Assessment of chip forming patterns with tool wear progression in machining via neural networks

Abstract

This paper presents a new method of employing techniques of neural networks to quantify the complicated interrelationship between the change of chip breakability and that of comprehensive wear states, including major flank, crater and minor flank wear. Chip breakability under unworn cutting tools is first predicted through a fuzzy rating system, then updated dynamically as tool wear develops. Change in surface finish with tool wear progression is assessed via the neural networks and finish conditions were used where both chip breakability and surface finish are primary concerns. Both training and testing results show that the method is not only valid and effective, but also provides a feasible means for in-process prediction of chip breakability and surface finish in automated finish-machining systems.