基于LMD-SVD和极限学习机的滚动轴承故障诊断方法研究

The Rolling Bearing Fault Diagnosis Method Based on LMD-SVD and Extreme Learning Machine

针对局部均值分解(LMD)得到的PF分量对于分类方法的输入而言过大,提出了一种基于局部均值分解(LMD)-奇异值分解(SVD)和极限学习机(ELM)的故障诊断方法。首先,通过LMD将非线性非平稳的原始振动信号分解为一系列乘积函数,从而得到具有物理意义的瞬时频率;然后,采用SVD处理PF分量以压缩特征向量尺度并获得更加稳定的特征向量值;最后,基于提取的特征向量,应用运算效率和分类精度更高的ELM对轴承故障状态进行分类。试验结果表明,该方法能有效的对风机轴承在变工况条件下进行自适应诊断。

The PF component obtained by Local Mean Decomposition(LMD)is too large for the input of the classification method.This paper proposes a fault diagnosis method based on LMD-singular value decomposition(SVD)and extreme learning machine(ELM).Firstly,the nonlinear non-stationary original vibration signal is decomposed into a series of product functions by LMD to obtain instantaneous frequencies with physical significance.Then,the PF component is processed by SVD to compress the feature vector scale and obtain a more stable feature vector value.Finally,based on the extracted feature vector,an ELM with higher computational efficiency and classification accuracy is applied to classify the bearing fault status.Test results show that this method can adaptively diagnose the fan bearing under variable working conditions.