Reliability analysis of a collective decision-making scheme by Co-operation of NPP operators and automatic diagnosis system

With the continuous improvement of hardware equipment performance, human error of operation team has become the major factor threatening the safe operation of nuclear power plants. The argument of the extent to which computerized automatic diagnosis can reduce human error of operation team in decision-making has become a major concern in the actual engineering practice. This paper presents a decision-making scheme of HPR1000 nuclear power plant under accident conditions. An automatic diagnosis (AD) system is designed to assist the operation team in emergency decision-making under accident conditions. The fault modes and effects of the AD system are identified. Aiming at the fault modes of the AD system, a collective decision-making scheme coordinated by the AD system and the operation team is presented. The reliability of the proposed scheme is analyzed by using the Probabilistic Safety Assessment (PSA) approach. The analysis results are compared with the other two schemes. The weakness of the proposed scheme and the corresponding countermeasures are identified. On the premise of operator error probability and decision-making mode set in this paper, the analysis results show that the reliability of the collective decision-making scheme proposed in this paper is higher than that of the traditional team decision-making scheme in the main control room. While ensuring the stability of diagnosis results, the proposed scheme reduces the dependence of decision-making on personal roles. Quantitative analysis results also show that when the failure probability of an AD system is lower than 0.01, it will bring a positive benefit to reduce the probability of wrong decision-making. It also implies an acceptable standard of the failure probability of a fully automatic decision-making system.