frequency peaks in modal frequency response

[frmx]

Hi,

doing a frequency response, I requested the output to be calculated also at the eigenfrequency of my structure by mean of a FREQ3 card.

Looking at the results, I noticed that the eigenfrequencies reported were slightly different that those I could calculate directly from a modal analysis.

Could someone explain? Is this a matter of accuracy?
I thought that it should be the same since the NASTRAN method is based on a modal frequency response. Could the residual vectors improve the results?

Thanks in advance for your answers.

 

[ESPcomposites]

Could this be due to damping? Try to run a model with zero damping and see if they are they same.

Brian

http://www.espcomposites.com

 

[frmx]

Hi,

sorry for the late answer but I was out of this office for some weeks. Thanks for the answer, I will look at it. This could be an explaination.

Cheers.

 

[ESPcomposites]

The solver first uses modal analysis to solve eigenvalues. If a damped solution is requested, then you will get the damped natural frequency, which is different than the original undamped values. Not sure if that is your problem, but something to think about.

Brian

http://www.espcomposites.com

 

[kashm]

Do you have any frequency dependent springs?
If yes, then the eigenfrequency will differ. Becasue they don't have any influence in Modal Analysis.