【作者】 张雨英；

【导师】 杨晨；

【作者基本信息】 重庆大学 ， 热能工程， 2009， 博士

【摘要】 现代先进能源利用系统涉及机电、动力、化工、生物、环保、计算机、自动控制等众多研究领域,复杂性、非线性和时空多尺度是其固有特性。多尺度现象是复杂性科学的核心,对于复杂系统如化工、能源、经济、生态、天体、地貌等系统中出现的多尺度问题,由于没有统一的哈密顿量,所以不能像平衡系统那样进行统计和重整化,这也是目前多尺度科学的最大困难所在。描述这种客观的复杂性和多样性,传统的连续介质模型和单一尺度的建模与仿真已显示出了较大的不足。本文结合多尺度科学和热力系统动态学的研究方法,提出了一种多尺度协同仿真方法,并进行了应用研究,主要研究内容及结论如下:①提出了一种多尺度建模方法,这是本文第一个创新点。基于复杂系统的组合建模法、“描述型”的多尺度模拟方法,和研究能源利用系统多尺度问题的界面耦合法的基本思想,给出了多尺度建模的主要步骤,对多尺度模型进行了分类,分析了各尺度下建模方法的选择,概括了多尺度建模顺序的几种方式,构建了各尺度模型之间信息交换和尺度模型之间耦合的方式。多尺度方法在能源领域的研究与应用还处于起步阶段,本文提出的多尺度建模方法抓住系统多尺度特性进行分析,使建模过程既满足认识系统内部非线性、多尺度等复杂内在本质特性的需要,也能使多尺度模拟的计算量在可控制的范围内,从而使传统模拟方法的深度和广度得到了延伸。②建立了一种基于TCP/IP网络的协同仿真框架,这是本文第二个创新点。基于点对点式结构的协同仿真模式和时间同步机制,结合现有的Windows的Socket通讯机制、进程线程机制和TCP/IP通讯协议,提出了多个尺度模型之间进行数据交换的一种基于网络的协同仿真框架。该框架将多个尺度模型两两组合为一系列数据交换对,采用重构和压缩算子实现各尺度模型间信息的转换,并采用保守或乐观时间同步协议实现多尺度模拟的同步和并行仿真,实现了模型的可重用性和计算并行性,从而更深入地揭示系统的多尺度本质特征。用一个具有典型Stiff特性的一维Brusselator模型的串行仿真和并行协同仿真的比较分析,验证了该方法的可行性与有效性。③建立了管式SOFC的宏观CFD和介观LBM两个尺度的模型,进行多尺度协同仿真。建立了管式SOFC宏观尺度的基于质量、动量、组分和能量守恒以及电化学反应动力学方程模型,并进行离散化求解;针对管式SOFC电极/电解质界面建立了介观尺度的基于电化学扩散反应的LBM模型;应用上述协同仿真框架完成了两个尺度模型的耦合、数据交换及并行协同仿真;首次采用多尺度观点对燃料电池进行仿真研究,揭示了电池内部更真实的物理化学特性。④对美国Capstone公司的微型燃气轮机进行了试验研究,揭示了其主要性能,为SOFC-MGT系统的建模提供了依据。⑤首次以多尺度的观点对固体氧化物燃料电池/微型燃气轮机(SOFC-MGT)混合发电系统进行了多尺度协同仿真研究,这是本文第三个创新点。分析SOFC-MGT系统的结构和主要控制机制,将其划分为系统尺度的过程模型、SOFC宏观尺度的CFD模型和SOFC介观尺度的电化学扩散反应模型三个层次;分别采用过程模型的模块化建模、CFD模型的流场模拟和电化学扩散反应的LBM(Lattice Boltzmann Method)模拟,以从中间开始的建模顺序完成了SOFC-MGT系统的嵌入式多尺度模型的建立;利用压缩和重构算子完成不同尺度下参数之间的耦合,通过基于TCP/IP的网络协同仿真框架交换耦合参数,集成各个尺度模型完成了SOFC-MGT系统的多尺度协同仿真,保守时间同步协议保证了数据交换的正确性,并进行了仿真实验。多尺度协同仿真方法在SOFC-MGT系统中的成功应用,初步验证了多尺度协同仿真相对于传统单一尺度仿真的优势,主要表现在:1)抓住了复杂系统的多尺度本质特性进行分析,将系统划分为不同尺度下的子系统,分析其相应的控制机制,加入尺度间的相互作用,能够更深入地揭示复杂系统的内在机理;2)弥补单一尺度仿真时所作简化和假设带来的误差,能够获得系统更真实、准确的结果;3)可以同时得到相同条件下多个尺度模型的计算结果,从多个层次出发,更系统、更深入地分析对象的性能与特性,有利于揭示对象的本质特性。更多还原

