Contact or Shared Topology

Hey,

     I am doing a static structural analysis of sandwich plate with pyramidal core and laminated facesheets. I want to setup my core(beam elements) to be bonded to the facesheets( shell elements).

I tried that by choosing firstly only a shared topology in the geometry design phase and afterwards only with two contacts, one bonded contact where master was the core and slave was the upper facesheet, and another one for the lower facesheet.

The two methods gave different results, although in the manual it is said that they should be the same..

I also tried to do both, shared topology and contacts, and again I had another different result in respect to the aforementioned two methods.

Furthermore I also have different results based on which design tool I use, Spaceclaim or Design Modeler..

I cannot evaluate which from all these results is the correct one, since all of them have a big difference from the analytical computation based on the theory/literature.

Can somebody please help me?

Gratefully,

Achilleas Milios

Comments

  • peteroznewmanpeteroznewman Member
    edited August 2018

    Hey Achilleas,

    My preference for a sandwich plate is to use Shared Topology.  Here is one relevant post on honeycomb sandwich.

    I don't like contact elements because depending on which formulation is used, there can be some shear flexibility added by the contact element.

    It shouldn't matter whether the geometry is prepared in SpaceClaim or DesignModeler.

    I can look at your model if you Attach a Workbench Project Archive .wbpz file to the post above.

    Regards,
    Peter

  • AchilleasMilAchilleasMil Member
    edited August 2018

    Thanks for the reply,

     

    Ok I'll attach a file where I have made three structural analyses of the same geometry and load/bc case, one with spaceclaim shared topology, one with spaceclaim and contacts and one with DM and contacts.

    I couldn't set shared topology on DM for some reason I had multiple pivot warnings which by setting contacts I got rid of, so I guess that the shared topology was not succesfull in DM.

    So you think that the shared topology fetaure is better, please let me know if the model seems ok in Spaceclaim.

    Thanks a lot!

     

  • peteroznewmanpeteroznewman Member
    edited August 2018

    Hello Achilleas,

    I have started to look at your file in ANSYS 19.1 so if you are using a different version, let me know.

    I'm looking at your SpaceClaim shared Topology model and that seems to have worked correctly.

    I opened the DM block and in DM, I can see you haven't done what is required to use Shared Topology.

    You have this, which shows 3 Parts, 3 Bodies.  That is not shared topology.

    Select all three parts and RMB to get the menu, then Form New Part

    Then you will have Shared Topology available.

    In your case, it is Edge Joints, but for solid bodies that touch, it would say Shared.

    This works well for solids, but it does not seem to have made a mesh that has shared nodes.

    So DM is a bust. It doesn't work. Luckily, SC worked properly.

    Regards,

    Peter

  • peteroznewmanpeteroznewman Member
    edited August 2018

    This post is a more critical look at shared topology versus bonded contact.

    I was able to get bonded contact to work after I picked vertex on the contact side and face on the target side.

     

    Below is the deformation with Shared Topology:

    Below is with Bonded Contact using MPC formulation.

    The reason it is a little less is that the contact spreads out a spider web to connect to more than one node, while the shared topology connects at exactly one node.

    Here is a graphic of the spider elements for the bonded contact.

    Now as you make the mesh elements smaller, the spiders will get smaller and the solutions will converge to a similar result.

    Regards,

    Peter

Sign In or Register to comment.