微型流化床因其区别于大型流化床的众多特点及潜在特定应用受到了越来越多的关注，按照实验、模拟、应用的思路，本论文系统研究了微型流化床内气体返混特性。通过研究微型流化床内气体返混程度，揭示了微型流化床内气体流动可实现最大程度接近平推流的特征，首次界定了“微型流化床”概念；借助CFD模拟，验证了实验结果；进而分别选取Geldart B和Geldart A类流化颗粒，在MFBRA中进行了气固反应测试，提供了微型流化床在最适条件下开展反应分析的有效应用案例。论文的主要研究内容和结果如下：1. 微型流化床系统的气体返混特性研究。考察由气体混合、流化床和检测部分构成的微型流化床系统利用FCC作为流化颗粒时随床径D，颗粒静床高Hs和床内表观流化气速Ug而变化的气体停留时间分布（RTD）。结果表明，随着床径，静床高以及表观气速的提高，操作参数对微型流化床系统中气体的返混程度影响增大，造成微型流化床以外部分与流化床部分的气体流动间差异增大，整个系统内气体RTD按照轴向扩散模型分析时产生的偏差增大。基于分析轴向扩散模型对系统内气体流动描述的适用性，定性判断了微型流化床系统内气体返混程度较小的操作条件范围，为微型流化床内气体流动接近平推流的判读提供了提供参考依据。2. 微型流化床中气体返混特性研究。结合反卷积数学方法和轴向扩散模型，定量分析了不同内径微型流化床内气体停留时间分布（RTD）的变化特性。考察了不同类型和大小的流化颗粒、静床高以及表观流化气速对RTD的影响，确定了停留时间分布函数方差σt2和峰高E(t)h之间的量化关系曲线。利用Geldart B类颗粒作为微型流化床颗粒介质时，σt2小于0.25，E(t)h大于1.0，E(t)h随着σt2的减小而指数式增大，σt2逐渐逼近0；此时Pea,g>20，结果表明气体流动基本接近平推流。当利用Geldart A类颗粒作为流化颗粒时，σt2在0.25和5.0之间，E(t)h在1.0和0.25之间，当且仅当静床高较小（Hs=20mm）时，方差与峰高能很好关联，Pea,g>10，气体流动的返混程度相对最小。这些结果为描述微型流化床内气体流动状态、判断和分析气体返混的特征提供了基础和依据，从而首次定义了什么是微型流化床。3. 微型流化床气体返混特性CFD模拟。采用双流体欧拉模型对微型流化床内气体停留时间分布进行研究，系统考察了颗粒静床高、表观气速及流化床内径对气体停留时间分布的影响规律。获得结果表明：模拟预测的RTD与实验曲线基本吻合，验证了实验中所运用的实验方法的准确性以及反卷积数学处理方法、轴向扩散模型的可靠性，同时验证了基于轴向扩散模型确定的RTD方差σt2和峰高E(t)h的关系曲线。运用稳定连续示踪法对微型流化床内气体返混程度开展进一步验证，明确了气体返混量通常小于2%，定量说明了微型流化床内气体返混程度。4. 利用B类颗粒的微型流化床反应研究。在MFBRA中进行以石英砂为流化颗粒的活性焦燃烧等温动力学研究，考察了不同操作条件对MFBRA求算的该反应动力学参数的影响。表明，所用流化气速太低或流化颗粒较小时，反应受外扩散影响较大，得到的反应活化能值较小。流化气速太高或流化颗粒太大时，反应颗粒与流化颗粒不能完全混合，导致气固传热传质效率降低，活化能值减小，活化能最高点出现的位置向转化率大的方向偏移。当反应器直径D减小时，计算的活化能值增大，证实了流化床内气体返混程度与床径的关系。针对反应分析的最佳操作条件，应当选用合适大小的流化颗粒，操作气速通常应为3倍的流化颗粒Umf，既能尽可能消除外扩散对反应动力学的影响，同时保证气固间的传热传质效率。5. 利用A类颗粒的微型流化床床反应研究。选取最优的MFBRA操作条件，利用α-Al2O3作为流化颗粒，开展了气固催化反应：甲烷催化裂解（CMD）的动力学研究，比较了三种不同方法，即MFBRA法，微型流化床中气体切换法和固定床中气体切换法获得的CMD动力学行为。结果表明，MFBRA方法使催化剂与流化颗粒充分流化混合，增大了催化剂的比表面积及与甲烷分子充分接触机会，得到了分散性更好的碳纳米管，且反应氢气开始生成的时间明显提前。此方法类似于工业上稳定进料的甲烷催化裂解流化床反应，表明用该方法得到的动力学参数对于实际工业流化床中的甲烷催化裂解反应更具应用指导价值和意义。;Micro fluidized bed (MFB) has received a lot of attention because of many traits different from traditional fluidized beds and potential specific applications. This work systematically studied the gas back-mixing characteristics in micro fluidized bed (MFB) through experiments, simulation and verification by application. First, it revealed the fact that the gas flow in MFBs can maximally approach the plug flow and then defined what is “micro fluidized bed” for the first time. Computational Fluid Dynamics (CFD) simulation was in turn performed to verity the experimental results. At last, gas-solid reaction research was carried out in the micro fluidized bed reaction analyzer (MFBRA) using Geldart A and Geldart B particles as bed materials to offer the effective application examples of reaction analysis using MFB under optimized conditions leading to gas plug flow.1. Characteristics of gas back-mixing in micro fluidized bed system. The influence on gas residence time distribution (RTD) in MFB system was investigated with respect to the inner diameters D of MFB, static bed height Hs of FCC partciles and superficial gas velocity Ug. Here, the MFB system mainly includes gas mixing part, micro fluidized bed and detecting part. The results show that the impact of operating parameters on gas back-mixing in MFB system becomes greater as Hs and Ug increase. This causes great deviation between experimental RTD of the system. Analyzing the applicability of ADM for gas flow in MFB system clearly identified the range of operation leading to limited gas back-mixin, and this in turn evidences the plausible judgement of MFB as a vessel to allow gas plug flow.2. Characteristics of gas back-mixing in micro fluidized bed. Here the influence of entrance and exit was removed from the gas RTD measured above. The variance of gas RTD in micro fluidized beds with different inner diameters was quantitatively investigated by combining the mathematical method of deconvolution and axial dispersion model (ADM). The resulting RTD for MFB was studied with respect to different particles, static particles height Hs and gas velocity Ug. We for the first time found that the RTD of gas is subject to an unique quantitative correlation between the E(t)h