(May 17, 2002) Documents of the Davis-Besse incident reveal a strong possibility that there are large numbers of undetected cracks in the world's pressurized water reactors, particularly in the US. This suspicion has been confirmed by the discovery of an additional crack at Davis-Besse itself.
(568.5402) NIRS/WISE Amsterdam - The cracks are nothing new: they were first discovered in Bugey-3, France in 1991 1. Soon afterwards, similar cracks were found in 10 other reactors in France, 5 in Germany and at least one in Switzerland. As time went by, cracks were discovered elsewhere, including Japan 2 and Spain 3. In 2001, similar cracks were found in the US, first in Oconee-3 4 and subsequently in a total of 13 reactors 5.
The cracks in question are located in reactor vessel head penetrations: the tubes where control rods and other equipment (such as thermocouples) enter the top of the reactor vessel. These are often made of an alloy - Alloy 600 - which is susceptible to "stress corrosion cracking": in other words, cracks form because the alloy is under stress and in contact with hot, corrosive primary coolant under high pressure.
The phenomenon is so widespread that Nuclear Regulatory Commission (NRC) has a special section on its web site devoted to Alloy-600 cracking 6. Cracks can occur in various parts of these reactor vessel head penetrations. They can occur in the tubes themselves ("nozzles"), the control rod drive mechanism casing, or at welded joints 7.
The cracks grow until they go right through the metal, at which point primary coolant starts to leak out of the crack. The primary coolant consists of a solution of boric acid in water, the boric acid being used to control the rate of nuclear fission in the reactor. When a leak occurs, the water turns into steam, leaving behind crystals of boric acid. Up until recently, these crystals were seen as a useful indicator that a leak has occurred rather than a major problem in itself.
However, this changed after a large hole was found in the lid of the reactor pressure vessel (RPV) at the Davis-Besse Nuclear Power Station near Toledo, Ohio 8. At Davis-Besse, so much reactor water had leaked from the cracks that piles of boric acid crystals, weighing as much as 900 pounds (400kg), had built up on the reactor vessel head itself. The borated primary coolant water released through cracks then flashed to steam before cooling and dripping down onto the 605 degree Fahrenheit (317 degree Celsius) vessel head. Upon contact with the hot vessel head the water content again flashed to steam leaving the growing accumulation of boric acid crystals. With a melting point of 340 degrees Fahrenheit (171 degrees Celsius), some of the crystals became molten and the concentrated acid aggressively ate a hole through more than 6 inches (15 cm) of carbon steel making up the vessel head, leaving only the corrosion-resistant stainless steel liner intact. This liner bulged under the high pressure in the reactor, which it was not designed to resist.
All PWR operators around the world must be concerned that there are no more incidents like the hole at Davis-Besse. Now, however, there is a new question: are reactors being checked for cracks thoroughly enough, or are many cracks going undetected?
4 out of 5 cracks undetected?
The evidence that the cracking problem may be far more widespread began with a curious statement buried in FirstEnergy's Root Cause Analysis Report for the Davis-Besse incident. The statement 9 indicated that more than 5 times as many cracks had been found in French reactors as in the rest of the world.
Curious as to why this may be, WISE Amsterdam forwarded the statement to French contacts, asking if they had any idea why the problem seemed to be much worse in France. They in turn contacted the French nuclear safety authority, who said that they thought this was because French reactors were inspected more thoroughly, using eddy current testing 10.
If the French authorities are correct, this means that for every crack detected in reactors outside France there could be another 4 which remain undetected because of deficiencies in the testing methods used.
These extra cracks remain undetected because they are very small. As such, they are generally less dangerous than the larger cracks already found. However, as the reactor vessel heads age, the small cracks grow into larger cracks, which can eventually cause leaks.
New crack discovered at Davis-Besse as new corrosion problem looms
On 7 May 2002, FirstEnergy engineers met with staff at NRC Headquarters in Rockville, Maryland to review the Davis-Besse Root Cause Analysis of the RPV Head Degradation. As reported by the Ohio-based Cleveland Plain Dealer, "NRC officials were startled yesterday when FirstEnergy revealed that for the first time that there was one more crack in one of the control rod sleeves than previously been reported, prompting them to question the integrity of the entire FirstEnergy report"11. In fact, FirstEnergy had just been informed by Framatome a few days earlier that Nozzle #2 had an additional vertical crack that was originally missed in the interpretation of an Ultrasonic Testing (UT) exam. FE said they would be issuing a corrective action document and going back to review the UT data to see if additional cracks were also missed. In addition, Electric Power Research Institute identified a significant uncertainty factor for sizing crack length in these components. FE officials acknowledged that industry does not have good models to test the accuracy of crack measurements for vessel head nozzle penetration cracking as has come with significantly more experience in the pressurized water reactor steam generator tube cracking. In both cases, crack growth rate remains largely an unknown.
