[新版]4.en_cn 真空制盐蒸发器中蒸发室流体力学剖析

[新版]4.en_cn 真空制盐蒸发器中蒸发室流体力学剖析 真空制盐蒸发器中蒸发室流体力学分析及其对固体悬浮的影响 王学魁，武首香，沙作良 (天津市海洋化学与资源重点实验室，天津科技大学，天津 3000457) 摘要:蒸发室的流体动力学状态对蒸发结晶过程具有重要的影响。本文使用计算流体力 学(CFD)的方法，对在真空制盐中常用的两种蒸发时结构的流体力学状态进行分模拟分析， 探讨不同的操作参数队蒸发室内流体力学状态的影响，以及对不同粒径悬浮状态的影响， 从 而分析其对蒸发结晶过程的影响。 关键词: 计算流体力学; 两相流; 蒸发; 结晶 Fluid Dynamic and Its Effect on Solid Suspension in Evaporation Chamber of Evaporation Crystallizer of Salt Production WU Shou-xiang，WANG Xue-kui，SHA Zhuo-liang，TANG Na (Tianjin Key Laboratory of “Marine chemistry and resources”, Tianjin University of Science & Technology, Tianjin 300450, China) Abstract: Fluid dynamics in evaporation chamber have strongly effects on the evaporation crystallization process. In this work, the fluid dynamics in two types of evaporation chamber were simulated by computational fluid dynamics method. The effect of the operation conditions on the fluid dynamics in evaporation chamber have been studied, and the effect of the fluid dynamics on the solid suspension and the evaporation crystallization process were analyzed. Key words: CFD simulation;two-phase flow;evaporation;crystallization 1(前言 作为一种传统技术，多效蒸发被广泛使用在制盐、制糖、以及许多化工产品 的生产中。多年来，在蒸发技术的研究中主要以增强过程的传热效率，提高系统 的热利用率为目标，以降低消耗和提高设备的利用率。然而，对蒸发器中的重要 部分-蒸发室很少有人注意到其结构、操作方式及其流体动力学状态对蒸发过程 的影响。另一方面，蒸发过程的最主要目的之一是去除溶剂，使溶液到达过饱和， 溶质以晶体的形式析出。因此蒸发室除具有汽液分离作用之外，更重要的是对晶 体产品的质量控制，晶体产品的粒度分布起着重要的作用。同时，如果产品粒度 控制不好，会影响产品的质量，引起过程消耗过高等问题。根据我国真空盐生产中广泛使用的强制循环蒸发器中的蒸发室结构，使用计算流体力学(CFD)的方法，探讨蒸发室内的流体动力学状态，从而分析其对固体悬浮状态的影响，为研究结晶过程奠定基础，同时分析流体动力学状态对汽液分离和传热效果的影响，为改善蒸发室结构和操作提供理论依据。 在最近的几年内，不同的作者针对蒸发结晶器设计以及结晶器中不同位置的 [1]固体密度分布进行了不同的研究。周全根据不同进料方式下垂直蒸发室断面速度分布，分析了结晶过程，并比较了不同进料方式、不同蒸发结晶结构NaCl晶体产品的粒径分布，结果认为逆循环轴向进料带育晶器的蒸发结晶装置能有效地 [2]生产不同粒度的NaCl晶体。魏宗胜讨论了蒸发器设计的技术关键，并根据切向进料方式下蒸发室不同高度横截面上固相浓度分布情况，结合轴向进料正、反循环时晶体粒度分布和短路温差损失，改进了蒸发结晶器的设计。然而，前者的研究都很难获得在整个容器中的有价值的固体悬浮密度分布值和流场分布信息。计算流体力学(CFD)，作为一种研究手段被广泛的应用于过程分析和设备结构 [3-6]的改造。CFD模拟可节约时间和成本，具有传统试验方法不可比拟的优势。其所预测的结果能直观地显示出三维空间内的流体信息和过程参数分布，这对于指导设备的设计和优化结晶操作具有非常重要的意义。 本研究使用ANSYS CFD 软件，模拟蒸发结晶器中的正循环径向进料和切向进料方式、不同循环速度下蒸发室的颗粒悬浮密度和流场情况。 2(模型 CFD使用有限体积法求解描述流体流动过程中的动量、热量和质量传递偏微分方程。在流动区域内，通过对流动空间的分割，把偏微分方程简化为有限的代数方程，通过适当的求解方法可得到流动区域的流场信息和不同过程参数的分布信息。为了解流场对固体颗粒分布的影响，在进行蒸发室流体动力学模拟中，使用多相流场模型。描述多相流场的主要方程可用下式示: ,()[()],,,,U,，,,,,,,,,,,,,,,,at, NN (1)pp,, Scmm()(),，,,,，,,,,,,,,,,,,,,,,,,11,,,, 方程中下标α、β和γ分别表示不同的相，相的数目由Np来表示。每个相的体积分数由φ来表示。是过程参数，可以是任何量。项描述αc(),,,,,,,, cc,c,0和β相间参数的相间传递通量。因而，。因此，所有相的相,,,,,,, ,, mm,,,,,,,间转移之和为0。项仅仅在相间发生质量传递时才会存在。,, 描述质量平衡的连续性方程为: ,()[]0,,,,U，,,, (2) ,,,,at, 描述流体运动的动量方程为: ,T()[((())],,,,,UUUUU，,,,,,，,,,,,,,,,,,aefft, Np (3)(L)ScUU(),，,,,,,,,,,,1, κ,ε湍流模型用于描述流体的湍流状态，在湍流状态下方程(3)中的粘度项可表示为: ,,,,， (4) effT,,,,, 根据κ,ε模型，湍流粘度为: 2,,,C,, (5) T,,,,,, 在ANSYS CFD中有许多描述流体湍流流动的模型，例如低雷诺准数κ,ε模型和κ,ε模型。我们使用κ,ε模型，由于它应用广泛，简单且适用于高雷诺准数的均质流体。在壁面上，设置速率为零，在边界附近，其速度分布用对数边 界层，由下式定义: 1，，,uEylog() (6) k 由于流体流动，相间的接触面会发生动量传递，主要是表面摩擦力和形体阻力。全部的阻力可根据无因次的阻力系数估计: D (7) ,CD12,UA2 ()L因此，动量传递系数为: C,, C3()LD,,,,,UUUU() (8) ,,,,,,C,,4L 对于在牛顿不可压缩流体中，运动的固体颗粒的阻力系数C，仅仅与雷诺数有D 关:，在这里μ是连续相的分子粘度。 