Research Journal of Applied Sciences, Engineering and Technology 4(7): 717-723, 2012

ISSN: 2040-7467
© Maxwell Scientific Organization, 2012
Submitted: September 25, 2011 Accepted: November 04, 2011 Published: April 01, 2012

Using Matlab with Quadrilateral Finite Elements in Analysis of Multilayered

Nonhomogeneous Soils under Strip Footing

Ahmed Qasim Obaid Al-Ramthan

Department of Engineering Affairs, Baghdad University, Iraq

Abstract: Natural soils are often comprised of separate layers. Foundations of engineering structures are
designed to transfer and distribute their loading to the underlying soil and/or rock. Therefore, the designer must
ensure that the structure does not suffer from excessive displacements. In this research, it is intended to make
a generalized numerical solution through a computer program has written in MATLAB to analyze a
multilayered nonhomogeneous (Gibson-type) soil, where the stiffness contrast exists between layers, by
determining the displacements and stresses under strip footing during applied incremental loading sequence.
This research presents a plane strain two-dimensional finite element method. In this method, the soil is divided
into several 4-node quadrilateral elements. The general behavior of a multilayered soil profile and the influences
of strip foundations resting on soils having homogeneous (constant modulus with depth) to Gibson-type
(linearly increasing modulus) profiles are studied. The influences of foundation size and embedment,
incremental loading, and soil stratifications are considered in this study.

Key words: Finite element method, MATLAB, multilayered soils, nonhomogeneous soils, strip footing

INTRODUCTION Desai (1968) considered the load deformation

problem of a circular footing resting on homogeneous and
In nature, soil deposits are often formed in discrete two layered clay systems. Employing a plane strain
layers. As a result, footings are usually supported by approach, Radhakrishnan (1969) investigated the behavior
multilayered soil profiles, which influence the value of the of a continuous strip footing on homogeneous and layered
settlements of the foundation. If a footing is placed on the clays.
surface of a layered soil and the thickness of the top layer Rowe and Booker (1982) presented a finite layer
is large compared with the width of the footing, then the approach for the analysis of both layered and
displacement behavior of the footing can be estimated, to continuously varying elastic soil. The approach rested on
sufficient accuracy, using only the properties of the upper the assumption that the soil deposit may be regarded as
layer (Poulos et al., 2001). consisting of a number of distinct horizontal layers.
It is commonly believed that the settlement criterion Douglas (1986) reported the existence of more than
is more critical than the bearing capacity one in the 40 different methods for estimating settlements in
designs of shallow foundations, especially for foundation granular soils. All these methods recognize that the
width greater than 1.5 m, which is often the case. By applied pressure, soil stiffness and the foundation width
limiting the total settlements, differential settlements and are the three most important variables affecting the
any subsequent distresses to the structure are limited. settlements in granular soils.
Generally, the settlements of shallow foundations such as Since many soils exhibit stiffness increasing with
pad or strip footings are limited to 25 mm (Terzaghi et al., depth because of the increase in overburden stress, the
1996). displacements and stresses will be evaluated for a Gibson
In general, the magnitude and distribution of the type soil. A footing resting on a nonhomogeneous elastic
displacements and stresses in soil are predicted by using medium with modulus increasing with depth is a more
solutions that model soil as a linearly elastic, generalized problem (Stark and Booker, 1997).
homogeneous and isotropic continuum. From the The solutions of displacements and stresses for
standpoint of practical considerations in engineering, various types of applied loads to homogeneous and
anisotropic soils are often modeled as orthotropic or nonhomogeneous isotropic/anisotropic full/half-spaces
isotropic medium. Besides, the effects of deposition, have played an important role in the design of
overburden, desiccation, etc., can lead geotechnical foundations. However, it is well known that a strip load
media, which exhibit both nonhomogeneity and solution is the basis of complex loading problems for all
anisotropic deformability characteristics. constituted materials. A large body of the literature was

