Linear discriminant function analysis of acoustic emission signals for cutting tool monitoring

Abstract

A principal setback to automation of the machining process is the inability to completely monitor the condition of the cutting tool in real time. Whereas several of the techniques developed to date are useful in specific applications, no universally applicable sensor is yet available.Acoustic emission is one of the most promising techniques to be recently developed for on-line cutting tool monitoring. However, signal analysis is still an area that requires further investigation to enhance the potential of acoustic emission. For this purpose, frequency-based pattern recognition concepts using linear discriminant functions have been used in analysing acoustic emission signals generated during machining to distinguish between different signal sources, specifically chip formation, tool fracture, and chip noise. Five features were used for classification in the frequency range of 100 kHz to 1 MHz, with each feature consisting of a 20 kHz bandwidth, and were selected using the class mean scatter criterion. The coefficients of the discriminant functions were obtained by training the system using signals generated by each of the sources of interest. An AISI 1018 steel was machined using a titanium carbide-coated cutting tool. Cutting speeds ranged from 200 to 800 ft/min (1 to 4 m/sec) with feed rats of 0[middle dot]0005 to 0[middle dot]0075 in/rev (0[middle dot]0133 mm/rev to 0[middle dot]191 mm/rev) and depth of cut 0[middle dot]17 in (4[middle dot]32 mm). The results show a successful classification rate of 90% for tool breakage, while those for chip formation and chip noise were 97 and 86% respectively.