Missouri University of Science and Technology

Scholars' Mine
International Conferences on Recent Advances in 1991 - Second International Conference on Recent
Geotechnical Earthquake Engineering and Soil Advances in Geotechnical Earthquake Engineering
Dynamics & Soil Dynamics

Mar 11th - Mar 15th

Seismic Earth Pressures Against Structures with

Restrained Displacements
P. Ortigosa
IDIEM, University of Chile, Chile

H. Musante
IDIEM, University of Chile, Chile

Follow this and additional works at: https://scholarsmine.mst.edu/icrageesd

Part of the Geotechnical Engineering Commons

Recommended Citation
Ortigosa, P. and Musante, H., "Seismic Earth Pressures Against Structures with Restrained Displacements" (1991). International
Conferences on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics. 10.
https://scholarsmine.mst.edu/icrageesd/02icrageesd/session04/10

This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in International
Conferences on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics by an authorized administrator of Scholars' Mine. This
work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder.
For more information, please contact scholarsmine@mst.edu.
( \ Proceedings: Second International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics,
~ March 11-15, 1991, St. Louis, Missouri, Paper No. 4.8

Seismic Earth Pressures Against Structures with

Restrained Displacements
P. Ortigosa H. Musante
Professor of Civil Engineering, IDIEM, University of Chile Professor of Civil Engineering, IDIEM, University of Chile

SYNOPSIS A simplified kinematic method to compute seismic earth pressures against structures with
restrained horizontal displacements is described. The retaining structure has a linear elastic
behaviour and the retained soil is considered linear as well as a nonlinear material. Seismic
pressures for different maximum free field accelerations were obtained by applying free field
horizontal displacements at the base of the interaction springs conecting the soil and the retain-
ing structure. For perfectly rigid structures without horizontal displacements and linear retained
materials, seismic pressures obtained with the simplified kinematic method are compared with those
obtained using elastic models; for nonlinear retained soils comparisons are established with results
obtained using the finite element method. Finally, a parametric analysis was performed using the
kinematic method with nonlinear soils for both perfectly rigid and flexible retaining structures
with restrained horizontal displacements.

INTRODUCTION the threshold acceleration to trigger tilting.

The analysis of seismic earth pressures against
retaining structures is commonly divided into
two main groups:
- Retaining structures with enough flexibi-
lity or with foundations able to undergo
some amount of horizontal sliding or
tilting during an earthquake.
- Retaining structures with large flexural
rigidity or with rigid lateral supports,
founded on rigid soils, where lateral
displacements are restrained to small
levels.
First Group
For the first group, displacement of the retain-
ing wall is assumed to be sufficient to produce
a state of plastic equilibrium in the retained
soil, resulting in minimum active pressures.
Probably the earliest design method to carry out
with this approach is the classical Mononobe and
Okabe (M-0) formulation, which is an extension
of Coulomb theory. In the M-0 method the
failure plane in the soil is assumed to pass
through the toe and be inclined at some angle,
so the soil wedge acts as a rigid body with
uniform vertical and horizontal accelerations.
By using the earth thrust given by M-0 formula
and applying the Newmark's approach, Richards
and Elms (1979) established an expression to
compute permanent horizontal displacements when
the earthquake acceleration pulses at the wall
base are larger than a threshold horizontal
acceleration. This threshold acceleration is
defined as the minimum value to trigger horizon-
tal slide at the wall base (or other critical
section) including inertia forces on the wall
itself. This definition assumes a lower
threshold acceleration to trigger sliding than
At this extend, to compute bending moments and
shear forces in the retaining structure, the
maximum horizontal inertia force to be applied
at the M-0 wedge and the retaining structure
seems to be the threshold value. For flexible
retaining structures where permanent displace-
ments at the wall base are constrained to small
levels, it seems that the maximum horizontal
acceleration to be applied is the maximum
earthquake acceleration.
More recently, Nadim and Whitman (1982) used
the finite element method (F.E.M.) to compute
seismic pressures and permanent displacements
on rigid retaining walls. To model the failure
plane in the soil and soil-wall interfaces they
use contact elements with elasto-plastic beha-
viour. For the retained soi 1 nonlinear
elements were introduced, with shear modulus
and damping factors depending on the maximum
shear strain induced by an earthquake record
applied at the wall base (actually they use an
equivalent linear behaviour following the
procedure outlined by Seed and Idriss, 1970).
Many cases of damage reported in the literature
can be attributed to increased lateral pres-
sures during earthquakes inducing sliding or
tilting on the retaining structures, or both.
A recent example of this behaviour was observed
at Valparaiso Harbour during the 1985 Chilean
earthquake. Figure 1 illustrates typical
profiles of the gravity retaining walls at
berths 1 to 5, built during 1913-1924 using
concrete blocks without shear keys. Due to the
foundation soil characteristics at berths 1
through 4, failure was triggered as a permanent
sliding between concrete blocks at 12m depth;
due to poorest soil conditions below the
retaining wall at berth 5, a permanent rotation
at the wall base was measured. Figure 2 shows
the horizontal displacements at the top of the
walls.