717
 Res. J. App. Sci. Eng. Technol., 4(7): 717-723, 2012

Fig. 1: Variation of soil modulus with depth

devoted to the calculation of displacements and stresses METHODOLOGY

in isotropic media with the Young’s or shear modulus
varying with depth according to the linear law, the power Generalized “Gibson-Type” profile: The elastic
law, and the exponential law, etc. (Wang et al., 2003). modulus for a generalized Gibson-type soil (Gibson,
For the purpose of analysis of natural soils, it may be 1967) is expressed by:
convenient and reasonable to assume that the soil within
each layer is homogeneous. If a footing is placed on the E s = E o + kz (1)
surface of a layered soil and the thickness of the top layer
is large compared with the width of the footing, the
where Es = the elastic soil modulus increasing linearly
displacement behavior of the footing can be estimated to
with depth; Eo = Young’s modulus of elasticity of soil
sufficient accuracy using the properties of the upper layer
directly underneath the foundation base; k = linear rate of
only. However, if the thickness of the top layer is
comparable to the footing width, this approach introduces increase of elastic modulus with depth (units of E per unit
significant inaccuracies and is no longer appropriate. This depth); and z = depth.
is because the zone of influence of the footing, including Two layers of Gibson-type soils are considered in this
the potential failure zone, may extend to a significant study, as shown in Fig. 1.
depth, and thus two or more layers within that depth range
will affect the behavior of the footing (Wang and Carter, Finite element analysis by MATLAB: The finite
2002). element method, which can handle very complex layer
Estimation of foundation settlement (immediate, patterns, has been applied to this problem and reliable
consolidation and creep) depends strongly on the some research had been undertaken investigating the
calculated stresses in the underlying soil mass due to probabilistic analysis of the settlement of foundations
footing pressure. In theory, accurate stresses can be supported on single-layered soil profiles for the calculated
calculated for any geometry using numerical methods stresses (e.g., Boussinesq method).
such as three-dimensional finite element or finite The well-known Boussinesq equations apply strictly
difference analysis (Hazzard et al., 2007). only to homogeneous soil deposits. The general theory of
The type of elastic nonhomogeneity is a useful stresses and displacements in a two-layer system is
approximation for modeling certain problems of developed in order to provide the engineer with a useful
geotechnical interest (Selvadurai, 1998). tool, which is more directly applicable to the analysis of
Griffith and Fenton (2001) observed that the actual conditions encountered in soil deposits (Burmister,
performance of foundations is considerably affected by 1945).
the inherent spatial variability of the soil properties. The case in the present paper is studied as a plane
In practice, most foundations are flexible. Even very thick strain two-dimensional problem. The shape of elements
ones deflect when loaded by the superstructure loads used is the quadrilateral element because of its suitability
(Bowles, 1996). to simulate the very important behavior of soils under
In this study, an elastic static loading problem for a strip footing.
continuously nonhomogeneous isotropic medium with The size of the finite element mesh in the horizontal
Young’s modulus varying linearly with depth is relevant. direction is adjusted according to the width of the footing.

718
 Res. J. App. Sci. Eng. Technol., 4(7): 717-723, 2012

Table 1: Comparison with the theoretical results

Smith and Griffiths
Item considered (1988) Results Author results
Ver. Disp. of Node 1 -0.8601E-05 - 8.6005657E-006
Ver. Disp. of Node 4 -0.3771E - 05-3.7707115E-006
Hor. Disp. of Node 4 -0.5472E - 07-5.4720226E-008
Ver. Stress at Elem. 1 0.8284 8.2847569E-001
Hor. Stress at Elem. 1 0.29075 2.9074293e-001

