ML041910255
| ML041910255 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 03/31/2004 |
| From: | John Miller Southern California Edison Co |
| To: | Office of Nuclear Reactor Regulation |
| References | |
| RSC-04-02NP, Rev 0 | |
| Download: ML041910255 (124) | |
Text
Enclosure 5 San Onofre Nuclear Generating Station Probabilistic Safety Enclosures Assessment, Evaluation of Risk Significance of ILRT Extension
[Non-Proprietary Version]
Southern California Edison RSC 04-02NP San Onofre Nuclear Generating Station Probabilistic Risk Assessment Evaluation of Risk Significance of ILRT Extension Revision 0 March 2004 Principal Analyst Jeffrey Miller NON-PROPRIETARY DOCUMENT This document has been reviewed and proprietary information removed. It may be freely distributed as a complete document only.
RSC Ricky Suummitt Consulting, Inc.
8351 E. Walker Springs Lane, Suite 401 Risk and Reliability Engineering Knoxville, TN 37923 USA Telephone +865.692A012 Telefax +865.6924013
Evaluation of Risk Significance of JLRT Extension Table of Contents Section Page Main Report: Evaluation of Risk Significance of Integrated Leak Rate Test (ILRT) Extension
1.0 INTRODUCTION
.................................. I 1.1
SUMMARY
OF THE ANALYSIS ................................... 1 1.2
SUMMARY
OF RESULTS/CONCLUSIONS .................................. 1 2.0 DESIGN INPUTS ................................... 4 3.0 ASSUMPTIONS ................................... 7 4.0 CALCULATIONS .................................. 8 4.1 CALCULATIONAL STEPS .................................. 8 4.2 SUPPORTING CALCULATIONS ................................. 10
5.0 REFERENCES
................................ 23 Appendix A: Alternative Calculation for ILRT Evaluation A.0 INTRODUCTION ................................ I A.1 ANALYSIS APPROACH ................................. 2 A.2 DEFINITION OF ACCIDENT SEQUENCES ................................ 3 A.3 CALCULATION OF INCREASE IN TYPE-A RELATED LEAKAGE ........................... 3 A.4 ESTIMATION OF IMPACT ON LERF ................................................... 7 A.5 MODIFIED MODEL EVALUATION ................................................... 9 A.6 REFERENCES .................................................. 10 Appendix B: Surrogate Person-Rem Methodology (RSC 01-44)
1.0 INTRODUCTION
.7 2.0 METHODOLOGY .7 3.0 DEVELOPMENT OF RADIONUCLIDE RELEASE TO PERSON-REM RELATIONSHIP .. 7 RSC 04-02NP i Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension 3.1 DATA ASSESSMENT EXISTING ..................................... 7 3.2 DATA INTERPRETATION .................................... 10 3.4 APPLICATION WITH MAAP .................................... 12 3.5 QUALITATIVE UNCERTAINTY ASSESSMENT .................................... 13 3.5.1 Qualitative Evaluation of Predictive Dose . .......................................................... 13 3.5.2 Results Predictability ........................................................... 14 4.0
SUMMARY
AND CONCLUSIONS ............................................................ 14
5.0 REFERENCES
........................................................... . 14 Appendix C: Estimation of Intact Containment Population Dose C.0 ESTIMATION OF INTACT CONTAINMENT POPULATION DOSE ........................... 1 C.1 METHODOLOGY ............................................. 1 C.2 LICENSING BASIS INFORMATION ............................................................. I C.3 DOSE SCALING FACTOR ............................................................ 2 C.4 CALCULATION OF POPULATION DOSE ........................................................ . 3 C.5 ALTERNATIVE SOURCE TERM CALCULATION OF POPULATION DOSE ............ 4 C.6 REFERENCES ............................................................. 4 Appendix D: Sensitivity Study with SONGS Calculation N-6060-002 Data D.0 INTRODUCTION ............................................................. 1 D.1
SUMMARY
OF THE ANALYSIS ..................................... 1.......................I D.2
SUMMARY
OF RESULTS/CONCLUSIONS . .......................................................... 1 I D.3 DESIGN INPUTS ............................................................ 4 D.4 CALCULATIONS ............................................................ 6 D.5 SUPPORTING CALCULATIONS . ............................................................ 7 D.6 REFERENCES ............................................................ . 20 Appendix E: ILRT Sensitivity Study (Inclusion of SONGS External Events)
RSC 04-02NP ..
Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension E.0 INTRODUCTION ................................ I E.1
SUMMARY
OF THE ANALYSIS ................................. 1 E.2
SUMMARY
OF RESULTS/CONCLUSIONS ............................... 1I E.3 DESIGN INPUTS ................................ 4 E.4 CALCULATIONS ................................. 6 E.5 SUPPORTING CALCULATIONS ................................ 7 E.6 REFERENCES ................................ 20 Appendix F: Response to USNRC Request for Additional Information for Degredation of the Embedded Side of the Steel Drywell Structure F.0 INTRODUCTION .
F.1 ANALYSIS APPROACH .I F.2 ANALYSIS RESULTS .
F.3 REFERENCES .5 Appendix G: Evaluation of Relevant SONGS ILRT Experience G.0 INTRODUCTION .
G.1 ANALYSIS RESULTS .I G.2 REFERENCES .2 Appendix H: SONGS PRA Quality Discussion H.0 INTRODUCTION .
H.1 SONGS PRA QUALITY STATEMENT .1 H.2 REFERENCES .2 RSC 04-02NP iii Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension List of Tables Table Page Main Report: Evaluation of Risk Significance of Integrated Leak Rate Test (ILRT) Extension Table I Summary of Risk Impact on Extending Type A ILRT Test Frequency ................................ 2 Table 2 SONGS Plant Damage States ...................................................... 4 Table 3 Release Category Radionuclide Fraction ...................................................... 7 Table 4 Containment Failure Classifications (from Reference 7) .................................................... 10 Table 5 SONGS PRA Release Category Grouping to EPRI Classes (as described in Reference 7) .11 Table 6 Baseline Risk Profile ..................................................... 15 Table 7 Risk Profile for Once in Ten Year Testing ..................................................... 17 Table 8 Risk Profile for Once in Fifteen Year Testing ..................................................... 19 Table 9 Comparisons of Release Class Doses ..................................................... 21 Table 10 Impact on LERF due to Extended Type A Testing Intervals ............................................ 22 Table 11 Impact on Conditional Containment Failure Probability due to Extended Type A Testing Intervals ................................................. 23 Appendix A: Alternative Calculation for ILRT Evaluation Table A.1 Probability of Type A Leakage Given a Testing Interval .................................................. 4 Table A.2 Type A Leakage Frequency ...................................................... 5 Table A.3 Comparison of Release Class Doses ...................................................... 6 Table A.4 Type A LERF Contribution ...................................................... 8 Table A.5 Calculation of the Change in LERF ...................................................... 8 Table A.6 Person-Rem Comparisons ....................................................... 9 Table A.7 Integrated Person-Rem Estimates ..................................................... 10 Appendix B: Surrogate Person-Rem Methodology (RSC 01-44)
Table I Release Split Fraction to Dose Conversion Factors ..................................................... 12 RSC 04-02NP iv Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table 2 Mapping of Method Variables to MAAP Output Variables ................................................ 12 Appendix C: Estimation of Intact Containment Population Dose Table C.1 Predicted Dose Rates taken from Reference 1........................................................ 1 Table C.2 Calculation Parameters taken from Reference I ........................................................ 2 Table C.3 Calculation Parameters taken from Reference 1........................................................ 3 Table C.4 Calculation Parameters taken from Reference 5 ........................................................ 4 Appendix D: Sensitivity Study with SONGS Calculation N-6060-002 Data Table D.1 Surnmary of Risk Impact on Extending Type A ILRT Test Frequency ............................ 2 Table D.2 SONGS Plant Damage States ........................................................ 4 Table D.3 Release Category Radionuclide Fraction ........................................................ 6 Table D.4 Containment Failure Classifications (from Reference 7) .................................................. 7 Table D.5 SONGS PRA Release Category Grouping to EPRI Classes (as described in Reference 7)8 Table D.6 Baseline Risk Profile .................... 12 Table D.7 Risk Profile for Once in Ten Year Testing ....................................................... 14 Table D.8 Risk Profile for Once in Fifteen Year Testing ....................................................... 16 Table D.9 Comparisons of Release Class Doses ....................................................... 18 Table D.10 Impact on LERF due to Extended Type A Testing Intervals ......................................... 19 Table D. 1 Impact on Conditional Containment Failure Probability due to Extended Type A Testing Intervals ................................................... 20 Appendix E: ILRT Sensitivity Study (Inclusion of SONGS External Events)
Table E.1 Summary of Risk Impact on Extending Type A ILRT Test Frequency ............................. 2 Table E.2 SONGS Plant Damage States ........................................................ 4 Table E.3 Release Category Radionuclide Fraction ........................................................ 6 Table E.4 Containment Failure Classifications (from Reference 7) ................................................... 7 Table E.5 SONGS PRA Release Category Grouping to EPRI Classes (as described in Reference 7)8 Table E.6 Baseline Risk Profile ....................................................... 12 RSC 04-02NP v Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table E.7 Risk Profile for Once in Ten Year Testing ............................................................. 14 Table E.8 Risk Profile for Once in Fifteen Year Testing ............................................................. 16 Table E.9 Comparisons of Release Class Doses ............................................................ 18 Table E. 10 Impact on LERF due to Extended Type A Testing Intervals ......................................... 19 Table E. I Impact on Conditional Containment Failure Probability due to Extended Type A Testing Intervals ......................................................... 20 Appendix F: Response to USNRC Request for Additional Information for Degredation of the Embedded Side of the Steel Drywell Structure Table F.1 SONGS Liner Corrosion Risk Assessment Results Using CCNP Methodology ................ 1 Table F.2 Changes Due to Extension from 10 Years (current) to 15 Years ....................................... 4 Table F.3 Changes Due to Extension from 3 Years (baseline) to 15 Years ........................................ 4 Appendix G: Evaluation of Relevant SONGS ILRT Experience Table G.l SONGS ILRT Resultant Leak Rates ............................................................. 1 RSC 04-02NP vi Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension List of Figures Figure Page Appendix B: Surrogate Person-Rem Methodology (RSC 01-44)
Figure 1 Sequoyah Release Fraction Cases (Reference 2) ............................................... 8 Figure 2 Unpublished PWR Release Fraction Cases (Reference 3) ................................................ 9 Figure 3 Oconee IPE Release Fraction Cases (Reference 4) ............................................... 9 Figure 4 Seabrook Release Fraction Cases (Reference 5) ............................................... 10 Figure 5 Relationship between Release Fraction and Dose .............................................. 10 Figure 6 Comparison of Equation Results and Reported Dose Values ............................................ 11 Figure 7 Variation of Equation to Reported Dose ............................................... 13 RSC 04-02NP vii Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Main Report:
Evaluation of Risk Significance of Integrated Leak Rate Test (ILRT) Extension RSC 04-02NP Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
1.0 INTRODUCTION
The purpose of this calculation is to evaluate the risk of extending the Type A integrated leak rate test interval beyond the current 10 years required by 10 CFR 50, Appendix J, Option B at the San Onofre Nuclear Generating Station for both unit 2 and unit 3.
1.1
SUMMARY
OF THE ANALYSIS 10 CFR 50, Appendix J allows individual plants to extend Type A surveillance testing requirements and to provide for performance-based leak testing. This report documents a risk-based evaluation of the proposed change of the integrated leak rate test (ILRT) test interval for the San Onofre Nuclear Generating Station (SONGS). The proposed change would impact testing associated with the current surveillance tests for Type A leakage, procedure S02-V-3.121 for unit 2 and procedure S03-V-3.12 2 for unit 3. No change to Type B or Type C testing is proposed at this time.
The evaluation for SONGS is consistent with similar assessments performed for the Indian Point 3 (UP3) plant, which was approved by the NRC 3' 4 and for the Crystal River 3 (CR3) plant 5 . This assessment utilizes the guidelines set forth in NEI 94-0 16, the methodology used in EPRI TR-104285 7 and the regulatory guidance on the use of Probabilistic Risk Assessment (PRA) findings in support of a licensee request to a plant's licensing basis, RG 1.1748.
This calculation evaluates the risk associated with various ILRT intervals as follows:
- 3 years - Interval based on the original requirements of 3 tests per 10 years.
- 10 years - This is the current test interval required for SONGS.
- 15 years - Proposed extended test interval, similar to IP3 request.
The analysis utilizes the latest SONGS probabilistic risk assessment (PRA) results 9 . The PRA was initially developed for the SONGS individual plant examination (IPE) to estimate the baseline core damage and plant damage states. Several updates to the SONGS level 1 analysis have been incorporated since the IPE, these updates also included an update to the Level 2 information. Therefore, this information represents the most recent analysis documented for SONGS.
The release category and person-rem information is based on design basis leakage evaluations and extrapolation of the release category information using a modeling framework that develops the person-rem estimates based on the relative release fractions of radionuclides. The framework is described in Appendix B. Intact containment release information is developed using the approach presented in Appendix C.
1.2
SUMMARY
OF RESULTS/CONCLUSIONS The specific results are summarized in Table I below. The Type A contribution to LERF is defined as the contribution from Class 3b.
RSC 04-02NP I Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table 1 Summary of Risk Impact on Extending Type A ILRT Test Frequency Risk lmpact for 3-yeats "Risk'Impact for 10- -Risk Impact for 15-,'
(baseline) , years (current - - years rquirement)
Total integrated risk (person-rern/yr) 80.0561 80.0566 80.0568 Type A testing risk (person-rern/yr) 0.0054 0.0060 0.0063
% total risk (Type A /total) 0.0068% 0.0075% 0.0078%
Type A LERF (Class 3b) (per year) 3.74E-07 4.1 1E-07 4.30E-07 Changes due to extension from 1O years (current)
A risk from current (person-rernfyr) 2.55E-04
% increase from current (A risk /total risk) 0.0003%
LERF from current (per year) 1.87E-08 ACCFP from current - 0.104%
Changes due to extension from 3,years (baseline)
Arisk from baseline (person-rem/yr) 7.66E-04
% increase from baseline (A risk /total risk) 0.001ID0h A LERF from baseline (Per year) ' - 5.60E-08 ACCFP from baseline . 0.312%
RSC 04-02NP 2 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Based on the analysis and available data the following is stated:
- The person-rem/year increase in risk contribution from extending the ILRT test frequency from the current once-per-ten-year interval to once-per-fifteen years is 0.0003 person-rem/year.
- The risk increase in LERF from extending the ILRT test frequency from the current once-per-10-year interval to once-per-15 years is 1.87E-08/yr.
- The change in conditional containment failure probability (CCFP) from the current once-per-10-year interval to once-per-15 years is 0.104%.
- The change in Type A test frequency from once-per-ten-years to once-per-fifteen-years increases the risk impact on the total integrated plant risk by only 0.0003%. Also, the change in Type A test frequency from the original three-per-ten-years to once-per-fifteen-years increases the risk only 0.0010%. Therefore, the risk impact when compared to other severe accident risks is negligible.
- Reg. Guide 1.174 provides guidance for determining the risk impact of plant-specific changes to the licensing basis. Reg. Guide 1.174 defines very small changes in risk as resulting in increases of CDF below 10-6/yr and increases in LERF below 10-7/yr. Since the ILRT does not impact CDF, the relevant criterion is LERF. The increase in LERF resulting from a change in the Type A ILRT test interval from aonce-per-ten-years to a once per-fifteen-years is 1.87E-08. Since guidance in Reg. Guide 1.174 defines very small changes in LERF as below 10-7/yr, increasing the ILRT interval from 10 to 15 years is therefore considered non-risk significant. In addition, the change in LERF resulting from a change in the Type A ILRT test interval from a three-per-ten-years to a once per-fifteen-years is 5.60E-08/yr, is also below the guidance.
- R.G. 1.174 also encourages the use of risk analysis techniques to help ensure and show that the proposed change is consistent with the defense-in-depth philosophy. Consistency with defense-in-depth philosophy is maintained by demonstrating that the balance is preserved among prevention of core damage, prevention of containment failure, and consequence mitigation. The change in conditional containment failure probability was estimated to be 0.104% for the proposed change and 0.312% for the cumulative change of going from a test interval of 3 in 10 years to 1 in 15 years. These changes are small and demonstrate that the defense-in-depth philosophy is maintained.
RSC 04-02NP 3 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension 2.0 DESIGN INPUTS The SONGS PRA is intended to provide "best estimate" results that can be used as input when making risk informed decisions. The PRA provides the most recent results for the SONGS PRA.
Appendix H provides a detailed discussion concerning the SONGS PRA qumlity.
The inputs for this calculation come from the information documented in the SONGS PRA and the level 2 update (References 10 and 16). The SONGS plant damage states are summarized in Table 2.
