Recently one user indicated that he was getting different results using Tochnog when meshing with triangles and quadrilaterals. I responded by indicating that normally for non dense meshes, 4-node quadrilaterals normally perform better than 3-noded triangles, as indicated in this link:
I tried to attached to this mail but it was bigger than permitted by the tochnog site.
To show that mesh density is important, I am attaching a cantilever beam example and a plate with a hole in one corner, using triangles and quadrilaterals. The files were run using the total formulation but similar results are obtained using the updated formulation.
The results shown below are for a cantilever beam 5 inches tall, 1 inch thick and 40 inches long with a 200 lbf load applied at the end. The solution using beam theory