- Clean fit up against a bus (yes, I put nuclear processing on my bus)
- Low cost to build (relatively speaking)
- High efficiency (respectable neighbor bonus)
- Auto-start (no priming necessary)
- Automated control (no fuel wastage)
- Manual overrides (for debugging/servicing)
- Deep throttling (can throttle between 0~480MW without energy loss)
- Failsafes (handles low/no fuel, power outages, contaminated belts, etc. gracefully)
- Customization (interface at steam/water pipes)
As recorded, the blueprint is configured for belting use, but it can easily be converted for bots by removing the belts, replacing the two lower iron chests with fuel chests, and placing depleted cell collection chests under the appropriate long-arm inserters.
I've intentionally excluded turbines and pumps from the blueprint, because I want flexibility in how to pipe in water, and I don't want to invest in 80-something turbines right off the bat. (I usually don't need that much power to start.) I've therefore made the water and steam pipes accessible, and as long as 5 pumps are connected to any 2 sides of this thing, water flow appears stable for at least a few full cycles. It's also possible to lop off at least 20 heat exchangers and their heat pipes for further build cost reductions, without compromising the overheat protection, so long as you don't try to tax the plant above its new power limit (-10MW per exchanger excluded).
I did design and test for all the weirdness that may occur during normal operation of a base (fuel starvation, blackouts, brownouts, waste backup, etc.), and this setup will simply stop whenever a blocking condition occurs, and resume operation normally once that condition is removed. Finally, the logic is designed to start from the blueprint state without priming, so I think I've achieved my goal of set-and-forget; once you hook this up, you shouldn't have to futz with it. (See circuitry section for details.)