A fuzzy-nets in-process (FNIP) system for tool-breakage monitoring in end-milling operations

Abstract

The fuzzy-nets in-process (FNIP) system is proposed for monitoring tool breakage in end-milling operations. The FNIP system consists of two components: (1) the fuzzy search classifier (FSQ, which maps a state vector into a recommended action using fuzzy pattern recognition; and (2) the fuzzy adaptive controller (FAQ, which maps a state vector and a failure signal into a scalar grade that indicates state integrity. The FAC also produces the output action value, p, to upgrade FSC mapping according to the variation of the input state. By coupling fuzzy logic control systems and neural networks into the fuzzy-nets system, a self-learning capability (ability to generate rule bases and to fine-tune the membership functions of each linguist variable to the appropriate level of granularity) was developed. With this on-line learning capability, the fuzzy rule bases of FSC and FAC are established by fine-tuning the parameters in the FNIP system. After establishing all the fuzzy rule bases, the performance of the FNIP system is examined for an end-milling operation. Experiments have shown that the FNIP system is able to detect tool breakage in the end-milling operation “on-line”, approaching a real-time basis.