Dual-Continuum Models of Lithium-Ion Batteries are Fast and Accurate Alternatives to the Doyle-Fuller-Newman Approach: I. Derivation and Validation

One of the main reasons for the Doyle–Fuller–Newman (DFN) model’s success in simulating lithium-ion batteries is also one of its main limitations—its hybrid micro/macro formulation. On one hand, this captures the slow diffusion of lithium within the active material particles. On the other hand, it makes the retention of realistic particle geometries computationally prohibitive, thereby provoking strongly simplifying assumptions on their shape. Dual-continuum models employ a fully macroscale description of the battery, and thus can potentially circumvent these challenges. Despite their widespread use in other fields, they have seen little application in battery modelling. In this work, we derive a dual-continuum model for lithium-ion batteries using the volume-averaging technique. A novel mapping between the volume-averaged and surface-averaged active material concentrations is introduced, based on the microscale source terms that generate concentration fluctuations. Unlike the DFN model, this approach makes no assumptions about particle shape but instead relies on a closure problem solved on the electrode microstructure. The resulting model is validated against a detailed microscale model, the DFN model, a recent dual-continuum formulation, and experimental data. Across all cases considered, our dual-continuum model reproduced cell-voltage data more accurately than the DFN approach while requiring 70–80% less computation time.