Evaluation of cyclic shear stress-strain using inverse analysis techniques in dynamic centrifuge tests

Evaluation of shear stress-strain characteristics in soils is paramount to the fundamental understanding of soil behavior in dynamic centrifuge tests. Three inverse analysis techniques for evaluating cyclic shear stress-strain response using data from accelerometer arrays are introduced and key factors that require consideration in the applied condition are discussed. Using a series of dynamic centrifuge tests on soft soil sites with stiff ground reinforcement, the data process for back-calculated displacements from acceleration records is presented. The back calculated displacements are compared to the recorded displacements in several cases with excellent agreements. The three inverse analysis techniques are used to estimate the dynamic shear stress-strain responses in the free field and with in the reinforced ground in two centrifuge tests with different shaking motions of varying magnitudes. The results demonstrate that as long as accelerometer spacing and sampling frequency are sufficient to the active mechanics, the calculated shear stress-strain responses from three inverse techniques are consistent and therefore likely accurately reflect the basic characteristics of interest. In these tests it is observed that shear stiffness decreased at shallower depths, consistent with the expected effects of confining pressure on shear stiffness. It was also observed that the amplitudes of shear strain in the free field are considerably larger, and stronger nonlinear features were observed in the stress-strain loops as compared to the motions with in the reinforced soil. With comparisons among the three inverse analysis techniques, the evaluation from cubic spline approach was more sensitive and the linear and weighted residual techniques produced more reasonably consistent results.