Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Potentially serious situation averted at Davis-Besse NPP

  1. Mar 15, 2013 #1
    A potentially serious situation was narrowly averted after a very large hole was discovered on the lid of the reactor pressure vessel at the Davis-Besse nuclear power plant in Ohio. Apparently this hole was caused by acid corrosion and was only about one centimeter away from eating all the way through the lid, which would have perforated the RPV.

    Hypothetically, what would happen if there was a hole or breach in a reactor pressure vessel during full operation?

    http://events.nace.org/library/corrosion/NuclearIndustry/nuclear-accident-2.asp [Broken]
     
    Last edited by a moderator: May 6, 2017
  2. jcsd
  3. Mar 15, 2013 #2
    It was ten years ago, and discussed many times already.
     
  4. Mar 15, 2013 #3
    The plant safety analysis considers the effects of large holes in the reactor coolant system (called LOCAs). The only good things we can say about the Davis Besse hole are that (1) it was smaller than the analyzed hole, and (2) if you're going to have a LOCA, a hole in the vessel head is probably the best place to have it (least likely to cause fuel damage and radioactive releases).
     
  5. Mar 15, 2013 #4
    High pressurized coolant would be ejected directly at the control rod drive. So I wouldn't consider this to be the "best place" to have a leak.

    In the worst case you have a loss of coolant accident AND troubles with inserting (damaged) control rods.
     
  6. Mar 15, 2013 #5

    NUCENG

    User Avatar
    Science Advisor

  7. Mar 15, 2013 #6

    jim hardy

    User Avatar
    Science Advisor
    Gold Member
    2016 Award

    it is fun to sound exciting but it is not really honest to exaggerate .


    Leaks generally start small, like a crack or pinhole and progress. That's why accident analyses assume a clean guilliotine break which is more severe.

    What happens with a small leak is humidity in the containment starts to rise, operators notice increase in normal makeup, and containment sump pump runs more often. If there are TV cameras in containment they may see steam or note water droplets on the lens.

    Head leaks are not without precedent. They spew steam so boric acid quickly builds up in the vicinity - it looks like snow. They're easy to spot and not terribly difficult to clean up after.
     
    Last edited: Mar 15, 2013
  8. Mar 15, 2013 #7

    QuantumPion

    User Avatar
    Science Advisor
    Gold Member

    No, the control rods are not exposed in that way. The rods themselves are internal to the vessel. Any damage to the control rod drive mechanisms would cause the rods to disengage and insert into the core.
     
    Last edited: Mar 15, 2013
  9. Mar 15, 2013 #8
    In a large break LOCA the reactor is shutdown on the voiding (all the moderator is gone). The safety injection flow that refills the core is borated. We dont credit rod insertion.
     
  10. Mar 15, 2013 #9

    Astronuc

    User Avatar
    Staff Emeritus
    Science Advisor

    Had the stainless let torn (cracked) or the CRDM let loose, it would have been a small-break LOCA. The reactor would have tripped, and in theory, the safety systems would have done their job. As it was, the stainless steel liner held. This was an example of leak-before-break, and it didn't break.

    The plant management got dinged because there had been signs of some kind of leak during the previous outage, about 20 months earlier. They didn't think it was a big deal.

    It was in a difficult location to simply just look because there is an array of CRDM housings on top of the RPV head.
     
  11. Mar 16, 2013 #10
    What is "LOCA" an acronym for?

    And how did this hole evade detection by NRC inspectors for so long?
     
  12. Mar 16, 2013 #11

    Astronuc

    User Avatar
    Staff Emeritus
    Science Advisor

    LOCA = loss of coolant accident.

    The hole was small and concealed by the equipment on top of the reactor vessel head. The area is enshrouded to cover the electrically powered drive mechanisms that raise and lower, or otherwise suspend the control rods above the core.

    http://www.nucleartourist.com/images/reactor.gif

    Boric acid had been detected in a previous outage, and it should have been investigated since it doesn't belong there under normal situations. It should have been interpreted as a sign of leakage, but then the primary system wasn't losing much coolant, or rather the loss was below action limits.
     
