ML070370619
| ML070370619 | |
| Person / Time | |
|---|---|
| Site: | Oyster Creek |
| Issue date: | 01/31/2007 |
| From: | Webster R Rutgers Environmental Law Clinic |
| To: | Advisory Committee on Reactor Safeguards |
| SECY RAS | |
| References | |
| 50-219-LR, ASLBP 06-844-01-LR, RAS 12990 | |
| Download: ML070370619 (6) | |
Text
RUTGERS ENVIRONMENTAL LA W CLINIC 123 Washington Street Rutgers, The State University of New Jersey Newark, NJ 07102-3094 School of Law - Newark Phone: (973) 353-5695 Fax: (973) 353-5537 January 31, 2007 DOCKETED VIA E-MAIL AND US-MAIL USNRC January 31, 2007 (1:47pm)
The Advisory Committee on Reactor Safeguards United States Nuclear Regulatory Commission OFFICE OF SECRETARY Washington, DC 20555-0001 RULEMAKINGS AND ADJUDICATIONS STAFF Re: Safety Evaluation Report for Oyster Creek Nuclear Power Plant Docket No. 50-219-LR
Dear Committee Members:
I am writing on behalf of STROC, the citizen's coalition including Nuclear Information and Resourc'e Service (NIRS), Jersey Shore Nuclear Watch, Inc., Grandmothers, Mothers and More for Energy Safety, New Jersey Sierra Club, New Jersey Environmental Federation (NJEF) and New Jersey Public Interest Research Group (NJPIRG). I am writing because since the last meeting of the Plant License Renewal Subcommittee on January 18, 2007, important new information has come to light that confirms the need for Exelon to significantly revise its aging management programs for the drywell shell at Oyster Creek Nuclear Power Plant ("Oyster Creek") if it is to provide a reasonable assurance of safety during any extended period of operation.
At the January meeting, it became clear. that the recent modeling by Sandia National Laboratory ("Sandia") invalidated Exelon's current approach to aging management for the drywell shell. This is because the modeling by General Electric ("GE") relied upon by Exelon could not model the first 9 modes of buckling and did not use the capacity reduction factor that Sandia considered that the ASME code required. However, Sandia also warned that its predictions were only relative and not absolute. The Sandia study of the degraded shell predicted a factor of safety during refueling of between 1.95 and 2.15 for the buckling analysis. This represented a reduction in safety factor of around 43% and shows that the shell is, at best, very close to the minimum code requirement. of a safety factor of 2. Because all of Exelon's proposed acceptance criteria for the extended licensing period are based on the GE modeling, the Sandia study shows that Exelon's approach.to acceptance is invalid.
At the January meeting, I pointed out that the Sandia model of the degraded shell is overly optimistic for at least two reasons. First, the model assumed only one locally thin area in two bays directly under the downcomier. This assumption is contrary to the data that show there are more than one locally thin areas of greater than one square foot in area in each of bays -1 and Carter H. Strickland, Jr., Esq.+ Julia L. Huff, Esq.*+ Kathleen J. Shrekgast, Esq.# Richard Webster, Esq.+
ActingDirector :. ý Staff Attorney Staff Attorney Staff Attorney cstricklandgkinoy.rutgsers.edu jhiiff@kinoy.rutgers.edu kshrekgast@kinoy.rutgers.edu nvebster@kinoy.rutgers.edu
- Admitted in New Jersey Pursuant t61:21-3(c) + Also admitted in New York # Also admitted in Pennsylvania TeP =*15ry... QAl9 j-zisc4' 0 .:'
R UTGERS ENVIRONMENTAL LA W CLINIC 13 placed away from the downcomers. In addition, the model assumed no further thinning of metal in the sandbed region since 1992, which is contrary to the thinning of the shell observed in the last outage.
Unfortunately, at the January meeting I was only able to present preliminary findings because we did not have complete data from the last outage. We now have the data and have also obtained a statistical analysis of that data carried out by Exelon, part of which is attached.
