Boundary conditions for a Fan geometry

saisanthoshm88 · January 28, 2011, 05:46

I have a fan to be modeled in CFX – pre . I’m not clear about the kind of Boundary conditions to be imposed on the fan. The image attached to the post shows the Fan geometry. The fan is a part of a Heat oven geometry and is used to impart a forced convection inside the oven.
The fan has a thickness .It has a surface mesh and the internal volume of the fan is not meshed. So I wonder if the conception of treating the upper surfaces of the fan as inlet and the lower surfaces as the outlet holds good at all.
Can someone please help me out.

ghorrocks · January 28, 2011, 07:01

Last time I saw a fan the flow did not go into one side of the blade and out the other. Why do you think this is an appropriate flow boundary for a fan?

saisanthoshm88 · January 28, 2011, 07:05

I'm not exactly aware of the kind of BCs by which a fan is modeled in CFX so I regret if my conception is misleading but please suggest upon the usual approach in which the fan shown in the image can be modeled in CFX.

I found it in the CFX help that the Fans and blowers are treated similar to liquid pumps.

ghorrocks · January 28, 2011, 15:12

Quote:

I found it in the CFX help that the Fans and blowers are treated similar to liquid pumps.

That's right.....

There are examples of how to model this sort of thing in the CFX tutorials which are installed by default.

saisanthoshm88 · January 31, 2011, 02:19

Sorry, I couldn't find the right tutorial from the documentation . can you please guide me upon that (or) is it something like-- Aerodynamic & structural performance of a centrifugal compressor.

ghorrocks · January 31, 2011, 06:11

It is best to do all the tutorials. It will be time well spent.

But for your case the rotating ones are the most important - flow in an axial rotor/stator is one example.

saisanthoshm88 · February 6, 2011, 05:38

Well Glenn, I've gone through stuff like : MRF method and moving mesh method which are the conventional approaches of Modeling a Fan for CFD analysis.

But what about an approach like meshing the internal volume of the Fan and using that volume to be a rotational domain.

(or) what if I choose the Fan to be a inlet, outlet (or) both as suggested in the CFX documentation and specify the velocity in terms of some cylindrical components so that I can mention the swirl component too.

ghorrocks · February 6, 2011, 16:39

Looks like you need to sort out what you want.

I think your second line is asking can you put the fan in as a solid object, I presume so you can do CHT on analysis. In V13 yes, you can do this. It did not work properly in previous versions. But if you are not doing CHT then CFX has been able to do it for years using the rotating frames of reference approach.

Your final line is suggesting you don't care about the details of the flow in the fan, just the flow it generates. If this is the case then forget about rotating frames of reference and model the fan as a momentum source. You can include swirl components. You need the fan performance curve to do this. There have been many posts on the forum for how to do this.

saisanthoshm88 · February 6, 2011, 19:17

You are right Glenn, I don't need the details of flow in the fan and i'm only interested in the flow it generates and of course i'm not doing the CHT analysis.

But I don't have the fan performance curve and the only information I have is that the Fan rotates at a speed of 2,300 r.p.m. I'm currently considering the MRF method . As far as I know it, in the MRF method fan blades are modeled stationary, but since the fluid domain surrounding them is in a rotating frame, the pressure jump and swirl will be given by the presence of the fan blades as walls.

I've defined the fluid domain surrounding the Fan to be a rotary domain but now I couldn't model the Fan surfaces as stationary walls. Because the CFX - pre doesn't allow to define a wall to be stationary in a rotating domain .

ghorrocks · February 7, 2011, 04:33

Quote:

the only information I have is that the Fan rotates at a speed of 2,300 r.p.m.

Then I guess you better model it.

Quote:

As far as I know it, in the MRF method fan blades are modeled stationary, but since the fluid domain surrounding them is in a rotating frame, the pressure jump and swirl will be given by the presence of the fan blades as walls.

