Repeatability of centrifuge tests containing a large, rigid foundation subjected to lateral spreading

Four centrifuge tests involving liquefaction-induced lateral spreading of loose sand against a large, rigid foundation are summarized with regard to the repeatability of ground motions, porewater pressures, and lateral displacement magnitudes. The base input motions delivered to the models exhibited a large coefficient of variation (COV) of 0.27, chiefly as a result of variations in the shaker hydraulic system, rather than differences in model construction (which exhibited a COV of 0.07 for the relative density of the loose sands). Variations in ground acceleration response, porewater pressure response, and lateral displacement magnitudes were dominated by variations in base motions, with COV values ranging from about 0.01 to 0.47. Nevertheless, more than 90% of the COV values were less than 0.30, consistent with or less than the COV of the base motions. In addition, the measured COV values are consistent with COV values of other geotechnical measurements reported in the literature. This is especially encouraging because centrifuge tests are relatively complex boundary tests, and many COV values reported in the literature involve element tests. Based on these findings, one must carefully monitor and control base motions (input to the soil container) in order to obtain repeatable results in a centrifuge testing program.