SIW Simulation using HFSS

I am trying to simulate a Substrate Integrated Waveguide structure in HFSS at 2Ghz frequency. I need some help to set up the excitation and boundary conditions. I have a microstrip to SIW transition, SIW structure followed by a SIW to microstrip transition. Right now when I use a driven terminal approach with lumped ports I get the error that there is only one conductor in one of the ports which I kinda understand since the top and bottom planes are shorted through the via holes. So do I use only Driven modal with waveport excitation? Also do I need any air box or radiation boundary definitions when simulating this structure?

As long as you are using a waveport, be sure the excitation plane is on the edge of your simulation. A waveport is a port that simulates an infinitely long waveguide as the source. Could you include a picture of your simulation thus far? If you have vias or slits, then I would think you would want to include an airbox larger than your waveguide and assign there a radiation boundary.

As far as your error, I'm pretty sure you shouldn't get an error for the conductor being shorted elsewhere, that would prohibit simulating a lot of common structures. You probably will get an error without an integration line leading from one conductor to another on the plane of your excitation.

I removed the microstrip to SIW transition and was able to excite the SIW part alone using regular wave ports. That gives me the right results. So I guess the problem is with the microstrip to SIW transition. So I need to figure out a way to get that transition right. Find the images of both the setups I am using below

Also when I am simulating a guided structure like a waveguide what is the meaning of including a radiation boundary or an air box. Ideally the structure should not radiate right? The boundary conditions should take care of everything. I did not define it when simply simulating the SIW part alone which is just like a waveguide and it works just fine.

You won't have waveports on all sides of your simulation domain. What is above the above the upper surface?

While for a guided mode, the difference between a radiation boundary and a PEC can be small (perhaps negligibly so), generally one wants to avoid a situation where it may be a critical difference, and err on the side of caution.

The structure is a dielectric with two layers of copper on bottom and top faces. The bottom and top layers are shorted through the dielectric by array of metallic vias. I have defined waveports on the two faces (perpendicular to Y axis on the first pic) and defined modes. I did not define an airbox/radiation boundary, but the results were as expected. I had defined the via material as copper and the bottom and top layers are also copper. I selected all vias and set them as radiation boundary and that also as expected gave the same results.

Don't you have a MS to SIW transition? If you only actually have a completely covered dielectric, then yes, you don't need a radiation boundary at all, assuming the waveport covers the entire face on either end.

I'm not sure what you mean "as expected". I wouldn't expect much from doing this... was there a particular reason?

Yeah.. I need to insert a MS to SIW transition. That's when its behaving weird.. So I do have to define a radiation boundary when I include the MS to SIW transition right?

Yes, that is correct.