Table 2 SONGS Plant Damage States Plant Damage Representative Sequence :requency
'Stt~(y)
PDS I Transient with loss of secondary heat removal 6.81E-06 Transient with loss of secondary heat removal, and loss of containment spray PDS 2 recirculation 1.71E-07 Transient with loss of secondary heat removal, and loss of containment heat PDS 3 removal 2.66E-07 PDS 4 Transient/SSL with loss of HPSI in recirculation 1.5 IE-09 Transient/SSL with loss of HPSI in recirculation, and loss of containment heat PDS 5 removal 0.OOE+00 Transient/SSL with loss of HPSI in recirculation, loss of secondary heat PSD 6 removal, and loss of containment spray recirculation 2.18E-07 Transient with loss of HPSIILPSI injection and loss of containment heat PDS 7 removal 1.20E-06 Transient with loss of HPSI/LPSI injection, loss of secondary heat removal PDS 8 and loss of containment heat removal 2.17E-07 PDS 9 Small LOCA with loss of containment spray recirculation 3.65E-10 PDS 10 Small LOCA with loss of containment beat removal 7.38E-07 PDS II Small LOCA with loss of secondary heat removal 2.19E-07 PDS 12 Small LOCA with loss of HPSI recirculation 2.43E-07 RSC 04-02NP 4 Printed: 05!12/2004
Evaluation of Risk Significance of ILRT Extension Table 2 (Continued)
SONGS Plant Damage States Plant Damage Representative Sequence Frequency
.State - (Iyr)
Small LOCA with loss of HPSI recirculation, and loss of containment spray PDS 13 recirculation 5.32E-06 Small LOCA with loss of HPSI recirculation, and loss of containment heat PDS 14 removal 1.56E-08 Small LOCA with loss of HPSI recirculation and loss of secondary heat PDS 15 removal 4.40E-09 Small LOCA with loss of HPSI recirculation, loss of secondary heat removal, PDS 16 and loss of containment spray recirculation 6.24E-08 Small LOCA with loss ofHPSI recirculation, loss of secondary heat removal, PDS 17 and loss of containment heat removal 5.99E-07 PDS 18 Small LOCA with loss of HPSIILPSI injection 1.92E-07 PDS 19 Large/medium LOCA with loss of core heat removal 1.28E-06 Large/medium LOCA with loss of core heat removal, and loss of containment PDS 20 spray recirculation 1.08E-09 PDS 21 Large/medium LOCA with loss of core and containment heat removal 4.06E-09 PDS 22 Large/medium LOCA with loss of HPSI recirculation 2.OOE-07 Large/medium LOCA with loss of HPSI recirculation, and loss of PDS 23 containment spray recirculation 1.59E-09 Large/medium LOCA with loss of HPSI recirculation, and loss of PDS 24 containment heat removal 8.40E-09 PDS 25 Large/medium LOCA with loss of HPSI/LPSI injection 1.93E-09 PDS 26 Transient/LOCA with loss of containment isolation and heat removal 1.14E-08 PDS 27 Interfacing system LOCA (ISLOCA) initiating event 5.61E-08 Steam generator tube rupture (SGTR) initiating event, no stuck open relief PDS 28 valve (SORV) 1.48E-08 Steam generator tube rupture (SGTR) initiating event, with stuck open relief PDS 29 valve (SORV) 1.25E-08 RSC 04-02NP 5 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table 2 (Continued)
SONGS Plant Damage States plant Damage' - Representative Sequence: Frequency -
State (/) -
TOTAL I .79E-05 In order to develop the person-rem dose associated with each plant damage state it is necessary to associate each plant damage state with an associated release of radionuclides and from this information to calculate the associated dose.
The IP3 submittal (Reference 3) utilizes a multiplication factor to adjust the design basis leakage value (La) that is based on generic information that relates dose to leak size. The CR3 submittal (Reference 5) utilized plant-specific dose estimates based on the predicted level 2 analysis results.
The SONGS PRA (References 10 and 16) contains the necessary information to convert the plant damage states to release categories. Using this information the plant damage states are mapped to the six release categories: B, D, G, L, T, and W. In addition, the fraction of intact containment cases is determined using the split fraction information contained in References 9 and 16.
Since the SONGS PRA contains the necessary release fraction information, an approach similar to the CR3 submittal is utilized that better reflects the specific release conditions for SONGS.
The SONGS PRA (Reference 9) release categories are defined by the release fraction of major radionuclides. These are extrapolated to dose using the approach presented in Appendix B. This approach has been presented in other licensing submittals (References 14 and 15) and is consistent with the method used in the CR3 submittal (Reference 5). The intact containment dose is developed in Appendix C and is consistent with the approach used in Reference 15. The release category dose information is presented in Table 3.
RSC 04-02NP 6 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table 3 Release Category Radionuclide Fraction Release, Frequency Category (yr) Noble Gas' Iodine' Cesium' Tellurium' Strntium' Total Dose IC-1 (S) 1.50E-05 NA2 NA NA NA NA 2.22E+02 3 B4 5.06E-07 6.50E-02 1.OOE-03 1.OOE-03 6.OOE-09 3.OOE-07 1.98E+05 D5 1.25E-08 9.74E-01 5.50E-02 3.30E-02 7.OOE-03 2.OOE-06 5.90E+06 G6 1.56E-07 5.48E-01 2.00-02 2.OOE-02 2.1OE-02 6.00E-04 2.93E+06 1! 1.24F,06 1.OOE+00 4.OOE-03 7.OOF-03 2.00E-04 2.OOE-06 2.05E+06 5.61E-08 9.98E-01 8.42B01 8.42E-01 7.70E-02 2.OOE-03 8.61E+07 9.83E-07 l .OOE+O0 1.20E-01 1.21E-01 4.OOE-05 3.OOE-07 1.36E+07
- 1. Contributing fission product groups are discussed in Appendix B.
- 2. Release fractions not necessary for this calculation.
- 3. Intact containment representing design basis leakage (developed in Appendix C).
- 4. Release category B is defined by containment bypassed with less than 0.1% of volatiles released.
- 5. Release category D is defined by containment bypassed with up to 10% of volatiles released.
- 6. Release category G is defined by early or isolation failure, containment failure prior to or at vessel failure with up to 10% of volatiles released.
- 7. Release category L is defined by late containment failure with up to 1% of volatiles released.
- 8. Release category T is defined by containment bypassed with greater than 10% of volatiles released.
- 9. Release category W is defined by late containment failure with more than 10% of volatiles released.
Other inputs to this calculation include ILRT test data from NUREG-149311 and the EPRI report (Reference 7) and are referenced in the body of the calculation.
3.0 ASSUMPTIONS
- 1. The maximum containment leakage for EPRI Class 1 (Reference 7) sequences is 1Xig (Type A acceptable leakage) because a new Class 3 has been added to account for increased leakage due to Type A inspections.
- 2. The maximum containment leakage for Class 3a (References 3 and 5) sequences is OxIL based on the previously approved methodology (References 3 and 4).
- 3. The maximum containment leakage for Class 3b (References 3 and 5) sequences is 35xLa based on the previously approved methodology (References 3 and 4).
- 4. Class 3b is conservatively categorized LERF based on the previously approved methodology (References 3 and 4).
- 5. Containment leakage due to EPRI Classes 4 and 5 are considered negligible based on the RSC 04-02NP 7 Printed: 05/lZ/2004
Evaluation of Risk Significance of ILRT Extension previously approved methodology (References 3 and 4).
- 6. The containment releases are not impacted with time.
- 7. The containment releases for EPRI Classes 2, 6, 7 and 8 are not impacted by the ILRT Type A Test frequency. These classes already include containment failure with release consequences equal or greater than those impacted by Type A.
- 8. Because EPRI Class 8 sequences are containment bypass sequences, potential releases are directly to the environment. Therefore, the containment structure will not impact the release magnitude.
4.0 CALCULATIONS This calculation applies the SONGS PRA release category information in terms of frequency and person-rem estimates to estimate the changes in risk due to increasing the ILRT test interval.
The changes in risk are assessed consistent with the previously approved methodology used by Indian Point 33.4 and Crystal River 35. This approach is similar to that presented in EPRI TR-1042857 and NUREG-1493 11 . Namely, the analysis performed examined SONGS PRA plant specific results in which the containment integrity remains intact or the containment is impaired.
The detailed calculations performed to support this report were of a level of mathematical significance necessary to calculate the results recorded. However, the tables and illustrational calculation steps presented may present rounded values to support readability.
4.1 CALCULATIONAL STEPS The analysis is based on guidance provided in Reference 7 and uses risk metrics presented in Reference 8 to evaluate the impact of a proposed change on plant risk. References 3 and 5 utilize several measures in their assessments. These measures are: change in release frequency, change in risk as defined by the change in person-rem, the change in LERF and the change in the conditional containment failure probability.
Reference 8 also lists the change in core damage frequency as a measure to be considered. Since the testing addresses the ability of the containment to maintain its function, the proposed change has no measurable impact on core damage frequency. Therefore, this attribute remains constant and has no risk significance.
The overall process is outlined below:
- Define baseline plant damage states and person-rem estimates
- Calculate baseline Type A leakage estimate to define the analysis baseline
- Modify Type A leakage estimate to address extension of the Type A test frequency
- Compare analysis metrics to estimate the impact and significance of the increase related to those metrics RSC 04-02NP 8 Printed: 05/12t2004
Evaluation of Risk Significance of ILRT Extension The first step in the analysis is to define the baseline plant damage states and person-rem dose measures. Plant damage state information is developed using the SONGS PRA (References 9, 10, and 16) results. The plant damage state information and the results of the containment analysis are used to define the sequences. The population person-rem dose estimates for each key plant damage states are based on the application of the method described in Appendix B and design basis information 13 The product of the person-rem for the key plant damage states by the frequency of the key plant damage state estimates the annual person-rem estimate for the plant damage state. Summing these estimates produces the annual person-rem dose based on the sequences defined in the PRA.
The PRA plant damage state definitions consider isolation failures due to Type B and Type C faults and examine containment challenges occurring after core damage and/or reactor vessel failure. These sequences are grouped into key plant damage states. Using the plant damage state information, bypass, isolation failures and phenomena-related containment failures are identified.
Once identified, the plant damage state was then classified by release category definitions specified in Reference 7. With this information developed, the PRA baseline model is completed.
The second step expands the baseline model to address Type Aleakage. The PRA did not directly address Type A (liner-related) faults and this contribution must be added to provide a complete baseline. In order to define leakage that can be linked directly to the Type A testing, it is important that only failures that would be identified by Type A testing exclusively be included.
Reference 7 provides the estimate for the probability of a leakage contribution that could only be identified by Type A testing based on industry experience. This probability is then us ed to adjust the intact containment category of the SONGS PRA to develop a baseline model including Type A faults.
The release, in terms of person-rem, is developed based on information contained in Reference 7 and is estimated as a leakage increase relative to allowable release I. defined as part of the ILRT.
The predicted probability of Type A leakage is then modified to address the expanded time between testing. This is accomplished by a ratio of the existing testing interval and the proposed test interval. This assumes a constant failure rate and that the failures are randomly dispersed during the interval between the test.
The change due to the expanded interval is calculated and reported in terms of the change in release due to the expanded testing interval, the change in the population person-rem and the change in large early release frequency. The change in the conditional containment failure probability is also developed. From these comparisons, a conclusion is drawn as to the risk significance of the proposed change.
Using this process, the following were performed:
- 1. Map the Level 3 release categories into the 8 release classes defined by the EPRI Report (Reference 7).
RSC 04-02NP 9 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- 2. Calculate the Type A leakage estimate to define the analysis baseline.
- 3. Calculate the Type A leakage estimate to address the current inspection frequency.
- 4. Modify the Type A leakage estimates to address extension of the Type A test intervaL
- 5. Calculate increase in risk due to extending Type A inspection intervals.
- 6. Estimate the change in LERF due to the Type A testing.
- 7. Estimate the change in conditional containment failure probability due to the Type A testing.
4.2 SUPPORTING CALCULATIONS Step 1: Map the Level 3 release categories into the 8 release classes defined by the EPRI Report EPRI Report TR-104285 defines eight (8) release classes as presented in Table 4.
Table 4 Containment Failure Classifications (from Reference 7)
Failure Classification -;. - Description;; I.-:nterpretation or Assigning SONGSRelease
_- .a-'tego I Containment remains intact with Intact containment bins containment initially isolated 2 Dependent failure modes or common Isolation faults that are related to a loss of cause failures power or other isolation failure mode that is not a direct failure of an isolation component 3 Independent containment isolation Isolation failures identified by Type A testing failures due to Type A related failures 4 Independent containment isolation Isolation failures identified by Type B testing failures due to Type B related failures 5 Independent containment isolation Isolation failures identified by Type C testing failures due to Type C related failures 6 Other penetration failures Other faults not previously identified 7 Induced by severe accident phenomena Early containment failure sequences as a result of hydrogen bumn or other early phenomena 8 Bypass Bypass sequence or SGTR Table 5 presents the SONGS release category mapping for these eight accident classes. Person-rem per year is the product of the frequency and the person-rem.
RSC 04-02NP 10 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table 5 SONGS PRA Release Category Grouping to EPRI Classes (as described in Reference 7)
erson7-Class - Description Release Category -F}requency PersonRemi - Rem/yr 1 No containment failure IC-1 (S) 1.50E-05 2.22E+02 3.33E-03 2 Large containment None E' isolation failures 3a Small isolation failures None Not O.OOE+OO (liner breach) addressed 3b Large isolation failures None Not i O.OOE+OO (liner breach) addressed Small isolation failures - N ;
failure to seal (type B)
Small isolation failures- N -
failure to seal (type C)
Containment isolation 6 failures (dependent failure, G 1.56E-07 2.93E+06 4.57E-01 personnel errors)
Severe accident 7 phenomena induced failure L, W 2.22E-06 1.36E+07 3.01E+01 (early and late) 8 Containment bypass B, D, T 5.75E-07 8.61E+07 4.95E+0I Total 1.795E-05 8.00510E+01
- 1. e represents a probabilistically insignificant value.
Step 2: Calculate the Type A leakage estimate to define the analysis baseline (3 year test interval)
As displayed in Table 5 the SONGS PRA did not identify any release categories specifically associated with EPRI Classes 3, 4, or 5. Therefore each of these classes must be evaluated for applicability to this study.
RSC 04-02NP 11l Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Class 3:
Containment failures in this class are due to leaks such as liner breaches that could only be detected by performing a Type A ILRT.
Reference 4 states that a review of experience data finds that Type A testing identified only 4 leakage events of the 144 events identified. Thus about 3% (0.028) of containment leakage events are identified by the ILRT. The remaining events were identified by a local leak rate test (LLRT) (Type B and C testing) and are not included in the analysis. This probability, however, is based on three tests over a 10-year period and not the ore per ten-year frequency currently employed at SONGS (References 1 and 2). The probability (0.028) must be adjusted to reflect this difference.
For this estimation, the question on containment isolation was modified consistent with the previously approved methodology (References 3 and 4), to include the probability of a liner breach (due to excessive leakage) at the time of core damage.
Class 3 is divided into two classes using this approach. Class 3a is defined as a small liner breach and Class 3b is defined as a large liner breach.
Calculation of Class 3b probabilit' To calculate the probability that a liner leak will be large (Class 3b), use was made of the data presented in NUREG-1493 (Reference 11). One data set found in NUREG-1493 reviewed 144 ILRTs. The largest reported leak rate from those 144 tests was 21 times the allowable leakage rate (L). Since 21 xL1 does not constitute a large release, no large releases have occurred based on the 144 ILRTs reported in NUREG- 1493.
To estimate the failure probability given that no failures have occurred, a conservative estimate is obtained from the 95th percentile of the x2 distribution. This is consistent with the Indian Point 3 (Reference 3) and Crystal River 3 (Reference 5) templates. In statistical theory, the x2 distribution can be used for statistical testing such as goodness-of-fit tests (Reference 12). The x2 distribution is really a family of distributions, which range in shape from that of the exponential to that of the normal distribution.
Each distribution is identified by the degrees of freedom, v. For time-truncated tests (versus failure-truncated tests), an estimate of the probability of a large leak using the x2 distribution can be calculated using the following equation:
p(a) = 2 (2F +2,a) (Eq. 1) 2N Where: N is the number of events, F is the number of events (faults) of interest, and a is the percentile distribution (typically assumed to be the 95 0/o-tile). The result of 2F+2 defines the degree of freedom.
RSC 04-02NP 12 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Given that there have been no large leaks F = 0, therefore v =2) in 144 events (N = 144) the value of X2(2, 0.05) is equal to 5.99. Solving for the 95th percentile estimate of the probability of a large leak yields 0.021 as presented below:
7.2(2,005) 5.99 PCI.3B = - = 0.021 (Eq. 2) 2.144 288 Calculation of Class 3a probability The data presented in NUREG-1493 (Reference 11) is also used to calculate the probability that a liner leak will be small (Class 3a). The data found in NUREG-1493 states that 144 ILRTs were conducted. The data reported that 23 of 144 tests had allowable leak rates in excess of l.0xL. However, of the 23 events that exceeded the test requirements, only 4 were found by an ILRT, the others were found by Type B and C testing or errors in test alignments.
Therefore, a best estimate for the probability of leakage is -0.03 (4-of-144). However, the Class 3a probability is estimated using the conservative x2 distribution approach described previously.
This is consistent with the approach taken in References 3,4 and 5.
The x2 distribution is calculated by F=4 (number of small leaks) and N=144 (number of events) which yields a solution as shown below:
2X(10,0.05) 18.307 =(Eq3)
PCI.BSA 2=14 28 =0.064(E.3 Therefore, the 95th percentile estimate of the probability of a small leak (Class 3a) is calculated as 0.064.
The probability of liner failures must then be multiplied by an appropriate accident frequency to determine the Class 3a and Class 3b frequencies. The EP3 (Reference 3) and CR3 (Reference 5) submittals utilized the entire core damage frequency when developing the contributions for Classes 3a and 3b and then adjusted the Class 1 contribution.
