Photonics

Photonics

Topics related to Lumerical and more.

Is it possible to optimize for 2 input modes simultaneously?

    • Brian West
      Subscriber

      Consider the following example....I want to design a y-branch that is simultaneously optimized for TE0 and TE1 inputs.  I can certainly find an optimal design for TE0, but it may be poor for TE1, and vice versa.  So the design I seek has an FOM that incorporates both TE0 and TE1 simulations.

      I can't figure out how to accomplish this.  Ideally, it would involve one of the following:  (1) each trial design has FDTD run with TE0 input, and the results saved, then a second FDTD run with TE1 input, and the results saved.  Then both sets of results are incorporated into the FOM.  Or, (2) an S-parameter scan is used rather then just a single FDTD, and the single set of S-parameters contains all the required data for the FOM.

      However, it seems that 2 FDTDs is not allowed, and s-parameter scan within optimization is not allowed.  Is this true?  And if so, do you have any suggestions?  Thanks in advance. (edit....I know lumopt is a good tool for this, and it specifically has an example that requires 2 FDTDs (the robust coupler superoptimization)...but I find that lumopt is more restrictive in how the FOM can be defined.  So a solution not using lumopt would be beneficial.)

    • GWANG
      Ansys Employee

      Hi Brian West,

       

      Thank you for the post and sorry for the late reply. For this request we actually recommend using the Python API as you mentioned. To use FDTD only for the simulation, I am thinking that you can ran the several sweeps for the TE0 mode and TE1 mode, then export and add the sweep results to pick the lagest transmission point. SHould this sound like doable? 

    • Brian West
      Subscriber

      This is what I am hoping to do (one TE0 simulation and one TE1 simulation).  I can't figure out how to do 2 FDTD simulations for each trial within the optimization framework.  Is this possible, and if so, how?

    • GWANG
      Ansys Employee

      Hi Brian West,

      I am thinking you can use a nested sweep for this, or just two seperated sweep and get the result from each of them and use script to add the result up.

    • Brian West
      Subscriber

      Guanhui,

      Two separated FDTD simulations are what I am trying to do.  As I wrote in my initial message, I can't figure out how to add a second FDTD simulation.  I just get an error when I try to do that.  Can you provide a script example that does this?  Or does it have to be two separate S-parameter sweeps?

      Thanks,

      Brian

    • Brian West
      Subscriber

      Just following up on this...can you provide an example that allows for two FDTD simulations within each trial geometry?  Or an example that uses an S-parameter sweep within the optimization framework?

       

      Thanks,

      Brian

Viewing 5 reply threads
  • The topic ‘Is it possible to optimize for 2 input modes simultaneously?’ is closed to new replies.