基于改进磷虾群算法优化Elman神经网络的PEMFC电堆建模

质子交换膜燃料电池(PEMFC)电堆的优化控制需要有精确的电堆模型。现有的基于Elman神经网络建立的PEMFC电堆模型已具有较好的精度,但是此种电堆模型仍然存在容易陷入局部极值、结果无法重现等问题。考虑可以将自适应莱维飞行和偏好随机游动两种机制引入基本磷虾群(Krill Herd,KH)算法,得到一种改进的磷虾群(Improved Krill Herd,IKH)算法用以优化神经网络的初始参数,进而建立基于IKH-Elman网络的PEMFC电堆模型。仿真结果表明,IKH算法用于优化神经网络可同时保证更高的寻优精度和更快的收敛速度;在预测的精度和稳定性上此种电堆模型也具有一定的优势。

PEMFC STACK MODELING BASED ON ELMAN NEURAL NETWORK OPTIMIZED BY IMPROVED KRILL HERD ALGORITHM

Optimal control of the proton exchange membrane fuel cell(PEMFC)stacks needs a precise stack model.The existing PEMFC stack model based on Elman neural network has already had good precision,but this kind of model still has some problems such as easy to fall into local extremum and the result cannot be reproduced.Considering that the adaptive Levy flight and the biased random walk can be introduced into the basic krill herd(KH)algorithm to obtain an improved krill herd(IKH)algorithm to optimize the initial parameters of the neural network,and then,a PEMFC stack model based on IKH-Elman network can be established.The simulation results show that IKH algorithm can ensure better searching precision and faster convergence speed when used to optimize neural network.In terms of accuracy and stability of prediction,this kind of stack model also has certain advantages.