Interaction and Multiscale Mechanics, Vol. 4, No.

1 (2011) 17-34 17
DOI: http://dx.doi.org/10.12989/imm.2011.4.1.017

Numerical simulation of soil-structure interaction

in framed and shear-wall structures
M. Dalili , A. Alkarni , J. Noorzaei* , M. Paknahad , M.S. Jaafar and B.B.K. Huat
1,2 3 1,2 1,2 1 1

1
Civil Engineering Faculty, Universiti Putra Malaysia, Malaysia
2
Institute of Advanced Technology, Universiti Putra Malaysia, Malaysia
3
Civil Engineering Faculty, King Saud University, Riyadh, Saudi Arabia
(Received September 13, 2010, Accepted January 26, 2011)

Abstract. This paper deals with the modeling of the plane frame structure-foundation-soil system. The
superstructure along with the foundation beam is idealized as beam bending elements. The soil medium
near the foundation beam with stress concentrated is idealized by isoparametric finite elements, and
infinite elements are used to represent the far field of the soil media. This paper presents the modeling of
shear wall structure-foundation and soil system using the optimal membrane triangular, super and
conventional finite elements. Particularly, an alternative formulation is presented for the optimal triangular
elements aimed at reducing the programming effort and computational cost. The proposed model is
applied to a plane frame-combined footing-soil system. It is shown that the total settlement obtained from
the non-linear interactive analysis is about 1.3 to 1.4 times that of the non-interactive analysis. Furthermore,
the proposed model was found to be efficient in simulating the shear wall-foundation-soil system, being
able to yield results that are similar to those obtained by the conventional finite element method.
Keywords: triangular element; shear wall structure; super element; drilling degree of freedom.

1. Introduction
The interaction among structures, their foundations and the soil medium below the foundations
considerably alters the behavior of the structure compared with the case where the structure is
considered alone. Thus, a reasonably accurate model for the soil-foundation-structure interaction system
with computational validity and efficiency is needed to improve the design of important structures.
Extensive research was conducted by Viladkar and co-workers on 2D and 3D framed structure-
foundation-soil interaction problems, as well as shear wall foundation-soil systems (Viladkar,
Godbole et al. 1991, Viladkar, Godbole et al. 1992, Viladkar, Godbole et al. 1994, Viladkar, Noorzaei
et al. 1994, Viladkar, Sharma et al. 1992). They employed the isoparametric beam bending element
to represent the superstructure and the coupled finite and infinite elements to model the near and far
fields of soil media, respectively. The superstructure materials were assumed to obey the Hookean
law, while the stress-strain relationships of the soil were considered as nonlinear elastic. In the study
by (Pandey, Kumar et al. 1994), an iterative approach was proposed for the soil-structure interaction
* Corresponding author, Ph.D., E-mail: [email protected]
18 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

