture interaction wherever deemed feasible or perform a comprehensive Finite Element
analysis. As a prelude to such analysis the present model would give a very good esti-
mate on how the loads are affecting the dam and whether it is worth such elaborate
analysis. In many of the cases the present analysis would sufce for a comprehen-
sive interaction analysis where further sophistication may not be necessary. We have
already established previously that for a fluid medium extending to innity in hori-
zontal direction of height H
w
the equivalent spring stiffness and lumped mass can be
expressed as K
f
=
2
B
m
/8 and M
f
=
w
H
2
w
/2g, respectively. To couple this with the
dam structure the most intuitive model would be to couple directly fluid stiffness to
the dam stiffness by static coupling which we are so conversant with. If we however
do this we would actually be grossly wrong!!
A B
Ra=2W/3 L/3 2L/3 Rb=W/3
W
Figure 3.4.8 A simple supported beam supporting triangular load.
u
f
6g
H
m
f
2
w w
=
8
B
K
f
m
2
=
u
s
3
I
g
BH
m
s
=
1
3
4
1
3
I
H 12
EB
K
s
=
Figure 3.4.9 Mathematical model for fluid structure interaction.
Housner et al. have shown that when an infinite fluid vibrates, a part of the fluid
mass get locked with the wall and vibrates in same phase as the wall (called convective
mass and the balance fluid vibrates on its own called the sloshing mass.
Now let us look at the mass expression it is obvious that the loading is triangular
(hydrostatic) in nature.
For a beam of span L having triangular load at as shown in Figure 3.4.8, it is
elementary to show that equivalent lumped mass at A is mL/3 and that at B is mL/6.
Here m is the mass per unit length and L the span. Since for the dam the base is
considered fixed the term ml/3 goes to the foundation and is not effective and the
active fluid participation is only ml/6. Thus based on this principle the mathematical
model for the dam and the fluid is expressed as in Figure 3.4.9.
3.4.2.9 Model for f luid-structure interaction of the dam
and reservoir
In the above mathematical it is thus assumed that two third of the weight of water
gets transfered to the dam foundation and one third of the water vibrates on its own
(sloshing mode). The free vibration equation of the coupled fluid structure system can
now be expressed as
_
m
s
0
0
m
f
3
_ _
u
s
u
f
_
+
_
K
s
+K
f
K
f
K
f
K
f
_ _
u
s
u
f
_
= 0 (3.4.71)
where, K
s
= stiffness of the dam structure; K
f
= stiffness of the fluid, M
s
and M
f
=
mass of the dam and the equivalent fluid respectively.
The eigen-solution of the above problem can now be carried out by the standard
expression
_
K
s
+K
f
m
s
2
K
f
K
f
K
f

m
f
3

2
_ _
u
s
u
f
_
= 0 (3.4.72)
It is evident from above that the frequency vis--vis the time period of the system
will get modified. Neither the fluid mode will have a frequency @ 4H
w
/c nor the
fundamental time period of the structure will remain T
1
= 4.68
H
2
B
_
c
/gE for B
t
/B =
0.1, say. The dynamic pressure and the response of the dam will now have to be
obtained based on this coupled frequency as derived above.
Let
1

2
be the two modified natural frequencies of the fluid and structure and
let the corresponding eigen vectors be expressed as [] =
_
11

12
21

22
_
;
The displacement of the dam is then given by
u
i
=
i
_
L
ni
M
ni
_
_
S
ai
2
i
_
(3.4.73)
Here M
ni
= [
i
]
T
[M] [
i
] and L
ni
= [
i
]
T
[M] [I] (3.4.74)
I is an identity matrix.
Thus, the maximum displacement for a particular mode is given by
u
i
=
i
_
L
ni
M
ni
_
_
S
ai
2
i
_
f () at 1.0 (3.4.75)
Once displacement are observed the maximummoment and shear for the two modes
are obtained from the expression
M = EI
d
2
u
dz
2
and V = EI
d
3
u
dz
3
(3.4.76)
Thus, M
max
=
E
c
I
0
(1 + )
3
2
1
H
2
u
max
and V
max
=
E
c
I
0
(1 + )
3
3
1
H
3
u
max
(3.4.77)
Comparison of displacement Rigourous Vs
Lumped Mass approach
0
5
10
15
0 0. 2 0. 4 0. 6 0. 8 1
z/H
D
i
s
p
l
a
c
e
m
e
n
t
(
m
m
)
Disp. Analytic
solution
Disp. Lumped
Mass
Figure 3.4.10 Comparison of displacement analytical versus lumped mass approach.
The final design moments and shear are then
M
R
=
_
M
2
1
+M
2
2
and V
R
=
_
V
2
1
+V
2
2
(3.4.78)
Once the moment and shear are known they may be multiplied by the coefficients
given in the Table 3.4.1 to get the values along the height of the dam.
You might wonder how correct would be the result?
Just for comparison a 100 m high dam 70 m wide with B
t
/B = 0.1 was analysed by
the rigorous semi-analytic method and the lumped mass approach, the displacement
plot is compared in Figure 3.4.10.
It can be observed that the values are in excellent agreement and gave almost same
results.
The method derived here is simple and practical. It does not need an elaborate
expensive analysis and can well be carried out in a spread sheet. This would give a far
more realistic result than what is suggested in IS-1893 (1984).
One major advantage now is that coupling the soil stiffness to consider the
foundation interaction effect becomes relatively simple.
3.4.2.10 Consideration of f luid- structure-foundation interaction
Having established the validity of the lumped mass model proposed above,
the model can be further extended to couple the soil stiffness as shown in
Figure 3.4.11.
u
f
H
M
f
w w
6g
2
=
8
2
= K
f
u
s
=
3
I
g
BH
M
s
h
1
3
4
1
3
12
I
H
EB
K
s
=
K
x
K

B
m
Figure 3.4.11 Lumped mass stiffness model of dam.
3.4.2.11 Lumped mass stiffness model of dam with
soil-f luid-structure interaction
We had shown in Chapter 1 (Vol. 2) that the equivalent stiffness of the soil structure
system is given by
1
K
e
=
1
K
s
+
1
K
x
+
h
2
K
(3.4.79)
where, K
e
= equivalent stiffness of the soil structure system; K
s
= stiffness of dam as
derived earlier; K
x
= lateral stiffness of the dam foundation; K
= rocking stiffness of
foundation, and,

h = center of gravity of the dam from bottom.
Based on above the equation of motion of free vibration gets modified to
_
m
s
0
0
m
f
3
_
_
u
s
u
f
_
+
_
K
e
+K
f
K
f
K
f
K
f
_ _
u
s
u
f
_
= 0 (3.4.80)
The rest of the steps now remain same as previous. This shall be elaborated latter
by a suitable numerical example.
3.4.2.12 Finite element analysis of concrete dam
Finite element stress analysis of concrete dams is the most usual practice in industry.
As stated in Chapter 2 (Vol. 1) the first application of FEM in civil engineering was
carried out for concrete dam(Norfolk Damanalysis by Clough). Since then it has been
a common practice that the stress analysis for such dams are invariably carried out by
applying FEM software. This is quite justified considering most of the concrete dams
are massive in nature and, secondly in many cases a number of cooling water pipes
run through its cross-sections to control the heat of hydration during the setting of
concrete. The heat generated is significant in such cases and if not properly controlled
can result in thermal cracks. In many cases tunnels run through the sections that are
used to carry water pipes and also used for maintenance. Often one is interested to
know the stress concentration around such openings and ensure no cracks generate
around them.
Figure 3.4.12 shows a conceptual finite element model for a concrete dam. For
static analysis the problem is quite straightforward and does not pose any problem.
For uncoupled vibration analysis also, the solution is not complicated.
One does the eigen value and modal analysis for the dam only and applies the
hydrodynamic pressure as point loads on the surface of the dam and combines the
stresses generated by these load cases to finally arrive at the dynamic stresses induced
in the dam body.
However, for coupled fluid structure interaction, the analysis poses some difficulties
that are discussed as hereunder:
The reservoir in question is unbounded. As such to ensure that no waves are
reflected back fromthe finite element fluid boundary one has to take the boundary
significantly away fromthe dam. This greatly increases the model size and at times
can make the analysis quite laborious and expensive.
Fluid Poissons ratio is = 0.5, this at times makes the matrix singular and poses
severe difficulties in eigen value solutions in certain cases.
Majority of the Finite Element software that are used commercially for struc-
tural analysis (except ANSYS and NASTRAN) do not have fluid elements in
their library as such trying to simulate 2D fluid elements by using normal struc-
tural elements (mostly by trying to use a Poissons ratio of 0.49) does not
often work.
Thus if one has not developed a special purpose source code circumventing the
above mentioned problem the task is indeed difficult to carry out a complete fluid
structure interaction analysis based on normal finite element structural software that
are available in the market.
Free surface of water
Figure 3.4.12 Conceptual finite element model of a concrete dam.
Free surface of water
Fluid mass @ M
f
/3
Fluid spring @
2
B
m
/8
Figure 3.4.13 Conceptual finite element model of a concrete dam with fluid modeled as spring and
lumped mass.
We now propose a simple procedure based on which the above problem can well
be avoided and show how we can do such analysis while carrying out a FEA.
It was shown earlier that the reservoir fluid can effectively be represented by equiv-
alent spring and lumped mass as furnished in Equation (3.4.51). These are in effect
stiffness and mass contribution per unit area. Thus if we multiply this value by the
effective contributing area for each node at wall-water surface we get an effective
spring contribution for each node along the depth of the dam. It is apparent that sum
of all such springs at each node must be equal to (
2
B
m
)/8.
For the mass distribution as explained earlier 2M
f
/3 of the mass goes to foundation
can be ignored for dynamic analysis of the dam and the balance lumped mass M
f
/3
can be connected to the top of the dam.
Based on above argument Figure 3.4.12 gets modified to Figure 3.4.13.
The problem has now simplified considerably. We do not have to worry about the
infinite fluid domain or about the numerical difficulties encountered in trying to model
a mock fluid in 2D. The spring and lumped mass adequately models the fluid and
coupled analysis for the dam can now be easily carried out in any of the commercially
available software like SAP 2000, GTSTRUDL, etc.
Example 3.4.1
A concrete dam of height 100 m, base width 70 m has top width of 7 m. The
maximum height of water it contains is 97.5 m. The dam is built on site of hard
rock and is in zone IV as per IS-1893 (2002). The grade of concrete used is M25.
Consider unit weight of water as 10 kN/m
3
and that of concrete as 25 kN/m
3
.
Bulk modulus of water (B
m
) = 2.1110
6
kN/m
2
. Shear wave velocity of
soil =750 m/sec. Unit weight of soil =19 kN/m
3
. Perform Earthquake analysis
based on
The IS-code method.
Proposed dynamic analysis.
Dynamic analysis with fluid structure interaction.
Dynamic analysis with coupled soil fluid structure interaction considering
20% overall damping.
Solution:
For the dam being in zone IV:
As per IS-code Z = 0.24, I = 3.0, R = 1.5 (Assuming the dam to be
un-reinforced).
This gives =
ZI
2R
= 0.24.
IS-code method
The time period is given by, T = 5.55
H
2
B
_

c
gE
s
.
Here, E
c
= 28500000 kN/m
2
, H = 100 m B = 70 m,
c
= 25 kN/m
3
g =
9.81 m/sec
2
.
This gives, T = 0.237 sec for which S
a
/g for 5% damping = 2.5.
The weight of the dam, W =
70+7
2
100125 = 96250 kN per meter
width
Thus base shear is given by, V
B
= 0.6
h
W V
B
= 0.6 (S
a
/g) W =
34650 kN.
The base moment is given by, M
B
= 0.9
h
W
h
Here

h = 36.36 m from base; thus, M
b
= 0.9 (S
a
/g) W

h =
1889811 kN m.
The Moment and shear force at various depth are given as
Z/H C
m
C
s
Moment (kN m) Shear (kN)
0 1 1 1889811 34650.00
0.2 0.675 0.875 1275622 30318.75
0.4 0.35 0.65 661433.9 22522.5
0.6 0.15 0.4 283471.7 13860
0.8 0.05 0.18 94490.55 6237
1 0 0 0 0
The hydrodynamic pressure is given by, p = C
s
w
h
Here as suggested in IS-1893, we take same
h
as that of dam i.e S
a
/g = 2.5
for T = 0.237 sec.
The hydrodynamic pressure is given hereafter in a tabulated form as
Z/H C
s
p
dyn
(kN/m
2
)
0 0 0
0.2 0.3528 206.388
0.4 0.5292 309.582
0.6 0.645519 377.6288
0.8 0.712875 417.0319
1 0.735 429.975
Considering, V
z
= 0.726p z and M
z
= 0.299 p z
2
.
The moments and shears are calculated at various depth as
Z/H Moment Shear
0 1.23 10
+06
3.04 10
+04
0.2 9.51 10
+05
2.36 10
+04
0.4 6.46 10
+05
1.60 10
+04
0.6 3.53 10
+05
8.77 10
+03
0.8 1.18 10
+05
2.92 10
+03
1 0.00 10
+00
0.00 10
+00
Thus total design moment (due to dam itself and hydrodynamic pressure) as
per IS code is
Z/H Moment (IS-code) Shear (IS-code)
0 3.12 10
+06
6.51 10
+04
0.2 2.23 10
+06
5.39 10
+04
0.4 1.31 10
+06
3.86 10
+04
0.6 6.37 10
+05
2.26 10
+04
0.8 2.12 10
+05
9.16 10
+03
1 0.00 10
+00
0.00 10
+00
Proposed dynamic analysis
Here B
t
/B = 0.1, this gives, T = 4.68
H
2
B
_

c
gE
s
.
Substituting the values, we have, T = 0.20 sec and that gives S
a
/g = 2.5.
Corresponding eigen vectors for the first mode is
{}T = 0.986 0.167 0.020
Hence,
M(z) = 0.0335
_
S
a
g
_
c
BH
2
_
1 +
z
H
_
3
(3.466f
1
(z) 3.679f
2
(z) +1.234f
3
(z))
where f
i
(z) =

2
i
H
2
_
sin
i
z
H
sin h
i
z
H
+
i
_
cos
i
z
H
+cos h
i
z
H
__
and = 0.9.
Similarly,
V(z) = 0.0335
_
S
a
g
_
c
BH
_
1 +
z
H
_
3
(6.5f
1
(z)
17.27f
2
(z) +9.693f
3
(z))
Substituting the values, we have Dynamic Moments and Shears as follows
Z/H Moment dynamic (kNm) Shear Dynamic (kN)
0 1.65 10
+06
7.58 10
+03
0.2 1.04 10
+06
1.25 10
+04
0.4 6.86 10
+05
1.11 10
+04
0.6 2.88 10
+05
2.38 10
+03
0.8 3.30 10
+04
2.33 10
+03
1 0.00 10
+00
8.01 10
04
We observe here a very interesting thing while the Bending moment is hardly
affected by the higher modes, the shear profile is significantly influenced by it.
The free field time period of fluid is given by T =
4H
w
c
where B
m
=
w
c
2
.
Substituting the value,
w
= 10/9.81 = 1.02 kN-sec
2
/m, we have T =
0.27 sec.
Considering 15% damping for fluid we have S
a
/g = 25 and taking Response
reduction factor
32
R = 15, we have = 0.24.
p =
12
( +2)
S
a
w
H
w
g
sin
z
2H
and considering V
dyn
= 0.6455p
z
z
M
dyn
= 0.4014 pz z
2
, we have
Dynamic pressure Moment Proposed Shear Proposed
Z/H (kN/m
2
) Dynamic (kNm) Dynamic (kN)
0 0 1.69 10
+06
2.71 10
+04
0.2 134.2987 1.29 10
+06
2.06 10
+04
0.4 255.4514 8.22 10
+05
1.32 10
+04
0.6 351.5987 3.98 10
+05
6.38 10
+03
0.8 413.329 1.05 10
+05
1.68 10
+03
1 434.5999 0.00 10
+00
0.00 10
+00
Thus the total moments and shears due to proposed dynamic analysis and
hydrodynamic pressure is given by
Moment proposed Shear proposed
Z/H dynamic (kN m) dynamic (kN)
0 3.34 10
+06
3.47 10
+04
0.2 2.33 10
+06
3.31 10
+04
0.4 1.51 10
+06
2.43 10
+04
0.6 6.86 10
+05
4.00 10
+03
0.8 1.38 10
+05
6.58 10
+02
1 0.00 10
+00
8.01 10
04
32 We have assume R = 1.5 considering the dam as unreinforced.
Response considering fluid structure interaction semi analytic approach
Though the method is analytic we will hardly use any analysis here, but would
refer extensively to the design tables furnished earlier.
Based on the tables furnished we have the following design parameters.
For H/B = 100/70 = 1.43, B
t
/B = 0.1 H
w
/B = 1.4
1
= 1.875, I
1
= 1.13, I
2
= 1.855, I
3
= 0.509, I
4
= 0.369, I
5
= 1.067,
m
=
0.272,
s
= 0.259 and C
fs
= 0.629;
Now considering
T = 6.19
H
2
B
_
c
C
FS
E
c
g
we have T = 0.21 secs S
a
/g = 2.5
Considering,
i
=
I
4
+
m
I
5
I
3
+
m
I
2
, we have
1
= 0.65
The displacement thus in mm is given by
u(z)
max
= 1.01175
i
c
C
FS
B
2
E
c
_
S
a
g
_
H
4
u(max) = 0.012 m
E
c
I = 28500000(70)
3
/12 = 8.14610
11
kN/m
2
Thus, considering M
max
=
E
c
I
0
(1+)
3
2
1
H
2
u
max
and V
max
=
E
c
I
0
(1+)
3
3
1
H
3
u
max
.
We have, M
max
= 3.46910
6
kN/m
2
and V
max
= 4.77 10
4
kN, at = 0
Multiplying by the coefficients furnished in the table, the variation with depth
is as shown hereunder: renewcommand
Z/H Moment FSI Shear FSI
0 3.469 10
+06
4.77 10
+04
0.2 1.387 10
+06
2.60 10
+04
0.4 4.193 10
+05
1.15 10
+04
0.6 7.762 10
+04
3.45 10
+03
0.8 4.865 10
+03
4.62 10
+02
1 0.000 10
+00
6.77 10
03
Fluid structure interaction lumped mass model
In this case for the first mode the dam stiffness is given by
K
s
=
EB
3
4
1
12H
3
I
1
= 1.13810
7
kN/m
The fluid stiffness is given by
K
f
=

2
B
m
8
= 2.60310
6
kN/m
The mass of dam in first mode is given by
M
s
=
BH
g
I
3
= 9080 kN-sec
2
/m
The mass of fluid is given by, m
f
=

w
H
2
w
2g
=
1097.5
2
29.81
= 4845 kN-sec
2
/m.
The free vibration equation is
_
M
s
0
0
1
3
M
f
_
_
u
s
u
f
_
+
_
K
s
+K
f
K
f
K
f
K
f
_ _
u
s
u
f
_
= 0
or
_
9.08 10
3
0
0 1615
_ _
u
s
u
f
_
+
_
1.39810
7
2.60310
6
2.60310
6
2.60310
6
_ _
u
s
u
f
_
= 0
On eigen value solution, we have
{} =
_
895.02
2256
_
{} =
_
29.917
47.502
_
rad/sec
The corresponding Eigen vectors are given as
[] =
_
0.445 0.40
1 1
_
The time period of the system
33
is thus
{T} =
_
0.21
0.132
_
sec
Here for the first mode
M
n1
= [
1
]
T
[M
1
]
Here [
1
]
T
= [0.445, 1] and this gives M
n1
= 3411.
L
n1
= [
1
]
T
M[I] here I = [1, 1]
T
and this gives L
n1
= 5653.
33 Observe here that under uncoupled condition the time period of the dam was T = 0.20 sec
and that of fluid was T = 0.27 sec. this is now different. While this matches exactly with semi-
analytic method @0.21 sec.
L
2
n1
/M
n1
= 9369; thus percentage of modal mass participation (MMP) factor is
given by
MMP =
9369100
9080 +1615
= 87.603%
For T = 0.21 sec, S
a1
/g = 2.5, thus design acceleration is given by
S
a1
= 2.5 g = 2.50.249.81 = 5.886 m/sec
2
.
S
v1
= S
a1
/
1
= 0.197 m/sec.
Thus
max
=
1
_
L
n1
M
n1
__
S
v1
1
_
f (); at = 1.0 (i.e. z = H), f () = 2.7242.
max
=
_
0.013
0.030
_
m
Thus it is apparent that for the first mode, displacement of the dam is 13 mm
while that of the water is 30 mm
34
.
Considering M = EI d
2
u/dz
2
and V = EI d
3
u/dz
3
we have
M
max
= EI
0
2
1
H
2
max(struct)
or M
max
= 2850000028580[(1.875)
2
/
(70)
2
] 0.013 = 3.78210
6
kN m at base and V
max
= 28500000
28580[(1.875)
3
(70)
3
] 0.013 = 7.09110
4
kN. at base
Similarly for second mode
M
n2
= [
T
2
M[
2
].
Here [
2
]
T
n2
= 3068. L
n2
= [
2
]
T
M[I] and
this gives L
n2
= 2017. L
2
n1
/M
n1
= 1326; thus percentage of modal mass
participation (MMP) factor is given by
MMP =
1326100
9080 +1615
= 12.397%
For T = 0.132 sec, S
a2
/g = 2.5, thus design acceleration is given by S
a2
=
2.5 g = 2.50.249.81 = 5.886 m/sec
2
. S
v2
= S
a2
/
2
= 0.124 m/sec.
Thus
max
=
2
_
L
n2
M
n2
__
S
v2
2
_
f () at = 1.0 f () = 2.7242.
34 It is to be noted that based on semi-analytic method the displacement obtained for the dam
is 12 mm thus variation with lumped mass model is of the order of 7% only.
max
=
_
1.86910
3
4.672 10
3
_
m
Thus for the second mode, displacement of the dam is 1.8 mm while that of
the water is 4.6 mm.
2
u/dz
2
and V = EI d
3
u/dz
3
we have
M
max
= EI
0
2
1
H
2
max(struct)
M
max
=2850000028580[(1.875)
2
/(70)
2
] 1.86910
3
=5.351
10
5
kN mat base and V
max
= 2850000028580[(1.875)
3
(70)
3
] 1.869
10
3
= 1.00310
4
kN m at base
The above now can multiplied by the coefficients given in design table to
obtain the values as shown hereunder
Z/H Moment (1st mode) Shear (1st mode)
0 3.78 10
+06
7.09 10
+04
0.2 1.51 10
+06
3.87 10
+04
0.4 4.57 10
+05
1.71 10
+04
0.6 8.46 10
+04
5.13 10
+03
0.8 5.30 10
+03
6.86 10
+02
1 0.00 10
+00
1.01 10
02
Soil fluid-structure interaction
In this case as calculated previously
K
s
=
EB
3
4
1
12H
3
I
1
= 1.13810
7
kN/m and K
f
=

2
B
m
8
= 2.60310
6
kN/m.
Considering foundation contact area = 701 = 70 m
2
, equivalent circular
footing radius is r = 4.72 m.
For rocking mode considering base moment of inertia of the damas I = B
3
/12
equivalent radius r
= 13.812 m.
Dynamic shear modulus =
_
19
9.81
_
750
2
= 1089450 kN/m
2
.
Poissons ratio = 0.45
K
h
=
32G(1 )r
7 8
=2.66210
7
kN/m; K
=
8Gr
3
3(1 )
=1.39210
10
kN/m.
h = 36.36 m from the base,

1
K
e
=
1
K
s
+
1
K
h
+
h
2
K
K
e
= 4.53610
6
kN/m
The free vibration equation is
_
M
s
0
0
1
3
M
f
_
_
u
s
u
f
_
+
_
K
e
+K
f
K
f
K
f
K
f
_ _
u
s
u
f
_
= 0
_
9080 0
0 1651
_ _
u
s
u
f
_
+
_
7.13910
6
2.60310
6
2.60310
6
2.60310
6
_ _
u
s
u
f
_
= 0.
On eigen value solution we have
{} =
_
403.76
1994
_
{} =
_
20.094
44.667
_
rad/sec
The corresponding eigen vectors are given as
[] =
_
0.749 0.237
1 1
_
The time period of the system
35
is thus {T} =
_
0.313
0.141
_
sec.
Here for the first mode
M
n1
= [
1
]
T
M[
1
]
Here [
1
]
T
n1
= 6716. L
n1
= [
1
]
T
M[I] here
I = [1, 1]
T
and this gives L
n1
= 8420. L
2
n1
/M
n1
= 10560; thus percentage of
modal mass participation (MMP) factor is given by
MMP =
9868100
9080 +807.5
= 98.719%
a1
/g = 2.5 thus design acceleration is given by S
a1
=
0.6 2.5 g 0.62.50.249.81 = 3.532 m/sec
2
considering 20% damping
of the soil.
S
v1
= S
a1
/
1
= 0.176 m/sec.
Thus
max
=
1
_
L
n1
M
n1
__
S
v1
1
_
f (); at =1.0 (i.e. z =H), f ()=2.7242.
35 Observe here that under uncoupled condition the time period of the dam was T = 0.199 sec and
of fluid was T = 0.27 sec. this is now different.
max
=
_
0.022
0.030
_
m
Thus for the first mode, displacement of the dam is 22 mm while that of the
water is 30 mm.
2
u/dz
2
and V = EI d
3
u/dz
3
, we have
M
max
= EI
0
2
1
H
2
max (struct)
M
max
= 2850000028580[(1.875)
2
/(70)
2
] 0.022 = 6.41310
6
kN
m at base and V
max
= 2850000028580[(1.875)
3
(70)
3
] 0.022 = 1.202
10
5
kN at base
Similarly for second mode
M
n2
= [
T
2
[M
2
].
Here [
2
]
T
n2
= 2126. L
n2
= [
2
]
T
M[I] and
this gives L
n2
= 539.81. L
2
n2
/M
n2
= 137.033; thus percentage of modal mass
participation (MMP) factor is given by
MMP =
137.033 100
9080 +807.5
= 1.281%
a2
/g = 2.5 thus design acceleration is given by
S
a2
= 0.625 g = 0.6 2.5 0.24 9.81 = 3.532 m/sec
2
.
S
v2
= S
a2
/
2
= 0.079 m/sec.
Thus
max
=
2
_
L
n2
M
n2
__
S
v2
2
_
f () at = 1.0 f () = 2.7242.
max
=
_
2.906 10
4
1.225 10
4
_
m
Thus for the second mode, displacement of the dam is 0.29 mm while that of
the water is 1.225 mm.
2
u/dz
2
and V = EI d
3
u/dz
3
we have
M
max
= EI
0
2
1
H
2
max (struct)
or M
max
= 2850000028580[(1.875)
2
/(70)
2
]2.90610
4
= 8323 kN m
at base and V
max
= 28500000 28580 [(1.875)
3
(70)
3
] 2.906 10
4
=
1561 kN at base.
The above now can multiplied by the coefficients given in design table to
obtain the values hereunder
Z/H Moment (1st mode) Shear (1st mode)
0 6.41 10
+06
1.20 10
+05
0.2 2.57 10
+06
6.55 10
+04
0.4 7.75 10
+05
2.89 10
+04
0.6 1.44 10
+05
8.69 10
+03
0.8 9.00 10
+03
1.16 10
+03
1 0.00 10
+00
1.70 10
02
The values computed above are now compared in Figs. 3.4.14 and 15.
Comparison of Bending Moment along Dam Height
-1.00E+06
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
0 0.2 0.4 0.6 0.8 1
z/H
M
o
m
e
n
t
9
k
N
.
m
)
Moment (ISCode)
Moment dynamic
Moment FSI
Moment lumped
mass
Moment SFSI
Figure 3.4.14 Comparison of bending moment by various methods.
Comparison of Shear force
-2.00E+04
0.00E+00
2.00E+04
4.00E+04
6.00E+04
8.00E+04
1.00E+05
1.20E+05
1.40E+05
0 0.2 0.4 0.6 0.8 1
z/H
S
h
e
a
r

k
N
Shear (IS code)
Shear dynamic
Shear FSI
Shear lumped
Shear SFSI
Figure 3.4.15 Comparison of shear force by various methods.
3.5 ANALYSIS OF EARTH DAMS AND EMBANKMENTS
3.5.1 Dynamic earthquake analysis of earth dams
In countries like India, USA, Australia where a major part of the economy is largely
dependent on agriculture and consists of a number of rivers and canals, for the purpose
of irrigation and flood control (specially in rural areas) dams made out of natural earth
are very popular.
The system usually consists of carefully chosen natural soil with a clay core (to
control the seepage) and has been successfully utilized for effective utilization of water
stored thereby. We shall deal with the earthquake analysis of such dams in this section.
Unlike concrete dams, which are far more rigid, the behaviour of earth dams are
different under earthquake.
Though theoretically concrete dams also do not behave exactly as rigid structures
they are usually assumed to behave as rigid without much practical error. In such case
it is assumed that the motion at the base is same in all parts of the concrete dam. On the
contrary for earth dams the constituent material being much softer it mostly behaves
as flexible structure where the acceleration induced within the dam varies with height
and could be different at different points of the dam.
3.5.2 Mononobes method for analysis of earth dam
One of the earliest analyses of such earth dam was proposed by Mononobe (1936)
in his classic paper, considering the dam as an isosceles triangle having uniform mass
density. As width of the base of the dam is considered as far greater then the height
of the dam, it is assumed that the shear deformation is predominant and bending
deformation which is secondary in nature may be ignored.
With respect to Figure 3.5.1, for a strip dz, the horizontal shear force and the inertial
shear is given by
(az)dz
_
2
X
t
2
_
=
s
z
dz az
_
2
X
t
2
_
=
s
z
(3.5.1)
where, X is the horizontal amplitude of displacement and s is the shear force.
The shear modulus
36
of the soil medium is given by
G =
Shear stress
Shear strain
=
s/az
dX/dz
which gives, s = Gaz
dX
dz
(3.5.2)
Differentiating the above with respect to Z we have
s
z
= Ga
X
z
+Gaz
2
X
z
2
from which we have,
36 This is considered constant over the depth in this case.
Figure 3.5.1 Shear force in a triangular wedge shaped earth dam.
z
2
X
t
2
= Ga
X
z
+Gaz
2
X
z
2
(3.5.3)
which is the basic equation of shear vibration of the triangular wedge.
For solution of the above equation considering, z = z
H, where H is the height of

the dam, and considering, X = x(z
) sin t, we have
2
X
t
2
= x
2
sin t and
X
z
=
x
z
sin t
or
X
z
=
1
H
x
z
sin t and