This analysis confirms that the October 2006 external thickness measurements show a thinning of the drywell by an average of around 0.02 inches compared to 1992. In addition, the analysis attempted to estimate whether there was any systematic error in the 1992 results compared to the 2006 results and concluded that there might be a bias of around 0.012 inches. Thus, Exelon's own analysis has confirmed that the shell in the sandbed region is now thinner than was measured in 1992 and that it is unlikely that all of this thinning can be explained by systematic measurement error. Thus, to establish the current margin, the Sandia model or something similar needs to be rerun using the latest results for the sandbed region. In addition, the uncertainties in the output of the model need to be explicitly evaluated. After the model is rerun, if the uncertainties are too high to provide reasonable assurance that the drywell liner meets code requirements, then the model would need to be refined and more data might need to be gathered.
Once there is a sufficiently detailed model to provide reasonable assurance that the drywell currently meets code requirements, Exelon would then need to show reasonable assurance that Oyster Creek would continue to meet the code requirements during any extended period of operation. One approach to this would be to use the Sandia model or something similar to determine how much general corrosion would be acceptable. The amount of permissible general corrosion could then be compared to potential corrosion rates to establish permissible intervals for the thickness measurements, At minimum, the Sandia Study and the October 2006 results show that Exelon has so far failed to provide a reasonable assurance of safety for any extended period of operations. In fact, it appears that we are now lacking reasonable assurance of current safety. We therefore urge this committee to recommend that the Safety Evaluation Report not be issued until rigorous studies are done that evaluate the current margin with reatsonable assurance and then establish a monitoring regime with valid acceptance criteria that ensures that any existing margin will be maintained throughout any extended period of operation.
Yours sincerely, Richard Webster c.c. Senator Robert Menendez Congressman Christopher H. Smith Congressman Jim Saxton
R UTGERS ENVIRONMENTAL LAW CLINIC Congressman Robert E. Andrews Congressman Rush Holt Congressman Frank Pallone, Jr.
Congressman Bill Pascrell, Jr.
Governor Jon Corzine Commissioner Lisa Jackson, New Jersey DEP Donnie Ashley, License Renewal Project Manger ASLB Service List
November 9, 2006 To: OC-12 Files From: George Licina
Subject:
Statistical Analysis of Oyster Creek Drywell Thickness Data
Background
In 1988, Oyster Creek experienced a problem with coriosion of the exterior of their drywell at the "sand cushion". The problem at that time was the sand cushion got wet and stayed wet, and the painted carbon steel drywell began to corrode. They removed all of the sand, did an enormous amount of calculation to prove they didn't need the sand cushion to disperse the loads from the drywell to the ground, sealed off the steel-concrete interface on the exterior of the drywell to make sure it stayed dry, jack-hammered several trenches in the concrete inside the drywell to permit them to do UT thickness measurements of the steel from the inside, etc. etc. Now that they are applying for license renewal, the issue of the condition of the drywell steel has been reopened. At the most recent the refueling outage (October 2006), they found that the concrete in the trenches was wet (one had 5" of standing water) so the question ofthe condition of the steel embedded in the concrete comes up once again.
The drywell (see Figure) is a huge (30' diameter or more where it intersects the concrete) but thin steel structure. The portion that is embedded in concrete (much of it has concrete on its interior as well) is basically a hemisphere. The drywell structure itself is shaped like a light bulb (upside down) with the reactor vessel, pumps, piping, etc. inside. The drywell is the secondary or tertiary containment structure for radionuclides (fuel cladding, then the reactor vessel, then the containment). Obviously, the containment and drywell get lots of regulatory scrutiny and attention from the public.
Discussions with Don Warfel, and later with Wayne Choromanski from Exelon indicated that a thorough and statistically based look at the data is required. For example, the UT thickness methods applied in 1986, 1992, and 2006 are all different; the prior examinations (1986 and 1992) were done on bare steel while the 2006 examination was done with a different technique and was done through the coating, plus the questions that always come up regarding whether the exact locations were examined at the different points in time. Further, the limited data from Zone 4 (above the 12'4" elevation; should never have been wet) appears to exhibit a thinning between the 1992 and 2006 inspections. A specific question asked by Don Warfel was whether a bias, based upon the apparent delta (t2006 - t 1992) in that zone, can reasonably be subtracted from all of the deltas to account for the technique differences.
