heat transfer validation problem

messbalint · March 31, 2012, 04:42

Hi!

I try to validate the heat transfer part of one of my cfx model. I use an article where they blew a thin hot plate by an axisymmetric air jet and they get the Nusselt numbers radial distribution.

I use SST turbulence model on an axisymmetric structured mesh.

I have three kind of problems with two configurations.

First I tried to use a simple wall boundary condition with heat flux to model the plate. It overestimates the Nusselt number (about 150 percent).

Secondly I tried to model the plate as a solid body (because of the radial conduction) with volume energy source (by using a subdomain) but at this case there were only really small part of heat flux coming through the solid-fluid interface.

And thirdly I defined wall with heat flux source at the lower side (so not on the interface) of the plate. In this case it seemed to be working but the model underestimated the Nusselt numbers.

I tried the next: tetra mesh, total energy model, thermal energy model and I checked my y+ which is under 1.

I do not use any heat transfer interface model because I do not have any thin material on my plate and I do looking for the heat transfer coefficient.

Am I right? What kind of turbulence model should I use?

Thank you for your help!

messbalint

messbalint · March 31, 2012, 04:44

I forgot that the Reynolds number is 1.5e4, my nozzle diameter is d=20 mm and the plate-to-nozzle distance is 12*d.

ghorrocks · March 31, 2012, 07:27

Quote:

I checked my y+ which is under 1.

This does not imply your mesh is adequate. Do a mesh sensitivity study to be sure.

Assuming you have set the simulation up correctly: One issue I can see is that your assumption of axial symmetry might be wrong. In the Re range you are modelling what do you expect the jet to do? If you are getting into the region where you get the jet moving about then you will need a 3D transient simulation, and do a time average to get the heat transfer rate.

Also be aware that errors of 50% are typical for CFD heat transfer simulations.

messbalint · March 31, 2012, 09:05

Thank you for your answer Glenn Horrocks!

I think it should be axisymmetrical but I will try the transient 3D as well (I did not write, but my simulations were steady states).

Do you know anything about the volume source? This is my first time I use it and it seems to be simple, so I don't know where could I make my mistake.There wasn't so much about it in the help so I just made a subdomain in the solid domain and defined a volume source. Is there any special option?

messbalint

ghorrocks · March 31, 2012, 16:14

Before doing the 3D transient simulation I would check the literature and see if you expect 3D transient structures at this flow regime. That will tell you if this is required or not.

Volume sources are pretty straight forwrd and your description sounds correct. It can be tricky developing the equations as the units of the source term are not always obvious.

March 31, 2012, 04:42	heat transfer validation problem	#1
messbalint New Member Mess Balint Join Date: Oct 2010 Posts: 6 Rep Power: 15	Hi! I try to validate the heat transfer part of one of my cfx model. I use an article where they blew a thin hot plate by an axisymmetric air jet and they get the Nusselt numbers radial distribution. I use SST turbulence model on an axisymmetric structured mesh. I have three kind of problems with two configurations. First I tried to use a simple wall boundary condition with heat flux to model the plate. It overestimates the Nusselt number (about 150 percent). Secondly I tried to model the plate as a solid body (because of the radial conduction) with volume energy source (by using a subdomain) but at this case there were only really small part of heat flux coming through the solid-fluid interface. And thirdly I defined wall with heat flux source at the lower side (so not on the interface) of the plate. In this case it seemed to be working but the model underestimated the Nusselt numbers. I tried the next: tetra mesh, total energy model, thermal energy model and I checked my y+ which is under 1. I do not use any heat transfer interface model because I do not have any thin material on my plate and I do looking for the heat transfer coefficient. Am I right? What kind of turbulence model should I use? Thank you for your help! messbalint

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Natural Convection heat transfer problem	srinivasa	FLUENT	21	November 11, 2016 06:08
simple heat transfer problem	openfoam1	OpenFOAM	2	January 27, 2010 14:17
Wall heat transfer coefficient (HTC) problem	Mohamed khamis	CFX	1	January 15, 2010 23:12
Conjugate heat transfer problem with porous media	piko	Siemens	1	April 17, 2009 15:41
How can I increase Heat Transfer at Domain Interf?	B.Simon	CFX	3	October 28, 2008 18:53

March 31, 2012, 04:44		#2
messbalint New Member Mess Balint Join Date: Oct 2010 Posts: 6 Rep Power: 15	I forgot that the Reynolds number is 1.5e4, my nozzle diameter is d=20 mm and the plate-to-nozzle distance is 12*d.

March 31, 2012, 09:05		#4
messbalint New Member Mess Balint Join Date: Oct 2010 Posts: 6 Rep Power: 15	Thank you for your answer Glenn Horrocks! I think it should be axisymmetrical but I will try the transient 3D as well (I did not write, but my simulations were steady states). Do you know anything about the volume source? This is my first time I use it and it seems to be simple, so I don't know where could I make my mistake.There wasn't so much about it in the help so I just made a subdomain in the solid domain and defined a volume source. Is there any special option? messbalint

March 31, 2012, 16:14		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Before doing the 3D transient simulation I would check the literature and see if you expect 3D transient structures at this flow regime. That will tell you if this is required or not. Volume sources are pretty straight forwrd and your description sounds correct. It can be tricky developing the equations as the units of the source term are not always obvious.