2
X
z
2
=
1
H
2
2
x
z
2
sin t (3.5.4)
Substituting the above in the basic equation of shear vibration, we have
G
_
x
z
+z

2
x
z
2
_
+ H
2
z
x
2
= 0 (3.5.5)
The above is the Bessels equation, whose solution is given by the expression
x = AJ
0
__
H
_
G
_
z
_
(3.5.6)
where, A is a constant and J
0
is the Bessel function of the first kind and zero order.
Implementing the boundary condition at the base of the dam (z
= 1), at x = 0, we
have
J
0
__
H
_
G
__
= 0 (3.5.7)
The above can be expanded as
__
H
_
G
__
n
1
= n
1
4
+
0.0507
4n 1
+ , where n = 1, 2, 3. . . (3.5.8)

n
=
v
s
H
_
n
1
4
+
0.0507
4n 1
_
(3.5.9)
Considering, T = 2/
n
we have, T =
2H
v
s
_
n
1
4
+
0.0507
4n1
_ (3.5.10)
Thus, for the first three modes, we have time periods as shown in Table 3.5.1.
The IS-code only furnishes the first fundamental mode as mentioned above for the
calculation of time period.
We started the derivation of the time period equation based on the equilibrium of
the elemental strip given by
(az)dz
_
2
X
t
2
_
=
s
z
dz and

2
X
t
2
= x
2
sin t (3.5.11)
which gives maximum value of

2
X
t
2
as (x
2
)
Again considering,
n
=
v
s
H
_
n
1
4
+
0.0507
4n1
_
for n = 1;
n
= 0.767
v
s
H
, which
gives,
_
2
X
t
2
_
max
= x(0.767)
2
2
v
2
s
H
2
(3.5.12)
Substituting the value of x derived previously, we have
_
2
X
t
2
_
max
= AJ
0
_
H
_
G
z
0.767

H
_
G
_
(0.767)
2
2
v
2
s
H
2
(3.5.13)
proportional to
v
2
s
H
2
J
0
(0.767z
);
the numerical constants being considered within the constant proportionality.
Table 3.5.1 Time period of earthen dams for first
three modes.
Mode Time period (sec)
1 2.607 H/v
s
2 1.138 H/v
s
3 0.726 H/v
s
Table 3.5.2 Variation of acceleration along the height of the dam as per Mononobe.
z
= z/H 0.0 0.1 0.3 0.5 0.7 0.9 1.0

Maximum acceleration , J
0
(2.41z
) 1.00 0.986 0.874 0.668 0.406 0.127 0.00

For a particular dam as the
v
2
s
H
2
is a constant quantity, hence maximum acceleration
_
2
X
t
2
_
max
is proportional to J
0
(2.41z
).
The value of J
0
(2.41z
) for various values of z
are as shown in Table 3.5.2.

The above values may be used to determine the inertial force at different heights of
the dam.
When resonance occurs the deformations would tend to infinite value without any
internal friction, however due to internal friction the maximum deflection is restricted
to a finite value.
Mononobe calculated the ratio of top and bottom deflection for G and and also
when the parameters vary linearly with depth. Considering linear variation with depth
the acceleration is expressed as
=
0
_
1 +
d
0
d
d
H z
H
_
(3.5.14)
where is the acceleration at any depth z below top,
0
is the ground level acceler-
ation and d
0
is the maximum displacement at the top. The ratio of top and bottom
displacement came to 2.5 and 3.5 respectively from which Mononobe concluded that
maximum acceleration at the top may be 2.5 to 3.5 times the acceleration at the base.
3.5.3 Gazetas method for earth dam analysis
Gazetas (1982) developed solutions to the shear beamwave equation for the case where
the shear modulus value varies with the depth of the dam given by G(z) = G
av
(z/H)
,
where G
av
is the average shear modulus of material constituting the dam.
The nth natural frequency of the dam considering the dam to be triangular in shape
(i.e. h/H = 1) is given by
n
=
v
s
H
n
8
(4 + )(2 ), (3.5.15)
where v
s
= average shear wave velocity of the soil in the dam and
n
is a function
given below in Table 3.5.3 for the first three modes; h is the height of the dam to it s
crest, and H is full height of the triangle.
The mode shape for the nth natural frequency is given by
n
(z) = (z/H)
/2
J
q
[
n
(z/H)
1/2
] (3.5.16)
Table 3.5.3 Frequency coefficient for earth dam for first three modes
as per Gazetas (1982).

1

2

3
0 2.404 5.520 8.654
1/2 2.903 6.033 9.171
4/7 2.999 6.133 9.273
2/3 3.142 6.283 9.525
1 3.382 7.106 10.174
where J
q
= Bessels function of the first kind of order q = /(2 ) and can be
evaluated from the expression
J
q
(x) =
k=0
(1)
k
k!(q +k +1)
_
x
2
_
q+2k
(3.5.17)
where (x) is the gamma function and is given by the expression
(x) =
_
0
e
x
x
n1
dx. (3.5.18)
3.5.4 Makadisi and Seeds method for analysis
of earth dam
A simplied procedure proposed by Makadisi and Seed (1977) taking into considera-
tion the strain dependent degradation of shear modulus and damping of soil possibly
remains the most popular method for dam analysis in design office and was developed
in similar line of Mononobes method as explained earlier.
Based on the shear beam equation as derived earlier they found the acceleration at
any level z as a function of time is given by
u(z, t) =
n=1
2J
0
(
n
z/H)
n
J
1
n
V
n
(t) (3.5.19)
where, J
0
, J
1
= Bessel function of first kind of order zero and one;
n
= the zero value
of frequency equation, J
0
(H
_
/G) = 0;
n
= natural frequency of the nth mode =
n
V
s
/H.
V
n
(t) =
t
_
0
u
g
e
n
(t)
sin[
dn
(t )]d (3.5.20)
where,
dn
=
n
1
2
the damped natural frequency and
n
= critical damping
ratio.
Thus, u(z, t) =
n=1
n
(z)
n
V
n
(t) (3.5.21)
where,
n
(z) =
n=1
2J
0
(
n
z/H)
n
J
1
n
= modal participation factor.
Considering the first three modes of vibration, the corresponding values of
n
are
always
1
= 2.4, 5.52 and
3
= 8.65 which gives the first three natural three natural
frequencies as shown in Table 3.5.4.
At the crest of the dam (z = 0) the corresponding values of mode participation
factors
n
(0) for the first three modes are given in Table 3.5.5.
Thus the crest acceleration at each mode is given by
u(0, t) =
n=1
n
(0)
n
V
n
(t) (3.5.22)
when based on response spectrum analysis the maximum expected acceleration is
given by
u(0, t) =
n
(0)S
an
(3.5.23)
Thus for the first three modes the maximum acceleration is given by
u
1
max = 1.6S
a1
; u
2
max = 1.06S
a2
, and u
3
max = 0.86S
a3
Table 3.5.4 Natural frequencies for first three modes
for earth dams as per Makadis and Seed
(1977).
Mode Frequency
1
1
= 2.4v
s
/H
2
2
= 5.52v
s
/H
3
3
= 8.65v
s
/H
Table 3.5.5 Modal participation factor for first three modes for earth
dams as per Makdisi and Seed (1977).
Mode Modal participation factor
1
1
= 1.6
2
2
= 1.06
3
3
= 0.86
The maximum SRSS values of the acceleration is thus given by the expression
u
n
max =
3
n=1
_
( u
n
max)
2
(3.5.24)
To estimate the strain compatible material properties an expression for the average
shear strain over the entire section should be determined. From the shear slice theory,
the expression for shear strain at any level in the dam is given by
(z, t) =
n=1
2J
1
(
n
z/H)
H
n
J
1
(
n
)
V
n
(t) =
H
v
2
s
n=1
2J
1
(
n
z/H)
2
n
J
1
(
n
)
n
V
n
(t)
=
H
v
2
s
n=1

n
(z)
n
V
n
(t) (3.5.25)
where
n
(z) =
2J
1(
n
z
H
)
2
n
J
1
(
n
)
= shear mode participation factor.
It is recommended that the contribution of higher modes being small it is sufficient
to consider the contribution of the first mode only over the entire depth of the dam
for calculation of the average shear strain.
Thus, the maximum average shear strain is obtained as
max
(z) =
H
v
2
s

1
S
a1
(3.5.26)
It has been shown that the average value of the first factor is given by,
1avg
= 0.3
Assuming an equivalent cyclic shear strain as approximately 65% of the average
shear strain
avg
(max) = 0.65 0.3
H
v
2
s
S
a1
(3.5.27)
Having obtained a new value for the average shear strain a new set of modulus and
damping value can obtained from the expressions
37
G =
G
max
(1 +

r
)
and
D
c
D
m
=
/
r
1 + (/
r
)
(3.5.28)
The iterations are carried out till the values become constant with respect to the
earlier cycle. It has been observed that the system generally converges by 3 cycles.
37 For further explanation of how to modify the damping and shear modulus with respect to shear strain
refer Chapter 1 (Vol. 2) under the section titled Geo-technical Consideration for Dynamic Soil Structure
Interaction.
Water Line
y
W
W
Failure line
Figure 3.5.2 Slope failure of dam under earthquake force.
3.5.5 Calculation of seismic force in dam and its
stability
Once the modal acceleration and its SRSS values is obtained, for assessing the damage
incurred in a damshown in Figure 3.5.2, the usual design office practice is to undertake
a pseudo-static analysis. For major damage observed in a damis due to stability failure
specially on the upstream side.
To asses this, a slope stability analysis is usually carried out to evaluate the factor
of safety which should be more than unity under the seismic force in addition to all
other forces working on the dam.
For stability analysis, any standard method like Slip-Circle, Fellenius or Bishops
method
38
may be applied which needs to be modified to cater for the applied seismic
force.
For instance the factor of safety (F) based on slip circle method is modified to:
F =
n
i=1
_
N U
y
T
_
tan +
n
i=1
cS
n
i=1
_
T +
y
N
_ (3.5.29)
where, F = factor of safety; N = normal component of the weight of the soil; T =
tangential component of the weight of the soil; = friction angle resistance of soil;
c = cohesion of soil; S = width of segments of slices considered for analysis;
y
=
seismic acceleration coefficient (S
a
/g) in the horizontal direction obtained from code
based on the fundamental time period as derived earlier, and U =force due to internal
Pore pressure within the dam.
3.6 ANALYSIS OF EARTH RETAINING STRUCTURES
3.6.1 Earthquake analysis of earth retaining structures
Retaining walls supporting earth make an important component in infra-structure
works like highways, roads, ports, bridge abutment etc.
It has not been uncommon that during post earthquake relief operations
39
, relief
could not be sent in time, for damage to retaining wall itself has cut off the link
38 For details of these methods refer any standard text book on soil mechanics.
39 Specially in mountainous region.
road or tilting of the same has resulted in secondary collapse of bridge girders thus
making some part of relief area inaccessible thus compounding the problems of the
relief workers who have anyway a tough job to execute.
Thus, understanding the behaviour of such structures under earthquake load is
of paramount importance to civil engineers undertaking such tasks. It is important at
times that such structures remain functional even after a severe earthquake maintaining
the vital link among two places.
One of the earliest methods for earthquake analysis of such retaining wall has been
proposed by Mononobe (1926) and Okabe (1926) which still remain the backbone of
analysis in almost all design office and code of practice.
3.6.2 Mononobes method of analysis of retaining wall
Mononobe developed a pseudo static analysis for estimation of earth pressure on a
gravity type retaining wall during an earthquake.
He basically extended the original Coulombs theory to develop the soil pressure
behind a retaining wall (Figure 3.6.1) catering to the horizontal and vertical component
of the ground acceleration.
Coulomb (1776) derived the equation of active and passive earth pressure as
P
a
=
1
2
K
A
H
2
and P
p
=
1
2
K
P
H
2
, where (3.6.1)
K
A
=
cos
2
( )
cos
2
cos( + )
_
1 +
_
sin(+) sin(i)
cos(+) cos(i)
_1
2
_
2
= coefficient of active earth pressure, and
K
P
=
cos
2
( + )
cos
2
cos( )
_
1
_
sin(+) sin(+i)
cos() cos(i)
_1
2
_
2
= coefficient of passive earth pressure,
in which, = unit weight of soil; H = height of the retaining wall; = friction angle
of the soil; = angle of friction between the wall and soil; = angle subtended with
the back of the wall with vertical, and i = slope of soil with respect to the horizontal.
Mononobe and Okabe modified the above coefficient of earth pressure considering
the horizontal and vertical seismic coefficient earthquake force
h
and
v
to
K
AE
=
cos
2
( )
cos cos
2
cos( + + )
_
1 +
_
sin(+) sin(i)
cos(++) cos(i)
_1
2
_
2
(3.6.2)
C
A
i
h
W
v
W
W
H
N
Pa
Fr
F
B
Figure 3.6.1 Mononobes force diagram for gravity type retaining wall under earthquake.
where, = tan
1
_

h
1
v
_
similarly the coefcient of passive earth pressure under
earthquake is given by
K
PE
=
cos
2
( + )
cos cos
2
cos( + )
_
1
_
sin(+) sin(+i)
cos(+) cos(i)
_1
2
_
2
(3.6.3)
Mononobes solution was based on the following assumptions:
The failure in soil takes place along the plane BC
The movement of the wall is sufficient to produce active plane
At failure full shear strength along the failure plane is mobilized
The soil mass behind the wall behaves a rigid body

h
and
v
are maximum peak acceleration in horizontal and vertical direction.
It will be observed that in the above calculation no reference to time period is made,
which clearly shows the analysis to be pseudo static in nature.
The
h
and
v
values are considered as peak acceleration is quite justified in this
case because of the following two reasons.
Response spectrum as per Japanese Code for Medium soil
0.000
0.200
0.400
0.600
0.800
1.000
1.200
0
0
.
2
0
.
4
0
.
6
0
.
8 1
1
.
2
1
.
4
1
.
6
1
.
8 2
Time period in second
S
a
/
g
Figure 3.6.2 Response spectrum for medium type soil as per Japanese code.
At the time when Mononobe worked out the solution, major retaining walls were
gravity type wall being massive was considered to be infinitely stiff i.e. have time
period T 0. This invariably makes the structure stiff attracting more force to it.
As per Japanese code (Figure 3.6.2), for time period up to nearly 0.8 second, it is
a practice to consider maximum acceleration on the structure.
Mononobe considered the dynamic pressure distribution as parabolic in nature and
for active case it was assumed that the force is acting at a height H/2 from the bottom
of the wall. While for passive case again considering a parabolic distribution the force
was assumed to be acting at a height of 2/3 H from the base of the wall.
The above procedure has almost universally dominated the design office practice
in almost all countries
40
though characteristics of retaining wall have undergone
signicant change with advent of Reinforced Concrete structures.
While in the early thirties when Mononobe developed this theory almost all retaining
walls were either made of Masonry or plain concrete and were usually massive in
nature. It sustained its stability by its own weight. However with advancement in
design techniques in RCC, they have progressively become much thinner and fit for
the purpose, globally.
Presently, the common practice is to go for a cantilever retaining wall for a height
up to 6.0 meter beyond which engineers normally provide counterfort retaining wall
which would provide overall economy in design. The typical sketches of such walls
are as shown in Figure 3.6.3.
It is obvious from the above figure that present day retaining walls have become
much more flexible as such considering maximum acceleration as inducing the design
pressure may not be true in all cases. Moreover, considering its flexibility the time
period will surely play a dominant role in dynamic response of the system where a
pseudo static analysis only could become insufficient.
In-spite of the above limitation Mononobes pseudo static method is in extensive
use irrespective of whether the wall is a cantilever or a counterfort or if it is thin or
thick.
40 Including India where IS code follows the same procedure.
Cantil ever retaining wall Counter fort retaining wall
Figure 3.6.3 Present day cantilever and counterfort retaining walls made of RCC.
Before we try tackle the above limitations as posed in Mononobes method we
describe herein other methods which are also in practice for evaluation of pressure
due to earthquake force.
3.6.3 Seed and Whitmans method
Seed and Whitman (1970) developed an expression which may also be used to deter-
mine the horizontal pseudo static force acting on a retaining wall. As per their
calculations the active earth pressure may be expressed as
P
AE
=
3
8
a
max
g

s
H
2
(3.6.4)
According to this work the location pseudo static force is assumed to act at height
of 0.6 H above the base of the structure.
3.6.4 Arangos method
Arango
41
(1969) developed this method which is deemed practical and simple by
extending the Coulombs theory of static earth pressure. As mentioned previously the
41 Dr. Ignacio Arango is Head of the Geotechnical division of Bechtel San Francisco office. A Bechtel
fellow, who has contributed significantly in many areas of Geotechnical engineering specially in the area
of Liquefaction Potential of soil.
static pressure is expressed as
P
a
=
1
2
K
A
H
2
where, (3.6.5)
K
A
=
cos
2
( )
cos
2
cos( + )
_
1 +
_
sin(+) sin(i)
cos(+) cos(i)
_1
2
_
2
The active earth pressure can be expressed in the form
P
a
=
1
2
1
cos
2
H
2
A
c
, (3.6.6)
where, A
c
=
cos
2
( )
cos( + )
_
1 +
_
sin(+) sin(i)
cos(+) cos(i)
_1
2
_
2
= K
A
cos
2
Mononobes expression on the other hand is given by

P
AE
=
1
2
H
2
(1
v
)K
AE
=
1
2
H
2
(1
v
)
1
cos cos
2
A
m
where, (3.6.7)
A
m
= K
AE
cos cos
2
=
cos
2
( )
cos( + + )
_
1 +
_
sin(+) sin(i)
cos(++) cos(i)
_1
2
_
2
Comparing A
m
and A
c
it shows that A
m
can be determined from the solution for
A
c
by redefining the slope of the back of wall as

where,

= + and

i = i +
Thus, A
m
= A
c
(

,
i) = K
A
(

,
i) cos
2

and
P
AE
=
1
2
H
2
(1
v
)
1
cos cos
2
K
A
(

,
i) cos
2

=
1
2
H
2
(1
v
)K
A
(

,
i)F where, F =
cos
2

cos cos
2
. (3.6.8)
So far we have explained the various pseudo-static methods available for analysis
for retaining wall.
The first attempt to explain dynamic characteristics of such wall was proposed by
Steedman and Zeng (1990) based on a pseudo-dynamic method.
3.6.5 Steedman and Zengs method
For the fixed base cantilever as shown in Figure 3.6.4, if subjected to ground
acceleration a
h
the acceleration at depth z is given by
a(z, t) = a
h
sin
_
_
t
H z
v
s
__
(3.6.9)
Considering the pressure on the wall is resulting from the triangular wedge only
being at a state of incipient failure, mass of a thin strip of depth dz within the soil
wedge is given by
m(z, t) =
(H z)
g tan
dz where, is the unit weight of soil. (3.6.10)
The total inertial force P
h
acting on the wall can thus be expressed as
P
h
(t) =
H
_
0
m(z)a(z, t) =
a
h
4
2
g tan
[2H cos + (sin sin t)] (3.6.11)
in which, =
2
v
s
is the vertically propagating shear wave length and = t

H
v
s
.
For a special case when the failed wedge act as a rigid block (i.e. v
s
), we have
[P
h
(max)]
limv
s
=
H
2
a
h
2g tan
=
a
h
g
W =
y
W (3.6.12)
The above can be stated to be equivalent force as assumed in the Mononobes
method.
P
h
z
W H
Ps
/ 2 45 =
Figure 3.6.4 Force diagram of cantilever retaining wall as per Steedman and Zengs (1990) method.
The total static plus dynamic thrust can then be obtained by resolving the force in
the wedge and is given by
P
AE
=
P
h
(t) cos( ) +W sin( )
cos( + )
(3.6.13)
The total earth pressure is obtained by differentiating above expression with respect
to z thus
p
AE
=
P
AE
z
=
z
tan
sin( )
cos( + )
+

h
z
tan
cos ( )
cos ( + )
sin
_
_
t
z
v
s
__
(3.6.14)
The first term in the above equation represents the static pressure acting at a height
of H/3 from base while the second term represents the dynamic pressure where the
thrust point is found to be varying with time and is given by
h
d
= H
2
2
H
2
cos +2H sin
2
(cos cos t)
2H cos + (sin sin t)
(3.6.15)
For very low frequency motion the dynamic thrust is found to at H/3 (this is when
H/ is small and the backfill moves in the same phase). For higher frequency motions
the point of application is found to move higher up on the wall.
3.6.6 Dynamic analysis of RCC retaining wall
So far we have explained the various pseudo static and pseudo dynamic methods avail-
able for analysis of retaining walls under earthquake force. We have also explained
briefly some of the limitations of the pseudo-static method which is perhaps more
appropriate for gravity type of wall. We now present herein a procedure for dynamic
response of cantilever and counter fort type of retaining wall.
While studying the dynamic response of structure previously we had seen that the
acceleration it would be subjected to depends on the time period of the structure, then
considering a suitable damping ratio we can find out the earthquake response based
on code prescribed response spectrum.
Based on the above, we proceed to explain the investigation which has been carried
out by the present authors.
3.6.7 Dynamic analysis of cantilever and counterfort
retaining wall
This method applies standard modal response technique based on Rayleigh Ritz
Method to evaluate the dynamic response of the soil wall system (Chowdhury and
Dasgupta 2004). Though the response herein has been obtained based on IS-code
A B
Possible Failure Line
H Wall
= 45 / 2
D
X
Figure 3.6.5 Cantilever retaining wall with soil behind it.
(IS-1893 2002 to be specic) can very well be adapted to any international code
provided the response history of the site in question is available either directly or
as prescribed in the national code of practice (like UBC, Eurocode etc), We start with
a simple case of retaining wall (Figure 3.6.5) with soil prole as shown above. It is
assumed here like Mononobes case that
the soil profile under active case is at incipient failure when the failure line makes
angle = tan(45 + /2) as shown in the above figure.
Since soil profile is already under failed condition under static load, the soil will
not induce any stiffness in the overall dynamic response but will only contribute
to the inertial effect.
Since the cantilever wall is relatively thin the mass contribution of the wall itself
may be ignored compared to that of the soil. The wall thus contributes only to the
stiffness of the overall soil-structure system.
The retaining wall is fixed at the base and foundation compliance has been ignored
for the time being.
It will be observed that the assumptions made are identical to what Mononobe or
Steedman and Zang has assumed in their analysis. Based on the above assumption the
mass distribution of soil along the height of the wall is as shown hereafter.
As shown in Figure 3.6.6, is the mass distribution of the failed wedge ABD. For an
elemental strip dz in vertical direction mass distribution is given by
m(z) =
zdz
g tan
(3.6.16)
Z
A B
=
g tan
zdz
) z ( m ;
2
= 45 +
D
Y
Figure 3.6.6 Mass distribution of the failed soil wedge under active soil pressure.
For the analysis of time period as a first step, we develop the stiffness and equivalent
mass contributing to the dynamic response of the system.
For this we use the Rayleigh Ritz method to obtain the stiffness and mass of the
wall-soil system.
We had already shown earlier while deriving the time period of chimneys that for a
flexural beam the stiffness and mass matrix may be obtained from the expressions
m
ij
=
_
_
L
_
0
m(z)
i
(z)
j
(z)dz
_
_
for i, j = 1, 2, 3. . . n and,
k
ij
=
L
_
0
EI(z)
d
2
i
(z)
dz
2
d
2
j
(z)
dz
2
dz for i, j = 1, 2, 3. . . n (3.6.17)
where m(z) is as defined above; E = Youngs modulus of the wall material, and I(z) =
Average moment of inertia of the wall.
Considering the wall as a cantilever flexural member the displacement y(z, t) can
represented trigonometrically by
Y
m
= sin
m
z
H
sin h
m
z
H

m
_
cos
m
z
H
cos h
m
z
H
_
(3.6.18)
where, m = number of modes 1, 2, 3, . . . ;
m
= 1.875, 4.694,7.855,
2m1
2
, for
m = 1, 2, 3. . . m, and
m
=
sin
m
+sin h
m
cos
m
+cos h
m
.
Thus for failure wedge of the soil, the mass contribution for an element dz for the
first mode is given by
[M]
ij
=
H
_
0
z
g tan
i
(z)
j
(z)dz =

g tan
H
_
0
zi
i
(z)
j
(z)dz (3.6.19)
Considering, the shape function as
i
= sin

i
z
H
sin h
i
z
H

i
_
cos
i
z
H
cos h
i
z
H
_
and
j
= sin
j
z
H
sin h
j
z
H

j
_
cos
j
z
H
cos h
j
z
H
_
. (3.6.20)
The double derivative of the above is given by
i
=

