The cross sections of five different model retaining walls used in this study are shown
in Figure 1. The broken lines show the initial locations and the hatched zones show the
displaced locations observed after failure of the model retaining wall. The total height
of the walls was 530 mm except for the reinforced soil-type wall which was 500 mm
high. The width of the base of the cantilever-type and gravity-type walls was 230 mm,
and it was reduced to 180 mm for the leaning-type wall. To the measure normal stress,
σ, and shear stress, τ, components of the earth pressures acting on the back of the facing
and the base of the walls, several two-component load cells were installed (see Figure
2 for the load cell locations in a typical cantilever-type wall). The details of the twocomponent
load cells were described by Tatsuoka et al. (1989). The load cells were located
within a 150 mm width along the center line of the wall to alleviate the effects
of side wall friction. To form a rigid structural body for the model wall, wooden blocks
(used for dummy walls) were stacked on both sides of the center section that was instrumented
with load cells and the blocks were reinforced with steel bars.
In conducting static loading tests on similar small-scale model walls, Tatsuoka et al.
(1989) lubricated the inside of the sand box by using a sheet of latex membrane smeared
with a thin layer of silicone grease. However, this method of lubrication was not
employed in this study because it was not expected to be effective for dynamic tests.
Instead, the width of the sand box in the direction of the wall facing was enlarged by
a factor of 1.5 to 600 mm, the earth pressures were measured at the center line as mentioned
in the previous paragraph. To adjust the dead load of the gravity- and leaningtype
model walls, extra weights were added near the center of gravity of these walls.
For model the reinforced-soil retaining walls, a grid of phosphor-bronze strips was
used as the model reinforcement (Figure 3). Each strip was 3 mm wide and 0.1 mm thick
and had a bending stiffness, EI , of approximately 0.003 N-m2 (Tatsuoka et al. 1989).
These materials and dimensions were selected to enable the tensile forces acting on the
model reinforcement to be measured, by attaching strain gauges at several locations on
the reinforcing strips. To form a lattice-shaped layer of model reinforcement that simulates
geogrids in actual field conditions, strips were glued together at 50 mm intervals in the direction parallel to the side wall and at 100 mm intervals in the direction normal
to the side wall. To mobilize friction between the reinforcement and the backfill soil,
as mobilized by actual geogrids, sand particles were glued to the surface of the strips.
Ten layers of model reinforcement having a length of 200 mm were used for the reinforced
soil, Type 1 model walls. On the other hand, the length of the top and fourth layers
were increased to 800 mm and 450 mm, respectively, for the reinforced soil, Type
2 model wall in order to increase the stability against overturning, as is the common
practice in Japan. It should be noted that any internal failure mode of the reinforcement,
such as a breakage or an excessive elongation, as well as that of the overall wall structure,
is not within the scope of the current paper