Density problem in fluent

flow_CH · December 3, 2013, 12:51

Hello

I am modeling a 3d spillway in Fluent.(vof multiphase)
Depths and speeds and pressures are correct. (in free surface volume fraction is 0.5) But in Density or Air concentration (C) i have a problem that it is confused me.
In free surface(volume fraction=0.5) of my model density is about 500 kg/m3 (air concentration is C=50) !!!
I mean in upper than free surface it is going to 1.225 (air density) (C=90 or 95 or 99).
I mean in upper than free surface the density should be about 1.225 (air concentration C=100) .
I mean after free surface it should be just air with density=1.225

Please help
Thanks

pakk · December 3, 2013, 14:40

I am sorry, but I don't understand what you are saying... What is your problem? Which thing is not what you thought it should be?

Zaktatir · December 3, 2013, 15:38

???? Your pic and profile are correct! What do you want? You are plotting the density of mixture and that is why you are getting averaged values

By the Way do you mean with C the void fraction

flow_CH · December 3, 2013, 16:08

yes c is void fraction.
in free surface density should be 1.225 or upper than but in fluent it is 500.
please see attached image

Zaktatir · December 3, 2013, 16:27

You are using VOF have u read something about phase average of the navier stokes equations. You will never get your 998 at the free surface from any programm because you are solving the phasic average NS-Equations: Your free surface is of a finite thickness this thickness does depend on your mesh your discretisation order and the volume fraction recontrusction say tracking or caprturing. Your result are OK. At the free surface you have 50% probability to have air as same as water that is why you will get there the arithmetic average almost

flow_CH · December 5, 2013, 07:36

Thanks for your replies.

I want to understand this problem for my thesis.
If in defence meeting, the judges ask me, why in free surface the air concentration is 0.5, (because in common and nature it should be 0.9 or upper) how can i explaine this problem? What justifications i should offer? If i say my modeling is OK and it is just related to Fluent VOF, they will not accept.

(because in common and nature air concentration should be 0.9 or upper)

pakk · December 5, 2013, 08:29

You define free surface as (volume fraction=0.5), so that is what you get. If you would define free surface as (volume fraction=0.9), you would get a different result. This answer might be a bit too simple, because maybe I don't understand your question.

Is your question: "why is the density on the free surface, which I define as having a liquid volume fraction of 50%, around 500 kg/m3, while I would expect it to have a liquid volume fraction of 90% and therefore have a density of 900 kg/m3?"?

flow_CH · December 5, 2013, 12:10

I am defining free surface in volume fraction=0.5 because it is usual in many papers. With defining free surface in volume fraction=0.9 the depth of flow is not same with experimental data.(it is larger than main depth)
Yes. I want with defining free surface in volume fraction=0.5, the density in free surface be about 1.225 (close to air). If it is not possible and my simulation is OK, so how can i doing with this problem? I need an explanation for this problem.

pakk · December 5, 2013, 12:28

In theory guide (Fluent version 14.5), section 1.3.2, equation (1-15):

$\rho_m=\sum_l \alpha_l \rho_l$

Here $\rho_m$ is the mixture density, $\alpha_l$ is the volume fraction of phase l and $\rho_l$ is the density of phase l.
In your case, both alpha's (volume fractions) are 0.5, so:

$\rho_m=0.5 \cdot (\rho_{water} + \rho_{vapor})$

And just put in the numbers, this value is 500 kg/m3 in your case. This can easily be calculated without simulation, so your simulation is giving the correct numbers.

The concept is not so difficult to understand: in your 'free surface', half of the cell is filled with water, the other half with vapor. The density will therefore be exactly in the middle. You can only get a different free surface density if you define your free surface in a different way.

Zaktatir · December 5, 2013, 12:35

You have first of all to be familiar with the theory behind what you are calculating say multiphase flows and how to get the underlying physics from the conservations equations of mass, momentum and energy. We helped there more than necessary it is now up to you to learn or not.

Keep it real

flow_CH · December 5, 2013, 15:27

I am familiar with the theory behind what i am calculating multiphase flows.

Can i equivalent density=500 with density=1.225 ? I mean, moving upper density from free surface to down.

pakk · December 5, 2013, 15:45

If you desparately want the free surface to have a density of 1.225, you should change the density of water to 1.225. It will not make any sense, but if it makes you happy, just do it.

