RESEARCH ON PROBABILISTIC SAFETY ASSESSMENT (PSA) METHOD OF SAFETY CLASS DISTRIBUTED CONTROL SYSTEM (DCS) IN NUCLEAR POWER PLANT BASED ON FUZZY MARKOV PROCESS

Safety class DCS(Distributed Control System) in nuclear power plant is characterized by multi-state interaction, complex function and unknown fault mechanism, which limits the use of tradition PSA(Probabilistic Safety Assessment) method such as event tree and fault free. An improved dynamic probabilistic security simulation method is proposed based on fuzzy Markov process to quantitative evaluation the risk of Safety class DCS in nuclear power plant, which resolved the failure uncertainty problem without in-depth understanding of failure mechanism. Firstly, the fuzzy fault free reliability evaluation model is established based on fuzzy theory and FTA(Fault Tree Analysis), which describes the probability and degree of failure with fuzzy number. Then, the fuzzy fault tree is transformed into fuzzy Markov chain, this model is described mathematically, and the reliability indexes are simulated and analyzed with Monte Carlo simulation method, the change of system safety performance in long-term operation is obtained. Finally, the scientificity and effectiveness is verified by the example of redundant analog output module of safety class DCS in nuclear plant.