Experimental investigation and prediction of ECDM parameters on fiber reinforced SiC composite using hybrid ERNN-based Sparrow Search Optimization

Abstract

The Silicon Carbide (SiC) fiber-reinforced SiC ceramic matrix composites have proved their outstanding performance on high thermal applications such as gas turbine engine parts, turbo-pumps, nozzles, and various other aerospace/aero-propulsion system parts. Regardless of their enhanced strength properties, the SiC fiber-reinforced SiC ceramic composite materials are difficult to process using conventional machining methods. Also, SiC fiber-reinforced SiC ceramic promotes a higher tool wear rate while doing conventional drilling operations. Therefore, this study investigates the Electrochemical Discharge Machining (ECDM) parameters on such SiC fiber-reinforced SiC ceramic composite. Besides, the levels of most influencing ECDM input parameters Voltage, Electrolytic Concentration and Duty Cycle are optimized using a novel hybrid machine learning optimizer called Elman Recurrent Neural Network-based Sparrow Search Optimization (ERNN-SSO). The experiment is planned using Response Surface Methodology based on Box Behnken Design (RSM-BBD) and verified using ANOVA. The Material Removal Rate (MRR), Tool Wear Rate (TWR) and Overcut (OC) are considered output performances during this study. The proposed ERNN-SSO has optimized ECDM input parameters at a maximum of 0.03 root mean square error (RMSE), approximately 33 % and 95 % lower than the conventional ERNN and RSM optimization techniques. This study’s suggested ECDM input parameters for drilling SiC fiber-reinforced SiC ceramic composite are 95 V voltage supply, 10 %Wt. of electrolytic concentration, and 50 % duty cycle. Because, such optimum ECDM parameters produced a maximum MRR of 370 μg/min with minimum TWR and OC of 283 μg/min and 161 µm during this study.