RC panels in rsa after transfer from rst

[josh.southall]

I've sent a structure from rst to rsa, looks right, geometry in the right place etc. but when it comes to meshing it goes a bit mental. It creates thousands of nodes along some edges, and then say 10 nodes along the opposite edge on the panel (which looks like it should do). It also displays the mesh on the centreline of the panels instead of the surface like it would do normally (toggling the thickness displayed option).
Anyone know what might be going on and how to fix it?

Another problem, I assume is related somehow, is that when running a verification on a locally generated mesh to test, it rsa claims there is no material associated with the panel... I can't work out why or how it could come up with this error as materials have all been defined (imported from rst and checked).

Help greatly appreciated!!

Thanks

[pawelp]

Hi,

The panel's attributation will be set automatically when calculations will be run.
As you described, at this stage the pure mesh has been proposed only, so other information will be associated automatically later on.

for the other question, please send a screen capture

regards

[josh.southall]

aha, thanks for the first part.

Here's a screen shot of the "No Material" error... Although from what you just said about the mesh, I could infer that the same thing applies in this case. This is taken from a verification run straight after "local mesh generation" on the panel shown.
It can be seen in the property window on the bottom left that the default concrete material is applied (CONCR).

Thanks

[romanich]

josh.southall, just save rtd file, close and open again.

[bauskas]

Josh,

have you managed to solve your problem yet?

I've experienced similar problems when using local mesh generation to 'glue' several converging plates correctly on edges. RSA refused to apply material from plate entity to its associated FEs whan local mesh generation for a separate plate was used.
A simple analysis restart eventually solved the problem.

[mcl222062]

Have anyone worked out how to solve this problem?

When I export my model from rst to rsa it gives me some 'Incoherent mesh on edges'-warnings when running a verification, but the solver will run despite of these warnings. I can fix the problem by creating local meshes on some panels, but then I get an error that there is no material associated with the panels (which there is - I have tried defining the materials again, but that doesn't help either). If I try to run calculations for the model, despite the 'no material'-error, Robot crashes with errorcode 1370 (which means 'Internal error in frontal analysis, too big problem'...that doesn't really help me in regards to how I can solve the problem).

It's very frustrating at this point, that I can calculate a flawed model, but I can't fix the inaccuracies, cause then Robot crashes when asked to calculate. I have tried sending the model from rst to rsa again and also tried only exporting small parts of the model, but I end up in the same mess.

Is there a solution for this?

[pawel.pulak]

Explaining some effects you observed:
1/ 'Incoherent mesh on edges' is treated as a warning and not an error when kinematic constraints are active in Job Preferences - in such case it is possible to continue analysis
2/ local mesh generation is only generating FE mesh without full assignement of properties&materials and without merging overlapped nodes on the edges panels. If verification is run after it the messages about "no material" or overlapped nodes are justified. Materials are assigned and nodes are merged during model generation which is run by default prior to the solver when running analysis.
3/ It seems that your model after corrections is too big to be run using frontal solver. Please check the number of nodes, elements and equations in the window of solver (bottom left part of it). It it is bigger than expected please check for existence of stange refinements of FE mesh along some edges (as mentioned in the first post of this thread) - they usually correspond to inaccuracies betwen adjoining panels.
4/ Please check analysis method in Job Preferences>Structure Analysis - if it is Frontal please change to SparseM or Multi-threaded solver. These solvers are more efficient for big models
Regards,
Pawel Pulak
Autodesk Product Support