Application of A Derivative-Free Data-Driven Optimization Algorithm on Li-Ion Battery Parameter Identification

Accurate electrochemical models are crucial for the design, control, and state estimation of lithium-ion batteries, yet parameterizing these models, such as the Single Particle Model with electrolyte and thermal effects (SPMeT), remains a significant challenge. This paper presents a comprehensive parameter identification methodology for a SPMeT model applied to a commercial lithium-ion cell. The study focuses on identifying five key electrochemical parameters: effective electrode area <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(A_{eff})$</tex>, solid-phase active material volume fractions in the anode <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\epsilon_{s,n})$</tex> and cathode <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\epsilon_{s,p})$</tex>, and initial solid-phase diffusion coefficients for the anode <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(D_{s,n0})$</tex> and cathode <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(D_{s,p0})$</tex>. Critically, the Open Circuit Potential (OCP) curves for both the anode and cathode are also characterized as part of the identification process. The Bayesian Adaptive Direct Search (BADS) algorithm is employed to optimize these parameters by minimizing the error between simulated and experimental data. The efficacy of the proposed approach is demonstrated through rigorous validation against a dynamic Highway Fuel Economy Test procedure, where the parameterized model achieved a low Root Mean Square Errorof only 19.8mV in predicting the cell voltage. This work establishes a robust framework for developing high-fidelity lithium-ion battery models capable of accurate performance prediction under realistic operating conditions.