Re/,,,ULa 3(模拟 3.1 强制蒸发器的模拟结构简化，及模拟网格 本工作的主要目的是模拟蒸发室的流体动力学状态，从而分析流体动力学状态对蒸发结晶过程的影响。 模拟采用我国真空盐生产中广泛使用的强制循环蒸发器中的蒸发室结构，简化为:圆柱部分高2.9m，直径2m;圆锥形底部与柱体 角;出料口直径为0.3m;进料口距顶部1.3m，直径为0.45m;径向进料结呈60º 构的网格单元数为161816，切向进料结构的网格单元为162477，如图1所示。 径向进料 切向进料 图1 蒸发器简化结构及模拟网格 3.2(模拟条件和模拟方法 ,32模拟基于两流体模型进行，液体被定义为连续相，溶液粘度为10N s/m， 3密度为1000kg/m，初始体积分数为0.95。固体颗粒被定义为分散相，分散相以颗粒的粒径为特征参数， 在不同的模拟过程中改变颗粒粒径，可改变分散相的流体动力学特征及其相关参数的分布。 在考察流体动力学状态对颗粒在蒸发室 3内分布的模拟中，密度为2300 kg/m，初始体积分数为0.05。模拟初始条件和进料状态完全一致，且悬浮密度为均匀分布。模拟区域以强制蒸发中循环蒸发器为基础，因只考察蒸发室内流体动力学状态对蒸发结晶过程的影响，其循环部分被简化为进口和出口。进料方式分别为径向进料和切向进料两种方式，蒸发器内的湍流状态使用κ,ε湍流模型， 如上所述。 4(模拟结果分析 4.2 径向进料模拟 使用上述的模拟基本定义，分别对进料速度为1.0m/s、1.5m/s和2.0m/s，分散相的粒径为100、300和500μm 时，蒸发室内的流体动力学状态，颗粒在蒸发室内的分布情况进行了模拟。图2 给出在同一颗粒尺寸、不同循环速度情况下，液相流速分布的模拟结果。由图可知，不同的进料速度对流场分布具有显著的影响。当进料速度为1.0m/s时，从进料口进入蒸发室的热流体，主要在进口位置以下形成循环，循环的液体很难达到蒸发表面，因而其气化效果较差。因此料液经加热管产生的温升在蒸发室内不能通过气化来降到该室压力下料液的平衡温度，偏高的温度将导致短路，造成温差损失，从而减小了传热有效温差，降低了蒸发器的生成能力。当进料速逐渐升高到1.5m/s时，流体可部分到达整个表面。由模拟结果可知，进料速度为2.0m/s时的流场分布情况比较理想。 100μm 1.0m/s 100μm 1.5m/s 100μm 2.0m/s 图2 径向进料时不同进料速度下的流场分布 图3,5为固体颗粒为100μm时不同进料速度下固体颗粒的体积分数分布情况。由图中可以看出，流场分布对固体颗粒的悬浮密度有较大的影响。虽然不同流速下进料口下部的悬浮情况都比较好。但由于在蒸发结晶器中，消除和产生过饱和度均在蒸发室内完成，而蒸发发生在表面，若沸腾区固相浓度低,沸腾时形成的过饱和度不能被足够的晶体表面提供的成长所消耗,蒸发表面的过饱和度会过高而引起局部初级成核，产生过量晶核，使产品粒度偏小，固液分离困难，造成干燥过程能量消耗高，产品质量的下降。比较100μm时不同进料速度的颗粒体积分数分布可知，在进料速度为2.0m/s时，固体颗粒充满整个结晶器，悬浮情况比较理想。 图3 颗粒尺寸100μm，进料速度1.0m/s 图4 颗粒尺寸100μm，进料速度1.5m/s 图5 颗粒尺寸100μm，进料速度2.0m/s 图6、图7分别为进料速度为2.0m/s时，尺寸为300μm和500μm的颗粒体积分布分数。由图可知，颗粒尺寸越大，流体对颗粒的作用力也就越小。虽然流速为2.0m/s时，液体流场分布已比较均匀，但当颗粒尺寸较大时，颗粒自身的重力作用就比较明显，大颗粒固体不能达到在蒸发表面， 因而不能起到很好的过饱合消除作用。从而，大颗粒仅参与循环而成长的机会较少。在循环过程中，会产生二次成核，而影响晶体产品粒度。因此，蒸发室的径向进料不适合结晶过程，对蒸发过程也不能达到理想状态。 图6 颗粒尺寸300μm，进料速度2.0m/s 图7 颗粒尺寸500μm，进料速度2.0m/s 4.3 切向进料模拟 图8 给出切向进料速度为1.0m/s和2.0m/s 时的液体在蒸发室内的运动轨迹。由图可知，切向进料时，由于离心作用使进入蒸发室的过热料液与蒸发室的料液迅速混合。当进料速度为1.0m/s时，进入蒸发室的过热料液并不能到达液面，其状态和对蒸发结晶过程的影响与径向进料相同。当进料速度达到2.0m/s时，从图中可以看出，部分液体会经过蒸发表面。 1.0m/s 2.0m/s 图8 切向进料时不同进料速度下的流场分布 100μm 2.0m/s 300μm 2.0m/s 500μm 2.0m/s 图9 不同颗粒尺寸对颗粒体积分数分布的影响 图9为不同颗粒尺寸在同一进料速度下的体积分数分布情况。总体来看，切向进料时颗粒多集中分布在结晶器器壁部分，尤其在锥形底部分更为显著。这是由于晶核受离心运动所产生的离心力，使固体颗粒积聚在罐壁部分，并下滑参与循环。 对于尺寸最小的100μm颗粒，其颗粒在蒸发器内分布较均匀。但当颗粒尺寸逐渐增大时，悬浮密度的分布不均匀性呈明显下降。颗粒尺寸为500μm时，颗粒不能在蒸发室内很好的悬浮，因而不能进一步增长。因此，就结晶过程而言， 切向进料使固液分离，影响颗粒在溶液中的悬浮状态，而不利于结晶过程。 5(结论 本文采用计算流体力学的方法对不同蒸发室结构的流体力学状态、及其对蒸发结晶过程的影响进行了分析。由模拟结果得知，径向进料的操作方式在较高的的进料速度下，对一定的颗粒粒径可达到较好的悬浮状态，但存在着因热短路问题而造成过热度不能完全消除的情况。切向进料，可使液体较好地混合，但是形成固液分离，使固体的悬浮状态不利于结晶过程。因此，对结晶过程的友好型结晶器，有待进一步研究。 参考文献: [1] 周全，粗粒盐的结晶环境及蒸发结晶器结构的讨论[J]. 海湖盐与化工,1999,29(1):14-18. [2] 魏宗胜.制盐蒸发器设计的技术关键及改进[J].化工设计,2000,10(1):31-3 4. [3] 张亚军.CFD技术在化工机械设计中的应用[J].贵州化工,2006,31(3):47-5 0. [4] 薛兆鹏,徐燕申,牛文铁.基于流体分析的工业结晶器搅拌桨结构优化[J]. 西 南交通大学学报,2003,38(5):501-504. [5] 程刚,孙会,潘家祯.用CFD计算双向组合桨的流场[J]. 设计与制造,2003, (3):14-16. [6] 周国忠,王英琛,施力田. 用CFD研究搅拌槽内的混合过程[J]. 