I think you are asking the wrong question. You should be asking for yourself:
- What is a free surface exactly?
- Why do I think that the free surface should have a density of 1.225, and is that correct? (Hint: "because I read it in an article" is not sufficient, you should understand it)

Your questions show that you are not familiar with the theory behind calculating multiphase flows. That is not something to be ashamed of, I am not familiar with many theories.

Zaktatir · December 7, 2013, 06:08

My last pot here:

if you consider the surface of water: and you introduce a probe to capture the surface and the water is moving the probability to get the surface or water is 50% that is you in the phasic average NS you get an average density. In the realtiy you only have either water or air this leading to discontinuity on the surface but if you want to calculate that you have to smooth the gradient this yields to an average quantity..

flow_CH · December 7, 2013, 06:37

Does the mesh effects in distance from contours?
I mean if my mesh be smaller than present, then the contour of density=1.225 is closer to density=500?

Zaktatir · December 7, 2013, 07:33

generally the free surface has to be fine meshed let say 10-15 cells mesh resolution would affect the solution especially if you want to resolve ther free-surface. Talking now about the density i still do not know why you think you have a problem. Your problem is not the simulation rather your understanding of what you expect from the simulation

flow_CH · December 8, 2013, 15:43

Just tell me.
Can density of mesh effects on distance between contours?
I mean if i use denser mesh, the distance between contour of 0.5 and contour of 0.9 is fix or can change in relative to big mesh?

Zaktatir · December 8, 2013, 15:57

if you soliton is grid independent then you will get nor change otherweise some changes will occur. But you will still get the average density of 500 at the free surface..

flow_CH · December 10, 2013, 04:22

If volume fraction=0.5 is free surface then can you see (without considering VOF and theories), what are the upper volume fractions? Are they wave or entrapped air or entrained air?

ghost82 · December 10, 2013, 11:22

Try to select/deselect "node values" when plot contours or on 2d chart. I think you will get what you want.

Daniele

flow_CH · December 10, 2013, 15:38

I tried but it is not effecting more in my results . What is "node values"?

December 3, 2013, 12:51	Density problem in fluent	#1
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	Hello I am modeling a 3d spillway in Fluent.(vof multiphase) Depths and speeds and pressures are correct. (in free surface volume fraction is 0.5) But in Density or Air concentration (C) i have a problem that it is confused me. In free surface(volume fraction=0.5) of my model density is about 500 kg/m3 (air concentration is C=50) !!! I mean in upper than free surface it is going to 1.225 (air density) (C=90 or 95 or 99). I mean in upper than free surface the density should be about 1.225 (air concentration C=100) . I mean after free surface it should be just air with density=1.225 Please help Thanks Last edited by flow_CH; December 13, 2013 at 16:48.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Fluent - license problem.	Marcin	FLUENT	3	April 13, 2018 17:33
A problem about density in liquid air definition	alloveyou	CFX	2	June 14, 2012 15:20
[ICEM] [FLUENT] 2D Geometry problem when exporting to Fluent - Unwanted walls	MikeTichondrius	ANSYS Meshing & Geometry	1	February 9, 2011 14:31
Problem about Fluent on Linux	hbinma	FLUENT	3	July 6, 2008 11:49
Fluent Vs CFX, density and pressure	Omer	CFX	9	June 28, 2007 05:13

December 3, 2013, 14:40		#2
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	I am sorry, but I don't understand what you are saying... What is your problem? Which thing is not what you thought it should be?

December 3, 2013, 15:38		#3
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	???? Your pic and profile are correct! What do you want? You are plotting the density of mixture and that is why you are getting averaged values By the Way do you mean with C the void fraction

December 3, 2013, 16:27		#5
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	You are using VOF have u read something about phase average of the navier stokes equations. You will never get your 998 at the free surface from any programm because you are solving the phasic average NS-Equations: Your free surface is of a finite thickness this thickness does depend on your mesh your discretisation order and the volume fraction recontrusction say tracking or caprturing. Your result are OK. At the free surface you have 50% probability to have air as same as water that is why you will get there the arithmetic average almost

December 5, 2013, 07:36		#6
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	Thanks for your replies. I want to understand this problem for my thesis. If in defence meeting, the judges ask me, why in free surface the air concentration is 0.5, (because in common and nature it should be 0.9 or upper) how can i explaine this problem? What justifications i should offer? If i say my modeling is OK and it is just related to Fluent VOF, they will not accept. (because in common and nature air concentration should be 0.9 or upper)

December 5, 2013, 08:29		#7
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	You define free surface as (volume fraction=0.5), so that is what you get. If you would define free surface as (volume fraction=0.9), you would get a different result. This answer might be a bit too simple, because maybe I don't understand your question. Is your question: "why is the density on the free surface, which I define as having a liquid volume fraction of 50%, around 500 kg/m3, while I would expect it to have a liquid volume fraction of 90% and therefore have a density of 900 kg/m3?"?

