ML17229A092
| ML17229A092 | |
| Person / Time | |
|---|---|
| Site: | Saint Lucie  |
| Issue date: | 10/24/1996 |
| From: | FLORIDA POWER & LIGHT CO. |
| To: | |
| Shared Package | |
| ML17229A091 | List: |
| References | |
| NUDOCS 9610280067 | |
| Download: ML17229A092 (62) | |
Text
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 Florida Power and Light Company St. Lucie Vnit 1 Steam Generator Run Time Analysis Cycle 14 9610280067 96i024 PDR, ADOCK 05000335 P PDR ~'
~ i e
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 TABLE OF CONTENTS Title P~ae Table of Contents ~............. ~ ~ ~ ~ ~ ~ ~
1.0 Abstract ~..............
. ~ ~ ~ ~ 2 2.0 Description ....... ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 3
'.0 Evaluation................., ~ ~ ~ 7 4 .0 Conclusions ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 16 5.0 References 17 Figure 1- Steam Generator Sludge Removal ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 19 Table 1 - Steam Generator ECT and Tube Plugging History..... ... 20 Appendix A - Request For Additional Information Responses ............. A-1
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 1.0 ABSTRACT The purpose of this report is to describe efforts completed by FPL to address degradation of steam generator (SG) tubing due to Outside Diameter Stress Corrosion Cracking (ODSCC) at St. Lucie Unit 1 with respect to Cycle 14 operation. Based on the results of extensive Eddy Current (ECT) examinations and analyses described in this report, degradation mechanisms exhibit a predictable growth rate and future steam generator tube performance can be evaluated. From an analytical perspective, Cycle 14 operation is justified for St. Lucie Unit 1 based on end of Cycle 13 inspection results and this evaluation. This evaluation is structured to support continued operation for up to 15 effective full power months (EFPM - defined as operation with T,above 500'F) following restart on July 25,1996.
It should be noted that the original length of 22 montlis for Cycle 14 has been reduced to 1S months due to results of the run time analysis'ompleted by APTECH Engineering Services, at which time SG replacement is planned.
The run time analysis has been used to determine the duration wliich the St. Lucie Unit 1 steam generators can be safely operated without exceeding the criteria of NRC Generic Letter (GL) 95-05'. However, since St. Lucie Unit 1 does not have Westinghouse design steam generators, and does not apply a voltage based repair criteria, the run time analysis was completed using GL 9545 guidance and a physically based statistical model to verify that 15 EFPM is acceptable for Cycle 14 operation. The Conditional Probability of Tube Burst, for all degradation mechanisms, is to the draft Regulatory Guide X.XX'imit of Sx10 ~. However, the limit of Conditional Probability of Tube 4.4x10'ompared Burst for an individual mechanism is above the GL 9545 and draft RG X.XX limit of 1x10'. As a result, a safety assessment was prepared in accordance with the guidance provided in GL 9545.
Further, qualitative support for at least 13.9'EFPM of operation is provided by the end of Cycle 13 inspection results and repairs completed. Prior to Cycle 13 which operated for 13.9 EFPM, all tubes with degradation 40% or greater through wall penetration based on eddy current examination were plugged. In-situ pressure testing of 17 bounding defects identified aAer Cycle 13 operation'emonstrated tliat the structural guidance contained in draft NRC RG 1.121'as not exceeded. As Cycle 14 was origuially scheduled for 22 montlis, more conservative plugging criteria were used prior to initiation of the cycle.
The baseline core damage frequency (CDF) calculation used a probability of burst (POB) of 9.78x10'/Yr. Since this calculated POB is less than the POB assumed for the baseline CDF the change in CDF is considered as risk insignificant. Steam generator tube rupture (SGTR) followed by main steam line break (MSLB) contributes to a core damage probability (CDP) of 1.4x10'ver the 15-month operating period. The large early release probability (LERP) increase is conservatively estimated to be 1.4x10~. Both the CDP and LERP are considerably below the risk significance criteria for a temporary plant change (1.0x104 for CDP and 1.0x15'or LERP) established in EPRI TR-105396.
In addition, the calculated leakage after a postulated main steam line break (MSLB) is less than 4 GPM which is a small fraction the leakage calculated by Scientific Applications International Corporation 'SAIC) to meet the 10 CFR 100 and GDC 19 radiation exposure limits. Therefore, Cycle 14 operation for the St. Lucie Unit 1 steam generators, from their startup date of July 25, 1996 through October 23, 1997, will be acceptable in accordance with the guidance contained in GL 95-05 and Draft RG X.XX.
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1
2.0 DESCRIPTION
FPL provided a response to an NRC request for additional information regarding steam generator inspection and repair criteria for St. Lucie Unit 1 in letter L-96-166, June 25, 1996'. In response to NRC Request 7, FPL committed to "..complete a run time analysis to demonstrate compliance with NRC GL 95-05, Voltage-Based Repair Criteria for Westinghouse Steam Generator Tubes Affected by Outside Diameter Stress Corrosion Cracking" and to "..present the results to the NRC within 90 days following the St. Lucie Unit 1 startup from the current refueling outage (SL1-14)."
The purpose of this evaluation is to fulfillthe commitment to provide the run time analysis. In addition, this evaluation responds to additional NRC Requests for Additional Information (RAI) dated July 16, 1996 (Appendix A). The run time analysis and the RAI response are due to be submitted to the NRC by October 23, 1996.
Steam Generator Description The St. Lucie Unit 1 design includes two recirculating Combustion Engineering (CE) Series 67 Steam Generators (SG) which are vertical U-tube and shell heat exchangers. The SGs are designed and fabricated as Class A vessels as defined in ASME Code, Section III through Winter 1967 Addenda. Each SG contains 8,519 Alloy 600 high temperature mill annealed tubes which are 3/4 inch OD and have a nominal wall thickness of 0.048 inches. The tubes are explosively expanded into the tubesheet for the entire tubesheet thickness. The tubes are arranged in rows, with all tubes in a given row having the same length. The rows are staggered, forming a triangular pitch arrangement. The shorter tubes, which have 180'ends, are at the center of the tube bundle in the first 18 rows. All subsequent rows have double 90'ends, The vertical tube lengths are supported by six full diameter eggcrates (lattice bars), two partial diameter eggcrates and two partial diameter drilled support plates. The bend and horizontal lengths are supported by batwings and vertical lattice supports respectively.
History of Tube Degradation Tube degradation was first detected in the U-bend of rows 8-11 during ECT in the 1981 refueling outage. The cause of the corrosion was a design flaw associated with these rows which created a localized dry out region due to steam blanketing. Due to inadequate inspection techniques for detection and sizing of degradation in this bend region, these tube rows were preventatively plugged in 1984.
Subsequent inspection results indicated that tube degradation had progressed to other regions of the tube bundle. Destructive examination of six tube sections removed from service in 1985 identified acidic sulfate intergranular attack and intergranular stress corrosion cracking as the primary cause of the ECT indications'. Several areas of transgranular stress corrosion, were also identified during the destructive examination.
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 Water Chemistry The corrosion attack initiated prior to routine monitoring of blow down chemistry for the presence of sulfates. The effluent from the make up water treatment plant demineralizers was identified as a potential source of sulfates'. Early plant operating procedures required the demineralizers to be rinsed after regeneration to achieve a specific conductivity level resulting in the effluent having a sulfate level of 70-100 PPB. Revisions to the demineralizer regeneration procedures have reduced the sulphate concentration to less than 2 PPB. Further improvements in secondary water chemistry control have resulted in significant reduction in the amount of sludge lanced from the SGs (Figure 1), and removal of copper bearing feedtrain components has significantly reduced the copper to iron ratio of the sludge.
St. Lucie Unit 1 has operated with all volatile chemistry since startup in December 1976 and has accumulated approximately 14.7 effective full power years through the end of Cycle 13 operation on April 29, 1996. Approximately 20 condenser tube leaks occurred during the early cycles of operation. The condensers are seawater cooled and the feedwater system has a condensate polisher system which is used during startup. A change in operation was implemented several cycles after startup to reduce the amount of sulfate introduced into the system immediately after regeneration of the demineralizers. Improvements in secondary 'chemistry control have paralleled industry developments with implementation of the EPRI PWR Secondary Water Chemistry Guidelines'nd subsequent revisions. A secondary side boric acid addition program was implemented in the Spring of 1990 to mitigate the ODSCC attack. Current levels of contaminants in the feedwater and steam generator blowdown are typically below the values that INPO uses in calculating the Chemistry Performance Index.
