Effective Mass Mode Selection for CMS

[Antonio Zamora]

Hello,

I am trying to perform the following:

I have an assembly which two main parts, where one is way more massive than the other (about 50 times) and I would like to perform a modal analysis in with the number of modes expanded are based on the Effective Mass of the lighter component.

I do not know if there is a direct way to do it (I have performed the modal analysis based on the Effective Mass of both components, but not focusing on the lighter one). If not, I have thought generating a superlement for the lighter part, using CMS (substructuring with Craig-Bampton Method) but based on the Effective Mass. 

I have seen that CMSOPT allows to select the number of modes to be used (NMODE) but I would like to select them based on the Effective Mass (as it is done with MXPAND (MODM) in a Modal Analysis).

I would appreciate any idea or suggestions.

Best regards,

Antonio Zamora.

[dlooman]

If the lighter body has higher frequency modes, you could set FREQB on the CMSOPT command to be just below these higher frequencies.  You can do a modal analysis with the master dof fixed to determine what the frequencies of the lighter body are.  There isn't a more precise way.

[Antonio Zamora]

Thank you very much for your reply but I am afraid this is not the case. The lighter part is a metallic structure (a telescope indeed) with the first mode at ~8 Hz and it overlaps with the massive one, which is the building where the telescope is mounted. In both systems there are optical components, which locations are the scope of the analysis (the modal analysis is performed to generate a space state model with these locations as outputs). As there are optical components in both parts, is why I need to sort the modes independently.

I will try to find other alternatives.

[dlooman]

One alternative is to specify master dof throughout the telescope structure.  That will make the modes less important and also shift the frequencies higher, because they are computed with the master dof fixed (assuming you are using the default cmsopt,fixed)

[Antonio Zamora]

Do you refer to generate a superelement of the building with the telescope-building interface as master dof? I have thought about it but does not solve that the number of modes in the superlement are truncated but not ordered in importance (and there are a lot of modes of the building, considering up to 90 % of the effective mass in the range of interest, which is 0-100 Hz).

The solution I was thinking is to not using a superelement or, incase of using it, considering a large number of modes which contains all the frequency range of interest. Then perform a modal expansion selecting the modes by ranges using MEXPAND,%modes_to_expand%. The ranges of modes can be estimated by standalone modal analyses of the telescope and the building. After that the space state model is generated with SPMWRITE, containing just the desired modes.

The problem there is that the modes of both systems separatedtly will not remain in the same location when both are assembled together so I will need to be carefull to ensure that they are selected.

Again, thank you very much for you reply.