Prediction and Modeling of Tool Wear with Cutting Force and Fine Gaussian Support Vector Machine in Drilling

The tool wear is a vital problem in machining which affects the accuracy, geometry of the workpiece and also productivity. This paper presents a cutting-force-based monitoring method with a Gaussian Support Vector Machine for tool wear prediction in a drilling process. The experimental results show that the tool wear has a significant effect on the cutting forces. Statistical features in time and frequency domains are extracted from the measured force signals and are used as inputs to the Support Vector Machine (SVM). The machining parameters speed and feed are also taken as inputs for tool wear prediction as they greatly influence the cutting force. The fine Gaussian SVM algorithm is able to model the tool wear with a prediction accuracy of 88% and can be used for online diagnosis of tool wear. The results of the Gaussian SVM algorithm are compared with standard algorithms like ANN to prove its efficiency.