[mech136984]

I am trying to validate a work which involves multiphase flow along with a porous media. I am very close to the validation but a problem is occurring for which i need certain help from you guys.
Actually after I had initialized and patched the fluid flow domain with 0.185 solid volume fraction the problem initialization is exactly similar to that given in the paper. But when I am running the solution my porous media is not being recognized i.e whole of the particles phase red strip is moving down and eventually going into porous media zone.
I am attaching a reference paper part of the work which i am trying to validate and also attaching the result which i am getting.

Image 102 is the work which I am trying to validate. Image 104 (1) and 104 (2) are of my solution.
100.JPG

101.JPG

102.JPG

104 (1).PNG

104 (2).PNG

[vinerm]

The model of the porous domain is implemented only as a momentum sink. It will not stop anything from going through. It is not very clear whether the inlet is at the top or the bottom. If it is at the top, since it is mentioned that the outlet is at the bottom, then the fluid is pushing it down. There is no reason it should not move over to and pass through the porous domain. The only thing will be the resistance offered to the momentum.

If you do not want it to pass through, then some condition has to be applied at the interface (not mesh interface) of the mud and the porous domain that acts like a wall.

[mech136984]

What should I do to make sure that porous zone do not slide outside the domain? I have applied different solid resistance (low to high value), but still it moved out. What conditions should I implement?

[vinerm]

Porous zone does not move anywhere; resistance coefficients are applied at the cells and cells are stationary until you use mesh motion or dynamic mesh, which you certainly are not using. So, the issue is not porous zone sliding out, rather the fluid going through the porous zone. As mentioned earlier, porous zone cannot stop it; even in reality. It can only slow it down. That's exactly what Fluent is doing. So, if you run a steady-state case, everything will pass through the porous zone because that is the steady state of the system. In a transient simulation, it will take time. Hence, the very first requirement you need to fulfill is to run the simulation in transient.

[mech136984]

bro i had tried with the transient case also but the problem is in velocity vectors calcium carbonate (secondary granular phase) appears not to penetrate inside porous zone and only water passing through porous zone which is exactly what i want but the problem is volume fraction contours are not appearing similar to paper and other thing is paper does not mention anything about the time step , courant number and time of flow.And please help me one more thing that whether we need to do any specific thing with the interface?

[vinerm]

By interface do you imply mesh interface or phasic interface? As far as phasic interface is concerned, you just need to provide exchange coefficients, such as, drag. No special treatment is needed.

If you do not have coming out from the bottom along with the water, then where is it going? This implies that there is mass conservation issue. With Euler-Euler models, that's not uncommon but not of this order. Instead of looking at velocity vectors, look at the volume fraction of secondary phases at the bottom outlet. In case it does not show any Calcium Carbonate, then it is good but surprising.

[mech136984]

i was talking about the mesh interface only.And the volume fraction contours are the main problem like vectors shows that particles are not crossing the porous zone and only water is crossing but particles should have deposited and accumulated on the top of porous zone and the volume fraction must have had increased there and result must be similar to paper but the contours is behaving as i had attached in my query.

[vinerm]

What is the objective of having a mesh interface in your domain? Mesh interfaces are required only if you want a moving mesh or if the model is very complex and user still wants hexahedral mesh everywhere. Your model should not have any mesh interface at all.

Secondly, I suppose you are using Euler-Euler or Mixture model. There are no particles in the domain then. Anyway, the problem is that you have a phase disappearing. I am not sure how it is done in the paper but as far as fluid flow modeling is concerned, you have to somehow model the phenomenon of stopping the granular phase from falling through. As such, it would fall through and that's what Fluent is doing. I suggest you go through the article again and figure out what extra modeling are they doing to ensure that it does not flow through the porous zone.

[mech136984]

Thanks for your help. I had successfully validated the first part of papers work which was the problem that i shared with you. Now I am trying to validate the second part in which there is nothing but changed geometry dimensions and changed boundary conditions. But the problem is paper had given the result in four different frames of simulation. Can anyone explain me what does that means? Also the results are looking inappropriate as for our simulation the cake is getting compressed from top and attaining a constant thickness and solid volume fraction after sometime. But in second case by paper the result shows cake build up from bottom to top and also the confusing thing is that when the cake thickness is least it must have had the highest value of volume fraction but the case is just opposite. At lowest thickness i.e at T=1/6 the thickness is least and volume fraction is also least =0.801. Please help me with the same. Thank you

[vinerm]

Since the scale used for all four images is different, you cannot directly compare those. In this scenario, most likely, there is more mass coming in. Otherwise, the results would have'd been as you expected, i.e., maximum volume fraction for least thickness. Higher thickness with higher volume fraction means more mass or reduction in volume of the other phase.

[mech136984]

Hi vinerm.
Can I please catch up with you on skype or something for discussing the problem in detail. It would be very thankful of you if you can view it once on teamviewer or we can discuss it on video call.You may mail me your skype id or hangout request on mail id @tyagimukul1994@gmail.com.
Thank you

[vinerm]

I am afraid I don't use any of these tools nor do I prefer to. I'd suggest you share the paper, if you can. I will go through it and share my opinion. Looking at the results you outlined, it does not appear to be well-written.

[mech136984]

Ok I am attaching the paper please see it once if you could help it would be very thankful of you.

[vinerm]

I suppose you forgot to attach the paper.

[mech136984]

Yaa actually it required to be compressed below 195kb.

[vinerm]

I will look into it.

[mech136984]

Bro did you get any idea?

[vinerm]

I did go through the whole document and could not find the details about inlet. It is not uncommon for the authors to hide or miss many details; the paper provides no information about what is entering from the top. However, looking at the results, it is NOT water alone entering from the top in the second scenario. Solid is entering as well otherwise the volume fraction will not increase along with thickness.

[mech136984]

Yaa bro I also have the same doubt. The other issue is that in the first case I am able to validate the result close enough to the papers result but there compaction starts from top to bottom. In the second case how come is it growing from bottom to top?

[vinerm]

The only way that can happen is if solid is coming along with water from top. There is no other explanation for this.