simulate cross section of 3D fully Devolped poiseuille rectangular closed channel

amin144 · July 2, 2012, 17:57

Hi Dear FOAMers

I want to simulate one cross section (2D) of a 3D rectangular closed channel in fully developed poiseuille flow.
I know should use cyclic BC but I have no idea to use it.

Anybody can upload me a case for study and learning ?

amin144 · July 2, 2012, 18:11

I don't know if cyclic BC is useable for poiseuille flow?
How should I implement pressure gradient?
Pressure is not cyclic in this flow!

amin144 · July 3, 2012, 03:28

NoBody has done some case like this?

Ohbuchi · July 3, 2012, 04:20

Hi, Amin.

Please check "channelFoam" source code and "channel395" tutorial.
The channelFoam can handle pressure gradient and cyclic boundary in streamwise direction.
If you want to simulate duct flow, the boundary condition for span direction should be changed to non-slip wall from cyclic boundary.

amin144 · July 3, 2012, 07:16

Hello Dear Masashi

Tanx for your quick answer
What should I do with pressure BC?

Thank you very much...

Ohbuchi · July 3, 2012, 20:38

Hi, Amin,

Pressure gradient is coded in solver source code as follows.

channelFoam.C
00063 fvVectorMatrix UEqn
00064 (
00065 fvm::ddt(U)
00066 + fvm::div(phi, U)
00067 + sgsModel->divDevBeff(U)
00068 ==
00069 flowDirection*gradP <-- here
00070 );
00071
00072 if (momentumPredictor)
00073 {
00074 solve(UEqn == -fvc::grad(p));
00075 }

So the pressure boundary condition at inlet-outlet should be cyclic.

amin144 · July 4, 2012, 03:54

Hi Dear Masashi

Tanx for your answer
It seems this way doesn't work for another solver? am I right?

So what should I do if I want simulate one cross secttion normal to basis axis of a fully developed rectangular duct with 2d mesh? (my flow is pressure driven)

Best wishes for you

July 2, 2012, 17:57	simulate cross section of 3D fully Devolped poiseuille rectangular closed channel	#1
amin144 Member Amin Shariat KHah Join Date: Apr 2011 Location: Shiraz Posts: 86 Rep Power: 15	Hi Dear FOAMers I want to simulate one cross section (2D) of a 3D rectangular closed channel in fully developed poiseuille flow. I know should use cyclic BC but I have no idea to use it. Anybody can upload me a case for study and learning ?

July 3, 2012, 07:16		#5
amin144 Member Amin Shariat KHah Join Date: Apr 2011 Location: Shiraz Posts: 86 Rep Power: 15	Hello Dear Masashi Tanx for your quick answer What should I do with pressure BC? Thank you very much... Last edited by amin144; July 3, 2012 at 09:20.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Neumann pressure BC and velocity field	Antech	Main CFD Forum	0	April 25, 2006 02:15
pressure driven flow by pressure correction method	justentered	Main CFD Forum	0	December 29, 2003 23:52
Gas pressure question	Dan Moskal	Main CFD Forum	0	October 24, 2002 22:02
pressure gradient term in low speed flow	Atit Koonsrisuk	Main CFD Forum	2	January 10, 2002 10:52
Hydrostatic pressure in 2-phase flow modeling (long)	DS & HB	Main CFD Forum	0	January 8, 2000 15:00

July 2, 2012, 18:11		#2
amin144 Member Amin Shariat KHah Join Date: Apr 2011 Location: Shiraz Posts: 86 Rep Power: 15	I don't know if cyclic BC is useable for poiseuille flow? How should I implement pressure gradient? Pressure is not cyclic in this flow!

July 3, 2012, 03:28		#3
amin144 Member Amin Shariat KHah Join Date: Apr 2011 Location: Shiraz Posts: 86 Rep Power: 15	NoBody has done some case like this?

July 3, 2012, 04:20		#4
Ohbuchi Member Masashi Ohbuchi Join Date: Oct 2009 Posts: 74 Rep Power: 16	Hi, Amin. Please check "channelFoam" source code and "channel395" tutorial. The channelFoam can handle pressure gradient and cyclic boundary in streamwise direction. If you want to simulate duct flow, the boundary condition for span direction should be changed to non-slip wall from cyclic boundary.

July 3, 2012, 20:38		#6
Ohbuchi Member Masashi Ohbuchi Join Date: Oct 2009 Posts: 74 Rep Power: 16	Hi, Amin, Pressure gradient is coded in solver source code as follows. channelFoam.C 00063 fvVectorMatrix UEqn 00064 ( 00065 fvm::ddt(U) 00066 + fvm::div(phi, U) 00067 + sgsModel->divDevBeff(U) 00068 == 00069 flowDirection*gradP <-- here 00070 ); 00071 00072 if (momentumPredictor) 00073 { 00074 solve(UEqn == -fvc::grad(p)); 00075 } So the pressure boundary condition at inlet-outlet should be cyclic.

July 4, 2012, 03:54		#7
amin144 Member Amin Shariat KHah Join Date: Apr 2011 Location: Shiraz Posts: 86 Rep Power: 15	Hi Dear Masashi Tanx for your answer It seems this way doesn't work for another solver? am I right? So what should I do if I want simulate one cross secttion normal to basis axis of a fully developed rectangular duct with 2d mesh? (my flow is pressure driven) Best wishes for you