Loading...
Developing and identifying physically based Li-ion battery models to inform real-time control applications
Weddle, Peter J.
Weddle, Peter J.
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2019
Date Submitted
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
Abstract
The enclosed thesis develops and identifies physically based battery models. The physically based battery models bridge fundamental chemical, electrochemical, and transport scales to system-level applications. The fundamental physics captured in such detailed models provide qualitative and quantitative insight for effective battery design and implementation. Because fundamental models are too computationally expensive to run at scale, systematic reduction techniques are introduced to identify low-order models that can be run in real time. Many techniques are implemented throughout the thesis to inform battery/pack level models of fundamental electrochemical physics. The techniques presented not only upscale important battery dynamics, but also can be used for practical implementation in battery management systems. The thesis is generally formatted to (1) introduce and derive common approaches to modeling important battery physics at the porous media scale (i.e., transport/kinetics between current collectors), (2) develop upscaling techniques to inform full-cell battery models of underlying microscale (electrode) composition and mesoscale (current collector wrapping) architecture, (3) formulate a systematic procedure for identifying reduced, low-order, linear state-space models that accurately describe full-cell battery dynamics, (4) implement the identified state-space models into model-predictive control algorithms, and (5) introduce and analyze phase-transformation physics in such a way as to identify important electrochemical impedance realizations. This research has at least four significant contributions to the field. The first novel contribution is developing upscaling techniques that inform full-cell models of mico- and meso-scale compositions. The physically informed full-cell models indicate that the influence of heterogeneous upscaled parameters are most significant during abuse scenarios. The second novel contribution is a systematic procedure to extract linear-state space models from complex battery models. Because the extraction procedure treats the physically informed battery model as a black-box, the same procedure can be extended to real batteries. The third novel contribution are pioneering model-predictive control algorithms that implement the extracted state-space models and respect both measurable and unmeasurable constraints. The final novel contribution of the present thesis is to interpret phase-transformation physics at the electrode particle ensemble scale. The fundamental analysis indicates that phase-transformation electrodes will exhibit history-dependent state-of charge and subsequently a history-dependent electrochemical impedance spectra. Phase-transformation electrode charge history-dependent impedance is validated with experimental results.
Associated Publications
Rights
Copyright of the original work is retained by the author.