3 weeks ago
I am trying to run a Modal Analysis (SOL103 Real Eigenvalues) on a model with many parts linked together by means of Surface to Surface glues and Edge to Edge glues.
Analysis has run for 17 hrs now and if I check graphs for Eigenvalues extracted, line is constant on 0 values.
On monitor I have these messages :
''
USER INFORMATION MESSAGE 5010 (LNCILS)
STURM SEQUENCE DATA FOR EIGENVALUE EXTRACTION.
TRIAL EIGENVALUE = 1.557409E-01, CYCLES = 6.280893E-02 NUMBER OF EIGENVALUES BELOW THIS VALUE = 2962
*** USER INFORMATION MESSAGE 5010 (LNCILS)
STURM SEQUENCE DATA FOR EIGENVALUE EXTRACTION.
TRIAL EIGENVALUE = -8.214753E-01, CYCLES = -1.442505E-01 NUMBER OF EIGENVALUES BELOW THIS VALUE = 0
''
Those have been showed over and over again, since analysis started, only trial eigenvalue and cycles values changed, while the last number was 2962 and 0 alternatevely....
I think analysis is not running properly but I don't get which is the problem and how to solve it.
Has anyone faced this kind of issue?
Thank you.
3 weeks ago - last edited 3 weeks ago
how many modes are you requesting? or frequency range specifed?
if not glued properly you could have quite a few rigid body modes
3 weeks ago
3 weeks ago - last edited 3 weeks ago
nastran tried a (low) eigenvalue and found nearly 3000 of them lower than this very low value.
Maybe you have a lot of rigid body modes - import the glue-preview and check everything was created the way you wanted!
maybe some units wrong somewhere
can sometime "help" nastran by specifying a low f1 value (on the LANCZOS card) , say -1
3 weeks ago
Mari4,
If all you are requesting is the first 10 modes then I don't think it is a problem with your glue constraints. Whenever I use glue constraints in a SIM, I verify them by first running a SOL103 setting the Number of Desired Modes = 10 and the Frequency Range - Upper Limit = 1.0 Hz. This way, if all my meshes are properly glued together it will solve fast because there won't be any rigid body modes below 1.0 Hz. And if there are some unglued meshes, it will still solve fast because it will only compute the first 10 rigid body modes.
Regards,
James
3 weeks ago
At the end it was a glue issues due to wrong search distance.
I had to run a 101 with bailout -1 in order to obtain a displayable result.
Thank you all for helping me!