After doing some research I've found some problems with using an RTG as an AIP system, the weight. The best RTG today produces one kW of energy per 200 Kg of equipment. This is because thermoelectric generators today have very low efficiency (very little energy radiated from the nuclear fuel is converted to wattage). Lets see how this stacks up against the power requirements of a modern sub, a Kilo class submarine's electric motor produces 4,400 kW worth of torque on the screw. 4,400 * 200 = 880000 Kg, 880000 / 1000 = 880 tons of weight. Thats quite a bit of equipment on a 3000 ton sub. If the whole battery and diesel engine system was swapped out for RTG then it might work but there is another problem getting the fuel. RTGs run on Pu-238 (Plutonium 238) while nuclear reactors run on U-235 (Uranium 235). Uranium is naturally occurring (you dig it out of the ground as an ore) while Plutonium is man made (its made out of Uranium in a nuclear reactor) Am-241 (Americlum) which is also used in RTGs is also man made so time and expense becomes a problem.
So a sub powered by RTGs is possible but expensive and wasteful compared to the alternatives. RTGs are better for powering things that need power for decades even centuries which at this time is only deep space probes. Maybe future sea floor research platforms might use RTGs but I think it would be simpler to use hydrogen fuel cells for those if a effective way to distill hydrogen from sea water becomes feasible, or just run a cable to the surface for power.
Go check out the beginning design page over at projectrho.com's page on atomic rockets for the numbers I've used for my calculations on RTGs.
__________________
 
|