A tool wear predictive model based on SVM

Tool wear monitoring is an integral part of modern CNC machine control. This paper presents a new tool wear predictive model by combination of workpiece surface texture analysis and support vector machine with genetic algorithm (SVMG). Firstly, the column projection method and the Gabor filter method are proposed to extract texture features of machined surfaces. Then, SVMG-based tool wear predictive model is constructed by learning correlation between extracted texture features and actual tool wear. The effectiveness of the proposed predictive model and corresponding tool wear monitoring system is demonstrated by experimental results from turning trials. After simulated and compared with the predictive model based on BP neural networks, the method shows much better performance on the predictive precision and the intelligent adjusting parameters.