Tool wear monitoring in face milling using force signals

Abstract

The primary objective of this research is to monitor tool wear in face milling on line. In this paper, two approaches to monitoring tool wear in face milling are presented. The first approach adopts neural network techniques to identify the tool wear conditions. The inputs to the neural network are the mean values of cutting forces and other known cutting parameters such as feed rate, and workpiece geometry. The neural network is trained to estimate the average flank wear on cutter inserts. The other approach uses a regression model to estimate tool wear. The regression model is established based on data obtained from experiments. It is confirmed experimentally that the tool wear can be well estimated by both approaches when cutting aluminum with a multi-tooth cutter and different workpiece geometries.