基于卷积稀疏滤波的轴承微弱故障检测方法

A Novel Weak Bearing Fault Diagnosis Method Based on Convolutional Sparse Filtering

提出了一种基于卷积稀疏滤波和Hilbert包络谱的齿轮微弱故障检测方法。该方法通过稀疏特征学习,提取强噪声样本中的微弱故障信息,提高故障信号的信噪比,最后通过时域波形和Hilbert包络谱的特征频率及其谐波,判断轴承的故障信息。通过仿真和试验信号,验证了该方法的有效性,与经典的MED算法相比,提出的方法具有更强的噪声适应能力。

This paper proposes a novel weak signature detection method based on convolutional sparse filtering and Hilbert envelope spectrum for rolling bearing fault diagnosis.The proposed method can extract the weak fault information from samples with strong noise interference through the sparse feature learning process.The signal to noise ratio(SNR)of the filtered signal is obviously improved.Finally,we can detect the fault information using the time domain waveform and envelope spectrum of the filtered signal.The proposed method is verified using simulated and experimental rolling bearing fault data.The results show that the proposed method has been found to be a promising tool for impulsive feature enhancement.Compared with MED,the proposed method performs superior noise adaptability.