ML20206C946
| ML20206C946 | |
| Person / Time | |
|---|---|
| Site: | Farley  |
| Issue date: | 04/30/1999 |
| From: | Cullen W, Malinowski D, Pitterle T WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML20206C944 | List: |
| References | |
| SG-99-04-001, SG-99-4-1, NUDOCS 9905030258 | |
| Download: ML20206C946 (111) | |
Text
__
SG-99-04-001 Farley-1: Final Cycle 16 Freespan ODSCC Operational Assessment April 20,1999 Prepared By:
ra.F##
6 Su-reMj T. A. Pitterle V. Srinivas R. F. Keating Consulting Engineer Principal Engineer Consulting Engineer Signature on file Signature on file D. D. Malinowski J. Begley T. Begley Consultant E-Mech Technology E-Mech Technology Reviewed By:
W. K. Cullen Senior Engineer Westinghouse Electric Company LLC Nuclear Service Division 9905030258 990423 PDR ADOCK 05000348 p
PDR Qtubeintda98T1 Final Freespan OA-1. doc-04/21/99
SG-99-04-001 Farley-1: Final Cycle 16 Freespan ODSCC Operational Assessment April 20,1999 Prepared By:
f(f gg Su-"
w'@h T. A. Pitterle V. Srinivas R. F. Keating
/
Consulting Engineer Principal Engineer Consulting Engineer "h
k a
Signature on file Signature on file D. D. Malinowski
.i. Begley T. Begley Consultant E-Mech Technology E-Mech Technology Reviewed By:
W. K. Cullen Senior Engineer Westinghouse Electric Company LLC Nuclear Service Division Q tubeint\\ala987IFinal Freespan OA l. doc-04/21/99
Farley-1: Final Cycle 16 Freespan ODSCC Operational Assessment Table of Contents S2bject Pag,e
1.0 INTRODUCTION
1-1 2.0
SUMMARY
AND CONCLUSIONS 2-1 2.1.
Overall Conclusions 2-1 2.2.
Summary 2-1 3.0 EOC-15 INSPECTION RESULTS FOR FREESPAN ODSCC 3-1
.3.1.
Unplanned Outage (8/98) Inspection Results 3-1 3.2.
1R15 Inspection Results 3-2 4.0 IN SITU AND PULLED TUBE DESTRUCTIVE EXAM RESULTS 4-1 4.1. In Situ and Pulled Tube Destructive Exam Results for R25C51, SG B 4-1 4.2. In Situ Test Results for 1R15 4-3 4.3. Destructive Examination of SG C Tube R6C10 4-5 5.0 EVALUATION OF FIELD AND PULLED TUBE RESULTS 5-1 5.1. POD Development 5-1 5 2. NDE Uncertainty Development 5-6 5.3. Correlation of Bobbin Volts with Maximum and Average Depth 5-7 5.4. New Indication Average Depth and Maximum to Average Depth Ratio Distributions 5-8 5.5. - EOC-15 Crack Profiles for Probabilistic Analyses 5-8 5.6. EOC-15 Crack Depth Distributions for Deterministic Analyses 5 5.7. Crack Length Distributions for New Indications in Probabilistic Analyses 5-9 6.0 GROWTH RATE EVALUATION 6-1 6.1. Growth Rates for Large Indications 6-1 6.2. Method of Developing Growth Rate Distributions 6-2 6.3. Average Depth Growth Distribution 6-3 6.4. Maximum Depth Growth Distribution 6-4 7.0 ESTIMATION OF NUMBER OF FREESPAN ODSCC INDICATIONS FOR CYCLE 16 7-1 7-1 General Description of Multi-Cycle Monte Carlo Simulation 7-1 7-2 AnalysisInput Data 7-2 7-3 Projection of EOC Freespan ODSCC Conditions 7-3 7-4 Structural Margins and Leakage Integrity Projections 7-3 8.0 DETERMINISTIC OPERATING CYCLE LENGTH ASSESSMENT 8-1 8.1. Burst Margin Requirements 8-1 8.2. Structural Acceptance Limit 8-3 8.3.
Sensitivity Analyses for Cycle Lengths as a Function of Confidence Levels on POD and Growth 8-4 8.4. Acceptable Operating Cycle Lengths Based on Largest Projected Indications
- Q:tunciariala98\\FIFinalFreespanoA l. doc 44/21/99 -
Utilizing Monte Carlo Analyses 8-5 8.5.
Sensitivity Analyses for Operating Cycle Length 8-7 8.6.
Estimated Full Cycle Average Depths and SLB Burst Probabilities 8-7 8.7.
Conclusions for Deterministic /Probabilistic Cycle Length Evaluation 8-8 9.0 PROBABILISTIC ASSESSMENTS FOR ACCEPTABLE OPERATING CYCLE LENGTHS 9-1 9.1.
Probabilistic Analysis Methods 9-1 9.2.
Input Distributions for Probabilistic Analyses 9-4 9.3.
Analysis Results for Partial Cycle Lengths 9-5 9.4.
Analysis Results for Full Cycle 16 Operation 9-5 9.5.
Leak Before Break Considerations for Full Cycle Operation 9-6 9.6.
Conclusions 9-7
10.0 REFERENCES
10 1 o tubeinnain9swirin Frmpen oA.I. doc 44/21/99
1.0 INTRODUCTION
This report provides the Farley-1 steam generator (SG) final operational assessment for freespan ODSCC following the 1R15 inspection. The condition monitoring and operational assessment was finalized for all other degradation mechanisms in Reference 10.1 and no indications other than freespan ODSCC were found to challenge structural and leakage integrity. The Reference 10.1 report supported i
full cycle operation for all degradation mechanisms except freespan ODSCC and identified the need for a final operational assessment for freespan ODSCC. The objective of this report is to define an.
acceptable operating cycle length until the next inspection. Two cycle length evaluations are performed including a deterministic approach based on maintaining 3APso (three times normal operating primary to s:condary pressure differential) and a probabilistic approach based on probability of tube burst at APsta (steam line break primary to secondary pressure differential).
During Cycle 15, freespan ODSCC had resulted in operational leat age leading to an unplanned outage in August 1998. Following the unplanned outage, leak rates of < 22 ppd were found until the EOC-15 inspection in November,1998. The planned full cycle length for Cycle 16 of 1.12 EFPY is shorter than the 1.18 EFPY to the August 1998 shutdown and corresponds to the point in Cycle 15 operation where primary to secondary leakage of about 29 gpd was measured. At the unplanned and EOC-15 outages, indications were found that did not satisfy a structural margin guideline of 3APso although all indications had burst pressures exceeding that associated with postulated accidents for which the steam line break (SLB) is the limiting event. Following stanup for Cycle 16 operation, leakage has been < 2 gpd and has not changed in the three months following startup until preparation of this report.
_ Operating cycle lengths are evaluated in this report against a deterministic guideline of 3APuo burst f
margin for the largest projected indication and against probabilistic guidelines on tube burst probability and leak rate at SLB conditions. Three Farley-1 pulled tubes destructively examined for freespan ODSCC provide support for the development of probability of detection (POD) as a fetion of depth and NDE uncertainties for sizing the Farley SG freespan indications with the + Point coil. Growth rates in average and maximum depth are developed from the IR14 and IRIS inspection results supplemented by pulled tube depth data where available. The reference methodology used for the probabilistic analyses is based on projecting crack depth profiles including allowances for POD, NDE uncertainties and growth to the end of cycle conditions. This methodology provides "first principle" analysis techniques in that burst and leakage are calculated from the crack profile rather than the more conventional simplifying assumptions of average depth orjudgmental crack shapes. The application of this more detailed analysis methodology is permitted by improvements in + Point coil depth sizing for which the associated NDE uncertainties are developed on a plant specific basis in this report. The methodology applied for evaluation against the deterministic structural limit is based on defining a bounding crack length and projecting average depths to EOC conditions for comparisons with a length / depth based structural limit that satisfies 3APso burst margins at 95% probability and 95%
confidence. A probabilistic multi-cycle analysis is used to estimate the number of freespan indications expected at EOC-16 and to provide an independent analysis method for assessing SLB burst probabilities and leak rates. The multi-cycle and average depth based probabilistic analyses provide independent methods for comparisons with the reference probabilistic analysis results.
Inspection results from the unplanned 8/98 unplanned outage and the IRIS refueling outage are
' described in Section 3. Section 4 describes in situ test results obtained during the two outages and also summarizes the three pulled tube destructive examination results. The evaluations of field and pulled tube results to develop the input data to the tube integrity assessments are described in Section 5. This section includes development of the depth dependent POD, NDE uncertainties and crack depth / profile Q:tubein6ala98\\FIFinal Freespan oA.1. doc-04/21/99 I-1
f distributions. The growth rate evaluation from Cycle 15 data is described in Section 6. The multi-cycle analyses to develop the number of expected EOC-16 freespan indications and to obtain independent probabilistic analyses are given in Section 7. The evaluation for the operating cycle length that satisfies the deterministic 3APuo burst margin is discussed in Section 8. The probabilistic assessments to detennine an acceptable operating cycle length are given in Section 9. References are included in Section 10. Section 2 provides the summary and conclusions from this report.
I l
l l
Q:tubeint\\ala98\\FIFinal Freespan oA-1. doc-04/21/99 1-2
~ 2.0
SUMMARY
AND CONCLUSIONS 2.1. OverallConclusions Overall, it is concluded that the probabilistic analyses support full Cycle 16 operation of 1.12 EFPY.
The SLB burst probability of 1.1x10-2 for SG B is less than the allowable limit of 2.5x10-2 per NEI 97-06 requirements. The results of two alternate probabilistic analysis methods also support full cycle operation.
. The probabilistic SLB leak rate of <0.01 gpm at 95%/95% confidence found for SG B after full cycle operation is much less than the freespan ODSCC allocated limit of 3.6 gpm based on not exceeding a small fraction (10%) of 10 CFR 100 limits (total leakage of 11.8 gpm for all degradation mechanisms).
The SLB leak rate at 95%/95% confidence conservatively satisfies the requirement of 1.0 gpm for a nominal leak rate to support severe accident considerations and is within the FNP current licensing basis.
The operating time for Cycle 16 that permits satisfaction of the detemtinistic structural limit guideline of 3APuo = 4308 psi was determinal to be 0.73 EFPY. This operating period is based on the largest projected indication satisfying the deterministic structural limit. For operation beyond 0.73 EFPY, the SLB burst probability results meet NEI 97-06 requirements. It can be noted that the largest indication after full cycle operation may have a burst pressure at 95%/95% confidence less than the determmistic structural limit.
Operating leak rates during Cycle 16 operation are to be administratively lowered below the Technical Specification limits to enhance the likelihood ofleak before break in the low likelihood event that larger than expected indications occur during plant operation. The following operating leak rate limits are administratively implemented for Farley-1 during Cycle 16 operation:
Plant to be shutdown within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a leak rate spike confirmed by chemistry samples to o
exceed 75 gpd.
Plant to be shutdown if the leak rate following a spike above 60 gpd does not decrease to less than o
60 gpd following the leak rate spike. Successive leakage spikes above 60 gpd, but less than 75 gpd, within a seven day span do not require shutdown since the 60 gpd limit is intended to be a longer term leak rate without spikes.
Plant to be shutdown within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a progressive (with spikes) increase in the leak rate o
to greater than 60 gpd confirmed by chemistry samples.
2.2. Summary Probability ofDetection (POD)
The probability of detection for Farley SG freespan ODSCC was developed from Farley-1 pulled tube data. Farley pulled tube data from three tubes with 21 indications are used to develop the POD. One tube was pulled in 1997 and two in 1998 including tube R25C51 which had throughwall indications leading to the leaker outage in August,1998. The original field bobbin calls and a field analyst reevaluation at the end of the IRIS outage are used for detection calls at the depths detennined from the
- Q:tubeinnais9:WIFinal Freespen oA I Acc-04/21/99 2-1
e pulled tube destructive examinations. PODS are developed as a function of average depth and of maximum depth. The burst effective average depth (part of total crack length that results in the lowest burst pressure for the flaw) from the destructive exam profile is used for the POD and tends to be about 5 to 10% deeper than the average depth for the total flaw length.
It can be noted that the bobbin vcitages for the pulled tube indications used to develop the POD are less than 1.0 volt indications with the exception of the two throughwall indications in R25C51. Ten of the indications are < 0.25 bobbin volts and 17 of the indications are < 0.75 bobbin volt. Thus, the indications would be expected to be in the relatively poor signal to noise ratio region for detection.
However,it should be recognized that the bobbin voltage at a given depth for freespan ODSCC is typically lower than found for ODSCC at TSP intersections since the freespan indications are dominated by a single flaw whereas the TSP indications typically hpve more than one flaw around the circumference. In addition, the pulled tubes show a wide range of uncorroded ligaments in the burst flaws used for the POD. The indications in R6C28 had 5 or less ligaments while the indications in R6C10 had 23 to 52 ligaments and the indications in R25C51 varied from 3 to 19. ODSCC indications at TSP intersections typically have about 5 or fewer uncorroded ligaments. Detectability and bobbin voltage tend to be reduced with increasing numbers of uncorroded ligaments. Thus, the pulled tube indications used for the POD provide a wide range ofligament affects and tend toward the more difficult range for detection. Overall, it is concluded that the indications applied for the POD development tend toward the more difficult range of detection relative to voltage amplitudes and ligament effects and the resulting POD would tend to be conservative.
Three enhancements for bobbin detection were implemented during the 8/98 leaker outage and the IR15 inspection. The three changes implemented were: 1) Increased analyst training to emphasize the importance ofidentifying and calling small freespan indications; 2) The tube regions near sleeve ends, found to be blind for bobbin detection, were inspected with + Point probes including 100% of these areas in the IRIS inspection; and 3) Based on a review of prior cycle data at the IRIS inspection following identification of the leaker in SG B R43C32, bobbin analysis techniques were revised for calling indications near the edges of the TSPs where the signals were relatively large but did not follow the expected behavior for a flaw. From a review of the largest 8 indications found during the 8/98 leaker outage and the IRI5 inspection,5 of the 8 largest indications resulted from relatively large (0.98 to 2.57 volt) indications near the edges of the TSPs that were not called at the 1R14 inspection and would be called based upon the revised calling techniques. This change in the flaw detection technique resulted in a total of 41 indications being called near TSP edges in the IR15 inspection compared to only 5 flaws in the IR14 inspection. One of the 8 largest indications was due to the blind area for bobbin detection near a sleeve end which would have been detected with the implemented + Point inspection. Only two of the S largest indications (0.44 and 0.97 BOC volts) are dependent upon item 1) above for enhanced analyst training. Thus, it can be expected that the changes implemented to enhance bobbin detection would significantly improve the POD for the IRIS inspection compared to that at the 1R14 inspection.
Databasefor Tube Integrity Analyses The database for input to the tube integrity analyses is developed from inspection and pulled tube destructive exam results. Crack length versus depth profiles are obtained from + Point analyses. Farley pulled tube data from three tubes, supplemented by additional data for sludge pile indications, are used to develop NDE sizing uncertainties. Growth rates utilize + Point sizing and pulled tube results for EOC-15 indications. Since + Point inspection data are not available for indications at the prior inspection (all confirmed indications were repaired), a correlation between bobbin voltage and depth was developed from the pulled tube results to provide depth estimates at EOC-14 for the growth rate evaluation.
Qtubeinfela9871 Final Froespan oA-l. doc-04/21/99 2-2 L
A probabilistic multi-cycle analysis was used to develop the expected number of freespan indications at EOC-16 for input to the reference deterministic and probabilistic analyses. This methodology adjusts the formation of new indications such that in combination with POD and growth considerations, the number ofindications and typical depths for the prior three inspections are reproduced from the analyses. This technique then projects the number of Cycle 16 indications using the methods and data that simulated the prior three inspections. The multi-cycle analyses were also applied to estimate the SLB tube burst probability and leak rates at EOC-16 as an independent methodology for comparison with the reference analysis methods.
DeterministicAnalyses Analyses were performed to determine the cycle length at which the projected largest indication would satisfy the deterministic structural limit guideline of a 3APso burst margin. These analyses are based on average depth distributions developed from + Point depth profiles as the partial or effective length that results in the lowest burst pressure for each indication. This method assumes that all average depths have a bounding crack length representing the largest effective crack length (part of total crack length that results in the lowest burst pressure) obtained from the EOC-15 + Point data and the Farley-1 pulled tube destructive exam depth profiles. The Farley-1 freespan POD as a function of depth is applied to the EOC-15 etTective depth indications to define the BOC-16 distribution similar to the POD application methods of NRC GL 95-05. NDE uncertainties and growth are applied to the BOC-16 distribution using Monte Carlo methods to obtain the projected EOC distribution. The large depth tail of the projected Monte Carlo distribution is integrated to one whole indication to define the largest EOC-16 indication.
The cycle length was then adjusted such that the large::t average depth satisfies the 3APuo burst margin structural limit obtained at 95% probability and 95% confidence on burst correlation and material property uncertainties using the Westinghouse partial depth burst correlation. Conclusions from the deterministic /probabilistic cycle length evaluations of this report are:
o A cycle length of at least 0.73 EFPY is acceptable for Cycle 16. At the rad of the 0.73 EFPY operating period, the largest projected average depth satisfies the 3APwo = 4308 psi burst margin requirement of an average depth of 67.0% for a bounding burst crack le tgth of 0.80 inch. The 0.80 ir!ch length bounds the largest burst effective lengths found for structureily challenging indications at the EOC-15 inspection.
The structural limit of 67.0% depth for a crack length of 0.80 inches is based on a burst correlation e
developed by correlating measured burst pressures with average depths from crack length profiles obtained from available pulled tube and laboratory specimen PWSCC and ODSCC axial crack data.
The correlation fits analytical formulations of burst pressures to the measured data. The resulting correlation then permits calculation of burst pressures from length / average depth or crack profiles with defined uncertainties for the correlation. The defined uncertainties then permit confidence -
statements to be made for the calculated burst pressures.
o Projections of average depths to full EOC-16 operation indicates that average depths would satisfy a 1.43APsts burst pressure margin at 95%/95% confidence and the burst probability would be less than the performance criteria requirements of NEI 97-06.
Q1ubeint\\als9871 Final Freespan oA l. doc-04/21/99 2-3
Probabilistic Analysis Requirements Based on satisfying NEI 97-06 requirements as modified for Farley-1 freespan ODSCC, the conditional probability limits for freespan ODSCC tube burst due to postulated accident conditions are given by:
< 2.4x10-2 per year that one or more tubes burst during an accident o
< l.2x10-2 per year that 2 or more tubes burst during an accident o
These requirements apply after reductions of the NEI 97-06 5x10-2 requirement for unknown or unanalyzed mechanisms and for the projected EOC-16 burst probability for ODSCC at TSP intersections.
Farley Nuclear plant currently has a leakage limit of 11.8 gpm approved for ARC at TSP intersections.
This limit is based on not exceeding a small fraction of 10 CFR 100 limits (10%) in the event of a main steam line break. Based on FNP's current projected SLB leakage from ODSCC at TSP intersections at EOC-16 of 8.2 gpm, margin exists of 3.6 gpm that could be applied to an allowable SLB leak rate limit for freespan ODSCC without changing current NRC approved offsite dose limits. For the probabilistic rpproach, the projected SLB leak rate at EOC-16 for freespan ODSCC must be less than this value when evaluated at 95% probability,95% confidence. An additional requirement is imposed on the freespan SLB leak rate for severe accident considerations. This requirement limits the nominal SLB leak rate from all freespan indications to less than or equal to 1.0 gpm.
Given that the Farley-1 PORVs satisfy the NRC Generic Letter 95-05 requirements for operability, the SLB pressure differential is based on the PORV setpoint of 2350 psia with an allowance of 3% for uncertainty. This leads to APst.s = 2405 psi.
SLB Tube Burst Probability It is concluded that the probabilistic analyses support full Cycle 16 operation of 1.12 EFPY. The SLB burst probability of 1.1x10-2 for SG B based on the reference analysis methods of this report is less than the allowable limit of 2.4x10-2 per NEI 97-06 requirements. The reference methodology used for the probabilistic analyses is based on projecting crack length versus depth profiles to end of cycle conditions including allowances for POD, NDE uncertainties and growth. This methodology provides "first principle" analysis techniques in that burst and leakage are calculated from the crack profile rather than the more conventional simplifying assumptions of average depth orjudgmental crack shapes. The application of this more detailed analysis methodology is permitted by improvements in + Point coil depth sizing for which the associated NDE uncertainties are developed on a Farley SG plant specific basis in this report The full cycle result of1.1x10 2 for SG B SLB burst probability based on using crack depth profiles is more conservative than the Section 8 result of 3.8x10'3 based on use of average depths and the Section 7 result of 4.6x10'8 based on multi-cycle, complete probabilistic analyses. Consequently, the results of three independent analysis methods support full cycle operation.
Comparisons were made of the in situ and destructive exam pressure test results with predictions from the NDE profiles and burst correlation applied in the analyses of this report. The results show that the methods lead to conservatively low predictions of burst pressures. These results indicate that the operating cycle length and burst probability results, which are based on + Point profiles and the burst correlation, are conservative predictions.
Q:tubeintiala98'ElFinal Freespan OA-1. doc-04/21/99 2-4 l
u_____________.____.__.____._.______.
SLB Leak Rates The SLB leak rate of <0.01 gpm at 95%/95% confidence for full cycle operation is much less than the dose based allowable limit of 3.6 gpm for freespan ODSCC. The SLB leak rate at 95%/95% confidence conservatively satisfies the requirement of 1.0 gpm for a nominal leak rate to support severe accident considerations.
Operating Leak Rate Limits The operating cycle length evaluations are based on 95% probability at 95% confidence that an acceptable structural limits will be maintained at the end of full cycle operation. It is desirable to have leak before break for the lower likelihood condition that the predicted structural margins are not maintained over the full cycle of operation. For leak before break considerations, acceptable structural integrity for full cycle operation of 1.12 EFPY in this assessment is based on preventing burst at SLB conditions. A shorter operating cycle length of about 0.73 EFPY is predicted to provide 95/95 confidence that a burst pressure exceeding 3APso is maintained. Thus, there is a reasonable likelihood that indications exceeding SLB conditions but less than a 3APso burst margin will result at EOC-16 conditions. Therefore, it is appropriate that leak before break is established to reduce the likelihood of a burst at SLB conditions rather than to maintain a burst margin of 3APso.
Based on providing some conservatism relative to an 85 gpd acceptable spike demonstrated to provide margin against burst at SLB conditions, the operating leak rate limits described in Section 2.1 above are recommended for Farley-1 during Cycle 16 operation.
I Q:tubeint\\ala98\\FIFinal Freespan oA-1. doc-04/21/99 2-5
3.0 EOC-15 INSPECTION RESULTS FOR FREESPAN ODSCC 3.1 Unplanned Outage (8/98) Inspection Results The initial visual inspection of the bottom of the tubesheet identified the tube at R25C51 in SG B as the leaking tube leading to the unplanned outage. By subsequent eddy current examination this tube was found to have two large freespan ODSCC axial indications with one located just above the top of a tubesheet sleeve installed in 1992; the second indication was about 8 inches above the top of the' sleeve.
Based on + Point sizing of the indications, both indications appear to have throughwall segments with total crack lengths of about 0.8" and 1.2" inch (without reductions for coil lead-in and lead-out effects) and the second indication penetrating throughwall may have contributed to the operational leak rate spikes found prior to plant shutdown.
It was determined during the unplanned outage that bobbin and Cecco inspections are essentially " blind" to potential indications near the top of the sleeve due to signal responses from the end of the sleeve
' masking potential flaw indications. The bobbin probe geometry, when used from the cold leg of the tube, leaves a gap in the detection region of the tube of approximately % inch from the end of the sleeve.
A corrective action was implemented to inspect all sleeved tubes in SG B with a + Point coil near the ends of the sleeve. In addition, all analysts were retrained to identify and call small freespan indications with the bobbin coil probe.
An inspection plan was developed following the EPRI NDE guidelines (Rev. 5). A critical area for potential large indications and a buffer zone to bound the critical area were defined. The critical area in SG B was defined to include potential missed indications based upon a review of the IR14 bobbin inspection results and the " blind" area near the free-span ends of sleeves installed prior to the IR14 outage which require a + Point inspection. A 100% inspection was performed for the critical area in SG B. The buffer zone was defined to be a 20% inspection including randomly selected tubes and tubes selected to be biased toward a potential for large indications. The biased tube selection included tubes with sleeves installed in IR14, bobbin indications found in IR14 but not confirmed by + Point inspection, tubes with the same heat as R25C51, sleeved tubes and the neighboring tubes to R25C51.
The inspection would have been expanded if freespan ODSCC growth rates larger than expected, based on prior operating history, were found by + Point inspection. This threshold for expansion was a 2.0 volt
+ Point indication. Since there was no operational leakage in SGs A and C and the operating period to complete Cycle 15 operation was less than eight weeks, it was not necessary to inspect these SGs unless required by the expansion criteria defined for SG B indications detected during the inspection.
No large indications were found in the SG B inspection and it was not necessary to funher expand the inspection.. A total of 27 + Point indications in 13 tubes were identified in the inspection and these tubes were plugged. The largest freespan ODSCC indication found was a 0.85 + Point volt indication. The largest sludge pile ODSCC indication found in the inspection was 1.25 + Point volts. No indications were found in sleeved tubes which tends to indicate that the large indications found in tube R25C51 are likely to be independent of the presence of the sleeve. The leaking tube, R25C51 in SG B, was inspected with both bobbin and + Point probes. A Cecco probe inspection of the sleeve and an in situ leak test of the leaking tube were also performed. The results of these inspections and tests are described below.
I Bobbin CoilInsoection ofR25C51 Q.tubeintiala\\ala98Pinal oAfinal rneespan oA-3,4. doc 04/21/99 3-1
The bobbin coil inspection of the tube between the top of the sleeve and the 1" TSP (tube support plate) identified a 12.7 volt indication at about 95% depth using a 0.720" probe. With a 0.640" probe, the I
bobbin voltage was found to be 15.1 volts with a 94% depth. The indication was located about 8" above the top of the sleeve. Small bobbin potential indications were also identified as 0.42 volt at 5.1",0.29 volt at 6.9",0.14 volt at 11.6", four potential indications of 0.19 to 0.46 volt between 15.9" and 17.9",
and three potential indications of 0.14 to 0.35 volt between 22.1" and 46.5" above the top of the sleeve.
No indication was detectable by the bobbin inspection near the top of the sleeve where the response from the end of the sleeve tends to mask any potential indications. Since the bobbin coil is apparently'" blind" to indications in the parent tube within about a % inch above or below the end of the sleeve, it was concluded that this region had not been adequately inspected since the sleeve was installed in 1992.
The bobbin data from the 1997,1R14 inspection were reevaluated for potential precursor indications.
This review identified 3 closely spaced indications in the area of 6" to 8" above the sleeve. The bobbin voltage for the indications ranged from a few tenths of a volt to 0.96 volts. It was concluded that, based on the bobbin guidelines in place for the IR14 inspection, this indication should have been called and was classified as a missed indication. This cluster ofindications apparently coalesced to grow to the 12.7 volts indication. Other indications identified above and below this elevation did not show any significant growth from 1R14 to the unplanned outage.
Bobbin data from the 1992 inspection, the time the sleeve was installed, were also reviewed for precursor signals. A small precursor signal of about 0.2 volt could be seen at the elevation of the 12.7 volt indication found in the unplanned outage and at the elevation above the sleeve end. These signals were too small to be classified as missed indications but indicate that low level indications were likely present as early as 1992.
+ Point Insoection Inspection with a + Point coil identified two large axial indications including the indication found by the bobbin inspection and an axial indication extending down to near the top of the sleeve. In addition, four small indications of 0.23 to 0.37 volt were found between 4.2" and 6.0" and five small indications of 0.32 to 0.46 volt were found between 14.5" and 27.7". These + Point indications confirm the small indications found by the bobbin inspection. No circumferential indications were found in the inspection.
The lack of circumferential indications would tend to indicate that the indications found were not associated with stresses resulting from the sleeving operations.
3.2 1R15 Inspection Results Following the unplanned leaker outag., a3 B had an operational leak rate of about 20 gpd that remained approximately constant until the IR15 outage. The leaking tube was identified to be R43C32 and resulting from an axial indication just above the 1" TSP.
Table 3-1 summarizes the number of freespan ODSCC + Point indications and tubes found in the last three inspections and the 8/98 unplanned outage. The number ofindication calls is seen to significantly increase from 1R14 to 1R15. The increased numbers ofindications, particularly within 1.5" of a TSP, represent an inspection transient resulting from changes in the bobbin analysis techniques and associated training. Three changes were implemented: (1) Increased analyst training was implemented for the 8/98 and IRIS outages to emphasize the importance ofidentifying and calling small freespan indications; (2)
Q tubemtWala98Pinal oA\\ Final Freespen oA-3Adoc-O#W99 3-2
the tube regions near sleeve ends, found to be blind for bobbin detection, were inspected with + Point probes including 100% of these areas in the IRIS inspection; and (3) based on the review of the prior cycle inspection data for R43C32 in SG B, bobbin analysis techniques were revised at IRI5 for calling indications near the edges of the TSPs. It is seen from Table 3-1 that this change to the bobbin analyses led to a substantial increase in the number ofindications found near the edges of the TSPs. These additional indications would not have been called using the calling criteria in place at IR14 and the indications an: not considered to be missed calls at 1R14.
The inspection identified a few relatively large indications in each SG and the remaining indications were small with low growth rates. The larger indications include: SG A - R32C17 near 1H, R43C32 near lH and R8C2 near 2H; SG B - R43C32 near 1H (R25C51 also from 8/98 inspection); SG C -
R10C2 and R15C68. Sizing of these indications is further addressed in the growth rate analyses of Section 6.
Indications which were candidates forin situ testing were also UT inspected to characterize the d: gradation. Some tubes in peripherallocations and above sleeves were not accessible for UT inspection. The large indications UT inspected include: SG A - R32C17 near lH, R43C32 near 1H and R8C2 near 2H; and SG B - R43C32 near 1H. The SG B R25C51 and SG C R10C2/R15C68 indications could not be accessed by UT.
QatenMaWa98Fid oAEM rmspan oA 3,Uoc&m/99 3-3
Table 3-1. Farley 1 HL Freespan ODSCC Historical Trend Outages 1R131R14-August '98 -1R15
+ Point Confirmed Indications for tubes >7.5" ATS or above TSPs SG Location 1R13 1R14 8/98 Outage 1RIS Axials Tubes Axials Tubes Axials Tubes Axials Tubes A
TSP + 51.5" 0
0 1
1 13 12 TSH TSP +>1.5" 1
1 13 4
24 12 B
TSP + s1.5" 0
0 1
1 0
0 17 12 TSH, TSP +>1.5" 4
4 17 17 17*
13 41 25 C
TSP + s1.5" 0
0 3
1 11 10 TSH TSP +>1.5" 8
8 14 12 55' 41 Totals TSP + 51.5" 0
0 5
3 0
0 41 34 TSH, TSP +>1.5" 13 13 44 33 17*
13 120' 78
+ Point indications. A single + Point indication can include multiple bobbin coil indications. The above table in the condition monitoring repon counted number of bobbin indications confirmed by + Point, which results in larger numbers.
1 t
Q:tubeint\\ala\\ala98\\ Final OA\\ Final Freespan OA-3,4. doc-04/21/99 3-4 j
a
4.0 IN SITU AND PULLED TUBE DESTRUCTIVE EXAM RESULTS 4,1 In Situ and Pulled Tube Destructive Exam Results for R25C51, SG B 4,1.1 In Situ Test Results for R25C51 During 8/98 Unplanned Outage Due to the presence of the sleeve in R25C51 in SG B, a conventional in situ test using a mandrel located at the flaw elevation could not be performed. A whole tube test was necessary using seals at the' inlet and outlet of the tube. For a whole tube test, the pressure test can be limited by the flow rate capacity of the test equipment since a bladder cannot be located over the flaw. This was found to be the case for R25C51 as discussed below.
Leak rate measurements were performed in increments of 200 to 300 psi starting at 1200 psi primary to secondary pressure differential. The leak rates measured were 0.24 gpm at 1200 psi,0.36 gpm at 1400 psi,0.42 gpm at 1700 psi,0.84 gpm at 2000 psi and 1.93 gpm at 2300 psi. Sufficient pump flow could not be maintained to increase the pressure above 2300 psi and testing was terminated.
The measured room temperature leak rates can be adjusted to operating temperatures and extrapolated to SLB conditions (2560 psi pressure differential) using the EPRIleak rate adjustment procedure of EPRI Repon NP-7480-L, the database report for ODSCC at TSP intersections. The resulting operating temperature leak rates are 0.23 gpm at 1400 psi (approximately normal operating conditions) and 2.5 gpm at SLB conditions.
The normal operating condition leak rate of 0.23 gpm (331 gpd) obtained in the in situ test exceeds the maximtur. operating leak rate at a spike of 95 gpd prior to the plant shutdown. This increase in leak rate tends to indicate that ligaments in the crack may have tom during the plant shutdown process to result in a larger throughwall crack length in one or both of the deep cracks. Neither the leak rate data nor the eddy current depth profiles permit a firm conclusion on whether one or both cracks were leaking.
However, the eddy current depth profiles (Figures 4-1 and 4-2) and voltage levels for both indications would imply that both indications would likely have been leaking at the time of the plant shutdown.
This conclusion is supported by the destructive exam results described below although the upper indication appears to have dominated the leak rate. It is possible that throughwall penetration of both cracks contributed to the spikes in the operational leak rate.
