CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Rotating Boundary, Wheels (http://www.cfd-online.com/Forums/cfx/117559-rotating-boundary-wheels.html)

 ChrisR May 10, 2013 12:24

Rotating Boundary, Wheels

I tried to simulate a regid body with a diffusor and rotating wheels. But it seems like it dont work.

Without wheels my simulation is covergent after some iterations but i dont know how to setup the wheels.

I meshed it like in this picture :

Then I define the wheel as a no slip wall with a rotating velocity.

But i get an error that i have a significant normal component at one or more faces and that the angle between the velocity and the surface is almost 90 degrees.

Whats wrong? The ground is also moving.

Should i define the wheels as complete new domain and let the domain rotate?

And guidance would be great!

Ps: I use CFX Pre and ICEM

 evcelica May 10, 2013 13:44

Does global X go right through the center of all four wheels? No, It can't, so my guess is you will have to define the proper axis of rotation.

 Far May 10, 2013 13:49

Your wheels (geometry) is not modeled correctly.

 evcelica May 10, 2013 14:03

Quote:
 Originally Posted by Far (Post 426575) Your wheels (geometry) is not modeled correctly.

Oh wow, I didn't see that, I figured they were circles.
I think you have a flat tire so your car is losing control during the simulation :)

 Far May 10, 2013 14:10

In reality they are somehow flat at the intersection with road, but not as modeled in pics in above post. They look like a extruded part. Correct method would be to intersect wheels with road and give corresponding velocity to road (convert circular to rectilinear)

 ChrisR May 10, 2013 15:17

I heard that it wont work if i mesh the tires tangential to the road. Somebody told me that it would produce a "singularity" and wont work.

I could just model the tires to the road without the vertical transition and specify the "wheel wall" as wall boundary with a rotating velocity (with right axes ;) )and the "ground" as moving wall?

Would this work?

I'm thankful for your help :-)

 Far May 10, 2013 15:20

Tangential boundary will create problems for meshing or solver? If meshing, I can show you how to handle. But solver should not have any problem.

Quote:
 with right axes
I saw this type of setting (i.e. giving different rotation axis to different parts) in Fluent but don't know how to activate this options in CFX.

 ChrisR May 10, 2013 16:39

I could also do it in fluent if its easier to setup.
I still wonder if i should define it as wall in one fluiddomain or should i do three bodys- For each wheel a solid and the fluid domain?

I think the solution with one domain would suit better, but if the wheel is tangential to the ground i create 3 domains? :confused:

 ghorrocks May 10, 2013 23:38

Do not model rotating tyres with a rotating frame of reference. This can all be modelled in the single stationary frame of reference with a tangential velocity (which happens to rotate).

Also, the funny little bits at the bottom of the tyres to eliminate the cusp point will make it MUCH easier to mesh this. If you are not interested in little details like this you will find this assists immensely.

 Far May 11, 2013 01:04

 ChrisR May 11, 2013 03:54

I think i first try to manage the right axis and if this dont work i change the mesh.

But how can i find out "where" the middle of the tyre is? I have no cad the next few days. Is there a good way in icem or cfx to do this?

Should the transition bits also rotate?

 Far May 11, 2013 04:10

In ICEM put two points on periphery of tyre at location 0.25 and 0.75 or 0.0 and 0.5 and create midpoint

 ghorrocks May 11, 2013 06:44

Just to check - are you doing this simulation with a rotation frame of reference or with just rotating boundary conditions?

Noe that if you are doing it with rotating BC then as the effective frame of reference for the simulation is translating with the vehicle, then the rotation axis for the wheel is through its centre, with the road surface having a tangential velocity backwards.

 Far May 11, 2013 11:30

You may find it interesting : Post # 14 and image no 3

http://www.cfd-online.com/Forums/ans...rmula-one.html

For boundary conditions, now it is clear that you have to specify road a tangential velocity = omega * R (radius of wheel) in back ward direction.

 ChrisR May 14, 2013 04:39

It is my "backup" plan to mesh it tangential.
But i will first try it with the bits.

But i think i cant find proper points to define my axis of rotation

Front tyre:

Start: 1200 | 515.392105 | -175
Oriantation : 800 | 515.392105 | -175

Back tyre :

Start : 1 100 | -656,65112 | -175
Oriantation : 800 | -656,65112 | -175

I attached my geometry, could someone who is better than me in icem(i think almost everybody) check the geometry for the proper middle of the tyres? That would help me alot.

http://dfiles.eu/files/2awsgnn1y

If i mesh it like this http://www7.pic-upload.de/thumb/14.0...zitq33aend.png

do i have to calculate the rotating speed with 175 mm or with 160 mm?

 ghorrocks May 14, 2013 08:02

If you are trying to get the rotational speed of the tyres correct based on the sag, then the correct answer is in between these two values. But I would not worry about minor details like this, it is unlikely that this small difference in the rotational velocity makes a difference anyway. If you do not believe me run them both and compare them - I bet they are the same.

 ChrisR May 14, 2013 08:52

Then my axes are probably false?

 Far May 14, 2013 12:31

175 would be the better choice. But as indicated by glenn, difference would be very minor.

 All times are GMT -4. The time now is 15:29.