ScholarWorks@경상국립대학교

초록

Heteroatom-doped graphene supercapacitors have attracted considerable attention owing to their high specific capacitance, excellent rate capability, and superior cycling stability. In this study, an artificial neural network (ANN) model was developed to predict the specific capacitance (F g-1) of heteroatom-doped graphene electrodes based on a dataset comprising 646 data points collected from 55 published studies. Eight input parameters were considered, including the nature of the electrolyte (acidic or basic, represented by binary encoding: 0 = absent, 1 = present), specific surface area (m2 g-1), Raman intensity ratio (ID/IG), nitrogen content (N%), oxygen content (O%), sulfur content (S%), and current density (A g-1). The optimized ANN architecture (8–36–36–36–1) achieved a low root mean square error (RMSE = 0.0002) and a high coefficient of determination (R² = 0.85) for the testing dataset. Single- and two-variable sensitivity analyses were carried out to elucidate the individual and combined effects of structural and electrochemical variables on capacitance. Furthermore, quantitative estimations and an index of relative importance (IRI) were employed to rank variable influence, providing interpretability and physical insight. The developed ANN-GUI framework enables accurate prediction and mechanistic understanding of heteroatom-doped graphene supercapacitor behavior, offering a powerful tool for the rational design of next-generation energy storage materials. © 2026 Elsevier Ltd.

키워드