L-2018-234, Subsequent License Renewal Application Safety Review Request for Additional Information (RAI) Set 5 Response 4.3.5-2 Revision

Subsequent License Renewal Application Safety Review Request for Additional Information (RAI) Set 5 Response 4.3.5-2 Revision

ML18362A146
Person / Time
Site:	Turkey Point
Issue date:	12/21/2018
From:	Maher W Florida Power & Light Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
L-2018-234
Download: ML18362A146 (11)

Text

L-2018-234 10 CFR 54.17 December 21, 2018 U.S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, D.C. 20555-0001 Re: Florida Power & Light Company Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 Turkey Point Units 3 and 4 Subsequent License Renewal Application Safety Review Request for Additional Information (RAI) Set 5 Response 4.3.5-2 Revision

References:

1. FPL Letter L-2018-004 to NRC dated January 30, 2018, Turkey Point Units 3 and 4 Subsequent License Renewal Application (ADAMS Accession No.

ML18037A812)

2. FPL Letter L-2018-082 to NRC dated April 10, 2018, Turkey Point Units 3 and 4 Subsequent License Renewal Application - Revision 1 (ADAMS Accession No.

ML18113A134)

3. FPL Letter L-2018-175 to NRC dated October 17, 2018, Turkey Point Units 3 and 4 Subsequent License Renewal Application - Safety Review Requests for Additional information (RAI) Set 5 Responses (ADAMS Accession No.

ML18292A642)

On April 10, 2018, Florida Power & Light Company (FPL) submitted to the NRC Revision 1 of the subsequent license renewal application (SLRA) for Turkey Point Units 3 and 4 (Reference 1), as well as supplemental information for the SLRA Environmental Report (ER) (Reference 2).

The purpose of this letter is to provide, as an attachment to this letter, the revised response to safety review Set 5 RAI 4.3.5-2 regarding the conditional probability of failure (PoF) for reactor coolant pump (RCP) flywheels. This revised response supersedes the corresponding RAI response provided in Reference 3.

If you have any questions, or need additional information, please contact me at 561-691-2294.

Florida Power & Light Company 700 Universe Boulevard, Juno Beach, FL 33408

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 L-2018-234 Page 2 of 2 I declare under penalty of perjury that the foregoing is true and correct.

Executed on December 21, 2018.

--~~--

Sincerely, William Maher Senior Licensing Director Florida Power & Light Company WDM/RFO

Attachment:

FPL Revised Response to NRC RAI No. 4.3.5-2 cc:

Senior Resident Inspector, USNRC, Turkey Point Plant Regional Administrator, USNRC, Region II Project Manager, USNRC, Turkey Point Plant Plant Project Manager, USNRC, SLRA Plant Project Manager, USNRC, SLRA Environmental Ms. Cindy Becker, Florida Department of Health

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 1 of 9 NRC RAI Letter Nos. ML18260A242 and ML18260A243 dated September 17, 2018 RAI 4.3.5-2

Background:

To support its TLAA disposition of§ 54.21 (c)(1 )(i), the applicant included the PWR Owners Group 'report, PWROG-17011-NP, Rev. 0, August 2017 in Enclosure 4 of the SLRA. PWROG-17011-NP provides the generic SLR methodology for deterministic and risk-informed analyses related to integrity of Westinghouse RCP motor flywheels.

PWROG-17011-NP is not approved by the NRC for use in SLR applications. In order to complete its review of this TLAA the staff must determine whether the applicant's proposed implementation of the generic SLR flywheel methodology in PWROG-17011-NP is acceptable for demonstrating, per§ 54.21 (c)(1 )(i), that the CLB analyses of the PTN 3 and 4 flywheels will remain valid for the subsequent period of extended operation (SPEO).

Issue:

The risk assessment in PWROG-17011-NP, Section 3 used the probabilistic fracture mechanics (PFM) analysis methodology to generate conditional probability of failure (PoF) for reactor coolant pump (RCP) flywheels for the 80-year risk assessment. The staff noted that the PoFs for the case when ISi was performed every 4 years for only the first 10 years of operation in PWROG-17011-NP for 80-year SLR terms are lower than the corresponding PoFs in WCAP-15666-A for 60-year initial LR terms. The reason for this is not clear. In theory, when a selected flaw is given 20 more years to grow without any ISi, the PoF should be higher.

Request:

Please explain why the PoFs for the .ISi case documented above are lower for the 80-year SLR analysis in PWROG-17011-NP compared to the corresponding PoFs for this ISi case for the 60 year LR analysis in WCAP-15666-A.