December 5, 2013, 12:10		#8
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	I am defining free surface in volume fraction=0.5 because it is usual in many papers. With defining free surface in volume fraction=0.9 the depth of flow is not same with experimental data.(it is larger than main depth) Yes. I want with defining free surface in volume fraction=0.5, the density in free surface be about 1.225 (close to air). If it is not possible and my simulation is OK, so how can i doing with this problem? I need an explanation for this problem.

December 5, 2013, 12:28		#9
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	In theory guide (Fluent version 14.5), section 1.3.2, equation (1-15): $\rho_m=\sum_l \alpha_l \rho_l$ Here $\rho_m$ is the mixture density, $\alpha_l$ is the volume fraction of phase l and $\rho_l$ is the density of phase l. In your case, both alpha's (volume fractions) are 0.5, so: $\rho_m=0.5 \cdot (\rho_{water} + \rho_{vapor})$ And just put in the numbers, this value is 500 kg/m3 in your case. This can easily be calculated without simulation, so your simulation is giving the correct numbers. The concept is not so difficult to understand: in your 'free surface', half of the cell is filled with water, the other half with vapor. The density will therefore be exactly in the middle. You can only get a different free surface density if you define your free surface in a different way.

December 5, 2013, 12:35		#10
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	You have first of all to be familiar with the theory behind what you are calculating say multiphase flows and how to get the underlying physics from the conservations equations of mass, momentum and energy. We helped there more than necessary it is now up to you to learn or not. Keep it real

December 5, 2013, 15:27		#11
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	I am familiar with the theory behind what i am calculating multiphase flows. Can i equivalent density=500 with density=1.225 ? I mean, moving upper density from free surface to down.

December 5, 2013, 15:45		#12
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	If you desparately want the free surface to have a density of 1.225, you should change the density of water to 1.225. It will not make any sense, but if it makes you happy, just do it. I think you are asking the wrong question. You should be asking for yourself: - What is a free surface exactly? - Why do I think that the free surface should have a density of 1.225, and is that correct? (Hint: "because I read it in an article" is not sufficient, you should understand it) Your questions show that you are not familiar with the theory behind calculating multiphase flows. That is not something to be ashamed of, I am not familiar with many theories.

December 7, 2013, 06:08		#13
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	My last pot here: if you consider the surface of water: and you introduce a probe to capture the surface and the water is moving the probability to get the surface or water is 50% that is you in the phasic average NS you get an average density. In the realtiy you only have either water or air this leading to discontinuity on the surface but if you want to calculate that you have to smooth the gradient this yields to an average quantity..

December 7, 2013, 06:37		#14
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	Does the mesh effects in distance from contours? I mean if my mesh be smaller than present, then the contour of density=1.225 is closer to density=500?

December 7, 2013, 07:33		#15
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	generally the free surface has to be fine meshed let say 10-15 cells mesh resolution would affect the solution especially if you want to resolve ther free-surface. Talking now about the density i still do not know why you think you have a problem. Your problem is not the simulation rather your understanding of what you expect from the simulation

December 8, 2013, 15:43		#16
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	Just tell me. Can density of mesh effects on distance between contours? I mean if i use denser mesh, the distance between contour of 0.5 and contour of 0.9 is fix or can change in relative to big mesh?

December 8, 2013, 15:57		#17
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	if you soliton is grid independent then you will get nor change otherweise some changes will occur. But you will still get the average density of 500 at the free surface..

December 10, 2013, 04:22		#18
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	If volume fraction=0.5 is free surface then can you see (without considering VOF and theories), what are the upper volume fractions? Are they wave or entrapped air or entrained air?

December 10, 2013, 11:22		#19
ghost82 Senior Member Rick Join Date: Oct 2010 Posts: 1,016 Rep Power: 27	Try to select/deselect "node values" when plot contours or on 2d chart. I think you will get what you want. Daniele

December 10, 2013, 15:38		#20
flow_CH Senior Member FHydro Join Date: Jan 2013 Posts: 197 Rep Power: 13	I tried but it is not effecting more in my results . What is "node values"?