See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/291527386

Counterfort retaining walls under earthquake force

Article · January 2004

CITATIONS READS
4 846

2 authors:

Indrajit Chowdhury Shambhu Dasgupta

Independent Researcher Indian Institute of Technology Kharagpur
73 PUBLICATIONS   295 CITATIONS    43 PUBLICATIONS   364 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Dynamic soil structure interaction in elalstic domain and some solutions under earthquake force. View project

Seismic response of rectangular liquid retaining structures resting on ground considering coupled soil-structure interaction View project

All content following this page was uploaded by Indrajit Chowdhury on 13 November 2016.

The user has requested enhancement of the downloaded file.

 Seismic Response of a Counterfort
Retaining Wall With Generalized
Backfilled Soil
Indrajit Chowdhury
Head of the Department
[email protected]

Jitendra P. Singh
Technical Specialist
[email protected]
Civil and Struct. Engg., Petrofac International Ltd., Sharjah, U.A.E.

Shambhu P. Dasgupta
Professor, Dept. of Civil Engineering
Indian Institute of Technology, Kharagpur, India.
[email protected]

ABSTRACT
Wall used to retain soil of height more than 6.0 meter is often provided with
counterforts to economize its design. Behavior of such stiffened wall under
earthquake forces is still uncertain and practically no solution exists till date in terms
of their dynamic behavior except the dynamic pressure concept proposed by
Mononobe and Okabe (1924) based on pseudo static analysis. The seismic coefficient
method that forms the backbone of M-O method has though been obsolete for other
type of structures (where more rational techniques like modal and time history
analysis have been in vogue) has however continued to be in practice for the last 86
years, though found to give a considerable lower bound solution to the problem
(Ostadan 2004, Chowdhury et al 2007). The present paper is an attempt to re-visit
this long standing practical problem and proposes a more realistic analysis based on
modal response technique considering a generalized back filled soil having both c
and φ value with overburden- an oft faced condition in practice for which again M-O
method is deficient.
KEYWORDS: Amplitude, Counterforts, Modal analysis, Moment, Plate theory,
Retaining Wall, Shape functions, Shear, Weighted Residual Method.

INTRODUCTION
Retaining walls deployed to retain soil height more than 6 meters, normal practice is to
provide counterforts on the soil side to achieve economy in the overall design. Behavior of these

- 1329 -
 Vol. 16 [2011], Bund. R 1330

stiffened walls are still plagued by significant uncertainty, especially under earthquake and
prevailing practice is to use the dynamic pressure concept developed by Mononobe and Okabe
(1924, 1929) [M-O method] for its design.

In Mononobe and Okabe’s method, stiffness of the wall is ignored, as it is assumed to be a

gravity wall of infinite stiffness and peak ground acceleration is used to determine the dynamic
pressure on the wall. Seismic pressure thus obtained is pseudo static in nature and does not take
into cognizance stiffness vis-à-vis dynamic property (the time period effect) of the wall. This
trend has remained unaltered for years starting from Okabe (1924), Mononobe & Matsuo (1929),
Seed & Whitman (1970), Whitman et al (1990, 1991), Das & Puri (1996), to some recent
researches in India like Ghosh et al (2010,2008, 2007), Saran et al (1968, 2003), Choudhury et al
(2002,2004,2006) etc.

Shown in Fig. 1 are Gravity and RCC retaining walls with counterforts. A Gravity retaining
wall resists the lateral earth pressure by its weight in contrast to the cantilever and counterfort
retaining walls where the pressure is resisted by their bending action. It is apparent that
counterfort retaining walls are far more flexible and the basic assumption made in Mononobe and
Okabe’s method that the wall is infinitely stiff cannot be justified for analyzing counterfort
retaining walls. Thus the pseudo static approach used for determining the dynamic pressure on
flexible retaining walls under seismic load may not be justified in this case. The counterfort
retaining walls do have a finite stiffness, and this will influence the dynamic response of the walls
under earthquake force.

Figure 1: Gravity retaining walls and counterfort reinforced concrete

retaining wall used to retain soil

Chowdhury & Dasgupta (2003) derived a solution based on modal response technique for
such counterfort retaining wall considering an eigenvector basis of beam shape functions adapted
for plates based on improved Rayleigh-Ritz technique. However, the solution is valid only for
cohesion less dry backfill (c = 0) and cannot be used for other soil types like c-φ soil, partially
saturated back fill, effect of overburden to name some of the often faced conditions in reality.

Recent research carried out by Ostadan & White (1997), Ostadan (2004), Chowdhury &
Dasgupta (2007, 2011) has also shown that M-O based methods significantly under predict the
dynamic pressure under seismic loads, to the extent that nuclear regulatory board of USA has
now stopped using any of the M-O based methods for determining earth pressure for any of their
structures.

- 1330 -
 Vol. 16 [2011], Bund. R 1331

Present paper is an attempt to re-evaluate this long standing problem for which no solution exists
in practice and seek answer to many of the open issues as cited above.

PROPOSED METHOD
To start with we take simplest of the case as shown below. Shown in Fig. 2 is a counterfort
retaining wall with dry sandy backfill and the ground has no inclination like in Figure 1. It is to be
noted that the same can also be derived from a more general soil condition but has been
considered first for brevity and also to use it as a benchmark for more generalized cases that will
be taken up subsequently.

Figure 2: A Counterfort retaining wall with dry cohesionless backfill (c = 0)

While performing the analysis it is assumed here that:

1. Soil profile under active case is at incipient failure when the failure line makes an
angle α = 450 ± φ/2 as shown in Fig. 2;
2. Since soil profile is already under failed condition under static load, it will not induce any
stiffness to the overall dynamic response but will only contribute to the inertial effect;
3. Since the counterfort wall is relatively thin, mass contribution of the wall itself may be
ignored compared to that of failed soil mass. The wall thus contributes only to the stiffness
of overall soil-structure system;
4. The retaining wall is fixed at base and at the edge of counterforts;
5. Foundation compliance has been ignored for the present analysis;
6. The wall essentially behaves as a plate having three sides fixed and one side free as its
boundary condition.

It will be observed that most of the assumptions (except the last three) made above are
identical to what Mononobe or Steedman and Zeng (1990) has assumed in their analysis. Based
on above assumptions the analysis is carried out as elaborated hereunder.

- 1331 -
 Vol. 16 [2011], Bund. R 1332

Dynamic Response of Wall having Dry Cohesionless Back Fill (c = 0)

Considering the wall behaving as a two way slab of aspect ratio r = H/b and essentially
behaves as a plate element. Considering φ as internal angle of friction of soil mass, active and
passive coefficient of pressures KA and KP may expressed as
1 − sin φ 1 + sin φ
KA = ; and K P = (1)
1 + sin φ 1 − sin φ
Now imagining a strip of element of width dx, in vertical direction (marked by hatch line in
Fig. 2) and γs being the unit weight of soil mass, the static pressures acting on this strip of element
can be expressed as
pay = K A .γ s . y.dx (2)

p py = K P .γ s . y.dx (3)

For the strip considered as a cantilever beam fixed at the base slab, differential equation of
static equilibrium under active soil condition can be expressed as
d 4u
EI 4 = K A .γ s . y.dx (4)
dy
where u is displacement of the strip element, E = Young’s Modulus of the RCC wall and I is
moment of inertia of the strip considered I = (1 / 12).(dx × t 3 ) , here t is thickness of the wall (can
be considered as average thickness for variation between top and bottom thickness of the wall).