Steam Generator Tube Inspections SG tube inspections at St, Lucie Unit 1 have consisted of full length bobbin coil examination of all active tubes since 1984. Motorized Rotating Pancake Coil (MRPC) techniques have been routinely used since 1987 to further characterize degradation. MRPC techniques have been extensively employed for tube expansion transitions and dented tube support intersections since 1991. Improvements in inspection techniques have resulted in the detection of greater numbers of degraded tubes. ODSCC has been detected at drilled hole and eggcrate type supports, tubesheet expansion transitions, sludge pile regions and in the upper bundle free span at or immediately below the bend tangent. The orientation of ODSCC is axial at these locations except for the tube expansion transition, where it is typically circumferential. The most recent SG inspection was completed at the end of Cycle 13 in July 1996.
A history of ECT inspections and tube plugging is provided in Table 1. The most recent steam generator tube inspections and repair efforts are discussed below:
Eggcrates and Sludge Pile - FPL has utilized bobbin coil techniques for detection and sizing of ODSCC indications since 1986. FPL recently completed a bobbin coil technique qualification in
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 accordance with Appendix H", Performance Demonstration for Eddy Current Zramination, for ODSCC indications at eggcrate type supports and sludge pile regions. The results of this qualification exceed the requirements of Appendix H, and were submitted to the NRC in FPL letter, I 96-166, June 25, 1996. The qualification includes all available data for tube flaws removed from sludge pile and eggcrate supports in CE design SGs. The FPL technique has an 80% probability of detection at a 90% confidence level for flaws penetrating 35% or greater through wall. The root mean square sizing error was 17.5% with a correlation coefficient of .83 for ECT depth versus metallographic depth. Further discussion on the use of these results for the run time analysis is provided in APTECH analysis.
Drilled Support Plates - A drilled support adversely effects the ECT signal for detection and sizing of ODSCC indications. Therefore, drilled supports were not included in the bobbin coil technique qualification discussed above. FPL has dispositioned ODSCC indications at drilled supports in accordance with Supplement I, "Guidelines for Disposition of Bobbin Coil Indications Attributed to ODSCC at Non-dented and Drilled Tube Support Plates", of EPRI NP-620.
However, Supplement I of EPRI NP-620 was deleted in a subsequent revision in June 1996. This resulted in a change in repair criteria for the end of Cycle 13 inspection. As a result, all ODSCC indications at drilled support plates were removed from service upon detection using MRPC techniques. The change in criteria resulted in repair of approximately 800 additional tubes prior to Cycle 14 operation.
Free Span Indications - Subsequent to a tube rupture in March of 1993 at Palo Verde Unit 2 (CE System 80) it was determined that upper bundle deposits may act as a precursor to free span axial cracking. Thermal hydraulic evaluations of steam generators, completed since the tube rupture, indicate that certain upper bundle regions have higher potential for deposit accumulation. Axial indications have recently been reported in the upper bundle free span region of two operating CE Series 67 design units.
Industry experience, and a thermal hydraulic evaluation of steam generators (ATHOS) for St.
Lucie Unit 1 and 2", was used to determine tube bundle regions most susceptible to deposit accumulation in FPL SGs. FPL implemented inspection plan changes to include MRPC inspection of selected regions in both SGs. MRPC inspections were expanded to all upper free span regions in the hot leg of each SG after several axial indications were detected. Approximately 13,000 tubes were inspected and a total of 44 axial indications were detected and removed from service.
In addition, approximately 40 volumetric free span indications, which were not present on the preservice baseline inspection, were also removed from service.
Circumferential Indications - Several CE design units began reporting circumferential ODSCC at the tube expansion transition prior to 1990". Circumferential ODSCC was first detected at the tube expansion transition in the hot legs of the St. Lucie Unit 1 SGs during the 1991 inspection.
MRPC inspection has been routinely performed for this region since that time. FPL data analysis guidelines have been continually upgraded to remain consistent with EPRI and industry practices.
Approximately 165 circumferential indications were stabilized and removed from service in this
H il St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 insp'ection. Ten of the indications were on the cold leg side of the steam generators. FPL has applied EPRI sizing techniques to verify that these indications are being detected and removed from service prior to exceeding the structural requirements of draft NRC RG 1.121. All active hot leg and cold leg tube expansion transitions were inspected during the most recent inspection.
Conservative Tube Plugging - Cycle 14, initially scheduled for 22 EFPM, was to be the longest operating period planned for St. Lucie Unit 1. Based on prior inspection results, approximately 3500 active tubes contained ODSCC indications penetrating 20-39% through wall at the end of Cycle 13 inspection'4. To reduce the risk of leakage during Cycle'14, a more conservative analysis of ECT data was implemented". The added conservatism resulted in removing approximately 1100 tubes from service. Also, as mentioned above, a change in repair criteria for indications at drilled supports added approximately 800 additional tubes to the repair list.
In-Situ Pressure Testing - In-situ pressure testing was conducted during the most recent inspection to verify that draft NRC RG 1. 121 structural margins were maintained. Seventeen in-situ pressure tests were completed using a full tube hydrostatic or defect-specific hydrostatic test method. The full tube method was performed on 3 tubes in SG A to test free span axial indications which were not suitable for testing with the defect-specific method.
The defect specific method was used for 14 defects at various elevations in both SGs. Defect-specific testing included circumferential indications at tube expansion transitions, and axial and volumetric indications at eggcrate supports, drilled supports and the sludge pile region. ABB Combustion Engineering Report 00000-OSW-16 Rev. 00, In-Situ Pressure Test Results for St.
Lucie Unit J Spring l996 Outage, provides detailed results for each test.
Candidate tube defects for in-situ pressure testing were screened for several criteria. The selection process included review of bobbin coil and MRPC data for maximum through wall depth, maximum voltage and growth. Percent degraded area (PDA) was also calculated for candidate circumferential and axial defects to assure that defects tested are bounding. Finally, lead ECT data analyst personnel recommendations were included.
The in-situ pressure test results demonstrate adequate structural margins existed for St. Lucie Unit 1 SG tubes. There were no catastrophic tube failures (burst), None of the defects tested leaked at normal operating differential pressure, and three defects leaked at or below MSLB differential pressure. The results of this testing are applied to conditional probability of burst analyses and ihtegrated leakage assessments as part of the run time analysis for St. Lucie Unit 1 as discussed in the APTECH analysis.
In-situ test pressures were increased by 13% for all defects to compensate for test conditions at room temperature. Pressures were increased an additional 8.5% (21.5% total) for circumferential defects to compensate for locked support plate conditions. Additional adjustments were made for pressure gauge calibration correction (50 psig) and for test system pressure drop in the event that substantial leakage was encountered. If leakage was encountered at main steam line break
I St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 (MSLB) or lower pressures, leak rate data was recorded at normal operating and MSLB differential pressures. Tube defects were then tested at 3 times normal operating differential pressures to assess structural conditions.
Test pressures were determined using very conservative assumptions for MSLB differential pressure. That is, differential pressures were calculated assuming design pressure for the tubes with no secondary pressure under MSLB conditions. This resulted in a conservative differential pressure of 2500 psid (nominal - uncorrected) compared to normal operating conditions where the differential pressure is 1435 psid. During transient conditions, design basis documents anticipate a maximum differential pressure of 1600 psid. A review of accident transients confirms the pressures presented in the St. Lucie design basis documents". Therefore, in-situ pressure test results are considered conservative.
3.0 EVALUATION NRC Generic Letter 95-05, Voltage-Based Repair Criteria for Westinghouse Steam Generator Tubes A+ected by Outside Diameter Stress Corrosion Cracking, was issued to give guidance to licensees who may wish to implement alternate steam generator tube repair criteria. Although St.
Lucie Unit 1 does not have Westinghouse steam generators, does not use voltage-based repair criteria and does not plan to request a licensing amendment to implement alternate steam generator tube repair criteria, this document is used as guidance to evaluate steam generator tube performance and predict plant operating capabilities as discussed in the APTECH analysis, In order to generally follow the guidance identified in GL 95-05, this evaluation will use the format and criteria delineated in the generic letter. These criteria and the applicable FPL compliance are described below (Note that GL Attachment 1 identified in the following criteria refer to Attachment 1 of GL 95-05):
- 1) "Implementation of the applicability requirements discussed in Section 1 of GL Attachment 1. The applicability requirements ensure that the repair criteria are applied only to those intersections for which the voltage-based repair criteria were developed."
