Assembly

[transall]

Hello,

I would like to create an assembly between two tubes but I don't know how to do.
I created a simple analysis with two tubes and I would like to create an assembly as if the two tubes was a unique part.The tubes are aligned according their longitudinal axis and there is a small gap between them. I have to create a contact between two circular area.

I tried to couple the DOFs,I have a good result on displacement but the Maximum equivalent Von Mises stress are not good. I tried to use contact manager but I have rigid body motion because of the gap.
To sum up, I would like the tubes assembly gives the same result as one tube with the same length.

Thank you for your Help,

Regards,

Mickaël

 

[cbrn]

Hi,
let me say that what you need to do doesn't seem very "physical":
if there is a gap, the coupling will be correct only if the tube is loaded axially (traction or compression), otherwise the rigid connection between distant nodes will lead to stresses' overshoots.
Why not open this model in Classical, add a cylindrical volume to "join" the two cylinders, glue them together, and then mesh? If you have troubles while meshing in Classical, you always have the solution to send it to Workbench via CDWRITE,solid,...
Can't you get rid of the gap at all? Do you have access to a solid modeler?

Regards

 

[transall]

Hi,

I use Ansys classical only. I use a gap to have a simple example (bending problem), indeed, my objectif is then to assemble two bonded tubes, but one is horizontal and the other is welded to the exterior surface of this horizontal so that there is an angle between them. The surface I have to link are not directly compatible (one is the exterior area of a tube, and the other is the slanting end of a tube). Maybe I can create a solid between them,but I don't how to obtain such a suitable solid.

Thank you,
Regards

 

[meshparts]

Transall, I would still use the CP command.

Define 2 cylindrical coordinate systems (local,,1), each centered about each of the tubes ends and with the z axis parallel with the tube symmetry axis.

Then rotate the nodes of each tube along each cylindrical coordinate system.

The next part is much more simple, if the meshes of each tube are identical, at least at the interface.

If the meshes are identical, you can write a simple APDL code to select each corresponding nodes an couple them together with the CP command. Of course you can also do it by hand, if the mesh is not too fine...

Hope this helps!

 

[cbrn]

Hi,
... on the opposite side, I would rather try to resolve the "problem" geometrically:
if I understand it correctly, you can model the slanted tube longer than real, thus compenetrating the "horizontal" one. Then, with the boolean operations of ANSYS it is easy to "cut" the slanted tube to match the other one. OK, this may not be what the tubes are in reality (in this kind of construction, there is a gap indeed, filled subsequently by the welding...), but I feel it could be a reasonable approximation without the need to explicitly model the welding. I also feel that this method would be faster, BUT it is based on the assumption that the boolean doesn't fail for some reason...

Regards

 

[transall]

Thank you all!