【作者】 杨杰；

【导师】 李箭； 李曦；

【作者基本信息】 华中科技大学 ， 材料学， 2009， 博士

【摘要】 固体氧化物燃料电池(Solid Oxide Fuel Cell,简称SOFC)是一种高效、清洁、安静、可靠的能源转换装置,它可以将化学能直接转换为电能。SOFC具有燃料来源灵活、热电联供效率高、抗毒化性能优越、固态组件稳定性高、较高的功率密度和环境友好等自身优点,被认为是最有发展前途、最具有商业化价值的燃料电池技术之一。目前SOFC的工作温度已经下降到750℃左右,然而这样的高温环境依然带来很大挑战。因为高温及温度变化将会给电堆机械性能带来负面作用,导致电池堆热疲劳、分层甚至断裂;此外温度变化也会给SOFC电特性带来影响,例如影响电池电压、电流密度分布和电功率等,所以对温度的管理和控制显得尤为重要。同时,国内外对SOFC性能的研究也从材料、结构和模型等逐步向电堆、系统等技术方向发展。电堆的可靠性和安全性需要高效的控制系统来实现,从而在负载波动条件下,电堆气体的温度和压力保持稳定,输出性能良好,使用寿命延长。此外,外部功率负载跟踪也是电池堆技术达到一定水平后需要考虑和分析的。本文以中温平板式SOFC电堆和系统为研究对象,探讨电堆热管理的控制策略和外部负载跟踪方法,以实现优化的温度控制和系统供能。论文首先介绍了SOFC电池的发展,特别是目前电堆和系统研究状况。随后论文讨论了建模过程中采用的基本理论、研究方法和算法。在详细分析了模型各个子系统的结构和组成之后,采用Matlab软件建立起SOFC电堆仿真模型。对SOFC电堆热管理系统进行建模,设计并实现一种离线的改进Takagi-Sugeno(T-S)模糊模型来代替物理模型,仿真结果验证了算法的良好逼近性能。运用基于在线的优化T-S模糊模型建模和离散优化的非线性模型预测控制(Model Predictive Control,简称MPC)算法,来实时控制电堆的工作温度,仿真结果显示了较好的模型仿真性能。运用参考模型法跟踪功率负载变化,研讨了功率负载动态跟踪的一种方法。论文的主要工作体现如下:(1)搭建千瓦级的SOFC系统模型。系统包含燃料电池堆、热管理子系统(电堆热管理和外部热管理)、燃料处理子系统、电力电子子系统和系统控制子系统等五大部分。着重分析了SOFC电堆建模过程,详细阐明了SOFC电堆的四个子系统,即电特性子系统、阴极流子系统、阳极流子系统和电堆热管理子系统。分析了电池堆四个子系统搭建所依据的基本原理和方法,探讨了子系统模型结构以及在Matlab程序中的实现。此外,讨论了外部热管理系统的构成和工作原理。(2)采用改进的T-S模糊模型仿真分析SOFC电堆中热管理子系统。依据质量守恒和能量守恒定律建立起温度控制的焓平衡方程,完成焓入和焓出等计算,从而构建起温度控制的物理模型。系统的两个输入为空气流和氢气流,输出为电堆温度。采用改进的T-S模糊模型建模,运用所搭建物理模型产生的数据来训练和辨识改进的T-S模糊模型。结果显示,改进的T-S模糊模型的温度响应值与物理模型产生的温度响应值相似度很高。(3)模型预测控制(MPC)建模在线实时控制SOFC电堆的温度。首先建立起MPC模型框架,然后构建一个基于优化的T-S模糊模型来仿真SOFC电堆,详细分析优化的T-S模糊模型的计算步骤和方法。阐明了MPC模型控制量离散寻优的方法及在SOFC电堆温度控制中的应用。从仿真结果看,SOFC电堆温度的模型预测控制输出和电堆实际输出之间的跟踪曲线非常逼近。(4)探讨外部功率负载动态跟踪方法。针对外部负载通常在不同的时刻启动、加载、卸载、制动等动态变化,提出参考模型法来实现外部功率负载的跟踪。搭建一个负载动态跟踪模型,把SOFC电堆电特性模型做参考模型。通过模型构建的功率与氢气流量二者之间对应关系,实现外部功率负载的跟踪功能。更多还原