FPL has utilized bobbin coil techniques for ODSCC indications since 1986, FPL recently completed a bobbin coil technique qualification in accordance with Appendix H, Performance Demonstration for Eddy Current Examination, of EPRI NP-6201 for ODSCC indications at eggcrate type supports and sludge pile regions. The results of this qualification exceed the requirements of Appendix H, and were submitted to the NRC in FPL letter, L-96-166, dated June 25, 1996. The qualification includes all available data for tube flaws removed from sludge pile and eggcrate supports in CE design SGs. The FPL technique has an 80% probability of detection at a 90% confidence level for flaws penetrating 35% or greater through wall. The root mean square sizing error was 17.5%
with a correlation coefficient of .83 for ECT depth versus metallographic depth. Further
St, Lucie Unit 1 8 Docket No. 50-335 L-96-273 Enclosure 1 discussion on the use of these results for the run time analysis is provided in the APTECH analysis.
- 2) "Implementation of the inspection guidance discussed in Section 3 of GL Attachment 1. The inspection guidance ensures that the techniques used to inspect steam generator tubes are consistent with the techniques used to develop the voltage-based repair criteria."
As stated in Item 1, FPL has utilized bobbin coil techniques for ODSCC indications since 1986. FPL recently completed a bobbin coil technique qualification in accordance with Appendix H, Performance Demonstration for Eddy Current Examination, of EPRI NP-6201 for ODSCC indications at eggcrate type supports and sludge pile regions. The results of this qualification exceed the requirements of Appendix H, and were submitted to the NRC in FPL letter, L-96-166, dated June 25, 1996. The qualification includes all available data for tube flaws removed from sludge pile and eggcrate supports in CE design SGs. The FPL technique has an 80% probability of detection at a 90% confidence level for flaws penetrating 35% or greater through wall. The root mean square sizing error was 17.5% with a correlation coefficient of .83 for ECT depth versus metallographic depth.
Further discussion on the use of these results for the run time analysis is provided in the APTECH report.
- 3) "Calculation of leakage according to the guidance discussed in Section 2.b of GL Attachment 1. This calculation, in conjunction with the use of licensing basis assumptions for calculating offsite and control room doses, enables licensees to demonstrate that the applicable limits of 10 CFR Part 100 and GDC 19 continue to be met. This calculation is performed using the projected EOC voltage distribution for the next cycle of operation. If it is not practical to complete this calculation prior to returning the steam generators to service, the measured EOC voltage distribution can be used (from the previous cycle of operation) as an alternative (refer to Section 2.c of GL Attachment 1) for the purposes of determining whether the reporting criteria of Section 6.a.1 apply."
A leakage limit calculation has been completed for St. Lucie Unit 1 by Scientific Applications International Corporation (SAIC). Steam Generator Degradation Specific Management (SGDSM) Leakage Limit Calculation for St. Lucie Unit 1, SAIC 05-5049 6734-500, September 1996, Rev. A is provided as part of the run time analysis. The report demonstrates that for a postulated MSLB with induced SG tube leakage of up to 200 gpm, the resulting dose is well within the 95/95 boundary confidence level, considered acceptable in the pending SG Rule making criteria, for not exceeding 10 CFR part 100 and GDC 19 accident dose limits. The report concludes that results show the leakage level needed to produce dose limiting conditions up to the 95/95 confidence level at St. Lucie Unit 1 is much greater than the industry recommended upper bound of 200 gpm.
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 The screening results indicated the potential for large margins on the leakage which was confirmed with a detailed site assessment. At the 95/95 confidence level the detailed assessment indicated the most limiting case was the pre-accident spike with the control room and EAB about equal at 1600 gpm. For a main steam line break and steam generator tube leak (MSLB/SGTL) of 200 gpm the confidence level is approximately 99/99 that the regulatory 'dose limit at the St. Lucie site would not be exceeded.
Since dose is unlikely to be the limiting constraint on leakage, plant physical features can be used in determining the leakage limit for the postulated MSLB. Examples of plant physical constraints that can be considered as limits are the combined charging pump capacity for the St. Lucie plant, or the sum of the charging pump and high pressure safety injection make up capacity.
The projected end of cycle leak rates at postulated accident conditions were reasonably small and not markedly sensitive to run time. After 15 months of operation, the total projected 95% upper bound leak rate is less than 4 gpm. Site-specific analyses of dose rate consequences for various leak rates demonstrate a leakage limit far in excess of this projected value.
Further, in-situ pressure test results of bounding defects observed in the most recent inspection demonstrate that SG tube leakage would be limited during postulated MLSB conditions to a fraction of the industry recommended upper bound of 200 gpm. It is, therefore, concluded that SG tube integrity will be maintained within required limits.
- 4) "Calculation of conditional burst probability according to the guidance discussed in Section 2.a of GL Attachment 1. This is a calculation to assess the voltage distribution for the next cycle of operation. The results are compared against a threshold value. This calculation is performed using the projected voltage distribution for the next cycle of operation. If it is not practical to complete this calculation prior to returning the steam generators to service, the measured end of cycle (EOC) voltage distribution can be used (from the previous cycle of operation) as an alternative (refer to Section 2.c) for the purposes of determining whether the reporting criteria of Section 6.a.3 apply."
The significance of corrosion degradation to the performance of generator tubing at St.
Lucie Unit 1 was evaluated for the final cycle of operation of the unit. Replacement of steam generators will begin in the Fall of 1997. Probabilistic methods were applied to make end of cycle projections of the structural and leakage integrity of the steam generator tubing experiencing corrosion degradation.
Over the past 12 years of operation, the steam generator tubing at St. Lucie Unit 1 has experienced substantial corrosion degradation. Eddy current inspection data and pulled tube examinations show the degradation to be combinations of intergranular attack ( IGA)
H
~
U
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 and stress corrosion cracking (SCC) on the outer diameter of the tubing. The five types of locations where this degradation occurs and the modes of corrosion are:
Circumferential degradation at explosive transitions at the top of the tubesheet Axial degradation at the top of the tubesheet in the sludge pile Axial degradation at lattice type tube support structures Axial degradation at drilled tube support plates Axial degradation at upper bundle freespan locations Several probabilistic run time models were employed. Modeling included scenarios of both plug on detection and plug on sizing, depending on the mode and location of the degradation. Circumferential degradation at the top of the tubesheet, as well as axial degradation at freespan and drilled tube support locations, was modeled using a plug on detection scenario coupled with an MRPC scheme. Modeling of axial degradation at other locations employed a bobbin probe inspection with a plug on size scenario. Probabilistic computations of the conditional probability of burst and the magnitude of leakage under accident conditions were developed. Calculational procedures were benchmarked by comparing predictions of the number and severity of eddy current indications with actual observations and comparing predicted versus observed leakage during in-situ testing.
The dominant contributor to the conditional probability of burst is axial degradation in the sludge pile and at lattice type supports. This is basically due to the number of indications involved. Degradation growth rates are largely independent of location.
The conditional probability of burst under postulated accident conditions is the limiting concern. When all corrosion mechanisms are considered, the projected end of cycle conditional probability of burst is .044 for a run time of 15 months. This run time provides for the conditional probability of tube burst below the historical figure of merit of 0.05. As the projected EOC failure distribution for individual mechanisms is above the GL 95-05 and draft RG X.XX limit of 1 x 10', a safety assessment was performed and provided below in response to GL 95-05 criterion 7.
- 5) "Implementation of the operational leakage monitoring program according to the guidance discussed in Section 5 of'GL Attachment 1. The operational leak rate monitoring program is a defense-in-depth measure that provides a means for identifying leaks during operation to enable repair before such leaks result in tube
. failure."
St; Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 St. Lucie Plant Off-Normal Operating Procedure (ONOP) 1-0830030, Steam Generator Tube Leak, has been revised to incorporate the criteria for shutdown based on primary to secondary leak rates as established in EPRI TR-104788, PWR Primary to Secondary Leak Guidelines, dated May 1995. Also, additional operator training concentrating on steam generator tube ruptures is being performed. The training and procedural changes are described below:
SIMULATORTRAINING- In addition to two hours per operator specifically targeted at Unit One performance after the tube plugging, the licensed operator two-year requalification program ending December 15, 1996, provides 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> of steam generator tube leak/rupture training to each licensed operator on the simulator. Additionally, each requalification week includes a simulator evaluation (exam) which, over the course of the program, may include steam generator tube leak/rupture, These scenarios typically begin with a small Tube Leak requiring identification of the ruptured steam generator, chemistry samples and health physics area monitoring. Upon the determination of excessive leakage, a plant shutdown is commenced. One block of simulator training was carried out over a four hour period to addre'ss the long term actions expected during a steam generator tube leak/rupture.
PROCEDURE CHANGES - The last two changes to St. Lucie Unit 1 Plant Emergency Operating Procedure, 1-EOP-01, Standard Post Trip Actions, chart 1 (diagnostic flow chart) were made to aid the operator in diagnosing a steam generator tube rupture. One change directs the operator to assess parameter based symptoms and any indicated trend in those parameters and the other encompasses symptoms of a steam generator tube rupture not previously evaluated.
The St. Lucie Plant Off Normal Operating Procedure, 1-0830030, Steam Generator Tube Leak, has been changed to include a flow chart that directs increased monitoring or plant shut down based on rate of change of leakage and absolute leakage (plant shut down commenced at .1 GPM). This meets the requirement of 150 gpd stated in GL 95-05.
- 6) "Acquisition of tube pull data according to the guidance discussed in Section 4 of GL Attachment 1."
FPL has utilized bobbin coil techniques for ODSCC indications since 1986. FPL recently completed a bobbin coil technique qualification in accordance with Appendix H, Performance Demonstration for Eddy Current Examination, of EPRI NP-6201 for ODSCC indications at eggcrate type supports and sludge pile regions. The results of this qualification exceed the requirements of Appendix H, and were submitted to the NRC in FPL letter, L-96-166, dated June 25, 1996. The qualification includes all available data for tube flaws removed from sludge pile and eggcrate supports in CE design SGs. The
St. Lucie Unit 1 12 Docket No. 50-335 L-96-273 Enclosure 1 FPL techmque has an 80% probability of detection at a 90% confidence level for flaws penetrating 35% or greater through wall. The root mean square sizing error was 17.5%
with a correlation coefficient of .83 for ECT depth versus metallographic depth. Further discussion on the use of these results for the run time analysis is provided in the APTECH report.
- 7) "Reporting of results according to the guidance discussed in Section 6 of GL Attachment 1."
Per Section 6.a.3 of GL Attachment 1, if the calculated conditional probability of rupture under postulated MSLB conditions based on the projected EOC failure distribution for a single mechanism exceeds 1 x 10', licensees should notify the NRC and provide an assessment of the significance of this occurrence. The projected 'EOC failure distribution for all failure mechanisms combined equals 4.4 x 10'hich meets the requirements of draft Reg Guide X.XX. However, the projected EOC failure distribution for individual mechanisms are above the GL 95-05 and draft RG X.XX limit of 1 x 10'. Therefore, a review of the safety significance is provided below:
Safety Assessment The safety significance associated with the operation of Unit 1 for a period of 15 EFPM during Cycle 14 has been evaluated by FPL. The results of an extensive inspection program conducted during the Cycle 13 refueling outage have been analyzed to determine the impact on the operation in Unit 1 during Cycle 14. As guidance, this assessment uses the standards defined in 10 CFR 50.92 for determining whether a significant hazard consideration exists, A significant hazard is not involved if the condition would not: (1)
Involve a significant increase in the probability or consequences of an accident previously evaluated; or (2) create the possibility of a new or different kind of accident from any accident previously evaluated; or (3) involve a significant reduction in the margin of safety. A discussion of these standards as they relate to Unit 1 operation during Cycle 14 is as follows:
Does the Unit 1 steam generator operation involve a significant increase in the probability or consequences of an accident previously evaluated' FPL has determined via the analyses presented in this report, that operation of the Unit 1 steam generators does not represent a significant increase in the probability or consequences of an accident previously evaluated in the St. Lucie Unit 1 UFSAR. It has been determined that there is a low probability of generating a defect in excess of the structural margins specified in Regulatory Guide 1.121.
Although conditional probability of burst given a MSLB for individual mechanisms may exceed the limit of 1.0x10'rovided in GL 95-05 and draft RG X.XX, it was
St. Lucie Unit 1 13 Docket No. 50-335 L-96-273 Enclosure 1 calculated to be 4.4x10'or all degradation mechanisms combined, which is less than the limit of Sx10'rovided in draft RG X.XX.
FPL has also assessed leakage potential, and the impact of operation with steam
.generator tube defects on core damage probability. The analyses contained in this report indicate that there is low probability of tube leakage and any resulting off-site doses, given a main steam line break (MSLB), are significantly within 10CFR100 limits. Leakage observed during end of Cycle 13 in-situ pressure testing is considered conservative and has been reflected in leakage modeling in the APTECH analysis.
The following summarizes the Core Damage Frequency (CDF) assessment as a result of the revised probability estimate of the St. Lucie Unit 1 steam generator tube rupture (SGTR) events:
I. Case 1: SGTR Under Normal Operating Conditions The run time analysis calculated a Probability of Burst (POB) on the order of low 10'/Yr under normal operating conditions. The baseline CDF calculation used a POB of 9.78x10'/Yr. Since this calculated POB is less than the POB assumed for the baseline CDF the change in CDF is considered as risk insignificant.
II. Case 2: SGTR following a main steam line break (MSLB)
In the original PSA, no CDF was attributed to the SGTR induced by MSLB. The SGTR followed by MSLB (assuming a burst probability of 5.0x10'or the 15th month and 2.0x10'or the first 14 months) contributes to a core damage probability (CDP) of 1. 17x10~ for the first 14 months and 2.08x10'or the period between fourteenth and fifteenth months, a total CDP of 1.4x10~ over the 15-month operating period. The large early release probability (LERP) increase is conservatively estimated to be 1.4x10~. Both the CDP and LERP are considerably below the risk significance criteria for a temporary plant change (1.0x10~ for CDP and 1.0x10'or LERP) established in EPRI TR-105396.
Alternatively if the risk significance criteria for a permanent plant change are used, the CDF increase is approximately 1.1x10~/Yr or 0.05% of the baseline CDF.
Similarly, the Large Early Release Frequency (LERF) increase is approximately 1.1x10~/Yr or 0.4% of the baseline LERF. Both the CDF and the LERF are considerably less than the criteria which define risk significance for a permanent plant change, i.e. 4.8x10~/Yr for CDF and 5.5x10'/Yr for LERF.
Because of the low likelihood (5.0x10~/Yr) of the MSLB events, it is of interest to note that the most limiting risk significance criteria based on LERP for a
St. Lucie Unit 1 14 Docket No. 50-335 L-96-273 Enclosure 1 temporary change would yield a 16% conditional burst probability above which the risk increase is risk significant.
It is noted that two factors offset the estimated risk increase associated with the continuous run of 15 months. One is the risk associated with shutting down the plant and perform a mid-cycle inspection. Another factor is that not all core damage scenarios associated with SGTR leads directly to large early release.
In summary, the risk increase associated with the degraded steam generator configurations is considered risk insignificant for a continuous run time of 15 months. If the risk of shutting down the plant to perform mid-cycle inspection is considered, the option to run for 15-month may be risk neutral.
Additional actions have been taken to assure safe plant operation. Administrative limits on primary-to-secondary leakage provide additional assurance that an orderly shutdown would be conducted prior to a through-wall leak propagating to a rupture, Also, additional operator training concentrating on steam generator tube ruptures is being performed. The training and procedural changes are described below:
SIMULATOR TRAINING - In addition to two hours per operator specifically targeted at Unit 1 performance after the tube plugging, the licensed operator Requalification program ending December 15, 1996 provides 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> of steam generator tube leak/rupture training to each operator on the simulator.
Additionally, each Requalification week includes a simulator evaluation (exam) which, over the course of the program, may include steam generator tube leak/rupture.
These scenarios typically begin with a small tube leak requiring identification of the ruptured steam generator, chemistry samples and health physics area monitoring. Upon the determination of excessive leakage, a plant shutdown is commenced. One block of simulator training was carried out over a four hour period to address the long term actions expected during a steam generator tube leak/rupture.
