ML20212B169
| ML20212B169 | |
| Person / Time | |
|---|---|
| Site: | Farley  |
| Issue date: | 10/20/1997 |
| From: | Dennis Morey SOUTHERN NUCLEAR OPERATING CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20212B177 | List: |
| References | |
| GL-95-05, GL-95-5, NUDOCS 9710270275 | |
| Download: ML20212B169 (6) | |
Text
_ _ _ _ -
D:ve levey 8:uthern Nuclear a
Vite hepent Operating Company Iarley htyect P0.Bn1795 Omingham. Alatiama 35201
+
Tel 205 932 513)
SOUTHERN Octouer 20, 1997 COMM Energy to Serve krWid" Docket No.: 50 348 i
50 364 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk l
Washington, DC 20555 l
Joseph M. Farley Nuclear Flant Steam Generator AlternatclepJdLQiteria(ARC) Analygs Ladies and Gentlemen:
By letter dated August 29,1997, Southern Nuclear submitted a 90 day report for implementation of the ODSCC ARC at tube support plates as required by Generic Letter 95 05. As noted in the cover letter, it was expected that the results of a Farley Unit I pulled tube could have significant impact on the probability of burst or steam line break leakage calculations. On September 11,1997, Southern Nuclear informed the NRC Staff of the results of the calculations for probability of burst and steam line break leakage in a conference call.
Southern Nuclear subsequently requested a reduction in the steady state dose equivalent iodine limit to.15 pcurie/ gram and a corresponding reduction in the transient dose equivalent iodine limit for both Farley Unit I and Farley Unit 2.
Although the reduction in dose equivalent iodine limits is adequate to justify the current operating cycle, this approach has significant implications for the next Unit I steam generator inspection. Based on conservative estimates of the steam generator inspection results, hundredt ofindications could require repair below the current 2 volt repair limit with 19tle, if any, increase in safety. Furthermore, current schedules will result in steam generator replacement on Unit 1 in Spring 2000. Unit 2 steam generator replacement is scheduled for Spring 2001. Consequently, Southern Nuclear requests that the NRC Staff allow Farley Unit 1 l/
and Unit 2 steam line break leakage projections to be calculated based on a database composed of all domestic pulled tubes and model boiler specimens available as proposed by the industry [/jjf in an NEl letter dated September 18,1996. Should the Staff find this approach unacceptable, Southern Nuclear proposes an alternate approach of allowing steam line break leakage projection calculations using the full database and allowing the use of the available steam line break leakage versus bobbin voltage correlation.
9710270275'971020 "
PDR ADOCK 05000348 ILillllylli,OlljlJiLII
U S. Nuclear Regulatory Commission Page 2 Southern Nuclear requests approval of one of the proposed approaches by February 28,1998 in the event it is needed for the Farley Unit 2 steam generator inspection scheduled for spring
'98. provides a brief description of the problem and providesjustification for the Southern Nuclear proposal. Enclosure 2 provides a discussion of the most recent Farley Unit I pulled tube examination results and the implications ofinclasion of the data point in the pulled tube databases. Please note that Southern Nuclear is not requesting NRC appro'.nl of the specific calculations in Enclosure 2. They are provided as supporting evidence that the calculations as required by Generic Letter 95 05 are excessivdy conservative.
If there are any questions, please advise.
Respectfully submitted, SOUTilERN NUCLEAR OPERATING COh1PANY f
lW]U ~
i Dave hiorey REhi/maf:FNPIPULL. DOC
Enclosures:
- 1. lirief Description of Leak Rate Calculation issue 2, Farley 1 SG ARC Analyses in Support of Full Cycle Operation cc:
hir. L. A. Reyes, Region 11 Administrator hir. J.1. Zimmerman, NRR Project hianager e
hir. T. hl. Ross, Plant Sr. Resident inspector hir. T. A. Reed, NRR hiaterials and Chemical Engineering Branch w
r lef Description of Leak Rate Calculation Issue
- Issuet Generic Letter 95 05, Section 2.b.3(2), Conditional Leakage Rate Under Main Steam Line lireak Conditions, states the following:
Use of the linear regression fit of the logarithm of the conditional leak rate to the logarithm of the bobbin voltage is subject to demonstrating that the linear regression fit is valid at the 5 percent level with a "p value" test. If this conditior is not satisfied, the linear regression fit should he assumed to have zero slope (l.c.
the linear regression fit should be assumed to be constant with voltage).
