INTERACTIVE ANALYSIS MODELS

FOR
SOIL AND STRUCTURES

BY
Prof. P. C.Vasani
Applied Mechanics Department
L. D. College of Engineering
Ahmedabad- 380015
Ph.(079) 7486320 [R]
E-mail:pcv-im@eth.net

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 INTERACTIVE ANALYSIS MODELS FOR SOIL AND
STRUCTURES

1. INTRODUCTION.
1.1 GENERAL.
Soil - structure interaction plays an important role in the behavior of foundations.
For structures like beams, piles, mat foundations and box cells it is very essential to
consider the deformational characteristics of soil and flexural properties of foundations.
It can be seen that when interaction is taken into account, the true design values
arrived-at may be quite different from those worked out without considering interaction.
In general in most of the cases interaction causes reduction in critical design values of the
shear and moments etc.. However, there may be quite a few locations where the values
show an increase . Because of these possibilities have their own roles to play in economy
and safety of structure.
Several studies have indicated that the maximum bending moment in a
foundation raft or beam could be substantially affected by interaction with superstructure.
Reduction as high as 80% is reported in certain cases.
The rigidity of foundation raft relative to soil is of extremely high values of
bending moments in relative rigid rafts as compared to those in flexible rafts. An elasto-
plastic analysis also indicates a similar trend, although to a much lesser degree.
Unequal settlement is the severest cause for cracking and even failure of
superstructures. On the other hand, rigidity of superstructure helps in reducing differential
settlements. of course, to realize this, only interactive analysis has to be carried out.

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 2. COMPONENTS OF INTERACTIVE ANALYSIS AND
METHODS APPLIED TO INTERACTION.
2.1 Soil Foundation Interaction Problems.

2.1.1 Introduction :
The study of the interaction between foundation and supporting soil media is of
fundamental importance to both geotechnical and Structural engineers. Results of such
study can be used in the structural design of the foundation and in the analysis of stresses
and deformations within the supporting soil medium.
The phrase 'soil-foundation interaction' may be defined as " the influence of the
behavior of the soil immediately beneath and around the foundation on the response of
the sub-structure subjected to either static or dynamic loads".
The foundation designer must consider the behavior of both structure and soil and
their interaction with each other. the interaction problem is of importance to many civil
engineering situations and it covers a wide spectrum of problems. These include the study
of shallow and deep foundations, floating structures, retaining wall-soil systems, tunnel
lining, buried structures, earth structures etc..

2.1.2 Behavior of soil :

In-situ soils are commonly anisotropic and non-homogeneous and display
markedly non-linear, irreversible and time dependant characteristics. The behavior of
such soils is expected to be influenced by following factors (25).

(i)The shape, size and mechanical properties of the individual soil particles.
(ii)The configuration of the soil structure.
(iii)The intergranular stresses and stress history.
(iv)The presence of soil moisture, the degree of
saturation and the soil permeability.

The solution of any interaction problem on the basis of all above factors is very
difficult, laborious and impracticable. Realistic and purposeful solutions can achieved by
idealizing the behavior of the soil by considering specific aspects of its behavior. The
most simple idealization of response naturally occurring soils assumes linear elastic
behavior of the supporting soil medium. This idealization also assumes the surface of the
soil medium to form the soil-foundation interface and the soil medium is represented by
an elastic medium occupying a half-space region. Though these assumptions are not
always rigorously satisfied by in-situ soils, these considerably simplifying the solution

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 and provide useful information to number of practicable problems in geotechnical
engineering.
Various idealized soil behavior models will be introduced afterwards.

2.1.3 Behavior of foundation :

The flexural behavior of foundations can be adequately described by modeling
appropriately foundations as beam and plates, and using the convenient respective
theories of beams and plates. The theories may be modified by incorporating the effects
of shearing deformations or by taking in to account the three-dimensional state of stress.
It may be noted that the time dependant behavior of the foundation itself can have a
significant influence on the modeling and the end results of a soil-foundation interaction
analysis.

