Model Tests of Seismic Stability of Several Types of Soil Retaining Walls - Reinforced Soil Engineering: Advances in Research and Practice - page 316

time scale were adjusted so that the base acceleration has a prescribed maximum
amplitude with a predominant frequency of 5 Hz. Each model was subjected to
several shaking steps, where the maximum amplitude of the base acceleration
was initially set to 100 gal and increased at increments of 100 gal. Shaking was
terminated when the wall displacement became considerably large. During
shaking, the deformation of the wall and the surrounding sand layers was
monitored up to the ultimate failure state through the side wall of the sand box by
means of a digital video camera. Stresses acting on the facing, displacements of
the wall, response accelerations of the wall and the backfill, and tensile forces
acting in the reinforcements were also recorded.
Results from these irregular shaking tests were compared with the previous
test results (Koseki et al., 1998a, 1999), where seismic loads were applied either
by tilting the sand box to simulate pseudo-static loading conditions or by shaking
the sand box with a sinusoidal base acceleration at a frequency of 5 Hz (Fig. 5b) .
In the tilting tests, the sand box was tilted continuously at a rate of approximately
1.0 /min until a considerable displacement of the wall was observed. Based on

the pseudo-static approach, the observed seismic coefficient k h in the tilting tests
was defined as
k h ¼

8

tanu ð 1 Þ

where uis the tilting angle of the sand box. In the sinusoidal shaking tests on the
cantilever-type wall model, the amplitude of the base acceleration was initially
set to 25 gal and increased at an increment of 25 gal. For the other models, the
initial base acceleration was set first to 50 gal, and the increment was also doubled
to 50 gal in order to minimize possible effects of the previous shaking history on
the behavior at the subsequent loading stages. At each acceleration level, the
same amplitude of base acceleration was maintained for about 10 sec. In this
study, effects of previous shaking histories on the test results were assumed to be
insignificant. The observed seismic coefficient k h in the shaking table tests was
defined as

Þ
where a max is the single amplitude of maximum base acceleration at the active
state (i.e., when the inertia force of the backfill is acting outward) for each
shaking step, and g is the gravitational acceleration.

k h ¼ a max = g ð 2

6 TEST RESULTS AND DISCUSSIONS

6.1 Failure Pattern

Figure 6 shows the residual displacement of the wall and the residual deformation
of the backfill, which were observed at the end of irregular shaking step when a