Prediction of Machining Force and Surface Roughness in Ultrasonic Vibration-Assisted Turning Using Neural Networks

Abstract

In present study, neural networks have been employed for studying the ultrasonic vibration-assisted turning (UAT) process and for predicting the machining force and workpiece's surface roughness. Extensive experiments were carried out using different values of UAT parameters such as vibration amplitude, depth of cut, feed rate and cutting speed. The tests were implemented on the basis of full factorial design of experiments for three different levels of each UAT parameter. The machining force and workpiece's surface roughness were measured as the responses of the experiments and were subsequently modeled with the aid of back propagation multilayer perceptron neural network for 1.1191 steel. The nonlinear relation existing between the aforementioned UAT parameters and the machining force and workpiec's surface roughness could effectively be modeled by the developed networks and the responses error could be kept less than ten percent. This was verified by further experiments different from those carried out for developing the network.