Showing results for 
Search instead for 
Do you mean 
Reply

Modal frequency response analysis with large mass and direct method


Hello,
I'm learning to perform frequency response analysis with FEMAP 9.31 and NX Nastran
5. I would like to better understand the difference between large mass and direct
method to apply an acceleration excitation. I studied the tutorial 17 of seismic
loading in FEMAP, which explains both methods, but the results are different, and
they should be the same. Moreover, tutorial 17 suggests to use param,resvec with
direct method, while tutorial 18 says it should not be used with latest versions
of FEMAP. Is there anyone expert in this kind of analysis? Which method is the best
and should I use param,resvec or not?
Thanks and regards.
Marco
5 REPLIES

Re: Modal frequency response analysis with large mass and direct method


Hello Marco,
By direct method, I think you are referring to the SPCD method for defining enforced
motion loading. There are two SPCD methods - the constraint mode method (introduced
in NXN 4.1) and the absolute displacement method. The default is the absolute displacement
method. If you want to use the constraint mode method, you need to use the following
system cell setting: NASTRAN ENFMOTN=0.
The large mass method and the SPCD constraint mode method are similar and do not
require the use of residual vectors. The SPCD absolute displacement method is different
and it is essential that residual vectors be used or you can get very poor results.
Mark


"marco" wrote:
>
>Hello,
>
>I'm learning to perform frequency response analysis with FEMAP 9.31 and NX
>Nastran
>5. I would like to better understand the difference between large mass and
>direct
>method to apply an acceleration excitation. I studied the tutorial 17 of seismic
>loading in FEMAP, which explains both methods, but the results are different,
>and
>they should be the same. Moreover, tutorial 17 suggests to use param,resvec
>with
>direct method, while tutorial 18 says it should not be used with latest versions
>of FEMAP. Is there anyone expert in this kind of analysis? Which method is
>the best
>and should I use param,resvec or not?
>
>Thanks and regards.
>
>Marco

Re: Modal frequency response analysis with large mass and direct method


Hi Mark,
thanks for your reply. With direct method I'm referring to the one described in
FEMAP tutorials, yes it writes SPCD entries on .dat file. So as far as I understand
I just apply acceleration loads to right nodes (after constraining them) then it
works (consider that I'm using FEMAP to set up the .dat file, not manual editing
because I don't know Nastran language enough...). Am I right? What I don't understand
is that you say default is absolute method, so I should turn on resvec, while on
tutorial it says resvec is not to be used with latest FEMAP/Nastran versions, so
I suppose it is by default constraint mode method...
Marco

"Mark Donley" wrote:
>
>
>Hello Marco,
>
>By direct method, I think you are referring to the SPCD method for defining
>enforced
>motion loading. There are two SPCD methods - the constraint mode method (introduced
>in NXN 4.1) and the absolute displacement method. The default is the absolute
>displacement
>method. If you want to use the constraint mode method, you need to use the
>following
>system cell setting: NASTRAN ENFMOTN=0.
>
>The large mass method and the SPCD constraint mode method are similar and
>do not
>require the use of residual vectors. The SPCD absolute displacement method
>is different
>and it is essential that residual vectors be used or you can get very poor
>results.
>
>Mark
>
>
>
>
>
>"marco" wrote:
>>
>>Hello,
>>
>>I'm learning to perform frequency response analysis with FEMAP 9.31 and NX
>>Nastran
>>5. I would like to better understand the difference between large mass and
>>direct
>>method to apply an acceleration excitation. I studied the tutorial 17 of

>seismic
>>loading in FEMAP, which explains both methods, but the results are different,
>>and
>>they should be the same. Moreover, tutorial 17 suggests to use param,resvec
>>with
>>direct method, while tutorial 18 says it should not be used with latest versions
>>of FEMAP. Is there anyone expert in this kind of analysis? Which method is
>>the best
>>and should I use param,resvec or not?
>>
>>Thanks and regards.
>>
>>Marco

>

Re: Modal frequency response analysis with large mass and direct method


Marco,
I am not that familiar with how Femap sets up this type of solution. I agree that
it sounds like that it is using the SPCD method. But unless the system cell setting
NASTAN ENFMOTN=0 or NASTRAN SYSTEM(422)=0 is used, then the absolute method is being
applied by default and residual vectors need to be used.
It may very well be that the Femap tutorial needs to get updated. If possible, can
you post the input file that gets written by Femap. I can look at it and will work
with the Femap development team if updates are needed.
Regards
Mark



