Parametrisation and estimation of surrogate critical surface concentration in lithium-ion batteries

Abstract

In this paper a surrogate electrochemical lithium-ion battery model, presented and discussed in Di Domenico et al. (2008a) and Di Domenico et al. (2008b), is parametrised and validated through experimental data by a 10 cell 37 V 6.8 Ah Li-ion battery pack. Following past results (Zhang et al., 2000; Smith, 2010), the model is based on an approximate relationship between the electrode-averaged Butler-Volmer current and the solid-electrolyte interface concentration of a surrogate single particle for each cell electrode. Equally-spaced radially-discretised diffusion dynamics of the surrogate single particle for the positive electrode are then used to emulate the lithium concentration evolution in the cell. The surface concentration of the surrogate single particle, defined as surrogate Critical Surface Concentration (sCSC) is then used to predict the cell terminal voltage. The resulting model is as compact as an equivalent-circuit model but its underpinnings are lumped lithium-ion diffusion dynamics. A few parameters of the lumped lithium concentration dynamics are tuned using an iterative optimisation procedure with continuous and pulsed current profiles. The single particle lithium concentration profile and the surface concentration are then estimated using a 4th order Extended Kalman Filter (EKF) and the voltage predictions are compared with data.