Surface roughness modeling in high speed turning of Ti-6Al-4V – Artificial Neural Network approach

Abstract

Surface roughness modeling is considered as a complicated task when uncontrollable parameters come into existence. This study deals with the modeling of surface roughness parameters Ra and Rt by considering machining parameters like cutting speed, feed rate and depth of cut along with uncontrollable parameters like tool flank wear and cutting tool vibrations in high speed dry turning of Ti-6Al-4V. Three Artificial Neural Network (ANN) techniques namely Multi Layered Perceptron (MLP), Radial Basis Function Neural Network (RBFNN) and Summation Wavelet – Extreme Learning Machine (SW-ELM) have been implemented using experimental data obtained from turning experiments using uncoated carbide inserts. A comparison has been done among the three techniques. SW-ELM outperformed with its prediction accuracy, MSE and execution time when compared to MLP and RBFNN.