Prognostics of the Lithium-Ion Battery Based on Deep Learning Algorithms: Comparative Study

Precise predictions of the remaining useful life (RUL) of batteries are becoming increasingly vital in intelligent battery health management systems. Progress in deep learning offers novel data-driven methodologies to address this issue. This paper presents a comprehensive deep learning methodology for forecasting the remaining useful life of batteries through the utilization of deep neural networks (DNNs). Initially, data are acquired from the NASA Li-batteries. A deep neural network is trained using the remaining useful life prediction model for batteries. This methodology is utilized across multiple deep learning models, and the comparative analysis of the results indicates that the CNN-LSTM model yields superior accuracy, necessitating further efforts to enhance these models for improved remaining useful life prediction.