Electrochemistry-Based Equivalent Circuit Model via Model Approximation
Loading...
Date
2021-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
Physics-based electrochemical models and empirical equivalent circuit models (ECMs) are two well-established and widely used techniques to predict the current-voltage behavior in lithium-ion cells. While physics-based models are typically very accurate and require relatively little experimental data to calibrate, they suffer from high mathematical and computational complexity. Conversely, empirical models are more computationally efficient and mathematically simpler, making them well-suited for applications in controls, diagnosis, and state estimation of lithium-ion battery packs. However, ECMs are not predicting the physical and chemical processes occurring in the cell, hence they are less accurate and require extensive and costly experimental campaigns to properly calibrate. This research bridges the gap between these two classes of models, incorporating some of the physical insight of electrochemical models into the calibration process of empirical models. To achieve this, the Extended Single-Particle Model (ESPM), an electrochemical model, was chosen as a foundation for the work. A Padé Approximation -- a frequency-domain approximation method -- was used to simplify the solid and liquid phase diffusion equations of the ESPM. From here, the ESPM was further simplified via linearization such that the mathematical structure of the ESPM matched the one of a second order ECM, making ECM calibration straightforward. The newly calibrated ECM had minimal loss in accuracy compared to the ESPM, with only a 7% increase in RMS error across various validation test profiles. The new ECM is calibrated using simple parameter definitions, eliminating the need for the additional testing typically required for ECM calibration. Additionally, since the ECM is constructed of the physical constants of the ESPM, the calibrated ECM has physical meaning, allowing for novel methods of battery performance exploration in future work, including an extension of the work to incorporate battery aging.
Description
Keywords
battery modeling, electrochemistry, equivalent circuit model, Pade Approximation, calibration