# Simulation of convective heat loss from cylinder

 Register Blogs Members List Search Today's Posts Mark Forums Read

April 27, 2015, 07:07
Simulation of convective heat loss from cylinder
#1
New Member

buidu
Join Date: Dec 2014
Posts: 27
Rep Power: 12
Hi all,

I am running a simulation with flow passing a finite cylinder in order to get the overall covective coefficient of the cylinder.

Unstructural mesh was adopted for the calculation domain, and the elements near the cylinder is much denser than other parts. The distance of the inlet, top wall and downwall to the cylinder is 5 times of diameter, and the distance of the outlet to the cylinder is 10 times of diameter.

I ran the simulation with inlet air velocity of 0.2 and 3m/s using RSM model in steady condition.

Comparing with empirical euqation that calculating the Nu number, the result agree well in lower speed but much higher overestimated in case of 3m/s, around 20%.

Does anyone knows how should I improve the calculation result. Thank you very much!!!

Regards,
Hao
Attached Images
 domain.jpg (74.4 KB, 20 views)

 April 27, 2015, 12:20 #2 Senior Member   Lane Carasik Join Date: Aug 2014 Posts: 692 Rep Power: 14 Did you refine your mesh with increasing speed/Reynolds number?

April 28, 2015, 03:57
#3
New Member

buidu
Join Date: Dec 2014
Posts: 27
Rep Power: 12
Quote:
 Originally Posted by lcarasik Did you refine your mesh with increasing speed/Reynolds number?
Hi Lane,

I did not refine my mesh. But I checked the Y+, they are below 1 in all the speed on the surface of the cylinder.

Hao

 April 28, 2015, 04:02 #4 Senior Member   Lane Carasik Join Date: Aug 2014 Posts: 692 Rep Power: 14 What wall treatment are you using?

April 28, 2015, 04:53
#5
New Member

buidu
Join Date: Dec 2014
Posts: 27
Rep Power: 12
Quote:
 Originally Posted by lcarasik What wall treatment are you using?
I used the enhaced wall treatment, the calculated Y+ were all below 1 at varied speeds.

 April 28, 2015, 04:58 #6 Member     Christian Join Date: Sep 2013 Location: Germany Posts: 88 Rep Power: 12 I think you need to refine further downwind. Make a cutplot of the velocity with visible mesh.

 April 28, 2015, 05:10 #7 New Member   Marion Join Date: Apr 2013 Posts: 15 Rep Power: 12 Which heat-flux model are you using? And I agree with Christian, you need to refine your mesh downstream

April 28, 2015, 05:55
#8
New Member

buidu
Join Date: Dec 2014
Posts: 27
Rep Power: 12
Quote:
 Originally Posted by Chris_321 I think you need to refine further downwind. Make a cutplot of the velocity with visible mesh.
Yes, I will have a try. But why the wake will affect the convection on the cylinder?

I think the turbulence intensity will also affect the final result. When I ran with TI in 5%, it can convergence very well. However, when I reduce the TI to 0.1%, it showed that the turbulence viscousity ratio exceed 100000. You can see the contour in attachment. The high ration zone happened in front of and behind the cylinder. How do you think I should solve this problem?

Thank you.
Attached Images
 tu.jpg (13.9 KB, 6 views)

April 28, 2015, 06:02
#9
New Member

buidu
Join Date: Dec 2014
Posts: 27
Rep Power: 12
Quote:
 Originally Posted by cymbourne Which heat-flux model are you using? And I agree with Christian, you need to refine your mesh downstream
Thank you. I will do that. But do you think the turbulence intensity will also affect the results. I tried to run the case in low TI. However, the turbulence viscosity ratio reached its limitation in the zone near the cylinder. Do you think if I need to change other parameter of setting besides refine the mesh?

 April 28, 2015, 08:11 #10 Member     Christian Join Date: Sep 2013 Location: Germany Posts: 88 Rep Power: 12 Check the visosity ratio after refining your mesh. Regards