.png){kind=icon}
.png){kind=icon}
Return
Model-based battery thermal parameter optimization using symbolic regression
.png){kind=icon}
PRE
AI
DOI：10.1016/j.est.2023.109243.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Lithium-ion battery packs are commonly used in electromobility and distributed energy storage. However, keeping both cell temperature and inter-cell temperature differences within operating ranges is important to avoid a negative impact on performance and lifespan. For this reason, modeling the thermal behavior of battery packs in different operating conditions is crucial for both design and practical use. In this work, we optimize the parameters of a battery thermal model using a symbolic regression method based on evolutionary algorithms. We obtain power-type expressions for the drag coefficient, friction factor, and heat transfer correlation (Nusselt number), as a function of the Reynolds number and the cell spacing factor, plus the Prandtl number in the case of the Nusselt number. The model is adjusted with CFD simulations of battery modules containing up to 102 cells to obtain the steady-state temperature distribution. Compared to CFD simulations, the proposed model gets a mean absolute percentage error of 2.39 % in estimating the temperature in a 53-cell battery pack. The adjusted parametric thermal model is validated with experimental data for cooling down a battery pack and a single cell, obtaining a mean absolute percentage error of 1.26 % and 4.68 %, respectively, in estimating the dynamic temperature balance. The adjusted parametric thermal model and the mathematical expressions obtained for the drag coefficient, friction factor, and Nusselt number, can be useful for designing battery thermal management systems in electromobility and distributed energy storage.
Keywords:
Lithium-ion battery cooling
Battery thermal model
Symbolic regression
Journal
IF:
9.8
Papers:
2.2W
Citations:
10.1W