Indirect method of tool wear measurement and prediction using ANN network in machining process

Abstract

In the present scenario, ensuring the quality of products is a major factor in the metal cutting industries and it demonstrates an overall success factor for businesses. Tool wear is one of the major factors which affect the quality of the product, production time, and also the manufacturing expenses in the metal cutting industry. The indirect method for tool condition monitoring system is very important for the increasing requirements of cost reduction and quality management in modern manufacturing industries. The indirect method of determining tool wear using Artificial Neural Network (ANN) has been proposed by evaluating force and vibration during dry turning. For the experiments, the cemented carbide insert and EN-8 (medium carbon steel) is utilized as workpiece material. Total nine experimental runs based on the Taguchi L 9 orthogonal array are conducted. Cutting speed, feed, and depth of cut is selected as machining parameters. Both vibration signal and cutting force measured during machining by measuring devices and its measured signal stored in the data acquisition system. The validity of the data used to predict the tool wear using the ANN model has been assessed by comparing with the manual measurement of tool wear. It has been found close co-relation between the predicted and measures tool wear during the machining. The study proves that the ANN is capable of predicting tool wear in machining for the considered machining parameters.