CHAPTER 2

LITERATURE REVIEW

2.1 GENERAL

Multistorey frames are simple discrete structures and their analysis has
been a part of structural engineering ever since. However classical and finite
element analysis of multistorey frames incorporating the modeling of soil and
its interaction with the structure have been in practice for only a few years.
The literature available in this field has been reviewed and presented here.

2.2 SOIL STRUCTURE INTERACTION

Lee and Harrison (1970) have conducted studies on structure —

foundation interaction and proposed two analytical methods for the analysis
of combined footing and two dimensional raft foundation. The methods take
into account the effect of the rigidity of the superstructure on the distribution
of forces and moments transmitted to the foundation. In the first method
rotations and sways at the column foundation junctions have been treated as
unknowns. The superstructure alone has been analysed and the displacements
and rotations obtained at each junctions have been equated to the
corresponding values obtained by considering the foundation to be a beam on
Winkler medium subjected to a system of forces and moments. The second
method involved the successive modification of an assumed contact pressure
distribution. The superstructure and foundation have been treated as a single
compatible unit and column forces and moments were evaluated from a
conventional structural analysis. In the next step, the foundation has been
isolated from the superstructure and structural analysis repeated with a new
estimate of contact pressure distribution. The procedure has been repeated till
acceptable accuracy has been achieved.
Lee and Brown (1972) have conducted a comparative study on a three
bay multistorey frame resting on raft foundation. Interaction studies have been

l0
done using two models (l) the conventional method using the Winkler
‘spring’ concept assuming the foundation to be rigid (2) the linear elastic
model. An advancement on the conventional method called the ‘soil - line —
method’ which considers the stiffness of the foundation relative to the soil in
addition to the Winkler model and loading for a rigid foundation, when
calculating moments, shearing forces and deflections has also been evaluated.
It has been observed that the difference in maximum bending moment in
foundation by the use of the Winkler model and linear elastic models have
been relatively small, and the maximum moment has decreased when the
flexibility of the foundation increased. It has also been observed that for
structures with more than three bays, the difference in maximum foundation
bending moment is large which necessitates a thorough interaction study.

Krauthammer and Chen (1988) have investigated the relationships

between the type and accuracy of the free field input motion generation and
the resulting effects on the corresponding structures. The general finite
element computer program ADINA has been utilized for obtaining the
numerical information. Three types of free field simulations have been
employed for the analysis of three typical soil-structure configurations namely
no embedment, partial embedment and full embedment. The results have
showed that simulation accuracy is a critical factor in such studies

Gazetas et al. (1993) have outlined a general methodology for a

complete seismic- soil — pile — foundation — structure interaction analysis. A
Beam — on - Dynamic —Winkler foundation simplified model and a Green’s
function based rigorous method have been utilized in determining the dynamic
response of single piles and pile groups. A systematic parametric study has
been conducted on the effect of pile group configuration upon dynamic
impedances of piles embedded in homogeneous as well as heterogeneous
soils. It had been shown that the cross interaction between piles in different
rows controls the dynamic response of a rectangular pile group and that
increasing the number of piles in a line group has very little effect on the
dynamic stiffness and damping factors. It has been demonstrated that the
predictions by the static interaction factor method are acceptable only for

ll
static and low frequency cases and that they may be very conservative or very
unsafe at higher frequencies. It had been concluded that at intennediate and
high frequencies it is better to ignore pile to pile interaction altogether than to
use static interaction factors.

Viladkar et al. (1994) have analysed a plane frame combined footing

soil system subjected to biaxial loading. An isoparametric interface / joint
element has been used to model the interface characteristics of the foundation
beam and the soil medium. It has been observed that interface elements are
essential for understanding the realistic nature of a laterally loaded structure
and that bending moments are not only relieved but also reversed due to the
interacting behavior of the framed structure footing soil system.

Yang et al. (1996) have demonstrated that the condensation technique

in structural mechanics can in reality be employed to fonnulate the soil­
structure interaction problems. The method has been used to calculate the
equivalent seismic forces exerted by the far-field soil on the near-field soil.

Wu (1997) has utilized a single — degree — of freedom replacement

oscillator to represent an SS1 system with SDOF structural model. A
methodology is then proposed to determine the equivalent fixed — base models
of general multi degree of freedom SSI system using simple system
identification techniques in the frequency domain.

Lu (2002) canied out comparative study on the non-linear behaviour of

reinforced concrete multistorey structures. Experimental and numerical
analysis on a scaled model has showed that the distribution of storey shear
over strength is an indicator of the general inelastic behavior of the frames.
Regular base frame, discontinuous-column frame, partially masonry infilled
frame and a wall-frame system have been used for the study.

