基于机械故障诊断的算法及小波基选择研究

针对小波分析在机械故障诊断应用中的快速算法和小波基选择存在的问题,对常用的正交、半正交、双正交小波提取信号特征的能力进行了分析比较,研究表明半正交B样条小波因具有线性相位和采用较长的分解序列而具有较好的频率局部化特性和较小的变换误差,具有比较好的综合性能。推导了正交小波包的图形显示算法,利用该算法解决了小波变换快速算法中存在的数据长度随分解层数成倍减小的问题。研究结果为机械故障诊断中选择小波基进行信号处理提供了依据和途径。

Research Base on Algorithm of Mechanical Failure Diagnosis and Choosing Wavelet Basis

This article aims at the issue that celerity algorithm of wavelet analysis in mechanical failure diagnosis and in choosing wavelet basis. Analysed and compared the power of distil signal character of common orthotropic, half orthotropic, double orthotropic wavelet. Research indicated that half orthotropic B spline wavelet have better frequency localize feature and less transform error and better synthetical capability, because which have linearity phase and better disassemble sequence. Deduced the graphic display algorithm of orthotropic wavelet package. Use this algorithm, we settle the problem of data length germinate min-ish with the number of decompose layer in wavelet transform celerity algorithm. Result of our research provide a bases and a approach of choose wavelet basis for signal processing in mechanical failure diagnosis.