Surface Roughness Prediction Modelling for WEDM of AA6063 Using Support Vector Machine Technique

Abstract

This research work presents an incorporated approach to modelling of WEDM of AA6063 (armour applications) using support vector machine technique. The experimental investigation has been carried out with four input variables namely pulse-on-time (Pon), pulse-off-time (Poff), servo-voltage (VS) and peak-current (IP). Surface roughness is measured as response parameter. The experimental runs are designed according to 3k full factorial design (k is number of input variables). It is apparent from this study that values anticipated by developed model are found closer to experimental results. Thus, it ensures appropriateness of model for prediction purpose and smart manufacturing. Machined surfaces are also examined by SEM to critically evaluate the process.