When periodicity should not be used?

shivasluzz · November 6, 2016, 14:38

Is there any situation where periodicity should not be applied even when the model axisymmetric?
(Like can we apply periodicity for multiphase problem involving cavitation)

ghorrocks · November 6, 2016, 18:39

For 2D axisymmetric simulations with no swirl along the axis it is usual to apply symmetry boundaries on the top and bottom face, not periodic boundaries. You use periodic boundaries when you have a 2D axisymmetric geometry but the flow is swirling in the third dimension.

Red Ember · November 7, 2016, 01:45

Well, I crossed somewhere that simulating jets reuires full volume to predict Coanda effect if it takes place

ghorrocks · November 7, 2016, 01:51

This comment appears to have nothing to do with the original question.

Many fluid flows require a 3D to model them.

Red Ember · November 7, 2016, 02:50

If I got it right, shivasluzz had asked about examples of tasks with axisymmetric fluid domain that imply 360-degrees statement for correct solution... And jets simulation requires such statement, not 2D or sector one...

Pardon my English if I just misunderstood the question

ghorrocks · November 7, 2016, 05:59

The original question is not very clear, I am not completely sure what it is asking. So it is up to Sluzzer to state what he is asking more clearly.

shivasluzz · November 7, 2016, 07:27

Thanks for the replies ghorrocks and Red Ember..
Here I try to clarify my question.
For a given problem, how to decide whether periodic model can be used or not?
For example,
Air is flowing axial between two concentric cylinders, where outer cylinder is stationary and inner cylinder is rotating about its axis.
For this problem, I should use 360 degree model or can I use a sector with periodic interfaces?
I read that jet flow with low velocity requires full model (360 degree model) to be analysed (because of coanda effect).
Is there any other limitations or conditions are there to use sector model with periodic interfaces!?
Whether flow nature should be axisymmetric?
For example, if I expect some eddies only at particular region (in theta direction), where model is axisymmetric, can I use a sector with periodic interfaces or should I use the full 360 degree model?

ghorrocks · November 7, 2016, 17:42

You are talking about Taylor-Couette flow, and this is a standard test case in fluid dynamics. It has been studied and simulated by many researchers.

https://en.wikipedia.org/wiki/Taylor...93Couette_flow
https://www.math.vt.edu/people/renar...ouette_jfm.pdf
ftp://ftp.nist.gov/pub/itl/div891/thesis-final.pdf

The answer to your question of whether it can be modelled as a segment or full 3D, and do you use periodic boundaries is - it depends! You can see that for some flow regimes the flow is 2D rotational. This can be modelled with a 2D mesh segment with periodic boundaries on top and bottom faces. But in other regimes the flow generates rolling vorticies, and these will require a 3D segment to model. In higher regimes as well the vorticies have 3D characteristics so you would need to model the full circumference for this - and finally it goes unstable and you need a full 3D transient model for that.

shivasluzz · November 10, 2016, 08:50

Thanks for the explanation ghorrocks..

My question is not about choosing between 2D and 3D model, but between sector model (for example, 30 degree model) and full 360 degree model.

And just came to know about Taylor-Couette flow from your answer and it is so interesting!
Actually my problem is about analysing leakage flow rate of "Labyrinth seal", in which shaft with teeth rotates against a stationary wall. As the air flow over the teeth, vortex gets generated and the pressure will drop and so leakage will reduce.

Is this phenomenon is same Taylor-Couette flow?
Can it be analyzed using sector (30 or 45 degree) model with periodic interfaces?

Thanks

ghorrocks · November 10, 2016, 17:12

It is a type of Taylor-Couette flow, yes. It is important to be aware of basic fluid mechanics benchmark cases like this and the various flow regimes they have as you will find many general fluid flow cases are similar in nature to the benchmark cases, and that will allow you to estimate what flow is likely to be generated and therefore what is required to simulate it.

My post #8 does describe whether a sector model or a full 360 degree model is required. The answer is "it depends on which flow regime you are in". Low flows are axisymmetric so can be modelled 2D. Moderate flows have 3 dimensional rolling vorticies which theoretically can be modelled as sectors but in practise this constrains them too much so you really need to model them as 3. High flows have full 3 dimensional structures which can only be modelled as full 360 degree geometries.