2.1.4 Behavior of Interface :

The complete solution of the interaction problem necessitates prior assignment of
a particular type of mechanical behavior to the soil-foundation interface. The interface
conditions associated with the elastic continuum behavior of the soil medium are assumed
to range from the completely smooth to the completely frictional interfaces. The factors
which are expected to significantly influence the conditions at the soil foundation
interface are :

(i) The presence of pore water which can alter the

magnitude and distribution of the frictional
forces throughout the consolidation process.
(ii) The distribution and character of the external
loads on the foundation.
(iii) The relative flexibility of the foundation, and
(iv) The time dependant effects.

Frictional effects at the interface are expected to acquire importance when dealing
with the interaction of highly flexible foundations resting on compressible soil media. It
is suggested that proper interface conditions be formulated only after obtaining adequate
data from field observations. In the absence of such data interface may be assumed as
smooth and such assumption can serve as a usual first approximation. The assumption of
the smooth contact considerably simplifies the analysis of the interaction by retaining
only the normal component of the contact stress (25).

2.2 Idealized Soil Behavior Models.

2.2.1 General :
The natural complexity in the behavior of in-situ soils has led to the
development of many idealized models of soil behavior based on the classical theories of
elasticity and plasticity for the analysis of Soil-Foundation interaction problems.

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 Although, the generalized stress-strain relations for soils, don’t represent even the gross
physical properties of a soil mass, the idealized models are observed to provide a useful
description of certain features of soil media under limited boundary conditions. The
idealized soil behavior particularly reduces the analytical rigor spent in the solution of
complex problems in geotechnical engineering. The idealization will depend on a variety
of factors such as:
(i) The type of soil.
(ii) The soil conditions,
(iii) The type of foundation,
(iv) The nature of external loading,
(v) The method of construction,
(vi) The purpose and life span of the structure and
(vii) The economical considerations.

Some important idealized models of soil-foundation interaction are briefly

presented in following articles.
The character of each model is typified by the surface deflection it experiences
under the action of a system of forces, and these surface deflection in-general represent
the displacement characteristics of the soil-foundation interface, and forms a significant
part of the soil foundation interaction analysis.

2.2.2 Elastic Models:

Models of soil behavior which exhibit purely elastic characteristic are considered
here. The simplest type of idealized soil response is to assume the behavior of supporting
soil medium as a linear elastic continuum. The deformations are thus assumed as linear
and reversible. Applications of these models to soil-foundation interaction have been
subject of extensive research and significant developments have been made in obtaining
exact and approximate solutions.

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 (1) The Winkler's Model:
Winkler assumed that the surface displacement of the soil medium at every point
is directly proportional to the stress applied to it at that point and completely independent
of stresses or displacements at other, even immediately neighboring, point of the soil-

X P X

(a) Z (b) Z

P q
X X

(c) (d)
foundation interface.

Fig.2.1. Surface displacements of the Winkler model due to (a) Non-uniform

load, (b) A concentrated load, (c) A rigid load, (d) A uniform flexible load.
Winkler's idealization of the soil medium can be physically represent as a system
of closely spaced spring elements (See Fig 2.1) each of all will be deformed by the stress
applied directly to it while the neighboring elements remains unaffected. The
characteristics features of this representation of the soil medium is discontinuous behavior
of the surface displacement. The deflection, y, of the soil medium at any point on the
surface is directly proportional to the stress, q, applied at that point and is independent of
stresses applied at other locations, i.e.,

q (x,y) = k y(x,y) … … … … .Eq (2.1)

Where 'k' is termed as the modulus of subgrade reaction with units of stress per
unit length. Equation (2.1) is usually the response function or the kernel function for the
Winkler's model. Physically, Winkler's idealization of the soil medium consists of a
system of mutually independent spring elements with spring constant 'k'.
One important feature of this soil model is that the displacements occurs
immediately under the loaded area and outside the region the displacements are zero. See

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 Fig. 2.1. Also, the displacements of a loaded region will be constant whether the soil is
subjected to an infinitely rigid load or a uniform flexible load.
There are many engineering problems (excluding soil-foundation interaction) for
which Winkler's Model represents a very accurate idealization of actual operating
conditions.

