In Figure 14, the observed critical tilting angles that caused outward displacements
at the facing top as large as 5% of the total wall height are also compared with the predicted
critical tilting angles which resulted in a factor of safety of unity against overturning.
The tilting angles were converted to seismic coefficients by using Equation 2. For
the same model wall type with the same surcharge, the predicted critical seismic coefficients
for shaking and tilt table tests are the same according tothe pseudo-static analysis.
For the same wall type, however, the observed critical seismic coefficient is smaller
for the tilt table tests than for the shaking table tests (Figure 14). This may have been
caused by an essential difference in the testing conditions; i.e. the seismic horizontal
force was simulated pseudo-statically in tilt table tests, while it was applied periodically
for approximately ten seconds in the shaking table tests. For the same seismic coefficient,
the loading condition was more severe with respect to wall stability for the tilt
table tests than the shaking table tests; however, the opposite was true when the effects
of amplification were large in the shaking table tests. It seems that for the test cases presented
in the current paper, the effects of the former factor are more predominant than
the latter factor.
Analyses of the recorded data, on the effects of amplification and phase difference
in the response acceleration during shaking, are in progress. These analyses will be extended
to the effects of frequency and irregularity of the actual earthquake motion for
practical applications.