Model Tests of Seismic Stability of Several Types of Soil Retaining Walls - Reinforced Soil Engineering: Advances in Research and Practice - page 337

This peculiar behavior could be explained qualitatively by considering the
sliding of the soil wedge along the failure plane as follows:

1. The broken curve in Fig. 21b is the base acceleration. When the soil
wedge started sliding (after point A in Fig. 21), its horizontal response
acceleration became smaller than the base acceleration. At the same
time, it slid down along the failure plane with negative (downward)
vertical acceleration (between points A and B in Fig. 21a).
2. Since reversal of the base acceleration took place, the sliding of the
soil wedge was terminated eventually (at point C in Fig. 21a). Before
the termination, the sliding movement was decelerated with positive
(upward) vertical acceleration (between points B and C in Fig. 21a).
3. The point B 0 in Fig. 21b is the point after which the horizontal response
acceleration of the soil wedge became larger than the base acceleration
(i.e., when the relative horizontal acceleration of the soil wedge to the
base was reversed). It was slightly different from the point B (when the
vertical acceleration of the soil wedge was reversed) in Fig. 21a,
possibly because the horizontal response acceleration in the underlying
nonsliding soil mass was not equal to the base acceleration. Similarly,
the point that corresponds to the point C in Fig. 21a (after which the
horizontal response acceleration of the soil wedge became equal to the
base acceleration) could not be clearly defined in Fig. 21b.

In the case with Fig. 21, the peak horizontal response acceleration was
mobilized while the sliding movement was decelerated (between points B and C
in Fig. 21a). In some of the other cases, however, the peak horizontal response
acceleration was mobilized while the sliding movement was accelerated
(between points A and B in Fig. 21a).
In Fig. 22 , correction for the effects of horizontal and vertical responses of
the soil wedge during the irregular shaking was made on the seismic coefficient k h
and the measured resultant force P a respectively; the k h -value was evaluated from
the (d h ) max -value; the P a -value was obtained at the moment when the (a h ) max -
value was mobilized, and it was corrected by dividing with a factor of “1

a v /g”,
where a v is the vertical acceleration of the soil wedge obtained at the same
moment as above (defined as positive when it induces downward inertia force).
The corrected relationships are represented by using open symbols in Fig. 22. For
reference, measured relationships between uncorrected k h - and P a -values that
were obtained at the moment when the base acceleration became its peak (i.e.,
when the a max -value was mobilized) are plotted by using solid symbols, and the
aforementioned theoretical relationships are also shown. It can be seen that, by
making a correction to the response of the soil wedge, the measured relationships
became much closer to the theoretical ones, in particular, in the region at high
seismic loads.

þ