On successive integration of equation (4) one can have

d 3u K A .γ s y 2
EI = dx + C1 (5)
dy 3 2

d 2u K A .γ s y 3
EI 2 = dx + C1 y + C2 (6)
dy 6

du K A .γ s y 4 y2
EI = dx + C1 + C2 y + C3 (7)
dy 24 2

K A .γ s y 5 y3 y2
EIu = dx + C1 + C2 + C3 y + C4 (8)
120 6 2
For the given strip (Fig.2) we consider the boundary conditions
d 3u
1) At y = 0 = 0  C1=0 ( 9)
dy 3

d 2u
2) At y = 0 = 0  C2 = 0 (10)
dy 2

- 1332 -
 Vol. 16 [2011], Bund. R 1333

du K .γ .H 4
3) At y = H = 0  C3 = − A s dx (11)
dy 24

K A .γ s .H 5
4) At y = H u = 0  C4 = dx (12)
30
Thus equation (8) can be expressed as

K A .γ s y 5 K .γ H 4 y K γ H5
EIu = dx − A s dx + A s dx (13)
120 24 30
Equation (13) can be further expressed in generalized form as

2 K A .γ s .H 5 η 5 5η 
u= − + 1
5 Et 3  4
(14)
4 
where η = y / H a non dimensional term that varies between 0-1.

From equation (14) one can conclude that for a strip element having base fixed and the other
end free, the generic shape function of displacement can be expressed as

η5 5η
f (η ) = − +1 (15)
4 4
Proceeding in identical fashion it can be shown that in horizontal direction considering a strip
element of width dy having both end fixed and subjected to a uniformly distributed load w,
generic shape function for displacement in natural co-ordinate can be expressed as
f (ξ ) = 16ξ 4 − 32ξ 3 + 16ξ 2 (16)

where ξ = x / b a non dimensional term that varies between 0-1.

The shape functions derived in equation (15) and (16) will generically satisfy plate
equilibrium equation having boundary conditions of three side fixed and one side free but may
still have residual error(Re) as they are not derived from exact analysis of the fourth order partial
differential of an isotropic plate expressed as (Timoshenko 1987)

∂ 4u ∂ 4u ∂ 4u q
+ 2 + = (17)
∂x 4 ∂x 2∂y 2 ∂y 4 D
where u is displacement of the plate under a pressure load q. D is flexural stiffness of plate
expressed as D = Et 3 / 12(1 − ν 2 ) , and ν the Poison ratio.

In natural coordinate, as expressed in equation (15) and (16), Equation (17) can be expressed
as

1 ∂ 4u 2 ∂ 4u 1 ∂ 4u q
+ 2 2 + 4 = (18)
b ∂ξ
4 4
H b ∂ξ ∂η
2 2
H ∂η 4
D

- 1333 -
 Vol. 16 [2011], Bund. R 1334

where displacement u is expressed as u = Δ. f (ξ ). f (η ) (19)

Substituting equation (19) in (18) we have

Δ ∂4 2Δ ∂4 Δ ∂4 q
f (ξ ). f (η ) + f (ξ ). f (η ) + f (ξ ). f (η ) − = 0 (20)
b ∂ξ
4 4
H b ∂ξ ∂η
2 2 2 2
H ∂η
4 4
D
The shape functions f (ξ ) and f (η ) will generically satisfy equation (20) as they conform
to boundary condition of the given plate, however may have residual error Re which may be
expressed as ,

Δ ∂4 2Δ ∂4 Δ ∂4 q
Re = f (ξ ). f (η ) + f (ξ ). f (η ) + f (ξ ). f (η ) − (21)
b ∂ξ
4 4
H b ∂ξ ∂η
2 2 2 2
H ∂η
4 4
D
The residual error Re is now minimized by Galerkin’s (Chowdhury & Dasgupta 2008)
weighted residual method based on which
1 1
 Δ ∂4 2Δ ∂4 Δ ∂4 q
   b
0 0
4
∂ξ 4
f (ξ ). f (η ) + 2 2
H b ∂ξ ∂η
2 2
f (ξ ). f (η ) + 4
H ∂η 4
f (ξ ). f (η ) −  f (ξ ) f (η )dξdη = 0 (22)
D 

Equation (22) can thus be expressed as

Δ 2Δ
1 1 1 1

 f ' ' ' ' (ξ ) f (ξ ) f (η ) dξ .dη + 2 2   f ' ' (ξ ) f ' ' (η ) f (ξ ) f (η )dξ .dη +
2

b4 0 0 H b 0 0
(23)
Δ
1 1 1 1
1
 f (ξ ) f (η ) f ' ' ' ' (η )dξdη −   q f (ξ ) f (η )dξ .dη = 0
2

H4 0 0
D00

Considering q = K A .γ s . y equation (23) can be expressed in a simplified form as

H 4 2H 2  K .γ .H 5
Δ 4 X 1 + 2 X 2 + X 3  = A s X4 (24)
 b b  D
Considering r = H/b, the aspect ratio of the plate, equation (24) can be expressed as
K A .γ s .H 5
X4
Δ= 4 D (25)
r X 1 + 2r 2 X 2 + X 3
where X1, X2, X3 and X4 are integral functions that can be solved numerically or explicitly and
are as expressed in Table 1.

Table 1: Values of integral functions X1 to X5 for cohesion less soil(c=0)

Integral Functions Values

1 1
∂ 4 
X1 =    ∂ξ
0 0
4
f (ξ ) f (ξ ) f (η ) 2 dξdη

55.4113

- 1334 -
 Vol. 16 [2011], Bund. R 1335

1 1
 ∂2  ∂2 
X2 = 
0 0
 2 f (ξ )  2 f (η ) f (ξ ) f (η )dξdη
 ∂ξ   ∂η 
-0.6772

1 1
 ∂4 
X3 =   4 f (η ) f (η ) f (ξ ) dξdη
2
1.4512
0
 ∂η
0 
1 1
X4 =  η f (η ) f (ξ )dξdη
0 0
0.0635

The displacement u can thus be expressed as

K A γ s .H 5  X4 
u=  4  f (ξ ). f (η ) (26)
 r X 1 + 2r X 2 + X 3 
2
D
Here f(ξ) and f(η) are as expressed in equations (16) and (15) respectively.

