基于EMD和SVM的缸盖加工颤振在线检测
2020-03-11
作者：密思佩1，徐锦泱1，明伟伟1，陈明1，陈龙2
单位：1上海交通大学；2上海交大智邦科技有限公司
摘要：切削载荷下加工系统的颤振现象直接影响加工过程的效率和性能。本文介绍了一种基于经验模式分解（Empirical Mode Decomposition，EMD）的机床刀具颤振分析方法。通过对机床主轴的振动信号进行综合分析，并对异常颤振信号进行EMD分解以获得本征模函数，采用Hilbert变换得到其包络信号，计算包络谱，提取噪声信号的特征频率，对特征频率进行支持向量机（Support Vector Machine，SVM）颤振判别学习，通过现场信号验证，证明该方法能有效检测加工颤振。
关键词：颤振；信号分析；经验模态分解；支持向量机；在线检测
中图分类号：TG5；TH161.6文献标志码：ADOI：10.3969/j.issn.1000-7008.2020.02.019
Insitu Detection of Machining Chatter of Cylinders
Based on EMD and SVM Methods
Mi Sipei，Xu Jinyang，Ming Weiwei，Chen Ming，Chen Long
Abstract:Chatter phenomena of machining systems under cutting loads have a direct impact on the efficiency andperformance of various manufacturing processes.A method of machine tool chatter analysis based on Empirical Mode Decomposition (EMD) is introduced.Through the comprehensive analysis of vibration signals of the machine tool spindle during the manufacturing process，the EMD decomposition of the abnormal chatter signals is performed to obtain the eigenmode functions，then the envelope signals are obtained by Hilbert transform.The envelope spectrum is calculated and the characteristic frequency of the noise signals is extracted.Support Vector Machine (SVM) flutter discriminant learning is also performed on the characteristic frequency.Through the insitu verification of signals，the method can effectively detect the machining chatter.
Keywords:chatter；signal analysis；Empirical Mode Decomposition (EMD)；Support Vector Machine(SVM)；insitu detection