PROCEDURE CHANGES - The last two changes to St. Lucie Unit 1 Plant Emergency Operating Procedure, 1-EOP-01, Standard Post Trip Actions, chart 1 (diagnostic flow chart) were made to aid the operator in diagnosing a steam generator tube rupture. One change directs the operator to assess parameter based symptoms and any indicated trend in those parameters and the other encompasses symptoms of a steam generator tube rupture not previously evaluated.
0 I
St. Lucie Unit 1 15 Docket No. 50-335 L-96-273 Enclosure 1 The St. Lucie Plant Off Normal Operating Procedure, 1-0830030, Steam Generator Tube Leak, has been changed to include a flow chart that directs increased monitoring or plant shut down based on rate of change of leakage and absolute leakage (plant shut down commenced at .1 GPM).
These actions and analyses assure that the expected end-of-cycle steam generator condition will not significantly increase the probability or consequence of a previously analyzed accident.
Does the proposed operating interval for the Unit 1 steam generators create the possibility of a new or different kind of accident from any accident previously analyzed?
The degradation identified only affects the steam generator tubing. Currently there exist UFSAR safety analyses which address the failure of the steam generator tubing and the subsequent consequences. The analyses contained in this report demonstrate with high probability that steam generator tubing structural integrity is maintained for the proposed operating interval. It concludes that in the unlikely event of a main steam line break with consequential tube rupture, with current administrative limits on reactor coolant system dose equivalent iodine, the resulting offsite doses are less than 10 CFR 100 limits. The analyses indicate that the probability of tube ruptures either as an initiating event or the consequence of a MSLB is 4.4x10'. A risk assessment indicates that there is a negligible effect (1.0x104) on the core damage frequency (CDF) associated with operation of the Unit 1 steam generators for 15 EFPM. The baseline core damage frequency is 2.32x10'er reactor year, and therefore CDF is negligibly increased (much less than 1%) for consequential tube ruptures due to the predicted propagation of steam generator tube axial cracks.
Therefore, as the failure only affects steam generator tubing which has been analyzed in the UFSAR and the consequences a failure of the existing tubing are bounded by those analyses, the possibility of a new or different kind of accident than currently analyzed in the St. Lucie Unit 1 UFSAR is not created for the proposed operating run.
Does the proposed operating interval involve a significant reduction in the margin of safety?
I The analyses contained in this report indicate that the structural integrity margins specified in Regulatory Guide 1. 121 have been satisfied to a high probability and confidence level. Additional actions have been taken to assure safe plant operation. More restrictive administrative limits for primary-to-secondary leakage adds significant margin over limits currently specified in the St. Lucie Unit 1
St. Lucie Unit 1 16 Docket No. 50-335 L-96-273 Enclosure 1 Technical Specifications. Leakage observed during end of Cycle 13 in-situ pressure testing is considered conservative and has been reflected in leakage modeling in the APTECH report. These conservative limits provide additional assurance that an orderly shutdown will be conducted prior to a through-wall leak propagating to a rupture.
These measures provide reasonable assurance that there are no reductions in the required safety margins, and that the St. Lucie Unit 1 can be safely operated for a period of 15 EFPM during Cycle 14.
- 8) "Submittal of a technical specification (TS) amendment request that commits to the preceding actions and provides TS pages according to the guidance discussed in GL Attachment 2, including the associated "no significant hazards consideration" (10 CFR 50.92) and supporting safety analysis."
FPL is currently meeting the criteria in the St. Lucie Unit 1 Technical Specifications with regard to steam generator tube repair. As a result, submittal of a technical specification amendment and its supporting analyses are not required at this time.
4.0 CONCLUSION
S This assessment has concluded that continued operation of St. Lucie Unit 1 steam generators in the existing condition is acceptable since this evaluation has shown that no accidents or safety concerns outside of those analyzed in the FSAR are generated, and that operation within the current Technical Specifications will continue to provide appropriate detection and control parameters. The margin of safety for the plant, as defined in the basis of any technical specification, has been evaluated and is not significantly reduced.
The analyses and evaluations contained in this report demonstrate that the operating, inspection and repair program for the St. Lucie Unit 1 steam generators permit safe operation for the specified Cycle 14 operating period. FPL has implemented industry recommended secondary chemistry controls to mitigate further initiation and propagation of ODSCC. FPL meets EPRI action levels for sulfate by requiring reduced power operation or shutdown for elevated sulfate levels. Improved operational response has been
. implemented by providing plant operators the ability to detect and respond to changes in steam generator primary to secondary leakage, and shutdown the unit prior to a significant leak or tube rupture should tube degradation exceed expected values. State of the art probabilistic models have been developed to demonstrate that operating cycle lengths will maintain the safety margins specified in Draft NRC RG 1 121. A probabilistic leakage
~
model has been developed to assess end of cycle leakage as a result of postulated accident
. conditions, Leakage observed during end of Cycle 13 in-situ pressure testing is considered conservative and has been reflected in leakage modeling in the APTECH report. The leakage model demonstrates that operation with the existing condition of the tubing will
St. Lucie Unit 1 17 Docket No. 50-335 L-96-273 Enclosure 1 not result in off site releases in excess of 10CFR100 limits should a MSLB event occur during Cycle 14 operation.
It is FPL's position that this evaluation demonstrates that the operation, inspection and repair program described in this evaluation constitutes a conservative approach which ensures that adequate structural and leakage integrity is maintained for normal operating, transient and postulated accident conditions for the specified Cycle 14 operating period for the St. Lucie Unit 1 steam generators. The analyses contained within this evaluation allow operation of St. Lucie Unit 1 for 15 EFPM (at T,greater or equal to 500'F) during, the current Cycle 14 operating cycle. Therefore, Cycle 14 operation for the St. Lucie Unit 1 steam generators, from their startup date of July 25, 1996 through October 23, 1997, will be in accordance with the guidance contained in NRC GL 95-05.
5.0 REFERENCES
APTECH Report AES 96052749-1-1, Analysis of ODSCC/IGA at Tubesheet and Tube Support Locations at St, Lucie Unit 1, APTECH Engineering Services, October 1996.
P NRC Generic Letter 95-05, Voltage-Based Repair Criteria for Westinghouse Steam Generator Tubes A+ected by Outside Diameter Stress Corrosion Cracking, Aug. 3, 1995.
- 3. Draft Regulatory Guide X.XX, Steam Generator Tube Integrity, U. S. Nuclear Regulatory Commission.
- 4. CE Report 00000-OSW-16 Revision 0, In-situ Pressure Test Results for St. Lucie Unit. 1 Spring 1996 Outages, ABB Combustion Engineering, 1996.
- 5. Draft NRC RG 1. 121, Bases for Plugging Degraded PWR Steam Generator Tubes, U, S.
Nuclear Regulatory Commission.
- 6. SAIC Report SAIC 05-5049-05-6734-500, Steam Generator Degradation Specific Management (SGDSM) Leakage Limit Calculation for St. Lucie Unit, Scientific Applications International Corporation, September 1996.
- 7. FPL Letter I 96-166, June 25, 1996, J. A. Stall, St, Lucie Plant Vice President to U.S.
Nuclear Regulatory Commission, Steam Generator Tube Inspection - Request for Additional Information (RAI) Response.
- 8. EPRI NP-5397-LD, Laboratory Evaluation of Steam Generator Tubes 120/12, 79/9 and 59/95 from St. Lucie Unit 1, Electric Power Research Institute, August 1987.
9 ASME Paper 89-JPGC/NE-6, St. Lucie Unit 1 Steam Generator Integrity Program, presented at the Joint ASME/IEEE Power Generation Conference, Dallas, Texas - October 22-26, 1989.
St. Lucie Unit 1 18 Docket No. 50-335 L-96-273 Enclosure 1
- 11. EPRI NP-6201, Rev. 3, PWR Steam Generator Tube Examination Guidelines, Electric Power Research Institute, November 1992.
- 12. ABB/CE Report CR-9417-CSI93-1129, Rev. 1, Assessment of Deposit Potentials in Florida Power & Light Company's St. Lucie Units 1 and 2 Steam Generator Tube Bundles, ABB Combustion Engineering, February 1994.