The maximum static deflection will occur at ξ = 0.5 and η = 0 which gives

K A γ s .H 5  X4 
u static =  4  (27)
 r X 1 + 2r X 2 + X 3 
2
D
Time period of the wall for the fundamental mode which is most critical, can be expressed as

ustatic
T = 2π (28)
g
Substituting equation (27) in (28) under active failure we have

K Aγ s H 5  X4 
T A = 2π  4  (29)
 r X 1 + 2r X 2 + X 3 
2
Dg
For modal analysis, maximum amplitude (Sd) is expressed as (Clough, 1984)
Sa
Sd = (30)
ω2
where Sa= Spectral acceleration and ω = 2π / T , natural frequency of the wall. In terms of Code,
equation (30) can be expressed
Sa
S d = κβ (31)
ω2
n n
where κ = Modal mass participation factor and is expressed as  miϕi /  miϕi 2 ; β = a code
i =1 i =1
factor expressed as ZI/2R where Z= zone factor; I= importance factor and R= response reduction
factor.
Thus based on equation (31), the dynamic amplitude of the wall can be expressed as

- 1335 -
 Vol. 16 [2011], Bund. R 1336

Sa
u d = κβ T A f (ξ ). f (η )
2
(32)
4π 2
Equation (32) can be finally expressed as
K Aγ s H 5  Sa   2 η 5η 
[ ]
5
X4
u d = κβ    4  16ξ − 32ξ + 16ξ  −
4 3
+ 1 (33)
  r X 1 + 2r X 2 + X 3 
2
D  g 4 4 
Equation (33) gives a complete dynamic displacement profile of the wall between limits 0-1
in both X and Y direction.

2.1.1 Dynamic Bending Moment and Shear Force

Bending moment and shear force induced in the wall can be expressed as (Timoshenko,
1987)

 ∂ 2u ∂ 2u 
M x = −D 2 +ν 2  (34)
 ∂x ∂y 

 ∂ 2u ∂ 2u 
M y = − D ν 2 + 2  (35)
 ∂x ∂y 

 ∂ 3u ∂ 3u 
Qx = − D  3 +  (36)
 ∂x ∂x.∂y 2 

 ∂ 3u ∂ 3u 
Qy = −D 2 + 3  (37)
 ∂x ∂y ∂y 
Transferring the above equations in natural co-ordinate and substituting in equation (33),
dynamic moments and shears can be calculated using following equations. To find out moments
and shear, η is replaced with (1- η) to keep the reference coordinate identical to that of the
displacement, which gives
 S   2
M x = κ .β .K A .γ s .H 3  a   4
X4
[
 r f ' ' (ξ ) f (1 − η ) + νf (ξ ) f ' ' (1 − η ) ] (38)
 g   r X 1 + 2r X 2 + X 3 
2

 S  
M y = κ .β .K A .γ s .H 3  a   4
X4
[
 f (ξ ) f ' ' (1 − η ) + νr f ' ' (ξ ) f (1 − η )
2
] (39)
 g   r X 1 + 2r X 2 + X 3 
2

 S   3
Q x = κ .β .K A .γ s .H 2  a   4
X4
[
 r f ' ' ' (ξ ) f (1 − η ) + rf ' (ξ ) f ' ' (1 − η ) ] (40)
 g   r X 1 + 2r X 2 + X 3 
2

 S   2
Q y = κ .β .K A .γ s .H 2  a   4
X4
[
 r f ' ' (ξ ) f ' (1 − η ) + f (ξ ) f ' ' ' (1 − η ) ] (41)
 g   r X 1 + 2r X 2 + X 3 
2

The modal participation factor κ can be expressed in this case as

- 1336 -
 Vol. 16 [2011], Bund. R 1337

n 1

m ϕ i i K A γ s H 2η f (ξ ) f (η ) dξdη
κ= i =1
n
= 1
0
(42)
m ϕ K
2
i i A γ s H η f (ξ ) f (η ) dξ dη
2 2 2

i =1 0

1 1
 η 5 5η 
(
0 0 η 16ξ − 32ξ + 16ξ  4 − 4 + 1dξ .dη
4 2 2
)
κ= 2
(43)
2 2 η 5η 
1 1
( )
5

0 0 η. 16ξ − 32ξ + 16ξ  4 − 4 + 1 dξ .dη

4 2

Equation (43) on computation gives a value of κ = 2.8

For passive pressure case replace KA by KP in equations (26) through (43) as expressed in
equation (1).
Dynamic amplitude, bending moment and shear can also be expressed in a simplified form
for quick computation as
K Aγ s H 5  S a 
u d = Coeff d × κ .β g (43a)
D  
S 
M x = Coeff mx × κ .β .K A .γ s .H 3  a  (43b)
g
S 
M y = Coeff my × κ .β .K A .γ s .H 3  a  (43c)
g
S 
Q x = Coeff qx × κ .β .K A .γ s .H 2  a  (43d)
g
S 
Q y = Coeff qy × κ .β .K A .γ s .H 2  a  (43e)
g
Typical values of such design coefficients for aspect ratio H/b = 1.5 and Poisson’s ratio ν =
0.25 are furnished in Table 2 to 6.

Table 2: Coeffd for estimation of amplitude for various values of η(vert.) and ξ (horz.)

η.ξ 0.0 0.2 0.4 0.5 0.6 0.8 1

0.0 0.00000 0.00009 0.00021 0.00023 0.00021 0.00009 0.00000
0.2 0.00000 0.00007 0.00016 0.00017 0.00016 0.00007 0.00000
0.4 0.00000 0.00005 0.00011 0.00011 0.00011 0.00005 0.00000
0.5 0.00000 0.00004 0.00008 0.00009 0.00008 0.00004 0.00000
0.6 0.00000 0.00003 0.00006 0.00006 0.00006 0.00003 0.00000
0.8 0.00000 0.00001 0.00002 0.00002 0.00002 0.00001 0.00000
1.0 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000

- 1337 -
 Vol. 16 [2011], Bund. R 1338

Table 3: Coeffmx for estimation of Mx for various values of η and ξ

η.ξ 0 0.2 0.4 0.5 0.6 0.8 1

0.0 0.00000 0.00012 0.00026 0.00028 0.00026 0.00012 0.00000
0.2 0.00134 0.00011 -0.00046 -0.00053 -0.00046 0.00011 0.00134
0.4 0.00442 0.00020 -0.00189 -0.00215 -0.00189 0.00020 0.00442
0.5 0.00000 0.00012 0.00026 0.00028 0.00026 0.00012 0.00000
0.6 0.00824 0.00034 -0.00361 -0.00410 -0.00361 0.00034 0.00824
0.8 0.01230 0.00049 -0.00541 -0.00615 -0.00541 0.00049 0.01230
1.0 0.01639 0.00066 -0.00721 -0.00820 -0.00721 0.00066 0.01639

Table 4: Coeffmy for estimation of My for various values of η and ξ

η, ξ 0 0.2 0.4 0.5 0.6 0.8 1

0.0 0.00000 0.00047 0.00105 0.00114 0.00105 0.00047 0.00000
0.2 0.00034 0.00025 0.00039 0.00041 0.00039 0.00025 0.00034
0.4 0.00110 0.00014 -0.00026 -0.00031 -0.00026 0.00014 0.00110
0.5 0.00000 0.00047 0.00105 0.00114 0.00105 0.00047 0.00000
0.6 0.00206 0.00011 -0.00084 -0.00096 -0.00084 0.00011 0.00206
0.8 0.00307 0.00013 -0.00134 -0.00153 -0.00134 0.00013 0.00307
1.0 0.00410 0.00016 -0.00180 -0.00205 -0.00180 0.00016 0.00410

