A neurofuzzy classification network and its application

Abstract

An important element of the automatic machining process control function is the online monitoring of cutting tool wear and fracture mechanisms. The paper presents an intelligent tool condition monitoring system. Cutting tool condition monitoring is a very complex process hence sensor fusion techniques and artificial intelligence based signal processing algorithms are employed. The tool condition monitoring system is equipped with four kinds of sensors, signal transformation and collection apparatus and a microcomputer. Multi-sensor signals reflect the tool condition comprehensively. Redundant signal features are removed by using a fuzzy clustering feature filter. A unique neurofuzzy classification network has been developed to carry out the fusion of multi-sensor information and tool wear classification. It combines the transparent representation of fuzzy systems with the learning ability of neural networks hence the algorithm has strong modelling and noise suppression ability. Successful tool wear classification can be realized under a range of machining conditions. A large number of monitoring experiments suggests that the proposed intelligent data processing algorithm is practical and reliable.