适用于Living PSA快速求解最小割集的算法研究  

FaultTree analysis algorithm for Living PSA

作　　者：胡文军[1] 喻宏[1] 任丽霞[1] 钱鸿涛[1] 宋维[1] 乔雪冬[1] 

机构地区：[1]中国原子能科学研究院,北京102413

出　　处：《核科学与工程》2009年第1期39-42,共4页Nuclear Science and Engineering

摘　　要：Living PSA是现今核电厂安全分析的热点之一,核电厂本身是一个非常复杂的系统,其相应的Living PSA模型也相当庞大,这样在Living PSA的应用中,计算软件的分析速度成为限制其发展的瓶颈之一。本文基于现行常用的PSA分析算法,设计了一个能够快速求解故障树的算法。该算法首先将生成的故障树转化成用于计算的标准故障树,然后对标准故障树进行模块化,生成四类基本的独立子树,最后调用优化过的最小割集算法并采用多叉树方法进行割集的合并和吸收。该算法目前已应用于核电站快速风险分析软件平台NFRISK的开发中,并通过应用于中国实验快堆的一些系统和设备的故障树分析对其进行了初步验证。Living PSA was one of focuses in the safety analysis of nuclear plant. Because of the complexity of nuclear plant system, the corresponding model of Living PSA was very complicated. To increase the ability of analysis software, a new algorithm was designed. Firstly, fault tree was converted into standard fault tree for calculation. Second- ly, the fault tree was modularized and four kinds of modules was gained. At last, cut sets were united and simplified by the new algorithm, then the minimum cut set was get. The algorithm was used in the platform NFRisk, and was validated by the basic system.

关 键 词：LIVING PSA 故障树算法 NFRisk 

分 类 号：TL43[核科学技术—核技术及应用]