基于BP神经网络的PEMFC稳定性及动态性预测

Steady and dynamic prediction of PEMFC based on BP neural network

研究了PBI/H_3PO_4体系高温质子交换膜燃料电池(PEMFC),将其在不同压力、温度、阴极气体和负载下的稳态电位响应数据及该体系对压力变化、温度变化的动态电位响应数据作为训练数据,建立以Matlab/Simulink和BP神经网络为基础的高温下PEMFC的稳定性能和动态性能的预测系统。通过所建立的神经网络模型对电池的稳态电位输出和动态电位响应进行模拟,结果表明,所建立的模型可以对电池的稳态及动态行为进行准确模拟,这为PBI/H_3PO_4体系高温PEMFC的控制及性能预测提供了一定的参考。

High temperature proton exchange membrane fuel cell （PEMFC） with PBI/H3PO4 system was studied. The data of the steady potential response under different pressure, temperature, cathode gas and load and the dynamic potential response of the system under different pressure and temperature were studied. A prediction system for the stability and dynamic performance of PEMFC at high temperature was established based on Matlab/Simulink and BP neural network. Through the established neural network model, the steady state potential output and the dynamic potential response of the battery were simulated. The results show that the model can accurately simulate the steady and dynamic state behavior of the battery. This provides some reference for the control and performance prediction of high temperature PEMFC with PBI/H3PO4 system.