【Abstract】 Solid Oxide Fuel Cell (SOFC) is a device of chemical energy conversion electrical energy with high efficient, clean, quiet and reliable. SOFC have many advantages, such as fuel source flexibility, combined heat and power, high performance, superior poison tolerance, high stability of solid state components, high power density, environmental friendly, etc. It is considered as one of the most promising and having commercial value for the fuel cell technology.The high temperature environment also brings great challenge regardless of the fact that the current operating temperature for SOFC has dropped to about 750℃. The high temperature and temperature change will affect the stack mechanical properties which lead to thermal fatigue, delaminating or even break. Besides, the temperature also impact to the electrical properties of SOFC, such as voltage, current density distribution and electric power, etc, therefore, the management and control of temperature is especially important for SOFC. At the same time, the study for SOFC at home and abroad gradually develops from materials, structures and models to the stack and system technology. The high efficient control system could operate the stack so reliable and secure that it brings the gas temperature and pressure stability, high performance of the output, and long life for the stack on load fluctuating. Furthermore, an external power load tracking could be researched and analyzed when the stack technology has advanced a certain extent. In this thesis, the intermediate temperature planar SOFC stack and system are considered as the research object. This thesis explores the control strategy of the thermal management of SOFC stack and the method of external power load tracking in order to achieve optimum temperature control and optimization of energy supply.The thesis firstly introduces the development of SOFC, in particular the current research status of the stack and system. Then it discusses the basic theory, research methods and algorithm used in modeling process. After analyzing the structure and composition of the various subsystems of model in detail, a simulation model of the SOFC stack is established by the software-Matlab. An off-line improved Takagi-Sugeno (T-S) fuzzy model is designed and implemented for replacing the physics model on the basis of modeling the thermal management system of SOFC stack, and the simulation results validate that the algorithm has high approximation. Moreover, an algorithm based on the on-line modeling of optimizing T-S fuzzy model and discrete optimization of nonlinear Model Predictive Control (MPC) could finish the real-time control of the stack temperature, and the simulation results show good performance. Finally, a method for dynamic power load tracking is discussed using a reference model law. The conclusions of the thesis are listed as below.(1) A model of kW class SOFC system is built. The SOFC system has five parts, including SOFC stack, thermal management subsystem (including stack thermal management and external thermal management), fuel processing subsystem, power electronics subsystem and control subsystem. The modeling process of SOFC stack is analyzed, and the SOFC stack consists of electrical characteristics subsystem, the cathode flow subsystem, the anode flow subsystem, and the stack thermal management subsystem. This paper discusses the structure and realization of four subsystems in Matlab program according to the basic principles and methods of modeling. Furthermore, the component and working principle of external thermal management system are described.(2) The thermal management subsystem of SOFC stack is simulated by an improved T-S fuzzy model. Enthalpy balance equation for temperature control is set up on the basis of mass conservation law and energy conservation law, and they can fulfill the calculation of enthalpy in and enthalpy out so that the physics model for temperature control is constructed. Actually, the model has two inputs, the air flow and hydrogen flow, and one output, the stack temperature. Then, a model built by an improved T-S fuzzy model is trained and identified through the data which the physical model produces. Finally, the result show that the temperature response value of the improved T-S fuzzy model is very close to the value of the physical model generated.(3) The Model Predictive Control (MPC) modeling is made for on-line and real-time the temperature control of SOFC stack. Firstly, a MPC frame is established. Thus, the optimized T-S fuzzy model which is constructed is simulated for a real SOFC stack, and its computation step and formula are introduced. The way of discrete optimization of control variable is explained and is applied to the temperature control of SOFC stack. Finally, the simulation result show that the temperature tracking curve between the MPC output and the actual output of the physical model approaches extremely.(4) The thesis discusses the method of external power load dynamic tracking. According to the change of external load starting, loading, unloading, and braking at different times, a reference model method is provided to implement the external power load dynamic tracking. Hence, a model of load dynamic tracking is set up, and the reference model is the model of SOFC electric characteristic. The relations between an external power load and the corresponding hydrogen flow are established, and the function of the external load tracking can realize.更多还原