【作者】 吴玉新；

【导师】 岳光溪；

【作者基本信息】 清华大学 ， 动力工程及工程热物理， 2007， 博士

【摘要】 为系统地采用数值方法研究气流床煤气化炉的运行特性,本文基于简化PDF方法对气流床煤气化炉建立三维数值模型,并对一台Texaco气化炉和清华大学分级气化炉进行了数值计算。通过对主要子模型进行评估、对预测结果同测试值进行比较、以及定量分析采用单组分简化PDF模型带来的误差,完成了对三维模型的模型验证和数值计算检验(Validation and Verification)。针对雷诺时均法(RANS)在耦合湍流-颗粒运动、湍流-化学反应时存在的理论欠缺,本文采用低维度湍流模型结合RANS方法解决这一困难,为此在一维湍流模型(ODT)基础上提出一种颗粒-涡团作用模型,模拟了平面自由射流中颗粒在亚尺度网格下的扩散过程。全文主要结论如下。(1)选取六面体贴体网格以及1/4炉体作为计算域能够达到节省计算时间和保证计算精度的最优效果;在三维复杂网格情况下,RSM模型受网格质量影响较大,Realizable k-ε模型适合用于对气流床煤气化炉冷态流场的计算。(2)基于单组分简化PDF方法的三维数学模型合理预测了一台Texaco气化炉和清华大学分级气化炉的出口组分及运行特性。气化炉中碳转化率>80%的区域内,简化PDF模型由于平衡假定、预设PDF形状以及煤气化过程中各元素均匀释放假定所造成的计算误差在10%以内。(3)ODT模型可再现平面自由射流的自相似性以及雷诺应力Reuu和Revv的分布特性,能够准确预测射流半宽及轴向速度衰减;预测结果显示涡团集中分布在射流剪切层上,涡团发生概率最大点的分布同射流半宽分布相吻合。(4)对颗粒在平面自由射流中扩散过程的计算结果表明,颗粒在扩散过程中会发生团聚现象;粒径与颗粒扩散程度成反比,小颗粒的质量流率在近喷嘴处为Gaussian分布,大颗粒质量流率在近喷嘴处呈现双峰分布,接近充分发展段才为Gaussian分布。这是因为模型中考虑了射流卷吸作用以及涡团扰动两种影响颗粒扩散的机理,在近喷嘴处射流卷吸作用对大颗粒运动起决定性作用。(5)受颗粒自身惯量和涡团作用的影响,St≈1的颗粒对涡团作用最为敏感。更多还原

【Abstract】 In order to make a systematic analysis for the performance of entrained flow coal gasifiers using numerical method, a three-dimensional numerical model for coal gasification process was developed based on presumed PDF (Probability Density Function) method. Numerical simulations applied this model, were carried out for a Texaco gasifier and a staged entrained flow gasifier of Tsinghua University. Validation and verification were carried out, which included the comparison between predicted results and the industrial measurements; the analysis of the reliability of the sub-models in the code; and a quantitative error estimation of the presumed PDF model. Due to the limitation of RANS (Reynolds Averaged N-S equation) method in solving turbulence-particle motion and turbulence-reaction coupling problems, a methodology was proposed based on the combination of RANS method and low dimensional turbulence model. A new particle-eddy interaction model with two-way coupling for particles and gas phase was developed and integrated into the ODT (One Dimensional Turbulence) model. A simulation case was also performed for a particle laden planar jet. The conclusions were drawn as follows.(1) Optimistic performance of the computational cost and accuracy of the results was obtained by using combination of a hexahedron body-fitted mesh and the calculation domain of 1/4th gasifier. The Reynolds Stress Model (RSM) is highly sensitive to mesh quality when adopting three dimensional body fitted mesh. The Realizable k-εmodel is a good fit for the simulation of the cold flow field of the entrained flow gasifiers.(2) Simulations were carried out for a Texaco gasifier and a staged entrained flow coal gasifier by using the proposed three-dimensional model. A good agreement at the outlet of the gasifiers was achieved for temperature and the composition of the synthesis gas between the simulation results and industrial measurements. A reasonable trend of gasification performance was also observed when adjusting the operation parameters. In this 3-D model, the principal error was introduced by the three assumptions made in the one stream presumed PDF method, which are equilibrium assumption,βPDF assumption, and the assumption of the same release speed of all species during coal gasification process. However, the error is no more than 10% when coal conversion is larger than 80%.(3) A numerical simulation was conducted for a turbulent planar jet with ODT model. The self-similarity of the planar jet and the distributive characteristics of Reuu and Revv were successfully repeated. A good agreement was achieved between the simulation and measurements of the jet half width and the decay of axial velocity at the center line. The result showed that most of eddies lied in the shear layer of the jet. The distribution of the points which has the highest probability of eddy formation on each ODT line agreed well with the spreading profile of jet half width.(4) The result of the particle laden jet showed that the agglomeration of particles occurred during particle dispersion in the planar jet. Particle diameter had an important negative effect on particle dispersion. The mass flux of micro particles showed an obvious Gaussian distribution near the nozzle, while that of macro particles had a bimodal distribution near the nozzle and the Gaussian distribution far from the nozzle. The analysis of the results revealed that two mechanisms had effect on the particle dispersion in the jet flow: one is the engulfment of the macro flow, and the other is eddy interaction. The engulfment of the macro flow dominated the macro particle dispersion near the nozzle.(5) Particle dispersion was affected by its inertia and the intensity of eddy interaction, which made those particles with St of roughly 1 the most sensitive to the eddy interaction.更多还原