Data-Driven Approaches for Estimation of EV Battery SoC and SoH: A Review
Shahid Gulzar Padder, Jayesh Ambulkar, Atul Banotra, Sudhakar Modem, Sidharth Maheshwari, Kolleboyina Jayaramulu, Chinmoy Kundu
Abstract
Electric vehicle (EV) technologies have marked a staunch foundation in the transportation industry. The precise assessment of State of Charge (SoC) as well as State of Health (SoH) is essential for problems like range anxiety and unanticipated breakdown in EVs. In that regard, we have examined various methodologies, including traditional methods like Coulomb Counting (CC) and Open Circuit Voltage (OCV), advanced filter-based approaches, and contemporary data-driven methods. An extensive evaluation of different methods, along with the identification of strengths and weaknesses, is discussed. Data-driven estimation using Machine learning algorithms demonstrates superior accuracy and adaptability in sophisticated battery management systems. External battery parameters such as voltage, current, time, and temperature (V.C.T.T) and internal battery parameters such as impedance and ultrasonic data are the principal constituents of the Data-driven approaches. In this study, machine learning algorithms exhibited substantial enhancements in predicting and maintaining the lifespan of electric vehicle batteries. Nevertheless, there remains a requirement for ongoing advancement in battery systems to up-hold environmentally friendly transportation and incorporate pioneering estimation techniques to improve the reliability and lifespan of batteries.