基于离散动态事件树的风险指引的安全裕度分析方法研究

风险指引的安全裕度是近十年来核工业界提出的新的安全理念。本文阐述了基于离散动态事件树的风险指引的安全裕度分析方法，给出该方法下核燃料包壳失效概率均值和标准差的数学表达式。针对简化压水堆模型下的全厂断电事故，提出了基于离散动态事件树的风险指引的安全裕度计算流程，计算了两种离散动态事件树分支规则下燃料包壳失效的风险指引的安全裕度及其不确定性。计算结果表明，不同的分支规则、模型参数分布、系统程序最大时间步长对核燃料包壳失效概率均值和标准差均有显著影响。提出了一种改进的可变概率阈值的分支方法，以更好地平衡风险指引的安全裕度分析过程中计算精度与计算资源的匹配问题。

Research on Analysis Method of Risk-informed Safety Margin Based on Discrete Dynamic Event Tree

Risk-informed safety margin is a new safety concept of nuclear power industry in recent ten years. By quantifying the risk informed safety margin, it may provide a new technical idea to support the management decisions for nuclear power plant life extension, long term operation, power uprate and some other safety problems. Methodologically, the calculation method of risk informed safety margin is to generate a series of risk based scenarios according to the state of knowledge, use the system model to calculate the specific safety parameters value of each scenario, then analyze the relationship between the load and capacity distribution of the safety parameters, and statistically calculate the probability that the load exceeds the capacity of safety parameters. There are several methods to generate the scenarios, which include Monte Carlo method, discrete dynamic event tree method, combination of Monte Carlo and discrete dynamic event, adaptive sampling and so on. The analysis method of risk informed safety margin based on the discrete dynamic event tree was described in this paper, which can better reflect the time characteristics of accident evolution. The mathematical expressions of mean and standard deviation of failure probability of nuclear fuel cladding were proposed. Under the simplified pressurized water reactor (PWR) model, the process of risk informed safety margin for station black out accident based on discrete dynamic event tree was proposed, then the mean value and uncertainty of nuclear fuel cladding failure probability under two kinds of branching rules were calculated. The influence of the branching rules, key model parameter distribution and maximum time step of system program were analyzed. The calculation results show that the risk informed safety margin can be automatically calculated by the discrete dynamic event tree method and computer programming for specific safety parameters, the distribution of key parameters and the setting of system program have a significant impact on the mean value of failure probability of fuel cladding. With different branching rules, the calculation results of the mean value of fuel cladding failure probability are different. The finer the branching is, the smaller the mean value of fuel cladding failure probability is, the longer running time and the larger storage space of computer are required. Finally, an improved branch method of variable probability threshold was proposed to balance the matching problem between calculation accuracy and calculation resources. It is necessary to further study the uncertainty evaluation of the calculation method for risk informed safety margin, the verification of the calculation results and the improved efficiency algorithm to balance the calculation accuracy and resources.