Post-liquefaction stability and deformation analysis of Duncan Dam

Abstract

Duncan Dam is a 39 m high zoned embankment and is founded on a deep deposit of unconsolidated sediments consisting of sands, silts, and gravels that are susceptible to liquefaction during earthquake loading. A comprehensive program of field, laboratory, and analytical investigations was conducted to evaluate the liquefaction potential of the foundation soils and performance of the dam during an earthquake. This paper describes the evaluation procedures and analyses and presents results of the post-liquefaction stability and deformation analysis of the dam. Analyses were carried out based on the total stress approach using two methods, namely the Lab method and Seed's method. The Lab method is a site–specific (direct), laboratory-based approach utilizing soil parameters from laboratory tests on undisturbed soil samples obtained in situ after freezing the ground. Seed's method is indirect and is based on field penetration data and past experience from earthquakes. The Lab method indicates that the residual strength of the liquefied sand is dependent on the effective confining stress, whereas Seed's method does not. Post-liquefaction deformations of the dam were computed using a pseudodynamic finite element procedure applying pre- and post-liquefaction stress–strain relationships and gravity and inertia forces satisfying the work–energy theorem. The Lab method predicts factors of safety of greater than 1.3 for the post-liquefaction limit equilibrium stability and acceptable deformations. On the contrary, Seed's method predicts factors of safety equal to or less than 1.0 and large deformations indicative of a flow slide. The merits of the two methods are discussed. Key words : sand, liquefaction, residual strength, performance, deformation.