Topics related to Lumerical and more

FDTD 3D convergence

    • Chinna Devarapu

      Hi, thanks in advance for your attention.
      I am using FDTD 3D in Lumerical to model and analyze a Multi-Mode Interferometer coupler but the simulation stops for divergence when it is 60/70% of the total simulation time.
      I have tried to change the mesh setting by using a dense mesh and a less dense mesh, or by increasing the distance of the boundaries in the simulation region but still same errors occur.
      I will show some pictures about the setting and the structure implemented.

      Best regards
      Thanks in advance for your availability.

    • Guilin Sun
      Ansys Employee

      Please do not specify the mesh size. Instead, switch back to the deault "auto non-mesh", and use mesh accuracy 2 for initial test. The override mesh seems too coarse for me. Please add another one, just round up the mesh size from its default values. If the geometry is not resolved by at least 2 meshes, decrease the mesh size.

      Once it does not diverge, you can gradually increse the mesh accuracy to 3 or 4, and at the same time, increase the number of PML layers to 24, 48 or so.  For higher mesh accuracy, you will need to use Custom PML with more PML layers.


      I would suggest to always use the default settings with minor modifications such as round up the values if you do not have a specific reason to modify them. 


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