This is somewhat conservative since it does provide the maximum possible contributions due to the extension of the ILRT testing interval. This approach is maintained for the SONGS analysis, in order to be consistent with the approved methodology.
Therefore the frequency of a Class 3b failure is calculated as:
FREQd..b = PROBE. x CDF = 0.021 x 1.79E-05/yr = 3.74E-07/yr (Eq. 4)
Therefore the frequency of a Class 3a failure is calculated as:
FREQ. = PROR...B x CDF = 0.064 x 1.79E-05 = 1.14E-06/yr (Eq. 5)
RSC 04-02NP 13 Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension Class 4:
This group consists of all core damage accident accidents for which a failure-to-seal containment isolation failure of Type B test components occurs. By definition, these failures are dependent on Type B testing, and Type A testing will not impact the probability. Therefore this group is not evaluated any further, consistent with the approved methodology.
Class 5:
This group consists of all core damage accident accidents for which a failure-to-seal containment isolation failure of Type C test components occurs. By definition, these failures are dependent on Type C testing, and Type A testing will not impact the probability. Therefore this group is not evaluated any further, consistent with the approved methodology.
Class 6:
The Class 6 group is comprised of isolation faults that occur as a result of the accident sequence progression. The leakage rate is not considered large by the PRA definition and therefore it is placed into Class 6 to represent a small isolation failure and identified in Table 5 as Class 6.
FREQCSS6 = 1.56E-07/yr Class 1:
Although the frequency of this class is not directly impacted by Type A testing, the PRA did not model Class 3 failures, and the frequency for Class 1 should be reduced by the estimated frequencies in the new Class 3a and Class 3b in order to preserve the total CDF. The revised Class 1 frequency is therefore:
FREQc1assi = FREQcisi - (FREQcIss3a + FREQcass3b) (Eq. 6)
FREQcIassl = 1.50E-05/yr - (1 .14E-06/yr + 3.74E-07/yr) = 1.35E-05/yr Class 2:
The SONGS PRA did not identify any contribution to this group above the quantification truncation.
Class 7:
The frequency of Class 7 is the sum of those release categories identified in Table 5 as Class 7.
FREQciass 7 = 2.22E-06/yr Class 8:
The frequency of Class 8 is the sum of those release categories identified in Table 5 as Class 8.
FREQcIasss = 5.75E-07/yr RSC 04-02NP 14 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension Table 6 summarizes the above information by the EPRI defined classes. This table also presents dose exposures calculated using the methodology described in Appendix B For Class 1, 3a and 3b, the person-rem is developed based on the design basis assessment of the intact containment as developed in Appendix C The Class 3a and 3b doses are represented as lOxL, and 35xL, respectively. Table 6 also presents the person-rem frequency data determined by multiplying the failure class frequency by the corresponding exposure.
Table 6 Baseline Risk Profile Class - ;Descrption ',Frequency Person-rem 'Person-rem (r)(calculated)1 (fr~om 14(/r)
.~~ -;. uA I No containment failure 1.35E-05 2.22E+02 2 3.OOE-03 Large containment -
isolation failures 4
Small isolation failures 1.4 22E0 25E0 aOiner breach) 2.22 2.54E-03 3b Large isolation failures (liner breach) !7.78E+03 2.90E-03 Small isolation failures -
failure to seal (type B)
Small isolation failures -
failure to seal (type C)
Containment isolation '
6 failures (dependent failure, 1.56E-07 2.93E+06 4.57E-01 personnel errors)
Severe accident 7 phenomena induced failure 2.22E-06 1.36E+07 6 3.01E+01 (early and late) 8 Containment bypass 5.75E-07 8.61E+07 6 ; 4.95E+01 Total 1.795E-05 - . 8.00561E+01
- 1. From Table 3 using the method presented in Appendix B.
- 2. I xt, dose value calculated in Appendix C.
- 3. E represents a probabilistically insignificant value.
- 4. 10 times I..
- 5. 35 times L-.
- 6. Maximum dose from contributing release categories.
RSC 04-02NP 15 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension The percent risk contribution due to Type A testing is as follows:
%RiskE =[(Class3a~u. + Class3bsE) / TotaLASE] x 100 (Eq. 7)
Where:
Class3a.A1 E= Class 3a person-remlyear = 2.54E-03 person-rem/year Class3bB4 sE = Class 3b person-rem/year = 2.90E-03 person-rem/year Total3 ASE = total person-rem year for baseline interval = 8.00561 E+01 person-rem/year (Table 6)
%Risk,. = [(2.54E-03 + 2.90E-03) / 8.00561E+01] x 100 = 0.0068%
Step 3: Calculate the Type A leakage estimate to address the current inspection interval The current surveillance testing requirements as proposed in NEI 94-01 (Reference 6) for Type A testing and allowed by 10 CFR 50, Appendix J, Option B is at least once per 10 years based on an acceptable performance history (defined as two consecutive periodic Type A tests at least 24 months apart in which the calculated performance leakage was less than l.OxL).
According to NUREG-1493 (Reference 11), extending the Type A ILRT interval from 3-in-10 years to 1-in-10 years will increase the average time that a leak detectable only if an ILRT goes undetected from 18 to 60 months. Multiplying the testing interval by 0.5 and multiplying by 12 to convert from "years" to "months" calculates the average time for an undetected condition to exist.
Since ILRTs only detect about 3% of leaks (4/144) that are not detected by other local tests, the increase for a 10-yr ILRT interval is the ratio of the average time for a failure to detect for the increased ILRT test interval (60 months) to the baseline average time for a failure to detect of 18 months (i.e., 0.03 x 60/18 = 0.10). References 3 and 5 indicate this is a 10% increase in the likelihood of a Type A leak.
Risk impact due to 10-year test interval Based on the previously approved methodology (References 3 and 4), the increased probability of not detecting excessive leakage due to Type A tests directly impacts the frequency of the Class 3 sequences. Consistent with Peference 3 and 5 the risk contribution is determined by multiplying the Class 3 accident frequency by the increase in the probability of leakage (1.1 x Class 3 baseline). The results of this calculation are presented in Table 7 below.
RSC 04-02NP 16 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension Table 7 Risk Profile for Once in Ten Year Testing Class Description Frequency (Iyr) ;Person-rem Person-rem (Iyr)
I No containment failure 2 1.33E-05 2.22E+02 2.96E-03 2 Large containment isolation failures 3a Small isolation failures (liner 1.26E-06 2.22E+03 2.79E-03 breach) 3b Large isolation failures (liner 4.1 IE-07 7.78E+03 3.20E-03 breach)
Small isolation failures - failure -
to seal (type B)
Small isolation failures - failure 5
to seal (type C)
Containment isolation failures 1.56E-07 2.93E+06 4.57E-01 6 (dependent failure, personnel errors) 7 Severe accident phenomena 2.22E-06 1.36E+07 3.01E+01 induced failure (early and late) 8 Containment bypass 5.75E-07 8.61E+07 4.95E+01 Total 1.795E-05 8.00566E+01
]1.n,~
FoTable6
- 2. The IPE frequency of Class I has been reduced by the frequency of Class 3a and Class 3b in order to preserve total CDF.
- 3. e represents a probabilistically insignificant value.
Using the same methods as for the baseline, and the data in Table 7 the percent risk contribution due to Type A testing is as follows:
%Riskzo =[(Class3a,. + Class3b,.) I Totalo] x 100 (Eq. 8)
Where:
Class3a, 0 = Class 3a person-rem/year = 2.79E-03 person-rem/year Class3b 1 . = Class 3b person-rem/year = 3.20E-03 person-rem/year Total 0 = total person-rem year for current 10-year interval = 8.00566E+01 person-rem/year (Table 7)
RSC 04-02NP 17 Printed: 0511212004
Evaluation of Risk Significance of ILRT Extension
%Risk, 0 = [(2.79F,03 + 3.20E-03) / 8.00566E+01] x 100 = 0.0075%
The percent risk increase (ARisk,.) due to a ten-year ILRT over the baseline case is as follows:
ARisk,. = [(Total 0 - TotaLn) / Totalms] x 100.0 (Eq. 9)
Where:
TotalsE = total person-rem/year for baseline interval = 8.00561 E+01 person-rem/year (Table 6)
Total. = total person-rem/year for 10-year interval = 8.00566E+01 person-rem/year (Table 7)
ARisk,. = [(8.00566E+01 - 8.00561 E+01) / 8.00561 E+01] x 100.0 = 0.0006%
Step 4: Calculatethe Type A leakage estimate to address extended inspection intervals If the test interval is extended to 1 in 15 years, the average time that a leak detectable only by an ILRT test goes undetected increases to 90 months (0.5 x 15 x 12). For a 15-yr-test interval, the result is the ratio (0.03 x 90/18) of the exposure times. Thus, increasing the ILRT test interval from 10 years to 15 years results in a proportional increase in the overall probability of leakage.
The approach for developing the risk contribution for a 15-year interval is the same as that for the 10-year interval. References 3 and 5 indicate that the increase is a 50% increase from that for the 10-year interval or a 15% increase from the baseline. Different values are provided for the probability of leakage. In addition, the containment leakage used for the 10-year test interval for Class 3 is used in the 15-year interval evaluation (L.15 x Class 3 baseline). The results for this calculation are presented in Table 8.
RSC 04-02NP 18 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table 8 Risk Profile for Once in Fifteen Year Testing Class Description - Frequency (lyr) e n Person-rem (Iyr)
I No containment failure 2 1.33E-05 2.22E+02 2.95E-03 2 Large containment isolation failures 3a Small isolation failures (liner 1.31E-06 2.22E+03 2.92E-03 breach) 3b Large isolation failures (liner 4.30E-07 7.78E+03 3.34E-03 breach)
Small isolation failures - failure - -
to seal (type B)
Small isolation failures - failure to seal (type C)
Containment isolation failures 1.56E-07 2.93E+06 4.57E-01 6 (dependent failure, personnel errors) 7 Severe accident phenomena 2.22E-06 1.36E+07 3.01E+01 induced failure (early and late) 8 Containment bypass 5.75E-07 8.61E+07 4.95E+01 Total 1.795E-05 8.00568E+01
- 1. From Table 6.
- 2. The IPE frequency of Class I has been reduced by the frequency of Class 3a and Class 3b in order to preserve total CDF, however, due to the number of reported decimal places, the number appears the same as in Table 7.
- 3. E represents a probabilistically insignificant value.
Using the same methods as for the baseline, and the data in Table 10, the percent risk contribution due to Type A testing is as follows:
%Risk13 =[( Class3a., + Class3b.,) / Total,] x 100 (Eq. 10)
Where:
Class3a,s = Class 3a person-rem/year = 2.92E-03 person-rem/year Class3b., = Class 3b person-rem/year = 3.34E-03 person-rem/year Total, = total person-rem year for 15-year interval = 8.00568E+01 person-rem/year (Table 8)
%Risk1 , = [(2.92E-03 + 3.34E1-03) / 8.00568E+01] x 100 = 0.0078%
RSC 04-02NP 19 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension The percent risk increase @%Risk,,) due to a fifteen-year ILRT over the baseline case is as follows:
A%Riskis = [(Totals - TotaL.) / TotatASS] x 100.0 (Eq. 11)
Where:
TotaL 4SE = total person-rem/year for baseline (3 per 10 years) interval = 8.00561E+01 person-rem/year (Table 6)
Total, = total person-rem/year for 15-year interval = 8.00568E+01 person-rem/year (Table 8)
ARisk,s = [(8.00568E+01 - 8.00561 E+01) / 8.00561E+01] x 100.0 = 0.0010%
Step 5: Calculateincrease in risk due to extending Type A inspection intervals Based on the previously approved methodology (References 3 and 5), the percent increase in risk (in terms of person-rem/yr) of these associated specific sequences is computed as follows.
%Risk, 0.os = [(PER-REMas - PER-REM, 0 ) / PER-REM,.] x 100 (Eq. 12)
Where:
PER-REM, 0 = person-rem/year of ten years interval (see Table 7, Classes 1, 3a and 3b) = 8.948E-03 person-rem/yr PER-REM,, = person-rem/year of fifteen years interval (see Table 8, Classes 1, 3a and 3b) 9.203E-03 person-rem/yr
%Risk,.0 ,s= [(9.203E 8.948E-03) / 8.948E-03] x 100 = 2.85%
The percent increase on the total integrated plant risk for these accident sequences is computed as follows.
%Total o-is = [(Totals - Total.) / Total.] x 100 (Eq. 13)
Where:
Totaho = total person-rem/year for 10-year interval = 8.00566E+01 person-rem/year (Table 7)
Totah5 = total person-rem/year for 15-year interval = 8.00568E+01 person-rem/year (Table 8)
% Totalio.s = [(8.00568E+01 - 8.00566E+01) / 8.00566E+01] x 100 = 0.0003%
Step 6: Calculate the change in riskin terms of large early releasefrequency (LERF)
The risk impact associated with extending the ILRT interval involves the potential that a core damage event that normally would result in only a small radioactive release from containment RSC 04-02NP 20 Printed: 05!12/2004
Evaluation of Risk Significance of ILRT Extension could in fact result h a larger release due to failure to detect a pre-existing leak during the relaxation period.
From References 3 and 5, the Class Ia dose is assumed to be 10 times the allowable intact containment leakage, LI (or 2,220 person-rem) and the Class 3b dose is assumed to be 35 times L (or 7,780 person-rem). The dose equivalent for allowable leakage (L.) is developed in Appendix C. This compares to a historical observed average of twice L (Reference 11).
Therefore, the estimate is somewhat conservative.
Based on the previously approved methodology (References 3 and 5), only Class 3 sequences have the potential to result in large releases if a pre-existing leak were present. Class I sequences are not considered as potential large release pathways because for these sequences the containment remains intact. Therefore, the containment leak rate is expected to be small (less than 2xL). A larger leak rate would imply an impaired containment, such as classes 2, 3, 6 and 7.
Late releases are excluded regardless of the size of the leak because late releases are, by definition, not a LERF event. At the same time, sequences in the SONGS PRA (Reference 10) that result in large releases, are not impacted because a LERF will occur regardless of the presence of a pre-existing leak. Therefore, the frequency of Class 3b sequences is used as the increase in LERF for SONGS, and the change in LERF can be determined by the differences.
References 3 and 5 identify that Class 3b is considered to be a contributor to LERF. The assumed dose for this class is compared to other LERF sequences to determine if it truly represents an increase in LERF. In order to be a LERF sequence, it must be both early in time and large in population dose. The first condition is met since the failure represents an existing isolation failure. However, the dose is small compared to other early sequences. Table 9 compares the doses for this and several other cases.
Table 9 Comparisons of Release Class Doses Release ClassI -Po t Dos- (Person- re)
Class 3b (Table 6) 7,780 Class 8 (Table 6) 86,100,000 Class 7 (Table 6) 13,600,000 The table shows that even a conservative estimate for the release (person-rem) is found to be less than 1.0 percent of that obtained from other early release classes. On a best-estimate basis the average expected leakage would be less than 444 person-rem (as developed in Appendices A and C) and would be less than 1.0 percent of the other classes associated with large early release.
The conclusion can be drawn from this data that the potential consequence of a Type A leakage event is not large and the proposed change has no impact on LERF. However, conservatively Class 3b is considered to be an estimate for the change in LERF to be consistent with the RSC 04-02NP 21 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension accepted methodobgy (References 3 and 5). Table 10 summarizes the results of the LERF evaluation assuming that Type 3b is indicative of a LERF sequence.
Table 10 Impact on LERF due to Extended Type A Testing Intervals ILRT inspection Interval 3 Years (baseline) Yea s10 -5 Years Class 3b (Typc A LERF) 3.74E-07 4.11E-07 4.30E-07 ALERF (3 year baseline) 3.74E-08 5.60E-08 ALERF (10 year baseline) . -. 87E-08 Reg. Guide 1.174 (Reference 8) provides guidance for determining the risk impact of plant-specific changes to the licensing basis. Reg. Guide 1.174 defines very small changes in risk as resulting in increases of core damage frequency (CDF) below L.OE-06/yr and increases in LERF below L.OE-07/yr. Since the ILRT does not impact CDF, the relevant metric is LERF.
Calculating the increase in LERF requires determining the impact of the ILRT interval on the leakage probability.
Since guidance in Reg. Guide 1.174 defines very small changes in LERF as below 1.007/yr, increasing the ILRT interval to 15 years (1.87E-08/yr) is non-risk significant. It should be noted that if the risk increase is measured from the original 3-in-10-year interval, the increase in LERF is 5.60E-08/yr, which is also below the 1.01E07/yr screening criterion in Reg. Guide 1.174.
Step 7: Calculate the change in conditionalcontainmentfailureprobability(CCFP)
The CCFP is defined as the probability of containment failure given the occurrence of an accident. This probability can be expressed using the following equation:
CCFP- = [_f F ] (Eq. 14)
Where ftnc]) is the frequency of those sequences which result in no containment failure. This frequency is determined by summing the Class I and Class 3a results, and CDF is the total frequency of all core damage sequences.
Therefore the change in CCFP for this analysis is the CCFP using the results for 15 years (CCFP.) minus the CCFP using the results for 10 years (CCFP.). This can be expressed by the following:
A CCFP,( 15 = CCFI, -CCFJo 0 (Eq. 15)
Using the data previously developed the change in CCFP from the current testing interval is calculated and presented in Table 11.
