Advancing Wire Electrical Discharge Machining of AISI P20+Ni Steel using Hybrid Deep Belief Neural Network - Search and Rescue Optimization Algorithm

Abstract

AISI P20+Ni steel is utilized in vehicle structures for components requiring high strength and toughness, such as injection molds, die casting dies and body panels. Its exceptional machinability and impact resistance ensure durability and safety, contributing to the structural integrity of automotive assemblies. The present research work examines the machinability of AISI P20+Ni steel using Wire Electrical Discharge Machining (WEDM) process. Various cutting wires are investigated, including zinc-coated brass wire, cryogenically treated zinc-coated brass wire and brass wire with ultrasonic vibration (UVBW). Optimal machining conditions for high productivity in terms of material removal and surface quality in terms of minimum roughness value are determined using Response Surface Methodology (RSM) and the Search and Rescue Optimization Algorithm (SAR), considering factors such as servo voltage, pulse off time, pulse on time and peak current. Experimental findings identify UVBW as the most effective wire electrode for machining of AISI P20+Ni, achieving a desirability of 0.722 through RSM and after 87-92 iterations with SAR. Furthermore, a comparison of prediction techniques, including Deep-Belief-Neural-network (DBN), hybrid DBN-SAR and RSM methods, illustrates the superior accuracy of the hybrid DBN-SAR approach in predicting WEDM process parameters, thus enhancing machining process efficiency and precision.