621
 L
a max
Berth
1' 2, 3
Rock Depth
26m
Retained
soi 1:
Tz
4, 5 31m Gcz
y
'(
12
Base
16
E
20
-"'
.c
0.

0 24 Fig. 3 Earth Retaining Structure used in the

Kinematic Method.
28
KINEMATIC METHOD
32 Figure 3 shows a retaining structure with
linear elastic behaviour. The retained soil
36 can be considered as a linear or nonlinear
0 20 1,0 60 80 material, with a length L ~ 4.0H in order to
avoid border effects (Wood, 1973). Other
N (blows/foot) parameters are as follows:
Fig. 1 Gravity Retaining Walls at Valparaiso
Harbour. H height of the retaining structure.
h span between horizontal supports
(struts, slabs, etc).
140
t wall thickness.
120 E Young modulus of the retaining wall.
E
~ 100 I moment of inertia of the wall cross
c: section per unit length.
"' 80
E maximum free field horizontal accelera-
"'u 60 tion expressed as a fraction of g.
~
0.
VI
a = horizontal acceleration at the base
0 40 expressed as a fraction of g.
Gcz shear modulus of the retained soil at a
depth, z.
maximum free field horizontal displace-
Horizontal Length at Berth 5, (m) ment at depth, z, relative to the base.

Fig. 2 Seismic Horizontal Displacement at 'Ycz maximum free field shear strain at
the Top of the Retaining Walls- depth z.
Valparaiso Harbour.
v Poisson modulus of the retained soil.
Second Group
'Y unit weight of the retained soil.
For retaining structures with restrained lateral
displacements, most of the methods assume a KH horizontal spring representing soil-
linear elastic behavior (i.e. Arias, 1982 and structure interaction.
wood, 1973). Nadim and Whitman (1982) applied
the F.E.M. to compute seismic pressures on Ke rocking spring representing soil-
perfectly rigid walls without horizontal displa- structure interaction.
cement and nonlinear retained soils, but
computer time and efforts to produce the input kz modulus of horizontal subgrade reaction
data can preclude solutions using this approach, at depth, z, between the structure and
especially for parametric analysis. That is why the retained soil.
a simplified kinematic method was developed to
compute seismic pressures against linear elastic For simplicity, in the analysis and results
structures with both linear and nonlinear that follow values of h, t, E, v and 'Y are
retained soils. The method was used to perform considered to be constant with depth, KH is
a parametric analysis on perfectly rigid and assumed to be very large and no dashpots are
flexible retaining structures with restrained introduced to take into account soil structure
horizontal displacements. radiational damping.

622
The simplified kinematic method computes the
seismic pressures, as, by applying the maximum ( 7)
free field horizontal displacements at the base 3
of the subgrade springs. Hence, values of as
can be expressed as:
(1- O.Ol67z) ( 8)

as ( l)

