Numerical prediction of undrained cyclic triaxial experiments on saturated Kasai river sand using two constitutive models of liquefaction

The paper presents a numerical simulation of stress-controlled undrained cyclic tri-axial tests conducted on local sand at various relative densities and CSR values. Two constitutive models for liquefaction, namely, Finn-Byrne model and PM4 sand model are used for the simulation. The results from both the material models are compared with the cyclic test results in terms of the stress strain loops, excess pore pressure ratio, and effective stress path. The chosen models have shown good predicting capabilities for predicting cyclic responses of a sand in terms of stress path and pore water pressure generation but lacks in predicting the stress strain behavior. Also, in terms of simulating the behavior of sand in pre- and post-liquefaction regime, it is found that PM4 sand model is more efficient than Finn model for all the relative densities. Hence, it is recommended that Finn model may be only utilized when a sand fails due to flow liquefaction, whereas PM4 sand model can predict both, flow liquefaction and cyclic mobility, accurately. Finally, the liquefaction resistance curves for different relative densities are obtained using experimental values and PM4 sand model. These curves are thus proposed for any sand with index properties similar to the particular sand considered for the study.