Air-Cooled Battery Pack Thermal Management via Control Barrier Proximal Dynamics

Efficient thermal management in lightweight electric mobility systems is critical for battery safety, performance, and longevity. However, challenges remain due to cost and weight constraints and limited onboard computational resources. In battery packs composed of hundreds of cells, lumped-parameter models fail to capture spatial temperature variations, while high-fidelity models render real-time, optimization-based control computationally intractable. This work addresses these challenges through: 1) a modular, conservation-based electrothermal graph model that balances fidelity and tractability; 2) a computationally scalable control strategy, called control barrier proximal dynamics (CBPD), which closely approximates safe and optimal control barrier function (CBF)-based controllers with major improvements in computational efficiency. We further extend CBPD to ensure safety under input disturbances and modeling uncertainties via an input-to-state safe (ISSf) and robust formulation. As a candidate system, we consider a fan-cooled lithium-ion battery pack in the University of Minnesota’s solar electric vehicle (EV), comprising 1296 cells. We validate the thermal model on this system and apply the CBPD controller on the model under realistic operational scenarios. The results demonstrate strict safety constraint satisfaction, performance comparable to intensive optimization-based methods, computational efficiency on par with classic controllers, and robustness to parameter variations.