November 6, 2016, 14:38	When periodicity should not be used?	#1
shivasluzz Senior Member sluzzer Join Date: Jul 2014 Posts: 146 Rep Power: 12	Is there any situation where periodicity should not be applied even when the model axisymmetric? (Like can we apply periodicity for multiphase problem involving cavitation)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ICEM] How define periodicity correctly in ICEM	FJSJ	ANSYS Meshing & Geometry	18	January 11, 2023 12:07
[ANSYS Meshing] Periodicity problems in icem	zeeshu	ANSYS Meshing & Geometry	0	April 17, 2016 21:59
[ICEM] ICEM periodicity specification	robboflea	ANSYS Meshing & Geometry	18	November 11, 2014 15:51
[ICEM] Two kinds of periodicity in one project	PS_KA	ANSYS Meshing & Geometry	3	January 22, 2014 05:09
[ICEM] Hybrid Meshing and Periodicity	RalfBecker	ANSYS Meshing & Geometry	0	October 31, 2013 07:31

November 6, 2016, 18:39		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	For 2D axisymmetric simulations with no swirl along the axis it is usual to apply symmetry boundaries on the top and bottom face, not periodic boundaries. You use periodic boundaries when you have a 2D axisymmetric geometry but the flow is swirling in the third dimension.

November 7, 2016, 01:45		#3
Red Ember Member Alex Join Date: Feb 2016 Posts: 81 Rep Power: 10	Well, I crossed somewhere that simulating jets reuires full volume to predict Coanda effect if it takes place

November 7, 2016, 01:51		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	This comment appears to have nothing to do with the original question. Many fluid flows require a 3D to model them.

November 7, 2016, 02:50		#5
Red Ember Member Alex Join Date: Feb 2016 Posts: 81 Rep Power: 10	If I got it right, shivasluzz had asked about examples of tasks with axisymmetric fluid domain that imply 360-degrees statement for correct solution... And jets simulation requires such statement, not 2D or sector one... Pardon my English if I just misunderstood the question

November 7, 2016, 05:59		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	The original question is not very clear, I am not completely sure what it is asking. So it is up to Sluzzer to state what he is asking more clearly.

November 7, 2016, 07:27		#7
shivasluzz Senior Member sluzzer Join Date: Jul 2014 Posts: 146 Rep Power: 12	Thanks for the replies ghorrocks and Red Ember.. Here I try to clarify my question. For a given problem, how to decide whether periodic model can be used or not? For example, Air is flowing axial between two concentric cylinders, where outer cylinder is stationary and inner cylinder is rotating about its axis. For this problem, I should use 360 degree model or can I use a sector with periodic interfaces? I read that jet flow with low velocity requires full model (360 degree model) to be analysed (because of coanda effect). Is there any other limitations or conditions are there to use sector model with periodic interfaces!? Whether flow nature should be axisymmetric? For example, if I expect some eddies only at particular region (in theta direction), where model is axisymmetric, can I use a sector with periodic interfaces or should I use the full 360 degree model?

November 7, 2016, 17:42		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	You are talking about Taylor-Couette flow, and this is a standard test case in fluid dynamics. It has been studied and simulated by many researchers. https://en.wikipedia.org/wiki/Taylor...93Couette_flow https://www.math.vt.edu/people/renar...ouette_jfm.pdf ftp://ftp.nist.gov/pub/itl/div891/thesis-final.pdf The answer to your question of whether it can be modelled as a segment or full 3D, and do you use periodic boundaries is - it depends! You can see that for some flow regimes the flow is 2D rotational. This can be modelled with a 2D mesh segment with periodic boundaries on top and bottom faces. But in other regimes the flow generates rolling vorticies, and these will require a 3D segment to model. In higher regimes as well the vorticies have 3D characteristics so you would need to model the full circumference for this - and finally it goes unstable and you need a full 3D transient model for that.

November 10, 2016, 08:50		#9
shivasluzz Senior Member sluzzer Join Date: Jul 2014 Posts: 146 Rep Power: 12	Thanks for the explanation ghorrocks.. My question is not about choosing between 2D and 3D model, but between sector model (for example, 30 degree model) and full 360 degree model. And just came to know about Taylor-Couette flow from your answer and it is so interesting! Actually my problem is about analysing leakage flow rate of "Labyrinth seal", in which shaft with teeth rotates against a stationary wall. As the air flow over the teeth, vortex gets generated and the pressure will drop and so leakage will reduce. Is this phenomenon is same Taylor-Couette flow? Can it be analyzed using sector (30 or 45 degree) model with periodic interfaces? Thanks

November 10, 2016, 17:12		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	It is a type of Taylor-Couette flow, yes. It is important to be aware of basic fluid mechanics benchmark cases like this and the various flow regimes they have as you will find many general fluid flow cases are similar in nature to the benchmark cases, and that will allow you to estimate what flow is likely to be generated and therefore what is required to simulate it. My post #8 does describe whether a sector model or a full 360 degree model is required. The answer is "it depends on which flow regime you are in". Low flows are axisymmetric so can be modelled 2D. Moderate flows have 3 dimensional rolling vorticies which theoretically can be modelled as sectors but in practise this constrains them too much so you really need to model them as 3. High flows have full 3 dimensional structures which can only be modelled as full 360 degree geometries.