As company officials and regulators scramble to analyze and bound the many uncertainities associated with the cracking phenomenon and the very aggressive boron corrosion, still more questions were raised at an 11 April 2002 meeting with NRC's Advisory Committee on Reactor Safeguards (ACRS). John Grobe, NRC director of Region III Division of Reactor Safety told the committee that liquid boric acid had been found pooled in the corrosion cavity of the reactor head 12.
According to Grobe, UT examinations have also revealed what is described as "debonding": an extremely narrow gap between the thin stainless steel inner liner and the thicker outer layer of carbon steel that make up the reactor vessel. The gap extends beyond the known damage around the #3 vessel head control rod drive mechanism penetration (identified as a 5" wide by 7" long by 6 3/4" deep cavity.) Grobe told the ACRS that the "debonding" extends into the vicinity of vessel head penetration #11. After conversations with NRC staff, NIRS suspects this "debonding" to be more extensive corrosion undercutting of the ferritic iron resulting from capillary action of the molten boric acid. FirstEnergy's Root Cause Analysis dated 15 April 2002 does not address this yet-to-be-explained deterioration of the vessel integrity.
In the meantime, FirstEnergy has enlarged its original 13" plug repair concept to a 17" diameter repair that will now cut into the vessel and remove the adjacent area to include Nozzle 11 along with Nozzle 3 while plugging Nozzle 2. Under the utility's unprecedented repair option, which is not covered by any code or standard under the American Society of Mechanical Engineers, a forged plate of Inconel 690 would then be welded in to plug the cut hole. The three affected control rod drive mechanisms would be relocated to spare penetrations in the vessel head.
Alternately, First Energy revealed on 10 May that the company had signed a letter of intent with the owners of the unfinished Midland nuclear power station in Michigan for purchase of their unused Babcock and Wilcox reactor pressure vessel head as a possible replacement option. To accomplish the replacement, a hole must be cut in both containment buildings to remove the replacement component before installing it at Davis-Besse. The replacement option still requires regulatory approval and a safety evaluation on possibly challenging modifications necessary to adapt the replacement component to the reactor. In December, 2001 FirstEnergy additionally ordered fabrication of a new reactor vessel head from a Japanese company which will take an estimated 2 years to complete.
By a letter dated 29 April, NRC convened a NRC special Oversight Panel under Inspection Manual Chapter (IMC) 0350, "Oversight of Operating Reactor Facilities in a Shutdown Condition with Performance Problems" for Davis-Besse to evaluate restart issues and schedules for the stricken reactor. On 24 April, NIRS along with the Union of Concerned Scientists and thirteen other regional environmental and consumer groups petitioned NRC to order a Verification by Independent Party (VIP) of FirstEnergy's Root Cause Analysis for apparent problem identification and resolution issues, possible damage to additional components in containment from aerosolized boric acid, competency of FirstEnergy actions in response to other NRC generic communications and possible deferment of other plant modifications and corrective action programs.
And it's not just cracks
At Davis-Besse, it was not just the cracks which caused the leakage: the flanges of the control rod drive mechanism also caused leakage, which collected on the vessel head and masked the nozzle leakage. Other U.S. plants use canopy seals instead of flange seals, but these too have been reported to leak 13. In France, the vessel head at Tricastin-4 was replaced in 1996, but two years later, boric acid from a leaking canopy seal caused superficial corrosion of the new vessel head 14. So even replacing the vessel head will not eliminate the problem entirely.
References:
- WISE News Communique 385.3770: "Alarm over faulty design in European PWRs".
- WISE News Communique 461.4583: "Japan: Radioactive water leak in Sendai-1"
- WISE News Communique 408.4043: "New cracks found at Zorita NPP"
- WISE News Communique 553.5309: "US: NRC ignores widespread safety flaw for decade".
- WISE News Communique 560: "In Brief".
- www.nrc.gov/reactors/operating/ops-experience/alloy600.html
- www.nrc.gov/reactors/operating/ops-experience/alloy600/overview.html
- WISE/NIRS Nuclear Monitor 565.5385: "Millimeters from disaster".
-
"As of February 2000, about 6.5% of all EdF nozzles inspected had been found to contain cracks and about 1.25% of inspected nozzles in other plants worldwide had been found to contain cracks greater than the minimum measurable depth of about 2mm (0.08 inches)".
Root Cause Analysis Report, FirstEnergy Nuclear Operating Company, 18 April 2002, 3.0 Data Analysis p.9. - Matthieu Schuler, e-mail to Bella Belbéoch,, 29 April 2002
- John Funk, "NRC wants reactor top replaced," The Cleveland Plain Dealer, 8 May 2002.
- Daniel Horner, "Davis-Besse Proposes RPV Head Repair Concept As New Promblems Emerge", Inside NRC, 22 April 2002.
- E-mail from Dave Lochbaum, Union of Concerned Scientists, 14 May 2002
- Rapport d'activité 1998, Autorité de Sûret Nucléaire.
Contact: WISE Amsterdam; Paul Gunter at NIRS