4.1.2 Destructive Examination Results for R25C51 Relevant background for evaluation of the indications in R25C51 of SG B includes the following:
About 0.2 volt bobbin indication in 1992 (time of sleeving) at elevation just above sleeve end o
(elevation oflower large flaw at 8/98 outage)
About 0.96 voh bobbin indication (missed call upon look-back) in IR14 at elevation about 8" above o
sleeve (elevation of upper large flaw at 8/98 outage)
Leaking tube leading to 8/98 unplanned outage -large flaws found just above sleeve and about 8" o
above sleeve. SG secondary side pressurized for led testing during outage.
Tube pressurized to 2300 psi pressure differential during in situ testing at 8/98 outage o
Tube plugged at 8/98 unplanned outage (l.182 EFPY) o Tube deplugged and pulled at IR15 outage (1.29 EFPY). SG secondary side pressurized for leak o
testing during outage.
Q:tubeintWainla98Pinal oA71nal Freespan oA 3.4. doc-04/21/99 4-I
Tube R25C51 was cut below the 1" TSP for removal. The tube was sectioned to obtain 3 pieces with field identified flaws including the lower large flawjust above the sleeve, the upper large flaw about 8 inches above the sleeve and a smaller flaw at a higher elevation about 16 inches above the sleeve. Post-pull visual examination of the tube cc,uld identify the upper large flaw but not the lower or higher flaws.
Diameter measurements found that the tube is ovalized about 4 to 5 mils radially at both the lower and upper flaws whem measurements were made. This magnitude of ovalization is very unusual for a straight section of tubing although within drawing tolerances. All flaws were found near the major diameter of the elliptical shape on one side of the tube. The ovalization likely contributed to crack initiation and growth which would tend to support occurrence of only a few large indications at ovalized tubes.
The laboratory leak rate tests were performed in the Westinghouse high leak rate facility at room temperature conditions to provide a leak system capability up to about 10 gpm. Leak rates for the upper flaw were measured between 1502 and 2170 psi AP. The leak facility capacity was reached on the upper flaw at 2170 psi with a room temperature leak rate of 9.95 gpm. This compares to the in situ test result of 1.93 gpm at 2300 psi. The large increase in leak rate for the post-pull test at a lower AP than the in situ test show that damage to the crack occurred between the in situ test and post-pull test. The crack damage likely occurred as tearing ofligaments during plant heatup/cooldown, secondary side pressurization, tube removal and/or ligament loss / crack growth while plugged for 0.11 EFPY.
The measured leak rates adjusted to operating temperatures are tabulated in Table 4-1. Calculations were made using the CRACKFLO code to estimate the required throughwall length to obtain the in situ and post-pull leak rates. This analysis indicates tnat the throughwall crack length increased from about 0.55" at the 8/98 outage to about 0.76" at the post-pull leak test. To obtain a 0.76" throughwall length, all ligaments and crack depths above about 85% depth (based on destructive exam depth profile) would have to tear or grow to throughwall. The destructive examination results show a throughwall length of 0.53" for the upper flaw, which is consistent with the calculated 0.55" at the 8/98 outage. Thus, the increased leak rate for the post-pull leak test is most likely due to tearing ofligaments. This conclusion is further supported by the increased bobbin voltage from a field value of 12.7 volts to 24.7 volts in the post-pull laboratory examination. The post-pull leak test cannot be considered to be a valid test due to the obvious crack damage following the 8/98 outage and the in situ test result of 2.5 gpm is applied for the R25C51 SLB leak rate. This result demonstrates the value ofin situ testing even when tube removal is planned in that the in situ results permit assessment of potential damage from tube removal operations.
At AP = 1550 psi for the post-pull test of the lower indication, the leak rate was only 0.00024 gpm and was an intermittent leak. This indicates that this crack was not significantly plastically opened by the 2300 psi pressurization from the in situ test or significantly damaged during the pulling operations.
When the post-pull measured leak rates (Table 4-1) are adjusted to standard SLB conditions, the SLB leak rate for the lower indication is 0.62 gpm. This result indicates that the in situ leak test results would be dominated by leakage from the upper crack.
Following leak testing, the indications were burst tested at room temperature conditions. The resulting burst pressures were 2889 psi and 4775 psi for the upper and lower flaws, respectively. When corrected to operating temperatures, the burst pressures are 2633 psi and 4352 psi. The upper indication burst pressure exceeds the SLB value of 2560 psi and the 3APuo = 4308 psi. Table 4-3 compares the measured burst pressures with predictions from the + Point depth profiles and the destructive exam depth profile. Calculated results are provided for the latest Westinghouse burst model that is used for Qtubein6alaiala98\\Pinal oA\\ Final Freespan oA-3,4. doc-04/21/99 4-2
the operating cycle length evaluations in Sections 8 and 9 of this report. For the upper flaw, the measured burst pressure is slightly underestimated using both the destructive exam and + Point crack depth profiles. For the lower flaw, the measured burst pressure is significantly underestimated by the NDE profile. The destructive exam profile for the lower flaw leads to an underestimate of the burst pressure by about 400 psi. As noted below, the underestimate of the burst pressure for the lower flaw is attributable to NDE overestimates of the crack length.
To funher assess the potential changes in the indications between the 8/98 outage and the post-pull tests,
+ Point depth profiles from the 8/98 inspection (before in situ test) can be compared with the post-pull and destructive exam profiles. These comparisons for the two large indications are shown in Figures 4-1 and 4-2. For the upper flaw, the results show an increase in the deep length of the flaw for the post-pull data which is consistent with the increased leak rate. The destructive exam data for the upper flaw show a deep crack over about 1.2 inch with shallow tails particularly at the upper end of the crack where the shallow cracking extends for about 0.7 inch. The destructive exam data in the figures are corrected for i
uncorroded ligaments (14 ligaments) and averaged over 0.1 inch, which is less than the effective coil fi:Id of about 0.16 inch for the + Point coil. The average depth for the upper flaw is obtained for a 1.2 inch length. Burst pressure calculations indicate that the minimum burst pressure occurs over cpproximately 0.78 inch of the crack length. Overall, the agreement between the NDE and destructive exam data are reasonable over the deep part of the flaw.
For the lower flaw, the post-pull + Point data of Figure 4-1 show a small decrease in depth which is consistent with no damage since the 8/98 inspection. These results support the conclusion that crack damage occurred to the upper flaw only and are consistent with expectations from the leakage data. The destructive exam results show a nearly uniform throughwall indication with only two ligaments in the 0.553" throughwall length. The uniformly deep crack with few uncorroded ligaments tends to indicate that the flaw may have been present as a deep indication for an extended period of time. It would generally be expected that this indication would leak more than found in the tests. It is possible that constraints from the sleeve to tube weld just below the indication restricted the crack opening for this indication during plant operation and the in situ tests. The NDE results also predicted a nearly uniformly deep crack but significantly overestimated the length of the crack with the 0.590" length estimated near 0.8" by + Point analyses. It is expected that the NDE overestimates oflength are due to coil lead-in and lead-out effects with the coils sensing the deep flaw about 0.1" from each end of the flaw. For the upper flaw with the ends of the crack not as deep as the lower flaw, the coil end effects do not increase the length of the crack compared to destructive exam results.
Destructive exam depth profiles were obtained for three additional sections of R25C51 in order to support POD evaluations. A section about 16" above the sleeve top was burst tested and the depth profile measured. Two additional sections were bent open to obtain the depth profiles. These sections were 4.3" and 6.0" above the sleeve top. The results of these examinations are given in the POD development of Section 5.
4tt In Situ Test Results for 1R15 Farley I had primary-to-secondary leakage of 20 gpd in SG B at the time the unit was shutdown for the IR15 outage. The source of the leakage was determined to be throughwall cracking degradation in the tube at location R43C32 on the hot leg side of the SG at an elevation just above the first tube support plate. To estimate the expected performance of the tube during postulated accident conditions and to Q:tubeintWaWa98\\ Final oA\\ Final Fnespan oA-3.4. doc-04/21/99 4-3
i ascertain whether desired structural margins against burst during either normal or accident conditions were present, the tube was subjected to in situ leak rate and pressure proof tests. Leak rates were measured at pressures near the steam line break (SLB) differential pressure. A proof test pressure of 3628 psi was reached before failure of the bladder being used to apply the internal pressure to the tube.
A video inspection of the indication following the bladder failure was performed to characterize the indication and further assess the burst pressure capability of the indication. The evaluation of the burst pressure based on the visual results indicated that the burst pressure, based on a crack extension requirement for a burst, could have been significantly higher than the measured 3628 psi value although likely less than the 3APuo guideline of 4308 psi.
In situ leak rate measurements were performed on tube R43C32 up to and beyond the SLB pressure i
differential of 2560 psi. The room temperature leak rate measurements were then adjusted to operating temperatures using the EPRIleak rate adjustment procedure. The operating temperature leak rates are i
0.023 gpm (33 gpd) at the normal operating pressure differential of about 1450 psi and 0.61 gpm at SLB conditions.
In addition to in situ testing of F.43C32 in SG D, other freespan indications were tested that exceeded the EPRI guidelines for selecting tubes for in situ testing. The NDE thmshold parameters of the EPRI guidelines are plant specific. For Farley-1 at the IR15 outage. the Parley threshold parameters for leak testing are either a + Point maximum voltage of 1.5 volts and a maximum depth > 75% or a maximum depth > 75% for a length > 0.1". For pressure testing, the threshold parameters are a crack length >
0.44", a maximum depth > 60% and an average depth that is a function of the crack length. Table 4-2 summarizes the NDE data for the freespan indications tested and a few of the larger indications not tested. A total of 20 free span indications were in situ leak tested including 6 in SG A,6 in SG B and 8 in SG C.
Five indications, including R43C32 in SG B, were found to have leakage in the in situ tests. The 5 indications that leaked had maximum + Point voltages ranging from 4.7 to 15 volts. The 15 indications that did not leak had voltages ranging from 0.79 to 3.19 volts. Table 4-1 provides the temperature corrected normal operating and SLB leak rates for the 5 indications that leaked. With the exception of R43C32, the operating leak rates for the 4 leakers ranged from 0.001 to 0.003 gpm and the SLB leak rates ranged from 0.006 to 0.045 gpm. These leak rates are small such that only R25C51 from the 8/98 unplanned outage and R43C32, both in SG B, represent significant contributions to leakage at SLB conditions. The 15 non-leaking tubes tested support a conclusion that freespan leakage would be limited to the 6 indications given in Table 4-1.
The in situ measured leak rate of R43C32 in SG B of 33 gpd at normal operating conditions is higher than the operating leak rate of about 20 gpd up to the IRI5 outage. No operating leakage was detected i
in SGs A and C prior to the outage although the in situ test results of the Table 4-1 would indicate leak rates > 5 gpd in these SGs.
A total of 15 indications exceeded the thmshold parameters for pressure testing and 14 indications were tested. Tube R8C2 in SG A exceeded the threshold parameters but was not tested due to the high personnel exposure required to test this peripheral tube location. The NDE parameters for this indication
(+ Point 2.95 volts) were exceeded by other indications that did not leak and passed the pressure test. To conservatively bound the 3APuo pressure differential of 4308 psi at operating temperature, the target pressure for the room temperature in situ pressure testing was 4909 psi. The target pressure was satisfied with no bladder ruptures or excessive leakage for tubes tested without a bladder for 12 of the 15 Q:tubeint\\ala\\ala98\\ Final oA\\ Final Freespan oA-3,4. doc-04/21/99 4-4
tubes tested. Tube R43C42 in SG B is discussed above. The two other tubes for which the test equipment did not permit reaching the 4909 psi target were R10C2 and R15C68 in SG C. Both of these tubes had to be pressure tested as a full tube test without a bladder. Tube R10C2 is a peripheral location which did not permit adequate clearance for movement of the sealing bladder to the flaw location and the indication in R15C68 is above a tubesheet sleeve. The pressures attained in these full tube tests were Emited by the leakage capacity of the test equipment to near 2 gpm. The pressures reached were 3525 psi for R10C2 and 4250 psi for R15C68.
Comparisons of the in situ and destructive exam pressure test esults with predictions from the NDE profiles are given in Table 4-3. It is seen for the 5 indications in SGs A and B that were pressure tested with a bladder in the destructive examination or in situ tests, the predicted burst pressures from the NDE profiles significantly underestimate the measund test pressures. For SG C, the predicted burst pressure is in good agreement with the leak limited pressure (lower than burst pressure) reached in the in situ test for tube R10C2. For R15C68, the predicted burst pressure exceeds the leak limited pressure test result which would indicate that the burst pressure is much higher than the leak limited test pressure. Overall, the results of Table 4-3 show that the combined use of + Point depth profiles and the burst correlation lead to conservatively low predictions of burst pressures. This results supports a conclusion that the operating cycle length and burst probability results of Sections 8 and 9, which are based on + Point profiles and the burst correlation, are conservative predictions.
4.3 Destructive Examination of SG C Tube R6C10 Tube R6C10 from SG C was pulled at the 1RIS inspection to obtain destructive exam data in support of the development of a POD for Farley-1 freespan ODSCC. The tube was cut just below the 2"dTSP such that two freespan' lengths were obtained. The freespan sections were further cut into ten sections for destructive examination. Each section was burst tested and the depth profile was obtained for each of the burst sections to provide 10 profiles for the POD evaluation. These data together with the 5 depth profiles obtained from R25C51 and the 6 data points obtained from tube R6C28 pulled at the IR14 inspection provide the database for the POD evaluation. Tne destructive exam results for R6C28 are
. given in the POD development in Section 5.
The corrosion on R6C10 generally occurred in narrow axial bands of short axial racks that ran continuously to semi-continuously along the tube. Occasionally, the corrosion was more wide spread, occurring over a complete quadrant. However, the circumferential positions of the narrow axial bands of corrosion varied considerably with elevation along the tub and occasionally, a given elevation had two distinct axial bands of corrosion. The burst openings including more uncorroded ligaments (between 23 i
and 52 for the ten sections) than typically found for ODSCC. The large number of uncorroded ligaments could be expected to reduce detectability by NDE. The general orientation of corrosion in R6C10 was very similar to that found for R6C28. However, the burst openings for R6C28 had fewer ligaments 1
(typically 5 or less) and detectability for shallow indications would be expected to be better for R6C28 i
than for R6C10. The corrosion for R25C51 occurred as a single narrow axial band of short axial cracks that ran semi-continuously along the tube. The number of ligaments in the cracks examined for R25C51 varied from 3 to 19. Together, these tubes can be expected to span the range of uncorroded ligaments
- and provide a wide range of ligament influence for the evaluation of bobbin coil detectability.
Qtbeint\\ala\\ala98Winal oAWinal Freespan oA.3,4. doc-04/21/99 4-5
l!
e r
f u
od =
t ecP l
u oAi s
s r
p
)
2 e
1 3
6 5
r pt m *5 a
2 1
0 0
4 e
4 1
t t
6 6
0 0
0 0
s n m0 t
p o0 0
e e p3 g
2 N
0 0
0 0
t mg 2
(
a n
k i
t s
ois a s 5 =
t eis p e u9 P lu nip t
l j
d 01 ed 9 A 5
s i
r af e
Ln6 u e o a t
o5 =
R mC2 t
t a
t r al e n.
t r
t a
=
e u a s
ek s,
pk s r e
t ap ma ek T
S e
r e
a rD c.
)
el t
a r
t s e k
F B
P u
mR I
l e
a 6
o m
t e
1 h
8 0
8 oP e L
6
/
6 5
8 4
3 2
oE r p 3
0 r
u g s
5 1
3 9
gta3 1
e n,
1 u n r r
1 9
t e
b t
i is s e l
u p1 i
u g
(
n md T
T
. u t h I
er e a d
t s
t e
a n a mh l
e o r c
e oi l
)
)
)
)
i t
p oh t
)
8 u a
)d d
d d
d d
a r
my w 9P R
p p
p p
p p
c 9
6 g
g g
g g
e r t i
1 d k
)
g
- 0 d
t o s 3
4 3
3 9
n r n a
m 3
5 0 t
e d
at i
0 3
l 4
4 2
e r e
p 3
70 e a L
g
(
r u
(
(
(
(
(
t e a ot 5
0 3
1 3
3 2
wvbi b u d
(
0 2
0 0
0 0
o a a s e
2 0
0 0
0 0
mit L n l
t 0
l e
1 eS s
0 0
0 0
0 d
i v
u d
n 4 o n j
a n
e a i
a n s
a gh I
d eN-A s p ei uP 6
i r
a t
t l
s b
n p0 p p o
p oA a1 s l
4 ai05 u08 r Tt n 5 7 o
mta4
/
6 9 e ni rr1 =
m5 i
8 w o2 o
oe n
r Ut Np=
r f =
gl c
)
n t
e O p s,
u epi a y p 3
aA w o
g e
D e
s c
h 8
5 8
5 7
ot l
1 1 2
2 5
7 3
k t
l r n a
P,
/
6 70 a al r aI 1
5 0
0 0
0 eBf k s
5 o
0 F
r l
e e
a t
f L e e ga O
oS gl i
a momh t
t g
u u
ai s d O
ed h s ak h t
t r
nl c c M
e e o a u k
s p m
r 1
2 e
c
/
a 4
a r
e pt o a 0
r t
e x o
c L
B o
r e e t
d u s h s 4,
d G
t ci d d 3
n S
s i
i t
a e
h s s e A
a T
wpel o
,e t
t n
5 v k B
A A t 0 d s
a t
1 s e a
s 0i e p
C s
ee G
G e3 l
t C
n Tl r
G e
t al e
2 e
S G
S S
u vl F
s I
r u
kS e
= a p l
O m
aT b oo
+,
n n S
C t
a i
n s
i
+
+
Pd e
ui E
c i
t F
i l e Tt a H
+
G H H nA e s
\\
t e
v i
r o A
a H
S P
u o d c c I
et eP o
1 1
S ib edd P
H P
P b ad l
t i i 1
a u mi S a l
s in n n S P
S S
l i
T S
S t
p n s F
ns ui i
T T
T T
e g o p 8
\\
I n P r r l
9 o
e e 2 2
7 o3 c0 la 8
pw 3
2 3
1 1
1 i
t 7
a 6
h 5
5 t
CU1p4 C C C C C sW1 N2 la 9 o1
\\
a a
Cc e
5i 5
3 0
5 3
2
\\
t t
2d 2
4 1
1 4
3 o
ie n
n RI R *.
R R
R R
R N1 2
b u
t.Q
i 1lI
)l l
y n
n le0l 4
a i
i r
s 6 n r
7 o
o ot ig f
s f
l uo r 3"
h h
t c
a t
t n
s sb s
1 p
p n a m e
t t
q m
e o
q 0
e e
n r e
d d
i r
m o
r t
i ahf l
l o
ct l
l e
a 5
5 5
C a
i id ws 7
7 7
e s
o s
v d
n e
idd d
e e
e e
e 5 ch. e 6"
l c u 5
0 s
xt Ta gt c u s
xt 0
1 4
S pi s
Ei UsI e Ei 0
0 0
T l
s t
s t
lu 0
s 0
e d 5 d 0 n2 n0 R
s s
t a5.
a 2
i i
t 1
a s 8
d t
t l t
s s u 0 mp 0 a p l
n u
e s 0
0 m1.
t t
t t
s t
a e Te pI t
f g4 s f p4 s k s k s k s k s n R R
o 8
r r
r r
r r
u o g8 u a
a u a
a 22 u
u u
b k
2 b c
e e
e t
k a
b b
b b
s 1
9 I
l l
l a0 o
e 0.d o oo oo oo oo r.d l
a e
P T n
n 0 aN L0 aN NN NN NN NN I
t u
s t
e iS s
T e
d 9
d 9
9 9
9 n
r n
0 n
e i
i i
i t
r 9 os0 o
s0 s0 s0 t
I t
u s ui a a 9
a a p9 p9 o
o p9 p9 es s u0 t
t t
sk k014 t
t s
4 4
k01 s9 k1 kI k1 4 4
Te i
k1 04 t
- a. 5 8 ai a04 e4i a4 a a4 a4t a4 S
r t
t t
r s
e58 r s
a a
a r8 e8 e8 e8 P
22P p L22P ap I
t n
2P I 2P L2P L2P t
I C
u t t 7 C
SsL i
/
L L
L L
L Ll e S
TP o
e
.l b
D h
P N
P P
P P
Pf u u
O' t
t h
t a
g p
n 62 33 03 67 6
61 99 e
67 53 86
- 4. 4 8
- 4. 3 1 9 7
s e
L 00 00 00 00 0
00 0
4 1
erF
.h gt f
vp 1
9 3
2 9
0 00 o
AeD 1
3 9
2
~
y n 4
22 1
t 8
8 7
6 6
8 r
i a
o P
m
.h 0
0
+
xt m
ap 0
0 0
0 Me
%1
%1
%7
%9
%0 D
7 90 70 00 2
90 8
u 2
S 9
98 98 85 9
97 2 <
1
-y s
9 t
9 e
U l
l o
8 5
0 3
6 7
94 V
r 1
2 V
79 9
0 2
0 1
- 7. 7 4
a F
45 2
8 00 c
1 1
1 o
d 2
n
(,
s i
bl 3
t 4
b o 1
2 4
4 oA e
oV 4
0 8
4 5
1 B
7 6
n l
1 1
1 1
ia b
m aT 9 5 iF 2
6 9
6 s
- 2. 1 1
n
- 9. k 1
7 3
8 n
0
+4 m
1 o
0c 0
1 1
i
+ a
+
+
+
+
, 4, A
ta
,r 0 6 o
8 9 c 2
3 0
5 v
1 1 l
e 6
- 3. 5 3
3 5
3 0
an lE 0
0 -
0 0
0 0
3 E
+T
+
+T
+T
+T l
+
+
8 I
1 i
9 1
I 1
I T
T 1
a 1
I U
2U I
1 U
I2U S
W U
1 1
I 1
a 8
M e
2 7
7 8
5 n
b 3
1 1
1 3
5 k
u
/
2
/
/
/
/
3
/
2 6
3 D
T 4
8 3
3 3
4 5
W u
1 SG A
A A
A A
A A
ll11
k k
s 5 cTc k P s
a TcorUr a
e c
"9 v aS s
t c mne5
- 2. b cT gca o
r UI 2
- r. dnk oe o&
n n
0a 2i s
s i
t r
,s Tth n
e a cf e 4
e k 0 a s s 4 k 5
Ui n
a r y yas c
7>
w s cal k
1 a
en rk we3 c k
r C
ga nai o a c 0 8
I d
c e
r a h
o c
k t
r s"
t d
5 cl nC 0 c w3 g
5 n 7
5 n r a gl c
7> a 7>
Td o0 n
ias aiP kUn h e
t s
e c n s0 l
2" wiiRcd r
'9 v o
P e
1 l
pot c a
c TdnS u
- 2. b 1
1 l
o l
r h e n
. r n e Os vi 1
c a s UaT F
0 a 0
0 s
t lu t
1 s
a e
m2 o6 7
e t
R p
0 r
.d u
s g1 f
s t
0 e t
o p
d t
t l d 0 s t
e s u 1
l u
g n6 uB n
u e s 1
r a
e Te a2jL s
i t
t t
t a
0t d r s k s k
k s k s k s k
f n R R
r r
r r
aS e p a
a a
a a u a
o0 a
.d u
u u
8 eb e
eb eb eb e
a t
k5 mk0 mda2 l
l l
l l
l s
a2 a6 lP T et pr5 pl 6 oo o
oo oo oo o
e l a gI 2 gB3 NN N
NN NN NN N
t u
s t
e iS s
0 e
a0 i9 i
T 1
1 9 i9 1 9 s
t 40 4
40 40 40 pp n
r a2 0 89 8
89 89 89 u1 t
u sui d1 5 24 2
2 4 2 4 24 4
I es s t
sk k m2 ot o
ot ot ot a
7 k 8 Te i
a r
o t
a t
t a
t a
t a
2 S
i P
rr os LP L
I t n
I C
u l
r l
e l
t t 7 l
C isL u
u d
f Se/
f e ed L
L L
L S
nTP
- b - b a o
o D
LtuPtulb P
L N
N P
P P
L I
O n
ht a
g 7
9 1 1 26 469 79 9
4 08 1 32 73 58 8
p ne 9253 532
- 4. 1 2 5
66 058 53 31 s
L 0000 000 00 0
0 00 00 00 00 4
e l
erF
.h gt f
vp 1 6 35 9
0 9
0 55 9
9 o
AeD 35 77 8
2 2
96 4
3 1
t 88 78 7
7 5
6 64 6
7 y
n r
i a
o P
m
.h 00 0
m ap 00 0
0 0
%0 5
0
+
xt Me 0%1
%%9 %1
% 7- %07
%9 %6 1
D 0900 490 70 4
5 80 u
0 05 90
%65 1
S 988 997 98 9
7 85 87 84 1
1-1 51 9
y s
t 5 9 50 9
e
/
1 l
1 3 21 2
l o
r 9
5 1
3 38 6
0
/
V t t t
t 4
a PP PP 1
1 1
8 09 8
8 0-F
+ +
+ +
3 1
1 2
0 0
0 c
1 o
d.
2 n
4 s
3 i
bt A
l 4
b o 8
9 6
8 8
5 7
4 o
oV 6
6 5
1 1
u 8 8
2 u 8
n e
B 2
3 2
2 7C 4
0 2C 0
w lba F.
T 3
3 6 0 8
5 la n
5 07 8
1
- 2. 3
- 9. k 7 in F
C 0
0 0 c 0
o l
1 1
1
\\
Co
+ +
+D
+,
+0
+ a +
o A
it
,+
,r a
7
+ t 94 6
4D 8
1 6
1 c 0 l
v 3
a 6.0.C C 02 43 8
- 5. N 4 7
- 4. 2
- 4. I 4 n
e 01 00 0
0 0
0 0 0 0 -
0 i
l F
E
+TT HHT
+
8
+ +
+T +T +
+
+T +T l
H+HT 9
H H
H T HH H
H H
a UU U
U 2U 2U U
U U
a I I 1 1 1
I 1 I
\\
1 1
ha
\\
e 2
3 9
tn b
3 3
3 2
5 3
ic u
/
/
9 9
/
3 3
/
3 u
b 3
2
/
/
5
/
/
1
/
t T
4 4
6 9
4 9
9 1
6 Q
SG B
B B
B B
B B
B C
2 on d
la n
e n
ra a0 C
v i
l l 5 E
e g
s e
t e
r ua T
t l
l a
n l i s
x u U
e b
m e a o l
S m
i a
s c
b O
T, ro s0 aP s
s s
e f
e1 nS N
o c
C c
e ahiT e-t a
s ci v
5 t
o o
d d wd n e
7 i
t ni e
n h
nh
.h i
l i
s 5
T gt c etc is Tt e
ai a S
1 T r
U Ht Upcp T
0U s
t lu e
d 0 e
s 0
g 5
g n
e
&5 a
a6 a
6 k
1 k
R 41 a
6 a
s s
t 0
.t e.
e e
e e
0 mpe ys g
i t e
sl 0 a s sl e
t d
- 0. taie t
t l i
l s u i
n u e s 0 mpteb p a
a a
a 0
s ys g
g g
g t
p a
s Te p1 eb c
e f
p1 r 5
k k
k k
f g4 n R R
o g4 ude2 a
a a
a o
8 u e5 a
d 0 k
r t 5 e
e e
e k 42 s k
8 s 2
e e
ts a0 2 em3 o o
o o
t si l
l l
l 3
si l
l r 0. d em4 o
a e
1 P T e.&Pl o N N
N N
r nr t
I 0 aPl o N
l 0 i
i t
t a
s t
i e S s
i i
i T
e i
b o r.
e n
r i
i os s
s s
s i
u /o r b
u s ui os0 o p op op o
t p
os0 I
t e sa p9tu/ e t
p t
e t
t t
t wd p9 t
wd sk k1 9l d
t k1 k1 k1 k1 9l d
Te k
1 k
1 i
a4 4l t n
4 a4 a4 a4 a4 4l t S
r a
4
- us a
- u s a P
e8 e
8 r8 e8 r8 r8 e
8 l
n l 2PFteb L2 l
I 2 I 2 I 2 I 2PFteb L2 I
C u
t t 7 C
sL iS e/
L S
nTP o
L, D
I P
L L
N L
L P
L O
n h
t a
g p
n 8
73 06 9
36 53 29505 2
9 e
7 1 6 32 4
22
- 6. 4 6 5. 2. 7.4 4
s L
0 00 00 0
00 00 00000 0
4 eerF
.h gt f
vp 6
3 3
6 6
o AeD 5
1 6
7 7
t 8
8 7
7 6
y n
r i
a o
P m
.h
+
xt m
ap 0
5 0
0 Me 0
%6 %9
%9 %9 0
D 0
50 45 9
u 0
%7 70 73001 0
S 94 97 9
77 79784 1
1 1-m y
s e
t l
i i
o 8
7 3
4 4
8 04562 7
/
2 r
V 7
0 3
9 8
1 8.9.6 6 5 8
4 c
0 F
8 1
2 0
1 2
281 1 1 0
co d.
4, n
2 is 3
t bl A
4 b o 8
4 9
6 2
6 4
o.
oV 3
7 7
7 8
3 21 2
e B
1 3
2 2
1 2
92 0
lb m
e 9
P 6
22.005 l".
T 9
S 7
0 7
s 9753 1
i a
1 1
0 T
7 9
0
- 5. k n
21 1 1 1 6
n
+
c F
a
+, +, + + +
2 n 0 0
+
0 rP o
\\
A
+
+
4
,cS 4.4.7 3.0
+
o i
5 i
t a
3 7 h 7
1 3
6 2T 08643 6
l v
c a
e 0
- 4. t 4 4
5 1
21 1 1 1 n
5 1
l a
n
+ + + + +
2 iF E
0 0 p 0
0 0
0 i
\\
+
+
+
+
+
+
h HHHHH
+
9 8
H H T-HT H HT HTi TTTTT H
l t
a U iU U 3Uw SSSSS a
1 I
1 1
1
\\ala
\\t 2
8 5
i e
2 n
1 b
2 6
4 3
3 1
6 3
e
/
/
/
/
/
/
/
/
b h
0 2
0 3
1 5
5 2
u t
T 4
1 2
1 1
4 Q
1 1
SG C
C C
C C
C C
C
I t
t t
t s
t s
t of s
or or n
e a
s n owe x
x nu nu h
h n
dt ot a
a b
b t
t f
d d
t sl m
m i
e ie i
f o
e d r db d b k
n m
u os c
o t
t t
t a
a t
u t
u a
i sbt n n s t st e r a
m e
t t
t e
e n eoe s
s n
n h c c
t o
ui m t
t i
r r
i i
t t
r i
s C
ut d a r
i l t ip u
u ut ut k e d
t t
e uei b
b l
t u
s u i l s u c w n
i s sb mq oP oP s
s a o i
e NA NA n e ne rl r
u neui I
r I
r c
e I r t l s
e s
rh l
l e
et a
r m
pr m
oE po s
P d
ts rD uF a
e rf
)
W e.
.d t
e r 9
0 7
t 0
8 3
0 s
c u 0
1 h
8 8
9 9
8 0
6 4
t r
iBr o 1
%%t e 1
4 8
0 2
7 5
2 3
2 9
u d
u 2
4 3
9 2
3 3
4 3
4 o
c r84 a osE et B
r sD 2
3 F32i PHe d
rN 99 d n e
P ef f i oo l
t a
i a
f hht ot a
l t
W r ppd u
.h pe eE c
xt ap 0
0 h dd l
eD 9
0 8
0 0
0 e
7 7
9 3
a Me 9
0 9
0 9
9 9
9 p ggN t
D C
s 1
1 1
1 e
t a a d r rd l
d u
)
mv va e e n n
se
.h a aa a
R gt n
x nn o s
vp 4
2 3
- 0. )
1 1
3 9
6 6
e aai n Ae 4
2 t
9 3
5 7
e 3
1 t
t D
8 2
9 8
8 7
8 8
6 vhh a 1
l 8
9 i
u o
t t
c iii o
t s
P c wwl e
urhh e R
+ h)
W t
c cb t h 9
9 1
6 0
3 8
9 s
nn u t
gc 0
0 7
9 9
6 8
5 7
5 eii t s
nn 2
0 5
d 0
0 0
0 0
0 75 a l
1 e
ei T
L(
0 e75 r 1
h e
00h t
e f p mf r
d ooi 0
o o
rd rd o
r u
o e
Ne Ne e e e e rhh e 1
t s
uh ur gd.
r t
p di dt f
t e e t
t r
r s
t o e i
r,u r,u d m d gg 4
r t
et et ai am d
s e
ie d a e d u n ni h
ki d
r SM e
t d
a m a p d p d p e eh l l l i t
t u d u d u b
bl al l t
P n
a i
k k
e WbLl Br aR aR
/o a /o a ukk g l
t l
l i
I s u
e c c cn T
B B
wle wr a ai l
l t
l I
a r r t
iS c cc se l l t
s a a n
t t
i i
i i
i eii r i
a ia s
s s
s s
r t t o
I s
s pt pt p
p p
p p
u r r f aa s
s s
1 t e 6 ri 6 ri 6
4 4
3 3
s p pd s r 9 u s e
9 u s 0
7 7
1 7
maare ye Tu 0 b p 0 b p 63 44 44 32 38 i
2l 3 2l 2 03 7 4 74 3
7 3 pr r u m
1 ooq 3
4 4
2 8
t l
l l
t u3 u5 r
o u
3 > 4 > 4 >
3 > 3 >
sf f
u e
p6 p3 a
r a
t udd F
HP iSt S t t
r t
t t
t t
- 2 i - 4 s
s s
s s
e ee r
r r
r r
b m
s s
t t
r no n o u
u u
u u
l l
h uu u 1
P B
B B
B B
I P I
t s s s 3
2 i
e e o
/
wr 4
r p 0
4 h eex i
i c
s s
r r e o
e pt pt pu ul d.
t s
s l
t r
r b
s e 0 u 0 u d
e s s e 4
s s a
e r 0 bi 0 bi 8
9 9
e d e e n
3 5
0 T
Ts 3
3 u
s s
t r r n A
d
.s 2l p 2l p 2
0 0
s 2
5 ppo o
l l
6 9
9 e
n s
5 2
u9 u5 ar u p8 up7 3
4 4
T 3
4 s s xe at t r
na e
p t
t t
h r r MP t
u u p s
t 8 7
o i
i e
t S s2 S s4 N
gbb h e
r n ot n ot n
g lF P a edd I P a I
i a
e eh i
l n
t t
mla a F
o
\\
r r
l
+
a u ut o
A v
ek ek
+
+
+
+
+
H x c c e2 l
e pc wc H
H H
H H
el uC a
a l
l pr or S
aa a
E Uc L c T
ec cd0 in 2
I 1
I i
1 F
v mmdeR
\\3 i
t 9
c uut r la s
e 2
2 7
8 u mmeo
\\a 1
a 2
b 5
3 3
1 2
/
6 r
t f
/
/
/
/
/
0
/
t ii u
5 3
3 2
8 5
s n nt t
la
\\
T 2
4 4
3 1
eii oa t
1 eDmmNh i _
n s
e t
t b
u o
t.
