zapperzero said:
The uncertainties associated with annealing aside, many things which are very improbable on a 40-year timescale and were left out of the design basis, are by definition certain to happen at least once, on an infinite timescale.
Annealling doesn't restore the full strength of the vessel. There's a good graph of it out there (that I can't find at this moment), that basically shows how you gain some life back, but its never going to be as good as the first dozen EFPY (Effective full power years) on the vessel. You can probably get most pressure vessels to 80 if you anneal, and you have been very kind on your vessel.
Assuming the plant was well designed and has no other major concerns like ASR (Seabrook station), and they haven't done an excessive amount of freeze seals on their ASME class 1 pipe, the most limiting component for life of the plant is the beltline of the reactor pressure vessel. This is the region around the active fuel. When a reactor vessel is built, it is rated with a certain number of EFPY (Effective full power years). For plants in the US, it was assumed that you would have 32 EFPY out of the vessel. This is the equivalent of an 80% capacity factor for the life of the plant, or running 32 years at 100% capacity factor. EFPY looks at the effects of radiation/neutron fluence to the vessel itself.
All vessel EFPY estimates include a certain number of normal, abnormal, emergency, and faulted cycles. Normal cycles are things like boltup, normal heatup, normal cooldown, hydrostatic pressure tests, daily power reduction, and a certain number of scram cycles. Abnormal looks at things like loss of some feedwater heating, turbine trips with bypass, rapid heatup and cooldown, cold water shocks. Emergency conditions look at things like turbine trip w/out bypass, total loss of feedwater + ECCS injection, MSIV (main steam valve) fast closure, maximum cold water injection shock. Faulted looks at DB LOCA and things with rapid depressurization, rapid pressurization, and things which could severely fault the vessel. There are engineers that keep track of every heatup/cooldown of the plant, and they have to tally these cycles against the number assumed in the original design of the reactor.
Each one of the events has a tabulated number of cycles. When reactor vessels are designed, they assume a poorly operating plant for 40 years, and that's how they come up with how many cycles they will assume. They then add these cycles to the total fatigue curves for the vessel end of life, (before they even happened). It's kind of like writing a check that you can cash later. As long as the number cycles I put my reactor through is less than what we assumed originally, the vessel is still good for use. If I go over the number of assumed cycles, I have to evaluate it. So for example, my plant assumes 10 total loss of feedwater heating events, where we go from full to no feedwater heating. As long as I don't use all 10 of those cycles, my vessel's EFPY curves are still conservative/bounding and my vessel is acceptable for use. If a plant uses up all of 1 type of cycle, let's say I used all my turbine trips or core spray injection cycles, I am allowed to do some limited substitution of a "Worse" event. So for example, high pressure core spray plants typically assume 10 HPCS starts over the life of the plant, and if I used most of those and I feel I need more margin, I can take some turbine trip without bypass cycles and convert them into HPCS injection cycles. The "exchange rate" (so to speak) is never favorable, as the goal is to ensure that your EFPY fatigue curves for your vessel are always bounding. I should make a note, that for the faulted conditions, like reactor emergency blowdown, only 1 cycle is assumed for the reactor, due to the extreme stresses it puts on the vessel. For people who work in GE plants, there's a set of prints that show these assumed cycles. For everyone else, if you look at a US plant's FSAR (Final Safety Analysis Report), generally somewhere in chapter 3. An example of this is in the following link from LaSalle station's FSAR (US BWR plant, BWR/5 Mark II)
http://pbadupws.nrc.gov/docs/ML0813/ML081330054.pdf
If you go to section 3.9, they talk about thermal-mechanical transients. If you go to table 3.9-24 they talk about each type of transient, the temperature changes it causes, how many cycles are assumed. You guys may like looking through all of chapter 3, as it discusses seismic criteria, wind/tornado flooding, etc.
As ASME code evolved, as we've removed those test capsules from the reactor to check for neutron embrittlement, as more refined computer analysis have been developed, we've learned a lot about radiation/neutron fatigue and its effects on the vessel. One thing we learned, is we overestimated neutron damage a significant amount. Another thing we learned, is changing core design allows us to reduce neutron leakage from the core, which further reduces neutron damage to the vessel beltline. This is the reason why many plants were able to perform EPU (extended power uprate), increase their power by 20%, and their original 32 EFPY fatigue curves were STILL bounding.
As plants age and have to go through license extension, these fatigue cycles, the EFPY fatigue curves, are all looked at. To get a license extension in the US, the licensee has to demonstrate that after another 20 years of life that the vessel will either A: still be bounded by existing analysis, B: that an updated analysis using new methods/codes and data shows that the plant is still bounding, or C: repair the vessel (annealing) such that it can withstand the extension period. A plant is allowed to not demonstrate this at the time of license extension as long as their is a commitment to do one of the previous options prior to reaching the original calculated end of life on the vessel. Palisades nuclear plant (CE PWR in the US) has chosen to go this route. There's also option D, shut down the plant, which is an economic decision if the plant does not want to anneal and option B isn't going to give them enough margin to extend the license.
Many plants are bound by A if they were "Good" to their vessels during the first 25-30 years of life. Some plants have to use B, like Palisades. I do not know of any US plant planning to anneal. Oyster creek is an example of a plant that had to use a new analysis, and now their 32 EFPY curves were extended to 38 EFPY. In order to extend these curves, they had to take penalties to the vessel minimum temperature for criticality, the number of allowable cycles they have, and they had to use new computer codes.
This brings up one more point. The whole basis behind the beltline fatigue, is the fact that the reactor needs to survive 1 emergency blowdown with a coldwater shock without breaking. The vessel is assumed to break if its Reference nil-ductility temperature after the accident is at or above 200 degrees F (this means the vessel will not be ductile while steaming is in progress). For those who are not mechanically based, Nil-Ductility Temperature is the temperature where something goes from brittle to ductile. When you are below NDT, an object will shatter, while an object above the NDT will bend and flex. In the US, there is a safety limit applied to the NDT (I think you take the vessel NDT and add 70 deg F to it for BWRs. I'm not sure how this works in PWRs). The vessel cannot go critical if it is below this safety limit NDT. As the vessel is fatigued, one option to extend the life of the vessel, is to raise the temperature allowed for criticality and pressurization. A typical BWR is required to be above around 120 deg F before going critical. A plant can opt to raise this temperature (no higher than 200 deg F) if necessary to maintain adequate safety margin to the NDT and extend the life of the reactor vessel. PWRs can do the same thing, raise the minimum temp for criticality, however I do not believe PWRs have the same flexibility with these limits as PWRs do. PWRs are likely to suffer much worse damage due to cold water shock events, and need to have larger safety limits.
This whole thing I'm talking about is why reactor vessels have a normal limit of heatup/cooldown of 100 deg F per hour, and why Fukushima unit 1 was cycling its Isolation Condenser on/off. Obviously if the operators knew they were about to lose the ability to cycle the IC, they would have left it on and chosen to violate their cooldown limit, rather than cycle it where they lost the IC.
I know this was a bit wordy, but with the number of comments on vessel life and fatigue, and previous discussions on the isolation condenser, I thought this might be a good thing to put up.