Modeling and optimization of Wire –EDM parameters for machining of Ni54.1Ti45.9 shape memory alloy using hybrid approach. (October 2022)

• Record Type: Journal Article
• Authors: Gupta, Deepak Kumar; Dubey, Avanish Kumar
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 236:Number 5(2022)
• Page Start: 2176
• Page End: 2186
• Publication Date: 2022-10
• Subjects: Wire-EDM -- artificial neural network -- surface roughness -- material removal rate -- genetic algorithm -- shape memory alloy; Mechanical engineering -- Periodicals; Production engineering -- Periodicals; Manufacturing processes -- Periodicals; 621.05
• DOI: 10.1177/09544089221085144
• Languages: English
• ISSNs: 0954-4089

Abstract

Shape memory alloys are a distinct family of alloys recognized for their unique features, such as extreme anti-fatigue, exceptional specific strength, corrosion resistance, biocompatibility, and pseudo-elasticity. Due to strain-hardening effects displayed by the latter during a conventional machining process, these features make shape memory alloys particularly functional for machining utilizing non-traditional methods rather than conventional ones. The alloy's original qualities are at risk as a result of this consequence. Nickel-Titanium shape memory alloys with 54.1 and 45.9% Nickel-Titanium blends were employed in this investigation (Ni54.1 Ti45.9 ). The surface characteristics of Ni54.1 Ti45.9 alloy are investigated using the wire electrical discharge machining (WEDM) technique with a brass tool electrode (zinc-coated) and the influence of process parameters on it. Variations in wire feed rate, wire tension, pulse on and off duration, and peak current were used as input factors to investigate variations in surface roughness and material removal rate. Surface roughness and material removal rate data from L27 orthogonal array (OA) studies were used to create artificial neural network (ANN) models. To improve the quality attributes, a model combining a hybrid version of the genetic algorithm (GA) and an ANN has been presented. The ANN models were found to accurately predict the results, which matched the experimental data. After the adjustment, the quality features also improved significantly.