计算流体动力学和动态耦合热力学毕业论文中英文资料外文翻译文献

计算流体动力学和动态耦合热力学软件
在顶吹转炉中的应用
Mikael ERSSON, Lars HÖGLUND, Anders TILLIANDER,
Lage JONSSON and Pär JÖNSSON
应用程序冶金部，皇家技术学院（KTH）
SE-10044，瑞典首都斯德哥尔摩。

（2007年11月8日收到，2007年12月10日接受）
一种新的建模方法被提出，这种建模方法是使用计算流体力学软件相连结热力学数据库以获取动态模拟冶金过程的现象。

这种建模方法已被应用在一个基本的氧气顶吹转炉模型。

通过各种气体之间的反应研究。

结果表明,大量的表面气体的流通是完全受对流控制的。

此外，在这个过程中大量产生的CO脱碳可能会放慢从浴缸喷出的液滴的脱碳率。

在目前的模拟反映实验室的实验条件下，这点也被证实在这个过程中所产生的炉渣（FeO和/或SiO2）接近于零，即只产生的气体（二氧化碳CO2）就好比是氧气射流击中钢液。

它也说明如何从几秒钟的采样推算脱碳速度，只要是含碳量足够的高可以在后期的时候做含碳量的模拟，从而得到的碳含量的粗略估计。

总的结论是，通过Thermo-Calc的数据库和CFD软件的动态耦合来达到冶金动态模拟是有可能的。

关键词：转炉;计算流体力学，热力学建模;炉渣和动态模拟。

介绍
在许多涉及氧气喷射撞击到钢液面的冶金过程中，为了优化涉及动力学的部分，如脱碳，底层流体动力学是需要的。

现在对于这个问题已经几个实验报告和一些数值或计算流体动力学（CFD）的报告。

Szekely and Asai已经介绍了一种将液体的冲击射流表面的计算模型。

Ngyen and Evans通过使用这种方法计算溶池喷嘴直径比液体表面所造成变形的冲击射流的影响，张等人模仿了一种同时使用顶吹和底吹复合吹炼的情况。

Odenthal等人展示了一种顶吹转炉
多相CFD模型，在这种顶吹转炉中由于冲击射流以及底部和顶部转换混合时存在飞溅现象。

Nakazono等人描述了铁液表面的超音速氧气喷射冲击时的含碳量的两阶段的数值分析。

通过计算表明在真空和表面处理条件下天然气和钢铁之间的变化。

该模型采用稳态方法无飞溅等。

在其他文献中还其他非顶吹CFD模型介绍，琼森等。

提出了硫精炼耦合计算流体力学和热力学模型
热力学是成立的在CFD程序中作为一个自定义的子程序，尤其是专门为调查系统所写的子程序。

图中1可以看到这样的做法一个示意图。

图1。

合并计算流体力学和热力学建模与数值模型方法的示意图。

由文献（14）
最近包含顶吹系统CFD模型已经出现并和实验数据进行比较，这里提出，这个模型是包括气/液/渣的流体体积（VOF）等反应扩展的一个多相流模型，这个模型适用于顶吹系统。

这个模型以及各个阶段的扩展已经获得批准使用这一方法所建的装备使得终于有一个可以说明基本的顶吹转炉模型结果。

2。

数值模型
目前的建模方法如图2所示。

图2。

合并数值模型方法的示意图
为了解决新的研究可以轻易的合并到这个模型的问题建立了以模块化的方式，这也意味着，当从一个系统变到另一个系统是只需要一个很小的重编程程序，只要是热力学数据在目前的热力学数据库中是存在的以及不超过CFD软件的运算能力。

