[lihuang]

Dear all,

I am a new user of FLUENT. Currently, I am trying to study the turbulence generated by two way traffics (i.e. two cars passing but moving in opposite directions). From tutorials, I learned that this case should be modeled using dynamic mesh in FLUENT. But now i am having some difficulties with using dynamic mesh. Hope someone can help me out.

I started from a simple case of only one tire in contact with road surface, and treated the system to mimic a wind tunnel environment. I followed a previous post (http://www.cfd-online.com/Forums/ans...e-meshing.html). Thanks for the great discussions in there, I got similar results. But when I moved on to use dynamic mesh (rigid body movement for tire but stationary cond. for all other parts) I got errors.

Specifically, when I set rigid body for tire and stationary for int-air-field, I did not see movement of tire with respect to the ground surface. If I set rigid body for tire and stationary for ground, I got an error of negative volume.

I guess the tricky part is the treatment of the contact surface/line between tire and ground surface. Would it sound reasonable by using sliding mesh at the contact surface/line? Is there special treatment should be done during the meshing procedure? I used the Octree method for volume mesh.

Any suggestion is greatly appreciated!

Li

[lihuang]

I have done dynamic mesh preview for the moving tire without any contact with the road surface. Of course, this is not eventually what I want. But it demos the dynamic mesh approach in FLUENT. In the attached figures, you can see the tire (flying over the ground) moved with respect to a vertical plan I created.

But when the tire comes into contact with the ground or even very close to the ground, dynamic mesh fails with negative volumes found. Would it be possible to use non-conformal interface on ground surface to work around this problem? Can I use ICEM to prepare non-conformal interface for FLUENT? Or is it possible to re-mesh the surface of ground in fluent while tire moves forward at each time step using local face/region face options in Mesh Method Settings?
Could anyone help me out?

[Amir]

Hi,
why didn't you use remeshing option in FLUENT?
before using that you should enable remeshing for all kinds of grids in TUI.

[lihuang]

Originally Posted by Amir

Hi,
why didn't you use remeshing option in FLUENT?
before using that you should enable remeshing for all kinds of grids in TUI.

Thanks for your reply.
But I believe I did use remeshing option in FLUENT. I have done the following for the two pictures attached previously. I first meshed it using ICEM to obtain volume mesh, and output to FLUENT. In FLUENT, I set up model and boundary conditions, then Dynamic Mesh -> Smoothing and Remeshing (with some settings)-> Dynamic Mesh Zones (define Rigid Body motion of Tire using UDF). See attached pic for a summary of what i have done.

Again, I don't know what's the proper way to let FLUENT remesh the contact region between Tire and Ground.
Please help!

[lihuang]

I just gave it another try. About the geometry: the tire lies above the ground for about 0.2 cm (not direct contact on the ground). And it moves in the y+ direction only with a speed of 28 m/s (a value close to 100 km/hr). The distance of the center of tire to Right Boundary is about 2.34 m. I reduced the time step from 0.001 to 0.0001 s. The mesh moves fine within the first 200 time steps. Pictures of initial position and when negative volumes found are attached.

Besides, the finer portion of surface mesh on ground (around the initial "contact" region) did not move with the movement of tire. Maybe this is the cause of remeshing failures? Any suggestion is highly appreciated.

[Amir]

I think the best choice is using remeshing and smoothing simultaneously. in remesh setting panel you should set min & max of length scales.
if you set these 2 methods correctly, founding negative volumes is impossible.

[lihuang]

Originally Posted by Amir

I think the best choice is using remeshing and smoothing simultaneously. in remesh setting panel you should set min & max of length scales.
if you set these 2 methods correctly, founding negative volumes is impossible.

Thank you, Amir, for your help.
But in my case, i did use both smoothing and remeshing at the same time. And in the remesh setting panel, I set min & max of length scales both to 0. I am hoping that this setting triggers FLUENT to mark all cells to be remeshed regardless of their size, and to improve skewness of cells. AM I right here?

When negative volumes found, I checked current mesh. A number of left-handed faces were also found on stationary boundaries and in fluid zone. I am not sure about what zones should be defined in "Dynamic Mesh Zones". I always defined tire as rigid body. Sometimes, I tried to include other boundaries (such as ground) and fluid zones. No luck. What's the criteria of defining zones in the list of dynamic mesh zones?

[Amir]

Originally Posted by lihuang

I set min & max of length scales both to 0. I am hoping that this setting triggers FLUENT to mark all cells to be remeshed regardless of their size, and to improve skewness of cells. AM I right here?

I guess that setting both values to 0 disable this feature, I propose you to change these and check.

What's the criteria of defining zones in the list of dynamic mesh zones?

you should define tire as a dynamic zone. your tire moves and consequently near meshes change with remesh or smoothing algorithms.

[xrs333]

The tire is not tangential with the ground, also it is not round. You rebuild the geometry like this and can be saved from all the troubles.

[lihuang]

[QUOTE=Amir;298060]I guess that setting both values to 0 disable this feature, I propose you to change these and check.
According to this document (http://eps.fluent.com/5903/500000762/20060904/6DOF.pdf), I set both values to 0. As you suggested I changed them to a few combinations, but still with no luck. The skewness of fluid cell kept increasing to 1, and negative volumes found after approximately the same time steps. Is there a way to display cells with high value of skewness in FLUENT?

[lihuang]

Originally Posted by xrs333

The tire is not tangential with the ground, also it is not round. You rebuild the geometry like this and can be saved from all the troubles.

Thanks! I rebuilt the geometry, and it works like a magic. I set tire as moving rigid body, and the ground as deforming surface (its surface cell skewness needs to be improved). I attached pic of meshes at the beginning and the end of preview. Fine mesh elements moved along with the contacting surface between tire and the ground.
But I am wondering why the changes in geometry solved the problem? If I define a mesh density around tire or at its wake region, would it follow the motion of the tire? What should I do during the meshing stage to archive a high quality dynamic mesh?
I really appreciate your help on it, xrs333.

Cheers.