2
i
H
2
_
sin
i
z
H
sin h
i
z
H
+
i
_
cos
i
z
H
+cos h
i
z
H
__
and
j
=
2
j
H
2
_
sin
j
z
H
sin h
j
z
H
+
j
_
cos
j
z
H
+cos h
j
z
H
__
(3.6.21)
Before performing the integration we change the above to generalized co-ordinate
by considering, =
z
H
when, d =
dz
H
and as z 0, 0 and as z H, 1.
Based on the above, we can now express the double derivative as
f
()
i
=

2
i
H
2
_
sin
i
sin h
i
+
i
_
cos
i
+cos h
i
__
and
f
()
j
=
2
j
H
2
_
sin
j
sin h
j
+
j
_
cos
j
+cos h
j
__
(3.6.22)
k
ij
=
EI
2
i

2
j
H
3
1
_
0
f
()
i
f
()
j
d (3.6.23)
and, mass of the system is given by
m
ij
=
H
2
g tan
1
_
0
f ()
i
f ()
j
d where i, j = 1, 2, 3, . . . . . . . . .m. (3.6.24)
where = unit weight of soil; g = acceleration due to gravity; E = Youngs modulus
of the wall material, and, I = moment of inertia of the wall.
Thus for the first three modes the stiffness matrix
42
is given by
[K]
ij
=
EI
H
3
_
_
_
_
_
_
_
_
_
_
_
4
1
1
_
0
f
()
2
1
d
2
2
2
1
1
_
0
f
()
2
f
()
1
d
4
2
1
_
0
f
2
()
2
d
2
3
2
1
1
_
0
f
()
3
f
()
1
d
2
3
2
2
1
_
0
f
()
3
f
()
2
d
4
3
1
_
0
f
()
2
3
d
_
_
[M]
ij
=
H
2
g tan
_
_
_
_
_
_
_
_
_
_
1
_
0
f ()
2
1
d
1
_
0
f ()
2
f ()
1
d
1
_
0
f
2
()
2
d
1
_
0
f ()
3
f ()
1
d
1
_
0
f ()
3
f ()
2
d
1
_
0
f ()
2
3
d
_
_
(3.6.25)
The above integrals can very easily be solved based on Numerical analysis between
the limits 1 to 0 when we have
[K]
33
=
EI
H
3
_
_
22.936 0.002 0.006
0.002 468.044 0.11
0.006 0.11 3812.81
_
_
[M]
33
=
H
2
g tan
_
_
1.496 0.205 0.028
0.205 0.573 0.188
0.028 0.188 0.533
_
_
(3.6.26)
generalized Jacobi technique
43
we have
[] =
EIg tan
H
5
_
_
15.317 0 0
0 846.507 0
0 0 8235
_
_
(3.6.27)
42 The matrix is symmetric about is diagonal.
43 The technique has been worked in detail in Chapter 5 (Vol. 1).
and the corresponding eigen vectors are given as
[] =
_
_
1 0.139 0.031
6.833 10
3
0.989 0.353
1.161 10
4
0.046 0.935
_
_
_
_
_
f
1
()
f
2
()
f
3
()
_
_
_
(3.6.28)
since [] =
2
and T =
2
we have
[T] =
_
_
1.6049 0 0
0 0.2198 0
0 0 0.0743
_
_
_
H
5
EIg tan
(3.6.29)
Considering moment of inertia of the wall as I =
1
12
t
3
having one meter width we
can modify the fundamental time period as
[T] =
_
_
5.556 0 0
0 0.761 0
0 0 0.257
_
_
_
H
5
Et
3
g tan
(3.6.30)
In terms of response spectrum analysis maximum displacement amplitude S
d
is given
by S
d
=
S
a
2
, expressing the above in terms of codal formulation we may express it as
S
d
=
i
ZI
2R
S
a
2
(3.6.31)
where,
i
= modal participation factor and is given by
i
=
_
_
H
_
0
m
i
i
d
_
H
_
0
m
i
2
i
d
_
_
=
_
_
1
_
0
f
i
()d
_
1
_
0
f
i
()
2
d
_
_
(3.6.32)
in which, Z = zone coefficient; I = importance factor, and, R = ductility factor.
Integration of the mass participation factor within limits 1 to 0 for the first three
modes gives the values as shown in Table 3.6.1.
Now considering, =
ZI
2R
an IS-code factor we can write the time dependent
function of displacement as,
S
d
=
i
S
a
2
. (3.6.33)
Thus, for the first mode, we have
S
d
= 0.619
S
a1
H
5
1
EIg tan
(3.6.34)
Table 3.6.1 Modal participation factor for cantilever walls for the first
three modes.
Mode number Mass participation factor (k
i
)
1 0.619
2 0.224
3 0.09
Substituting the value of = 15.317, calculated earlier, we have
S
d
= 0.485
S
a1
H
5
Et
3
g tan
(3.6.35)
where t = the average thickness of the wall.
The complete displacement function is thus given by
w(z, t) = (z) q(t) and for first mode
w(z, t) = 0.5387
S
a1
H
5
Et
3
g tan
[f
1
() 6.833 10
3
f
2
() +1.161 10
4
f
3
()]
(3.6.36)
Ignoring the effect of second and third mode which is found very small
w(z, t) = 0.485
S
a1
H
5
Et
3
g tan
f
1
() (3.6.37)
Considering, EI
d
2
w
dz
2
= M
z
, we get
M(z, t) = 0.0404
S
a1
H
3
g tan
[
2
1
f
1
)]; V(z, t) = 0.0404
S
a1
H
2
g tan
[
3
1
f
1
()]
(3.6.38)
The above can now be represented as
w(z, t) = Coeffd
i

S
a1
H
5
Et
3
g tan
; M(z, t) = Coeffm
i

S
a1
H
3
g tan
V(z, t) = Coeffv
i

S
a1
H
2
g tan
(3.6.39)
where, Coeffd
i
, Coeffm
i
and Coeffv
i
are the dynamic amplitude, moment and shear
coefficients for the first three mode (i = 1, 2, 3).
Table 3.6.2 Factors of dynamic amplitude moments and shears for cantilever retaining wall.
Z/H coeff d
1
coeff d
2
coeff d
3
Coeff m
1
Coeff m
2
Coeff m
3
Coeff v
1
Coeff v
2
Coeff v
3
0 0.00000 0.00000 0.00000 0.38681 0.01143 0.00135 0.53200 0.05400 0.01062
0.05 0.00567 0.00016 0.00002 0.36018 0.00870 0.00082 0.53190 0.05389 0.01052
0.1 0.02216 0.00058 0.00006 0.33358 0.00599 0.00031 0.53123 0.05320 0.00993
0.15 0.04866 0.00117 0.00011 0.30703 0.00333 0.00016 0.52945 0.05147 0.00857
0.2 0.08437 0.00187 0.00016 0.28061 0.00080 0.00053 0.52607 0.04843 0.00640
0.25 0.12851 0.00260 0.00019 0.25441 0.00154 0.00079 0.52063 0.04396 0.00360
0.3 0.18029 0.00327 0.00020 0.22854 0.00362 0.00089 0.51272 0.03807 0.00048
0.35 0.23893 0.00384 0.00018 0.20313 0.00537 0.00083 0.50197 0.03093 0.00258
0.4 0.30367 0.00425 0.00014 0.17835 0.00673 0.00064 0.48805 0.02280 0.00514
0.45 0.37376 0.00446 0.00008 0.15434 0.00767 0.00033 0.47065 0.01405 0.00688
0.5 0.44849 0.00444 0.00001 0.13130 0.00815 0.00003 0.44953 0.00510 0.00754
0.55 0.52717 0.00418 0.00006 0.10941 0.00819 0.00040 0.42447 0.00356 0.00707
0.6 0.60913 0.00367 0.00012 0.08888 0.00780 0.00072 0.39527 0.01145 0.00556
0.65 0.69376 0.00293 0.00016 0.06992 0.00705 0.00094 0.36178 0.01809 0.00328
0.7 0.78050 0.00198 0.00017 0.05274 0.00600 0.00104 0.32389 0.02303 0.00063
0.75 0.86883 0.00085 0.00015 0.03758 0.00476 0.00101 0.28149 0.02589 0.00193
0.8 0.95829 0.00043 0.00010 0.02464 0.00342 0.00086 0.23452 0.02633 0.00390
0.85 1.04849 0.00180 0.00002 0.01418 0.00214 0.00063 0.18293 0.02411 0.00484
0.9 1.13913 0.00324 0.00008 0.00641 0.00103 0.00040 0.12669 0.01908 0.00440
0.95 1.22997 0.00472 0.00019 0.00158 0.00026 0.00022 0.06578 0.01112 0.00232
1 1.32086 0.00620 0.00030 0.00010 0.00004 0.00020 0.00021 0.00020 0.00153
It will however be shown based on a numerical problemsubsequently that moments
and shears developed in the wall for the higher modes (i = 2, 3, etc.) have insignificant
contribution i.e. their SRSS values almost same as the moment, shear and amplitude
values for the first mode.
Table 3.6.2 showing values of these coefficients are furnished for reference.
It is to be understood that once the time period and subsequent acceleration value
S
a
/g is known from the code one has to only multiply the displacement, moment, and
shear expressions by the values in Table 3.6.2 to arrive at the dynamic design data
without resorting to any elaborate dynamic analysis.
Analysis for passive case
For passive case the analysis remains the same except the fact that the angle becomes
tan 45 /2 in lieu of what has been shown earlier.
We now explain the above theory based on a suitable numerical problem.
Example 3.6.1
A retaining wall of height 5.8 m has top thickness of wall as 200 mm and
bottom thickness as 500 mm. The unit weight of soil it retains has a value of
22 kN/m
3
. The angle of friction of soil is 28
. the unit weight of the concrete

wall is considered as 25 kN/m
3
. Consider the wall is in zone IV as per IS-code
(1893, 2002) and is resting on Hard soil. Determine the earthquake force acting
on the wall.
Solution:
Based on the above theory,
Average thickness of wall =
200+500
2
= 350 mm; Unit weight of concrete =
25 kN/m
3
; Thus, Youngs modulus of concrete = 5700
_
f
ck
1000 kN/m
2
=
2.85 10
7
kN/m
2
.
Angle (active case) = 45+
1
2
28 = 59
; Angle (passive case) = 45

1
2

28 = 31
.
As per IS-code consider Importance factor as 1.0 and ductility factor as
R = 2.0.
Considering =
ZI
2R
we have, =
0.241.0
22
= 0.06
Now considering
[T] =
_
_
5.556 0 0
0 0.761 0
0 0 0.238
_
_
_
H
5
Et
3
g tan
, we have
[T] =
_
_
5.556 0 0
0 0.724 0
0 0 0.238
_
_
_
22 5.8
5
2.85 10
7
0.35
3
9.81tan 59
which gives
{T} =
_
_
_
0.473
0.065
0.022
_
_
_
secs, for active case for the first three modes.
Similarly for passive case considering = 31 degree, we have
[T] =
_
_
5.567 0 0
0 0.724 0
0 0.238
_
_
_
22 5.8
5
2.85 10
7
0.35
3
9.81tan 31
, which gives
{T} =
_
_
_
0.787
0.108
0.036
_
_
_
sec for passive case
Corresponding to the time periods for the first three modes the S
a
/g values for
active and passive case are shown hereafter. The values are obtained considering
7% damping for the RCC wall i.e. codal values scaled by a factor 0.9.
S
a
/g (passive case)
Mode S
a
/g (active case) m/sec
2
1 1.9 1.144
2 1.773 2.25
3 1.173 1.355
The displacement shear and moment factors are as shown below:
w(z, t) =
S
a
H
5
Et
3
g tan
, M(z, t) =
S
a1
H
3
g tan
and V(z, t) =
S
a1
H
2
g tan
For the first three modes, we have
Amplitude coefficient Amplitude coefficient
Mode (active case) (passive case)
1 0.00809 0.013499
2 0.00755 0.02655
3 0.00499 0.015992
Moment coefficient Moment coefficient
1 294.46 490.06
2 274.61 964.421
3 185.02 596.64
Shear coefficient Shear coefficient
1 50.768 84.493
2 47.345 166.279
3 31.9 102.87
Multiplying the above factors by the coefcients as furnished earlier in
Table 3.6.2, the dynamic displacement moment and shear for active and passive
case are obtained as shown below
Dynamic amplitude (mm), moment (kN m) and shear (kN) for active case
Z/H D
1
D
2
D
3
M
1
M
2
M
3
V
1
V
2
V
3
0.000 0.000 0.000 0.000 113.899 3.138 0.250 27.009 2.557 0.339
0.050 0.046 0.001 0.000 106.060 2.388 0.152 27.004 2.552 0.336
0.100 0.179 0.004 0.000 98.225 1.644 0.057 26.970 2.519 0.317
0.150 0.394 0.009 0.001 90.408 0.916 0.029 26.879 2.437 0.273
0.200 0.683 0.014 0.001 82.628 0.220 0.099 26.708 2.293 0.204
0.250 1.040 0.020 0.001 74.913 0.423 0.145 26.431 2.081 0.115
0.300 1.459 0.025 0.001 67.295 0.994 0.164 26.030 1.803 0.015
0.350 1.933 0.029 0.001 59.815 1.475 0.154 25.484 1.464 0.082
0.400 2.457 0.032 0.001 52.516 1.849 0.118 24.777 1.080 0.164
0.450 3.024 0.034 0.000 45.447 2.105 0.062 23.894 0.665 0.219
0.500 3.628 0.034 0.000 38.662 2.238 0.006 22.822 0.242 0.241
0.550 4.265 0.032 0.000 32.217 2.248 0.074 21.549 0.169 0.225
0.600 4.928 0.028 0.001 26.172 2.143 0.133 20.067 0.542 0.177
(continued)
Z/H D
1
D
2
D
3
M
1
M
2
M
3
V
1
V
2
V
3
0.650 5.613 0.022 0.001 20.589 1.936 0.175 18.367 0.857 0.105
0.750 7.029 0.006 0.001 11.065 1.306 0.186 14.291 1.226 0.062
0.800 7.753 0.003 0.000 7.257 0.940 0.159 11.906 1.247 0.124
0.850 8.482 0.014 0.000 4.175 0.587 0.117 9.287 1.142 0.154
0.900 9.216 0.024 0.000 1.888 0.284 0.074 6.432 0.903 0.140
0.950 9.950 0.036 0.001 0.464 0.070 0.041 3.340 0.526 0.074
1.000 10.686 0.047 0.002 0.028 0.012 0.036 0.010 0.010 0.049
Dynamic amplitude (mm), moment (kN m) and shear (kN) for passive case
Z/H D
1
D
2
D
3
M
1
M
2
M
3
V
1
V
2
V
3
0.000 0.000 0.000 0.000 189.559 11.020 0.807 44.954 8.980 1.092
0.050 0.077 0.004 0.000 176.512 8.388 0.491 44.946 8.962 1.082
0.100 0.299 0.016 0.001 163.473 5.773 0.185 44.889 8.846 1.021
0.150 0.657 0.032 0.002 150.463 3.215 0.093 44.739 8.559 0.881
0.200 1.139 0.051 0.003 137.515 0.773 0.318 44.453 8.054 0.659
0.250 1.735 0.070 0.003 124.675 1.486 0.469 43.993 7.310 0.371
0.300 2.434 0.088 0.003 111.997 3.491 0.530 43.325 6.332 0.049
0.350 3.226 0.104 0.003 99.548 5.179 0.498 42.417 5.144 0.265
0.400 4.100 0.115 0.002 87.401 6.493 0.381 41.240 3.792 0.529
0.450 5.046 0.121 0.001 75.637 7.393 0.200 39.770 2.336 0.707
0.500 6.055 0.120 0.000 64.345 7.860 0.018 37.986 0.849 0.776
0.550 7.117 0.113 0.001 53.618 7.896 0.239 35.867 0.592 0.727
0.600 8.223 0.099 0.002 43.557 7.525 0.430 33.400 1.904 0.572
0.650 9.366 0.079 0.003 34.265 6.799 0.563 30.570 3.009 0.337
0.700 10.537 0.053 0.003 25.847 5.788 0.621 27.368 3.830 0.064
0.750 11.729 0.023 0.002 18.414 4.586 0.601 23.786 4.305 0.199
0.800 12.937 0.012 0.002 12.077 3.302 0.512 19.817 4.379 0.401
0.850 14.155 0.049 0.000 6.948 2.061 0.379 15.458 4.010 0.498
0.900 15.378 0.088 0.001 3.142 0.996 0.237 10.705 3.172 0.452
0.950 16.605 0.127 0.003 0.772 0.248 0.133 5.559 1.849 0.239
1.000 17.832 0.167 0.005 0.047 0.041 0.116 0.017 0.034 0.158
If we compare the SRSS value with that of the first mode the values compared
as given hereafter.
Z/H Moment (1st mode) Moment (SRSS) Shear first mode Shear SRSS
0.000 189.559 189.881 44.954 45.855
0.050 176.512 176.712 44.946 45.843
0.100 163.473 163.575 44.889 45.764
0.150 150.463 150.497 44.739 45.559
0.200 137.515 137.518 44.453 45.181
0.250 124.675 124.685 43.993 44.598
0.300 111.997 112.053 43.325 43.785
0.350 99.548 99.684 42.417 42.728
(continued)
Z/H Moment (1st mode) Moment (SRSS) Shear first mode Shear SRSS
0.400 87.401 87.642 41.240 41.417
0.450 75.637 75.997 39.770 39.845
0.500 64.345 64.823 37.986 38.003
0.550 53.618 54.197 35.867 35.880
0.600 43.557 44.205 33.400 33.459
0.650 34.265 34.937 30.570 30.720
0.700 25.847 26.495 27.368 27.635
0.750 18.414 18.986 23.786 24.173
0.800 12.077 12.531 19.817 20.299
0.850 6.948 7.258 15.458 15.977
0.900 3.142 3.305 10.705 11.175
0.950 0.772 0.822 5.559 5.863
1.000 0.047 0.132 0.017 0.162
It is observed that the difference in the values are insignificant as such for these
type of structure performing a first mode analysis based on the fundamental time
period only should suffice for practical designs.
3.6.8 Some discussions on the above method
The method as shown above is an approximate method for evaluation of the time
period of the structure
44
, however is far more realistic then considering a maximum
value of 0.9 2.5 = 2.25 m/sec
2
, which would result in significant over design of the
retaining wall thus making it a more expensive proposal. Another major advantage
is that unlike the pseudo static or dynamic case one need not estimate the point of
application of this load, the moments and shear expression are directly derived from
the amplitude expression.
Based on the above numerical problem it is seen that fundamental time period
is the most signicant mode. Thus arriving at a closed form expressions for more
complex cases like soil sloped at an angle or a counter fort retaining wall where
the wall acts like a plate with appropriate boundary conditions become much
simpler
45
.
3.6.9 Extension to the generic case of soil at a slope i
behind the wall
In this case, shown in Figure 3.6.7, we need to take the additional soil mass of the
triangle ABE converted it into a form so that it fits into the integration scheme we had
proposed. Thus for having same mass as triangle ABE we take en equivalent trapezium
44 We should remember that the integrations performed where based on numerical analysis and not a
closed form one as such errors due to truncation is surely to be expected.
45 This we will see subsequently as we take them up subsequently.
C E
O
B
H
A
D
Figure 3.6.7 Retaining wall with soil sloped at an angle behind the wall.
ABECequal to triangular area ABE, when equating the area of ABECand triangle ABE
it can be proved that
AC = H(
_
1 + 1) where =
cos sin i
sin( i)
Now considering H
= H +AC = H
1 +
[M]
ij
=
H
_
0
z
g tan
i
(z)
j
(z)dz =

g tan
H
_
0
zi
i
(z)
j
(z)dz
m
ij
=
H
2
g tan
1
_
0
f ()
i
f ()
j
d where i, j =1, 2, 3, . . . . . . . . . m
in natural co-ordinate
=
H
2
(1 + )
2
g tan
1
_
0
f ()
i
f ()
j
d where i, j = 1, 2, 3, . . . . . . . . .m. (3.6.40)
Since we had seen earlier that the first mode analysis suffice for these type of structure
restricting the above expansion to the first mode, we have
m
1
=
H
2
(1 + )
2
g tan
1
_
0
f ()
2
1
d and k
1
=
EI
4
1
H
3
1
_
0
f
()
2
i
d (3.6.41)
Considering =
2
and T =
2
, we have
Fundamental time period as
[T] = 5.556(1 + )
_
H
5
Et
3
g tan
(3.6.42)
Based on above we can find out the fundamental time period of the system and
subsequently the expression S
a
/g from the code for an appropriate damping of the
system.
It will be observed that for = 0 the time period value reduces to the fundamental
time period for the case where the soil plane is considered parallel to the base.
The modal mass participation is given by the expression
1
=
H
_
0
z
g tan
1
(z)dz/
H
_
0
z
g tan
2
1
dz (3.6.43)
1
=
1
_
0
f
1
()dz/
1
_
0
f
2
1
dz and remains the same as in the earlier case.
Thus, for the fundamental mode, S
d
= 0.619
S
a1
H
5
(1+ )
2
1
EIg tan
.
Substituting the value of we have
S
d
= 0.485
S
a1
H
5
(1 + )
2
Et
3
g tan
(3.6.44)
the complete displacement function is thus given by
w(z, t) = (z) q(t) which gives, w(z, t) = 0.485
S
a1
H
5
(1 + )
2
Et
3
g tan
F
1
()
M(z, t) = 0.0404
S
a1
H
3
(1 + )
2
g tan
2
1
F
1
()
V(z, t) = 0.0404
S
a1
H
2
(1 + )
2
g tan
3
1
F
1
() where, the value
1
= 1.875
(3.6.45)
Plan view of counter fort retaining wall Side view of the wall
Figure 3.6.8 Plan view and elevation of counterfort retaining wall.
The coefficients are as furnished in Table 3.6.2.
3.6.10 Dynamic analysis of counterfort
retaining wall
When retaining walls are more then 6.0 meter high normally for economic reason
counterforts are provided to reduce the moment in the supporting wall. In such cases
unlike cantilever, the wall does not behave as pure flexural member but behaves as
plate having three sides fixed and one side free.
To the best of our knowledge there exists no solution to this problem in terms of
earthquake induced dynamic force.
Present state of art is to develop the additional pressure based on Mononobes coef-
cient add it too the static pressure (hydrostatic in nature) and solve for the moments
and shears based on coefcients as furnished in IS-3370 Part IVor similar literature. In
other words neither the time period nor the spatial distribution of the load is accounted
for, considering its mode shape which again is a function of its boundary condition
like free, fixed etc.
For such walls when subjected to force due to an earthquake solutions have to
be sought based on plate vibration in lieu of beam vibration as shown previously
and the solution becomes downright tricky. However, since we have seen earlier
that it is fundamental mode which basically governs the design with a little bit of
intelligent mathematical manipulation it is not too difcult to arrive at the time
period vis-a-vis dynamic amplitude shear and moments for such type of retaining
wall.
A typical counterfort retaining wall is shown in Figure 3.6.8.
It is evident that the wall between the buttress in this case will behave as plate instead
of a beam having boundary condition of three sides continuous and one side free.
Shown in Figure 3.6.9, is the wall of a counterfort retaining wall of height H and
width b. Also, b is the dimension between the two consecutive counterforts. The failed
soil mass under active and passive pressure is shown by the triangular wedge which
contributes its inertial effect on the wall.
H
Fixed edge(typ.)
b
Figure 3.6.9 Wall with three edges fixed with soil mass under earthquake.
Here the displacement function of the wall can be represented by
w(z, y, t) =
n
i=1
i
(z, y)q
i
(t) (3.6.46)
The kinetic energy of the system is given by
T =
1
2
H
_
0
b
_
0
m(z, y)
_
w(z, y, t)
t
_
2
Substituting the displacement function in above equation we have
T =
1
2
H
_
0
b
_
0
m(z, y)
_
n
i=1
i
(z, y) q
i
(t)
_
_
_
n
j=1
j
(z, t) q
j
(t)
_
_
=
n
i=1
n
j=1
q
i
(t) q
j
(t)
_
_
_
H
_
0
b
_
0
m(z, y)
i
(z, y)
j
(z, y)
_
_
(3.6.47)
From above we deduce that the mass coefficient is given by
m
ij
=
_
_
_
H
_
0
b
_
0
m(z, y)
i
(z, y)
j
(z, y)
_
_
dy dz for i, j = 1, 2, 3. . . . . . . n.
The strain energy equation of the plate is given by (Timoshenko and Woinowsky-
Krieger 1958)
V =
1
2
H
_
0
b
_
0
D(z, y)
_
_
_
2
w
z
2
+

2
w
y
2
_
2(1 )
_
2
w
z
2
2
w
y
2

_
2
w
zy
_
2 _
_
_
(3.6.48)
where, w(z, y, t) =
n
i=1
i
(z, y)q
i
(t).
It can be proved based on Lagranges equation that on differentiation of the above
equation with respect to q
j
(t) the stiffness coefficient of an isotropic plate is given by
Kij =
H
_
0
b
_
0
D
__
2
i
i
z
2
+