Osinov (2003) has presented a mathematical model for the defonnation

of soil under irregular cyclic loading in the simple shear condition. The model
incorporated (l) the possible change in the effective pressure in saturated soil

l2
due to the cyclic shearing, (2) the reciprocal influence of the effective pressure
on the response of the soil to the shear loading and (3) the pore pressure
dissipation due to the seepage of the pore fluid

Kokusho (2003) has focused on the mechanism involved in the void re

distribution and water film effects in layered sand deposits to study the lateral
flow mechanisms during liquefaction. lt has been found that sand deposits in
the field consist of sub layers with different particle sizes and permeability;
and these readily developed water films by post liquefaction void
redistribution at sub layer boundaries.

Takewaki et. al. (2003) have presented a simple and fast evaluation
method of soil structure interaction effects of embedded structure via cone
model. The cone model has been used to evaluate the impedances and the
effective input motion at the bottom of an embedded foundation.

Takewaki(2005) has developed a new critical excitation method for

soil-structure interaction system. The input energy to the soil-structure
interaction system during an earthquake has been introduced as a new measure
of criticality. Two kinds of input energy have been defined, one to the overall
soil-structure interaction system and the other to the super-structure only. The
differences between these two energies indicate the energy dissipated in the
soil or that radiating into the ground.

Tokimatsu et al. (2005) have conducted studies on inertial and

kinematic forces on pile stresses based on large shaking table tests on pile —
structure models with foundation embedded in dry and liquefiable sand
deposits. The horizontal subgrade reaction acting on the pile and the earth
pressure acting on the embedded part of the foundation have been treated as
kinematic forces. An artificial ground motion called Rinkai, produced as an
earthquake in southern Kanto district in Japan, has been used as an input base
acceleration to the shaking table. The maximum values of bending moment
and displacement of the pile and soil has showed considerable variations for
dry and liquefiable sand deposits. The bending moments after liquefaction

13
have been considerably larger than that before. The results have showed that if
the natural period of superstructure is less than that of the ground, the
kinematic force tend to be in phase with the inertial force from the
superstructure, increasing the stresses in the piles and if the natural period of
the superstructure is greater than that of the ground, the kinematic force tend
to be out of phase with the inertial force, restraining the pile stresses from
increasing. Further in the study, a pseudo - static analysis based on Beam — on
Winkler — springs method has been conducted to examine its effectiveness in
estimating pile stresses in the large shaking table tests. It has been observed
that the estimated bending moment and deformation of the pile from the
pseudo — static analysis have been in good agreement with the observed values
and hence the method has been suggested for estimating pile stresses and
deformation mode with accuracy.

Yang et al. (2005) have observed that direct integration of the ground
acceleration data is firstly base line — corrected in the time domain using the
latest square curve fitting technique, and then processed in the frequency
domain using a windowed filter to remove the components that cause long
period oscillations in the desired displacement.

Krishna et.al (2006) assessed the liquefaction mitigation of ground

treated by granular piles. Pore pressure generation and dissipation accounting
for both densification and drainage effects of granular piles have been
considered. It has been observed that both the coefficients of volume change
and permeability are affected by densification.

Foundation impedance functions provide a simple means to account

for soil-structure interaction when studying seismic response of structures.
Impedence functions represent the dynamic stiffness of the soil media
surrounding the foundation. Impedence functions have been frequency
dependent and hence it is difficult to incorporate SS1 in standard time-history
analysis software. Safak(2006) has introduced a method to convert frequency
dependent impedance functions into time domain filters. The method has been
based on least-squares approximation of impedance functions by ratios of two

14
complex polynomials. Such ratios are equivalent, in the time domain, to
discrete- time recursive filters, which are simple finite difference equations
representing the relationship between foundation displacements and forces.

2.3 SEISMIC SOIL STRUCTURE INTERACTION

Hoshiya and Ishii (l983) have used a stochastic model to evaluate the
kinematic interaction of embedded rectangular foundations by the random
vibration theory. The formulation has been based on the fact that statistical
correlation of ground motions at different points decreases as the distance
between the points increases when components of high frequency are
contained in the ground motion. For the stochastic model, earthquake records
at a large scale inground tank and a foundation, made of cement — mixed soil
improved — ground has been used as examples of deep and shallow embedded
foundations. It has been observed that the foundation slab, which is relatively
stiff compared with the soil, constrain the ground motions and hence short
period components of the ground motion whose wave length has been less
than the dimension of the slab are weakened. Hence the kinematic interaction
effect of the slab has been like a low pass filter on the ground motions.

