Liquefaction induced lateral spreading in large-scale shake testing
MetadataShow full item record
Liquefaction and lateral spreading of gently sloping ground remains a puzzling problem. Data collected from sites damaged due to previous earthquakes, scaled experiments and computational modeling have given a general picture of how the soil behaves during liquefaction and lateral spreading but many questions are yet to be answered. A large scale 1-g laminar box system capable of simulating response of flat ground as well as gently sloping ground conditions of up to 6m depth subjected to shaking has been developed by the researchers at University at Buffalo. The system involves a laminar box made of 24 laminates stacked together, a shaking base, computer controlled high speed actuators, strong floor, strong wall, dense array advanced instrumentation, and a novel laboratory hydraulic fill method to construct the sand simulating underwater deposition with density control. It is suitable to study liquefaction response and its effects on foundations in saturated loose sands at a relative density of 35~50% or less. Two large scale liquefaction experiments, involving nearly 5m deep sand deposit, one simulating level ground and the other simulating gently sloping ground, are presented. The tests were conducted using 6.2m high laminar box with base dimensions of 2.75m wide and 5m long. The soil deposit was prepared using hydraulic fill method at 35~50% range. A dense array of accelerometers, pore pressure sensors, potentiometers (for displacement measurements), high speed video cameras, and a novel MEMS type sensor array were used to monitor the soil response. Liquefaction occurred in level ground test with little lateral deformation. Liquefaction and large lateral displacements occurred in the sloping ground inclined at 2 degrees. A comparison of the soil response in level ground test and sloping ground test revealed distinctly different soil behavior. The effect of initial static shear in SG-1 was highly responsible for the change in behavior. Failure in the level ground condition was reached when the soil fully liquefies, whereas in the sloping ground condition, failure was reached at two sliding planes, at times when the soil at those elevations was not fully liquefied, but was softer because it had built significant pore pressures, with pore pressures ratios ranging between 40 and 70%. Two nearly distinct sliding planes were observed in the sloping ground test. The residual shear strength of the soil was evaluated for each one of the two sliding events that took place during the test. The stress-strain path remarkably resembles that of an undrained sandy soil with contractive/dilative behavior, suggesting a contractive/dilative response of the soil layers at the sliding planes. Residual shear strength of the soil layers at the sliding planes follows the trend of the quasi-steady state shear strength for the relevant range of confining pressures. Its value corresponds to 80% of the driving shear stress. Vertical ground settlement data points towards occurrence of void redistribution in some regions of the soil deposit. The experimental results presented herein demonstrate the effectiveness of full-scale testing to investigate the mechanism of lateral spreading and confirm the advantage of using the University at Buffalo's geotechnical laminar box to simulate and understand the physics of even large and complex problems such as soil-structure interaction.