(2) Elastic Half-Space (Elastic continuum) Models:

In the case of in-situ soils surface deflections will occur certainly under and
around the loaded region. Applicability of Winkler Model is only limited to such soil
media, which possesses cohesion or transmissibility of applied forces. In attempts to
account for this continuous behavior, soil media have often been idealized as three-
dimensional continuos elastic solids or elastic continua.
The distribution of displacements and stresses in such media remain continuous
under the action of external force systems. Generally, the application of the continuum
theory of classical elasticity to soil-foundation interaction presents a complex
mathematical problem. A number of solutions to various boundary value problem of soil
foundation have been attempted by various investigators.
The isotropic elastic continuum model can be effectively employed in the
analytical treatment of soil-foundation interaction problems.

P P'
X r

u u

w w

(a) (c)
Z Z
2b
2a
q q'
X r

(b) (d)
Z Z

Fig. 2.2. Typical surface displacement profiles of an elastic half-plane subjected to (a) A line
load P, (b) A uniform load q of width 2b, (c) A concentrated load P’ , (d) A uniform load q’ of radius ‘a’ .
The response function for the linear elastic half space-space/plane is characterized
by the deflected shape of the surface of the elastic medium subjected to a concentrated

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 force or a uniform stress of finite extent. In Fig. 2.2 a state of plane strain(two
dimensional problem) exists in the X, Z plane. For such a state of strain, the displacement
component v in the Y direction is zero and the displacements u and w in the X and Z
directions, respectively, are purely functions of these co-ordinates.
In case of axisymmetric problems are discussed the class of three-dimensional
deformations which are symmetrical about the Z- axis of the cylindrical polar co-ordinate
systems r,θ,z. See Fig. 2.2 .
The anisotropic continuum model and the non-homogenous continuum models
are of particular interest to interaction problems. Generally 21 independent elastic
constants are necessary to characterize the elastic behavior of an anisotropic material. A
homogenous anisotropic body with three mutually perpendicular planes of elastic
symmetry passing through every point is called an orthotropic material. Nine independent
elastic constants are necessary to define such a material. A transversely isotropic elastic
material is a special case of the generally orthotropic elastic material, where there exists a
plane through every point in the body for which all directions are equivalent with respect
to the elastic properties, and elastic constants for a particular case reduces to six.
Several solutions for surface loading conditions associated with multi-layered
elastic solids of infinite lateral extent are available in literature as follows:
(i) A single elastic layer underlain by a rigid stratum
(ii) An elastic layer underlain by an elastic half-space
(iii) Multilayered elastic systems.

(3) Two Parameter Elastic Models:

The deficiency of the Winkler's Model in describing the continuous behavior of
real soil masses and the mathematical complexities of the elastic continuum has lead to
the development of many other simple soil behavior models. These models posses some
of the characteristics features of continuous elastic solids. The term "Two Parameter"
signifies that the model is defined by two independent elastic constant. The development
of these models has been approached along following different lines.
(i) The First type stems from the discontinuous Winkler's model and eliminates its
discontinuous behavior by providing mechanical interaction between the individual
spring elements by either elastic membranes, elastic beams or elastic layers capable
of purely shearing deformations (i.e. Filonenko-Borodich, Hetenyi, Pasternak and
kerr).
(ii) The Second approach proceeds from the elastic continuum model and introduces
constraints or simplifying assumptions with respect to the distribution of
displacements and stresses (Reissner, Vlazov and Leontiev).

The following discussion is related to salient features of some typical two

parameter models.