Table 5: Coeffqx for estimation of Qx for various values of η and ξ

η, ξ 0 0.2 0.4 0.5 0.6 0.8 1

0.0 0.00000 0.00525 0.00262 0.00000 -0.00262 -0.00525 0.00000
0.2 -0.01209 -0.01423 -0.02041 -0.02417 -0.02793 -0.03411 -0.03626
0.4 -0.03975 -0.05452 -0.07098 -0.07950 -0.08801 -0.10448 -0.11925
0.5 0.00000 0.00525 0.00262 0.00000 -0.00262 -0.00525 0.00000
0.6 -0.07414 -0.10346 -0.13328 -0.14828 -0.16328 -0.19310 -0.22242
0.8 -0.11066 -0.15488 -0.19916 -0.22131 -0.24346 -0.28775 -0.33197
1.0 -0.14752 -0.20653 -0.26554 -0.29505 -0.32455 -0.38356 -0.44257

Table 6: Coeffqy for estimation of Qy for various values of η and ξ

η, ξ 0.0 0.2 0.4 0.5 0.6 0.8 1

0.0 0.00000 0.00140 0.00315 0.00341 0.00315 0.00140 0.00000
0.2 -0.01210 0.00041 0.00734 0.00823 0.00734 0.00041 -0.01210
0.4 -0.01783 -0.00021 0.00898 0.01015 0.00898 -0.00021 -0.01783
0.5 0.00000 0.00140 0.00315 0.00341 0.00315 0.00140 0.00000
0.6 -0.01996 -0.00057 0.00929 0.01053 0.00929 -0.00057 -0.01996
0.8 -0.02046 -0.00076 0.00913 0.01036 0.00913 -0.00076 -0.02046
1.0 -0.02049 -0.00082 0.00902 0.01024 0.00902 -0.00082 -0.02049

- 1338 -
 Vol. 16 [2011], Bund. R 1339

Similar design charts can be developed for various values of aspect ratio (H/b) for quick
computation of design moments and shears in a design office environment.

2.2 Soil inclined at an angle i with the horizontal

In this case, as shown in Fig. 3, expressions presented vide equations (26) to (43) remains
valid except that in this case the active and passive earth pressures coefficients
are expressed as

cos i − cos 2 i − cos 2 φ

K A = cos i × (44)
cos i + cos 2 i − cos 2 φ

cos i + cos 2 i − cos 2 φ

K P = cos i × (45)
cos i − cos 2 i − cos 2 φ

Fig. 3: Retaining wall with Fig. 4: Retaining wall with general c-φ soil as backfill
inclined backfill at an angle i

2.3 Dynamic Response of Wall with c- Backfill

For a c-φ soil the active earth pressure is expressed [Murthy (1984)] as
pa = γ s z cot 2 α − 2c cot α where, α = 450+φ/2 (46)

Shown in Fig. 4 is a retaining wall with pressure induced on it as per equation (46) for c-φ
soil.
It is apparent from the figure that due to cohesion a portion of soil of height (2c / γ s ) tan α
cracks and separates away from the wall, static pressure due to this portion (as per practice) is
ignored and the wall pressure is derived based on the net area shown by the hatched line in Fig. 4.

For the above case displacement at point O in Fig. 4 where the pressure is zero is expressed
as

- 1339 -
 Vol. 16 [2011], Bund. R 1340

5
 2c 
K Aγ s . H − tan α 
 γs   X4 
uo =  4  (47)
 r X 1 + 2r X 2 + X 3 
2
D
It is to be noted that in this case f(ξ) and f(η) is as expressed in equations (16) and (15). The
integral functions X1, X2, X3 and X4 are as furnished in Table 1.
2c tan α
Considering a dimensionless term, μ = . (48)
γ sH
Maximum deflection at the tip of the wall is expressed as

K Aγ s .H 5 (1 − μ )    5  μ 
5
X4
u max =  4  1 +  
 r X 1 + 2r X 2 + X 3   4  (1 − μ ) (49)
2
D

Time period of the wall can now be expressed as

K Aγ s H 5 .(1 − μ )    5  μ 
5
X4
T A = 2π  4  1 +  
 r X 1 + 2r X 2 + X 3   4  (1 − μ ) 
2
Dg
(50)
Considering equation (32) the dynamic amplitude can now be expressed as

κ β K A γs H 5 ( 1- μ )  Sa     5  μ 
5
X4
ud =    4  1+   f(ξ)f(η) (51)
 g   r X 1 + 2r X 2 + X 3   4  ( 1- μ ) 
2
D

Dynamic bending moment and shear force are expressed as

3  S    5  μ  2
M x = κβK Aγ s H 3 (1 − μ )  a   4
X4
 1 +  [
 r f ' ' (ξ ) f (1 − η ) + νf (ξ ) f ' ' (1 − η ) ] (52)
 g   r X 1 + 2r X 2 + X 3   4  (1 − μ ) 
2

 S    5  μ 
M y = κβK Aγ s H 3 (1 − μ ) 3  a   4
+
X4
2
+
 1 + 
( − μ )
[
 f (ξ ) f ' ' (1 − η ) + νr f ' ' (ξ ) f (1 − η )
2
] (53)
 
g r X 1 2 r X 2 X 3  4  1 

2  S    5  μ  3
Qx = κβK Aγ s H 2 (1 − μ )  a   4
X4
 1 +  [
 r f ' ' ' (ξ ) f (1 − η ) + rf ' (ξ ) f ' ' (1 − η ) ] (54)
 g   r X 1 + 2r X 2 + X 3   4  (1 − μ )
2

 S    5  μ  2
Q y = κβK Aγ s H 2 (1 − μ ) 2  a   4
+
X4
2
+
 1 +  [
 r f ' ' (ξ ) f ' (1 − η ) + f (ξ ) f ' ' ' (1 − η )
4  (1 − μ )
] (55)
 
g r X 1 2 r X 2 X 3 

In this case design coefficients furnished in Table 2 also could be used provided r = 1.5 and ν =
0.25.

The modal mass participation may be expressed as

1 1

μ μ (
η 5η
) 
5
η 16ξ 4 − 32ξ 2 + 16ξ 2  − + 1dξ .dη
4 4  
κ= 1 1 2 (56)
2 2 
μ μ η.(16ξ ) η 5η
5

4
− 32ξ + 16ξ
2
 4 − 4 + 1 dξ .dη
 

- 1340 -
 Vol. 16 [2011], Bund. R 1341

Values of modal participation factor κ for various values of μ varying from 0 to 0.6 are as
furnished in Table 7. Intermediate values can be linearly interpolated.

Table 7: Values of Modal Participation factor

Sl. No. μ κ
1 0 2.8
2 0.1 2.89
3 0.2 3.086
4 0.3 3.45
5 0.4 4.205
6 0.5 5.841
7 0.6 9.785

The above derivation is for a general soil that has finite value of c and φ. When the soil is
purely cohesion less i.e. c = 0, μ → 0 equations (51) to (56) degenerates to equations (38) to
(43). This shows the correctness of the derivation of the above expressions.
For passive case the earth pressure coefficient KA to be replaced by KP. For the case of pure intact
clay (φ = 0) value μ = 2c/γsH when α = 450, with all other terms above remains the same.

Dynamic response of wall with c-φ backfill and surcharge q

This problem is often faced by engineers in brown field industrial plants and construction in
dense urban region and remains a serious problem under earthquake. No solution exists till date
for this problem and engineers have to often resort to Finite element analysis (FEM) to arrive at a
workable solution.

