A CUSTOMIZED NEURAL-NETWORK FOR SENSOR FUSION IN ONLINE MONITORING OFCUTTING-TOOL WEAR

A customized neural network for sensor fusion of acoustic emission and force in on-line detection of tool wear is developed. Based on two cr itical concerns regarding practical and reliable tool-wear monitoring systems, the maximal utilization of ''unsupervised'' sensor data and t he avoidance of off-line feature analysis, the neural network is train ed by unsupervised Kohonen's Feature Map procedure followed by an Inpu t Feature Scaling algorithm. After levels of tool wear are topological ly ordered by Kohonen's Feature Map, input features of AE and force se nsor signals are transformed via Input Feature Scaling so that the res ulting decision boundaries of the neural network approximate those of error-minimizing Bayes classifier. In a machining experiment the custo mized neural network achieved high accuracy rates in the classificatio n of levels of tool wear. Also, the neural network shows several pract ical and reliable properties for the implementation of the monitoring system in manufacturing industries.