SRF (Single Rotating Frame)

sasanghomi · November 6, 2018, 05:03

Dear friends,

Is it possible (in CFX) to simulate a rotating domain without any stationary region?
In other words, can I simulate a rotating domain by using SRF without any stationary domain or boundaries?

Best Regards
Sasan Ghomi

Gert-Jan · November 6, 2018, 06:42

Yes. For example, you can setup a simple cilindrical volume and let it rotate around its central axis. It can contain parts that rotate with the cilindrical volume.

On top of of that, you can set the outer wall to be counter rotating. In such a case it is standing still, effectively. Then you have a stationary object in your rotating domain. So, without having a stationary domain, you can have a stationary object. But you don't need that, if I read it correctly.

sasanghomi · November 6, 2018, 07:59

Thank you.

Quote:

Originally Posted by Gert-Jan

On top of of that, you can set the outer wall to be counter rotating. In such a case it is standing still, effectively. Then you have a stationary object in your rotating domain.

What do you mean by this?
Is there any need for stationary boundaries?

Gert-Jan · November 6, 2018, 08:36

Strictly there is no need. But then you only get a solid body rotation. You don't need CFD for that.

So what do you want to simulate?

sasanghomi · November 11, 2018, 00:36

Thank you.
Regarding my simulation,
I am simulating the fluid flow inside a pipe which is rotating around an axis.
Can I define the whole domain as a rotating domain? In doing so, there would not be any need for frozen rotor interfaces.
So, I suppose that such simulation is called Single Rotating Frame (SRF).

Best Regards

ghorrocks · November 11, 2018, 03:07

Yes, this model is possible. Yes, you define the whole fluid region as a rotating domain. Which axis is the pipe rotating about?

sasanghomi · November 11, 2018, 03:45

Does it make any difference? The pipe is rotating around an axis parallel to it own axis.
Please let me know if there is any difference!!

ghorrocks · November 11, 2018, 15:59

In this case you have to be careful to define inlet and outlet boundary conditions to account for the frame change. But it should work fine as long as that is sorted.

Alisa_W · January 22, 2019, 06:30

It is said that SRF can only simulate steady problem, so I am not sure if it is suitable fot your problem.

Quote:

Originally Posted by sasanghomi

Does it make any difference? The pipe is rotating around an axis parallel to it own axis.
Please let me know if there is any difference!!

Gert-Jan · January 22, 2019, 07:00

That is not true. It is possible to run a transient case with single rotating of reference. I did it yesterday,.......

AtoHM · January 22, 2019, 07:07

Out of curiosity, I wonder if it is possible to break it down to a stationary case? If the rotation rate is very high, maybe gravitational effects can be neglected even if the axis does not coincide with the height-axis. Then the rotation can be accounted for by a constant "radial" force which should be easy to implement with another coordinate system coinciding with the rotation axis. This would allow for easier definition of other BCs. Maybe I miss something though.

Alisa_W · January 22, 2019, 08:41

Actually, I am running the case. I used SRF and got a steady result. Then, when I change the simulation type to transient. the scaled residual line seemed unchanged. However, when I changed the 'cell zone condition' to 'Mesh Motion', the line seemed more normal(lines ups and downs). So I thought SRF may be not suitable for unsteady simulation.

My software is FLUENT and case is rotating channel. For SRF, 'cell zone condition' is 'Frame Motion'.

Quote:

Originally Posted by Gert-Jan

That is not true. It is possible to run a transient case with single rotating of reference. I did it yesterday,.......

Alisa_W · January 22, 2019, 08:46

It makes sense in theory. However, When I calculated a rotating channel using SRF(got a steady result). When I change the simulation type to transient. The scaled residual line seemed unchanged(abnormal). However, when I changed the 'cell zone condition' to 'Mesh Motion', the line seemed more normal(lines ups and downs).

Quote:

Originally Posted by AtoHM

Out of curiosity, I wonder if it is possible to break it down to a stationary case? If the rotation rate is very high, maybe gravitational effects can be neglected even if the axis does not coincide with the height-axis. Then the rotation can be accounted for by a constant "radial" force which should be easy to implement with another coordinate system coinciding with the rotation axis. This would allow for easier definition of other BCs. Maybe I miss something though.

November 6, 2018, 05:03	SRF (Single Rotating Frame)	#1
sasanghomi Senior Member Sasan Ghomi Join Date: Sep 2012 Location: Denmark Posts: 292 Rep Power: 14	Dear friends, Is it possible (in CFX) to simulate a rotating domain without any stationary region? In other words, can I simulate a rotating domain by using SRF without any stationary domain or boundaries? Best Regards Sasan Ghomi vavnoon likes this.

November 11, 2018, 03:07		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,719 Rep Power: 143	Yes, this model is possible. Yes, you define the whole fluid region as a rotating domain. Which axis is the pipe rotating about? __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

November 11, 2018, 15:59		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,719 Rep Power: 143	In this case you have to be careful to define inlet and outlet boundary conditions to account for the frame change. But it should work fine as long as that is sorted. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Opening frame type (rotating, stationary)?	urosgrivc	CFX	0	July 3, 2017 05:24
Boundary conditions for one rotating and one stationary frame	mrswordf1sh	FLUENT	2	May 7, 2015 02:23
Convert DFT of pressure in rotating frame to stationary frame	CFD-123	FLUENT	0	February 12, 2015 17:51
Rotating Frame Stationary Vaneless Diffuser?	khoopes	CFX	2	July 24, 2014 10:20
question about governing equation in CFX using rotating/non rotating reference frame	rystokes	CFX	0	January 12, 2010 06:14

November 6, 2018, 06:42		#2
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,835 Rep Power: 27	Yes. For example, you can setup a simple cilindrical volume and let it rotate around its central axis. It can contain parts that rotate with the cilindrical volume. On top of of that, you can set the outer wall to be counter rotating. In such a case it is standing still, effectively. Then you have a stationary object in your rotating domain. So, without having a stationary domain, you can have a stationary object. But you don't need that, if I read it correctly.

November 6, 2018, 08:36		#4
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,835 Rep Power: 27	Strictly there is no need. But then you only get a solid body rotation. You don't need CFD for that. So what do you want to simulate?

November 11, 2018, 00:36		#5
sasanghomi Senior Member Sasan Ghomi Join Date: Sep 2012 Location: Denmark Posts: 292 Rep Power: 14	Thank you. Regarding my simulation, I am simulating the fluid flow inside a pipe which is rotating around an axis. Can I define the whole domain as a rotating domain? In doing so, there would not be any need for frozen rotor interfaces. So, I suppose that such simulation is called Single Rotating Frame (SRF). Best Regards

January 22, 2019, 07:00		#10
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,835 Rep Power: 27	That is not true. It is possible to run a transient case with single rotating of reference. I did it yesterday,.......

January 22, 2019, 07:07		#11
AtoHM Senior Member M Join Date: Dec 2017 Posts: 642 Rep Power: 12	Out of curiosity, I wonder if it is possible to break it down to a stationary case? If the rotation rate is very high, maybe gravitational effects can be neglected even if the axis does not coincide with the height-axis. Then the rotation can be accounted for by a constant "radial" force which should be easy to implement with another coordinate system coinciding with the rotation axis. This would allow for easier definition of other BCs. Maybe I miss something though.