Heated Walls

vitor · August 12, 2010, 09:56

Hi,

I'm performing a simulation both in OpenFOAM, Fluent and Code-Saturne with the same conditions and the same mesh.
The case is very simple: an axial flow in a straight pipe with heated walls. The conditions are:

Inlet: 0,05 m/s; 313K ; pressure: zeroGradient
Wall: 500K or 2000K; pressure zeroGradient
Outlet: Zero gradient for all; except for pressure: fixedValue = 0.

I initialized turbulence using the equations in the User Guide (Lid-Driven cavity tutorial).

The issue is that, although Code-Saturne and Fluent give me the same results, I'm getting different results from OF. As you can see below:

Wall Temperature Outlet averaged Temp (Fluent) Outlet averaged Temp (OF)
[K] [K] [K]

500 315,7 319,4
2000 336,2 370,0

I'm using the bouyantBouusinesqSimpleFoam. Can anybody give me a hint about what might be happening?
If you really want to help, please try to run the simulation yourself (it's very quick) and check if you get the same solution.

Thank you.

BR,
Vitor.

KrisT · August 13, 2010, 03:21

What are the dimensions of the pipe? Is turbulence needed, the speed is so low, isn't the flow laminar?

vitor · August 13, 2010, 08:11

The diameter is 6m (about 20ft) and the lenght 20m (65.6ft), it gives a Reynolds number (for water) equal to 0,05*6/(1e-6) = 3e5 , wich is a lot greater than 2300, so the flow must be turbulent.

Thank you.

KrisT · August 14, 2010, 04:12

Big pipes.

One thing you could do would be to do the simulations on a smaller domain, where the flow is laminar. If you get the same results with the different solvers then you know it is related to the turbulence model.

vitor · August 14, 2010, 13:05

Good idea.

I think I`ll give it a try.

Thanks.

vitor · August 16, 2010, 14:47

What puzzles me is that I've performed the same simulation with a heat flux at the wall instead of temperature at the wall, and it gave me the same results as Fluent and Code-Saturne.

But in the heat flux case I specified the cp (specific heat) of the water, and in the other case I only specified the Prandtl number.

August 12, 2010, 09:56	Heated Walls	#1
vitor New Member Vitor Braga Join Date: Oct 2009 Posts: 28 Rep Power: 16	Hi, I'm performing a simulation both in OpenFOAM, Fluent and Code-Saturne with the same conditions and the same mesh. The case is very simple: an axial flow in a straight pipe with heated walls. The conditions are: Inlet: 0,05 m/s; 313K ; pressure: zeroGradient Wall: 500K or 2000K; pressure zeroGradient Outlet: Zero gradient for all; except for pressure: fixedValue = 0. I initialized turbulence using the equations in the User Guide (Lid-Driven cavity tutorial). The issue is that, although Code-Saturne and Fluent give me the same results, I'm getting different results from OF. As you can see below: Wall Temperature Outlet averaged Temp (Fluent) Outlet averaged Temp (OF) [K] [K] [K] 500 315,7 319,4 2000 336,2 370,0 I'm using the bouyantBouusinesqSimpleFoam. Can anybody give me a hint about what might be happening? If you really want to help, please try to run the simulation yourself (it's very quick) and check if you get the same solution. Thank you. BR, Vitor.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Pipe with heated walls	vitor	OpenFOAM Running, Solving & CFD	3	July 15, 2010 16:20
wallHeatFlux for heated walls in laminar channel flow	svens	OpenFOAM Post-Processing	0	June 20, 2009 20:33
Radiation between heated rods	lego	Main CFD Forum	1	November 13, 2003 08:39
Radiation between heated rods	lego	CFX	0	November 12, 2003 08:44
thermal conditions of walls	Stefan	FLUENT	2	March 23, 2003 06:17

August 13, 2010, 03:21		#2
KrisT Member Join Date: May 2009 Posts: 32 Rep Power: 16	What are the dimensions of the pipe? Is turbulence needed, the speed is so low, isn't the flow laminar?

August 13, 2010, 08:11		#3
vitor New Member Vitor Braga Join Date: Oct 2009 Posts: 28 Rep Power: 16	The diameter is 6m (about 20ft) and the lenght 20m (65.6ft), it gives a Reynolds number (for water) equal to 0,05*6/(1e-6) = 3e5 , wich is a lot greater than 2300, so the flow must be turbulent. Thank you.

August 14, 2010, 04:12		#4
KrisT Member Join Date: May 2009 Posts: 32 Rep Power: 16	Big pipes. One thing you could do would be to do the simulations on a smaller domain, where the flow is laminar. If you get the same results with the different solvers then you know it is related to the turbulence model.

August 14, 2010, 13:05		#5
vitor New Member Vitor Braga Join Date: Oct 2009 Posts: 28 Rep Power: 16	Good idea. I think I`ll give it a try. Thanks.

August 16, 2010, 14:47		#6
vitor New Member Vitor Braga Join Date: Oct 2009 Posts: 28 Rep Power: 16	What puzzles me is that I've performed the same simulation with a heat flux at the wall instead of temperature at the wall, and it gave me the same results as Fluent and Code-Saturne. But in the heat flux case I specified the cp (specific heat) of the water, and in the other case I only specified the Prandtl number.