Insights into the estimation of capacitance for carbon-based supercapacitors

Area of Science:

• Materials Science
• Electrochemistry
• Machine Learning

Background:

• Carbon-based materials are crucial for high-power-density electric double-layer capacitors (EDLCs).
• Existing computational studies often focus on equilibrium conditions, limiting practical application in energy system design.
• A predictive model is needed to link material properties to EDLC performance.

Purpose of the Study:

• To establish a correlation between structural features of carbon materials and EDLC performance.
• To develop a predictive model for EDLC capacitance using experimental data.
• To identify key structural features influencing EDLC capacitance.

Main Methods:

• Extracted experimental data from over 300 published papers on carbon-based EDLCs.
• Employed an optimized support vector machine (SVM) with a grey wolf optimization (GWO) algorithm.
• Utilized structural features (pore size, specific surface area, N-doping, I_D/I_G ratio, potential window) as input variables.

Main Results:

• Sensitivity analysis revealed specific surface area as the most critical factor affecting EDLC capacitance.
• The proposed SVM-GWO model achieved a high accuracy with an R² value of 0.92.
• The SVM-GWO model outperformed other machine learning models for EDLC capacitance prediction.

Conclusions:

• The study successfully correlated carbon material structure with EDLC performance using an advanced ML approach.
• Specific surface area is identified as the dominant factor for optimizing EDLC capacitance.
• The developed SVM-GWO model offers a powerful tool for designing high-performance carbon-based supercapacitors.