Tips for Using FEM
Several tips are presented to reduce runtime and memory consumption when using hybrid methods involving the finite element method (FEM).

 Decouple the FEM and MoM parts if the coupling is small

If the FEM part of the model does not have a significant influence on the MoM part of the model, then the MoM and FEM can be decoupled. This removes the coupling matrix and reduces computational resources.Note: For FEM models containing dielectric surfaces (excluding surfaces for modal ports) on the outer boundary of the model, add an air layer with a thickness of at least $\frac{\mathrm{3\lambda}}{10}$ ^{1} around the model when using the decoupling method.
The air layer reduces the number of surface triangle elements on the boundary between the FEM and the MoM. Resource requirements are reduced when the number of triangles decreases.

 Use a surrounding dielectric air layer to reduce the number of surface triangles
 If the outer surfaces require a mesh size much finer than $\frac{\lambda}{10}$ , add a surrounding dielectric layer of air (relative permittivity of 1). This surrounding air dielectric creates a new outer surface that can be meshed with a size of roughly $\frac{\lambda}{10}$ . The computational resources will be reduced with the reduced number of triangles on the outer FEM surface.

 Model internal free space regions as PEC FEM regions
 Set internal free space regions to perfect electric conductor (PEC) and set its solution method to FEM. This setting reduces the coupling matrix and as a result reduces the computational resources.

 Use first order basis functions if the geometry requires a very fine mesh

If the FEM part of the model requires very fine meshing compared to the wavelength in the dielectric (for example, geometrically complex objects) then using first order basis functions can reduce the computational resources and often improve the convergence as well.
If the MoM part is electrically large, then use the MLFMM / FEM instead.