FUZZY CLASSIFICATION OF MILLING TOOL WEAR

Modern advanced machining systems in the 'unmanned' factory must posse ss the ability automatically to change tools that have been subjected to wear or damage. This can ensure machining accuracy and reduce the p roduction costs. A practical on-line tool wear monitoring and classifi cation system is needed by industrial users and this paper presents an intelligent system for intermittent machining processes, such as mill ing. The system is fitted with multi-sensors to collect different sign als from the machining process and the data are processed by the use o f intelligent techniques. Different types of transducers were initiall y investigated during a large number of experiments and as a result, f our sensors, ie load, force, vibration and acoustic emission, were cho sen. Frizzy pattern recognition techniques have been used to accomplis h sensor fusion and tool weal stale classification.