化工学报,2 003,54(7):886-890. Hydrodynamics and Solid Suspension in Chamber of Force circulated evaporative crystallizer WANG Xuekui, WU Shou-xiang，SHA Zuo-liang， (Tianjin Key Laboratory of “Marine Chemistry and Resources”, Tianjin University of Science & Technology, Tianjin 300450, China) Abstract: Evaporation crystallization process strongly depends on the fluid hydrodynamics in evaporation chamber. The fluid hydrodynamics in two types of evaporation chamber were simulated by computational fluid hydrodynamics method in this work. The effect of the operation conditions on the fluid hydrodynamics in evaporation chamber has been studied. The effect of the fluid hydrodynamics on the solid suspension and its effect on evaporation crystallization process were analyzed. Key words: CFD simulation;two-phase flow;evaporation;crystallization The design of multi-effect solid suspension distribution. The evaporation process was mostly aimed Computational Fluid Hydrodynamics to improve the heat-transfer efficiency, (CFD) can simulate the multiphase flow to reduce the energy consumption and to field through the method of numerical increase the specific production rate of calculation. It is known as the powerful equipment. However, a few attention tool to simulate the crystallization was put on the structure of evaporation process. CFD method has been used to chamber, operation style and the analyze the crystallization process and influence of fluid hydrodynamics on improve the equipment structure by evaporation process. On the other hand, several authors [3-6]. The results of the functions of evaporation chamber do simulation have significant guiding for not only separate the steam from liquid, optimizing crystallization operation and but also provide the crystallization equipment design. environment. In this work, commercial software Recently, several authors have CFD of ANSYS 10 was employed to studied the solid volume fraction simulate the evaporation crystallization distribution in different position of process in different structures of crystallizer. Zhou Quan[1] had analyzed evaporation chamber. The fluid the crystallization process according to hydrodynamics and crystal suspension vertical velocity distribution of distribution in different feeding mode evaporation chamber in different feeding were analyzed. modes. Results show that the crystallizer 1 MODELING with reverse circulation and axis feeding can product large size crystals Using Eulerian-Eulerian approach, effectively. Wei Zong-sheng[2] had the transport equations of all parameters discussed the key technique of in multiphase flow system can be crystallizer design, and improved the expressed as follows: structure of evaporation crystallizer. However, all studies were not able to obtain the information of flow field and ,()[()]rU,,,,，,,,,,,,,,,,,,,,at, NNpp,, rScmm()(),，,,,，,,,,,,,,,,,,,,,,,11,,,, (1) Phases are labeled by subscripts α, β =c. Hence, the sum over all the cαββα and γ, and the number of phases is phases of all the inter-phase terms is ,,denoted by Np. The volume fraction of mm,,,,,,,,,zero. The term only each phase is denoted by r, and. The arises if inter-phase mass transfer takes term c(Φ-Φ) describes the αββαplace. inter-phase transfer of variable Φ The continuity equation of the between phases α and β. Thus, c=0 and αα phase