Not much thought does go in, 95% of the time you keep things reasonably standard and that makes stuff so much easier, its just if you want to break away into slightly more unusual arrangements that are still permitted by BS7671 that you start having to do calculations, etc.
Basically the Adiabatic compares the let-through energy of the device (I²t) I = current, t= operating time, to the amount the cable can safely take (K²S²) k = special constant from tables taking into account resistivity, maximum temperature rise, specific heat capacity, etc, S= CSA of conductor. It appears arranged in serveral different ways, but basically its just about making sure the cable withstand is higher than the let-through as I said.
Yes, its firmly based on P=I²r, r is factored into the K factor though
You have to consider it for the worst case I²t, generally fuses will give the worst case I²t at the low fault level end, and breakers at the high fault level end, but to be safe, work it through for both extremes.
Most of the time we don't bother with any of this though, there are other ways you can design that don't require this working out, although thicker cables may be needed[/i]