Artificial Intelligence-Driven Optimization Of Supercapacitor Performance: A New Frontier In Energy Storage

Abstract

Background: Supercapacitors are increasingly becoming relevant in energy storage due to their performance characteristics (high power density, rapid charge and discharge cycles, and long lifespan). However, optimizing their performance hinges on the selection of advanced materials, electrolyte type, and charge transfer efficiency. Optimizing supercapacitor efficiency with AI (Artificial Intelligence) is a step in the right direction, but systematic evaluation of impact is still lacking. Objective: This study endeavours to determine the effects of AI optimization on supercapacitor performance by looking at the charge transfer efficiency as a mediator, along with the structural composition of the electrodes, the type of electrolyte used, and AI algorithms applied as primary independent variables that increase power and energy density within certain environmental conditions. Methods: A quantitative approach was adopted and data collection was accomplished using a structured questionnaire completed by using a Likert scale to elicit responses from participants. Responses were sought from 355 selected purposively sampled experts in energy storage and AI research. Simple descriptive analysis, reliability analysis with Cronbach's Alpha, normality with Shapiro-Wilk, correlation with Spearman's Rank, and linear regression analysis were conducted to test the relationships between the variables. Furthermore, inter-rater agreement was examined using Fleiss’ Kappa, and structural equation modelling (SEM) was used to test for mediation and moderation. Results: The dataset was marked as non-parametric based on the normality test. The analysis confirmed the strong internal consistency of the survey. Energy density was only minimally affected by electrode material composition, electrolyte type, and AI algorithm selection. Thus, it can be assumed that factors like charge transfer efficiency and the surrounding environment also have a considerable impact. Test-retest reliability demonstrated through Spearman’s correlation showed that responses over time were moderately stable, with some responses varying due to changing industry views. Conclusion: Although there is evidence suggesting the use of AI for optimizing supercapacitor performance, the arguments are not statistically significant enough to overlook other possibilities. Further optimization of supercapacitors' performance forecasting would benefit from more sophisticated AI models, hybrid energy storage solutions, and validation of models against operational data. These advancements would serve as a bridge toward suggested improvements in supercapacitors’ performance. These insights will be beneficial for the academic field, industry, and policymakers regarding the application of AI in energy storage systems.

Keywords: Artificial Intelligence, Supercapacitor Optimization, Energy Storage, Machine Learning, AI-driven Material Selection, Charge Transfer Efficiency, Renewable Energy Systems.