Accurate forecasting of the lifetime and degradation mechanisms of lithium-ion batteries is crucial for their optimization, management, and safety while preventing latent failures. However, the typical state estimations are challenging due to complex and dynamic cell parameters and wide variations in usage conditions. Physics-based models need a tradeoff between accuracy and complexity due to vast parameter requirements, while machine-learning models require large training datasets and may fail when generalized to unseen scenarios. To address this issue, this paper aims to integrate the physics-based battery model and the machine learning model to leverage their respective strengths. This is achieved by applying the deep learning framework called physics-informed neural networks (PINN) to electrochemical battery modeling. The state of charge and state of health of lithium-ion cells are predicted by integrating the partial differential equation of Fick’s law of diffusion from a single particle model into the neural network training process. The results indicate that PINN can estimate the state of charge with a root mean square error in the range of 0.014% to 0.2%, while the state of health has a range of 1.1% to 2.3%, even with limited training data. Compared to conventional approaches, PINN is less complex while still incorporating the laws of physics into the training process, resulting in adequate predictions, even for unseen situations.

The concept of Digital Twin (DT) is widely explored in literature for different application fields because it promises to reduce design time, enable design and operation optimization, improve after-sales services and reduce overall expenses. While the perceived benefits strongly encourage the use of DT, in the battery industry a consistent implementation approach and quantitative assessment of adapting a battery DT is missing. This paper is a part of an ongoing study that investigates the DT functionalities and quantifies the DT-attributes across the life cycles phases of a battery system. The critical question is whether battery DT is a practical and realistic solution to meeting the growing challenges of the battery industry, such as degradation evaluation, usage optimization, manufacturing inconsistencies or second-life application possibility. Within the scope of this paper, a consistent approach of DT implementation for battery cells is presented, and the main functions of the approach are tested on a Doyle-Fuller-Newman model. In essence, a battery DT can offer improved representation, performance estimation, and behavioral predictions based on real-world data along with the integration of battery life cycle attributes. Hence, this paper identifies the efforts for implementing a battery DT and provides the quantification attribute for future academic or industrial research.