Engineering evaluation of deterministic ground-motion simulation based on the dynamic response of Pine Flat dam

Abstract

Despite significant advances in deterministic ground-motion simulations, their application in engineering practice is still constrained due to reliability concerns. Current studies on deterministic ground-motion simulations typically evaluate simulated ground motions by comparing single parameters like peak ground acceleration (PGA), peak ground velocity (PGV), and spectral acceleration. However, limited research has explored whether these simulations can accurately replicate the dynamic responses of complex structures—a critical requirement for practical engineering applications. This study evaluates the feasibility of using fully deterministic numerically simulated ground motions for dam dynamic analysis. Taking the dynamic response of a gravity dam as a benchmark, we propose several structural response parameters to quantitatively assess the reliability of simulated ground motions. The 1994 Northridge earthquake is simulated with a newly proposed multidimension source model and the refined velocity structures. The simulated ground motions, along with recorded data at 32 stations, are used as input ground motions for the seismic response analysis of the Pine Flat gravity dam. Results from the linear-elastic seismic analysis show that the principal stress distribution in the dam due to simulated ground motions is generally comparable to that observed with recorded ground motions. Furthermore, the nonlinear seismic analysis shows that the maximum relative displacements and damage of the dam subjected to both the recorded and simulated ground motions are in good agreement at most stations. The comprehensive validation results demonstrate that deterministic broadband ground-motion simulations offer a promising and reliable approach for performance-based seismic analysis of engineering structures.