unv 2414 not working? about feconv HOT 12 CLOSED

Hi,

I am glad to hear that FEconv is useful for you. Let me check your files and I will answer after that.

from feconv.

HI,

Just a question. In test2.unv, the mesh has 41 nodes but the field is only defined for 20 of those nodes. Is that intended? I am not saying it is incorrect, I am trying to understand the example.

from feconv.

Hi again,
In fact there are 20 nodes in the mesh, but numbering is not consecutive (1 2 3 6 7 9 ...). I have to check whether UNV reader is ready to receive non-consecutive node numbering.

from feconv.

Hi,
There is a check to control that the maximum node numbering (in test2.vtu is 41) is equal to the number of field entries (in test2.vtu is 20). When I hide this check I got a field representation
Here you can see the field magnitude, is that correct?

The UNV reader was intended to work with consecutive node numbering (yes, that should be indicated in the Help) but, since the latter case appears to work, I will change the check from an error to a warning. Let's see what happens...
See this commit

from feconv.

Hi again!

Brilliant, now it works fine with several elements (the first was just a classical example of the single element test)!

from feconv.

Hi,

One last test before closing the issue. I see that test2.unv was created with Patran. Is it posible that you try the following steps?

• To export/save in Patran the mesh to BDF format.
• Use FEconv to convert from BDF to VTU.

The reason is that I believe that we implemented the BDF reader taking into account non-consecutive nodes.

from feconv.

from feconv.

from feconv.

Sorry, but I do not see any attached BDF file!

from feconv.

Ok, here it is
fluid.bdf.txt

from feconv.

Thank you, with your last example I have checked that BDF reader of Feconv is assuming consecutive numbering.
In fluid.bdf, line 1533, the first CHEXA has Element Identification Number (EID) equal to 2000. Since the total number of elements is 960, FEconv uses 960 as dimension of the subdomain reference array NSD (see line 342 of module_patran.f90), so when the first CHEXA is read (in line 582 of module_patran.f90), the error arises.
I will try to adapt FEconv to non-consecutive numbering in the future. Meanwhile, I will indicate this restriction in the Help.

from feconv.

from feconv.