Evaluating Lithium-Ion and Nickel-Metal Hydride Batteries in EVs: A Data-Driven Machine Learning Approach

The performance and longevity of Electric Vehicle (EV) batteries are critical for advancing sustainable transportation. This study presents a data-driven machine learning approach to systematically compare Lithium-Ion and Nickel-Metal Hydride batteries using real-world data from the U.S. Department of Energy (DOE) Battery Data Hub. Unlike previous research, which primarily focuses on isolated performance metrics, our study integrates multiple factors such as energy density, charge rate, cycle life, and thermal stability. We employ advanced machine learning models— LSTM, XGBoost, and Random Forest—to predict battery state-of-charge and state-of-health with high accuracy, achieving R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> values above 0.97 in SOC estimation using LSTM. Our findings confirm that Li-ion batteries outperform NiMH in key performance areas, exhibiting higher energy density, slower degradation rates, and better thermal stability. The results offer crucial insights for battery selection in EV applications and highlight the potential of ML-driven approaches in battery management systems (BMS) to enhance real-time monitoring and predictive maintenance. By optimizing battery performance and reducing premature failures, this study contributes to the sustainable adoption of EV technology.