Applying drag force in the fluid model

CanadaFlanker · May 5, 2014, 17:06

Hi,

I am simulating a very small hair (diameter of 10 micrometers) in the fluid flow. The flow around it is Stokes flow as the Reynolds number based on the diameter is less than 5. It is unrealistic to make the grid smaller than the diameter in order to capture boundary layer around the hair and obtain drag force; an alternative way is to manually put a drag force on the points where the hair should be, as the drag force can be calculated from Stokes flow. Is there a way to apply a force at a point in CFX? I have searched about sources but it seems that source points don't have force or momentum, and the only place where I can put momentum source is subdomain. If anyone knows whether CFX can do it please let me know, otherwise I will switch to other software such as OpenFoam. Thank you.

ghorrocks · May 5, 2014, 21:06

I think you have to define a volume for this and use a momentum source. Make the source small so it is only one element across - don't forget that source "points" act on a control volume anyway, so you are doing the same thing as a source "point" by doing this.

CanadaFlanker · May 5, 2014, 22:03

Quote:

Originally Posted by ghorrocks

I think you have to define a volume for this and use a momentum source. Make the source small so it is only one element across - don't forget that source "points" act on a control volume anyway, so you are doing the same thing as a source "point" by doing this.

Hi Glen,

Thanks for the reply. By defining a volume, do you mean define a volume as a subdomain to present the hair? Does this volume have to be physically there (acting like a wall, which I don't think so since it generate other drag forces) or just one overlapped in the fluid flow?

Also, I am actually trying to do a FSI between the hair and the flow, is it possible to make this volume deformable? Thank you!

ghorrocks · May 5, 2014, 22:30

That changes things....

How are you going to do FSI on this? What motion does the hair have? Are you coupling it with ANSYS mechanical or are you modelling the motion some other way?

CanadaFlanker · May 5, 2014, 22:38

The hair is modeled as a cantilever beam that can interact with the flow. Originally I was trying to do FSI using Workbench, coupling with ANSYS mechanical, then I realized it is impossible to resolve the small boundary layer and thus drag force on the hair since it is too small. Then I am thinking of replacing the hair with a series of drag forces since they can be calculated from Stokes flow theory, and somehow simulate the interaction between the drag forces and the flow.

Could you first tell me how the volume with momentum should be modeled? Is it physically presenting like a wall or just overlap on the fluid? Thank you!

ghorrocks · May 5, 2014, 22:44

The momentum source is simply a region which puts a momentum sink on the flow where the momentum sink equals the momentum loss caused by the hair.

Does the hair have significant inertia? In other words is its location simply a function of the current drag, or does the inertia from its history have a significant effect?

CanadaFlanker · May 5, 2014, 23:00

So the volume is not a physical wall but just a region containing momentum sink, right?

Ideally I am simulating the fluid flows over the hair, causing the hair to bend, which may bounce back and imposes force back on the flow. Essentially a FSI problem. Eventually I will simulate a carpet containing hundreds of hairs on a surface over which the fluid flows.

If I use the momentum sink approach, I am thinking of modeling the hair as a 2D beam, with a series of momentum sinks along its length, just like a series of small elements, and somehow simulate the interaction between the momentum sinks and the flow. Does it make sense? Thanks a lot for your replies!

ghorrocks · May 6, 2014, 00:13

Sounds like you need to read the documentation about source terms. You have to understand them if you are going to use them. The velocity equations are conservation of momentum - the source terms are momentum added or removed "magically" by the user. By "magic" I mean it does not come from any other physics, the user can specify anything they like.

The question about inertia is critical. If inertia is not significant then your hair displacement function can be implemented in basic CEL and nothing tricky is required. If inertia is important then the hair displacement will require another approach - probably user fortran and that is quite a bit harder.

May 5, 2014, 17:06	Applying drag force in the fluid model	#1
CanadaFlanker New Member Join Date: Apr 2014 Posts: 6 Rep Power: 12	Hi, I am simulating a very small hair (diameter of 10 micrometers) in the fluid flow. The flow around it is Stokes flow as the Reynolds number based on the diameter is less than 5. It is unrealistic to make the grid smaller than the diameter in order to capture boundary layer around the hair and obtain drag force; an alternative way is to manually put a drag force on the points where the hair should be, as the drag force can be calculated from Stokes flow. Is there a way to apply a force at a point in CFX? I have searched about sources but it seems that source points don't have force or momentum, and the only place where I can put momentum source is subdomain. If anyone knows whether CFX can do it please let me know, otherwise I will switch to other software such as OpenFoam. Thank you.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Error message: Insufficient Catalogue Size	Paresh Jain	CFX	33	August 16, 2024 06:09
Use of k-epsilon and k-omega Models	Jade M	Main CFD Forum	40	January 27, 2023 08:18
Calculation of Drag Coefficient, Help Please	teek22	CFX	1	April 26, 2012 19:41
Fluid Force _output	jiko	FLOW-3D	2	September 29, 2010 13:11
Drag force on scaled models.	arunjingade	Main CFD Forum	6	July 1, 2010 09:54

May 5, 2014, 21:06		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	I think you have to define a volume for this and use a momentum source. Make the source small so it is only one element across - don't forget that source "points" act on a control volume anyway, so you are doing the same thing as a source "point" by doing this.

May 5, 2014, 22:30		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	That changes things.... How are you going to do FSI on this? What motion does the hair have? Are you coupling it with ANSYS mechanical or are you modelling the motion some other way?

May 5, 2014, 22:38		#5
CanadaFlanker New Member Join Date: Apr 2014 Posts: 6 Rep Power: 12	The hair is modeled as a cantilever beam that can interact with the flow. Originally I was trying to do FSI using Workbench, coupling with ANSYS mechanical, then I realized it is impossible to resolve the small boundary layer and thus drag force on the hair since it is too small. Then I am thinking of replacing the hair with a series of drag forces since they can be calculated from Stokes flow theory, and somehow simulate the interaction between the drag forces and the flow. Could you first tell me how the volume with momentum should be modeled? Is it physically presenting like a wall or just overlap on the fluid? Thank you!

May 5, 2014, 22:44		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	The momentum source is simply a region which puts a momentum sink on the flow where the momentum sink equals the momentum loss caused by the hair. Does the hair have significant inertia? In other words is its location simply a function of the current drag, or does the inertia from its history have a significant effect?

May 5, 2014, 23:00		#7
CanadaFlanker New Member Join Date: Apr 2014 Posts: 6 Rep Power: 12	So the volume is not a physical wall but just a region containing momentum sink, right? Ideally I am simulating the fluid flows over the hair, causing the hair to bend, which may bounce back and imposes force back on the flow. Essentially a FSI problem. Eventually I will simulate a carpet containing hundreds of hairs on a surface over which the fluid flows. If I use the momentum sink approach, I am thinking of modeling the hair as a 2D beam, with a series of momentum sinks along its length, just like a series of small elements, and somehow simulate the interaction between the momentum sinks and the flow. Does it make sense? Thanks a lot for your replies!

May 6, 2014, 00:13		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,854 Rep Power: 144	Sounds like you need to read the documentation about source terms. You have to understand them if you are going to use them. The velocity equations are conservation of momentum - the source terms are momentum added or removed "magically" by the user. By "magic" I mean it does not come from any other physics, the user can specify anything they like. The question about inertia is critical. If inertia is not significant then your hair displacement function can be implemented in basic CEL and nothing tricky is required. If inertia is important then the hair displacement will require another approach - probably user fortran and that is quite a bit harder.