Development and Parameterization of a Physics-Based Ageing Model for Li-Ion Battery Cells

Although significant progress has been made on developing electrochemical models of Li-ion batteries performance, there is a significant gap in predictive, physics-based modelling of the degradation mechanisms. In this work, we perform a systematic experimental and modelling study to explore the potential of predictive battery ageing models. A commercial NMC pouch cell is initially characterized in detail using tear-down analysis, electrical and electrothermal tests to obtain electrochemical model parameters and validate its fidelity in a large range of operating conditions in terms of temperature, state-of-charge and load. The cell is then exposed to accelerated ageing operating conditions and its performance is monitored regularly to obtain its degradation rate in terms of capacity and resistance. The aged cell is also characterized by tear-down and optical techniques. The experimentally obtained test database is used to develop and validate the mathematical models that describe the ageing processes, using a fully physics-based approach implemented in the commercial software GT/AutoLion®. The ageing is attributed to side, competitive electrochemical reactions characterized by their own reaction mechanisms and kinetic parameters. The latter are obtained iteratively by fitting the model results to experimental data. We introduce a model that accounts for solid-electrolyte interphase (SEI) growth formation kinetics, reversible and irreversible lithium plating-stripping dynamics, and loss of active material (LAM) evolution through interdependent rate equations. The test protocols are specifically engineered to isolate and quantify the individual contributions of the degradation mechanisms. By deconvoluting the effects of SEI formation, lithium plating/stripping, and LAM, the methodology enables the extraction of distinct, mechanism-and electrode specific parameter sets. This work is contributing to the development of a comprehensive framework for systematically parameterizing electrochemical battery aging models. Such a framework could enable high-fidelity predictions of battery behavior under real-world conditions and accelerate the optimization of battery management system (BMS) functionalities.