Shown in Figure 5 is a retaining wall with generic c-φ soil as the backfill, and also having a
surcharge q on the top. In horizontal direction the shape function f(ξ) remains as
f (ξ ) = 16ξ 4 − 32ξ 3 + 16ξ 2 , while in vertical direction it has been shown by Chowdhury &
Dasgupta (2011) that the shape function is expressed as

η 5 5η η4 4ξ ζη 4 4ζη 
f (η ) =  − + 1 −ψ +ψ −ψ + − +ζ  (57)
4 4 3 3 3 3 

- 1341 -
 Vol. 16 [2011], Bund. R 1342

Figure 5: A counterfort retaining wall with c-φ backfill and surcharge load q

15 H c 15q
In equation (57), ψ = tan α and ζ = are both dimensionless parameters and
4H 4γ s H
H c = 2c / γ s . For analysis of this case it is assumed that the overburden pressure is sufficiently
high so that the top portion of soil does not separate away from the wall and remains in contact
with the wall, in other words ζ > ψ .

Based on above, the modified Galerkin’s equation can be expressed for this case as
Δ 2Δ
1 1 1 1

b4 
0 0
f ' ' ' ' (ξ ) f (ξ ) f (η ) 2 dξ .dη +
H 2b 2   f ' ' (ξ ) f ' ' (η ) f (ξ ) f (η )dξ .dη +
0 0 (58)
Δ
1 1
K γ H5 1 1
K A qH 4 1 1

H4 
0 0
f (ξ ) 2 f (η ) f ' ' ' ' (η )dξdη − A s
D 0 0 η f (ξ ) f (η ) dξ .dη −
D   f (ξ ) f (η )dξdη = 0
0 0

H4 2H 2  K .γ .H 5  q 
 Δ 4 X1 + 2 X 2 + X 3  = A s X4 + X5 (59)
b b  D  γ sH 
Considering u= Δ f(ξ).f(η), the static deflection is thus expressed as

K Aγ s .H 5  X 4 + λX 5 
u=  4  f (ξ ). f (η ) (60)
 r X 1 + 2r X 2 + X 3 
2
D

q
Where λ = and r=H/b the aspect ratio.
γ sH
In this case the integral functions, X1, X2, X3, X4 and X5 are not unique, but are functions of
both ψ and ζ. The explicit values of these integral are furnished in Table 8.

Table 8: Explicit values of integrals X1 through X5 for general c-φ soil with overburden

Integral Functions Values

- 1342 -
 Vol. 16 [2011], Bund. R 1343

1 1
 ∂4 
X1 =   4 f (ξ ) f (ξ ) f (η ) dξdη
2
52.59(ψ 2 + ζ 2 ) + 107.93(ζ − ψ ) − 105.18ψ .ζ + 55.41
0

0
∂ξ 

1 1
 ∂2  ∂2 
X2 =   2 f (ξ )  2 f (η ) f (ξ ) f (η )dξdη ( )
− 0.677 − 0.93 ψ 2 + ζ 2 + 1.858ψ .ζ + 1.625(ψ − ζ )
0 0  ∂ξ   ∂η 
1 1
 ∂4 
X3 =   4 f (η ) f (η ) f (ξ ) dξdη
 ∂η 
2
(
1.451 − 2.6ψ .ζ + 2.709(ζ − ψ ) + 1.3 ψ 2 + ζ 2 )
0 0 
1 1
X4 =  η f (η ) f (ξ )dξdη
0 0
0.0635 + 0.06(ζ − ψ )

1 1
X5 = 
0 0
f (η ) f (ξ )dξdη 0.222 + 0.213(ζ − ψ )

Maximum static deflection is expressed at ξ = 0.5 and η = 0 as

K Aγ s .H 5  X 4 + λX 5 
umax =  4  (1 −ψ + ζ ) (61)
 r X 1 + 2r X 2 + X 3 
2
D
The time period is thus expressed as

K Aγ s H 5  X 4 + λX 5 
TA = 2π  4 [1 − ψ + ζ ] (62)
Dg  rc X 1 + 2rc X 2 + X 3 
2

The dynamic amplitude is then expressed as

κβK Aγ s .H 5  S a   X 4 + λX 5 
udyn =  g   r 4 X + 2r 2 X + X  (1 − ψ + ζ ) f (ξ ) f (η ) (63)
D   1 2 3

Dynamic bending moment and shear force are expressed as

 S  X 4 + λX 5  2
M x = κβK Aγ s H 3 (1 −ψ + ζ ) a   4 [
 r f ' ' (ξ ) f (1 − η ) + νf (ξ ) f ' ' (1 − η ) ] (64)
 g   r X 1 + 2r X 2 + X 3 
2

 S  X 4 + λX 5 
M y = κβK Aγ s H 3 (1 − ψ + ζ )  a   4 [
 f (ξ ) f ' ' (1 − η ) + νr f ' ' (ξ ) f (1 − η )
2
] (65)
   r X 1 + 2r X 2 + X 3 
2
g

 S  X 4 + λX 5  3
Q x = κβK Aγ s H 2 (1 − ψ + ζ ) a   4 [
 r f ' ' ' (ξ ) f (1 − η ) + rf ' (ξ ) f ' ' (1 − η ) ] (66)
 g   r X 1 + 2r X 2 + X 3 
2

 S  X 4 + λX 5  2
Q y = κβK Aγ s H 2 (1 − ψ + ζ )  a   4
+ 2
+
[
 r f ' ' (ξ ) f ' (1 − η ) + f (ξ ) f ' ' ' (1 − η ) ] (67)
 
g r X 1 2 r X 2 X 3

Equations (63) to (67) give the displacement, moment and shear under seismic loading for the
most general condition of soil. When there is no overburden i.e.,η → 0 , the formulas converges

- 1343 -
 Vol. 16 [2011], Bund. R 1344

to the case of c-φ soil only. When again,ψ → 0 the equations converges to the case of pure
cohesion less soil(c=0).

Finally an interesting case is observed, when,ψ → 0 ,η → 0 and even = 0(i.e. K A → 1) ,

the values obtained in equations (63) to (67) are not zero and finite and converges to a hydrostatic
case. Thus if we take density of soil in this particular case ( sat- w) it reflects a case when the
soil has undergone liquefaction.

The modal mass participation factor in this case is expressed as

1 1 1 1

 ηf (ξ ) f (η )dξdη +  λf (ξ ) f (η )dξdη

κ= 0 0
1 1 1 1
0 0
(68)

ηf (ξ ) f (η ) +  λf (ξ ) f (η ) dξdη
2 2 2 2

0 0 0 0

Explicit value of the above integral in terms of ψ, ζ and λ is expressed as

0.0635 + 0.0593(ζ −ψ ) + 0.222λ + 0.213λ (ζ −ψ ) (69)

κ=
(
0.0226 + 0.0432(ζ −ψ ) + 0.0206 ψ 2 − ζ 2
)− 0.0412ψ .ζ + 0.1099λ − 0.2086λψζ + 0.1043λ (ψ 2
+ζ 2
)+ 0.214λ (ζ −ψ )

Dynamic response of wall with c-φ backfill partially

submerged below water
Shown in Fig. 6 is a retaining wall having partially submerged c-φ soil as backfill. The soil is
partially submerged to a height of H2. In this case, considering the complex nature of the pressure
diagram it would become difficult to perform the integration with appropriate boundaries. As
such one can approach the problem differently for this particular case.

