【作者】 刘欣；

【导师】 郝晓弘；

【作者基本信息】 兰州理工大学 ， 控制理论与控制工程， 2013， 博士

【摘要】 在不可再生资源锐减、生态环境恶化的现状下,能源多样化的研究已将燃料电池这种高效、环保的产能模式推到了可再生能源研究的前沿,其中作为燃料来源多、转化效率高、容量范围宽、环境更友好的固体氧化物型燃料电池(Solid Oxide Fuel Cell, SOFC)是最具商业化发展潜力的能源之一。SOFC系统正常运转需要800-1000℃甚至更高的内部环境温度,电堆内高温、高速率的化学反应过程给系统在材料加工、结构设计、过程控制等方面开展研究带来了诸多困难。全封闭的工作环境使得过程参数难以实测获取,直接影响了对其电特性、热管理、稳定度的分析与研究。目前,以数值模拟与仿真分析为主、实验测试与参数拟合为辅的低成本、高效率的研究方法,仍然是加速SOFC本体及其发电系统技术发展的最有效手段。为了了解SOFC微观状态下的稳态性能以及关键参数的分布情况,研究中首先对板式结构SOFC单体建立了多物理场耦合模型,进行数值模拟与性能分析,在此基础上又对电池电堆展开了动态特性的分析与建模,采用时间尺度分析的方法简化模型。在控制方法研究过程中,首先对电堆的机理模型进行简化并建立了用于温度控制的非线性模型,开展温度控制策略的研究；其次对包含内部燃料重整环节和尾气循环利用的SOFC发电系统提出了模型预测控制的控制策略。论文的主要工作体现在：1.在对SOFC工作原理与特性开展电化学、热力学理论推导的基础上,为SOFC单电池建立了多物理场耦合模型并进行了稳态性能的分析。建模过程中,充分考虑了电池内部多种影响传热传质的因素,对不同物理域尺度下的基于流体力学、电化学、热力学的守恒方程采用设定模型边界耦合条件的方式进行了综合建模。该方法解决了由于模型属性差别、量级不同而无法匹配的问题,最终建立了稳态工作点下电池单体的三维仿真模型,能够采用3D图像的形式描述单电池极板、电解质层内部燃料与氧化剂的气体质量、流量、压力和电势、电流的分布状况。通过对不同工况下的仿真结果进行分析与比较,获得更全面的SOFC单体稳态性能描述。2.在数值模拟工作的基础上,结合实验测试获得的暂态性能分析结果围绕电池电堆的集总动态建模方法展开了研究。首先将电堆内部的动态行为等效为阻性、容性、感性元件在电路中带来的影响,重点考虑了电容性元素在燃料电池建模过程中的识别与数学表述。接下来引入了时间尺度分析的方法,将动态行为表述形式予以分类,或为常值,或为函数形式,使得最终得到的模型能够充分表征电池电堆内这些潜在的阻性、容性和感性元件在工作过程中参数的变化、反应速率的变化、耦合参数的调整等瞬态行为,为接下来控制方法的研究提供准确有效的数学模型。3.分析了引起电堆内部温度变化的参变量间直接或间接的影响关系,针对之前搭建的系统模型本身具有强非线性、单一控制策略无法达到控制目标等问题,结合控制系统设计的需要,将电堆的温度模型与NARIMAX模型进行联合,得到了包含电堆温度和输出电压的辨识模型。分别采用常规广义预测控制方法和改进型广义预测控制方法对入口气体流量、组分和温度进行实时控制,改进中采用了基于最小二乘原理的Levenberg-Marquardt算法作为参数预测优化的方法。通过对两种控制效果的比较,提出的算法降低了计算过程陷入局部极小值的可能性,并且使系统对由负载电流变化引起的扰动具有迅速、有效的抑制作用。4.为了使SOFC电堆对多种燃料具有更好的适用性,在之前的电堆模型基础上加入了燃料重整环节和尾气循环燃烧室两个重要的辅助部件,将之前建立的SIMULINK仿真模型扩展为可采用CH4、CO、H2、N2等多成分混合气体燃料的模型并以非线性状态空间方程的方式予以表述。针对扩展后的系统分别采用基于白适应UKF参数估计的非线性模型预测控制与鲁棒模型预测控制方法,从直接能量最大化、损耗能量最小化的角度提出控制的优化目标和约束条件。提出的控制方案在处理SOFC系统的控制目标值与边界约束发生冲突时能够表现稳定,当目标与约束条件发生冲突时,能迅速得出最接近于目标的控制输入的次优值。本课题的研究工作对于推进SOFC电堆及其集成系统在材料制备、结构设计、能效分析、控制策略方面的研究具有理论应用和实践指导的意义,为进一步研究和设计SOFC-GT系统的控制方案、系统故障诊断等方面的研究奠定了基础。更多还原