As a result of the inclusion of the last pulled tubes from Farley in the leakage database, the mean of the zero slope linear regression required to be used increased from 12.9 liters / hour to 18.1 liters / hour, an increase of 40%. The standard deviation increased from 20.6 liters / hour to 34.6 liters / hour, an increase of 68%. As a result, all Daws, regardless of bobbin voltage, are projected l
to leak with a mean of 18.1 liters / hour and a standard deviation of 34.6 liters / hour.
l This requirement could potentially require the repair of hundreds ofindications below two volts in I
order to meet conservative leakage limits at the next Farley steam generator inspection with little, if any, imptovement in safety.
l Discussion: The supporting bases for using a voltage dependent steam line leak rate correlation l
for 7/8 inch tubes is provided below:
l
- 1. Use of a zero slope linear regression as required by Generic Letter 95 50 results in unres sonable leakage trends. As indicated in Figure 1, the majority of steam line break leakags for Farley Unit 1 is attributed to numerous flaws with bobbin voltages less than l
2.8 vohs. Ilowever, the smallest voltage indication to leak at conservative steam line break differential test pressure is 2.8 volts. Over 35 additional pulled tubes with voltages less than 2.8 volts (ranging from 2.8 volts to 0.8 volts) have been leak tested at steam line break differential pressure and have failed to leak. The result of current analysis l
requirements is that the large voltage indications, which are expected to have large leak rates as demonstrated by testing, dominate the voltage independent, zero slope leak rate projection, Furthermore, these large leak rates are assigned to all indications, even those below 2.8 volts. The result is a larger projected leak rate with the calculation driven by indications less than 2.8 volts, as shown in Figure 1. For example, the use of a "zero slope" correlation, as required by Generic Letter 95 05, results in a projected steam line l
break leak rate of 20.4 gpm. This approach assigns the same leakage characteristics to all Daws, even those below 2.8 volts which have not leaked in tests of pulled tubes.
Calculations based on the industry recommended database using a voltage dtpendent leak rate calculation yield a projected steam line break leak rate of 3.8 gpm.
i
O Brief Description of Leak Rate Calculation Issue Page3
- 2. A voltage dependent steam line leak rate correlation has been approved for 3/4 inch tubes. While this does not justify the use ef the linear correlation for 7/8 inch tubes, it demonstrates that the Stafragrees that a voltage dependent correlation is valid for steam generator tubes.
- 3. Use of an nrhitrary 5% p value test requirement results in excessive consenatism with little, if any, increase in safety. The p value represents the probability of obtaining the observed results even if there is no correlation between steam line break leakage and bobbin voltage. Following analysis of the latest Farley Unit 2 pulled tube, the p value for the leakage database had been reduced to 5.6%. Ilowever, inclusion of the latest Farley Unit I pulleci tube resulted in a p value of 7.6%.110 wever, the bobbin voltage and steam line break leak rate for this pulled tube may have been impacted by pressure pulse cleaning based on the pulled tube metallurgical examination results. To require the use of a zero slope correlation based on a p-value of 7.6% instead of a voltage dependent correlation based on a p value of 5.0%, when coupled with other Generic Letter 95 05 conservatisms, results in unreasonable leakage trends as discussed above.
The two data points which are primarily responsible for failing to meet the 5% p value requirement are a 31 volt French pulled tube which leaked at 0.13 liters / hour at steam line break ditrerential pressures and a model boiler specimen of 50 vohs which leaked at 0.86 liters / hour at steam line break differential pressures. If these tubes had developed leak rates on the same order as the Farley Unit I pulled tube,164 liters / hour, the 5.0% p-test requirement would be satisfied and a leak rate correlation could have been used. If these tubes had not leaked at all during the leak test at steam line break diffe ential pressure, the 5.0% p test requirement would be satisfied since they would not be included in the leakage database. Only by including the data points at very small, insignificant leak rates relative to the bobbin voltages will the data points result in the p-value test requirement not being met.
Proposed Resolution: Southern Nuclear requests that the NRC Staff allow Farley Unit I and Unit 2 to make steam line break leakage projections based on a database conr=~d of all domestic pulled tubes and model boiler specimens available as proposed by me in: ustry il m bel letter dated September 18,1996. As an alternative, Southern Nuclear propoq iaking the calculations using the full database; however, allowing the use of the available steam line break leakage versus bobbin voltage correlation.
Figure i
Farley Unit Leak Rate for Steam Generator C Percentage Contribution from Voltage Bins i
14.0 %
EOC-15 Projection 12.0% -
9 i
- o-EOC-14 Actual
,I i
t 10.0% -
Il os
,i S
E It I,I a
j q 8.0% -
9
,I i
Q gt It Y t gi Ii s
l V
t 8i y 6.0%-
Leakage has not occurred in steam
'~'
i a
~
8 g
line break difTerential pressure tests ss d
i 1 I i
b ofover35 pulled tubes with 8 8 i _,'_l _
indications less than 2.8 volts.
,' _'s A
4 cy, _
g g
6 s' t gp i
l q
E I
I h
20%-
s s 4
p gg g9 8/
-~T i
r t
,\\
s u
9 1
1 s
i a
0.0*/.
, n'
/
cr s
t 8
2 ; 2 E 2 3 3 5 % 0 % % % %0 I % % 0 0 ; O Bobbin Volts S
_ _ _ - _ _ _, =
Farley l SG ARC Analyses in Support of Full Cycle Operation SG 9710 004 October 1997 l
l
,rcw y-m w
_y w.-
- -