[esi1520]

Hi every body

Please help me
I have big problem!!!!!!

I am trying to simulate combustion in a furnace and use two mixture. I have two zone and two mixture for each zone.

After initialization, fluent automatically set mass fraction of last component of species in list but physically this is wrong. In fact this component is for another mixture and zone. After some iteration initial mass fraction of this component almost constant. How to fix this problem?

Also i read Fluent user document. In that emphasize this note " The order of the species in the Selected Species list is very important. ANSYS Fluent considers the last species in the list to be the bulk species. You should therefore be careful to retain the most abundant species (by mass) as the last species when you add species to or delete species from a mixture material."

According this note last specie of component list is bulk species that we can not set up solution control, solution method, multigrid type and another factor for simulation.

How we can set up of bulk ccomponent

Thnaks

[LuckyTran]

Are you sure? The mixture is attached to the zone and you should not have a mass fraction belonging to a different zone. Secondly, the mass fraction can only be interpreted in terms of a mixture, so you cannot have a mass fraction belonging to a different mixture.

The last species does not need any solution method because it is calculated by using 1 minus all the other species. The last species should be the one with the highest mass fraction for numerical precision reasons.

Also why is "FLUENT" in quotes? I don't understand the reason for the emphasis. Is this a reference to some obscure joke?

[esi1520]

Originally Posted by LuckyTran

Are you sure? The mixture is attached to the zone and you should not have a mass fraction belonging to a different zone. Secondly, the mass fraction can only be interpreted in terms of a mixture, so you cannot have a mass fraction belonging to a different mixture.

The last species does not need any solution method because it is calculated by using 1 minus all the other species. The last species should be the one with the highest mass fraction for numerical precision reasons.

Also why is "FLUENT" in quotes? I don't understand the reason for the emphasis. Is this a reference to some obscure joke?

Dear Lucky Tran, thanks for your reply,

I'm sure. I know mass fraction of last specie of list automatically is 1- minus ... but this component with its mass fraction retain in some iteration with no change. Is necessary to any set up for this bulk mixture component?

[LuckyTran]

Are you saying you have found a mass fraction of a species that does not exist? Or is your problem that you expect this mass fraction to change but it isn't? Well then you know that it is calculated from 1-minus the rest, so the problem is in the settings for the species that you do solve.

[nguyentbtuyen]

Hi LuckyTran, and experts

In the case I have only one solute and one solvent, the initial solute concentration is 0.9 so what should be the order of the species?

I tried the abundant species - the solute - as the last species, I could not get the right answer which is from the work of Voller et al.;

Please help!

Thank you

Tuyen

Originally Posted by LuckyTran

Are you saying you have found a mass fraction of a species that does not exist? Or is your problem that you expect this mass fraction to change but it isn't? Well then you know that it is calculated from 1-minus the rest, so the problem is in the settings for the species that you do solve.

[vinerm]

The objective of keeping the most abundant species as the last one is to reduce the error. A small error for a specie with 0.9 mass fraction may not be noticeable but such an error for a specie with 0.01 mass fraction may lead to wrong results. Otherwise, it does not matter.

Which article by Voller are you trying to replicate?

[nguyentbtuyen]

Hi Vinerm,
Thanks for your reply.
I am working on the NH4Cl-H2O of Voller Brent Prakash 1989 where the initial liquid concentration was mentioned 0.1 but it might be the solute’s instead (initial T is 600K).
So I placed the solute (0.1) and the solvent in the species. I have been working on this for weeks but could not reproduce the solute concentration figure at 3000s in that paper (the lid frac and stream func seems to be fine). i am sure I have input the right data into Fluent. The max solute conc is about ~0.15 instead of ~0.21 in that paper. If you know well that paper could you advise please?
Kind regards
Tuyen

[vinerm]

I do not have access to this paper, however, ammonium chloride and water will react. Are you modeling the reactions properly?

[nguyentbtuyen]

Hi Vinerms,

No, I don't think there would be any reactions. Also, there is no reactions mentioned in the paper. I just try to play with the material properties a bit more to see.

