CAJ | 학술논문

转辙机故障对铁路运输安全和效益影响重大，针对转辙机故障原因与现象之间的复杂不确定性关系，提出一种基于粗糙集约简的高效贝叶斯网络故障诊断方法。首先，建立故障诊断决策表，利用改进的差别矩阵算法剔除对结果影响较小的属性，得到最简故障诊断决策表。其次，根据表中故障现象与故障类型连接关系建立贝叶斯网络模型，利用推理算法求解各类故障发生的概率。算法通过约简属性简化贝叶斯网络结构，降低算法复杂度，加快计算速度。最后，用某车站转辙机故障实例验证该智能故障诊断方法的正确性。
전철궤고장대철로운수안전화효익영향중대，침대전철궤고장원인여현상지간적복잡불학정성관계，제출일충기우조조집약간적고효패협사망락고장진단방법。수선，건립고장진단결책표，이용개진적차별구진산법척제대결과영향교소적속성，득도최간고장진단결책표。기차，근거표중고장현상여고장류형련접관계건립패협사망락모형，이용추리산법구해각류고장발생적개솔。산법통과약간속성간화패협사망락결구，강저산법복잡도，가쾌계산속도。최후，용모차참전철궤고장실례험증해지능고장진단방법적정학성。
The impact of switch machine on the safety and efficiency of railway transportation is significant.A senior Bayesian network fault diagnosis method based on rough set theory was proposed according to the complex uncertainty relation between the fault reasons and phenomena.Firstly,the fault diagnosis decision table that used the improved discernibility matrix algorithm to eliminate attribute that had little effect on the result was formed to obtain the most simple fault diagnosis decision table.Secondly,the Bayesian network model was established ac-cording to the relationship between failure phenomenon and failure type,and the probability of kinds of failure was solved using reasoning algorithm.The algorithm could simplify Bayesian network structure and decrease cal-culation speed through reducing attribute.Finally,the correctness of the intelligent fault diagnosis method was verified through a case example of a service section switch machine fault.