Research On Dynamic Probabilistic Safety Assessment Methodologies Of Nuclear Power Plant For Digital Instrumentation And Control Systems

Dynamic probabilistic safety assessment (PSA) methodologies are defined as those that can account for the coupling between the triggered or stochastic logical events in system reliability modeling through explicit consideration of the time element in system evolution. Dynamic PSA can overcome the shortcoming of traditional PSA methodologies and effectively validate the reliability and safety of digital instrumentation and control systems. Two methodologies, i.e., Markov/Cell-to-Cell Mapping Technique (Markov/CCMT) and Dynamic Flowgraph Methodology (DFM) have been investigated, which are two most appropriate techniques for the reliability and safety modeling of digital instrumentation and control systems that were proposed by Nuclear Regulatory Commission report NUREG/CR-6901.Suppose that a small-break loss of coolant accident of main feed-water piping happens when the system is under full power operating, we conduct the PSA for the AP1000main feed-water system with the Markov/CCMT methodology. In the process of PSA, a main feed-water pump and a main feed-water regulating valve have been chosen as the failure components. Firstly, the transition probabilities of the component state combinations have been calculated. Then all the possible failure scenarios have been considered according to the system states and simulate these scenarios with system model built on3KeyMaster simulation platform. The next step is to calculate the cell transition probabilities. At last, the probabilities of the top events, steam generator at low level and steam generator at high level have been obtained. Through the Markov/CCMT methodology, some important failure cases can be avoided from being lost, the competition of top events can be solved, and the accurate time for failures can be obtained as well.In this study, the impact on steam generator water level at full power level has been considered due to components failure in main feed-water system of CPR1000. And the DFM methodology has been conducted to implement PSA analysis. Firstly, choose the critical system physical variables, and construct decision tables and discrete node states according to the data produced by the simulation platform of CPR1000nuclear power plant, and then the deductive analysis and inductive analysis according to the constructed DFM model have been conducted. Through deductive analyze all the prime implicants of the top event, i.e., steam generator at high-high level can be obtained. The prime implicants can interpret all possible factors which lead to top events, meanwhile, when combined with different component failure models, can instruct operator to take some appropriate strategies to avoid the accidents. The inductive analysis is similar to failure model and effect analysis (FMEA), which can obtain the effect on system under different initial conditions.Through the research of this study, the theory and application of Markov/CCMT and DFM methodology have been determined, which will provides a theoretical basis for developing the dynamic PSA software with proprietary intellectual property rights.