Convergence testing is the method of varying the settings used in the simulation to ensure
that the simulated results are accurate.
The main sources of error for the EME method are the discrete number of modes used in the
modal expansion which was discussed in the Solver Physics section, and the discretization
of the structure using cells along the propagation direction.
The error in the modal expansion can often be reduced by increasing the number of modes
found in each cell, and the error due to using a discrete number of cells is reduced as you
increase the number of cells used in regions where the structure cross section is smoothly varying.
Since increasing the number of modes and cells increases the time that it takes to run the simulation,
the recommended simulation method is to first set up the simulation using low accuracy settings,
with a small number of modes and few cells in each cell group region,
and this will allow you to run initial simulations quickly and get some rough results to make
sure that there are no major problems with the setup.
Then, convergence testing by increasing the number of modes and cells can be performed
to determine the settings required to get accurate results.
The EME solver region also returns some error diagnostic results which can be used to troubleshoot
problems in the simulation setup, and these can be used if the results aren't converging
well from the initial convergence testing of the number of modes and cells.
The error diagnostics results will also be covered later on in this section.
Besides convergence testing, another way to validate the accuracy of the simulation results
is to simulate the same device using a different solver, such as FDTD.