Artificial Neural Network-Based Prediction of Cutting Parameters from Tool Vibration and Forces

摘要

Automation in machining process becomes crucial in successfully maintaining high-quality and low-cost production. In automated machining process, the tool condition and the cutting parameters affect the quality of product. So it is required to monitor cutting parameters without interruption of machining operation. This can be done by monitoring the tool vibration continuously and predicting the cutting parameters using artificial neural network (ANN). This work aims at monitoring vibration of tool and feeding it as input to ANN for predicting the cutting parameters. This will help in reducing the production time by not interrupting the operation. Experiments were conducted to monitor tool vibration in conventional lathe for machining mild steel using HSS tool based on design of experiments. Tri-axial accelerometer was mounted on the tool to acquire vibration data using National Instruments Data Acquisition systems (NIDAQ). The force data is measured using a dynamometer attached to the lathe. The experimental values were used for developing a feed-forward backpropagation ANN model. The cutting parameters were predicted using the trained network, compared and found to be very close with the experimental values. It is concluded that the proposed ANN model is able to predict the cutting parameters which helps in monitoring the tool condition for good product quality.