A bit late, but here's what I derived from the mod's code:
- Fuel rods produce heat equal to Reactivity * 2.25MW
- Fuel rods gain Reactivity based on current Reactivity^2, Temperature, (Power Output)^2, and a small base amount.
- Boron Control Rods work by periodically multiplying the Reactivity in each connected Fuel Rod by a multiplier between 0.8376 (at I=3,000) and 1 (at I=0). While the exact formula is
(1 - 0.83*I/3,000)^0.1, this is more-or-less linear for our purposes in the range that matters.
- Don't use graphite rods. They could probably be used to warm up your reactor faster, but the control is way more complicated.
- The amount of energy removed from the fuel cell each tick is scaled down by Fuel Rod Efficiency, which peaks around 16 between 625° and 920°. This is much better than normal reactors, which peak at 4 with a +300% neighbor bonus.
- Fuel Rods produce a minimum of 1.8MW but radiate away a small amount of heat, scaling with temperature. So long as you damp them fully (I>=3,000), they'll reach 1,100° or so but won't ever explode.
So, to make a reactor:
- Surround a single Boron Control Rod with up to 24 Fuel Rods, with it at the center of a 5x5 grid.
- Wire all the fuel rods together and divide everything by the N signal (the total fuel rod count) to get an average.
- If our target output per rod is 80MW, then subtract 80,000 from the KW signal, divide by some damper constant, and feed this into the Control Rod on the red wire as signal I. I went with /23 after completely mis-understanding how the reactivity code functioned, but it seems to work just fine.
- This will hold the power output steady at 80MW, but now there are 2 problems:
- At max power, the reactor will settle at 500° (400° with Coolant Heaters instead of Exchangers) and won't run at max efficiency.
- Far more importantly, if the Exchangers/Heaters draw <80MW of heat per rod, the temperature will instead increase indefinitely until everything explodes.
- To fix this, we set a target Temperature (I went with 625°). Subtract from the average T signal, multiply by a scaling factor (250 seems to work nicely; higher keeps the temperature closer to target but produces larger oscillations), and feed this back into the input of the previous combinator on the KW signal.
- In short, if the reactor's too cold, increase the power target above 80MW until it warms up. If it's too warm, decrease the power target and let the Exchangers cool it down. It won't ever quite cool down to the target temp, but my 1.92GW reactor only hits 965° at 60MW output (2.5MW/rod). I'm sure this could be done better, but I haven't quite gone far enough down the rabbit hole to build or tune a proper PID controller for any of this.
Of course, all of this starts with the assumption of 80MW per fuel rod, a number I arbitrarily pulled out of the changelog. You can absolutely generate hundreds of MW per rod, but with the above control circuitry they tend to get very hot when severely underloaded.
