1567
Accumulated Stress Based Model for Prediction of Residual Pore Pressure
Étude et développement du modèle pour le pronostic sur l’excès de pression hydrostatique
interstitielle causé par les contraintes accumulées
Park D., Ahn J.-K.
Department of Civil and Environmental Engineering, Hanyang University
ABSTRACT: Even though the important influence of pore pressure rise under cyclic loading on seismic wave propagation is
recognized, effective stress analysis is rarely performed due to difficulties in selecting the parameters for the pore pressure model. In
this paper, a new numerical model for predicting pore pressure under cyclic loading is developed. The advantages of the model are
that it requires only the CSR – N curve in selecting its parameters and it can be can be used for any loading pattern. The accuracy of
the model is validated through comprehensive comparisons with measurements.
RÉSUMÉ : L’importance de l’excès de la pression interstitielle causé par les contraintes accumulées de la propagation des ondes
sismiques est bien reconnue, mais l’analyse de contrainte effective est rarement pratiquée en raison de difficultés à évaluer les
paramètres pour le modèle de la pression interstitielle. Le présent article concerne le développement du nouveau modèle numérique
pour le pronostic sur l’excès de pression interstitiellecausé par les contraintesdans le sol. Les avantages de ce modèle sontque nous
pouvons déterminer tous les paramètres avec la courbe CSR-N et qu’il peut s’appliquer aux diverses formes de contraintes. La
précision du modèle est contrôlée par comparaison avec le résultat du test.
KEYWORDS: pore water pressure, damage parameter, cyclic stress ratio, accumulated shear stress, time-domain analysis.
1 INTRODUCTION
Build-up of residual excess pore water pressure in sands and
silts during seismic loading causes reduction in stiffness and
strength of soils and can lead to liquefaction. It may greatly
influence the characteristics of ground motion propagation,
stability of embankments, and seismic performance of structures
such as tunnels and bridges. The importance of predicting the
pore pressure has been well recognized and the characteristics
of pore pressure generation for sands and silts have been
extensively studied (Booker et al. 1976, Carraro et al. 2003,
Derakhshandi et al. 2008, Lee and Albaisa 1974, Polito et al.
2008, Xenaki and Athanasopoulos 2003).
Various empirical models have been developed in the past to
predict the generation of pore pressure under cyclic loading.
The earliest models are based on the concept of cyclic stress
approach, where the seismic loading is presented as uniform
cyclic shear stress and the liquefaction potential is characterized
by the amplitude of cyclic shear stress and number of loading
cycles (Seed and Lee 1966, Seed et al. 1975b). The laboratory
test that best fits the cyclic stress approach is stress controlled
cyclic test. The result of a stress-controlled cyclic test is often
presented in the form of
CSR – N
curve, where the
CSR
represents the ratio of shear stress (shear stress normalized by
the effective confining pressure in a cyclic triaxial test and
effective vertical stress in a simple shear test) that triggers
liquefaction at the given number of cycles,
N
. While the stress
controlled cyclic triaxial test is still the most popular method,
the problems of the test procedure have been identified, which
include difficulty in defining the exact state at which the
liquefaction initiates, specimen non-uniformity, abrupt build-up
of pore pressure at high pore pressures, different state of
stresses compared to the field (Kramer 1996).
Consequent laboratory tests have shown that the controlling
factor of the build-up of excess pore pressure is not cyclic shear
stress, but cyclic shear strain. Strain-controlled cyclic tests,
especially simple shear tests, have been increasingly used to
measure the excess pore pressure under cyclic loading.
Numerical models that predict pore pressure as a function of
accumulated shear strain have been proposed (Dobry et al.
1985a, Dobry et al. 1985b, Finn and Bhatia 1982, Ivsic 2006).
While the advantages of strain-controlled test procedure and
strain-based pore pressure model are well recognized, it should
be noted that the stress-controlled cyclic tests are still the most
widely used laboratory procedure for evaluating the liquefaction
potential. In the absence of the strain-controlled test
measurements, the input parameters for a strain-based model
cannot be determined. The difficulty in selecting the input
parameters for the strain-based models is one of the reasons
responsible for the seldom use of effective stress dynamic
analysis in practice. In the absence of strain-controlled test data,
it seems logical that an alternative pore pressure model that only
requires the
CSR – N
curve obtained from the stress-controlled
test in selecting its parameters be used.
This study proposes such a pore pressure model and presents
guidelines for selecting its input parameters. A method for
constructing the empirical
CSR – N
curve from in-situ
penetration resistance in case the measured
CSR – N
curve is
not available is also outlined. The applicability of the model is
validated through comparisons with laboratory test data selected
from literature and also non-published test data.
2 PORE PRESSURE MODEL
One of the earliest pore pressure model, developed by Seed
et al. (1975b), is defined as follows:
1/
1
1 1
sin 2
1
2
u
L
N
r
N
(1)
where,
= residual pore pressure normalized to the initial
effective confining stress,
N
= equivalent number of cycles,
N
L
= number of cycles required to cause liquefaction,
= empirical
u
r
