Abstract: The system of PEMFC stack is too complicated to model, and the parameters of the existing mathematical model are very different because the construction and materials used in the stack are different.There is no general model with high accuracy to describe all stacks and meet the need for designing the PEMFC control system. In this paper, the knowledge of heat transport and liquid kinetics is used to establish a stack model.Some parameters that have great influence on the accuracy of the model cannot be obtained from a simple experiment.Improved genetic algorithm proposed in the paper possesses excellent global and local searching capacity and is able to identify the parameters accurately.The results of simulation and experiment validate the practicability and reliability of the algorithm.

Key words: 质子交换膜燃料电池, 简单遗传算法, 改进遗传算法

摘要： 针对质子交换膜燃料电池（PEMFC）系统过于复杂、难以建模，而已建立的数学模型由于电堆使用的材料、结构不同，模型参数相差甚远，无法统一给出一个具有代表性的高精度的PEMFC电堆模型，难以满足PEMFC控制系统设计和应用的要求.本文从实际应用角度出发，利用传热学、流体动力学知识建立了一个千瓦级的PEMFC电堆温度模型.由于模型中一些关键参数无法由实验精确测出，但它们对模型精度的影响极大，所以需要找出一个能够高精度辨识模型参数的方法来解决这个问题.文章中提出的改进遗传算法具有良好的全局和局部的搜索、优化能力，能够高精度地辨识这些参数.仿真和实验结果验证了模型的实用性和可靠性.

关键词: 质子交换膜燃料电池, 简单遗传算法, 改进遗传算法