Equation (5) gives the maximum shear strain

where Osz = horizontal displacement of the
structure at depth z. Extreme condition for a
perfectly rigid structure without horizontal
displacement (6sz = 0) can be reached by setting
very large E and Ke values. On the other hand,
two independent conditions must be satisfied
when solving the problem:
( 2)
( 3)
First condition was introduced when dealing with
nonlinear retained soils. It means that seismic
plus initial static horizontal pressure, oh, at
a given depth can reach a limit value given by
the passive soil pressure at that depth, op.
Passive resistance was computed using conven-
tional static expressions becouse maximum
seismic pressures act when inertia forces go
from the soil towards the retaining structure;
static horizontal pressures were computed as
oh = K0 ·ov, where K0 is the at rest coeffi-
cient and ov the vertical pressure expressed
as y. z. Values of K0 = 1 -sencj> were used,
where cj> represents the angle of internal
friction of the retained soil. For compacted
granular fills the initial static horizontal
pressure is computed using the procedure
outlined by Broms (1971).
Second condition means no relative horizontal
displacement at any depth between the structure
and the soil immediately close to the wall. This
condition was introduced to maintain approxima-
tely the at rest free field condition used to
compute 6z values.
Finally, it is necessary to point out that the
kinematic method was set for small inertia
forces coming from the structure (i.e. mostly
buried structures, underground structures with
light weight aerial members, etc).
using a reduction coefficient, rd. According
to Seed and Idriss (1971) and others analyses,
a simplified expression to obtain rd is given
by equation (8).
Equation (6J applies for cohesionless granular
soils and gives the shear modulus in ton/m2 if
the mean initial confining pressure, oc, is
expressed in ton/m2. The shear coefficient,
K2, used in that equation is strain independent
when dealing with a linear elastic behavior.
However, for nonlinear granular soils, K2
depends on the soil characteristics as well as
on the shear strain, Ycz· Figure 4 gives the
strain dependent variation of K2/K2max for
granular soils, where K2max is the shear
coefficient obtained for small shear strains
(i.e. Yc<l0-4%). For cohesive soils equations
(6) and (7) are replaced by direct correlations
between Gc/Su and Yc• where Su is the undrained
strength of the soil (Seed and Idriss, 1970).
Accordingly, if the soil is characterized by
K2max or Su, it is possible to obtain the shear
modulus as a function of the shear strain level
(actually a secant equivalent linear modulus).
By combining equations ( 4) through ( 8), an
iterative computation process was used to
obtain free field horizontal displacements.
Modulus of Subgrade Reaction
The horizontal modulus of subgrade reaction was
computed using the equation given by Gonzalez
and Ortigosa (1977), which applies for walls
rotating about the top, about the base or
subjected to horizontal translation:
2 Gcz 1
kz ( 9)
1 - v

Free Field Displacements

Maximum free field horizontal displacements were
computed using a simplified approach given by
the following equations:

z N 0.4
:,::
f Ycz · dz ( 4)
0.2
Range for
sandy gravels
H
(Ortigosa 1 1987)
0~------~--------~--------~------~
10- 4 10- 3 10- 2 10-1
Ycz ( 5)
Gcz · g Shear strain I 'lc ("f.)

Fig. 4 Strain dependent Variation of the Shear

( 6) Coefficient for Granular Soils.

623
 0 r----,-----r---. ...----rrO--e.- -, - - Wave equation
Arias (1982)
F E.M. rough
0.2 wall, Wood (1973)
FE.M. smooth
wall, Wood (1973)
0.4 ---+-- Kinematic
~ method
H -o-- FE.M. smooth
N 0.6 wall, Kern County
..c
~
•
3~---~~----~-------~-------,--­
earthquake
Nadim-Whitman
Q.o
(1982)
0 0.8 Arias, Wood and
Kinematic method use
constant acceleration
lO
0 0.8 l2 1.6
Cr =Cis JL
d'H a
Fig. 6 Dimensionle ss Seismic Earth Pressures
against Rigid Walls without Horizontal
Fig. 5 Horizontal Modulus of Subgrade Reaction Displacemen t: Constant Elastic Shear
Tunnel Section within the Santiago Sandy Modulus with Depth.
Gravel.

