节点文献

管式搅拌反应器流动特性与混合特性的CFD数值模拟

Numerical Simulation of Flow and Mixing Characteristics in Tubular Stirred Reactor Using CFD Method

【作者】 曹晓畅

【导师】 张廷安;

【作者基本信息】 东北大学 , 有色金属冶金, 2009, 博士

【摘要】 管式反应器广泛应用于化工冶金领域,根据我国一水硬铝石矿的管道化溶出中效率不高、管壁结疤现象严重等特点,本实验室自主开发了一种带搅拌的叠管式反应器,目前已通过冷态物理模拟的实验方法对这种新型搅拌反应器的流动和混合特性进行了初步研究,但有关的理论及设计方法仍不完善。近年来,利用计算流体力学(CFD)的方法研究搅拌设备内的流动和混合特性逐渐发展起来。CFD技术以其卓越的优势,对搅拌设备的研究、开发与设计提供新的研究方法。本文以自主开发的单、叠管式搅拌反应器为研究对象,利用商业CFD软件FLUENT6.3和并行计算系统的图形工作站,在已有的研究基础上对其进行了CFD的系统研究,并与实验数值进行对比,验证了利用CFD对管式搅拌反应器进行流动和传质特性研究的可行性,本研究能为反应器的进一步设计和改善提供新的研究思路,也可以作为实验研究和进一步模拟的基础。首先,采用雷诺时均方程(RANS)的标准k-ε湍流模型和多重参考系(MRF)法,对单、叠管式搅拌反应器进行数值模拟,根据传统(无搅拌)管式反应器与新型带有搅拌装置的管式反应器流场的比较,并通过对管式搅拌反应器速度场、湍流动能场的分析,带搅拌的管式反应器起到了增加管内湍流区,增强混合效果等作用。此外本文还讨论了不同转速对管内速度分布、压力分布、湍流动能分布的影响。在用RANS法对单、叠管式反应器的流场研究的基础上,采用速度场和浓度场分开求解的方法,计算了该管式反应器的停留时间分布,其模拟计算结果与实验测量结果吻合较好。通过对管式搅拌反应器内的停留时间分布模拟计算的结果显示,反应器内接近活塞流,带搅拌装置通过增加扰动、避免死区等来改善反应器的性能,还讨论了进口流量及搅拌转速对平均停留时间、方差及RTD曲线的影响。本文采用标准k-ε湍流模型和多重参考系(MRF)法,对单管式搅拌反应器的混合时间及功率准数Np进行了数值模拟,其模拟值与实验值吻合较好,并考察了不同搅拌转速以及不同监测点的位置对混合时间的影响规律。鉴于搅拌设备的复杂性,RANS法模型对于桨叶区湍流预测偏低,影响管内流体的物质交换的预报,本文采用大涡模拟(LES)的方法,用滑移网格(SM)法来处理搅拌桨区,对单管式搅拌反应器的流场进行数值模拟,并用该方法计算了停留时间分布和混合时间。结果表明LES法比标准k-ε模型更能有效解决该管式搅拌反应器之间的物质交换问题,尤其是在搅拌转速较高时模拟得到的平均停留时间及混合时间更趋于实验值,其中平均停留时间的平均相对误差降低了5%,混合时间模拟值的平均相对误差降低了13%。在单管式搅拌反应器内采用玻璃珠-水体系对固-液两相流场进行数值模拟,从CFD角度研究反应器内固-液悬浮状况下速度分布、浓度分布,并考察了不同搅拌转速的影响规律,同时,在固体体积2%和7.5%两种体系下,采用不同搅拌转速反应器内固相体积分率的的最大值来判断固体颗粒临界悬浮转速Njs,其值约为35rpm。

【Abstract】 Tubular reactor is widely applied in the chemical metallurgy field. To solve the problems of low efficiency and scarring at the tube wall in diaspore tube digestion techniques, a new-style tubular reactor with a stir is developed to decrease or even eliminate the scab in tubular reactors. Although the primary researches on the flow and mixing characteristic of this new-style reactor were investigated by the cold state physical modeling method, the study of mixing and the design methods are not developed well up to now. In recent years, the study of the flow and mixing characteristic using Computational Fluid Dynamics (CFD) has been gradually developing. With its superior advantage, the CFD method provides new approach for the design and optimization of the stirred reactors.In this study, based on the single and multi pipe tubular stirred reactor, a systematic CFD study was carried out using the commercial software FLUNET 6.3 on parallel computing graphic workstation. By comparing the simulation and experimental data, the feasibility of CFD method was verified to simulate the flow and mass transfer characteristics. The results can provide CFD research thought for further design, optimization and simulation.First, the flow field in the tubular stirred reactor was numerically simulated by using the Reynolds-averaged Navier-Stokes(RANS) approach with standard k-εmodel and Multi-Reference Frame(MRF). According to the analysis of the flow field, velocity field and turbulent kinetic of traditional(without agitation) tubular reactor and the stirred tubular reactor, the tubular reactor with a stirrer could increase turbulent kinetic, enhance mixing effect and avoid the dead zone. Furthermore, effects of different rotating speeds on the velocity, pressure and turbulent kinetic distribution were investigated Based on the flow field study, Residence Time Distribution (RTD) of the single/multi pipe tubular stirred reactor was simulated by using RANS, and the velocity field and the concentration field were simulated respectively. The result predicted by CFD model showed good agreement with experimental data. The CFD prediction showed that the flow pattern approached to the plug flow. According the tubular reactor without agitation, the tubular reactor with a stirrer could improve the flow profile by creating high Reynolds numbers and avoiding the dead zone. The effects of inlet mass flow and rotating speed on the mean residence time and variance were investigated numerically as well.The mixing time and power number Np in a single pipe tubular stirred reactor were simulated by CFD technique, and standard k-εmodel and MRF method were adopted for the simulation. The calculated mixing and power number were consistent with experimental data. The law of the influence of different rotation speeds and detecting positions on the mixing time was also discussed.In the light of the complexity of the stirred reactor, RANS approach could not predict the turbulence near the impeller properly, which would affect the mass transfer prediction. A new approach, the Large Eddy Simulations(LES), was used to predict the fluid field in the single pipe tubular stirred reactor, while the Sliding Mesh(SM) method was adopted for the rotation impeller. The RTD and mixing time were also modeled in this simulation. The results showed that LES can solve more effectively the mass transfer problem in the tubular stirred reactor than standard k-εmodel. The mean.residence time and mixing time predicted by LES were closer to experimental data, especially for high impeller speed. The average relative error of the mean residence time reduced by 5%, and a reduction of 13% was seen of that of the mixing time.The Solid-Liquid flow flied was simulated by the use of CFD, and sand particles were chosen as the dispersed phase. The particle volumetric concentration was 2% and 7%. The law of the influence of different rotation speeds and detecting positions on the velocity and concentration distribution was also investigated. In addition, the maximum solids concentration criterion was used to predict Njs. the just-suspended speed of the impeller. The Njs of these two particle volumetric concentration systems were similar, which was about 35rpm.

  • 【网络出版投稿人】 东北大学
  • 【网络出版年期】2011年 05期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络