A self-organizing approach to the prediction and detection of tool wear

Abstract

The decision tree method and the group method of data handling (GMDH), due to their self-organizing capability for sensor integration, diagnostic reasoning and decision making, were adopted for the reconition and prediction of the tool wear state in a turning operation using acoustic emission and cutting force signals. The decision tree approach was utilised to stimulate human intelligence and generalize heuristic rules from learning examples and was demonstrated to be able to make reliable inferences and decisions on tool wear classification. As a process modeling tool, the GMDH algorithm determines a representation of the real-time machining system interrelationship between tool flank wear and the quantitative measure of sensor variables involved. The derived model was used to predict the tool wear from the in-process sensor output features. A high accuracy of the prediction was obtained (within 5% accuracy of the measured values). The predictive performance of tool wear in machining using the GMDH approach has been proved to be superior to predictions using conventional analysis.