This video is taken from the FDTD Learning Track on Ansys Innovation Courses.
Now that we’ve seen the different settings that can be used for setting up the simulation
mesh using the Mesh settings tab of the solver region and mesh override regions, we’ll
go over some tips for setting up the mesh.
Using a combination of the auto non-uniform mesh type with mesh override regions is recommended
since it allows you to control the mesh based on a target number of mesh points per wavelength,
and also allows you to specify a finer mesh where needed to resolve fine geometric features
or regions that are sensitive to meshing such as interfaces between metal and dielectric
When using the auto non-uniform mesh, the mesh accuracy setting of 2 which corresponds
to 10 mesh points per wavelength is considered reasonable for the FDTD method, and mesh accuracy
4 or 5 which corresponds to 18 or 22 mesh points per wavelength is considered high accuracy.
It is worth running the simulation at mesh accuracy 1 or 2 for initial simulations in
order to be able to run simulations more quickly, before refining the mesh for high accuracy
results since the simulation memory scales as 1/dx^3 and simulation time scales as 1/dx^4.
For the mesh refinement, as mentioned earlier, conformal variant 0 applies conformal meshing
to all materials except metals and conformal variant 1 applies conformal meshing to all
materials but it can sometimes lead to numerical artifacts if the mesh is too coarse, so a
simple rule of thumb is to always use the default conformal variant 0 method, and only
use conformal variant 1 when convergence testing of the mesh step size will be performed for
getting high accuracy results.
Comparing the results obtained using conformal variant 1 and conformal variant 0 is also
recommended to make sure there are no major discrepancies between the result when using
conformal variant 1.
After setting up the simulation mesh, the mesh that is generated can be checked by using
the “View simulation mesh” button to show the mesh in the CAD view ports, and you can
also use the refractive index monitor to view the meshed structure.
The mesh that is generated will try to meet several constraints including placing a mesh
line at the position of 2D objects, placing mesh lines at the edges of the mesh override
regions, grading the mesh between fine and coarse mesh regions, and meeting the minimum
mesh step size.
Sometimes it is not possible to meet multiple constraints at the same time, so it can be
important to check the mesh that is generated and make sure that it is reasonable and that
all the structures are adequately resolved.
When simulating periodic structures, it’s best to use a mesh with an integer number
of mesh cells per period in the direction of periodicity in order to make sure that
each period is meshed the same way.
This will ensure that the simulated periodicity of the structure will have the same periodicity
as the physical structure, and this can be important for devices such as cavities or
resonators where the quality factor and resonance wavelength of the cavity is sensitive to the
periodicity of the structure.