α is expressed by: ,()[]rrUS,,，,,,,,,,,at, (2) The momentum equation for the phase α is: ,T()[((()))]rUrUUUU,,,，,,,,,，,,,,,,,,,,,aefft, Np(d)rPcUUrS(),,,，,，,,,,,,,,,1, (3) In the multiphase flow field, continuous phase. Only drag force is interaction between the phases is the considered in this model. The total drag main effect of the flow field of different force is most conveniently expressed in phases. There are several interface terms of the dimensionless drag forces between dispersed phase and coefficient: D,CD12,(,)UUA,,,2 (4) ()dC,,For spherical particle, the momentum transfer coefficient may be express as: C3()LD,,,,UU,,,,C,,4d (5) For Newtonian incompressible fluid, resistance coefficient, CD, depends only on Reynolds number: 24/Re(Re1) ,C,,D5,,，)0.44(1000Re1-210, (6) Where, the Reynolds number was defined by: ,UUd,a,,,Re, ,, (7) Mostly, the flow in industrial employed because it was widely used in crystallizer is turbulence. Several industrial process simulation. According models can be used to describe turbulent to the standard κ-ε model, the turbulent flow, for instance, the κ-ε model, k-ω viscosity is defined as: model and the Re-stress model. In this work, the standard κ-ε model was 2,,,C,,T,,,,,, (8) 2 SIMULATION Gridding 2.1 Configuration and Simulation The aim of this work is to simulate the fluid hydrodynamics in evaporation adopted in CFD simulation. The chamber and to analyze the flow field dimensions and the mesh of the and its influence on evaporation crystallizer are listed in Table 1 and in crystallization process. The structure of Fig.1. evaporation chamber in forced circulation evaporator which was Table 1 Geometry of evaporation chamber cylinder cylinder conical inlet inlet distance outlet diameter height bottom joint diameter /m to top /m diameter /m /m /m angle 2 2.9 0.45 1.3 0.3 60? (a) radial feeding (b) tangential feeding Fig.1 The geometry and meshes of evaporation chamber 2.2 Simulation Condition and conditions in simulation were set as Simulation Methodsame as feed. The simulation was based on the assumption of full suspension. Because of the aim of simulation is to The simulation is based on the analyze the influence of fluid two-fluid model mentioned above. hydrodynamics on evaporation Water was defined as continuity phase, crystallization process in evaporation -3 2the viscosity is 10N s/m, the density chamber, circulation was simplified as 3is 1000kg/m, and the initial volume inlet and outlet. Radial and tangential fraction is 0.95. The particle was feeding was adopted in structure of considered as dispersed phases, the evaporation chamber. characteristic parameter is particle size. Fluid hydrodynamics of dispersed phases and correlating parameters in 3 RESULTS AND DISCUSSION evaporation chamber will change with the change of particle size. The density of the crystal is 1980 kg/m3 and the 3.1 Radial Feeding initial volume fraction is 0.05. Initial The fluid hydrodynamics of evaporated well so that the high different phases