机械信号连续小波系数的统计特征研究

通过对机械信号的连续小波分解，利用小波滤波器良好的时频特性，针对机械信号的特征，论述了机械信号经过连续小波分解后的统计特征。在机械信号连续小波系数的特征提取中，引入机械系统的动力学维数，将连续小波系数按尺度方向进行了研究。利用尺度与频率的对应关系，应用小波系数的容积维数来研究其统计特性，并提取信号特征。实践证明，这种方法能够简单有效地提取机械信号的振动特征，并在诊断机械故障中得到了应用。

Study on the Statistic Characteristic of Continuous Wavelet Coefficients in Processing Mechanical Signal

The continuous wavelet coefficients of mechanical signal possess the feature of mechan-ical signal. The dynamic dimension of mechanical system can describe the characteristic of the system. The wavelet scale corresponds to the frequency feature of the detected signal. Using continuous wavelet transform to decompose and denoise the mechanical signal, the dynamic dim-ension of continuous wavelet coefficients along the scale can be calculated. According to the dynamic dimension of the continuous wavelet coefficients, the feature of the mechanical signal can be extracted. Practice proved that the method could be used to extract mechanical feature and pattern recognition.