I also reviewed a Tech Eval prepared by Oyster Creek and reviewed by Steve Leshnoff.
That Tech Eval includes data in various forms from 1986, 1992, and 2006. It focuses on OCLROO01 5503
present thickness with a lesser emphasis on the trends. Most of the evaluation is for data collected for Bays 5 and 17, where the trenches are. The tech Eval concludes that "the Drywell Vessel in the region below the concrete floor at elevation 10'3" may have been corroding at a rate of.002 to .003 inches per year between 1986 and 2006. UT readings below the concrete floor at Elevation 10'3" confirm that all locations meet the required thickness criteria."
I looked at the data as many ways as I could think of to sort out anything systematic (e.g.,
a bias) between measurements, differences among zones, among bays, and any oddities or obvious outliers. I also developed fits of the data to test for the most appropriate distribution to use and to determine coefficients that would enable quantitative analysis of the statistics.
All data are included in multi-page spreadsheet OC Data-1991-2006-GJL-Rl.xls. That spreadsheet processed data assembled and checked by Wayne Choromanski. The second version of the data submittal from Wayne was used. The second version was more complete and also corrected some errors in the reported elevation (Zone) that were included in the initial transmittal.
My original focus Was on the deltas. I looked at all deltas as a function of "original" (1992) plate thickness and by zone. If the UT technique had a bias, I would have suspected that different absolute values of thickness would show different effects. I also looked at the distribution of delta by zone and by bay. One thing that was clear was the mean delta varied by bay and by zone and that the distribution of deltas looked very much like a normal distribution centered at a small negative value (small metal loss).
I also looked at thickness, primarily at the thickness in 2006. The main thing that I found was that thickness was a strong function of the bay and much less a function of zone.
Finally, I tried to evaluate the statistical distribution. To do that, I ordered the deltas from smallest to largest, and applied a look-up table to assign a parameter PHI. PHI is related to where in a normal distribution the point lies, based on the point's rank. For example, the point that is in the exact middle of the distribution (F = 0.50000; see the CDF tab of the attached spreadsheet) is at the mean (i.e., PHI = 0; which means 0 standard deviations from the mean). The first (lowest value) point defines the extreme of the data we have and will be in the lower tail of the distribution (PHI will be a relatively large negative*
number). Similarly, the largest value will correspond to a relatively large positive PHI.
When the data are plotted as PHI vs. delta, the data generate a reasonably straight line.
The better the straight line, the better the fit to the normal distribution. The mean of the distribution is where PHI = 0 and the breadth of the distribution (i.e., how large the standard deviation is) can be determined by how horizontal the cuirve is. For example, if all of the values were at exactly the same value, that value would obviously be the mean and the standard deviation would be zero (no variation in the data). The CDF plot for the deltas produced a very nice straight line over much of the population, however, the larger negative deltas were the values that destroyed the fit. The best fit line had an R2 value (a 0CLR00015.504.
perfect fit has R2 = 1.000) was 0.83; not bad but not great I also drew in my eyeball fit to the well behave data.
When I took the same approach to check whether the 2006 plate thickness data were described by a normal distribution, I got a very nice straight line (R2 = 0.98). The 1992 thickness data also gave a similarly good ,traight line and that fit showed that the 1992 measurements were thicker at all values of PHI than those from 2006 (i.e., the drywell apparently lost thickness between 1992 and 2006 as would be expected). At the mean (PHI = 0), that difference was about 20 mils of thinning. At PHI = 3 (3 standard deviations from the mean, approximately the 99h percentile, the thickness difference was about -29 mils. At PHI = -3, approximately the 1t percentile, the difference was about 12 mils. That latter observation suggests that the measurements made in 2006 were systematically lower than the those in 1992 by about 12 mils. It can probably be argued that the actual thickness differences based upon subtracting the 2006 thickness from the 1992 thickness actually average 12 mils less than the values reported.
Note that this analysis doesn't say whether the 1992 measurements are better than the 2006 measurements or vice versa; only that the difference between the two has a bias in it.
OCLR000I15505