Seismic Probabilistic Risk Assessment of Reservoir Structures Considering Earthquake Hazard and Updated Fragility by Bayesian Inference

摘要: 1.1 BackgroundSeismic probabilistic risk analysis (SPRA) is an integral part of the risk assessment of critical engineering structures such as nuclear power plants, dams, bridges. The aims of SPRA are to estimate the frequencies of occurrence of earthquake-induced accidents that may lead to different levels of damage and to identify the key risk contributors so that necessary risk reductions may be achieved (Kennedy and Ravindra, 1984). However, it is critical that engineers be able to assess quantitative of seismic risk to structures, and the associated risks due to seismic uncertainties present in many aspects of the assessment, to a high degree of confidence.The basic elements of SPRA are seismic hazard analysis, seismic fragility evaluation and the system analysis and quantification. The seismic fragility of the structural system is the pivotal component of seismic risk assessment which expresses the relationship between a ground motion intensity and the corresponding probability of failure in specific performance criteria. The commonly used method for fragility evaluation is the lognormal model base on design limit state (Kennedy et al., 1980). However, the conventional method relies a large number of data from past studies, experimental results and numerical analysis which may not be available for all structure. On the other hand, a large number of dynamic analyses of structure require significant computational time. Bayesian Inference technique with Markov Chain Monte Carlo (MCMC) simulation is used in this investigation to reduce the total number of time history analysis by adding prior information of fragility curve with simulation data …