Reply #78: After going back to basics, I agree.

[GliderGuider]

pD is the probability of one reactor having core damage in any given year = 0.0001
pF is the probability of containment failure following core damage = 0.1
nY is the period in years = 25
nR is the number of reactors = 440
pLER is the probability of an LER in nY years

pLER(1) is the probability of a LER from one reactor in nY years
pLER(nR) is the probability of a LER from nR reactors in nY years

pLER(1) = (1-(1-pD)^nY)*pF
pLER(nR) = pLER(1)*nR

pLER(1) = (1-(1-.0001)^25)*.1 = 0.00025
pLER(nR) = 0.00025*440 = 0.11

So I agree that assuming the accuracy of pD=0.0001 and pF=0.1, the current Gen II fleet has an 11% probability of an LER over the next 25 years.

I have a few questions perhaps you could help with:

Do you know whether the "L" in "LER" is quantified? Does it refer to a release on a Chernobyl scale, or a TMI scale, or is it relatively undefined? It makes a difference in terms of how exercised we should get over that 11%.
How did they come up with the 1/10,000 for pD?
How did they come up with the 0.1 for pF?
I saw core damage estimates on the order of 10^-7 and 10^-8 for Gen III reactors. Thoughts?
I get the impression that the Gen IVs when they are built will be intrinsically safer than either Gen2 or Gen3. Maybe it's time to skip a generation?

In any event, it looks as though the answer from a safety perspective is to build out the more advanced designs as fast as possible, retiring the older reactors in the process.

As with nuclear weapons, we are actively addressing the risk factors in nuclear power, and making impressive strides doing so. The probability of a carbon catastrophe within the next 100 years, on the other hand, is rising, and is rapidly approaching 1.0.