Convergence problem FDTD

    • pierre_fehlen


      We are simulating a plasmonic antenna working in the mid-infrared range, with a 2D simulation. We have a problem of convergence with our simulation. We have tried to correct this, and we have found out that the problem could be in the boundary conditions settings. In particular, by playing with the pml settings, we observe that the “alpha” coefficient is crucial for convergence. In “standard settings” alpha is kept at 0, but we see that if we increase it to 1 the simulation converges. However, we also see that the modes of the antenna (which are recorded with a reflectance monitor) spectrally shift if we keep increasing alpha (until a value of 20, then they are stable). If we play with other parameters, the modes shift as well. We would like to understand how we can set the parameters in order to have convergence and a meaningful result.


    • Guilin Sun
      Ansys Employee
      When PML is close to the waveguide core, or the mode has long tails the PML can affect the mode calculation.
      Do you mean the spectrum shifted due to PML? it seems most likely the PML has some strong reflection. Please increase the PML distance. You can view the mode profile in log scale, and check which side has larger intensity at the boundary, then increase its distance .
      Please visit this post Ansys Insight: Choosing the most appropriate profile for PML boundary conditions for some info regarding to PML parameters.
Viewing 1 reply thread
  • You must be logged in to reply to this topic.