2
j
y
2
__
i
y
2
+

2
j
z
2
_
(1 )
_
i
z
2
j
y
2
_
2
i
zy
j
zy
_
dy dz (3.6.49)
where, D =
Et
3
12(1
2
)
the flexural stiffness of thin plate; E = Youngs modulus of
concrete; t = thickness of the wall; = Poissons Ratio of concrete, and
i
,
j
=
mode shapes at different modes whose values have to be chosen for the appropriate
boundary condition and shall be a function of both z and y.
To select the appropriate mode shape let us imagine a thin strip of element in vertical
and horizontal direction and try to visualize how they will deflect under a dynamic
loading.
It is obvious that while vertical strip behaves as a cantilever beam, the horizontal
strip shall behave as a beam whose ends are fixed. Based on the above assumption the
shape function for the mode are considered as
(z, y) = F(z) F(y), where (3.6.50)
F(z) = sin
z
H
sin h
z
H

z
_
cos
z
H
cos h
z
H
_
F(y) = sin
3y
2b
sin h
3y
2b

y
_
cos
3y
2b
cos h
3y
2b
_
For first mode = 1.875
z
=
sin +sin h
cos +cos h
and
y
=
sin
3
2
sin h
3
2
cos
3
2
cos h
3
2
(3.6.51)
3.6.10.1 Calculation of mass coefficient
For first mode analysis as
i
(z, y) =
j
(z, y), we have
m =
H
_
0
b
_
0
z
g tan
2
dy dz =
H
_
0
b
_
0
z
g tan
F(z)
2
F(y)
2
dy dz, (3.6.52)
where F(z) and F(y) are as defined above.
For ease of analysis we convert the integration basis fromlocal to natural co-ordinate
when
z =
H
2
( +1) and y =
b
2
( +1) which implies
dz =
H
2
d and dy =
b
2
d
Thus, F() = sin
( +1)
2
sin h
( +1)
2

z
_
cos
( +1)
2
cos h
( +1)
2
_
and F() = sin
3( +1)
4
sin h
3( +1)
4

y
_
cos
3(+1)
4
cos h
3(+1)
4
_
.
Substituting the above functions in mass equation, we have
m =
1
_
1
1
_
1
H
2
b
8g tan
F()
2
F()
2
dd (3.6.53)
the above can be further simplified to the expression
m =
H
2
b
8g tan
C
1
(3.6.54)
where, C
1
=
1
_
1
1
_
1
F()
2
F()
2
dd.
3.6.10.2 Derivation of stiffness coefficient
For derivation of stiffness coefficient for the first mode as
i
(z, y) =
j
(z, y), we have
from the stiffness expression furnished above as
K =
H
_
0
b
_
0
D
_
_
_
z
2
+

2
y
2
_
2
(1 )
_
_
_
2
z
2
y
2
2
_

2
zy
_
2
_
_
_
_
_
dy dz
(3.6.55)
Using, (z, y) = F(z) F(y), we have
z
2
=

F(z)F(y)
y
2
=

F(y)F(z) and

2
yz
=

F(y)
F(z).
Substituting the above in the stiffness equation and on expansion, we have
K =
H
_
0
b
_
0
D
_
_
F(y)
F(z)
_
2
+
_
F(z)
F(y)
_
2
+2
_
F(y)F(z)
F(y)
F(z)
_
+2(1 )
_
F(y)
F(z)
_
2
_
(3.6.56)
where F(z) and F(y) are as defined previously.
For further derivation of the stiffness matrix it is essential to derive the first and
second derivatives of F(z) and F(y) and are as shown hereafter
F(z) = sin
z
H
sin h
z
H

z
_
cos
z
H
cos h
z
H
_
F(z) =

H
_
cos
z
H
cos h
z
H
+
z
_
sin
z
H
+sin h
z
H
__
F(z) =

2
H
2
_
z
_
cos
z
H
+cos h
z
H
_
sin
z
H
sin h
z
H
_
.
Transforming the above in natural co-ordinate, we have
F() = sin
( +1)
2
sin h
( +1)
2

z
_
cos
( +1)
2
cos h
( +1)
2
_
= f (). say
F() =

H
_
cos
( +1)
2
cos h
( +1)
2
+
z
_
sin
( +1)
2
+sin h
( +1)
2
__
=

H
f
(), say
F() =

2
H
2
_
z
_
cos
(1 + )
2
+cos h
(1 + )
2
_
sin
(1 + )
2
sin h
(1 + )
2
_
=

2
H
2
f
(), say.
Similarly for y direction on differentiation and transformation to natural co-ordinate
we have
F() = sin
3( +1)
4
sin h
3( +1)
4

y
_
cos
3( +1)
4
cos h
3( +1)
4
_
= f ()
F() =
3
2b
_
cos
3( +1)
4
cos h
3( +1)
4
+
y
_
sin
3( +1)
4
+sin h
3( +1)
4
__
=
3
2b
f
()
F() =
9
2
4b
2
_
y
_
cos
3(1 + )
4
+cos h
3(1 + )
4
_
sin
3(1 + )
4
sin h
3(1 + )
4
_
=
9
2
4b
2
f
()
Substituting the above expressions in the stiffness equations, we have
K =
DHb
4
_
_
4
H
4
1
_
1
1
_
1
_
f ()f
()
_
2
dd +
81
4
16b
4
1
_
1
1
_
1
_
f
()f ()
_
2
dd
_
_
+
DHb
4
_
_
9
2
2b
2
H
2
1
_
1
1
_
1
_
f ()f ()f
()f
()
_
dd
+
9(1 )
2
2
2b
2
H
2
1
_
1
1
_
1
_
f
()f
()
_
2
_
_
(3.6.57)
Considering,
C
2
=
1
_
1
1
_
1
[f ()f
()]
2
dd; C
3
=
1
_
1
1
_
1
[f ()f
()]
2
dd
C
4
=
1
_
1
1
_
1
[f ()f ()f
()f
()]dd and C
5
=
1
_
1
1
_
1
[f
()f
()]
2
dd.
(3.6.58)
We have,
K =
DHb
4
_
4
H
4
C
2
+
81
4
16b
4
C
3
+
9
2
2b
2
H
2
C
4
+
9(1 )
2
2
2H
2
b
2
C
5
_
K = D
_
4
b
4H
3
C
2
+
81
4
H
64b
3
C
3
+
9
2
8Hb
C
4
+
9(1 )
8Hb
C
5
_
(3.6.59)
3.6.10.3 Derivation of the time period
Considering the time period as T = 2
_
m
K
, substituting the value of mand k obtained
above, we have
T = 2
_
H
2
bC
1
8Dg tan
_
4
b
4H
3
C
2
+
81
4
H
64b
3
C
3
+
9
2
8Hb
C
4
+
9(1)
2
2
8Hb
C
5
_
Considering the value of D as given previously and considering H/b = r as the
aspect ratio the above on some manipulation and simplification may be expressed as
T = 2
_
12 H
5
(1
2
)
8Et
3
g tan (X
1
r
4
+X
2
r
2
+X
3
r
2
+X
4
)
(3.6.60)
where X
1
, X
2
, X
3
, X
4
are constants whose values are mentioned below.
It will be observed that here that the time period expression is very similar to the
what we got for beam except the fact that for plate element the aspect ratio (r) factor
in the denominator affect the equation including the Poissons ratio ().
3.6.10.4 The integration constants
The integration constants for C
1
to C
5
are solved numerically based on Simpsons
method and are furnished in Table 3.6.3 for reference.
Similarly the constants X
1
, X
2
, X
3
, X
4
is given by
X
1
= 75.3121451, X
2
= 3.2405562, X
3
= 17.688041, X
4
= 1.91589362
Table 3.6.3 Integration constants for time period calculation.
Coefficients Integral function Value
C
1
1
_
1
1
_
1
F()
2
F()
2
dd 12.444
C
2
1
_
1
1
_
1
[f ()f

()]
2
dd 7.716
C
3
1
_
1
1
_
1
[f ()f

()]
2
dd 7.60
C
4
1
_
1
1
_
1
[f ()f ()f

()f

()]dd 1.033
C
5
1
_
1
1
_
1
[f

()f

()]
2
dd 5.638
3.6.10.5 Calculation of nodal forces under dynamic load
Once the time period is calculated and the response acceleration is obtained from
IS-1893, the shear force/or the nodal force is obtained from the expression
V
i
=
i
ZI
2R
S
a
n
i=1
m
i
i
where (3.6.61)
1
=
n
i=1
m
i
i
/
n
i=1
m
i
i
2
is the modal mass participation factor
for the present plate problem it can be represented by
1
=
1
_
1
1
_
1
(1 + )f ()f ()dd
1
_
1
1
_
1
(1 + )f
2
()f
2
()dd
(3.6.62)
The above on integration gives a unique value of
1
= 0.423.
Substituting this value in the above equation, we have
V
i
= 0.423S
a
n
i=1
m
i
i
, where =
ZI
2R
and the equation can be further transformed
into
Vi =
0.423S
a
H
2
b
8g tan
1
_
1
1
_
1
(1 + )f ()f ()dd. (3.6.63)
In the above expression the term S
a
/g is a function of the time period as derived
above. The above on integration will give the dynamic shear force.
However this is not required and the nodal forces can be found out for various values
of and for boundary +1 to 1 to obtain the nodal force coefcient. Once these
coefcients are known they can multiplied by the constant term to obtain dynamic
force imposed by the soil at various points of the plate. The summation of all these
force over the surface of the wall will give the total shear induced on the wall due to
the earthquake.
Thus, V
i
(z, y) =
0.423S
a
H
2
b
8g tan
(, ) (3.6.64)
The values of (, ) are as given in Figure 3.6.10, as coefficients on the plate.
These coefficients, when multiplied by the term
0.423S
a
H
2
b
8g tan
, will give the nodal
force at various points of the plate.
It will be interesting to note from the surface envelope generated below that as the
load coefficients are dependent on the assumed shape function there generic curve
-1 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
1 0 0 .44 1.44 2.56 3.40 3.72 3.42 2.60 1.48 0.47 0.0
0.8
0 0.342 1.12 1.98 2.64 2.89 2.65 2.02 1.153 0.365 0
0.6
0 0.255 0.84 1.485 1.98 2.16 1.988 1.507 0.863 0.273 0
0.4 0 0.182 0.547 1.06 1.407 1.54 1.417 1.074 0.614 0.194 0
0.2 0 0.121 0.399 0.707 0.941 1.03 0.948 0.719 0.411 0.130 0
0.0 0 0.074 0.245 0.434 0.578 0.632 0.581 0.441 0.253 0.079 0
-0.2 0 0.041 0.133 0.235 0.313 0.342 0.315 0.238 0.136 0.043 0
-0.4 0 0.020 0.060 0.104 0.134 0.152 0.140 0.106 0.061 0.019 0
-0.6 0 0.005 0.020 0.032 0.043 0.047 0.043 0.033 0.018 0.006 0
- 0.8 0 0.0007 0.003 0.004 0.006 0.006 0.006 0.004 0.003 0.0007 0
-1.0 0 0 0 0 0 0 0 0 0 0 0
Figure 3.6.10 Load coefficients (, ) for plates with three edge fixed and one edge free under.
follows the same shape as that of the assumed mode that is, it varies like a cantilever
in vertical direction and a beam fixed at edge in horizontal direction. This is surely
logical and is in variance to equivalent static load where hydrostatic force is profile is
assumed. The variation is more profound in horizontal direction for in normal analysis
this is considered as constant (like a rectangular shape) while in reality it is hyperbolic
in nature with zero at the edge and maximum at the centre.
Thus if the wall is spanning in one direction (i.e. a one way slab) when major
load spans horizontally along the shorter span, present state of art of arriving at
the Shear Force and Bending Moment could be signicantly in variation to the
reality.
Once the force are known it can be put directly as an input in a FEM analysis of a
plate subdivided up into meshes as shown in Figure 3.6.11 and applying the forces as
nodal loads we get the shear and bending moment in the wall.
1
0
.
6
0
.
2
-

0
.
2
-

0
.
6
-
1
S1
S6
S11
0
0.5
1
1.5
2
2.5
3
3.5
4
f(P)
Height
Width
Nodal Load function
Figure 3.6.11 Surface envelope of nodal loads of counterfort wall under first mode.
Else the amplitude, dynamic moments and shear can be obtained by the method as
shown here after.
3.6.10.6 Calculation of dynamic amplitude
As shown at the outset the displacement is defined by
w(z, y, t) =
n
i=1
i
(z, y)q
i
(t) (3.6.65)
For structural systems under earthquake the dynamic amplitude can be calculated
from the expression
w(z, y) =
i
(z
i
y
i
)S
d
(3.6.66)
where, S
d
=
S
a
2
; S
a
= acceleration spectrumand is a function of time period, and
2
=
square of the natural frequency.
Based on the codal provision the displacement spectrum can be expressed as
S
d
=
S
a
2
(3.6.67)
Considering
2
=
4
2
T
2
, substituting equation of time period we have
2
=
8Dg tan (X
1
r
4
+X
2
r
2
+vX
3
r
2
+X
4
)
H
5
(3.6.68)
where D= flexural stiffness as expressed earlier.
S
d
=
S
a
H
5
8Dg tan (X
1
r
4
+X
2
r
2
+ X
3
r
2
+X
4
)
(3.6.69)
which gives,
w(z, y) =

i
S
a
H
5
8Dg tan (X
1
r
4
+X
2
r
2
+ X
3
r
2
+X
4
)
F(z)F(y) (3.6.70)
where, F(z) and F(y) are as defined in Equation (3.6.51).
Considering (r) = X
1
r
4
+ X
2
r
2
+ X
3
r
2
+ X
4
, a function of aspect ratio above
expression can be modified and written as
w(z, y) =

i
S
a
H
5
8Dg tan (r)
F(z)F(y) (3.6.71)
1
0
.
4
-
0
.
2
-
0
.
8
S1
S8
0.000
1.000
2.000
3.000
4.000
f(d)
Height
Width
Displacement function f(d) for r=2
Figure 3.6.12 Amplitude function of wall for aspect ratio H/b = 2.
Now considering, A =

i
S
a
8g tan
, a constant for a system for a particular mode.
Expression w(z, y) can now be expressed as
w(z, y) =
A
D
F(z)F(y)
(r)
H
5
(3.6.72)
It is thus observed that the dynamic amplitude w(z, y) is function of the flexural
stiffness of the plate, (r) the aspect ratio of the plate, and the shape function F(z)
and F(y).
Shown in Figure 3.6.12 is the displacement envelope of the plate with typical aspect
ratio of r = 2 and Poissons ratio of 0.25.
The coefficients are scaled to 1000. The values when multiplied by the term A (and
divided by 1000) will give the dynamic amplitude for the first mode.
3.6.10.7 Calculation of dynamic moments and shear
The bending moment of a thin plate is given the expression
M
z
= D
_
2
w
z
2
+
2
w
y
2
_
; M
y
= D
_
2
w
y
2
+
2
w
z
2
_
;
Q
z
= D

z
_
2
w
z
2
+

2
w
y
2
_
; Q
y
= D

y
_
2
w
z
2
+

2
w
y
2
_
(3.6.73)
where, w(z, y) =
A
D
F(z)F(y)
(r)
H
5
.
Substituting the value of w(z, y) above we have
M
z
=
A
(r)
_
2
H
2
F
(z)F(y) +
9
2
4b
2
F(z)F
(y)
_
H
5
where F
(z) and F
(y) are as defined in previously in local and natural co-ordinates.

The above expression of M
z
can be further modified to
M
z
=
A
(r)
_
2
F
(z)F(y) +
9
2
r
2
4
F(z)F
(y)
_
H
3
. (3.6.74)
From the above the expression we observe that moment varies as the cube of the
height and is a function of the basic shape function and their derivative, Poissons ratio
and the aspect ratio.
Shown in Figure 3.6.13 is a typical envelope for aspect ratio H/b = 2 and Poissons
ratio=0.25.
-
1
-
0
.
4
0
.
2
0
.
8
S1
S7
-0.1
-0.05
0
0.05
0.1
0.15
f(Mz)
Height
Width
Function f(Mz) for r=2
Figure 3.6.13 Variation of dynamic moment in vertical (Z) direction for H/b = 2.
Similarly for y direction we have
M
y
=
A
(r)
_
2
F
(z)F(y) +
9
2
r
2
4
F(z)F
(y)
_
H
3
(3.6.75)
where, substituting the values of z in either local and nodal co-ordinates moment
coefficient and envelope can be plotted.
We show in Figure 3.6.14 a typical envelope of Moment (M
y
) for aspect ratio
r(H/b) = 2 and Poissons ratio =0.25.
Similarly for shear force in Z and Y direction can be expressed as
Q
z
=
A
(r)
_
3
F
(z)F(y) +
9
2
r
2
4
F
(z)F
(y)
_
H
2
and
Q
y
=
A
(r)
_
3
2
r
2
F
(z)F
(y) +
27
3
r
3
8
F(z)F
(y)
_
H
2
(3.6.76)
Looking at above expressions we see that the shear force varies as the square of the
height and is a function of the aspect ratio and the basic shape function.
The above procedure gives a comprehensive solution for walls with three side fixed
and one side free subjected to dynamic earth pressure under earthquake load where
the moments, shears and amplitude are dependent on the time period, geometry of the
wall, its boundary condition as well as the material and engineering property of the soil.
-
1
-
0
.
4
0
.
2
0
.
8
S1
S8
-0.4
-0.2
0
0.2
0.4
0.6
f(My)
Height
Width
Function f(My) for r=2
Figure 3.6.14 Variation of dynamic moment in vertical (Y) direction for H/b = 2.
3.6.11 Soil sloped at an angle i with horizontal
In this case like we did in cantilever wall let H
= H + x = H
1 + , where
=
cos sin i
sin(i)
46
.
Thus the mass coefficient can now be considered as varying between, H
to 0
m =
H
_
0
b
_
0
z
g tan
F(z)
2
F(y)
2
dy dz (3.6.77)
The above on transformation to natural co ordinate can be represented as
m =
1
_
1
1
_
1
H
2b
8g tan
F()
2
F()
2
dd =
H
2
b(1 + )
8g tan
C
1
(3.6.78)
The time period is thus given by
T = 2
_
12 H
5
(1 + )(1
2
)
8Et
3
g tan (X
1
r
4
+X
2
r
2
+X
3
r
2
+X
4
)
(3.6.79)
46 Refer to the figure we have drawn for the cantilever retaining wall with sloped soil surface.
It will be observed that for i = 0, = 0 the above equation converges to equation (for
time period with wall having soil parallel to the ground. The constants X
1
, X
2
, X
3
, X
4
remains same as mentioned earlier.
The modal mass participation factor shall remain same as earlier i.e.,
1
=
_
1
1
_
1
1
(1 + )f ()f ()dd
_
1
1
_
1
1
(1 + )f
2
()f
2
()dd
= 0.423 (3.6.80)
Thus the nodal load can now be expressed as
V
i
=
0.423S
a
H
2b
8g tan
1
_
1
1
_
1
(1 + )f ()f ()dd
=
0.423S
a
H
2
(1 + )b
8g tan
1
_
1
1
_
1
(1 + )f ()f ()dd (3.6.81)
Thus we see that the constant term is multiplied by an additional factor (1 + ) and
the integration constants (, ) remain same as expressed earlier.
It may again be noted that for i = 0, = 0 when the above expression converges to
expression shear equation for the soil parallel to the ground.
From above we can safely deduce from mathematical similarity that for this case.
The term A as expressed in earlier can be now be expressed as
A =

i
S
a
(1 + )
8g tan
(3.6.82)
The rest of the steps remain same as explained earlier.
3.6.11.1 Design moments and shear coefficients
Observing the values of moments and shear as derived above it can be inferred that
they can be expressed in the form
M(z) = Coeff
z