Figure 5 shows measured kz values along the wall

of a tunnel section at the Santiago subway,
located within a very dense sandy gravel
deposit. Predicted kz values were obtained by
means of equation ( 9 l using a Poisson ratio 0.4
from triaxial tests on large size undisturbed z
samples. The shear modulus vs depth used into H FE. M. smooth wall
equation (9) for predicting kz values exhibited Q6 Kern County quake1959
a well defined profile either from triaxial Nadim-Whitma n (1982)
tests, building settlement measurement s or
horizontal plate tests (Ortigosa, 1987 l.
Predicted values compare pretty well with those
measured, so the kinematic method used equation
( 9 l for computing kz values. Final Gcz values
given by the iterative process when computing
maximum free field horizontal displacement s were Cr = Cis _g_
introduced into that equation. tH a
Fig. 7 Dimensionle ss Seismic Earth Pressures
Water Table against Rigid Walls without Horizontal
Displacemen t: Variable Elastic Shear
For simplicity, water table in granular soils Modulus with Depth.
was handled assuming no excess pore pressures
(i.e. dense or high permeability soils behind
the retaining structure). Saturated unit weight v=0.30, L= 5H and a constant horizontal accele-
was introduced into eq. (5) when dealing with ration, a, acting simultaneous ly in all points
soil layers below the water table and bouyant of the retained material. Figure 6 also
unit weight was used for Oc computations . For includes maximum seismic pressures, Osr
saturated cohesive soils a similar approach was reported by Nadim and Whitman for a rigid wall
used when using eq. (5), but shear modulus was with H= 10m, subjected at its base to the S69E
obtained using direct correlations between Gc/Su acceleration record of the Kern County 1952
VS Yc·
earthquake, scaled to represent a peak accele-
ration of 0.4g. To set Nadim and Whitman
results into a dimensionles s factor, Os values
RESULTS were divided by the peak acceleration of the
earthquake record. Results under similar
Linear Behaviour conditions are given in Fig. 7 using a shear
modulus Gcz = K ~. where K is a constant.
The kinematic method was applied for retained
materials with elastic linear behavior and According to these results, for linear elastic
perfectly rigid structures without horizontal materials the simplified kinematic method
displacement s (6sz = 0). Figure 6 compares the predicts seismic pressures with enough engi-
dimensionle ss seismic pressure, cr, obtained neering accuracy when compared with more
using a constant shear modulus with depth, sophisticate d solutions.

624
 0 --.;;:::,
---~
r-=-::.: ---.::
1.0

2
- - Kinematic Method
~ r--
......

''
' r-.
-v N"
0.8

E 0.6
......
c
• /
•
~eosured

•
4 - - - F.E.M. smooth wall / 0
_, 0--. .~ r--...
E
N
Kern County quake 1959
Nodim-Whitmon ( 1982)
/
l( 0.4
['--.--'
I ~.

~
I 0.2
6
v I Kinematic

8 v /
/
/
0
0
Method, omax=0.013g

10 15 20
\ 25
/ ...... ..- Depth z(m)

10
~
~ --- Fig. 9 Comparison between Predicted
Measured Seismic Earth Pressures.
and
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Cr = ~ _g_ basement walls during the earthquake of 12 June
~H Omax. 1978 (magnitude 7.4 and epicentral distance
Fig. 8 Dimensionless Seismic Earth Pressures 380km) which caused a maximum acceleration at
against a Rigid Wall without Horizontal the 2nd basement of about 12.5 gals (0.013g).
Displacement: Comparison for a Dry Sand The soil profile comprised a thick alluvial
with Nonlinear behavior. deposit of soft saturated silt with y = 1.52
ton/m3, K0 =0.58 obtained from measured static
pressures and Su ranging 1.25 to 4.7 ton/m2,
Nonlinear Behaviour: Rigid Structures which reached the hard support layer of a
diluvial deposit at a depth of 22 to 28m. The
As a first step, the kinematic method was used kinematic method was applied using the average
on perfectly rigid structures without horizontal Gc/Su vs Yc correlation for saturated clays
displacement (Osz = 0). The retained soil was a proposed by Seed and Idriss, a Poisson modulus
dry sand with ~=30°, y=2ton/m 3 , v=0.3, K2max=50, v=0.5, a perfectly rigid retaining wall going
average K2/K2max vs Yc given in Fig.4 and a down 25m and a maximum free field acceleration
maximum free field acceleration amax=0.4g. of O.Ol3g. Figure 9 shows a more or less good
Fig. 8 compares the dimensionless seismic comparison in spite of ignoring inertia forces
pressures with those reported by Nadim and coming from the building and uncertainties due
Whitman using F.E.M. for a rigid wall with to unknown free field accelerations.
H=lOm, the same sand and the S69E Kern County
acceleration record scaled to a peak accelera- A parametric study followed using the kinematic
tion of 0.4g applied at the wall base. To set method on perfectly rigid structures with Osz=O
Nadim-Whitman's results into a dimensionless and granular retained soils. Table I summa-
factor, maximum Os values given by these authors rizes parameters used in such analysis, which
were divided by the peak acceleration of the are referred as Basic Cases, and Fig. 10 shows
earthquake record. Similar comparisons are in typical results.
progress using wall heights ranging 4 to 13m and
other acceleration records such as El Centro To see the influence of the K2/K2max vs Yc
1940 and Chilean earthquake records on rock and curve on seismic pressures, some Basic Cases
deep sand deposits, scaled to represent a peak with H=4m and H=lOm were selected using
acceleration ranging O.lg to 0.4g. K2max=50 and 150, and the upper and lower
bounds for sands given by Fig. 4~ or K2max=310,
Comparisons were also established with measured the average curve for sands and the lower bound
seismic pressures on the basement walls of a for sandy gravels. Typical results are
building in Yokohama (Ikuta et al, 1979). The illustrated in Fig. 11 showing a maximum
Yokohama Tenri Building has two basement floors variation of the average seismic pressure along
and 27 stories above ground level. The founda- the wall height of ± 7%. This percentage
tions comprise cast-in place piles supporting increases up to ± 12% when considering the
the central core and basement walls extended to whole set of selected cases. The influence of
26-28m from ground level, forming a continuous the angle of friction was also analyzed using
piling wall supporting the perimeter. Maximum ~=38° and 45° on some selected Basic Cases,
seismic pressure records were taken on the proving to be less than 10%.
TABLE I. Parameters for Basic Cases: Rigid Walls with Nonlinear Granular Retained Soils
Soil Characterization H amax K2 v y
- - - vs Yc
K2max ~ (m) K2max (tonjm3)
50 38° 4 O.lOg Average 0.30 2.0 (dry)
150 42° 7 0.15g curve 2.25 (water
310 45 ° 10 0.20g for sands table at the
13 0.30g retained soil
0.45g surface)

