基于粗糙集和图论的ZPW-2000A轨道电路故障诊断模型

FAULT DIAGNOSIS MODEL OF ZPW-2000A NON-INSULATED TRACK CIRCUIT BASED ON ROUGH SET AND GRAPH THEORY

目前,对于ZPW-2000A无绝缘轨道电路的判别方法主要是依赖人工测试和分析,存在主观性强、判别效率低等缺陷。为了解决上述问题,将粗糙集理论和图论知识相结合引入到ZPW-2000A无绝缘轨道电路的故障诊断中。通过构建故障诊断决策表以及加权多部决策表图及其邻接矩阵,获得一种基于邻接矩阵的属性约简方法以及识别核属性的方法,并利用故障规则的分级以及故障信息覆盖度完成故障规则的提取。研究表明,该算法条理清晰,计算简便,能够有效减少算法复杂度。

At present, the method of judging ZPW-2000A non-insulated track circuit mainly relies on manual testing and analysis, which has strong subjectivity and low discrimination efficiency. In order to solve the above problems, this paper introduced rough set theory and graph theory into the fault diagnosis of ZPW-2000A non-insulated track circuit. By constructing fault diagnosis decision table, weighted multi-part decision table graph and its adjacency matrix, we obtained an attribute reduction method based on adjacency matrix and a method of identifying core attributes, and fault rules were extracted by using the classification of fault rules and the coverage of fault information. The research shows that the algorithm is clear, easy to calculate and can effectively reduce the complexity of the algorithm.