0.423 S
a
8g tan
(1 + )H
3
; M(y) = Coeff
y

0.423 S
a
8g tan
(1 + )H
3
Q(z) = Coeff
z

0.423 S
a
8g tan
(1 + )H
2
; Q(y) = Coeff
z

0.423 S
a
8g tan
(1 + )H
2
.
(3.6.83)
The coefficients are furnished in Table 3.6.4 to 3.6.6
Table 3.6.4 Moment coefficients in vertical direction (M
z
).
Aspect ratio H/b 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0
1.0 0.1016 0.0548 0.0102 0.0273 0.0526 0.0618 0.0533 0.0285 0.0087 0.0530 0.0999
0.8 0.0877 0.0473 0.0089 0.0234 0.0452 0.0531 0.0457 0.0244 0.0076 0.0458 0.0861
0.6 0.0737 0.0398 0.0077 0.0192 0.0374 0.0440 0.0379 0.0201 0.0066 0.0386 0.0724
0.4 0.0601 0.0326 0.0067 0.0149 0.0295 0.0348 0.0303 0.0156 0.0058 0.0316 0.0590
0.2 0.0469 0.0256 0.0059 0.0106 0.0216 0.0255 0.0234 0.0111 0.0052 0.0248 0.0460
2.0 0.0 0.0345 0.0191 0.0052 0.0063 0.0139 0.0166 0.0176 0.0066 0.0047 0.0186 0.0339
0.2 0.0234 0.0133 0.0046 0.0022 0.0067 0.0083 0.0122 0.0024 0.0043 0.0129 0.0230
0.4 0.0139 0.0084 0.0043 0.0015 0.0003 0.0009 0.0073 0.0014 0.0042 0.0082 0.0136
0.6 0.0065 0.0046 0.0042 0.0046 0.0051 0.0053 0.0080 0.0046 0.0043 0.0046 0.0064
0.8 0.0017 0.0022 0.0044 0.0071 0.0092 0.0100 0.0005 0.0072 0.0045 0.0023 0.0017
1.0 0.0000 0.0015 0.0050 0.0088 0.0117 0.0128 0.0118 0.0089 0.0051 0.0016 0.0000
1.0 0.0804 0.0433 0.0081 0.0216 0.0416 0.0489 0.0421 0.0225 0.0069 0.0420 0.0790
0.8 0.0694 0.0374 0.0070 0.0185 0.0358 0.0420 0.0362 0.0193 0.0060 0.0362 0.0681
0.6 0.0583 0.0315 0.0061 0.0153 0.0297 0.0350 0.0301 0.0160 0.0052 0.0305 0.0573
0.4 0.0475 0.0257 0.0052 0.0120 0.0236 0.0278 0.0242 0.0126 0.0045 0.0249 0.0467
0.2 0.0371 0.0202 0.0045 0.0087 0.0175 0.0207 0.0187 0.0091 0.0039 0.0196 0.0364
2.25 0.0 0.0273 0.0150 0.0038 0.0055 0.0117 0.0139 0.0140 0.0058 0.0034 0.0146 0.0268
0.2 0.0185 0.0104 0.0033 0.0024 0.0062 0.0076 0.0097 0.0026 0.0030 0.0101 0.0182
0.4 0.0110 0.0065 0.0029 0.0003 0.0014 0.0020 0.0058 0.0002 0.0028 0.0063 0.0108
0.6 0.0051 0.0035 0.0028 0.0026 0.0026 0.0027 0.0056 0.0026 0.0028 0.0034 0.0050
0.8 0.0013 0.0015 0.0028 0.0044 0.0056 0.0061 0.0005 0.0044 0.0029 0.0016 0.0013
1.0 0.0000 0.0009 0.0031 0.0055 0.0073 0.0080 0.0073 0.0056 0.0032 0.0010 0.0000
1.0 0.0652 0.0351 0.0066 0.0175 0.0338 0.0396 0.0342 0.0183 0.0056 0.0340 0.0641
0.8 0.0562 0.0303 0.0057 0.0150 0.0290 0.0341 0.0294 0.0157 0.0048 0.0293 0.0552
0.6 0.0473 0.0255 0.0049 0.0125 0.0242 0.0284 0.0245 0.0130 0.0042 0.0247 0.0465
0.4 0.0385 0.0208 0.0042 0.0099 0.0193 0.0227 0.0197 0.0103 0.0036 0.0202 0.0379
0.2 0.0301 0.0163 0.0035 0.0072 0.0145 0.0171 0.0153 0.0076 0.0030 0.0158 0.0295
2.5 0.0 0.0221 0.0121 0.0029 0.0047 0.0098 0.0117 0.0114 0.0050 0.0026 0.0118 0.0217
0.2 0.0150 0.0084 0.0024 0.0024 0.0055 0.0067 0.0079 0.0025 0.0022 0.0081 0.0147
0.4 0.0089 0.0052 0.0021 0.0003 0.0018 0.0023 0.0047 0.0003 0.0020 0.0050 0.0087
0.6 0.0042 0.0027 0.0019 0.0015 0.0013 0.0013 0.0041 0.0015 0.0019 0.0027 0.0041
0.8 0.0011 0.0011 0.0019 0.0028 0.0036 0.0039 0.0005 0.0028 0.0019 0.0011 0.0011
1.0 0.0000 0.0006 0.0020 0.0036 0.0048 0.0052 0.0048 0.0037 0.0021 0.0007 0.0000
1.0 0.0539 0.0291 0.0054 0.0145 0.0279 0.0328 0.0283 0.0151 0.0046 0.0281 0.0530
0.8 0.0465 0.0251 0.0047 0.0124 0.0240 0.0282 0.0243 0.0130 0.0040 0.0243 0.0457
0.6 0.0391 0.0211 0.0040 0.0103 0.0200 0.0235 0.0203 0.0108 0.0034 0.0204 0.0384
0.4 0.0319 0.0172 0.0034 0.0082 0.0160 0.0189 0.0164 0.0086 0.0029 0.0167 0.0313
0.2 0.0249 0.0135 0.0028 0.0061 0.0121 0.0143 0.0127 0.0064 0.0025 0.0131 0.0244
2.75 0.0 0.0183 0.0100 0.0023 0.0041 0.0084 0.0099 0.0095 0.0043 0.0020 0.0097 0.0180
0.2 0.0124 0.0069 0.0019 0.0022 0.0049 0.0059 0.0065 0.0023 0.0017 0.0067 0.0122
0.4 0.0074 0.0042 0.0016 0.0005 0.0019 0.0024 0.0039 0.0006 0.0015 0.0041 0.0072
0.6 0.0034 0.0022 0.0014 0.0009 0.0006 0.0005 0.0031 0.0008 0.0013 0.0021 0.0034
0.8 0.0009 0.0009 0.0013 0.0019 0.0024 0.0025 0.0004 0.0019 0.0013 0.0009 0.0009
1.0 0.0000 0.0004 0.0014 0.0025 0.0033 0.0036 0.0033 0.0025 0.0014 0.0005 0.0000
1.0 0.0453 0.0244 0.0046 0.0122 0.0235 0.0276 0.0238 0.0127 0.0039 0.0236 0.0445
0.8 0.0391 0.0211 0.0040 0.0105 0.0202 0.0237 0.0204 0.0109 0.0034 0.0204 0.0384
0.6 0.0329 0.0177 0.0034 0.0087 0.0169 0.0198 0.0171 0.0091 0.0029 0.0172 0.0323
0.4 0.0268 0.0145 0.0028 0.0070 0.0135 0.0159 0.0138 0.0073 0.0024 0.0140 0.0263
0.2 0.0209 0.0113 0.0023 0.0052 0.0103 0.0121 0.0107 0.0055 0.0020 0.0110 0.0205
3.0 0.0 0.0154 0.0084 0.0019 0.0035 0.0072 0.0085 0.0080 0.0037 0.0017 0.0081 0.0151
0.2 0.0104 0.0058 0.0015 0.0020 0.0043 0.0052 0.0055 0.0021 0.0014 0.0056 0.0102
0.4 0.0062 0.0035 0.0012 0.0006 0.0018 0.0023 0.0032 0.0007 0.0011 0.0034 0.0061
0.6 0.0029 0.0018 0.0010 0.0005 0.0002 0.0001 0.0024 0.0005 0.0010 0.0017 0.0028
0.8 0.0008 0.0007 0.0009 0.0013 0.0016 0.0017 0.0004 0.0013 0.0009 0.0007 0.0007
1.0 0.0000 0.0003 0.0010 0.0017 0.0023 0.0025 0.0023 0.0018 0.0010 0.0003 0.0000
Table 3.6.5 Moment coefficients in horizontal direction (M
y
).
Aspect ratio H/b 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0
1.0 0.407 0.219 0.041 0.109 0.210 0.247 0.213 0.114 0.035 0.212 0.399
0.8 0.351 0.189 0.035 0.094 0.181 0.213 0.184 0.098 0.030 0.183 0.344
0.6 0.295 0.159 0.030 0.079 0.152 0.179 0.154 0.083 0.025 0.154 0.290
0.4 0.240 0.129 0.024 0.064 0.124 0.146 0.125 0.067 0.021 0.125 0.236
0.2 0.187 0.101 0.019 0.050 0.096 0.113 0.098 0.052 0.016 0.098 0.184
2.0 0.0 0.138 0.075 0.014 0.036 0.070 0.083 0.071 0.038 0.012 0.072 0.136
0.2 0.093 0.051 0.010 0.024 0.047 0.055 0.048 0.025 0.009 0.049 0.092
0.4 0.055 0.030 0.006 0.014 0.027 0.032 0.027 0.014 0.005 0.029 0.055
0.6 0.026 0.014 0.004 0.005 0.011 0.013 0.011 0.006 0.003 0.014 0.026
0.8 0.007 0.004 0.002 0.000 0.001 0.001 0.001 0.000 0.002 0.004 0.007
1.0 0.000 0.000 0.001 0.002 0.003 0.003 0.003 0.002 0.001 0.000 0.000
1.0 0.3217 0.1733 0.0324 0.0864 0.1665 0.1955 0.1686 0.0902 0.0275 0.1678 0.3160
0.8 0.2774 0.1495 0.0280 0.0744 0.1436 0.1686 0.1454 0.0777 0.0237 0.1447 0.2726
0.6 0.2334 0.1258 0.0236 0.0626 0.1207 0.1417 0.1222 0.0653 0.0200 0.1218 0.2293
0.4 0.1901 0.1025 0.0193 0.0508 0.0981 0.1153 0.0994 0.0531 0.0163 0.0992 0.1867
0.2 0.1483 0.0800 0.0151 0.0395 0.0764 0.0897 0.0773 0.0413 0.0128 0.0774 0.1457
2.25 0.0 0.1092 0.0589 0.0113 0.0289 0.0559 0.0657 0.0566 0.0301 0.0096 0.0571 0.1073
0.2 0.0739 0.0400 0.0078 0.0192 0.0374 0.0440 0.0379 0.0201 0.0067 0.0387 0.0727
0.4 0.0439 0.0238 0.0049 0.0110 0.0216 0.0255 0.0219 0.0115 0.0042 0.0231 0.0431
0.6 0.0205 0.0112 0.0026 0.0045 0.0093 0.0110 0.0094 0.0047 0.0023 0.0109 0.0202
0.8 0.0054 0.0031 0.0012 0.0003 0.0012 0.0016 0.0012 0.0003 0.0011 0.0030 0.0053
1.0 0.0000 0.0002 0.0008 0.0014 0.0018 0.0020 0.0018 0.0014 0.0008 0.0003 0.0000
1.0 0.2608 0.1405 0.0263 0.0700 0.1350 0.1585 0.1367 0.0731 0.0223 0.1361 0.2562
0.8 0.2249 0.1212 0.0227 0.0604 0.1164 0.1367 0.1179 0.0630 0.0192 0.1173 0.2210
0.6 0.1892 0.1020 0.0191 0.0507 0.0979 0.1149 0.0991 0.0530 0.0162 0.0987 0.1859
0.4 0.1541 0.0831 0.0156 0.0412 0.0796 0.0935 0.0806 0.0431 0.0132 0.0804 0.1514
0.2 0.1203 0.0648 0.0122 0.0321 0.0620 0.0728 0.0628 0.0335 0.0104 0.0628 0.1182
2.5 0.0 0.0885 0.0478 0.0091 0.0235 0.0454 0.0534 0.0460 0.0245 0.0077 0.0463 0.0870
0.2 0.0600 0.0324 0.0063 0.0157 0.0305 0.0358 0.0309 0.0164 0.0054 0.0314 0.0589
0.4 0.0356 0.0193 0.0039 0.0090 0.0177 0.0209 0.0179 0.0094 0.0033 0.0187 0.0350
0.6 0.0167 0.0091 0.0020 0.0038 0.0078 0.0092 0.0079 0.0040 0.0018 0.0088 0.0164
0.8 0.0044 0.0025 0.0009 0.0004 0.0012 0.0015 0.0013 0.0004 0.0008 0.0024 0.0043
1.0 0.0000 0.0002 0.0005 0.0009 0.0012 0.0013 0.0012 0.0009 0.0005 0.0002 0.0000
1.0 0.2157 0.1162 0.0217 0.0579 0.1116 0.1311 0.1130 0.0604 0.0184 0.1125 0.2119
0.8 0.1860 0.1002 0.0188 0.0499 0.0963 0.1130 0.0975 0.0521 0.0159 0.0970 0.1827
0.6 0.1565 0.0843 0.0158 0.0420 0.0809 0.0950 0.0820 0.0438 0.0134 0.0816 0.1537
0.4 0.1274 0.0687 0.0129 0.0341 0.0659 0.0773 0.0667 0.0356 0.0109 0.0665 0.1252
0.2 0.0995 0.0536 0.0101 0.0266 0.0513 0.0602 0.0519 0.0277 0.0086 0.0519 0.0977
2.75 0.0 0.0732 0.0395 0.0075 0.0194 0.0376 0.0442 0.0381 0.0203 0.0064 0.0382 0.0719
0.2 0.0496 0.0268 0.0052 0.0130 0.0253 0.0297 0.0256 0.0136 0.0044 0.0259 0.0487
0.4 0.0294 0.0159 0.0032 0.0075 0.0148 0.0174 0.0149 0.0079 0.0027 0.0154 0.0289
0.6 0.0138 0.0075 0.0016 0.0033 0.0065 0.0077 0.0066 0.0034 0.0014 0.0073 0.0135
0.8 0.0036 0.0020 0.0007 0.0004 0.0012 0.0014 0.0012 0.0005 0.0006 0.0020 0.0036
1.0 0.0000 0.0001 0.0003 0.0006 0.0008 0.0009 0.0008 0.0006 0.0004 0.0001 0.0000
1.0 0.1813 0.0977 0.0183 0.0487 0.0939 0.1102 0.0950 0.0508 0.0155 0.0946 0.1781
0.8 0.1564 0.0843 0.0158 0.0420 0.0809 0.0950 0.0819 0.0438 0.0134 0.0816 0.1536
0.6 0.1315 0.0709 0.0133 0.0353 0.0681 0.0799 0.0689 0.0368 0.0112 0.0686 0.1292
0.4 0.1071 0.0577 0.0108 0.0287 0.0554 0.0650 0.0561 0.0300 0.0092 0.0559 0.1053
0.2 0.0836 0.0451 0.0085 0.0223 0.0431 0.0507 0.0437 0.0233 0.0072 0.0436 0.0821
3.0 0.0 0.0616 0.0332 0.0063 0.0164 0.0317 0.0372 0.0321 0.0171 0.0053 0.0321 0.0605
0.2 0.0417 0.0225 0.0043 0.0110 0.0213 0.0250 0.0216 0.0115 0.0037 0.0218 0.0410
0.4 0.0247 0.0134 0.0026 0.0064 0.0125 0.0147 0.0126 0.0067 0.0023 0.0130 0.0243
0.6 0.0116 0.0063 0.0013 0.0028 0.0056 0.0066 0.0056 0.0029 0.0012 0.0061 0.0114
0.8 0.0030 0.0017 0.0005 0.0004 0.0011 0.0013 0.0011 0.0005 0.0005 0.0017 0.0030
1.0 0.0000 0.0001 0.0002 0.0004 0.0006 0.0006 0.0006 0.0004 0.0003 0.0001 0.0000
Table 3.6.6 Shear coefficients in vertical direction (Q
z
).
Aspect ratio H/b 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0
1.0 0.5596 0.3016 0.0564 0.1502 0.2897 0.3401 0.2933 0.1569 0.0478 0.2920 0.5498
0.8 0.5588 0.3006 0.0547 0.1529 0.2931 0.3438 0.2967 0.1595 0.0460 0.2910 0.5490
0.6 0.5534 0.2973 0.0528 0.1539 0.2935 0.3441 0.2972 0.1605 0.0441 0.2877 0.5437
0.4 0.5393 0.2894 0.0502 0.1521 0.2890 0.3385 0.2925 0.1586 0.0418 0.2800 0.5299
0.2 0.5134 0.2751 0.0467 0.1468 0.2777 0.3251 0.2811 0.1530 0.0386 0.2662 0.5044
2.0 0.0 0.4728 0.2530 0.0419 0.1371 0.2583 0.3023 0.2615 0.1429 0.0344 0.2448 0.4646
0.2 0.4157 0.2221 0.0357 0.1227 0.2300 0.2688 0.2327 0.1278 0.0290 0.2149 0.4085
0.4 0.3406 0.1815 0.0278 0.1031 0.1918 0.2240 0.1941 0.1073 0.0223 0.1756 0.3347
0.6 0.2466 0.1308 0.0181 0.0781 0.1435 0.1673 0.1452 0.0812 0.0141 0.1264 0.2423
0.8 0.1331 0.0696 0.0066 0.0478 0.0850 0.0985 0.0859 0.0496 0.0044 0.0672 0.1308
1.0 0.0000 0.0021 0.0068 0.0121 0.0161 0.0176 0.0162 0.0123 0.0070 0.0022 0.0000
1.0 0.4428 0.2386 0.0446 0.1189 0.2292 0.2691 0.2321 0.1241 0.0378 0.2310 0.4351
0.8 0.4422 0.2380 0.0436 0.1205 0.2313 0.2714 0.2341 0.1258 0.0367 0.2304 0.4344
0.6 0.4379 0.2354 0.0423 0.1209 0.2311 0.2710 0.2339 0.1261 0.0354 0.2278 0.4302
0.4 0.4267 0.2292 0.0404 0.1192 0.2270 0.2661 0.2298 0.1243 0.0338 0.2218 0.4193
0.2 0.4062 0.2179 0.0378 0.1147 0.2177 0.2551 0.2204 0.1195 0.0314 0.2109 0.3991
2.25 0.0 0.3741 0.2005 0.0341 0.1068 0.2022 0.2367 0.2046 0.1113 0.0282 0.1940 0.3676
0.2 0.3289 0.1761 0.0292 0.0952 0.1795 0.2100 0.1817 0.0992 0.0241 0.1703 0.3232
0.4 0.2695 0.1440 0.0231 0.0796 0.1492 0.1744 0.1510 0.0829 0.0188 0.1393 0.2648
0.6 0.1951 0.1038 0.0155 0.0598 0.1109 0.1295 0.1123 0.0623 0.0123 0.1004 0.1917
0.8 0.1053 0.0555 0.0064 0.0358 0.0646 0.0750 0.0653 0.0372 0.0046 0.0536 0.1035
1.0 0.0000 0.0013 0.0043 0.0076 0.0100 0.0110 0.0101 0.0077 0.0044 0.0014 0.0000
1.0 0.3590 0.1935 0.0362 0.0964 0.1858 0.2182 0.1882 0.1006 0.0306 0.1873 0.3527
0.8 0.3585 0.1930 0.0355 0.0974 0.1871 0.2196 0.1895 0.1017 0.0299 0.1868 0.3522
0.6 0.3550 0.1909 0.0346 0.0975 0.1867 0.2189 0.1890 0.1017 0.0290 0.1848 0.3488
0.4 0.3460 0.1859 0.0332 0.0959 0.1831 0.2147 0.1854 0.1000 0.0278 0.1800 0.3399
0.2 0.3293 0.1768 0.0311 0.0921 0.1754 0.2055 0.1775 0.0960 0.0260 0.1711 0.3236
2.5 0.0 0.3033 0.1627 0.0282 0.0856 0.1626 0.1905 0.1646 0.0893 0.0235 0.1575 0.2980
0.2 0.2667 0.1429 0.0243 0.0761 0.1441 0.1687 0.1459 0.0794 0.0201 0.1383 0.2620
0.4 0.2185 0.1169 0.0193 0.0634 0.1195 0.1398 0.1209 0.0661 0.0159 0.1131 0.2147
0.6 0.1582 0.0844 0.0132 0.0474 0.0884 0.1033 0.0895 0.0493 0.0107 0.0816 0.1554
0.8 0.0854 0.0452 0.0058 0.0279 0.0508 0.0591 0.0514 0.0290 0.0044 0.0436 0.0839
1.0 0.0000 0.0009 0.0028 0.0050 0.0066 0.0072 0.0066 0.0050 0.0029 0.0009 0.0000
1.0 0.2969 0.1600 0.0299 0.0797 0.1537 0.1804 0.1556 0.0832 0.0253 0.1549 0.2917
0.8 0.2965 0.1596 0.0294 0.0804 0.1545 0.1814 0.1565 0.0839 0.0248 0.1545 0.2913
0.6 0.2936 0.1579 0.0288 0.0803 0.1540 0.1806 0.1559 0.0838 0.0242 0.1529 0.2884
0.4 0.2861 0.1538 0.0277 0.0789 0.1509 0.1769 0.1527 0.0823 0.0232 0.1489 0.2811
0.2 0.2723 0.1463 0.0260 0.0756 0.1443 0.1692 0.1461 0.0789 0.0218 0.1416 0.2676
2.75 0.0 0.2508 0.1347 0.0237 0.0702 0.1337 0.1566 0.1353 0.0732 0.0197 0.1303 0.2465
0.2 0.2205 0.1183 0.0205 0.0623 0.1183 0.1386 0.1198 0.0650 0.0170 0.1145 0.2167
0.4 0.1807 0.0968 0.0164 0.0518 0.0979 0.1146 0.0991 0.0540 0.0135 0.0937 0.1775
0.6 0.1308 0.0699 0.0113 0.0385 0.0722 0.0844 0.0730 0.0401 0.0092 0.0676 0.1285
0.8 0.0706 0.0375 0.0052 0.0223 0.0411 0.0478 0.0415 0.0232 0.0041 0.0362 0.0694
1.0 0.0000 0.0006 0.0019 0.0034 0.0045 0.0049 0.0045 0.0034 0.0020 0.0006 0.0000
1.0 0.2496 0.1345 0.0252 0.0670 0.1292 0.1517 0.1308 0.0700 0.0213 0.1302 0.2452
0.8 0.2492 0.1342 0.0248 0.0675 0.1298 0.1523 0.1314 0.0704 0.0209 0.1299 0.2449
0.6 0.2468 0.1328 0.0243 0.0673 0.1292 0.1515 0.1308 0.0702 0.0205 0.1286 0.2425
0.4 0.2405 0.1294 0.0234 0.0660 0.1265 0.1483 0.1280 0.0689 0.0197 0.1252 0.2363
0.2 0.2289 0.1231 0.0221 0.0632 0.1209 0.1417 0.1224 0.0660 0.0185 0.1191 0.2249
3.0 0.0 0.2109 0.1133 0.0201 0.0586 0.1119 0.1311 0.1132 0.0612 0.0168 0.1096 0.2072
0.2 0.1854 0.0995 0.0175 0.0520 0.0989 0.1159 0.1001 0.0542 0.0145 0.0963 0.1822
0.4 0.1519 0.0815 0.0140 0.0431 0.0817 0.0957 0.0827 0.0449 0.0116 0.0788 0.1492
0.6 0.1100 0.0588 0.0097 0.0319 0.0601 0.0703 0.0608 0.0332 0.0080 0.0569 0.1080
0.8 0.0594 0.0316 0.0046 0.0183 0.0339 0.0396 0.0343 0.0191 0.0037 0.0305 0.0583
1.0 0.0000 0.0004 0.0014 0.0024 0.0032 0.0035 0.0032 0.0024 0.0014 0.0004 0.0000
Table 3.6.7 Shear coefficients in horizontal direction (Q
y
).
Aspect ratio H/b 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0
1.0 3.7646 3.7055 3.3533 2.5797 1.4184 0.0269 1.3690 2.5424 3.3319 3.6985 3.7640
0.8 3.2466 3.1947 2.8905 2.2235 1.2225 0.0232 1.1799 2.1913 2.8721 3.1886 3.2460
0.6 2.7311 2.6847 2.4278 1.8670 1.0263 0.0195 0.9906 1.8399 2.4122 2.6796 2.7305
0.4 2.2244 2.1820 1.9708 1.5146 0.8323 0.0158 0.8033 1.4926 1.9581 2.1776 2.2236
0.2 1.7359 1.6961 1.5285 1.1732 0.6444 0.0122 0.6219 1.1562 1.5185 1.6925 1.7351
2.0 0.0 1.2781 1.2394 1.1122 0.8517 0.4673 0.0089 0.4510 0.8393 1.1048 1.2366 1.2770
0.2 0.8654 0.8265 0.7351 0.5603 0.3067 0.0058 0.2960 0.5520 0.7301 0.8242 0.8640
0.4 0.5138 0.4733 0.4118 0.3101 0.1687 0.0032 0.1628 0.3055 0.4088 0.4715 0.5121
0.6 0.2404 0.1968 0.1579 0.1133 0.0601 0.0011 0.0579 0.1114 0.1564 0.1953 0.2385
0.8 0.0631 0.0148 0.0106 0.0180 0.0125 0.0002 0.0121 0.0180 0.0112 0.0134 0.0608
1.0 0.0000 0.0550 0.0776 0.0710 0.0421 0.0008 0.0407 0.0703 0.0778 0.0564 0.0027
1.0 3.3510 3.2984 2.9849 2.2963 1.2625 0.0239 1.2186 2.2631 2.9659 3.2922 3.3505
0.8 2.8899 2.8439 2.5733 1.9795 1.0883 0.0206 1.0504 1.9508 2.5568 2.8385 2.8894
0.6 2.4311 2.3904 2.1620 1.6627 0.9141 0.0173 0.8822 1.6386 2.1482 2.3859 2.4306
0.4 1.9800 1.9437 1.7562 1.3500 0.7420 0.0141 0.7161 1.3304 1.7450 1.9399 1.9794
0.2 1.5452 1.5121 1.3639 1.0474 0.5754 0.0109 0.5554 1.0322 1.3551 1.5090 1.5446
2.25 0.0 1.1377 1.1067 0.9949 0.7627 0.4186 0.0079 0.4040 0.7516 0.9884 1.1043 1.1369
0.2 0.7703 0.7404 0.6610 0.5048 0.2766 0.0052 0.2669 0.4974 0.6566 0.7385 0.7693
0.4 0.4573 0.4273 0.3751 0.2839 0.1548 0.0029 0.1494 0.2797 0.3724 0.4259 0.4562
0.6 0.2140 0.1826 0.1511 0.1104 0.0592 0.0011 0.0571 0.1087 0.1497 0.1815 0.2126
0.8 0.0562 0.0220 0.0030 0.0045 0.0043 0.0001 0.0042 0.0047 0.0025 0.0210 0.0546
1.0 0.0000 0.0386 0.0545 0.0499 0.0296 0.0006 0.0286 0.0494 0.0547 0.0397 0.0019
1.0 3.0187 2.9713 2.6889 2.0686 1.1373 0.0216 1.0978 2.0387 2.6718 2.9658 3.0182
0.8 2.6033 2.5620 2.3182 1.7833 0.9805 0.0186 0.9464 1.7575 2.3035 2.5572 2.6029
0.6 2.1900 2.1538 1.9482 1.4984 0.8237 0.0156 0.7951 1.4767 1.9357 2.1497 2.1896
0.4 1.7837 1.7518 1.5834 1.2173 0.6691 0.0127 0.6458 1.1997 1.5732 1.7484 1.7832
0.2 1.3920 1.3637 1.2308 0.9455 0.5195 0.0099 0.5014 0.9318 1.2228 1.3609 1.3915
2.5 0.0 1.0249 0.9992 0.8994 0.6900 0.3788 0.0072 0.3656 0.6799 0.8936 0.9971 1.0243
0.2 0.6939 0.6700 0.5998 0.4587 0.2515 0.0048 0.2427 0.4520 0.5958 0.6684 0.6932
0.4 0.4120 0.3889 0.3435 0.2608 0.1425 0.0027 0.1375 0.2569 0.3410 0.3877 0.4111
0.6 0.1928 0.1693 0.1429 0.1057 0.0570 0.0011 0.0550 0.1040 0.1417 0.1684 0.1918
0.8 0.0506 0.0255 0.0108 0.0033 0.0005 0.0000 0.0004 0.0031 0.0104 0.0248 0.0494
1.0 0.0000 0.0282 0.0398 0.0364 0.0216 0.0004 0.0209 0.0361 0.0399 0.0289 0.0014
1.0 2.7460 2.7029 2.4460 1.8817 1.0346 0.0196 0.9986 1.8545 2.4304 2.6979 2.7456
0.8 2.3682 2.3306 2.1089 1.6223 0.8920 0.0169 0.8609 1.5989 2.0955 2.3263 2.3678
0.6 1.9922 1.9596 1.7726 1.3634 0.7495 0.0142 0.7235 1.3437 1.7613 1.9558 1.9918
0.4 1.6226 1.5942 1.4412 1.1081 0.6091 0.0116 0.5879 1.0921 1.4320 1.5911 1.6222
0.2 1.2663 1.2415 1.1211 0.8614 0.4734 0.0090 0.4569 0.8489 1.1138 1.2391 1.2658
2.75 0.0 0.9323 0.9105 0.8203 0.6296 0.3458 0.0066 0.3337 0.6204 0.8150 0.9086 0.9318
0.2 0.6313 0.6116 0.5485 0.4199 0.2304 0.0044 0.2223 0.4138 0.5449 0.6102 0.6307
0.4 0.3748 0.3564 0.3162 0.2406 0.1316 0.0025 0.1270 0.2371 0.3140 0.3553 0.3741
0.6 0.1754 0.1573 0.1345 0.1003 0.0543 0.0010 0.0524 0.0988 0.1334 0.1565 0.1746
0.8 0.0461 0.0271 0.0152 0.0079 0.0034 0.0001 0.0032 0.0077 0.0149 0.0265 0.0452
1.0 0.0000 0.0212 0.0299 0.0274 0.0162 0.0003 0.0157 0.0271 0.0300 0.0218 0.0010
1.0 2.5184 2.4788 2.2432 1.7257 0.9488 0.0180 0.9158 1.7008 2.2289 2.4742 2.5180
0.8 2.1718 2.1375 1.9342 1.4879 0.8181 0.0155 0.7896 1.4664 1.9218 2.1334 2.1715
0.6 1.8270 1.7973 1.6259 1.2506 0.6876 0.0130 0.6636 1.2325 1.6155 1.7939 1.8267
0.4 1.4880 1.4624 1.3223 1.0168 0.5589 0.0106 0.5395 1.0021 1.3138 1.4596 1.4877
0.2 1.1613 1.1393 1.0291 0.7909 0.4347 0.0082 0.4195 0.7795 1.0225 1.1371 1.1609
3.0 0.0 0.8550 0.8361 0.7538 0.5787 0.3179 0.0060 0.3068 0.5703 0.7489 0.8343 0.8546
0.2 0.5789 0.5623 0.5050 0.3870 0.2124 0.0040 0.2050 0.3813 0.5017 0.5610 0.5785
0.4 0.3437 0.3287 0.2925 0.2231 0.1221 0.0023 0.1178 0.2198 0.2905 0.3278 0.3432
0.6 0.1608 0.1465 0.1265 0.0949 0.0515 0.0010 0.0497 0.0935 0.1256 0.1459 0.1602
0.8 0.0422 0.0275 0.0177 0.0107 0.0051 0.0001 0.0049 0.0105 0.0174 0.0270 0.0415
1.0 0.0000 0.0163 0.0231 0.0211 0.0125 0.0002 0.0121 0.0209 0.0231 0.0168 0.0008
It will be observed from above that these coefficients are a function of the aspect
ratio of the slab. Normally the ratio of H/b for counterforts varies between 2 to 3.
We furnish below the coefficients for moments and shears for various values between
2 to 3 in increment of 0.25
47
.
We now explain the above with a typical numerical problem.
Example 3.6.2
A counter fort retaining wall of height 7.5 m has counter forts spaced at 3.0 m.
The average thickness of the wall is 300 m having RCC grade as M25. The wall
is resting on hard soil in zone IVas per IS-1893-2002. Unit weight of the backfill
soil is 22 kN/m
3
having friction angle of 28
. The soil is sloped to the horizontal

plane at angle I = 15
o
. Consider, E
conc
= 2.85 10
7
kN/m
2
and Poissons
ratio of concrete as 0.3.
Determine the time period of the wall and find out the horizontal dynamic
moments and shears in the wall under active earth pressure?
Solution:
Here we have, = 28
; i = 15
= 45 +0.5 28 = 59
.
Considering, =
cos sin i
sin(i)
=
cos 59sin 15
sin 44
= 0.192, for the problem.
T = 2
_
12 H
5
(1 + )(1
2
)
8Et
3
g tan (X
1
r
4
+X
2
r
2
+X
3
r
2
+X
4
)
= 2
_
12 22 7.5
5
(1 +0.192)(1 0.3
2
)
8 2.85 10
7
0.3
3
9.81tan 59(75.3 2.5
4
3.25 2.5
2
+17.68 0.3 2.5
2
+1.72)
= 0.03 sec
Thus,
S
a
g
= 1 +15T = 1.45, for hard soil; here, =
ZI
2R
=
0.241
4
= 0.06.
In the horizontal direction
M(y) = Coeff
y