- 13. CEOG Task 888 Report, Rev. 1, CEOG Support of Utility Responses to NRC Generic Letter 95-03: Circumferential Cracking of Steam Generator Tubes, ABB Combustion Engineering Owner's Group, October 1995.
- 14. FPL Document CSI-ET-96-11, Rev. C, May 1996 Eddy Current Examination Plan for Steam Generator Tubing at St. Lucie Unit 1, Florida Power and Light Company, 1996.
- 15. St. Lucie Unit1 Eddy Current Data Analysis Guideline And Performance Demonstration, Florida Power and Light Company, May 1996.
~ ~ ~
16.~ JPN-PSL-SEMP-96-066, Rev. 0, Engineering Evaluation of the 1996 St. Lucie Unit 1
~
Spring Outage In-Situ Pressure Test Results, Florida Power and Light Company, July 1996.
St. Lucie Unit 1 19 Docket No. 50-335 L-96-273 Enclosure 1 Figure 1- Steam Generator Sludge Removal Raised 11/QS PSL Steam Generator Sludge Removal History UNIT 1 4,000 3 900 3,500 3,000 2,500 2,000 1,500 1,100 1,000 590 496 500 400 295 295 167 1984 1985 1987 1988 1990 1991 1993 1994
+ 9/G Bundle Rush 'A'- 67 'A' 160
++ Dry Weight 'B' 100 'B' 135
St. Lucie Unit 1 20 Docket No. 50-335 I 96-273 Enclosure 1 Table 1 - Steam Generator ECT and Tube Plugging History
";jEXAM$
O'::::DAYE': '"
Preservice Sho Plu 10%
1978 10%
1979 10% 21 DrBIed Su rt Rim Cut 1981 A -28% 103 U-Bend Cracking Rows 8-11 B -38%
1984 Preventive Plugging Rows 8-11 Random Indic. - Stud e Pile k, Su rts 1985 (8XI) Random Indic. - Sludge Pile 8c Supports Pulled 3 Tubes 1986 id cle PostOuta eECI Review@,Plu 1987 ~
(50 Bobbin Indications 76 Random Indic. - Slud e Pile dc Su rts 1988 <50 Bobbin Indications Random Indic. - Slud e Pile Ec Su rts 1990. 97 Bobbin Indications 180 Random Indic. - Slud e Pile dc Su rts 1991 100% HL Expan.Trans.
3% CL Expan. Trans, 98 Circ. Indications - HL Expan. Trans.
650 Bobbin Indications 1993 100% HL Expan. Trans. 130 Random Indic. - Sludge Pile dc Supports 3% CL Expan. Trans, 250 Bobbin Indications 12 Circ. Indications - HL Ex . Trans.
1994 100% HL Ex pan. Trans. Random Indic. - Sludge Pile Jc Supports 3% CL Expan. Trans. 14 Circ. Indications - HL Expan. Trans.
100 Bobbin Indications 1996 100% HL 8c CL Expansion 2081 Random Indic. - Sludge Pile 8c Supports.
Transition. Allupper Freespan 160 Circ. Indications (10 CL). -800 Tubes Regions plugged due to revised criteria for driued support plates. 44 Axial freespan indications u r bundle .
TOTAL TUBES Each = 8519 PLUG MARGIN - 2555 Tubes (30%) each S/G TOTAL PLUGS S/G A 2159 (25.3%) +/- 7% Asymtnetry S/G B 1834 21.5%
psllhist.ect
,d St. Lucie Unit 1 A-1 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Request For Additional Information Responses (NRC Letter dated July 16, 1996)
NRC REQUEST 1 Clarify how free span indications which are not crack-like and which can not be traced back to the preservice baseline inspection were dispositioned.
FPL RESPONSE Free span indications which are not crack-like and which can not be traced back to the preservice baseline inspection were removed from service without regard to depth sizing.
NRC REQUEST 2 Discuss the extent to which the data used in the sizing qualification program have similar morphologies and eddy current responses (e.g., extent of intergranular attack, cracking, voltage responses, noise levels, etc.).
Discuss the basis for including data from different frequencies in the qualification program (4$ kHz and 560 kHz) given that the ability to reliably size defects will, in part, depend on the frequency.
FPL RESPONSE SIZING QUALIFICATIONSAMPLES Qualification data was provided to NRC staff in FPL letter L-96-166, June 25, 1996. The data set used to establish sizing qualification was constructed using pulled tube data from operating plants with IGA/SCC and supplemental laboratory IGA/SCC samples. The hboratory samples were made by Battelle-Pacific Northwest Laboratories using fabrication techniques identical to samples used by NRC research to establish NDE system performance. From a.metallographic perspective, both the lab and plant pulled tube samples are described as having IGA/SCC. In this sense the samples are indistinguishable.
IGA/SCC morphology for the samples is predominantly axial with some of the St. Lucie 1 pulled tube samples exhibiting volumetric IGA and minor amounts of transgranular cracking..
The table below provides a listing of the qualification samples with attributes that describe their condition from a metallographic and eddy current perspective. The samples are
St. Lucie Unit 1 A-2 Docket No. 50-335 I 96-273 Enclosure 1 Appendix A approximately equally distributed as to their origin. The average maximum crack lengths are comparable while the St. Lucie 1 average maximum depth is somewhat lower compared to the lab and other plant samples. This difference is attributed to the presence of near or through wall cracks in the lab and other plant samples. Ifnear or through wall cracks from lab or other plant samples are excluded, comparable values (in parenthesis) are obtained. 'Oe degraded angular extent for all three sample sets is comparable in that the predominant orientation is axial.
The average bobbin coil amplitude for the various sample sets ranges from 0.86 volts to 6.11 volts (data provided in letter FPL L-96-166, June 25, 1996). A plot of individual sample voltages versus depth is provided below. As can be seen, higher voltages are contributed by samples with near or through wall cracks. Ifthese values are excluded, the average amplitudes for the lab samples are reduced significantly, while average values for St. Lucie 1 and other plant samples remain largely unchanged. Average through wall depth sizing error ranges from +1% for the lab samples to -10.8% for other plant samples. The average sizing error of -6% through wall for St. Lucie 1 samples is approximately halfway between the lab and other plant sample sizing error. The sizing error standard deviation for the lab, St. Lucie 1 and other plant data is 10.75, 6.32 and 31.4 respectively. The error contributed by the lab samples is conservative in that it is larger than the St. Lucie 1 sample set value. Sizing error contributed by other plant data is roughly twice that of the St. Lucie 1 data set. Therefore, inclusion of lab and other plant data is conservative in that it contributes both towards systematic sizing error and larger standard deviation.
Signal-to-noise ratios for the qualification data set are controlled by the voltages of the individual indications and the local noise level established by the coil field-of-view. From the qualification data provided in FPL L-96-166, June 25, 1996, only two of the eighteen samples exhibit comparatively high voltages. These voltages were contributed by lab samples with through wall cracks. The voltages for the remainder of the samples are viewed as comparable. Potential noise sources include electronic or instrumentation noise, tube noise, and environmental noise. A review of the eddy current graphics shows no electronic noise. Tube noise is usually caused by pilgering or local variations in tube geometry, e.g., dings, denting, ID chatter, etc. A review of the eddy current graphics for the qualification sample set shows no evidence of tube noise. Sources of environmental noise include tube deposits and secondary-side support structures. Review of the eddy current graphics does not show evidence that tube deposits are influencing the measurement process at the primary analysis frequency. Support structures are present for eight of the eighteen samples. While the edges of the support structure are visible in the overall signal pattern, the edges do not strongly influence the measurement process. This minor influence is within the bounds of the qualification program which addressed sizing for free-span degradation and degradation roughly centered within the support structure typical of plant experience.
St. Lucie Unit 1 A-3 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A BASIS FOR INCLUDING DATA ACQUIRED AT DIFFERENT FREQUENCIES For appropriately normalized data acquisition frequencies in the context of phase angle spread, sizing accuracy is independent of test frequency numeric value. The essential test variable is phase angle spread between two reference discontinuities, usually 100% and 20% through-wall holes, and not the numerical value of the coil driving frequency. A coil drive frequency of 400 kHz is typically used for 0.048 "4.050" wall tubing while a higher frequency, e.g., 550 kHz', is used for thinner 0.043" wall tubing. The phase angle spread between the two reference holes in each of the two different wall thickness tubes. is similar; accordingly, the test conditions are comparable.
