Monitoring tool wear by visual pattern recognition

This paper reports the results of acquiring vibration readings during the cutting operation of a centre lathe and analysing the data off-line with a view to determining the state of the tool tip during the cutting procedure. The process of using visual pattern recognition to measure the signature analysis of the data and to distinguish differing classes of tool wear will be discussed. The viability of such a system on the shop floor and the potential future applicability will be illustrated. The results show that the system is capable of classifying variant degrees of tool tip flank wear from the vibration readings, however, the process is too unwieldy and requires too much acumen for it to be transferred easily to the shop floor.