(540c) Optimization of Energy Density in Supercapacitors By Utilizing a Hybrid Artificial Neural Networks-Genetic Algorithm Based Optimization Algorithm

Supercapacitors are promising energy storage devices with long cycle life and high-power density. Nonetheless, they have limited energy density. In this work, we utilize a hybrid Artificial Neural Network-Genetic Algorithm (ANN-GA) model to optimize the energy density of supercapacitors while maintaining their high power density. We extract 2370 data points from the literature on carbon-based supercapacitors, and analyze the dataset using simple descriptive statistics to show the general trends in the field. Then, we perform preprocessing and feature selection, and apply machine learning algorithms to predict energy storage performance. Artificial neural networks yield the best accuracy. Consequently, we use the trained neural network as the fitness function for genetic algorithm to identify a set of optimal parameters that yield improved energy and power performance. Our findings indicate that obtained optimal design parameters are consistent with the literature. In a nut shell, this study takes a step towards the rational design of supercapacitors by implementing a hybrid ANN-GA as an optimization tool to improve energy storage performance.