of tall buildings having closely spaced independent footings. In order to verify the proposed
approach, a finite element analysis was carried out for a 10 storey plane frame. The computational
effort was shown to be less than that needed for the conventional finite element analysis.
A comprehensive review on the modeling of soil-structure-foundation system was presented in Dutta
and Roy (2002), in which an attempt was made to review the possible models available in the literature
for soil structure interaction analysis. Major attention was focused on the physical modeling of the soil
media. The authors concluded that the Winkler theory, despite its obvious limitations, yields reasonable
performance and it is easy to exercise. Therefore, for practical purpose, such an idealization should be
employed, instead of using structures with fixed base. Furthermore, the effect of material nonlinearity,
non-homogeneity and anisotropy of the supporting soil medium should be taken into account.
The method presented by Fan et al. (2005), applicable for the analysis of superstructure-foundation
interaction, was a method combining the analytical solution with finite element method. This method
enables us to obtain the nodal forces of the superstructure and pile–soil system, bending and torsional
moments, and the normal and shear stresses at any point in the raft foundation.
A new method was proposed by Badie and co-workers for analyzing shear wall structures on
elastic foundations. The soil is modeled using a three-nodded quadratic element that includes the
vertical sub grade reaction technique and soil shear stiffness (Badie and Salmon 1996, 1997, Badie,
Salmon et al. 1997). In their studies, two types of soil were assumed, namely, dense sand and soft
clay, and the flexibility for the base of shear walls was modeled by the Winkler model. It was
observed that the maximum drift is about six times that of a rigid foundation for dense sand and
about three hundred times that of a rigid foundation for soft clay.
Nadjai and Johnson (1998) investigated the importance of base flexibility on the behavior of planar
shear walls subjected to lateral loading. A simple analytical model with basic assumptions was
proposed and compared the results obtained with the recorded response (Boroschek and Yáñez 2000).
Super element was presented for the analysis of shear walls with opening (Kim and Lee 2003).
Soil-pile-structure interaction of a two bay single storey frame was carried out considering a
group pile embedded in clayey soil supporting a flexible cap to study the effect of pile cap-soil
interaction and piles configuration and spacing on the maximum displacement at the top of the
frame and on the moments in columns of the superstructure (Chore et al. 2009). Influence of soil-
pile interaction has been found considerable on the behavior of pile cap-soil media.
The above study has been further extended by Chore et al. (2010a, b) where authors introduced a
modified finite element idealization through idealizing the soil, pile and pile cap with linear spring
element, one dimensional twenty-nodded iso-parametric continuum beam element and two dimensional
twenty-nodded iso-parametric continuum plate element respectively. The three major focus points
considered in their uncoupled analysis have been pile spacing, configuration of the piles and
diameter of the piles on the behavior of the superstructure.
Two group piles were considered one including two piles and another one with three piles. With
the proposed idealization method the effect of interaction was found to be significant with a
considerable amount of superstructure displacement compared to the fixed based structure.
It is clear from the above literature review that so far no attempt has been made to consider: (a)
numerical simulation of framed structures pile foundation-soil on the slope, (b) numerical model
that includes the super element or drilling element and conventional finite elements for simulation
of the shear wall foundation and soil media as a single compatible unit. The present study is a
continuation of authors’ previous works (Noorzaei, Alkarni et al. 2010, Paknahad, Noorzaei et al.
2007, Paknahad, Noorzaei et al. 2009), who analyzed a shear wall structure using the optimal
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 19

membrane triangular element. Hence the objective of this study is focused on:
(i) Modelling of framed structures-pile foundation – soil media built on inclined strata.
(ii) Modelling of shear wall-foundation using modified optimum triangular element with drilling
degree of freedom.
(iii) Modelling of soil systems with conventional finite element and super element to reduce the
computational time and effort without sacrificing the accuracy.
(iv) Examination of the effect of different types of soil (soft to stiff) on the overall interactive
behaviour of the shear wall foundation soil system.
The formulation of the Modified Optimal Triangular (MOPT) element and generation of the super
element are presented in subsequent sections.

2. Formulation of optimal membrane elements

Felippa and co-workers are the key contributors in the development of elements with drilling
degrees of freedom (Felippa 1996, 2003, Felippa and Alexander 1992, Felippa and Militello 1992)
They have developed elements of this sort that are appropriate for the two dimensional
representation of shear wall structures.
In this study, the Natural Deviatoric Strain (ANDES) formulation and the Natural Strain (ANS)
method is applied (Park and Stanley 1986). ANDES is a variant of ANS that exploits the
fundamental decomposition of the stiffness as follows
KR = ( Kb + α Kh ) V (1)
Here K is the basic stiffness, which satisfies consistency and K is the higher order stiffness
b h

which deals with stability (rank sufficiency) and accuracy and α is scaling coefficient with positive
magnitude. The basic stiffness matrix K is constructed by the standard procedure, i.e. for the
b

