摘要
质子交换膜燃料电池膜电极组件表面的温度分布会影响质子交换膜燃料电池的性能、寿命和可靠性。为探究质子交换膜燃料电池传热规律,本文提出了一种基于神经网络的质子交换膜燃料电池膜电极组件温度分布的预测模型。本研究选取径向基函数神经网络(RBF)和广义回归神经网络(GRNN)两种神经网络,以电流密度、温度点的位置作为网络输入,不同位置的温度作为网络输出,对平行流道质子交换膜燃料电池、蛇形流道质子交换膜燃料电池分别建立了神经网络预测模型。结果显示,RBF神经网络预测的均方根误差平均为0.464、平均绝对百分误差为1.179%,GRNN神经网络预测的均方根误差平均为0.7155、平均绝对百分误差为2.27%;相较于GRNN神经网络,RBF神经网络精度更高;基于RBF神经网络的平行流道质子交换膜燃料电池膜电极组件温度分布预测模型预测值与96%的实验值的相对误差在5%以内。基于RBF神经网络的蛇形流道质子交换膜燃料电池膜电极组件温度分布预测模型预测值与95%的实验值的相对误差在5%以内。
The temperature distribution on the surface of the membrane electrode assembly of the proton exchange membrane fuel cell will affect the performance,life and reliability of proton exchange membrane fuel cell.In order to investigate the heat transfer of proton exchange membrane fuel cell,a prediction model of the temperature distribution of the membrane electrode assembly based on neural network is proposed.This study selected the radial basis function(RBF)neural network and generalized regression neural network(GRNN),two kinds of neural network,with the location of the current density and temperature point as network input,the different position as network output,the temperature of the parallel port of proton exchange membrane fuel cell,serpentine flow proton exchange membrane fuel cell neural network prediction model is established,respectively.The results showcase the average root mean square error of RBF neural network prediction is 0.464,the average absolute percentage error is 1.179%,the average root mean square error of GRNN neural network prediction is 0.7155,the average absolute percentage error is 2.27%.Compared with GRNN neural network,RBF neural network has higher accuracy.The relative error between the predicted value and the experimental value of 96%for the temperature distribution prediction model based on RBF neural network is within 5%.The relative error between the predicted value and the 95%experimental value of the temperature distribution prediction model based on RBF neural network for the membrane electrode assembly of PEMFC is less than 5%.
作者
李志浩
彭浩
陈亚琴
LI Zhihao;PENG Hao;CHEN Yaqin(Engineering Research Center of Shipping Simulation(Ministry of Education),Shanghai Maritime University,Shanghai 201306,China)
出处
《储能科学与技术》
CAS
CSCD
北大核心
2021年第6期2053-2059,共7页
Energy Storage Science and Technology
基金
上海市自然科学基金项目(19ZR1422300)。
关键词
质子交换膜燃料电池
膜电极温度分布
人工神经网络
proton exchange membrane fuel cell
membrane electrode temperature distribution
artificial neural network
