液固搅拌槽中流动和混合过程的数值模拟与实验研究
文献类型：学位论文
| 作者 | 王涛 |
| 学位类别 | 博士 |
| 答辩日期 | 2011-12-14 |
| 授予单位 | 中国科学院研究生院 |
| 导师 | 杨超 |
| 关键词 | 搅拌槽 错位桨 向心桨 斜叶桨数值模拟 镜像流体法 |
| 其他题名 | Numerical simulation and experiment on hydrodynamic and mixing in single phase and liquid-solid stirred tanks |
| 学位专业 | 化学工程 |
| 中文摘要 | 化工反应器的设计和放大是化学反应工程研究的核心内容。搅拌槽是过程工业中最常见的反应器之一，它具有相际接触面积大、传热和传质效率高、操作弹性大等特点，在化工、石化、生化和冶金等领域有着广泛的应用。在对搅拌槽反应器的设计和放大中，由于缺乏机理性的数学模型和工程实用的数值计算方法，研究者多采用经验或半经验关联式来进行逐级放大，这种方法不但费时，而且成本高、风险大。针对这种情况，本论文从实验研究和数值模拟两个方面对多相搅拌槽反应器内的流体动力学和混合特性进行研究。实验部分首先提出了两种新搅拌桨构型：错位桨和向心桨。以空气-水-石英砂三相体系为研究对象，通过与传统搅拌桨在功耗、混合时间和气体循环能力三个方面的比较，表明错位桨提高了径向流叶轮的轴向混合能力，混合时间低于传统的Rushton桨和斜叶桨，并且气体分散能力更强，通气对其功耗影响更小。而向心桨能利用叶片与圆盘法线的偏角，推动流体向心流动，产生了除三种传统流动（即径向流、周向流和轴向流）以外的新的流动方式，经过实验证明：向心流动有利于降低能耗、减少混合时间和改善液固混合状况。其次，对搅拌槽反应器中常用的三种搅拌桨加上错位桨进行组合，研究了双层桨体系的流体动力学，证明了通气对于固体悬浮不利，采用轴流桨的双层桨组合在能耗及稳定性方面均优于径流桨组合，但径流桨的宏观混合能力强于轴流桨。针对搅拌桨产生的不同流型提出了两种混合模式：强扩散弱剪切和先破碎再分散，进而解释了不同桨组合在混合时间方面的差异。随着计算流体力学（CFD）理论的发展和计算科学的进步，对搅拌槽内多相体系的流动进行数值模拟方面的研究已成为热点。斜叶桨具有良好的轴向混合能力，但相对于叶片形状比较规则的Rushton桨，斜叶桨的数值处理是一个难点。形状复杂的叶片一般采用近似台阶法、贴体坐标或浸入边界法处理，以上三种方法均有各自的优缺点。本论文提出了一种处理不规则叶片搅拌桨的新方法，首次将镜像流体法（Mirror fluid method，MFM）结合局部加密桨区，应用于湍流条件下（单）多相斜叶桨无（有）挡板搅拌槽的数值模拟中。镜像流体法把求解域扩展到固体区域，真实边界自然淹没在扩展的流场中。这种新方法可以准确地预测出斜叶桨搅拌槽的流动参数和相含率分布，模拟结果优于传统的k-ε模型，计算量也不会明显增加，对网格数要求不高（105数量级），而且对Re数也没有限制（Remax=1.9×105），是一种可行的处理不规则叶片搅拌桨的新方法。 |
| 英文摘要 | Reactor design and scale up is one of the core topics in chemical reaction engineering. Stirred tanks are most frequently employed in chemical industry, petrochemical processing, biochemical engineering and hydrometallurgy as a result of their relatively large contacting area, high mass and heat transfer efficiency, and flexible operation. Until recently, semi-empirical and empirical correlations have been mainly adopted for design and scale up of stirred tank reactors. Without convincing mechanismic models and engineering computational methods as the guide, reactor design is quite time-consuming and costly. Consequently, this dissertation is devoted to investigating the hydrodynamic and mixing characteristics in multiphase stirred tanks both experimentally and by numerical simulation. In the experimental section, two new impellers are proposed. One is named alternative blades disk turbine (ABDT), and the other is centripetal disk turbine (CDT). Relative to traditional impellers, ABDT is featured with blades arranged alternatively above and underneath the disk, and has been proved to have lower power consumption, shorter mixing time and stronger gas dispersion ability in multiphase systems. For CDT, the blades pitch forwards by an angle at the disk edge, which causes the generation of a new type of flow, i.e., centripetal flow. Such flow helps to decrease power consumption, shorten macro mixing time and improve solid suspension in multiphase stirred tanks. Then, the hydrodynamic performances of five dual-impeller combinations (including ABDT) are examined in a multiphase stirred tank. It is found that the configuration having an axial impeller consumes less energy than that with a radial one in all multiphase systems. Power number reaches a low limit with the increase of Reynolds number in liquid-solid systems. Aeration is proved to disfavor solid suspension. The experimental results show that the mixing time of the double radial impeller configuration is the shortest, while the double axial impeller combination performs the worst. Judging from the axial solid particle concentration distribution and overall suspension uniformity, the combination taking a pitched blade turbine downflow (PBTD) as the lower impeller achieves the best solid suspension. In numerical simulation section, mirror fluid method (MFM) combined with local grid refinement is proposed to deal with the numerical simulation of turbulent flow in a pitched-blade turbine stirred tank. By such a novel method, the domain occupied by the impeller is assigned suitable flow parameters explicitly by the mirror relation, so that the correct shear and normal forces on the fluid side of an interface segment is eventually guaranteed. Thus, the whole domain including the real fluid and the inside mirror fluid can be solved altogether by a set of equations. Moreover, it is beneficial to refine the local meshes in the impeller swept region so that the accuracy of computation in the high shear zone and the computational efficiency may be well compromised. Such methodology is proved to be sufficiently accurate and efficient through comparison of the present simulation with reported experimental data both in unbaffled or baffled stirred tanks with single or two-phase systems. The location of the interface in the multi-frame of references (MFR) is discussed, and the recommended criterion is that the inner region should be as small as possible. With such principle a satisfactory residual error level and efficient convergence rate could be achieved. Since the Navier-Stokes equations are solved in a fixed and regular grid for the whole expanded domain, the computational effort does not increase sharply as that employing grid regeneration procedure or using boundary-fitted coordinates (BFC). Compared with the extremely large number (with the order of magnitude of 106) of cells adopted frequently in the recent literature, the agreement between our predictions and the reported experiment results is satisfactory, though a total number of only about 105 nodes is utilized in this study. Another advantage is that no Reynolds number limitation exists in the use of MFM (Remax=1.9×105), which makes it more practical than the well-known immersed boundary method (Remax=7280) in the simulation under complete turbulent condition |
| 语种 | 中文 |
| 公开日期 | 2013-09-24 |
| 源URL | [http://ir.ipe.ac.cn/handle/122111/1751]  |
| 专题 | 过程工程研究所_研究所（批量导入） |
| 推荐引用方式 GB/T 7714 | 王涛. 液固搅拌槽中流动和混合过程的数值模拟与实验研究[D]. 中国科学院研究生院. 2011. |
入库方式： OAI收割
来源：过程工程研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。