Application of cutting power consumption in tool condition monitoring and wear prediction based on Gaussian process regression under variable cutting parameters

Tool wear is inevitable in actual manufacturing, especially in extreme processing conditions for machining difficult-to-cut materials. The monitoring of the tool state has an important influence on the surface quality and dimensional accuracy of the precision parts. In the previous studies, the original total power consumption is usually used to predict tool wear while ignoring the cutting power consumption accounts for a small proportion of the total power consumption of machine tools. Therefore, the accuracy is difficult to achieve the expected target. For better prediction results, a novel prediction method based on net cutting power consumption by Gaussian process regression (GPR) with ARD Matern 5/2 kernel is proposed in this study. Firstly, the physical model of net cutting power consumption is established. Then, tool wear under fixed working conditions is predicted by using the net cutting power consumption and GPR, and the advantage of the proposed method in this study is verified by comparing it with the existing methods. Finally, the proposed method is verified to obtain better prediction performance with variable cutting parameters than using total power consumption with the neural network. This study reveals that low-cost sensors like power meter can be used as an important supplement to monitoring tool conditions in the industry and also provides a research basis for predicting tool wear under different cutting conditions.