摘要
分析了故障传播的机理并推断出其有限步扩散性,提出了一种K步故障扩散定位算法。对系统故障有向图进行了分片处理,结合故障传播概率矩阵对故障嫌疑节点进行故障扩散模拟寻找最佳扩散路径和最大嫌疑节点。使用C++编程和Matlab工具实现了KFL算法并给出部分具体工程源代码,对实际工程项目故障定位进行了仿真模拟。该算法可同时模拟诊断单点故障和多点故障,与传统故障定位的逆向算法不同,该算法是一种新型的正向思维的图匹配算法。
The fault pervasion mechanism is analyzed and limited-steps propagation is deduced. A k-steps fault pervasion and localization algorithm is put forward, which fractionalizes the fault directional graph and finds the optimal pervasion route or the most suspicion nodes with the fault propagation probability matrix. KFL algorithm is realized with C＋＋ and Matlab, which can also locate failures of actual project. This algorithm can simulate and diagnose the single node fault and many node faults. Comparing with the traditional converse fault detection algorithms, it is a late-model straight thinking and graph matching algorithm.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第8期1900-1902,1993,共4页
Computer Engineering and Design
关键词
故障传播
故障有向图
扩散定位算法
概率矩阵
仿真
fault propagation
fault directional graph
pervasion and localization algorithm
probability matrix
simulation