't should be noted that the frequency of 560 kHz is incorrectly stated in previous submittal material. The correct value is 550 kHz and represents the condition in which data were acquired.
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A FPL RESPONSE (NRC REQUEST 2 CON'T)
PSI 1 Sizing Qualification Data Set Description Sample Average Range in Average Axial Average Average Origin Samples maximum maximum maximum /Circ. amplitude, slzlllg Length, in depths, % depth,% Volts error, %
TW 0.383 16-100 55 6.11(1.0&)*
37.8 4 PSL-1 0.566 13-72 37.5 Axial 1.46 Other 0.666 49-100 0.86 Plants 0.68 4
- Excluding samples with near or 100% through-wall cracks
~ ~ ~ r ~ ~ ~ 1 a
Cl c 1 8
4+
Oe .
8
( ~
- ~
C
, ~
~
RO CO 60 80 100 Maximum Metallographic Depth, % TW PSL-1 Sizing Qualification Data Set Voltage versus metallographic maximum percent through wall
St. Lucie Unit 1 A-5 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A NRC REQUEST 3 In the June 25, 1996, response to NRC Request 3, it was indicated that the burst pressure correlation that will be used relies on average crack depth. The qualification data for the sizing program relies on maximum crack depth. Discuss how the uncertainties associated with converting from maximum to average crack depth will be addressed (e.g., using a curve similar to that in Figure 3 of Attachment A to the June 25, 1996 letter). Provide the supporting metallographic and eddy current data to support this approach.
. Discuss how the uncertainties in the predicted burst pressure will be accounted for in the probabilistic methodology (i.e., the predicted burst pressure does not exactly match the observed burst pressure as is illustrated in Figure 1 of Attachment A).
In Figures 4, 5, and 6 of Attachment A to your June 25, 1996 letter, information regarding the distribution of material properties expected in the St. Lucie Unit 1 steam generator tubes is provided. Discuss if this data is from steam generators tubes that have been in service. Ifnot, provide the distribution of material properties based on destructive examination of pulled tubes from similarly fabricated Combustion Engineering steam generators. Provide the mean and standard deviation of this data along with the 95/95 confidence value (i.e., the lower tolerance limit).
Discuss how the growth rate distribution was developed. Discuss how the average growth rate was determined.
Clarify whether a lower 95% prediction interval curve for the burst pressure versus crack size correlation was used in the deterministic run-time evaluation. Use of the lower 95%
prediction interval curve is consistent with the approach in Generic Letter 95-05. Please provide for comparative purposes a deterministic run-time evaluation assuming a lower 95% prediction interval curve for the burst pressure correlation adjusted for lower bound (95/95) material properties based on the destructive examination of inservice steam generator tubes. This analysis should also assume a 95% cumulative probability value for non-destructive examination uncertainty.
FPL RESPONSE Burst pressures are calculated from combinations of structurally significant depth and structurally significant length. Structurally significant lengths are obtained in probabilistic calculations by sampling from a distribution. Based on pulled tube data, the measured distribution of axial degradation lengths obtained from RPC eddy current inspections is a conservative representation of the distribution of structurally significant lengths.
I The Framatome equation acting in concert with the structural minimum method is an appropriate bounding burst pressure equation (see circled points in attached Figures 3 and
St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A 4). Strengthening effects of ligaments between axial crack segments are ignored. These strengthening effects are evident by the burst data which lies to the right of the burst line in Figure 3. A maximum strengthening effect of a 40% increase in burst pressure was observed relative to the calculated burst pressure. Note that the only ligament in the present burst model is the ligament in the depth direction of a single planar crack. Actual degraded, regions are composed of several axial crack segments. Ligament between segments are an important unmodeled element of conservatism which makes use of the Framatome equation a conservative approach. A blind statistical fit to all of the pulled tube burst data is technically unsound. In contrast to pure correlation approaches, there is no physical basis to infer low side scatter to the physically based burst pressure relationship from high side scatter provided by crack segment ligaments.
Variations in crack morphology are treated probabilistically. Pulled tube data from Plant A is used to sample crack morphologies. Observed ratios of maximum depth to structurally significant depth are used to generate variations in physical crack shape which then impact detection, sizing, burst, and leak calculations.
The growth rate distribution was obtained from measured cycle to cycle changes in crack depth. An analysis has been conducted showing that the extremes of the measured growth rate population bound the extremes of the actual growth rate population. Point by point discrepancies can be large but the data set is extensive and the extremes in actual growth rates are bounded by the measured growth rates.
In a correlation approach, such as burst pressure versus bobbin voltage, there is no true relationship of the dependent variable to the independent variable. There is only a vehicle to describe scatter. Hence, it is imperative to deal statistically with a lower bounding limit. This is not the case with a physically based approach. For a given crack geometry, there is a true, unique burst pressure, which is, in principle, knowable. For a given bobbin voltage, there is virtually a limitless possible array of crack geometries. If only bobbin voltage is known, there is no unique associated burst pressure. It cannot be calculated, even in principle. At a given bobbin voltage, it can only be stated that the past set of burst pressure measurements can be analyzed statistically to make statements regarding the probability of observing a burst pressure of some selected magnitude. In contrast, there is a true, unique burst pressure for a given crack geometry in an Alloy 600 steam generator tube. The Framatome equation in concert with the structural minimum method is demonstrated by test data to be very close to this true relationship for the relatively long crack lengths of interest.
NRC REQUEST 4 In the June 25, 1996, response to NRC Request 4, EPRI report TR-104788 was cited.
EPRI report TR-104788 contains guidelines for primary-to-secondary leakage monitoring programs. In this report, deviations from the guidelines are considered acceptable
St. Lucie Unit 1 A-7 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A following plant-specific evaluations. Please clarify any significant deviations taken from these guidelines. In addition, specifically address the leakage limits/action levels to be implemented at St. Lucie Unit l.
FPLRESPONSE No significant deviations are taken from the EPRI Leak Rate Monitoring Guidelines.
Specific leakage limits/action levels implemented via St. Lucie Plant procedures are as follows:
NORMAL OPERATION < = to 5 gallons per day (GPD).
INCREASED MONITORING > 5 GPD but < 30 GPD.
ACTION LEVEL 1 > = 30 GPD and < = 150 GPD with
< = 60 GPD/Hr increase.
ACTION LEVEL 2 > 150GPDor
> 60 GPD/Hr Increase NRC REQUEST 5 In the June 25, 1996, response to NRC Request 5, it was indicated that undetected flaws will be projected to give the end-of-cycle (EOC) distribution of through-wall cracks.
Discuss whether the detected flaws will also be included in this projection. Discuss the need to include in the leakage analysis indications which are near through-wall and will pop-through the wall under postulated accident conditions. Discuss how nondestructive examination (NDE) uncertainty will be accounted for in the projection of the EOC distribution.
The stafF has accepted a value for the probability of detection which is independent of flaw depth of 0.6. Clarify the value of the probability of detection that will be used in your analyses for predicting the EOC distribution of indications. If different than the 0.6 value discussed above, provide a sensitivity study using this value.
Clarify whether the analyses to be performed to determine the EOC distribution will start from a beginning of cycle distribution which has been adjusted for the probability of detection and the number of indications repaired similar to the methodology described in Generic Letter 95-05.
St. Lucie Unit I Docket No. 50-335 L-96-273 Enclosure' Appendix A FPL RESPONSE
. Detected and undetected flaws are specifically treated in the initiation, growth and detection probabilistic model. This is a major advantage of a multi-cycle model which specifically. includes initiation.
Pop-through events are included in leakage calculations. Consideration of the fracture stability of cracks with through wall segments is a recent refinement. This is the most advanced treatment of pop-through and tearing of through wall axial crack segments in any run time model.
Sizing errors are probabilistically treated in the analysis. Both actual and perceived (NDE measured) crack depths are tracked. The real distribution of EOC crack depths is computed along with the "perceived" distribution of EOC crack depths. The model tracks four cycles of operation.
The use of a constant POD value of 0.6 obviates the need to deal specifically with crack initiation. It also confines reasonable predictions to a single cycle of operation and adds a severely conservative level of degradation to the BOC population. A multi-cycle, physically based model, which specifically deals with initiation and is bench marked against post NDE observations must include consideration of the physically based fact that deep crack like degradation is more detectable than shallow crack like degradation. This provides for realistic numbers of indications growing into the detectable range over multiple cycles of operation. Initiation, growth and detection are inextricably woven together to provide reasonable simulation of actual NDE observations over more than one cycle of operation. Unnecessary or arbitrary conservatism in any of these areas distorts the entire model.
