Force reaction problem

[TeoMV]

I have designed the assembly shown in the picture.
In image no1 (file2)

http://img526.imageshack.us/img526/8600/file2k.jpg

you can see the force (2000N) which is put upwards on the blue surface.

In image no2(file3)

http://img688.imageshack.us/img688/5505/file3.jpg

its a cylindrical support with fixed radial and axial movements but free tangential movement so the bar can rotate while loaded.

In image no3 (file4)

http://img29.imageshack.us/img29/4456/file4y.jpg

its a fixed support on the blue surface.

The scope of the analysis is to analyze the force reaction on the fixed support.

I start with default meshing which gives the meshed model of image no4 (file5)

http://img40.imageshack.us/img40/2889/file5.jpg

The result is shown in image no5 (file6)

http://img230.imageshack.us/img230/7676/file6.jpg

where the force reaction is 130N and the deformation of the left blade is quite normal too.

When i set mesh sizing and lowering the mesh size the results are getting very very different!

The new mesh is shown in the following image (file7)

http://img524.imageshack.us/img524/4250/file7.jpg

The force reaction now is 2882N which is way different to the previous result! Which leads me to believe that there is something wrong in my analysis!!!

Here is an image from the force reaction result (file8)

http://img691.imageshack.us/img691/6980/file8.jpg

Please help me because i am very much confused!
Thank you very much for any help!

 

[JohnyGluebag]

Your first result does not make much sense to me. If you check your problem first by hand, with a free body diagram, your reaction force should be 2000N. So, you should expect a similar result from Ansys.

Open the help file and read about 'mesh convergence'. There is a feature where you can tell the software to remesh & solve until the deformation does not change by more than a certain percentage...say 5%.

Depending on what element you use, you can get a variety of results.

I am not sure what you want out of this analysis. Seems like you have alot of load on a thin part, deflection could be large, do you need accurate through thicknes data? Representing a thin plate with tet brick elements not always a good idea.

You might have to turn on Large Deflection.

Sorry I have more questions than answers. Maybe others will reply with help.



____________
JohnyGluebag