625
 0 0

~'\
a max =0.45 g
/

2 2
0.1g
,\ \
I
"
o.39
'\ 2

~ 3 4
3
E E 0.2g E
4 N 4 1 N Dry Soil
N 6
K2max =150
I
Iv
5 H =10m
K2max =150 - - - Upper bound
8
H =7m for sands.

IMv
6 6
- - Lower bound
7 10
2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 4 8 12 16 20
Cis (ton/ m2) <Ts(ton/m2) G's(ton/m2)

Fig. 10 Seismic Earth Pressures against a Rigid Wall Fig. 11 Rigid Wall: Influence of the
without Horizontal Displacement: Basic Ca- Shear Modulus Strain depen-
ses for a Dry Nonlinear Granular Soil. dent Variation.

Values of the dimensionless average seismic
pressure along the wall height, cr, obtained
from the Basic Cases analysis are plotted in
Fig. 12. These values can vary up to ± 12% due
to changes on the K2/K2max vs Yc curves as
pointed out before. Dimensionless seismic
pressures in Fig. 12 were computed using
averaged results for H=4m to 13m, so additional
variations on Cr must be considered due to a non
perfect normalization with the wall height,
ending with a maximum final percentage of
variation of the order of ±15%. As a reference,
Fig. 12 includes the dimensionless average
seismic pressures along the wall height given by
the M-0 formula. The M-0 average pressures were
plotted as a function of K2max• using K2max vs ~
correlation established for the Basic Cases as
shown in Table I.
1.4
--
----
1.2
\ fixed support (Ke=~J.
~~
1.0
Results in Fig. 12 show a significant increase
of seismic pressures when going from yielding
retaining structures, where M-0 formula ap-
plies, to rigid walls without horizontal
displacement. In spite of assuming no pore
pressure increments, soil bouyancy also produce
an increase of seismic pressures for both
yielding and rigid structures. However,
increase due to wall rigidity is significantly
greater than the increase due to soil bouyancy.
Nonlinear Behaviour: Flexible Walls
Retaining structures similar to the basement
wall shown on Fig. 3 were analyzed using the
kinematic method. For simplicity, the analysis
was performed assuming a constant wall thick-
ness, t, and a constant span lenth, h, between
rigid horizontal supports. Soil-structure
interaction at the wall base was modeled using
Dry ( t = Yl
a very large horizontal stiffness (KH=~l and a
Submerged ( Y = J1 sat l rocking spring representing a hinge (Ke=O) or a
The retained soil was a
_Q.45 g
dry or a submerged nonlinear granular material.
Accordingly, a parametric analysis was perfor-
I RIGID med using the soil and wall characteristics

