基于交叉验证法优化参数的Morlet小波消噪方法  

Parameter optimized Morlet wavelet de-noising method based on cross validation method

作　　者：汤宝平[1] 刘文艺[1] 蒋永华[1] 

机构地区：[1]重庆大学机械传动国家重点实验室,重庆400044

出　　处：《重庆大学学报（自然科学版）》2010年第1期1-6,共6页Journal of Chongqing University

基　　金：国家高技术研究发展计划(863计划)资助项目(2009AA04Z411);国家自然科学基金资助项目(50875272);长江学者和创新团队发展计划资助项目(IRT0763);重庆大学'211工程'三期建设研究生开放实验室支持项目(S-0916)

摘　　要：针对机械监测故障信号的非平稳性特点,提出一种基于交叉验证法优化参数的Morlet小波消噪方法。选择与机械冲击振动波形相似的Morlet小波,对Morlet母小波进行改进,增加了波形调整参数。通过交叉验证方法设计了改进Morlet小波的波形参数和变换尺度。对信号进行连续小波变换(CWT),实现对含噪信号的滤波消噪。将该方法应用于齿轮故障检测中,对比2种传统的小波消噪方法,验证该方法能够提取出强噪声背景下的有效信号特征成分,具有较好的滤波消噪效果。A new de-noising method based on parameter optimized Morlet wavelet is put forward. The Morlet wavelet is chosen as the mother wavelet because its shape is similar to the mechanical shock signals. The mother Morlet wavelet is improved By adding two parameters which decide the shape of the mother wavelet in time domain. The added parameters and the appropriate scale parameter for the wavelet transformation are designed by the cross validation method. Finally, the useful components of the signal can be obtained by the improved Morlet wavelet de-noising method. The gear fault diagnosis experimental result shows that the proposed method has a good de-nosing performance and it is effective in fault feature extraction

关 键 词：小波变换 小波消噪 改进Morlet小波 故障诊断 交叉验证法 

分 类 号：TN911.7[电子电信—通信与信息系统]