基于信号奇异性分析的轴承故障检测方法

Bearings Fault Detection based on Signal Wavelet Singular Analysis

针对轴承振动信号利用小波单奇异点检测无法克服噪声影响的不足,提出利用小波模极大值分析信号奇异性变化进而进行轴承故障检测的方法。实验中对信号的模极大分形指数、模极大分形指数熵、Lipschitz指数以及Lipschitz指数熵等奇异特征进行分析比较,实验结果表明这些特征都能有效克服噪声影响实现故障检测,但模极大曲线数最能体现故障特征且检测效果最好。最后将本方法同基于小波包能量谱特征和小波单奇异点检测的方法进行比较,实验结果表明本文中建议的方法在检测时间及检测率上都有显著提高。

To solve the problem that the wavelet- singular point detection is more sensitive to the noise, the scheme of bearings fault detection based on the number of singularity point using wavelet transform modulus maximum with the constant length is proposed. Wavelet transform modulus maximum fractal spectrum, Wavelet transform modulus maximum fractal spectrum entropy, Wavelet transform modulus maximum Lipschitz spectrum and Wavelet transform modulus maximum Lipschitz spectrum entropy are analyzed and compared, results show the number of singularity point with the constant length can effectively reflect the characteristics of bearings faults and copy with the noise effect. The proposed method is compared with the method based on wavelet packet energy spectrum and wavelet- singular point detection - based method in the experiments. The results show that the number of singularity point using wavelet transform modulus maximum with the constant length is particularly well adapted to describe fault characteristics and fault diagnosis, which outperforms the method based on wavelet packet energy spectrum in terms of detection time and detection rate.