Parameter estimation of an electrochemistry-based lithium-ion battery model

Abstract

Parameters for an electrochemistry-based Lithium-ion battery model are estimated using the homotopy optimization approach. A high-fidelity model of the battery is presented based on chemical and electrical phenomena. Equations expressing the conservation of species and charge for the solid and electrolyte phases are combined with the kinetics of the electrodes to obtain a system of differential-algebraic equations (DAEs) governing the dynamic behavior of the battery. The presence of algebraic constraints in the governing dynamic equations makes the optimization problem challenging: a simulation is performed in each iteration of the optimization procedure to evaluate the objective function, and the initial conditions must be updated to satisfy the constraints as the parameter values change. The ε -embedding method is employed to convert the original DAEs into a singularly perturbed system of ordinary differential equations , which are then used to simulate the system efficiently. The proposed numerical procedure demonstrates excellent performance in the estimation of parameters for the Lithium-ion battery model, compared to direct methods that are either unstable or incapable of converging. The obtained results and estimated parameters demonstrate the efficacy of the proposed simulation approach and homotopy optimization procedure. • Presenting a high-fidelity model of an electrochemistry Lithium-ion battery. • Introducing a simulation scheme, robust in parameter estimation process. • Applying homotopy optimization in parameter estimation of the battery model. • Satisfying algebraic constraints during the optimization process. • Updating initial conditions during optimization by the numerical algorithm.