SG B
B A
A A
C C
N1 2
Q
0 0
3 ls 0
9 0
5 E
0 0
5 1
5 ia D
1 0
9 T
E w
D 1
5 l
lo d
a 8
0 6
t l
u ne.
o 0
8 0
l l
la etc_
P 9
3 9
u P
hw 1
0 8
Sl me.
P a
t r.
-e s
F a
r r
o E
E ig o.-
f e
P P
D Do LC_
r 8
0 4
8 P
8 9
9 9
7 6
1 0
9
^-
)
)
)
x (*
a a
h
(
(
)
,s
. h M "'P t
h t
tp g
.t xt x
a io a e A*
n ).
g p a
en ve
- 4. D MV MD R"
L(
AD i
0 m
i
~
a
+
x TE 9
L I
- 0 I
e L
S v i
(
t 1 c 8
I1 u
0 k r cta s n
r e 7
t C Da 0
D 1
d 4
nR 6
1 a 0
e5 D
rC E N
)
ug5D t
5 n
i2N-i I
0
(
FR 1
i h
4 l
l t
e un g
b pi 4
n 0
e u -
Ts L
t
^
l o
d 3
a eP i
0 x
l A
luE PD 2
N 0
1 l
yle u f.
1 99 l
/
r p
(.
1
- .5 0 2
s
/
a d
F r I
.g:. !: :'
4 e
0 r
P l
A
+
l
.?.=
0 4
3
,'*,*,5*,*,
A O
n O*,
1 aps 0
e r
D l
iF an 2
F i
\\
0 A
O g
0 0
0 0
0 g
0 0
0 0
l g
9 8
7 6
5 4
3 2
1 an iF
\\8 abi o 5jo@yH ta m
9 f
a
\\a la
\\tn ie bu t
Q
0 0
6
_E 0
0 0
7 0
0 2
D_
1 1
1 1
8
~
~
s 0
2 l
t ia s
6 0
3 1
T E
o 4
0 1
6 1
1 w
D R_
1 8
l lo d
t l
l l
ne u
a et
=
7 0
lu P
hw c
r 9
4 3
9 Sla me D
6 9
0 4
P t
r 1
8 e
s F
ar r
o E
igo f
P P
Do LC
)
)
)
(
x(
(
a O
i O
.h a
h h
t
.h t
. s
)
xt x p Mt g) t t
a a lo a e p
n.
gp n
A e ei v e MV MD
~
5 D RD L(
AD 8
7 m
8
~+
Q 7
ax 8
-g 6
8 He S v G
5 i
8
(
t 2 c 4
D, u
8 k r 3
- .*,a ct 8
a s a
e 2
r t
C Da 8
- d D I,.d 2
1 8
4 nE 1 a
.='.b_.
8 e5 D
^
rC E N
= 9 2
)
u g5 D t
(
l
)
7 n
g 1
1 l
t i
2 Nn r1 y
4 8
i
(
l FR i
7 h
l o
T_
t 7
l e uP g
7 b
p+
n 6
e u -
7 L
t Ts
^
o 5
la d
7 eP 4
i x
l A
l 7
uE 3
PD 7
1 N 2
m l
7 yl I
e u 1
2/
l 7
4 r p 0
c a
7 o
e d.
F r 4
9 3
P a
O 6
i 6
A 8
n l
(
n a
6 ps e
ll
=*.E 7
re
- .0 F
I
. f.*
- ;a I
6 lan 6
iF 6
\\A O
5 l
6 an 0
0 0
0 0
0 0
0 0
0 0
F i
0 9
8 7
6 5
4 3
2 1
\\8 1
9 a
ta
^b gh q
c
\\
e j [==
la e
a
\\tn ie bu t
Q
5.0 EVALUATION OF FIELD AND PULLED TUBE DATA 5.1. POD Development POD Database Tables 5-1 to 5-3 summarize the destructive examination results for crack length, depth and axial elevation. Table (a) for each of these tables gives the destructive exam results and table (b) provides the destructive exam results aligned with the elevations of the NDE results for each indication. Two field bobbin coil analyses were performed including the original field analysis and a reevaluation by field analysts at the end of the EOC-15 inspection. The bobbin detection techniques and analyst training were enhanced for detection at the 8/98 inspection and again at the 11/98 inspection. To incorporate these detection enhancements in the POD evaluation, the field analysts at the end of the 11/98 inspection conducted a field review of the pulled tube bobbin analyses. This review led to the detection results given in the field reevaluation column of each of the (b) tables. The original and reevaluated field data comprise two independent bobbin analyses for inclusion in the POD evaluation. Use of the original field data in the POD, particularly for the IR14 and 8/98 inspections, provides some conservatism in the POD evaluation for the IRI5 inspection since these inspections did not include the updated detection techniques implemented for IRI5. Tables (b) include a column identifying the fraction (2/2,1/2,0/2) detected for use in the POD evaluation. This column also includes the bobbin voltage assigned to the indication for developing correlations of flaw average and maximum depth with bobbin voltage in Section 5.3. Detectability is assigned on a one to one basis for cracks that were destructively examined, typically as the burst crack although a few of the indications were bent open for destmetive exam due to the short tube length ovailable. To assess detectability along the length of a few of the indications, detectability over partial lengths of the crack length destructively examined were evaluated for R2SC51 (cracks 4 to 6 total to burst crack 7) and for R6C10 (cracks 3a and 3b total to burst crack 3). Of these 5 sub-lengths. only Crack 3b of R6C10 was not detected in the original field analyses. It is conservative in the POD evaluation to include only the two burst cracks rather than the sub-lengths. Crack 1 of R6C10 is a sludge pile indication. Since Parley-1 applied the + Point probe for detection in the sludge pile (up to about 7.5 inches above 'ITS), Crack 1 is not included in the POD evaluation for bobbin coil detection.
As noted in Section 4, the three pulled tubes show strong variations in the number of uncorroded ligaments in the burst crack profiles. Uncorroded ligaments reduce the signal amplitude and tend to
-increase the difficulty in bobbin coil detection of the indications. The indications found in tube R6C28 had only a few uncorroded ligaments as shown in Table 5-1a, while the indications in tube R6C10 had a large number of uncorroded ligaments as shown in Table 5-3a. The number of uncorroded ligaments for tube R25C51 is intermediate between the other two tubes. As would be expected for R6C10 with the large number of uncorroded ligaments, the detectability of the indications by the bobbin coil was the poorest for this tube. Consequently, the three pulled tubes provide an excellent sample of tubes for detectability evaluations since the variations in uncorroded ligaments span the range of that potentially found for ODSCC indications. In addition, all the indications used in the POD except the throughwall indications in R25C51 have bobbin voltages less than one volt, ten indications have < 0.25 volt and 17 are less than 0.75 volt. These indications would be expected to be in the relatively poor signal to noise r tio region for detection. Overall, it is concluded that the indications applied for the POD development tend toward the more difficult range of detection relative to effects of voltage amplitudes and j
uncorroded ligaments and the resulting POD would tend to be conservative.
A significant improvement in bobbin detection is demonstrated by the field reevaluation results which incorporate the bobbin analysis technique and training updates from the unplanned outage and scheduled outage in 1998. In Table Ib for the 1997 pulled tube R6C28, two additional indications were found in mwnema.m.wu r-.,-omrm o+s*c-ounm 5-1 j
~
5.0 EVALUATION OF FIELD AND PULLED TUBE DATA 5.1. POD Development POD Database Tables 5-1 to 5-3 summarize the destructive examination results for crack length, depth and axial elevation. Table (a) for each of these tables gives the destructive exam results and table (b) provides the destructive exam results aligned with the elevations of the NDE results for each indication. Two field bobbin coil analyses were performed including the original field analysis and a reevaluation by field analysts at the end of the EOC-15 inspection. The bobbin detection techniques and analyst training were
- enhanced for detection at the 8/98 inspection and again at the 11/98 inspection. To incorporate these detection enhancements in the POD evaluation, the field analysts at the end of the 11/98 inspection conducted a field review of the pulled tube bobbin analyses. This review led to the detection results given in the field reevaluation column of each of the (b) tables. The original and reevaluated field data comprise two independent bobbin analyses for inclusion in the POD evaluation. Use of the original field data in the POD, particularly for the IR14 and 8/98 inspections, provides some conservatism in the POD evaluation for the IR15 inspection since these inspections did not include the updated detection techniques implemented for 1R15. Tables (b) include a column identifying the fraction (2/2,1/2,0/2) detected for use in the POD evaluation. This column also includes the bobbin voltage assigned to the indication for developing correlations of flaw average and maximum depth with bobbin voltage in Section 5.3 Detectability is assigned on a one to one basis for cracks that were destructively examined, typically as the burst crack although a few of the indications were bent open for destructive exam due to the short tube length available. To assess detectability along the length of a few of the indications, detectability over partial lengths of the crack length destructively examined were evaluated for R25C51 (cracks 4 to 6 total to burst crack 7) and for R6C10 (cracks 3a and 3b total to burst crack 3). Of these 5 sub-lengths, only Crack 3b of R6C10 was not detected in the original field analyses. It is conservative in the POD evaluation to include only the two burst cracks rather than the sub-lengths. Crack 1 of R6C10 is a sludge pila indication. Since Farley-1 applied the + Point probe for detection in the sludge pile (up to about 7.5 inches above 'ITS), Crack 1 is not included in the POD evaluation for bobbin coil detection.
As noted in Section 4, the three pulled tubes show strong variations in the number of uncorroded ligaments in the burst crack profiles. Uncorroded ligaments reduce the signal amplitude and tend to increase the difficulty in bobbin coil detection of the indications. The indications found in tube R6C28 had only a few unconoded ligaments as shown in Table 5-la, while the indications in tube R6C10 had a large number of uncorroded ligaments as shown in Table 5-3a. The number of uncorroded ligaments for tube R25C51 is intermediate between the other two tubes. As would be expected for R6C10 with the large number of uncorroded ligaments, the detectability of the indications by the bobbin coil was the poorest for this tube. Consequently, the three pulled tubes provide an excellent sample of tubes for detectability evaluations since the variations in uncorroded ligaments span the range of that potentially found for ODSCC indications. In addition, all the indications used in the POD except the throughwall indications in R25C51 have bobbin voltages less than one volt, ten indications have < 0.25 volt and 17 are less than 0.75 volt. These indications would be expected to be in the relatively poor signal to noise ratio region for detection. Overall, it is concluded that the indications applied for the POD development tend toward the more difficult range of detection relative to effects of voltage amplitudes and uncorroded ligaments and the resulting POD would tend to be conservative.
A significant improvement in bobbin detection is demonstrated by the field reevaluation results which incorporate the bobbin analysis technique and training updates from the unplanned outage and scheduled outage in 1998. In Table Ib for the 1997 pulled tube R6C28, two additional indications were found in om==m.ww om.u om.--
5-1
the field review. No differences in detection between the original field and field review analyses were found for tube R25C51 given in Table 2b. In Table 3b, the original field analyses resulted in good detection for the first span below TSP 1 but detection was not as good for the span above TSPl.
POD Analysis Methodoloey In order to meaningfully deal with the POD data, certain assumptions may be made with respect to the alternatives for utilizing the data in analysis models. For example, the data may be evaluated similar to the analysis of the probability ofleak (POL) data for the ARC for ODSCC at TSP intersections..
Following that approach, the POD for each individual indication is either not detected or detected, i.e.,
zero or one. A regression analysis of the POD on the depth of the indication is performed for specified functions considered to be appropriate for describing the data. Basically the process of developing the
. model is one of examining the data, and potential boundary conditions, and finding a mathematical function that best fits the data. Considerations that were made to arrive at a desirable model involved consideration of the form or shape of the describing function and the treatment of boundary conditions.
In addition, some consideration was also given to the potential for estimating the statistics of the parameters of the model. It was judged that the POD should nominally be a monotonically increasing function, such as a cumulative prob 2bility distribution function, of the depth of the indications. There were also additional features which were considered to be desired of the model as described in the following:
1)
If the relative depth, h, i.e., the ratio of the depth, d, of the indication to the thickness, t, of the tube,is small, so should be the probability of detection. This is panicularly important for applications in which the detection of shallow indications is important to the tube integrity assessment. For these applications, the probability of detection for a zero depth indication should be zero. The probability of detection for indications with depths less than some specified value, say 5 to 10%, should be very small or zero. The Farley-1 single cycle analyses of Sections 8 and 9 have been shown to be essentially independent of the number of undetected shallow indications. However, the multi-cycle analyses of Section 7 are dependent upon undetected shallow indication growth over many cycles and obtain improved results with a realistically small POD at shallow depths.
2)
For relative depths greater than some specified limit, as supponed by the data, the probability of detection should be unity. POD data generally show a trend of achieving a POD of one for indications in the range of 40 to 60%. This leads to the conclusion that a bound should be selected for using a POD of one in the model and is important for applications in which the detection of deep indications is important to the tube integrity assessment, which is generally the case.
3)
The POD function does not need to asymptotically approach the upper and lower bounds.
For example, a straight line between the lower bound at a POD of zero and the upper bound at a POD of one could be sufficient.
A number of distribution shapes were examined including the beta, logistic, log logistic, normal, Gumbel, Frechet, Weibull, and Kunin distributions. The later four distributions represent Types I
- through IV of the extreme value distributions, i.e., describing the stochastic behavior of the minimum and maximum values of other distributions. Hence, for each extreme value distribution form, there are individual distribution equations for the distribution of minima and maxima. Funhermore, each of the distributions was also considered based on evaluating the POD as a function of the logarithm of the depth ratio. The following conclusions were reached relative to the potential use of each of the distributions.
emur-.we,
=w%mw--
5-2
1)
The logistic, normal, and Gumbel distributions are unbounded at each extreme, however, the depth ratio cannot be less than zero nor greater than one. Hence, the logistic, normal, and Gumbel distributions were judged to be undesirable for describing the POD as a function of the depth ratio when detection of shallow indications is imponant to the problem. For the current Farley-1 analyses, the detection of shallow indications is not important for single cycle analyses and the logistic function is further considered for the POD evaluation. The normal and Gumbel distributions offer no advantages over the logistic function and were not further considered for the POD function.
2)
Transforming to the logarithm of the depth ratio can be performed to address the first issue, but, cannot account for the upper extremes being unbounded. Hence the logistic, normal, and Gumbel distributions werejudged to be undesirable for describing the POD as a function of the logarithm (i.e., log logistic POD) of the depth ratio. Since detection of the deep indications is important for the Farley-1 analyses, the log-logistic function, which results in lower PODS at deep depths than the logistic function, is not used for the POD function in single cycle analyses where the BOC distribution is a function of(1/ POD -1). The log-logistic function leads to unrealistic, non-detection for the large indications. However, this conservatism for the log-logistic can be accepted in the multi-cycle model of Section 7 and the log-logistic is applied in this section.
3)
The Fr6chet distribution of maxima is limited at the lower extreme and unbounded at the upper extreme. Thus, it too was judged as being not directly suitable for describing the POD as a function of either the depth ratio or the logarithm of the depth ratio. Alternatively, the distribution of minima is limited at the upper extreme and unbounded at the lower extreme.
Using the log of the depth instead of the depth would result in a distribution form bounded at both extremes. However, the lower bound would be restricted to zero depth and in practice the POD may be zero for sore small value of the depth, say 5%. Based on these considerations, the Fr6chet distribution was not considered for the Farley-1 POD.
4)
The Weibull distribution of maxima is unbounded at the lower extreme and bounded at the upper extreme. This distribution was treated as potentially suitable if the logarithm of the i
depth ratio was used to achieve a POD of zero corresponding to a depth ratio of zero. The j
distribution of minima is bounded at the lower extreme and unbounded at the upper extreme and was considered unsuitable for use. The Weibull function did not appear to offer any advantages for the Farley-1 POD and was not further evaluated.
{
5)
The distributions described by Kunin for extreme values bounded at the lower and upper end, i.e., Type IV extreme value distributions, were examined. These last distributions have the advantage of allowing consideration of a non-zero depth value for which the POD is zero and a non-infinite depth value for which the POD is one. All of the other distributions are either not bounded or bounded only at the lower end.
The conclusion is that the Kunin distributions appear to be the most desiracle from qualitative considerations and the logistic function appears to be acceptable for Farley-1 single cycle analyses and the log-logistic appears acceptable for multi-cycle analyses. The Kunin minima and maxima distributions and the logistic and log-logistic distributions are evaluated below to obtain the Farley-1 POD for freespan ODSCC. Selection of the POD for Farley-1 applications is based on consideration of the distribution with the smallest uncertainties for the Farley data.
ev-nuw. # ir o4 sir ir oa so.,_ouzim 5-3
F The logistic cumulative distribution function is symmetric about some mean value of the independent variable x. The cumulative probability of x, i.e., the probability of detection, is given by the following
- equation, P= 1 + e* +"2
- An examination of the equation shows that P-+0 as x-+- = and P-+1 as x-+ =. This means that a non-zero probability of detection exists for locations with no indications, but may be acceptably small and would be conservative in simulation evaluations even if zero depth was not considered per se. At the upper extreme, the use of the Logistic distribution conservatively results in a finite probability of non-detection for 100% throughwall indications. The log-logistic distribution utilizes the log of the depth variable in the above expression. This change leads to a POD approaching zero as depth approaches zero but increases the probability of non-detection for 100% throughwall indications.
Using the Kunin Type IV distributions, the cumulative distributions of the probability of detection can be written as,
' B-h '*1
~
' h-A '*1 P = e ~ <h-A >
and P = 1-e B-h, depending on whether the form for maxima or minima is used respectively. Here, h is the ratio of the depth of the crack to the thickness of the tube, A is the lower bound depth for which the POD is zero, e.g., this may be set to zero or some relatively small positive depth, say,5%, and B is the upper bound depth for which the POD becomes unity, e.g.,60%. It is assumed that these values are given or suppoited by other considerations. The parameters of the distribution are ao nd af. By making the a
transformations z equals the common logarithm of the fraction term, and bo is the common logarithm of ao, where b equals ai, the cumulative probability of the Kunin distributions for maxima and minima, i
respectively, may be written as, P = e-'
and 1 - P = e-'*
The maximum likelihood estimate of the parameters bo and b can be obtained from an iteratively re-l i
weighted least squares analysis because the distribution conforms to the requirements for a generalized linear model (GLM). Generalized linear model regression analyses were performed considering the cumulative probability of the logistic and Kunin distributions as the probability of detection for the l
crack characteristics ofinterest. Because the Kunin distribution evaluated was considered to be bounded by values greater than zero at each end, the solution obtained using the logarithm of the depth ratio retums the same PODS as the solution using the depth ratio directly. Finally,it is noted that the parameters ao nd ai must both be positive for the Kunin function to represent a cumulative probability a
for maxima, however, the same is not true to use the function as a representation of the probability of detection. It is possible for the function shape to be such that the maximum bound is approached with increasing slope as opposed to the decreasing slope usually associated with a asymptotic cumulative distribution function.
Calculation of the POD As noted above, the logistic, log-logistic and Kunin distributions are further evaluated for selection of the POD for Farley-1 freespan ODSCC. The log-logistic distribution is considered only for the POD as om:uomum. mar., owar., onu-ouum 5-4
a function of average depth for use in multi-cycle analyses. Comparisons of uncenainties between the different functions are used to select the POD for tube integrity analyses. The selections of the depth values for PODS of 0.0 and 1.0 for the Kunin function are based on general trends for the POD shape while trying to minimize the uncenainty in the POD. Alternate depth values for POD = 0 and POD = 1 were examined. No difference in the best fit function was found for POD = 0 at 0% or 5% for average depth and 0% to 10% for maximum depth. The larger 5% and 10% values for average and maximum depth, respectively, were used for the Kunin function. For POD = 1, the range of 65% to 100% average depth and 70% to 100% maximum depth were examined. No significant differences in the standard error were found for the full range examined for POD = 1 and 70% average depth, 75% maximum depth were selected for the Kunin PODS. The logistic function does not permit selection of the POD intercepts at 0 or 1 and provides only a single option for a given data set. The resulting logistic and Kunin POD functions are shown in Figure 5-1 for average depth and Figure 5-2 for maximum depth.
The following table summarizes the resulting uncertainties for the logistic, log-logistic (only average depth POD evaluated), Kunin maxima and Kunin minima functions where Pearson standard deviation (near 1.0 desired) and mean square error (smallest value desired) are used to estimate the uncertainties.
Uncertainty Parameter Logistic Log-Logistic Kunin Maxima Kunin Minima Average Depth o Pearson standard dev.
1.003 1.007 1.055 1.034 Mean square error 1.046 1.038 1.191 1.213 o
M ximum Depth o Pearson standard dev.
0.996 NA 1.033 1.013 Mean square error 1.036 NA 1.121 1.133 o
From the above, it is seen that the logistic or log-logistic distribution is the preferred POD function for the Farley-1 freespan data based upon the smaller uncertainties tnan obtained for the Kunin distributions. The log-logistic is unacceptable for single cycle analyses due to the extreme conservatism ofleaving a large indication in service due to the unrealistically low POD for large indications. The single cycle analyses divide the number ofindications by POD to determine the number of undetected indications at the same size (length, depth) as the detected indication. For multi-cycle analyses which are totally statistical in methodology, the impact of the low POD for large indications is not as important j
and the log-logistic POD can be used for this methodology (Section 7). In the multi-cycle analyses, l
lengths are randomly assigned to indications and an inspection at each outage is simulated for each SG sample to identify the undetected or hidden population. With this methodology, the occurrence of long and deep flaws in the undetected population is a lower frequency occurrence than for the single cycle analysis with a direct 1/ POD adjustment for undetected indications. The POD below about 25% average depth has little influence on the Farley-1 single cycle, tube integrity analyses as these indications cannot grow sufficiently to challenge tube integrity at the end of the operating cycle. From Figure 5-1,it is seen that the logistic and log-logistic PODS for average depth are higher than obtained from the Kunin functions between about 30% and 65% average depth and a few percent lower above 65%. It appears that the steeper S-shape permitted by the logistic permits a better fit to the data with lower uncenainties in the range of the destructive exam data from 25% to 60% depth. The logistic and log-logistic functions show the expected differences for very small and very large depths. The logistic function shows a POD < 1 up to 100% depth compared to the expected value of unity closer to 70% average depth and the log-logistic has an even lower POD above 70% depth. Figure 5-2 shows that the differences between the logistic and Kunin distributions are smaller for maximum depth than for average depth in Figure 5-1. The differences between the logistic and Kunin functions for maximum depth are QE*BEINT\\ALA\\Als 90Fmal Feupon OA\\FIFmal Freespen OA 5 duc - mm 5-5
not as large as found for average depth but the logistic function continues to have smaller uncertainties as shown above.
Based on the smaller uncenainties for the logistic distribution and implied better fit to the destructive exam data, the logistic function is selected for the Parley-1 freespan ODSCC POD 3 for both average and maximum depth. The logistic PODS are used in the tube integrity analyses of Sections 8 (deterministic analyses) and 9 (probabilistic analyses) of this repon. For the multi-cycle analyses of Section 7, both the logistic and log-logistic PODS are used and it is shown that the log-logistic leads to improved predictions of the prior indication history for Farley-l.
5.2 NDE Uncertainty Development NDE + Point depth profile sizing is used to size the EOC-15 indications, which are used to define the undetected indications at BOC-16 for the tube integrity analyses, which follow the GL 95-05 formulation for BOC indications. For Farley SG indications with less than 0.5 + Point maximum volts, the correlation of crack depth versus bobbin voltage is used rather than the + Point profiles due to the difficulty in sizing the small indications with low signal to noise ratios. The correlation of bobbin voltage to crack depth is specific to the Farley pulled tubes and may not be applicable for other plants.
For tube integrity analyses, it is necessary to define the combined + Point and correlation sizing uncenainties for axial ODSCC. Since sizing of axial ODSCC profiles has not been Appendix H qualified, a study was performed to obtain the NDE uncenainties. Pulled tube data for Farley-1 freespan (above sludge pile) and sludge pile (includes Plant P-1 data) locations are used for the destructive exam deph profiles. It is expected that the + Point sizing uncenainties would be approximately the same for each of these ODSCC locations and the combined data provides an adequate pulled tube database for the sizing assessment. The Parley-1 freespan ODSCC data of Tables 5-1 to 5-3 are the principal database supplemented by sludge pile pulled tubes that have field + Point inspections. All indications were analyzed for + Point profiles using the same techniques used for the Farley-1 analyses. As noted, the depth to bobbin voltage correlation of Section 5.3 is used for freespan indications with less than 0.5 +
Point maximum volts. This correlation does not include sludge pile indications and is not applied to the sludge pile data. Since the tube integrity analyses are based on the effective crack length that results in I
minimum burst pressure, the NDE and destructive exam profiles were analyzed to determine the effective crack lengths and associated average depths to develop the NDE sizing uncenainties.
1 Appendix H recommends that the destructive exam profiles be averaged over the NDE coil field length which is about 0.16 inch for the + Point coil. For conservatism in defining the maximum depth, the destructive exam profiles were averaged over a length of 0.1 inch. For pulled tubes that had the uncorroded ligament areas sized in the destructive examinations (all Farley-1 freespan ODSCC), the depth profiles are corrected for the uncorroded ligament areas prior to comparisons with the NDE results.
Table 5-4 shows the NDE and destructive exam data used for the NDE sizing uncenainties. The tube identification and type of degradation (freespan - FS or sludge pile - SP) are given in the table. The NDE and destructive exam data are burst adjusted data. The burst adjusted data represent the lengths, average depths and maximum depths for the effective length resulting in the lowest burst pressure. The burst adjusted data are obtained from a search for the lowest burst pressure over the total length of the indication. NDE uncenainties are defined for the burst adjusted data.
At the end of Table 5-4, the NDE uncenainties in terms of mean error, standard deviation and the upper 95% confidence value for the burst adjusted data are defined for length, average depth and maximum j
depth. Figures 5-3 to 5-4 show the trends of destructive exam vs. NDE for average and maximum i
depth, respectively. Acceptable agreement between destructive exam and NDE data are obtained for QTTunstwnALAwn 9s$ mal Finaapan oaf FmalFaspan oA4h ocim 5-6
i aver ge and maximum depths, while the length uncertainty is not as good. Since the analyses are based j
on using EOC Irngths for larger indications, the trend to underpredict the lengths of shallow indications l
does not significantly impact the accuracy of the tube integrity analyses.
The NDE sizing uncertainties obtained from this evaluation for use in the tube integrity analyses are:
Average depth: Mean error of 0.29%, standard deviation of 7.30%
o l
Maximum depth: Mean error of-2.07%, standard deviation of 9.4%
o l
53 Correlation of Bobbin Volts with Maximum and Average Depth i
To develop crack growth rates in average and maximum depth, the crack depths at EOC-15 and EOC-14 are required. NDE depths from + Point sizing are available only at EOC-15 since the indications were not detected by bobbin and were not + Point inspected at EOC-14. EOC-14 bobbin voltages for the EOC-15 indications were obtained from reanalysis of the EOC-14 bobbin data. To obtain the growth rates in terms of crack depth, relations between bobbin volts and average / maximum depth are required.
These relations were developed by correlating the Farley-1 freespan ODSCC destructive exam data of l
Tables 5-lb to 5-3b with the bobbin voltages given in the tables for each destructive exam location.
Burst adjusted average depths are used for the correlations as the tube integrity analyses are based on burst analyses from the crack profiles.
Figures 5-5 and 5-6 show the resulting correlations of average depth and maximum depth with bobbin 2
coil voltage. Both correlations show high R values indicating the adequacy of the correlations. The p-l l
values for these correlations are much smaller than the 5% NRC guideline for use of a correlation. As noted in Section 5.1, the bobbin voltages used in the correlations were obtained from multiple bobbin analyses for the indications. In most cases, the field bobbin voltages were used. When one of the other analyses provided a voltage closer to the maximum depth from destructive exam, the other analysis was l
l used. Although not used to develop the correlation, the maximum depth plot of Figure 5-6 shows data points from in situ test results for indications which leaked and indications which did not leak. The correlation predicts throughwall penetration at about 5.3 volts, which exceeds the < 5 volt indications which did not leak and is less than the indications which leaked. Thus, the correlation developed from the pulled tube data is consistent and conservative relative to predicting deep indications as well as consistent with the dominantly more shallow pulled tube indications.
Ov:rall, the good correlations obtained support the adequacy of the correlations for estimating crack depths from bobbin voltages for Farley SGs. The correlations are inherently tied to the Farley SG freespan degradation, as found in the pulled tubes, and are not generic for applications other than the Farley SGs The correlations then permit development of crack growth in depth from the EOC-15 +
Point data and the EOC-14 bobbin voltages. As shown in Section 5.2 above, the + Point depths for indications with < 0.5 + Point maximum volts show higher uncertainties in depth sizing than the larger l
indications due to the low signal to noise ratio. For this reason, the depth / volt correlations are used to l
size all indications with < 0.5 + Point maximum volts for the EOC-15 growth analyses and for the EOC-15 indications used in the tube integrity analyses. This selection provides consistency in defm' ing the depths for small indications for both the growth and tube integrity analyses. This consistency in splitting the total EOC indication size between missed indication size and growth is essential to minimizing uncertainty in the tube integrity calculatioas. If the missed indication size is underestimated, the calculated growth increases to compensate the error in the missed indication size. Similarly, if the missed indication size is overestimated, the growth rate decreases for error compensation.
Q4TUREIMRALA\\Ala 95fmal Frecapan oaf 1Fmal Freespaa OAddoc - DU2H99 5-7
5.4 N:w Indic-tion Average Depth and Maximum to Average Depth Ratio Distributions In the probabilistic tube integrity analyses, new indications must be included such that the analyses j
account for the expected total number of new indications. In the Monte Carlo analyses, the new indications sizes at BOC are obtained by sampling crack length, average depth and maximum depth distributions. New indications must account for indications below the minimum detection threshold (not accounted for by POD adjustments to the measured EOC-15 crack distributions for undetected indications) and truly new indications. In these analyses, it is appropriate to obtain the best estimates for average depth distributions and for the ratios of maximum to average depths. The new indications would have shallow depths below the smallest detected indications. Figure 5-5 indicates that these indications would have average depths less than about 25% (smallest pulled tube depth) which conesponds to less than about 0.1 bobbin volt indications, the minimum detection threshold. The new indication, average depth distribution can be represented as a triangular distribution (results in s-shaped cumulative probability distribution) between average depths of 5% and 25%. The resulting cumulative probability distribution is shown in Figure 5-7. In Monte Carlo analyses, this distribution is sampled to obtain an average depth for the new indication.
The ratio of maximum to average depth can best be obtained from the Farley-1 pulled tube data. Since this distribution applies to the shallow new indications, the pulled tube data for < 60% average depth is used to define the conelation. However, the correlation does not significantly change if data above 60%
average depth are included in the correlation. The resulting correlation from the pulled tube data is shown in Figure 5-8. Following sampling of Figure 5-7 to obtain the average depth, the correlation of Figure 5-8, including uncertainties, is sampled to obtain the maximum depth. The crack length distribution for the new indications is developed in Secdon 5.7.
Figure 5-8 is obtained using only the Farley-1 pulled tube data. This ratio of maximum to average depth can be compared with that obtained from the maximum and average depth distributions developed below in Section 5.6. The ratio from the depth distributions of all indications is shown in Figure 5-9.
While there is more spread in the NDE developed ratios of Figure 5-9 than found in the pulled tube data of Figure 5-8, the regression fits to the data differ by < 3%. This agreement supports the general consistency in the maximum to average depth ratio, particularly when the burst effective average depth is used as for the current analysis.
5.5 EOC-15 Crack Profiles for Probabilistic Analyses The EOC-15 crack depth profiles for the probabilistic analyses for SLB burst probability are obtained from + Point analyses of the indications found in the inspection. The + Point profiles provide depth versus length distributions. Due to the uncertainty in sizing the small indications, the profiles for indications with maximum voltage less than 0.5 volts were normalized to the depth from the average depth versus bobbin voltage correlation of Figure 5-5 for structural analyses or to the depth from the maximum depth versus bobbin voltage correlation of Figure 5-6 for leakage analyses. All depths in the profile are multiplied by the ratio of the average or maximum depth from the correlation to the average or maximum depth from the original + Point profile.
QM'BEWT\\ALAWs 9ff mal fwespen oaf 1Fmal Freespan O A-5. doc - W21m 5-8
16 EOC-15 Cr ck Depth Distributions f:r Deterministic Analyses The EOC-15 average and maximum depth profiles for the deterministic analyses (Monte Carlo analyses for EOC-16 depth distributions) are obtained from the + Point profiles analyses. In this case, the average and maximum depths for indications with less than 0.5 maximum + Point voltage are obtained from the depth versus bobbin voltage correlations of Figures 5-6 and 5-7. For pulled tubes R25C51 and R6C10, the destructive exam depth profiles are used rather than the + Point data for the depth profiles.
