Health-conscious fast charging of Li-ion batteries via an electro-chemical battery aging model
H.A.M.R. Sewailem (TU Delft - Mechanical Engineering)
A.J.J. van den Boom – Mentor (TU Delft - Mechanical Engineering)
Robinson Medina Sanchez – Mentor
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
The transition towards a more sustainable future has become increasingly necessary due to rising greenhouse gas emissions. With the growing size of modern cities, transport has become a major contributor, with road transport accounting for up to 70% of total emissions [3]. As a result, Electric Vehicles (EVs) are increasingly adopted as a replacement for Internal Combustion Engine Vehicles (ICEVs), which are the main source of emissions in road transport.
However, EVs still face two main bottlenecks: significantly slower charging times compared to refuelling ICEVs, and Li-ion battery degradation over time, which affects lifespan. Therefore, a charging strategy is required that mitigates these effects. This leads to the research question: finding a model-based real-time control charging strategy that reduces charging time and degradation.
To develop such a strategy, accurate battery and degradation models are required to capture internal states such as State of Charge (SOC) and degradation mechanisms such as Solid Electrolyte Interphase (SEI) growth and lithium plating. Three electrochemical Li-ion battery models are considered: the Pseudo 2-Dimensional (P2D) model, the Electrolyte Enhanced Single Particle Model (SPMe), and the Extended Single Particle Model (ESPM). The P2D model is a full-order model with high accuracy but high computational cost. The SPMe is a simplified version with reduced accuracy but lower computational cost. Both models are unsuitable for real-time control. The ESPM is developed as a further simplification to achieve real-time feasibility while maintaining accuracy.
The P2D and SPMe models are implemented using PyBaMM, while the ESPM is implemented in MATLAB. Results show that the ESPM maintains over 90% similarity with reference models for currents up to 2C, making it suitable for control applications.
A degradation model including SEI growth and lithium plating is also implemented in MATLAB, calibrated to match capacity fade of the LGM50 battery. These degradation mechanisms are incorporated into the control framework to be minimised during charging.
Finally, a Nonlinear Model Predictive Control (NMPC) strategy is developed to optimise the trade-off between charging time, SEI growth, lithium plating, and current constraints. The resulting health-conscious fast charging strategy achieves a charging time of 34 minutes with an estimated battery lifespan of approximately 800 cycles. This charging time is comparable to DC fast charging standards while reducing degradation effects, saving around 100 cycles over the battery lifetime.