PEMFC堆的正交实验分析与建模

Analysis and modeling of PEMFC stack based on orthogonal experiment

采用正交实验法研究操作参数对PEMFC电堆性能的影响.实验结果表明:电堆性能受工作温度、气体压力和空气过量系数影响较大,而氢气过量系数以及阴、阳极湿度在60%～100%时对性能的影响较小.神经元网络可以避开电堆内部的复杂关系,在正交实验基础上能实现快速建模.采用改进型BP算法对多参数影响下的电堆性能进行建模,仿真结果表明:方法可行,并且模型具有较好的精度和泛化能力.

The effect of the operation parameters on the performance of the PEMFC stack was studied by the orthogonal experiment method. The experiment results showed the working temperature, gas pressure and air excess ratio affected the stack performance remarkably. The influences of the hydrogen excess ratio, anode and cathode humidity in the range of 60% to 100% were less. As the neural network modeling could avoid the complicated stack inner mechanisms, the rapid modeling of stack was possible. By means of the improved BP algorithm, the performance of stack under the effect of different parameters was modeled. The simulation results showed the model had high accuracy and good generalization ability.