RSC 04-02NP 22 Printed: 05/1212004
Evaluation of Risk Significance of 1LRT Extension Table 11 Impact on Conditional Containment Failure Probability due to Extended Type A Testing Intervals
-.ILRT Inspection Interval l, lears 0 Y'ars(baselin 15 Years Xncf) 1.4626E-05 A.4589E-05 1.4570E-05 Ancf),!CDF 0.815 0.813 0.812 CCFP 1.85E-01 1.87E-01 1.88E-01 ACCFP (3 year baseline) , 0208% 0.312%
ACCFP (1O year baseline) . 0.104%
5.0 REFERENCES
- 1. San Onofte Nuclear Generating Station Unit 2, Containment Integrated Leak Rate Test, Rev. 4, Procedure S02-V-3.12.
- 2. San Onofre Nuclear Generating Station Unit 3, Containment Integrated Leak Rate Test, Rev. 4, Procedure S03-V-3.12.
- 3. Indian Point 3 Nuclear Power Plant, "Supplemental Information Regarding Proposed Change to Section 6.14 of the Administrative Section of the Technical Specification",
Entergy, IPN-01-007, January 18,2001.
- 4. Indian Point Nuclear Generating Unit No.3 - Issuance of Amendment Re: Frequency of Performance-Based Leakage Rate Testing (TAC NO. MB0178), United States Nuclear Regulatory Commission (USNRC), April 17, 2001.
- 5. Evaluation of Risk Significance of ILRT Extension. Revision 2, Florida Power Corporation, F-0I-000 I June 2001.
- 6. Industry Guideline for Implementing Performance-Based Option of 10 CFR Part 50, Appendix J Revision 0, Nuclear Energy Institute, NEI 94-01, July 26, 1995.
- 7. Gisclon, J. M., et al, Risk Impact Assessment of Revised Containment Leak Rate Testing Intervals Electric Power Research Institute, TR-104285. August 1994.
- 8. An Approach for Using Probabilistic Risk Assessment in Risk-Informed decisions on Plant-Specific Changes to the Licensing Basis, U.S. Nuclear Regulatory Commission (USNRC), Regulatory Guide 1.174, July 1998.
RSC 04-02NP 23 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- 10. San Onofre Nuclear Generating Station Living PRA. SONGS 2/3, PRA Level II Analysis Report, IPE-LEVEL2-000.
- 11. Performance-Based Containment Leak-Test Program. USNRC, NUREG-1493. July 1995.
- 12. PRA Procedures Guide- A Guide to the Performance of Probabilistic Risk Assessments for Nuclear Power Plants, American Nuclear Society and the National Institute of Electrical and Electronic Engineers, NUREG/CR-2300, January 1983.
- 13. Southern California Edison, San Onofre Nuclear Generating Station 2&3 FSAR.
Updated June 2003.
- 14. Summitt, R., Assessment of Safety Benefit for Installation of a Generator Disconnect Switch at Robinson Ricky Summitt Consulting (RSC), Inc., RSC 98-19 June 1998.
- 15. Sumrnitt, R., Comanche Peak Steam Electric Station Probabilistic Safety Assessment.
Evaluation of Risk Significance of ILRT Extensiom RSC, Inc., RSC 01-47/R&R-PN-110, November2001.
- 16. San Onofre Nuclear Generation Station WinNUPRA/WinNUCAP Model Update, January 2004.
RSC 04-02NP 24 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension Appendix A.
Alternative Calculation for ILRT Evaluation RSC 04-02NP Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension A.0 INTRODUCTION In addition to the approved template, other analyses are possible that rely on a best-estimate approach to avoid some potential weaknesses identified below:
- The calculation of the increase in the probability of Type A leakage is incorrect and equates a probability with a percentage increase.
- The approach utilizes upper bound (950/o-tile) values and not best-estimate mean values when developing the estimates for the likelihood of Type A leakage events although adequate data exists.
- The approach arbitrarily increases the probability of leakage for the intact containment cases by a factor of two although the Type A leakage is assessed separately and the intact containment cases would still be isolated.
- The approach increases the intact containment source term by factors of 1.1 and 1.15 for the 10-and 15-year cases although the source term is a physical process that would not be altered by changing the statistic of testing.
- Although historical data indicates that the mean value for leakage is on the order of twice allowable (2xL9 ) the analysis arbitrarily utilizes factors of lOTxl and 35xL' which tends to overestimate the initial dose and thereby mask some of the predicted increase in risk when considered in terms of delta change.
- The analysis assumes that Type 3a and Type 3b are independent when developing the upper bound estimates and utilizes the 144 events as two separate populations. Actually the Type 3b leakage events are a proportion of the Type 3a events and should be calculated in a dependent manner. The embedded independence assumption results in an arbitrary increase in the likelihood of leakage by over a factor of two when compared to actual historical evidence. This overestimation inflates the baseline calculation and reduces the net change predicted in a non-conservative manner.
The approach documented in this appendix addresses these weaknesses by providing a best-estimate approach. Although the approach is similar to the template there are key differences that are highlighted as appropriate. The same input data (References 7, 9, 10 and 11) are used to generate release category frequency information.
The person-rem information is based on the approach found in Appendix B.
Section A.1 of the document presents a summary of the analysis steps. Section A.2 presents the baseline analysis. Sections A.3 and A.4 develop the impact of the increased testing interval on the analysis metrics. Section A.5 presents a summary of the analysis.
RSC 04-02NP A.1 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension A.1 ANALYSIS APPROACH The analysis is based on guidance provided in Reference 7 and uses risk metrics presented in Reference 11 to evaluate the impact of a proposed change on plant risk. Finally, Reference 8 suggests two measures be utilized in the assessment, core damage frequency (CDF) and the large early release fraction (LERF). It is these two metrics that are assessed first.
These measures are considered in the determination of the impact of testing extension. The San Onofre Nuclear Generating Station (SONGS) is currently considering an extension from 10 years to 15 years.
Since the testing addresses the ability of the containment to maintain its function, the proposed change has no measurable impact on core damage frequency. Therefore, this attribute remains constant and has no risk significance.
The change in testing interval could impact the ability of the containment to perform its function and this could impact the LERF attribute. Therefore, the estimated change in LERF is addressed.
The change in risk, as defined by the change in annual population person-rem dose is calculated.
This metric provides a means to identify the hicreased risk posed by the change in testing interval.
The basic analysis steps are outlined below:
- Define baseline plant damage states and person-rem estimates
- Calculate baseline Type A leakage estimate to define the analysis baseline
- Modify Type A leakage estimate to address extension of the Type A test frequency
- Compare analysis metrics to estimate the impact and significance of the increase related to those metrics The first step in the analysis is to define the baseline plant damage states and person-rem dose measures. Plant damage state information is developed in Section 2.0 of the main report. The plant damage state information and the release fraction information (Table 3 of main report) are used to develop the population person-rem dose estimates.
The SONGS PRA (Reference 9 and 10) plant damage state definitions include isolation failures due to Type B and Type C faults and examine containment challenges occurring after core damage and/or reactor vessel failure. These sequences are grouped into plant damage states.
Using the plant damage state information, bypass, isolation failures and phenomena-related containment failures are identified. Once identified, the sequence was then classified by release category definitions specified in Reference 7 and summarized in Table 5 of the main report.
The second step expands the baseline model to address Type A leakage. The SONGS PRA did not explicitly include Type A (liner-related) faults and this contribution must be added to provide a complete baseline. In order to define leakage that can be linked directly to the Type A testing, RSC 04-02NP A.2 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension it is important that only failures that would be identified by Type A testing exclusively be included.
Reference 11 provides the estimate for the probability of a leakage contribution that could only be identified by Type A testing based on industry experience. This probability is then used to adjust the intact containment category of the SONGS PRA to develop a baseline model including Type A faults.
The release, in terms of person-remn, is developed based on information contained in Reference 11 and is estimated as a leakage increase relative to allowable release L defined as part of the ILRT.
The predicted probability of Type A leakage is then modified to address the expanded time between testing. This is accomplished by a ratio of the existing testing interval and the proposed test interval. This assumes a constant failure rate and that the failures are randomly dispersed during the interval between the test.
The change due to the expanded interval is calculated and reported in terms of the change in population person-rem. In addition, the change in large early release frequency is predicted and compared to the acceptance criteria presented in Reference 8.
From these comparisons, a conclusion is drawn as to the risk significance of the proposed change.
A.2 DEFINITION OF ACCIDENT SEQUENCES The SONGS PRA (References 9 and 10) provides the baseline core damage bin frequency information for the contributing accident sequences. The assessment includes internal initiating events and the total core damage frequency is estimated to be 1.79E-05/yr. Table 2 in the main report presents a summary of this information.
A.3 CALCULATION OF INCREASE IN TYPE-A RELATED LEAKAGE In order to determine the impact of the change in testing interval, it is first necessary to define a baseline probability for Type A leakage events and then to adjust this probability to account for the proposed change in testing interval.
Reference 11 states that a review of experience data finds that Type A testing identified only 4 leakage events of the 144 events identified. Thus about 3% (0.028) of containment leakage events are identified by the ILRT and that the remaining events are identified by the LLRT that is not being evaluated for change. This probability, however, is based on three tests over a 10 year period and not the one per ten year frequency currently employed at SONGS (References 1 and 2). The probability (0.028) must be adjusted to reflect this difference.
The impact of relaxing the Type A penetration test interval will increase the average time that a leak that could only be detected by the Type A test would possibly be present. The increase in risk is proportional to the increase in the duration between containment tests. The historical data RSC 04-02NP A.3 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension is based on testing three times per 10 years (120 months). This equates to a mean time between tests of 3.3 years or 40 months. The SONGS testing interval is once per 10 years (120 months).
The increase in the exposure time will influence the probability of leakage.
To calculate this impact, two assumptions are made consistent with standard practice and are listed below:
- A constant rate for Type A leakage events;
- The potential for leakage is equally distributed across the period of interest such that the exposure time is reduced by one-half the period.
With these assumptions, the increase can be determined by a ratio of the proposed to the prior exposure times multiplied by the known rate for the prior probability of failure. The equation is shown below:
plos= P10.3 05oExrs 011 (Eq. l) 0.5 *EXp 1 013 Substituting the values for pot3 (0.028) and the exposure times (ExplQ'I = 120, Explo' 3 =40) yields a value for the probability of leakage of 0.0833. This value serves as the baseline probability of Type A leakage for the analysis.
The proposed change would increase the duration between tests by decreasing the number of tests from once per 10 years to once per 15 years. Therefore, the total time between Type A testing will increase from ten years (120 months) to 15 years (180 months). The same equation is again utilized with the variables altered to reflect the specific bounds as shown below:
= 0 0.5 . ExP0 (Eq. 2)
Substituting yields a value for the probability of Type A (ILRT) detectible leakage events for the relaxed testing interval of 15 years. This probability represents the probability of a leakage path that could only be identified by a Type A test. The results are summarized below in Table A. 1.
Table A.1 Probability of Type A Leakage Given a Testing Interval Case - Probabilityo-ePr bi XO Baseline (once per 10 years) 0.0833 15-year testing 0.125 The baseline analysis must include the consideration being assessed in order to preclude biasing the results. The existing analysis does not account for Type A faults. The model is expanded to RSC 04-02NP AA Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension encompass Type A faults. The intact containment cases are adjusted to include these isolation failures. Since other sequences represent failure they are not adjusted.
For the 15-year period case, he intact containment frequency for the baseline (1.50E-05) is multiplied by the potential for Type A leakage (0.125) for the baseline case to generate the frequency contribution associated with Type A leakage 0.88E-06/yr). The intact containment contribution is then reduced by this value to maintain the overall frequency (1.31E-05). Table A.2 summarizes the results for the extended case.
Table A.2 Type A Leakage Frequency Variable 10-year penod case 15- A Testing interval (years) 10 15 Intact containment frequency (Class 1) l.50E-05 l.50E-05 Baseline probability Type A fault 8.33% 8.33%
Extension (years/test) 0 5 Modified probability of Type A faults 8.33% 12.5%
Contribution from Type A faults 1.25E-06 1.88E-06 Revised intact containment frequency (Class 1) 1.38E-05 1.3 1E-05 In order to develop the estimate for person-renm it is necessary to determine the magnitude of the expected release for Type A leakage. Information in Reference 11 indicates that the typical leakage from a Type A failure is on the order of 2xL, where La represents the allowable leakage rate.
Information in Appendix C estimates the intact containment contribution as 2.22E+02 rem. The dose rate from this release category is selected and then doubled to define the expected release for the Type A leakage of 4.44E+02 person-rem (2 x 2.22E+02).
This new release category must be compared to other sequences to determine if it represents an increase in the large early release fraction In order to be a LERF sequence, it must be both early in time and large in population dose. The first condition is met since the failure represents an existing isolation failure. However, the dose is small compared to other early sequences. Tabe A.3 compares the doses for this and several other cases.
RSC 04-02NP A.5 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table A.3 Comparison of Release Class Doses i;-;Release Class - Population Dose (person-rem)
Class 7 (Section 4.2, Table 6) 13,600,000 Class 8 (Section 4.2, Table 6) 86,100,000 Type A (Class 3b) 444 The Type A person-rem estimate is shown to be substantially lower than the other early releases (less than 1.0 percent). Therefore, the potential consequence is not a LERF sequence and the proposed change has no impact on LERF.
No examples of Type A leakage sufficient to be considered LERF are identified in the historical data. However, an estimate for LERF is calculated for comparison using the Reg. Guide 1.174 risk criterion for change in LERF and is developed. The information provided in References 7 and 11 indicate that the historical leakage rates are not sufficient to result in a situation defined as large-early release. Therefore, the current data supports the supposition that the relaxed testing interval will not have a measurable impact on LERF.
However, the data does not preclude events that may occur at a lower probability than would be supported by the data collected to date. As a sensitivity study, the probability of larger leakage rates is conservatively estimated. Using the estimated probability and the plant damage state information presented in the SONGS PRA, an estimate of LERF is defined for both the baseline model and the 15-year frequency.
The LERF contribution is based on estimation of the frequency through the use of a chi-square distribution to develop an upper estimate as defined by Reference 12. The chi-square distribution can provide an upper bound given no events. The general equation is presented below:
(a) x 2 (2F +2,a) q 3) 2N(q.3 where: N is the number of events, F is the number of events of interest, a is the percentile distribution (assumed to be the 95 0/o-tile).
This equation is used to estimate the probability that, given a leak, it will be sufficiently large to represent a LERF contributor. For this estimate, the following is supplied: N=144 events, F=0 LERF events.
Substitution yields the following:
RSC 04-02NP A.6 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension P(a) 2(2,0.05) (Eq. 4)
For the baseline or 10-year case, solving for the probability yields a value of 0.02 (5.991288).
This probability represents a probability that given a leak event, it will be sufficiently large to contribute to LERF. The probability of a Type A failure sufficient to contribute to LERF is found by multiplying the probability of a Type A leak (0.0833) and the probability that the leak will be sufficiently large to generate a LERF contributor (0.02). This equates to a probability of a Type A LERF contributor of 1.67E-03 for the baseline case. For the case involving a 15-year interval, the probability of a leak increases to 0.125. Thus, the probability of a Type A LERF contributor is estimated as 2.50E-03 for the 15-year case.
With the LERF contributor estimated, the next section defines the baseline LERF frequency.
The baseline LERF frequency is defined by collecting any frequency for plant damage states (PDS) that:
- Involve an early containment isolation failure
- Involve a bypass failure
- Involve early containment failure at or near reactor vessel failure Since the potential for containment challenges is addressed in the containment event tree (CET),
the last item must be addressed later in the definition of release categories. However, the LERF frequency attributed to the first two conditions can be defined by the PDSs.
Isolation failures must be sufficiently large to preclude any future challenges. For example, an isolation failure must be of sufficient size to mitigate pressure challenges that could occur at reactor vessel failure, e.g., loads from high pressure melt ejection. If this condition is not met, the failure is not sufficient large to meet the definition of a LERF contributor. References 10 and 14 define a baseline LERF frequency of 7.3 1E-07/yr. This is the sum of the frequencies for release categories B, D, G, and T in Table 3 of the main report.
A.4 ESTIMATION OF IMPACT ON LERF The LERF contribution defined in the SONGS PRA does not include the impact of Type A leakage events. This contribution must be added to provide a basis comparison.
The Type A LERF contribution is determined by multiplying the probability of a LERF sequence by the intact containment contribution in the same manner as the Type A sequence contribution was developed. The LERF probability (or split fraction) is determined by multiplying the probability of Type A leakage (0.0833 for the 10-year case) by the probability that the leakage will be large (0.02). The result, 1.67E-03, represents the LERF fraction. This value is then multiplied by the intact containment frequency to obtain the result 0.50-05 x 1.67E-03=
2.60E-08). Table A.4 summarizes the calculations for the 10- and 15-year cases.
RSC 04-02NP A.7 Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension Table A.4 Type A LERF Contribution Variable .asel ne 15-year peod case Testing interval (years) 10 15 Intact containment frequency (Class 1) 1.50E-05 1.50E-05 Probability Type A LERF fault, p(LERF) 0.02 0.02 Probability of Type A leakage, p(TYPEA) 0.0833 0.125 LERF fraction, LF = p(LERF) x p(TYPEA) 0.001666 0.0025 Type A LERF frequency, Class 1 x LF 2.60E08 3.90E-08 The calculated Type A LERF frequency is added for the baseline and the 15-year case and the difference calculated. This represents the increase in LERF due to the relaxation of the testing interval. The results are presented in Table A.5.
