Ryan,
I am in agreement on the general point of calculating the fault current to see if it is sufficient, and presumably calculating the energy delivered to the conductors in a fault situation to make sure that the OCPD protected the conductors. However I believe that the point of this exercise is to size the _parallel_ EGCs so that they correctly function even when _not_ effectively in parallel.
In the example situation that started this thread, 4 parallel 1/0 conductors were used as the EGC for a 1200A panel. Table 250.122 calls for a single 3/0 conductor. The total cross section of the parallel 1/0 conductors is in excess of 2.5x the cross section of the 3/0 conductor; so _taken as a set_ the 4 1/0 conductors are clearly a better EGC than the single 3/0 conductor. Clearly then, the issue is 'What happens when these conductors are _not_ effectively in parallel?'
The situation that I was suggesting requires separate calculation is as follows:
Equipment ground bonded at _both_ ends of the metallic raceway.
Phase to ground fault _inside_ of the raceway.
Now you cannot consider the two conductors to be in parallel, but would instead have to consider the following two paths in parallel: the 'short' path through the EGC to the supply end of the raceway, and the 'long' path through the EGC to the load end of the raceway, and then back through the raceway to the supply end.
-Jon