Table 2: The properties of soils

Parameters Layer 1 Layer 2
Eo (kN/m2) 12500 9000
k (kN/m2/m) 2000 500 Fig. 2: Sign convention and element numbering
< 0.4 0.3
The results obtained by the program in this research
It is generally recognized that the mesh boundaries should were compared with results presented by Smith and
be set at a distance at least five times the loaded area to Griffiths (1988). In all comparisons, excellent agreement
ensure that boundary effects do not influence the results was found between the present program results and those
(Desai and Abel, 1972). published, as shown in Table 1.
In the present study, 4-noded quadrilateral elements
with two degree of freedom at each node have been used Problem geometry: The basic problem chosen for the
to model soil. In each increment of the analyses, the parametric study shown in Fig. 5a, which involves soil
stress-strain behavior of the soil is treated as being linear, strata of clayey soils, 30.0 m width, of a first layer (3.0 m
and the relationship between stress and strain is assumed thick) over a second layer (18.0 m thick) underlaid by
to be governed by the generalized Hooke's law of elastic bedrock and loads by strip sequence loadings (40, 45, 55,
deformations, which may be expressed as follows for 70 kN/m2) with base width equal to 3.0 m.
conditions of plane strain case: The finite element mesh (Fig. 5b) used consists of
2623 nodal points and 2520 quadrilateral two-dimensional
⎡ v ⎤ elements. The nodal points along the bottom boundary of
⎢ 1 0 ⎥ the mesh are assumed to be fixed both horizontally and
⎧∆σ x ⎫ ⎢ 1− v ⎥ ⎧∆ε x ⎫
⎪ ⎪ E s (1 − v ) ⎢ v ⎪ ⎪ vertically. The nodes on the right and left ends of the
⎨∆σ y ⎬ = 1 0 ⎥ ⎨∆ε y ⎬ (2)
⎪∆σ ⎪ (1 + v)(1 − 2v) ⎢ 1 − v
⎢ ⎥ mesh are fixed in the horizontal direction while they are
⎪ ⎪
⎩ xy ⎭ 1 − 2v ⎥ ⎩∆ε xy ⎭ free to move in the vertical direction. All interior nodes
⎢ 0 0 ⎥
⎢⎣ 2(1 − v ) ⎥⎦ are free to move horizontally and vertically.
The first layer properties (Küçükarslan, 2007) and the
where )Fx, )Fy and )Jxy = the increments of stress during second layer properties (Das, 2006) are reported in
a step of analysis; Dgx, Dgy and )(xy = the corresponding Table 2. The behavior of soil materials is a
increments of strain; Es = the value of Young's modulus; nonhomogeneous elastic medium with modulus
and < = the value of Poisson's ratio. increasing linearly with depth.
The author using MATLAB writes a computer
program that used in the finite element analysis carried RESULTS AND DISCUSSION
out during this research. The soil model that is considered
in this work is nonhomogeneous, isotropic on primary In this study, a model of two-layered system was
loading with a different modulus. So, the behavior of the analyzed under uniformly flexible strip loading with soil
soil can be approximated by Gibson model (Gibson, modulus increasing linearly with depth. In order to
1967). develop more knowledge about the behavior of
The sign convention for the stresses and the multilayered soils under strip loading problems, a
convention for numbering the nodes of elements are parametric study is performed by varying the basic
shown in Fig. 2. The program presents the results of problem parameters and comparing these results with the
analysis as the displacements of the nodal points, and the original basic problem results.
value of stresses developed at the centre of each element The results of increasing the incremental loading,
at the end of each solution increment. soils having homogeneous (constant modulus with depth)
Figure 3 is a flowchart that illustrates the main to Gibson-type (linearly increasing modulus) profiles,
features of the solution procedure adopted in the finite foundation size (B) and embedment (Df), and soil
element computer program used. stratification are presented as follow:
For uniformly flexible strip loading area at surface,
Verification of the computer program: The author has the vertical displacement along the surface of the model
used this program in a different problem (Fig. 4) is shown in Fig. 6 and the contact settlement under the
presented by another researcher (Smith and Griffiths, strip footing is shown in Fig. 7. The settlement at the
1988). center is much larger than the settlement at the edge of the

