Dynamic Centrifuge Tests to Evaluate Reinforcing Mechanisms of Soil-Cement Columns in Liquefiable Sand

Abstract

Four centrifuge tests were performed to investigate the reinforcing mechanisms of soil-cement columns in liquefiable sand. Two unimproved baseline models and two models improved with soil-cement columns were subjected to sine sweep and earthquake base motions of varying intensities to observe acceleration, pore pressure, lateral displacement, and settlement responses. The dynamic records were processed to derive the effective natural frequency of the profiles and to obtain the dynamic stress-strain responses for unimproved and improved soil. It was found that the shear reinforcement mechanisms of columns was not effective in reducing cyclic stress ratios in the treated soil; liquefaction triggering occurred nearly at same time for both unimproved and improved soil cases and the magnitude of the resulting soil settlement was not significantly reduced. When the bases of the columns were free to rotate, the columns rocked within the soil and produced negligible shear stiffening of the soil profile. When the bases of the columns were fixed against rocking, the columns deformed primarily in shear and flexure but the observed increase in natural frequency of the improved profile was much smaller than predicted by assuming shear strain compatibility between the columns and soil. Design equations that account for shear strain incompatibility are shown to give reasonable estimates for the observed natural frequency of the improved soil profile, and these equations suggest that the associated reduction in cyclic shear stress in the improved soil is small. The soil-cement columns did, however, remain intact and would have provided a means for supporting overlying structures even after liquefaction was triggering in the soil.