摘要
针对目前S700K电动转辙机故障识别手段相对落后这一问题,本文通过分析说明其运行状态可通过功率曲线间接反映。根据微机监测系统存储的常见故障下的功率曲线建立故障诊断特征集,利用神经网络的高度并行运算能力,将灰色关联分析和神经网络技术相结合,建立灰色神经网络,计算待检功率曲线和各故障曲线之间的灰色关联度值,根据该值的大小判断转辙机的当前运行状态,实现S700K转辙机的故障诊断。从微机监测系统获取多组S700K转辙机动作功率数据作为测试样本集,对其进行验证计算,所得结果均与现场检修结果一致。
To target the issue that the fault identification method for the S700K electric switch machine is rela-tively backward,this paper showed that the running state of the switch machine can be indirectly reflected through the power curve.A fault diagnosis feature set was established based on the power curves of the com-mon faults stored in the computer monitoring system.Based on the highly parallel computing ability of the neural network,the grey correlation analysis was combined with the neural network technology to establish the grey neural network,to calculate the grey correlation degree between the test power curve and various fault curves.The diagnosis result for the S700K switch machine was obtained according to the grey correlation de-gree.Finally,multiple groups of the operating power data of the S700K switch machine,obtained from the computer monitoring system as the test sample set,were verified through calculation.The results were con-sistent with the field inspection results.
作者
王瑞峰
陈旺斌
WANG Ruifeng;CHEN Wangbin(School of Automatic & Electrical Engineerings Lanzhou Jiaotong U niversity?Lanzhou 730070, China)
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2016年第6期68-72,共5页
Journal of the China Railway Society
基金
甘肃省自然科学基金(1310RJZA046)