基于模糊神经网络PID的质子交换膜燃料电池温度控制策略

舒晨，王桂华^*，白书战，朱思鹏

山东大学核科学与能源动力学院，山东 济南  250061

摘要：为解决质子交换膜燃料电池（proton exchange membrane fuel cell，PEMFC）电堆在动态负载下易出现温度波动、区域温差偏离与控制鲁棒性不足等问题，建立燃料电池电堆–冷却回路联合仿真模型，基于Simulink构建以燃料电池冷却液进、出口温差为核心的闭环控制框架，设计模糊比例微分积分（proportional integral derivative，PID）及模糊神经网络PID控制器；采用阶跃电流工况，对比分析传统PID、模糊PID、模糊神经网络PID 3种控制策略下，燃料电池出、入口水温及冷却液流量的动态响应特性。仿真结果表明，相比传统PID、模糊PID，采用模糊神经网络PID控制策略时，电堆出、入口水温与温差的控制效果和精度均有显著提高，阶跃电流工况下能够显著减小温度响应超调量、缩短调节时间并保持更好的鲁棒性，从而有效实现燃料电池热管理系统的高性能控制。

关键词：PEMFC；热管理；模糊控制；神经网络

The temperature control strategy for proton exchange membrane fuel cell based on fuzzy neural network PID

SHU Chen, WANG Guihua^*, BAI Shuzhan, ZHU Sipeng

School of Nuclear Science, Energy and Power Engineering, Shandong University, Jinan 250061, China

Abstract: Aiming at the problems of obvious temperature fluctuation, regional temperature deviation and insufficient control robustness of proton exchange membrane fuel cell (PEMFC) stack under dynamic load conditions, a joint simulation model of fuel cell stack and cooling circuit is established. A closed-loop control framework taking the temperature difference between inlet and outlet cooling liquid as the core is constructed on the Simulink platform, and fuzzy proportional integral derivative（PID）and fuzzy neural network PID controllers are designed. Under step current operating conditions, the dynamic response characteristics of inlet and outlet water temperature as well as coolant flow rate are comparatively analyzed under three control strategies, namely traditional PID, fuzzy PID and fuzzy neural network PID. The simulation results show that, compared with traditional PID and fuzzy PID, the fuzzy neural network PID strategy significantly improves the control accuracy and comprehensive performance of stack inlet-outlet water temperature and temperature difference. Under step current conditions, it effectively reduces the temperature overshoot, shortens the adjustment time and maintains superior robustness, which realizes high-performance control of the fuel cell thermal management system.

Keywords: proton exchange membrane fuel cell; thermal management; fuzzy control; neural network