  13. Mar 16, 2013 #12
    Theoretically, what would have happen if the hole breached the RPV while it was running? Could it have caused a serious accident or damaged the core/released radiation?
     
  14. Mar 16, 2013 #13

    Astronuc

    User Avatar
    Staff Emeritus
    Science Advisor

    There would be a steam jet. It would be a small-break loss-of-coolant event. The reactor would be shutdown and normal core cooling operations would commence. Some radiation could be dispersed into containment and actions would be implemented according to those circumstances.
     
  15. Mar 16, 2013 #14
    But it probably wouldn't have damaged the core or resulted in a meltdown?
     
  16. Mar 16, 2013 #15
    Right. But this was taken as a very serious event within the industry. It is one of the events we teach our new engineers about, there are many "lessons learned" from this.
     
  17. Mar 16, 2013 #16

    Astronuc

    User Avatar
    Staff Emeritus
    Science Advisor

    Correct. The reactor protection systems are design for large-break LB-LOCAs, e.g., if the primary system hot legs - highest temperature water, or cold legs - cooler water being returned to the core - break. Then there is the class of small-break (SB-LOCAs) LOCAs, which are analyzed.

    The analysis tests how the system would respond and what core damage might be expected. Systems are designed to provide emergency cooling based on analysis. The models in the analysis are based on separate effects experiments. The numerical models can be incorporated in simulators and the operators can then test 'what if'. It is incumbent upon those devising the simulations to ensure that the 'what ifs' encompass possible scenarios - e.g., seismic, LOCA and loss of power or station blackout (SBO).
     
  18. Mar 17, 2013 #17
    It was a serious event indeed. I have a few questions.

    Why reactor head wasn't examined every refueling - the head is removed and reinstalled anyway, it should be relatively easy to inspect it?

    Why (it seems) only the inner surface of the head is clad in stainless steel? Can't the entire head be made from a better steel?
    If it really is not practical, why at least the internal surfaces of all penetrations aren't clad in stainless steel, making this scenario impossible?
     
  19. Mar 17, 2013 #18
    There are ASME code requirements per 10CFR50.55 do to inspection of a percentage of vessel components on a regular basis. These tests look for general trends, since flaws tend to develop slowly and across the entire vessel. Additionally anytime the RCS is taken apart (reactor coolant system), prior to restart, pressure tests have to be done to look for leaks. This leak probably was not a through-wall leak yet during the previous cycle.


    With regards to this penetration and weld issue, there was a lot of evidence that there were problems, for over 12 years, and it was mostly ignored. The biggest piece was when they had to change out their containment air filters every few days rather than every few months (as this was when the leak reached its worst point, and the discoloring and clogging of the filters was metal from the vessel accumulating).

    With regards to other people's questions about LOCA (loss of coolant accident), this SB LOCA would have been well enveloped by the intermediate break or interfacing systems break LOCA, which are far worse accidents. In the event this leak actually punched through, the leak likely would not have been so bad as to cause a draindown of reactor coolant. The normal CV charging pumps (chemical volume control pumps) would likely have been able to maintain reactor coolant inventory. The operators would have recognized a high unidentified leak rate in the containment though, based on the containment sump accumulating water combined with increased charging flow without an increased letdown flow. Station technical specifications (not sure what the PWR tech spec is, in BWRs with standard technical specifications its in section 3.6) would direct them to have the plant cooled down within a 36 hour timeframe, and the leak would be found by inspection.