Neuss and Maison (1984) have presented a matrix formulation to

account for p — A effects in computer seismic analysis of multistorey
buildings. The method has employed a linear solution approach requiring no
iteration which can be used for static and dynamic — elastic analyser.

Veletsos and Prasad (1989) have made a study of soil structure

interaction for seismically excited simple structures considering both
kinematic and inertial interaction effects. The system investigated has been a
linear structure which was supported on a circular mat foundation at the
surface of a homogeneous elastic half space. The structures have been
presumed to have one lateral and one torsional degree of freedom in their
fixed base condition and have been excited by obliquely incident, horizontally
polarized, incoherent shear waves. The temporal variation of the free field
ground motion has been expressed stochastically by a local power spectral

15
density function(psd), and its special variability has been specified by a cross
psd function. The response quantities examined included the ensemble means
of the peak values of the lateral and torsional components of the foundation
input motion and the corresponding structural deformations. It has been
observed that like kinematic interaction, inertial interaction may affect
significantly the responses of systems in the medium and high frequency
spectral regions and that the effects of the latter are more important. It has also
been reported that unlike kinematic interaction, which generally reduces
lateral response, inertial interaction may increase the corresponding response
of tall, slender structures in the high frequency region of the response
spectmm. The inertial interaction effects for low frequency structures have
been negligible because such systems consider the half space as a very stiff,
effectively rigid medium. lt has also been observed that reliable estimates of
the effects of kinematic interaction on the peak values of structural response
may be obtained from the knowledge of the corresponding values of the
acceleration, velocity and displacement traces of the foundation input motion.
These quantities may be computed from analysis of the response of the
massless foundation to the free- field ground motion. Insofar as the mean
maximum values of the responses have been concerned, the kinematic
interaction effects due to ground motion incoherence are similar to those due
to wave passage and the two effects may be interrelated.

Guin and Banerjee (1998) have developed a methodology for the

dynamic analysis of soil- pile -— structure system using a generalized coupled
finite element boundary element formulation for the entire problem domain.
The formulation has been done in the frequency domain and the excitation is
defined through a rock outcrop motion causing vertically propagating S­
waves. Linear dynamic analysis has been conducted on two super structural
systems namely, a bridge and a multistory frame. It has been observed that
coupling of the problems facilitates in the preparations of transfer functions
for various degrees of freedom in the structure, including the effects of
interaction.

16
 Wolf and Song (2002) have formulated a criterion for the presence of
radiation damping in a site. The procedures for the analysis of dynamic soil
structure interaction have been out lined. The procedures have included simple
physical model (cones, spring — dashpot — mass representations) for the soil,
the damping — solvent extraction method, the rigorous forecasting method and
the scaled boundary finite element method.

Spyrakos and Xu (2003) have studied the seismic response of massive

flexible strip foundation embedded in layered soils subjected to seismic
excitation. The foundation has been treated with a finite element formulation,
while the difficulty in modeling the infinite extend of the soil has been
overcome by a boundary element formulation. System responses have been
investigated with the help of boundary element — finite element coupled
formulations by enforcing compatibility and equilibrium conditions at the soil
foundation interface.

Davenne et al. (2003) have developed numerical tools for the modeling
of reinforced structures for the non linear transient analysis of RC structures.
A multifiber beam element has been used to describe the response of structural
components and a macro element to account for soil structure interaction. The
method has been applied for various boundary conditions and incorporating
soil-structure interaction.

Spyrakos and Loannidis (2003) have conducted studies on the effect of

soil structure interaction on seismic analysis and design of bridges. The
significance of soil structure interaction on a model with geometric stiffness
and seismic response of a bridge with integral abutments has been established.
It has been reported that the role of soil structure interaction is of great
importance for the post tensioned modular integral bridge system.

Gen-shu and Jin-qiao (2005) have examined the seismic force and
modification factor R based on elastic-plastic time -— history earthquake
analysis of single degree of freedom systems. The constitutive hysteresis
models that have been used are elastic- perfectly -—plastic, elastic —linearly­

17
hardening and bilinear-elastic. It has been concluded that R increases linearly
with ductility and energy dissipating capacity in short period ranges.

Anandarajah et al.(2005) have demonstrated that soil parameters

needed for simplified dynamic analysis of a single pile may be back calculated
from the dynamic response of the pile measured in the field. Two methods
have been proposed, the first based on Winkler foundation approach and the
second based on the equivalent- linear finite element approach, with non
linearity of shear modulus and damping accounted for by employing
degradation relationships.

