熔融碳酸盐燃料电池热启动过程的人工神经网络建模

Artificial Neural Network Modeling of Hot Start Processes of Molten Carbonate Fuel Cell

熔融碳酸盐燃料电池(MCFC)是燃料电池研究领域的一个难点,其严格的热启动过程对电池性能和寿命的影响至关重要。针对这一问题,建立了基于人工神经网络的熔融碳酸盐燃料电池热启动过程模型,并详细给出了采用改进BP算法的熔融碳酸盐燃料电池热启动过程的模型结构、算法、训练和仿真。MATLAB仿真结果证明其快速准确,为熔融碳酸盐燃料电池热启动过程的控制提供了实际工程应用模型。

At present, the molten carbonate fuel cell is a difficult topic in the research field of fuel cells. Its hot starting process, governed by stringent requirements, has a critical effect on the cell＇s performance and its service life. Focusing on this problem, an artificial neural network based model, of MCFC＇s starting process, has been elaborated. The structure, the algorithm used, and the model＇ s way of training and simulating with improved BP algorithm, are explained in detail. MATLAB simulation results vindicate the model＇s worthiness in swiftness and accuracy which make it practically suitable for control purposes of hot starts of molten carbonate fuel cells.