Integrating physics-based modeling with machine learning for lithium-ion batteries. (1st January 2023)

• Record Type: Journal Article
• Authors: Tu, Hao; Moura, Scott; Wang, Yebin; Fang, Huazhen
• Is Part Of: Applied energy. Volume 329(2023)
• Publication Date: 2023-01-01
• Subjects: Hybrid modeling -- Physics -- Machine learning -- Lithium-ion batteries; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2022.120289
• Languages: English
• ISSNs: 0306-2619

Abstract

Abstract: Mathematical modeling of lithium-ion batteries (LiBs) is a primary challenge in advanced battery management. This paper proposes two new frameworks to integrate physics-based models with machine learning to achieve high-precision modeling for LiBs. The frameworks are characterized by informing the machine learning model of the state information of the physical model, enabling a deep integration between physics and machine learning. Based on the frameworks, a series of hybrid models are constructed, through combining an electrochemical model and an equivalent circuit model, respectively, with a feedforward neural network. The hybrid models are relatively parsimonious in structure and can provide considerable voltage predictive accuracy under a broad range of C-rates, as shown by extensive simulations and experiments. The study further expands to conduct aging-aware hybrid modeling, leading to the design of a hybrid model conscious of the state-of-health to make prediction. The experiments show that the model has high voltage predictive accuracy throughout a LiB's cycle life. Highlights: New frameworks integrate physical models with machine learning for accurate voltage prediction. Novel hybrid models combine electrochemical and equivalent circuit models with neural networks. The hybrid models offer high predictive accuracy at relatively low computational costs. Experiments and simulations show the models are highly accurate across broad C-rate ranges. An aging-aware hybrid model as an extension shows accurate prediction throughout a cell's life.