摘要
为进行ZD6直流转辙机故障快速在线诊断,根据定性趋势分析算法原理,设置固定时间窗对动作电流数据进行平均分割,并对窗口内的数据进行线性拟合,将每个趋势片段识别为上升、下降或不变,最终获得一串简化的趋势序列。建立故障诊断规则知识库,并提出了计算趋势序列的匹配度,可有效区别不同故障状态。最后通过实例进行应用分析,实验结果表明,该方法可100%准确区分无故障状态与有故障状态,也能很好地区分几种特定的故障状态,具有较高的准确性与较低的复杂度。
For the rapid online fault diagnosis of ZD6 DC switch machine, according to the qualitative trend analysis algorithm principle,to set a fixed time window for the average segmentation of action current data,and the data in the window of the linear fitting,each fragment identified as the ascending and descending trend or unchanged,finally get a simplified trend sequence. The knowledge base of fault diagnosis rules is set up,and the matching degree of computing the trend sequence is put forward,which can distinguish different fault states effectively. Finally,through the analyses of application examples,the experimental results show that this method can accurately distinguish the fault free state and fault state,can well distinguish the fault state of certain kinds,with high accuracy and low complexity.
作者
胡锦添
蔡俊涛
张滔
HU Jin-tian;CAI Jun-tao;ZHANG Tao(Guangzhou Research Institute of O-M-E Technology,Guangzhou 510663,China;Guangzhou Metro Group Co.,Ltd.,Guangzhou 510330,China)
出处
《机电工程技术》
2018年第5期13-16,119,共5页
Mechanical & Electrical Engineering Technology
基金
广州市科技计划项目(编号:201508030038)
广州经济技术开发区科技计划资助项目(编号:2016S-P130)
