793
Numerical Investigation of The Mobilization of Active Earth Pressure on Retaining
Walls
Enquête numérique de la mobilisation de la pression de la terre active sur les murs de retenue
Sadrekarimi A., Damavandinejad Monfared S.
Western University, Department of Civil and Environmental Engineering, Spencer Engineering Building, London, Ontario,
Canada. N6A 5B9.
ABSTRACT: The correct estimate of lateral earth pressure is important for the design of earth retaining structures. This study
presents an investigation into the lateral earth pressure distribution on a wall and in particular the effect of arching at deeper levels of
backfill for both at-rest and active conditions. Three-dimensional numerical simulations are performed using the ABAQUS finite
element software. The effect of wall displacement, wall-backfill interaction, subsoil-wall interaction, subsoil-backfill interaction, soil
modulus and friction angle on the mobilization of an active condition are investigated. The results of these simulations indicate that
the true lateral stress distribution on a wall is non-linear and backfill arching increases by wall displacement and backfill-subsoil
friction while increasing friction between the backfill and wall or subsoil and wall has no substantial effect on arching. The results are
further compared with those from physical model tests. An active state is mobilized at wall displacements smaller than those
suggested by Terzaghi’s physical model tests. By increasing backfill-subsoil friction and backfill stiffness, the active state is
mobilized at smaller wall displacements.
RÉSUMÉ: L'estimation correcte de la pression latérale des terres est importante pour la conception de structures de soutènement.
Cette étude présente une étude sur la distribution de la pression latérale des terres sur un mur et en particulier l'effet d’arche en partie
inférieure du remblai à la fois au repos et en poussée. Des simulations tridimensionnelles numériques sont réalisées en utilisant le
logiciel ABAQUS d’éléments finis. L'effet du déplacement de la paroi, de l’interaction mur - remblai, de l’interaction sol-mur, du
module du sol et de l'angle de frottement sur la mobilisation de la poussée sont étudiés. Les résultats de ces simulations montrent que
la distribution réelle des contraintes latérales sur un mur est non linéaire et que l’effet d’arche dans le remblai augmente avec le
déplacement de la paroi et le frottement entre remblai et sous-sol, alors que l’augmentation du frottement entre le remblai et le mur ou
le sous-sol et le mur n'a pas d'effet substantiel sur cet effet d’arche. Les résultats sont ensuite comparés avec ceux d'essais sur modèles
physiques. Un état de poussée est mobilisé pour des déplacements inférieurs à ceux suggérés par les essais de Terzaghi sur des
modèles physiques. En augmentant le frottement entre le sous-sol et le remblai et la raideur du remblai, l'état actif est mobilisé pour
des déplacements de la paroi plus petits.
KEYWORDS: finite element, retaining wall, active earth pressure, arching, displacement, numerical modeling.
1 INTRODUCTION
Estimating lateral earth pressure has been one of the earliest
concerns in civil engineering and designing retaining structures.
The most widely used theories of earth pressure are those of
Coulomb (1776) and Rankine (1857) that are both based on the
limit equilibrium theory. These classical methods have been
used widely because of their simplicity. However, they provide
little information regarding the distribution and magnitude of
lateral earth pressures produced by different magnitudes of wall
displacement. These methods are only valid for the limiting
condition of sufficient ground and wall movements to mobilize
an active state and do not provide any information for the
conditions prior to the active state. Thus, several experimental
(Terzaghi 1934; Sherif et al. 1984) and numerical (Clough and
Duncan 1991; Mei et al. 2009; Salman et al. 2010) studies have
been performed in order to evaluate the contributions of these
factors on the lateral earth pressure distribution. This study
presents a finite element numerical modeling investigation of
the lateral earth pressure distribution and impact of wall
displacement, wall-backfill interaction, subsoil-backfill
interaction, backfill modulus and internal friction angle on the
mobilization of an active condition. The numerical modeling
results are then compared with experimental data of Terzaghi
(1934) and Sherif et al. (1984).
2 NUMERICAL MODELING
Analyses are carried out using the ABAQUS finite element
code. A model is developed for a 3 m wide by 10 m high
retaining wall with plane strain boundary conditions that are
chosen to minimize container boundary effects on the backfill
sand. The wall and soil are modeled using 3D solid elements.
The concrete wall is modeled as an elastic material using a
linear isotropic elastic model. The extended Drucker-Prager
plasticity model is used with a non-associated flow rule in this
study for non-linear analyses of the backfill sand behavior. The
parameters of this model are based on triaxial compression tests
on Ottawa quartz sand (Sadrekarimi 2009). A non-dilatant flow
is assumed (
= 0) to model a loose backfill sand. The choice of
zero dilatancy angle was selected based on the extensive
experimental experiences of the first author. For loose
contractive sands (for which their state lies above the critical
state line), the mobilized friction angle becomes equal to the
critical state friction angle or in other words there is no negative
or positive dilatancy angle (Manzari and Dafalias 1997; Been
and Jefferies 2004). Accordingly, since our analyses simulate a
loose contractive backfill, we use the critical state friction angle
(32
o
) with zero dilatancy to model the loose backfill.The
properties of the backfill/foundation soil and wall are
summarized in Table 1.
