Seismic earth forces against embedded retaining walls

This paper provides results from carrying out two-dimensional dynamic finite element analyses to determine the applicability of simple pseudo-static analyses for assessing seismic earth forces acting on embedded cantilever and propped retaining walls appropriate for New Zealand. In particular, this study seeks to determine if the free-field Peak Ground Acceleration (PGAff) commonly used in these pseudo-static analyses can be optimized. The dynamic finite element analyses considered embedded cantilever and propped walls in shallow (Class C) and deep (Class D) soils (NZS 1170.5:2004). Three geographical zones in New Zealand were considered. A total of 946 finite element runs confirmed that optimized seismic coefficients based on fractions of PGAff can be used in pseudo-static analyses to provide moderately conservative estimates of seismic earth forces acting on retaining walls. Seismic earth forces were found to be sensitive to and dependent on wall displacements, geographical zones and soil classes. A reclassification of wall displacement ranges associated with different geographical zones, soil classes and each of the three pseudo-static methods of calculations (Rigid, Stiff and Flexible wall pseudo-static solutions) is presented. The use of different ensembles of acceleration-time histories appropriate for the different geographic zones resulted in significantly different calculated seismic earth forces, confirming the importance of using geographic-specific motions. The recommended location of the total dynamic active force (comprising both static and dynamic forces) for all cases is 0.7H from the top of the wall (where H is the retained soil height).