Modelling and prediction of tool wear using LS-SVM in milling operation

This article focuses on the least squares support vector machine (LS-SVM), which can solve highly nonlinear and noisy black-box modelling problems, and tool wear model based on LS-SVM for ball-end milling cutter is established by considering the joint effect of machining conditions. In the established model, machining parameters and position parameter of ball-end cutter are considered as input and the output of the proposed model is tool wear of cutting edge position. The experimental measured tool wear is used to train the established model, and the interconnection relationship between input and output parameters is determined after training. The analysis and comparison of predicted performance are given by taking different tuning parameters and data regularisation. Some interesting analysis results are deduced from the established LS-SVM-based tool wear model. In order to further show the effectiveness of LS-SVM-based tool wear model, the verified comparison between LS-SVM-based and ANN-based model is given. Finally, the discussion of interactional effect of machining parameters on tool wear estimation is used to evaluate prediction performance of LS-SVM-based model. The verification shows that the LS-SVM-based tool wear model is suitable to predict tool wear at certain range of cutting conditions in milling operation.