Discussion Forum

You said, " I am assuming that the mechanism to convert f1 to f2 is already defined in the model (using port for example)."

I am not aware of any obvious way to convert between frequencies using a port. Can you post an example of the model in which you are trying to do that?

If the physics that links your two frequencies are entirely separated except in one place/manner, then you may be able to solve this as two problems in sequence, where the output of one serves as the input to the next. Is your problem of that type?

If by convert, you are referring to modeling some kind of non-linear physical process (which would be typical in the case of a change in frequency), then you should probably model this in the time domain.

Alternatively, there may exist physical equations (partial differential equations with supporting boundary conditions, constraints, etc.) that can be (with effort!) defined for the conversion/generation of waves of multiple frequencies in some particular problem. Comsol Multiphysics supports fairly general PDEs, both linear and non-linear, but the mathematical sophistication needed to use this capability effectively can be greater than that associated with using the more specialized/popular "modules." But, at least in theory, one might be able to program those custom equations into Comsol Multiphysics and then solve them using (very-carefully-configured) solver settings. You would need to have a very good understanding of all the equations involved, all the conditions involved, and the solution methods. It would also likely increase your odds of success if you obtained specialized assistance from Comsol personnel directly. If you have an actively-supported license, you might consider asking them for help!

You said, " I am assuming that the mechanism to convert f1 to f2 is already defined in the model (using port for example)."

I am not aware of any obvious way to convert between frequencies using a port. Can you post an example of the model in which you are trying to do that?

If the physics that links your two frequencies are entirely separated except in one place/manner, then you may be able to solve this as two problems in sequence, where the output of one serves as the input to the next. Is your problem of that type?

If by convert, you are referring to modeling some kind of non-linear physical process (which would be typical in the case of a change in frequency), then you should probably model this in the time domain.

Alternatively, there may exist physical equations (partial differential equations with supporting boundary conditions, constraints, etc.) that can be (with effort!) defined for the conversion/generation of waves of multiple frequencies in some particular problem. Comsol Multiphysics supports fairly general PDEs, both linear and non-linear, but the mathematical sophistication needed to use this capability effectively can be greater than that associated with using the more specialized/popular "modules." But, at least in theory, one might be able to program those custom equations into Comsol Multiphysics and then solve them using (very-carefully-configured) solver settings. You would need to have a very good understanding of all the equations involved, all the conditions involved, and the solution methods. It would also likely increase your odds of success if you obtained specialized assistance from Comsol personnel directly. If you have an actively-supported license, you might consider asking them for help!

Thanks for your response, Yes, as you mentioned by ‘’converting’’ I am referring to a nonlinear process to create different harmonics around a center frequency f1 (let’s consider only up to 1st order terms f1, f1+f2 and f1-f2). So, what I am trying to do is to excite with frequency f1 passing through a port and convert to f1+f2 or f1-f2 (I am not sure if it is possible with ports or not even with physical equations for this conversion). Since this process only happens at one place as you pointed out, is there any chance for me to solve this like two problems in sequence? Best, Reza

If you want to capture/model the actual physics of the conversion process within the model, then you will need that physics to be described by the PDEs involved. And I expect you would need customized PDEs for that, or at least customized constraints imposed on the equations.

If you have your own separate (separate from the PDEs) equations that describe how separate solutions (to Comsol-provided PDEs) at single frequencies at some location (point, line, surface, or volume) are expected to interact, then you may be able to manage that in separate steps, combining variously separate models and post-processing stages. Regardless, I think you need to supply a lot more detail about the physics problem you are trying to solve, so that others here can understand the challenges involved and can help you.