    If for some unknown reason, the leak was large enough to actually cause an increase in drywell pressure or a reduction in reactor coolant system inventory, a LOCA trip would have shut the reactor down and the ECCS system would have injected to ensure core inventory was maintained. If the leak was large enough to cause this situation, the reactor would depressurize through the leak slowly, and the reactor would be brought to a hot shutdown condition (mode 4), then decay heat removal would bring it to cold shutdown (mode 5), with no core/fuel damage. The big issue in this case, is the reactor would have been blown down faster than 100 degrees F per hour, which is a safety limit on reactor vessels, and extensive analysis would have to be performed prior to restart to ensure the vessel did not embrittle beyond safety margins. There's always a small possibility that the vessel may never be used again.

    In all cases, the plant would have to be shut down and the management and operational causes for the issue corrected prior to restart, as we have already seen at Davis Besse. Leaks in the ASME Class 1 reactor coolant system pressure boundary are not allowed.

    What really got Davis Besse in trouble about this, was that they were told by the NRC to shut down to do reactor head inspections, based on indications other plants have seen of cracking. They said their head inspection program had no indications and that they sampled (most/all) of the head penetrations recently and none had indications. They asked for permission to wait a months or two until their next shutdown, and the NRC agreed based on their statements. This was a false statement, and constituted deliberate and willful misconduct, since the NRC made a safety decision based on apparent false information. Several people were indicted, and at least one is banned from US nuclear activities (and is currently having a lot of trouble getting an engineering job ANYWHERE).

    Ultimately in all cases, the impact to the public would have been negligible under design basis conditions.

    I suggest those who are interested go online and read some more. Everyone in the US nuclear industry gets trained on this every 2-3 years (along with TMI, Chernobyl, and soon Fukushima). There is a LOT of info available publicly.
     
    Last edited: Mar 17, 2013
  20. Mar 17, 2013 #19

    Astronuc

    User Avatar
    Staff Emeritus
    Science Advisor

    It is not easy to inspect the top of the head because of the array of CRDMs and housing, which gives limited access. The reactor is on 24 month cycles, so it was in operation when a notification came out in 2001.

    Davis Besse went into a refueling outage Feb 16, 2002. The cavity on the RPV was found in early March, 2002.

    Ref: http://www.nrc.gov/reading-rm/doc-collections/nuregs/brochures/br0353/br0353r1.pdf

    The unit actually went into an outage about 1 month early.

    Pressure vessels are made of carbon steel, grades SA 533 B or 508 B. The inner surface was clad in an austenitic stainless steel, e.g, SS304L. The penetrations for the control rods are typically Inconel (typ Inconel 600). The problem was with the welds, some of which (made with alloy 82/182) were susceptible to intergranular stress corrosion cracking (IGSCC). There has been an ongoing program to monitor and evaluated IGSCC and the materials in the primary systems. There are separate vessel integrity programs for PWRs and BWRs. There has also been a vessel head replacement program.

    https://secure.inl.gov/VHTRRDTR12/pres/HighTemp/05_Rabin_A508533Pressure.pdf [Broken]
    http://www.nrc.gov/reading-rm/doc-collections/nuregs/contract/cr6875/cr-6875-discussion-end.pdf

    http://www.nrc.gov/reading-rm/doc-collections/gen-comm/bulletins/2001/bl01001.html
    http://pbadupws.nrc.gov/docs/ML0227/ML022700373.pdf

    http://www.nrc.gov/reading-rm/doc-collections/gen-comm/bulletins/2002/bl02001.html
    http://www.nrc.gov/reactors/operati...d-degradation/plant-specific-information.html
     
    Last edited by a moderator: May 6, 2017
  21. Mar 18, 2013 #20
    Stainless is not "better" in every respect. It is "better" in terms of resistance to corrosion, but the carbon steel used to form the head and the reactor vessel is "better" in terms of toughness (resistance to fracture). The materials chosen for an application are always a compromise between the alternatives.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Potentially serious situation averted at Davis-Besse NPP
  1. NPP fire codes (Replies: 14)

  2. NPP Works on 75% (Replies: 7)

  3. South-Ukraine NPP (Replies: 7)

Loading...