Figure 6: A counterfort retaining wall with c-φ backfill partially submerged

Considering the first step in the process is to determine the static deflection and this is
obtained based on the loading on the retaining wall, the net horizontal loading on the wall can be
expressed as [Murthy (1984)].

- 1344 -
 Vol. 16 [2011], Bund. R 1345

PA =
1
2
[ ]
cot 2 α γ s H 1 2 + (γ sat − γ w )H 2 2 − 2c cot αH + γ s H 1 .H 2 cot 2 α +
2c 2 1
γs
+ γ w H 2 2 (70)
2

where γ s = Dry unit weight of soil of height H1

γ sat = Saturated unit weight of soil of height H2
γ w = unit weight of water.

Now if we consider an equivalent dry back fill of density γs which imposes the same load on
the wall the displacement of the wall will be same as that as would be induced by load as
expressed in equation (70).

Thus considering
1
2
1
[ ]
K AE γ s H 2 = cot 2 α γ s H 1 + (γ sat − γ w )H 2 − 2c cot αH + γ s H 1 .H 2 cot 2
2
2 2

(71)
2c 2 1
+ + γ wH2
2

γs 2

We have

K AE = cot 2 α
[γ Hs 1
2
+ (γ sat − γ w )H 2
2
]− 4c cot α
+
2 H 1 .H 2 cot 2 α
γ s (H 1 + H 2 ) 2
γ s .(H 1 + H 2 ) (H1 + H 2 ) 2 (72)
4c 2
γ wH2 2

+ +
γ s (H1 + H 2 )
2 2
γ s (H1 + H 2 ) 2

where KAE= An equivalent coefficient of active earth pressure when considering a pressure
diagram of pay = K AE .γ s . y over the height H will give same deflection as that produced by PA in
equation (70).

Thus for the present case the problem now gets simplified considerably when we have

K AE γ s H 5  X4 
T A = 2π  4  (73)
 r X 1 + 2r X 2 + X 3 
2
Dg

K AE γs H 5  Sa  X4  2 η
5
5η 
ud = κβ  
 4  16ξ 4
- 32ξ 3
+16ξ  - +1 (74)
2  
D  g   r X 1 + 2r X 2 + X 3  4 4 
 S   2
. M x = κ .β .K AE .γ s .H 3  a   4
X4
[
 r f ' ' (ξ ) f (1 − η ) + νf (ξ ) f ' ' (1 − η ) ] (75)
 g   r X 1 + 2r X 2 + X 3 
2

- 1345 -
 Vol. 16 [2011], Bund. R 1346

 S  
M y = κβK AE γ s H 3  a   4
X4
[
 f (ξ ) f ' ' (1 − η ) + νr f ' ' (ξ ) f (1 − η )
2
] (76)
 g   r X 1 + 2r X 2 + X 3 
2

 S   3
Q x = κ .β .K AE .γ s .H 2  a   4
X4
[
 r f ' ' ' (ξ ) f (1 − η ) + rf ' (ξ ) f ' ' (1 − η ) ] (77)
 g   r X 1 + 2r X 2 + X 3 
2

 S   2
Q y = κ .β .K AE .γ s .H 2  a   4
X4
[
 r f ' ' (ξ ) f ' (1 − η ) + f (ξ ) f ' ' ' (1 − η ) ] (78)
 g   r X 1 + 2r X 2 + X 3 
2

The modal participation factor κ will remain same as expressed by Equation (43) cited earlier.

VALIDATION OF THE THEORY

Using a polynomial shape function that conforms to the boundary condition of a differential
equation (linear or partial), and minimizing the residual error for solution of this equation based
on various weighted residual techniques (like Sub domain, Galerkin, Least square etc) has been
an acceptable and established technique in realms of numerical methods in engineering,
especially by finite element method.

For rectangular plate elements assuming polynomial functions based on Pascal triangle
(Szilard 2004) or an approximate assumed boundary satisfying shape function and applying
weighted residual technique has been often used to derive approximate solution of a plate element
(Chandrashekara 2001).

Tocher (1962) first developed an approximate solution based on non conforming polynomial
shape function, the stiffness of a rectangular plate element. This was followed by Adini, Clough
and Melosh (1964) who derived an explicit expression of stiffness matrix for a general plate
element for finite element analysis.

The authors are however yet to come across any work where shape functions have been
derived based on explicit boundary conditions of plates considering a strip of beam element
having similar boundary conditions and then subsequently using it in a plate equation to arrive at
a solution based on Galerkin’s technique of minimizing the error as presented herein.

So naturally question arise as to how accurate is the technique?

To this end a simply supported plate subjected to a load w is solved by the present technique
and the displacement coefficient thus obtained is compared to the analytical solution as cited in
Timoshenko (1987).

For a simply supported plate considering the shape functions as (shape functions derived
from solution of differential equation of a simply supported beam by identical procedure as cited
at the outset)
16 4 32 3 16
f (ξ ) = ξ − ξ + ξ (79)
5 5 5

- 1346 -
 Vol. 16 [2011], Bund. R 1347

16 4 32 3 16
f (η ) = η − η + η (80)
5 5 5
The static deflection is expressed as

w.H 4  X4 
u=  4  f (ξ ). f (η ) (81)
 r X 1 + 2r X 2 + X 3 
2
D

The above can be expressed as

w.H 4
u = Coeff × (82)
D

 X4 
where Coeff =  4  f (ξ ). f (η ) (83)
 r X 1 + 2r X 2 + X 3 
2

The coefficient thus derived above is compared to Timoshenko’s data at ξ=0.5 and η=0.5 and
are shown hereafter for different aspect ratio of the plate.

Coefficient of displacement-simply supported plate vs

proposed method

0.012
0.01
Coefficient

0.008
Proposed coefficient
0.006
Timoshenko
0.004
0.002
0
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Aspect Ratio

Figure 7:_Comparison of coefficient of displacement proposed versus analytical

solution by Timoshenko.

From Fig. 7 it is observed that the values are in excellent agreement. Maximum variation is
about 5% at r = 1.9 and minimum is 1.8% at r = 1.0. Thus it can be concluded that considering its
accuracy for this benchmark problem the proposed technique should also be valid for other
boundary conditions like free, fixed simply supported etc and suffice for practical application as
presented in this case.

- 1347 -
 Vol. 16 [2011], Bund. R 1348

Efficiency of the method lies in using such higher order polynomial is the minimum
computational effort required to arrive at an acceptable solution of sufficient accuracy without
having to resort to finite element analysis or tedious computation with trigonometric functions
that are encountered in classical analysis. The solution can well be carried out in a spread sheet or
generic utility software like MATHCAD which are common computational tool available in any
design office environment.