The resulting average and maximum depth distributions are shown in Figures 5-10 and 5-11 for SGs B and C, respectively. The burst effective average depths are used for these distributions to better. reflect the burst capability of the indication when used in conjunction with a bounding burst effective length in Section 8. From these figures,it would be expected that SG B would be the more limiting SG for Cycle 16 operation, which is shown to be the case in the tube integrity analyses of Sections 8 and 9.
Crack shapes for the new indications are sampled from five Farley-1 destructive exam crack profiles that vary in shape from a single peak to multiple peaks to a fairly uniform profile.
5.7 Crack Length Distributions for New Indications in Probabilistic Analyses Since growth data for crack length is not available from the inspection results, the crack length distribution used for new indications is the EOC-15 crack distribution from the + Point analysis. No growth is then applied to the crack lengths. The EOC-15 crack lengths are conservative for EOC-16 since Cycle 16 has a shorter operating cycle than Cycle 15. The crack length distribution is shown in Figure 5-12 as developed for the combined data from SGs B and C. Separate distributions are shown for the total NDE crack length and for the burst effective crack lengths. The shorter lengths for the burst effective lengths are expected since they represent the part of the total crack length that results in the lowest burst pressure. When crack depth profiles are used for the tube integrity analyses and the BOC J
profiles are grown in depth over the cycle (probabilistic analyses of Section 9), the total crack length distribudon is used for the new indications since the length contributing to burst can change with the additional crack growth in depth. When average depths (Section 7) are used for the tube integrity analyses, the appropriate length distribution is the burst effective length.
01WBEINTMLA%is 9thhal Fmespan OA71Fmal Fearspan OA-3 doc - M21m 5-9
no iS ev e
l aT v
t t
io o
oI n
cT bI e
5 "2
c i
9 bP o e o
aP 1
a Lv
" S 0
9 6
S Ph 1
t o
- 8. T "6. T
+
2 3
4 t
1 sb o
r 4
t t uA 2
r 1
od B
h e s ra yp l
Mm i c x
1 e m a
a P
v o t
a S
t T
a D
v e f e o v e
r v 5
2 1
t o
A 8
v ei t r
oa N
3 4
I.
8 o
s a s
nb
/
nv 9
6 b
or 1
9 i
2 3
4 a
ce u
tah s
B vp "6
t r
n e
e e 5
v o d
c lED 2
o n
s si a
h
- e. e r
d b u h) t 8*
4 *2 5 *4 4 *9 pe us m
gh A
t 1
eli t
s 9
e
- 0. 1 64 I.
1 c
/
5 0. 2 f p h p 1
l
. 03 d p e r a
d n n N 2
0 ei x
w e
l 0
0 0
0 oa I
L(
t t
E a
t F
e gg ms c
.e l
r v a ou e
r t
r
.h frb ar v
s o
gt A
e
- 8. 2 52 84 0. 7 64 gv t
u C vp
/
m t
i r
na u
Ae N
9 9
4 7
cu 1
c B
D 2 3 24 2 3 5
2 8 ng nm i
t 1
u n
3 3
3 2
2 nn r
e ui oi n m
r n it im t
n c
s a
.h n
xt A
u e
e g
55
- 0. 0 77 66
- 2. 2 or s n ap
/
D L
Me N 00 88 99 1 1 55 d e t i i
s D
44 33 33 33 33 eh s
f et s t t l o
at b a e u h
h s y
e r
og A
kt t
r i
/
4 2
1 i
1 c
r aa NL N
a o ot a r s f
l m
h c".
rt o1 e
5 m atgh A
8 4
5 4
h) c0 b h r
l t
a mg 1
e u on n t
c
/
0 6
I.
0 0
mn u n lbS Tei N
2 0
a e
t h nl 0
1 l
l a:
L(
st 1
ml r
a TB u e 5
r at
.h b o x
o gt A
t G
vp
/
9 9
9 3
0 nm ef i
eo S
Ae N
9 9
4 7
2 ao v
D 2
2 2
1 2
es ih rl t t 8
a a c g t s un 2
.h ni tr e laxt A
C cap
/
eh sl 8
0 0
0 0
mtg di el a 6
oMe N
7 5
4 8
9 L
D 4
4 4
3 4
an st R
ige r
i a
l l
r p e
d d
e b
e e
dl b -
r 8
0 3
2 5
ee di mt t t u
asu 2
1 4
5 9
f s
l r s o
u r T
u us 4,
8, 7,
9, 4,
r e nu r v b
cBit 0
0 0
1 1
oi d
d l
1 1
1 1
1 ct r
a P
n c n o m
e C
e of i
l uf c
l m._
f u
r eh P
e o e st fl
,p r
0 0
0 0
0 0
u )g 0
0 0
0 0
0 d o n e i
t 7
r si ec ed s
9 us s 5,
4, 7,
4, 7,
8, c
ve s
n g
B r (p i
e 1
0 0
0 0
0 ui g
o 9
db a
1 1
1 1
1 1
1 P
eb. r r s ee r o eb v 1
sb b ma i
)
e h
eo ud y
e 1
t v
ph r w
P o
eTpn n
e n
7 5
S b
d v
a e
5 2
2 T
aI
.h sh lr o e t
i l 6
t r
a aS A
o 4
P g"
oig o
t t
"0 S v1 b
/
t o
I f
F ve N
o o
t w
n "5
T a0r s
t 1
i e ev t
6 7"
"5 o
l l
o df f r Eo 2
2 no ek b
3 t
a 4
a sb c i.
1 hh a
mr r.
(
1 mt t
gguc unn nl n
mlee a
k c
)
l 8
ixdd ci e
e l
2 m o
B A
B A
t ai w
- al Bt A
l r
e er r c
5 n 3
4, 5,
3 4
Ci eMfiiCC m
s i
f 6 e o
2 3
4 5
t w
Rp c
1 o
m
(
S N1 23 o
79t
',i lt t
t t
t loo o
l t t t
t n
l l
l l
o o
o o
o l
l o
4 o v
v v
v v
vv v
v D
D D D
1 P 6
9 5
5 09 8
5 1
1 1
2 D 22 D D a
R+
2 1
D 1
1 t
1 0
0 0
0 0 N 00 N N 0 0
N aD
',bt nn 8 bio l
i t
l s
a 9
u t
5Bl o
o l
l "t
loo o
o r C a
o u t
t t
t t
l l
l lt t
t l
l o
o o
B n
1 a
v v
v v
v vv v
v r
Rldv D
D D D e 3 9
6 7
5 02 3
9 a
E r
1 e 1
D
- 3. 2 D D D e 3 1
1 2
1 1
i u
FR 0
0 N 0 0
0 00 N N N 0 0
D C N
y d
7 d
d 9
n b
o o
o l
n E
',ldi t
t t
t l
l l
o a
t 4 eb v
v v
v k
1 i
D D D
D D D D D
Fo 9
m e
R B
2 D D 7 0
5 i
1 3
D 3 D D D D D
a F 1
0 N N 0 0 N 0 N N N N N
xE l
l l
l e
' nd t
t t
t oe o
o t
v o
o l
"D t it c
2 v 2 v 2 o 2 v 2 v i
ce
/
/
/v
/
/
tc Orat 2 7 25 0
1 3 1 9 7
e 1
1 1
1 u
P FD 0
0 1
0 0
r 0
t s
k e
s c
D m
n ar u
io.
c s
f t
m al s
t t
o uo
. r i
b x
v e u l
y ah vna a
pb 1
t r
Mp 5
0 7
6 2
b ah eit e 5 a
/e dbt e
0 8
9 1
5 em m
t gD leoa n 4
3 3
3 3
a
/
/
/
/
/
bd o a
f l
x r
i bm hd h E
e 2
2 4
7 4
1 S
cu v
t l et 1
I i p
TS e
3 4
3 5
8 Rwf e 5
i v
A 3
3 3
2 2
h ed I
i t
h t
c d pt m B
u n ef u n ).
r t
on ad om G
4 k ni s
i i S
e t (a 0
0 6
c ox 1
D v $
4 8
8 3
Rai a I
rt P
c a m 8
1 1
8 e
n 9
7 1
2 E T
- 0. S yeluf l
1 2
3 4
bt a o 3
1 C
kv o
e o
o o
T a
t t
t o
nl vt 6
c o 6
5 2
t e ioee n t
a 1
6 3
7 v T ei R
rb t r r o o
1 8
6
- 0. b co p
C A ecy e
9 2
3 4
2 a t
r e b
eooht dt a t
u B B t
f d r o s
oeot T
s A
3 3
B A A A B sbt m
ru mua A
3 s
d A A 3
c c 3 4
4 4
4 b
a l
e e
3 3
e e
e e
e e
o s ao l m.
ue s
I e
e umo c
p p
c c
c c c n
s c
s s
l e
c e
p p
p p
p p
l l
c v
c t
r u
o e
e p
s s
s s
s s
s s
n vf h
t t
1 e
P b
p p
s t
r r
t t
t t
t P
esd g 5
s u
u s
s s
s s
s t s
s t
r r
r r
r S
ene a A
a s
r t
t b
b t
7 n
s s
r u
u u
u u
u T
r enl O
r r
u b n
n b b
b b
b psi o e
9 o
u u b o
o e
eet v 9
a s
9 i
n n
n n
n n v
r r
t b
b pb n w
o o o o o oA d
o o l
1 a
n o e
T 5
3 v
n n
o "
b 5 e r ei i
2 7" "9 "0 1
e o
o "
4 8
5 a
8 a
t t
cnga W
a l
2 5
2 3
3 4
4 y
E "4 9 "4
c e ea 1
1 p
e vl ic 1
e t
t W
3 7
8 3
6 8
9 7
d A
e 3
8 9
o i
2 3
3 3
4 2 s d g n
O 1
1 1
l vi n
r n e e
io gi i 9
nn a
n abb w
F 8 e t
t r
cl bb 2 m
- r a ooo Ci 2
3 4
5 eFVBB s
I 6 ce t
Rp o
m S
N1 234 i.
w w
m me
n t oee si v v ye rt oe 4
5 "6
v a
6 l
u cb e BOAS a
le ei 0
7 vt 1
1 l
i i
L
.t v
f a r o
a n v e r
t o r s p
a i e n ta s o m
D leon c a
e 2
f h r
s x
ts ot v 9
5 r ci e
r npe 6
7 ord e
ce e
u o el 7
0 7
6 4
5 6
vi e v
5 9
e DS 5
B a
.e 0
5 8
o 2
8 3
5 8
s ol t
i i
o t
p c
t t
8 5
6 6
5 us p u
2 5
vxv 0
9 oh 5
1 1
1 1
a r
d eao 3
2 h
t t
n EMb I.
7 n e g e
gpe s
l a
A 0
i d a d
t r
sf e
e e
E eo v r
h h
Wea u
t D
8 s
r 9 *5 3 *7 4 "6 0 *7 4 *4 6 *6 6 *6 3 *3 5 *9 g g s
o ga N
g n r n e
f 1
d 2
55 52 37 27 77 65 32 3. 3 38 i v e
r ei p
d e
0 0
0 n
0 0
0 0
1 1
1 s
0 0
0 0
0 0
0 0
0 ua n t
m w
c u
c t
t sr a
a e
d gr u
e l
r nn e b
p ai x F r
x nh m
e o
t E
s n
r v
0 04 7. %
ts C n
t t
ua u
a
- 4. 1 27 00 52 7 22 53 o
1 r h m
h e
0 1 6 2
5 0
4 0
3 i nt t
v n
A 0
u 1
t t
2 9 2 4 7 2 0 8 3 25 5 2 5 9 2 0 4 3 c
n i
B e
9 5
3 9
2 5
5 2
5 e oo imh.d i
m rd r
s t
n c
o oe".
a t
u g
s1 ng y c b 0 n e a
r i
i s eb.
t t
L a
5 00 33
- 0. 0 00 66
- 9. 9 4. 4 5. 5 44 n
1 1 44 88 66 88 ek n t l s
3 00 88 99 00 e
l s
a ut r e
2 66 33 33 66 66 22 66 mcah s r 1 1 1 1 s un D
s o a r t e u si 9
igce r
b et f
og 1
2 or t
a 1
2 2
4 2
3 5
r l
i 1
1 1
1 r o at o
NL h c npu m
h l
ay t
a t i t a
os v iwmo h Pr e h) 2 r l
8 atgh 9
3 4
0 4
6 6
3 5
s t
u 5 a t
g c
st 1
l on n 2
5 5
3 2
7 6
3 3
3 er a n+b s
s em Tei 0
0 0
t 1
us eh el 0
0 0
0 l
1 1
o i
l t h L(
f bi bm v
l i
o h
a l
t r nt wd s
2 t
pi g oh eu 1
au
.h mee an s s TS gt t
t r
5 f ga vp 3
I.
3 2
7 5
0 7
4 0
arl on e e Ae 0
6 3
8 6
5 7
6 1
8 x a r v d
e c D
2 9
4 2
8 2
5 5
2 4
et hl B
gt inh l
ne t
t e ef s
p i
nr t v
G
.h me e
e ua d xt c a v l bh i
t S
ap 0
0 7
5 6
8 6
4 8
uigi t
d l
2 ams n t
Me 4
7 4
3 7
7 3
7 t
0 0
r l c
D i
s a 1
1 1
e t
a e sdf r
5 e ef a xrD d d e pet C
sO d
ol e
mri. - v pP 5
e e
or o ets o
i 2
or t t r) 0 0
5 r
ocb s t as uis 1
3 1
u us p 3
7 8
d ninouu h o r c f c R
l r
f s
r b
ed r
e cBt(
4 2
6 e ub p t
rb r s v e l
t h oei s b
a a
a oot fd s u s
u C
P luf b o e
m T
lcd ebh gr s
t i
e a eh r pinp iz m7 s
c i
c uTfoe r r
d o
a t
d pok 9
3 sd
.s e
sui
)
e e
c c 5
A A
8 8
e r s s r "1 h gm a
e l
us p 7
r N
N 8
9 u
t h r a
l e
7 2
6 s s ga o gC u
Br(
s 4
sh0 n r ci e
4 P
P o.
et f e rh t v
s pol o
r r
af a 8
pe d
u r
dhl d sdB 9
9 4
7 3
6 6
r e gi nh a t
s t
e e
9 r
v 7
9 5
7 5
0 0
4 7
7 ugnf at h.
i.
g y
1 ne 0
6 5
6 9
7 6
6 6
6 b aelih n el eel ot e v n i
oe) 0 0
4 6
7 8
rl r
1 il 1
1 1
1 r vh c gl a o r
t S h a
c o
o o
o o
o o
o o
o u aeeek n
o.
n h
m y
ven t
t t
t t
t t
t t
t tadfihl cyo e
evi 9
4 2
3 5
6 1
7 3
1 a ai r
t o(
3 0
0 3
7 9
4 0
4 4
e nt k r r,
l l
pa n oc cmta r
Eb 1
1 4
5 6
7 5
6 6
5 i
t a
mmod rd nm a
A 4
0
)
r I
1 1
1 F
e uP e ceor i.
rl nio t
m+a ai t f r
5 mi pcb an m
n t
ci 1
r i mi i.
e o x omiod t
"6 r
Cm 6'
7 o o a o r o
a
)
eRMh cCCInF 0
s m
i 8
9 2
3 "4
45 6
a 1
s t
5 c 1
2 e t
s t
o i
Rp 4 N1 2 3456.
m S
tseno
89
/t s
4 42 1 n t
t t
t t
a 5P v
v v
v 0 o lt 1 i l
l l
l
- 3. l 33 o
o o o
o o
D D
,v v
00 a
m t
D a
t 1
8 7
7 4
62 2
a 62 a
R+
d 6
2 3
3 D
D 43 3
D 43.l t
x o
m e
D 1
6 0
0 5
N N
00 0
N 00 v a
e x
v E
l nn e
i l
t D
a 8 o
e c
i C 9bi v
u N
/
t t
r 1 b a 5
55 icd.
t l
t t
n 1 o u o
lt t
t s
t t
- 8. l t
85.l u n e
v d
e
,Bl "d
- v
,v D l
l l
)
o o
o t
o o
00 o r e d
"d v 0v v
o
,v s e m
n r
S a
l t
r I d v v
a u Re e i
li 2
6 l
n n
83 D 2
93 5
3 933 d e o
4 5
40 e v 4
5 1
1 e
8 5
C i
5 l
r m
FR B
B 0
0 I
e 7
N 0
00 0
0 000 hl f
1 y
t s
E a
d iwm D
x d
d E
E l*
l l
l l
l 1
l 53 vf f
9 05 e o N
t t
t t
t t
r en o
o o
o o
i o
e d Fi "d "d e
0, 00 e e o
b v
v v
v v
v l
D v
e 8b n
i 2
9 7
D 1
4
- 5. l 3 41 l e o io s
l 4
6 0t 5
69t s n n
9 o i
n i
i c
8 l
4 2
2 t
F /B B
B 0 0
l o
o v p N
0 0
0 v 0
00 v s
. r t
1 a
u o
e s r
b r n a a
op nd d
d d
l 6 n a t
l ai e oe lt lt t
s 2'-
t l
t e
e e
t s
o o
o e
it A A 2 v 2 v 2 v 2 o o
l
- 2 U *2 D
"D t s
s s
gu l
U 2 v 7i c
U a
ce
/
/
/
/
/
/ v 2
Ore N N 2 2 29 27 0
/ t
/t
/
2 5 s
s l i
/
a x at t
n
.t ov a 4
8 2
2 2 o 2 o 2 o 5
os f
a o
P FD 0
0 1
N N
N 0
i l t
e w
o t
1 0
es t
y a o a ut t
u vl l
e f d o u.
ra m
a n e e vdk l
i vb gvs c
b m
u eb a eud a 2
m d ol es er t
m t
c b ol r i
l s e c 5
x e
vf vt et u
ah s
f h i
r t e o e
eS Mp 0
3 0
0 6
9 4
5 4
5i t
g h
u d
l b
de
% 0 8
9 0
wrd b
t 1
4 8
6 8
a nps 3
l a:
m gD 4
3 6
6 2
6 Rhl e eae 6
3 1
1 1
d h
T B
/
/
/
/
/
/
/
/
/
I t
2 pnr dt 5
d e igad e n a
/
x ie
% 4 1
1 7
0 2
7 2
3 ndl endi G E v
0 2
7 0
3 5
2 9
0 3
k
)
a a nau h
S e A 2
9 5
3 9
2 5
5 2
5 8 c od Dct l
v np 9 at e 1
t
/ r h sOie i
c uPt d c
8 c 5
u nh ne n r
o
)
n nm C
r i
t i o
t oc a
of u
y 5
s in nl 1 icdleim t
t e
ai 9
oe =
2 D v(
7 4
7 1
r 6
6 r ye e
i R
es 0
9 0
4 7
7 c ob eutivx t
s t
a 6
5 7
0 e c l
i 6
6 6
6 am 0
1 1
1 1
e od iv d s e
E r 0
5 6
0 t
b k e 1
o 4 6
8 7
d t 0
et rf o
o o
o s n7 eo 8
t t
t t
f d ai s
u c v 9
o o
o t
oe e ok nt a o 3
4 t
t t
o e
1 7
n Pc oi 1
I s r T
r 0
2 3
9 t
n 4
0 4
4 muc aCo b
1 C
1 0
3 6
0 o
u e d (+
e r
d A
0 5
6 6
5 1
1 1
1 s u C. p 0
4 5
6 8
N sdt ne e
e l
s a n es oh n
l tn vol ri t e
l i
t u
es i pu ao e
e t
s s t a pB c t P
v e n va
.d s
i t ne r e 8
oe) 0 4
8 8
pslet yne 3
e e e ol i s il 9
t S h 0
6 6
6 6
r r nnt o n
a c
1 7
0 0
0 7
pn osl o
9 ven 0
1 1
I 1
1 d
0 5
7 8
8 1
e o n ec 1
ev(
0 o
o o
o e ri ions i
e o
o o o
o o
o l
t t
oo g t
t t
t 1
Eb t
c nat t
t t
t t
t l
0 0
4 0
e e c ci t c a y
A 1
1 5
0 5
0 7
5 6
6 5
e id vi ea t
t w
0 sl te mi o r
e 0
3 5
6 8
8 1
1 1
1 d gn id r t v
r m
l n e ol n
o gi innf on iabbinvi o
a 1
)
e b
t F
5 6t k ct bb r4b Cm 0
l a
os c r
r oooo1 o 8
9 2
3 i
I a
s 4
5 6
4 u eFVBBF0B t
5 c 1
r 2 e
( BC t
w Rp 7
o m.
S N1 234567 m
snw e
mo
nr s
S i
F e
oo e
e e
e e
e e
T v
v v
v v
v v
+h i
e t
I aS v
t o
o o
o e
e oI o
o o
t cTP o
oI v
v bS bS bs b s bP aP aP aP ot bI bI bI r
S o
o aF aF a
a i
i a
L 'e T S " S S " S hd
- 7. T
- 6. T ".i
".I 2. T 8. T 2". T 5. T b
b s e 4
r t v a
a 1
ya s o 6
2 7
2 1
r 1
1 l
p ub 6
1 2
2 3
i 2
4 e
8 1
v m B A 2
9 i
o ta c vr e I
a xS e v S
S s
S s
S s
S P
P P
si I
I I
P M 'T i
F i
F i
a F
F t
t F
S n a T
T T
T T
a S
S S
T T
T T
T nc r I
I i
.ov D
e e
e e
e e
e vP e
e v
v v
v v
e e
e v
e or e.
fo t
oS v
v v
v v
v s
s nbT o
o o
o o
o o
o o
o b
o a v n e
i o
t b
b b
b b
b b
b a
b o
b b
o r
oa r a
a a
a a
a a
a le c
i a
a a
r s u
ah o "8
"8 "6
"6 t
s "9
7" "4
"9 rhi 8
B vp 0
1 t
ot 6
6 3
5 9
4 9
5 2
c pd e e 6
6 6
2 2
e ie d
1 6
1 1
t l
3 2
sdl ED 3
9 2
2 8
1 2
4 1
1 1
n r
f p
.e u
a b op r
a u
h) m "*
m
". *6
". *4 e e m
7 "9 0 *9 "4 "5 s
"6 "2 "
4" 7"
2" ge t
h s
s 9 "9 7" I "
g u a g e
9" 1
c L(
0 0
0 0
0 0
- 0. 47 7. 6 68 5
a 7
8 1
6 x
w d n n 77 6 92 2
66 67 43 9
9 h v e 7
r a r
8 o e r p
1 1
ei 2 0 0
e 1
1 1
1 1
1 1
1 1
1 0
0 0
0 ga v E
a t
ts l
c F
e n
a r
g e
r i n g u
t t
r n
b s
v s
i r
o
.h e inin m
t u C vp 80 09 79
- 2. %
gt 7 59 Wn 9 1 5 96 30 0. 8 20 52 1
u n u
c B u
n Ae 0
0 6
5 2
0 4
5 0
7 0
5 t
r u m
t D 3 5 5 3 2 2 2 2 3 5 3 6 21 23 4 4 3 3 3 6 3 5 s
r r
e 3
5 4
4 3
4 3
4 5
4 3
4 e n i
t o e n
t r.
a h
i s
m e
l d t e
a get
.h isb t g
xt n s a n
D i
1
- 2. 2 22 ah L
ap
- 0. 0 77
- 4. 4 44 33
- 0. 0 22 1
99 i
1 1
f Me s
55 00 55 55 1 1 99 99 1 1 44 00 44 33 u
o o
D 44 77 55 55 44 55 33 55 66 55 44 55 d c u
t k
l s
sn na".
s y
a r1 e o r
og 8
3 3
4 8
7 2
2 5
4 1
3 s o0 c
ri t
aa NL 2
2 4
2 t
a i
1 3
5 4
3 2
4 2
nr au 3 m ocn hl i a a t
a t
5 m h) c mh h v l
e at h "0
4 1
t e
g 7
9 t
2 4
7 2
9 r t e
gs e u on n 7 1
9 1
7 6
6 4
4 7
6 8
or o 7
1 c
t r s r nt lbS Tei l
0 c u e o 4
2 a:
L(
b m 1
l 1
1 1
1 1
1 1
1 l
v 1
t l
TC nno a s 5
t ei s o u
.h mal a t
s gt G
vp 6
5 9
6 7
4 0
9 8
6 0
1 a e f er r s S
Ae 2
1 8
6 3
3 6
6 4
8 3
6 igai ov D
3 5
2 2
3 3
2 2
4 3
3 3
l t h h h t
0 h nt t
e g gtp n
n e w me i
1
.h axt l
ed al C
cap l
0 s gd d
5 0
3 3
3 8
7 l
0 4
0 7
ei a n 6
oMe 5
8 6
6 4
7 4
3 7
5 5
5 i di l
l l L
D 5
e i a R
t f
r f
oe aD e
rd e p
b d
pov O
P e
e ri 1
2 9
6 mr t r 0
6 9
6 2
5 t
u t
t 0
4 0
7 2
a o c s r as u 8
T e
r o 9
4 M
0 5
x cf l
r s
8 4
3 4
e nf uf 0
u us 0
0 5
7 0
1 d
cBe 8
9 8
9 9
ue b d 1
1 1
1 1
e a
r e
l l
r e C
P v ri os m
oo f u t
l it f l
c c
2/
h s 8
u ud n P
e r ei pi 0
t 3
r cb e
e t
0 0
0 s
ts u )g 6 0
6 9
9 0
5 ub 5
0 0
A A
7 2
7 0
2
.dk n
8 e
r si 9
0 d
d o 3
0 B r (p 9 6
0 N
N 6
9 4
6 0
5 eb s e c 5
3 d
9 us s 9
6 2
a e
8 9
8 8
9 mr egr A
9 O
1 I
1 s eb aC 1
P oh h o r
ne et g
Tr r
t 1
f pv s e
p r
S s
d d e a
w e
.h u
tT 1
I i
S s
S ed".b y
v)
S d
F T
F I
dB P
P P
P t
s I
i t
1 a e on in e
oP T
T T
i S
S S
I 1
1 g0 a.
F b
l l
a T
T T
T u
S I
r aS T
5 1
1 af rh ly F
6 6
4 1
I.
8
\\
c r T
A a
oot 9
6 5
9 3
4 2
3 3
e
(
6 1
7 a
f gn O
l F
nr 5
o n
0 5
c vh 1
s n oo 4
2 2
3 9
3 4
at 8
a 1
1 p
1 1
s gh e n s
o o
o i
o o
o o
o o
o ed e
l o
w taS o
t t
n gk i t
t t
t t
o r n T
t t
F t
9 t
t e
t eT 7
6 6
4 6
9 5
9 3
6 6
1 1
2 4
3 3
ual n c a t
v t
e s
e a m F'm 1
5 s mldl r l
4 6
1 1
cr m
E 1
5 1
2 2
9 e
2 3
r u ed l o 1
i l 9
1 pmf ieaf a
n t i t f c n lA
\\
i A
0 e s x n it r
l l
ai r
A m
"b 4
5 6
7 8
9 u
o o i o 1
i r
"a 0
eBMPcCF m
Ci 2
3 s
6 c 1
n 1
e 3
3 r
t Mp o. 2 34.
m S
N1.
m o
f
}
)
l
2 4
t l
lt 1
2 lo o t
t l
o o
0, 0,
,v v
n v
v 8
9 0
8 e
1 1
2 52 0
/ t t
5 8
- 2. l 8. 3 2 m
5 2
t t
1 n o
l lo o
0, 0,
1 i i
0, 0
0, v 00 0,
o v
c v
5P 4
4 7
2 6
0 95 96 3
p e
t t
R+
7 1
1 0
O o
l 4
1 2
5
- 2. 3 2
2 s
l 1
o v
v 0
O 0,
0 00 00 0
e k.h 1
t 6
7t 0
7 6
1 1
t 75 39 3
c 1
r 4
- 2. l 5 2
1 0
4 3i
- 4. 3 2. 4 a
o o
r o 1
a 0
0v 0
0 0
0 0
0 v 00 00 0
cf t
t k a
s c r
D a
l t
t lt t
u r l
l l
,o l
b c a
o o
o E C 8 nn v
v 2
9 0 v v
et o
s i
h r 9
2 2
1
- 7. 4 9 t
9bi D
1 tn bt 2
lt t
u
/
2 6
0 Ol 03 1
f 1
a o
b N
e 1 o u 0,
o o
0, 0,
- ,v
, 0, 0, v
r 2
e r
,Bl h h d
u 5 a
0 0
5 1
5 49
- 2. 1 5 t
d v pt t
t 3l 4
lt lt 3
2 6
22 02 1
n C Re e o
l 1
l en a
ie v
0 o o
o
,v 0, 0,
0 0
0, 0 *,0 0 di v
v y
1 FR 9
72 2
0 2
7 D
h 08 39 0
mt d
m d
1
- 4. 3 4 4
4 1
D 3
3l 42
- 2. 3 2 u
3 1 t p
o e
a E 0
00 0
0 0
0 N
0 0 v 00 00 0
md x
i d
x m E
l a
t e 8 n l
m m u
o i
i e F 9b v
2
/
v b
oi
)
6 x
1 1 o 0
t i
(
a t
t t
c
,B 2
l l
l t
0 t
t m
l s
o o
o o
e 0,
5 v
v v
D D
D oo u
1 d D
v sf D
D r
Re D
4 4
4 D
3 D
3l D
D 7
D ct l
6 8
3t 0
I 1
l t
i 1
o n
s F
N 0
0 0
N 0
N 0 v N
N 0
N ni e
oo D
ch p
i t
e
)
)
a e f
nd i
"t l
d d
t "t
l t
"t sd e
e o
l lt l
l i
t t
t o
l
.i oe P
o t
o o
o o
l n
)
y "D t
' s s
l it c Ae 2 o 2 v (2U 2 U 2 v 2 o 2 v 2 v 2 v 2 v 2 o o
n
/
/ v
/
/
/
/
/v i
n o ce
/g
/v
/
1 3
a ioi t
Orat Nd 2
21
/
/
2 8 0
t t
2 2 3 5
r 4
2 3 1
t t
1 1
8 4 2 o o
3 3
6 2
2 ns t
e b a P FD u
aie u
5 2
ub s 0
N N
0 1
0 0
0 0
3 m S
0 0
0 a
l l
b o la v. v ol 5 m e s eBc et g
eu h
rl s. n t
ot l
d vl ei b S p
ope 5
l e
TC ie nv ab c:
D 1
f i t
b s n
a m
Sb ut o 5
os a p I
G u
Rb rd u S
m h vdd e
I i
t l
e x
dih e s nwtpf a i
a a
0
'mM 0
7 4
4 3
0 2
1 9
I.
2 2
l dh y b
h e e 1
dt a
/
C x
e 5
0 5
5 1
6 E
g 4
7 5
5 4
5 3
5 6
5 4
5 ieped f 9
9 1
4 0
4 3
t l l
f d n a
/
/
/
/
/
/
/
/
/
/
/
/
o c a
R e
r 5 k a i
np v
e 0
9 9
9 5
6 0
8 0
2 7
9 cDoy I
e v
i RaOi t
5 8
2 2
5 6
t 1
3 4
3 6
5 b
c A 3
5 4
4 3
4 3
4 5
4 3
4 cP a n r
t I
u u
u r
yerl o t
bt oai T
t s
af va e
n nlo ed el d
D o r i r e
e r e r r r
o) e i
tcosuyo im tash 5
8 9
3 8
2 ec l
l r c c
2 6
6 4
3 8
0
- 6. I 2
- 4. I eoi oh v
i r
t u
e s
r n 5 S 2. s 2I e
- 9. I 2 P dt Ti 7
s 7 6
2 s 7 S 3
s t
o P 2 P P
l P
3 at
- 4. F 0
- 9. S E
(
S fd r p 2
2 F 3
1 2
1 1
1 8
S 2 S 4 i
i e
i i
8 k vP 3 T
oT T
oT oek oe c
1 T
oT oi i
o i
o o
o 1
n s cbd a
9 coS o
o t
t t
t t
o t
t o
t t
e t
e 6 e 9
6 2 e 8 e e
t e
e t
e 3 e mua a/
-s 9
abT 1
1 rl e A
2 v 2 v 6 v v 2 v 6 v 8 v 9 v v
v g
ueCmt o.
o o
o 4 6
7 o o
o 1
rA o
o 3o 5 o s u a
C
- 7. b b
5b 6
5 0b 6b 1
b 4. b 0b b
0b s a s 1
l 1
t ol 1
a 8 a 1
a 1
1 2 a 2 a 3 a 7 a 1
a 2 a 4 a t
r o 1 -
n vr f v n
ee u
s y
es I
I st Bd r i.
S S
e r
P P
en e o i
1 r e
.nf r
l n
o)
T F
P S
S ps yiad r
a iodh T
T 6
1 1
S 2
T T
e el r on i
c 9
5 9
s F
ci n
6 1
S 2
2 8
T 8
r r t
e A
nb s t
i 1
I o.
3 3
po t
S S
3 4
P 0
5 e r n ee s
e aTi 5
oT oF oF 3
1 d
ou e
(
4
- 6. T
- 6. T T
9 oS 3
4 t
Sv t
t t
1 e
o e
t e vP o
t cnigg
- 4. T I.
o e
o o
o o
e et aa r,
eEoS sl t
t t
5 9
5 v at t i.
t t
t t
bT 7
2 1
s 2 9
e imloo s
l 6
T 3
3 d
1 1
gr vv A
4 2T 3
1 1
i r
1 9
s n eo e.