0.423 S
a
8g tan
(1 + )H
3
and Q(y) = Coeff
y

0.423 S
a
8g tan
(1 + )H
2
M(y) =
0.423 0.06 22 1.45 7.5
3
1.192
8 9.81 tan 59
Coeff(y) = 30.58 Coeff(y)
Q(y) =
0.423 0.06 22 1.45 7.5
2
1.192
8 9.81 tan 59
Coeff(y) = 4.077 Coeff(y)
47 Intermediate values may be linearly interpolated.
Thus, multiplying by the appropriate coefficient, we can have M
y
and Q
y
as
given hereunder
Horizontal Moment M
y
H/b 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0
1.0 7.975 4.298 0.804 2.141 4.128 4.847 4.180 2.235 0.681 4.161 7.836
0.8 6.878 3.706 0.694 1.846 3.560 4.180 3.604 1.927 0.587 3.588 6.757
0.6 5.786 3.118 0.584 1.551 2.993 3.514 3.030 1.620 0.495 3.019 5.685
0.4 4.712 2.540 0.477 1.261 2.434 2.859 2.465 1.317 0.404 2.459 4.630
0.2 3.678 1.983 0.374 0.981 1.895 2.226 1.919 1.024 0.318 1.920 3.613
0.0 2.708 1.461 0.278 0.718 1.389 1.632 1.407 0.749 0.237 1.414 2.660
0.2 1.833 0.990 0.192 0.479 0.932 1.096 0.944 0.501 0.164 0.959 1.801
0.4 1.088 0.589 0.119 0.276 0.542 0.638 0.549 0.289 0.102 0.571 1.069
0.6 0.509 0.278 0.063 0.117 0.237 0.281 0.240 0.122 0.055 0.269 0.500
0.8 0.134 0.076 0.027 0.012 0.038 0.047 0.038 0.013 0.025 0.074 0.131
1.0 0.000 0.005 0.016 0.028 0.037 0.040 0.037 0.028 0.016 0.005 0.000
Horizontal Shear Q
y
H/b 1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0
1.0 12.307 12.114 10.963 8.434 4.637 0.088 4.476 8.312 10.893 12.091 12.305
0.8 10.614 10.445 9.451 7.271 3.997 0.076 3.858 7.165 9.391 10.426 10.612
0.6 8.929 8.781 7.943 6.109 3.358 0.064 3.241 6.020 7.892 8.764 8.927
0.4 7.272 7.142 6.455 4.963 2.728 0.052 2.633 4.891 6.414 7.128 7.270
0.2 5.675 5.560 5.018 3.855 2.118 0.040 2.044 3.799 4.986 5.549 5.673
0.0 4.179 4.074 3.667 2.813 1.544 0.029 1.491 2.772 3.643 4.065 4.176
0.2 2.829 2.732 2.445 1.870 1.025 0.019 0.990 1.843 2.429 2.725 2.826
0.4 1.680 1.585 1.400 1.063 0.581 0.011 0.561 1.047 1.390 1.581 1.676
0.6 0.786 0.690 0.582 0.431 0.232 0.004 0.224 0.424 0.578 0.687 0.782
0.8 0.206 0.104 0.044 0.013 0.002 0.000 0.002 0.013 0.042 0.101 0.202
1.0 0.000 0.115 0.162 0.149 0.088 0.002 0.085 0.147 0.163 0.118 0.006
3.7 UNYIELDING EARTH RETAINING STRUCTURES
3.7.1 Earthquake Analysis of rigid walls when the soil
does not yield
In many cases when a retaining wall is sufficiently rigid, or a basement wall of a
building or an underground tank is restrained at top by a rigid slab, the wall becomes
unyielding as such the triangular profile of failure which usually generates during the
static case does not take place. During such case (under static loading) we design it for
a condition of earth pressure at rest.
Now we have a catch! Since the soil is not under incipient failed condition, the
assumption made for dynamic analysis of cantilever and counterfort retaining wall as
shown previously is not valid in this case (where we ignored the stiffness of the soil)
and a completely different approach has to be adopted to obtain its dynamic response.
The major problem it boils down to is what will be the dynamic pressure induced
on the wall in such case?
Building
a
H z
Basement
BedRock

x
Figure 3.7.1 Schematic sketch of building with basement wall.
A procedure is proposed herein for evaluation of the same (Chowdhury and
Dasgupta 2007).
Shown in Figure 3.7.1, is a typical basement of a tall building (usually used for car
parking and sundry building services). The objective is to find the dynamic pressure
on the wall of the basement due to earthquake waves propagating through the soil
medium.
Normally for deep basement wall there will be built in intermediate floor slabs,
which make the wall sufficiently rigid due to which the soil in contact with the wall
does not yield and a plain strain condition prevails.
The soil medium and the basement is assumed to be resting on stiff soil considered
as the bedrock the level from which the ground acceleration propagates.
Let the depth of the basement is H and the soil medium in horizontal direction has
been considered having a finite dimension a, where lima , is a reality.
Having defined the problem with basic conditions we can argue that for shear
waves propagating through the soil medium the wave propagation equation can be
represented by
2
u(x, z, t)
x
2
+

2
u(x, z, t)
z
2
=
1
V
2
s
2
u(x, z, t)
t
2
(3.7.1)
where V
s
=shear wave velocity of the soil medium; u(x, z, t) =the displacement
function and can be considered as u = H(x)Q(z)G(t); three independent functions.
Without getting into the details of theory of partial differential equation it can be
shown that Equation (3.7.1) can be broken up into three ordinary differential equations
of second order, given by (Kreyszig 2001)
d
2
G
dt
2
+
2
G = 0, where, = v
s
i where i is a constant. (3.7.2)
d
2
H(x)
dx
2
+k
2
H(x) = 0, where, k is another constant. (3.7.3)
d
2
Q(z)
dz
2
+p
2
Q(z) = 0, where, p
2
= i
2
k
2
. (3.7.4)
The general solution of Eqns. (3.7.3) and (3.7.4) are given by
H(x) = Acos kx +Bsin kx : Q(z) = Ccos pz +Dsin pz (3.7.5)
We now impose the following boundary condition:
At x = 0, u = 0, or H(x) = 0, which implies A = 0.
At x = a (where a is very large), u = 0 or H(a) = 0 which implies, H(a) =
Bsin ka = 0, implying,
k = m/a. (3.7.6)
Thus, H
m
(x) = sin
mx
a
(3.7.7)
At the free surface i.e. the superstructure base interface we have,
At z = 0, shear strain
u
z
= 0 or
dQ(z)
dz
= 0, which implies D = 0.
At z = H, displacement, u = 0, i.e., Q(H) = 0 Ccos pH = 0 cos pH = 0, which
implies,
p =
(2n 1)
2H
. (3.7.8)
and hence, Q(z) = cos
(2n 1)z
2H
(3.7.9)
Thus, the eigenvector of the problem can be established as
(x, z) = H(x)Q(z) = sin
mx
a
cos
(2n 1)z
2H
where, m, n = 1, 2, 3. . .
(3.7.10)
For calculation of eigen values, we have, p
2
= i
2
k
2
and = v
s
i
thus, p
2
=

2
Vs
2
k
2
or = v
s
_
p
2
+k
2
.
Substituting the value of p and k from Eqns. (3.7.8) and (3.7.6) we have
= v
s
_
m
2
a
2
+
(2n 1)
2
4H
2
.
For the fundamental mode, considering m, n = 1 and, lima , we have
= = v
s
_
0 +
1
4H
2
=
v
s
2H
. (3.7.11)
Considering, T =
2
, we have, T =
4H
v
s
which is basically the free field time period
in one dimension for the site.
For lim a , we drop the first term of eigen function (in the x direction) in
Equation (3.7.10) to determine the displacement and pressure at wall face and consider
the eigen function as only
(z) = cos
(2n 1)z
2H
(3.7.12)
Based on modal response technique the maximum amplitude function can be
defined by
S
d
= S
a
/
2
where S
d
= maximum displacement, and S
a
=acceleration which is the function
of time period, 4H/v
s
, and can be read off from the normalized response given in
the code.
Considering, =
ZI
2R
, the code factor, we can write, u(z) =
i
S
a
2
(z).
Now substituting the value of from Equation (3.7.10), we have
u(z) =
4
S
a
H
2
v
2
s
cos
z
2H
=
4
S
a
s
H
2
Gg
cos
z
2H
(3.7.13)
where G = v
2
s
;
s
=unit weight of soil; g =acceleration due to gravity, and
i
=modal
mass participation factor =
m
i
i
/
m
i
2
i
.
The modal participation factor can thus be considered as
i
=
m
i
m
i
2
i
=
H
_
0
z cos
z
2H
dz
_
H
_
0
z cos
2
z
2H
dz (3.7.14)
The above, on integration by parts gives,
i
=
8
+2
.
The strain within the soil body is given by
xx
=
u
x
= 0;
zz
=
u
z
which gives,
zz
=
2
S
a
s
H
Gg
sin
z
2H

zz
=
16
( +2)
S
a
s
H
Gg
sin
z
2H
. (3.7.15)
The constitutive stress-strain relationship under plain strain condition is given by
_
_
_
xx
zz
xz
_
_
_
=
2G
(1 2)
_
_
1 0
1 0
0 0
12
2
_
_
_
_
_
xx
zz
xz
_
_
_
(3.7.16)
Thus,
xx
=
2G(1 )
1 2
xx
+
2G
1 2
zz
As
xx
= 0, in this case we have,
xx
=
2G
12
zz
, which gives the dynamic pressure
on the wall as
p
dyn
=
16
( +2)
2
1 2
S
a
s
H
g
sin
z
2H
(3.7.17)
Where, negative sign indicates that the pressure is acting in the direction of the wall.
p
dyn
=
32
( +2)
S
a
s
H
g
sin
z
2H
where
v
=

1 2
(3.7.18)
The above can be further simplified to
p
dyn
(z) = Coeff
S
a
s
H
g
, (3.7.19)
where the coefficients may be read off for different values of Poissons ratio (0.25, 0.3
and 0.4) as shown in Figure 3.7.2.
3.7.2 Ostadans method
Ostadan (2004) conducted extensive study on the subject based on analysis carried
out in SASSI 2000 (Lysmer et al. 1999) with various soil properties and Poissons
ratio value and came with a normalized dynamic pressure coefficient curve given by
the expression
p(z) = 0.0015 +5.05z 15.84z
2
+28.25z
3
24.59z
4
+8.14z
5
. (3.7.20)
The above normalized pressure coefcient when compared with the analytical solu-
tion proposed in Equation 3.7.19 gives quite closely matched value as shown in
Figure 3.7.3.
Variation of dynamic coefficient of pressure with depth
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Z/H
P
r
e
s
s
u
r
e

c
o
e
f
f
i
c
i
e
n
t
n=0.25
n=0.3
n=0.4
Figure 3.7.2 Variation of dynamic pressure coefficient with depth.
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
0
0
.
2
0
.
4
0
.
6
0
.
8
1
Z/H
P
r
e
s
s
u
r
e

c
o
e
f
f
i
c
i
e
n
t
Closed
form
Solution
Ostadan's
curve
Figure 3.7.3 Comparison of normalized pressure on the wall. (Proposed & Ostadan)
Based on the above equation, Ostadan proposed a simplied method for determi-
nation of dynamic pressure on such unyielding walls whose steps are as mentioned
hereunder:
Perform free field soil column analysis and obtain the ground response at the
depth corresponding to the base of the wall in the free field. The response motion
in terms of acceleration response spectrum at 30% damping should be obtained.
The free field soil column analysis may be performed using Computer programlike
SHAKE with input motion specified either at the ground surface or at the depth of
foundation base-mat. The choice for location of control motion is an important
-500
0
500
1000
1500
2000
0 3 6 9 12 15 18 21 24 27 30
Height in feet
D
y
n
a
m
i
c

P
r
e
s
s
u
r
e

(
p
s
f
)
Closed
form
Simplified
Method
Figure 3.7.4 Comparison of dynamic pressure for the 30 ft basement wall with V
s
= 1000 ft/sec.
decision that needs to be made consistent with development of the design motion.
The location of input motion may significantly affect the dynamic response of the
building and seismic soil pressure amplitude.
Obtain total mass of the soil body, m = 0.5 H
2

v
, for the present case
v
= 2/[(1 )(2 )]
0.5
, here is mass density of the soil and is Poissons
ratio.
Obtain the lateral seismic force from the product of the total mass obtained above
and the acceleration value of the free field response at the soil column obtained at
the depth of the bottom of the wall.
Obtain the maximum seismic soil pressure at the ground surface level by dividing
the lateral force obtained above by a factor 0.744 H (which actually the area under
curve for the equation furnished by him as mentioned above).
Obtain the pressure profile by multiplying the peak pressure from the above step
by the pressure distribution relationship as furnished earlier.
Ostadans method has been compared with analytical method proposed here
48
for
a basement wall which is 30 feet deep having shear wave velocity of soil as 1000 ft/sec.
The density of soil considered as 125 lbs/ft
3
.
The wall is considered to be in a zone subjected to severe earthquake where Z =
0.24, I = 1.2 and R = 2.0. The out come of the results are shown in Figure 3.7.4.
It will be observed that variation is not too wide and well within the acceptable limit
of civil engineering design.
3.8 EARTHQUAKE ANALYSIS OF WATER TANKS
3.8.1 Analysis of water tanks under earthquake force
Water tanks resting on ground, underground or on staging at a height form an impor-
tant part of infrastructure and township development. This is shown in Figure 3.8.1.
48 Personal communications with Dr. Ignacio Arango and Dr. Farhang Ostadan April 2005.
Figure 3.8.1 Typical overhead water tank modeled as an inverted pendulum.
Even in industrial sector like power, petrochemical industry it forms an important
ingredient of process engineering or fire ghting. As such in post earthquake scenario
many of the tanks storing water it becomes essential that they remain functional with
nominal damage.
In India water tanks resting on ground were usually given the reprieve of any
earthquake analysis not many years ago
49
.
While overhead tanks were mostly treated as an inverted pendulum where the time
period is derived from the expression
T = 2
_
stat
g
where,
stat
=
W
K
(3.8.1)
where, W = weight of water in tank + weight of tank + 1/3rd wt of staging frame,
and, K=lateral stiffness of the frame.
And water tanks both resting on ground as well as overhead have found to have
undergone damage during earthquakes in India and abroad. In India the damage is
possibly due to the fact that the IS-code recommendation on the pressure induced on
the tank wall is incomplete.
IS-1893-1984 for instance only suggests the pressure coefficients for impulsive force
and does not cater to the convective or hydro-dynamic sloshing force, whose effect
could be significant, especially if the tank is shallow. Nor does the code have any
provision for calculation of the sloshing height for which if sufficient free board is not
kept could result in damage of the roof slab.
49 For most of the case it was observed that impulsive force suggested by IS-1893 1984 provided with
forces of small magnitude unless and until the tank was situated in a place where earthquake has a
severe intensity.
Sloshed liquid
Sloshing Mass
h
Impulsive
Mass
Sa
D
Figure 3.8.2 Typical cylindrical tank with liquid divided into impulsive and sloshing mass.
Connected by flexible spring to wall
h

sl
Connected by rigid link to wall
h
i
Figure 3.8.3 Typical mathematical model of water tank with liquid impulsive and sloshing mass.
It is unfortunate that code committee did not update this previously though superior
mathematical model for analysis of such liquid retaining vessel has been in existence
since 1957 (Housner 1957).
Before we delve into the details of such mathematical model for analysis of such
water tanks, let us see how the fluid behaves, when acceleration is induced at the base
of such tanks.
Let us consider a cylindrical vessel of diameter D containing liquid of height h
(Figure 3.8.2). When the vessel is subjected to an acceleration S
a
at its base a part of
the liquid (called the impulsive mass) moves along with the container as a rigid body
at the bottom of the tank. Balance mass at the top of the tank acts in more flexible
manner and induces a convective or a sloshing force on the tank wall. What part of
the mass would act as an impulsive mass and sloshing mass depends
50
upon the aspect
ratio h/D.
In simplistic mechanical analogy the above can be represented as shown in
Figure 3.8.3. Haroun and Housner (1981) derived the values of sloshing and impulsive
mass considering the wall of the tank as rigid and later derived the same for exible
wall.
50 The total impulsive and sloshing mass constitutes the full mass of liquid in the tank.
We will not go into details of the derivation of the same from the fundamentals,
which are given elsewhere (Housner 1963) but will deal with final results only, both for
circular and rectangular tanks which can be directly used for computation of pressure
in a tank either manually or through a computer.
In Table 3.8.1,
Table 3.8.1 Design parameters for dynamic analysis of tanks with fluids.
Sl. No. Rectangular tank Circular tank
1
W
i
W
=
tan h
_
3
L
H
_
3
L
H
W
i
W
=
tan h
_
3
R
H
_
3
R
H
2 h
i
=
3
8
H (Excluding base pressure) h
i
=
3
8
H (Excluding base pressure)
3
h
i
H
=
3
L
H
2 tan h
_
3
L
H
_
1
8
h
i
H
=
3
R
H
2 tan h
_
3
R
H
_
1
8
(Including base pressure) (Including base pressure)
4 P
i
= S
a0
W
i
g
P
i
= S
a0
W
i
g
5
W
sl
W
= 0.527
L
H
tan h
_
1.58
H
L
_
W
sl
W
= 0.318
R
H
tan h
_
1.84
H
R
_
6
h
sl
H
= 1
cos h
_
1.58
H
L
_
1
1.58
H
L
sin h
_
1.58
H
L
_
h
sl
H
= 1
cos h
_
1.84
H
R
_
1
1.84
H
R
sin h
_
1.84
H
R
_
(Excluding base pressure) (Excluding base pressure)
7
h
sl
H
= 1
cos h
_
1.58
H
L
_
2
1.58
H
L
sin h
_
1.58
H
L
_
h
sl
H
= 1
cos h
_
1.84
H
R
_
2.01
1.84
H
R
sin h
_
1.84
H
R
_
(Including base pressure) (Including base pressure)
8 K
S
=
1.58
L
tan h
_
1.58
H
L
_
W
sl
K
S
=
1.84
R
tan h
_
1.84
H
R
_
W
sl
9
h
= 1.58
Sa
2
L
tan h
_
1.58
H
L
_

h
= 1.534
Sa
2
R
tan h
_
1.84
H
R
_
10 P
sl
= W
sl
h
sin t P
sl
= 1.2W
sl
h
sin t
11 d
max
=
0.527L cot h
_
1.58
H
L
_
g
h
L
1
d
max
=
0.408R cot h
_
1.84
H
R
_
g
h
R
1
W =Total weight of liquid in the tank;
W
i
= Weight of impulsive fluid assumed to fastened rigidly with the tank wall at
a height h
i
above the tank bottom;
h
i
= Height from the bottom of the tank where the impulsive force acts on the wall;
P
i
= Impulsive force which acts on the tank wall;
S
ao
= Acceleration at zero time period to be considered = g (9.81 m/sec
2
) as per
IS-1893 2002;
W
sl
= Sloshing weight of the fluid oscillating at the top part of the tank;
h
sl
= Height from the bottom of the tank where the sloshing force acts on the wall;
Ks = Equivalent spring stiffness of the sloshing fluid;
sl
= Angular amplitude of free oscillation at the free fluid surface;
P
sl
= Sloshing force acting on the tank wall;
d
max
= Sloshing Height of the liquid at free surface;
L = Half width of rectangular tank of width 2 L;
R=Radius of circular tank;
H =Height of fluid in the tank;
= natural frequency of the sloshing mass =
_
K
s
g
W
sl
.
The total moment acting on the base of the tank (when base pressure is included)
and on the wall (when base pressure is excluded) is given by
M = P
i
h
i
+P
sl
h
sl
(3.8.2)
Above is the original form in which Housner presented a solution to the problem
and has been the backbone of further research on this topic for next 30 years.
The basic assumption in Housners hypothesis which may not be always true (though
a conservative estimate) was the impulsive time period considered to be zero in his
analysis. The assumption was justified in the above case for Housner assumed the
tank to be infinitely rigid but in reality the tank could also be flexible when the time
period may have a finite value (albeit low compared to the sloshing time period).
Further researches by Veletsos & Young (1977) and Veletsos (1984) have defined
the impulsive time period when the wall is not rigid.
3.8.2 Impulsive time period for non rigid walls
Codes fromdifferent countries use different formulas for derivation of the time period.
We present here the most practical one which is easy to apply and amenable to
electronic computation.
Eurocode 8 (Part IV, 1998) suggests the formula proposed by Veletsos for circular
tank, which is given by
T
i
=
2R
C
i
_
H
E
w
t
where, C
i
= 0.01675
_
H
R
_
2
0.15
_
H
R
_
+0.46 (3.8.3)
Here R=radius of the tank; H =height of the fluid in tank; =mass density of
the fluid; E
w
= Youngs modulus of the wall, and t =thickness of the wall. [See
Figure 3.8.4].
w
e
P
h
H
w
Figure 3.8.4 Load deflection diagram of tank wall.
For rectangular tank, formula proposed in Eurocode 8 is given by
T
i
= 2
_
d/g (3.8.4)
where d is deflection of the wall due to an equivalent uniformly distributed load of
w
e
=
0.5W
i
+W
w
BH
(3.8.5)
where,
W
i
=weight of the impulsive fluid;
W
w
=weight of wall perpendicular to the direction of the earthquake force;
B=inside width of the tank;
H =height of fluid, and
w
e
=equivalent udl acting on the wall.
To calculate d, following steps may be considered, c.g. of the u.d.l., we may be
calculated from the expression
h =
0.5W
i
h
i
+W
w
H
w
0.5W
i
+W
w
(3.8.6)
In which, H
w
=height of the wall, An equivalent concentrated load P may be
calculated from the expression, P = w
e
H, where H =height of fluid in the tank.
The deflection at the level of fluid surface is thus given by
=
P(
h)
3
3EI
where, I =
t
3
12
(3.8.7)
It should be noted that the above is valid for tank walls which are free at top and
have an aspect ratio L/H > 2.0 i.e. it behaves as a one way slab. This may not be
valid for walls with other boundary conditions but is what is in vogue at present.
3.8.3 Sloshing time period of the vibrating f luid
The sloshing time period of the vibrating fluid for a circular tank may be obtained
from the expression
T
sl
= 2
_
D/g
_
3.68tan h
_
3.68H
D
_
(3.8.8)
where, D=diameter of the tank.
Similarly for rectangular tank,
T
sl
= 2
_
2L/g
_
3.16tan h
_
3.16H
2L
_
(3.8.9)
where, 2L=inside length of tank parallel to the direction of the earthquake.
3.8.4 Calculation of horizontal seismic force for tank
resting on ground
Once the time-period for the impulsive and sloshing modes is known the corresponding
accelerations may be obtained from the chart as furnished in the IS-code. The seismic
coefficient may then be obtained from expression
A
hi
=
ZIS
ai
2Rg
(3.8.10)
for impulsive force, and
[For RCC tank damping factor considered is usually 5% and for steel tank this is
considered as 2%, as such corresponding damping factor from the code should be
read].
A
hsl
=
ZIS
asl
2Rg
, (3.8.11)
for sloshing force where damping considered for fluid is normally 0.5%.
Here, Z=zone factor as explained earlier; I =importance factor, and R=ductility
factor.
The newIS-1893 (2002) is yet to arrive at the importance factor to be recommended
for liquid retaining structures; in absence of such data recommendations as followed
in UBC 97 may be followed:
For non important tanks consider, I = 1.0;
For tanks supplying water to public community or meant for fire ghting in impor-
tant industry like power plant or petrochemical plants or containing liquid having
nominal hazard I = 1.25;
For tanks containing hazardous or toxic liquid higher importance factor between
I = 1.5 to 1.75 may be considered.
The assessment of ductility factor is far more complex as the basis of S
a
/g as given in
UBC and that IS-1893 does not correspond one to one and a comparison of base shear
for similar structures has to be seen, which is obviously a topic of research. In absence
of such data presently, a conservative value of R between 1.5 and 2 may be used for
tanks resting on the ground. For overhead tanks similarly, a value of R = 1.21.3 may
be used for non ductile detailing and R = 22.2 may be used for tank frames with
ductile detailing.
3.8.5 Calculation of base shear for tanks resting
on ground
Once the seismic coefficients are known, the impulsive and sloshing shear force may
be obtained from
Vi = A
hi
(W
i
+W
w
+W
r
) and V
sl
= A
hsl
W
sl
(3.8.12)
where, W
i
=weight of sloshing fluid; W
w
=weight of wall, and W
r
=weight of roof
coming on the wall.
The resultant shear may either be considered as
V = V
i
+V
sl
(Absolute sum)
52
V =
_
V
2
i
+V
2
sl
(SRSS value). (3.8.13)
3.8.6 Calculation of bending moment on the tank
wall resting on the ground
The bending moment for impulsive and sloshing force at the base of the wall is given by
M
i
= A
hi
(W
i
h
i
+W
w
H
w
+W
r
H
r
) and M
sl
= A
hsl
W
sl
h
sl
(3.8.14)
where, h
i
= height of impulsive mass from the bottom of the tank as explained earlier;
H
w
=height of the c.g. of the wall mass; H
r
=height of the c.g. of the roof mass,
and
H
sl
=height of the sloshing mass from the bottom of the tank.
In this case the height h
i
and h
sl
shall be calculated for the case Excluding base
pressure.
For moment in the base of the tank (i.e. the pressure induced in the soil) the same
expression as above may be used with exception that the height h
i
and h
sl
shall be
calculated for the case Including base pressure.
52 This is the recommendation of Euro code 8.
The resultant Moment can now be obtained based on expression
M = M
i
+M
sl
(Absolute sum) M =
_
M
2
i
+M
2
sl
(SRSS value) (3.8.15)
3.8.7 Calculation of sloshing height
In absence of any recommendation form IS code presently sloshing height may be
calculated based on Euro-code model
= 0.84 A
hsl
L for rectangular tank,
= 0.84 A
hsl
radius of tank for circular tank. (3.8.16)
We now further elaborate the problem based on a suitable numerical example.
Example 3.8.1
A rectangular RCC fire water tank (Figure 3.8.5) is resting on ground having a
size of 7.5 m 7.5 m 6.5 m is constructed in a refinery site which is classified
as zone IV as per IS-1893 2002. The average thickness of wall is considered to
be 450 mm. Grade of concrete used for constructing the tank is M30. The tank
is covered by a roof slab which is simply supported on the four walls having
thickness of 200 mm. Nature of ground on which it is resting is considered hard.
Calculate the seismic force at the wall base and on the foundation.
500
6000

