【作者】 蒋建华；

【导师】 李曦；

【作者基本信息】 华中科技大学 ， 控制科学与工程， 2013， 博士

【摘要】 固体氧化物燃料电池(Solid Oxide Fuel Cell, SOFC)以其安静、环保、高效的优点而成为21世纪最具发展前景的供电技术。其普及应用对保护环境、缓解能源危机具有重大的意义。随着SOFC技术的发展，电堆在实验室环境下已经具有良好的工作性能，但SOFC要走向应用，就必须脱离实验台的环境成为独立发电系统。因此，实现SOFC独立发电系统稳定、高效、长寿命地运行是其产业化应用的必经之路。为此，本文就SOFC独立发电系统的温度约束、功率跟踪和系统效率三个关键问题，从热电耦合建模、稳态性能优化、系统综合控制三个方面进行了深入的研究。由于本文具有很强的工程项目背景，所以本文的理论及方法研究成果均以能真正指导实际应用为检验标准。根据项目的推进计划，本文具体的研究对象为5kW平板式纯氢气SOFC独立发电系统以及水蒸气重整器，为后续集成重整器的SOFC独立发电系统的理论、方法研究及应用打下坚实的基础。首先，本文搭建了带有冷空气旁路阀的SOFC系统热电耦合物理模型。为了保证该模型的准确性，我们利用大量实验数据对模型进行验证。结果表明本文搭建的模型能够准确地反映实际电堆的工作特性，确保了本文研究工作的基础有效性。基于物理模型，我们对SOFC独立发电系统进行深入的稳态性能分析、优化，揭示了系统的内在机理，理清了系统各输入变量对温度约束、功率跟踪与系统效率三要素的影响关系，并获得了系统的最优操作点。研究发现，旁路阀装置对电堆的温度控制效果与系统效率都有显著的提高。然后，基于最优操作点，我们研究了系统的开环动态响应特性，为SOFC安全、快速、高效的动态控制奠定了基础。由于在实际应用中，SOFC内部的温度分布无法在低成本、不影响电堆性能的情况下获得，所以，基于最少的易测变量设计能够准确观测SOFC温度分布的观测器是进行SOFC最大工作温度、最大温度梯度控制的关键环节。因此，我们首先对物理模型采用温度层化简和准静态假设的方法对模型进行降阶简化，然后在工作点附近进行线性化获得状态空间模型。并基于“全可测最小维状态空间”的方法设计快速、准确的SOFC空间温度分布观测器。最后，本文分别使用“基于T-S模糊模型的约束广义预测控制策略”与“基于最优操作点的温度约束前馈、功率跟踪反馈控制策略”实现了对重整器系统与SOFC系统的有效控制。由于重整器样机已经实现组装与控制调试，在获得大量实验数据的基础上可以对物理模型进行校正，设计面向工程应用的“基于T-S模糊模型的约束广义预测控制策略”。该控制策略采用在线T-S模糊模型来校正CARIMA模型的参数获得准确的系统预测输出，并且利用拉格朗日乘算子法处理输入约束，这样在保证控制效果的基础上可以极大地减少计算量，有利于工程实现。仿真结果表明该算法比传统PID控制方法具有更好的控制性能。对于还尚处于组装过程中的SOFC系统，没有获得大量的实验数据来建立校正模型，以获取系统准确的阶次及时滞系数，从而无法设计面向工程应用的基于模型的广义预测控制算法。因此，以工程应用为导向，基于稳态分析与观测器，提出“基于最优操作点温度约束前馈、功率跟踪反馈控制策略”对SOFC的约束、功率、效率三要素进行协同控制。仿真结果表明，该控制策略能够有效地抑制温度约束的震荡，并具有快速的功率跟踪性能。最终实现了对SOFC安全、快速、高效的控制，将SOFC控制中温度约束、功率跟踪、系统效率三个相互耦合的关键控制要素进行了有效的协调管控。更多还原

【Abstract】 Due to the advantages of quiet operation, environmental friendly and high efficiency,SOFC (Solid oxide fuel cell) becomes the most promising power technology in21century.Although the SOFC stack achieves a good performance in laboratory conditions, thestand-alone SOFC system still faces the challenges of long-life time, high efficiency andload following in the process of the large-scale implementation. In order to achieve anstable and long-life operation for a stand-alone solid oxide fuel cell system, thetemperature constraints, load following and high efficiency are studied in this dissertationfrom the aspects of thermo-electrical modeling, steady-state analysis, observer design andcontroller development. Based on an engineering project, a steam reformer and a5kWscale pure hydrogen planar SOFC stand-alone system is studied in this dissertation.Firstly, a high-fidelity physical model of a improved SOFC system comprising aco-flow SOFC stack, a tail-burner, two heat-exchangers, a blower and a bypass valve, isdeveloped to capture both steady state and transient behavior of the system as well as thetemperature distributions in SOFC along the direction of gas flow. In order to ensure theaccuracy of the system model, the electrical characteristics of stack are validated by plentyexperimental data from two SOFC stacks (22and24cells) assembled in lab. The resultsconfirm the effectiveness of the work in this dissertation.Based on this model, the steady state performance of the system is analyzed andoptimized, which gives insights into the sensitivity of input variables to systemtemperature constraints, load following and system efficiency. And then, the systemopen-loop response between two optimal operation points are investigated. Moreover, thesimulation results show that the bypass valve in SOFC system can be used to improve thesystem efficiency and manage spatial temperature distribution both. As the spatialtemperature distribution profile is critical for the safe operation of SOFC system, andcannot be obtained by practical directly measuring at an acceptable cost, an accurate andfast linear observer is developed according to a novel sensor configuration estimationmethod to estimate the spatial temperature distribution profile in SOFC. Finally, two control strategies named “T-S fuzzy model based constrained generalizedpredictive control (TS-CGPC)” and “optimal operation points based temperatureconstraints feed forward and load following feedback control (OOPFF-FB)” areimplemented to steam reformer and the5kW SOFC system separately. Due to the steamreformer model can be modified by experimental data from the steam reformer prototype,a “T-S CGPC” controller is developed for practical implementation. In order to achievefast calculation, the CARIMA model in GPC is revised with the consequent parameters ofonline T-S fuzzy model and the input constraints is handled by the Lagrange multipliertechnique. The simulation results shows the reforming process can be well managed byTS-CGPC controller, which is better than PID controller. However, the SOFC system isunder assembly process, and the SOFC system model cannot be modified. The systemorder and time-delay coefficients are not available for the development of TS-CGPCcontroller for SOFC system. Therefore, based on the steady-state analysis and temperatureobserver, the control strategy of “OOPFF-FB” is proposed for the cooperative control oftemperature, power, efficiency in SOFC system. The simulation results demonstrate thatthe temperature variation of SOFC is effectively restrained during fast load following, andthe system efficiency is guaranteed by the optimal operation point. In a word, the SOFCsystem achieves high system efficiency and fast load following capability withouttemperature constraints violation.The modeling and control analysis results in this work can be extended to any SOFCsystems with different configurations and other similar nonlinear system, and provides avaluable solution for the control of SOFC system to achieve high efficiency, longlife timeand fast load following capability.更多还原