Topics related to Lumerical and more

The oscillation of auto-shutoff level in long FDTD calculated region

    • Kenny Chen

      I conducted simulations of an antenna composed of a straight Si waveguide (red) with periodic SiN block (cyan).

      When set a shorter FDTD range (e.g., 150 um) with the identical structure, the auto-shutoff level consistently satisfied the auto-shutoff threshold (default: 10^-6) without oscillations.

      However, when the FDTD simulation region was expanded to 1mm (x-axis), the auto-shutoff level displayed oscillations, even the overall trend consistently showed a gradual decrease in the auto-shutoff level.

      The FDTD boundary conditions were established as follows, with the Y_min boundary exhibiting anti-symmetry due to the structural symmetry along the y-axis.

      I want to understand that why auto-shutoff level display a oscillation when FDTD calculated region is extended.

      Thank you.

    • Amrita Pati
      Ansys Employee

      Hi Kenny,

      As you know the auto shutoff level gives a rough estimate of the total energy left in a simulation as a fraction of the total injected energy. If the device is operating close to the resonant frequency, it takes a relatively longer time for the energy to decay to the default auto shutoff level and it can also keep oscillating for a long time as you have seen. As long as this level reduces (and doesn't exponentially increase), it should not be a concern. Also, this doesn't always depend upon how big the structure is but how close it is to the resonant frequency. Sometimes even when really small structures are excited by their resonant frequency, they tend to take a lot longer to simulate and you can also see this oscillatory behavior. This can be most commonly observed in resonators.


Viewing 1 reply thread
  • You must be logged in to reply to this topic.