Artificial neural networks and multi response optimisation on EDM of aluminium (A380)/fly ash composites

Abstract

Aluminium metal matrix composites reinforced with fly ash particles of three different particle size ranges [(53-75) µm, (75-103) µm and (103-125) µm] were fabricated using stir casting technique. Electrical discharge machining was employed to machine the composite materials with copper electrode. The influence of EDM process parameters namely peak current, pulse-on-time, pulse-off-time, particle size and the percentage fly ash on material removal rate, tool wear rate and surface roughness were investigated. Artificial neural network model was employed to predict the material removal rate, tool wear rate and surface roughness of the composites. The experimental values coincide with the predicted values of the proposed networks. The process parameters are then optimised using desirability-based multi response optimisation technique to maximise the MRR and minimise both TWR and SR. Increase in peak current and pulse-on time increased the MRR while increase in pulse-off time, percentage fly ash and fly ash particle size decreased the MRR. The experimental results along with the ANN model and multi response optimisation will serve as a technical database for aerospace, automotive, military and commercial applications.