Lithium-ion Battery Aging Prediction with Electrochemical Models: P2D vs SPMe

Abstract

Battery aging models are essential tools when predicting how much a battery will age under certain working conditions, which is key when sizing a battery pack and controlling its operation. Nowadays, mostly empirical battery aging models are used, which require a high amount of long degradation experiments, and a lot of facilities are needed to perform these tests (cyclers, climate chambers … ). Due to the better predictability of physics-based models (PBMs), its use could reduce the costs of this process by decreasing the number of experiments. For that, an appropriate physics-based aging model must be selected. Hence, in this work we have compared two of the most used PBMs: the pseudo-two-dimensional (P2D) model and the single particle model with electrolyte dynamics (SPMe). We have analyzed their battery aging prediction accuracy as well as the computational cost. The results show that the SPMe can predict capacity fade with high accuracy compared to the P2D model, while the computational cost is reduced significantly. However, some gradients of internal mechanisms cannot be captured with the SPMe, which may generate differences when predicting internal aging variables.