Abstract: The power operation of the solid oxide fuel cell needs stack cell as well as many other auxiliary components in order to obtain the maximum efficiency of system output. To make the SOFC stack have better applicability for the fuels other than the pure hydrogen，two important auxiliary components are added - the fuel direct internal reforming device and combustion chamber. The nonlinear model predictive control strategy is based on system modeling. This can make the output of gas composition，temperature，pressure，concentration and flow rate more effective and satisfy the needs of the normal operation of fuel cell stack. It discusses the effectiveness and applicability of two kinds of different control scheme of the LMPC and NMPC，respectively.

Key words: fuel cell, solid oxide fuel cell, internal reformer, linear model predictive control, non-linear model predictive control

摘要： 固体氧化物燃料电池（SOFC）发电系统运行除了电堆本体外还需要包含诸多其他辅助组件以期获得系统输出的最大效率，为了使SOFC电堆能够对纯氢以外的燃料具有更好的适用性，加入了燃料内部重整装置和燃烧室两个重要辅助组件。文中在对系统展开建模的基础上提出了采用非线性模型预测控制策略，能够更有效地使输出燃料气体的组分、温度、压力、浓度和流率满足燃料电池堆正常运行的需要，通过仿真分别论证了线性模型预测控制和非线性模型预测控制两种不同控制方案的有效性和适用性。

关键词: 燃料电池, 固体氧化物燃料电池, 内部重整装置, 线性模型预测控制, 非线性模型预测控制