Study on the Identification Method of Tool Wear State Based on SLLE and SVM in NC Machining
暂无分享,去创建一个
Aiming at the nonlinear features of Acoustic Emission reflecting the tool wear state, an identification method of tool wear state is proposed based on SLLE and lib-SVM. SLLE can overcome the problems of redundant parameters of the original nonlinear algorithms, low convergence rate, and tiny local areas in the reflecting process, etc. It can realize the reflecting from the high-dimensional data points into a global low dimensional coordinate system. It can also maintain the original topological structures of the lowered dimensional data. By using KC9125 cutting tools to cut 40CrNiMoA alloy steel to collect the acoustic signals, the collected data is vector space reconstructed. Then the high-dimensional space data are reflected into low-dimensional space data by means of SLLE in order to extract the features of the tool wear state. And LIB-SVM classification machine is used to identify and classify the tool wear state. The result shows: this method can accurately identify the tool wear state, and should be extended to a larger prevalence.
[1] D. D. Ridder,et al. Locally linear embedding for classification , 2002 .
[2] L. Saul,et al. An Introduction to Locally Linear Embedding , 2001 .
[3] S T Roweis,et al. Nonlinear dimensionality reduction by locally linear embedding. , 2000, Science.
[4] Nie Peng,et al. Prediction of tool VB value based on PCA and BP neural network , 2011 .