Table A.5 Calculation of the Change in LERF dVarible .aseline 15 - pe case; Baseline LERF 7.311-07 7.31E-07 Type A LERF frequency 2.60E-08 3.90E-08 Total LERF 7.57E-07 7.70E-07 Delta LERF 1.30E-08 Reference 4 defines a set of risk significance criteria. The following summarizes the criteria:
- If the calculated increase is very small, which is taken as being less than 10- per reactor year, the change is typically considered to be an insignificant increase in risk.
- If the increase is in the range of 10-7 per reactor year to 10-6 per reactor year, proposed change will be considered only if it can be reasonably shown that the total LERF is less than I05 per reactor year.
RSC 04-02NP A.8 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- If the result shows an increase above 10-6 per reactor year, the proposed change would not normally be considered.
A comparison of the results to these criteria indicate that the change in LERF frequency is below the level differentiating risk significance and therefore the net change is not risk significant. This result is based on assuming a conservative estimate for the potential for a Type A failure resulting in a LERF sequence. A more realistic value would most likely result in a further reduction in the change in LERF and further support this conclusion.
A.5 MODIFIED MODEL EVALUATION The information provided in Section 4.0 develops an estimate of the increase in the likelihood of a containment isolation failure given that the Type A ILRT testing interval is extended. An increase in the testing interval to once per 15 years increases the probability of a Type A detectible leakage by 0.04. This increase is used to adjust the baseline model to determine the estimated person-rem The baseline model results must be adjusted to address this increased likelihood of increased containment leakage. Only certain sequences would be impacted by this increase since many sequences already involve an impaired containment or isolation failure. Basically only intact containment scenarios need be addressed.
Sequences that already represent release sequences are excluded. The intact containment sequences are combined with the increased probability of leakage (0.04) to define a new contribution to increased leakage. The resulting change in person-rem is summarized below in Table A.6.
Table A.6 Person-Rem Comparisons Case Person-re r Delta Peron-Rem Per t Increase Baseline with Type A 80.05661 ; -
15-year 80.05692 0.0003 0.0003%
- 1. From Table 7 of the main report.
- 2. From Table 8 of the main report.
The net increase in person-rem per year is estimated to be 0.0003% (0.000003) for the 15-year case. The results indicate an increase in integrated person-rem of approximately 4.17E-03 person-rem if the 15-year interval is adopted. The integrated results are presented for all cases in Table A.7.
RSC 04-02NP AS9 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table A.7 Integrated Person-Rem Estimates
]ntegrated Person-Case - Person-Remlyr Delta Person-Rem Ren Increase, Baseline with Type A 80.0566 15-year 80.0569 0.0003 4.17E-03
- 1. Net increase relative to the period being assessed for extension.
A.6 REFERENCES
- 1. San Onofre Nuclear Generating Station Unit 2. Containment Integrated Leak Rate Test, Rev. 4, Procedure S02-V-3.12.
- 2. San Onofre Nuclear Generating Station Unit 3, Containment Integrated Leak Rate Test, Rev. 4, Procedure S03-V-3.12.
- 3. Indian Point 3 Nuclear Power Plant. "Supplemental Information Regarding Proposed Change to Section 6.14 of the Administrative Section of the Technical Specification",
Entergy, IPN-01-007, January 18,2001.
- 4. Indian Point Nuclear Generating Unit No.3 - Issuance of Amendment Re: Frequency of Performance-Based Leakage Rate Testing (TAC NO. MB0178), United States Nuclear Regulatory Comnmission, April 17, 2001.
- 5. Evaluation of Risk Significance of ILRT Extension Revision 2, Florida Power Corporation, F-01-0001, June 2001.
- 6. Industry Guideline for Implementing Performance-Based Option of 10 CFR Part 50, Appendix J Revision 0, Nuclear Energy Institute, NEI 94-01, July 26, 1995.
- 7. Gisclon, J. M., et al, Risk Impact Assessment of Revised Containment Leak Rate Testing Intervals, Electric Power Research Institute, IR-104285, August 1994.
- 8. An Approach for Using Probabilistic Risk Assessment in Risk-Informed decisions on Plant-Specific Changes to the Licensing Basis. U.S. Nuclear Regulatory Commission (USNRC), Regulatory Guide 1.174 July 1998.
- 10. San Onofre Nuclear Generating Station Living PRA. SONGS 2/3. PRA Level II Analysis Report IPEFLEVEL2-000.
RSC 04-02NP A.10 Printed: 05!12/2004
Evaluation of Risk Significance of ILRT Extension
- 11. Performance-Based Containment Leak-Test Program, USNRC, NUREG-1493, July 1995.
- 12. PRA Procedures Guide- A Guide to the Performance of Probabilistic Risk Assessments for Nuclear Power Plants American Nuclear Society and the National Institute of Electrical and Electronic Engineers, NUREG/CR-2300. January 1983.
- 13. Southern California Edison, San Onofre Nuclear Generating Station 2&3 FSAR, Updated, June 2003.
- 14. San Onofre Nuclear Generating Station WinNUPRA/WinNUCAP Model Update, January 2004.
RSC 04-02NP A.11 Printed: 0511212004
Evaluation of Risk Significance of ILRT Extension Appendix B:
Surrogate Person-Rem Methodology (RSC 01-44)
RSC 04-02NP Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension Southern California Edison RSC 04-02NP Evaluation of Risk Significance of ILRT Extension Revision 0 March 2004 Principal Analyst Jeffrey Miller NON-PROPRIETARY DOCUMENT This document has been reviewed and proprietary information removed. It may be freely distributed as a complete document only.
Ricky Surnmitt Consulting, Inc.
RS 7 Risk and Reliability Engineering 342 Ebenezer Road, Knoxville, TN 37932 Telephone 865.692.4012 Fax 865.692.4013 RSC 04-02NP B.1 Printed 05112/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology RSC Document Configuration Control Form FORM NO.: RSC-RPT-STD99-04, Rev. 4 Report Number: RSC 04-02
Title:
Evaluation of Risk Significance of ILRT Extension Revision: Revision 0 Author: Jeffrey Miller Date Completed: August 6, 2001 Location on Server of Report Files: //rscsvr1/sc.jnternalreports Location of Report and Files RSC Site: NA Date of Record for all Models and August 6, 2001 Analysis Software and Version Used: Word (doc) Version 95, Version 97, Version 2000 (bold all that apply)
Excel (xIs): Version 95, Version 97, Version 2000 Access (mdb): Version 97, Version 2000 Designer (ds4/f): Version 7 CAFTA: Version 3.2b ETA Version 3.2b MAAP BWR Version 3.0B R9, R1 1, Version 4.0 MAAP PWR Version 18, 19, 20, Version 4.0 RSC Software: PRAMS, SIP, TIFA, BAYESUPDATE RSC RSC STD RPT R3 RSC 04-02NP B.2 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology Report Review and Resolution Form FORM NO.: RSC-RPT-RVROO-02Rev. 2 Preparer: Jeffrey Miller RSC Reviewer R. Summitt Date: August 6, 2001 RSC Approver: R. Summitt Date: August 6, 2001 Abstract (brief statement of purpose): Document methodol gy for converting radionuclide release fractions to dose. NOTE: Document grandfathered and does not require independent review since prior client review.
Documentation Retrieval Information:
Keywords: Level 2 Analysis Other Calculation MAAP Analysis 0 Amends / EJ Superceeds I D Supplements RSC Document(s): PSA Paper (Reference 1).
Verification and Review Method:
El Detailed Review El Alternative Calculation El Qualification Testing 0 Other (specify: Grandfathered)
General Documentation Requirements Acceptable Reviewer Comments Introduction - provides summary of l purpose, scope, and principle tasks required to meet objective Methodology - description of process and H supporting methodology that is sufficient to understand approach and to support peer review Analysis and Results - detailed Li documentation of the implementation of the methodology and task steps that may be supported by report appendices and includes intermediate and final results Conclusions and Recommendations - L NA concise presentation of results of the analysis that answers the objectives of the study and should include any important assumptions and/or findings Editorial Review:
0 Spell Checked 0 Grammer Checked 0 Tables and Figures Checked 0 Sections Checked Sufficient References to Reproduce Results: Yes RSC RSC STD RPT R3 RSC 04-02NP B.3 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology Resolved all Comments: NA Incorporated Resolutions From Review: NA Reviewer Comment Resolution of Comment
- 1. NA I
4.
Editorial or illustrative comments are attached to this review sheet to complete the review package.
RSC RSC STD RPT R3 RSC 04-02NP B.4 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology Table of Contents Section Page
1.0 INTRODUCTION
..................................................................... 7 2.0 METHODOLOGY ..................................................................... 7 3.0 DEVELOPMENT OF RADIONUCLIDE RELEASE TO PERSON-REM RELATIONSHIP 7
3.1 DATA ASSESSMENT EXISTING ................................................................... 7 3.2 DATA INTERPRETATION ........................................... 10 3.4 APPLICATION WITH MAAP ........................................... 12 3.5 QUALITATIVE UNCERTAINTY ASSESSMENT ........................................... 13 3.5.1 Qualitative Evaluation of Predictive Dose ........................................... 13 3.5.2 Results Predictability ........................................... 14 4.0
SUMMARY
AND CONCLUSIONS ............................................ 14
5.0 REFERENCES
........................................... . 14 RSC OI-44NP Printed 08/06/2001 RSC 04-02NP B.5 Printed 05/12/2D04
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology List of Tables Table Page Table 1 Release Split Fraction to Dose Conversion Factors ............................................. 12 Table 2 Mapping of Method Variables to MAAP Output Variables ............................................. 12 List of Figures Figure Page Figure 1 Sequoyah Release Fractions (Reference 2) ............................................. 8 Figure 2 Unpublished PWR Release Fractions (Reference 3) .............................................. 9 Figure 3 Oconee IPE Release Fractions (Reference 4) ............................................. 9 Figure 4 Seabrook Release Fractions (Reference 5) ............................................. 10 Figure 5 Relationship between Release Fraction and Dose ............................................. 10 Figure 6 Comparison of Equation Results and Reported Dose Values ............................................ 1 Figure 7 Variation of Equation to Reported Dose ............................................. 13 RSC 0I-44NP Printed 08/06/2001 RSC 04-02NP B.6 Printed 05/12/2004
1.0 INTRODUCTION
The current industry emphasis is on applying the PSA t assist in plant operational decision-making. Most of the IPE submittals stop at the frequency of containment release and do not address offsite consequences. Since public safety is a primary consideration, it is important to have a tool that provides insights into how potential changes will impact public health risk.
Although a primary measure currently being proposed examines changes in the large early release fraction (LERF), the total effect should also be considered when evaluating changes.
The total whole body person-rem released is one measure to address the change in public health risk due to a proposed change to plant configuration. This quantity is considered one possible measure of merit and is traditionally calculated for the Level 3 PSA.
Given that most PSAs stop at containment release, additional effort is needed. To generate the person-rem release in order to expand the evaluation it is necessary to develop a model for extrapolating the existing information in the PSA to person-rem.
One approach to accomplish this task is to expand the existing PSA into a Level 3 PSA. This requires information on meteorological conditions, population densities, and evacuation planning. This information is then input into an offsite analysis code and results generated. The effort required to develop this detailed model may not be necessary for most cases.
A surrogate model can be used to estimate the change in whole body person-rem based on existing analyses'. The process used to develop the model is present in this report.
2.0 METHODOLOGY The basis for the surrogate model is the development of a relationship between the radionuclide release fractions and the predicted whole body person-rem. To make the model useful, this relationship is developed at a release category level and in terms of a minimal set of radionuclide release fractions that, based on prior studies, can be shown to control the various aspects of offsite doses. This is accomplished by examining several prior studies that included measures of offsite consequences.
3.0 DEVELOPMENT OF RADIONUCLIDE RELEASE TO PERSON-REM RELATIONSHIP The understanding that the dose values must be considered in terms of the "fence post" dose is key to the model development. In other words, the dose that the envelop around the plant would receive. This allows the results to be independent of evacuation and meteorological considerations. The result may be somewhat conservative, but it provides a measure that can be applied across plant sites uniformly.
3.1 DATA ASSESSMENT EXISTING The results of the Level 2 IPE assessment are typically provided in terms of release category frequencies and radionuclide release fractions. Therefore, any method must utilize these two RSC 01-44NP Printed 08/06/2001 RSC 04-02NP B.7 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology characteristics to form the basis for estimating the offsite consequence from release sequences to be useful.
To determine this relationship, available published and unpublished Level 3 PSAs were reviewed to determine a range of release fractions and corresponding doses. The release fractions identified in these PSAs for the following radionuclides: noble gases, iodine, cesium, tellurium, strontium, ruthenium, lanthanum, cerium, and barium. The relative release fractions for each were collected as identified in the PSAs.
These radionuclides are most reported in the literature and provide the majority of offsite dose.
The release fractions for each of the release categories is cataloged (each release category is defined as a case) along with the associated whole body person-rem. Figures A. 1 tfrough A.4 graphically presents the results for four PSAs as examples of this effort.
- Noble Gas A Iodine
- Cesium X Tellurium
- Strontium 1 Ruthenium a Lanthanum
- Cerium
- Barium I- .
0
- a 0.9 J .
U 0.8 - .
0.7 -
C U 0.6- . 4
- a X 05 - 4 OA.
U a 03 02
- x
& x 0.1 x x
- W W W i X .6C * -
0: : He s A_-A____
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Case Figure 1 Sequoyah Release Fraction Cases (Reference 2)
RSC 01-44NP Printed 08/06/2001 RSC 04-02NP B.8 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology E]
Figure 2 Unpublished PWR Release Fraction Cases (Reference 3)
- Lanthanum U Cerium *Barium l ***********l*.
0.9 0.8
° 0.7
" 0.6
,, 0.5
& OA 03 Aa 0.2 0.1 1 3 5 7 9 11 13 15 17 19 Case Figure 3 Oconee IPE Release Fraction Cases (Reference 4)
RSC 0-44NP Printed 08/06/2001 RSC 04-02NP B.9 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology
- Nobl Gas a Iodine
- Cesium X Tellurium
- Strntitan a Ruthenium a Lanthanum E Cerimn Barium 0.9 0.8 z 0.7 0.6 t 0.D
~ 0.4 X 9 02 0.1 '
1 2 3 4 5 6 7 8 9 10 1I 12 13 Case Figure 4 Seabrook Release Fraction Cases (Reference 5) 3.2 DATA INTERPRETATION From these studies, a total of 56 unique release categories, defining radionuclide fractions and person-rem were plotted on a normalized plot to determine the type of relationship that existed between dose and release fractions. Five of the more important radionuclides were used to develop the release fraction value. These five radionuclides, noble gases, [ ],[1]["], and [ ],
are all considered important contributors to offsite dose.
Noble gas releases were chosen to represent the "baseline" dose. Most studies indicate that if a release occurs, the vast majority of noble gases will be released. The others were chosen based on their relatively important biological effects and tend to be significant release contributors.
Figure A.5 shows how the dose essentially maps the release fraction.
I - Dose -- - Release Fraction C.
z 11 21 31 41 51 Cast Figure 5 Relationship between Release Fraction and Dose RSC 01-44NP Printed 08/06/2001 RSC 04-02NP B.10 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology Although a clear linear relationship does mt exist between the two functions, it is clear that a trend is found between the fraction released and the resulting dose. This is hardly a revelation since the dose exposure is a function of the radionuclides released. The simplicity of the relationship, [ ], is somewhat of a surprise. Given this relationship, a set of 56 [ ] equations was developed. For each case, the equation took the formn Di=[ ]
where: Di = dose for case i Xni = the release fraction for the key radionuclide n and case i A,B,C,D, and E are constants.
These equations were setup as a series of simultaneous equations and the constants varied until an optimal solution to all equalities was determined. The correlation was obtained by matching the values generated by the equation to the whole body dose reported in the literature. Figure A.6 presents the correlation for the 56 cases obtained for the final solution.
- Equation Value a Whole body dose 1.OOE+09 I.OOE+08, o- i _1 I.OOE+07_
6 ~~ E ~ ~ ~h 4m E ft_ ^
I.OOE+05 I.OOE+D4 1 11 I1 I 1II1 1M ...........
1 6 11 16 21 26 31 36 4146 51 Dose Case Figure 6 Comparison of Equation Results and Reported Dose Values RSC 01-44NP Printed 08/06/2001 RSC 04-02NP BAIl Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology The factors used to serve as constants that provide the best solution are presented in Table l.
Table 1 Release Split Fraction to Dose Conversion Factors
'Constant -' dioiucide Grop, 'aue a -
A Noble gases B
C D
E 3.4 APPLICATION WITH MAAP The MAAP code provides radionuclide release fractions for significant radionuclides given a failure of containment. The release fractions can be used along with the method presented in this document to estimate the person-rem release.
In order to perform the calculation it is necessary to define what radionculide categories, as defined by MAAP, are needed. Table 2 lists the radionuclide categories utilized and how these radionuclides are mapped to the variables in the methodology.
Table 2 Mapping of Method Variables to MAAP Output Variables
'iqatio'n Variabe MAA Output'Variables:'-
Xi Noble gas X2 X3 [*]
X4 X5 Several of the surveyed PSAs utilized MAAP results to define the release category source term and the correlation has shown to be applicable if these MAAP variables are utilized.
