Crash and Odd behaviour of cyclic symmetry

FatmacFatmac
While playing around with a simple contact model I had multiple crashes with mecway (very rare for me to have any). I was using a 2d mesh I revolved into a wedge and then tried to apply cyclic symmetry BC. All worked ok while testing with static linear tests and constraint equations. However as soon as I started to get into NL analysis and get more ambitious with the contacts the model crashes. See pictures. While trying to debug I also ran into an odd situation where the analysis ran a static analysis with cyclic symmetry supressed but when unsuppressed there was an error mentioning temperature issues! as far as I know no temp loads are present. I also notice an odd thing about the model - when trying to select one flat face on the lower rectangular wedge it selects more than one face even with edge threshold set to zero - see pic 2.
so perhaps there is an issue with my mesh? I have noticed this phenomenon when I have revolved other 2D meshes. Any insights into what I have set up incorrectly here most welcome

Comments

  • SergioSergio
    Pay attention that you have colapsed hexas at the center line of the model, it should be wedges.

    Use the tool Merge nodes witi a small tolerance such as 0.001 mm to stich the nodes and convert the colpased hexas in wedges. With that at least the face selection error is fixed


  • SergioSergio
    @Victor, could be that the program shows the colapsed elements with the same red cross symbol as the distorted elements? In that way we could figurate quickly the error.
  • FatmacFatmac
    thank you Sergio for pointing that out -certainly something to watch out for when when revolving surfaces when one edge is on the axis or rotation. I've had very little experience with wedges - do they generally behave behave ok - even if they have no mid surface nodes? in this situation they just happen to be right where all the action is!
    I've ratified this issue in my simple model and it does clear up the edge selection issue but I still have the same crashes and odd mention of a temperature issue when i play around with the contact options
  • SergioSergio
    I use wedges for structural simulations and they behave well. This model can be solved easily with frictionless supports in the simmetry faces and high contacts stiffness



  • SergioSergio
    Attached the model solved
  • FatmacFatmac
    thank you for looking at that. It certainly solves nice and easily in that wedge with the frictionless supports. I'm keen to master the cyclic symmetry so I'm really interested to know why when i take this model and try and apply cyclic symmetry it crashes
  • VictorVictor
    edited February 2025
    A few other issues besides the collapsed hexes Sergio identified:

    1. Nodes of the slave surface, which includes the axis nodes, typically can't have constraints. They don't need them anyway because they're tied to the master surface.

    2. An exception to 1. is an axial displacement constraint on the axis, for some reason. But that's also the one place where it's useful.

    3. Contact's MPCs seem to be interfering with the cyclic symmetry MPCs. Use Elastic contact instead.

    Yes, ideally, it would merge the axis nodes automatically.

    The error reading the .frd file looks like the file was truncated and the final line isn't finished. Maybe CCX exited improperly?
  • SergioSergio
    edited February 2025
    I have used very few times the cyclic symmetry, attached some examples that I did years ago (was not implemented in GUI so I wrote the cards manually), hope it help you.

    https://drive.google.com/file/d/1rUnfCIAdxrY3Nx6fUAcNqFWH_GCyIufH/view?usp=sharing
  • SergioSergio
    I think that with a model with more element in thickness should work well the cyclic symmetry. Another issue is that Mecway cannot separate components in the results if you ask for the symmetric results.
  • FatmacFatmac
    Thank u Victor for comments and serio for the example files -i think I have set up the cyclic symmetry correctly as my model runs ok and gives expected results in static linear and non linear runs until I start playing around with contacts then i get those strange crashes. I will try more elements in the thickness to see if that helps
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