719
 Res. J. App. Sci. Eng. Technol., 4(7): 717-723, 2012

Fig. 3: Simplified flow chart of the finite element program

Fig. 4: Mesh and data for different problem (Smith and

Griffiths, 1988) (a)

loaded area. These results agree with the results founded

by Wu (1974) and Das (2006). Also, the vertical
displacement increases in direct proportion to the pressure
of the loaded area, as shown in Fig. 6 and 7, which agrees
with that reported by Craig (1987).
The vertical stress contours throughout the soil under
the strip loadings (40, 45, 55, 70 kN/m2) with base width
(B) equal to (3.0 m) are shown in Fig. 8. It can be seen
that the vertical stress values along the depth of the layer
decrease throughout the layer for each increment and
increase throughout the loading sequence stages. (b)
From the settlements at surface (Fig. 9), it can be
seen that the settlements for soils with modulus increasing Fig. 5: The basic problem for the parametric study

720
 Res. J. App. Sci. Eng. Technol., 4(7): 717-723, 2012

0
linearly with depth are less than the settlements for soils
with constant modulus (Es, soil1 = Eo, soil1 and Es, soil2
-0.005
= Eo, soil2). The results agree with that mentioned by
Settlement (m)

-0.010 Terzaghi in Wu's book (Wu, 1974).

-0.015 The immediate settlement at the center of the loaded
-0.020
W = 40 kPa area decreases when the thickness of the first (upper)
W = 45 kPa
W = 55 kPa layer increases, as shown in Fig. 10, where the top layer
-0.025 W = 70 kPa is stronger than the bottom layer. In addition, the effect of
-0.030 the stiff (top) layer will be to reduce the stress
0 2 4 6 8 10 12 14 16 20 22 24 26 28 30
Horizontal distance(m)
concentration in the lower layer, as shown in Fig. 11.
These results agree with that discussed by Das (2008 and
Fig. 6: Settlements along the top of model under the strip 2009).
loadings (40, 45, 55, 70 kN/m2) with base width equal to The immediate settlement at the center of the loaded
(3.0 m) area is reduced when the strip footing is placed at some
-0.010
W = 40 kPa W = 55 kPa depth (Df # B) in the ground, as shown in Fig. 12. These
W = 45 kPa W = 70 kPa
results agree with that mentioned by Fox in Bowels' book
Contact settlement (m)

-0.015 (Bowles, 1996).

The vertical displacement (immediate settlement)
-0.020 increases in direct proportion to the width of the loaded
area (size of the footing) at surface, as shown in Fig. 13,
-0.025 which agrees with that reported by Wu (1974) and Craig
(1987).
-0.030
13.5 14.0 14.5 15.0 15.5 16.0 16.5
Horizontal distance under footing (m)
CONCLUSION

Fig. 7: Contact settlements under the strip loadings (40, 45, 55, The results obtained from this study can lead that a
70 kN/m2) with base width equal to (3.0 m) generalized numerical solution through a computer

21 21
20 20
19 40 19 45
18 18
17 35 17 40
16 16
15 30 15 35
14 14
13 25 13 30
12 12 25
11 20 11
10 15 10 20
9 9
8 10 8 15
7 7
6 8 6 10
5 5
4 6 4 8
3 3
2 5 .5 2 6
1 5 1 5.5
0 0
0 1 2 3 4 5 6 7 8 9 10111213141516 1718192021222324 25 2627282930 0 1 2 3 4 5 6 7 8 9 10111213141516 1718192021222324 25 2627282930

(a) for w = 40 kN/m2 (b) for w = 45 kN/m2

21 21
20 20
19 55 19 70
18 18 65
50
17 17 60
16 45 16
15 15 55
40
14 14 50
13 35 13 45
12 12
11 30 11 40
10 25 10 35
9 9 30
8 20 8
7 15 7 25
6 6 20
5 10 5 15
4 8 4
3 3 10
2 7 2 9
1 6.75 1 8.5
0 0
0 1 2 3 4 5 6 7 8 9 10111213141516 1718192021222324 25 2627282930 0 1 2 3 4 5 6 7 8 9 10111213141516 1718192021222324 25 2627282930

