基于J-EEMD的刀具磨损状态特征提取技术  

Feature Extraction Techniques of Tool Wear States Based on J-EEMD

作　　者：陈洪涛[1] 傅攀[1] 李晓辉[1] 钟成明[2] 

机构地区：[1]西南交通大学机械工程学院,成都610031 [2]东方汽轮机有限公司,德阳618000

出　　处：《机械科学与技术》2014年第6期849-853,共5页Mechanical Science and Technology for Aerospace Engineering

基　　金：中央高校基本科研业务费专项资金项目(SWJTU12CX039)资助

摘　　要：在实际刀具状态监测的过程中,通过传感器所直接测得的数据都包含了大量的噪声信号,因此难以从中获取刀具磨损状态的变化规律,这样显然不利于进行模式识别。应用近似联合对角化下的集合经验模态分解(J-EEMD)对观测信号进行处理,基于信号本身特征,自适应地将切削加工中检测得到的振动和声发射信号分解为多个内蕴模式函数(IMF),然后根据各个IMF之间的能量比对变换,提取出了不同磨损状态下的刀具状态特征。实验证明:在该方法对测得数据进行处理的基础上,能够很好地识别出刀具磨损程度的不同状态。In the monitoring of cutting tool state,a large number of acoustic noise is contained in the sensor signal.Therefore,it is difficult to get the tool wear state,which is obviously not conducive to pattern recognition. Observed signals were processed using the method of ensemble empirical mode decomposition based joint approximate diagonalization of eigenmatrice（ J-EEMD）. In this method,based on the characteristics of the signal itself,the signals of vibration and acoustic emission were adaptively decomposed into several intrinsic mode functions（ IMF）;and then transform the energy ratio between the IMF and the original signals of vibrations and acoustic emission;finally the tool state characteristics under different wearing can be extracted. The test result showed that the method could preferably help in carrying out the pattern recognition to the different states of tool wear.

关 键 词：噪声 特征提取 分解 模式识别 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]