If there is an error in the PoF analysis for this ISi case, please provide the correct PoF calculations for the 60-year and/or 80-year analyses, as needed. Please revise the PWROG-17011-NP report, as needed, to show the correct PoF calculations.

FPL Revised Response:

This revised RAI response supersedes in its entirety the RAI response provided in of Reference 6 [6].

PWROG-17011-NP [1] is a generic topical report that supports reactor coolant pump (RCP) Flywheel 20-year inspection intervals for plants with Westinghouse original equipment manufacturer (OEM) RCPs. This report is an update to WCAP-14535A (40-year) [2] and WCAP-15666-A (60-year) [3], both reports receiving a Safety Evaluation Report (SER) from the Nuclear Regulatory Commission (NRC), as PWROG-17011 extends the evaluation period of the flywheel analysis to 80 years of operation. This

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 2 of 9 work was done under the purview of the Owners group and performed to support the fleet as plants begin to apply for extending their license from 60 to 80 years.

As in the 40 and 60 year evaluations in [2 and 3], the analysis performed in PWROG-17011 consists of both a deterministic stress and fracture evaluation and a probabilistic risk assessment. The risk assessment is comprised of two parts: probabilistic fracture mechanics (PFM) evaluation to determine probability of RCP flywheel failure, and a core damage evaluation. The PFM analyses used a Westinghouse developed computer program, RPFWPROF, that calculates the difference in RCP flywheel failure probabilities with and without In-Service Inspections (ISi). All three reports, WCAP-14535A, WCAP-15666-A, and PWROG-17011-NP include both the deterministic and probabilistic analyses. The NRC's SER forWCAP-14535A was solely based on the deterministic evaluation and granted an ISi interval of 10 years. The NRC's SER for WCAP-15666-A reviewed both the deterministic and probabilistic evaluation and granted an ISi interval of 20 years.

In seeking to address the NRC's RAI above, Westinghouse engineers determined that the RPFWPROF executable file used in PWROG-17011-NP cannot reproduce the results of WCAP-15666-A when run on original computer platforms. This confirms that Westinghouse had lost configuration control on the PFM executable program, RPFWPROF.

To correct for this, Westinghouse has performed a review of the deterministic aspects of the analysis to ensure their continued appropriateness and then also re-established configuration control of the RPFWPROF program and determined revised RCP flywheel probabilities of failure for 40, 60, and 80 years of operation.

Engineering Assessment of the Deterministic Evaluation The function of the RCP in the reactor coolant system (RCS) of a pressurized water reactor plant is to maintain an adequate cooling flow rate by circulating a large volume of primary coolant water at high temperature and pressure through the RCS. A concern over overspeed of the RCP and its potential for failure led the NRC to issue Regulatory Guide (RG) 1.14 [5] in 1975. RG 1.14, Revision 1, Section C, subsection 2, provides the NRC regulatory guidance for flywheel design. Key provisions of this position are to analyze for ductile failure, analyze for non-ductile failure, demonstrate compliance with excessive deformation failure criterion, and demonstrate compliance with loss of coolant accident (LOCA) overspeed criterion.

WCAP-14535A Since the NRC granted the 10-year ISi interval based solely on the 40-year deterministic analyses in WCAP-14535A [2], the loss of configuration control of the RPFWPROF program does not affect the NRC staff decision. The deterministic evaluation contained in the report remains valid and meets requirements of RG 1 .14.

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 3 of 9

. WCAP-14535A classifies flywheels into 16 groups according to their material and geometric information, and selected six flywheel groups for evaluation, which encompass the range of domestic flywheel dimensions covered by the report. In Section 4.2 of the report [2], a linear elastic stress analysis was performed and demonstrates that the flywheel structure can resist ductile failure with sufficient margin of safety during faulted conditions by meeting the faulted condition criteria of ASME B&PV Code Section Ill. The calculated critical speed due to ductile failure satisfied the RG 1.14, Section C, items 2.f for normal speed of 1200 revolutions per minute (RPM),

and item 2.g, LOCA overspeed of 1500 RPM. The analysis performed and documented in WCAP-14535A to evaluate for ductile failure has been reviewed and confirmed to remain valid.

