Optimization-Based Hybrid Intelligent Model for Decision Making on Electrical Discharge Machining (EDM) Process of A6061/6%B4C and A6061/9%SiC Composite Materials

Abstract

This study intends to suggest an improved deep learning decision-making model on the die-sinking electrical discharge machining (EDM) process of A6061/6%B4C and A6061/9% silicon carbide (SiC) composite materials. The proposed model consists of feature extraction and prediction stages. An extraction of statistics and much more advanced statistical characteristics is the first phase in the feature extraction process. The hybrid classifier that mixes convolutional neural network (CNN) and neural network (NN) is then fed the features. The final projected outcome is attained by averaging the results from both CNN and NN. An upgraded Sea Lion with Customized Levy Flight (SL-CLF) is suggested as a significant improvement to the classifier prediction accuracy. Finally, a number of indicators are employed to demonstrate the superiority of the SL-CLF strategy. At last, several metrics are used to show the superiority of SL-CLF approach. The mean standard error (MSE) of the proposed work for dataset 1 is 4. 965, which is 9. 7%, 38%, 59. 89%, and 59. 89% better than CNN, NN, Taguchi Coupled grey relational analysis (GRA), and ANN.