奇异值分解结合频率切片小波变换的轴承故障提取方法

Bearing Fault Extraction Method Based on Singular Value Decomposition and Frequency Slice Wavelet Transform

针对频率切片小波变换在强背景噪声条件下故障特征识别能力不足的缺点,提出了奇异值分解和频率切片小波变换相结合的故障特征提取方法。首先利用原始信号构造Hankel矩阵,根据奇异值差分谱单边极大值原则确定阶次并进行降噪处理;继而利用频率切片小波对降噪信号进行分析,得到全频带时频图后,对能量集中的时频区域进行细化分析;通过频率切片小波逆变换得到相应的重构信号;最终可以从重构信号的波形图中提取出轴承故障特征频率信息。仿真信号和实测信号分析表明,该方法能够实现滚动轴承运行状态的准确判别,对实际工程应用具有重要意义。

The capability of fault feature recognition of the frequency slice wavelet transform （ FSWT） is weak under the condition of strong background noise, in order to solve this problem, a fault feature extraction method by combing the singular value decomposition （ SVD） and frequency slice wavelet transform is proposed. Firstly, the Hankel matrix is constructed using the original signal, then the order is determined based on the principle of the single side maximum value of the singular value differential spectrum, and the de - noising process is implemented. Secondly, the de - noised signal is further analyzed by FSWT and the time - frequency map of the full frequency band is obtained, the detail analysis is performed on the time - frequency area with concentrated energy, then the corresponding reconstructed signal is acquired by FSWT.Finally,the fault feature frequency information of the bearing is extracted effectively from the reconstructed signal. The analysis results of the simulation signal and the measured signal show that the accurate judgment of the running state of the rolling bearing can be realized by utilizing the proposed method, and it is of great important significance for the practical engineering application.