【Abstract】 The study of the various energy sources has brought the fuel cells, which are clean, pro-environment and efficient, to the forefront of the renewable energy research, when the non-renewable resources have dropped sharply and environment degraded. Of them, the Solid Oxide Fuel Cell (SOFC) which possesses the quality of resourcefulness, high convertibility, wide range and friendly environment has become one of the most potential energy resources for commercial development. It needs800-1000℃even higher temperature in the enclosed environment for the work properly. It causes difficulties for the studies on material processing, structural design and other aspects of the process control because of its higher temperature and faster speed in the chemical reaction process. In its enclosed working environment, part of the unmeasurable process parameters directly affects the analysis and study of its electrical characteristics, thermal management, and stability. At present, the low-cost, high-efficiency methods with numerical simulation analysis supplemented by experimental testing and parameter fitting have become the main means of speeding up the development of SOFC and the power generation.In order to understand the steady-state characters of SOFC in the microscopic and the distribution of the key parameters, this thesis adopts multiple physical field coupling models to carry out the numerical simulation for the planar structure of SOFC. Based on this, it builds the dynamic modeling of the stack, and simplifies the model using time scales. In the process of control method research, first, it simplifies the model and establishes the nonlinear model for controlling the temperature; second, it puts forward the predictive control strategy of SOFC power system which contains internal fuel reforming and exhaust gas recycling. The main work of the paper is presented in the following:1. Based on the analysis of the principle and characteristics of SOFC electrochemistry, thermodynamics theory, multi-physics coupling model is established for SOFC single cell and its steady state is analyzed. In the establishment, the various factors of affected internal battery of heat and mass transfer are fully considered. The model contains series of conservation laws of fluid mechanics, thermodynamics, electrochemistry which adopts the method of setting boundary coupling condition in different physical domains, the scales. This method solves the problems of the difference model attributed and unmatched magnitude. Based on this, the work has built the three-dimensional simulation model of the single battery under the steady-state operating point, describing the distribution of oxidizing agent quality, gas flow rate, gas pressure and electric potential under the multiple steady-state operating points inside the single cell using3D images so as to master the working point of the steady state performance.2. On the working of the numerical simulation, the stack’s lumped model is studied on the basis of analysis of the results obtained from the experiment. First, the dynamic behavior of the internal stack attributed to resistive, capacitive, inductive elements in the circuit, which causes the impact on it. Of them, the focus is on the identification and equivalent of the element of capacitance in the modeling. Then the time scale analysis method is introduced, the expression of the dynamic behavior is classified as constant value or as a function, so that the resulting model can adequately characterize the cell stack within these potential resistive, capacitive and inductive element in the work process parameter variations. All of these provide the accurate, visible, and controllable mathematical model for the following work of control method of the research.3. It first analyzes the relationships of the changes influenced by the internal temperature of the stack parameters directly or indirectly, and then combined the lumped model with NARIMAX model the system according to the system control design, which obtains the identification model containing the stack temperature and output voltage. It also uses conventional methods and improved one, respectively, to control inlet gas flow rate, composition and temperature in real-time. In the improvement, the Levenberg-Marquardt algorithm is applied as a parameter prediction. By comparison of the two control effect, the proposed algorithm reduces the computation process into the possibility of local minima and make system changes caused by the load current disturbance has rapid and effective inhibition.4. In order to make the SOFC pile have better applicability for composite fuel, the two important auxiliary parts-fuel reforming and exhaust gas recirculation-are added to the established pile model, expending the SIMULINK simulation model to use CH4, CO, H2, N2etc mixed gases nonlinear model and express in nonlinear state space equation method. For the extended system, the nonlinear predictive control system of the self-adaptive UKF parameters and robust model are used respectively, then put forward the optimization objectives and constraints of control and robust model directly from the point of energy maximization, loss minimization. All of these result better effectiveness. The control method performs steadily when the target value conflicts with the boundary constraints in SOFC system and quickly get the suboptimal control closest to the target control input of the optimal value.This research work has its practical guidance and application significance to develop the SOFC pile and the integrated system in material preparation, structure design, performance analysis, and the control strategy, has laid the solid foundation for the further establishment of SOFC-GT system in the control design and system fault diagnosis.更多还原