这个模拟过程已经在包含6个节点的集群的Linux PC上执行。

现实生活中的模拟时间是和运算系统的数量相关的，所以没有固定和典型的模拟时间;它可能会在一小时和10天的实时变化亦或是10秒的时间。

在图3中可以看出其物理域和数值域的原理。

图3。

顶吹转炉示意图。

a）物理域b）数值域。

所涉及的边界条件为入口速度，出口压力，无滑墙壁和对称轴。

所有的墙壁都用标准壁面函数表示，合理的k–e模型已使用在所有的例子中。

域的宽度为0.075米，高度是0.13米。

最开始，1500°C的15.6公斤距顶枪0.01米以上的钢熔液用来试验，大量的气体从入口吹入相当于25升/分钟的纯体积流量氧气。

用于模拟的表1中的各种参数可以看出和表2中的初始浓度的不同。

表1.不同阶段密度和扩散系数
表2.初始浓度
2.1。

差价
ANSYS软件同时也被使用，这是一个为计算求解流体体积、质量的有限元分析软件，
这样动量守恒方程中所涉及的质量、体积都得到了解决。

根据湍流模型同时使用了一些额外的守恒方程作为补充,例如湍流动能守恒,k,和动荡能量耗散,e,都是用标准的k-e 模型，下面的公式用来计算任何形式的
这里r是密度,u是指速度矢量,当基于雷诺兹平均使用湍流模型,G是扩散系数,正如从表3中看到的一样方程(1)和表3描述了质量、动量、湍流动能,动荡能量耗散、能量、物质和体积分数。

表3守恒方程参数
2.2热计算
为了获得准确描述了热力学特征，软件Thermo-Calc被使用使用。

这是一个通用的软件方案的多组分的相平衡计算。

它使用一种技术,它允许非常灵活的设置条件的平衡状态,从而适用过程模拟。

问题的解决方法是以下。

第一,质量和热量含量分别计算每个阶段。

然后,总质量和热量的内容是总结。

该系统是氢原子核装备中的。

最后,程序将计算各阶段温度,新的成分和含量。

能与热力学的软件应用程序编程接口使用TQ操作。

这个接口是一个接口
Thermo-Calc中可用的软件包,使其能要推广实施不同系统(例如组件和元素,该系统由),而无需更改代码。

2.3. 耦合和假设
其目在于CFD-package和Thermo-Calc数据库软件两个软件的耦合,是创建一个通用的计算模型,包括化学反应冶金系统。

在下面的文本描述的耦合。

主要的假设是局部均衡、每个时间步可以到达每个计算单元的过程中,。

软件之间的接口CFD-package和热力学数据库分别被编码在C和FORTRAN。

2.4．多相考虑
所有的散装阶段的建模为不可压缩;有统一的密度,见表1。

与一个精化的模型应该有可能使用一个理想气体定律假设对气态和一些温度对钢密度模型的依赖。

简要地说明一些细节的功能假设一摩尔的氧和碳反应形成的融化2摩尔的一氧化碳。

在计算单元在考虑将会有一个扩张的气相到约两倍于原体积。

这反过来将很可能意味着气体向外扩张的计算单元。

如果再加上CFD,扩张将发生在以下步骤中,作为额外的气体质量被添加到细胞源项。

经过计算细胞已经达到平衡它也将会有一个特定的温度平衡温度。

3.结论
3.1.基本顶级吹转炉
在图5显示气体射流向量的情况,正如水落在钢铁溶液中所显示的那样。

这是看到射流失去其轴向冲击就好像其滴落在刚溶液表面。

从射流中来的低流量给出了一个相对较小的渗透率在钢铁溶液中。

图6说明了流场的钢铁浴引起碰撞射流。

目前的最高速度的射流冲击的钢液面和面积。

流体措施阻止了渗透区向墙壁,然后进入到钢液中。

一个大型的循环,循环中形成的冲击流(即轴和墙之间)和几个小的循环回路。

应该指出的是,在更大的循环级速度是非常小的。

这是由于低流量来自高层的射击流。

更高的流率呈现较大的渗透和较强的循环。

图5.在气体速度矢量。

矢量有固定长度;他们是由色彩反映速度级(m / s)。

图6.速度向量在钢液中的变现，矢量有固定长度;他们速度级有色彩反映
图7展示了碳在1、2、3和5s时在钢液里的浓度。

首先气体的射流量应该注意,只有在相当狭窄的范围内才能看到的结果。

这是故意为了使少量浓度的差异在一定范围内可见。

同时可以看到图7(a),碳浓度
梯度存在于一个小区域接近自由面面积以及在一个更大的地区的右边的渗透区当仔细观察图7时。

7(a)-7(d),它的变化就会更明显,规模更大的向炉壁发展混合,且随着时间变得更大。

从图之间可以看到碳浓度在不断地减小。

从7(a)-7(d)看来,钢液实在表面混合运动(参见图6),随后由于循环模式出现在中下部出现,而且呈现紊流的混合。

碳浓度高的钢的深度只有几毫米的渗透区;然而真正的大梯度出现在一个更薄的地区靠近自由表面。

渐渐地,随着流场的发展,再循环趋向炉壁。

图7a.钢中碳的质量分数.1 s模拟时间
图7 b.钢中碳的质量分数.2 s模拟时间
图7 c.钢中碳的质量分数.3s模拟时间
图7 d.钢中碳的质量分数.5s模拟时间
图8.CO的质量分数5s模拟时间。

