Publication

Abstract : Lithium-ion batteries (LIBs) are an important component in applications such as electric vehicles, energy storage systems, and consumer electronics due to their high efficiency and high energy density. Accurate State of Health (SoH) prediction is critical to ensure battery performance, safety, and longevity. Traditional methods often struggle to capture the complex nonlinear relationships inherent in battery behavior. This paper investigates the use of Recurrent Neural Networks (RNNs) and their variants, Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU) for SoH computation. The comprehensive datasets are collected in the battery parameters, including voltage, current, temperature, and capacity. These data were preprocessed to address noise and inconsistencies of the lithium-ion batteries. RNN models, including LSTM and GRU, were trained and evaluated using the preprocessed dataset. The performance of the models was assessed based on metrics such as Mean Squared Error (MSE) and R-squared (R2). The results show that RNN algorithms capture the non-linearities of battery parameters well, improving SoH prediction accuracy. For example, the RNN model achieved the lowest Mean Squared Error (MSE) of 0.000014 and the highest R2 of 0.974455 for the B46 cells, demonstrating its robustness and accuracy. This study highlights the potential of deep learning algorithms to enhance battery management systems (BMS) and extend the service life of lithium-ion batteries.