基于截断高阶奇异值分解的电机多轴承故障诊断

Multi-Bearing Fault Diagnosis of Electric Motor based on Truncated High-Order Singular Value Decomposition

电机在运行过程中轴承承担着高频载荷运动,单一通道信号识别无法保障故障识别精度。为了进一步降低轴承多通道信号滤波干,设计了一种基于截断高阶奇异值分解(HOSVD)的电机多轴承故障诊断方法。通过多通道故障降噪方法对多通道信号实施滤波处理。三个轴承故障实例表明,域内特征无法针对轴承的健康状态进行有效的识别。该方法可以同时提取大量通道中的外圈和内圈故障特征,比较理想地提取出大量通道轴承外、内的故障特征,所提方法的有效性得到验证。

The bearing of the motor bears the high-frequency load movement during operation,and the single-channel signal recognition cannot guarantee the fault identification accuracy.In order to further reduce the filter interference of bearing multi-channel signal,a multi-bearing fault diagnosis method of motor based on the truncated high-order singular value decomposition(HOSVD)is designed.The filtering processing is carried out for the multi-channel signal by multi-channel fault denoising method.Three bearing fault examples show that the health state of bearings cannot be effectively identified by the features in the domain.The proposed method can simultaneously extract the fault features of the outer ring and the inner ring in a large number of channels,and it is ideal to extract the fault features of the outer ring and inner ring in a large number of channel bearings.The effectiveness of the method has been verified.