Experimental investigations of electrical discharge micro-drilling for Mg-alloy and multi-response optimization using MOGA-ANN

Abstract

Electric Discharge Drilling (EDD), a spark erosion drilling process, is designed to produce small to large size holes in high-strength and heat-resistant materials with the greatest precision. Magnesium (Mg) based alloy is potential material for biodegradable medical implants, and EDD can be employed to create perforated implants for tissue engineering. The present study was carried out to investigate the EDD process for Mg-alloy using a face-centered central composite design (FCCD) and the effect of input process variables such as discharge current (I p ), spark-on time (T on ), spark-off time (T off ) and work height (WH) have been evaluated on response characteristics; drilling rate (DR), tool wear ratio (TWR) and hole overcut (HO). Analysis of Variance (ANOVA) suggests that Ip, T on , T off and WH are the most significant factor affecting the response characteristics under 95% confidence level. Based on the response surface methodology (RSM), regression equations have been developed between the input process variables and response characteristics. Further, an artificial neural network (ANN) model is developed, and a multi-objective genetic algorithm (MOGA) was utilized to optimize the multi-response objectives. It has been found that the best combination of input parameters while using RSM is: Ip; 4 amp, T on ; 30 µs, T off ; 48 µs, WH: 10 mm, and the predicted values of DR, TWR and HO were Ip; 129.22 mm/min, TWR; 0.142 and HO; 87.26 µm respectively whereas in the Multi-objective Genetic algorithm artificial neural network (MOGA-ANN) hybrid technique, optimal balance results achieved are DR of 142.527 mm/min, with TWR and HO 0.168, 98.258 µm respectively at optimal input parameters. The optimal results obtained with MOGA-ANN are validated experimentally.