Error in the references after remeshing using MMG2D

#1 (348.6 KB)

After remeshing the attached example with MMG2D the references I am getting do not correspond with the references in the original mesh (most of the references are 0, instead of 6,7,8,9).

Merci beacoup par votre attention



To keep the references of your boundaries, you have to define the references on the edges (not only on the vertices).

In your file “out_step=1.mesh”, you define the edges only one side of your square thus mmg2d put a reference equal to 0 on all the other sides.


Thanks a lot, I will try. Thanks again


Hi, there is another way to define in which mesh a node belongs without the need of specify in which edge it belongs?

The other approach is very troublesome and in some cases can be even inconsistent with what I am trying to do.



What do you mean by “in which mesh a node belongs” ? Do you have several subdomains in your mesh ?

If not, can you explain why it is a problem to define the edges ?


OK, sorry, maybe in my last message I didn’t explain myself correctly.

What I meant is that if a want to remesh a problem where I have repeated some nodes and elements in different meshes, because I use each mesh to organize which way I assign properties or I impose BC. Right now each time I assign I reference to an element and it remeshed the resulting mesh “keeps” the original references to the meshes that I assigned originally. This is not the case for the nodes.

According to the first reply I need to assign an edge to the node to be able to recognize in which mesh it belongs. This could work for me, but it requires some extra work and I am asking if it is possible to work directly with the nodes, or it is necessary to assign these edges always.

Thanks again



It is not possible with Mmg to recover the new nodes references.

Thus you have 2 solutions:

  • Add edges and get back a node reference from the edges references
  • Use the element references to assign node references. For exemple, a node may have the maximal reference of its ball (the list of the elements to which the node belongs)

Best regards,



Merci beaucoup, I will try with the two actions. I will post it if I succeed.


Done, it worked, not 100% but it is not trivial



I’m glad to know that it works.

Thank you for the feedback.