基于MEEMD排列熵的高速列车轮对轴承故障诊断

Fault Diagnosis of High Speed Train Wheelset Bearing Based on MEEMD Permutation Entropy

针对目前应用原始振动信号的排列熵单一尺度域分析对高速列车轮对轴承故障在特征提取研究方面的局限,提出了基于改进聚合经验模态分解排列熵的特征分析方法。该方法首先对原始信号进行改进聚合经验模态分解,得到一系列窄带本征模态函数;然后对原信号和本征模态函数分别计算排列熵,组成高维特征向量;最后,将高维特征向量输入最小二乘支持向量机状态识别分类。台架试验数据分析结果表明:该方法针对高速列车轮对轴承故障实现了较高的识别率,验证了通过改进聚合经验模态分解排列熵对高速列车轮对轴承故障诊断的有效性。

Aiming at the limitations of the permutation entropy single-scale domain analysis of the original vibration signal, the feature analysis method based on the improved ensemble empirical mode decomposition permutation entropy is proposed. Firstly, the original signal is improved by ensemble empirical mode decomposition, and a series of narrow-band intrinsic mode functions are obtained. Then, the permutation entropy is calculated for the original signal and the intrinsic mode function to form a high-dimensional eigenvector. Finally, the high-dimensional eigenvector is input into the least squares support vector machine state recognition classification. The analysis results of the bench test data show that the method achieves a high recognition rate for the high-speed train wheel bearing faults, and verifies the effectiveness of the ensemble empirical mode decomposition permutation entropy for the fault diagnosis of high-speed train wheelset bearings.