A computationally efficient model for performance prediction of lithium-ion batteries

The pseudo-two dimensional electrochemical model is capable of accurately predicting the transient behavior of the batteries. However, since the numerical complexity of this model prohibits its application in real-time simulations, many efforts have been made toward reducing the model order to develop fast and reliable battery modeling methods. In this paper, a computationally efficient electrochemical-based model is proposed to predict the temporal and spatial distributed processes inside the battery. By considering some reasonable assumptions, complex partial differential equations (PDEs) are simplified so they can be analytically solved. Applying Green's function method, the electrolyte concentration distribution is obtained, and solid concentration is approximated using a simplified expression. Verifying the results with the previous full order model shows high precision in low C-rates; even in high applied currents (4C), the model can deliver reasonable accuracy. Compared to the CFD method, this model is highly efficient and significantly reduces computing time because of utilizing linear algebraic equations.

Automated summary

The paper proposes a computationally efficient electrochemical model that accurately predicts the temporal and spatial distributed processes inside batteries by reducing model order and simplifying partial differential equations. Compared to the previous full order model, this model demonstrates high precision at low C-rates and reasonable accuracy even at high applied currents. The model is highly efficient and significantly reduces computing time compared to the CFD method by utilizing linear algebraic equations.