~ ~--
WALL Jsz=Of
....' ' given in Table II. Typical results are

- --- --- -
0.30g

~li 0.8
~
illustrated by Fig. 13 for one and three span
::::.::::...:!.;), - walls with hinged support at the wall base.
1 This figure show a larger seismic pressure
amax =0.1g .... attenuation as long as the span flexural

j'I
0.6
--I ~-r --- 0.45g flexibility, Fs, increases. This parameter was
defined as:
~""
0.3 g\ lM-0
~-
"L.
0.4
- -- -4· 1
lu
l:;.z:::~
Fs = I ( 10)
0.2
- - EI 12
amax =0.1g
The whole set of results from the parametric
I analysis were plotted as As vs Fs, where As is
100 200 300 the seismic pressure attenuation coefficient
defined as Cf/Cr· Values of cr represent the
K 2max dimensionless average seismic pressure along
the wall height as defined previously for a
Fig. 12 Dimensionless Average Seismic rigid retaining wall (Fs=O) without horizontal
Pressure for Granular Soils from displacement (Osz=O), and Cf is the correspon-
Basic Cases. · ding dimensionless average seismic pressure for

626
 TABLE II. Parametric Analysis: Flexible Walls with Nonlinear Granular Retained Soils

Soil Characterization Number of h t Ka E a max K2 v r

- - - vs "Yc
K2max q, spans (m) (m) (ton/m2) K2max ( ton/m3)
50 38 ° 1 2 0.15 0 2.sx1o6 O.lg Average 0.30 2.0 (dry)
150 42° 2 3 0.20 0.2g curve 2.25 (water
310 45 ° 3 4 0.25
"' 0.3g for sands table at
0.30 0.4g the retai-
0.45 ned soil
0.60 surface)

0
-~ Rigid
0
I .,._._ ~igid

ll' ~ 2
r-- 3 Span
"" ~
~'\

J \ " f-- h = l.m ~

-~
fll K2max = 150
E6 .///

!ve
2
I N \ I
1-- 'V
3
.... c::,·
I 8
t = 0.1.5m .#
lj / 1 Span
::;.l ~V/
10 .....
_j_ ~v
h =I. m
K2max =150 ~ ~
....
rc- 0.60
I. 12
0 2 3 0
<Is (tonfm2) " 8

Fig. 13 Seismic Earth Pressures against Flexible Walls with Hinged Support: Dry Nonlinear Granu-
lar Soils with amax = 0.3g.

the same wall with some degree of flexural 0.002H and O.OOlH for K2max=50, 150 and 310,
flexibility (Fs>O), KH="' and Ka=O or "'· respectively). These displacements seem to
Figures 14 and 15 summarize the As vs Fs be lower than those required to reach a full
correlation for Ka=O and h=2 and 4m, respective- state of plastic equilibrium in the retained
ly. Quite similar results were obtained soil, which is an implicit condition when
for submerged granular soils. Some important using the M-0 formula.
features related to Figs. 14 and 15 must be
pointed: For a given span flexural flexibility, seis-
mic pressure attenuation increases as long
Each curve has been plotted until reaching a as the soil stiffness, typified by K2maxr
limit value of the span flexural flexibility, increases.
Fsmax• to satisfy the condition given by
equation(3); for Fs>Fsmax at least one point
on the retaining wall exhibits an horizontal FINAL REMARKS
displacement greater than the free field
seismic displacement ( 6sz>6z). Greater A simplified kinematic method was developed to
attenuation coefficients can be obtained if compute seismic pressures against retaining
condition given by equation ( 3) is not structures with restrained horizontal displa-
satisfied at some points along the wall. cements. The method can be applied using
linear and nonlinear behavior for granular and
Each curve represents an average correlation cohesive soils under dry and submerged condi-
between As and Fs, with a maximum variation tions. Submerged granular soils are assumed
of ±12% for K2max=50 and ±8% for K2max ~ 150. with no pore pressure increments. The method
These variations arise from a non perfect is suitable for parametric analysis using a
normalization with the wall height and the microcomputer, but it is constrained to solve
maximum free field acceleration. problems with small inertia forces coming from
the structure. More empirical evidences are
As a reference, it is shown the seismic needed to validate results given by the kinema-
pressure attenuation coefficient using Cf tic method or other analytical approach.
values from the M-0 formula.
Attenuations for the limit span flexural ACKNOWLEDGEMENT
flexibility are lower than those obtained
with the M-0 formula. This is in accordance The writers wish to acknowledge their apprecia-
with the maximum wall displacement associated tion for the finantial support offered by the
with that limit (i.e. for one span wall, National Commission for Science and Technology
displacements are of the order of O.OlH, of Chile (Research Project FONDECYT 89-964)