图8展示了在气相中CO气体浓度呈现的规律。

射流所覆盖的钢液的表面几乎完全充满氧气所以几乎没有CO的存在,除了一个薄层旁边的钢液表面。

CO气体数量则会变得更明显的逐渐减少距离炉壁附近。

4.结论
一个新的建模方法,提出了采用CFD软件已经耦合热力学数据库(Thermo-Calc)使用自定义子例程获得动态模拟冶金过程的现象。

gas-steel之间的反应,gas-slag,steel-slag 和gas-steel-slag一直被认为在一个基本的模型顶吹转炉。

最后的结论是,它可能是一个动态的耦合Thermo-Calc数据库和CFD软件进行动态模拟的冶金过程如顶吹转炉。

具体的结论来自高层的吹转炉模拟包括:
(1)紊流扩散的气体中不可忽视的要考虑射流冲击面积的影响。

(2)大量一氧化碳脱碳期间可能会减慢钢液中钢的脱碳速度。

(3)可以使用脱碳速度的推断,对一个几秒钟的仿真,得到的粗略估计碳含量在随后阶段过程中只要碳含量相对较高(比较下一点)。

(4)对当前系统来说，大约3%碳在钢的初始的渣中。

FeO或二氧化硅了接近于零即只要气体(COCO2)在钢液中的含量足够的多。

找出浓度的结合,流动速率和温度下,呈现最高效脱碳,未来会有更进一步的参数研究。

确认
这个工作是由瑞典财政支持战略研究基金会(SSF)和瑞典钢铁行业通过热力学计算中心(CCT)进行的。

参考文献
1) E. T. Turkdogan: Chem. Eng. Sci., 21 (1966), 1133.
2) N. A. Molloy: J. Iron Steel Inst., (1970), Oct., 943.
3) T. Kumagai and M. Iguchi: ISIJ Int., 41 (2001), S52.
4) A. Nordquist, N. Kumbhat, L. Jonsson and P. Jönsson: Steel Res. Int., 2 (2006), 82.
5) B. Banks and D. V. Chandrasekhara: J. Fluid Mechanics, 15 (1963), 13.
6) A. Chatterjee and A. V. Bradshaw: The Interaction Between Gas Jets and Liquids,
Including Molten Metals, 314.
7) M. Ersson, A. Tilliander, M. Iguchi, L. Jonsson and P. Jönsson: ISIJ int., 46 (2006),
No. 8, 1137.
8) J. Szekely and S. Asai: Metall. Trans, 5 (1974), 464.
9) A. Nguyen and G. Evans: 3rd Int. Conf. on CFD in the Minerals andProcess
Industries CSIRO, Melbourne, Australia, (2003), 71.
10) J.-Y. Zhang, S.-S. Du, S.-K. Wei: Ironmaking Steelmaking, 12 (1985), 249.
11) H.-J. Odenthal, U. Falkenreck and J. Schlüter: European Conf. onComputational
Fluid Dynamics, the Netherlands, (2006).
12) D. Nakazono, K.-I. Abe, M. Nishida and K. Kurita: ISIJ Int., 44 (2004), 91.
13) M. Ersson, A. Tilliander, and P. Jönsson: Proc. Sohn Int. Symp Advanced
Processing of Metals and Materials, ed. by F. Kongoli and R. G. Reddy, TMS, San diego, USA, Aug 27–31, (2006), p. 271.
14) L. Jonsson, D. Sichen and P. Jönsson: ISIJ Int., 38 (1998), 260.
15) L. Jonsson: PhD Thesis, Dept. of Metallurgy, KTH, Sweden, (1998).
16) L. Jonsson, P. Jönsson, S. Seetharaman and D. Sichen: Proc. of 6th Japan–Nordic
Countries Steel Symp., ISIJ, Tokyo, (2000), 77.
17) C. W. Hirt and B. D. Nichols: Comput. Physics, 39 (1981), 201.
18) B. E. Launder and D. B. Spalding: Comp. Meth. Appl. Mech. Eng., 3 (1974), 269.
19) Fluen t User’s Manual, (2007).
20) T.-H. Shih, W. W. Liou, A. Shabbir, Z. Yang and J. Zhu: Computers Fluids, 24 (1995), 227.
21) J.-O. Andersson, T. Helander, L. Höglund, P. Shi, and B. Sundman: Calphad, 26
(2002), 273.
22) A. Nordquist, A. Tilliander and P. Jönsson: Proc. 5th European Oxygen
Steelmaking conf., Aachen, Germany, (2006), 519.
Dynamic Coupling of Computational Fluid Dynamics and Thermodynamics Software: Applied on a Top Blown Converter Mikael ERSSON, Lars HÖGLUND, Anders TILLIANDER, Lage JONSSON and Pär
JÖNSSON
Division of Applied Process Metallurgy, Royal Institute of Technology (KTH), SE-100 44 Stockholm, Sweden.(Received on November 8, 2007; accepted on December 10, 2007)A novel modeling approach is presented where a computational fluid dynamics software is coupled to thermodynamic databases to obtain dynamic simulations of metallurgical process phenomena. The modeling approach has been used on a fundamental model of a top-blown converter. Reactions between gas–steel, gas–slag, steel–slag and gas–steel–slag have been considered. The results show that the mass transport in the surface area is totally controlled by convection. Also, that a large amount of CO produced during the decarburization might slow down the rate of decarburization in droplets ejected from the bath. For the present simulation conditions reflecting laboratory experiments, it was also seen that the amount of slag (FeO
and/or SiO2) created is close to zero, i.e. only gas (CO_CO2) is created as the oxygen jet hits the steel bath. It was also illustrated how an extrapolation of the decarburization rate, sampled from a few seconds of simulation, could be done to get a rough estimate of the carbon content at a later stage in the process as long as the carbon content is relatively high. The overall conclusion is that it is possible to make a dynamic coupling of the
