Application of Machine Learning Techniques for Predicting the State of Health of Lithium-Ion Batteries

This research applies machine learning techniques to predict the state of health (SOH) of lithium-ion batteries, crucial in modern electronics and sustainable energy systems. It compares various machine-learning methods using the NASA Prognostics Center of Excellence dataset, adopting a unique approach of dividing the dataset based on entire batteries rather than individual charging cycles, coupled with meticulous hyperparameter optimization for each model. It explores MLP Neural Networks, CNN, CatBoost, and XGBoost, evaluating their performance based on Mean Squared Error (MSE), R-squared values, grid search and prediction times. This study offers a comparative analysis of machine learning methods in battery health assessment, highlighting their potential in enhancing battery management systems and advancing lithium-ion battery sustainability and efficiency.