Intelligent process supervision for predicting tool wear in machining processes

An intelligent supervisory system supported on a model-based approach is presented herein. The application for predicting tool wear in machining processes is selected as a case study. A model created using artificial neural networks and able to predict the process output is introduced as a means of dealing with the characteristics of such an ill-defined process as machining. This model describes the output's dynamic response to changes in the process-input command (feed rate) and process parameters (depth of cut). In order to predict tool wear, residual errors are used as the basis for a decision-making algorithm. Based on the model and the weighted sum of squared residuals method, the procedure continuously checks whether a given index (tool condition) exceeds a critical threshold. In the chosen application, an over-the-threshold index is interpreted as indicating unacceptable tool wear necessitating immediate tool replacement. Experimental tests are run in a professional machining centre under different cutting conditions using real-time data and new, half-worn and worn tools. The results show this supervisory system's suitability and potential for industrial applications. (C) 2003 Elsevier Science Ltd. All rights reserved.