Viliam Oleksandrovych Zaloha

ANFIS Building Methodology for the Task of Cutting Tool Condition Diagnosis Using Matlab Software

Machining of materials with cutting still covers a significant part of shapegenerating operations in manufacturing process. One of the most important tasks in research touches upon the area of cutting is the development of a method which can insure: required productivity, high accuracy of machining, optimal usage of cutting tool and machine-tool resource, automation of manufacturing process, reduction of machinetool down time and cutting tool costs. In this connection cutting tool condition diagnosis (CTCD) becomes an important requirement for the realization of computer-aided manufacturing. The real-time CTCD allows improving the efficiency of machining with the opportune cutting tool replacement and the prior prevention of its catastrophic wear or breakdown. The main goal of this investigation is the verification of the ANFIS working capacity for the description of the relation between the flank wear of cutting tool and the power of vibration signal received during CTCD in turning by means of Matlab software. Consequently, the methodology of building the adaptive neuro-fuzzy inference system (ANFISnetwork) for the task of cutting tool condition diagnosis is worked out.