Battery control via adaptive modeling and predictive control
There will be a tremendous shift in road transport from fossil to electric, in which the availability of capable batteries is a major issue. In the absence of physical limits, existing battery systems are inappropriately designed and prematurely aged.
This project aims to conduct fundamental and multidisciplinary research and significantly advance battery management technology for safe and optimal use. To enable online manipulation of internal battery electrochemical states, this project will first develop a novel modelling framework that incorporates physically meaningful circuit models for a range of battery applications. Theoretical and practical contributions will be made in model simplification, identification and adaptivity. Then, a robust multi-timescale observer will be designed, that for the first time synthesizes model-based and machine learning algorithms, and to estimate unmeasurable yet crucial battery states in different timescales.
Finally, based on the battery models and estimator, a hierarchical optimal control methodology will be proposed that optimally trades off different objectives, systematically handles constraints on input and states and ensures real-time implementability. As a result, the battery lifetime will be increased, the charging will be faster, and the energy and resource efficiency will be improved. This will make electric vehicles more attractive to potential drivers and significantly contribute to a more sustainable vehicle fleet.
Changfu Zou (contact)
Assistant Professor at Chalmers, Electrical Engineering, Systems and control, Automatic Control
Professor at Chalmers, Electrical Engineering, Systems and control, Automatic Control
Swedish Research Council (VR)
Funding Chalmers participation during 2020–2023
Related Areas of Advance and Infrastructure