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 (a) Filonenko - Borodich Model:
This model requires continuity between the individual spring elements in the
Winkler's model by connecting them to a thin elastic membranes under a constant tension
T.

P
T Elastic Membrane T X T T X

(a) Z (b) Z

T P T T q T
X X

(c) Z (d) Z

Fig. 2.3. Surface displacements of the Filonenko-Borodich model.(a) Basic

model, (b) Concentrated load, (c) Rigid load, (d) Uniform flexible load.
Typical examples of surface displacement profiles of this model due to
concentrated, flexible and rigid external loads are shown in Fig.2.3. The consideration of
the equilibrium of the membrane-spring systems, show that for three dimensional
problems (e.g. rectangular or circular foundations) the surface deflection of the medium
due to a pressure q is given by,
q (x,y) = k w(x,y) - T ∇2 w(x,y)… … Eq (2.2).

Where ∇2 = δ2 + δ2
δx2 δy2

is Laplace's differential operator in rectangular Cartesian co-ordinates.

F For two-dimensional problems (strip foundation) above equation reduces to

q (x) = k w(x) - T d2 w(x) … … … … … .Eq (2.3).

dx2

The two elastic constants necessary to characterize the soil model are k and T.

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 (b) Hetenyi Model:
The interaction between the independent spring elements is accomplished by
incorporating an elastic plate in three-dimensional problems, or an elastic beam in two-
dimensional problems. The response function for this model is given by,

q(x,y) = k w(x,y) - D ∇4 w(x,y) … … … ...Eq (2.4).

Where D ( = Ep h3/12(1-ν p2) ) is the flexural rigidity of the plate.

F For two-dimensional problems above equation becomes,

q (x) = k w(x) - D d4 w(x) … … … … … .Eq (2.5).

dx4

(c) Pasternak Model :

The model proposed by Pasternak assumes the existence of shear interaction
between the spring elements. This condition is accomplished by connecting the spring
elements to a layer of incompressible vertical elements, which deform in transverse shear
only. The deformations and forces maintaining equilibrium in the shear layer are shown
in Fig. 2.4.

The response function for this model is,

q(x,y) = k w(x,y) - G ∇2 w(x,y) … … … … Eq (2.6).

It can be seen that this is similar to Filonenko-Borodich equation if T is replaced

by G. The surface deflection profiles for this model are very similar to those obtained for
the Filonenko-Borodich model. With the two-parameter models described so far, the
Winkler case can be record as a limiting case, as T, D and G tend to zero.

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 ShearLayer x x+dx x
q(x,y)
X
w
w + ∂w dx
∂x
τ xz
τxz + ∂τxz dx
∂z

(a) Z (b)
z

q
x Ny
y Nx
z

(c)
Nx + ∂Nx dx
Ny + ∂Ny dy r0 ∂x
∂y

Fig.2.4. The Pasternak model: (a) Basic Model, (b)Stresses in the shear layer, (c) Forces
acting on the shear layer.

(d) Vlazov and Leontiev model:

The model of soil response proposed by Vlazov presents an example of the
second type of two-parameter elastic model, which is derived by introducing
displacement constraints that simplify the basic equations for a linear elastic isotropic
continuum. Vlazov's approach to the formulation of this soil model is based on the
application of a variational method. By imposing certain restrictions upon the possible
distribution of displacement in an elastic layer, he was able to obtain a soil response
function similar in character to Eq (2.5) and Eq (2.6). The details of the general
variational methods of analysis together with solution to many practical problems are
given in Vlazov and Leontiev (1966).

(e) Reissner model:

The Reissener's model is derived by introducing displacement and stress
constraints that simplifying the basic equations for a linear elastic isotropic continuum.
The response function for the soil model is given by,

C1 w - C2 ∇2 w = q - C2 ∇2 q … … Eq (2.7).
4C1

Where H is the thickness of soil layer,

w is the vertical displacement of the surface of the elastic layer,

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 z = 0, and q is the external load.
The constants C1 and C2 characterizing the soil response (Eq 2.7) are related to
E and n by
C1 = E/H and C2 = HG/3,
where E and G are the elastic modulus and shear modulus, respectively of the soil
layer. Also, it may be noted that for a constant or linearly varying stress, after
redefining C1 = k and C2 = G, Eq (2.7) becomes identical to Eq (2.5) and Eq
(2.6).