2 3
o gf nn n i
ibbniioi s
i a A
t t
cl rbbt u
0m a o oooc A
1 r
e Ccin
't "b
4 5
6 7
8 9
0 eFVFBBs n
s 2
3 a
1 w
6 ee 1
t 3
3 u
Rp o
s S
N1 2345 ruo
2 0
0 4
63 5
6 8
3 0
1 7
3 9
?
8 0
0 5
6 0
?
7 0
3 5
0 5
5 9
0 3
%h 0
8 1
4 1
8 4
5 7
7 0
2 9
2 9
5 7
1 0
8 9
2 3
9 8
0 8
0 7
3 2
3 2
E tp 0 2
0 3
3 1
1 0
2 0
1 4
2 3
2 1
0 1
6 4
0 5
8 9
0 0
0 8
3 2
0 7
1 D
g e
1 1
1 1
2 1
N vD A
5 ts u %h 8
9 4
9 0
0 0
0 3
2 7
2 8
4 6
5 0
9 5
7 5
0 0
5 1
0 9
3 0
2 7
4 5
0 r
2 5
0 4
0 0
0 0
4 5
9 7
2 5
9 4
1 0
0 3
7 8
0 7
9 0
2 4
0 8
0 9
3 t
B o
.x 0
0 1
4 1
2 1
1 7
0 0
2 1
0 3
4 1
1 0
3 1
0 2
6 7
3 6
2 8
6 2
1 c
1 3
1 2
1 s
aD i
t E
h D
lo 0
0 0
0 5
0 5
5 5
7 5
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 3
4 2
t h
V s
0 1
3 1
6 0
5 7
6 3
4 0
6 0
8 3
2 6
9 8
2 9
1 4
4 0
4 5
7 2
1 2
5 tg )n.
r 2
0 0
5 0
5 2
2 1
1 2
2 2
2 2
6 3
3 2
0 0
0 0
2 1
4 1
2 0
1 0
0 0
u n
t 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 B
e(
i n
L io P
%h e
0 2
2 4
9 4
4 7
7 3
1 1
9 9
9 6
6 8
0 2
2 7
9 5
4 5
7 0
1 3
u t
+
o 0
.g 3
4 3
5 2
2 3
3 3
7 0
8 8
2 6
6 3
4 3
3 6
5 2
7 8
2 4
8 9
l 0
E c
r vD 3
3 3
2 9
9 9
9 5
5 3
5 5
4 4
4 4
5 4
4 3
4 6
6 7
3 4
4 3
a 1
n V o D A
f d
o io r
e e
r c
t
%h 0
0 0
0 r
a n
n t
0 9
1 1
2 2
7 9
0 7
6 0
2 E iv e
s 5
0 7
6 4
3 0
7 7
4 0
0 1
t o
u p
0 0
0 0
0 e 0 8
9 1
8 8
9 0
0 5
9 9
1 4
0 0
4 3
9 6
8 5
7 2
7 n e d i
j xa 0
0 0
0 0
t d
6 7
8 3
4 5
4 a D i
a A
iD 4
3 3
3 6
6 3
7 7
5 5
5 5
6 5
5 4
5 f
1 1
1 1
1 n
h e
o l
M d t
e s
t c
r r
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
S h
r u
t B
g )n.
1 2
4 1
5 5
7 7
9 7
6 9
9 5
7 7
8 1
6 6
6 6
0 7
9 7
1 4
6 1
o n
1 4
5 3
5 5
7 7
8 2
7 6
6 2
6 6
3 7
9 9
8 9
4 6
3 4
8 3
6 3
5 i
C e
(
1 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
9 L
s t
. s 6
9 4
1 8
0 0
0 9
7 1
3 7
6 3
9 9
6 3
6 4
8 6
l x
t o
al 2
1 1
1 8
0 3
3 2
2 2
7 7
2 5
3 2
4 5
4 1
6 8
o V
ih V 0
0 0
0 6
7 9
9 0
0 0
0 0
0 0
0 0
0 0_ 0 0
2 9
n E ib D %h t
3 4
3 3
6 3
2 3
5 3
1 7
0 2
6 1
0 9
2 3
7 7
7 6
7 5
5 0
9 6
b N
p
.g 3
1 3
9 5
4 5
4 0
6 1
4 6
6 9
5 3
2 7
8 7
0 1
2 6
8 3
2 1
1 e
o d
vD 3
3 3
2 9
9 9
9 5
5 4
5 5
4 4
4 4
5 2
3 3
4 9
5 5
6 5
5 6
4 e
B A
ts o
u t
d
%h j
7 6
7 1
0 0
0 0
0 8
9 0
0 9
0 6
2 8
0 7
9 0
0 0
0 0
0 0
0 0
t h
A a
.x t
t e 0 8
0 6
9 8
9 9
1 7
9 8
2 5
3 4
2 3
0 6
5 4
8 3
9 1
2 0
0 4
aD 4
3 4
3 9
9 9
9 6
64 6
7 5
6 5
5 6
4 4
4 5
9 6
6 9
7 6
8 5
s D
r i
c u
h B
6 D
h 1
0 0
0 0
5 0
5 5
5 3
5 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
tg )n.
1 1
6 2
1 5
2 4
2 3
1 9
3 5
9 4
6 5
7 8
4 7
1 1
3 7
7 9
3 9
5 s
4 4
1 4
6 6
2 5
7 9
4 4
7 2
9 1
n 7
3 2
3 7
7 8
8 3
2 2
4 4
1 e
I e(
d L
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
u lc in s
t 7
5 7
3 7
7 5
2 9
8 8
1 8
8 3
5 4
4 3
3 n
b 5
8 2
5 5
4 3
3 3
6 2
2 2
2 l
I b o 1
1 1
1 2
2 oV 0
0 0
0 1
1 0
0 0
0 0
0 0
0 0
0 0
0 0
0 B
no t
i s
9 9
9 9
9 6
9 6
9 9
9 6
9 9
6 9
6 9
6 9
9 9
9 9
9 9
9 9
9 9
t E y 9
9 9
9 9
3 9
3 9
9 9
3 9
9 3
9 3
9 3
9 9
9 9
9 9
9 9
9 9
9 a
D a 9
9 9
9 9
5 9
5 9
9 9
5 9
9 5
9 5
9 5
9 9
9 9
9 9
9 9
9 9
9 l
u N n 9
9 9
9 9
4 9
4 9
9 9
4 9
9 4
9 4
9 4
9 9
9 9
9 9
9 9
9 9
9 A D D D D D P D P D D D P D D P D P D P D D D D D D D D D D D lav k.
E c
0 m
a o 1
2 3
4 1
1 2
2 7
8 9
2 2
3 4
4 6
7 8
8 9
1 2
4 5
1 2
1 1
t y
rN 1
2 t
C W
0 n
ia n
co t
io d
5 S
S S S S S S S S S S S S S S S S S S S S S P P P P P
P P P r
t A
a e
c F F
F F
F F
F F
F F
F F
F F
F F
F F
F F
F F
S S S S S S S S O
c o
en n
L 9
m U
. 2 2
2 2
5 5
5 5
5 5
5 1
1 1
1 1
1 1
1 1
1 1
4 4
4 4
2 2
3 3
lF 8
8 8
8 1
1 1
1 1
1 1
0 0
0 0
0 0
0 0
0 0
0 7
5 5
5 8
8 0
6 a
l E
o n
C i
f D
lF
\\A N
w 6 6
6 6
5 5
5 5
5 5
5 6
6 6
6 6
6 6
6 6
6 6
9 2
2 2
6 6
8 8
O n
o 2
2 2
2 2
2 2
2 1
1 1
1 1
1 1
a R
g 4
m F
G 2 2
2 2
2 2
2 2
2 2
2 3
3 3
3 3
3 3
3 3
3 3
1 1
1 1
1 1
1 1
l a
5 2
2 2
3 3
3 3
n S
i e
\\
F 8
l r 7 7
7 7
8 8
8 8
8 8
8 8
8 8
8 8
8 8
8 8
8 8
7 6
6 6
6 6
6 6
9a b
a 9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
lA e 9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
9 9
\\
a Y
A 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 T
I A
\\T W
tn A A A A A A A A A A A A A A A A A A A A A A A W W W W W W W F
B a L L
L L
L L
L L
L L
L L
L L
L L
L L
L L
L L
L L
L L
L L
L L
U r
D D D D D t
l T
)
P A
A A
A A A A
A A
A A
A A A A
A A
A A
A A A A
D O
\\
)\\!lll l\\1 1i ll\\{fil!
0 0
1
- co 0
9 e
s 0
a b
a a a t
mm a
c h
Di in i
x t
t e ai i
0 s
p c
e nMMc co 8
i D
mnntsL 0
eii i
f n n gc e
e u u oo g
RKKLL ga re a"
0 v
7
~
A 0
.sv Y D
O 0
e 6
g PcJ[
0 a
o r
n e
i v
b n 0
A b
?
o m
0 B
a 5
x C
0 E
Cc 7
e c
v S n i
7 D:
7;-
t 1
3 c
O
[
5 u
0 r
4 n
7 ts 0
e a
pc
{
g'-
D s
ey e
j
/'.
C r
~.
O. 3 0
F
]
M0 1
f[".Y yc Nf[_f, 5
J i
rc m
F f
0 2f
[f 2
49
/
0 r
1
/
e J
S 5
!'J3 s'
s e
eap r
e u
/
0 r
g 1
d.
iF
/
r 0
ir sa Tr o.
- '0 r.
~
0 w
0 w
0 0
0 0
0 0
0 0
0 0
0 u
0 9
8 7
6 5
4 3
2 1
w 1
m i
w a f'1o T 3jg2' mo l1
0 0
1 m_
0 m_
9 es h
a 0
t ba a a p
4, t
mm a
e Di D
x n i
e ai c
0 m
mmc 8
n i
e n n 0
t u
r s
ii e
m w
e u u o i
f n n g M
fp-ix RKKL a
C 0
0 7
3 f 0
s h
v t
c p
D K
D e
O 0
6 m
P :
d,/
0 u
-v,*
n 7
m i :
A-ix b
(1/
a b c c
M
- 0 B 3 C,
J C
'J/
- 5 m
0 ax f,
E C
0 j
e 8
S v
l DC f
0 t
5 i
c 4
u OD j'
E 0
r C
l t
n A/
s
(
e a a c
D
?
p
/r-2 s
e.
c0 e
x 3
e 3
r s/
0 F
f K x/
f 1
ye
[
m 0
u 2
l u
r
/
0 O
c F
/
K, h
~
~
A
/
O 2
/
n.
0 5
l(
L 1
r 0
a e
/
ir r
w u
o.
g i
~
F r
00 wr 0
w w
0 0
0 0
0 0
0 0
0 0
0 9
8 6
5 4
3 2
1 ua 1
nu E. U;o o b5aa#2w m
t mo
1l i!
\\l l!),{lil 00 1
0 09 h
tp 0
e 8
D egarev 07 A
+
ev i
tc e
)
f 0
f 6
E
(
+
t
+
h t
s
+
p r
e u
+
+
D B
0 e
- 4 5
g 9 a
1 y
r t
+
e 5 n
vA ia E
t 8
+
r 8
0 D
e 7
4 3
+
N c
9 n
1 1
U
+ 88 E
x8 0 D
9
=
+
0 N
7 'R
/.
3 9
0
+
m x
i 3-
=
m 5
y o
/
e r
s u
0
-a 2
g y
o.
iF
/
m r
/
wm w
0 o.
1
.r w
/
w w
u 0
a m.
0 0
0 0
0 0
0 0
0 0
n 0
9 8
7 6
4 3
2 1
w 1
u l:
o a & a & l > < ! T i. * ! L' a l
0 0
9 1
/'i 09
/
///
08
/
h
/-
tp e
=
D 0
7 7
m
[
y+
u m
+
)
i x
a o
06
(
/
M h
t p
+ '
e y
+
D t
n m
0 ia e
5 u
t
/
m 0
- /,
r 2
i e
x 5
c a
/
M n
92 U
0 m
E 0 5 0
E 4
D 3
D
+ 0 N
N x 8
/+
5 0 7
=
%. *R
/
4-0 0
5 3
e
=
7 r
- y u
m g
1 2W i
O F
-e 0
du 2
-5 A
D
/
le T
e F
I 0
WA 1
D n
/
l a
sp w
F a
eW 89 lsA O
\\AL 0
0 0
0 0
0 0
0 0
0 0
A
\\
0 9
8 7
6 5
4 3
2 1
N T
1 IEBU a EaE M h l wW b.B'i-
\\T Q
0 0
1
)a t
aD e
a b
ta u
D T
e d
b e
l u
l T
u P
d
(
a e
.g l
t l
a e
u o
g P
L D
a t
Cl X
o g_
C V S
n Di Ob 0
b 1
n o
aB p
p 5
s s
2 e
u 9
e s
r r
5 0
F e J~
e v
+ 7 g
e 5
a bh
/,
t
)
1 l
x 9
o ut p
V
(
n 0 T
e L
n 1
d D
=
2 i
7 b
5 62 b
e e 4 R o
l g
l 1
B u a 0
P re
=
1 v
y yA e
d lr e
r, at 1
s F
u g
.j 5 d A
m
-5 a
m.
t e s r
r u
u gB x
5+
xF, i
o F
x w
x*/
w X/
wo
/
o X
w x/
i.
/ %x w.
x w
1 u
m 0
m 0
0 w
0 g.S&* yk<
m 1
1
0 0
~
1
~
eg a
)
k a
t d
a a
e e D k L a
e o b a rN t
u a
I T
D d
e t
t b
s s
e l
u e
e l
_ T T T u
P e
d u
u
(
t t
e i i S S g
l l
n n
o u
a e
P I
I l
ta g
n Di X
- O l
C oo C V-S n e 0
Di 4
1 b
Ob oJ naB ps s
u e
s f a
e e
r g
r e
4 a
F v p t
2 lo eh 17 V
b t u
p y%
0 n
2 e
i 2
T D 1
b
+
6 b
5
)
0 o
d m
x 9
B e
(
o u
n 0 ll L
=
f u
mo Pi 1
2 *R 1
x 7
a
/
1 eM y
0 1
-(
=
lrd oO a
e y
Ft su m
6.j
~ }, <
t 2
d r40 5 A Y
w et r
s d
s A
ur On g u a
p F B i
r e
s_/
la e
F n
O k 4 I
iF F
\\AO e
p ee
/*
la rF i
's N'
^
tA 19 a
~
1 A
\\
L 0
A
\\
N T
0 0
E 0
( j$.6,EE B
1 U
T 1
\\
Q'
5 2
7 02 sno i
tac idn I
weN 5
r 1
o f
n o
h it tp u
e b
D ir e
3 g
2 t
s a
5 r
i e
D vA h
t 0
p
/-
1 eD egarevA m
1 2
7
/40 5
cu d
5 3
e A
r O
n u
s ep g
w i
lF F
e u
F I
FtA O
ne sp e
F teh F
\\39 la M
~
0 A
0 9
8 7
6 5
4 3
2' 1
0 W
T N
1 0
0 0
0 0
0 0
0 0
O IE B
- yo.5 g E 5E.2A#%3 eau m
Q
07 0
- /. A j
6 h
t p
e g
D
/
e gt 0
5 a s r
r u
e vB h
Ar o
f g #-
sfu) r%
s e0 e
v6 c
0 4
n e
/
m (<
h lu t
r
^
p u
n s
e mh Co D
- 4. y, t
e ix p g
4 a
2 a e 5
r MD 9
e 5
G v
e 0
A e g 3
b a u r e
Tv dA
/.,
e e l
v lui t
P c 0
e x
2
.f 9 5 8 f 3 7 E
9 m
1
-5 2 0 m
1 e
=
.o
= *R r
u y
g u
i o
F 0
1 w-
-o w
i.
w 0
n E
0 0
0 0
0 m
0 0
0 0
0 9
8 7
u 5
4 3
2
(
u 1
1 m
ek a.gA E=EiE u.
me
0 0
1 0
n n
- Oa is s
0 o o 9
i_
s s e e P
r r
g g h
e e t
R R_
p 41 e
B cB C 0
D a
G G G G_
8 j
S sS S e
g
+
., o +
a
+o-r e
v A
0
.s x~
7
+
v 9
h 2
t 8
1
- +4 p
e 1
D i
=
0 m
6 y
u h
m t
p ix a
e a
o D
M 8
e
^
o 0
^
g C
0 5
ar 5 G
e 2 v
S a
g 5
A g
x 9
B 3
9 0
4 G
4 S
1 1
5 a
1 a
=
\\y C
O 0
/
E 3
a 1-m ye t
2
/
l 4
r i
0 a
c F
0 o
d 2
5 A
9 O
aa 5
psee e
F r
r tsn u
iF g
0 F
I t
i A
F 1
O nepnee rF W
W hA 0
M W
0 0
0 0
0 0
0 0
0 0
0 M
0 9
8 7
6 5
4 3
2 1
U 1
B W
5gO !
E
h h t
tp p e e D D
.g x v a AM E0 s
0 n
0 o
1 i
tu 6
b l1 9
i r
ts 2
l 9
iD h
8 8
tpe 4
D 1
8 ev c
0 i
1 8
tce ff c
6 E
1 7
ts 2
r 7
u B
)%
8
(
5 1
6 h
6 1
t p
2-C 4
e 5
l 6
O D
E t
0 s
r B
6 u
G B
S rI 6
i 5
1-2 y
l 5
e lra 8
F IIIlIl 4
m i
n 4
0 e
IIl I
4 1
u
-5 s
0 n
e 4
o rug 6
r-l 3
i.
iF m.
u 2
lI 3
o.
8
.r i.
r.
2 w
4
^
m 1
u u
mustuo
4
)l ll h h t
tp p e e DD
~
t 1
u b
- 7 i
- 9 r
ts
- 4 i
9 D
1 h
9 t
pe 88 D
e 58 v
i tc 28 e
ff 9
E 7
t s
6 r
7 u
B 37
)
5
(
1 0
7 7 h 2
C t
5 p
O 7 e 6
E D
t 4
s 6 r C
u B
G 1
6 S
85 1-y 55 m
e u
lr 2
v a
5 O
F c
9 u
d 4
5 1
A O
1 e
6 ep 5
4 s
e lF e
3 e
r m
4 F
u I
g
\\F A
0 O
iF l 4 neps e
7 F
3 lan iF M
4 iM a
3 A
W 4
2 0
8 6
4 0
T 1
1 1
fM EB
- y5o i
U T
Q
rj r:
~ nsO 3~
3 G
G" cc O c 0lf3 0
1 C
l
.?CC I
d~
e 58 2 4
E y
i 07 2 9
i 55 2 g
o8 i
04 2 r
nc a
52 2 i
g m
01 2 O
m s'
59 1 O
O t
08 1 M
W m
56 1 g
3l:
N oo g
h 051 c y
E 53 1 y
,g h
al 8F L
{
U I
N 02 1 o
"B o
M 50 1
.,b 09 0
""4 1
a
'h 57 0
$e a
l m
11-
'E 06 0 t
{
.y
.m ,,
54 0
?
y 03 0
[
8 N
.e **,... -
c 5
o ot
,..s 51 0 j
y l
8 E
E E
8
$E 5
o o
noitubirtsiD evitalumuC g
C
6.0 GROWTH RATE EVALUATION This section describes the method used to obtain growth rates for the average and maximum depths and presents the growth data for Cycle 15.
6.1 Growth Rates for Large Indications The largest values in the growth distribution for freespan ODSCC indications have the most significant influence on determining the largest EOC-16 indications that result in the lowest burst pressures.
Therefore, it is important to ensure that the largest growth values are determined accurately. Growth rates observed during Cycle 15 are applied to project the EOC-16 condition. This section examines growth rates for the largest indications observed during the EOC-15 inspection as they are likely to also define thelargest growth values.
The growth distribution ofinterest while projecting the burst pressure at the end of a cycle is the average crack depth growth distribution. The average depth for the freespan ODSCC indications observed during the EOC-15 inspection, as well as for those detected and repaired in SG-B during the unplanned outage in July 1998, were determined from the crack depth profiles obtained by sizing the + Point coil signal. The shallow ends of a crack do not affect the burst pressure of the crack. Therefore, instead of averaging the crack depth over the entire length of the crack, the depth averaged over the critical section of the crack that yields the lowest burst pressure is used to estimate the burst pressure at the EOC-16 condition. More details on the estimation of average crack depth used for growth rate evaluations are provided in the following section.
Tube R25C51 in SG-B had two of the 8 larger freespan ODSCC cracks observed in all 3 SGs during Cycle 15. As discussed in Section 5.1, these two cracks were destructively examined in the laboratory and their actual depth profiles are available for the growth evaluation. Therefore, the destructive exam depth values (over the critical segment yielding lowest burst pressure) are used to establish the Cycle 15 growth rates for the 2 cracks in tube R25C51.
The BOC-15 average depths for all indications were obtained by applying a correlation between the bobbin voltage and average depth described in Section 5.3. This correlation was developed using actual depth profiles from destructive examination of pulled tubes. The BOC-15 average depth for the second crack found above the sleeve in R25C51 was estimated using the bobbin voltage measured in 1992 (0.2 volt) prior to sleeve installation. Growth to EOC-15 occurred over nearly 4 full cycles of operation.
The average depth of the crack was assumed to grow unifonnly from the value corresponding to 0.2 volt in 1992 (about 36% average depth based on the bobbin voltage to average depth correlation) to that obtained by destructive exam for the tube corresponding to the August 1998 inspection (92.4%). The BOC-15 depth was obtained by interpolating between the 1992 and August 1998 values. The indication in R15C68 was not reported during the EOC-14 inspection; however, a reanalysis of the EOC-14 bobbin signal yielded 0.57 volts, and this voltage was used to estimate BOC-15 average and maximum depth using the bobbin voltage to depth correlations presented in Section 5.3.
Growth data for the largest 8 indications found during the EOC-15 inspection and the August 1998 inspection are shown in Table 6-1. A significant conclusion from Table 6-1 is that the largest growth values occur for indications with BOC voltages below 1 volt (R43C32 in SG-A, R25C51 in SG-B and R15C68 in SG-C). This result would tend to indicate that large growths are not likely to occur on the largest undetected indications at the EOC-15 inspection.
Q tubeint\\ala\\nla9tEinal FreespanoA\\FIFinal Freespan OA-6. doc 6-I
A comparison of the length and depth data in Table 6-1 for partial cracks that yield the lowest burst pressure with those for the corresponding whole crack shows that the adjustment for burst pressure results in significantly shorter effective crack length and higher average depth. The maximum depth for the partial length is generally the same as that for the whole crack since the deepest segment of the crack generally has the lowest burst pressure. The use of the burst-adjusted crack length eliminates the shallow tails at the ends of the crack that do not influence the burst capability of tne indication. The extent of the shallow tails that are not included in the burst effective length vary from crack to crack and is determined by the partial crack length identified from the burst adjustment. The burst adjusted or
" critical" crack lengths, average depths and maximum depths are used for the growth rates and tube integrity analyses of this report. This is important for freespan ODSCC assessments as the use of average depths over the full length of the crack can lead to significant underestimates of the burst pressures for the indications. The longest critical crack lengths are 0.78 inch for R43C32 and 0.77 inch for R25C51. The largest crack length is bounded by 0.80 inch which is applied for the deterministic cycle length assessments given in Section 8.
6.2 Method of Developing Growth Rate Distributions This section describes the approach used to develop growth distributions for both the average depth and maximum depth of the crack population. The average crack depth growth distribution is needed to project the burst probability at the end of a cycle, and the maximum depth growth distribution is needed to predict leakage. In this report, growth rates observed during Cycle 15 are applied to project EOC-16 crack depths. Both the average and maximum growth values for SG-A free span indications are significantly below those for SGs B and C based on historical and Cycle 15 data. Therefore, only growth distributions for SGs B and C are presented herein. This section describes the method utilized to predict the growth rates for average and maximum depth of freespan ODSCC indications.
Growth during Cycle 15 is the difference between the crack size at the EOC-15 and EOC-14 conditions.
The average depth for freespan ODSCC indications observed during the EOC-15 inspection, as well as for those detected in SG-B during the August 1998 outage, are determined from the crack depth profiles obtained by sizing the + Point coil signal. The shallow ends of a crack do not affect the burst pressure of the crack. Therefore, instead of averaging the crack depth over the entire length of the crack, the depth averaged over the critical section of the crack that yields the lowest burst pressure is used to estimate burst pressure at the EOC condition. The critical section of a crack was established by systematically calculating the burst pressures for all possible contiguous segments of the crack and selecting the segment that yields the lowest burst pressure. The burst pressure was calculated using a correlation developed in Reference 10.4 for partial depth axial cracks. The correlation in Reference 10.4 relates measured burst pressures for pulled tube and laboratory specimens with those predicted by applying analytical models such as the Framatome correlation (Reference 10.2) and the EPRI throughwall burst correlation. Available data for ODSCC and PWSCC specimens were used to develop the conelation. The application of this correlation results in a best estimate burst pressure with defmed uncertainties that can be used to establish confidence levels on burst pressure predictions.
The + Point signal for all freespan ODSCC indications found during the EOC-15 inspection, as well as in the August 1998 inspection, were sized to obtain their depth profiles. For indications in tubes that were pulled during the August 1998 outage (R25C51 in SG B) and in the EOC-15 outage (R6C10 in SG-B),
actual depth profiles from destmetive examination were used. Growth data for 5 indications in R25C51, which was pulled during the August 1998 outage, are included in the growth distribution established for SG-B. Similarly, pulled tube data was used for the indications in SG-C, R6C10. The average crack Q:tubeintiataWa98 final FreespanoA\\FI Final Freespan oA.6. doc 6-2
depths at EOC-15 were established by identifying the weakest segment of the cracks as described above.
For cracks less than 0.5 volt by RPC, the depth profile generated by sizing the + Point signal is not considered reliable (See Section 5.2). Therefore, the average depths for cracks with less than 0.5 RPC volt were obtained using their bobbin voltages and a correlation between the average crack depth and bobbin voltage. This correlation was developed considering critical segments in actual depth profiles me sured during destructive examination of pulled tubes and is described in detail in Section 5.3. The BOC-15 average depths for all indications were obtained by applying the correlation between bobbin voltage and average depth to the BOC-15 voltage.
As discussed in Section 6-1, the BOC-15 average depth for the second crack foundjust above the sleeve in SG-B R25C51 was estimated by interpolating between the 1992(0.2 volt) and the bobbin voltage for a throughwall indication in August 1998. The crack in SG-C R15C68 was not reported during the EOC-14 inspection and not identified in the initial reevaluation at the EOC-15 outage. The tube was sleeved at EOC-14 and the indications had to be carefully aligned between the two outages. The reanalysis of the EOC-14 data yielded 0.44 volts, and this voltage was used to estimate BOC-15 average and maximum depth using the bobbin voltage to depth correlations presented in Section 5.3. For other indications considered NDD at EOC-14, a bobbin voltage of 0.12 volt (28% average depth) was assigned at BOC-15 which represents the threshold for detectability for freespan ODSCC indications based on the smallest indications reported in the EOC-15 inspection.
63 Average Depth Growth Distribution Growth distributions during Cycle 15 for average depths of freespan indications in SGs B and C were established using the average depths at EOC-15 and BOC-15 (same as EOC-14) data obtained using the procedure described above. The average depth growth data in the form of cumulative probability distribution function (CPDF) is presented in a graphical form on Figure 6-1. Also shown in Figure 6-1 is e composite distribution obtained by combining the individual data for SGs B and C and smoothening the resulting CPDF curve. The 95th percentile growth value during Cycle 15 from the composite distribution is 23.5 %/EFPY. The growth distribution shows a long tail with the upper 10% of the average depth growth rates between 18%/EFPY and 35%/EFPY. As the average growth rate is only 5.7M, the majority of the indications show very small growth rates.
Since the individual growth distribution for SGs B and C are based on relatively small populations (58 and 67 indications, respectively), the composite growth distribution obtained by combining the data for both SGs B and C is expected to be more reliable. Therefore, the composite growth distribution was j
applied both SGs B and C to perform the leak rate and burst probability evaluations at the EOC-16 conditions. Since the high growth values for SGs B and C are similar in number and frequency, the high growth tail of the distribution is not significantly affected by combining growth data for the two SGs as shown in Figure 6-1.
Due to IRl5 tube plugging, the average T oi or Cycle 16 at uprated power conditions is about 2.1 F h f higher than that for Cycle 15. Since corrosion rates, and therefore crack growth rates, increase with temperature, the Cycle 15 growth values shown above were corrected for the temperature increase.
Using the Arrenhius equation to represent the growth dependency on temperature and an activation energy of 30 kcal/ mole, the increase in growth rates due to uprating is estimated to be 5%. Therefore, growth rates used for the Cycle 16 predictions were increased by 5% relative to the Cycle 15 growth rates.
Q tubcint\\ ale \\ala98'Jinal FreespanoA\\FIFinal Freespan oA4 doc 6-3
6.4 Maximum Depth Growth Distribution Cycle 15 growth distributions for free span maximum depth in SGs B and C were also calculated using the procedure described above, and the results in the form of CPDFs are presented in Figure 6-2. Figure 6-2 also shows a composite distribution obtained by combining the individual data for SGs B and C and smoothening the resulting CPDF curve. For indications close to throughwall penetration at EOC-15, the following adjustment was made to account for potential growth rate underestimation due to throughwall penetration occurring prior to reaching EOC-15. If the maximum depth growth rate as calculated above is found to be smaller than for average depth, the maximum depth growth rate is set equal to the average depth rate. The 95th percentile for the maximum depth growth during Cycle 15 is about 27.9 %/EFPY (based on the composite distribution). The growth distribution shows a long tail with the upper 10% of the maximum depth growth rates between about 21%/EFPY and 37%/EFPY. As the average value is only 7.1%, the majority of the indications show very small maximum depth growth rates. As with the average depth growth data, a composite growth distribution obtained by combining SGs B and C data was also applied for the maximum depth. Again, the maximum depth growth rates used for EOC-16 predictions are 5% above the Cycle 15 growth values to account for about a 2.1 F increase in the Thot temperature due to tube plugging effect at uprated conditions.
i Q tubcint alaWa98' Final FreespanoA'ElFinal Freespan oA-6. doc 6-4
i
')
')
')
')
')
')
')
h 9
8 e
t
.h v
w xt 9
7 4
3 9
2 5
7 8
9 2
o ap 2
2 2
2 2
3 p
1 r
(
o
(
(
(
(
(
(
Me 5
r D
3 m
G 3
8 4
4 5
0 4
5 0
ei 3
d 3
9 7
6 5
0 5
0 rao 1
e 3
2 2
1 3
3 3
3 t
t s
h a
ed t
m
')
')
ge w
t
.h a c o
i t
gt s.
t e
r s vp 3
8 4
4 5
0 4
3 5
0 lop g
=
E Ae 3
9 7
6 5
3 0
1 5
0 vx f
D 3
2 2
1 1
3 3
3 3
r e o
3 i
t
(
(
e e s
a gr e
r a c
W a
c l
i 0
y 5
.h h
Y d
xt l
t r1 3
6 7
1 7
8 1
w P c
1 ap o
n C
fC F
1 O Me y
7 7
7 8
7 8
7 7
o 6
I D
5 r E s
n eR a
5 B
G r t
1 e
ad m
d m
p mn s
C-e a
u h t
t i
a
.h m
t c
O m vp 0
2 7
1 8
4 7
2 t
s1 gt e
w E5 i
l Ae 9
x o
C E
f i
1 3
3 8
9 4
7 t
D 5
6 6
7 5
7 6
4 85 a
r e
s 5
r t
E M G s
2 2 1
8 e
m C R 9
g u
6 s C
9 r
')
')
')
')
m Re O
1
%% %% %% %% % '% %% ') % %% %'%
)')
')
')
b
)
a
.h eu E
o xt i
L ap 0
- 0. 0 0. 0 0 0 t
0 x
bt t
0.0 0 00 0 0 0 0. 0 7
a ud a
h 2
Me 97 7 7 9 9 00 900 8
0 0 0 3 3 M
e n
w D
t t
0 91 1
9 9 9 9 9 1
9(1 9 9 1
1 1
dl o
l h
(
(
(
((
(
(
(
l o
(
eu t
i t
r t
l p
a G
l r
n upm c
i f:
o h i
o d
P t
mf n
2 s
p
')
')
')
)*)
')*)
')
r
)
)
o 9
+
e i
c r a n
9 t
5 D 1 3 3. 0 9. 1 1 5 4.2. 8 304
- 6. 2 6. 5 f t 1
a a
s 1
1 2
91 31 3 9 42 3 21 2 5 7 7 8 d d f
n g
8 R
C v 9
7 9 8 9 8 8 88 (9 9 99 8 9 6 7 e
(
((
(
(
t n
o i
(
(
(
(
h O A a o e
n t
E mi g
o t
i c
d t
5 i
w ts e e
a
-6 et l
l o
h
')
e w
a l
')
5 8 9 )0 7 9 )'5 6 )5 0) 3)
r t
)
8) 9)
sd 1
G n
- 6. 4 8. 4 5. 4 9. 7 0.2 7 7.5 5
- 7. 5 5. 4 g
t 9
2 0
d o
s o a n n 5
l l
n e
0 0 0 0 0 0 0 0 0
0 00 0 0 0 0 h n k oi 1
1 n
L i
(
(
(
(
( (
((
(
(
s o h c v t e a
ei t
t e e p
ri i
hd wtee s
t d e
w l
sd s nA e
4 W
o nno 8
4 5
7 7
5 4
oit r
s 1
C 9
1 3
5 9
2 4
4 i
t b
t F
i s
r l
0 ct ab o 2
0 0
t 0
o O 1
1 1
V e ce uoi t
E ef b r l
n df a
p y
r e
v on i
e b
e i
b 5
r e ef o 1
8 3
3 8
ag e
r ei c
r o
8 0
c B C 3
0 0
8 7
e 7
2 nl.
t t
1 s a t
t e vc o
6 2
d o
3 ee F
O 7
6 1
9 oos ae p 6
1 2
N v erl s A
1 1
i E
t t
a s n o
1 cd ah emt oi n
d si wt 2 a
i l
6 n
i o
7 8
3 4
S 0
9 p
n at o e9 s
e s e
5 i
r e r s9 e
4 7
8 8
Tv 1
i e
go1 2
t t
r
+
Dc e b
a 0
0 0
0 l
F e e 1
Dingmcm la v
+
+
+
vl
+
e l
e e
H H
H oS 1
1 i
a n
b S
Nnt o o o i
T E
l 1
1 F
I 2
1 l
r or I
I 1
A T
o of r
t f F
osvd nd A
\\
f t nCteoe O
2 7
2 1
8 s
it e
a n
moat 2
e e
3 1
2 3
5 6
a a
gBmc m b
C C
C C
C C
C a
t ps 0
i i i e
u 3
2 8
3 5
5 md et dt e
T 4
3 R
4 2
1 r
s s
F 1
iuh EInE l
R R
R R
R R
sjt t
a s E.
n i
234.