RSC 01-44NP Printed 08/06/2001 RSC 04-02NP B. 12 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology 3.5 QUALITATIVE UNCERTAINTY ASSESSMENT The objective of this activity is to develop a realistic tool for estimation of person-rem. The process must not introduce excessive or unpredictable uncertainty. Two aspects of uncertainty that impact the analysis are the uncertainty in the generated magnitude and the consistency of the overall predictions.
3.5.1 Oualitative Evaluation of Predictive Dose In addition to choosing the best fit for the 56 cases, the variation of the result for each unique case was examined. Figure A.7 plots the variation from the reported value for each of cases.
The range represents a deviation of a factor of two (2) in either direction.
INX 0.20 aK 0.60
- 0.40 . . .
020' a:0.00 - tW W WWl l1 11 11 11 1;11111 111 I ;;llI ! ;;tla I11 ttt tll 1 1
-0.2D
-0.40
-0.60' O
.t.00 Figure 7 Variation of Equation to Reported Dose As shown, most calculated values do not vary from the reported value by more than 50%. Given that the most likely use of this evaluation is to perform an assessment of relative change and that large uncertainties are already present in the PSA, errors of this magnitude (less than a factor of
- 2) are not significant.
The equation, however, was found to significantly over predict dose for cases involving intact containment leakage rates. In these cases, the offsite dose was less than 1.OE+5 person-rem and the variation approached a factor of 50. Thus, the equation may not be appropriate for intact containment cases. The cause of this error is the noble gas contribution. A basic assumption for impaired containment cases is that essentially 100% of noble gases are released such that the noble gas release is essentially a baseline dose as stated earlier. This is not the case for intact containments and the constant chosen for the noble gas contribution is significantly overestimated. This limitation, however, does not affect the use of this model since any assessment would be based on results for impaired containment events. Existing licensing basis analyses can cover intact containment doses and it is this data that is the support for the intact containment release category.
RSC OI-44NP Printed 08/06/2001 RSC 04-02NP B.13 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology 3.5.2 Results Predictability To have confidence in the method it is necessary for the analysis to be internally consistent. This does not preclude generating conservative or non-conservative results. It does require that the results generated are not bimodal resulting in significant differences in the trend of the results.
For example, if one release category is underestimated and another overestimated the importance of the two release categories will be incorrect. If both are slightly overestimated the relative importance will be maintained.
An evaluation of the results (see Figures A.6 and A.7) indicates that the model consisting estimates a value slightly greater than the reference value. For intact containment cases, however, this was not the case. The value was significantly overestimated and again this supports not using this approach for intact containment cases. Figure A.7 also shows several cases when the values were slightly under predicted. This was a single plant with an older evaluation of source term not representative of the current state of knowledge and the underestimation is appropriate and more representative of expected source term. Again the analysis is internally consistent. The method is consistent to provide predictable results and the uncertainty from this aspect is small.
4.0
SUMMARY
AND CONCLUSIONS A simplified model for addressing offsite risk is possible using existing PSA information and can be based on relatively few radionuclides. The development of this model can provide a useful tool to evaluate potential plant configuration changes and improvements.
The use of this model to calculate the impact of proposed changes can be used to assess the impact of procedural changes, operating status, or other modifications on a relative change in whole body person-rem.
It is important to mention that person-rem is only one of the factors that should be considered and that it is not usually the most restrictive when evaluating total risk. The lost plant investment and replacement power costs must also be considered internally in the decision process. The use of a health risk measure such as person-rem, however, does provide a type of regulatory perspective on potential changes in plant status or configuration.
5.0 REFERENCES
- 1. Summitt, R., Development of a SurrogateRisk MeasureforRisk Benefit Assessment, PSA 96, September 29-October 3, 1996.
- 2. Benjamin ct al., Evaluation of Severe Accident Risks and the Potential for Risk Reduction: Seauovah Power Station. Unit 1, United State Regulatory Commission, NUREG/CR-4551 Vol. 2, in press.
- 3. Unpublished PSA.
RSC 01-44NP Printed 08/06/2001 RSC 04-02NP B.14 Printed 05/12/2004
Evaluation of Risk Significance of ILRT Extension Surrogate Level 3 Evaluation Methodology
- 4. W. Sugnet et al, Oconee PRA, A Probabilistic Risk Assessment of Oconee Unit 3 The Nuclear Safety Analysis Center and Duke Power Company, NSAC-60 June 1984.
- 5. Garrick et al., Seabrook Station Probabilistic Safety Assessment, Pickard, Lowe, and Garrick, Inc., PLG-0300 December 1983.
RSC 0 1-44NP Printed 08/06/2001 RSC 04-02NP B.15 Printed 05/1212004
Evaluation of Risk Significance of ILRT Extensbn Appendix C:
Estimation of Intact Containment Population Dose RSC 04-02NP Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extensbn C.0 ESTIMATION OF INTACT CONTAINMENT POPULATION DOSE This appendix documents the development of an estimate for whole body population dose given that although the core has melted, the containment is intact with safeguards functioning. This is the typical licensing basis assessment used for judging the acceptability of the containment. Two cases are presented. The first case addresses the existing San Onofre Nuclear Generating Station (SONGS) licensing basis assumptions. A second case provides the same information but utilizes a proposed refined alternative source term methodology employed for SONGS.
C.1 METHODOLOGY The overall methodology utilizes a scaling factor to estimate the person-rem dose for the population within 50-miles of SONGS. The dose rates for the exclusion area boundary (EAB) and the low population zone (LPZ) are used to define a distance scaling factor. This scaling factor is then used to estimate the dose for distances beyond the LPZ up to the 50-mile radius.
An average person-rem dose is predicted assuming a uniform distribution of radionuclides that decreases with increased distance. A uniform distribution of the surrounding population is then combined to calculate the final total dose. The analysis depends on inputs from the licensing basis analysis1 to arrive at the EAB dose rate, LPZ dose rate, LPZ total man-rem dose and population data2 .
C.2 LICENSING BASIS INFORMATION The information contained in References 1 and 2 provide the dose rates for the EAB and LPZ and the population data. The EAB is defined as the circular area within a radius of 576 meters
(-0.36 miles) from the containment (each containment has a similar sized area). The LPZ extends the radius to 3,140 meters (-1.95 miles). The population is collected for a distance (radius) of 50 miles from the SONGS site.
Analysis presented in Reference 1 provides the dose rates listed in Table C.1, below, for the EAB and LPZ which are derived at a time of two hours following the initiation of the release.
Table C.1 Predicted Dose Rates taken from Reference 1
-Location Dose Ratc'(person-rem/hr)@ 2 hr EAB 3.33E-1 LPZ 9.45E-3 In addition, a 30-day exposure of 0.1 134 person-rem is estimated for the LPZ RSC 04-02NP CA1 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Reference 2 provides the population data surrounding SONGS. This data indicates that there are no persons within the EAB and approximately 1,201 persons within the LPZ. The LPZ is defined as an area within a 1.95 mile radius of the plant site. Reference 2 also indicates that the population within a 50-mile radius is approximately 8.32E+06 if the outer areas of San Diego and Los Angeles are included.
C.3 DOSE SCALING FACTOR The calculation of the necessary scaling factor is based on a relationship of dose rate and distance. This is consistent with the Indian Point 3 submittal (Reference 3). The scaling equation is based on a ratio of the LPZ dose to the EAB dose. The equation is shown below:
Y= c (Eq. 1)
Where:
Y = LPZ dose rate X = EAB dose rate dLPZ= distance for LPZ dE" = distance for EAB c = scaling constant This equation assumes that the dose rate is decreasing in a constant manner with distance and is consistent with the Comanche Peak submittal (Reference 4). Solving for the equation yields a value for the scaling constant (C). The input data is listed below in Table C.2.
Table C.2 Calculation Parameters taken from Reference 1 Parameter Val-e its - .
X 3.33E-01 (person-rem/hr)
Y 9.45E-03 (person-rem/hr) dEAB 576 (meters) dLPZ 3140 (meters)
Solving yields a value of 2.1 for the constant, C. The next step involves extrapolation of this information to the 50-mile radius.
RSC 04-02NP C.2 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension C.4 CALCULATION OF POPULATION DOSE To calculate the population dose for the 50-mile radius, Equation 1 is employed using somewhat different parameters. The LPZ total dose is extrapolated to calculate a population dose, it is important to estimate the dose out to 50 miles (although dose rates decrease significantly with distance, the population is much greater as distance increases) in order to account for the total exposed population. Consistent with Reference 4, the value at 25 miles can be used to represent an average dose for the 50-mile distance since it is the mid point between the plant and the 50-mile radius and the dose is assumed to decrease constantly with distance.
The values provided in Table C.3 are used to solve for the dose at 25 mile (Y).
Table C.3 Calculation Parameters taken from Reference 1
-Parameter Value units);.:
X (LPZ) 1.134E-01 (person-rem)
C 2.1 dLPZ 1.95 (miles) [3140 meters]
d25 25 (miles)
Solving Equation 1 yields a value for the person-rem dose of 5.34E-04 person-rem. The value represents an average individual dose.
The next step is to define the effected population. The estimated population is 8.32E+06 persons. However, it is usually assumed that 95% of the population will be evacuated prior to release such that only 5% of the population would be involved. Given a total population estimate of approximately 8.32E+06 persons, this equates to an exposed population of 4.16E+05 persons.
The population dose 4ojp is then developed using this information by employing the following equation:
dpo = ,,d *P (Eq. 2) where:
died is the dose calculated for a single individual (5.34E-04) p is the total effected population (4.1 6E+05)
Solving Equation 2 yields a total population whole body dose of 2.22E1-02 rem.
RSC 04-02NP C.3 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension C.5 ALTERNATIVE SOURCE TERM CALCULATION OF POPULATION DOSE Reference 5 presents a newer approach to estimate the source term and to calculate offsite dose.
This method, referred to as the alternative source term (AST) method includes newer information with regard to source term composition and is currently being evaluated for use at SONGS.
To address this potential change in the basic licensing approach, the calculated data presented in Reference 5 was used to perform a sensitivity analysis of the ILRT analysis to the implementation of this revised source term.
The same approach was employed to define the scaling factor. From this result the LPZ dose was extrapolated to 25 miles and a total population dose calculated. The information taken from Reference 5 is presented in Table C.4.
Table C.4 Calculation Parameters taken from Reference 5 Parameter Value (units)
X 4.309E-1 (person-rem/hr)
Y 1.22E-2 (person-rem/hr) dEAB 576 (meters) dLPZ 3140 (meters)
Using Equation 1 and this data, the value for C is again 2.1. The next step utilized the LPZ dose for 30 days (8.645E-2 person-rem) and Equation 1 to define the 25 mile dose. Substituting the values in to Equation 1 yields a value of 4.06E-4 person-rem. Using Equation 2, the total person dose is estimated to be 1.69E+02 rem.
This supports the general observation that the AST approach would reduce the potential leakage and decrease the potential for large early releases and improve the conclusions related to allowing implementation of the ILRT extension.
C.6 REFERENCES
- 1. Arastu, A., Calculation N-4061-002. Post-LOCA Containment Leakage-CR & Offsite Doses, Southern California Edison, Revision 0, October 2000.
- 2. Southern California Edison, San Onofre Nuclear Generating Station 2&3 FSAR.
Updated, June 2003.
RSC 04-02NP CAs Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- 3. Indian Point 3 Nuclear Power Plant. "Supplemental Information Regarding Proposed Change to Section 6.14 of the Administrative Section of the Technical Specification",
Entergy, IPN-01-007, January 18, 2001.
- 4. Summitt, R., Comanche Peak Steam Electric Station Probabilistic Safety Assessment, Evaluation of Risk Significance of ILRT Extension, RSC, Inc., RSC 01-47/R&R-PN-110, November 2001.
- 5. Schulz, J., Calculation N-6060-002, LOCA Containment Leakage. CR & Offsite Doses - AST, Southern California Edison, Revision 0, November 2003.
RSC 04-02NP C.5 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extenskbn Appendix D:
Sensitivity Study with SONGS Calculation N-6060-002 Data RSC 04-02NP Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extensibn D.0 INTRODUCTION The main report utilizes the current offsite intact containment dose of record contained in the San Onofre Nuclear Generating Station (SONGS) calculation N-4061-002 and SONGS 2 and 3 UFSAR (References 1 and 2 in Appendix C) as the basis and concludes that the additional risk is insignificant. The purpose of this appendix is to provide a sensitivity study using alternative data for the revised offsite intact containment dose in SONGS calculation N-6060-00217 for person-rem exposure at the exclusion area boundary (EAB) and the low population zone (LPZ) to evaluate the risk of extending the Type A integrated leak rate test interval beyond the current 1O years required by 10 CFR 50, Appendix I Option B at the San Onofre Nuclear Generating Station for both unit 2 and unit 3.
D. 1
SUMMARY
OF THE ANALYSIS This appendix is completed in the same manner as the San Onofre Nuclear Generating Station (SONGS) integrated leak rate test (ILRT) extension main report. Information from References 1 through 16 are used in the same manner as the SONGS ILRT extension main report.
The only difference in this sensitivity study is the person-rem exposure data uDntained in Appendix C. For this sensitivity study, information from SONGS calculation N-6060-002 (Reference 17) is used for the person-remr/hr exposure rate at the EAB and the LPZ From Reference 17, the EAB dose rate is given as 0.4309 person-rem/hr and the LPZ dose rate is given as 0.0122 person-rem/hr. In addition, the LPZ dose is given as 0.08645 rem whole body assuming a 30-day exposure to a single individual.
Since the methodology for this appendix is the same as for the SONGS ILRT extension main report, only a summary of the results are provided here. Refer to the main report for step by step details.
D.2
SUMMARY
OF RESULTS/CONCLUSIONS The specific results are summarized in Table D.1 below. The Type A contribution to LERF is defined as the contribution from Class 3b.
RSC 04-02NP D.] iPnted: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table D.1 Summaiy of Risk Impact on Extending Type A ILRT Test Frequency Rismpatfo -as :;Risk Impact for 0o- Risk Impact for 15-(baseline) years (current years requirement)-
otal integrated risk (person-rem/yr) 80.0541 80.0544 80.0546 Type A testing risk (person-rernlyr) 0.0041 0.0045 0.0048
% total risk (Type A / total) 0.0052% 0.0057P0 0.0059S%
Type A LERF (Class 3b) (per year) 3.74E-07 4.1 IE-07 4.30E-07 Changes due to extension from 10 years(curent)
Arisk from current (person-remryr) 1.94E-04
%increase from current (Arisk I total risk) 0.0002%
A LERF from current (per year) 1.87E-08 A CCFP from current 0.104%
Changes due to cxtension from 3 years (baseline)'l A risk from baseline (person-remnyr) - 5.82E-04
% increase from baseline (A risk / total risk) - 0.0007%
A LERF from baseline :
(per year) 5.60E-08 CCFP from baseline 0.312%
RSC 04-02NP D.2 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Based on the analysis and available data the following is stated:
- The person-rem/year increase in risk contribution from extending the ILRT test frequency from the current once-per-ten-year interval to once-per-fifteen years is 0.0002 person-rem/year.
- The risk increase in LERF from extending the ILRT test frequency from the current once-per-i 0-year interval to once-per- 15 years is 1.87B-08/yr.
- The change in conditional containment failure probability (CCFP) from the current once-per-10-year interval to once-per-iS years is 0.104%.
- The change in Type A test frequency from once-per-ten-years to once-per-fifteen-years increases the risk impact on the total integrated plant risk by only 0.0002%. Also, the change in Type A test frequency from the original three-per-ten-years to once-per-fifteen-years increases the risk only 0.0007%. Therefore, the risk impact when compared to other severe accident risks is negligible.
- Reg. Guide 1.174 provides guidance for determining the risk impact of plant-specific changes to the licensing basis. Reg. Guide 1.174 defines very small changes in risk as resulting in increases of CDF below 10-6/yr and increases in LERF below 10' 7/yr. Since the ILRT does not impact CDF, the relevant criterion is LERF. The increase in LERF resulting from a change in the Type A ILRT test interval from a once-per-ten-years to a once per-fifteen-years is 1.87E-08. ince guidance in Reg. Guide 1.174 defines very small changes in LERF as below 10-7/yr, increasing the ILRT interval from 10 to 15 years is therefore considered non-risk significant. In addition, the change in LERF resulting from a change in the Type A ILRT test interval from a three-per-ten-years to a once per-fifteen-years is 5.60E-08/yr, is also below the guidance.
- RG. 1.174 also encourages the use of risk analysis techniques to help ensure and show that the proposed change is consistent with the defense-in-depth philosophy. Consistency with defense-in-depth philosophy is maintained by demonstrating that the balance is preserved among prevention of core damage, prevention of containment failure, and consequence mitigation. The change in conditional containment failure probability was estimated to be 0.104% for the proposed change and 0.312% for the cumulative change of going from a test interval of 3 in 10 years to 1 in 15 years. These changes are small and demonstrate that the defense-in-depth philosophy is maintained.
RSC 04-02NP D.3 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension D.3 DESIGN INPUTS The inputs for this calculation, similar to the information documented in the SONGS ILRT main report, come from the SONGS PRA and level 2 update (References 10 and 16).
The SONGS plant damage states are summarized in Table D.2.
