It may also be seen from Figure 14 that the observed critical seismic coefficients for
the different model walls scatter over a large range, while the corresponding values predicted
by the pseudo-static stability analysis are within a relatively narrow range. This
result indicates that the pseudo-static stability analysis cannot evaluate important aspects
of the seismic stability of different types of walls. Particularly, the stability of the
reinforced soil-type model walls in the shaking table test is underestimated by the current
pseudo-static stability analysis in comparison with the gravity- and leaning-type
retaining walls.
In Figure 15, the observed critical seismic coefficients, kh-cr(obs) , are compared to the
predicted coefficients, kh-cr(cal) , against sliding for all of the models for δ = 3/4φ. It may
be seen that the predicted values, kh-cr(cal) against sliding for all of the model walls were
larger than the observed values, kh-cr(obs) , which is consistent with the fact that the observed
failure mode in all of the tests was overturning. However, these predicted values
are not very reliable, because the values are too sensitive to the interface friction angle
of the model walls except for the reinforced soil-type (Figure 13).
It is to be noted that for the reinforced soil-type model walls, although the reinforced
backfill was assumed to behave as a rigid body when the factors of safety were evaluated,
during the tests overturning of the wall accompanied simple shear deformation of
the reinforced backfill (Figures 1d and 1e). This behavior suggests that the horizontally
placed reinforcement layers do not effectively resist such simple shear deformation of
the reinforced backfill. In evaluating seismically induced residual displacement of geosynthetic-reinforced
soil retaining walls, therefore, simple shear deformation of the reinforced
backfill should be taken into account as stated by Tatsuoka et al. (1996a).