【Abstract】 Complexity, nonlinearity and temporal-spatial multi-scale characteristics are inherent mechanisms for modern advanced energy systems which involve mechanics and electrics, power, chemistry, biology, environmental control, computer, automatic control and many other research areas. Multi-scale phenomena are the core of science of complexity. Because there is no uniform Hamiltonian, the statistic and renormalization of the complex systems can not be achieved like the balance systems, such as chemical systems, energy systems, economic systems, ecosystems, astronomic systems, and geomorphic systems. It is also the largest difficulty of multi-scale science for now. The traditional continuum model and single-scale simulation reveal a huge limitations in describing the complexity and diversity. A multi-scale and collaborative simulation method and its application have been developed based on the ways of multi-scale science and thermo-dynamic system dynamics. The main research contents and conclusions of this paper are as follows:①Presenting a multi-scale modeling method for energy utilizing system The primary steps of multi-scale modeling have been proposed, the types of multi-scale model have been categorized, the modeling modes of scales have been chosen, the modes of multi-scale modeling order have been summed, and the patterns of information exchange among scale models and coupling scale models have been built, based on the combination modeling means of complex system, the descriptive modeling way of multi-scale simulation, and the coupling interface method to research the multi-scale problems of energy systems.The research and application of multi-scale method in energy field are still in its infancy. The methodology presented in this paper aims at the analysis of the system multi-scale characteristics, makes the modeling process meet the needs for recognizing complex and essential characteristics inside the system, such as nonlinear and multi-scale identities, makes the computation of multi-scale simulation among the controllable range, and extends the range of conventional simulation methods.②Building a framework of collaborative simulation based on TCP/IP Based on the point to point collaborative simulation and the time synchronization mechanism, combining the existing Socket communication mechanism, Windows course mechanism and TCP/IP communication protocol, a collaborative simulation framework for exchanging data among scale models has been presented. Each two scale models have been assembled a series of pairs to exchange data, and the conversions of information among all scale models have been achieved by reconstruction operator and compression operator, and the conservative or optimistic time synchronization protocol have been used to ensure the synchronization and parallel simulation of multi-scale model. With these steps, the reusability of established models and the parallel computation have been realized, and the multi-scale characteristics of system have been deeply revealed.The feasibility and validity of this way have been proved by the comparative analysis between serial and parallel collaborative solution of 1D Brusselator model having the typical Stiff characteristic.③Performing the multi-scale collaborative simulation of tubular SOFC concerning the macroscopic CDF model and mesoscopic LBM modelBased on the conservation of mass, momentum, species and energy, and the equations of electrochemistry dynamics, the macroscopic model of tubular SOFC has been built, discretized and solved. The mesoscopic LBM model aiming at the electrochemistry diffusion reaction of the electrode/electrolyte interface in tubular SOFC has been established. The coupling, data exchange and parallel computation of two models have been achieved by the collaborative simulation framework presented in this paper. The first simulation of fuel cell from a multi-scale point of view has been fulfilled, and the results have revealed the more realistic physical and chemical characteristics inside cells.④The experimental studies on Capstone micro gas turbine⑤The comprehensive simulation of solid oxide fuel cell/micro gas turbine (SOFC-MGT) hybrid generating system from a multi-scale point of view Based on the analysis of SOFC-MGT system configuration and dominate control mechanism, the SOFC-MGT system has been divided into three levels, namely, the process analysis at system scale, the flow field analysis of SOFC at macroscopic scale, and the electrochemistry diffusion reaction analysis of SOFC at mesoscopic scale. All of these have been completed using the modular modeling of process model, the flow field simulation of CFD model, and the LBM simulation of electrochemistry diffusion reaction respectively. The embedded multi-scale model of SOFC-MGT system has been built by using the mode of middle-out. With coupling the parameters of different scale models by the compression and reconstruction operators, and exchanging the coupling variable using the collaborative simulation framework based on TCP/IP, the multi-scale and collaborative simulation of SOFC-MGT system has been performed by integrating all of three scale models. The correctness of exchanged data has been ensured by the conservation time synchronization protocol, and the simulation experimentation on the system response to the step disturbance of fuel import quantity has been accomplished.The successful application of multi-scale collaborative simulation methodology in SOFC-MGT system has been preliminarily verified and shows that the multi-scale collaborative simulation has many advantages over the traditional single-scale simulation: 1) With analyzing the multi-scale characteristics of complex system, the system is divided into many subsystems in different scales. With the analyses of corresponding control mechanisms in all scales, the interaction among scales has been added into the models, and the outcomes have deeply revealed the inherent mechanisms of complex systems. 2) The errors from the simplification and hypothesis of single-scale simulation have been compensated by the multi-scale and collaborative modeling, and the more reality and accurate results can been acquired. 3) The multi-scale collaborative simulation can simultaneously obtain more comprehensive results from different scale models.更多还原