？SA计算引擎的研发  

Development of Fault Tree Analysis Engine

作　　者：唐国锋[1] 高巍[1] 温珂伟 张经瑜[1] 

机构地区：[1]上海核工程研究设计院,上海200233

出　　处：《核电工程与技术》2015年第1期6-10,共5页Nuclear Power Engineering and Technology

摘　　要：本文介绍一款概率安全评价计算引擎的研发。该引擎运用先进的二元决策图／零压缩二元决策图（BDD／ZBDD）算法进行故障树的求解：首先将故障树规范化。然后将规范化的故障树结构转换为BDD/ZBDD结构，最后求解定量化结果及最小割集结果。引擎还具有重要度和敏感性分析（包括RAW／RRW／FV／FC重要度指标）以及不确定性分析（包括参数抽样和事件抽样）的功能。经测试，引擎能够求解大规模高复杂度的故障树模型，在较短的时间内得到较为精确的定量化和最小割集结果，以及准确的重要度敏感性和不确定性分析结果。A quantification engine for Probabilistic Safety Risk Analysis（PSA） is developed in this pa- per. This engine applies advanced Binary Decision Diagram/ Zero-suppressed Binary Decision Diagram （BDD/ZBDD） method to solve fault tree analysis. Firstly, it regularizes the fault tree. Secondly, it trans- forms the regularized fault tree into BDD/ZBDD structure. Finally, it solves the quantification result and minimal cut sets result. Also,the engine has functions to process importance ＆ sensitivity analysis, in- cluding RAW/RRW/FV/FC, and uncertainty analysis, including parameter sampling and basic event sam- pling. It is shown by the test that the engine is able to solve fault tree models of large scale as well as high complexity, and obtain accurate risk, cutsets, importance ＆ sensitivity analysis and uncertainty anal- ysis results within a shorter period of time.

关 键 词：概率安全评价 BDD/ZBDD 计算引擎 最小割集 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]