Abstract:
This study presents an effective stress site response analysis method that has been refined for use without the need for complex and involved experimental testing. The effective stress model of the analysis requires the determination of five parameters for which there are no correlations to more general soil properties available. By using a trial and error fitting technique these parameters may be successfully determined without the need for testing.The technique proposed involves use of a cyclic shear simulation program incorporating the same effective stress model as the full site analysis program. By back fitting the cyclic shear simulation, and resulting liquefaction curve, to a liquefaction resistance curve determined for a particular sand, these parameters may be obtained by a trial and error approach. The liquefaction resistance curve may be determined either experimentally using a cyclic triaxial test or empirically, with experimentally being the preferred but probably the Iess likely option.Using empirically generated liquefaction curves, the full effective stress site response analysis method was investigated in relation to two recent large magnitude events. These case studies were Treasure Island in the San Francisco Bay, subjected to the 1989 Loma Prieta earthquake and the Edgecumbe, New Zealand earthquake of 1987. Both cases experienced significant liquefaction and hence were well suited to the effective stress approach. The predictions yielded close agreement with the observed situations in both cases, for both liquefaction prediction and in the prediction of general site response characteristics. These two case studies inspired confidence in the prediction of this analysis technique.The validity of a number of common empirical relationships was investigated for the New Zealand situation with varying results. These investigations Ied to the conclusion that relying totally on empirical data may not be advisable, particularly if no corroborating evidence is available.The influence of effective stress properties was investigated and compared to total stress solutions. The major influence of significant pore pressure generation is in the lengthening of the predominant response periods of a site and the general increased damping of the surface motion.