Dynamic mesh fidelity

montag dp · January 8, 2011, 11:12

I've been doing simulations of oscillating airfoils in Fluent using dynamic meshing. I usually use spring-based smoothing with a spring constant of 0 and local remeshing with as fine parameters as possible. However, the grid fidelity around the airfoil tends to degrade rather quickly using these techniques, no matter how fine I try to set the remeshing parameters.

Are there any suggestions for getting a better moving grid near the airfoil? For example, it would be nice if I could set up a boundary layer grid in Gambit and set that as the moving zone (with the airfoil, of course). I know I could do this if I made a separate wall to define the zone around the airfoil, but of course, I wouldn't want it to be a real wall. Are there any suggestions in dealing with this?

Thanks,
Dan

-mAx- · January 10, 2011, 01:34

in Gambit create a volume surrounding your airfoil and define it as separated fluid domain.
In fluent set the rigid body motion on your airfoil AND also on this fluid domain

montag dp · January 10, 2011, 09:48

Thank you! I saw another similar thread where you suggested something like that, but I wasn't quite clear what you meant. I will try this today.

montag dp · January 10, 2011, 15:37

Wow, it actually worked on the first try. Thanks again, mAx.

Some more details for anyone else trying to do this:
-Set the boundary type of your interface to "interior."
-Apply dynamic mesh motion to the real surface, the interior boundary, and the fluid domain between them.

BuilttoSpill · January 13, 2011, 10:52

Hi Dan
I would like to ask you a question.I have tried a similar simulation with an oscillating cylinder in 2D.The mesh was built with ICEM.
I've used the same tips(real boundary-interior and the interior boundary all moving with rigid motion),but when i look to the boundary layer i found inertial effects e.g. in the first cell of boundary layer the vector of velocity is not parallel to the surface but have a vertical component.

Did you noticed this effect or do you think that i've mede some mistake?

Thanks in advice

Paul

montag dp · January 13, 2011, 12:34

Paul,

Are you sure the normal component of velocity is incorrect? In a viscous flow there should be some normal component in the boundary layer due to the rotational nature.

Dan

BuilttoSpill · January 13, 2011, 16:21

In fact i'm not sure...i don't have a lot of CFD experience,but the fact that the viscous sub layer have a normal component it sounds me strange,flow shoud be tangential to the surface ...
Maybe is my mistake...
Thanks for your answer

montag dp · January 13, 2011, 17:07

I should probably clarify what I said. It depends on what your motion is and where you are looking. If you have separated regions, I would expect to see normal components of velocity there. If you're not looking at a separated region it should be fairly tangential and the boundary layer should be thin.

Atze · March 19, 2011, 08:01

Hi all,

is not creating a volume sourranding my body similar using sliding mesh? is not different using sliding or deforming meshes?

January 8, 2011, 11:12	Dynamic mesh fidelity	#1
montag dp Member Join Date: Jul 2010 Posts: 38 Rep Power: 15	I've been doing simulations of oscillating airfoils in Fluent using dynamic meshing. I usually use spring-based smoothing with a spring constant of 0 and local remeshing with as fine parameters as possible. However, the grid fidelity around the airfoil tends to degrade rather quickly using these techniques, no matter how fine I try to set the remeshing parameters. Are there any suggestions for getting a better moving grid near the airfoil? For example, it would be nice if I could set up a boundary layer grid in Gambit and set that as the moving zone (with the airfoil, of course). I know I could do this if I made a separate wall to define the zone around the airfoil, but of course, I wouldn't want it to be a real wall. Are there any suggestions in dealing with this? Thanks, Dan

January 10, 2011, 01:34		#2
-mAx- Super Moderator Maxime Perelli Join Date: Mar 2009 Location: Switzerland Posts: 3,297 Rep Power: 41	in Gambit create a volume surrounding your airfoil and define it as separated fluid domain. In fluent set the rigid body motion on your airfoil AND also on this fluid domain __________________ In memory of my friend Hervé: CFD engineer & freerider

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Dynamic Mesh on Pintle type injector.	herntan	FLUENT	16	September 4, 2020 08:27
Incylinder dynamic mesh with volumetric reaction	mas	FLUENT	4	May 3, 2012 10:22
dynamic mesh on a hexa grid	Manoj Kumar	FLUENT	0	August 21, 2007 07:41
Dynamic mesh + grid adapt = Crash! (Files included	BillH	FLUENT	4	July 24, 2007 15:31
Help: Is it possible to use dynamic mesh model and	Valery Ponyavin	FLUENT	1	September 20, 2003 14:18

January 10, 2011, 09:48		#3
montag dp Member Join Date: Jul 2010 Posts: 38 Rep Power: 15	Thank you! I saw another similar thread where you suggested something like that, but I wasn't quite clear what you meant. I will try this today.

January 10, 2011, 15:37		#4
montag dp Member Join Date: Jul 2010 Posts: 38 Rep Power: 15	Wow, it actually worked on the first try. Thanks again, mAx. Some more details for anyone else trying to do this: -Set the boundary type of your interface to "interior." -Apply dynamic mesh motion to the real surface, the interior boundary, and the fluid domain between them.

January 13, 2011, 10:52		#5
BuilttoSpill New Member Join Date: Aug 2009 Posts: 9 Rep Power: 16	Hi Dan I would like to ask you a question.I have tried a similar simulation with an oscillating cylinder in 2D.The mesh was built with ICEM. I've used the same tips(real boundary-interior and the interior boundary all moving with rigid motion),but when i look to the boundary layer i found inertial effects e.g. in the first cell of boundary layer the vector of velocity is not parallel to the surface but have a vertical component. Did you noticed this effect or do you think that i've mede some mistake? Thanks in advice Paul

January 13, 2011, 12:34		#6
montag dp Member Join Date: Jul 2010 Posts: 38 Rep Power: 15	Paul, Are you sure the normal component of velocity is incorrect? In a viscous flow there should be some normal component in the boundary layer due to the rotational nature. Dan

January 13, 2011, 16:21		#7
BuilttoSpill New Member Join Date: Aug 2009 Posts: 9 Rep Power: 16	In fact i'm not sure...i don't have a lot of CFD experience,but the fact that the viscous sub layer have a normal component it sounds me strange,flow shoud be tangential to the surface ... Maybe is my mistake... Thanks for your answer

January 13, 2011, 17:07		#8
montag dp Member Join Date: Jul 2010 Posts: 38 Rep Power: 15	I should probably clarify what I said. It depends on what your motion is and where you are looking. If you have separated regions, I would expect to see normal components of velocity there. If you're not looking at a separated region it should be fairly tangential and the boundary layer should be thin.

March 19, 2011, 08:01		#9
Atze Member Join Date: Apr 2010 Location: Pisa / Italy Posts: 62 Rep Power: 16	Hi all, is not creating a volume sourranding my body similar using sliding mesh? is not different using sliding or deforming meshes?