Figure 3.8.5 Elevation and plan view of the water tank with typical wall slab detail resting on
ground.
Solution:
Weight of water in tank = 7.5 7.5 6 10 = 3375 kN; Wt. of roof slab =
8 8 0.2 25 = 320 kN, and Wt. of one wall = 7.5 6.5 0.45 25 =
548 kN.
Here L = 0.5 7.5 = 3.75 m, and H = 6.0 m.
Based on Housners expressions, impulsive mass is given by,
W
i
=
tan h
_
3
L
H
_
3
L
H
W, =
tan h
_
3
3.75
6
_
3
3.75
6
3375 = 2475.8 kN
h
i
=
3
8
H = 2.25 m (EBP)
53
= H
_
_
3
L
H
2tanh
_
3
L
H
_
1
8
_
_
= 3.342 m (IBP)
Sloshing mass is given by
W
sl
= 0.527
L
H
tan h
_
1.58
H
L
_
W = 1097.6 kN.
h
sl
= 1
cos h
_
1.58
H
L
_
1
1.58
H
L
sin h
_
1.58
H
L
_H = 3.977 m (EBP);
h
sl
= 1
cos h
_
1.58
H
L
_
2
1.58
H
L
sin h
_
1.58
H
L
_H = 4.358 m (IBP)
Equivalent weight, w
e
=
0.5W
i
+W
w
BH
=
0.52476+548
7.56
= 39.7 kN/m
2
.
C.G. of the load,

h =
0.5W
i
h
i
+W
w
H
w
0.5W
i
+W
w
=
0.524762.25+5483.25
0.52476+548
= 2.56 m.
Equivalent concentrated force P, P = w
e
H 1 = 39.7 6 1 = 238 kN.
I =
t
3
12
=
1 (0.45)
3
12
= 7.59375 10
3
m
4
; E
conc
= 3.122 10
7
kN/m
2
.
d =
P(
h)
3
3EI
=
238 (2.56)
3
3 3.122 10
7
7.5938 10
3
= 5.614 10
3
m.
53 Here EBP means excluding base pressure and IBP = Including base pressure.
Calculation of impulsive time period
T
i
= 2
_
d
g
= 2
_
5.614 10
3
9.81
= 0.150 sec
a
/g = 2.5 as per IS-1893.
Calculation of sloshing time period
Considering, 2L = 7.5 m; H = 6 m; T
sl
= 2
_
2L
g
3.16tan h(
3.16H
2L
)
= 3.110 secs
which gives, S
a
/g = 0.321, for the sloshing mode with 5% damping and it needs
to be multiplied by a factor of 2.98 to convert it into an equivalent acceleration
with 0.5% damping for the fluid.
For zone IV consider Z = 0.24 and I = 1.25, thus
A
hi
=
ZIS
ai
2Rg
, for impulsive force
=
0.24 1.25 2.5
2 2
= 0.1875, and
A
hsl
=
ZIS
asl
2Rg
, for sloshing mode
A
hsl
=
0.24 1.25 0.321
2 2
2.98 = 0.072
(Mulitpiled by a factor 2.98 to cater to 0.5% of damping for fluid).
Calculation of Base Shear
Weight of roof = 320 kN, so assume 1/4th of the weight coming on each wall.
Hence,
Vi = A
hi
(W
i
+W
w
+W
r
) =0.1875 (2476 +548 +
1
4
320) =582 kN,
for impulsive force.
Vsl = A
hsl
W
sl
= 0.072 1098 = 79 kN for sloshing force.
Thus resultant shear is given by, V =
_
V
2
i
+V
2
sl
= 587 kN.
Calculation of Bending Moment at base of wall
M
i
= A
hi
(W
i
h
i
+W
w
H
w
+W
r
H
r
) = 0.1875 (2476 2.25 +548 3.25 +
80 6.6) = 1478 kN m, for impulsive mode
M
sl
= A
hsl
W
sl
h
sl
= 0.072 1098 3.977 = 314.40 kN m, for sloshing
mode.
Resultant Moment, M =
_
M
2
i
+M
2
sl
= 1511 kN m.
Calculation of Bending Moment at foundation
M
i
= A
hi
(W
i
h
i
+W
w
H
w
+W
r
H
r
) = 0.1875(24763.342+5483.25+
80 6.6) = 1984 kN m, for the impulsive mode.
M
sl
= A
hsl
W
sl
h
sl
= 0.072 1098 4.358 = 344.52 kN m, for the sloshing
mode.
Resultant Moment, M =
_
M
2
i
+M
2
sl
= 2013 kN m
Calculation of sloshing height
= 0.84 A
hsl
L = 0.84 0.072 3.75 1000 = 227 mm, less than the
free board of 500 mm provided.
3.9 MATHEMATICAL MODEL FOR OVERHEAD TANKS UNDER
EARTHQUAKE
3.9.1 Earthquake Analysis for overhead tanks
Water served to communities for daily use and even for industrial purpose in many
cases is stored in overhead tanks, so that it can be distributed under adequate pressure.
Thus in a post earthquake scenario it is essential that they remain serviceable to serve
the community and also mitigate secondary damages like fire.
Overhead water tanks come in different shapes like rectangular, Intze type, conoids
etc supported on frames constituting of beam columns or single circular shaft etc. as
shown Figure 3.9.1.
Figure 3.9.1 Typical overhead water tank with its staging system modeled as two mass system.
In this case also the liquid in the tank may be considered as a two mass lumped
system like the case of tank resting on ground constituting of impulsive and sloshing
mass. The impulsive and the sloshing mode may be treated as two uncoupled system
where the impulsive mass of the fluid as obtained by the Housners expression may be
added to the tank mass and 1/3rd of that of the staging and whose dynamic response
may be obtained from the expression
T = 2
_
W
i
+W
t
+
1
3
W
s
gK
s
(3.9.1)
where, W
i
= weight of impulsive fluid; W
t
= weight of the tank; W
s
= weight of the
staging, and K
s
= stiffness of the staging frame.
The sloshing time period may be obtained from the expression
T = 2
_
W
sl
gK
sl
(3.9.2)
where, W
sl
is the sloshing mass as per Housners expression and K
sl
is the fluid stiffness,
given by
K
sl
H
W
= 0.83266tan h
2
_
1.58
H
L
_
(3.9.3)
else they can be obtained from the graph as shown in Figure 3.9.2.
Impulsive,Sloshing mass and stiffness Ratio
0
0.2
0.4
0.6
0.8
1
1.2
0
.
1
0
.
4
0
.
7
1
1
.
3
1
.
6
1
.
9
2
.
2
2
.
5
2
.
8
L/H
D
e
s
i
g
n

R
a
t
i
o
Wi/W
Ws/W
KsH/W
Figure 3.9.2 Impulsive, sloshing mass and stiffness parameters for rectangular tank with different L/H.
In the above analysis everything is fine except the fact that the staging stiffness needs
to be evaluated. The easiest way it can be done is by modeling the frame in a computer
analysis program like STAAD-Pro, SAP 2000, GTSTRUDL, etc. and apply a unit load
at center of mass of the tank and the water and find out the deflection at the base of
the tank (top of staging).
Knowing the deflection the stiffness value K
s
may be obtained fromthe relationship,
P = K
s
d (3.9.4a)
where, d is the deflection and P is the applied load. Else, for a regular frames the
stiffness may be obtained from the formula
K
s
=
12nEI
jL
3
(3.9.4b)
where n is the number of coulumns in the frame and j is the number of storey and
L is the height of column per storey.
Once the time periods are established as per Equation (3.9.1) the calculation becomes
quite straight forward.
The shear force at the top of the frame, shown in Figure 3.9.3, is given by
Vi = A
hi
(W
i
+W
T
+
1
3
W
fr
) and V
sl
= A
hsl
W
sl
(3.9.5)
where, W
i
= weight of impulsive fluid; W
T
= weight of the tank; W
fr
= weight of the
frame, and W
sl
= weight of sloshing fluid.
Unit Load acting at the cg of tank+water
Rigid Links(Typical)
Nodes(typical)
Beam elements(typical)
Figure 3.9.3 Typical computer model for staging for determing the deflection/stifness of the frame.
And,
A =
ZI
2R
S
a
g
(3.9.6)
In which, S
a
= the accelerations due to impulsive and sloshing mode time periods
as calculated above.
The resultantant shear at the top of the frame is given by
V =
_
V
2
i
+V
2
sl
(3.9.7)
The overturning moment in impulsive mode at the base of staging is thus given by
M
i
= A
hi
_
W
i
_
h
i
+Hst
_
+
_
W
T
+
1
3
Wst
_
H
cg
_
(3.9.8)
where, h
i
= impulsive height of the fluid as per Housners expression considering the
Including base pressure case, H
st
= height of the staging frame, and H
cg
= height
from the base of staging to the c.g. of the tank + fluid.
The overturning moment in sloshing mode at the base of staging is thus given by
M
sl
= A
hsl
W
sl
(h
sl
+H
st
) (3.9.9)
where, h
sl
= sloshing height of the fluid as per Housners expression considering the
Including base pressure case.
Impulsive, sloshing and stiffness ratio for
circular tank
0
0.2
0.4
0.6
0.8
1
1.2
0
.
1
0
.
4
0
.
7
1
1
.
3
1
.
6
1
.
9
2
.
2
2
.
5
2
.
8
Radius/Height
D
e
s
i
g
n

R
a
t
i
o
Ws/W
KsH/W
Wi/W
Figure 3.9.4 Impulsive, sloshing mass and stiffness parameters for circular tank with different R/H.
For circular tank the steps remain exactly same except the stiffness value which gets
modified to
K
sl
H
W
= 0.58512 tan h
2
_
1.84
H
R
_
. (3.9.10)
The sloshing design parameters can also be obtained form the graph as furnished in
Figure 3.9.4.
3.9.2 Hydrodynamic pressure on tank wall and base
Other then knowing the overturning moement at the base of the tank and the maxi-
mum moment in the wall for big tanks, it is essential that we know the hydrodynamic
pressure on the wall and the base of the tank, without which curtailment of reinforce-
ment in the wall is not possible. The hydrodynamic distribution of pressure in the tank
is given in the following.
3.9.3 Hydrodynamic pressure for circular tank
The impulsive pressure is given by
p
hi
(dyn) =
i
(z)A
hi
gH cos (3.9.11)
where,
i
= 0.866
_
1
_
z
H
_
2
_
tanh
_
0.866
D
H
_
, in which, = mass density of the
liquid; = circumferential angle, and z = vertical distance of a point from the bottom
of the tank wall. Impulsive hydrodynamic pressure in vertical dirfection on a strip of
length b is given by
p
vi
(dyn) = 0.866A
hi
gH
sin h
_
0.866
x
h
_
cos h
_
0.866
b
h
_ (3.9.12)
The sloshing pressure for circular tank wall is given by
p
si
(dyn) =
s
(z)A
si
gD
_
1
1
3
cos
2
_
cos (3.9.13)
where,
s
= 0.5625
cos h(3.674
z
D
)
cos h(3.674
H
D
)
.
Sloshing pressure in vertical direction on the base slab is given by
p
vs
(dyn) =
sv
A
hs
gD (3.9.14)
where,
sv
= 1.125
_
x
D

4
3
_
x
D
_
3
_
sec h
_
3.674
H
D
_
.
where, x = horizontal distance of a point on base of tank in the direction of seismic
force from the center of tank.
3.9.4 Hydrodynamic pressure for rectangular tank
For rectangular tank the impulsive pressure is given by
p
hi
(dyn) =
i
(z)A
hi
gH (3.9.15)
where,
i
= 0.866
_
1
_
z
H
_
2
_
tan h
_
0.866
2L
H
_
, in which
Impulsive hydrodynamic pressure in vertical dirfection on the base slab is given by
p
vi
(dyn) = 0.866A
hi
gH
sin h(0.866
x
h
)
cos h(0.866
2L
h
)
(3.9.16)
The sloshing pressure on tank wall is given by
p
si
(dyn) =
s
(z)A
si
g (2L) (3.9.17)
where,
s
= 0.4165
cos h(3.162
z
2L
)
cos h(3.674
H
2L
)
.
Sloshing pressure in vertical direction on the base slab is given by
p
vs
(dyn) =
sv
A
hs
g (2L) (3.9.18)
where,
sv
= 1.25
_
x
2L