(c) for w = 55 kN/m2 (d) for w = 70 kN/m2

Fig. 8: Vertical stress contours for the soil model throughout the loading sequences

721
 Res. J. App. Sci. Eng. Technol., 4(7): 717-723, 2012

0 -0.023

Immediate settlement (m)

-0.01 -0.024
Settlement (m)

-0.02 -0.025

-0.03
E variable -0.026
E constant
-0.04
0 2 4 6 8 10 12 14 16 20 22 24 26 28 30 -0.027
Horizontal distance(m) 0 0.5 1.0
Depth of footing, Df (m)
Fig. 9: Settlements along the horizontal distance at surface
under the strip loading (70 kN/m2) Fig. 12: Immediate settlements at the center of the strip loading
-0.014 (70 kN/m2) with different depth of footing
Immediate settlement (m)

-0.008
-0.015

Immediate settlement (m)

-0.012
-0.016
-0.016
-0.017
-0.020
-0.018
-0.024
-0.019
0 1 2 3 4 -0.028
Thickness of the first layer (m) 1 2 3
Width of footing, B (m)
Fig. 10: Immediate settlements at the center of the strip loading
(40 kN/m2) with different thickness of the first (top) Fig. 13: Immediate settlements at the center of the strip loading
layer (70 kN/m2) with different width of footing
21 21
20 20
19 19 40
18
40
18
17 17 35
16 35 16
15 15 30
14 30 14
13 13 25
12 25 12
11 11 20
10 10
20
9 9 15
8 8
7 15 7 10
6 6
5 10 5 6
4 4
3 6 3 5
2 2
1 1 4.85
5
0 0
0 1 2 3 4 5 6 7 8 9 10 1112 1314 1516 17 18 19 2021 2223 24 2526 2728 2930 0 1 2 3 4 5 6 7 8 9 101112 131415 16 171819 202122 23 2425 26 272829 30

(a) thickness of first (top) layer = 0 m (b) thickness of first (top) layer = 4 m

Fig. 11: Vertical stress contours for the soil model for the strip loading (40 kN/m2)

program, written in MATLAB, can simulate the analysis beneath a strip footing resting on multilayered Gibson-
of multilayered nonhomogeneous (Gibson-type) soils that type soils.
had a soil modulus increasing linearly with depth and The immediate settlement at the center is much larger
loaded with incremental strip loading. than the settlement at the edge of the strip loaded area.
Displacements and stresses can be calculated with The immediate settlement increases in direct proportion
knowledge of soil stiffness beneath the footing, rate of to the pressure of the strip loaded area. The vertical stress
increase of soil stiffness with depth, soil Poisson’s ratio, values (stress bulb) under the strip loading area decrease
depth to an incompressible layer, and footing width. throughout the layers for each increment and increase
This study shows how the computer solutions may be throughout the loading sequence stages. The vertical
used to improve the prediction of settlements and stresses displacements for soils with moduli increasing linearly

