Neural Network (NN)-Based RSM-PSO Multiresponse Parametric Optimization of the Electro Chemical Discharge Micromachining Process During Microchannel Cutting on Silica Glass

Abstract

The production of miniature parts by the electrochemical discharge micromachining process ([Formula: see text]-ECDM) draws the most of attractions into the industrial field. Parametric influences on machining depth (MD), material removal rate (MRR), and overcut (OC) have been propounded using a mixed electrolyte (NaOH:KOH- 1:1) varying concentrations (wt.%), applied voltage ([Formula: see text]), pulse on time ([Formula: see text]s), and stand-off distance (SOD) during microchannel cutting on silica glass (SiO[Formula: see text]). Analysis of variances has been analyzed to test the adequacy of the developed mathematical model and multiresponse optimization has been performed to find out maximum MD with higher material removal at lower OC using desirability function analysis as well as neural network (NN)-based Particle Swarm Optimization (PSO). The SEM analysis has been done to find unexpected debris. MD has been improved with better surface quality using a mixed electrolyte at straight polarity using a tungsten carbide (WC) cylindrical tool along with [Formula: see text], [Formula: see text], and [Formula: see text] axis movement by computer-aided subsystem and combining with the automated spring feed mechanism. PSO-ANN provides better parametric optimization results for micromachining by the ECDM process.