直接甲醇燃料电池的数学建模与神经网络辨识建模

针对直接甲醇燃料电池(DMFC)非线性系统建模问题,提出了两种不同的建模方法(1)采用电化学、流体动力学、热力学等理论,建立了DMFC电池性能数学模型;(2)利用改进型BP神经网络建立DMFC电池性能辨识模型.结合DMFC实验数据进行仿真测试,结果表明这两种建模方法均合理、有效,建立的模型精度较高,从而为设计DMFC在线控制器奠定了基础.

Direct Methanol Fuel Cell Modeling Based on Mathematical Theory and Neural Networks Identification

Two different modeling methods were proposed for solving the modeling problem of direct methanol fuel cell(DMFC) nonlinear system. One is a mathematical model of DMFC performance, which is established by using theories of electrochemistry, hydrokinetics and thermodynamics. The other is an identification model of DMFC performance, which is established by the novel BP algorithm of neural networks. The simulation results show that the two modeling methods are correct, effective and the models have better accuracy. Moreover, the mathematical and neural networks modeling make it possible to design on-line controller of DMFC.