Reliability assessment of passive residual heat removal system of IPWR using Kriging regression model

Abstract

Passive safety systems have been incorporated into new design nuclear power plants to enhance the power plants’ inherent safety. However, due to the lack of the experimental data and operational experience, evaluating the reliability of passive systems is a challenging task. Many uncertainties are involved in the thermal hydraulic process during startup and operation. This may make the system unable to accomplish its expected function even though the hardware is available, known as functional failure. The evaluation of system functional failure probabilities based on the direct Monte Carlo method may be computationally impractical. In order to reduce the computational load, the Kriging regression model was constructed to avoid a large number of thermal hydraulics simulations. Compared with other metamodels, the Kriging regression model obtained better accuracy. The proposed method was applied to evaluate the reliability of a passive residual heat removal system of IPWR200 during a station blackout accident. The Kriging regression model not only ensures the accuracy of the reliability assessment, but also greatly reduces the runs of T-H codes. Furthermore, global sensitivity analysis aimed at determining the contributions of input parameters was carried out. Results indicate that the probability of passive residual heat removal system failing functionally is estimated to be 1.94e-4 and that insufficient decay heat removal is the most likely failure mode. Sensitivity analysis results identified five key uncertain parameters, which is significant for system performance. The sensitivity information can provide guidance for enhancing passive safety systems design and operation, which is of great importance to the safety and reliability of passive safety systems.