Grid Convergence

[Paria M.]

In the TI planning tool, I came across the statement, 'A grid convergence analysis was conducted, varying the resolution between 0.3-1.0mm, which confirmed a resolution of 0.5mm to be sufficient.' Could you please provide guidance on how to perform a grid convergence analysis like this?

[brown]

To perform such a study, define your region or quantity of interest, then parametrize it. After a quantity has been parametrized, you can run multiple simulations using the built in Sweeper feature or using the Sim4Life Python API.

There is an example in the Sim4Life manual titled 'Parametrized Patch Antenna' that shows how to use the Sweeper feature.

If using the Python API, the easiest way would be to create a baseline simulation by hand, then right-click on the simulation name and select To Python. Then you would find the quantity of interest in the auto-generated script and assign it to a variable instead of the hardcoded value. Using a simple loop in Python, you can create multiple simulation that have identical setup except for parameters of interest.

In your example, multiple simulations would be run with identical setup except for the grid resolution within a region of interest (e.g., a wire block surrounding a region of interest is placed in a manual grid settings folder, then Maximum Step is changed from 0.3 to 1.0 mm).

Then in Analysis you extract a quantity of interest and compare how this value changes as a function of grid resolution.

If the change in the value is small as you increase resolution, you can proceed in future simulations with a coarser resolution to save on run time. The exact convergence of the value needed depends on your application.