Thermo-Calc databases and a CFD software to make dynamic simulations of metallurgical processes such as a top-blown converter.
KEY WORDS: BOF; CFD; thermodynamics; modeling; slag and dynamic simulations.
1. Introduction
In many metallurgical processes involving an oxygen-jet impinging onto a steel bath surface, a good understanding of the underlying fluid dynamics is desirable in order to optimize the involved kinetics such as decarburization. There have been several experimental reports on the subject for instance1–7) and also some numerical or Computational Fluid Dynamics (CFD) reports.8–13) Szekely and Asai8) presented a computational model of a jet impinging onto a liquid surface. Ngyen and Evans investigated the effect the nozzle-to-pool diameter ratio had on the deformation of the liquid surface caused by an impinging jet, using a computational model.9) Zhang et al. modeled a combined blown case where a top jet as well as a submerged jet was employed. 10) Odenthal et al. showed a multiphase CFD model of a top blown converter where splashing phenomena due to the impinging jet was investigated as well as the mixing time in the converter due to bottom and top blowing.11) Nakazono et al. described a two-phase numerical analysis of a supersonic O2-jet impinging on a liquid iron surface containing carbon.12) The calculations were performed under vacuum and addressed surface chemistry between the gas- and the steel-phase. The model used a steady state approach without treatment of splashing, ripples etc. There are also other non top blowing CFD models presented in the literature that address chemical reactions in metallurgical systems, see for instance.14,15) Jonsson et al. presented a coupled CFD and thermodynamics model of sulfur refining in a gas-stirred ladle.14) The thermodynamics was incorporated in the CFD program as a custom subroutine specifically written for the investigated system. A schematic of such an approach can be seen in Fig. 1. Recently a CFD model consisting of a top blown system has been presented and compared to experimental data.13) Here an extension to this model is presented which includes reactions i.e. a gas/liquid/slag Volume of Fluid (VOF)17) multiphase model, for a top blown system. Reactions between all phases
have been allowed as well as expansion/ contraction associated with the creation or destruction of phases in the computational cells. The methodology of the setup is shown and finally some illustrative results of a fundamental top blown converter model are presented.
2. Numerical Model
The current modeling approach is seen in Fig. 2. It is built in a modular fashion in order to ease the incorporation of new research into the model. This also means that very little reprogramming is necessary when changing from one system to another, as long as the thermodynamic data is present in the thermodynamic database and the capabilities of the CFD software is not exceeded. The simulations have been performed on a Linux PC cluster containing 6 nodes. The real-life simulation time has been highly dependent on the number of Thermo-Calc calls performed so no typical simulation time can be given; it varies between one hour and ten days real-time for a 10 s
Fig. 1. Schematic of a numerical model approach with combined CFD and thermodynamics
modeling. From Ref. 14).
Fig. 2. Schematic of a numerical model approach with combined CFD and thermodynamics modeling. Modular approach that uses databases from the Thermo-Calc software.
Fig. 3. Schematic of top blown converter. a) Physical Domain b) Numerical Domain.
