Model Tests of Seismic Stability of Several Types of Soil Retaining Walls - Reinforced Soil Engineering: Advances in Research and Practice - page 334

Figure 19 Continued.

the backfill. The active failure state could be defined as the state either when the
failure plane is about to develop (for f ¼ f peak ), or after the failure plane has
developed in the backfill, where the f-values have dropped to f res .
Phase difference in the shaking tests in the vertical distribution of
horizontal response accelerations of backfill would be one of the reasons for the
different P a -values from the tilting tests. In addition, as discussed by Tatsuoka
et al. (1998), different from the case of tilting tests, the earth pressure acting on
the back of the facing in the shaking tests is controlled largely by an interaction
between dynamic response of the backfill and the wall structure. In fact, the
P a values defined under the three different conditions as mentioned above

were, in general, different from each other.
In Fig. 20 , the peak horizontal response acceleration (a h ) max in the backfill-
inducing outward inertia force was compared with the peak base acceleration
a max in the irregular shaking tests. The (a h ) max -values were evaluated based on
the records of an accelerometer located near the mass center of the soil wedge (in
the unreinforced zone for reinforced soil RWs; i.e., the B-wedge shown in Fig. 8b )
above the failure plane. With the increase in the shaking level, the (a h ) max. -value
became gradually smaller than the a max -value. In particular, after the failure
plane was formed in the backfill, the rate of increase in the (a h ) max -value was
significantly reduced, or even the increase stopped temporarily, due possibly to
sliding of the soil wedge along the failure plane.
For gravity-type and reinforced soil-type 1 RWs, results from sinusoidal
shaking tests are also shown in Fig. 20. Note that, using 20 cycles of sinusoidal