ML13151A075
| ML13151A075 | |
| Person / Time | |
|---|---|
| Site: | Beaver Valley |
| Issue date: | 05/16/2013 |
| From: | Robinson T Citizen Power |
| To: | Peter Bamford Office of Nuclear Reactor Regulation |
| References | |
| Download: ML13151A075 (5) | |
Text
6 CITIZEN POWER Public Policy Research Education and Advocacy May 16, 2013 Peter Bamford, Project Manager Office of Nuclear Reactor Regulation Mail Stop: 08B3 Washington, DC 20555-0001
Dear Mr. Bamford:
This purpose of this letter is to elaborate on concerns expressed by Citizen Power at the Beaver Valley Annual Assessment Meeting on April 11, 2013 regarding the results of the random and non-random examinations reports submitted by FirstEnergy Nuclear Operating Company ("FENOC"). Citizen Power believes that the methodology used by FENOC to determine whether a location has experienced a loss of material significant enough to be included into the corrective action program was flawed. Specifically, FENOC's use of nominal thickness as a proxy for the original thickness of the containment liner in order to determine the amount of wall loss is unsupported by the available evidence.
During the spring 2006 refueling outage of Unit 1, two areas of corrosion in the containment liner were identified during a steam generator replacement project. This was followed by the discovery of a 1 inch by 1 and 3/8 inch hole in the containment liner during the spring 2009 refueling outage of Unit 1. As a result of these discoveries, FENOC agreed to modify the License Renewal Commitments as a condition of the relicensing of Units 1 and 2. Specifically, License Renewal Commitment (LRC) 32 for Unit 1 and LRC 33 for Unit 2 were modified to include the inspection of 75 random locations by UT for each unit. In addition, LRCs 33-35 for Unit 1 and 34-36 for Unit 2 were added requiring 8 non-random UT examination locations for each unit, a summary of the random and non-random testing results to be documented in a letter to the NRC, and an evaluation if appropriate or applicable statistical methods exist to gain additional insight into potential liner degradation for each unit.
During the Unit 1 fall 2010 refueling outage (1R20), FENOC examined 38 random locations and 8 non-random locations. Additionally, 2 non-random locations were examined in May of 2010. According to their letter of February 14, 2011 (FENOC Letter L-1 1-00 1), there was no evidence of loss of material at these locations. During the Unit 2 spring 2011 refueling outage (2R15), FENOC examined 61 random locations and 8 non-random locations. According to their letter of June 27, 2011 (FENOC Letter L 181), there was no evidence of loss of material except at one non-random location (2RN-063), which had a lowest thickness reading of 0.31 inch. A subsequent autoscan of 2RN-063 indicated that an area of 2 inches by 10 inches had lower thickness than the surrounding area with a localized point with a thickness of 0.267 inch. FENOC then indicated that the minimum design thickness for the general wall at this location is 0.213 inches and that the minimum required thickness of the localized point is 0.032 inches.
During the Unit 1 spring 2012 refueling outage (1R21), FENOC examined 37 random locations. As indicated by their letter of July 11, 2012 (FENOC Letter L-12-204),
FENOC found no evidence of degradation at any location. Therefore, FENOC concluded that "license renewal commitments 32, 33, and 34 [for Unit 1] are considered closed". On December 18, 2012 (FENOC Letter L-12-437), FENOC changed measurements from two random locations at Unit 2 (2RN-039 and 2RN-074) and concluded that these changes did not modify their previous conclusion that there was no evidence of loss of material at the examination locations. On February 5, 2013, the NRC requested additional information regarding acceptance criteria and how "breadth and shape" was related to degradation in order to complete the review of the containment liner random and non-random reports for Unit 1. By their letter of March 4, 2013 (L-13-060), FENOC responded by listing the acceptance criteria and identifying the nine locations where the lowest thickness was lower than the nominal thickness. In eight of these locations there was no evidence of wall loss because there was no breadth and shape. In the ninth location (1RN-073), subsequent evaluations indicated that the low spots were "either partially fused nelson studs located on the concrete side of the liner or laminar reflections inherent in the rolled plate material used to fabricate the liner." FENOC has scheduled a reexamination of this location during 1R22 to assess whether these conclusions are correct. On April 30, 2013, the NRC concluded that FENOC completed licensing commitments 32 through 34 for Unit 1.
Citizen Power does not believe that the methodology used by FENOC to determine whether locations have experienced loss of material is correct and disagrees with the conclusion that license renewal commitments 32-34 for Unit 1 should be considered closed. According to FENOC Letter L-09-243, "FENOC characterizes a statistical failure at the concrete to liner interface as a random location with greater than 10% loss of material, displaying measurable breadth and shape; unless through engineering evaluation it is determined that indications are attributed to fabrication/erection practices or locations not indicative of corrosion." We agree that this is the correct way to determine the prevalence of corrosion not attributable to fabrication or construction. The relevant measurable is the "loss of material", which is the original thickness of the containment liner minus the current thickness of the containment liner.