simulation. In Fig. 3 a schematic of the physical and numerical domains can be seen. The boundary conditions used are velocity inlet, pressure outlet, no-slip walls and symmetry axis. Standard wall functions have been used for both walls.18,19) The realizable k–e model20) has been used in all examples. The domain width is 0.075 m and the height is 0.13 m. Initially, a 15.6 kg steel-melt is introduced with a temperature of 1 500°C. From the top lance, placed 0.01 m above the steel, a mass flow inlet was specified corresponding to a volumetric flow rate of 25 L/min of pure oxygen. In Table 1 various parameters used in the simulation can be seen and in Table 2 the initial concentrations of the different species are shown.
2.1. CFD
The Ansys Fluent software has been used which is a commercial finite volume solver used for computational fluid dynamics.19) Conservation equations of mass, momentum, energy and species are solved. Depending on the turbulence model used some extra conservation equations are added, for instance conservation of turbulence kinetic energy, k, and turbulence energy dissipation, e , as prescribed in the standard k–e model.18) The following form is used for transport of any property f :
where r is the density, u is the mean velocity vector, when using a turbulence model based on Reynolds Averaging, G is the diffusion coefficient and Sf is the source term, as can be seen in Table 3. Equation (1) and Table 3 describes the transport of; mass, momentum, turbulence kinetic energy,
turbulence energy dissipation, energy, species and volumefraction.
2.2. Thermo-Calc
In order to obtain an accurate description of the thermodynamics the software Thermo-Calc21) is used. This is a general software package for multi-component phase equilibrium calculations. It uses a technique that allows for a very flexible setting of conditions for the equilibrium state thus being suitable for use with process simulations. The method of solution is the following. First, the mass
and heat content in each phase is calculated separately. Then, the total mass and heat content is summed up. The system is thereafter equilibrated. Finally, the program calculates the temperature, new compositions and amounts of the phases. To communicate with the thermodynamic software an
application programming interface TQ21) is used. This interface is one of the interfaces available within the Thermo- Calc software package21) and makes it possible to generalize the implementation of different system (e.g. the components and elements that the system consists of) without changing the code.
2.3. Coupling and Assumptions
The aim with the coupling of the two softwares, the CFD-package and the
Thermo-Calc database software, is to create a general numerical model for metallurgical systems including chemical reactions. In the following text the coupling will be described. The major assumption is that local
equilibrium can be reached in each computational cell during the course of each time step. The software interface between the CFD-package and the thermodynamic databases is coded in C and FORTRAN, respectively.
2.4. Multiphase Considerations
All bulk phases have been modeled as incompressible; having uniform density, see Table 1. With a refinement of the model it should be possible to use an ideal gas law as- sumption for the gas phase and some temperature and composition dependent density model for the steel and the slag phases. To briefly explain some specifics of the functionality
assume that one mole of oxygen reacts with carbon in the melt to form 2 mole of carbon monoxide. In the computational cell under consideration there will be an expansion of the gas phase to roughly twice the original volume. This in turn will most likely mean that the gas expands outside
the computational cell. When coupled with CFD, the expansion will occur in the following time step, as the extra gas mass is added to the cell as source term. After a computational cell has reached equilibrium it will also have a specific temperature–the equilibrium temperature. Assumptions:
