Hello, I am a graduate student in Metallurgy and am doing research on Post combustion in EAF. I am using Fluent 5 for doing the fluid flow modeling. however, I am not very familiar with it and am not sure as to how should go about modeling the scrap and also take into account the heat transfer.

I would appreciate any suggestions regarding this and would be more than happy to give more details regarding my research if needed.

Thanks. Divyang.

hi,

here are some details:

I am trying to simulate the steelmaking process in the EAf

For the 2d model....I have CO coming in from bottom passing through porous bed of scrap and O2 coming in through the lances from the sides.The gases leave through the top.

Heat loss is due to convection and radiation...

I am not sure how to model the CO passing thru' the porous scrap....,.....once again I am using the commercially available Fluent 5 version!

Divyang.

What is the "scrap" portion of your problem?

During the early stages of the heat there is molten heel at the bottom of the furnace. After charging we have scrap in it. Now when the electrodes are lowered and the arc is struck the scrap starts melting. Thus, we have the scrap melting continuously and the gases going out through the top. A number of reactions occur at this time. What I am interested in is to look at the distribution of CO after the melting occurs....or I would say during that stage of heat when the gases coming out are maximum. The aim is to post combust this CO to CO2 and reuse that energy within the furnace.

I HOPE THIS HELPS.....MY PROBLEM IS THAT I DON'T KNOW HOW TO PUT SCRAP IN THE FURNACE( SAY 2/3RD FULL) AND LEAVE THE REST AS GAS SPACE!! ALSO, I AM NOT SURE HOW I CAN DEFINE THE COMPOSITION OF THE SCRAP IN FLUENT V5 !

Thanks for your interest and I hope u can give me some hints and/or suggestions.

Divyang.

(1). It is a very " messy " problem, because you don't know the original conditions and the full geometry. (2). So, you need to understand your problem first, about the exact geometry, the properties, and conditions. (3). You can build a small scale model and try to obtain the geometry, properties and flow conditions. The model can be a transparent one, which you can observe the geometry and the flow field. (4). Once you have a good picture of the model, you can try to transfer it to the CFD model. (5). So, it is not yet a CFD problem. You have to work hard to come up with a simplified working model. Otherwise, you will have to input all the detailed geometry into the computer program, which is almost impossible at this point. (6). Try to look into area of automobile catalytic converter ( for exhaust gas) CFD modeling. This is probably the best way to model your problem. In the catalytic converter, there are also chemical reactions taking place. You should be able to find many CFD papers published in this area.

(1). Thank you for sending me the e-mail. I am afraid that I am not going to have time to look at the problem in detail. (2). There are several steps you can take. (3). First is the geometry modeling. Try to simplify the geometry of the device so that you can define the computational domain , which should include the interior area, the walls, the inlets, and the exit. In the interior area, the porous material region needs to be defined also. This will define the extent of the porous material region, whether it is everywhere or sitting only in the middle, or on the bottom, is part of your geometry model. (4). From this exercise, you will reach a conclusion whether the model is 2-D, 2-D axisymmetric, or 3-D. This is important, because a 3-D problem will require much more resources than a 2-D problem. It will take much longer to solve. (5). Once this is done, you need to find out the properties of the porous material. (6). From this point on, you can start transfering the model to the computer program. (7). Since you are looking for the distribution of Co, you can split the problem into two parts, one is cold flow simulation, and the other is the reacting flow. Sometimes, the cold flow information is good enough to answer your questions. Then, there is no particular need to run reacting flow calculations, which will take much longer if the problem is 3-D. (8). So, the ideal place to start the CFD simulation is a 2-D cold flow calculation, where you can easily change the configuration of the porous material region and its properties. In this way, you can observe the Co gas distribution. (9). I don't know how to handle the heating and the melting processes. I also don't know how the Co is generated and distributed originally. If Co is linked to the heating and melting processes, then I think you need to model this part also. (10). Try to handle the problem in a step-by-step fashion. Try to solve each one at a time. The most important thing to remember is to keep the model as simple as possible. This will be easier for the code you are going to use. Good luck.

Thanks for the suggestions. I am actually partly doing what you have suggested. I have done a cold flow simulation. I am working for the 2d case and have defined the domain for it. I also have gases flowing as well as reacting.

What I can't figure out is how to put a porous bed of scrap sitting at the bottom.( 2d furnace- rectangle shaped, exit at top). I want to put the scrap in only 1/3rd part of the furnace and leave the rest 2/3rd as gas space at the top. What papers/books do you suggest I need to look at to better understand and simulate this!

I will take a look at the papers on catalytic converters.

Thanks a lot for your interest

Divyang.

(1). So, one third of the bottom computational domain will be porous medium, and the top two third will have nothing in it. (2). In this case, you have two regions in one computational domain. (3). It should be very similar to the flow in a catalytic converter, where only portion of the space is filled with catalytic material.( different sizes, shapes, and design) (4). You should be able to learn a lot from the modeling of the catalytic converter.

To keep the problem simple, if the scrap is melting and then releasing the CO, you may try a sublimation model. Fluent will automatically calculate the stoichiometry of the reaction and if you keep the geometry very simple, you may find success in observing the reaction from solid to gas. These individual experiments in Fluent may require several steps such as first run a porous model to get a feel for the flow field, then run a sublimation or melting model to see the heat transfer and chemical reactivity. Something else to keep in mind is the size of your original scrap-CO-CO2 experiment. If its at all possible to scale this down, then it will be easier to model the stages so that later you will be able to save time by having to focus on only one small part of the entire reaction system.

I am currently trying to run a porous model to get the flow field....my problem is I am not sure if Fluent is actually simulating it...

I am trying to put a porous bed of scrap and see the flow of gases when they pass through it......is there any way I can check........like, the porosity etc. in Fluent.

I also have to take into account the heat transfer which i am ignoring for now....when I put the heat tranfer coefficient and the free stream temperature etc. for the walls, I can't figure out the numbers which Fluent comes up with after iterating and getting a converged solution. Is there any way I can verify these answers?

do you have manuals? do you have tech support? have you done simple experiments? cfd, even fluent 5.0, is not a black box and relying on the answers without answering yes to the above questions may be problematic. my first recommendation would be to discuss your problem in detail with tech support at fluent. it is a very interesting and useful problem to solve, but it is also quite complex, so their specific fluent/numerical experience will be most helpful in answering your questions.

I do have the manuals....I had contacted the Tech people at Fluent but unfortunately their replies were too vague......

Fluent manuals and Tutorials have been of immense help during the early stages when I didn't know anything ......

They do say what can be done using Fluent but when it comes to implementation and the problem gets a bit complex the details are not enough!