Prognostics and health management of Lithium-ion battery using deep learning methods: A review

• Published in: Renewable & sustainable energy reviews Vol. 161; p. 112282
• Main Authors: Zhang, Ying, Li, Yan-Fu
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.06.2022
• Subjects: Deep learning; Lithium-ion battery; Prognostics and health management; Deep learning; Lithium-ion battery; Prognostics and health management
• ISSN: 1364-0321; 1879-0690
• DOI: 10.1016/j.rser.2022.112282

Prognostics and health management (PHM) is developed to guarantee the safety and reliability of Lithium-ion (Li-ion) battery during operations. Due to the advantages of deep learning on nonlinear modeling and representation learning, it gains considerable attentions in the PHM of Li-ion battery. To provide a comprehensive view of deep learning-based PHM of Li-ion battery, this paper summarizes these applications on the basis of current research. Deep learning-based PHM of Li-ion battery roughly involves three steps, namely data acquisition, deep learning methods and performance evaluation. Firstly, regular data types and public datasets are introduced. Secondly, brief introductions of deep learning methods and their applications to PHM of Li-ion battery are summarized. These deep learning methods include autoencoder, deep neural network, deep belief network, convolutional neural network, recurrent neural network and generative adversarial network. Thirdly, commonly-used evaluation metrics are presented. Finally, the paper draws a conclusion and presents the prospects of PHM of Li-ion battery with deep learning techniques. •This paper summarizes applications of deep learning to the PHM of Li-ion battery.•This paper summarizes the typical data used in the PHM of Li-ion battery.•This paper summarizes the public datasets used in the PHM of Li-ion battery.•Future works of deep learning-based PHM of Li-ion battery are given in seven aspects.