To evaluate for non-ductile failure, the analysis used the closed-form solution for a radial full-depth crack emanating from the bore of a rotating disk to calculate the applied stress intensity factor. The fracture toughness, Kie as a function of flywheel temperature and RTNoT, was calculated per the ASME Code,Section XI, A-4200, a lower bound curve. The assumed operating temperature of 70°F is conservative because the typical containment ambient temperature is 100°F to 120°F. Furthermore, the flywheel would receive additional heat from the pump motor due to friction and from reactor coolant through conduction through the pump itself. RTNoT was assumed to be 60°F for additional conservatism as the 1973 equipment specification requires RT NOT no greater than 10°F. The loads used in calculating the applied K1 were from an overspeed of 1500 RPM. As shown in Table 4-3 of [2], the critical crack length for flywheel overspeed of 1500 RPM are large, even considering the conservative flywheel temperature and RT NOT. As noted in the SER, the original submittal did not calculate a critical speed based on an assumed initial flaw as requested by RG 1.14. In response to this, allowable crack sizes were calculated in Table 1of [2] using the margins in IWB-3610 of Section XI of the ASME Code, and including the effect of shrink fit. The analysis performed and documented in WCAP-14535A to evaluate for non-ductile failure has been reviewed and confirmed to remain valid.

To demonstrate compliance with excessive deformation failure criterion, the analysis in

[2] used standard closed-form formulas for rotating disks to calculate the change of flywheel inner and outer radii at 1500 RPM. Results presented in Table 4-5 of [2]

demonstrate a maximum flywheel deformation of only 0.010 inches and conclude that this would not result in adverse conditions since the flywheel assemblies are typically shrunk fit to the flywheel shaft and the deformations are negligible. The analysis performed and documented in WCAP-14535A to evaluate for compliance with excessive deformation failure criterion has been reviewed and confirmed to remain valid.

WCAP-15666-A The purpose ofWCAP-15666-A [3] is to extend the ISi interval from 10 years [2] to 20 years for a 60 year operating lifetime. It was determined that Groups 1 and 2 bound the

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 4 of 9 11 groups evaluated in [2]. The deterministic fracture mechanics evaluation used the same methodology in [2] to demonstrate acceptance of RG 1.14 requirements regarding ductile failure, nonductile failure, fatigue crack growth and excessive deformation. The NRC staff agreed that RG 1.14 requirements were satisfied in the SER. Although the loss of configuration control of the PFM program RPFWPROF affects WCAP-15666-A, the PFM evaluation is completely separate from the deterministic stress and fracture mechanics. The RG 1.14 requirements were satisfied solely based on the deterministic evaluations. Therefore, there is no impact on the RCP flywheels from performing their intended safety function for 60 years of operation.

PWROG-17011-N P The purpose of PWROG-17011-NP [1] is to extend the evaluation and conclusion in [3]

for an extended operating period of 80 years. The only time dependent variable of the deterministic evaluation is the number of cycles of RCP start and shutdown. It was confirmed that the assumed 6000-cycle in [3] remains conservative and unchanged going from 60 to 80 years. Additionally, since the flywheel is not local or adjacent to the reactor core, RT NOT remains unchanged for 80 years. Therefore, the deterministic evaluation results and conclusion in WCAP-15666-A [3] remains applicable for [1]. The RG 1.14 requirements were satisfied, and the RPFWPROF program issue has no impact on the RCP flywheels from performing their intended safety function for 80 years of operation.

Review of Fleet Operating Experience In preparing [2], in 1995, a review was performed and identified 729 flywheel inspections have been performed, with only 43 resulting in recordable indications. None of these indications were determined to affect flywheel integrity.

A review of industry operating experience was also performed in 2017 and reported in

[1] as part of the work in extending the evaluation to 80 years. Since the inspection records reported in [2], there are 81 flywheel inspection records with only 4 resulting in recordable indications. All 4 of these indications were determined to be non-relevant to flywheel integrity, and are attributed to the disassembly and reassembly during inspection.

Conclusions from Deterministic Evaluation Review The deterministic work performed in [1 to 3] have been reviewed and confirmed. The deterministic work demonstrates that requirements of RG 1.14 have been satisfied with significant margins, ensuring the reactor can be safely shutdown, the core will be cooled per the LOCA licensing criteria, and the reactor will remain shut down. Additionally, significant fleet operating experience with flywheel inspections demonstrates, empirically, no indications to date that shown to have impacted RCP flywheel integrity.

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 5 of 9 Overview of Investigations into Probabilistic Fracture Mechanics Program, RPFWPROF As discussed above, the RAI questions the results (probabilistic results are -3% less than the WCAP-15666-A work) and requests FPL and Westinghouse to confirm that there are no errors in what was submitted. As a note: the 2003 report (WCAP-15666-A), listed Probability of Failures (POF) at 2.57E-7 for Flywheel Group 1 flywheels at 60 years (with ISi at 4 year intervals prior to 10 years and without ISi after 10 years) while PWROG-17011-NP shows POF at 2.49E-7 at 60 years; and 2.52E-7 at 80 years for this same flywheel grouping and same ISi assumptions.