and σt2, the peak and variance of RTD for MFB, respectively. For the bed fluidized with Geldart B particles, the σt2 of RTD is below 0.25, and the height of E(t) peak E(t)h is larger than 1.0. Also, E(t)h exponentially increases as σt2 decreases to 0. At this point, Pea,g is greater than 20 but the gas flow in the tested MFB is highly approaching to a plug flow. For Geldart A particles, the σt2 of RTD is between 0.25 and 5.0, and the height of E(t) peak is larger than 0.25 but smaller than 1.0. Its Pea,g is only higher than 10 when Hs= 20 mm. Thus, the gas flow in the MFB has certain, although limited gas dispersion. These results provide solid basis to analyze the gas flow characteristics and back-mixing extent in MFBs, and make also our definition for the first time of what is the distinctive MFB.3. CFD modeling of gas back-mixing in micro fluidized bed. Two-fluid model was used to investigate the RTD of gas in MFB. The influences of modeling parameter settings on CFD simulation results were systematically investigated and discussed by comparing the simulation and experimental data. The gas RTD obtained from CFD simulation are similar with that from experiments under different but typical operating conditions. This verifies the feasibility of using deconvolution method and the reliability of using the axial dispersion model. Meanwhile, the quantitative correlation between the peak height E(t)h and the variance σt2 is also validated through calculation from the axial dispersion model. Using the steady tracer response method in our CFD simulation aims revealed that the amount of gas back-mixing in micro fluidized bed is hardly over 2%. 4. Research analyses in MFBRA using Geldart B particles. The isothermal kinetics of active coke combustion using silica sand as fluidized particles was analyzed in MFBRA. The influence of operating conditions on kinetic parameters was investigated. When the gas velocity was too low or the fluidized particles were too small, the external diffusion affected the reaction, leading to the smaller reaction activation energy. On the contrary, the reactant and fluidized particles can not fully mix at very high velocity and for very large fluidized particles. This also caused the lower heat and mass transfer efficiency and thus the smaller reaction activation energy. The maximal activation energy appeared at larger conversion. The reaction activation energy increases with decreasing inner diameter of the reactor, which verifies the relationship between gas back-mixing and reactor diameter. The best operating conditions aiming at reaction analysis can be determined: gas velocity is 3 times Umf of fluidized particles, in which the heat and mass transfer efficiency is guaranteed and meanwhile the influence of external diffusion on reaction can be greatly suppressed.5. Research analysis in MFBRA using Geldart A particles. Under its optimized operating conditions, MFBRA was adopted to study the reaction kinetics of catalytic decomposition of methane (CMD) using α-Al2O3 as fluidized particles. The kinetic behavior of CMD obtained from three different methods was compared: MFBRA method, gas switching method in micro fluidized bed and gas switching method in fixed bed. The hydrogen production rate appeared much higher when catalyst is injected into an MFB reactor by pulse, in comparison with the other two methods taking the gas-switching method. The carbon nanotubes (CNTs) produced in an MFB are more dispersed than that in fixed bed due to the efficient mixing of solid particles and reactants. In addition, the injection method is similar with the continuous feeding of reactants in an industrial process. Therefore, the micro fluidized bed reaction analyzer (MFBRA) is proven to be an alternative effective tool for determining the kinetics of methane decomposition and offering the high reference values for CMD in industrial fluidized beds.Key words: micro fluidized bed, gas back-mixing, residence time distribution (RTD), CFD simulation, gas-solid reaction kinetics