101106-1442 EDT
bendomaniac:
Consider this approach.
The power dissipated in a resistance is I^2*R. Therefore wire temperature is a function of the current thru the wire.
The power dissipated in a transformer or a wire determines the temperature rise. Wire insulation fails based its absolute temperature. Temperature rise is the temperature above the ambient temperature. So a 105 deg C rated wire can handle less current at a high ambient temperature compared to a low ambient.
Now consider the pole transformer. Its secondary may be considered as one continuous coil of wire with a tap at the center. Neglecting some other considerations a rating of 100 A thru the secondary applies to anywhere in the secondary. So if you use only one half of the secondary by applying a 100 A load on one phase, the maximum based on the wire, then your maximum load is 120*100 = 12,000 VA. However, the transformer is rated for 24,000 VA if you fully load both sides of the secondary, 240*100 = 24,000 VA. If you put a 24,000 VA load on one half of the transformer secondary, then that secondary half would burn up. But no problem with a 24,000 VA load across the 240 V output.
You can use any distribution of loads on this transformer so long as you do not load either side more than 12,000 VA.
In actuality there is a little slack in the transformer when using only half of the secondary. However, if the above discussion is applied to the service lines, then it is a precise description.
Insulation does not just abruptly fail at its temperature rating. So operating at 104 deg C does not mean life time is infinite, nor does operation at 106 deg C mean immediate failure. Life of insulation is a function of its prior history, and its current stress level.
.