constant strain triangular (CST) element. An explicit form of the basic stiffness for the linear strain
triangle (LST) was obtained in 1984 and published the following year. It can be expressed as
K = V − L E L , where V = Ah is the element volume, A the area, h the thickness, and L is a 3 × 9
b
1 T

matrix that contains a free parameter α (Bergan 1985) b

y 23 0 x 23
0 x 23 y 23
1--- α y ( y – y ) 1--- α x ( x – x ) 1--- α ( x y – x y )
6 b 23 13 21 6 b 32 31 12 3 b 31 13 12 21
y 31 0 x 13
L = 1--- h 0 x 13 y 31 (2)
2 1 1--- α x ( x – x ) 1--- α ( x y – x y )
--- α b y31 ( y 21 – y32 )
6 6 b 13 12 23 3 b 12 21 23 32
y 12 0 x 21
0 x 21 y 12
1--- α y ( y – y ) 1--- α x ( x – x ) 1--- α ( x y – x y )
6 b 12 32 13 6 b 21 23 31 3 b 23 32 31 13
20 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

In the free formulation (FF) this is called a force-lumping matrix, If α b = 0 the basic stiffness
reduces to the total stiffness matrix of the Constant Strain Triangle (CST) element. For this case, the
rows and columns associated with the drilling rotations vanish.
The mean portion of the strains is left to be determined from the constant stress assumptions used
to develop K . The main advantage of ANDES over ANS is that elements constructed with the
h

former technique are guaranteed to pass the individual element test (Bergan 1985, Park and Stanley
1986). In the derivation of the higher order stiffness by ANDES, the natural strains play a key role.
The ANDES form of the higher order stiffness matrix K has been derived by some researchers as
h

(Felippa and Alexander 1992, Paknahad, Noorzaei et al. 2007)

K h = C fac T Tθ u K θ T θ u (3)
In which Kθ is the higher order stiffness in terms of hierarchical rotations and depends on nine
free dimensionless parameters β through β , as will be explained below. First, the matrix Q that
1 9 i

relates the natural strains i at corner i to the deviatoric corner curvatures θ of the triangle can be
given as follows
β1-
---- β2-
---- β3-
----
l 221 l 221 l 221
β4-
Q 1 = 2------A- ---- β5-
---- β6-
---- (4a)
3 l232 l 232 l 232
β7-
---- β8-
---- β9-
----
l 213 l 213 l 213
β9-
---- β7-
---- β8-
----
l 221 l 221 l 221
β3-
Q 1 = 2------A- ---- β1-
---- β2-
---- (4b)
3 l232 l 232 l 232
β6-
---- β4-
---- β5-
----
l 213 l 213 l 213
β5- ----
---- β6- ----
β4-
l 21 l 21 l 221
2 2

Q 1 = 2------A- ---- β9- ----

β8- ---- β7- (4c)
3 l32 l32 l232
2 2

---- β3- ----

β2- ---- β1-
l 13 l 13 l 213
2 2

of which the values at the midpoints are as follows

Q 4 = 1--- ( Q 1 + Q 2 ) (5a)
2
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 21

Table 1 Dimensionless parameter of OPT element

αb β0 β1 β2 β3 β4 β5 β6 β7 β8 β9
3/2 1/2 1 2 1 0 1 −1 −1 −1 −2

Q 5 = 1--- ( Q 2 + Q 3 ) (5b)
2
Q 6 = 1--- ( Q 3 + Q 1 ) (5c)
2
Then the matrix Kθ can be related to above quantities as
K θ = h ( Q T4 E nat Q 4 + Q T5 E nat Q 5 + Q T6 E nat Q 6 ) (6)
In which E is the natural stress-strain matrix. The stiffness matrix K can be given as follows
nat h