Thanks again,

Tuyen

[vinerm]

Water and Ammonium Chloride, if mixed even at room temperature, react and produce HCl. But I suppose the paper you are referring to deals with the solidification and solute rejection. Without either the reactions or the solute rejection or absorption, the concentration will not change.

If it is solidification you are trying to model, then you need to setup diffusion in solid properly. As far as I remember, the paper uses some finite back diffusion.

[nguyentbtuyen]

Hi Vinerm,

The paper dealed with two limiting cases i.e. infinite and zero duffusion in solid, no back diffusion I think. The first one is what I am trying to reproduce. The conditions are as follow: the left wall of 400K, three other walls insulated, initial T=600K, melting T=630. Teut=~250, Ceut=0.8. In my simulation the max. fraction of the 2nd phase is around 0.14 while it was shown 0.21 in the paper (though I can't find the max concentration at the intersection point - between the solidus and eutectic lines from the binary diagram).

Could you advise me about this deviation please?

Thanks Vinerm,

Tuyen

[vinerm]

I remember to have worked with this paper a long time ago and as far as I remember, we had to do something with back diffusion or solutal buoyancy to ensure the match. Since that was still Fluent 6.3.26, we had to write a UDF for the same. I might have this paper somewhere in my repository but if it is not much of an issue, could you share it?

[vinerm]

I have been able to find the article in my repository. I am also searching for the case files. Will update you by end of tomorrow.

[nguyentbtuyen]

Good morning Vinerm,

Thanks so much for your support. The phase diagram of the Nh4Cl-H2O may be useful as well - Fig. 3 of the paper https://pubs.acs.org/doi/10.1021/acs.jced.7b01062

Looking forward to your next post

Best

Tuyen

[vinerm]

I was also able to locate the case file, however, that is for Model C, i.e., with Scheil Rule. I looked at the results and found that we also did not match the numbers given in the paper, but we were closer. Model C shows 0.25, which is quite high. Do you have thermal as well as solutal buoyancy enabled in your case?

[nguyentbtuyen]

Hi Vinerm,

Yes I enabled both thermal and solutal boyancy as they are both drivers. Our Scheil model also gave max. concentration higher than the Lever one’s, though not matching with the paper’s. What is the driving force for this max concentration do you think?

Also, if you look at the diagram I sent, the “solute concentration” mentioned in that paper may not be appropriate? I think it is the minor species concentration instead.

Cheers

Tuyen

[vinerm]

Our usual conception is that in an aqueous solution water is the solvent; that's why aqueous. And the case is partially setup as per that. However, the paper takes it other way around. Water is taken as solute while ammonium chloride is the solvent; look at the melting point of pure material. If you setup the case that way, you will be able to improve the results.

Furthermore, there is no need to use viscosity UDF since Fluent does not use the methodology used in the paper. Solidification model in Fluent is based on Enthalpy-Porosity technique, so, no special treatment is required for the viscosity of solid.

Additionally, you do not need to setup commands to extract images. Fluent has animation option within Calculation Activities that can be used to create animation directly. It needs to be setup before starting the case.

[nguyentbtuyen]

Thanks Vinerm for following up this challenge!
Yes you (and me) are right about the solute vs solvent. In the phase diagram - as I sent you in the last post, at 600K the minor/2nd species of 0.1 is actually the liquid (Even the “solute” was named in my case file but I meant it the minor or secondary species).

Regarding the viscosity, I found that harmonic mean (lower limit) was actually used by the authors instead of arithmetic as mentioned in the paper. But I have played around with the vis already, from constant to variable, still not getting the same max concentration. So I think I no longer want to exactly reproduce the graphs, as a number of properties were hidden in the paper!

Regarding the execute commands - I used to like them but not now, memory saving but very time consuming. Good to know the animation that can do better job! Thanks for this. I am quite okay with CFD post than before.

I appreciate your share of knowledge

Best regards
Tuyen

[vinerm]

Yes, I observed the name but Fluent does not care about the name but the properties. With the current properties, if you plot Liquidus and Solidus, the numbers calculated are incorrect. Therefore, you need to change properties of the solute to those of water. Ensure that the solidus temperature is always 250 and liquidus is varying.

[nguyentbtuyen]

I will look closely all the properties and run again to see. Thanks