Importing partioning mesh

Partioning a mesh in Salome takes a lot of time. So I decided to create multybody part from CAD, meshing with full hexa in Salome and exporting as UNV. Mesh contains double nodes that are deleted in Mecway very fast with "merge nodes" command.
What I didn't understand is the presence of thick lines in Mecway if only one part is recognized . After merging double nodes some thick lines are interrupted.
Regards

2018-01-05_181400.jpg
936 x 818 - 192K
2018-01-05_182037.jpg
548 x 339 - 54K

Comments

  • Could it be that there are some nodes not merged still? Have you check it with Open Cracks view?
  • edited January 5
    Yes open crack!
  • Some glitch on the display maybe then, I have seen some also before, and with partitioned hexa meshes also so could be related.
  • edited January 5
    Merging nodes in Salome there are no problems.
    In Mecway some elements continue to be disconnected. In fact the node number is different:
    Merging double nodes in Salome 1260
    Merging double nodes in Mecway 1288
    2018-01-05_183825.jpg
    533 x 325 - 60K
  • Maybe you need to increase the tolerance for the merge (selecting only the near nodes to avoid mix some small elements in other regions).
  • Yes...is a tolerance problem but these nodes (double) have the same coordinates.
    Thanks
  • Merge nearby nodes is supposed to work with 0 tolerance if the coordinates are identical. Maybe the last digit got changed when importing through .unv? To confirm they're identical in Mecway, select all the same-coordinates nodes and open Node coordinates. All 3 coordinates show a value if they're exactly the same. Any that are blank are not identical from Mecway's point of view.
  • I will point out the problem with more clarity.
    On more complex models is difficult to manage double nodes because merging nodes with tolerance too high could cause mesh problems. For example the two models on picture: after merging nodes with 0.5 mm tolerance the total nodes are different from the real number. With 1 mm the node number become lower than the real one (and the elements which invalidates the mesh)
    Sincerly I don't know if it is possible a tolerance of more than few micron ( 3 digit on the rigth of a decimal point or 6 digit after decimal point for Mecway).

    Regards

    2018-01-06_013502.jpg
    1575 x 794 - 386K
    2018-01-06_014153.jpg
    1660 x 769 - 390K
    2018-01-06_014547.jpg
    412 x 476 - 98K
    2018-01-06_014719.jpg
    526 x 446 - 108K
  • Thanks for the pictures but I somehow still don't understand, sorry.

    If it's merging too many nodes and damaging the elements, you should be able to use a smaller tolerance. For meshes where I have no feeling for the scale, I usually zoom in to what looks like the smallest element and use the tape measure to check it size, then use a power of 10 smaller than that as the tolerance.

    If it's that the nodes you want to merge aren't really very close together compared to the element sizes - or other groups of nodes you don't want to merge - then Sergio's suggestion of selecting the nodes would limit the scope to the relevant ones.
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!