Stochastic Sensitivity of 3D-Elastodynamic Response Subjected to Random Ground Excitation

Abstract

The present study deals with structural sensitivity of dynamic response having uncertainties in design parameters subjected to random earthquake loading. Earthquake is modeled as stationary random process defined by Kanai–Tajimi power spectral density. The uncertain design parameters are modeled as homogeneous Gaussian process and discretized through 3D local averaging. Subsequently the Cholesky decomposition of respective co-variance matrix is used to simulate random values of design parameters. The Neumann expansion blended with Monte Carlo simulation (NE-MCS) is explored for computing response sensitivity in frequency domain. Application examples related to a building frame and a gravity dam are presented serving to validate the NE-MCS technique in terms of its accuracy and effectiveness compared to direct Monte Carlo simulation and perturbation method.