摘要
为消除Mallat算法存在频率混叠的缺陷,采用卷积型小波和小波包变换快速分解算法对信号进行分解.为克服Daubechise系列小波滤波器组幅频特性的不足,采用巴特沃斯滤波器构造了幅频特性更好的小波滤波器组,并导出了巴特沃斯小波.基于此,提出了一种结合卷积型小波和小波包变换快速分解算法与巴特沃斯小波的机械故障特征提取方法.仿真实验与实例分析表明:本方法比基于Daubechise系列小波与Mallat算法的机械故障特征提取方法在微弱故障特征提取方面更有优势.
To eliminate the frequency aliasing in Mallat algorithm, the fast decomposition algorithms of wavelet of convolution type and wavelet packet transformation were used. In order to overcome the defect of amplitude frequency response characteristic in Daubechise wavelets filter banks, the wavelet filter banks based on Butterworth filter were constructed and the Butterworth wavelets were derived. Then, a new method to extract the mechanical fault feature was proposed. The simulation and analysis of the fault data shows that the new method can detect the mechanical fault feature more effectively than conventional methods which are based on Daubechise wavelets and Mallat algorithm.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第10期68-73,共6页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(50775218)
国防预研基金资助项目
关键词
MALLAT算法
卷积型小波
共轭正交镜像对称滤波器组
巴特沃斯小波
特征提取
包络解调
Mallat algorithm
convolution type of wavelet
conjugate quadrature mirror filter banks（CQMFB）
Butterworth wavelet
feature extraction
envelope demodulation
