Abstract:
This research report proposes and validates the use of the LSDYNA finite element code as a rigorous approach for analyzing the seismic soil-structure interaction of spillway walls adjacent to embankments, moving beyond traditional pseudostatic methods. To benchmark the performance and accuracy of LSDYNA for this application, a 3-D finite element model was created to replicate a centrifuge experiment conducted at the University of California, Berkeley, which measured the seismic response of retaining walls with a sand backfill. The study specifically evaluates two of LSDYNA's nonlinear soil material models: Material Model 25 (a geologic cap model) and Material Model 16 (a Mohr-Coulomb yield surface model). The results from the finite element analysis, including acceleration time histories and wall moments, were compared with the experimental data, showing that LSDYNA can favorably model the complex behavior, with Material Model 16 providing a better prediction for the flexible wall response. Following the validation, parametric studies were conducted using the LSDYNA model to investigate the effects of reversing seismic record polarity and varying wall stiffness and mass. The report also details the creation of a full 3-D finite element model of Scoggins Dam to demonstrate the application of this validated methodology to a real-world, complex spillway structure with zoned soil materials.
