Clemson University
Optimal Management of Electric Vehicle Fast Charging Stations via Lyapunov Optimization and Reinforcement Learning
Abstract
dc:description.abstractThe transportation sector is responsible for 28% of global energy-related emissions, with road transport accounting for most of these emissions. While electric vehicles (EVs) are essential for decarbonizing transportation, widespread adoption depends on fast, efficient charging. Fast charging reduces charging time but accelerates battery degradation, whereas slow charging increases range anxiety. In addition, fast charging stations (FCSs) impose significant stress on distribution grids, making their operation a critical challenge. Optimizing FCS operation can conflict with EV fast-charging objectives, yet existing research largely treats these problems independently despite their strong interdependence. Current fast-charging methods focus on individual cells or small battery modules, neglecting battery pack behavior and FCS constraints. Likewise, FCS management relies on stochastic optimization over fixed planning horizons, requiring unrealistic forecasts of uncertain user behavior. This dissertation addresses these limitations through a unified framework based on Lyapunov optimization (LO), which decomposes large optimization problems into real-time per-time-step decisions using only current information. This approach naturally accommodates the different time scales of EV charging and FCS operation, eliminates the need for demand forecasting, reduces computational complexity, and provides provable near-optimal performance. The dissertation reformulates EV fast charging, FCS management, and electricity market participation within the LO framework and establishes convergence guarantees for each formulation. It also develops a reinforcement learning (RL)-based battery pack fast-charging strategy and introduces an RL-augmented LO framework that combines RL's ability to capture complex nonlinear battery dynamics with LO's constraint satisfaction and optimality guarantees, enabling practical, real-time optimization of EV charging, FCS operation, and market participation.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy (PhD)
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Automotive Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Abbasi, Mohammad Hossein
- Contributors dc:contributor
-
- Jiangfeng Zhang
- Venkat Krovi
- Benjamin Lawler
- Zheyu Zhang
Subjects
dc:subject × 2Identifiers
dc:identifier.*- Repository record dc:identifier
- https://open.clemson.edu/all_dissertations/4352