Prediction model of high-speed oblique cutting temperature based on LS-SVM

Abstract

High-speed oblique cutting temperature is an important factor in ensuring workpiece quality. In order to gain the temperature real time in the cutting process, a prediction method based on least squares support vector machine (LS-SVM) was proposed. To verify the feasibility of the method, firstly, the high-speed cutting temperature model was established based on LS-SVM, and the major operation parameters (cutting speed, feed rate, axial depth of cut, and radial width of cut) were chosen as the model input based on oblique cutting process analysis; secondly, the cutting experimental scheme was designed applying the Box–Behnken experimental design method for gaining more cutting temperature data and less experimental times. Then, a high-speed cutting temperature measurement system was established based on a MCV850 vertical machining center for testing the reliability of model prediction. Finally, the model prediction results based on LS-SVM and neural networks were compared. And the results show the prediction error of the model gained is less than 1 %, and taking two-group random parameters as test data with different with Box–Behnken experimental parameters designed before, the percentages of prediction data deviation measurement were 0.83 and 0.51 %, respectively. The results demonstrate the feasibility of applying the cutting temperature prediction model in predicting the main required processing parameters.