Electric discharge coating (EDC) process is one of the unique and notable surface modification processes that enhance the surface properties of substrates. The coating layer thickness and surface hardness developed by the coating process are the predominant reason for substrate protection and performance enhancement. To enhance the coating efficiency and develop the desired output, optimization of input parameters and forecasting of output parameters was required. In this work, Mg alloy substrates are coated with Cu–Ni green compact electrodes by an EDC process. During the EDC process peak current ( I), pulse duration ( Ton), and compaction load for the electrode ( L) are major influencing parameters of coating layer thickness which are considered as input parameters. The coating layer thickness and surface hardness of the coating substrate were estimated by an optical microscope and Vickers hardness tester, respectively, and also considered as a response. The observed data were analyzed and optimized through analysis of variance (ANOVA) and response surface morphology (RSM) which provides a static relationship between the input parameter and response. Additionally, a prediction model was developed with a machine-learning approach. Two different tools namely artificial neural network (ANN) and, adaptive neuro-fuzzy inference system (ANFIS), were employed for accurate prediction of the model. From the ANN and ANFIS analysis regression coefficient (R) values were 0.97991 and 0.9912, respectively. The present study provides machine-learning capabilities and supports optimizing the input parameter to the target output in the EDC process.
Read full abstract