Title:Optimal convergence in finite element fully discrete error analysis and a novel fast solver for the Doyle-Fuller-Newman model of lithium-ion batteries

Abstract:We investigate the convergence of a backward Euler finite element discretization applied to a multi-domain and multi-scale elliptic-parabolic system, derived from the Doyle-Fuller-Newman model for lithium-ion batteries. Our analysis establishes optimal-order error estimates for variables in the norms $l^2(H^1)$ and $l^2(L^2(H^q_r))$, $q=0,1$. To enhance computational efficiency, we introduce a novel scale-decoupled solver that balances rapid convergence with reduced memory requirements. Numerical experiments using realistic battery parameters validate the theoretical error rates and highlight the superior performance of the proposed solver compared to existing algorithms.