Westinghouse had noted the small difference in predictions between the 40 and 60 year cases between the two reports. At the time these differences were attributed to a change in computer platforms and a full set of interval calculations were performed with the available version of the RPFWPROF so that delta interval calculations would be self-consistent. In seeking to address the NRC's current RAI confirming Westinghouse's initial judgement and regarding the Kie model, Westinghouse engineers determined that the RPFWPROF executable file used in PWROG-17011-NP cannot reproduce the results of WCAP-15666-A when run on original computer platforms.

This confirms that Westinghouse had lost configuration control on the PFM executable program, RPFWPROF.

The RPFWPROF program was not re-written, but was used in what was believed to be the same form as that used in the 60 year POF analysis in the determination of the POF for 80 years [3]. In support of the PWROG-17011-NP report, the engineers retrieved the RPFWPROF executable file from local engineering group's shared folder, ran the RPFWPROF executable program, and compared RCP probability of failure results with those in WCAP-15666-A. Because the results were within -3% of WCAP-15666-A result, and the 80 year POF was greater than the 60 year POF, demonstrating that the POF would increase with_ time as expected, the eng_ineers considered the -3%

difference in results to be reasonable for a probabilistic evaluation and attributed the POF difference to the employment of different (upgraded) computer platforms between the work performed in 2003 and 2017.

In seeking to address the NRC's current RAI, and in particular the question regarding the basis for selected Kie values in the probabilistic assessment, an error was uncovered in the available hard copy of the original independently reviewed source code for RPFWPROF. Therefore, the assessment.of reasonableness was reviewed more closely, and Westinghouse engineers attempted to reproduce the original results of the probabilistic fracture mechanics program of WCAP-15666-A, RPFWPROF when removing the computer platform differences.

As part of this reproduction, benchmarking cases that were used to test the original RPFWPROF program were re-performed now, using the RPFWPROF executable program used for the PWROG-17011-NP work, but running the executables on older computer platforms. The purpose of this reproduction was to confirm that the

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 6 of 9 RPFWPROF executable program used in support of the PWROG-17011-NP report matched that used in the earlier reports - (removing the computer platform difference as a variable). The benchmarking case results could not reproduce the results from [3];

therefore, this confirms that Westinghouse has lost configuration control on the probabilistic fracture mechanics executable program, RPFWPROF.

Re-establishing Configuration Control of the RPFWPROF Program The RPFWPROF executable files were re-constructed to follow the original methodology. Additional lines of code were added to allow for validation of intermediate functional steps of the program, but all other aspects of the program were purposely left unchanged to ensure conformity with the RPFWPROF program detailed in Section 3.3 of [3]. Thus, parameter values listed in Tables 3-1 through Tables 3-6 of [3] remain unchanged. Results of the program were validated using hand calculations to ensure that the program was appropriately considering the physical phenomena being simulated, and also validated by making comparisons to the original program's benchmarking case results, as appropriate, as these are allowed validation methods in Westinghouse's validation and verification program for programs. The result of this effort is a revised RPFWPROF program that has been confirmed to appropriately calculate the probability of RCP Flywheel failures.

By this verification and validation process, using the 1995 independently reviewed version of RPFWPROF, the POF results listed in Table 3-8 ofWCAP-15666-A [3] can be exactly reproduced. However, it has been determined that there is an error in the RPFWPROF program that supports the WCAP-15666-A report. The program compares a stress intensity factor, K1, for a set of input conditions, to the fracture toughness, or critical stress intensity factor, Kie, and where K1 >Kie, the flywheel is considered to have failed for that set of conditions. Kie is calculated in the RPFWPROF program using the mean curve developed through regression analysis and documented in [4], as shown below.

Kie= 5S.l + 28.0exp(0.0214(T-RTNDT)) for T-RTNOT > -50°F (H-5b)

During verification and validation efforts, it was determined that there was a transcription error in the program code and the exponent 0.0214 was mistakenly written as 0.2014. This results in Kie increasing rapidly to the upper limit of 200 ksiv'm as the difference of service temperature minus the reference temperature for nil ductility transiti9n *(T-RT NDT) increases. For positive differences in (T-RTNDT) this error results in a higher allowable when determining whether a failure occurred. The opposite is true for predicted conditions with negative differences in (T-RTNoT). Thus, while the WCAP-15666-A results can be exactly reproduced, the calculated POFs must be corrected to account for this error.

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 7 of 9 It should also be noted that the RPFWPROF executable program ran in 2017 to support issuance of PWROG-17011-NP, Rev. 1 [1] is being discarded. While results from this executable program were within 3% of the WCAP-15666-A POF results, Westinghouse's present verification and validation efforts have determined that changes were made to the RPFWPROF executable program utilized for the PWROG-17011-NP, Rev. 1 report following the work performed in support ofWCAP-15666-A and were not managed through configuration control, resulting in calculations that should be discarded.

