| Summary: | Liquefaction-induced ground displacements resulting from earthquake shaking are
a major cause of damage to earth structures comprising of or underlain by loose saturated
sands. A number of dams have failed due to liquefaction-induced deformations. Examples
of these are the failures of eleven tailings dams in Chile during the March 1965
earthquake and the Mochikoshi tailings dams in Japan due to the 1978 earthquake. A
number of other dams have undergone large deformations but have not failed in as much
as the impounded water was not released. The classic example of this was the near failure
of the Lower San Fernando dam due to the 1971 earthquake. A liquefaction induced flow
slide occurred on the upstream side removing the crest of the dam and leaving only about
1.5 m freeboard. Of more interest from the analytical point of view was the behaviour of
the Upper San Fernando dam in which the crest of this dam moved about 1.5 m due to
earthquake induced liquefaction.
Of equal importance are the ground failures due to liquefaction-induced lateral
spreading. It occurs on gently sloping grounds and sometimes on almost flat grounds, but
usually occurs over a very wide area. Although this type of earthquake-induced ground
movement does not involve a flow failure where the static shear stresses exceed the
residual strength of soils, it is potentially damaging and it has caused over one hundred
million US dollars worth of damage in United States alone since the 1964 Alaska
earthquake.
The prediction of earthquake induced displacements of earth dams involving soils
whose properties change markedly during cyclic loading is a difficult problem. The
difficulty mainly arises from modeling the stress-strain relations of soils, particularly when
pore pressure rise and liquefaction occur. The strains required to trigger liquefaction are
generally small (<1%). Once liquefaction is triggered, however, large but limited
deformation may occur on soils whose undrained residual strengths are greater than the driving stresses. Such soils strain harden, and regain stiffness and strength as they deform,
so the displacements are limited. For soils whose residual strengths are less than the
driving stresses, unlimited deformation leading to catastrophic failures may occur.
Complex effective stress dynamic analyses procedures have been proposed to
predict such deformations but they are essentially research tools and not generally
appropriate for analysis of most earth structures in geotechnical engineering practice. It is
important, therefore, to develop a simple reliable method for predicting such
displacements, and this is the objective of this thesis.
The deformation analysis proposed here is essentially an extension of Newmark's
method from a rigid-plastic single-degree-of-freedom system to a flexible multi-degreeof-
freedom system. It takes into account the effects of inertia forces from the earthquake,
the softening of the liquefied soil, and the settlement following liquefaction. The method
is based on the concept that the deformations prior to liquefaction will be small and can
be neglected compared to those that occur after liquefaction is triggered. A key aspect of
the method is the post-liquefaction stress-strain response for which there is now
considerable laboratory data available. The proposed method employs a pseudo-dynamic
finite element method in which the additional displacements due to liquefaction and
inertia forces are accounted for by applying additional forces that satisfy energy
principles. The procedure has been validated by applying it to field case histories
involving both one-dimensional sloping ground as well as two-dimensional cases. These
case histories include the Wildlife and the Heber Road sites, the Lower and Upper San
Fernando dams, the Mochikoshi tailings dams, the La Marquesa and La Palma dams in
Chile. It was found that the predicted and observed results in those case histories are in
reasonable agreement in terms of both the magnitude and pattern of displacements.
|
|---|