wrote:
>
>Hi Mark,
>
>thanks for your reply. With direct method I'm referring to the one described
>in
>FEMAP tutorials, yes it writes SPCD entries on .dat file. So as far as I understand
>I just apply acceleration loads to right nodes (after constraining them) then
>it
>works (consider that I'm using FEMAP to set up the .dat file, not manual
>editing
>because I don't know Nastran language enough...). Am I right? What I don't
>understand
>is that you say default is absolute method, so I should turn on resvec, while
>on
>tutorial it says resvec is not to be used with latest FEMAP/Nastran versions,
>so
>I suppose it is by default constraint mode method...
>
>Marco
>
>
>
>"Mark Donley" wrote:
>>
>>
>>Hello Marco,
>>
>>By direct method, I think you are referring to the SPCD method for defining
>>enforced
>>motion loading. There are two SPCD methods - the constraint mode method (introduced
>>in NXN 4.1) and the absolute displacement method. The default is the absolute
>>displacement
>>method. If you want to use the constraint mode method, you need to use the
>>following
>>system cell setting: NASTRAN ENFMOTN=0.
>>
>>The large mass method and the SPCD constraint mode method are similar and
>>do not
>>require the use of residual vectors. The SPCD absolute displacement method
>>is different
>>and it is essential that residual vectors be used or you can get very poor
>>results.
>>
>>Mark
>>
>>
>>
>>
>>
>>"marco" wrote:
>>>
>>>Hello,
>>>
>>>I'm learning to perform frequency response analysis with FEMAP 9.31 and

>NX
>>>Nastran
>>>5. I would like to better understand the difference between large mass and
>>>direct
>>>method to apply an acceleration excitation. I studied the tutorial 17 of

>>seismic
>>>loading in FEMAP, which explains both methods, but the results are different,
>>>and
>>>they should be the same. Moreover, tutorial 17 suggests to use param,resvec
>>>with
>>>direct method, while tutorial 18 says it should not be used with latest

>versions
>>>of FEMAP. Is there anyone expert in this kind of analysis? Which method

>is
>>>the best
>>>and should I use param,resvec or not?
>>>
>>>Thanks and regards.
>>>
>>>Marco

>>
>

Re: Modal frequency response analysis with large mass and direct method


Dear Mark,
thanks for your help. Here I have attached a file written by FEMAP. I have done
a test with a frequency response analysy of a tower, taken from a FEMAP tutorial.
I have followed steps from tutorial, that says resvec is not necessary and acceleration
load is applied directly to the base nodes of the tower by a RBE2.
By the way, testing it with and without resvec on results are the same on my pc.
Regards.
Marco
"Mark Donley" wrote:
>
>Marco,
>
>I am not that familiar with how Femap sets up this type of solution. I agree
>that
>it sounds like that it is using the SPCD method. But unless the system cell
>setting
>NASTAN ENFMOTN=0 or NASTRAN SYSTEM(422)=0 is used, then the absolute method
>is being
>applied by default and residual vectors need to be used.
>
>It may very well be that the Femap tutorial needs to get updated. If possible,
>can
>you post the input file that gets written by Femap. I can look at it and will
>work
>with the Femap development team if updates are needed.
>
>Regards
>Mark
>
>
>
>
>
>
> wrote:
>>
>>Hi Mark,
>>
>>thanks for your reply. With direct method I'm referring to the one described
>>in
>>FEMAP tutorials, yes it writes SPCD entries on .dat file. So as far as I

>understand
>>I just apply acceleration loads to right nodes (after constraining them)

>then
>>it
>>works (consider that I'm using FEMAP to set up the .dat file, not manual
>>editing
>>because I don't know Nastran language enough...). Am I right? What I don't
>>understand
>>is that you say default is absolute method, so I should turn on resvec, while
>>on
>>tutorial it says resvec is not to be used with latest FEMAP/Nastran versions,
>>so
>>I suppose it is by default constraint mode method...
>>
>>Marco
>>
>>
>>
>>"Mark Donley" wrote:
>>>
>>>
>>>Hello Marco,
>>>
>>>By direct method, I think you are referring to the SPCD method for defining
>>>enforced
>>>motion loading. There are two SPCD methods - the constraint mode method

>(introduced
>>>in NXN 4.1) and the absolute displacement method. The default is the absolute
>>>displacement
>>>method. If you want to use the constraint mode method, you need to use the
>>>following
>>>system cell setting: NASTRAN ENFMOTN=0.
>>>
>>>The large mass method and the SPCD constraint mode method are similar and
>>>do not
>>>require the use of residual vectors. The SPCD absolute displacement method
>>>is different
>>>and it is essential that residual vectors be used or you can get very poor
>>>results.
>>>
>>>Mark
>>>
>>>
>>>
>>>
>>>
>>>"marco" wrote:
>>>>
>>>>Hello,
>>>>
>>>>I'm learning to perform frequency response analysis with FEMAP 9.31 and

>>NX
>>>>Nastran
>>>>5. I would like to better understand the difference between large mass