in the crystallizer are temperature liquid goes through the the external condition for crystallization. chamber and results in temperature loss The simulated fluid hydrodynamics of in heat transfer process. The production liquid phases with different feeding capacity of evaporation crystallizer will velocities are given in Fig. 2. It is shown be decreased. The liquid with high that the feeding velocity had strongly temperature can reach the surface of influence on flow field. The circulation evaporation partly when feeding of fluid with high temperature was velocity was 1.5m/s. The simulation located below of inlet, when the feeding result shown that the flow field velocity was 1.0m/s. In this situation, the distribution was ideal when feeding fluid with fed temperature is difficult to velocity was 2.0m/s. reach the surface of evaporation. As a result, the hot fluid can not be (a)1.0m/s (b)1.5m/s (c)2.0m/s Fig.2 The liquid velocity distribution at different feeding velocity (0.1mm) surface should be enough to consume The simulated volume fraction the supersaturation in the surface area, distributions of 100μm crystal at otherwise, the primary nucleation will different feeding velocity are given in occur and bring excessive thin crystal. Fig. 3. It is shown that the flow field This will result in more thin crystal in distribution strongly influences on final product and more energy consume crystal suspension. Although all crystal in dryness process. The simulation result can be suspended well below of inlet of shown that the crystal suspension the chamber, the elimination and condition was ideal when feeding produce of supersaturation should be velocity was 2.0 m/s for crystal size of occurred in whole chamber. The highest 100μm suspended well. supersaturation is in the surface of evaporation in the chamber, crystal (a) feeding velocity 1.0m/s (b) feeding velocity 1.5m/s (c) feeding velocity 2.0m/s Fig.3 The volume fraction distribution of crystal at different feeding velocity (crystal size 0.1mm) The simulated volume fraction crystals with small size have uniformed distributions of the crystals with sizes of distribution below of inlet in evaporation 300 μm and 500 μm are shown in Fig. 4 chamber. However, the volume fractions when feeding velocity is 2 m/s. It clearly of crystal with sizes of 300 μm and 500 shows that the volume fraction of the μm have low value in the upper region of evaporation chamber. There is nearly the evaporation chamber with radial no crystal in the region of evaporation. It feeding is not suitable for crystallization means that the superstaturation can’t be process. consumed well in those areas. Therefore (a) crystal size 0.3mm (b) crystal size 0.5mm Fig.4 The volume fraction distribution of different crystal size(feeding velocity 2.0m/s) 3.2 Tangential Feedingevaporation chamber with tangential feeding. When feeding velocity was 1.0m/s, the fed fluid couldn’t reach on The simulated moving contrails of surface of liquid as same as radial 