Optimal intensity measure selection and probabilistic seismic demand models for dam-reservoir-layered foundation system

Abstract

This study evaluates the optimal intensity measures (IMs) for the seismic fragility analysis of dam-reservoir-layered foundation (DRLF) system. Probabilistic seismic demand models (PSDMs) for the DRLF system are developed that are plotted between engineering demand parameters (EDPs) and intensity measures. Optimal IMs are evaluated based on the selection measures: efficiency, practicality, proficiency, sufficiency, and hazard computability. In addition, finest IM, based on coefficient of determination (COD) in linear regression analysis, is assessed. Prior to that, the numerical model of DRLF system is developed in finite element method software, Abaqus. The concrete damaged plasticity model and Mohr-Coulomb failure model are adopted to provide the non-linear behavior to dam and foundation, respectively. The screening experiments are performed to examine the modeling parameter's effect on the tensile cracking failure in the dam. Ground motions are selected using the conditional mean spectrum approach. Dynamic structural analysis is performed through the incremental dynamic analysis method and a total of 110 non-linear time-history analyses are performed. Further, based on damage index, an analytical approach is used to construct the seismic fragility curves. Various damage states, based on tensile crack failure, are proposed. The screening experiment shows that concrete damping, rock to rock elastic modulus of different layers, concrete elastic modulus, concrete mass density and rock to rock frictional angle have a significant effect on the seismic response of the dam. While investigating the uncertainties associated with the modeling parameters of DRLF system, the elastic modulus of different layers should be chosen with care. Further, based on the IMs selection measures, ASI, EDA, MIV, IP_I and PGD are recognized as optimal IMs. However, ASI and EDA are observed to have excellent predictability and show best correlation for the considered EDPs of the dam.