An algorithmic framework for full-order physics-based simulations of electrochemical impedance spectroscopy

Area of Science:

• Electrochemistry
• Materials Science
• Computational Modeling

Background:

• Simulated electrochemical impedance spectroscopy (sEIS) is crucial for non-invasive lithium-ion battery (LiB) analysis.
• Current physics-based LiB models, like the Electrochemical-Ageing-Capacitance (EAC) model, are complex, involving coupled partial and ordinary differential equations (PDEs/ODEs).
• Efficient simulation of these models is essential for advancing sEIS applications in parameterization and degradation characterization.

Purpose of the Study:

• To develop and validate a high-fidelity, efficient MATLAB-based solver for the full-order EAC model.
• To transform the EAC model's complex PDE/ODE system into an ODE-only system for improved computational efficiency.
• To provide an open-source tool for researchers to advance sEIS simulations in LiB analysis.

Main Methods:

• Developed a custom 'ODE+iterative' solver framework in MATLAB.
• Transformed the EAC model's governing equations from coupled PDEs/ODEs to a coupled ODE-only system.
• Benchmarked the solver against state-of-the-art solvers in MATLAB and PyBaMM for accuracy and performance.

Main Results:

• The proposed solver achieved less than 1% prediction error for most EAC model variables.
• Demonstrated a 4x improvement in solving performance compared to standard MATLAB solvers.
• Achieved competitive performance against PyBaMM for sEIS impedance spectra computation.
• Successfully applied the solver in a case study for quantitative degradation characterization.

Conclusions:

• The developed MATLAB solver offers an accurate and efficient tool for high-fidelity sEIS simulations of LiBs.
• The transformation to an ODE-only system and the custom solver framework significantly enhance computational performance.
• The open-source availability of the solver facilitates further research in LiB modeling, parameterization, and degradation analysis using sEIS.