Cyclic Symmetry issue

FammalFammal
Good afternoon folks,

This is probably a very stupid question; Apologies for that in advance;

I'm struggling with cyclic symmetry. I'm doing pre-loading of a torqued pipe thread. I've created a section of it so I don't have to solve the full 360 degree. I cannot seem to get the cyclic symmetry to work. I believe it's due to the nodes being both a support and a contact.

To solve this i created a small section on the side where the was no contact and set that up as a slave for cyclic symmetry and it still failed with the same error.

1) is using a frictionless support on each side an acceptable solution?
2) is an elastic support a decent option?
3) is the only option to remove the edge nodes on one side from the contact and is there an easy way of doing that? I'd really like to avoid having to click on each one and remove from named selection.

Comments

  • VictorVictor
    edited April 10
    If it's axially symmetric, not just cyclically symmetric, then yes, it would be better to use symmetry constraints on the two surfaces. Frictionless support is OK too but gets risky if there are shells or slight imperfections in the mesh because that alters the direction a little. Since you have contacts too, elastic support would be the safest way to simulate axial symmetry.

    If you do need cyclic symmetry, the easiest way is elastic support since it doesn't interfere with the cyclic constraints. Otherwise, extruding a thin layer from a surface is a way to keep the constraints separate. It sounds like that's what you already did though, so maybe some edges still had common nodes?

    If you're using the CCX solver, that allows constraints on the master surface of cyclic symmetry. The internal solver allows some axis-aligned constraints on either surface. See manual section 9.41 Cyclic symmetry.
  • FammalFammal
    edited April 10
    Yes - sorry i should have been more clear. It is axially symmetric along the X-Axis.

    What i'm trying to do is essentially run a 2d Axial Symmetric Solve but with non linear contact on the thread teeth.

    Since the solvers don't do non linear in 2D Axi - i made a 1 degree slice of the parts.

    What i'd like to do is solve for the pre-load torque. then apply the test load after that. I'm assuming i do that with quasi-static and load steps; but still reading the manual on that one.

    I need to put fixed supports on each end while the pre-load displacement resolves. Can i turn off the fixed support on one end and add a load step there after?

    Again sorry for the questions that are basic;
  • JohnMJohnM
    This might be useful. The uypull surface is postprocessed so you can see the deflection/load relationship. Make sure to multiply load by 360/sectorAngle, and don't ever design a bolt like this ;)
  • disladisla
    Since the solvers don't do non linear in 2D Axi


    ccx does. Just push the run button. Contact will automatically trigger the NLGEOM analysis.



  • JohnMJohnM
    edited April 11
    That's interesting @disla. I don't use that feature. We have a lot of casual users here, I try and keep the ways to do things to a minimum :-)
  • disladisla
    I try and keep the ways to do things to a minimum :-)


    I agree but I thought it's good to know users can get much more if they explore.
    Mecway has clearly prioritized reliability over the number of functions. This makes it a more robust program but can sometimes be limiting. If it appears in red it probably has some aspect that Victor found sensitive to incorrect result, convergence problem or undesired interactions with the rest of the implemented tools. In this case it is advisable to do some tests and validate some known problem before going further.

    NOTE: For example, there are some bugs reported for axysimmetric when reading reaction forces and contacts are involved.

    https://github.com/Dhondtguido/CalculiX/issues/98

  • FammalFammal
    Very good to know about 2D Axi-Symmetric. I just assumed because it highlighted in red it would not run. I should have clicked run just to see.

    Thank you for the reference bolt model. I've got my analysis running and solving. I'm not getting contact pressure at the torque load shoulder for some reason; but things are at least moving forward.
  • JohnMJohnM
    @disla I just tried it, very cool.

    Many of our sector models are looking at developing contact and friction. I think this approach would not apply, correct? It is Static, does not allow time stepping (I think).
  • disladisla
    edited April 11
    @Fammal
    Probably too late but recall to request for CDIS and CSTR to see contact results.
    Mecway doesn't add them automatically as it's supposed you are not unsing contact in Axy 2D.

    @JohnM

    If you enforce a maximum time increment your BC will be ramped ¿isn't it?.One could also add an Amplitude. You need custom cards.
    Please check it carefully and do your own validation.


  • FammalFammal
    Folks,

    managed to get to solve and verified the results with testing.
    Learned a few things along the way.

    Thank you very much for your input and help.
Sign In or Register to comment.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register