unv2abaqus issue

Hi, I was trying to convert a Salome mesh to Calculix, with unv2abaqus.
Salome and Calculix have a different node order, so with the node-to-node translator the elements get twisted.
Obviously this will not happen with a linear tet element, but with higher order elements it gets messy.
I guess the procedure would be to create a map between the 2 node order options, and reorder the nodes.
Maybe using a case types[typ] it would be possible to enter an ad hoc function for every element type.
I would gladly help, but my understanding of python is really poor, so before start coding I would like to understand if it is a feasible road.
Moreover there is a lst.pop(0). In my case I would have to access arbitrary nodes. It it possible to use a lst[index] to get the elements?
Thanks,
Massimo

Hi,
I was able to correctly convert the 10 nodes tet. I guess it will be rather easy to solve the same way the 20 nodes brick or any other element.
The problem is I started converting between the MED node order and the calculix one.
unfortunately the I-Deas unv file describes the elements in a third way, and I wasn't able to find the documentation about it.

Massimo

Hi,
I was able to insert in the translator a mapping function, and I was able to correctly map the 10 nodes tet elements.
I created a single element and estrapolated the map.
I have the unv documentation, but I wasn't able to find the node order convention.
Am I missing something? If it is present in the documentation it will be a rather easy task...

Massimo

I have similar problem using nastran to salome.
I cannot traslate the thikness materilas rigid links ant integrated points.

I only can see the model inside salome but I am no sure how to add at less material and thikness.


Maybe Is a better to traslate directly to code aster.
But I do not have the keywords for aster.

With this script at less is possible to se nastran models.

this script works for shells and solids