WEDM of complex profile of IN718: multi-objective GA-based optimization of surface roughness, dimensional deviation, and cutting speed

Abstract

Over the past decade, components featuring complex curved geometries have become a prerequisite for the high performance of aeroengines, thus necessitating stringent limits for excellent surface quality and high dimensional precision. As such, the manufacturing of complex features presents machining challenges. Wire electric discharge machining is a non-conventional process that has the potential to effectively mitigate these challenges, but further research is needed. Specifically, the coated wire electrodes that are incorporated in machining have been examined in terms of their impact on cutting speed, but their influence on the surface roughness and the dimensional accuracies is somewhat unexplored. In this experimental study, the influence of zinc-coated copper wire (BroncoCut-X) and corresponding process parameters (wire tension, pulse-on-time, pulse-off-time, servo voltage, and wire feed) were scrutinized to machine a complex profile in IN718 superalloy. The response characteristics associated with the complex profile, namely, surface roughness, dimensional deviation, and cutting speed, were thoroughly investigated. The machining results were analysed through main effect plots, analysis of variance, and scanning electron microscopy. The results revealed that BroncoCut-X was effective in machining the complex curved geometry profiles (convex, concave, straight, and inclined) and produced the best results for the concave geometrical feature. Considering the conflicting nature of response attributes, multi-objective genetic algorithm was employed to obtain the best combination of control parameters for the each geometry of the complex profile. The best optimal solution set the values of control variables at 2.6405 g wire tension, 2.9931 μs pulse-on-time, 22.4108 μs pulse-off-time, 54.620 V servo voltage, and 4.1730 mm/s wire feed. The exactness of multi-objective genetic algorithm was verified through confirmatory tests wherein the average percentage error was found less than 2%.