Table D.2 SONGS Plant Damage States Plant Damage RC nesentative Sequence Staite -§(/yr)
PDS I Transient with loss of secondary heat removal 6.B1E-06 Transient with loss of secondary heat removal, and loss of containment spray PDS 2 recirculation 1.71E-07 Transient with loss of secondary heat removal, and loss of containment heat PDS 3 removal 2.66E-07 PDS 4 Transient/SSL with loss of HPSI inrecirculation 1.51E-09 TransientlSSL with loss of HPSI in recirculation, and loss of containment heat PDS 5 removal O.OOE+00 TransientlSSL with loss of HPSI in recirculation, loss of secondary heat PSD 6 removal, and loss of containment spray recirculation 2.18E-07 Transient with loss of HPSI/LPS1 injection and loss of containnent heat PDS 7 removal 1.20E-06 Transient with loss of HPS1/LPSI injection, loss of secondary heat removal PDS 8 and loss of containment heat removal 2.17E-07 PDS 9 Small LOCA with loss of containment spray recirculation 3.65E-10 PDS 10 Small LOCA with loss of containment heat removal 7.38E-07 PDS 11 Small LOCA with loss of secondary heat removal 2.19E-07 PDS 12 Small LOCA with loss of HPSI recirculation 2.43E-07 RSC 04-02NP DA4 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table D.2 (Continued)
SONGS Plant Damage States Plant Damiage -Representative Sequee Frequncy,
- State (fYr)
Small LOCA with loss of HPSI recirculation, and loss of containment spray PDS 13 recirculation 5.32E-06 Small LOCA with loss of HPSI recirculation, and loss of containment heat PDS 14 removal 1.56E-08 Small LOCA with loss of HPSI recirculation and loss of secondary heat PDS 15 removal 4.40E-09 Small LOCA with loss of HPSI recirculation, loss of secondary heat removal, PDS 16 and loss of containment spray recirculation 6.24E-08 Small LOCA with loss of HPSI recirculation, loss of secondary heat removal, PDS 17 and loss of containment heat removal 5.99E-07 PDS 18 Small LOCA with loss of HPSI/LPSI injection 1.92E-07 PDS 19 Large/medium LOCA with loss of core heat removal 1.28E-06 Large/nedium LOCA with loss of core heat removal, and loss of containment PDS 20 spray recirculation 1.08E-09 PDS 21 Large/medium LOCA with loss of core and containment heat removal 4.06E-09 PDS 22 Large/medium LOCA with loss of HPSI recirculation 2.OOE-07 Large/medium LOCA with loss of HPSI recirculation, and loss of PDS 23 containment spray recirculation 1.59E-09 Large/nedium LOCA with loss of HPSI recirculation, and loss of PDS 24 containment heat removal 8.40E-09 PDS 25 Large/nedium LOCA with loss ofHPS1/LPSI injection 1.93E-09 PDS 26 Transient/LOCA with loss of containment isolation and heat removal 1.14E-08 PDS 27 Interfacing system LOCA (ISLOCA) initiating event 5.61E-08 Steam generator tube rupture (SGTR) initiating event, no stuck open relief PDS 28 valve (SORV) 1.48E-08 Steam generator tube rupture (SGTR) initiating event, with stuck open relief PDS 29 valve (SORV) 1.25E-08 RSC 04-02NP D.5 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table D.2 (Continued)
SONGS Plant Damage States PlantlDam e ersetative Sequence Frqnc
'State --
TOTAL I .79E-05 The release category dose information is presented in Table D.3.
Table D.3 Release Category Radionuclide Fraction
'Release Freq'uency
- ees ; ~'. qec : ::': - ',-':' . . . - -- ;'. '-. ; ,, ,r .,a,, ;,
Category, (/r) Noble Gs Iodine', Cesium' Telluriunm Stronti um' IC-1 (S) 1.50E-05 NA2 NA NA NA NA 1.69E+02 3 B' 5.06E-07 6.50E-02 1.OOE-03 1.OOE-03 6.OOE-09 3.001-07 1.98E+05 Ds 1.25E-08 9.74E01 5.50E-02 3.30E-02 7.001E03 2.OOE-06 5.90E+06 G6 1.56E-07 5.48E-01 2.OOE-02 2.00E-02 2.10E-02 6.00E-04 2.93E+06 X.24E-06 1.OOE+OO 4.OOE-03 2.OOE-04 2.001-06 2.05E+06 5.61E-08 9.98E1-01 8.42E-1 8.42E-01 7.70E-02 2.001-03 8.611E+07 9.83E-07 1.OOE+O0 1.20E-01 1.21E-01 4.OOE-05 3.00E-07 1.36E+07
- 1. Contributing fission product groups are discussed in Appendix B.
- 2. Release fractions not necessary for this calculation.
- 3. Intact containment representing design basis leakage (developed in Appendix C).
- 4. Release category B is defined by containment bypassed with less than 0.1% of volatiles released.
- 5. Release category D is defined by containment bypassed with up to 10% of volatiles released.
- 6. Release category G is defined by early or isolation failure, containment failure prior to or at vessel failure with up to 10% of volatiles released.
- 7. Release category L is defined by late containment failure with up to 1% of volatiles released.
S. Release category T is defined by containment bypassed with greater than I0% of volatiles released.
- 9. Release category W is defined by late containment failure with more than 10% of volatiles released.
D.4 CALCULATIONS Following the methodology presented in the main report, the following calculation steps were performed:
- 1. Map the Level 3 release categories into the 8 release classes defined by the EPRI Report (Reference 7).
RSC 04-02NP D.6 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- 2. Calculate the Type A leakage estimate to define the analysis baseline.
- 3. Calculate the Type A leakage estimate to address the current inspection frequency.
- 4. Modify the Type A leakage estimates to address extension of the Type A test interval
- 5. Calculate increase in risk due to extending Type A inspection intervals.
- 6. Estimate the change in LERF due to the Type A testing.
- 7. Estimate the change in conditional containment failure probability due to the Type A testing.
D.5 SUPPORTING CALCULATIONS Step 1: Map the Level 3 release categoriesinto the 8 release classes defined by the EPRP Report EPRI Report TR-104285 defines eight (8) release classes as presented in Table D.4.
Table D.4 Containment Failure Classifications (from Reference 7)
Failure Classification Descriptiointerpretation for Assigning SONGS Release I Containment remains intact with Intact containment bins containment initially isolated 2 Dependent failure modes or common Isolation faults that are related to a loss of cause failures power or other isolation failure mode that is not a direct failure of an isolation component 3 Independent containment isolation Isolation failures identified by Type A testing failures due to Type A related failures 4 Independent containment isolation Isolation failures identified by Type B testing failures due to Type B related failures 5 Independent containment isolation Isolation failures identified by Type C testing failures due to Type C related failures 6 Other penetration failures Other faults not previously identified 7 Induced by severe accident phenomena Early containment failure sequences as a result of hydrogen bum or other early phenomena 8 Bypass Bypass sequence or SGTR Table D.5 presents the SONGS release category mapping for these eight accident classes.
Person-rem per year is the product of the frequency and the person-rem.
RSC 04-02NP D.7 Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension Table D.5 SONGS PRA Release Category Grouping to EPRI Classes (as described in Reference 7)
Class Description Release Category FrequencyPersonFRe': Rem/yr 1 No containment failure IC-1 (S) 1.50E-05 1.69E+02 2.53E-03 2 Large containment None isolation failures 3a Small isolation failures None Not O.OOE+00 (liner breach) addressed 3b Large isolation failures None Not ,.OOE+OO (liner breach) addresse Small isolation failures -- N failure to seal (type B)
Small isolation failures - Nn ,
failure to seal (type C)
Containment isolation 6 failures (dependent failure, G 1.56E-07 2.93E+06 4.57E-01 personnel errors)
Severe accident 7 phenomena induced failure L, W 2.22E-06 1.36E+07 3.01E+01 (early and late) 8 Containment bypass B, D, T 5.75E-07 8.61E+07 4.95E+O1 Total 1.795E-05 8.00502E+Ol
- 1. E represents a probabilistically insignificant value.
Step 2: Calculate the Type A leakage estimate to define the analysis baseline (3 year test interval)
As displayed in Table D.5 the SONGS PRA did not identify any release categories specifically associated with EPRI Classes 3, 4, or S. Therefore each of these classes must be evaluated for applicability to this study.
RSC 04-02NP D.8 Printed: 0511212004
Evaluation of Risk Significance of ILRT Extension Class 3:
Containment failures in this class are due to leaks such as liner breaches that could only be detected by performing a Type A ILRT.
Reference 4 states that a review of experience data finds that Type A testing identified only 4 leakage events of the 144 events identified. Thus about 3% (0.028) of containment leakage events are identified by the ILRT. The remaining events were identified by a local leak rate test (LLRT) (Type B and C testing) and are not included in the analysis. This probability, however, is based on three tests over a 10-year period and not the one per ten-year frequency currently employed at SONGS (References 1 and 2). The probability (0.028) must be adjusted to reflect this difference.
For this estimation, the question on containment isolation was modified consistent with the previously approved methodology (References 3 and 4), to include the probability of a liner breach (due to excessive leakage) at the time of core damage.
Class 3 is divided into two classes using this approach. Class 3a is defined as a small liner breach and Class 3b is defined as a large liner breach.
Calculation of Class 3b probability To calculate the probability that a liner leak will be large (Class 3b), use was made of the data presented in NUREG-1493 (Reference 11). One data set found in NUREG-1493 reviewed 144 ILRTs. The largest reported leak rate from those 144 tests was 21 times the allowable leakage rate (L.). Since 21x 1a does not constitute a large release, no large releases have occurred based on the 144 ILRTs reported in NUREG-1493.
To estimate the failure probability given that no failures have occurred, a conservative estimate is obtained from the 95th percentile of the %2 distribution. This is consistent with the Indian Point 3 (Reference 3) and Crystal River 3 (Reference 5) templates. In statistical theory, the x2 distribution can be used for statistical testing such as goodness-of-fit tests (Reference 12). The x2 distribution is really a family of distributions, which range in shape from that of the exponential to that of the normal distribution.
Each distribution is identified by the degrees of freedom, v. For time-truncated tests (versus failure-truncated tests), an estimate of the probability of a large leak using the x2 distribution can be calculated using the following equation:
p(a) X2(2F+2,a) (Eq. 1)
Where: N is the number of events, F is the number of events (faults) of interest, and a is the percentile distribution (typically assumed to be the 95%/-tile). The result of 2F+2 defines the degree of freedom.
RSC 04-02NP D.9 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension Given that there have been no large leaks F = 0, therefore v =2) in 144 events (N = 144) the value of X2 (2, 0.05) is equal to 5.99. Solving for the 95th percentile estimate of the probability of a large leak yields 0.021 as presented below:
X 2 (2,0.05) 5.99 =
PClass3B 2
- 144 80.021 (Eq 2)
Calculation of Class 3a probability The data presented in NUREG-1493 (Reference 11) is also used to calculate the probability that a liner leak will be small (Class 3a). The data found in NUREG-1493 states that 144 ILRTs were conducted. The data reported that 23 of 144 tests had allowable leak rates in excess of I.Ox1. However, of the 23 events that exceeded the test requirements, only 4 were found by an ILRT, the others were found by Type B and C testing or errors in test alignments.
Therefore, a best estimate for the probability of leakage is -0.03 (4-of-144). However, the Class 3a probability is estimated using the conservative x2 distribution approach described previously.
This is consistent with the approach taken in References 3, 4 and 5.
The x2 distribution is calculated by F=4 (number of small leaks) and N=144 (number of events) which yields a solution as shown below:
2(l 0,0.05) = 18.307 - 0.064 (Eq. 3)
PcgsA 2.e144 2 88-Therefore, the 95th percentile estimate of the probability of a small leak (Class 3a) is calculated as 0.064.
The probability of liner failures must then be multiplied by an appropriate accident frequency to determine the Class 3a and Class 3b frequencies. The IP3 (Reference 3) and CR3 (Reference 5) submittals utilized the entire core damage frequency when developing the contributions for Classes 3a and 3b and then adjusted the Class I contribution.
This is somewhat conservative since it does provide the maximum possible contributions due to the extension of the LLRT testing interval. This approach is maintained for the SONGS analysis, in order to be consistent with the approved methodology.
Therefore the frequency of a Class 3b failure is calculated as:
FREQ.b = PROB. x CDF = 0.021 x 1.79E-05/yr = 3.74E-07/yr (Eq. 4)
Therefore the frequency of a Class 3a failure is calculated as:
FREQ.,,. = PROB..,. x CDF = 0.064 x 1.79E05 = 1.14E-06/yr (Eq. 5)
RSC 0¢-02NP D. 10 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Class 4:
This group consists of all core damage accident accidents for which a failure-to-seal containment isolation failure of Type B test components occurs. By definition, these failures are dependent on Type B testing, and Type A testing will not impact the probability. Therefore this group is not evaluated any further, consistent with the approved methodology.
Class 5:
This group consists of all core damage accident accidents for which a failure-to-seal containment isolation failure of Type C test components occurs. By definition, these failures are dependent on Type C testing, and Type A testing will not impact the probability. Therefore this group is not evaluated any further, consistent with the approved methodology.
Class 6:
The Class 6 group is comprised of isolation faults that occur as a result of the accident sequence progression. The leakage rate is not considered large by the PRA definition and therefore it is placed into Class 6 to represent a small isolation failure and identified in Table D.5 as Class 6.
FREQciass 6 = 1.56E-07/yr Class 1:
Although the frequency of this class is not directly impacted by Type A testing, the PRA did not model Class 3 failures, and the frequency for Class 1 should be reduced by the estimated frequencies in the new Class 3a and Class 3b in order to preserve the total CDF. The revised Class 1 frequency is therefore:
FREQcIlass = FREQcissl- (FREQclass3a + FREQciass3b) (Eq. 6)
FREQCIASl = 1.50E-05/yr-(l1.14E-06/yr + 3.74E-07/yr) = 1.35E-05/yr Class 2:
The SONGS PRA did not identify any contribution to this group above the quantification truncation.
Class 7:
The frequency of Class 7 is the sum of those release categories identified in Table D.5 as Class 7.
FREQcLw7 = 2.22E-06/yr Class 8:
The frequency of Class 8 is the sum of those release categories identified in Table D.5 as Class 8.
FREQcass 8 = 5.75E-07/yr RSC 04-02NP D.11 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table D.6 summarizes the above information by the EPRI defined classes. This table also presents dose exposures calculated using the methodology described in Appendix B For Class 1, 3a and 3b, the person-rem is developed based on the design basis assessment of the intact containment as developed in Appendix C. The Class 3a and 3b doses are represented as lOxLa and 35xL. respectively. Table D.6 also presents the person-rem frequency data determined by multiplying the failure class frequency by the corresponding exposure.
Table D.6 Baseline Risk Profile Class Description - requency ' ;Person-rem ' Person-ein m .Prson-rem
(/yr). (calculated)- -from (/yr)
.-- factors)
No containment failure 1.35E-05 1.69E+022 2.28E-03 Large containment -
isolation failures 3a Small isolation failures 1.14E-06 ;. 1.69E+03 4 l.93E-03 (liner breach) 3b Large isolation failures 374E07 : . - 91E+03 5 2.21E-03 (liner breach) . -
Small isolation failures - -
failure to seal (type B)
Small isolation failures -
failure to seal (type C)
Containment isolation 6 failures (dependent failure, 1.56E-07 2.93E+06 . 4.57E-01 personnel errors)
Severe accident 7 phenomena induced failure 2.22E-06 1.36E+076 3.OIE+01 (early and late) 8 Containment bypass 5.75E-07 8.61E+07 6 4.95E+O1 Total 1.795E-05 8.00541E+0l
- 1. From Table D.3 using the method presented in Appendix B.
- 2. 1xI dose value calculated in Appendix C.
- 3. E represents a probabilistically insignificant value.
- 4. 10 times I.
- 5. 35 times I..
- 6. Maximum dose from contributing release categories.
RSC 04-02NP D. 12 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension The percent risk contribution due to Type A testing is as follows:
%RiskwE =[( Class3a.E + Class3bBASE) I TotaL.,] x 100 (Eq. 7)
Where:
Class3a.3 ME = Class 3a person-rem/year = 1.93E-03 person-rem/year Class3b. = Class 3b person-rem/year = 2.21 E-03 person-rem/year TotalASE = total person-rem year for baseline interval = 8.00541E+01 person-rem/year (Table D.6)
%Risk^,E = [(1.93E-03 + 2.21E-03) / 8.00541E+01] x 100 = 0.0052%
Step 3: Calculate the Type A leakage estimate to address the currentinspection interval The current surveillance testing requirements as proposed in NEI 94-01 (Reference 6) for Type A testing and allowed by 10 CFR 50, Appendix J is at least once per 10 years based on an acceptable performance history (defined as two consecutive periodic Type A tests at least 24 months apart in which the calculated performance leakage was less than 1.0xL).
According to NUREG-1493 (Reference 11), extending the Type A ILRT interval from 3-in-10 years to 1-in-10 years will increase the average time that a leak detectable only if an ILRT goes undetected from 18 to 60 months. Multiplying the testing interval by 0.5 and multiplying by 12 to convert from "years" to "months" calculates the average time for an undetected condition to exist.
