(Warning: talking out of my left *censored* and well outside of my experience. Take the below thoughts and evaluate them using your own experience and knowledge!)
Which suggests that no additional capacity is needed for these new receptacles, since you are not adding any square footage to the rooms. With no additional capacity required, one could get away with 0kVA transformers and panels *grins*
Actually, as the original post is worded, I don't think that the NEC provides a useful method of calculating demand for these circuits.
The rooms are already fed with circuits, and presumably these feeders have already been properly sized for the square footage of the room. These new circuits are going to be fed from a separate panel with a separate transformer. How many square feet of room area do you use to size these new feeds? The full size of the room? 1/2 of the room? A weighted average of room area based on number of receptacles?
IMHO the approach should be to feed these new circuits from the same panels that currently feed the rooms. The feeders to these panels should already be sufficient for the square footage of the rooms. Redoing the demand calculation for the entire rooms using the current NEC would clue you in should the feeders need to be increased in size, but I bet that this isn't a problem.
Presumably you would need to use additional panels to make room for the additional circuits, so the only thing left in question is the size of the feeders to these new panels. The panel rating itself is not going to be in question; it is almost certainly going to be set by the number of individual branch circuits (though I bet if you hunt, you will find someone making a 30A 42 circuit panel
)
Now you've reduced the unknowns to a small portion of the job: the feeders to the new panels. Remember I am making the claim that no new transformers would be needed, since the transformers and feeders for the original panels were based upon the demand calculations for the entire square footage of the rooms. At this point you could oversize these feeders at minimal cost, perhaps even saving some money. For example, if the feeders are sized to match the feeders going into the source panel, then you probably eliminate a step of OCPD.
No new calculations (just repeating the old calculations), NEC compliant, without adding lots of what is almost certainly excess capacity.
-Jon