a) Thermodynamic equilibrium can be reached in each cell during any time step.
b) The densities of gas/steel/slag are constant in time and space.
c) Equilibrium needs only to be calculated in cells containing at least two phases. A schematic of the coupling and the solution procedure can be seen in Fig. 4.
Fig. 5. Velocity vector plot in gas. Vectors have a fixed length; instead they are colored by
velocity magnitude (m/s).
Fig. 6. Velocity vector plot in the steel. Vectors have a fixedlength; instead they are colored by
velocity magnitude(m/s).
Fig. 7a. Mass fraction of carbon in steel. 1 s simulation time. Fig. 7b. Mass fraction of carbon in steel. 2 s simulation time.
Fig. 7c. Mass fraction of carbon in steel. 3 s simulation time. Fig. 7d. Mass fraction of carbon in steel. 5 s simulation time.
Fig. 8. Mass fraction of CO in the gas phase. 5 s simulation time.
3. Results
3.1. Fundamental Top Blown Converter
In Fig. 5 vector plot showing the gas jet, as it impinges on the steel bath, is shown. It is seen that the jet looses its axial momentum as it hits the bath surface. The low flow rate from the lance gives a relatively small penetration in the steel bath. Figure 6 illustrates the flow field in the steel bath caused by the impinging oxygen jet. The highest velocities are present where the jet hits the bath and in the surface area. The fluid moves from the penetration zone towards the wall and then down into the bath. A large re-circulation loop is formed in the center of the bath (i.e. between the axis and the wall) and several small re-circulation loops are formed close to the surface. It should be noted that the velocity magnitude in the larger loop is very small. This is attributed to the low flow rate from the top lance. Higher flow rates render larger penetration and a stronger bath circulation.
Figure 7 illustrates the carbon concentration in the bath at times 1, 2, 3 and 5 s. First of all the plotting limits should be noted, where it can be seen that
the range of the plot is quite narrow. This is intentional in order to make the small concentration differences visible over the range of plotting
colors used. It can be seen from Fig. 7(a) that a carbon concentration gradient exists in a small region close to the free surface area as well as in a larger region to the right of the penetration zone. When examining Figs.
7(a)–7(d) it becomes evident that the larger region moves towards the wall of the converter and that it also becomes larger with time. The carbon concentration in the moving region slowly decreases between Figs.
7(a)–7(d). It seems that the decarburized steel is transported along the surface (see Fig. 6) and subsequently dragged down into the bath due to the recirculation
pattern present in the flow and because of turbulent
mixing. The depth of the decarburized steel is only a few millimeters in the penetration zone; however the really large gradients appear in an even thinner region close to the free surface. Gradually, as the flow field develops, the recirculation zone moves towards the wall. Figure 8 illustrates the CO gas concentration present in the gas phase. The region where the jet impinges on the steel surface is almost exclusively filled with O2 so no CO is present there, except for a thin layer right next to the bath surface. The CO gas amount then becomes gradually more pronounced as the distance to the wall decreases.
5. Conclusions
A new modeling approach has been presented where a CFD software has been coupled to a thermodynamic database (Thermo-Calc) using custom subroutines to obtain possible to make a dynamic coupling of the Thermo-Calc databases and the CFD software to make dynamic simulations of metallurgical processes such as a top-blown converter. Specific conclusions from the top blown converter simulations include:
(1) Turbulent diffusion of species can not be neglected when considering the species transport in the surface area.
