negative eigenvalues. solution not converging 1

[H choudhary]

Hi,

I am running a test on a ceiling mount monitor bracket. I have been run the job for hours now but it keeps aborting. the warning messages are negative eigenvalues. my step configuration are okay, you can see the picture, I think you'll agree. I have also checked all my BCs, interactions and material properties but still can't find the error.
Could someone please explain what exactly could be the problem and how can i tackle negative eigenvalues warnings.

any help is appreciated.

thank you

 

[FEA way]

Are there any other warning messages ? Run eigenfrequency extraction analysis to see if all parts are connected properly (if not, they will detach in mode shapes).

 

[H choudhary]

I ran the frequency analysis and these are all the error warning. they were same for static analysis. the frequency analysis is still runnnig. -

FOR *TIE PAIR (ASSEMBLY_beam's surf icw pipe-ASSEMBLY_pipe's surf icw beam), NOT ALL THE NODES THAT HAVE BEEN ADJUSTED WERE PRINTED. SPECIFY *PREPRINT,MODEL=YES FOR COMPLETE PRINTOUT.

FOR *TIE PAIR (ASSEMBLY_joint's surf icw bracket-ASSEMBLY_bracket's surf icw joint), ADJUSTED NODES WITH VERY SMALL ADJUSTMENTS WERE NOT PRINTED. SPECIFY *PREPRINT,MODEL=YES FOR COMPLETE PRINTOUT.

130 nodes have been adjusted. The nodes have been identified in node set WarnNodeAdjust.

The system matrix has 1 negative eigenvalues.

the number of eigenvalues keep fluctuating.

 

[FEA way]

Can you share the cae or inp file ? I would take a look at this model.

 

[H choudhary]

here's the .CAE file.

i checked the messages, it's says some nodes on part9-1 do not have defined boundary conditions.

 

[H choudhary]

actually it says

DEGREE OF FREEDOM 5 IS NOT ACTIVE ON NODE 543 INSTANCE PART-9-1 -
THIS BOUNDARY CONDITION IS IGNORED

and the message repeats for various nodes

 

[FEA way]

The problem is with contacts. If you change them to tie constraints or tied contacts, the analysis will work and then you can bring the nonlinear contact interactions back one by one and see which pair is causing the issue. You should also refine the mesh, it's too coarse, especially in the thickness direction (there should be at least 3-4 layers of elements).

 

[H choudhary]

yeah i read the message file. the problem is with 2 regions of contact. i had to give it up . it required too much computation. i decided to go another way. I decided to analyse the vertical beam now. i just modelled the pipe and horizontal beam as wire structure and modelled the bracket as a solid and used MPC beam b/w beams and MPC tie b/w bracket and horizontal beam. it gave similar results.

 

[H choudhary]

thank you for all the suggestions.