Heat transfer between solid rotating and stationary fluid domains

DCSERE · November 10, 2015, 03:31

Hi,

I would like to calculate the heat transfer coefficient between a solid rotating (rod) and stationary fluid domain (air) in CFX. I defined 600°C to the center surfaces of the rod (see the attached figure) and I want to get the temperature and the heat transfer coefficient distribution of the entire rod. A 20km/h velocity also was defined at the inlet.
The results don’t seem to be realistic because the temperature distribution of the rod is far from the axisymmetric. The airflow around the rod is not axisymmetric of course because 20 km/h also was considered. Since the rod rotates I expected an axisymmetric temperature distribution on its surface.
I also completed an analysis where the rotation of the rod wasn’t considered only the 20km/h, I got almost the same results. So it seems to me that the rotating domains is not accurate enough for the cases when we would like to calculate the heat transfer between a rotating solid domain and a stationary fluid domain.
I used frozen method, because only this option is available for fluid-solid interface in a steady state analysis.
Does anybody have any experience with the same problem?
Thank You in advance!

ghorrocks · November 10, 2015, 20:52

First of all, have you read the FAQ on accuracy? http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

What version of CFX are you using? Note that rotating solid domains is only supported from V15 onwards I think.

DCSERE · November 17, 2015, 04:43

Dear ghorrocks,

Thank you for your answer. I am using Ansys V15.
Finally I found the reason of the problem, which was the following:
-in the first case I defined the rod as a stationary solid domain, then I changed the domain to the rotating one and I also set the option of the frame change of the interface from none to frozen rotor. Unfortunately I got the same results, so the rotation of the rod wasn’t considered in spite of the frozen rotor and the rotation of the rod were activated.
Finally I deleted the interface between the rod and the air and I defined it again using frozen rotor. In this case it worked well.
The conclusion is the following: if you define an interface for the first time using option none then you change it to frozen rotor the solver will calculate with the option none (I don’t know why). So you have to delete the interface and define it again using frozen rotor immediately.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convective term in heat equation for rotating solids	vabishek	OpenFOAM Programming & Development	8	June 24, 2020 12:10
Can CFX do CHT simulations with a solid domain rotating in a stationary fluid domain?	acro	CFX	15	September 23, 2016 12:16
Wrong flow in ratating domain problem	Sanyo	CFX	17	August 15, 2015 07:20
An error has occurred in cfx5solve:	volo87	CFX	5	June 14, 2013 18:44
rotating solid zone with heat transfer	Sasha	FLUENT	1	March 24, 2009 17:55

November 10, 2015, 20:52		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	First of all, have you read the FAQ on accuracy? http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F What version of CFX are you using? Note that rotating solid domains is only supported from V15 onwards I think.

November 17, 2015, 04:43		#3
DCSERE New Member Sandor Sandor Join Date: Jan 2015 Posts: 3 Rep Power: 11	Dear ghorrocks, Thank you for your answer. I am using Ansys V15. Finally I found the reason of the problem, which was the following: -in the first case I defined the rod as a stationary solid domain, then I changed the domain to the rotating one and I also set the option of the frame change of the interface from none to frozen rotor. Unfortunately I got the same results, so the rotation of the rod wasn’t considered in spite of the frozen rotor and the rotation of the rod were activated. Finally I deleted the interface between the rod and the air and I defined it again using frozen rotor. In this case it worked well. The conclusion is the following: if you define an interface for the first time using option none then you change it to frozen rotor the solver will calculate with the option none (I don’t know why). So you have to delete the interface and define it again using frozen rotor immediately.