为实现刀具磨损状态准确快速的识别,开发了一套基于自动化设备规范(ADS)通信技术和经验模态分解(EMD)的刀具状态在线监控系统.运用EMD将振动信号分解成多个固有模态函数分量(IMF),综合使用相关系数法以及能量值法筛选了前6阶IMF分量的均方根值作为监测特征,然后将监测特征作为支持向量机的输入,建立监测特征与刀具磨损状态的关系模型.加工中的一定长度的振动数据经ADS技术传输到建立好的支持向量机(SVM)模型中,完成刀具状态的识别.使用TwinCAT和Matlab实现了整套系统的功能.经试验验证,刀具监控系统运行稳定,能对刀具状态进行准确快速的判断.%In order to recognize tool wear state accurately and quickly, a system based on Empirical Mode Decomposition(EMD) and Automation Device Specification(ADS) is developed.Development process include model building and online recognition.Firstly, vibration signal is decomposed into several intrinsic mode functions(IMF) by EMD method, using a combination of correlation coefficient method and energy method to figure out eigenvalues, the eigenvalues are the first six orders of IMF's root mean square.Eigenvalues are selected as inputs of support vector machine(SVM), then the relation model between eigenvalues and wear condition is build.Vibration data with a specified length are transported into SVM model, Finally, the tool wear condition is identified.Using Matlab and TwinCAT to implement the system's functions.Proved by test, the system works well and judge the tool wear state quickly and accurately.
展开▼
