质子交换膜燃料电池(Proton Exchange Membrane Fuel Cell，简称 PEMFC)是一种清洁能源，随着环境问题和能源问题的突出，PEMFC有望成 为解决问题的关键，针对现有质子交换膜燃料电池模型自适应性差的问题，本 文对PEMFC的建模进行了研究，介绍了PEMFC的面向系统设计分析的模型、 面向控制器设计模型和面向系统辨识的模型。建立了一种动态神经网络的 PEMFC模型。该方法基于3层前馈神经网络，根据隐含层神经元的敏感度来 确定隐含层神经元数量的增减，实现神经网络适应不同数据的需求。采用L M 训练法对神经网络中各神经元的权重进行调整，多次重复训练，对结果取 平均值，提高了神经网络的精度。

Proton Exchange Membrane Fuel Cell (PEMFC) is a kind of clean energy. As environmental and energy problems become more prominent, PEMFC is expected to become the key to solving the problem. It is adaptive to the existing proton exchange membrane fuel cell model. To solve the problem of poor performance, this paper studies the modeling of PEMFC and introduces the system design analysis-oriented model, controller design model and system identification oriented model of PEMFC. A PEMFC model of dynamic neural network was established. This method is based on a 3-layer feedforward neural network, and determines the increase or decrease in the number of hidden layer neurons based on the sensitivity of the hidden layer neurons, allowing the neural network to adapt to the needs of different data. The L-M training method is used to adjust the weight of each neuron in the neural network, repeat the training multiple times, and average the results to improve the accuracy of the neural network.