基于最小熵反褶积的动车组牵引电动机轴承故障诊断方法  

Fault Diagnosis Method for EMU Traction Motor Bearings Based on Minimum Entropy Deconvolution

作　　者：李伟 LI Wei(Xi'an CRRC Electric Research Institute, CRRC Yongji Electric Co., Ltd., Xi'an 710000,China)

机构地区：[1]中车永济电机有限公司西安中车电气研究院,西安710000

出　　处：《轴承》2021年第12期56-60,64,共6页Bearing

基　　金：山西省关键核心技术和共性技术研发攻关专项项目(2020XXX012)。

摘　　要：针对牵引电动机轴承故障信号冗余程度高,信号微弱且伴随有较强非高斯噪声,故障特征难以提取的问题,采用了一种最小熵反褶积算法对复杂信号进行处理,由于该算法中滤波步长对处理信号影响较大,采用包络谱熵作为适应度函数,使用量子行为粒子群优化算法对最小熵反褶积算法中的滤波步长进行优化,使最小熵反褶积算法达到最优。试验结果表明,优化后的最小熵反褶积算法能够准确提取牵引电动机轴承故障特征。In view of high degree of redundancy of fault signal of traction motor bearings,weak signal with strong non-Gaussian noise and difficult extraction of fault features,a minimum entropy deconvolution algorithm is used for complex signal processing.Because of filtering step in the algorithm has a greater impact on processed signal,the envelope spectral entropy is used as fitness function,and the quantum-behaved particle swarm optimization algorithm is used to optimize the filter step in minimum entropy deconvolution algorithm,so that the minimum entropy deconvolution algorithm is optimized.The test results show that the optimized minimum entropy deconvolution algorithm can accurately extract the fault features of traction motor bearings.

关 键 词：滚动轴承 牵引电动机 故障诊断 谱分析 熵 最小熵反褶积 量子行为粒子群优化算法 

分 类 号：TH133.33[机械工程—机械制造及自动化] U268.3[机械工程—车辆工程]