Assessment of the cyclic strain approach for evaluating liquefaction triggering

Abstract

The cyclic strain approach was proposed in the 1980s as a potential alternative to the stress-based simplified liquefaction evaluation procedure. However, despite its fundamental basis and many positive attributes, it has not been embraced by practice. One reason for this may be the need to perform cyclic laboratory tests on undisturbed/reconstituted samples to develop a relationship among excess pore water pressure, cyclic strain amplitude, and number of applied strain cycles. Herein an alternative implementation of the strain-based procedure is proposed that circumvents this requirement, using a strain-based pore pressure generation model in lieu of laboratory test data. To assess the efficacy of the alternative implementation, several hundred small strain shear wave velocity (Vs) and Standard Penetration Test (SPT) field liquefaction case histories are evaluated. The results are compared with both field observations and with predictions from the stress-based procedures. It was found that the stress-based approach yielded considerably more accurate predictions compared to the cyclic strain approach. One likely reason for this is the strain-based procedure's inherent and potentially fatal limitation of ignoring the decrease in soil stiffness due to excess pore pressure when representing the earthquake loading in terms of shear strain amplitude and number of equivalent cycles.