F e
\\
1 t
G A
B C
o 9
8 a
S N
la ia la
\\tn ie bu t
Q
h n
d e
t
.h w
xt eg i
ah aph t
o ar t t
Me w
e mvew r
r D
G a o o
h i
t r
d g
s ng r
e G
ei ih d
t sh a
s ut e m
Y i
t P
h nwa
.h i
i c
t gt t
oi mrd s vp F
g E Ae E n
)
gn D
e d ni 9
l e ar 2
s k hle
.i 1 h at 1
l "5
.h e
5 s a t
e l
m xt l
r s
ap c
Ce ee 1
s C
y 5 v b
l l O MeD c
2 o a nn R
t ei B
l h
u nt nht l
l wu p
d f
i i
e e
s s
t sd s e ta
.h c
n n er gt e
oe os m vp g
e n l
i f
i Ae e
a a ah a t
i t
D t
r r
ct c s
ec e n s E
i s
i rk d v d eu eipn a nrah h si e pt i
t l
r d
awH oe l
e an n
.h g
(i xt e
a ap l
%n%
h k o ).
i MeD wedt 0i0 e
al h 0 v ng g0 l
)
1 e6 an 1
t r e n
g2 r e orf i
a/el ao o h 7 pd h
t P
t ut pe t
hd p
oa uts ptpn
+
e 5 D dYf u e u ed d o eP oj 1
g n
d e
b C-g%C v
nF na O A aEot ts a0 ue0O i
r r l
s t
E p2 a r E
6 n1 n u bv1 u
ib a
r me 6
r l
m (e. fi op
(
h on Y<a t
rh r p gn P n x ut t hf ie s gnt u
s e
F L
Ed e s newen eemoc a ev rl t r n e 2 ri pk s ge v r t 8 uct cuf r 1
c us ea ajofh r r rd ct u f
4 1
s eb carih os s
1 d e edt C
yd e eh r t
t g
p l
cl r
o O e k t s a ee.
emht t
V E t d l
u yl h
t f
n el oo jb e a i
mtp r t l
f d
h i
r l
l s adt eue bb 5
1 as e - e sh md 1
5 whtus ciTie t t c
o t
o B C C h pb uf r eh xg d.
O 5 gei
.a a 6
t 2 udtrbf ph a et mr A
E Ro e
o e net d p rh goh o env r
n a
a ot act nmdi p
n f
rt r s r s
oea esumteo e
o h t vh vimua f e
i r
t t
oh F
a wh at e
o pd wh pxmrd i
i t
t l
v aih n an enat e xt u l
r F
E gd al pd mawf I
o f
e F
mmmfd minmo
\\
r n A
uuu o eu 5Ga O
mmmh gmh
.h n
1 t
t a
t a
e xgri wCt h
p i
i b
x x i
s s
u a a a n e x e
oOepe e
e v a r
t r
T MmMLAMGBd a
F r
lan i
.0 F
5 67891
\\8 G
9 S
laa
\\laa
\\t Q
5 0
J C
B s
G 4
S 0
dd Y
ee nh P
i t
F B.
C.
m b o Y
E G
G on P
/
h_
S S
CS F
t E
3 p
0
/
s e 5
nD o
32 e
i
'ni i i j
.1 l I j1 j! l - 1 *
! ! i,y" t g a
a c
r i
.._ /
d e n v A
2 Y I
Cn
. r 0
P Ci F
S h E
/
/
t Dw ht Oo w
o r
n aG G
r 1-pr S o
, 1 6
0 h e
f r
t
((</
p r e u e n e
7 gr o D
6
[ '- { ',e iFi t
e F - u g
a 5b I,ll r
i I:e1:
e 1
r t
v Cs 0 A
. 0 OiD E
-/
y 1
t i
t l i inb r
e.
Ua l_
b i.
t yo n_
. 0 d.
1 co e r e
rP c.
l 6
r e._
A a e O
Fv P_
n a
p i
t s
6_
e a
9_
e r
l F
u l
a m
i n
2 F
u 0
I F
C A
\\
O na ps cer F
l i
an F
3
\\8 0
9 a
0 8
6 4
2 0
l a
1 0
0 0
0 0
l
\\a a
\\t gia 1!.
5iiu 2%3Es0 ic n
bu t.Q
5 0
C B
s G
S 4
dd 0
Y ee nh P
Y i
t F
B C
b o P
mo F
E G
G om E
/h S
S CS
/
t M
9 p
e 7
2 snD 3
om w,
.1 0
it u
.l j1
! I l uj-l -
I. ! i - s*1 I l acm i i d x Y
na P
M F
I C
E n
/
Ci ht S h 2 w Dt
, 0 o Ow r
o G
s n r h
t,',
2 aG t
- p p
6 Sr e
e o
D f
ef r
8 m
u e n
-6 r
giFo u
e 1
m i
F t
0 i
u 5
x b
a 1
Ct y'
M i
r N
Ois ED 1 y
- t
.:./
t i
el i
l U b e
'a 0
ni 0
l a
i t
yb n
co eo e
d l
r c
4 r
rP A
a e
Fe P
O na v
p i
5 se t
9 er a
F 1
l l
a u
0 in m
I F
F u
\\
C
~
A O
~
nap
~
see
~
rF
~
lan iF 2
\\8 0
9a 0
8 6
4 2
0 t
a
\\
1 0
0 0
0 0
laa
\\
- ss2* $ 5 i iuAe>i =Ea0 ie C
t n
bu t
Q l
lllll\\
7.0 ESTIMATES OF NUMBER OF FREESPAN ODSCC EOC INDICATIONS This section describes the application, by E-Mech Technology Inc., of a multi-cycle Monte Carlo simulation approach to estimate the progression of freespan axial ODSCC at Farley Unit 1. A principal objective of this analysis is to estimate the number ofindications expected for Cycle 16 as input for the reference deterministic and probabilistic analyses of Sections 8 and 9. The indication distributions developed in determining the number ofindications also permit structural and leakage integrity analyses.
The tube integrity analyses using the multi-cycle methods of this section are more statistical than the reference methods of Sections 8 and 9, tend to be less conservative than the reference methods and are used as an independent methods assessment to support the results given in Sections 8 and 9. The burst probability and leak rate results of this section are provided for information and comparisons with the reference analyses of Section 9 but are not used to establish an acceptable Farley-1 operating cycle. The multi-cycle approach has been used with good success in the past, particularly with respect to instances of freespan axial ODSCC. It is the subject of numerous presentations and docketed submittals to the NRC (References 10.7 to 10.10). The following paragraphs summarize the methods of analysis, describe analysis input data, present projections of the number and severity of freespan axial ODSCC indications expected at the end of the current operating cycle and describe projections of expected structural margins and leakage integrity.
7.1 General Description of Multi-Cycle Monte Carlo Simulation Computer simulation of the processes of crack initiation, crack growth, eddy current inspection and sizing and repair decisions using Monte Carlo methods provides a powerful, insightful tool to estimate the progression of corrosion degradation of steam generator tubing. This approach allows projections of both the number and severity of degraded tubes and quantitatively treats both detected and undetected instances of degradation. Projections of the severity of degradation, for example, lengths, depths, and shapes of axial freespan OD stress corrosion cracks, allow calculation of degraded tube burst pressures and, in instances of through wall penetration, leak rates at both normal operating and accident conditions. Each simulation of an operating cycle provides a projection of a worst case degraded tube.
Hence information is available for comparison with deterministic structural performance requirements.
The results of many simulations provide data to calculate conditional probabilities of tube burst at any desired differential pressure. Hence probabilistic structural performance criteria can be evaluated.
Typically, leak rate distributions are calculated for EOC normal operating and postulated SLB conditions.
The Monte Carlo simulation approach employs a Weibull crack initiation function, a distribution of crack growth rates, and a probability of detection curve for the eddy current probe ofinterest to simulate the processes of crack initiation, growth and detection (or non detection) via eddy current inspection.
For axial cracking, crack shapes are selected from a distribution of shapes generated by destructive examination results on pulled tubes. The new pulled tube results described in this report are in excellent agreement with the database supporting the multi-cycle simulation model employed by E-Mech. A characteristic EOC distribution of crack lengths is used based on eddy current inspection data. Growth in crack lengths is an option in the simulation program but such growth is easily covered by using a I
conservative EOC length distribution. Once a simulated crack is detected, a sizing error is selected to f
develop projections of measured crack depths and lengths. The remaining element of variation is the tubing tensile properties. Burst pressure varies directly with the sum of yield and ultimate strength.
Tensile properties are selected from a known distribution.
Q:tubeint.ala\\ala9tfinal OA\\FI Final Freespan oA-7. doc-04/21/99 7-1 l
I The simulation program tracks the length, depth, shape, tensile properties, detection status and any associated NDE measurement errors of all initiated cracks for multiple cycles of operation. A plug on detection or plug on sizing repair criteria can be specified. Initiation function parameters are selected by comparing projected versus observed numbers ofindications for past operating cycles. Comparison of actual measured EOC depth and length distributions with calculated results serve as benchmarking items. When good benchmarking is obtained, projections for the next operating cycle can be made with confidence. The next section presents analysis input data.
7.2 AnalysisInput Data The input distributions ofinterest are the distribution of the sum of yield plus ultimate tensile strength at temperature, the distribution of crack growth rates, and the distribution of crack shapes. The probability of detection curve of the eddy current probe ofinterest defines NDE detection properties.
The distribution of the sum of yield plus ultimate strength is taken from the generic distribution for Westinghouse mill annealed Alloy 600 tubing with a diameter of 0.875 inches and a nominal wall thickness of 0.050 inches (Rev.10.11). The distribution is essentially normal with a mean of 137.56 ksi and a standard deviation of 6.34 ksi. Minimum and maximum values were truncated at 108 ksi and 165 ksi respectively.
Crack shape is defined as the ratio of maximum depth to structural average depth. A substantial pulled tube database for axial ODSCC leads to a mean value of 1.28 with a standard deviation of 0.15. New Farley pulled tube results fit well with the previous database.
The characteristic distribution of EOC crack lengths is taken from Farley Plus Point NDE data.
Historically Plus Point axial cracks lengths are significantly conservative with respect to the structural lengths which actually control the burst pressure. Figure 7-1 displays the cumulative distribution of crack lengths used in the multi-cycle analysis. A log normal fit to measured data was used as input to the computer program. The mean of the natural logarithm oflengths (in inches) is -0.837 with a standard deviation of 0.75.
The crack growth rate distribution ofinterest is the growth in average crack depth. The shape distribution then converts simulated average depth to a simulated maximum depth for leak rate calculation purposes. The results of a Westinghouse growth rate evaluation described in Section 6 were used as input. The cumulative distribution of average depth growth rates is shown in Figure 7-2.
Section 5 describes Farley pulled tube data allowing the calculation of probability of detection by the bobbin probe as a function of either maximum depth or average depth. The multi-cycle program takes the maximum depth POD curve as input and converts, for calculation of detection, between maximum depth and average depth by the mean shape ratio of 1.28. Figure 7-3 shows logistic POD curves. The R15 curve is from Section 5. The other two curves are estimated and reflect a historical trend of improved calling criteria over time. Figure 7-4 shows a log logistic fit POD curve using the same input data used to develop the R15 logistic POD curve. With a crack initiation function the more conservative log logistic curve fit provides better benchmarking results in terms of observed versus projected severity of crack depths. As seen in the next two sections, the number of projected cracks is not sensitive to the use oflogistic or log logistic POD curves but the projected severity of cracking is greater when log logistic POD curves are used.
Q.tubeintiala\\ala98'Jinal oA\\FIFinal Freespan oA.7 doc-04/21/99 7-2
7.3 Projection of EOC Freespan ODSCC Conditions
' Die output of the E-Mech Monte Carlo simulation program consists of a detailed description of the physical geometries of all degraded tubes and associated leak rate and burst behavior. One of the output details ofinterest in other sections of this report is the projected number of axial freespan ODSCC indications at EOC-16. Figures 7-5 through 7-8 illustrate comparisons of observed and projected numbers ofindications in previous inspections. The histograms illustrate the fact that each simulation of an operating cycle provides a separate estimate of the number of detected (as well as undetected) sites of degradation. The selected initiation function parameters provide an excellent match of actual observations with calculated results. This correspondence leads to confidence in projections for EOC-
- 16. Figures 7-9 and 7-10 present projected results for EOC-16 for operating cycle lengths of 0.79 and 1.12 EFPY respectively.
Table 7-1 summarizes observations and projections of numbers of axial freespan ODSCC indications for the operating history of Farley, Unit 1. The median, or best estimate, projections are listed along with the upper 95 percentile projection. The Table 7-1 results show the expected number ofindications for Cycle 16 to be about 104 freespan ODSCC indications. At 95% confidence, the number ofindications would be about 120 at EOC-16. The 120 indication number at 95% confidence is used in Sections 8 and 9 as the total number ofindications to be included in the tube integrity analyses.
Note that use oflogistic or log logistic POD curves leads to essentially the same number of projected indications. The next section discusses the fact that the projected severity of degradation is greater when log logistic POD curves are used in the analysis. Use oflog logistic POD curves generates results which are a better match with NDE measured distributions of crack depths. For the analyses of this
{
section, the improved results with the log logistic POD are due to the more realistic POD at shallow depths than obtained with a logistic POD. The lower POD permits more small indications to remain undetected and grow to larger indications over the multiple cycles of the analysis.
7.4 Structural Margins and Leakage Integrity Projections Burst pressures and leak rates are calculated from the projected physical geometry of each degraded tube simulated in the E-Mech Monte Carlo program. Burst pressure calculations are probabilistic in nature and follow the methodology described in the EPRI Flaw Handbook. (Reference 10.12) Leak rate calculations are performed using the leak rate equations presented in Rreference 10.13 and a crack opening area calculation based the formulation in the EPRI Ductile Fracture Handbook (Reference 10.14). The burst pressure and leak rate analyses of this section do not apply the Westinghouse burst i
pressure and leak rate correlations that are used in Sections 8 and 9 of this report. The results are provided for information and comparisons with the reference analyses of Section 9 but are not used to establish an acceptable Farley-1 operating cycle.
Each SG simulation generates a projected worst case degraded tube. This tube is simply the one with the lowest burst pressure in a given single simulation of an operating cycle. Hence, when many SG simulations are performed, a distribution of worst case degraded tube burst pressures is created. Since all degraded tube burst pressures are calculated, the global population of all degraded tube burst pressures also is calculated. The simulation program tracks the number of occurrences of EOC degraded tube burst pressures being less than SLB or 3AP requirements. Thus, data is available for calculations of conditional probability of burst at any specified differential pressure.
Q:tubeint\\ala\\ala987inal oATIFinal Freespan OA.7. doc-04/21/99 7-3
7 Figure 7-11 shows the distribution of projected worst case burst pressures for an EOC-16 cycle length of 1.12 EFPY, The best estimate of the worst case burst pressure is the median or 50* percentile value.
This is 5052 psi and meets the 3AP deterministic performance requirement of 4308 psi. This 3AP requirement is met by over 80% of the population of worst case degraded tube projections.
The global population of degraded tube burst pressures includes all degraded tube burst pressures for a given EOC from all simulations. Undetected as well as detected incidents of degradation are included.
Figure 7-12 presents this global population of degraded tube burst pressures for EOC-16 after a cycle length of 1.12 EFPY. The upper and lower boand curves are the actual boundaries of all simulation results. Using this curve and the expected number of degradation sites at EOC-16 (undetected and detected) structural margin figures of merit of almost any definition can be calculated. The distribution of worst case burst pressures can be calculated and the probability of burst at any specified differential
)
pressure also can be calculated. Naturally calculations of this type require a very accurate description of the lower tail of the global degraded tube burst pressure population.
Conditional probability of burst results are presented in Figure 7-13 versus EFPY. Best estimate and 95% confidence probabilities are shown for SLB and 3AP conditions. These results are based on use of log logistic POD curves. Note that the conditional probability of burst at SLB for EOC-16 at 17.11 EFPY (1.12 EFPY for Cycle 16) is less than the required value of 0.025 to satisfy NEI 97-06 requirements. It is important to note that the associated probability of burst at 3AP is 0.185. Thus, there is some non trivial chance of observing an EOC burst pressure less than 3AP even if the conditional probability of burst at SLB is less than 0.01.
Figure 7-14 presents conditional probability of burst results based on the use oflogistic POD cun'es.
These results are significantly lower than those of Figure 7-13. When the logistic POD cun'es are used in the Monte Carlo simulation, the significant probability of detection at very small degradation depths leads to satisfaction of benchmarking requirements for observed versus calculated numbers of indications with a high fraction of small crack depths. Thus, the BOC crack depth distributions are unduly skewed to small depths and there is a resulting poor agreement between calculated versus NDE measured crack depth distributions. While the number of projected indications is insensitive to the use oflogistic or log logistic POD curves, the projected severity of degradation is substantially affected.
The use oflog logistic POD curves is necessary when a crack initiation function is employed. When initiated or undetected degradation sites are estimated with procedures similar to the bobbin voltage ARC methodology, use of the logistic formulation is appropriate and necessary to meet benchmarking requirements. It is not unexpected that radically different calculation methodologies are sensitive to different details ofinput. The objective is to benchmark either approach to achieve similar degrees of conservatism.
Table 7-2 summarizes structural margin and leak rate figures of merit obtained from the multi-cycle Monte Carlo simulation approach. It is clear that projected 95/95 SLB leak rates are not an issue. In terms of structural margins, if conditional probability of burst at SLB is the figure of merit, an operating cycle length of 1.12 EFPY is appropriate. Deterministic structural performance criteria would also be satisfied if the worst case degraded tube is defined as the 50* percentile of the distribution of worst case burst pressures. The 3AP deterministic structural performance criteria is met by over 80% of the population of projected worst case degraded tubes for an EOC cycle length of 1.12 EFPY. This is sufficient from an engineering point of view. At the same time, it must be recognized that the chance of encountering a degraded tube with a burst pressure less than 4308 psi is not trivial.
i Q tubcint\\ala\\ala98; Final oA\\FIFinal Freespan oA-7. doc-04/21/99 7-4
\\
\\
\\
Table 7-1 l
Comparison of Projected and Observed Numbers ofIndications Log Logistic POD Logistic POD Cycle EFPY Number Median Upper 95th Median Upper 95th Indications Projection Percentile Projection Percentile Obsen'ed Projection Projection 13 13.77 18 19 26 19 27 14 14.70 37 37 47 35 46 Leaker 15.88 27 24 32 23 30 Outage 15 15.99 58 63 77 69 82 16 16.78
?
75 89 77 93 16 17.11
?
104 120 99 116 Q:tubeintala\\ala98' final OATIFinal Frecspan OA-7. doc-04G1/99 7-5
)5 i
9 s
/
p 4
5 0
9) 5 0
0 0
0 (T
0 4
0 eR 2
t a)d R(
k m i
aP s
e p
4 L (R 0
0 0
0 0
0 B
65 0
L 2
S ts i
r s
6 0
5 4
u p
4 1
1 0
B 5
0 0
0 0
f 0
0 0
0 0
t ir o
4 0
0 0
0 2
e y
M t
i l
i i
f b
s 5
8 0
6 no a
p 7
1 2
0 o
b 0
0 0
0 s
0 o
i t
e r
6 0
0 0
0 ar P
5 0
0 0
0 u
2 lug a
l mi n
F o
s 5
i i
i S e t
p 5
5 7
7 t
8 7
i o a d
8 0
1 l
n 0
1 0
0 0
rR o
3 0
0 0
2 a 7C k C
4 6-7 a
e ee eCs lb L
t
)
liOr e n
i r
a t
s od p
eh nEu 0
0 0
0 T
t 5
5 0
5 w5 el s Mn
(
a e
o cl s 6
2 1
1 9 r e
7 8
8 8
r L
ear e n u
P fP lci s
o y g s
r e
Ca r
P e
es iM l
t ia t
s r
r eh nC 7
9 0
4 t
lul a u
w5 et 6
0 4
4 t
Mr B
o9 c s 6
1 2
9 u
e L
r r 3
4 4
4 e o M
t b
PW I
c 2
u
/
u 4
T 0
r
-c t
d e
o S
e s
d.
d na 7-a aC 2
2 2
2 A
i r
5 1
9 2
O g
dt e
e s 0
5 5
1 n
r 5
5 5
6 a
D Mo pse W
erF lan c c c
c iF et i
i i
it I
Dvu gts s
W t
gts s
r Ou p oi A
oi i
i n
Lg g L g g
O P CI o o o o la L L L
L in T
8 h) 9 et Y la 2
2 9
9 a
lc gP
\\a ynF 1
1 7
7 la Ce 1
1 0
0 ti E
n L(
ie bu t
Q
r f
i 1
Cumulative Distribution Function VS NDE Adjusted Length l
1.000 9
w
,-y r -
i
/
h3. $,.
0.900 j
,/
0.800
' 1 0.700 j
1 I
/
e 0.600 E
i
\\
I j'
l l
g 0.500 i
i l
I g
3 E
$ 0.400 l
l 0.300 i
i i
l 0.200 i
0.100 m
\\
Jl
\\
f i
0.000 0
0.5 9
1.5 2
4.5 3
Length, Inches Figure 71. Arial Crack Length Distribution Q:Winrea'da98 Einal OA\\F1 Final Frecspan OA-7. doc-04/2169 7,7
1 O.9 -
0.8 -
5 0.7 -
a l0.6-u.
3 0.5 j0.4-
!0.3-
+
0 0.2 !
0.1 0
0 10 20 30 40 Growth Rate, %TW/EFPY Figure 7-2. Distribution of Average Depth Growth Rates Q:tubeintiala\\ala987inal OATIFinal Freespan OA-7. doc 04/21M 7-8
Logistic POD Curves 1
0.9 -
,.',s..-
}0.8-0
/
b
.7 g 0.6 -
% 0.5 R15 POD Curve i
.N 0.4 -.
i I
E Leaker Outage 0.3 -
,~
Previous to O'2 A
Leaker Outage l O.1 0
0 20 40 60 80 100 Maximum Crack Depth I
i 1
l Figure 7-3. Logistic POD Cun es Q1ubeintWa\\ala9Ffinal OA\\FIFinal Freespan OA-7. doc-04/21/99 7-9
Log Logistic POD Curves 1
0.9 -
~~
/..<,,,,... -
0.8 -
. r' 5 0.7 -
/
a g
~0 E
.6 -
0.5 -
R15 POD Curve 5 0.4 -
j Previous to Leaker 2 0.3 -
Outage n.
....... Leaker Outage 0.2 -
0.1 -
0 0
20 40 60 80 100 Maximum Crack Depth Figure 7-4. Log Logistic POD Curves Q:tubeint\\ala\\ala987inal OATIFinal Freespan OA-7. doc-04,0Im 7-10
Nin1ber of Sites Det.,1R13 Outage-13.77 EFP(Cycle 13 Hstogram CDF 180..
- ;.1.0 160 0.9 140
" 0.8
.. 0.7
.. 0.6 I E E._
0.0 I 0.5 0.4
[0.3 02 20
[o,1 OI : _
10 A 20 30 40 Mnter(persmkean)
Observed Figure 7-5 Distribution of Projected Number ofIndications for 1R13 Inspection Q.tubeintdaWa98\\ Final OA\\F 1 Final Frecspan OA-7. doc-04/2IM 7-i 1
i Mmber of Stes Det.,1R14 Outage - 1470 EFP(Ofcle 14 Hstogram CDF 300
- 1.0 i
10.9 L
.. 0.8 i.7 0
1 2001 i
10.6
{
l
[0.5 t
i
[0.4 f.3 1001 0
+
10.2 I.1 0
0: ; -
3-. _
0.0 20 30 A40 50 60 70 80 90 Nmter(persntiation)
Obser/ed Figure 7-6 Distribution of Projected Number ofIndications for 1R14 Inspection Q:tubeint.alatal:98 Einal OA\\F1 Final Freespan OA-7. doc-04/21/99 7-12
Number of Sites Det., Cycle 15 Laaker Outage-15.88 EFP((1.18 EFPfinto Cycle 15)
Hatogram CDF 15 10 1m[
!0.9 i
140{
- 0.8 I
I
^ O.7 120 [
l 0.6
+
100[
[
[0.5 mi I
10.4 i
@[l a
10.3 0.2 20
,7.1 0
0; am a _
i0.0 10 20 A
30 40 NLnt)er(;ersirniaban)
Observed i
Figure 7-7 Distribution of Projected Numbers ofIndications for 1R15 LO Inspection Q tubeint\\alawa98'Einal OA\\FIFinal Freespan OA-7. doc 44/21/99 7-13
Nmter of Sites Det.,1R15 Outage-15.99 EFP(Ofcle 15 Hstognun CDF 220 1.0
.::: : : : : : 7 L
200[
0.9 f0.8 180 I
[0.7 140.I e
0.6 120; L
0.5 100i i
10.4 80 4
[
eof
[0.3
[0.2 r
4 0~_
E__
d0 '
50 E
70 80 90 100 110 Mnter(persirnMon)
Obss.irved Figure 7-8 Distribution of Projected Numbers ofIndications for IRIS Inspection Q:tubeintala'ala98'Jinal OATIFinal Freespan OA-7. doc-04/21/99 7-14
Nurther d Sites Det.,1R16 Partial Cycle Outage-16.78 EFP((0.79 EFPY into Cycle 16)
Hstogmm CCF 220.
_, _1.0 200
[0.9 i
180..
$0.8
[0.7 140[i
%l
. 0.6 100i
.. 0.5 L
" 0.4 80 ;
60j
. 0.3 l
l 49
. 0.2 20 [
..0.1 0 - -
EE-0.0
' 50 "
00 70 A 80 90 100 110 120 Nriber(persmkiban)
Projected Figure 7-9 Distribution of Projected Numbers ofIndications for 1R16PC Inspection Q tubeintiala\\ala98' final OA\\FIFinal Freespan OA-7. doc-04/21/99 7-15
Number of Stes Det.,1R16 Outage - (17.11 EFPf-Cycle 16)
HEvao CDF 180
.1.0 160,
f0.9 140; f0.8 0.7 120j 100 y i
0.5 80 i i
40.4 60 L L
1 10.3 0.2 20--
- 0.1 OI
_ _ _ _ IO.0 70 80 90 100
110 120 130 140 finter(persir ulation)
Proje cted Figure 7-10 Distribution of Projected Numbers ofIndications for 1R16 Inspection Q:tubeint\\alaiala98'Einal OA'f1 Final Freespan OA-7. doc-04/21/99 7-16
MrL Blast Pressure,1R16 Partial Cycle Ottage - 16.78 EFP((0.79 EFPY irto Cycle 16) i-i: mea n CCF E
1.0 y
1 4
- 0.9 200
[0.8 10.7 150
[0.6 f
- 0.5 l
i 100i 10.4
{ 0.3 t
L 50 10.2
(
t a
to.1 L
0: :
- A-10.0 4
2 3
4 5
6 7
8 9
isi Figure 7-11 Distribution of Worst Case Burst Pressure Projections for EOC-16 Q:tubeint\\ala\\al:98'Jinal OA\\FI Final Frecspan OA-7. doc-04/21M 7-17
Burst Preestse Dist.,1R16 Partal Oyde Outage - 16.78 EFPf(0.79 EFPYirto Cycle 16)
CDF 1.0
,of 0.91 1
[/'
0.81
'/
l ll 0.7i
.O
/
L 0.6
./
mean 0.51 I wer bound
.y upper bound 0.4f
//
l
. ll 0.31
.k 1
7/
0.21
'/
1
/
0'11
.//
I
.//
0.01_,.,.,,
W 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 kai Figure 7-12 Global Distribution of Degraded Tube Burst Pressures at EOC-16 Q:tubemtWa'ala987inal OA'JIFinal Frecspan OA-7. doc-04/21M 7-I8
Probability of Burst Projections POB II t
e, 0.1 '
. r #'
POB at SLB (mean)
POB at SLB (95% conf.)
POB at 3DP (mean)
POB at 3DP (95% conf.)
0.01'
,s
.SN s-
._\\,
/
N,'
"f(, -
0.001 10 11 12 13 14 15 16 17 18 EFPY Figure 7-13 Conditional Probability of Tube Burst at SLB and 3AP versus EFPY, Log Logistic POD Curve Input I
Q tubeintWa'.ala987inal OA71 Final Freespan OA-7. doc-04/2im 7-19
e -
Probability of Burst Projections POB iz I
p...
' 's 0.1 -
c
,_,.c
_ _. ~ -
POB at SLB (mean)
POB at SLB (95% conf.)
0.01" POB at 3DP (mean)
{
POB at 3DP (95% conf.)
t
'N.,
.N
' '_\\
0.001-V,'
/
0.0001 10 11 12 13 14 15 16 17 18 EFPY 1
l l
Figure 7-14 Conditional Probability of Tube Burst at SLB and 3AP versus EFPY, Logistic POD Curve Input Q:tubeintialaWa98'Einal OA\\FIFinal Freespan OA-7. doc-04/21m 7-20 J
w
I
)
i i
Q:tubeintWs'ala987inal OA71 Final Freespan OA 7. doc-04/21/99 7-21 l
i l
)
8.0 DETERMINISTIC OPERATING CYCLE LENGTH ASSESSMENT This section describes a deterministic assessment of an acceptable operating length for cycle 16.
Deterministic is defined as satisfying a burst margin of 3APso for either a partial length cycle or at EOC-16. The analysis methods apply deterministic /probabilistic methods using the terminology of the EPRI tube integrity guidelines of Reference 10.3. The known quantities for the assessment are the EOC-15 depth distribution found in the inspection, POD as a function of depth, NDE uncertainties, growth rates as a function of depth and the burst pressure correlation for partial depth cracks.
Combining the distributions and associated uncertainties requires a probabilistic analysis. The most accurate deterministic analysis is obtained by Monte Carlo analysis of all of the distributions. In this section, Monte Carlo analyses are applied to obtain the EOC depth distribution for which the largest projected indication is compared to the structural limit. The Monte Carlo analysis methods for projecting the EOC distribution ofindications are the same (Reference 10,5) as used for analyses of ODSCC at TSP intersections per GL 95-05. Sensitivity analyses are provided that combine the growth and burst correlation distributions to define an acceptable BOC depth which can be compared with a POD for the acceptable depth from the POD correlation. These sensitivity analyses provide some physical insight relative to the size of an indication that could be missed and the associated confidence level on growth that would satisfy the structural limits at EOC.
The EOC-15 depth distributions used for the EOC-15 projections are developed in Section 5.6. The PODS as a function of average and maximum depth are given in Section 5.1 where the recommended POD function is a logistic distribution. NDE uncertainty estimates for + Point sizing of axial ODSCC are developed in Section 5.2 and used in Section 8.5. The growth distributions for average and maximum depth are developed in Section 6.
8.1.
Burst Margin Requirements The detenninistic analysis requirement is to satisfy a burst margin of 3APso at EOC-16. The steady state, full power condition is used to define APuo. Startup and low power operating conditions have a lower pressure differential. A few transienu classified under the ASMME Code as normal operation can have a primary to secondary pressure differential somewhat higher than the full power condition.
However, these are short term transients which have a low frequency of occurrence. Typically, the increased pressure differential occurs for power increases about full power such as a small step load increase. The historical development of the 40% depth limit has not addressed transients and it is not considered to be requied to include the transients in the 3APso determination. The primary system pressure is 2250 psia at the pressurizer. Based on this pressure, the primary pressure at the tubesheet is also about 2250 psia which includes the pressure difference from the pressurizer to the entrance of the SG nozzle of +7 psi and the pressure drop from the SG nozzle to the tube sheet of about -7 psi. The primary pressure at the top of the tubesheet would likely be about a psi or two below 2250 psia. Within the SG tubes, the primary pressure decreases more rapidly than the secondary side pressure increases as the hot leg elevation within the tube increases. For example, the primary pressure decreases by about 10 psi from the TTS to the midpoint of the tube bundle while the secondary pressure increases by only 5 psi. Therefore, the primary and steam pressures can conservatively be taken as that at the top of the tubesheet to define APso. The steam pressures for Farley-1 are reported at the exit of the SG flow limiter at the top of the SG. The secondary pressure drop between the top of the tubesheet and the exit of the SG is 24 psi. The reported steam pressure for Farley-1 cycle 16 is 790 psia at the SG outlet which leads to 814 psia at the top of the tubesheet. Therefore, APso is 1436 psia and 3APso = 4308 psi.
