For the record, all testing has been done with TMi up to this point.
Reactors running using additonal mods should probably have their own class. They dont necessarily have to play by the previous rules.
maybe mark-m (for modified)
When using buildcraft to inject ice, its important to remember the reactor fills from left to right, then top to bottom. So if there are cells depleting at the top and bottom, it will fill the top cells first, THEN the bottom ones after the top ones are full. If the cells at the top are depleting fast enough to absorb all the incomming Ice, the bottom ones will not be refilled.
Also, reactors will only pull one ice per stack per tick. so if your reactor is producing 1000 heat per tick, and you have all your ice in a single stack, its only going to pull them one at a time for 200 cooling per tick. Good luck dealing with the other 800 heat. This gets problematic when your reactor is only filling one of the stacks of ice, and is letting the other 4 empty out. I am pretty sure this is how I lost reactor 2.
I lost the first reactor due to a poorly placed mfsu. I was testing how high i could push the output. I finally put enough uranium to go over 512 output, it popped the cable and blew up a nearby mfsu. That took out the ice feeder lines, and things went bad before i could hook them back up.
reactor two lasted much longer and featured a design where if i popped an mfsu, iy wouldn't blow up anything else that was critical. but i didnt keep a good eye on it, and i am pretty sure it ran out of ice.
using equivalent exchange + build craft to convert lapis into groups of 16, 32 or 48 ice. This allows for MUCH faster refills off of a single pulling pipe. without it you need several chests pumping materials to get the necessary volume, and it gets laggy in the pipes.
Reactor 3 was when i figured out the lapis trick with eqex. It ran very well (748 continuous output)for many hours. Then the cells started to deplete and for reasons i cant identify, the whole thing exploded. But not before filling up 20 mfsu.
A question on "health bars" of things in the reactor. For cooling cells and plating, im pretty sure the health is how much heat they are currently holding. the closer they are to overheating, the less health they have.
But for uranium, i think its the amount of time left until they switch to "depleted" If i am wrong, it could explain a few things.