How does HFSS differentiate between signal and ground conductors?

Upon exciting a structure via wave-ports/lumped ports, how does HFSS differentiate between signal and ground conductors? How can one assign them separately? I thought, integration line was responsible for differentiating between these two. But, it is not!

For example, if one has a 2 plate concentric cylindrical conductor and excites them with a waveport, which one of them would carry the signal? Which would be the ground? How does HFSS decide this?

It doesn't decide pal. The excitation integration line can even be inverted and still giving the same results. it is not even important to decide that. The structure is excited with a field and the solution which includes fields and s parameters are calculated depending on Maxwell equations and boundary conditions. So as i know it isn't important on EM problem solution to know where is the ground and where is the signal.

The concepts of "ground" and "signal" lines are purely artificial. A TEM mode is supported by two or more conductors -- choosing to establish a reference conductor ("ground") is completely arbitrary, and only done out of convenience. In complex systems, we generally establish a global reference (think the conductor backing "ground" plane of a PCB with multiple microstrip traces), because then we only need to refer to each of the smaller traces when referring to a "signal". However, we could just as accurately choose one of the "signal" traces to be the reference and refer to the "signal" on the "ground" plane.

Thank you PlanarMetamaterials and moh.haroun!

More doubts follows!

1. Why is only TE10 mode considered in a waveguide (dominant mode; I know! Any particular reason?)? Why not TE20 or TM11? Is it because more connectors/probes are need for exciting/detection?

2. How does one excite the structure with higher order modes like TE20 or TM11 using HFSS? Draw 2 integration lines for TE20?

Yes, ease of excitation/detection is a large part of it. The TE10 mode, being dominant, is supported at the lowest frequency. This means that the for a given frequency, the waveguide can be physically smaller when using this mode - and space savings is usually equated with money savings.

No -- you simply instruct the waveport to support more modes. Each mode will have its own integration line (the direction of the E-field). HFSS will generally find modes in order of increasing cutoff frequency.

Is the dominant mode propagating and the other modes evanescent?

In general? No. The dominant mode is just the mode with the lowest frequency cutoff.