(2) The large amount of CO produced during the decarburization might slow down the rate of decarburization in droplets ejected from the bath.
(3) It is possible to use extrapolation of the decarburization rate, sampled from a few seconds of simulation, to get a rough estimate of the carbon content at a later stage in the process as long as the carbon content is relatively high (compare next point).
(4) For the current system, concentrations of about 3 mass% carbon in the steel yields no initial amount of slag. FeO and/or SiO2 created are close to zero i.e. only gas (CO_CO2) is created as the oxygen jet hits the steel bath. To find out the combination of concentrations, flow rates and temperatures that renders the most efficient decarburization, a future parametric study would be of interest.
Acknowledgements
This work was financially supported by the Swedish Foundation for Strategic Research (SSF) and the Swedish steel industry through the Centre for Computational Thermodynamics (CCT).
Nomenclature
D0 : Molecular mass diffusivity [m2 s_1]
Dt: Turbulent mass diffusivity [m2 s_1]
e: Turbulence dissipation [m2 s_3]
fC : Mass fraction carbon in steel
f˜
C: Mass-weighted average mass fraction carbon in
steel
g : Gravitational acceleration [m s_2]
k: Turbulence kinetic energy [m2 s_2]
ki
0 : Diffusion coefficient of species i [kgm_1 s_1]
kc
eff: Effective thermal conductivity [Wm_1K_1]
m : Molecular viscosity [kgm_1 s_1]
n : Kinematic viscosity [m2 s_1]
m˙ : Mass rate of change [kg s_1]
Ni : Number of phases
Nk : Number of scalars in the phase k
f : General transported property
f i
k : Scalar k in phase i
P : Pressure [Nm_2]
Pe : Cell Peclét number
Prt: Turbulent Prandtl number
r : Density [kgm_3]
Sf : Source term for general property f
Sct: Turbulent Scmidt number (here assumed_0.7)
ur : Radial velocity component [m s_1]
uz : Axial velocity component [m s_1]
G : Diffusion coefficient for general property f
REFERENCES
1) E. T. Turkdogan: Chem. Eng. Sci., 21 (1966), 1133.
2) N. A. Molloy: J. Iron Steel Inst., (1970), Oct., 943.
3) T. Kumagai and M. Iguchi: ISIJ Int., 41 (2001), S52.
4) A. Nordquist, N. Kumbhat, L. Jonsson and P. Jönsson: Steel Res. Int., 2 (2006), 82.
5) B. Banks and D. V. Chandrasekhara: J. Fluid Mechanics, 15 (1963), 13.
6) A. Chatterjee and A. V. Bradshaw: The Interaction Between Gas Jets and Liquids, Including Molten Metals, 314.
7) M. Ersson, A. Tilliander, M. Iguchi, L. Jonsson and P. Jönsson: ISIJ int., 46 (2006), No. 8, 1137.
8) J. Szekely and S. Asai: Metall. Trans, 5 (1974), 464.
9) A. Nguyen and G. Evans: 3rd Int. Conf. on CFD in the Minerals and
Process Industries CSIRO, Melbourne, Australia, (2003), 71.
10) J.-Y. Zhang, S.-S. Du, S.-K. Wei: Ironmaking Steelmaking, 12 (1985), 249.
11) H.-J. Odenthal, U. Falkenreck and J. Schlüter: European Conf. on Computational Fluid Dynamics, the Netherlands, (2006).
12) D. Nakazono, K.-I. Abe, M. Nishida and K. Kurita: ISIJ Int., 44 (2004), 91.
13) M. Ersson, A. Tilliander, and P. Jönsson: Proc. Sohn Int. Symp Advanced Processing of Metals and Materials, ed. by F. Kongoli and R. G. Reddy, TMS, San diego, USA, Aug 27–31, (2006), p. 271.
14) L. Jonsson, D. Sichen and P. Jönsson: ISIJ Int., 38 (1998), 260.
15) L. Jonsson: PhD Thesis, Dept. of Metallurgy, KTH, Sweden, (1998).
16) L. Jonsson, P. Jönsson, S. Seetharaman and D. Sichen: Proc. of 6th Japan–Nordic Countries Steel Symp., ISIJ, Tokyo, (2000), 77.
17) C. W. Hirt and B. D. Nichols: Comput. Physics, 39 (1981), 201.
18) B. E. Launder and D. B. Spalding: Comp. Meth. Appl. Mech. Eng., 3 (1974), 269.
19) Fluent User’s Manual, (2007).
20) T.-H. Shih, W. W. Liou, A. Shabbir, Z. Yang and J. Zhu: Computers Fluids, 24 (1995), 227.
21) J.-O. Andersson, T. Helander, L. Höglund, P. Shi, and B. Sundman: Calphad, 26 (2002), 273.
22) A. Nordquist, A. Tilliander and P. Jönsson: Proc. 5th European Oxygen Steelmaking conf., Aachen, Germany, (2006), 519.。