Strange behaviour: Ethanol solution flow on channel with interface to porous domain

khariel · December 10, 2014, 14:07

Hi all,

I'm modelling a ethanol solution (9% mass) flowing through a serpentine-like channel, and at the base of this channel there's an interface connecting to a porous domain.

I'm using a Permeability coefficient for the porous domain (2e-12 m²) and inputting the diffusivity of ethanol on the Fluid Model for each domain. There's an outlet at the base of the porous domain, as shown on the pictures below.

I'm getting these strange results as can be seen from the pictures. As the solution flows through the channel, the ethanol mass fraction goes down, but around the mid-section after the first turn it goes up. I don't believe this should be happening in reality. I also attached a mass fraction contour on the porous domain.
It is evident that both profiles sort of mimic each other. I think I'm missing something here. What I'd expect to happen is the concentration along the channels to be more or less the same, and a concentration gradient to be present only in the porous region.

Mass fraction along the flow channel (fluid domain):

Mass fraction in the porous domain:

Appreciate the help.

ghorrocks · December 10, 2014, 15:52

Isn't the ethanol just doing a short-circuit through the porous material? You can see it diffusing out in the bottom image.

khariel · December 10, 2014, 19:58

Quote:

Originally Posted by ghorrocks

Isn't the ethanol just doing a short-circuit through the porous material? You can see it diffusing out in the bottom image.

I thought about it, but checking the Z component of the velocity in the porous media only shows negative values, meaning that the ethanol is only going down towards the porous media outlet. Is there any other way to check if that's what's happening?

ghorrocks · December 10, 2014, 23:32

It seems pretty obvious to me this is ethanol short-circuiting through the porous material. Have a look at the ethanol velocity vectors in the porous material. Post an image of it on the forum.

khariel · December 11, 2014, 08:40

Quote:

Originally Posted by ghorrocks

It seems pretty obvious to me this is ethanol short-circuiting through the porous material. Have a look at the ethanol velocity vectors in the porous material. Post an image of it on the forum.

I tried to check the vectors but the arrows are so tiny in the porous region that it's really difficult to tell something. I tried all sorts of plot options and couldn't get a better visualization.

However, I found something interesting playing with a W Velocity contour very close to the interface. I could see regions of positive Velocity W, meaning the ethanol is really short-circuiting. Is that a definite conclusion? And if so, what are the reasons this could be happening? Not enough pressure in the flow channels?

ghorrocks · December 11, 2014, 17:02

If the vectors are too small then make them bigger. Go to the vector object, select the symbol tab and make the Symbol Size parameter larger.

If your W velocity is looking as blotchy as that I would not trust the accuracy of your results. I suspect it is one of more of the following:
* Inadequate mesh resolution
* Not fully converged
* numerical instability at the porous/fluid interface

khariel · December 11, 2014, 17:08

I had vector size at maximum size and still, the vectors in the region of interest were too scarce and too small, no matter how large the vector numbers was set as well. In some regions you could probably tell that there was some fluid going up back to the channel, but not really conclusive.

I'll check mesh dependency.

I checked convergence via imbalance and residuals and everything seemed to not be changing by the time the run was over. Is there another better way to do that?

How can numerical instability be checked?

December 10, 2014, 14:07	Strange behaviour: Ethanol solution flow on channel with interface to porous domain	#1
khariel New Member Leonardo Join Date: Nov 2014 Posts: 27 Rep Power: 11	Hi all, I'm modelling a ethanol solution (9% mass) flowing through a serpentine-like channel, and at the base of this channel there's an interface connecting to a porous domain. I'm using a Permeability coefficient for the porous domain (2e-12 m²) and inputting the diffusivity of ethanol on the Fluid Model for each domain. There's an outlet at the base of the porous domain, as shown on the pictures below. I'm getting these strange results as can be seen from the pictures. As the solution flows through the channel, the ethanol mass fraction goes down, but around the mid-section after the first turn it goes up. I don't believe this should be happening in reality. I also attached a mass fraction contour on the porous domain. It is evident that both profiles sort of mimic each other. I think I'm missing something here. What I'd expect to happen is the concentration along the channels to be more or less the same, and a concentration gradient to be present only in the porous region. Mass fraction along the flow channel (fluid domain): Mass fraction in the porous domain: Appreciate the help.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Error finding variable "THERMX"	sunilpatil	CFX	8	April 26, 2013 07:00
Question about creating porous domain interface for CFX by ICEM	lnk	CFX	2	July 21, 2012 06:48
Air-water interface in a channel flow using VOF method	Chocosoboro	FLUENT	0	April 6, 2011 10:04
Material identification between a flow domain and a porous domain	Ervideiro	CFX	1	June 1, 2010 09:50
fluid flow fundas	ram	Main CFD Forum	5	June 17, 2000 21:31

December 10, 2014, 15:52		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Isn't the ethanol just doing a short-circuit through the porous material? You can see it diffusing out in the bottom image.

December 10, 2014, 23:32		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	It seems pretty obvious to me this is ethanol short-circuiting through the porous material. Have a look at the ethanol velocity vectors in the porous material. Post an image of it on the forum.

December 11, 2014, 17:02		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	If the vectors are too small then make them bigger. Go to the vector object, select the symbol tab and make the Symbol Size parameter larger. If your W velocity is looking as blotchy as that I would not trust the accuracy of your results. I suspect it is one of more of the following: * Inadequate mesh resolution * Not fully converged * numerical instability at the porous/fluid interface

December 11, 2014, 17:08		#7
khariel New Member Leonardo Join Date: Nov 2014 Posts: 27 Rep Power: 11	I had vector size at maximum size and still, the vectors in the region of interest were too scarce and too small, no matter how large the vector numbers was set as well. In some regions you could probably tell that there was some fluid going up back to the channel, but not really conclusive. I'll check mesh dependency. I checked convergence via imbalance and residuals and everything seemed to not be changing by the time the run was over. Is there another better way to do that? How can numerical instability be checked?