Thank you for your comments,
I'm doing some calculations and simulations in order to verify if I
can bypass representation problem of s-parameters in other ways rather
than the Thevenin one.
There is one thing I'd like to submit to your attention, the known
formula
S11=(Zin-Z0)/(Zin+Z0),
where Z0 is the port impedance and Zin the load, does not works if you
consider Z0 not pure real (usually 50 ohm) but composed by a real and
an imaginary part i.e. Z0 = a+jb.
Are there other known formulas for S11 for Z0 real+imaginary?

Camelot