Artificial Neural Network and Multiple Linear Regression Approach for Optimization of Material Composition for Sustainable Super Capacitor

The proposed capacitor is made from natural materials like soil, river sand, and magnetite powder. The electrode is the capacitor’s major component, acting like a conductor of electric storage. In this experimental investigation, sustainable materials are used as the substrates of electrodes to improvise and increase the efficiency of the capacitor. To optimize the ideal materials machine learning is brought into this study. Regression and ANN models are the most popular techniques for predicting values and identifying similarities that are used to optimize the ideal materials for manufacturing a sustainable capacitor. In preliminary tests, the soil capacitor device demonstrated the best electrochemical performance. Mix ratios were determined from the preliminary study conducted on the materials used as dielectrics and substrates in capacitors. These low-cost, high-performance capacitors’ ideal combination was identified using the evolutionary algorithm. The optimization models were evaluated using matrix assessments such as MSE, and RMSE values. From the analysis, the optimization of materials, compositions, and performance were determined for the manufacturing of sustainable supercapacitors.