Meso-level intrinsic investigation of PMEDM process using waste oil blend underpinning cleaner production: experimental and machine learning approaches

Abstract Inconel 718 (IN718) is a nickel (Ni)-based superalloy that is extremely valuable for its vital uses in spacecraft, rocket engines, and nuclear reactors. It is notoriously hard to machine; as a result, a possible substitute is provided by electric discharge machining (EDM), which erodes the material by repeatedly releasing sparks in a dielectric fluid, usually kerosene oil (KO). Since KO releases toxic fumes during EDM, scientists have investigated different biodegradable oils to cut down on using KO while still getting good results. This study investigates the combinations of graphene nanoparticles (GNPs), waste oil blend (WOB), and surfactant (Tween80) to get better machining results. For the experiments, the Taguchi L18 design of experiments was employed. Significant processing parameters have undergone analysis of variance (ANOVA). Following that, a model of an artificial neural network (ANN) was used to accurately predict the response measures. Following the modelling process, the nondominated sorting genetic algorithm (NSGA-II), which is based on machine learning (ML), was used to determine the ideal set of parameters. The combined influence of peak current (I P ) and surfactant concentration (S C ) shows that the magnitudes of material removal rate (MRR), surface roughness (SR), electrode wear rate (EWR), and accuracy index (AI) improved by 5.93 times, 1.27 times, 1.08 times, and 1.55%, respectively, when comparing lower and higher parametric states. In the confirmatory experiment, SR and EWR decreased by 9.80% and 16.14%, respectively, while MRR and AI increased by 90.83% and 1.95%, according to the NSGA-II results. As a result, energy savings using WOB in the EDM process are computed, and the potential for CO 2 reduction is then evaluated. It has been found that the percentage reduction in CO 2 emissions for WOB has been recorded as 77.60 ± 3.48% in comparison to the KO. Therefore, WOB dielectric can be an ideal alternative to KO in EDM applications.