Description
This is a self managing powerplant. It has a bootstrapping routine and only needs fuel cells in the 4 chests. No need to manually insert a fuel cell into the core.
The plant is manged by a finite state machine. It takes these sensory inputs: core temperature, core fuel level, steam buffer amount, pump mode. Those values are interpreted and then transformed into a 4-digit number representing the current state. It does control the pumps and refueling.
I have separated the reactor core from the steam buffer (and from the turbines) by pumps. The reason: When heat exchangers can't produce steam, the whole core unit can't lose temperature. Thus, if enough steam is available, the turbines will work only on buffered steam.
The plant will go from one state into the next. In some states, one of 3 actions is taken: 1) refuel, 2) activate the pumps, 3) deactivate the pumps. Most states are reached on their own - by time and power demand. The plant will try to prevent overheating. And it will try to prevent cooling off.
I have set limits for core temperature and steam buffer. Those are arbitrary. Also, I have reasoned which states need what intervention. Those are encoded in the several constant combinators. That whole logic sits on the left - between the core and the buffer.
Top left is an emergency power system. In case the main grid does not provide electricity, a global warning sounds, the critical infrastructure is separated from the main grid and connected to the emergency system. Reconnection is done only, when the main grid is stable enough (meaning that the one accumulatur has more than 50% charge)
Top right is a watchdog that monitors the type of fluid in the system: Water and Steam. It won't allow refueling when wrong fluids are detected. Connected to this is the bootstrapping mechanism for the first fueling.
Diagonally are warning sensors when either not many fuel cells are left, or the chest with the burned fuel cells is nearly full. Those limits are hardcoded in the decider combinators. This system monitors each core on its own.
The whole system is carefully isolated from outside powerlines. Also, the circuit networks are carefully isolated. I have marked critical powerpoles (and everything else I deemed critial) with warning concrete. If you want a functioning emergency power supply, do not connect any powerlines to poles on warning concrete.
It was agony and fun to design. I use the "numbers display" blueprint from a reddit thread (shitbnag, Twellux and realboabab) to display the current state. You can remove that. And the "Power Failover Switch" from Nius Atreides.
I was very happy to see it boot itself up and just chugging along, going from one state to another all by itself. That was a magic moment.
I originally wanted to also manage the connection from the steam buffer to the turbines. But I won't do that. The idea was "be more modular", but also: this separation gives the plant the ability to produce more electricy than the core can generate - for a limited time. Imagine full steam buffes, or a full, large secondary buffer. You can provide steam to more than 83 turbines under full load, if you pump more than 6k steam in and have a reasonable large buffer. Those additional turbines could be activated when needed by having them under a separated and switchable power grid.
