增强型滤波及冲击性机械故障特征的提取  

Extracting Impact Feature of Machine Fault by Using Enhanced Filtering

作　　者：邵毅敏[1] 周晓君[1] 欧家福[2] 陈再刚[1] 周本学[2] 

机构地区：[1]重庆大学机械传动国家重点实验室,重庆400044 [2]中国汽车工程研究院,重庆400039

出　　处：《机械工程学报》2009年第4期166-171,共6页Journal of Mechanical Engineering

基　　金：国家自然科学基金(50675232);教育部科学技术研究重点(106139);重庆市自然科学基金重点(2007BA6021);“十一五”国家科技支撑计划(2006BAF01B02-06)资助项目。

摘　　要：机械故障所引起的时间序列的微弱变化往往被由机械自身结构特点所引起的振动信号和来自其他振源的振动干扰信号及白噪声等所组成的强背景噪声所淹没,尤其是在设备出现早期故障时,这种微弱的故障特征信号很难被识别。根据机械噪声及冲击性故障特征信号的特点,提出了基于进化论自适应滤波和小波降噪耦合的增强型滤波器新算法,即克隆法和匹配法的子代繁衍与重构小波系数耦合的算法。模拟计算和物理台架试验结果表明,该算法不仅可较大幅度提高信号的信噪比,且可处理强噪声环境下的非线性噪声,具有较强的提取微弱冲击性故障特征的能力,且该方法具有很强的实用性。The feature signal which can indicate whether the machine is in good condition or not is difficult to extract when the machine is operating under heavy noise environmental condition, specially early in the failure development. According to the properties of the mechanical noise and the impulse feature signal, a new method coupled by the evolutionary adaptive filtering and the wavelet denoising is proposed. The evolutionary adaptive filtering uses the cloning method and mating method to get the generations＇ evolvement so as to realize the global optimization, while the wavelet denoising is realized by the re construction of the wavelet coefficient. Simulation and experiment results show that the signal to noise ratio are greatly improved and the algorithm has an excellent effect on processing the non-linear noise and extracting the early feature signal.

关 键 词：增强型滤波 机械故障 特征提取 

分 类 号：TH165.3[机械工程—机械制造及自动化]