Research On State Recognition And Prediction Method Of Milling Tool Wear

Intelligent tool condition monitor technology is meaningful in advanced manufacturing technology.Tool wear monitor on line as its main content,can ensure the quality of products,improve production efficiency,reduce production costs,and reduce production time,avoid the problem of surface quality accuracy,parts scrap and machine fault caused by the failure of tool parts,realize the automation of machining process has very important significance.The paper focus on the cutting process of milling tool,launched a research about the state recognition and wear predicting of milling tool.The details are as follow:First of all,according to the relationship of tool wear and cutting force and vibration,the monitor signals based cutting force and vibration were build.The two signal energy of each band was extracted in different wear state and normalizing.The experiments result shows that harmonic wavelet packet has advantage in feature extracting of the non-stationary signal than wavelet packet.Secondly,a novel method based on harmonic wavelet packet and least squares support vector machine was build.Backtracking search optimization algorithm was proposed to search the optimal value of the kernel function parameter and error penalty factor which affect the precision of the LS-SVM significantly.The experiments result shows that harmonic wavelet packet was more effective and feasible than wavelet packet,and the proposed milling tool wear recognition method has higher accuracy than particle swarm optimization.Then,the observation prediction model based on the combination of characteristic value and observation value is established,the regression algorithm of LS-SVM was proposed to train model.The experiments result shows that the observation prediction model was higher accuracy than characteristic value model.The regression algorithm of LS-SVM can achieve good results.Finally,on the basis of the theoretical research,a system interface about tool wear monitor was designed;To realize visualized operation,the functions of normalizing,wear state classification and wear prediction were integrated into the system.