Q.tubemnal:0811 Final Freespan oA 8 doc-0401/99 8-1
I The NRC draft Regulatory Guide DG-1074 defines the deterministic burst margin requirement on the basis that the largest projected EOC indication must satisfy 3APso burst pressure at 95% probability and 95% confidence for an operational assessment. The difficulty in application of this requirement is to define what constitutes the largest projected EOC indication. The current drafts of the EPRI Tube Integrity Guidelines (Reference 10.3) do not include a recommended method for defining the largest indication. Since Monte Carlo analyses are necessary to project the EOC indication distribution, the definition oflargest indication is not uniquely defined. The Monte Carlo analyses project a distribution ofindication sizes where the large indication tail of the distribution represents a small fraction of an indication. The definition used in the analyses of this section is to define the largest indication as the integration of the tail of the distribution to one complete indication. Other definitions have been proposed. One proposed by ABB-CE defines the limiting indication from Monte Carlo analyses as the median of the distribution ofminimum burst pressures per SG sample. This definition does not uniquely define how the 95%/95% confidence levels are satisfied. It is shown by the analyses of this section that this definition is less conservative for the Farley-1 analyses than obtained using the proposed largest indication obtained by integrating the tail to a whole indication and comparing this indication with a structural limit at 95%/95% confidence. The proposed definition oflargest indication can readily be applied when burst is a function of a single variable such as the average depth applied in this section.
When burst is a function oflength and depth, it is difficult to define a largest indication in terms other than lowest burst pressure. Thus, a generic method for defining the largest indication at 95%/95%
confidence is required.
Calculations are also included in this section for the probability of burst per SG at steamline break (SLB) conditions. For Farley-1, the pressure operated relief valves (PORVs) are operable per the requirements of NRC GL 95-05 and provide the maximum pressure for the SLB event. With the PORV setpoint of 2350 psia and a 3% uncertainty on the opening pressure, the SLB pressure differential (APsts) is 2405 psi. This pressure differential is used for the calculations of SLB burst probability per SG in this report.
A demonstration of the adequacy of the Monte Carlo techniques for predicting the EOC conditions can be obtained by projecting the BOC-15 indications to EOC-15 and comparing the predicted average depths and throughwall indications with the actual results found in the inspection. BOC-15 depths for indications not detected at the EOC-14 inspection were obtained by a review of the EOC-14 bobbin data for indications found at EOC-15 and converting the bobbin voltages to depths using the depth / voltage correlation of Section 5. These depths are also used to develop the growth distribution given in Section
- 6. The BOC depth distribution properly combined with the growth distribution should predict the EOC-15 indications. The use of Monte Carlo analyses for this combination demonstrates the general adequacy of the techniques for combining depth distributions, NDE uncertainties and growth to predict the limiting EOC-15 indications. Table 8-1 compares the projected and actual outage results for the largest average depth and for the number of throughwall indications for SG B. The excellent agreement between the Monte Carlo projections and actual results supports the adequacy of the projection methods.
This is an expected result based on the wide acceptance and other demonstrations of adequacy for Monte Carlo techniques in tube integrity analyses.
Q tubeint\\ala98\\FI Final Freespan oA-8. doc-04/21/99 8-2
8.2. Structural Acceptance Limit A burst correlation for partial depth cracks was developed in Reference 10.4. The correlation is applied to develop a lower 95% confidence level, EOC structural acceptance limit (alternately can be called the structural limit) that includes the combination of burst correlation uncertainties and material properties.
This structural limit is then compared to the largest projected EOC indication that is developed from the BOC indications (undetected indications plus new indications), NDE sizing uncertainties and growth.
For Farley-1, a material property distribution was developed for the heats found to have freespan ODSCC indications. Pulled tubes with SCC indications have consistently been found to have higher flow stresses than the mean of the overall distribution of Westinghouse tubing and higher than the material certifications for the heats of the pulled tubes. That is, tubes that crack tend to have high flow stresses. A conservative (lower range of flow stresses) distribution of material properties for tubes subject to freespan ODSCC can be obtained as the flow stress distribution from the material certifications for heats found in Farley-1 to have freespan ODSCC. This conservatively ignores the increases in flow stresses found in pull tubes above that given for the material certifications. Based on the heats for tubes found to have freespan ODSCC at EOC-15 (60 heats), the mean flow stress is 79131 psi at room temperature with a standard deviation of 2645 psi and a LTL value of 74045 psi. This compares to the distribution for all Westinghouse heats in 7/8" tubing that has a mean of 75470 psi with a standard deviation of 3500 psi and a LTL value of 69225 psi.
For the assessments of this section, the acceptance or structural limit is defined as the lower 95%
probability at 95% confidence for the combined uncertainties of the burst correlation and material properties. This is approximately equivalent to the GL 95-05 definition that applies the burst correlation at the lower 95% prediction interval and lower tolerance limit (LTL) material properties. Statistically combining the uncertainties, as used in this report, leads to a slightly less conservative definition of the structural limit than the GL 95-05 application of separate confidence limits on the burst and material property correlations. The structural limits given for average depth as a function oflength were developed for a burst pressure of 3APso = 4308 psi and for 1.43APsts f 3657 psi (safety valves) and o
3436 psi (PORVs).
For the deterministic assessments of this section which are. based upon average depth, a bounding crack length is applied as the largest crack length found for EOC-15 indications that challenged structural integrity. Since the detection threshold for the EOC-15 inspection is lower than that for the EOC-14 inspection, the flaw sizes at EOC-16 can be expected to be smaller than that at EOC-15. As discussed in Sections 5 and 6, the applicable crack length and average depth for ODSCC are the effective length and average depth associated with the critical length that results in the lowest burst pressure for the crack.
The effective length and associated average depth are obtained by searching the total crack length for the partial length that results in the lowest burst pressure. The effective lengths for the largest indications found at EOC-15 are developed in Table 6-1. From this table, it is seen that a crack length of 0.78 inch bounds the effective crack length for the largest EOC-15 indications. Based on this result, a crack length of 0.80 inch is applied as the bounding crack length for burst at EOC-16. From Figure 5-12, a length of 0.80 inch is very close to 95% cumulative probability for the burst length distribution. Thus, the 0.80 inch length can be interpreted as bounding the largest EOC-15 structurally limiting indications and as the 95% cumulative probability of burst effective crack lengths from EOC-15.
For a crack length of 0.80 inches, an average crack depth of 67.0% satisfies the 3APso = 4308 psi burst margin requirement at 95%/95% confidence for detenninistic analyses. The structural acceptance limit Q:tubeint\\ala98\\FIFinal Freespan oA-8. doc-04/21/99 8-3
is then an average depth of 67.0% and a crack length of 0.80 inches. The 1.43APsts = 3436 psi stmetural acceptance lirnit is found to be 76.3% average depth.
8.3 Sensitivity Analyses for Cycle Lengths as a Function of Confidence Levels on POD and Growth Projections to EOC-16 are a function of the undetected indications at the EOC-15 inspection and the growth rate. The POD distribution as a function of depth is developed in Section 5 and the growth distribution is developed in Section 6. Since POD is a strong function of average depth, a singular detection threshold cannot be defined for the EOC-16 projections and a relatively simple EOC-16 projection cannot be performed. A statistical projection to EOC-16 is required and is performed by Monte Carlo analyses as described in Section 8.4. The general sensitivity of acceptable cycle lengths can be developed as a function of POD values with associated depths and confidence levels on the growth distribution. For example, at a POD of 0.90 (logistic POD), the maximum depth would be 63%
and the average depth would be 53% based on the POD correlations of Figures 5-1 and 5-2. For these depths, the times to grow to throughwall for potential leakage and to 67.0% average depth for satisfying 3APso = 4308 psi can be determined for assumed confidence levels on the growth distribution. By varying the POD value, the sensitivity of the cycle length to the POD value can be determined. For estimating the time to potential leakage, the same method is applied but using the POD for maximum depth and the growth rate distribution for maximum depth. This sensitivity assessment does not include NDE uncertainties and is only used to demonstrate that acceptable cycle lengths can be obtained if relatively low probability (5-10%) undetected indications are combined with low probability high growth rates.
The estimated times to throughwall indications for varying POD values and cumulative probability distribution (CPD) values for maximum depth growth are shown in Table 8-2. For Case I with a POD =
0.90, maximum depths less than 63% would be detected and this depth is combined with the growth at 95% CPD. This case leads to a time to potential leakage that exceeds the planned operating cycle length of 1.12 EFPY, From Case 2, it is seen that the time to throughwall is about a year if a POD = 0.95 is used in combination with growth at 95% CPD. Since there are only a few undetected indications above a 63% depth for POD = 0.90, there is statistically a very low likelihood that the largest growth would occur on the largest indications left in service. The statistically expected result would be that the largest growth would occur near the depths associated with the peak in the population ofindications at BOC conditions rather than at a large missed indication. For the shorter cycle 16 length (1.12 vs.1.29 EFPY for cycle 15), a throughwall crack with leakage would not be expected for an indication below about 70% maximum depth for growth at 95% CPD. The results for Cases 2 and 3 indicate that leakage is possible during the operating cycle but would require the very low likelihood of one of the largest growth values occurring on one of the largest undetected indications.
The corresponding estimates for the time to the structural limit are shown in Table 8-3. For Case I with a POD of 0.80 corresponding to 44% average depth and a growth CPD of 95%, the estimated time to the structural limit is about 0.98 EFPY. The case 2 and 3 results combining relatively high POD and growth probabilities are consistent with the largest EOC-16 indication projected by Monte Carlo analyses in Section 8.4.
As noted above, the analyses given in Tables 8-2 and 8-3 do not permit a cycle length calculation and are provided only to show sensitivity to missed indication levels and growth rates. The most reliable calculation is based upon Monte Carlo projections to obtain the largest EOC average depth.
Q:tubeintiala9871 Final Freespan oA-8. doc-04/21/99 8-4
l l
l 8.4 Acceptable Operating Cycle Lengths Based on Largest Projected Indications Utilizing Monte Carlo Analyses As discussed in Section 8.2, the 95* percertile EOC average crack depth to meet the 3APso burst margin is 67.0% for a bounding EOPC crack length for burst of 0.80 inch. For these analyses, EOPC is used to designate the end of the partial cycle length that satisfies the deterministic structural limit and l
may be an operating period shorter than the planned full Cycle 16 EFPY. The more conventional term l
mid-cycle is not used as the partial cycle length may be considerably longer than half the planned Cycle l
16 length of 1.12 EFPY. The best estimate of the EOC crack depths requires a statistical combination of the depth, NDE uncertainty and growth distributions which can best be obtained by Monte Carlo I
analyses. The projected EOC average depth is then compared with the structural limit. The cycle length is varied to obtain a projected EOC average depth equal to the structural limit which then defines an l
acceptable operating period. This methodology is a deterministic /probabilistic analysis using the terminology of the draft EPRI tube integrity guidelines of Reference 10.3. The Monte Carlo methodology described in Reference 10.5 to estimate EOC bobbin voltage for ODSCC indications at TSPs in accordance with Generic Letter 95-05 was applied. The calculation consists of adjusting the j
EOC-15 depti, distribution using POD to account for undetected indications, adding new indications appearing during a operating period, adjusting the resulting depth distribution for NDE uncertainties and projecting the indication growth over the operating period. Since indication growth is considered l
proportional to operating time, the limiting tube conditions occur at the end of any given time period. The following input quantities are used to perform the analysis.
l
- 1. EOC-15 depth distribution (Section 5.6, Figures 5-10 and 5-11)
- 2. Probability of detection (Section 5.1, Figures 5-1 and 5-2)
- 3. NDE uncertainty (Section 5.2, Table 5-4) l l
- 4. Total number ofindications expected in the operating period (Section 7)
- 5. Growth distribution (Section 6) 1 I
The number ofindications not detected during an inspection based on repair of all confirmed indications l
is determined using the following equation:
l 0
l N_s _ Nu-s
- Nn,.ns poo a
\\
l Where N4eteci,a represents number ofindications detected, POD represents probability of detection and l
N,, paired represents number ofindications repaired. POD is a function of depth so that the above equation
{
is eppW d to either a single indication or a narrow depth bin with more than one indication of nearly the same depth.
For simplicity, all new indications appearing during the operating period were assumed to appear at the beginning of the period and the size of the incipient indication was taken to be the detection threshold for the bobbin coil (about 0.12 volt, equivalent average depth 28%). This lower detection threshold i
Q tubcint\\ala98'JIFinal Frecspan oA-B. doc 04/21/99 8-5 1
f I
1 represents the smallest indications reported in the EOC-15 inspection. Indications larger than this l
threshold value are included as missed indications per the above equation. Since there are no detected indications below this threshold value, the POD adjustment of the above equation does not adjust this range for missed indications and the indications below the threshold are classed as new indications. The new indications, along with the missed indications, are assumed to grow over the entire operating period. The number of new indications included in the analysis was adjusted so that the total number of l
EOC-16 freespan ODSCC indications is equal to about 120, which bounds the expected number of l
indications in any SG at EOC-16 based on the Section 7 estimates. From Table 7-1, the number of l
indications expected at EOC-16 would be 120 at 95% confidence with a best estimate of about 104 l
indications. For a cycle length of about 0.79 EFPY, the number of EOPC indications would be expected to be about 89 at 95% confidence. The NDE uncenainties are developed in Section 5.2.
The recommended POD discussed in Section 5.1 is the logistic distribution and the logistic distributions of Figures 5-1 for average depth and 5-2 for maximum depth are used in the tube integrity analyses.
Growth rate distributions presented in Section 6.3 were used. The Monte Carlo analyses are performed for SGs B and C to determine the limiting SG at EOPC or EOC-16. The largest projected EOC-16 average depths projected using the Monte Carlo method are shown in Table 8-4. Since the Monte Carlo projections yield a tail for the depth distribution, the largest indication is defined as the tail integrated to I
one whole indication.
l Based on 67.0% EOC-16 average depth for meeting the 3APuo burst limit, the acceptable operating l
cycle lengths for the deterministic projections are:
l SG B -0.73 EFPY (265 EFPD,8.7 months) l o
SG C -0.79 EFPY (290 EFPD,9.5 months) o These results show that SG B is the limiting SG for Cycle 16 as was also the case based on measured indications at EOC-15. Cycle lengths in the Monte Carlo calculations were varied until the projected i
EOC average depths were at the structural acceptance limit of 67.0% to define the acceptable cycle length meeting the burst margin requirements. Figures 8-1 to 8-4 show the BOC-16 and projected l
EOPC (0.73 EFPY) and EOC-16 (1.12 EFPY) average depth distributions for SGs B and C. The BOC-
)
l 16 distribution is comprised of new indications introduced at BOC at 28% average depth and undetected indications resulting from the application of the POD to the detected indications at EOC-14 as described above. In SG B, there were 75 + Point confirmed indications identified for the EOC-14 inspection and i
application of the POD results in 18 undetected indications. Then,102 new indications were added to obtain the estimated 120 indications at 95% confidence for EOC-16. In SG C, there were 65 + Point confirmed indications identified for the EOC-14 inspection and application of the POD results in 16 undetected indications. Then,104 new indications were added to obtain the bounding 120 indications l
for SG C at EOC-16. Summing the largest BOC indications to one whole indication yields 53.3% and 51.8% BOC average depth indications for SGs B and C, respectively. The largest single projected EOPC average depths at 0.73 EFPY are 67.0% (cycle length set to yield this value) for SG B and 65.9%
for SG C. The largest single projected EOC-16 average depths at 1.12 EFPY are 75.6% for SG B and 75.0% for SG C. Thus, the accident condition detenninistic burst margin of 1.43APsLB is satisfied by the largest projected indication after full cycle operation.
I Q:tubeintWa9811 Final Freespan oA-8 doc-04Gl/99 8-6
Monte Carlo analyses were also performed to determine whether or not the full cycle operating period would lead to the largest indication reaching 100% depth as an estimate of potential leakage. These calculations were performed for SGs B and C and used maximum depth based EOC-15 distributions, POD, NDE uncenainties and growth rates. It is seen from Table 8-4 that the time for the largest indication to reach 100% depth exceeds the planned operating period of 1.12 EFPY. This result indicates a low likelihood of significant leakage in Cycle 16 for the POD and growth distributions developed in this report 8.5 Sensitivity Analyses for Operating Cycle Length The results given in Section 8.4 and Table 8.4 are based on the recommended growth distribution obtained by combining the SGs B and C growth distributions to obtain the best estimate growth distribution and the number of Cycle 16 indications found from the Section 7 multi-cycle analyses. This section provides sensitivity analyses for the operating cycle length dependence upon these two parameters. The results are given in Table 8.5. From this table, it is seen that both the use of SG B growth rates and increasing the number ofnew indications to obtain 180 total indications have a negligible influence on the projected average depth at 0.73 EFPY. As would generally be expected, the projected average depths are somewhat more sensitive to a change in the POD or the NDE sizing uncertainties as also shown in Table 8.5. Use of the Kunin maxima POD (See Section 5.1) and increases in the NDE uncertainties lead to increases in the average depth of about 3% at the end of 0.73 EFPY operation. These results show only modest sensitivity to these changes. The Kunin POD was shown in Section 5.1 to have larger uncertainties than the logistic POD, which was the basis for applying the logistic function for the POD. The increased NDE uncenainties are typical of that obtained using only + Point sizing rather than applying the bobbin voltage to depth correlation for indications with < 0.5 + Point maximum voltage. The reference POD and NDE uncertainties have a stronger technical basis than the attematives evaluated in Table 8-5.
8.6 Estimated Full Cycle Average Depths and SLB Burst Probabilities The reference and best estimate of the EOC-16 conditions after full cycle operation are given in Section 9 based on projecting crack depth profiles to the EOC. The use of average depths, as applied in this section, tends to be conservative due to the bounding assumption that all crack lengths have a 0.80 inch burst effective length at EOC conditions. Projections to EOC-16 using average depth distributions were made to obtain estimates of the average depths and SLB burst probabilities as an alternate methods comparison with the Section 9 results. Table 8-6 shows projections for both the 265 day and full operating cycles. Results are given for the largest average depth indication and for the SLB burst probability based on PORV availability. As previously discussed, an alternate deterrninistic acceptance criterion proposed by other organizations is based on applying the median of the minimum burst pressure per SG sample for comparison with the 3APuo burst margin guideline. The results of Table 8-6 show that the largest indicadon methodology of this section is more conservative than this alternate acceptance criterion in that the alternate criterion shows addition margin against 3APso = 4308 psi at the end of the 0.73 EFPY operating period.
Projected SLB burst probabilities given in Table 8-6 indicate about a 3.8x10 burst probability after full 3
cycle operation. This is below the acceptable value of 2.4x10 for Farley-1 based on satisfying NEI 97-2 06 performance criteria. Although 3APso burst margins are not obtained after full cycle operation, the Q tubemtiala98\\FIFinal Frecspan oA-8 doc-04/21/99 8-7
largest average depth of 75.6% is less than the 76.3% average depth structural acceptance value for 1.43APsts = 3436 psi at 95%/95% confidence as given in Section 8.2. These results indicate minimal risk of a tube rupture for operation to EOC-16.
8.7 Conclusions for Deterministic /Probabilistic Cycle Length Evaluation Conclusions from the deterministic /probabilistic cycle length evaluations of this section are:
'A cycle length of at least 0.73 EFPY is acceptable for Cycle 16. At the end of the 0.73 EFPY e
operating period, the largest projected average depth satisfies the 3APso = 4308 psi burst margin requirement of an average depth of 67.0% for a bounding burst crack length of 0.80 inch. The 0.80 inch length bounds the largest burst lengths found for structurally challenging indications at the EOC-15 inspection.
Projections of average depths to full EOC-16 operation (1.12 EFPY) indicate that average depths would satisfy a 1.43APsts burst pressure margin and the burst probability would be less than the performance criteria requirements ofNEI 97-06.
A cycle length of about 0.82 EFPY is consistent with a postulated missed indication of about 53%
average depth (corresponds to about 0.90 POD) with growth at the 90% cumulative pmbability level.
This result supports the 0.73 EFPY of operation to 67% average depth even with the low likelihood of a low probability growth combined with a low probability undetected indication.
A low likelihood is expected for the maximum growth to occur for one of the largest indications undetected at the EOC-15 inspection. Thisjudgement is consistent with typical Monte Carlo projections for which the largest growth tends to occur for indications near the highest frequency of the BOC distribution since the larger number ofindications has a higher likelihood of obtaining the low probability, high growth rates.
The structural limit of 67.0% depth for a crack length of 0.80 inches is based on a burst correlation e
developed by correlating measured burst pressures with average depths from crack length profiles obtained from available pulled tube and laboratory specimen PWSCC and ODSCC axial crack data.
The correlation fits analytical formulations of burst pressures to the measured data. The resulting correlation then permits calculation of burst pressures from length / average depth or crack profiles with dermed uncertainties for the correlation. The defined uncertainties then permit confidence statements to be made for the calculated burst pressures.
Q:tubeintiala9871 Final Freespan CA-8. doc-04/m99 8-8
Table 8.1. SG B Demonstration of Monte Carlo Results vs. Actual Results at EOC-15 Cciculation and Largest Indication Average Depth Number of ThroughwallIndications Applied Growth Distribution Projected Actual Projected Actual Average Depth.
98.7 %
93.8%'"
NA NA Recommended SG 97.2%*
A+B Composite Growth Average Depth.
95.8 %
NA NA SG B Growth Maximum Depth.
NA NA 3"'
3"'
Recommended SG A+B Composite Growth for Max.
Depth Notes:
- 1. Pulled tube R25C51 destructive exam result. Tube plugged at 8/98 refueling outage. The indication would have been slightly deeper ifleft in service to EOC-15.
- 2. Tube R10C2 in SG C over burst effective length of 0.57 inch based on + Point analysis at EOC-15.
- 3. Three indications predicted to be throughwall based on largest single indication, i.e. integration of the tail of the Monte Carlo distribution to one whole indication.
- 4. Three indications found to be throughwall with two indications in R25C51 and one indication in R43C42 confirmed by in situ testing. No other indications were found to leak during in situ testing in SG B.
i Q tubcintiala98\\FIFinal Frecspan oA-8. doc-04/21/99 8-9
Table 8-2. Conservative Deterministic Sensitivity Estimates of Time to Leakage BOC Depth Growth in Maximum Depth Time to 100%
Case Depth
- Comments POD Maximum Cum.
Depth /EFPY EFPY Months Depth Prob.
1 0.90 63 %
95 %
28 %
1.32 15.8 2
0.95 72 %
95 %
28%
1.00 12.0 3
0.95 72 %
100 %
46%
0.78 9.3 Very low likelihood combination of undetected indication and large growth.
Notes:
- 1. The time to grow to a throughwall crack is taken as the time to initial leakage. Leakage experience indicates that leakage would not occur until the throughwall length is at least 0.1 inch so the estimated time to throughwall is a conservative estimate of time to leakage.
Table 8-3. Conservative Deterministic Sensitivity Estimates of Time to Structural Limit (67.0% Average Depth)m Case BOC Depth Growth in Average Depth Time to Structural Limit
- POD Average Cum.
Depth /EFPY EFPY Months Comments Depth Prob.
1 0.80 44 %
95 %
23.5 %
0.98 11.7 2
0.85 48%
95 %
23.5 %
0.81 9.7 Estimated time 3
0.90 53 %
90 %
17 %
0.82 9.9 consistent with largest EOC-16 depth from Monte Carlo analysis 4
0.90 53 %
95 %
23.5 %
0.60
7.1 Notes
- 1. The structural limit is 67.0% average depth for a bounding EOC crack length for burst of 0.80 inch.
Q tubeint\\ala98\\FIFinal Freespan oA-8. doc-04/21/99 8-10
r Table 8-4 Farley Unit-1: Deterministic Cycle 16 Operating Period Based on Projected Free Span ODSCC Largest Average and Maximum Depths l
Cycle 16 SG Operating Largest Dep'th Comments l
Period Indication ( )
EFPY l Months Operating Period to Structural Limit (Average Depth)
Calculations include: a total of 120 indications-0 G
B 0.73 )
8.7(2) 67.0% )
including new indications, the logistic POD for 0
1.12 )
13.4(3) 75.6%(4) average depth, NDE uncertainties with a standard deviation of 7.3% on average depth and the combined SGs B and C average depth C
0.73(2) 9,4(2) 67.0%c) growth distribution.
0 1.12 )
13.4(3) 75.0%(d)
Maximum Depth at Ecd of Operating Period - Potential Throughwall Evaluation 0.73 8.7 78.4 %
Calculations include: a total of 120 indications 0
0 B
1.12 )
13.4 )
87.6 %
including new indications, the logistic POD for maximum depth, NDE uncertainties with a 0.73 8.7 77.4 %
standard deviation of 9.4% on maximum depth 0
0 and the combined SGs B and C maximum C
1.12 )
13.4 )
87.0 %
depth growth distribution.
Notes:
- 1. Largest indication is obtained by integrating the tail of the projected EOC depth distribution to a single (one whole indication) indication.
- 2. Time to structural limit of 67.0% for an EOPC burst crack length of 0.80 inch. This length is based on the longest structurally challerging burst crack length found at EOC-15,
- 3. Full anticipated operating length for Cycle 16 (1.12EFPY)
- 4. The largest projected single indication at EOC-16 satisfies a 1.43 APsts = 3436 psi structural limit i
Q:tubeintWa98'J1 Final freespan oA-8 doc 04/21/99 8-lI
Table 8-5. Farley-1: Deterministic Sensitivity on Input Parameters for SG B Average Depths at 0.73 EFPY Growth Rates and NDE EOPC POD Number of Uncertainties Average Comments Indications Depth
- Logistic
Reference:
Combined
Reference:
67.0%"'
Growth Mean = 0.29%,
120 Indications Std. Dev. = 7.3%
SG B Growth Reference 65.9 %
Negligible sensitivity of 120 Indications average depth on Reference Growth Reference growth distribution and 180 Indications 67.4 %
munber of new
- 89 Indications 66.9 %
indications Logistic POD set to Reference Reference 67.0% at 1.0 at 70% depth 0.76 EFPY Kunin Maxima Reference Reference 69.9 %
Logistic Reference Mean = 1.6%,
70.5 %
Std. Dev. = 9.0%
Table 8-6. Farley Unit-1: Deterministic and Probabilistic Results for Partial and Full Cycle Operation End of Operating Period Alternate Cycle Parameter Deterministic SG Operating Acceptance Comments Period Criteria
- Deterministic"'
Probabilistic"'
Median of Minimum Largest SLB Burst Burst Indication Probability Pressures B
0.73 EFPY 67.0 %
1.2x10 '
4663 psi All deterministic and probabilistic (265 days) criteria are satisfied.
C 0.73 EFPY 65.9 %
1.1x10
4781 psi B
1.12 EFPY"'
75.6 %
3.8x10-'
3789 psi Probabilistic criterion satisfied.
C 1.12 EFPY"'
75.0 %
3.3x10
3854 psi Deterministic results meet (409 days) 1.43APsLB average depth of 76%
and burst pressure of 3436 psi Notes:
- 1. Deterministic criterion is that largest indication must meet 3APso structural acceptance limit of 67.0% average depth.
- 2. Probabilistic criterion is that SLB tube burst probability must be s 2.4x10-2 for APsta = 2405 psi.
- 3. Alternate deterministic acceptance criterion for Monte Carlo analyses is that the median of the minimum burst pressures per SG sample should meet the 3APuo = 4308 psi burst margin
- 4. Full anticipated operating cycle length for Cycle 16 Q.tubeint.ala98\\FIFinal Freespan oA-8. doc-04/21m 8-12
n t
D seioC P
gaO el t g
f iincE F
- aSdt r
E L
n a I
6
~-
- U 5
6 2
9+
1 f
~
o C
~
d O
n
~$
B E
_ 9 M
)-[
n seioC e
t elgaO aP I
b t
g iincB r
D aSdt 9
L n a e
oF iE tc5 c
6 i
d 2 9
zt I
a e
Cn Ci
~
o
- + $
)
S t u
D b
r Oi r
(
t ss
- +h' h
ai pD t
p S h e
v D
3 et e p 1
r e e
m #+$'
FD g
8 a
B e r
g e
Ga v
r S e A
1 v
$+h' yA ed l
r e t
c c Fi
$+h'
.d e
1 r
- P 8
ua ed r n I
i
?'
g C
- ?+ q i
F O I
B 9
9
/
s 1
n 2/
io
$+
4 0
ta 1
co ic d.
I
$+h' d
8-n AO w
n e
ap N
se e
h' l
r F
a
$e*
la n
iF
,' J i -
FI 2
0 8
6 4
2 0
8 4
2 0
\\8 2
2 1
1 1
1 1
9a
\\tn Ujt 1Eg ie b
tu Q
t n
seioC elgaO t
gncE D
aSdt ii P
I 9+
L n a F
E U
5 6
9+#
6 2
1 f
o C
d O
n B
E
- +?
]
n 3
tseioC
[
elgaO t
mncB ii 9+ q ia_ S d t D
oF In a nP E
ita c5 i6 d2 9+*$
nt I
a Cn Coi St
,#+$ h
)
Du Oi b
n L_
r
(
ts ai t
pD p
- +$
g S
e h
D et e p r e e
4 FD 1
a Ce 8
r
- +$
g e
Ga v
r S e A
1 v A
y
$+$
ed lr e t
a c Fi d
. e 2 r r n I
% 4+?q
- P 8ed u a gC
=
i F O I
~
, ?+
H s
M
$+$
1 n
2/
io 40 ta
-c ic o
d.
d 8
In
-A
$+
O we na N
p
$+$
l see rF an iF 0
8 6
4 2
0 8
6 4
2 0
1 F
2 1
1 1
1 1
\\89taa
\\
.W $
ie tn bu t
Q
i' 1II l
{
1l) s eioC {_
n t
elgaO t
g
.C C
, $ +-
iincE aSdt I
D_
r
.6 6
1 L
n a
-_1
_O O
4+*
_B E
B O
_ d' + $
~
_ $+$
6
~
n1
. $+$
o -
iC t
n cO tseioC ciE elgaO t
aSdt g
d r incB
?'
t i
E. 9' 9 n
I
{_
na L
n a I
q Co Ci
~
t S ub Di
, $' 9*'
r Ot h
t s
ai p
rD e
p D
h e p
$' $e S
t g
e e rD
+'
a 5
F 1
r
$' $v e
e 8
B ga A
Gr S
+'
ev E, $' $
A I
sn yd o
e e it lrt a
c ei c
Fd id
.r I
. $+
e n
3P w
e 8 d N
5 4
-$+$
e n r a gC J,
u l
iFO 0
$ + ?'
B 1
m l
q 1
2
/40-c o
_ $+$
d.
8 A
O na
$+
l psee rF a
$+$
Ia n
iF 1W 8
0 8
6 4
2 0
8 6
4 2
o 9
2 1
1 1
1 1
ie tn b
E :: 0 5 3 eo I E u
t Q
t noC se it elgaO iincE
_ 4' gr 6
6 aSdt L
n a 1
1 I
C C
O O
4' #'
B E
3 t_
l 4
r 4' #'
~
~
~
6 n1 t
o -
a C
i t
cO se n t
eC elgi cE g
E, 9+?q r in aO t
i d
aSicB t
n a L
d I
Co In n
Ci t
S ub
)
=
Ot
~
, 9+*
Di r
(
~
s ni h
aD tp
,#$e p
S h D
t e p 6
e e e
rD 1
,# $r g
F 8
a e
Cg e
a Gr e vA S v s
A n
1
-yd io t
c e a
ci dic i
t r
c Fd i
m$ h n
e
. r w
4 P e
8 d N
m$ h e n r a gC
~
iO l
~
F 5
, 9+?q 1
B 0
99 1
l
/
2
/
0 4
-c
, 4' #'
o d.
8
-A O
4 n
a
, 4' # '
l p
se erF a
4 i
_ 4' #
n F
1T 8
2 0
8 6
4 2
0 8
6 4
2 0
9 l
2 2
1 1
1 1
1 a
h tn ie mjgii o IE Z b
i u
t Q
9.0 PROBABILISTIC ASSESSMENTS FOR ACCEPTABLE OPERATING CYCLE LENGTH This section describes analysis methods and results for probabilistic analyses of SLB burst probability and leak rates for Farley-1 freespan ODSCC. The acceptance criteria and methodology are described in Section 9.1 with the input to the analyses summarized in Section 9.2. The methodology used for the probibilistic analyses is based on projecting crack depth profiles including allowances for POD, NDE uncertainties and growth to the end of cycle conditions. This methodology provides "first principle" analysis techniques in that burst and leakage are calculated from the crack profile rather than the more conventional simplifying assumptions of average depth orjudgmental crack shapes. The application of this more detailed analysis methodology is permitted by improvements in + Point coil depth sizing for which the associated NDE uncertainties are developed on a plant specific basis in this report.
Probabilistic results for partial cycle operation are discussed in Section 9.3 and for full cycle operation in Section 9.4. Leak before break considerations for establishing the operational leak rate limit are addressed in Section 9.5. Conclusions from this section are given in Section 9.6.
)
\\
9.1.
Probabilistic Analysis Methods
]
9.1.1. Acceptance Criteria Burst Probability Limits Farley-1 is committed to satisfying the performance criteria of NEI 97-06 which include probabilistic acceptance criteria. NRC guidance on probabilistic criteria are currently a draft document (DG-1074) with the exception of GL 95-05 which defines criteria for ODSCC at TSP intersections. The NEI 97-06 conditional probability limits for tube burst due to postulated accident conditions and unknown degradation mechanisms are:
2
< 510 per year that one tube burst during an accident o
< 2.5 10 2 per year that 2 to 10 tubes burst during an accident o
4
< l 10 per year that more than 10 tubes rupture during an accident o
An allowance of 2.010-2 is conservatively provided for a single burst due to unknown and unanalyzed causes which then leaves an c.llowance of 3.1 10 2 per year for known causes. This allowance bounds the operating plant historical burst frequency for all causes. For Farley-1, sludge pile and circumferential cracking in the expansion transition are the more significant degradation mechanisms that are not analyzed for burst probability. Both of these mechanisms showed large margins against burst at the EOC-15 inspectionfReference 10.1) and would be expected to have collective burst probabilities much less than 10 In addition to freespan ODSCC, ODSCC at TSP intersections contributes to the burst probability. From the Farley-190 da7 report (Reference 10.6) for ODSCC at TSP intersections, the single tube burst probability is 0.610 for SG C and 0.3x10 2 for SG B. The 4
probability of more than one tube burst is < 10 for ODSCC at TSP intersections. The value for SG C is rpplied to the conditional probability limit for all SGs to obtain a single tube burst probability limit of 2.510-2 (statistical combined components,2.410-2 by subtraction) for freespan ODSCC. For additional conservatism, the single tube burst probability limit is applied as the limit for burst of one or more tubes, i.e., essentially the probability that any tube bursts. The limit on 2 to 10 bursts is set at one-half the limit for a single tube burst and the limit on more than 10 bursts is not relevant (does not occur) for these analysis.