While POD is modeled as being dependent on depth, probability of burst calculations are not highly sensitive to reasonable uncertainty in the location of this curve. Analyses of the simulation results in terms of the BOC crack depths, growth rates and mechanical properties of simulated burst events shows that large undetected cracks are not a dominant contributor to the frequency of simulated burst. Even modest detectability at large depths, ifapplied over multiple cycles of operation, effectively identifies and removes deep crack contributors to burst.
NRC REQUEST 6 Discuss and provide the qualification data for the leakage model that will be implemented.
Discuss how the results from the in-situ pressure tests will be factored into the leakage model.
St. Lucie Unit 1 A-9 Docket No. 50-335 I 96-273 Enclosure 1 Appendix A FPL RESPONSE The leakage model is based on crack opening areas calculated from equations in the Ductile Fracture Handbook and bench marked against measured axial and circumferential crack opening displacements. Leak rates through these crack opening areas are based on PICEP analyses. The methodology for circumferential cracks was developed as part of the industry wide EPRI circumferential cracking program.
In-situ pressure test results provide two benchmarks for the leakage model. The first is the number of leaking defects. In the in-situ tests, three defects showed leakage. This outcome can be compared with the simulation results for this operating period as shown in Appendix A Figure 1. As can be seen from the figure, the occurrence of three leaking defects under steam line loading is a reasonable expectation. The second benchmark is the total quantity of leakage which was on the order of 0.5-1.0 GPM. As can be seen from the simulation results shown in Appendix A Figure 2, the outcome has a probability of between 8% and 12% which constitutes a reasonable outcome in the probabilistic sense.
NRC REQUEST 7 Provide tabularized and graphical data for the distribution of indications detected (length and depth) for each steam generator, the distribution of indications repaired for each steam generator, the growth rate of indications (length and depth) for each steam generator, the material properties distribution, the NDE uncertainty models, and the burst pressure correlation. For the distribution of indications detected and repaired, the growth rate distribution, and the material properties distribution, provide the number of indications with depths/lengths/material properties within a given interval (e.g., 10 indications with depths between 35% and 40% through-wall). For the remaining distributions, provide the data in a format suitable for the staff to independently verify the results of the tube integrity analysis.
FPLRESPONSE Available tabulated and graphical data for indication distributions, indication growth rates, material properties distributions, NDE uncertainty models, and the burst pressure correlation are provided in Appendix A Figures 3 - 14. Numbers of indications within given intervals for depth and location are provided in Appendix A Figures 13 and 14.
Independent verification of the tube integrity analysis results by the staff would require access to APTECH Engineering proprietary source code which is not available.
St. Lucie Unit 1 A-10 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A NRC REQUEST &
In the June 25, 1996, response to NRC Request 7, it is stated that the "burst pressure will be treated deterministically." Clarify what is meant by this statement.
FPL RESPONSE As noted in other answers above, the axial crack burst pressure relationship is based on the Framatome equation and the structural minimum method. This procedure provides a reasonable lower limit curve. It is a conservative treatment of scatter in comparisons of measured versus calculated burst pressure. The strengthening effects of ligaments between segments of axial cracks is neglected. This effect provides the bulk of the noted data scatter. Correct treatment of this scatter would decrease the probability of burst, not increase it. There is a difference in kind between correlation approaches to burst pressure and physically based approaches.
St. Lucie Unit 1 A-11 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 1 - Predicted Number of Leakers SG-A 1994/1996 Operating Cycle 4000 3500 ---'-
3000 2500 2000
+00 1000 500 0
0 2 3 4 6 Expected NUMBER OF LEAKING DEFECTS AES96052749-1-1
~ ~ I I I I I
~ I ~ ~
I ~
~ ~, ~ ~
~ ~ I
~ >
~ I I ~
0 I C
St. Lucie Unit 1 A-13 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 3 - Predicted Burst Pressure Versus Normalized Observed Burst Pressure, Plant A Pulled Tube Data O4 4
0 0
St. Lucie Unit 1 A-14 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 4- Results of EDM Slot Machined Sample Burst Test Program Iaaao STRUCTURAL MINlMUM METHOD FRAMATOME EOUATlON 4
eaao 5oao g
sea o <oao @co oooo 4oe axe eooo vae axe Oooo soaao MEASURED ElJRST PRESSURE, P5l hES96052749-1 1
St. Lucie Unit 1 A-15 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 5 - Tubing Tensile Properties Used In Probabilistic Analyses PLANT B SG A = SG B 1800 1600 1400 1200 1000 Q
800 R
600 400 200 0
O O 00 O O O O O O lA O
tD O
la O ln O lA O
EA O
lA 0O lA O
O lA IA O
O M
O O
lA O
O lA O N W tu ce O CO <<Cl O Ol CV CV Ol Ol PJ 07 P7 CO SUM OF YlELD AND ULTlMATESTRENGTH AT TEMPERATURE, PSl
~ ~ ~ I I ~
~
C s . ~ s ~ ~
~ ~
s . ~ ~ ~
M II~R~IP,'M~llll maaarmmammraaaaam)
~WRSIIIII~WREllll,~M~II MRSHIIIIMRRSBF.lMHRHIIII MHSHHIIMHLU'IMHEHIIII
~WS&IIII~Mlflll~&MHIII
~~IIII balll II MSIIIIIIIWRII'illll
~
St. Lucie Unit 1 A-17 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 7 - Bobbin Probe Depth Calls Versus Maximum Depth Destructive Examination Results (FPL Sizing Procedures Applied)
APPROPRIATE FOR AXIALINDICATIONS SLUDGE PILE, AND LATTICE TYPE TUBE SUPPORT LOCATIONS 14 4
4 Sn . m aa m m ~' '
MAXIMUMCRACK OEPTHr % THROUGH WALL OEBTRUCTlVE EXAMINATION hES985 2749.1-1
A-18 St. Lucie Unit 1 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 8 - Bobbin Depth Calls Versus Maximum Depth Destructive Examination Results, Drilled Tube Support Plates EPRl Bobbin Coil Qualification for ODSCC at Drilled Support Plates FPL Sizing Technique (excludes NQI calls) t n 70 40 0 +
+
30 O
IQ 10 10 20 30 40 50 60 70 Maximum Crack Depth, % Through Ylall Destructive Examination
St. Lucie Unit 1 A-19 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 9- Bobbin Probe Log Logistic Probability of Detection Curve Used in Probabilistic Evaluations. Instances of Bobbin Probe Detected Degradation less than 40% Excluded O
gf, L4 0 70 N I0 10 20 3) 40 00 IO 0
IIAXWUMCRACIC CEPS, % THROUGH WALL AES960I27W-I
St. Lucie Unit 1 A-20 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 10 - RPC Probability of Detection Based On Plant A Pulled Tube Data MRFC PROBABKJTY OF DETECT10M VERSUS NOAE CRACK OEPTH PLANTA PULLED TUSE DATABASE 0.1 0 IO 20 Xl & 50 5) 70 IO N7 ljKI STj5tCTURALAVERAGECRACK DEPTH,5 TTIIC)UCjjlhNAL Apjcch Eaghear jag Saving, Ing. AES96QS2740.l-l
St. Lucie Unit 1 A-21 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 11 - Change in Percent Degraded Circumferential Area per EFPY ST. LUCIE UNIT 1 +X X +-S/GA S/GB~ X X ~ PLANT B 0
0@i 0
0 A7 0X R
O
)
IC X0 O X0 0
O I
D V
15 DELTA PDAIEFPY
St. Lucie Unit 1 A-22 Docket No. 50-335 L-96-273 Enclosure 1 Appendix A Figure 12 - Axial Degradation Growth Rates from Bobbin Indications at St. Lucie Unit 1 o Zoo 150 g e i o ep ac we o o r e g co o r GROWTH RATE, % THROUGH WALL/EFPY AES960$ B49-1 1
ggg, .
I
- I I I
KR I
I I
MIBM~~
~HM ME~
I
~
I. I:
I:
I:
I '
I I:
~e I:
I:
I GHZH~MMHG~KiHi ~~i~~is i
V Iy
(