However, FENOC compares the current thickness to the nominal thickness instead of the original thickness. The nominal thickness does not represent what the original thickness was, but can be more accurately defined as the minimum specification that the containment liner was designed to meet. Under FENOC's methodology, if a location's current lowest thickness is less than 90% of the nominal thickness then it is categorized as a statistical failure unless through further evaluation it is determined that the low thickness is attributable to fabrication/erection practices or the location does not indicate corrosion is present. This methodology would be correct if the nominal thickness was approximately the same as the original thickness. However, where the nominal thickness is significantly different than the original thickness, the amount of corrosion cannot be accurately determined by using the nominal thickness as a proxy. For example, if the original thickness of a location was.42 inch, the nominal thickness is listed as.375
inch, and the lowest point is now.34 inch; then the location would not be evaluated to determine if it is a statistical failure under FENOC's methodology even though there would have been a 20% loss of material at that location.
Although the original thickness cannot be accurately determined, the use of the nominal thickness as a proxy clearly underestimates the amount of material loss. There is ample evidence that the original thickness was much greater than the specified nominal thickness. Of the 75 random Unit 1 locations selected, 74 of them had a "general" thickness that was greater than the nominal thickness. In addition, only 9 of the 75 locations had a "lowest" thickness lower than the nominal thickness. If this sample is representative of the containment liner, it is clear that the current thickness in most locations is greater than the nominal thickness.
As a more accurate alternative, Citizen Power recommends that the median thickness reading for each examination location be used as an estimate of the original thickness. This methodology more accurately represents the original thickness since variations in the original fabrication process would be reflected in the data while the nominal thickness is unable to account for such variations. The wall loss measurement would then be calculated by comparing the median thickness to the lowest point at each examination location in order to determine if the 10% threshold had been passed. Since the reported measurements included the mode for each location, it is reasonable to assume that the median would also be available.
In addition, the lowest thickness reported by FENOC represents the lowest thickness found, not the lowest thickness that exists. Based upon the available information, it is clear that there can be a large discrepancy between the measured lowest thickness as a result of the random examinations and the actual lowest thickness. For example, at location 2RN-063, the lowest thickness found was.31 inch during the manual UT examination. However, the subsequent autoscan examination of the location identified a lowest thickness of.267 inch, which is almost 14% lower. It is unclear what caused this discrepancy. However, if this discrepancy is a result of the inability of the initial UT to identify the actual lowest point of an examination location, it would follow that the results do not necessarily reflect the amount of wall loss. Citizen Power believes that the reason for this discrepancy should be explained.
Also, the change of lowest thickness indicated by 2RN-039 is unexplained. In the original report the lowest thickness measurement was.377 inch. However, the December 18, 2012 correspondence from FENOC indicated that this number had been changed to
.404 and the reason for the change was entered in the FENOC corrective action program.
Citizen Power respectfully requests that the reason for this change be explained.
Finally, referring back to the FENOC letter of June 27, 2011, Citizen Power would like to know if FENOC's contention that the minimum required thickness of the localized point at location 2RN-063 is 0.032 inches is correct. If it is correct what are the calculations involved to determine the minimum required thickness and are the calculations influenced by either the breadth/ shape of the area of lower thickness
surrounding the localized point or the general condition of the containment liner?
Hypothetically, would it be allowable to have a dozen localized points in the containment liner with a thickness of less than 0.035 inch?
In conclusion, Citizen Power does not believe that the examination results, as reported by FENOC, answer the original question of whether it is 95% probable that 95%
of the liner is not degraded in either Unit 1 or Unit 2. Citizen Power believes that each location should be reanalyzed comparing the median thickness to the lowest thickness.
The methodology used by FENOC seems to attempt to answer the question of whether the containment liner is currently at more than 90% of the nominal thickness. While this is also a relevant inquiry, it is not clear that FENOC's methodology answers that specific question either. Specifically, in order for the methodology used by FENOC to be accurate, the distribution of wall thickness needs to be a normal distribution. It is unknown whether the variability in wall thickness of the original fabrication of the containment liner is normally distributed. An analysis of the thickness of the examination locations would be appropriate to determine if there is a normal distribution. In addition, it is unclear whether the UT measurements are accurate.
Sincerely, Theodore S. Robinson, Esquire Staff Attorney Citizen Power 2121 Murray Avenue Pittsburgh, PA 15217 Cc: Erin Bonney
Ted Robinson C.itizen Power 2121 Murray Avenue Pittsburgh, PA 15217
$0. 6-0 US POSTAGE FIRST-CLASS 062S000782 84 1 1 15217 Peter Bamford, Project Manager Office of Nuclear Reactor Regulation Mail Stop: 08B3 Washington, DC 20555-0001 I-......-