基于频率补偿小波的屏蔽主泵裂纹转子识别

针对转子裂纹振动信号不易被发现的问题，利用小波包分析方法，在无动不平衡频率的频带内进行分析，从中找出裂纹振动的特征频率。在分析中，针对小波算法存在的频率混淆、幅值失真、主频偏移的缺陷，在前人研究基础上，提出了一种频率补偿小波算法，利用该方法分析了核用屏蔽主泵转子裂纹振动信号，并得出了其振动的特征频率。结果表明，相比一般小波包，频率补偿小波包能识别出振动的特征频率，具有良好的实用价值。

Identification of Cracked Rotor of Canned Motor Pump Based on Wavelet Analysis of Frequency Compensation

This paper touches upon that the cracked rotor is hard to detective from vibration signal and provides a solution of it from wavelet packet analysis, finds the characteristic frequency of cracked vibration by analyzing within the band where dynamic-unbalance frequencies does not exist. To deal with the defects of frequency aliasing, distortion of amplitude, and main frequency shift in Mallat algorithm, in the analysis, a new method of wavelet analysis based on frequency compensation was supplied to analyze the cracked vibration signal of shielding main pump for nuclear use, and obtain the characteristic frequency of the vibration. The results show that compared with normal wavelet packet, frequency compensation wavelet packet can recognize the characteristic frequency of vibration, and it turns out to be a quite practical method.