Nonlinear and Adaptive Model Predictive Control Methods for Battery Thermal Management System

Battery packs with larger capacity and power demand today are more likely to have overheating problem, which may further lead to thermal run-away. Active Control Strategy for battery thermal management can improve the battery safety level by larger cooling capacity. However, conventional control methods like Rule-based control and PID control have response delay problem and consume too much energy. Nonlinear Model Predictive Control (NMPC) method and Adaptive Model Predictive Control (AMPC) method are adopted here to improve the temperature tracking ability and energy efficiency. Fast models of thermal management system including battery pack are built for NMPC and successively linearized for AMPC. Several interesting conclusions are shown in this research. Firstly, the comparison between the Model-In-the-Loop (MIL) results of NMPC and AMPC shows similar control ability. However, AMPC has an obvious advantage on simulation speed, which is suitable for embedded real-time control. Also, the predicted future information was fed to AMPC controller and the result shows that, with future information, the controller can pre-response to the sudden change of the operating conditions, which indicates the promising future of AMPC on the Intelligent Connected Vehicle (ICV). At last, Hardware-In-the-Loop (HIL) test is also conducted for AMPC and PID. The results show the advantage of AMPC on both the temperature tracking ability and energy efficiency.