Showing results for 
Search instead for 
Do you mean 
Reply

strange results in frequency response

Hello everyone,
I'm performing a freq response analysis of a wooden box, simulating a vibration test (resonance search), applying a constant 0.5 g acceleration at a base node connected to the box base surface by means of a RBE2 element. The box is modelled by CQUAD4 elements. I'm using FEMAP 10.0.2 with Nastran 6.1.

This is my problem:
if I run the analysis with modal frequency method, using a highest frequency value of 6000 Hz (i'm interested in results between 20-2000 Hz), the output accelerations of nodes near base start from values much lower than the 0.5 input. If I use direct integration method the solution is ok: output accelerations start from 0.5 g. I think this means I loose important modes with modal frequency method and the solution is approximate. I also tried raising highest frequency value to 50000 Hz and in this case the solution matches with direct integration. Why do I have to use such high frequencies? Usually this doens't happen...

Thanks

Marco
3 REPLIES

Re: strange results in frequency response


Hello Marco,
In NXN 6.1, if you use the default SPCD for enforced motion, you must remember to
turn on residual vectors (PARAM, RESVEC, YES). Otherwise you will get the effect
you are describing.
In NXN 7, we changed the default to a newer SPCD formulation that does not require
residual vectors. I have put a ppt file ( nxn_forum_enf_motion.pptx )on the scratch
ftp site that has more details.
Regards
mark





dtmtech wrote:
>
>Hello everyone,
>I'm performing a freq response analysis of a wooden box, simulating a
>vibration test (resonance search), applying a constant 0.5 g
>acceleration at a base node connected to the box base surface by means
>of a RBE2 element. The box is modelled by CQUAD4 elements. I'm using
>FEMAP 10.0.2 with Nastran 6.1.
>
>This is my problem:
>if I run the analysis with modal frequency method, using a highest
>frequency value of 6000 Hz (i'm interested in results between 20-2000
>Hz), the output accelerations of nodes near base start from values much
>lower than the 0.5 input. If I use direct integration method the
>solution is ok: output accelerations start from 0.5 g. I think this
>means I loose important modes with modal frequency method and the
>solution is approximate. I also tried raising highest frequency value to
>50000 Hz and in this case the solution matches with direct integration.
>Why do I have to use such high frequencies? Usually this doens't
>happen...
>
>Thanks
>
>Marco
>
>
>--
>dtmtech
>------------------------------------------------------------------------
>dtmtech's Profile: http://bbsnotes.ugs.com/vbulletin/member.php?userid=198072
>View this thread: http://bbsnotes.ugs.com/vbulletin/showthread.php?t=40174
>


Re: strange results in frequency response

[ Edited ]
Dear Mark,

I've tried to add resvec,yes too, but the problem is unchanged, unless I increase max frequency to extremely high values (like 50000 Hz!). How can I find your pptx file on ftp?
If you have any chance, I'm attaching my .dat file with modal frequency method and 10000 Hz set as high frequency. I set it through FEMAP, maybe you can tell me what is the problem? Thanks and regards,

Marco

Re: strange results in frequency response

Dear Mark,
I am intesrested in the file you mentioned "nxn_forum_enf_motion.pptx", could you please share it again? -- thanks!.
Best regards,
Blas.