基于RS和GCA-TOPSIS的S700K转辙机故障诊断方法

Fault diagnosis method for the S700K switch machine based on the RS and GCA-TOPSIS

针对目前S700K转辙机故障诊断效率低、准确性不高等问题,提出了一种基于粗糙集、灰色关联分析与理想排序法相结合的转辙机故障诊断方法。首先以微机监测系统(Maintenance and Monitor System,MMS)存储的常见转辙机故障功率曲线数据作为数据源,提取功率曲线在各工作区段的特征参数,构建故障特征集。然后针对冗余特征,采用粗糙集理论中的约简算法以属性重要度为选择标准,对特征集进行约简,降低特征集的维度。最后将灰色关联分析和逼近理想排序法相结合,计算待检样本与各故障类型间的曲线贴合度,判断待检样本与各故障类型间的紧密程度,将最大曲线贴合度对应的故障类型作为待检样本的诊断结果。实例分析表明,该方法能够准确地诊断出转辙机故障且诊断效率和准确性较高,能够满足铁路现场实际需要。

Aiming at heightening the efficiency and accuracy of the fault diagnosis of the current S700K switch machine,the present paper intends to propose a switch machine fault diagnosis approach by integrating the rough set,the grey correlation analysis method and technique for order preference by similarity to ideal solution. To achieve the aforementioned purposes,it would be necessary first of all to obtain the available data from the fault power curve calculation of the common S700K switch machine so as to keep up and monitor the needed material collections. And,the fault features of the switch machine should also be checked by extracting the fault featured materials needed for each power curve in the different working sections. And,secondly,there should also be existing redundancies for the extracted fault features,and the attributing reduction algorithm in the rough set theory to delete the redundant features based on the discretization of the said fault features. The algorithm to be adopted should be done based on the fault feature set to set up the class discernibility matrix among the fault types,with the size of attribute importance as the reduction set selection standard. At the same time,to reduce the fault feature set,it is also necessary to reduce the dimensions of the fault feature set and heighten the efficiency of the fault diagnosis. Thus,eventually,the curve fit degree has to be calculated and worked out between the sample to be tested and each fault type by joining the grey correlation analysis and the technique for order preference by similarity to ideal solution. And,simultaneously,the closeness between the test samples to be tested and all the fault types has to be confirmed and determined in accordance with the curve closeness based on the maximum curve passed. And,when the fault diagnosis result of the sample were to be tested,it is also necessary to adopt the fault type with the biggest degree of convergence taken. And,so,the diagnostic method can be accurately diagnosed and established through the sample or the example analysis,so that the faults of the S700K switch machine can be reduced to the minimalized whereas the diagnostic efficiency and accuracy be elevated so as to meet the actual needs of the railway sites.