基于双树复小波的重叠块阈值降噪方法  

Overlappling group thresholding denoising method based on dual-tree complex wavelet packet transform

作　　者：吴定海[1] 张培林[1] 杨望灿[1] 齐蕴光[1] 

机构地区：[1]军械工程学院车辆与电气工程系,石家庄050003

出　　处：《振动与冲击》2016年第10期162-166,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金资助项目(51305454)

摘　　要：针对机械早期故障信号受到强背景噪声影响导致故障特征不明显的问题,提出一种基于双树复小波包的重叠块阈值降噪方法。利用有限冗余双树复小波包变换对信号进行具有平移不变性的稀疏分解,结合双树复小波包变换系数所具有的稀疏集簇和邻域相关性特点,建立重叠块阈值估计模型,通过最小化包含块稀疏模型的适应度函数获得估计信号,分析了重叠块阈值降噪各参数对降噪效果的影响和参数优化原则,仿真与实测信号实验结果表明,该方法在不同信号和不同噪声水平下均有效地抑制了噪声干扰,提高了信噪比。In order to solve the problem that the early mechanical fault characteristics are usually indistinct due to strong background noise,an overlappling group thresholding denoising method based on dual-tree complex wavelet packet transform was proposed.The signal was sparsely decomposed by using the dual-tree complex wavelet packet which has the characteristic of shift-invariant and limited redundancy.The overlappling group thresholding denoising method based on dual-tree complex wavelet packet transform was developed which can capture the neighborhood correlation and large-amplitude coefficients form clusters,and a denoising model was built based on the minimization of a convex cost function incorporating with a mixed norm.Then the parameters optimization of the denoising model was discussed.The simulation and experimental results show that the noise reduction effect of the presented method is satisfactory for different signals with different level of noise,the background noise is restrained effectively and the signal noise ratio can be well improved.

关 键 词：机械振动 状态监测 重叠块阈值 双树复小波包 降噪 

分 类 号：TH165[机械工程—机械制造及自动化] TP206[自动化与计算机技术—检测技术与自动化装置]