An graduated from Beihang University with a Bachelor's degree in Aircraft Propulsion Engineering in 2019. Following the completion of undergraduate studies, An embarked on a career as a technical researcher at a reputable company. An pursued a Master's degree at Beihang University, focusing on the cooling and heat exchange systems within aircraft engines.
An employed entropy analysis to compute and fit multi-modal bleed air structure equations. Proficient in the analysis and modelling of thermophysical phenomena, An adeptly constructed systems of partial differential equations to calculate fitted formulas under specified boundary conditions. An possesses a strong passion for programming and aspires to create a personal gaming project in their spare time.
An's enthusiasm for mathematical analysis and programming led her to join the AAPS CDT, where she aims to establish a comprehensive lithium-ion degradation model. This endeavour aligns with An's aspiration to leave a profound mark in advancing the electrified future.
Lithium-ion batteries are deployed in a wide range of applications due to their low and falling costs, high energy densities and long lifetimes. Accurate prediction of lifetime would unlock new opportunities in battery use and optimization. An's research will apply both physics-based electrochemical-thermal battery model and data-driven neural network model to predict battery degradation.
The physics-based model is based on our in-house battery model, LionPower [Yuan et al., Int. J. Heat Mass Transf., 2021], which is developed for accurate and efficient modelling of electrified propulsion and has advantages over existing models including LIONSIMBA, GT-AutoLion, and PyBaMM in terms of accuracy and computational efficiency. Theoretical degradation models involving the loss of lithium-ions and other active materials will be integrated into LionPower to predict battery degradation represented by State of Health (SOH) and Remaining Useful Life (RUL). Meanwhile, advanced discretisation schemes and efficient numerical solvers will be investigated to further improve the performance of the physics-based battery degradation model.
The data-driven neural network model relies on published battery degradation datasets in the literature, which basically uses non-linear functions to emulate the physical processes of battery degradation. Novel structures of neural networks will be proposed to correlate charging parameters and/or curves with SOH and RUL. Besides, neural networks will be applied to predict critical properties, e.g., SEI thickness, electrode particle cracking, lithium plating, etc., during battery degradation to obviate the need for solving coupled PDEs and thus improve computational efficiency, and the prediction can be fed into the physics-based model for accurate estimation of SOH and RUL.
The output of this research will be a comprehensive battery degradation model that can be directly applied to electrified propulsion systems.
© Copyright 2024 AAPS CDT, Centre for Doctoral Training in Advanced Automotive Propulsion Systems at the University of Bath