2.2.3 Elastic-Plastic, Perfectly Plastic Models :

The Elastic soil models do not take into account any elastic-plastic or irreversible
behavior of the soil medium. The basic distinction between the purely elastic and elastic-
plastic models is that, in the latter case, the stresses that can be induced in the soil
medium are limited owing to the introduction of a yield or failure criterion. An example
of a purely mechanical type is that foundation model proposed by Rhines. This model
assumes that the shear layer interconnecting the spring elements of the Pasternak
foundation model is capable of sustaining finite shearing stresses. The shear stress-shear
strain relationship for the elastic layer is of an elastic-rigid plastic type. Using this
particular model, Rhines has investigated the distribution of contact stresses beneath a
rigid foundation, which is subjected to a symmetric load. The results indicate that the
inclusion of such yielding characteristics can alter both the magnitude and distribution of
contact stresses that are developed at soil-foundation interface.
In the second type of analysis, the elastic continuum model can be modified to
account for effects of plastic flow. If the soil medium possesses definite yield
characteristics of an ideal elasto-plastic material, failure will occurs at the zones of high
stress concentration. The plastic regions that are developed in the medium will satisfy
stress-strain relations generally different to those applicable to the elastic regions. Such
methods of stress analysis have received considerable attention owing to the development
of numerical methods.
Chen and Snitbhan (1977) discussed two elastic-plastic soil models : the perfectly
plastic Drucker-Prager model and the isotropic strain hardening model of Roecoe and
Burland. Explicit incremental constitutive expressions, suitable for numerical analysis,
have been presented for both models. It is stated that any numerical analysis using these
models can be obtained within the framework of the Finite-Element method and a step-
by-step integration procedure. Both models are viewed by the authors as useful
computational tools.

2.2.4 Time-Dependent Behavior.

When soil media exhibit consolidation or creep effects, it becomes necessary to
consider the time-dependent behavior of the soil-foundation interaction problem. The
stresses and strains in cohesionless soil masses are not likely to change substantially with
time except for the occurrence of plastic flow or a nominal amount of visco-plastic flow.
With predominantly cohesive, saturated soil masses, however, the stresses and

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com

 deformations could vary significantly with time. The representation of the deformation of
a saturated soil medium under load, in terms of consolidation due to the dissipation of
pore water pressures, was first formulated by Terzaghi. This theory was later generalized
by Biot to account for three-dimensional effect. The theory of primary consolidation
predicts that, under the action of external loads which remains constant with time, the
stresses and deformations in a soil medium reach a stable value. The analysis of soil-
foundation interaction, which is based on the theory of consolidation, is a mathematically
complex but straightforward problem. The deflections and stresses in structures resting on
such consolidating media could exhibit marked variations with time. The time dependent
deformations of soil masses, which occur beyond those, predicted by the theory of
primary consolidation. Secondary consolidation is assumed to be due to creep processes
in the structure of the soil media. Certain unsaturated soil and soft rocks may also exhibit
creep effects in the absence of a primary consolidation. The theory of linear visco-
elasticity can be adopted for the analysis of such creep effects. The analysis of soil-
foundation interaction can be performed by employing an elementary rheological model
approach (Maxwell model, Kelvin model or their combinations) or a viscoelastic
continuum approach (where the relationship between stress, strain and their time rates of
both dilatational and distortional behavior are represented by relationships of an integral
or a differential type). Such viscoelastic analyses have been applied by many authors in
the treatment of soil-foundation interaction.

PDF created with FinePrint pdfFactory trial version http://www.fineprint.com