100μm crystal at different feeding feeding. The simulation results shown velocities are given in Fig. 5. It is shown that some of the fed fluid would reach that the fed fluid with high temperature on the fluid surface at 2.0m/s feeding could be rapidly mixed with the fluid in velocity. (a) feeding velocity 1.0m/s (b) feeding velocity 2.0m/s Fig.5 The liquid velocity distribution at different feeding velocity (a) crystal size 0.1mm (b) crystal size 0.3mm (c) crystal size 0.5mm Fig.6 The volume fraction distribution of different crystal size (feeding velocity 2.0m/s) The simulated volume fraction distribution of different sizes crystals at References: same feeding velocity is given in Fig. 6. [1] Zhou Quan. Discussion on the It is shown that crystals are concentrated Crystallization Environment and on nearly the crystallizer wall, especially in the Structure Evaporative conical bottom. This is because of the Crystallizer for Coarse Salt. centrifugal force of the crystals. The [J]SEA-LAKE SALT & crystal suspension was not well CHEMICAL INDUSTRY, 1999, 29 considering the crystallization process.(1): 14-18. [2] Wei Zong-sheng. Bottleneck of Salt For the smallest size, 100μm, Evaporator Design and Technical crystals were well suspended. While Improvement [J]. CHEMICAL crystal size increases gradually, the ENGINEERING DESIGN, 2000, homogeneity of the crystal suspension is 10(1): 31-34. decreased correspondingly. Crystals [3] Zhang Ya-jun. Application of CFD could not be well suspended when on Designing Chemical crystal size was 500μm. As a result, Machinery[J]. GUIZHOU crystals were separated from liquid with CHEMICAL INDUSTRY, tangential feeding. Therefore, the 2006,31(3): 47-50. tangential feeding is not suitable for the evaporation crystallization operation. [4] Xue Zhao-peng, Xu Yan-shen, Niu Wen-tie. Optimized Design of the Industrial Crystal Propeller's 4 CONCLUSIONSStructure Based on CFD [J]. PRECISE MANUFACTURING & AUTOMATION, 2003, 38(5): In this work, the fluid 501-504. hydrodynamics was simulated in [5] Cheng Gang, Sun Hui, Pan Jia-zhen. evaporation chamber with different Calculation the Flow Field of Bidirectional Combined Oar by CFD feeding style, and its influence on [J]. CHEMICAL EQUIPMENT & crystallization process was analyzed. ANTICORROSION,2003,(3):14-16. Results show that the small size crystal [6] Zhou Guo-zhong, Wang Ying-chen, could suspend well with radial feeding Shi Li-tian. CFD Study OF Mixing style in higher feeding velocity. The Process in Stirred Tank [J]. problem is short-cut-circuiting of the JOURNAL OF CHEMICAL high temperature fluid. Tangential INDUSTRY AND ENGINEERING (CHINA), 2003, 54(7): 886-890.feeding style is not suitable operation for crystallization process because of Acknowledgement: crystals were separated from liquid. The The authors thank to the financial better structure of support of Doctoral Programs crystallization-friendly crystallizer need Foundation of Ministry of Education of to be further studied.China (No.20070057001).