Q:tubeint&da987inal oA7inal Freespan oA-9. doc-04/21N9 9-1
The conditional probability limits for freespan ODSCC tube burst due to postulated accident conditions are then given by:
e.< 2.5x10-2 per year that one or more tubes burst during an accident
< l.2x10 2 per year that 2 or more tubes burst during an accident e
The limiting postulated accident for Westinghouse Model 51 SGs is the steamline break (SLB) event.
To define the limiting pressure differential, it is assumed that operator actions are delayed until the primary pressure reaches the relief valve setpoints. With allowances of 3% for uncertainty on the valve opening pressures, the SLB primary to secondary pressure differentials (APsts) are 2405 psi for the PORVs and 2560 psi for the safety valves. For Farley-1, the pressure operated relieve valves are operable per the requirements of NRC GL 95-05 and provide the maximum primary pressure for the SLB event. With the PORV setpoint of 2350 psia and a 3% uncertainty, APsts = 2405 psi. This value is used as the accident condition maximum pressure differential for the calculations of SLB burst
- probabilities and SLB leak rates.
Allowable Leakage Limits Farley Nuclear plant currently has a leakage limit of 11.8 gpm approved for ARC at TSP intersections.
This limit is based on not exceeding a small fraction of 10 CFR 100 limits (10%) in the event of a main steam line break. Based on FNP's current projected SLB leakage from ODSCC at TSP intersections at EOC-16 of 8.2 gpm, margin exists of 3.6 gpm that could be applied to an allowable SLB leak rate limit for freespan ODSCC without changing current NRC approved offsite dose limits. For the probabilistic approach, the projected SLB leak rate at EOC-16 for freespan ODSCC must be less than this value when evaluated at 95% probability,95% confidence.
An additional requirement is imposed on the freespan SLB leak rate for severe accident considerations.
This requirement limits the nominal SLB leak rate from all freespan indications to less than or equal to 1.0 gpm. Freespan ODSCC is the only Farley-1 mechanism potentially contributing to this limit.
Therefore, it must be demonstrated that the nominal or best estimate for the total freespan SLB leak rate must be less than 1.0 gpm. It can be noted that this requirement is satisfied with a large margin if the SLB leak rate calculated for 95/95 confidence is less than 1.0 gpm.
9.1.2. Description of Probabilistic Analysis Methods The probabilistic analysis methods of this section are based on the use of crack depth versus length profiles obtained from + Point analyses of the freespan ODSCC indications. This methodology provides "first principle" analysis techniques in that burst and leakage are calculated from the crack profile rather than the more conventional simplifying assumptions of average depth orjudgmental crack shapes.
Since all + Point confirmed indications are repaired, the BOC population is comprised of undetected indications and new indications (truly new indications and indications below the bobbin coil detection threshold or smallest detected indication). The undetected population is obtained using the GL 95-05 methodology that leaves (1/ POD - 1) indication in service for each detected indication, where the POD is depth dependent. Monte Carlo analyses are used to incorporate NDE uncertainties and growth for i
projecting the EOC indication size, burst pressures and leak rates.
i l
The steps of the analysis process can be summarized as follows:
l Q:tubeinfela\\ala98\\ Final oA\\ Final Frecspan oA-9. doc-04/21/99 92
BOC population for undetected indications o
- Each + Point indication evaluated for effective average depth and maximum depth. PODS (developed as functions of effective average depth for burst and maximum depth for leak) evaluated at the depths of the indication. For burst pressure analyses, the POD for average depth is applied and for leakage analyses, the POD for maximum depth is applied.
- Application of(1/ POD-1) leads to a fraction of an indication or more than one indication for each EOC-15 detected indication. The undetected indications are assumed to have the same BOC crack profile as the EOC-15 indication for which the POD is applied. The assumption of the same crack length for the undetected indications is particularly conservative since the more likely missed indications at a given depth are the shorter indications. Since the lengths used are EOC lengths, there is no further need to grow the cracks in the length direction.
- Separate populations of crack profiles are carried through the analysis for burst pressure and for leakage due to differences in PODS, NDE uncertainties and growth rates between average and maximum depth. This separation permits more direct application to the analysis of the parameters affecting leakage and can be expected to improve the accuracy ofleakage predictions compared to applying PODS, NDE uncertainties and growth based only on average depth.
BOC population for new indications o
- The total number ofindications expected for Cycle 16 was developed as 120 indications based on the multi-cycle analyses of Section 7. The number of new indications is defmed to obtain a total of 120 undetected plus new indications. Each new indication crack profile is obtained by sampling distributions of average depths, crack length and crack shape. The crack shape distributions used are examples from the Farley-1 pulled tubes and have been selected to provide maximum to average depth ratios typical of the correlation developed in Section 5.4. A crack shape sample is adjusted by applying the samples for average depth and length.
- The sampling process defines BOC new indication crack profiles for each SG simulation.
Application ofNDE uncertainties o
- NDE uncertainties on average or maximum depth and crack length are applied to the BOC populations of undetected and new indications. The NDE uncertainty distributions developed in Section 5.2 are used.
- The NDE uncenainty on average depth (same process for maximum depth) is uniformly applied to each point in the crack depth profile. That is, the profile is assumed to grow in depth with the same shape as the initial profile.
- The NDE uncertainty on crack length is applied by increasing or decreasing the spacing of the two profile points near the ends of the crack since the length uncertainty occurs at the crack tips. A negative length uncertainty is conservatively applied by not decreasing the length by more than the spacing between the two end points of the profile.
Application of growth o
- For each indication, the cumulative probability distributions for growth in average depth and in maximum depth are sampled. The growth is then applied to each point in the crack depth profile.
- Separate profile populations are evaluated using average depth growth for burst analyses and maximum depth growth for leakage analyses.
- As noted above, growth in length is not applied since the crack profiles used are already EOC lengths and POD is conservatively assumed to be independent oflength.
Burst pressure calculations o
- The above steps lead to a projected EOC crack profile for each indication in the population.
Profiles based on average depth PODS, NDE uncertainties and growth are used for the burst analyses. The crack profile is then searched for the lowest burst pressure between any two points in the profile. The predicted burst pressure is then obtained by sampling the burst correlation at the length and effective average depth for the indication.
Q:tubcinnalsWa98' final oA'Einal Freespan oA-9. doc-04/21/99 9-3
1
- The predicted burst pressure is then compared to the allowable value (APsLB = 2405 psi for current analyses) to determine if the indication is a " burst" for inclusion in the burst probability analyses.
Leak rate calculations
- The above steps lead to a projected EOC crack profile for each indication in the population.
Profiles based on maximum depth PODS, NDE uncertainties and growth are used for the leakage analyses. The crack profile is then searched to determine the crack length at 98% throughwall for calculation ofleakage. The 98% depth provides an allowance of 2% wall thickness for tearing of the crack at SLB conditions. The 2% allowance is based on the largest found in extensive leak testing of ODSCC and PWSCC corrosion cracks. Although postulated crack shapes such as bathtub flaws would permit tearing of more than 2% wall thickness, the irregular shapes and ligaments of corrosion induced cracks have been found to limit wall thickness tearing. The 98%
corrosion depth is the smallest corrosion depth found to leak in more than 100 leak tests of corrosion cracks.
- Given the EOC "throughwall" length, the leak rate is calculated based on a correlation of corrosion specimen leak rates with calculated results from the Westinghouse "CRACKFLO" code. The correlation is developed for leak rates calculated for throughwall crack lengths. The CRACKFLO results adjusted to " truth" by the test to CRACKFLO regression curve represent the mean SLB leak rates and the uncertainties are obtained from the correlation of measured leak rates with the CRACKFLO results.
- Leak rates are summed over all Monte Carlo sample indications in a SG population to obtain a sample of the total SG leak rate. The multiple SG samples then lead to a distribution of SG oak rates.
Burst probability and SLB leak rate
- The SLB burst probabilities are calculated by counting the number of Monte Carlo SG samples that have 1,2 or more bursts (burst pressure < APsLB). The number of SGs with a burst divided by the number of SG samples adjusted for the confidence level (95%) yields the burst probability.
- The distribution of SG SLB leak rates is evaluated at 95% probability and 95% confidence to obtain the reported SLB leak rate.
9.2.
Input Distributions for Probabilistic analyses Input distributions for the analyses are described in Sections 5 and 7. The following input distributions are used in the probabilistic analyses:
EOC-15 + Point crack depth profiles - Section 5.5 Logistic POD distributions for POD as a function of average and maximum depth - Section 5.2, Figures 5-1 and 5-2 NDE uncertainties on average depth, maximum depth and length - Section 5.1, Table 5-4 Growth rates for average and maximum depth - Section 6, Figures 6-1 and 6-2 New indication average depth, length and shape distributions - Sections 5.4, 5.7 and 5.6, e
respectively Number of newindications-Section 7 As discussed in Section 8.2, the burst correlation for partial depth cracks developed in Reference 10.4 is used for the burst pressure analyses of this report. This correlation was developed by fitting functional relations for burst pressures as a function of average depth and length to measured burst pressures. The calculated burst pressures for this correlation are obtained by searching the destructive exam depth Q:tubeint\\ala\\ala98Tinal oATinal treespan oA-9. doc-04/21/99 9-4
. profile for the lowest burst pressure for correlation with the measured data. This correlation then permits determination of uncertainties in the burst pressure calculations for use in the Monte Carlo 3
analyses. A linear correlation between the measured and calculated burst pressures provides the best i
estimate burst pressure and the uncertainties are defined relative to this linear correlation.
The SLB leak rate model follows the same process as used for burst pressures. Leak rates calculated using the Westinghouse CRACKFLO Code for a given destructive exam throughwall length are linearly correlated to the measured leak rates for each specimen. The leak rate uncertainties are then defined relative to the correlation between measured and calculated leak rates. The development of this'SLB leak rate correlation is also given in Reference 10.4.
i 9.3.
Analysis Results for Partial Cycle Lengths Probabilistic analyses were performed for the 0.73 EFPY partial cycle shown in Section 8 to satisfy the 3APso burst margin guideline. The results are given in Table 9-1. It is seen that SG B is more limiting than SG C with a SLB (AP = 2405 psi) burst probability of 3.1x10'3 compared to 1.8x10 for SG C.
These results are in good agreement with the results given in Table 8-6 for which a SG B burst 3
probability of 1.2x10 was obtained. The Section 8 results are based on the assumption of a bounding 0.8 inch length for all indications and projecting average depth distributions for the burst probability analyses. The results of Table 9-1 are also consistent with the results of the multi-cycle analyses given 3
in Section 7 for which a SLB burst probability of 1.5x10 was obtained for a cycle length of 0.79 EFPY using a log logistic POD. It is expected that the multi-cycle analyses would be less conservative than the methods of this section since all distributions are applied probabilistically in the analyses.
I No SLB leakage is predicted at 95/95 confidence for the 0.73 EFPY operating cycle. This result is consistent with the Sections 7 and 8 results for multi-cycle and maximum depth analyses.
For the Farley-1 freespan ODSCC indication distributions, it is concluded that satisfying the 3APuo burst margin guideline at 0.73 EFPY corresponds to a SLB burst probability of about 3.1x10 and no leakage is expected at SLB conditions.
9 9.4.
Analysis Results for Full Cycle 16 Operation The results of the probabilistic analyses for full Cycle 16 operation of 1.12 EFPY are given in Table 9-1.
The differences in SLB burst probabilities and leak rates between SGs B and C are negligible for 1.12 EFPY. For SGs B and C, the burst probability at 95% confidence for one or more tube bursts is 1.1x10-2 at 2405 psi pressure differential. This result is below the allowable limit of 2.5x10-2 developed in Section 9.1. The burst probability for burst of 2 or more tubes per SG is about 8x10'5. The results of Table 9-1 for SG C are approximately the same as found for SG B.
The predicted SLB leak rate per SG at 95/95 confidence is <0.01 gpm for SG B. This result is well below the allowable limit of 3.6 gpm developed in Section 9.1 based on dosage considerations. Even though this result is at 95/95 confidence, it satisfies the allowable limit of 1.0 gpm on the nominal leak rate for severe accident considerations. The leak rate results for SG C are also very small and lower than those found for SG B. Consequently, the results show that the allowable leak rate limits are satisfied for full cycle operation of1.12 EFPY.
Q:tubeintala\\ala98\\ Final oA\\ Final Freespan oA-9. doc 04/21/99 9-5
The probability that a SG would have one or more indications with burst pressures less than 3APso =
4308 psi was also evaluated. The results for SGs B and C for 1.12 EFPY operation are 0.20 and 0.17, respectively. These results indicate that there is a significant likelihood (about 80%) that the 3APuo burst margin would be ratisfied for the limiting indications after full cycle operation.
Table 9-2 compares the results of the SLB analyses of this section based on crack profile analyses with the multi-cycle analyses of Section 7 and the average depth analyses of Section 8. The full cycle result of 1.1x10-2 for SG B SLB burst probability from the profile analyses is moderately more conservative than the Section 8 (Table 3-6) result of 3.810'3 and the Section 7 result (Table 7-2) of 4.610'3. ' ll A
analyses predict negligible SLB leakage after 1.12 EFPY The Monte Carlo analyses permit an assessment of the worst case burst pressures predicted per SG sample. In this case, worst case is the lowest burst pressure of any Monte Carlo sample in the SG. Statistically, this worst case has a very low likelihood of occurrence although not quantifiable. It has been suggested by some analysts that the median of the worst case burst pressures be used for comparisons with the 3APso burst margin guideline. The comments column of Table 9-2 provides information on the distribution of worst case burst pressures from the profile analyses of this section and the multi-cycle analyses of Section 7. It is seen that the median of the worst case burst pressures is >3APso = 4308 psi for both analysis methods.
The extreme case of the lower 95* percentile of the worst case burst pressures is El.43APst.s for both the profile and multi-cycle analyses, which further supports adequate margins against burst at SLB conditions.
Overall, it is concluded that the probabilistic analyses support full Cycle 16 operation of 1.12 EFPY.
The SLB burst probability of1.1x102 is less than the allowable limit of 2.5x10'2 per NEI 97-06 requirements and the SLB leak rate of <0.01 gpm at 95%/95% confidence is much less than the dose based allowable limit of 3.6 gpm for freespan ODSCC. The SLB leak rate at 95%/95% confidence conservatively satisfies the requirement of 1.0 gpm at nominal conditions for severe accident considerations.
9.5.
Leak Before Break Considerations for Full Cycle Operation The results described above show that SLB burst probabilities and leak rates are acceptable at the end of full cycle operation. The analyses are performed at 95% probability and 95% confidence such that there is a low likelihood of the acceptance limits being exceeded. While not statistically expected, there is a finite probability that the largest growth rates could occur on one of the largest undetected indications at j
the EOC-15 inspection. It is desirable for these low likelihood situations to have leak before break conditions lead to plant shutdown prior to reaching unacceptable structural integrity. For leak before break considerations, acceptable structural integrity for this assessment is based on preventing burst at SLB conditions. The Farley-1 basis for full cycle operation of 1.12 EFPY is an acceptable burst probability at SLB conditions. A shorter operating cycle length of about 0.73 EFPY is predicted to i
provide 95/95 confidence that the largest indication would satisfy a burst pressure margin of 3APso.
Thus, there is a reasonable likelihood that indications exceeding SLB conditions but less than a 3 APso burst margin will result at EOC-16 conditions. Therefore, it is appropriate that leak before break is established to prevent a burst at SLB conditions rather than to maintain a burst margin of 3APso. The Farley-1 Cycle 15 operating experience has demonstrated leak before break conditions as described below.
The SG B leak rate history for Cycle 15 is shown in Figure 9-1. Points of note from this history are given below:
Q:tubeint\\alainla987inal oA7inal Freespan oA-9. doc-04/21/99 9-6
Possible initial spike of about 2 gpd on July 3 at 1.06 EFPY o
o Additional small leakage spikes of < 10 gpd leading to a spike of about 29 gpd on July 26 at 1.12 EFPY A larger spike of about 59 gpd on August 7 at 1.157 gpd o
Two additional spikes with a peak of about 85 gpd on August 14 at about 1.176 EFPY o
Plant shutdown for unplanned leaker outage on August 16 at 1.182 EFPY o
At the unplanned leaker outage, the leaking tube, which was also the structurally limiting tube, was found to be R25C51. This tube, which was found to have two throughwall indications as described in Section 4, was pulled and burst tested in the laboratory. The upper flaw had a burst pressure of 2633 psi corrected to operating temperature and the lower flaw had a hot burst pressure of 4352. The upper flaw exceeded burst at SLB conditions of 2405 psi for PORV operation as well as 2560 psi for safety valve operation. The lower flaw burst pressure exceeded the Farley-13APso burst margin of 4308 psi.
Consequently, the operating experience up to the August shutdown indicates that shutdown following an operating leak rate spike of about 85 gpd is adequate to maintain a burst pressure exceeding SLB conditions.
Following the return to power after the unplanned leaker outage, Farley-1 had an operating leak rate of about 20 to 25 gpd after initial full power and this leak rate remained approximately constant until the IRIS refueling outage. The leaking tube was identified as SG B, R43C32, which was also the structurally limiting indication found at the 1R15 inspection. This indication was in situ pressure tested and found to have a bladder rupture at 3306 psi adjusted to operating temperatures. Visual exam' ation m
of the indication following in situ pressure testing did not show tearing at the crack tips and the burst pressure may have been higher than the measured 3306 psi. The minimum 3306 psi burst pressure following a 25 gpd leak rate significantly exceeds the SLB burst requirement of 2405 psi.
Consequently, a leak rate limit as low as 25 gpd is not necessary to prevent rupture at SLB conditions.
Based on the Cycle 15 leakage experience, an operating limit for a leakage spike of about 85 gpd is adequate to maintain a burst margin against SLB conditions. The trends of Figure 9-1 show a decrease I
in the leak rate of close to 15 to 20 gpd following the larger leakage spikes in August,1998. Under the assumption that a crack near SLB burst does not continue to grow or lose ligaments between crack segments, it is appropriate to establish a longer term or more steady state leak rate limit following a
{
spike.
l 1
Based on providing some margin relative to the demonstrated 85 gpd acceptable spike, the following operating leak rate limits are recommended for Farley-1 during Cycle 16 operation:
Plant to be shutdown within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a leak rate spike confirmed by chemistry samples to o
exceed 75 gpd.
Plant to be shutdown if the leak rate following a spike above 60 gpd does not decrease to less than o
60 gpd following the leak rate spike. Successive leakage spikes above 60 gpd, but less than 75 gpd, within a seven day span do not require shutdown since the 60 gpd limit is intended to be a longer term leak rate without spikes.
Plant to be shutdown within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a progressive (with spikes) increase in the leak rate o
to greater than 60 gpd confirmed by chemistry samples.
Q1tubeint\\ala\\als9FEinal oA\\ Final Freespan oA-9. doc-04/21/99 9-7 j
j
p The r.bove operating leak r:.te limits are to be administratively implemented for Cycle 16 operation only.
The SGs are to be %placed following Cycle 16 operation.
9.6.
==
Conclusion:==
Probabilistic Analysis Requirements Based on satisfying NEl 97-06 requirements as modified for Farley-1 freespan ODSCC, the coriditional probability limits for freespan ODSCC tube burst due to postulated accident conditions are given by:
< 2.5 10-2 per year that one or more tubes burst during an accident o
< 1.210-2 per year that 2 or more tubes burst during an accident o
These requirements apply'after reductions of the NEI 97-06 510-2 requirement for unknown or unanalyzed mechanisms and for the projected EOC-16 burst probability for ODSCC at TSP intersections.
Farley Nuclear plant currently has a leakage limit of 11.8 gpm approved for ARC at TSP intersections.
This limit is based on not exceeding a small fraction of 10 CFR 100 limits (10%) in the event of a main steam line break. Based on FNP's current projected SLB leakage from ODSCC at TSP intersections at EOC-16 of 8.2 gpm, margin exists of 3.6 gpm that could be applied to an allowable SLB leak rate limit for freespan ODSCC without changing current NRC approved offsite dose limits. For the probabilistic approach, the projected SLB leak rate at EOC-16 for freespan ODSCC must be less than this value when evaluated at 95% probability, 95% confidence. An additional requirement is imposed on the freespan SLB leak rate for severe accident considerations. This requirement limits the nominal SLB leak rate from all freespan indications to less than or equal to 1.0 gpm.
For Farley-1, the pressure operated relieve valves are operable per the requirements of NRC GL 95-05 and provide the maximum primary pressure for the SLB event. With the PORV setpoint of 2350 psia and a 3% uncertainty, an accident condition APst.s = 2405 psi is used as the maximum pressure differential for the calculations of SLB burst probabilities and SLB leak rates.
SLB Tube Burst Probability Overall, it is concluded that the probabilistic analyses support full Cycle 16 operation of 1.12 EFPY.
The SLB burst probability of 1.1x10-2 for SG B based on the reference analyses of this section is less than the allowable limit of 2.5x10-2 per NEl 97-06 requirements.
The methodology used for the probabilistic analyses is based on projecting crack depth profiles including allowances for POD, NDE uncertainties and growth to the end of cycle conditions. This methodology provides "first principle" analysis techniques in that burst and leakage are calculated from the crack profile rather than the more conventional simplifying assumptions of average depth or judgmental crack shapes. The application of this more detailed analysis methodology is permitted by improvements in + Point coil depth sizing for which the associated NDE uncertainties are developed on j
a plant specific basis in this report.
l Q1ubeint\\ala\\ala9871nal oA7inal Freespan oA-9. doc-04mS9 9-8
The full cycle result of1.1x10 2 for SG B SLB burst probability based on using crack depth profiles is
- moderately more conservative than the Section 8 result of 3.8x10 based on use of average depths and the Section 7 result of 4.6x10~3 based on multi-cycle, complete probabilistic analyses. Consequently, the results of three independent analysis methods support full cycle operation. It can also be noted that the Cycle 15 operating leakage was < 10 gpd until the spike of about 29 gpd at 1.12 EFPY. This low leakage level provides additional support for full cycle operation of 1.12 EFPY even assuming no significant improvements in the inspection at 1R15 over that at IR14.
SLB Leak Rates The SLB leak rate of <0.01 gpm at 95%/95% confidence for full cycle operation is much less than the dose based allowable limit of 3.6 gpm for freespan ODSCC. The SLB leak rate at 95%/95% confidence conservatively satisfies the requirement of 1.0 gpm for a nominal leak rate to support severe accident considerations.
Operating Leak Rate Limits Based on providing some margin relative to the demonstrated 85 gpd acceptable spike, the following operating leak rate limits are recommended for Farley-1 during Cycle 16 operation:
Plant to be shutdown within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a leak rate spike confirmed by chemistry samples to o
exceed 75 gpd.
Plant to be shutdown if the leak rate following a spike above 60 gpd does not decrease to less than o
60 gpd following the leak rate spike. Successive leakage spikes above 60 gpd, but less than 75 gpd, within a seven day span do not require shutdown since the 60 gpd limit is intended to be a longer term leak rate without spikes.
Plant to be shutdown within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a progressive (with spikes) increase in the leak rate o
to greater than 60 gpd.
The above operating leak rate limits are to be administratively implemented for Cycle 16 operation only.
The SGs are to be replaced following Cycle 16 operation.
Q:tubeintiala\\ala9871nal oA\\ Final Freespan oA-9. doc.04/2189 9-9
i sp 8
B g
0 t
i a
v s
3 d
G a
p 4
e S
6
=
ro t
s c
n f
3 o
4 u
i e
s d
a 0
3 P
t t
a n
e h
t 1
e
=
A r
e p
s R
m s
=
g a
e 3
ee e
D a
t a
c s
k m
gc l
s t
n e
k P
s e
h an e
O a
o k e b
a A
P d
c e
e C
ad P%
o c
e 3
A a
i L
ei o
0 o
u n
f n
y7 L
4 3
n r
P e
f l
t 4
o p
A d
d wo d
t 1
i>
d i
v 1
c p
3 f
e n
oc a
y%
ic t
s n
m a
n t
i 5
l a
o ih a
s 5
9 5
d s
c t
9 t
/
9 s
r5 e
np t
n
/
5
/
e e9 r
ee s
5 5
9 5
5 9
5 V>
P Sd r
e i
9 9
9 9
/
t u
m a
5 t
i t
o t
o t
l o
t o 9 B
m a
a i
a a
i u
u
)
u u
u b e%
)iP iP
)
)
t iP iP a
o s
a r
s A sA P
s A sA C
t 5 e o
b a9 f
p3 p3 A
p3 p3 o
o R/c u
7 3
0
> 0 > P i
1 n
s s
9 6>
r k%e p
0 0
t 2) 6)
4 4
4 9) 2)
A 1
1 P
a5d 0
0 3 s 3 s 3 s 4 s 3
l i
i i
i 1
t e9f 5
0 0
0 0
u i
( p
( p s
( p
( p s
s Lt n 0
s e0 e2 e
e 4
e2 e
4
=
~
e r
ao r
r i
r r
i u B 2
B LmC R
u7 u5 u0 u9 f
f s
s s1 s0 s7 s5 i
i S p s
s5 s5 t
s5 s5 t
e(
e(
a e(
e(
a B
g d
r r
s r
r s
o pe pe n
pe pe n
s s
s s
L t
a t a o
t a
t a
o h
s s
i s
s i
t r c r c S
t r c r c t
e e
ut ut a
ut ut a
f n
b s b s c
b s b s c
M o
eOt 0
0 0
0 r
r i
r r
i e o eo d
e o eo d
r sr 1
1 y
o s w w
n s
1 1
s w s w n
r nu x
x x
x s
a a
a a
i i
MaB i
5 4
2 3
s c n c n t
c n c n t
a i
t a
t a
s s
s h
2 1
l 1
t a
t a 0
y si si e
si si e
m p
T 1
l rd rd g
r d rd g
a oe oe r
o e oe r
9 m
0 a
a 6
n WM WM L WM WM L u
5 A
y
'0 2
2 2
r s
S ie o et t
0 0
0 e
r s 1
1 1
1 l
c eor x
x x
x a k m e
e e
e t
d d
d d
ib n nMu C ae 9
O B
5 8
5 n
ay c
c c
c t
t t
t C bd 4
l 2
1 r
eg e
e e
e 1
1 i
of e L 0
0 p
p p
p S
r n t
0 0
x x
x x
D P o l
Be E
E E
E C
e A Lt O
t a
t t
t t
s n
' 4 e
f SR o
o o
o ru%
rot 0
0 0
0 0
s o
N N
N N
n r
1 1
1 1
1 a B5 o nu x x
x x
x s
9 p Bt MaB 9
8 2
3 7
n h
s i
l 6
l 8
4 o
e La s
T s
e S p
y m
i r
5 r
t 4
4 4
1 i
F 0
a t
2 l
0 0
0 0
0 0
/
i s
4 4
p rb 1
1 1
1 1
1 0
'0 0'
0 0
0 m u a x
x x
x x
x c
2 1
2 r
s
~
o et b
1 o
o Bo 6
8 1
5 2
d.
y rs 1
1 1
1 1
e eor x
x x
x x
1 4
3 3
1 1
9 r
nMu C
P A
l O
B 1
1 8
1 2
r o
a 3
1 1
1 7
n F
2 ap s
9 e
e Y
e Y F
r 1
2 2
2 3
9 3
3 2
3 2
2 P
a 7
1 7
1 1
b F
1 1
1 7
7 7
l l
9 g
P n
n F
i 0
0 0
F 1
1 1
a E e id E
T A
0 0
1 1
1
\\
l t
o b
ai o
a rr la ee T
pP D
in s
s F
O P
5 9
5 9
9 h
i h
i
\\
6 0
6 0
0 f
s t
s 8
F 2
4 2
4 4
os y
p y
p l
t 9a e
e E
d s a
l d
i l
e e
n la l
i oly n
c n
c d
la a
a y
e a
y e
h a c
g e
i G
B C
B e
c g
n t
n e
t a
l i
a i
l i
S M A c
i t
r i
t r
b f
l f
l e
e u
o u
v o
u v
t r
r Q
P M
A P
M A
9
/
1 2
/
8 8
9
/7 s
1
/
r 8
o t
8 9
i n
/
o 3
1
/
M 8
n 8
o 9
/
i t
9
/
a 8
i d
a 8
9 R
/
5
/
8 m
ae 8
t 9
/
S 1
/
8 n
i a
8 9
M
/
8 e 2 m C
/
7 i T
/
I B
8 e t
1 o
9 a t
/
9 d
4 D A
la 7
t e
2
/
0 es 7
c o
s 8
t R
8
/
e 9
r 9
0
/
n 2
1 7
e
/
2 e
7
/
t 7
b i
8 n
n e.
/
o v
9 s
U at 6
e hl d
u 1
y 8 s
/
a 7
9 e e
m
/ r l
7 e
8 9
r t
2l 9
9 n s
/ p
/
U a
e e 7
m 2
2 F
mlp o
1 4
/
a
/
0 t
t sm s
7 r
c u a d s iob o
1 j
d.
r a 9
ab pr 8
9-g 9
A n a a
/
O r
t h 8
e iag at
/
n r
g o d i 7
a w
p u
t s
B Be ee g
0 e 0 e r
n 8
F i
7 i 7 r l
l
- g 9
a F
Rn Ra 4
i
/
n i
W
/
7 A
O 8
lan 9
i7
/0 8
3 9
0 9
8 7
6 5
4 3
2 6
la 0
0 0
0 0
0 0
0 0
0 /
a
\\a 1
d e
{9 I$.3 in e
bu t
Q
10.0 REFERENCES
10.1 Report NSD-E-SGDA-98-0450, "Farley-1: IRI5 Inspection Condition Monitoring and Preliminary Operational Assessment", December 22,1998, Westinghouse Electric Company 10.2 EPRI Report NP-6865-L, " Steam Generator Tube Integrity, Volume 1: Burst Test Results and Validation of Rupture Criteria (Framatome Data)," Final Repon, June 1991.
10.3 EPRI Draft Repon GC-107621, Rev. A2, " Steam Generator Tube Integrity Assessment Guidelines", October 1998 10.4 WCAP-15128, " Depth Based SG Tube Repair Criteria for Axial PWSCC at Dented TSP Intersections:, Westinghouse Electric Company, January,1999 10.5 WCAP-14277, Revision 1, "SLB Leak Rate and Tube Burst Probability Analysis Methods for ODSCC at TSP Intersections," Westinghouse Nuclear Services Division, December 1996.
10.6 Report SG-99-03-001, "Farley Unit-1 1999 Voltage Based Repair Criteria 90 Day Repon",
March 1999, Westinghouse Electric Company 10.7 Docketed Submittal to the NRC, Woodman, B. W., Begley, J. A., and Brown, S. D.,"Palo Verde Unit 2 Steam Generator Tube Regulatory Analysis for Axial Cracking", Packer Engineering Report B52076-R1, August,1995.
10.8 Docketed Submittal to NRC, Docket 50335, Begley, J. A., Woodman, B. W., Brown, S. D., and Brose, W. R., " Analysis of ODSCC/ IGA at Tubesheet and Tube Support Locations at St. Lucie, Unit 1, Aptech Engineering Services,Inc, Repon, AES 96052749-1-1, October,1996.
10.9 Docketed Submittal to NRC, Docket 50361, Begley, C. J. Begley, Woodman, B. W., and Begley, J. A.,"A Probabilistic Operational Assessment of Steam Generator Tube Degradation at SONGS Unit 2 for Cycle 9", Aptech Engineering Services, Report AES 97043057-1-1, September,1997.
10.10 Docketed Submittal to the NRC, Begley, J. A. Begley, Peck, M. G., and Begley, T. F., " An Operational Assessment of Steam Generator Tube Degradation at Crystal River Unit 3", Aptech Engineering Services Report, AES98033350-1-1, April,1998.
10.11 Begley, J. A. and Houtman, J. L.,"Inconel Alloy 600 Tubing-Material Burst and Strength 4
Properties, WCAP-12522, Wesinghouse Electric Corp. January,1990.
10.12 Keating, R. F., Begley, J. A., Brose, W. R., and Lagally, H. O.," Steam Generator Degradation Specific Management Flaw Handbook", EPRI Project S550-7, to be published.
10.13 EPRI TR-107197-P1, " Depth Based Structural Analysis Methods for Steam Generator Circumferential Indications," Electric Power Research Institute, Palo Alto, CA, Interim Report, November,1997.
10.14 Zahoor, A., " Ductile Fracture Handbook", EPRI Report, NP.6301-D, RP 1757-69, June,1989.
10.15 Letter NEL-99-0131, Docket Nos. 50-348, 50-364, Dave Morey to U. S. Nuclear Regulatory Commission, " Joseph M. Farley Nuclear Plant, Reactor Coolant System Specific Activity Technical Specifications Change Request", April 2,1999 i
Q:tubeint\\nla\\ala98irinal oA\\ Final Frecapan oA-10 doc 10-1 I