Abstract The aim of this study is to design a thermal management system to address the temperature regulation and fluctuation of an open-cathode proton exchange membrane (PEM) fuel cell. A model predictive control (MPC) approach is newly proposed to command of blowers in the fuel cell. First, a thermal management-oriented model of the fuel cell is set up, and linearized by using Taylor formula. Then, the closed-loop feedback MPC according to the linearized state-space model is developed. An experimental test rig via LabVIEW-based thermal control system prototype is configured. Finally, different load current tests were applied to verify the effectiveness, and the performance of the MPC controller was compared with a traditional PI controller. The results of this study demonstrate that the proposed MPC can effectively manipulate the stack temperature to track the reference trajectory under the constant current as well as dynamic load schedules. In particular, in a combined driving cycle equivalent test condition, the fuel cell temperature reached to the target after a settling time of 146.6 s, and temperature fluctuations were less than ±0.5 °C with the MAE value of 0.223 and the RMSE value of 0.346.

中文翻译：

---

一种开式阴极燃料电池热管理系统模型预测控制的设计与实现

摘要 本研究的目的是设计一种热管理系统，以解决开式阴极质子交换膜 (PEM) 燃料电池的温度调节和波动问题。新提出了一种模型预测控制 (MPC) 方法来控制燃料电池中的鼓风机。首先，建立了燃料电池的热管理导向模型，并利用泰勒公式对其进行了线性化。然后，根据线性化状态空间模型开发了闭环反馈 MPC。配置了基于LabVIEW的热控制系统原型的实验测试台。最后，应用不同的负载电流测试来验证有效性，并将 MPC 控制器的性能与传统的 PI 控制器进行比较。这项研究的结果表明，所提出的 MPC 可以有效地操纵堆栈温度以跟踪恒定电流和动态负载计划下的参考轨迹。特别是在联合驱动循环等效试验工况下，燃料电池温度在146.6 s的稳定时间后达到目标，温度波动小于±0.5°C，MAE值为0.223，RMSE值为0.346 .