4
3
_
x
2L
_
3
_
sec h
_
3.162
H
2L
_
.
3.9.5 Effect of vertical ground acceleration
Vertical gorund acceleration increases the effective weight of liquid which induces
additional pressure on the tank wall. The distribution of this pressure is taken as
similar to that of hydrostatic force.
Hydrodynamic pressure on tank wall due to vertical acceleration may be taken as
p
vw
=
2
3
A
h
gH (1 z/H) (3.9.19)
where, A
h
=
1.25ZI
R
as per IS-code
3.9.6 Pressure due to inertia of the wall
Pressure due to inertia of the wall itsdelf may be taken as
p
mw
(dyn) = A
hi
t
w
g (3.9.20)
where, t = average thickness of the wall, and
w
= mass density of the tank wall.
3.9.7 Maximum design dynamic pressure
The maximum dynamic design pressure may be obtained by taking the SRSS value of
the above pressures and be expressed as
p
des
(dyn) =
_
(p
hi
+p
mw
)
2
+p
2
si
+p
2
vw
(3.9.21)
Example 3.9.1
Shown in Figure 3.9.5 is an elevated rectangular water tank of capacity 500 m
3
resting on medium soil which is classified as falling in a zone of IV as per IS
Code. Grade of concrete used is M25. Based on dynamic analysis find out the
overturning moment on the top of the foundation. Shear at top of staging, and
hydrodynamic pressure on the tank wall and base. Code to be used is IS 1893-
2002. Col size 600 600, beam size 500 750.
9000 250(typ)
6350
750(typ)
4250(typ) 4450 4450
4450
20,000
4450
Plan view of column
Figure 3.9.5
Solution:
Capacity of water tank = 500 m
3
; Weight of water in tank = 500 10 =
5000 kN, and Height of water = 500/81 = 6.17 = 6.2 m (say).
Calculation of weight of tank
Weight of roof slab (150 mm thick) = 9.5 9.5 0.15 25 = 338.4 kN
54
Weight of side wall = 9.25 4 6.35 0.25 25 = 1468.4 kN
Weight of bottom slab(350 mm thk) = 9.5 9.5 0.35 25 = 789.7 kN.
Thus weight of empty tank = 338.4 + 1468.4 + 789.7 = 2596.5 kN.
Calculation of weight of staging
Height of staging = 20 m.
Size of column = 600 600.
Thus weight of 9# of column = 20 0.6 0.6 9 25 = 1620 kN.
Length of peripherial beam = 35.6 = 36 meter(app).
Size of beam = 500 750 mm
2
.
Weight of peripherial beams = 36 0.5 0.75 25 4 (levels) = 1350 kN.
Length of internal beams = 18 m (app).
Weight of internal beams = 18 4 (levels) 0.5 0.75 25 = 675 kN.
Total weight of staging = 1620 + 1350 + 675 = 3645 kN.
Calculation of impulsive weight of water.
As per Housner
W
i
W
=
tan h
_
3
L
H
_
3
L
H
Here, L = 4.5 m, H = 6.2 m, W = 5000 kN; thus, W
i
= 5000
0.8503
1.257
=
3381 kN.
Calcualtion of sloshing weight of water
W
sl
W
= 0.527
L
H
tan h
_
1.58
H
L
_
W
sl
= 5000 0.527
4.5
6.2
tan h
_
1.58
6.2
4.5
_
= 1864 kN
54 Unite Weight of concrete considered as 25 kN/m
3
.
Calculation of column stiffness
K
col
=
12nEI
jL
3
Here, E
c
= 2.85 107 kN/mm
2
; I = (1/12) 0.6 0.6
3
= 0.0108 m
4
EI = 307800 kN/m
2
.
Here, n = 9, and j = 4 and height of column, L = 5 m, and we have,
K
col
=
12nEI
jL
3
=
12 9 307800
4 5
3
= 66485 kN/m.
Calculation of impulsive time period
T = 2
_
W
i
+W
t
+
1
3
W
s
gK
s
= 2
_
3381 +2597 +
1
3
3645
9.81 66485
= 0.6598 sec.
For soil of medium strength, we have
S
a
/g = 1.36/T = 1.36/0.66 = 2.06 m/sec
2
for 5% damping.
Considering 7% damping for RCC design S
a
/g is multiplied by a factor 0.9,
as per IS-1893 2002
Sa/g = 2.06 0.9 = 1.855 m/sec
2
.
For zone IV, Z = 0.24; Importance factor = 1.5 (say), and R = 1.5, for non
ductile frame.
A
hi
=
ZI
2R
S
a
g
=
0.24 1.5
2 1.5
1.855 = 0.2226
Thus, impulsive shear at base slab level of the tank
V
hi
= 0.2226 (3381 +2597 +1/3 3645) = 1626 kN.
Calculation of sloshing time period
T
sl
= 2
_
2L
g
_
3.16tan h
_
3.16H
2L
_
and substituting the value of H =6.2 mand L =4.5 mas stated above, we have,
T
sl
= 3.43 secs.
Considering,
_
S
a
g
_
s
=
1.36
T
= 0.3965 m/sec
2
for 5% damping.
Considering, the fluid damping as 0.5%as per IS-18932002, the above value
gets modified to
_
S
a
g
_
s
= 0.3965 2.98 = 1.18 m/sec
2
.
Thus, A
hs
=
ZI
2R
S
a
g
=
0.241.5
21.5
1.18 = 0.1416
Thus sloshing shear at base slab level of the tank
V
hi
= 0.1416 1864 = 264 kN (W
sl
= 1864 kN, the sloshing weight of water)
Thus, total shear acting at bottom of the tanks (top of staging) =
_
(1626)
2
+ (264)
2
= 1647 kN.
Overturning moment at the base of the staging
h
i
H
=
3
L
H
2
tan h
_
3
L
H
_
1
8
; h
i
=
3
4.5
6.1
x6.2
2tan h
_
3
4.5
6.1
_
6.2
8
= 3.80 m;
H
cg
=
2597 23.30 +
3645
3
10
2597 +1215
= 19.06 m.
M
i
= A
hi
_
W
i
_
h
i
+Hst
_
+
_
W
T
+
1
3
Wst
_
H
cg
_
M
i
= 0.2226 (3381 (3.80 +20) + (2597 +
1
3
3645) 19.06)
= 34086 kN m
For the sloshing mode
h
sl
H
= 1
cos h
_
1.58
H
L
_
2
1.58
H
L
sin h
_
1.58
H
L
_
Substituting the values H = 6.2 m and L = 4.5 m, we have, h
sl
= 4.586 m
Thus, M
sl
= A
hsl
W
sl
(h
sl
+ H
st
) = 0.1416 1864 (4.586 + 20) =
6489 kN m.
Thus, the resultant overturning moment is given by
M =
_
34086
2
+6489
2
= 34698 kN . m
Calculation of hydrodynamic pressure
Based on the Equation 3.9.21 hydrodynamic pressure on wall for various cases are
furnished hereafter
Impulsive Sloshing pressure Presure due to Pressure due Resultant design
z/H pressure on wall wall vertical acceleration wall inertia pressure(kN/m
2
)
0 10.21798 1.02172 10
18
9.76 1.4125 15.18307
0.1 10.1158 3.91873 10
17
8.784 1.4125 14.49346
0.2 9.809262 3.00498 10
15
7.808 1.4125 13.67087
0.3 9.298363 2.30469 10
13
6.832 1.4125 12.70428
0.4 8.583104 1.7676 10
11
5.856 1.4125 11.58468
0.5 7.663486 1.35567 10
09
4.88 1.4125 10.30475
0.6 6.539508 1.03974 10
07
3.904 1.4125 8.858648
0.7 5.21117 7.97436 10
06
2.928 1.4125 7.241974
0.8 3.678473 0.000611599 1.952 1.4125 5.452368
0.9 1.941416 0.046906968 0.976 1.4125 3.493355
1 2.27E-15 3.5975604 1.08E-15 1.4125 3.864919
Similarly in vertical direction hydrodynamic pressure on base slab is
Sloshing pressure in Impulsive pressure in Design pressure
X/2L vertical direction vertical direction (kN/m
2
)
0 0 0 0
0.05 0.183688 0.58238833 0.61067
0.1 0.36369 1.167154385 1.222506
0.15 0.53632 1.756685596 1.836732
0.2 0.697892 2.35338885 2.454688
0.25 0.84472 2.959700315 3.077885
0.3 0.973117 3.578095387 3.708062
0.35 1.079398 4.211098795 4.347235
0.4 1.159877 4.861294911 4.99775
0.45 1.210868 5.531338298 5.662323
0.5 1.228683 6.223964552 6.344084
3.10 PRACTICAL ASPECTS OF EARTHQUAKE ENGINEERING
We are almost through with this chapter and take this opportunity to pacify those
hardened professionals whom we had perhaps bored to stupor with double integrals,
partial differentials and hyperbolic trigonometric functions.
Before we start with this topic, we would like to point out that perhaps we could
make the reader realize that it is not a very easy subject to grasp. It requires competence
in multifaceted subject like engineering geology, soil dynamics, structural dynamics,
fluid dynamics and finally applied mathematics which is not a very easy thing to
achieve in a nutshell. It is for this, specialists from different field convene together
to pool their knowledge and experience to develop a unified design policy which is
otherwise known as the code of practice.
55
55 And violations of the same, in the name of economy or financial budget is not uncommon.
A professional engineer undertaking a design is usually guided by three things
Cost involved
Engineering and construction schedule
Sociological importance of the structure in hand.
To what sophistication an analysis should be carried out depends a lot on the budget
the client has the engineering and construction schedule he has to meet and social
outcry it would create in case the structure undergoes damage during an earthquake.
For instance a commercial building or a hotel sustaining damage during an earth-
quake without collapse would cause a much lesser furor then a reactor building or
a heavy water container undergoing even a minor crack during an earthquake. For
radiation effects emanating from those cracks could have a catastrophic effect on the
surrounding and would possibly result in complete shut down of the plant till such
cracks are rectied. This would possibly result in power shortage in an area for months
and could result in a huge revenue loss for industries dependent on such power.
Chemical plants storing toxic and hazardous material if undergoes damage can again
have deadly consequence on the surrounding and can ravage the ecological balance so
badly that it could take years to restore the same.
Public building like hospitals, town halls, schools where people mostly take refuge
in the post earthquake scenario must remain functional for relief work to be effectively
carried out.
People could surely argue that whats the big deal? As the code suggests we take a
higher importance factor and design it for a higher force. It is indubitable a fact that
the argument do have some substance in it.
But it has been seen in many cases that though the force induced in the structure was
possibly lesser then expected, structures have undergone a spectacular failure while
there are structures which was subjected to a far higher force then it was designed for
and yet it has survived the shock with only minor damages.
Reasons attributing to such spectacular failures have been very simple.
The structures were inherently planned poorly making them generically weak
under earthquake force
56
And last but not the least improper detailing causing improper stress dissipation
path resulting in considerable damage to the structure.
We would like to re-emphasise at this point that irrespective of the most sophisticated
analysis one undertakes the most advanced software one may use if the same is not
followed up with well conceived structural arrangement and proper detailing can still
result in collapse.
On the other hand, analysis of structures based on simple seismic coefcient method
and plane frame analysis carried out by simple portal method, but detailed properly
and having robust structural conguration has been found to survive severest of the
shock.
56 Like irregular geometry in plan causing additional torsion not catered to properly etc.
One of the major limitations a civil engineer faces specially in the building industry is
that in many of the major buildings during conceptual stage he has very little control
on selection of material and planning of the functional space for this is principally
controlled by the architects
57
. But howwell the structure will behave under earthquake
depends a lot on these decisions. If the architect concerned does not have appreciation
of the problems earthquake could create, may lead to a situation of impassewhen in
extreme case can even result in replacement of the project civil engineer
58
.
While it is surely not the job of a structural engineer to put spanner in every aspiring
wheel an architect could conjure, yet if he sees something that could seriously mar the
performance of the building should be pointed out clearly if possible with compar-
ative numbers enabling an architect to make a quantitative assessment of the issue.
The bottom line is that it is necessary to have an unbiased continuous and an open
dialogue between the architect and the engineer to arrive at the most optimal shape
and configuration which is structurally sound and aesthetically pleasing.
During planning stage of structure at a location susceptible to severe earthquake if
some fundamental rules are followed and adhered to much of the risk of a collapse can
surely be significantly mitigated. We discuss a few of the important ones hereafter
Avoid the fundamental period of the building to be near the free field site of the
motion, equating the two as shown in Chapter 1 (Vol. 2) one can arrive at the
critical height of the building which an architect could be made aware to suppress
the seismic excitation of the building.
Avoid irregular geometry in plan these creates additional force. If not properly
taken care off can lead to significant damage of the building. We explain this
point with an example.
Shown in Figure 3.10.1 is a typical plan view of a school building with a play-
ground
59
. It is obvious that shear center of such building will be along the chain
dotted line as shown in the figure. Thus an earthquake force acting on the building
would act along this line and would invariably create additional torsion in the building
which if not properly catered for could result in severe cracking at the junction.
Now the point remains is that does this functional concept be rejected at the outset
citing it is dangerous?
For wearing an architects hat one can envisage a number of functional advantage
with this type of configuration as a school building. So, what are the other options a
structural engineer is left with? He can surely under take a detailed dynamic analysis
of such building and cater to the additional torsion or can simply do the following:
The trick is simple break up the building into three regular modules as shown
in Figure 3.10.2 by providing construction joints as shown which surely makes the
57 Not to mention, those mafias (with high level political nexus) who in the name of promoters today
control almost everything in building industry and have polluted the complete work ethics of the building
industry in India.
58 If the firm concerned is primarily an architectural firm.
59 A common feature one observes in many public schools in England and Germany, playground in front
is usually called a quadrangle.
Figure 3.10.1 Typical plan of a school building.
Construction joint
(Typ.)
Figure 3.10.2 Plan of a school building with construction joint.
configuration much simpler to handle while the functional configuration conceived by
the architect can still be maintained
60
.
Avoid rapid change in stiffness in vertical direction.
This is a common problem faced in many congested urban area where due to lack
of space the ground floor is completely kept open for the cars to park while the top
portion constitute of residential or office complex with usual curtain walls as shown
in Figure 3.10.3.
In such cases, the top portion of the building having higher stiffness would possibly
have a lowtime period which shows the structure is quite stiff. This would thus attract
a significant force which when gets transferred to the foundation, suddenly finds a level
which has a much poorer stiffness then rest of the building and if the members are not
sturdy enough to transfer this shear force would invariably result in a failure at the
column beam junction as marked as the weak zone.
60 A similar situation can happen in a pipe rack configuration too where due to process requirement and
pipe stress limitations the configuration cannot be changed. Again opting for a separation joint will do
the trick in such case.
Weakest link
Figure 3.10.3 Typical office building with open space for car parking in ground floor.
Figure 3.10.4 Office building with open shear wall and access cut-out.
It can well be envisaged that while we cannot reject the option
61
, yet try to arrive at a
solution which would make the building safe against earthquake. The easiest solution
would to provide one bay with a shear wall which would be stiff enough to absorb
the load as shown in Figure 3.10.4.
We had just cited a fewexamples to give some idea. All international codes including
IS-1893 (2002) has come up with do and donts in termof building planning and should
be adhered to as much as practicable.
Detailing is another aspect that needs to be given proper attention. Ductile detailing
as such is gaining importance more and more to attenuate the effect of earthquake
force. IS code has developed a special code, IS-13920 for the same which should be
adhered to. Ductility is an important aspect which safe guards a structure by dissipating
the energy induced in the body due to seismic force by cracking thus preventing a total
collapse. A detailed discussion is beyond the scope of this book and interested reader
may refer to a number of excellent reference like Park & Pauley 1975, Dowrick 2002.
61 For doing so might result in a rejection of the building plan itself citing provisions have not been kept
for adequate parking space and the client does not have budget to provide a basement parking.
A word or two for the rookies. . . . . .
In eagerness to carry out a clever and sophisticated analysis it is not uncommon to
see freshmen start with an elaborate computer model at the outset and finally get lost in
maze of numbers loosing sight of the basic issue as to how the structure is behaving in
general. Whether there is an inherent flawin the arrangement that generates additional
forces in one part of the system or not etc.
Study the plan carefully, if required seek help of more experienced people and dis-
cuss. Always start with a simple model like a stick model or a simple plane frame
model to have a first feel of the system behavior before launching into a more detailed
and elaborate analysis we can assure you this will eliminate a number of erroneous
design decisions vis-a-vis re-work at later stage of the project when cost implication
in terms of schedule can be more severe.
3.10.1 Epilogue
We are almost at the end of the road, for readers seeking more information on the
subject we would encourage him to read the literatures mentioned at the end of the
chapter. Read them, if you are really interested in this topic we can assure you that you
will enjoy them immensely. Considering this is not a handbook our intention is not to
work out design problems in completeness enabling one to follow them blindly. The
purpose was to provide you with the basic essence of the phenomena and encourage
you to understand the fundamental mechanics behind it.
Finally a word of apology, to the bridge engineers for not having addressed -such
an interesting topic.
The reasons were basically the following:
Bridge engineering being a topic by itself would significantly increase the volume of
this book
62
, finally private bridges in India is a rare commodity and most of the bridges
and flyovers are controlled based on legislation of IRC
63
and guidelines prescribed by
MOST
64
.
Even though detailed dynamic analysis is possible for such bridges however, as per
Indian practice, most of the dynamic loads coming on the bridges due to moving
vehicles or earthquake are catered for based on dynamic load factor or pseudo static
methods. Applying too much sophistication in their analysis may be construed as a
rude intrusion in their all is fine world of slumber and may not be approved by
the legislative body. In USA bridge codes put forward by AASHTO
65
have regular
provisions for design of bridges under dynamic earthquake loading are not officially
recognized in the country. Till such modifications are brought about by IRC and
MOSTwe thought it prudent not to venture in this otherwise a very interesting subject.
62 The intenion has never been to make international weightlifters out of the reader.
63 Indian Road Congress.
64 Ministry of Surface Transport of India.
65 American Authority of state highway official.
References
1 Alpan, I. 1961, Machine foundation and soil resonance, Geotechnique, Vol. 11.
2 Almuti, A.M. 1976, Large exible Turbine foundation, Methods of Structural Analysis,
ASCE, NY, pp. 707719.
3 Ambraseys, N. (1995), The prediction of earthquake peak ground acceleration in Europe.
J. Earthq. Eng. & Str. Dyn., Vol. 24, pp. 467490.
4 Aneja, I. 1975, Dynamic Response of SystemsTurbine generators on Various Founda-
tions, Proceedings of the American Power conference, Vol. 37, pp. 528540.
5 Applied Technology Council 1982, Tentative provisions for the development of seis-
mic regulations for buildings, ATC 3-06, NBS SP-510, NSF 78-8, National Bureau of
Standards USA.
6 Archer, J.S. 1963, Consistent Mass Matrix for Distributed Mass Systems, Jour. of Struct.
Division, ASCE, Vol. 89, No. STA4, pp. 161178.
7 Arya, S.C., ONeill, M.W. & Pincus, G. 1979, Design of Structures and Foundations for
Vibrating Machines, Gulf Publishing Co., Houston, Texas.
8 Arya, A.S. & Drewer, R. 1997, Mathematical modeling and computer simulation of
elevated foundations supporting vibrating Machinery, Tansaction of IMACS, Vol. XIX,
No. 4, Dec.
9 Bannerjee, P.K & Sen, R. 1987, Dynamic behavior of axially and laterally loaded piles
and pile groups, Chapter 3, Developments in Soil Mechanics and Foundation Engineering,
Vol. 3, ed., Bannerjee, P.K. and Buttereld, R., Elsivier Applied Science, London.
10 Baranov, V.A. 1967, On the calculation of excited vibrations of an embedded foundation,
Voprosy Dinamiki Prochnocti, No. 14, Polytech. Ins. Riga, pp. 195209.
11 Brooker E.W. & Ireland H.O. 1965, Earth pressure at rest related to Stress History,
Canadian Geotechnical Journal, Volume II, No. 1.
12 Barkan, D.D. 1962, Dynamics of Bases and Foundations, McGraw-Hill Book Co. NY.
13 Blake, T.F., Hollingsworth, R.A. & Stewart, J.P., eds. 2002, Recommended procedures
for implementation of DMG, Thomas F Blake (Fugro-West Inc., Ventura, Ca.): Special
Publication 117.
14 Bojtsov, G., Postnov, M., Paliy, V., Chuvikovsky, V.S. 1982, Dynamics and Stability of
construction, Vol. 3, Structural Mechanics of Ship Leningrad, Sudostronei (In Russian),
p. 317.
15 Broms, B.B. 1965, Design of Laterally Loaded Piles, J. of Soil Mechanics and Foundation
Engng. Div., ASCE, 91, No. SM3, pp. 7998.
16 Bycroft, G.N. 1956, Forced vibration of rigid circular plate on a semi innite half space
on an elastic stratum, Philosophical Transactions Royal Society of London, Series 248,
pp. 327368.
17 Chopra, A.K. & Chakrabarti, P. 1973, The Koyna Earthquake and damage to Koyna
Dam, Bull. Seis. Soc. America, 63, No. 2, pp. 38197.
606 References
18 Chowdhury, I. 1984, Dynamic soil-structure interaction of Turbo-Generator Foundation,
M.Tech Dissertation, Department of Civil Engineering, IIT Kharagpur.
19 Chowdhury, I. &Som, P.K. 1993, Dynamic Pile structure interaction of Boiler Feed Pump
Frame Foundation, Indian Geotechnical Conference, Vol. 1, pp. 411414.
20 Chowdhury, I., Ghosh, B. & Dasgupta, S.P. 2002, Analysis of Hammer foundation
considering soil damping, International Conference on Advancement in Civil Engineering,
Vol. 2, Indian Institute of Technology (Kharagpur) India, pp. 10191028.
21 Chowdhury, I. & Dasgupta, S.P. 2002, Earthquake Response of Soil-Structure System,
Indian Geotechnical Journal, 32(2), pp. 309328.
22 Chowdhury, I. & Dasgupta, S.P. 2004, Earthquake response of cantilever and counter-
fort retaining wall, Electronic Journal of Geotechnical Engineering, Vol. 8, [Online]
Available at: http:/www.ejge.com/
23 Chowdhury, I. & Dasgupta, S.P. 2006, Dynamic response of piles under vertical loads,
Indian Geotechnical Journal, Vol. 36, No(2), April.
24 Chowdhury, I. & Dasgupta, S.P. 2007, Dynamic Earth pressure on rigid unyielding walls
under Earthquake Force, Indian Geotechnical Journal, Vol. 37(2), pp. 8193.
25 Chowdhury, I. 2008, Some Myths and Facts about Dynamic Soil Structure Interaction and
its application to Building Design, keynote lecture, Conf. on Challenges and Application
of Mathematical Modeling in Building Science and Technology, Central Building Research
Institute of India, Roorkee, India.
26 Chowdhury, I &Dasgupta, S.P. 2008, Earthquake Analysis of Concrete Dams considering
uid structure and uid-structure soil interaction (communicated).
27 Chowdhury, I. & Dasgupta, S.P. 2008, Dynamic analysis of piles under lateral loads,
Indian Geotechnical Journal (In Press).
28 Cohen, M. & Jennings, P., Silent Boundary Methods For Transient Analysis, Compu-
tational Method in Transient AnalysisComputational Method in Mechanics, Vol. 1,
North Holland.
29 Coulomb, C.A. 1776, Essai sur une Applicaion des Regles de Maximis er Minims a
quelques Problemes de Satique, relatifsa lArchutecture-Mem. Roy des Sciences Paris,
Vol. 3, p. 38.
30 Craig R.H. 1981, Structural Dyanmics, John Wiley Publication, NY.
31 Crede, C.E. 1951, Vibration and Shock Isolation, John Wiley & Sons, NY.
32 Dasgupta, S.P. & Kameswara Rao, N.S.V.K. 1976, Some nite element solutions in the
dynamics of circular footings, Proc. 2nd International Conference on Numerical Methods
in Geomechanics, Blacksburg USA.
33 Dasgupta, S.P. & Kameswara Rao, N.S.V.K. 1978, Dynamics of rectangular footings by
Finite elements, Journal of GT Division, ASCE, Vol. 104(5).
34 Der Kiureghian, A. 1981, A Response Spectrum Method for Random Vibration Analysis
for MDF Systems, Earthquake Engineering and Structural Dynamics, 9, pp. 419435.
35 Design Criteria for Turbine Generator Pedestal, 1970, Journal of Power Division, ASCE,
Vol. 96, Jan, pp. 122.
36 Dobry, R. & Gazetas, G. 1988, Simple Method for dynamic stiffness and damping of
oating piles groups, Geotechnique, 38, No. 4, pp. 557574.
37 Dowrick, D.J. 2002, Earthquake Resistant Design of Structures a Manual for Architects
and Engineers, John Wiley & Sons, NY.
38 Dowrick, D.J. 2003, Earthquake Risk Management, John Wiley Publication, U.K.
39 Erden, S.M. & Stokoe, K.H. 1975, Effects of embedment on foundation response, Shock
and vibration problems in Geotechnical Engineering, ASCENational Convention, Denver,
Colorado.
40 Eurocode-8 1998, Design Provisions for Earthquake resistance of Structures Part 4- Silos,
tanks and pipelines. European Committee for Standardization Brussels.
References 607
41 Federation of Emergency Management Agency(FEMA)1997, Building Seismic Safety
Council, Washington D.C., USA.
42 Gazetas, G. 1982, Shear vibrations of vertically inhomogeneous earth dams, Int J. Num
and Analytical Methods in Geo-mechanics, Vol. 6, No. 1, pp. 219241.
43 Gazetas, G. &Tassoulas, A.L. 1987, Horizontal Stiffness of Arbitrarily shaped embedded
foundation, Journal of GT Division, ASCE, Vol. 113, No. 5.
44 Ghosh, D., Chakroborty, D. &Batavyal, H. 1984, Dynamic Response and Static Analysis
of RCC Frames supporting high speed centrifugal machines with Soil Structure Interac-
tion, International conference on case histories in geo-technical engineering, St Louis,
University of Missouri Rolla, USA.
45 Ghosh, D.K. & Batavayal, H.N. 1985, Analysis of Structural response to earthquake for
150Mhigh RCC Chimney with soil-structure interaction, Proc. National Seminar on Tall
Chimneys, Vigyan Bhavan, New Delhi.
46 Gupta, B.N. 1972, Effect of foundation embedment on the dynamic behavior of
foundation of the foundationsoil system, Geotechnique, 22(1), pp. 129137.
47 Hardin, B.O. &Richart, F.E., Jr. 1963, Elastic wave velocities in granular media, JSMFD,
ASCE, Vol. 89, No. SM1.
48 Hardin, B.O. 1965, The nature of damping in sands, J. Soil Mech. Found. Div., Am. Soc.
Civ. Eng. 91(SM-1), Part 1, pp. 6397.
49 Hardin, B.O. & Drnevich, V.P. 1973, Shear Modulus and damping in soils, Design
equations and curves, JSMFD, ASCE, Vol. 98, No. SM7, July.
50 Haroun, M.A. & Housner, G.W. 1981, Seismic design of liquid storage tank, Journal of
Technical Councils of ASCE, Vol. 107, No. TCI, pp. 191207.
51 Housner, G.W. 1957, Dynamic Pressure on Accelerated Fluid Containers, Bulletin of
Seismological Society of America, Vol. 47(1).
52 Housner, G.W. 1963, Dynamic Analysis of uids in containers subjected to accelera-
tion, Nuclear Reactors and Earthquakes Report No. TID 7024, U.S. Atomic Energy
Commission, Washington DC.
53 Hsieh, T.K. 1962, Foundation Vibrations, Proceedings of Inst. of Civil Engineers,
Vol. 22, pp. 211226.
54 Humar, J.L. 1990, Dynamics of Structure, Prentice-Hall of India, New Delhi.
55 Hurty, W.C. &Rubenstein, M.F. 1967, Dynamics of Structure, Prentice Hall of India Pvt.
Ltd., New Delhi.
56 Idriss, I.M. 1999, An update to the Seed-Idriss simplied procedure for evaluating lique-
faction potential, Proc. TRB Workshop on New Approaches to Liquefaction, January,
Publication No. FHWA-RD-99-165, Federal Highway Administration.
57 Indian Standards Institution 1984, 2002, Indian Standard Criteria for Earthquake
Resistant Design of Structures, IS: 1893 (Part 1), ISI, New Delhi, India.
58 Indian Standards Institution 1968, Indian Standard Code of Practice for Design and
Construction of Machine Foundations, Part I to Part IV, IS: 2974, ISI, New Delhi,
India.
59 Indian Standards Institution 1993, Indian Code of practice for ductile detailing of
Reinforced Concrete Structure Subjected forces, IS: 13920, ISI, New Delhi, India.
60 Ishibashi, I. & Zhang, K. 1993, Unied dynamic shear modulii and damping ratios of
sand and clay, Soils and Foundations, Vol. 33, No. 1, pp. 182191.
61 Jadi, H. 1999, Prediction of lateral dynamic response of Single Pile embedded in Clay,
Ph. D. Thesis, University of Missouri Rolla, Missouri.
62 Kardestuncer, H. ed., 1987, Finite Element Handbook, McGraw-Hill, NY.
63 Karnovsky, I. & Lebed, O. 2001, Formulas for Structural Dynamics, McGraw-Hill, NY.
64 Kaynia, A.M. & Kausel, E. 1982, Dynamic behavior of pile groups, 2nd Int. Conf. on
Numerical. Methods. in Offshore Piling, Austin, Texas.
608 References
65 Lasdon, L.S., Warren, A.D., Jain, A. & Ratner, M. 1978, Design and testing of
a generalized reduced gradient code for nonlinear programming, ACM Trans. Math.
Software, 4, pp. 3450.
66 Lysmer, J. & Richart, F.E. Jr. 1966, Dynamic response of footings to verti-
cal loading, J. Soil Mech. and Found. Div. Proc., ASCE, Vol. 92, No. SM1,
pp. 6591.
67 Jennings, P.C. & Bielek, J. 1973, Dynamics of building soil interaction, Bulletin of
Seismological society of America, 63,1,948.
68 Kameswar Rao, N.S.V. 1977, Dynamic soil structure systemA Brief Review, J. Struct.
Engg., India, Vol. 4.
69 Kameswara Rao, N.S.V. 1998, Vibration Analysis and Foundation Dynamics, Wheeler
Publishing, New Delhi.
70 Kasten, H.L. & Kirkland, W.D. 1970, Spring mounted Turbine Generator Spins Quitely,
Efciently, Electric Light and Power, E/D Edition, Nov., pp. 3840.
71 Kramer, S. 2004, Geotechnical Earthquake Engineering, Pearson Education Publication.
New Delhi, Section 7.5, Page 297.
72 Kreyszig, E. 2001, Advanced Engineering Mathematics, 8th Edition, John Wiley & Sons
Inc., NY.
73 Lysmer, J. & Kuhlemeyer, R.L. 1969, Finite Dynamic Model of Innite Media, J. EM.
Divn, ASCE, EM4.
74 Lysmer, J., Ostadan, F. & Chen, C. 1999, A System for Analysis of Soil Structure
Interaction, SASSI-2000, University of California, Berkeley.
75 Luco, J.E. 1986, Soil Structure interaction effects on seismic response of tall chimneys,
Soil Dynamics and Earthquake Engineering, Vol. 5(3), pp. 170177.
76 Major, A. 1980, Dynamics in Civil EngineeringAnalysis and Design, Vols. IIV,
Akademia Kiado, Budaapest and Collets Holding London.
77 Makadisi, F.I. & Seed, H.B. 1977, A simplied procedure for estimating earthquake
induced deformations in dam and embankments, EERC Report -77/19, University of
California Berkeley.
78 Mario, Paz. 1987, Structural Dynamics, CBS Publishers Ltd., New Delhi.
79 Meirovitch, L. 1967, Analytical Methods in Vibrations, Macmillan Co., NY.
80 Meirovitch, L. 1975, Elements of Vibration Analysis, McGraw-Hill Book Co. NY.
81 Mononobe, N. & Matsuo, H. 1926, On determination of earth pressure during
earthquakes, Proc. on World Engineering Congress, p. 9.
82 Mononobe, N. 1936, Seismic stability of earth dam, Proc. Second Congress on Large
Dams, Vol. IV.
83 Murray, D.A. 1967, Introductory Course on Differential Equations, Orient Longman
(India).
84 Murthy, V.N.S. 1991, Soil Mechanics and Foundation Engineering, Vols. I &II, Sai-Kripa
Technical Consultants, Bangalore(India).
85 Naeim, F. 1989, The Seismic Design Handbook, Van Nostrand Reinhold, NY.
86 Navarro, C. 1992, Inuence of soil exibility on seismic behavior of chimneys, Soil
Dynamics and Earthquake Engineering, Vol. 11, pp. 403409.
87 Newcomb, W.K. 1951, Principles of foundation design for engines and compressors,
Trans. ASME, 73, 307318.
88 Novak, M. & Beredugo, Y.O. 1972a, Vertical vibration of embedded footings, J. Soil
Mech. Found. Div., ASCE, 12, pp. 12911310.
89 Novak, M. & Beredugo, Y.O. 1972b, Coupled horizontal and rocking vibration of
embedded footings, Canadian Geotechnical Journal, 9, pp. 477497.
90 Novak, M. 1974, Dynamic stiffness and damping of piles, Canadian Geotechnical
Journal, Vol. 11, pp. 574598.
References 609
91 Novak, M. & El Sharnouby, B. 1983, Stiffness and damping constants for single piles,
J. Geotech. Engng. Div., ASCE, 109, pp. 961974.
92 Novak, M. & El Hifnawy, L. 1983, Vibration of Hammer Foundation, Journal. of Soil
Dynamics and Earthquake Engineering, 2(1), 4353.
93 Ohsaki, Y. & Iwasaki, R. 1973, On dynamic shear modulus and Poissons ratio of soil
deposits, Soils and Foundations, 13(4), pp. 6173.
94 Okabe, S. 1926, General theory of earth pressure, Journal of Japanese Society of Civil
Engineering, Vol. 12, No. 1.
95 Ostadan, F. 2004, Seismic Soil Pressure for Building Wallsan Updated Approach,
International Conference on Soil Dynamics and Earthquake Engineering (11th ICSDEE),
University of California, Berkeley, January.
96 Park, R. & Pauley, T. 1975, Reinforced Concrete Structures. 1st ed., John Wiley &
Sons, NY.
97 Paz, M. 1991, Structural Dynamics: Theory and Computation, Van Nostrand-
Reinhold, NY.
98 Peck, R.B., Hanson, W.E. & Thornburn, T.H. 1980, Foundation Engineering, 2nd edn,
Wiley Eastern Ltd., New Delhi.
99 Poulos, H.G. & Davis, E.H. 1980, Analysis and design of Pile foundations, John Wiley
Publication.
100 Randolph, M.F. & Poulos, H.G. 1982, Estimating Flexibility of Offshore Pile groups,
Proceedings of the 2nd International Conference on Numerical methods in Offshore Piling,
University of Texas, Austin, USA.
101 Rausch, E. 1959, Maschinen fundamente und andere dynamisch beanspruchte baukon-
structionen, VDI Verlag, Dusseldorf.
102 Richart, Jr. F.E., Hall, Jr. J.R. & Woods, R.D. 1970, Vibrations of Soils and Foundations,
Prentice-Hall, Englewood Cliffs, New Jersey.
103 Rosenblueth, E. 1951, A Basis for Aseismic Design, PhD thesis, Univ. of Illinois, Urbana.
104 Sadhegpour, N. & Chowdhury, I. 2008. Aerodynamic response of Tall Chimneys
considering foundation compliance- 12th International Conference in Advancement in
computational methods in Geo-mechanics-India.
105 Schmertmann, J.H. 1978, Use the SPT to measure Dynamic Soil Properties?_ Yes,
But . . . !, Dynamic Geotechnical Testing, ASTM, STP 654, ASTM, Philadelphia, Pa,
pp. 341355.
106 Seed, H.B. & Whitman, R.V. 1970, Design of Earth Retaining Structures under Dynamic
Loads, Proc. ASCE Specialty Conference on Lateral Stresses in Ground and Design of
Earth Retaining Structures, ASCE, pp. 103147.
107 Seed, H.B. & Idriss, I.M. 1970, Soil Moduli and Damping Factors for Dynamic Response
Analysis, Report No. 70-1, EERC, Berkeley, California.
108 Seed, H.B., Tokimatsu, K., Harder, L.F. &Chung, R.M. 1984, The Inuence of spt. proce-
dures in soil liquefaction resistance evaluations. Report No. UCB/EERC-84/15, University
of California at Berkeley, 1984.
109 Segol, G., Abel, J.F. & Lee, P.C.Y. 1975, Finite element Mesh Gradation of surface
waves, J. GT Division, ASCE, Vol. 101, GT 11.
110 Shames, I.H. & Dym, C.L. 1995, Energy and Finite Element Method in Structural
Mechanics, New Age International Publishers Ltd., New Delhi.
111 Shen, G.T. & Stone, N.E. 1975, Natural frequencies of turbine foundation, Structural
Design of Nuclear Plant Facilities, Vol. II, ASCE, NY, pp. 302330.
112 Srinivasulu, P. &Lakshmannan, N. 1978, Dynamic response of turbo-generator pedestal,
ASCE, Spring convention, Pittsburgh, Pensylvania, April, pp. 2428.
113 Srinivasalu, P. & Vadiyanathan, C.V. 1977, Handbook of Machine Foundations, Tata
Mcgraw-Hill, New Delhi.
610 References
114 Steedman, R.S. & Zeng, X. 1990, The Seismic Response of Waterfront Retaining Walls,
Proc Speciality Conference on Design and Performance of Earth Retaining Structures,
Speciality and Technical Publication 25, Cornell University, Ithaca, NY.
115 Steven, J. 1978, Prediction of pile response to vibratory loads, Proc. Xth Offshore
Technology Conference, Vol. IV, May, Houston, Texas.
116 Stevens, G.R. 1980, New Zealand Adrift, A.H. & A.W. Reed, Wellington.
117 Task Committee on Turbine Foundations 1987, Design of large steam turbine-generator
foundations, ASCE, NY.
118 Terzaghi, K. & Peck, R.B. 1967, Soil Mechnics in Engineering Practice, John Wiley &
Sons, NY.
119 Timoshenko, S. 1956, Strength of Material Vol. 2, Van Nostrand & Reinhold Publica-
tions, NY.
120 Timoshenko, S.P. & Woinowsky-Krieger (1958), Theory of Plates and Shells, McGraw-
Hill Kogakusha, Ltd., Tokyo.
121 Tschebotarioff, G.P. 1953, Performance records of engine foundation, ASTM Special
Technical Publication, No. 156.
122 Veletsos, A.S. & Meek, J.W. 1974, Dynamic behavior of building foundation system,
Journal of Earthquake Engineering and Structural Dynamics, Vol. 3(2), pp. 121138.
123 Veletsos, A.S. & Young, 1977, Earthquake response of liquid storage tank, Proceeding
of 2nd Engineering Mechanics specialty conference, ASCE, Raleigh, pp. 124.
124 Veletsos, A.S. 1984, Seismic response and design of liquid storage tanks, Guidelines for
seismic design of Oil and gas pipelines system, ASCE, NY, pp. 255370.
125 Verma, C.V.J. &Lal, P.K. ed., Treatise on the design, analysis and testing of High capacity
Turbo Generator foundation, Central Board of Irrigation and Power Publication #262.
126 Waas, G. &Hartmann, H.G. 1981, Piles subjected to dynamic horizontal loads-European
simulation meeting on modeling and simulation of large scale Structural Systems, Capri,
p. 17.
127 Weaver, Jr. W. &Gere, J.M. 1986, Matrix Analysis of Framed Structure, CBS Publishers &
Distributors, Delhi, India.
130 Wedpathak, A.V., Pandit, V.K. & Guha, S.K. 1977, Soil-Foundation interaction under
sinusoidal and impact type dynamic loads, Int. Symp. on Soil-Structure Interaction,
University of Roorkee, Roorkee.
131 Whitman, R.V. 1970, Soil Structure InteractionSeismic design for Nuclear power plants,
The MIT press, Cambridge, Massachusets.
132 Whitman, R.V. 1972, Analysis of Soil-Structure interactionA state of the art review,
Soil Publications No-300, Massachusetts Institute of Technology, April.
133 Whitman, R.V. 1973, Analysis of soil structure interactionA state of the art review,
Soil Publication, No 300, M.I.T.
134 Wilson, J.L. 2003, Aseismic design of tall chimneys, CICIND Research Report, Vol. 13.
135 Wolf, J.P. & VonArx, G.A. 1978, Impedance function of a group of vertical piles, Proc.
ASCE Speciality Conf. on Earthquake Engg. and Soil Dynamics, Pasadena, California,
pp. 10241041.
136 Wolf, J.P. 1985, Dynamic Soil Structure Interaction, Prenctice-Hall Englewood Cliffs, NJ.
137 Wolf, J.P. 1988, Dynamic Soil Structure Interaction in Time Domain, Prenctice-Hall,
Englewood Cliffs, NJ.
138 Wolf, J.P. 1994, Foundation Vibration Analysis: Using Simple Physical Model, Prentice-
Hall, Englewood-Cliffs, NJ.

Dynamics of Structures Volume 2 - Applications Chowdhury

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dynamics of Structures Volume 2 - Applications Chowdhury

Uploaded by

Copyright:

Available Formats

Dynamics of Structure and

Foundation A Unied Approach

, we build herein the basic concepts of dynamic soil structure

For theoretical background on these topics please consider Volume 1.

= mass and mass moment of inertia

= horizontal and rotational spring constant of the soil (IS-1893),

which can be further modified to

= translation and rotation spring stiffness of the total foundation

= damping ratio of the

] is non-proportional and not a diagonal matrix, and based

] becomes nearly a diagonal matrix

= 50000 kN/m for the soil-foundation.

17.104 = 4.135 rad/sec.

3.9666 = 1.9916 rad/sec;

very high the K

= 15% for the soil in rocking mode

= 5.5% for the soil in rocking

is given by (Terzaghi and Peck 1967).

= corrected SPT value for overburden, N

= corrected SPT value for

= gross overburden pressure in kN/m

= Corrected SPT value for overburden; N

= Corrected SPT value for

= Gross overburden pressure in kN/m

= 22 10.6 = 233.2 kN/m

12.22 = 14710.5 p.s.i. (101426 kN/m

12.22 = 11647 p.s.i. (80308 kN/m

(z) = Apsin pz +Bpcos pz = 0, at z = 0

which are otherwise known as coefficient of elastic uniform

This is in vertical Z direction

This induce sliding in horizontal

, This damping value is valid

Substituting the above values in the Lagrangian equation, we have

This is in vertical Z direction

This induce sliding in horizon-

as the case may be; G = Dynamic shear modulus of the soil;

27, a correction made to avoid approaching infinity as s tends to

= 9[12448 +79344 +33342.25 +21.55273] +14604

= 1/L which gives,

= foundation stiffness in horizontal direction; G

(z) is as expressed in Equation (2.5.201).

(z) in Equation 2.5.208, and carrying out

= 1/L which gives

=rotational stiffness of pile group; U =

1000679.61476.4 = 39974 kN sec/m

900611.631476.4 = 24446 kN sec/m

12511907.5218588.4 = 71375 kN sec/m

+ 1265 = 0.388sin 236t

+1265 = 0.388 sin 236t

where = the natural frequency of the structure.

from which we deduce,

and, G = dynamic shear

km = 0.1 31623 = 3162 kN sec/m

22560000224 = 68699 kN sec/m

= soil spring to be calculated from Barkan or Richarts formula in direction;

= damping of the soil to be obtained from Richarts formula in direction;

. Figure 2.11.3 shows that is nearly 180

for large values

, and the energy input is F

2. Isolation is possible only for

2 (refer to Figure 2.11.3). An un-damped spring is superior to a damped

r. Again, about two-third of

= is the farthest point from the center of gravity point G.