摘要
为实现地铁车辆走行部关键部件的不解体检测诊断,采用过车轨道振动来分析车辆平轮故障。试验采集了正常情况、剥离故障及擦伤故障等3种工况下的振动信号。首先对信号进行集合经验模态分解;然后,用相关系数法筛选分解产生的本征模态函数分量,再计算主分量的模糊熵熵值作为故障特征向量;最后,输入到由遗传算法优化的支持向量机分类器进行故障识别。试验结果表明,该方法可以实现地铁车辆平轮故障的准确识别。
To achieve disassembly detection and diagnosis of key components in metro vehicle running gear, the fault of flat wheel through the rail vibration is analyzed. This ex- periment collects vibration signals in three working conditions: nrmal condition, peeling failure and abrasion fault. Firstly, the vibration signal is adaptively decomposed by using the ensemble empirical mode decomposition into a series of intrinsic mode functions. Then, the correlation coefficient is calculated to sift out intrinsic mode functions (IMF) that have largest correlation coefficients with the original signal, and the fuzzy entropies of these IMFs constitute a high dimensional characteristic vector. Finally, the feature vector is put into the genetic-support vector machine for classification and identification. The experimental result shows that this method can achieve accurate iden- tification of the flat wheel fault.
出处
《城市轨道交通研究》
北大核心
2017年第3期80-84,共5页
Urban Mass Transit
基金
国家自然科学基金资助项目(51275093)
广东省科技厅科技项目(2013498A)
关键词
地铁车辆
轨道振动
集合经验模态分解
模糊熵
支持向量机
故障诊断
metro vehicle
rail vibration
ensemble empirical mode decomposition
fuzzy entropy
support vector machine
fault diagnosis