RESULTS AND DISCUSSIONS

To evaluate how the procedure works, an 8.0 meter high retaining wall with following
structural and soil properties is analyzed under earthquake force:
Height of wall = 8.0 meter; Thickness of wall = 0.25m; Counterfort spacing = 6.0m;
Thickness of counterfort = 0.4m; Grade of concrete = M30, Foundation width =4.0m; Foundation
thickness = 0.6m; Density of backfill = 22 kN/m3; Angle of internal friction = 28o when
considered as cohesionless soil (c = 0) and 10 degree for c-φ soil; Cohesive strength (c) = 10
kN/m2; Overburden (q) = 50 kN/m2. Earthquake zone = Zone IV as per IS-1893; Damping of soil
considered (average) =15%.
The results are worked out for following 7 cases:

1. When the soil is cohesion less with no overburden

2. When the soil is pure clay with no overburden
3. When both cohesion(c) and friction(φ) is present with no overburden
4. General c-φ soil with overburden(q)
5. φ soil with overburden
6. c soil with overburden
7. Soil liquefied under earthquake force (c & φ both considered zero) with overburden.
Dynamic parameters affecting the response for the seven load cases as mentioned above are
as presented in Table 9 below:

Table 9: Design dynamic parameters for seven load cases

Load Time Period Mode participation (Sa/g) for 15% Design acceleration (Sa), g [Sa =
Case (s) (κ) damping (ZI/2R)*(Sa/g)*g]
1 0.094 2.800 1.686 0.101g
2 0.124 2.910 1.750 0.105g
3 0.100 2.947 1.746 0.105g
4 0.185 1.517 1.750 0.105g
5 0.133 1.146 1.750 0.105g
6 0.221 1.442 1.750 0.105g
7 0.194 0.766 1.750 0.105g

It is observed from above that the time period and mass participation are different for each of
the seven cases. This is attributable to the soil parameters like c,φ and γs that heavily affect the
response. For load case 4 where c-φ soil is subjected to overburden the shape functions do not
remain independent of the soil properties and influence the response significantly. In other words
the analysis reflects a realistic interaction between the soil and wall as to how it affects the overall
behavior. It clearly shows that using M-O method – which again is only valid for sandy soil(c=0)

- 1348 -
 Vol. 16 [2011], Bund. R 1349

is far too simplistic and does not reflect the effect of time period as well as the plate behavior of
the wall which the proposed method does.

Typical displacement and dynamic moments and shear plots in 3D with colored contours for
Load case-1 are shown in Figures 8 to 10. Figs. 8 to 10 depict the dynamic response of the wall in
fundamental mode for φ−soil with no overburden. It is observed that considering the slab
behaving as a two way panel the major load is getting transferred to the shorter span which is
horizontal- this is an acceptable behavior of two way plates. It is apparent that for a counterfort
wall the aspect ratio of the wall (H/b) will heavily influence its dynamic response other than the
soil parameters as mentioned earlier. Finally unlike M-O method one need not use a suggestive
center of application of load either for the active or passive cases. These are automatically catered
for in applying the plate equation while deriving the dynamic moments and shears.

Based on above argument it may be concluded that the analytical method proposed is more
realistic. Figure (11) to (14) shows plots of Moments and shears at the center of the wall and at
the base of the wall where they are most critical.

Figure 8: Dynamic amplitude plot of the wall for Load case-1

- 1349 -
Vol. 16 [2011], Bund. R 1350

Figure 9: Dynamic Moments Mx and My of the wall for Load case-1

Figure 10: Dynamic Shears Qx and Qy of the wall for Load case-1

Mx (kN-m/m) at center of counterfort

-10 0 10 20 30 40 50 60 70
Normalized height of counterfort

0
0.1 Case 1
0.2 Case 2
0.3
Case 3
0.4
Case 4
0.5
0.6 Case 5

0.7 Case 6
0.8 Case 7
0.9
1

(a)

- 1350 -
 Vol. 16 [2011], Bund. R 1351

My (kN-m/m) at center of counterfort

-15 -10 -5 0 5 10 15 20
0
Normalized height of counterfort

0.1 Case 1
0.2 Case 2
0.3
Case 3
0.4
Case 4
0.5
0.6 Case 5

0.7 Case 6
0.8 Case 7
0.9
1

(b)

Figure 11: Dynamic Moments: (a). Mx and (b). My at the center of wall

80
60
Mx (kN-m/m) at bottom edge

40
20 Case 1
0 Case 2
-20 0 0.2 0.4 0.6 0.8 1 Case 3
-40 Case 4
-60 Case 5

-80 Case 6

-100 Case 7

-120
-140
Normalized width of counterfort

- 1351 -
Vol. 16 [2011], Bund. R 1352

(a)

20
15
My (kN-m/m) at bottom edge

10
5 Case 1
0 Case 2
-5 0 0.2 0.4 0.6 0.8 1 Case 3
-10 Case 4
-15 Case 5
-20 Case 6
-25 Case 7
-30
-35
Normalized width of counterfort

(b)

Figure 12: Dynamic Moments: (a). Mx and (b). My at bottom edge of wall

Qx (kN/m) at center of counterfort

-3.00E-15
0
Normalized height of counterfort

0.1
Case 1
0.2
Case 2
0.3
Case 3
0.4
0.5 Case 4

0.6 Case 5

0.7 Case 6
0.8 Case 7
0.9
1

(a)

- 1352 -
Vol. 16 [2011], Bund. R 1353

Qy (kN/m) at center of counterfort

-12 -10 -8 -6 -4 -2 0
0

Normalized height of counterfort

Case 1 0.1

Case 2 0.2
0.3
Case 3
0.4
Case 4
0.5
Case 5 0.6
Case 6 0.7

Case 7 0.8
0.9
1

(b)

Figure 13: Dynamic Shear: (a.) Qx and (b). Qy at the center of wall

150

100
Qx (kN/m) at bottom edge

Case 1
50
Case 2
Case 3
0
Case 4
0 0.2 0.4 0.6 0.8 1
Case 5
-50
Case 6

-100 Case 7

-150
Normalized width of counterfort

- 1353 -
 Vol. 16 [2011], Bund. R 1354

(a)

15

10
Qy (kN/m) at bottom edge

5 Case 1
Case 2
0
Case 3
0 0.2 0.4 0.6 0.8 1
Case 4
-5
Case 5
-10 Case 6
Case 7
-15

-20
Normalized width of counterfort

(b)

Figure 14: Dynamic Shear: (a). Qx and (b). Qy at bottom edge of wall

It clearly shows that soil with overburden significantly enhances the dynamic pressure and
engineers need to be careful on this issue while designing such retaining walls under seismic
force. The most critical of the cases are intact clay- subjected to overburden and φ soil with
overburden and high ground water (making the soil susceptible to liquefaction).In such case
though the structure inducing the overburden will ultimately settle/collapse due to liquefaction,
yet can induce a sufficiently high transient pressure that makes the wall slide or fail leading to a
larger disaster.

CONCLUSION AND REMARKS

A compreshensive analytical solution based on modal analysis is proposed herein that takes
into cognizance the effect of time period of the wall considering its actual plate behaviour, the
effect of which, is mostly ignored in research as well as in practise. It also gives solution to
almost all types of soil and loading conditions that may be expected in a real field design-
including liquefaction whose effect on the wall surely needs more research. It clearly reflects that
using the gravity wall based M-O model concept is far too simplistic and does not reflect the
actual condition as exists in reality especially for counterfort retaining walls whose behaviour is
completely different.

