基于WATV去噪的冲击特征提取方法在高速列车轴承故障诊断中的应用

提取高速列车轴承故障振动信号中的冲击特征,可以有效地对其进行故障诊断。利用"小波-全变差(Wavelet-Total Variation,WATV)"算法能够对信号进行稀疏引导的特点,提出了基于WATV去噪的冲击特征提取方法。该算法针对含噪声冲击特征的提取问题构建了目标优化函数,该函数融合了冲击特征的保真度度量算子以及惩罚因子。利用凸优化理论可对目标函数进行求解,从而增强信号在小波域和时域的稀疏性,使得特征提取结果最优化。通过构造一仿真信号对WATV算法的有效性进行了验证,并将该方法应用于高速列车齿轮箱轴承故障诊断中。结果表明,该方法能够很好地提取出信号中的冲击特征,并且频谱中的故障表征明显,能够有效地应用于高速列车轴承故障诊断中。

Application of an Impact Feature Extracting Method Based on WATV in Fault Diagnosis of High-speed Train Bearing

Extracting impact features from a fault vibration signal of high-speed train bearing is significant to some fault diagnoses.Based on the fact that wavelet-total variation(WATV)algorithm is capable of inducing sparsity,an effective impact feature extracting method with WATV was proposed.In this algorithm,the objective optimization function was constructed for the extraction of noisy impact features,which combined the fidelity measurement operator and penalty factor of impact features.The convex optimization theory could be used to solve the objective function,so as to enhance the signal sparsity in wavelet domain and time domain,and optimize the feature extraction results.The validity of WATV algorithm was verified by constructing a simulation signal,and the method was applied in the fault diagnosis of gearbox bearing of high-speed train.The results showed that the method could extract the impact feature of signal well,and the fault representation in spectrum was obvious,which could be effectively applied in the fault diagnosis of high-speed train bearing.