Revised RCP Flywheel POFs Cumulative Probability of Failure over 40, 60, and 80 Years With and Without lnservice Inspections (Values from Table 3-8 of WCAP-15666-A are shown italicized in parentheses below each newly calculated value for ease of comparison)

Cumulative Probability of Flywheel Failure Cumulative Probability of Flywheel with ISi at 4-year Failure with ISi at 4-year intervals prior to Design Intervals 10 Years, and without ISi after 10 years Limiting Flywheel Speed Over40 Over 60 Over 80 Group (RPM) Over 80 Years Years Years Years 1.99E-08 2.00E-08 2.01E-08 1 1500 2.02E-08 (2.45E-07) (2.50E-07) (2.57E-07) 1.26E-08 1.26E-08 1.36E-08 2 1500 1.37E-08 (1.43E-07) (1.45E-07) (1.47E-07) 5.88E-02 5.88E-02 5.88E-02 5.88E-02 1 3321 (1.01E-02) (1.01E-02) (1.02E-02) 1.66E-02 1.66E-02 1.66E-02 1.66E-02 2 3321 (0.91E-02) * (0.91E-02) (0.91E-02)

As shown in WCAP-15666-A, Table 3-7, the 60-year analysis used a sample size of 9,999. Due to the lower POF results of the 80-year analysis, the 80-year analysis used a sample size of 99,999 to ensure statistical stability of the POF results.

Conservatisms in the Analysis The probabilistic fracture mechanics evaluation, RPFWPROF, conservatively assumes that the probability of a flaw existing after preservice inspection is 10% and that the ISi flaw detection probability is only 50%. In reality, most pre-service and ISi flaws would be detected, especially for the large flaw depths that may lead to failure.

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 8 of 9 Furthermore, the core damage evaluation performed for the analysis considers the double-ended guillotine break (DEGB) coincident with instantaneous loss of offsite power event (3321 RPM with uncertainties that allows variations about 3000 to 3600 RPM) leading to no electrical braking to the RCP, resulting in likelihood of core damage to be well within RG 1.174 [5] guidelines. The risk assessment does not take credit for the leak before break (LBB) strategy that would easily detect and eliminate breaks, which would limit the maximum LOCA overspeed to 1500 RPM per the deterministic evaluations in [2 and 3].

During an accident_like the DEGB, the containment temperature would be much higher than the currently assumed flywheel temperatures. As discussed in [2 and 3], the flywheel equipment specifications require both Charpy and drop-weight tests to ensure the flywheel RTNDT to be no greater than 10°F. Therefore, the RPFWPROF input for RT Nor of 30°F as the median with 17°F uncertainty is conservative.

Conclusion The probabilistic fracture mechanics work performed in [1] has been reviewed and corrections are being made to [1]. The computer program, RPFWPROF, has been verified and validated through Westinghouse's software configuration control process and the corrected POFs are provided above. No pther Westinghouse documents are affected by the transcription error in the RPFWPROF executable program.

References:

1. PWROG Report, PWROG-17011-NP, Rev. 1 "Update for Subsequent License Renewal: WCAP-14535A, Topical Report on Reactor Coolant Pump Flywheel Inspection Elimination" and WCAP-15666-A, "Extension of Reactor Coolant Pump Motor Flywheel Examination," May 2018.

2. Westinghouse Report, WCAP-14535A, Rev. 0, "Topical Report on Reactor Coolant Pump Flywheel Inspection Elimination," November 1996.

3. Westinghouse Report, WCAP-15666-A, Rev. 1, "Extension of Reactor Coolant Pump Motor Flywheel Examination," October 2003.

4. NRC Policy Issue SECY-82-465, "Pressurized Thermal Shock," 11/23/1982 (ADAMS Accession No. ML16232A574)

5. NRC Regulatory Guide 1.14, Revision 1, "Reactor Coolant Pump Flywheel Integrity,"

August 1975 (ADAMS Accession No. ML003739936)

6. FPL Letter L-2018-175 to NRC dated October 17, 2018, Turkey Point Units 3 and 4 Subsequent License Renewal Application, Safety Review Requests for Additional Information (RAI) Set 5 Responses (ADAMS Accession No. ML18292A642)

Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-251 FPL Revised Response to NRC RAI No. 4.3.5-2 L-2018-234 Attachment Page 9 of 9 Associated SLRA Revisions:

None Associated

Enclosures:

None