Open Collections

An analysis of lithium-ion battery state-of-health through physical experiments and mathematical modelling

University of British Columbia

Lithium-ion batteries are ubiquitous in modern society. The high power and energy density of lithium-ion batteries compared to other forms of electrochemical energy storage make them very popular in a wide range of applications, most notably electric vehicles (EVs) and portable devices such as mobile phones and laptop computers. However, despite the numerous advantages of lithium-ion batteries over other forms of energy sources, their performance and durability still suffer from aging and degradation. The purpose of the work presented in this thesis is to investigate how different load cycle properties affect the cycle life and aging processes of lithium-ion cells. To do so, two approaches are taken: physical experiments and mathematical modeling. In the first approach, the cycle life of commercial lithium-ion cells of LiNiCoAlO₂ chemistry was tested using three different current rates to simulate low-, medium-, and high-power consuming applications. The batteries are discharged/charged repeatedly under the three conditions, all while temperature, voltage, current, and capacity are recorded. Data arising from the experiments are then analyzed, with the goal of quantifying battery degradation based on capacity fade and voltage drop. The results are then used to build two predictive models to estimate lithium-ion battery state-of-health (SoH): the decreasing battery V₀₊ model and the increasing CV charge capacity model. Furthermore, a simple thermal model fitted from the battery temperature profiles is able to predict peak temperature under different working conditions, which may be the solution to temperature sensitive applications such as cellphones. The limitation to physical experiments is that they can be costly and extremely time-consuming. On the other hand, mathematical modeling and simulation can provide insight, such as the internal states of the battery, that is either impractical or impossible to find using physical experiments. Examples include lithium-ion intercalation and diffusion in electrodes and electrolytes, various side-reactions, double-layer effects, and lithium concentration variations across the electrode layer. Thus, in the second approach, work focuses on implementing the pseudo-two-dimensional (P2D) model, the most widely accepted electrochemical model on lithium-ion batteries. The P2D model comprises highly-nonlinear, tightly-coupled partial differential equations that calculate lithium concentration, ionic flux, battery temperature and potential to significant accuracy. The unparalleled prediction abilities of the P2D model, however, are shadowed by the low computational efficiency. Thus, much of this work focuses on reducing model complexity to shorten effective simulation time, allowing for use in applications, such as a battery management system, that have limited computational resources. In the end, four model reductions have been identified and successfully implemented, with each one achieving a certain standard of accuracy.

Text

2018-08-23

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/66917

Mathematics

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/