This has been the subject of some debate. If you look closely at the description at the top of Table 310.16, one might conclude that it does not apply to your situation, for two reasons. First, you don?t have conductors in one of the three configurations it lists. They are not ?in raceway, cable, or earth (direct buried).? Instead, they are in raceway, and the raceway is buried in concrete (not direct buried in earth). Secondly, the table applies to ?not more than three current-carrying conductors.? Each raceway (conduit) has only three current-carrying conductors, but each is in the near vicinity of nine other such raceways. The heat generated by the conductors in each of the ten conduits will make it harder for the other nine conduits to reject their heat to the dirt that surrounds them all.
But Inspectors in some jurisdictions (Washington State is among them) recognize that the amount of load that is calculated per NEC 220 is going to be much higher than the actual load that will be seen on a day-to-day basis. So they view Table 310.16 as being applicable to underground installations, so long as the conductors are sized on the basis of an Article 220 calculation. However, if the load is determined on the basis of a measurement (such as might happen if you are adding load to an existing facility), it might be more difficult to convince the AHJ that Table 310.16 still gives a reasonable and conservative value for ampacity under the specific conditions of use.
You are right in saying the Annex tables are not enforceable. But if your AHJ will not accept Table 310.16, the NEC itself only gives one other option. You can have the ampacity determined ?under engineering supervision.? That would require a licensed Professional Engineer to perform a calculation, most likely using a software package that applies the Neher-McGrath analysis method. From the times I have performed such calculations, I can tell you that the calculated ampacity has always been much lower than the allowances in Table 310.16.