K h = 3--- β 0 T Tθ u K θ T θ u (7)
4
Where β0 is an overall scaling coefficient. Finally K , which is the fundamental decomposition of
R

stiffness, see Eq. (1), initially constructed by combining basic stiffness (K ) and the higher order
b

stiffness (K ), assumes a template form with 11 free parameters (Felippa 1996, 2003), i.e. α , β ,
h b 0

β ... β .
1 9

K R ( α b, β 0, β 1, …, β 9 ) = V –1 L E L T + 3--- β T Tθ u K θ Tθ u (8)
4
In which V is the element volume, i.e. V = Ah, L is the free-lumping matrix and Tθ stands for the
u

transformation matrix. The free dimensionless parameters α , β through β are determined from a
b 0 9

higher order patch test that tunes up the higher order stiffness of the triangular elements. Their
values are listed in Table 1.

3. A fast calculation of higher order stiffness matrix

In this study, an alternative formulation of the above element is presented, which turns out to be
more efficient than the optimal element proposed by Felippa (2003). Let us assume the LL matrix as
l 221 0 0
LL = 0 l 232 0 (9)
0 0 l213
With this matrix, we can rewrite the Q , T and E*nat matrices as
i e

Q i = LL–1 Q *i (10a)
Te = T *e LL (10b)
E *nat = T *e E T*e (10c)
22 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

Q*i and T*e are as follows

β1 β2 β3 β9 β7 β8 β5 β6 β4
2 A
Q i = ------- β 4 β 5 β 6
* 2 A
Q = ------- β 3 β 1 β 2
* 2 A
Q = ------- β 8 β 9 β 7
*
3 2
3 3
3
β7 β8 β9 β6 β4 β5 β2 β3 β1

y 23 y 13 y 31 y 21 y 12 y 32
1-
T *e = -------- x 23 x 13 x 31 x 21 x 12 x 32
4 A2
y 23 x 31 + x 32 y 13 y 31 x 12 + x 13 y 21 y 12 x 23 + x 21 y 32
In addition,
Q *4 = LL–1 Q 4 Q *5 = LL –1 Q 5 Q *6 = LL –1 Q 6
Substituting the new form of T and Q in the high order stiffness formulation Kθ, Eq. (6), and
e i

expanding, one has

T T
K θ = h ( ( LL–1 Q *4 ) ( T *e LL ) E ( T *e LL ) ( LL –1 Q *4 )
T T T T
+ ( LL–1 Q*5 ) ( T*e LL ) E ( T *e LL ) ( LL–1 Q*5 ) + ( LL –1 Q *6 ) ( T*e LL ) E ( T*e LL ) ( LL–1 Q*6 ) ) (11)
After simplification
K θ = h ( Q *4T E *nat Q *4 + Q *5T E *nat Q *5 + Q *6T E*nat Q *6 ) (12)
Hence the formulation of the matrix K with new notation is more efficient in terms of
h

computational time and effort compared with that of Felippa (2003). The same formulation has been
adopted for the calculation of stresses.

4. Generation of super element in shear wall system

The stiffness matrix of each super element is calculated independently and the super element is
subsequently joined to other elements in the overall analysis. Kim and Lee (2003) presented the
super element for the analysis of shear walls with openings, (Paknahad, Noorzaei et al. 2009).
Fig. 1 briefly shows the computational strategy used in the present study to generate the super
element, including the assembly and final configuration of a shear wall building. The procedure
enables us to systematically reduce the number of assembled equations on different substructure
levels. The process of the super element generation is as follows:
a. Selecting the type of super element based on geometrical configuration of the considered
structural system and fixing the position of super elements,
b. Generating the finite element mesh utilizing the eight node isoparametric elements,
c. Evaluating the stiffness matrix for each element,
d. Applying the condensation technique Eq. (13) and extracting the super element stiffness matrix,
e. Writing down the stiffness matrix of super elements on an auxiliary device and prepare for
assembly of stiffness matrix (Paknahad 2008, Paknahad, Noorzaei et al. 2009).
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 23

