Abstract
Centrifuge research conducted at Rensselaer Polytechnic Institute (RPI) is summarized focusing on the effect of sand permeability and of weak aftershocks on earthquake-induced lateral spreading. One-dimensional shaking tests were conducted in a laminar box at a centrifugal acceleration of 50 g, simulating a 10 m thick homogeneous layer of clean Nevada sand of relative density 40-45% inclined a few degrees to the horizontal. Seven centrifuge experiments were done using either water or a viscous pore fluid fifty times more viscous than water, thus varying the soil permeability by a factor of fifty and simulating either a coarse prototype sand or a fine sand in the field. These seven tests were subjected to the same shape of base input excitation but with peak input accelerations ranging from 0.18 g to 0.46 g between tests. An additional special centrifuge test of a similar model with viscous pore fluid was subjected to a strong base "main shock", followed after a few seconds by two "weak aftershocks". Excess pore pressures, accelerations, settlements and lateral ground deformations were measured. The results are discussed in detail and several correlations are presented between testing and measured parameters such as thickness of liquefied soil, ground surface settlement, lateral ground displacement and input peak acceleration. The results of the special test, together with other evidence, provide strong support to the hypothesis that reported continued ground deformations from lateral spreads in the field are often caused by continued weak vibration or aftershocks occurring after the main shock.
