Sounds reasonable, but with randomness thrown into the mix there would be some uncertainty, but I think that would be manageable. I like the data to be a bit obfuscated, so simple redstone connection loops aren’t enough and there is some challenge and use case for advanced redstone engineering or computers.
For the sensors itself, I would recommend being able to set a minimum and a maximum neutron count, the minimum signifying a redstone strength of one and the maximum (and above) a strength of 15, neutron counts between the two values get interpolated redstone strengths. This allows some degree of finer measuring, without making the sensor terribly more complicated.
I think completely different systems for different coolants may not be such a good idea, as it makes learning to build reactors very complicated, potentially bloats tool tips and is very hard to balance. There needs to be some common ground, the coolants only adding their unique twist onto the system.
Fuels having different factors across different systems (or not having a factor at all because the system is so different) would be very confusing, just because the current system doesn’t allow for factors greater than 1/4 (I checked it this time, 1/5 was false). That’s why I think changing how the durability for fuel rods works would be the best way, as it doesn’t change the system so that current setups would break, while also allowing for fuels with a factor greater than 1/4.
But things going critical too fast is a valid concern, as the neutron economy gets updated every tick. When having a fuel rod with an emission and self of 128 and a factor of 1/4 surrounded by reflectors, it would take two seconds for the neutron and therefore heat output to rise to 20608. After 47 seconds the heat would be so big, no pipe could transfer the coolant needed to cool the reactor.
Reducing the neutron counts growth artificially however would make high neutron economies have the same level of risk as low ones, because there is no exponential growth. A better way to solve this would maybe to change the update-rate of the reactor, similar to the IC2 reactor, which only updates once per 20 ticks. The update rate could also be a game mechanic the player can control, dependent on the coolant.
I meant the extra durability loss only to be added when the rod is active, when it normally loses durability, not when it’s inactive. Wires not being used don’t consume electricity either.