Reliability Assessment of NPP System for Risk Management Based on an Information Reasoning Methodology

Abstract

Abstract Risk management (RM) is an effective means of safety assessment and reliability management of nuclear power plant (NPP). During the operation of NPP, the main contents of RM work are as follows: 1. Real-time reliability assessment of NPP system, and evaluate NPP risk, so as to ensure that the NPP system can complete the specified functions to deal with the consequences of system failure. 2. Complete the diagnosis and mitigation of abnormal faults, so as to ensure that the NPP system can return to a safe state as soon as possible in the case of abnormal operation of NPP. With the continuous improvement of the requirements of online RM, LPSA (Living PSA) technology, as an effective tool of RM, has been developed and provides more powerful support for the safe operation of NPP. In order to evaluate the reliability of NPP system in real time, LPSA model needs to be updated to match the state of the system. Most of the existing LPSA model updating techniques are based on the premise that the system information is completely knowable and the information is accurate. However, in the process of NPP operation, the accessibility of system information and the accuracy of information are difficult to meet the requirements of updating the existing LPSA model. How to update the LPSA model in the case of incomplete and inaccurate information is one of the difficulties in the application of online LPSA technology. Especially when the NPP is in abnormal conditions, it is more urgent for operators to obtain correct system information in real time to recognize the available state of the plant system. In fact, the required system information needs to be obtained by the operator through his or her cognitive behavior process, such as the fault diagnosis process. At the same time, this information can be used to help operators complete the diagnosis of system status and the evaluation of system reliability. In this study, an information reasoning method is proposed to evaluate the probability of various failure modes under the information known to the operator through the modeling and analysis of the cognitive behavior process of the operator. The method can assist the operator to make the correct fault judgment to complete the fault diagnosis and fault mitigation. In addition, the information obtained by the operator in the process of cognitive behavior, and the results of information reasoning can be used to update the LPSA model in real time, so as to obtain the real-time reliability evaluation results of the NPP system.