Is there a way to define a localized incident TEM wave in the surface integral formulation of the moment method? This can easily be done in FDTD, but I have no idea if it is possible in MoM.
I now know the answer so I thought to post it here in order for others to know.
You can define a localized incident TEM wave but if only it is incident on a surface, you cannot ,for example, define an incident plane wave inside a coaxial cable, since it won't be incident on a surface and you will have no means of including it in the formulation.
I think I didn't well clarify my point, for example when you excite a wave in a microstrip line you use a some sort of localized TEM wave when formulating the problem using FDTD. This was possible because of volume meshing. However in the case of MoM a lumped port (as you said) may be needed to model such an excitation.
For the FDTD case you mentioned, it is a TEM wave, with well define boundary. You wouldnot want to add the term "localized" to it. As soon as you restrain/localize the source to a subregion, it is no longer a TEM wave.