【作者】 陈巨辉；

【导师】 陆慧林；

【作者基本信息】 哈尔滨工业大学 ， 热能工程， 2013， 博士

【摘要】 气固两相流的流动与反应现象普遍存在于能源、化工、电力等领域，一个典型的工业应用就是流化床反应器。流化床以其颗粒掺混性强、操作温度易控制、燃料适应性广等优势被广泛应用，深入认识和研究流化床内气固两相流动与反应机理有着重要的现实意义。计算机技术的发展使得数值模拟方法已经在气固两相流的预测与研究中广泛应用。然而，流化床内气相湍流、颗粒碰撞耗散以及气固相间多向耦合作用等问题增加了模拟的复杂性。另外，均相反应与气固异相反应的同时存在，反应过程的吸放热温度变化，使得气固相间的流动与反应过程相互影响，对研究模型精度的要求更高。因此，正确合理地构建和完善数学模型对流化床内气固两相流动与反应的研究有重要的指导意义。本文采用大涡模拟方法模拟气相湍流，推导了考虑气固相间作用影响的气相亚格子湍动能方程，建立了亚格子湍动能模型。颗粒相考虑了颗粒速度脉动各向异性，采用颗粒速度脉动二阶矩模型。对于气固相间作用，不但考虑了气固相间作用力影响，而且补充了气固相间二阶脉动能量作用的影响，类比Simonin模型，通过气相亚格子湍动能与颗粒相速度脉动二阶矩之间的脉动能量传递来封闭气固相间作用二阶模型。考虑气相湍流、颗粒相速度脉动各向异性及气固相间的一阶与二阶相互作用建立了气相大涡模拟-颗粒相二阶矩双流体模型（LES-SOM Model）。将LES-SOM模型拓展到反应领域，补充了组分方程和能量方程，同时考虑了反应过程中气相与颗粒相质量改变对流动的影响。将涡耗散概念（Eddy Dissipation Concept, EDC）反应模型思想应用于气相大涡模拟反应中，通过比较亚格子过滤尺度与微细结构尺度的大小关系，在微细结构上建立亚格子反应模型进行封闭。应用LES-SOM双流体模型分别对低质量流率和高质量流率提升管内气固两相流动特性进行数值模拟。与Jiradilok等低质量流率情况和Herbert等高质量流率情况的实验数据吻合均较好。模拟结果显示提升管内颗粒浓度中心低边壁高，颗粒速度中心高边壁低，呈典型的“环-核”流动结构。气相亚格子湍动能与亚格子能量耗散呈中心高边壁低的分布趋势。颗粒速度脉动各向异性明显，颗粒相轴向速度脉动二阶矩大于径向速度脉动二阶矩，分散颗粒的速度脉动二阶矩及速度脉动各向异性比颗粒聚团明显。颗粒相雷诺应力型二阶矩与气相雷诺应力型二阶矩分布一致，数值上小于气相雷诺应力型二阶矩。同时比较了不同气相大涡模拟亚格子模型、不同气固相间二阶作用模型下气固流动特性，分析了颗粒弹性恢复系数、气体表观速度及颗粒入口质量流率对气固流动的影响。随着气体速度增加和颗粒循环流量降低，颗粒轴向二阶矩与径向二阶矩的比值呈幂函数增加，其比值可达到4-4.5。颗粒浓度径向分布由“U型”变为“倒U型”，分布出现反转。获得了所研究的颗粒直径和密度为300m和2500kg/m~3时径向颗粒浓度分布出现反转、流动由“稀相流态化”变为“快速流态化”的界限气体表观速度与界限颗粒循环流量。应用LES-SOM双流体模型，对流化床内煤颗粒的流动燃烧过程进行数值模拟。建立了煤燃烧反应模型，比较了气相亚格子反应模型对燃烧过程模拟结果的影响，模拟得到的出口处各气体组分含量与Topal等实验数据吻合较好。模拟结果给出了煤燃烧过程中煤颗粒与脱硫剂颗粒的浓度、速度、颗粒速度脉动二阶矩及雷诺应力型二阶矩分布规律，并对不同组分速度脉动各向异性进行了研究。同时得到了煤热解、碳燃烧、挥发分燃烧、NOx气体排放及脱硫过程中气体及颗粒各组分分布及温度场分布特点。总结了提升管煤燃烧过程中，反应速率随温度和浓度的变化规律。亚格子反应速率随温度升高而增大。随颗粒浓度的增加，挥发分均相反应速率升高，而碳颗粒燃烧异相反应速率逐渐降低。考虑了高颗粒浓度下的摩擦应力影响，将LES-SOM双流体模型应用于鼓泡流化床内生物质木材颗粒的气化过程。建立了生物质气化反应模型，比较了气相亚格子反应模型对气化过程模拟结果的影响，模拟得到的出口处各气体组分含量在实验数据允许范围内。分析了鼓泡床内气泡的存在对流动与反应的影响，着重研究了粒径不同的两种碳颗粒在反应过程中浓度、速度、颗粒速度脉动二阶矩及雷诺应力型二阶矩变化特性。同时分析了焦油热解、燃气氧化和碳颗粒燃烧及气化过程特点，模拟结果给出了气体各组分分布以及温度场分布，并总结了流化床生物质气化过程中，反应速率随温度和浓度的变化规律。亚格子反应速率随温度升高而增大。随着颗粒浓度的增加，气体均相反应速率和碳的氧化反应速率先增加，达到最大值后，再逐渐降低。在低颗粒浓度时，碳还原反应速率随浓度增加而增加；在高颗粒浓度时，随颗粒浓度变化不大。水煤气反应和甲烷化反应与颗粒浓度基本无关。更多还原

