Physics-based and data-driven modeling of multi-active material electrode batteries

As the design of single-component battery electrodes has matured, the battery industry has turned to hybrid electrodes with blends of two or more active materials to enhance battery performance. Leveraging best properties of each material, these multi-active material electrodes open a new and complex design space that could be more efficiently explored through physics-based simulations, leading to a growing demand for improved battery simulation frameworks capable of accounting for parallel reactions and diffusion pathways, phase transformations, multi-scale heterogeneities, and interactions between individual active materials. However, existing open-source battery simulation frameworks based on porous electrode theory are tailored for single-component electrodes and cannot meet these requirements without extensive modification by users. In this thesis, I first introduce the implementation of Hybrid Multiphase Porous Electrode Theory (Hybrid-MPET), a newly developed open-source and modular simulation software for batteries with multi-active material electrodes. Building upon the theoretical foundations of the reaction kinetics, transport phenomenon, thermodynamics, and electrochemistry of lithium-metal and lithium-ion batteries, Hybrid-MPET models are capable of accurately simulating solid solution, conversion and multiphase active materials in the hybrid porous electrode at intra-particle and inter-particle scales. To demonstrate the practicality of this new framework, I next present a series of Hybrid-MPET models for different commercial battery applications with multi-active material electrodes to predict battery performance and explain experimental phenomena. I primarily focus on validating physics-based models against testing datasets of multi-active material electrode batteries powering Medtronic's implantable cardioverter-defibrillators (ICDs). I present a many-particle Hybrid-MPET model for medium-rate Li/silver vanadium oxide (SVO) battery that predicts the impact of reaction heterogeneity across particle populations on previously unexplained cell voltage transient behavior. As a natural extension, I then develop a Hybrid-MPET model for high-rate Li/carbon monofluoride (CFx)-SVO battery that accurately predicts cell voltage under low-rate background monitoring, high-rate defibrillation pulsing, and post-pulse relaxation by addressing solid-phase diffusion limitations of Li⁺ in SVO. In addition to the rate dependence of active material utilization, my insights are centered around active material interaction in the form of Li⁺ redistribution: I discuss its effect in multi-active material electrodes batteries under both near open-circuit equilibrium and far-from-equilibrium conditions, as well as its broader impact on cell operation protocols and design principles. Lastly, I explore integration of physics-based and machine learning models to more accurately predict pulsing performance of individual Li/ CFₓ-SVO cells and capture cell-by-cell performance variance. While the physics-based Hybrid-MPET model still predicts nominal cell performance, the machine learning models with additive or ensemble learning nature can account for extra information from battery production and early testing data responsible for the observed variability in the battery pulse voltages. The interpretability of the selected machine learning models also can offer insights to support future battery production as well as clues on how to refine physics-based models. It is hoped that the simulation framework, models, and insights presented in this thesis would not only assist medical battery designs, but also accelerate the development of future multi-active material electrodes battery applications.

Name thesis:degree_name

Doctoral

Department dc:contributor.department

Massachusetts Institute of Technology. Center for Computational Science and Engineering

Grantor dc:publisher

Massachusetts Institute of Technology

Year dc:date.issued

2024

Author dc:creator

• Liang, Qiaohao

Advisor dc:contributor.advisor

• Bazant, Martin Z.

Statement dc:rights

• In Copyright - Educational Use Permitted
• Copyright retained by author(s)

Licence dc:rights.uri

• https://rightsstatements.org/page/InC-EDU/1.0/

Handle dc:identifier.uri

https://hdl.handle.net/1721.1/155373

OAI identifier oai:identifier

oai:dspace.mit.edu:1721.1/155373