Publication

Abstract : This article provides a survey of recent studies that focus on the estimation of State of Charge (SoC), State of Health (SoH), and Reliable Useful Life (RUL) of batteries using traditional and machine learning methods. The accurate estimation of SoC and SoH is critical for the effective management of batteries in renewable energy systems and electric vehicles. Machine learning algorithms such as support vector machines, decision trees, artificial neural networks, and ensemble methods have been widely employed for SoC and SoH estimation. These techniques utilize data obtained from various sensors such as battery voltage, current, temperature, and other measurements to train models capable of accurately predicting SoC and SoH. The article reviews the advantages and disadvantages of these approaches and highlights the importance of precise SoC, SoH, and RUL estimation for ensuring battery performance, safety, and reliability. Finally, the paper proposes future research directions, including the integration of machine learning with physical battery models and the use of advanced sensor technologies for more precise data acquisition.