Wegner et al. (2005) have developed a numerical procedure for three­

dimensional dynamic-soil—structure interaction analysis. Scaled boundary
finite-element method has been used for modeling the unbounded soil and
standard finite element method is used for modeling the structure. The
dynamic response of tall buildings, with multi-level basements, subjected to
seismic excitations including P, SV and SH waves at various angle of
incidence have been arrived at.

Takewaki and Kishida (2005) have developed a method for the

analysis of pile-group effects on the seismic stiffness and strength of buildings
with pile foundations. A continuous model consisting of a dynamic Winkler­
type soil element and a set of pile has been used to express the dynamic
behavior of the structure-pile-soil system. The pile group effect has been
accounted through the influence of coefficients that have been defined for
interstorey drifts and pile-head bending moments. Pile group effect has
reduced the interstorey drift of buildings and increased the bending moments
at the pile head.

2.4 RESPONSE SPECTRUM ANALYSIS

Newmark (1973) has shown that the response of simple systems to

ground motion can be represented by idealizing the linear response spectrum
into constant acceleration, constant velocity and constant displacement

l8
response for a given damping factor. Browning (2001) has presented a simple
method for proportioning of regular, moderate-rise reinforced concrete
building structures. In this method the member sizes have been selected based
on the demand defined by the displacement spectrum and criteria specified in
relation to different responses.

Ghiocel and Ghanem (2002) have conducted studies on probabilistic

analysis of the seismic soil structure interaction problem. The procedure has
accounted for the uncertainties in the free field input motion, local site
conditions and structural parameters. The uncertain parameters have been
modeled using a probabilistic frame work as stochastic processes. The
earthquake ground acceleration has been represented by a probabilistic
acceleration response spectrum. The procedure has been then applied to the
seismic analysis of a nuclear reactor facility and has been observed to have
good co-relation with other deterministic methods of risk assessment of
hazardous facilities under dynamic loads.

Ambraseys and Douglas (2003) have presented strong motion

attenuation relationships for peak ground acceleration, spectral acceleration,
energy density, maximum absolute input energy for horizontal and vertical
direction and for the ratio of vertical to horizontal of these ground motion
parameters. The equations have been derived using a world wide data set of
186 strong — motion records that have been recorded with in l5 km of the
surface projection of earthquakes with magnitudes 5.8 to 7.8.

Yuan et.al. (2003) have studied the effect of asymmetry and

irregularity of the input seismic waves on the earthquake — induced differential
settlement of the buildings on natural subsoil. It has been concluded that these
are necessary factors that has to be considered in the evaluation for differential
settlement and other problems that have been related to the soil deformation
due to earthquakes.

Chintanapakdee and Chopra (2004) have conducted non-linear

response history analysis on vertically regular and irregular frames to study

l9
the storey drifts and floor displacements. Drift demands in the upper storeys
have been more sensitive to irregularities in the lower storeys than the
response of lower storeys due to irregularities in the upper storeys. Irregularity
in the base storey or lower storeys has significant influence on the height-wise
distribution of floor displacements.

Boore and Bommer (2005) have suggested methods for processing of

strong motion accelerograms which have been masked and distorted by noises.
It has been important to identify the presence of noise in the digitized time­
history and its influence on the parameters for any application of recorded
accelerograms in earthquake engineering.

Camiel et al. (2006) have investigated the application of the Singular

Spectrum Analysis (SSA) to improve the Nakamura technique. Nakamura
technique has been employed to estimate the dynamic characteristics of
surface layers by measuring solely the tremor at the surface. The SSA has
allowed the time series to be decomposed into different components, like
signal itself, various noise components etc.

2.5 COMMENTS

The incorporation of soil structure interaction effects in the

analysis and subsequent design of multistorey frames has been the subject
of research and, started getting attention in the late seventies. It is
observed that research attention has been focused on the soil structure
interaction effects of high risk structures like nuclear reactors, bridges
etc. However only limited research studies have been carried out on soil
structure interaction effects of multistorey building frames, which are
common, popular and simple in configuration. Soil structure interaction
studies of multistorey frames which generally are provided with pile
foundations, is tedious, voluminous and cumbersome due to the possible
involvement of large number of nodes and finite elements necessary to
model the soil mass. This may be the reason why such analysis are not
addressed and reported in considerable numbers. High rise buildings on

20
pile foundations have become part of regular building construction due to
population explosion. From the review of the reported studies it has been
felt that a complete soil structure interaction analysis of multistorey
building frames with pile foundations using appropriate numerical
methods is justifiable.

21