High confidence of low probability of failure (HCLPF) values of equipment, representing the seismic capacities of the equipment, are the fundamental ingredient in seismic probability safety assessment (SPSA) and seismic margin analyses (SMA). In this paper, two methods for calculating the HCLPF values of equipment were investigated, fragility analysis, and conservative deterministic failure margin (CDFM). These methods are linear methods. Based on these methods, HCLPF value of equipment can be computed conveniently by scaling the results of the existing seismic analysis. For a nonlinear systems, the HCLPF values based on these linear scaling methods are unrealistic. For a complicated nonlinear equipment or structure, a detail nonlinear model was used to derive the seismic capacity. The results by this method are realistic, but cost calculation time. In this paper, a nonlinear model of reactor coolant system coupled reactor building was built. This model includes the steam generator and considers the nonlinear factors of steam generator such as gap in the supports, plasticity of hot leg and cold leg. Forced motion was applied to the base of reactor building. And seismic response of the steam generator was calculated iteratively by scaling the ground motion level step by step. Based on these calculations, a curve of load on the supports VS peak ground acceleration (PGA) can be obtained. Then based on these curves and allowable load of supports of steam generator, which derived from stress analysis on support of steam generator, seismic capacity of the supports of steam generator was determined. Then the HCLPF Value of the supports of steam generator was obtained by this nonlinear time history analysis and was compared with the results based on the CDFM. The two results were different. Therefore, the HCLPF seismic capacity of equipment with nonlinearity, such as gap nonlinearity, should be calculated by nonlinear time history method.

This content is only available via PDF.
You do not currently have access to this content.