Fig. 1 Implementation of super elements

Table 2 Column dimensions

Grid A Grid B Grid C Grade Ec
Depth (mm)Width (mm)Depth (mm)Width (mm)Depth (mm)Width (mm) (N/mm2)
20 to Roof
th
600 600 675 600 675 600 C40 28
15 to 20
th th
600 600 900 600 825 600 C40 28
10th to 15th 675 600 1125 600 900 600 C40 28
5th to 10th 825 600 1425 600 1050 600 C40 28
G to 5th 1050 600 1650 600 1275 600 C40 28
LG3 to G 1200 600 1800 600 1500 600 C45 29

( Krr – Krc K–cc1 Kcr ) Ur = Rr – Krc K–cc1 Rc (13)

The K , K and K are stiffness matrix and R and R are external load vector corresponded to
rr rc cc r c

displacements U and U , where subscripts “r” and “c” indicate the retained and condensed degree
r c

of freedom respectively.

5. Analysis of 28 story building system

Let us consider a 28 storey, three bay plane frame, supported by piles of unequal lengths with
th

sockets embedded into the rock at a depth of 2 m. Pile A ends at a depth of 19 m, pile B at a depth
of 15 m and pile C at a depth of 11m. This frame structure-piles system is supported by four layers
of soil and a rock layer, with their nonlinear stress-strain characteristics taken into account. The
material properties, geometry and loadings of the superstructure are shown in Tables 2, 3 and Fig. 2,
respectively. The vertical and lateral loadings acting on the building are presented in Fig. 2.
The tangent modulus, E of the top to the most bottom soil layer is 52.49, 122.51, 187.25 and
i

205.37 kg/cm , respectively. The Finite Element Mesh configuration of the system is shown in Fig.
2

3. The soil boundary between the soil and rock is represented in the figure by a line passing
24 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

Table 3 Beam dimensions

Grid A to Grid B Grid B to Grid C Grade
Depth (mm) Width (mm) Depth (mm) Width (mm)
Roof 600 600 600 600 C30
11th to 24th 600 375 600 375 C30
1st to 10th 600 300 600 300 C30
LG2 to G 600 300 600 300 C35
LG3 600 450 600 450 C35

Fig. 2 Frame dimensions

through the soil strata. The symbol B denotes the width of the superstructure. The upper layer
consists of soil and the lower layer consists of rock. The detailed configuration of the superstructure
is shown in Fig. 2.
The following analyses were performed on the above structure:
a) Non-Interactive Analysis (NIA).
b) Linear Interactive Analysis (LINA).
5.1 Results and discussion
The displacements in various columns and the maximum displacements of the beams of the frame
obtained by the two methods, namely, the Non Interactive Analysis (NIA) and Linear Interactive
Analysis (LIA), were listed in Table 5. As can be seen, the displacement due to the interactive
analysis is higher compared with the non interactive analysis. This is due to the fact that NIA
ignores the effect of soil compressibility, as the base is assumed to be fixed to the ground, and vice
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 25

Fig. 3 Finite element mesh

Table 4 Maximum displacements of beams and columns by the two analyses

Max. Beam Displ. (mm) Max. Column Displ. (mm)
− A B C
Analysis NIA 12.241 64.447 64.525 64.576
LIA 14.926 83.233 83.20 83.094

Table 5 Soil properties for different types of soils (Selvadurai 1979)

No Soil types E (kN /m2) ν
1 Sandy clay 35000 0.32
2 Gravely sand 57600 0.30
3 Dense gravely sand 95000 0.25

versa for the interactive analysis. Moreover, the maximum lateral displacement of LIA is
approximately 1.3 to 1.4 times higher than that of NIA respectively. This is due to the fact that the
LIA considers the settlement of the soil.
The variation of lateral displacement of the piles by LIA is plotted in Fig. 4. It is clear that the
pattern of the displacement for each pile is rather similar, whereby the displacement diminishes as
the depth increase. This is due to the inclined strata of bedrock and unequal lengths of the piles.
The trend of displacement decreases nearly to zero for the pile end with socket embedded in the
bedrock. Since the rock is a very stiff material, the displacement in the rock is nearly zero. In
addition, the displacement of each pile becomes zero at different depths.
Fig. 5 shows the variation of lateral displacement along pile and soil depth obtained by different
analyses, from which it can be seen that the difference in displacement between the pile and soil
(rock) is approximately in the range of 6 to 8 mm.
26 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

