The objective of this research program is to evaluate several different methods for predicting pore water pressure buildup in saturated cohesionless soils when subjected to cyclic shear stresses. Such stresses are generated as a result of the strong ground motions resulting from a large earthquake. The buildup in pore water pressure can result in the failure of building foundations, as well as of underground lifelines such as water, sewer, power, and telecommunication lifelines. The damage results from the liquefaction of the soil when the pore pressure reaches a critical level. Thus, it is very important to be able to predict the generation of pore water pressure, and the potential for the soil to liquefy. Considerable effort has gone into the development of prediction methods, but the validation of these methods is far from complete due to a lack of data against which to check them. In April 12, 1986, the Molikpaq, a hydrocarbon exploration island in the Beaufort Sea, was subjected to ice loads which had not been predicted, and which caused the sand core to partially liquefy. The structural performance of this sand island was closely monitored by the owner, Gulf Canada, and as a result, an excellent data set exist against which to check the methods proposed for calculating pore pressure generation and liquefaction. This data set is being made available for research purposes, and a research program has been initiated by the following Canadian Government organizations: The Geological Survey, the Subcommittee on Marine Development of the National Research Council's Associate Committee on Geotechnical Research, and the Panel on Energy Research and Development. A similar study supported by the National Science Foundation is closely coordinated with the Canadian initiative. This joint effort between the Rutgers University and Prof. Prevost at Princeton University, is part of this study. Coupled dynamic field equations calculate pore fluid and soil skeleton response concurrently, accounting for dynamic interaction between the two phases. Multi-yield surface plasticity is used to accurately model the nonlinear, hysteretic behavior of the soil skeleton. The numerical calculations are performed using a general purpose, three-dimensional nonlinear finite element program. The results from both the Canadian and NSF studies will be presented at a workshop to be held in Vancouver in August, 1988.