>and
>>>>direct
>>>>method to apply an acceleration excitation. I studied the tutorial 17 of
>>>seismic
>>>>loading in FEMAP, which explains both methods, but the results are different,
>>>>and
>>>>they should be the same. Moreover, tutorial 17 suggests to use param,resvec
>>>>with
>>>>direct method, while tutorial 18 says it should not be used with latest

>>versions
>>>>of FEMAP. Is there anyone expert in this kind of analysis? Which method

>>is
>>>>the best
>>>>and should I use param,resvec or not?
>>>>
>>>>Thanks and regards.
>>>>
>>>>Marco
>>>

>>
>

Re: Modal frequency response analysis with large mass and direct method


Hello Marco,
The deck you provided is using the SPCD method using the absolute formulation. In
that case I think that residual vectors are important to use. But I ran this model
and agree that the use of residual vectors made little difference.
In general that may not be the case. So it is my suggestion that you always use
residual vectors with the absolute method to be safe.
I will make the Femap team aware of this so they can update their documentation.
Regards
Mark


"Marco" wrote:
>
>
>
>
>Dear Mark,
>
>thanks for your help. Here I have attached a file written by FEMAP. I have
>done
>a test with a frequency response analysy of a tower, taken from a FEMAP tutorial.
>I have followed steps from tutorial, that says resvec is not necessary and
>acceleration
>load is applied directly to the base nodes of the tower by a RBE2.
>
>By the way, testing it with and without resvec on results are the same on
>my pc.
>
>Regards.
>
>Marco
>
>"Mark Donley" wrote:
>>
>>Marco,
>>
>>I am not that familiar with how Femap sets up this type of solution. I agree
>>that
>>it sounds like that it is using the SPCD method. But unless the system cell
>>setting
>>NASTAN ENFMOTN=0 or NASTRAN SYSTEM(422)=0 is used, then the absolute method
>>is being
>>applied by default and residual vectors need to be used.
>>
>>It may very well be that the Femap tutorial needs to get updated. If possible,
>>can
>>you post the input file that gets written by Femap. I can look at it and

>will
>>work
>>with the Femap development team if updates are needed.
>>
>>Regards
>>Mark
>>
>>
>>
>>
>>
>>
>> wrote:
>>>
>>>Hi Mark,
>>>
>>>thanks for your reply. With direct method I'm referring to the one described
>>>in
>>>FEMAP tutorials, yes it writes SPCD entries on .dat file. So as far as I

>>understand
>>>I just apply acceleration loads to right nodes (after constraining them)

>>then
>>>it
>>>works (consider that I'm using FEMAP to set up the .dat file, not manual
>>>editing
>>>because I don't know Nastran language enough...). Am I right? What I don't
>>>understand
>>>is that you say default is absolute method, so I should turn on resvec,

>while
>>>on
>>>tutorial it says resvec is not to be used with latest FEMAP/Nastran versions,
>>>so
>>>I suppose it is by default constraint mode method...
>>>
>>>Marco
>>>
>>>
>>>
>>>"Mark Donley" wrote:
>>>>
>>>>
>>>>Hello Marco,
>>>>
>>>>By direct method, I think you are referring to the SPCD method for defining
>>>>enforced
>>>>motion loading. There are two SPCD methods - the constraint mode method

>>(introduced
>>>>in NXN 4.1) and the absolute displacement method. The default is the absolute
>>>>displacement
>>>>method. If you want to use the constraint mode method, you need to use

>the
>>>>following
>>>>system cell setting: NASTRAN ENFMOTN=0.
>>>>
>>>>The large mass method and the SPCD constraint mode method are similar and
>>>>do not
>>>>require the use of residual vectors. The SPCD absolute displacement method
>>>>is different
>>>>and it is essential that residual vectors be used or you can get very poor
>>>>results.
>>>>
>>>>Mark
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>"marco" wrote:
>>>>>
>>>>>Hello,
>>>>>
>>>>>I'm learning to perform frequency response analysis with FEMAP 9.31 and
>>>NX
>>>>>Nastran
>>>>>5. I would like to better understand the difference between large mass

>>and
>>>>>direct
>>>>>method to apply an acceleration excitation. I studied the tutorial 17

>of
>>>>seismic
>>>>>loading in FEMAP, which explains both methods, but the results are different,
>>>>>and
>>>>>they should be the same. Moreover, tutorial 17 suggests to use param,resvec
>>>>>with
>>>>>direct method, while tutorial 18 says it should not be used with latest
>>>versions
>>>>>of FEMAP. Is there anyone expert in this kind of analysis? Which method
>>>is
>>>>>the best
>>>>>and should I use param,resvec or not?
>>>>>
>>>>>Thanks and regards.
>>>>>
>>>>>Marco
>>>>
>>>

>>
>