Fig. 4 Variation of lateral displacement of piles (LIA)

Fig. 5 Variation of horizontal displacement along pile and soil depth (LIA)

6. Analysis of shear wall buildings with openings

The application of the proposed finite element model for the shear wall foundation soil system is
demonstrated in the analysis of an 8-story shear wall building. Fig. 6 shows the material and
geometry of the shear wall considered, which has a height of 32 m, width of 10 m and thickness of
0.4 m. The material property and locations of the openings along the height are also shown in this
figure. In the non-interactive analysis, two finite element models are used:
Model I: Constructed by eight-node conventional isoparametric element as shown in Fig. 6(b).
Model II: Constructed by two types of super element respectively for the main shear wall area
with 20 nodes and the connection beam with 6 nodes in Fig. 6(c).
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 27

Fig. 6 Geometry, material and mesh generation of shear wall with openings

Fig. 7 Computation time versus different stages of analysis

6.1 Discussion and comparison of the results

The computational time versus the different stages of finite element operation corresponding to
both structural models are presented in Fig. 7. It is clear from this plot that there is a considerable
reduction in the computational time in using the super element model as compared with the
conventional eight-node finite element model, the latter has a computation time of 350 % higher.
This is an indication of the efficiency of using the super element.
The variation of horizontal displacement along the height of the shear wall is shown in Fig. 8. As
depicted in this figure, the displacements of all the retained nodes of the super element model are
coincident with those obtained by the conventional finite element model (model I).
28 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

Fig. 8 Horizontal displacement of shear wall

Fig. 9 Finite element mesh of shear wall foundation with soil media (MOPT and conventional elements)

7. Linear interactive analysis via MOPT and conventional finite elements

The finite element mesh used in modeling has been shown in Fig. 9. The modified optimal
triangular element (MOPT) was used to model the super structure and foundation, while the eight-
node isoparametric element was employed to discretise the soil media. The effect of stiffness of the
soil was investigated by using three types of soil, of which the material properties are listed in Table 7.
7.1 Discussion of results
Figs. 10 (a) and (b), respectively, shows the horizontal and vertical displacements of the shear
wall along its height. From these figures, the effect of different stiffnesses of the soil on the
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 29

Fig. 10 Displacements of shear wall – foundation – soil media: (a) Horizontal, (b) Vertical

response of the shear wall building can be clearly observed. For the soil with Young’s modulus of
elasticity as E = 35000, 57600 and 95000 kN/m , the top horizontal displacement was calculated to
s
2

be 0.607, 0.387 and 0.254 m, respectively. Clearly, the maximum displacements occur at the top and
the minimum displacements at the base of the soil media. Also, the effect of soft soil on both the
vertical and horizontal displacements of the shear wall structure can be appreciated.
The normal stress across the width of shear wall at an elevation of 48.281 m below the
superstructure obtained by the interactive analysis is compared with that of the non interactive
analysis in Fig. 11. It can be seen that the support condition leads to a distribution of stress in the
shear wall. Furthermore, this figure indicates high concentration of stresses at the corners of wall
originated by flexibility of the soil, resulting in stress redistribution in the shear wall.
30 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

Fig. 11 Comparison of normal stress along the width of shear wall at elevation 48.281

Fig. 12 Finite element discretisation of shear wall-foundation-soil media using MOPT and super element