I do not know what you are talking about. Usually the fan rotates, so stick that in a rotating frame of reference, and the surrounding stuff is in a stationary frame of reference.

January 28, 2011, 07:05		#3
saisanthoshm88 Senior Member --------- Join Date: Oct 2010 Posts: 303 Rep Power: 17	I'm not exactly aware of the kind of BCs by which a fan is modeled in CFX so I regret if my conception is misleading but please suggest upon the usual approach in which the fan shown in the image can be modeled in CFX. I found it in the CFX help that the Fans and blowers are treated similar to liquid pumps. __________________ Best regards, Santhosh.

January 31, 2011, 02:19		#5
saisanthoshm88 Senior Member --------- Join Date: Oct 2010 Posts: 303 Rep Power: 17	Sorry, I couldn't find the right tutorial from the documentation . can you please guide me upon that (or) is it something like-- Aerodynamic & structural performance of a centrifugal compressor. __________________ Best regards, Santhosh.

February 6, 2011, 05:38		#7
saisanthoshm88 Senior Member --------- Join Date: Oct 2010 Posts: 303 Rep Power: 17	Well Glenn, I've gone through stuff like : MRF method and moving mesh method which are the conventional approaches of Modeling a Fan for CFD analysis. But what about an approach like meshing the internal volume of the Fan and using that volume to be a rotational domain. (or) what if I choose the Fan to be a inlet, outlet (or) both as suggested in the CFX documentation and specify the velocity in terms of some cylindrical components so that I can mention the swirl component too. __________________ Best regards, Santhosh.

February 6, 2011, 19:17		#9
saisanthoshm88 Senior Member --------- Join Date: Oct 2010 Posts: 303 Rep Power: 17	You are right Glenn, I don't need the details of flow in the fan and i'm only interested in the flow it generates and of course i'm not doing the CHT analysis. But I don't have the fan performance curve and the only information I have is that the Fan rotates at a speed of 2,300 r.p.m. I'm currently considering the MRF method . As far as I know it, in the MRF method fan blades are modeled stationary, but since the fluid domain surrounding them is in a rotating frame, the pressure jump and swirl will be given by the presence of the fan blades as walls. I've defined the fluid domain surrounding the Fan to be a rotary domain but now I couldn't model the Fan surfaces as stationary walls. Because the CFX - pre doesn't allow to define a wall to be stationary in a rotating domain . __________________ Best regards, Santhosh. Last edited by saisanthoshm88; February 7, 2011 at 02:30.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
hexagonal boundary conditions	s_h	Main CFD Forum	0	August 9, 2009 22:15
Boundary conditions for rotating fan	rohit	Siemens	3	October 4, 2008 04:18
Boundary conditions for rotating fan	rohit	Fidelity CFD	3	September 24, 2008 07:18
Water vapour condensation in CFX-5.7.1	hdj	CFX	1	November 27, 2005 07:15
Convective Heat Transfer - Heat Exchanger	Mark	CFX	6	November 15, 2004 15:55

January 28, 2011, 07:01		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	Last time I saw a fan the flow did not go into one side of the blade and out the other. Why do you think this is an appropriate flow boundary for a fan?

January 31, 2011, 06:11		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	It is best to do all the tutorials. It will be time well spent. But for your case the rotating ones are the most important - flow in an axial rotor/stator is one example.

February 6, 2011, 16:39		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,701 Rep Power: 143	Looks like you need to sort out what you want. I think your second line is asking can you put the fan in as a solid object, I presume so you can do CHT on analysis. In V13 yes, you can do this. It did not work properly in previous versions. But if you are not doing CHT then CFX has been able to do it for years using the rotating frames of reference approach. Your final line is suggesting you don't care about the details of the flow in the fan, just the flow it generates. If this is the case then forget about rotating frames of reference and model the fan as a momentum source. You can include swirl components. You need the fan performance curve to do this. There have been many posts on the forum for how to do this.