Research On State Recognition Of Tool Wear Based On CEEMD-WPT

With the development of automation, integration and unmanned technology in machine tools, how to guarantee the quality and efficiency of processing products are important.As a direct executor of the machining process, the wear phenomenon is inevitable. Therefore, in order to ensure the product quality and realize the efficient use of tools, it is necessary to study the tool condition monitoring technology.According to the characteristics of tool machining, tool to monitor the acoustic emission signal of acoustic emission monitoring technology is an effective nondestructive testing technology because of its high sensitivity, strong anti-interference ability, has the advantages of no downtime etc. has been widely used, but because of the collected acoustic emission signal of high frequency, large amount of data and frequency complex composition, unable to direct the cutting tool state recognition, in order to accurately grasp the qualitative tool wear in cutting process, the paper proposes a Complementary Ensemble Empirical Mode Decomposition(CEEMD) and Wavelet Package Transform(WPT) of the tool condition monitoring method. The first use of CEEMD acoustic emission signal is adaptively decomposed into several Intrinsic Mode Function(IMF) within each IMF contains different time scale characteristics of the original signal,the modal aliasing problem still exists in the IMF, the local processing capacity of WPT good corrected,so as to realize the accurate extraction of the characteristics of components, and then select the energy before several large IMF component value after correction, the proportion of the total energy of the feature vectors, finally input to the Support Vector Machine(SVM) for training and testing, which established by the 6 SVM two value of the classifier is composed of 4 kinds of cutting tool state recognition system.This article through the comparison with the CEEMD feature extraction method,Indicate feature extraction of CEEMD-WPT is more accurate, more representative,two kinds of time-frequency analysis methods, which effectively solves the modal decomposition of CEEMD still exists aliasing problems, but also eliminates the separate into WPT treatment effect, false frequency component the frequency aliasing, and accurately identify for later laid a good foundation for tool wear condition.