Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Can one solve a problem in frequency domain at a complex frequency?

Please login with a confirmed email address before reporting spam

Hello everyone!

I am looking for an efficient way to solve a wave optics problem with an exponentially decaying source signal of the form: exp (i w t - gam t). I know that I can use the Time Explicit or Transient interfaces. But it seems much more efficient to be able to set a complex frequency in the Frequency Domain interface and make it find the electric field complex amplitude distribution (up to an arbitrary constant). Unfortunately, this does not seem to be possible.. Or am I missing something?

Thank you in advance for your suggestions!


3 Replies Last Post 23 mar 2021, 00:09 GMT-4
Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 years ago 19 mar 2021, 18:10 GMT-4
Updated: 4 years ago 19 mar 2021, 18:14 GMT-4

Interesting question. The use of "frequency domain" to solve RF problems removes the time parameter from the problem, i.e., the differential equation becomes literally "time-independent." See the attached image. A time-domain exponentially decaying signal is not a "single-frequency" signal. And I can see that an attempt to specify a complex frequency fails. That said, a signal that decays spatially can certainly be modeled in the frequency domain. Perhaps one of the real Comsol experts here can offer a comment on whether there is any hope at all for doing what you want to do via using a frequency domain interface. I would suspect that aside from some kind of trick (such as employing various transform operations in post-processing, or adding some clever analytic feature to the model that would effectively allow an interpretation as if it were time domain), you will probably have to work with a time domain solver.

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Interesting question. The use of "frequency domain" to solve RF problems removes the time parameter from the problem, i.e., the differential equation becomes literally "time-independent." See the attached image. A *time-domain* exponentially decaying signal is not a "single-frequency" signal. And I can see that an attempt to specify a complex frequency fails. That said, a signal that *decays spatially* can certainly be modeled in the frequency domain. Perhaps one of the real Comsol experts here can offer a comment on whether there is any hope at all for doing what you want to do via using a frequency domain interface. I would *suspect* that aside from some kind of trick (such as employing various transform operations in post-processing, or adding some clever analytic feature to the model that would effectively allow an interpretation as if it were time domain), you will probably have to work with a time domain solver.


Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 years ago 20 mar 2021, 06:32 GMT-4

Robert is right, a decaying signal is not harmonic. On the other hand, your model has to be linear to be suitable for frequency domain. So any quantities that depend on the stimulus in the model will be following the stimulus in a linear manner. It might be good enough to add the decay in the post processing.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Robert is right, a decaying signal is not harmonic. On the other hand, your model has to be linear to be suitable for frequency domain. So any quantities that depend on the stimulus in the model will be following the stimulus in a linear manner. It might be good enough to add the decay in the post processing. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 4 years ago 23 mar 2021, 00:09 GMT-4

Dear Robert and Edgar,

Many thanks for taking time to share your thoughts. I agree with what you said and I do work with a time domain solver now. I could not think of an easy way to implement the decay I want in the post prosessing.

Still it would be interesting to read a comment from someone close to the development of comsol as to whether it is at all possible, what I asked, at least in principle.

Dear Robert and Edgar, Many thanks for taking time to share your thoughts. I agree with what you said and I do work with a time domain solver now. I could not think of an easy way to implement the decay I want in the post prosessing. Still it would be interesting to read a comment from someone close to the development of comsol as to whether it is at all possible, what I asked, at least in principle.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.