Can't find any mistakes. I think this is a great design. Combine it with an industrial centrifuge and a breeder like this one and you can produce 156.4 million EU per cycle (the plutonium cycle of 20,000 seconds) for just 9.5 uranium. This gives an efficiency of about 16.46 and it will produce 63 thorium cells extra in 50,000 seconds (the length of the thorium cycle), which can also be burned in 7 of these reactors. Burning the extra thorium will produce a total of 387.4 million EU and the efficiency will rise to 40.78. Combined output of the 9 reactors would be 632 EU/t with just 4 EU/t (on average, with normal config the centrifuge wouldn't even run with anything less then 5 EU/t for this recipe). Just some tin is needed to make empty cells and 64 copper is needed to make the dual-plutonium cells.
EDIT: 3 of these could also work (instead of the 7 chamber-less designs). This would boost the total EU/t output to 656 EU/t (minus the 4 for the centrifuge, so 652 EU/t) and bring the efficiency to ((387.4 million EU + 24 million EU) / 9.5) 43.31.
Oooh thanks for the advice!
I've already been centrifuging since I use GT 'hard mode' and the matter fab I usually have turned off since it's foolishly hooked to my main power line which only produces 380 eu/t at the moment (I only use it for the GT machines and a few macerators/compressors/etc) but the idea you have is great. I'll try to come up with a nice design using thorium and uranium as well! Your reactor design looks nice and cheap 
Never researched breeders before but I'll set this up and I assume it will re-enrich my cells?