【Abstract】 Reactive gas-solid two-phase flows are widespread in energy conversion,chemical and oil industries, electric power and other fields. A typical industrialapplication is fluidized bed reactors. Fluidized bed reactor has achieved a wideindustry uses over the past decades due to its advantages such as mixingefficiently of particles, ability to control temperature easily, fuel flexibility.In-depth understanding and grasp of the mechanisms of gas-solid flows has verypractical significance. With the advancement of computational fluid dynamics(CFD), numerical simulations have been widely utilized in the research ofgas-solid flows. However, challenges such as gas-solid turbulence, dissipation ofparticle collision and multi-coupling between gas and solid phases increase thecomplexity of numerical simulation. Besides, the coexistence of homogeneousand heterogeneous reactions and variation of temperature due to heat exchange ofreactions lead to severe coupling between the interdependent process of turbulentgas-solid flow and reactive process. Up to now, there is no universal numericalmodel applicable to the reactive gas-solid process in fluidized bed reactor in theopen literature. Current mathematical models from different researchers based onsome assumptions are only valid under certain circumstances. Therefore,development of mathematical models plays an important role in researchingreactive gas-solid flow process.Based on the Large Eddy Simulation (LES) of gas turbulence, the turbulentkinetic energy equations considering interaction of gas-solid phases are present,and the subgrid turbulent energy model is proposed. The transport equations forthe second-order moment (SOM) of particle fluctuating velocity are used,considering anisotropic characteristic of particle velocity fluctuating. For theinteractions between gas and solid phases, not only consider gas-solidinter-phase force, but the second order fluctuating energy are implementedtransferring between gas subgrid turbulent energy and second-order moment ofsolids. The LES-SOM model is presented considering gas turbulence, anisotropyof solid fluctuating velocity and first order and second order interaction betweengas and solid phases. With the extension of LES-SOM model into reactions, theequations of species and energy of gas and solid phases are presented. Applyingthe eddy dissipation concept (EDC) reaction model into LES simulation of gasreactions, by comparing subgrid scale and “fine structure” scale, the subgrid reaction model is presented within fine structure.Numerical simulations of hydrodynamics of gas-solid two-phase flows inrisers are performed based on LES-SOM model. The simulations are in goodagreement with low mass flux experiment results by Jiradilok et al. and highmass flux experiment results by Herbert et al. Simulations predict a typicalcore-annular flow structure of low solid concentration and upward flow in thecore, and high solid concentration and downward flow in the annular region. Thedistribution of subgrid turbulent kinetic energy and subgrid energy dissipation ishigh in the middle and low near the wall. The anisotropic behavior of fluctuatingvelocity of particles is obvious. The axial second-order moment of particle ishigher than radial second-order moment of particles. Dispersed particlesdemonstrate more anisotropic behavior of fluctuating velocity than particles inclusters. The distribution of solid Reynolds stress has the similar trend with thoseof gas phase, but the value is lower. Compare gas-solid flow behaviors withdifferent sub-grid LES model of gas phase and with different gas-solidinteraction model between phases. The effects of different restitution coefficients,superficial gas velocities and mass flow rates are analyzed. As the gas velocityincreases and particle circulating flow reduces, the ratio of axial and radialsecond-order moments is increased as a power function. The ratio can be up to4-4.5. The distribution of particle concentration along radial direction changesfrom "U" type to "inverted U" type, and the distribution is occurred reversal. Itobtained the critical gas superficial velocity and critical particle circulation flowwith diameter and density of particles of300m and2500kg/m~3.The LES-SOM model is applied to simulate the flow and combustionbehavior of coal particles in fluidized bed combustor. The comprehensive coalcombustion model is presented. By comparing simulating results with subgridreaction model of gas phase, the simulating gas components in the outlet agreewell with experiment of Topal et al. Simulations also take coal devolatisation,carbon particles combustion, volatile matter combustion and desulfurizationprocess into account. The results capture the distribution of concentration,velocity, second-order moment, and Reynolds stresses of coal particles anddesulfurizer particles. The distributions of gas components and temperaturefields are studied. The reaction rates of coal combustion process are alsoconsidered as a function of temperature and concentration. The subgrid reactionrates increase with temperature rising. With the increase of particle concentration,volatile homogeneous reaction rates increase, and carbon heterogeneous reaction rates decrease gradually.Considering solids frictional stress in high volume fraction of particles, theLES-SOM model is applied to simulate the gasification process of wood particlesin bubbling fluidized bed. The comprehensive wood gasification model ispresented. By comparing simulating results with subgrid reaction model of gasphase, gas species in the outlet are in agreement with experimental data withinacceptable error. The influence of bubbles on flow and reaction behavior isstudied. It is especially addressed that distribution of particle concentration,velocity, second-order moment and Reynolds stresses of two groups of particleswith different sizes in the gasification process.The process of tar devolatisation,wood gas combustion, char combustion and gasification are also analyzed.Simulation also presents the distributions of gas species molar fraction andtemperature field in the reactor. The reaction rates of biomass gasificationprocess are also considered the relation with temperature and concentration. Thesubgrid reaction rates increase with temperature rising. With the increase ofparticle concentration, homogeneous reaction and carbon oxidation rates increasefirst, reach the maximum, and then decrease gradually. At low particleconcentration, carbon reduction reaction rate increases with the increase ofconcentration, but at high particle concentration, little changes with theconcentration of particles. Water gas reaction and methanation reaction rates areundepedent upon particle concentrations.更多还原