Discussion Forum

Hi

I'm not sure it is that easy, car by essence the eigenfrequency analysis is a normalised view of a linearised frequency distribution. There are a few other ways to normalise the eigenvectors, than the one choosen by COMSOL, and you can recalculate those if you use matab on the results (do a search on the forum it has been discussed before) but anyhow, you have the isue of the eigenfrequency shapes, the peak high to width ratio is very "damping" dependent and in most of the analysis one hardly enter daping data. If you say you have a 1W energy density set in over a givn frequency band, then you should be able to integrate and renormalise the power spectral density you can estimate from the COMSOL run, but it will only be correct if you have entered and validated the damping values for your model.

This issue is an interesting one, and I would prefer too to see eigenfrequencay sweeps omlasied differently (or to have this as option in the postprocessing) currenty its up to you to finalise this on the raw data in matlab

Have fun Comsoling
Ivar

Hello Ivar,

Thanks for your comment. I understand the tricky part.

But, I don't mind even if the normalisation varies every time. But do you know where i can find the exact value of "input power" used for a specific eigenfrequency? I can then divide it with my parameter. It works for me.

I don't know how to find it. Do tell me if there is a way in postprocessing options.

Thanks,

hello

in fact there is no variation each time there is one norm and its stable. But there is no finite energy input variable either. I'm not even sure how to express it in RF oe EM, I'm used to do structural and there the tradition is to norme per total mass (which is a a way a normal energy over the given scan) but comosl has choosen to norm the vectors to 1, not relative to each other. Check the doc, and I would send a demand to "support". And keep us other informed as its of general interest.

Good luck
Ivar

Hi,
I am facing EXACTLY the same problem as yours - I am interested in getting result from the eigenfrequency analysis but with a defined power input.
Does anyone knows the solution?

Thanks,
Alex.

Hi

then you ned todefine some losses too, as an eigenfrequency (without losses) pushes the motin to "infinity".

If you add damping (or local energy exit) you can run a frequency scan with a given amplitude (= power) in and a coherent result out

--
Good luck
Ivar

Hi Ivar, thanks for the quick reply!

Actually I do have losses in the problem (I'm doing Optics with losses in silica and metalic absorbing structures).
your suggestion:
"run a frequency scan with a given amplitude (= power) in and a coherent result out"
sounds good, but the question is where do I define the given amplitude in the first place?

Thanks, Alex.