k_zurawski
i develop a bit, a bit in a disorder:
of course you can define many COMPORTEMENT within a single model
example wood with elastic and steel with plastic in a a single building
i do not know if PETIT on the discrete in your case has a strong influence on the results
anyhow it is the only possibility
commercial codes would do their own cooking without telling you much until you go for a deep reading of the message file
orphan nodes are detected when we have nodes not connected to any defined element
which is the case here as the POI1 is created afterwards
this a warning not an error
in your case the support nodes could have been at the corner of a plate element with the proper LIAISON_SOLIDE that would have avoided creating discrete element on them
about the way you create the rmed file:
do you need all the results for every component at every calculated step,
i do not think so
i create a list for the results which contains less step than the calculated ones
for the stress components many CMP are either null for the specified geometry and load
or do not mean much
i try to restrict the print to really useful values (and the same for printing in the base if i need one)
that makes for much much smaller file
COQUE_3d is not aware of perpendicular to the plane behavior and needs quadratic
in the same condition
3D modeling would not need many elements in the thickness
linear could do the job
with only 3 DOF per nodes calculation times would not be more, maybe less