8. Linear interactive analysis with MOPT and super element

As a further demonstration of the application of the super element, an attempt has been made to
represent the super structure by MOPT and the soil media by the super element. Fig. 12 shows the
finite element mesh used in modeling of the superstructure, foundation and soil system. In this
mesh, the MOPT element has been used to represent the superstructure and the foundation, and a
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 31

Fig. 13 Horizontal displacements of shear wall-soil media: (a) E = 35000, (b) E = 57600, (c) E = 95000 kN/m2
32 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

Fig. 14 Computational time versus different stages of analysis

super element with 44 nodes has been generated for the soil media. In the generation of the super
element, initially the soil was discretised by eight-node isoparametric element and then the
computational algorithm was activated to generate the super element. Effort has been made to
always maintain compatibility at the contact nodes between the foundation and soil media.
8.1 Discussion of results
An attempt has been made to compare the present results with those obtained by previous finite
element model. The variation of lateral displacements along the height of the shear wall for three
types of soil is illustrated in Fig. 13. It is clear from these plots that the maximum displacement of
250 mm occurs at the top and the minimum displacement of 37.2 mm at the foundation level.
However the relative displacement of 100.2% at the top and 78% at the bottom predicted by both
idealizations are almost the same. These figures show that the super element can provide a good
estimate of the displacement for the shear wall structure. Similar trend was observed for different
types of soil.
Fig. 14 shows the computational time taken at different steps for the analysis of the shear wall
building foundation soil system. This plot clearly illustrates the efficiency of the super element in
the analysis of shear wall - foundation -soil media. The computation time needed for analysis with
the super element case is around 6 times less than that of the conventional finite element method.

9. Conclusions
The major conclusions drawn from the element formulation and numerical analyses presented in
this paper are as follows:
• Optimal triangle element was formulated and its computational efficiency was found in
representation of shear wall system.
• The finite element idealization using the super element – finite element model is very
encouraging when the displacement and stress analysis of the super structure – foundation – soil
 Numerical simulation of soil-structure interaction in framed and shear-wall structures 33

media is of primary concern.

• The effect of interaction between the soil and shear wall is important and cannot be neglected to
ensure the accuracy of the result. The interaction analysis shows remarkable difference in the
displacements and stress distribution of the shear wall.
• Different types of soil have significant effect on the displacement and stress response of the
shear wall. There is a marked increase in the maximum deflection, say, 25%, of the shear wall–
foundation–soil due to inclusion of soil nonlinearity.
• Accurate results can be obtained by using plane strain soil element along with an infinite
element to account for the semi-infinite extent of the soil domain.
• As for interactive behavior, it has been realized that the total settlement obtained from the linear
interactive analysis is about 1.3 to 1.4 times that of the non-interactive analysis.

Acknowledgments
The financial support for this research is provided by King Saud University, Saudi Arabia under
grant number of 67001. This support is gratefully acknowledged and appreciated.

References
Badie, S. and Salmon, D. (1996), “A quadratic order elastic foundation finite element”, Comput. Struct., (3),
58

435-443.
Badie, S. and Salmon, D. (1997), “Three-noded curved isoparametric soil interface element”, Comput. Struct.,
65 (2), 205-212.
Badie, S., Salmon, D. and Beshara, A. (1997), “Analysis of shear wall structures on elastic foundations”,
Comput. Struct., (2), 213-224.
65

Bergan, C. (1985), “A triangular membrane element with rotational degrees of freedom”, Comput. Method. Appl.
M., (1), 25-69.
50

Boroschek, R. and Yáñez, F. (2000), “Experimental verification of basic analytical assumptions used in the
analysis of structural wall buildings”, Eng. Struct., (6), 657-669.
22

Chore, H., Ingle, R. and Sawant, V. (2009), “Building frame-pile foundation-soil interactive analysis”, Interact.
Multiscale Mech., (4), 397-411.
2