627
 1. 0.----,,---r-r--rrn-n-.,----r..-rTTTTl

0.9 +-----""1..::----+-r~r++-H---~

0. 8 +---f-t-+-+++++1---+-+-+--m+H

2 SPAN 3 SPAN
0. 6+---f-+-t--t+Htt--+--.-,--,rrrrrl

0. 5+---f-+-t--t+Htt--+--+-Hr-t+lti 0.5+--1--~~H+H--+-+4-H++H

Fs max (Typ)

~M-0 ~ = 38~-45~
4
0.
0. 3
10- 3 10- 2 10-1 10- 2 10-2
Fs ( m3 I ton) Fs (m3tton) Fs(m3/ton)

Fig. 14 Seismic Pressure Attenuation Coefficient for Flexible Walls with Hinged Support and h=2m
Dry Nonlinear Granular Soils with amax = O.lg to 0.4g.

1 SPAN
0.9 ~ r--,b, 0.9+--T-t""'k-t+-?rrt--+-+4'~

O.A 1\._\~ ~ ~ 0. 8+----1i--t-t-+tt++f----t--t~-++1f-H-i 0. 8+--t-t-+-t++-1+1--+-++-+++++1

1'\ ~
0.7+----1~'\-~'-o-+"<+~-'--&.uo'++--4.:\~ 0. 7 +----11--t-+-++++++----'---'-'---.L..L..If-H-i
2 SPAN
0. 7 +---lt--t-t-++++++--'---.L....I.--'-'-+f-H
3 SPAN
0.6 1\ IN \ 0. 6 +----11--t-+-++++++-----,----,,---.--rif-H-i

0.:>1;.-t-----lr--t--t--i 1 ~~ 0. 5+---t--f-+-++++++--+--+-+++++-H
Fs max CTyp)
0.4 ~\~'
~-0
0.3 3
10- 10-2 10-2
Fs(m3/ton) Fs ( m3 /ton)

Fig. 15 Seismic Pressure Attenuation Coefficient for Flexible Walls with Hinged Support and
h=4m: Dry Nonlinear Granular Soils with amax = O.lg to 0.4g.

REFERENCES Ortigosa, P. (1987), "Horizontal Plate Tests on

Gravels", Proc. VI I I Panamerican Conf. on
Arias, A. (1982), "Seismic Analysis of Retaining Soil Mech. and Foundation Eng., Cartagena,
Walls", Revista Ing. Sismica, N° 26, Sociedad Colombia, pp 597-606; see also pp 607-619.
Mexicana de Ing. Sismica.
Richards, R. Jr. and D. G. Elms (1979),
Brems , B. B. ( 1 9 71 ) , "Lateral Earth Pressures "Seismic Behavior of Gravity Retaining
due to Compaction of Cohesionless Soils", Walls", Journal Geotech. Eng. Div., ASCE,
Proc. 4th. Conf. on Soil Mech., Budapest, Vol. 105, N° GT4, pp 449-464.
pp 373-384.
Seed, H. B. and I. M. Idriss (1970), "Soil
Ikuta, Y., M. Maruoka, T. Mitoma and M. Naganou Moduli and Damping Factors for Dynamic
(1979), "Record of Lateral Pressure taken Response Analysis", Report N° EERC 70-10,
during Earthquake", Soils and Foundations, University of California, Berkeley.
Vol. 19, No 4, pp 85-92.
Seed, H. B. and I . M. Idriss (1971), "Simpli-
Nadim, F. and R. V. Whitman (1982), "A Numerical fied Procedure for Evaluating Soil Liquefac-
Model for Evaluation of Seismic Behavior of tion Potential", Journal Soil Mech. and
Gravity Retaining Walls", MIT Research Report Found. Div., ASCE, Vol. 97, N° SM9, pp 1249-
R82-33. 127 3.
Gonzalez, R. and P. Ortigosa (1977), "Lateral Wood, J. H. (1973), "Earthquake - Induced Soil
Soil Restriction on Piers", Revista del Pressures on Structures", Report N° EERL
!DIEM, Vol. 16, N° 2, pp 73-102. 73-05, CALTECH, Pasadena, California.

628