- 1354 -
Vol. 16 [2011], Bund. R 1355

REFERENCES
1. Adini A. & Clough R .W.(1964) “Analysis of plate bending by finite element method”,
Report G-7337 National Science Foundation Washington D.C.
2. Chandrashekara K. (2001) Theory of Plates, University Book Publication, New Delhi
India.
3. Choudhury D. & Subba Rao K.S. (2002) "Displacement - Related Active Earth Pressure",
International Conference on Advances in Civil Engineering (ACE-2002), January 3 - 5,
2002, IIT Kharagpur, India, Vol.2., pp. 1038-1046.
4. Choudhury D., Sitharam T.G. & Subba Rao K.S.(2004) "Seismic design of earth
retaining structures and foundations", Current Science, (ISSN: 0011-3891, IF:
0.694/2003) India, Vol. 87, No. 10: pp. 1417-1425.
5. Choudhury D. & Chatterjee S. (2006) "Displacement - based seismic active earth
pressure on rigid retaining walls", Electronic Journal of Geotechnical Engineering,
(ISSN: 1089-3032), USA, Vol. 11, Bundle C, paper No. 0660.
6. Chowdhury I. & Dasgupta S.P.(2003) “Counterfort Retaining wall under Earthquake
Force ”- ; Electronic Journal of Geo-technical Engineering Vol-9A.
7. Chowdhury I. & Dasgupta S.P. (2007) “Dynamic Earth Pressure on rigid unyielding
walls” Indian Geotechnical Journal Vol. 37 # 2. pp 81-93.
8. Chowdhury I. & Dasgupta S.P.(2008) Dynamics of Structures and Foundations –a
unified approach Volume-1, Taylor and Francis Publication, Leiden, Holland.
9. Chowdhury I. & Dasgupta S.P.(2011) “ An analytical solution to seismic response of
cantilever retaining wall with generalized backfilled soil” Electronic Journal of
Geotechnical Engineering Volume 16C USA.
10. Clough R.W.(1984) Dynamics of Structures, McGraw Hill Publications New York USA.
11. Das B.M. & Puri V.K.(1996) “Static and dynamic active earth pressure”, Geotechnical
and Geological engineering Vol-14, pp-353-356.
12. Ghosh S. & Saran S.(2007) “Pseudo static Analysis of Rigid Retaining wall for Dynamic
Active Earth Pressure” Cenem B.E.College Kolkata India
13. Ghosh S., Dey G.N. & Datta B.N.(2010) “Pseudo static Analysis of Rigid Retaining wall
for Dynamic Active Earth Pressure” 12th International Conference of International
Association for Computer Methods and Advances in Geomechanics.
14. Ghosh S. and Pal J.(2010) “Extension of Mononobe-Okabe expression for active earth
force on retaining wall backfilled with c-φ soil”14th Symposium on Earthquake
Engineering, Indian Institute of Technology Roorkee India Vol-1 pp 522-530.
15. IS-1893(2002) – Code for Earthquake resistant design of Structures; Bureau of Indian
Standard Institution, New Delhi, India.
16. Mononobe N. & Matsuo H. (1929) “On the determination of earth pressure during
earthquakes”, Proc. World Engineering Congress, Tokyo, Vol. 9, Paper 388.
17. Murthy V.N.S. (1984) Soil Mechanics and Foundation Engineering, Sai Kripa
Publication, Bangalore, India.

- 1355 -
 Vol. 16 [2011], Bund. R 1356

18. Okabe S. (1924) “General theory of earth pressures and seismic stability of retaining wall
and dam” J. Japanese Society of Civil Engineers, Vol. 12, No. 1.
19. Ostadan F. & W. H. White (1997) “Lateral seismic soil pressure, An updated approach”,
Bechtel Technical Group Report Los Angles USA
20. Ostadan F. (2004), “Seismic soil pressure on building walls-An Updated approach”, 11th
International Conference on Soil Dynamics and Earthquake Engineering. University of
California, Berkeley.
21. Saran S. & Prakash S. (1968) “Dimensionless Parameters for Static and dynamic earth
pressures behind retaining walls”, Indian Geotechnical Journal Vol. (72(3) pp 295-310.
22. Saran S. & Gupta R.P.(2003) “Seismic Earth Pressure behind retaining walls” Indian
Geotechnical Journal Vol. 33(3) pp195-213.
23. Seed H.B. & Whitman R.V. (1970) “Design of earth retaining structures for seismic
loads”, ASCE Specialty Conference on Lateral Stress in Ground and design of Earth
Retaining Structures, June.
24. Steedman R.S. & Zeng X. (1990) “The Seismic response of Waterfront Retaining walls”,
Proceedings on Specialty Conference on design performance of Earth Retaining
Structures, Special Technical Publication 25 Cornell University Ithaca New York pp 897-
910.
25. Szilard R. (2004) Theories and applications of plate analysis, John Wiley Publication.
26. Timoshenko S. & Krieger W. (1987) Theory of Plates and Shells, McGraw Hill
Publication, N.Y., USA.
27. Tocher J.L.(1962) “ Analysis of Plate bending using triangular element” PhD Thesis
University of California, Berkeley, USA.
28. Whitman R.V.(1990) “Seismic Design and Behavior of Gravity Retaining walls”,
Proceedings Specialty Conference on design and performance of Earth Retaining
Structures, ASCE, Cornell University, June18-21.
29. Whitman R.V. (1991) “Seismic design of Earth Retaining structures”, Proceedings 2nd
International conference on recent advances in Geotechnical Earthquake Engineering
and Soil Dynamics, St Louis USA, March 11-15.

LIST OF SYMBOLS
a, b = Dimension of the plate;
C1, C2, C3, C4 = Constants of integration;
c = Soil cohesion;
D = Et 3 / 12(1 − ν 2 ) = Flexural stiffness of the plate;
E = Young’s Modulus of the RCC wall;
g = Acceleration due to gravity;
H = Aepth of soil body;
I = Moment of inertia of the strip I
i =Inclination of the soil slope;
I, R and Z= Respectively, importance factor, response reduction factor and zone factor;
KA, KP = Active and passive earth pressure coefficients;

- 1356 -
 Vol. 16 [2011], Bund. R 1357

Mx, My = Dynamic moments in x and y-directions;

mi = The ith structure mass;
pay, ppy = Static pressures acting on the strip of element in x and y directions;
Qx,Qy = Dynamic shear forces in x and y-directions;
q = Pressure load;
r = H/b = Aspect ratio;
Re = Residual error Re;
Sd = Maximum amplitude;
Sa = Spectral acceleration;
T = Time period;
t = Thickness of the wall;
u = Displacement of the strip element;
ud = Dynamic amplitude of the wall;
ustatic = Statical displacement;
X1, X2, X3, X4, X5 = Integra functions;
x,y = Cartesian coordinates;
α = Failure line inclination, = 450+φ/2
β = Code factor, earthquake parameter;
γs = Bulk density of soil body;
γw = Bulk density of water;
ξ = x/b = Non-dimensional x;
η= y/H = Non-dimensional y;
κ = Modal mass participation factor;
ν = Poisson ration;
φ = Soil friction angle;
Φ = Eigenvectors as a function of x and y;
φi = Mode shape;
ω = Circular frequency of vibration.

© 2011 ejge

- 1357 -

View publication stats