Chore, H., Ingle, R. and Sawant, V. (2010a), “Building frame-pile foundation-soil interaction analysis: a
parametric study”, Interact. Multiscale Mech., (1), 55-79.
3

Chore, H., Ingle, R. and Sawant, V. (2010b), “Parametric study of pile groups subjected to lateral load”, Struct.
Eng. Mech., (2), 243-246.
36

Dutta, S. and Roy, R. (2002), “A critical review on idealization and modeling for interaction among soil-
foundation-structure system”, Comput. Struct., (20-21), 1579-1594.
80

Felippa, C. and Alexander, S. (1992), “Membrane triangles with corner drilling freedoms - III. Implementation
and performance evaluation”, Finite Elem. Anal. Des., (3-4), 203-239.
12

Felippa, C. and Militello, C. (1992), “Membrane triangles with corner drilling freedoms - II. The ANDES
element”, Finite Elem. Anal. Des., (3-4), 189-201.
12

Felippa, C. (1996), “Refined finite element analysis of linear and nonlinear two-dimensional structures”,
Structures and Materials Research, Dept. of Civil Engineering, Univ. of Calif., Berkeley, Report, 66-22.
Felippa, C. (2003), “A study of optimal membrane triangles with drilling freedoms”, Comput. Method. Appl. M.,
192 (16-18), 2125-2168.
Kim, H. and Lee, D. (2003), “Analysis of shear wall with openings using super elements”, Eng. Struct., (8),25

981-991.
34 M. Dalili, A. Alkarni, J. Noorzaei, M. Paknahad, M.S. Jaafar and B.B.K. Huat

Nadjai, A. and Johnson, D. (1998), “Elastic and elasto-plastic analysis of planar coupled shear walls with flexible
bases”, Comput. Struct., (1-3), 213-229.
68

Noorzaei, J., Alkarni, A., Jaafar, M.S. and Dalili, M. (2010), “Modelling of soil structure interaction in framed
and shear wall structures”, The First International Conference on Advances in Interaction and Multiscale
Mechanics, Jeju, Korea.
Paknahad, M., Noorzaei, J., Jaafar, M. and Thanoon, W. (2007), “Analysis of shear wall structure using optimal
membrane triangle element”, Finite Elem. Anal. Des., (11-12), 861-869.
43

Paknahad, M. (2008), Development of finite element codes for shear wall analysis, Ph.D Thesis, Civil
Engineering, Universiti Putra Malaysia.
Paknahad, M., Noorzaei, J., Jaafar, M. S. and Thanoon, W.A.M. (2009), “Nonlinear analysis of shear wall
system with Super element”, Struct. Eng., (5), 319-326.
36

Pandey, A.K., Kumar, G. and Shamra, S.P. (1994), “An interactive approach for the soil-structure interaction in
tall buildings”, Eng. Fract. Mech., (2), 169-176.
41

Park, K. and Stanley, G. (1986), “A curved C° shell element based on assumed natural-coordinate strains”, J.
Appl. Mech., , 278-290.
53

Viladkar, M., Godbole, P. and Noorzaei, J. (1991), “Soil-structure interaction in plane frames using coupled
finite-infinite elements”, Comput. Struct., (5), 535-546.
39

Viladkar, M., Godbole, P. and Noorzaei, J. (1992a), “Space frame-raft-soil interaction including effect of slab
stiffness”, Comput. Struct., (1), 93-106.
43

Viladkar, M., Sharma, R. and Ranjan, G. (1992b), “Visco-elastic finite element formulation for isolated
foundations on clays”, Comput. Struct., (2), 313-324.
43

Viladkar, M., Godbole, P. and Noorzaei, J. (1994a), “Modelling of interface for soil-structure interaction studies”,
Comput. Struct., (4), 765-779.
52

Viladkar, M., Noorzaei, J. and Godbole, P. (1994b), “Behaviour of infinite elements in an elasto-plastic domain”,
Comput. Struct., (4), 337-342.
51