Step for axisymmetric

We use SW for CAD. I can export Step file and do static . Great!
However all we do is round and symmetrical, so instead I want to do axisymmetric analysis form the step file. The step file is a (say) 30 degree segment of the complete model . I cannot for the life of me find a way to use just the XY plane surface of this for the analysis. If not possible is there another wAy. Cheers Dave

Comments

  • You can do a Surface mesh (in Meshing parameters) then delete all the surfaces except what's in the XY plane for axisymmetry.
  • That works great.
    However I want my team to be able to do fairly often repeats or a generically similar shape. (a very simple hydraulic cylinder.)
    If we do as a 3D static, we can set up the FEA loads, supports etc. on the relevant geometry surfaces. Then each iteration they save the STEP version from CAD, then reload and remesh, but everything else is already set up so its a quick process.
    If doing Axisym. anaylsis then each time, am i correct, that when reload the geom and remesh then have to reselect the edge elements for loads and constraints, as cant do this on geometry. So basically have to set up the FEA again each time?
    Any work around this? I would rather we did an axisym anaysis than 3D.
    Great FEA package by the way!
  • No way to keep edge loads, sorry.

    Another possibility is to use 3D solids but with a smaller segment size, such as 1 degree so that the solving is fast. Put frictionless support on both symmetry surface to enforce the axial symmetry boundary conditions.
  • Right. Didn’t realise I could do that. That will be perfect. Cheers
  • Hi,
    An alternative way to generate a surface mesh would be to generate a surface in SolidWorks and export that as a STEP file. I didn't realise you could do that until it was pointed out to me here and it seems to work well.
    But... as Victor says that doesn't help with being able to carry the loads over.
  • Hi,
    I'm currently working with a similarly axisymmetric item in 3D and using this process, which seems to work well but I do have a question...
    Is there a way to replace the geometry rather than reload it? What I mean is that I'd like the name in the outline tree to reflect the geometry that was actually used. I'm going through several iterations and variations of a design and would like to be able to save various different versions of the model and have the outline tree make it clear which geometry file was used. So I can be sure that when looking at the Mecway file I can work out which STEP file was used and trace it back that way.
    Maybe I'm missing something obvious but I've been copying and renaming the STEP files each time to match the name of the geometry in the Mecway file and trying to keep a separate record but that's got more scope for me to mess up the records.
    Any tips would be appreciated,
  • I was thinking in the same feature this week, a new option in the geometry to replace the part...
    You can made it just editing the liml file with Notepad, luckily the part definition is at the begining of the file.
  • ¿Is there a way to replace the geometry rather than reload it?

    This is probably not in the Best practice books but, when the model is not heavy, I put the different geometries all into the same file.

    This way:

    They can share the material definitions.
    BC can normaly be shared too. (Be carefoul with pinned)
    Differences between the geometries and results are clearly represented without need to jump from one file to other.

    I’m not sure if this is possible with step files but I don’t see why not. It would require to shift in space the model to avoid overlapping previous version when loaded.
  • This afternoon I've been editting the .liml file as per Sergio's suggestion and it has worked well so far. The surface definitions have all re-attached correctly, materials already assigned and the loads and constraints applied as required. Thanks for that suggestion, Sergio. By the way, I have been using Notepad++ rather than Microsoft Noetpad since it seems to cope with the large files more readily.
    I can see the utility of disla's method but I don't think it would work for me at this time. I'm iterating toward a solution (hopefully) which has multiple parts so it would be awkward to have to adjust the coordinate systems of the separate STEP files each time, so that they line up with one another but miss other models. But it's definitely a tip I'll remember for the future.
    Thanks for the help,
  • Editing directly the liml file can be very usefull, sometimes the GUI is slow or take several steps do a repetitive task as creating glue contacts or assign material when the model is big or has lot of components.
  • I think if you put a new step file with the same name in the directory and use the reload feature, it updates the geometry. I have found though that there is no guarantee that all of the surfaces will line up, but most of them often do.
  • Hi JohnM,
    Sorry for delay in replying. You're right and that's what I had been doing but I found that I was struggling to keep my records straight as to what geometry had been run in what analysis. What Sergio suggested is basically the same as you've said but with the added stage of changing the filename in the .liml file and saving the modified .liml file with a new name. When I load the new .liml and reload the geometry it loads the new geometry.
    That's a pretty poor explanation, I'm afraid.
    So far it seems to reconnect the named faces, loads, etc. as long as the geometry was generated from the same CAD. I've got a CAD file that I'm tweaking the geometry in but not adding or removing features. It seems the STEP file generated from this is structured in such a way that the faces appear to be identified the same. I do check though.
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!