On-Machine Measurement of Machining Accuracy Based on a Support Vector Regression (SVR) Model for Product Quality Monitoring

Measurement systems and sensor technologies are key to production engineering in Industry 4.0. On-machine measurement (OMM) is used to evaluate various factors, conditions, and work performed on machine tools. OMM systems have been increasingly used in recent years due to their advantages in online measurement, high efficiency, and high stability. In this study, an on-machine accuracy measurement based on a non-contact probing sensor is proposed using a support vector regression model (SVR) for product quality monitoring. The laser displacement sensor as a non-contact probing scans the complex surface contour of the workpiece being machined by a regression method, which shows that SVR with the RBF kernel produces the best performance (RMSE 0.053).