@article{Wang2025, 
author = {Jiafeng Wang and Xiangguo Xu},
title = {Dynamic Modeling and Model Predictive Control Study of Direct Cooling Thermal Management System for Batteries},
year = {2025},
journal = {Journal of Refrigeration},
volume = {46},
number = {6},
pages = {23-33},
keywords = {battery thermal management, direct cooling system, refrigeration system modeling, model predictive control},
url = {https://www.sciopen.com/article/10.12465/j.issn.0253-4339.2025.06.023},
doi = {10.12465/j.issn.0253-4339.2025.06.023},
abstract = {Battery thermal-management systems based on direct refrigerant cooling are characterized by multiple sources of disturbance, strong coupling, and nonlinearity, making regulation of the temperature difficult through traditional proportional-integral-derivative (PID) controllers. In this study, a dynamic step-down model of a direct cooling thermal-management system was first established based on the first principle, and its linearized model was verified using experimental data. Based on this model, a linear time-varying model-based predictive control strategy was designed and compared with PID control under two operating conditions: battery cooling and the worldwide harmonized light vehicles test cycle (WLTC). Under the condition of constant battery heat generation, the stabilization time for the battery temperature from 50 ℃ to 30 ℃ was 262 s for the model predictive control (MPC) strategy and 952 s for the PID control strategy. At the same time, the MPC reduced the energy consumption by 5.57%. Under the WLTC condition, the temperature fluctuation of the PID control was large, with a maximum deviation of 2.9 ℃, while the MPC was able to respond quickly to a heat load change and stabilize the control temperature at 30 ℃. The standard temperature deviations of the PID and MPC control strategies were 1.7 ℃ and 0.06 ℃, respectively. In summary, the MPC strategy was superior to the PID control in terms of temperature regulation speed, energy efficiency, and robustness.}
}