722
 Res. J. App. Sci. Eng. Technol., 4(7): 717-723, 2012

with depth (Gibson-type) are less than the vertical Küçükarslan, S., 2007. The Effect of a Transient Dynamic
displacements for the same soils with constant moduli, Loaded Pile on a Neihbouring Pile for
which leads to that the soils with (Gibson-type) Nonhomogeneous Soils. International Earthquake
modulus, are more approximate simulation for soil Symposium, Kocaeli, Turkey.
modulus. The immediate settlement decreases when the Poulos, H.G., J.P. Carter and J.C. Small, 2001.
thickness of the upper (stronger) soil layer overlaying a
Foundations and Retaining Structures -Research and
weaker soil layer increases and the stress bulb of strip
foundation reduces when the thickness of top (stronger) Practice. Proceeding the 15th International
soil layer increases. The immediate settlement of the strip Conference on Soil Mechanics and Geotechnical
loaded area decreases when the embedment of strip Engineering, Istanbul, Turkey, 4: 2527-2606.
footing increases and the immediate settlement increases Radhakrishnan, N., 1969. Solution of plane strain
with the increasing of foundation size. The results problems in soil mechanics using the method of finite
compare approvingly with available published analytical elements. Ph.D. Thesis, The University of Texas at
and numerical solutions. Austin, Texas, USA.
Rowe, R.K. and J.R. Booker, 1982. Finite layer analysis
REFERENCES
of nonhomogeneous soil. Journal of Engineering
Bowles, J.E., 1996. Foundation Analysis and Design. 5th Mechanics Division, Proceedings of the American
Edn., McGraw-Hill Companies. Society of Civil Engineers, ASCE, Vol. 108, No.
Burmister, D.M., 1945. The general theory of stresses and EM1.
displacements in layered systems. J. Appl. Phys., Selvadurai, A.P.S., 1998. Lan Q. Axisymmetric mixed
16(2): 89-94. boundary value problems for an elastic halfspace
Craig, R.F., 1987. Soil Mechanics. 4th Edn., Van with a periodic nonhomogeneity. Int. J. Solids
Nostrand Reinhold Co. Ltd, UK. Structure, 35: 1813-1826.
Das, B.M., 2006. Principles of Geotechnical Engineering. Smith, I.M. and D.V. Griffiths, 1988. Programming the
5th Edn., Thomson Ltd, Canada. Finite Element Method. 2nd Edn., John Wiley &
Das, B.M., 2008, Advanced Soil Mechanics. 3rd Edn.,
Taylor and Francis Group. Sons Ltd.
Das, B.M., 2009. Shallow Foundations: Bearing Capacity Stark, R.F. and J.R. Booker, 1997. Surface displacements
and Settlement. 2nd Edn., Taylor and Francis Group. of a non-homogenous elastic half-space subjected to
Desai, C.S., 1968. Solution of stress deformation uniform surface tractions. Inter. J. Numer. Anal.
problems in soil and rock mechanics using finite Methods Geomech., 21(6): 361-378.
element methods. Ph.D. Thesis, The University of Terzaghi, K., R.B. Peck and G. Mesri, 1996. Soil
Texas at Austin, Texas, USA. Mechanics in Engineering Practice. 3rd Edn., John
Desai, C.S. and J.F. Abel, 1972. Introduction to the Finite Wiley & Sons Ltd.
Element Method. Van Nostrand Reinhold, London. Wang, C.D., C.S. Tzeng, E. Pan and J.J. Liao, 2003.
Douglas, D.J., 1986. State-of-the-art. Ground Eng., 19(2):
2-6. Displacements and stresses due to a vertical point
Gibson, R.E., 1967. Some results concerning load in an inhomogeneous transversely isotropic half-
displacements and stresses in a non-homogeneous space. Inter. J. Rock Mech. Min. Sci., 40: 667-685.
elastic half-space. Geotechnique, 17: 58-67. Wang, C.X. and J.P. Carter, 2002. Deep penetration of
Griffith, D.V. and G.A. Fenton, 2001. Bearing capacity of strip and circular footings into layered clays. Inter. J.
spatially random soil: The undrained clay prandtl Geomechanics, 2(2): 205-232.
problem revisited. Geotechniqe, 51(4): 351-359. Wu, T.H., 1974. Soil Mechanics. 7th Printing, Allyn and
Hazzard, J.F., T.E. Yacoub, S. Vijayakumar and Bacon Inc.
J.H. Curran, 2007. Stresses Under Footings in
Multilayered Soils: A Comparative Study.
Proceedings the 60th Canadian Geotechnical
Conference and 8 th Joint CGS/IAH-CNC
Groundwater Specialty Conference; Ottawa, Canada.

723