Hi, I am encountering an issue related to the singularity of the dynamic matrix. I am modeling a pipe and conducting some tests. My goal is to discretize half the length of the pipe using beam elements and the other half using plate/shell elements. To achieve this, I tried using both the liaison_RBE3 modeling approach and the liaison_elem approach with the COQ_POU option. In both cases, I encounter the following error:
Degré de liberté physique associé au noeud N28 et à la composante DRZ.
Problème : la matrice est singulière ou presque singulière :
Lors de la factorisation de la matrice, on a rencontré un problème
(pivot nul ou presque nul) à la ligne 174 qui correspond au degré de liberté donné ci-dessus.
To better detail my modeling process, I provide an example below, taken from the document Connection hull-beam (r3.03.06), along with a screenshot of my model.


The node causing the error is highlighted in red in the image below. The node is part of the edge group when using liaison_elem, or part of the node group when using liaison_RBE3.

I am unsure how to resolve this issue. While searching the forum for similar problems, I noticed that some were resolved by applying concentrated stiffness to the node, using something like the K_TR_D_N operator. However, I am uncertain if this is the right solution for my case. I believe the error I am encountering might be due to the incorrect use of the liaison_RBE3 or liaison_elem operators.
Thank you in advance for your attention.