Boundary condition & reverse flows

zhenglun.wei · May 8, 2011, 15:02

I met a problem on my current simulation. It always give me a reverse flow problem.
It is a 3D problem with a major inlet on the front boundary. The flow is allowable to go out from the back boundary. There is a minor inlet on the top boundary which is very closed to the back boundary. The flow comes from this minor boundary is very small.
The problem is, if I use 'out flow' as the boundary condition of the back boundary, the FLUENT gives me a lot of warning about 'reverse flow' and the simulation will not converge to even 10^-3. Then, I tried the 'velocity inlet' and set a constant velocity 'outward' of the back boundary. It can converge to 10^-4, but I know it is not real, specially I put a turbulence model on the whole simulation. Then, I tried the 'pressure-outlet', it still gives me the 'reverse flow' warning, but it can converge to 10^-4. Personally, I prefer to work with 'pressure-outlet' boundary, since, I think, it is more physical and understandable. But, I do not quite understand this BC.
Any suggestions on this?

thanks in advance.

farshadexp · May 9, 2011, 12:32

Hi there....

you can also use outlet-vent in your simulation, remember that according to my experience P-outlet many times leads to R-flow. By the way, I have similar problem now, but in different simulation case and I'll be really happy if you find better answer and inform me

Regards...

zhenglun.wei · May 9, 2011, 14:35

Dear Farshadexp,
thanks for your reply. Fluent manual says that the P-outlet can help us to make the simulation converge better when RF exists. However, in my opinion, it does not say that the P-outlet can help us to eliminate the reverse flow. As I tested, the P-outlet indeed help me to converge the simulation to 10^-4 (outlet BC can not even converge to 10^-3); however, the RF warning is still showing over there. I'll do some tests on your suggestions and let you know if it can fit my problem.

best,

choonhwai · October 20, 2011, 14:10

Hi,

I have the same problem. I have 1 inlet and 4 outlets. Both outflow and outlet-vent BC did not correct the non-convergence and the slowly growing number of reverse flow faces. I tried refining the mesh and re-meshing with aggressive smoothing, but neither worked.

will appreciate any suggestions..

choonhwai · October 25, 2011, 18:38

Hi guys,

I have some update: I think meshing may be at least partially responsible for the non-convergence of the multiple out-flow BC simulations.

I meshed a model with 1 inlet and 4 outflow BC with ICEM CFD and with Gambit. The one from ICEM CFD does not converge (I tried re-meshing to get finer mesh and also tried outlet vent BC), but the one from Gambit has no stability problems.

Does anyone know what is the difference in the meshing approach of ICEM CFD and Gambit?

May 8, 2011, 15:02	Boundary condition & reverse flows	#1
zhenglun.wei New Member Join Date: Jan 2011 Posts: 25 Rep Power: 15	I met a problem on my current simulation. It always give me a reverse flow problem. It is a 3D problem with a major inlet on the front boundary. The flow is allowable to go out from the back boundary. There is a minor inlet on the top boundary which is very closed to the back boundary. The flow comes from this minor boundary is very small. The problem is, if I use 'out flow' as the boundary condition of the back boundary, the FLUENT gives me a lot of warning about 'reverse flow' and the simulation will not converge to even 10^-3. Then, I tried the 'velocity inlet' and set a constant velocity 'outward' of the back boundary. It can converge to 10^-4, but I know it is not real, specially I put a turbulence model on the whole simulation. Then, I tried the 'pressure-outlet', it still gives me the 'reverse flow' warning, but it can converge to 10^-4. Personally, I prefer to work with 'pressure-outlet' boundary, since, I think, it is more physical and understandable. But, I do not quite understand this BC. Any suggestions on this? thanks in advance.

October 25, 2011, 18:38	problem is due to meshing?	#5
choonhwai New Member Choon Hwai Yap Join Date: Oct 2011 Posts: 4 Rep Power: 15	Hi guys, I have some update: I think meshing may be at least partially responsible for the non-convergence of the multiple out-flow BC simulations. I meshed a model with 1 inlet and 4 outflow BC with ICEM CFD and with Gambit. The one from ICEM CFD does not converge (I tried re-meshing to get finer mesh and also tried outlet vent BC), but the one from Gambit has no stability problems. Does anyone know what is the difference in the meshing approach of ICEM CFD and Gambit?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Inlet turbulent boundary condition: external flows	ROOZBEH	FLUENT	3	August 28, 2015 16:34
External Radiation Boundary Condition (Two sided wall), Grid Interface	CFD XUE	FLUENT	0	July 8, 2010 07:49
problem with boundary condition???	smn	CFX	5	November 24, 2009 07:37
how to set up a wall boundary condition according to calculated wall shear stress?	gameoverli	OpenFOAM Pre-Processing	1	May 21, 2009 09:28
New topic on same subject - Flow around race car	Tudor Miron	CFX	15	April 2, 2004 07:18

May 9, 2011, 12:32		#2
farshadexp Member Farshad Join Date: Sep 2010 Posts: 36 Rep Power: 16	Hi there.... you can also use outlet-vent in your simulation, remember that according to my experience P-outlet many times leads to R-flow. By the way, I have similar problem now, but in different simulation case and I'll be really happy if you find better answer and inform me Regards...

May 9, 2011, 14:35		#3
zhenglun.wei New Member Join Date: Jan 2011 Posts: 25 Rep Power: 15	Dear Farshadexp, thanks for your reply. Fluent manual says that the P-outlet can help us to make the simulation converge better when RF exists. However, in my opinion, it does not say that the P-outlet can help us to eliminate the reverse flow. As I tested, the P-outlet indeed help me to converge the simulation to 10^-4 (outlet BC can not even converge to 10^-3); however, the RF warning is still showing over there. I'll do some tests on your suggestions and let you know if it can fit my problem. best,

October 20, 2011, 14:10		#4
choonhwai New Member Choon Hwai Yap Join Date: Oct 2011 Posts: 4 Rep Power: 15	Hi, I have the same problem. I have 1 inlet and 4 outlets. Both outflow and outlet-vent BC did not correct the non-convergence and the slowly growing number of reverse flow faces. I tried refining the mesh and re-meshing with aggressive smoothing, but neither worked. will appreciate any suggestions..