Artificial neural networks modeling of surface finish in electro-discharge machining of tool steels

Abstract

Electro-Discharge machining (EDM) is a thermal process with a complex metal removal mechanism that involves the formation of a plasma channel between the tool and the workpiece electrodes and the melting and evaporation of material resulted thus in the generation of a rough surface consisting of a large number of randomly overlapping craters and no preferential direction. EDM is considered especially suitable for machining complex contours, with high accuracy and for materials that are not amenable to conventional removal methods. However, certain phenomena negatively affecting the surface integrity of EDMed workpieces, constrain the expanded application of the technology. Accordingly, it has been difficult to establish models that correlate accurately the operational variables and the performance towards the optimization of the process. In recent years, artificial neural networks (ANN) have emerged as a novel modeling technique that is able to provide reliable results and it can be integrated into a great number of technological areas including various aspects of manufacturing. In this paper ANN models for the prediction of the surface roughness of electro-discharge machined surfaces are presented. A feed-forward artificial ANN trained with the Levenberg-Marquardt algorithm was finally selected. The proposed neural network takes into consideration the pulse current and the pulse-on time as EDM process variables, for three different tool steels in order to determine the center-line average (R a) and the maximum height of the profile (R t) surface roughness parameters. Copyright © 2006 by ASME.