基于IGA-BP神经网络的PEMFC供氢系统模型预测控制算法
MPC algorithm for PEMFC hydrogen supply system based on IGA-BP neural network
李岱泽 1熊树生 1姜琦 1吴占宽 1焦志筱 1程俊杰 1宋雅楠2
作者信息
- 1. 浙江大学,浙江 杭州 310012
- 2. 福州海卓马克机电设备有限公司,福建 福州 310100
- 折叠
摘要
对化石能源的过度使用导致了严重的环境问题和能源担忧.氢能作为一种清洁的能源,被认为是实现能源转型和可持续发展的重要资源.在此背景下,氢燃料电池作为一种将氢能高效转化为电能的技术,展现出了巨大潜力.本文以质子交换膜燃料电池阳极供氢系统为研究对象,以氢气计量比和阴、阳极压强差为控制目标,设计了基于神经网络的模型预测控制算法.首先基于MATLAB/Simulink搭建了面向控制的燃料电池集总参数机理模型,通过实验验证了模型的可靠性;然后通过免疫遗传算法优化神经网络的学习过程,实现了对燃料电池系统状态的精确拟合与预测;最后,将离线训练的神经网络应用于模型预测控制器,并验证了控制算法的有效性.
Abstract
In recent years,the excessive use of fossil fuels has led to serious environmental issues and energy concerns.Hy-drogen energy,as a clean source,is considered an important resource for achieving energy transformation and sustainable de-velopment.Against this backdrop,hydrogen fuel cells,as a technology that efficiently converts hydrogen energy into electrical energy,have shown great potential.This paper focuses on the anode hydrogen supply system of proton exchange membrane fuel cells(PEMFC)and designs a model predictive control(MPC)algorithm based on neural network with the hydrogen stoichiometric ratio and the pressure difference between the anode and cathode as the control targets.Firstly,a control-orien-ted PEMFC lumped parameter mechanism model is constructed by using MATLAB/Simulink,and its reliability is verified through experiment.Secondly,the learning process of the neural network is optimized by using the immune genetic algorithm(IGA)to achieve precise fitting and prediction of the fuel cell system state.Finally,the offline-trained neural network is ap-plied to the model predictive controller,and the effectiveness of the control algorithm is validated.
关键词
质子交换膜燃料电池/供氢系统/神经网络/免疫遗传算法/模型预测控制Key words
PEMFC/hydrogen supply system/neural network/immune genetic algorithm/MPC引用本文复制引用
基金项目
"十四五"国家重点研发计划"新能源汽车"重点专项项目(2022YFB2502403)
浙江省"尖兵""领雁"研发攻关计划项目(2023C01239)
出版年
2024
NETL
NSTL
万方数据