基于多源信息的高速列车走行部故障识别新方法.pdfVIP

基于多源信息的高速列车走行部故障识别新方法 朱建渠 1，2 ，金炜东 1 ，朱斌 1,3 （1.西南交通大学电气工程学院，四川 成都 610031 ； 2.重庆科技学院，重庆 401331 ； 3.长江师范学院，重庆涪陵 408100 ） 摘 要 ：针对高速列车安全性能跟踪监测点多、监测数据量大、走行部故障识别难的问题，提出了一种基于模糊证据理论 的多特征、多源信息融合的走行部故障识别新方法。该方法首先根据不同的传感器信息的某类特征属于不同故障模式下的隶属度 间的差异来计算传感器间的信息融合度，利用融合度来确定不同传感器在融合中的权重，从而得到同类特征不同传感器间信息融 合后的隶属度；然后由融合后的隶属度转化为基本概率分配函数；最后用证据理论对不同特征间信息进行融合。实验结果表明， 该方法能有效地识别列车正常、空气弹簧失气、抗蛇形减震器全拆、横向减震器全拆四种工况，同时在不同速度下均取得了满意 的识别率，验证了该方法的有效性。 关键词： 高速列车；信息融合；故障识别；熵特征 中图分类号 ：U266 文献标识码 ：A High-speed train running gear fault feature extraction and recognition based on harmonic wavelet packet transform 1,2 1 1,3 ZHU Jian-qu JIN Wei-dong ZHU Bin (1.School of Electric Engineering, Southwest Jiaotong University Chengdu 610031; 2. Chongqing University of Science and Technology, Chongqing, 401331; 3. Changjiang Normal University, Chongqing, 408100) Abstract : To solve the problem that the fault identification of high speed train is difficult, a fault feature extraction and identification method based on harmonic wavelet packet decomposition and resemblance coefficient was proposed. Harmonic wavelet packet transform is used to extract fault feature vector consisting of the band energy which can reflect the characteristics of various types of fault. By calculating and comparing resemblance coefficient of the signal feature vector and different types of fault feature vectors, the method can identify and classify the faults. Experimental results show that the faults, including the normal train, spring air of loss air, without anti-yaw shock