Since ILRTs only detect about 3% of leaks (4/144) that are not detected by other local tests, the increase for a 10-yr ILRT interval is the ratio of the average time ibr a failure to detect for the increased ILRT test interval (60 months) to the baseline average time for a failure to detect of 18 months (i.e., 0.03 x 60/18 = 0.10). References 3 and 5 indicate this is a 10% increase in the likelihood of a Type A leak.
Risk Impact due to 10-year test interval Based on the previously approved methodology (References 3 and 4), the increased probability of not detecting excessive leakage due to Type A tests directly impacts.the frequency of the Class 3 sequences. Consistent with Reference 3 and 5 the risk contribution is determined by multiplying the Class 3 accident frequency by the increase in the probability of leakage (1.1 x Class 3 baseline). The results of this calculation are presented in Table D.7 below.
RSC 04-02NP D.13 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table D.7 Risk Profile for Once in Ten Year Testing Class - Description Frequency (/yr) Person-rem Person-rem (Iyr)
I No containment failure2 1.33E-05 1.69E+02 2.25E-03 Large containment isolation --
failures 3a Small isolation failures (liner 1.26E-06 1.69E+03 2.12E-03 breach) 3b Large isolation failures (liner 4.1 IE-07 5.91E+03 2A3E-03 breach) 4 Smallisolationfailures-failure E to seal (type B)-
Small isolation failures - failure E to seal (type Q -::
Containment isolation failures 1.56E-07 2.93E+06 4.57E-01 6 (dependent failure, personnel errors) 7 Severe accident phenomena 2.22E-06 1.36E+07 3.01 E+OI induced failure (early and late) 8 Containment bypass 5.75E-07 8.61E+07 4.95E+Ol Total 1.795E-05 8.00544E+01
- 1. From TableD.6.
- 2. The IPE frequency of Class I has been reduced by the frequency of Class 3a and Class 3b in order to preserve total CDF.
- 3. E represents a probabilistically insignificant value.
Using the same methods as for the baseline, and the data in Table D.7 the percent risk contribution due to Type A testing is as follows:
%Riskio =[(Class3a, 0 + Class3b10 ) / Total.] x 100 (Eq. 8)
Where:
Class3a,. = Class 3a person-rem/year = 2.12E-03 person-rem/year Class3b,. = Class 3b person-rem/year = 2.43 E-03 person-rem/year Total. = total person-rem year for current 10-year interval = 8.00544E+01 person-rem/year (Table D.7)
RSC 04-O2NP D.14 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
%Risk,. = [(2.12E-03 + 2.43E-03) / 8.00544E+01] x 100 = 0.0057%
The percent risk increase (ARisk..) due to a ten-year ILRT over the baseline case is as follows:
A\%Risk,. = [(Total 0 - Totals) / Totale] x 100.0 (Eq. 9)
Where:
TotaLSE = total person-rem/year for baseline interval = 8.00541E+01 person-rem/year (Table D.6)
Totals = total person-rem/year for 10-year interval = 8.00544E+01 person-rem/year (Table D.7)
A%Risk,. = [(8.00544E+01 - 8.00541 E+01) / 8.00541 E+01] x 100.0 = 0.0005%
Step 4: Calculatethe Type A leakage estimate to address extended inspection intervals If the test interval is extended to 1 in 15 years, the average time that a leak detectable only by an ILRT test goes undetected increases to 90 months (0.5 x 15 x 12). For a 15-yr-test interval, the result is the ratio (0.03 x 90/18) of the exposure times. Thus, increasing the ILRT test interval from 10 years to 15 years results in a proportional increase in the overall probability of leakage.
The approach for developing the risk contribution for a 15-year interval is the same as that for the 10-year interval. References 3 and 5 indicate that the increase is a 50% increase from that for the 10-year interval or a 15% increase from the baseline. Different values are provided for the probability of leakage. In addition, the containment leakage used for the 10-year test interval for Class 3 is used in the 15-year interval evaluation (1.15 x Class 3 baseline). The results for this calculation are presented in Table D.8.
RSC 04-02NP D.15 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table D.8 Risk Profile for Once in Fifteen Year Testing Class - Description - requency (/yr)l Personi rem Person-rem (Iyr)
I No containment failure 2 1.33E-05 1.69E+02 2.24E-03 Large containment isolation :;
failures 3a Small isolation failures (liner 1.31E-06 1.69E+03 2.22E-03 abreach) 3b Large isolation failures (liner 4.30E-07 5.91 E+03 2.54E-03 breach)
Small isolation failures - failure -
to seal (typeB) -
Small isolation failures - failure :
to seal (type C)
Containment isolation failures 1.56E-07 2.93E+06 4.57E-01 6 (dependent failure, personnel errors) 7 Severe accident phenomena 2.22E-06 1.36E+07 3.01E+01 induced failure (early and late) 8 Containment bypass 5.75E-07 8.61E+07 4.95E+01 Total 1.795E-05 - 8.00546E+01
- 1. From Table D.6.
- 2. The IPE frequency of Class 1 has been reduced by the frequency of Class 3a and Class 3b in order to preserve total CDF, however, due to the number of reported decimal places, the number appears the same as in Table D.7.
- 3. E represents a probabilistically insignificant value.
Using the same methods as for the baseline, and the data in Table D.10, the percent risk contribution due to Type A testing is as follows:
%Risk,, =[( Class3a., + Class3bs) / Total,] x 100 (Eq. 10)
Where:
Class3a,, = Class 3a person-rem/year = 2.22E-03 person-rem/year Class3bs = Class 3b persor-rem/year= 2.54E-03 person-rem/year Totals = total persor-rem year for 15-year interval = 8.00546E+01 person-rem/year (Table D.8)
%Risk., = [(2.22E-03 + 2.54E-03) / 8.00546E+01] x 100 = 0.0059%
RSC 04-02NP D. 16 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension The percent risk increase (4%Risks) due to a fifteen-year ILRT over the baseline case is as follows:
ARisk,, = [(Totals - Total") / TotaLASE] x 100.0 (Eq. 11)
Where:
TotaL., = total person-rem/year for baseline (3 per 10 years) interval = 8.00541E+01 person-rem/year (Table D.6)
Total, = total person-rem/year for 15-year interval = 8.00546E+01 person-rem/year (Table D.8)
AORisk., = [(8.00546E+01 - 8.00541E+01) / 8.00541E+01] x 100.0 = 0.0007%
Step 5: Calculate increase in risk due to extending Type A inspection intervals Based on the previously approved methodology (References 3 and 5), the percent increase in risk (in terms of person-rem/yr) of these associated specific sequences is computed as follows.
%Risk,.u = [(PER-REM,, -PER-REM..) /PER-REM,.] x 100 (Eq. 12)
Where:
PER-REM,. = person-rem/year of ten years interval (see Table D.7, Classes 1, 3a and 3b) =
6.798E-03 person-rem/yr PER-REMs = person-rem/year of fifteen years interval (see Table D.8, Classes 1, 3a and 3b) =
6.992E-03 person-rem/yr
%Risk,.,, = [(6.992E 6.798E-03) / 6.798E-03] x 100 = 2.85%
The percent increase on the total integrated plant risk for these accident sequences is computed as follows.
%Totalto-is = [(Total, - Total,) / Total,] x 100 (Eq. 13)
Where:
Totalo = total person-rem/year for 10-year interval = 8.00544E+01 person-rem/year (Table D.7)
Totalis = total person-rem/year for 15-year interval = 8.00546E+01 person-rem/year (Table D.8)
% Totalio-1s = [(8.00546E+01 - 8.00544E+01) / 8.00544E+01] x 100 = 0.0002%
Step 6: Calculatethe change in risk in terns oflarge early releasefrequency (LERF)
The risk impact associated with extending the ILRT interval involves the potential that a core damage event that normally would result in only a small radioactive release from containment RSC 04-02NP D. 17 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension could in fict result in a larger release due to failure to detect a pre-existing leak during the relaxation period.
From References 3 and 5, the Class I dose is assumed to be 10 times the allowable intact containment leakage, L (or 1,690 person-rem) and the Class 3b dose is assumed to be 35 times L4 (or 5,910 person-rem). The dose equivalent for allowable leakage (La) is developed in Appendix C. This compares to a historical observed average of twice L9 (Reference 11).
Therefore, the estimate is somewhat conservative.
Based on the previously approved methodology (References 3 and 5), only Class 3 sequences have the potential to result in large releases if a pre-existing leak were present. Class 1 sequences are not considered as potential large release pathways because for these sequences the containment remains intact. Therefore, the containment leak rate is expected to be small (less than 2xT,). A larger leak rate would imply an impaired containment, such as classes 2, 3, 6 and 7.
Late releases are excluded regardless of the size of the leak because late releases are, by definition, not a LERF event. At the same time, sequences in the SONGS PRA (Reference 10) that result in large releases, are not impacted because a LERF will occur regardless of the presence of a pre-existing leak. Therefore, the frequency of Class 3b sequences is used as the increase in LERF for SONGS, and the change in LERF can be determined by the differences.
References 3 and 5 identify that Class 3b is considered to be a contributor to LERF. The assumed dose for this class is compared to other LERF sequences to determine if it truly represents an increase in LERF. In order to be a LERF sequence, it must be both early in time and large in population dose. The first condition is met since the failure represents an existing isolation failure. However, the dose is small compared to other early sequences. Table D.9 compares the doses for this and several other cases.
Table D.9 Comparisons of Release Class Doses Release Class Population Tse,(Personrem Class 3b (Fable D.6) 5,910 Class 8 (Table D.6) 86,100,000 Class 7 (Table D.6) 13,600,000 The table shows that even a conservative estimate for the release (person-rem) is found to less than 1.0 percent of that obtained from other early release classes. On a best-estimate basis the average expected leakage would be less than 338 person-rem (using a similar method for the person-rem leakage developed in Appendices A and C for the main report) and would be less than 1.0 percent of the other classes associated with large early release. The conclusion can be drawn from this data that the potential consequence of a Type A leakage event is not large and the proposed change has no impact on LERF. However, conservatively Class 3b is considered to RSC 04-02NP D. 18 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension be an estimate for the change in LERF to be consistent with the accepted methodology (References 3 and 5). Table D.1O summarizes the results of the LERF evaluation assuming that Type 3b is indicative of a LERF sequence.
Table D.1O Impact on LERF due to Extended Type A Testing Intervals
-ALRTlnspection Interval, .-3 Years (baseline) 10 Years -:55 .ears-Class 3b (Type A LERF) 3.74E-07 4.1 IE-07 4.30E-07 ALERF (3 year baseline) 3.74E-O8 5.60E-08 ALERF (10 year baseline) - I.87E-08 Reg. Guide 1.174 (Reference 8) provides guidance for determining the risk impact of plant-specific changes to the licensing basis. Reg. Guide 1.174 defines very small changes in risk as resulting in increases of core damage frequency (CDF) below L.OE-06/yr and increases in LERF below 1.013-07/yr. Since the ILRT does not impact CDF, the relevant metric is LERF.
Calculating the increase in LERF requires determining the impact of the ILRT interval on the leakage probability.
Since guidance in Reg. Guide 1.174 defines very small changes in LERF as below l.OE-07/yr, increasing the ILRT interval to 15 years (1.87E-08/yr) is non-risk significant. It should be noted that if the risk increase is measured from the original 3-in-I 0-year interval, the increase in LERF is 5.60E-08/yr, which is also below the l.OE-07/yr screening criterion in Reg. Guide 1.174.
Step 7: Calculatethe change in conditionalcontainmentfailureprobability (CCFP)
The CCFP is defined as the probability of containment failure given the occurrence of an accident. This probability can be expressed using the following equation:
CCFP=I[Ff(nf)] (Eq. 14)
Whereftncj) is the frequency of those sequences which result in no containment failure. This frequency is determined by summing the Class 1 and Class 3a results, and CDF is the total frequency of all core damage sequences.
Therefore the change in CCFP for this analysis is the CCFP using the results for 15 years (CCFP.) minus the CCFP using the results for 10 years (CCFP,). This can be expressed by the following:
ACCFj 15 , = CCF 5 -CCFIo0 (Eq. 15)
RSC 04-02NP D. 19 Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension Using the data previously developed the change in CCFP from the current testing interval is calculated and presented in Table D. 11.
Table D.1 1 Impact on Conditional Containment Failure Probability due to Extended Type A Testing Intervals ALRT lnspection Interval 3 Years (baseine) 10 Years 15 Years ftncf) 1.4626E-05 1.4589E-05 1.4570E-05 Ancf)/CDF 0.815 0.813 0.812 CCFP 1.85E-01 1.87E-01 1.88E-01 ACCFP (3 year baseline) 0.208% 0.312%
ACCFP (10 year baseline) - 0.104%
D.6 REFERENCES
- 1. San Onofre Nuclear Generating Station Unit 2. Containment Integrated Leak Rate Test, Rev.-4, Procedure S02-V-3.12.
- 2. San Onofre Nuclear Generating Station Unit 3. Containment Integrated Leak Rate Test, Rev. 4, Procedure S03-V-3.12.
- 3. Indian Point 3 Nuclear Power Plant, "Supplemental Information Regarding Proposed Change to Section 6.14 of the Administrative Section of the Technical Specification",
Entergy, IPN-01-007, January 18,2001.
- 4. Indian Point Nuclear Generating Unit No.3 - Issuance of Amendment Re: Frequency of Performance-Based Leakage Rate Testing (TAC NO. MB0178) United States Nuclear Regulatory Commission(USNRC), April 17,2001.
- 5. Evaluation of Risk Significance of ILRT Extension Revision 2, Florida Power Corporation, F-01-0001 June 2001.
- 6. Industry Guideline for Implementing Performance-Based Option of 10 CFR Part 50 Appendix J3 Revision 0, Nuclear Energy Institute, NEI 94-01, July 26, 1995.
- 7. Gisclon, J. M., et al, Risk Impact Assessment of Revised Containment Leak Rate Testing Intervals, Electric Power Research Institute, TR-104285 August 1994.
- 8. An Approach for Using Probabilistic Risk Assessment in Risk-Informed decisions on Plant-Specific Changes to the 'Licensing Basis, U.S. Nuclear Regulatory Commission (USNRC), Regulatory Guide 1.174. July 1998.
RSC 04-02NP D.20 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- 10. San Onofre Nuclear Generating Station Living PRA. SONGS 2/3. PRA Level II Analysis Report IPE-LEVEL2-000.
- 11. Performance-Based Containment Leak-Test Prom USNRC, NUREG-1493, July 1995.
- 12. PRA Procedures Guide- A Guide to the Performance of Probabilistic Risk Assessments for Nuclear Power Plants, American Nuclear Society and fth National Institute of Electrical and Electronic Engineers, NUREG/CR-2300 January 1983.
- 13. Southern California Edison, San Onofre Nuclear Generating Station 2&3 FSAR, Updated June 2003.
- 14. Summitt, R., Assessment of Safety Benefit for Installation of a Generator Disconnect Switch at Robinson RSC, Inc., RSC 98-19 June 1998.
- 15. Summitt, R., Comanche Peak Steam Electric Station Probabilistic Safety Assessment, Evaluation of Risk Significance of ILRT Extension RSC, Inc., RSC 01-471R&R-PN-110, November 2001.
- 16. San Omfre Nuclear Generation Station WinNUPRA/WinNUCAP Model Update, January 2004.
- 17. Schulz, J., Calculation N-6060-002, LOCA Containment Leakage. CR & Offsite Doses -
AST Southern California Edison, Revision 0, November 2003.
RSC 04-02NP D.21 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Appendix E:
ILRT Sensitivity Study (Inclusion of SONGS External Events)
RSC 04-02NP Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extensbn E.0 INTRODUCTION The purpose of this appendix is to provide a sensitivity study using alternative data for the plant damage state frequencies and the overall core damage frequency due to inclusion of seismic and fire contribution to release frequency to evaluate the risk of extending the Type A integrated leak rate test interval beyond the current 1 0 years required by 10 CFR 50, Appendix J, Option B at the San Onofre Nuclear Generating Station for both unit 2 and unit 3. This represents a significant extension of the risk evaluation not previously addressed in the integrated leak rate test evaluation process.
El1
SUMMARY
OF THE ANALYSIS This appendix is completed in the same manner as the San Onofre Nuclear Generating Station (SONGS) integrated leak rate test (ILRT) extension main report. Information from References 1 through 16 are used in the same manner as the SONGS ILRT extension main report.
The only difference in this sensitivity study is the frequencies of the plant damage state frequencies and the overall core damage frequency. Reference 10 contains the data for the plant damage state frequencies and the overall core damage frequency with the inclusion of seismic and fire events.
Since the methodology for this appendix is the same as for the SONGS ILRT extension main report, only a summary of the results are provided here. Refer to the main report for step by step details.
E.2
SUMMARY
OF RESULTS/CONCLUSIONS The specific results are summarized in Table E. 1 below. The Type A contribution to LERF is defined as the contribution from Class 3b.
RSC 04-02NP El Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension Table E.I Summary of Risk Impact on Extending Type A ILRT Test Frequency Risk mipact for 3-years Risk Inpacft for 5-:
(baseline) 'years (current years
.requirement)
Total integrated risk (person-rem'yr) 174.152 174.153 174.154 Type A testing risk (person-rem/yr) 0.0105 0.0115 0.0121
% total risk (Type A / total) 0.0060% 0.0066% 0.0069%
Type A LERF (Class 3b) (per year) 7.20E-07 7.92E-07 8.28E-07 Changes due to extension from 10 years (current) risk from current (person-rernIyr) 4.92E-04
% increase from current (A risk total risk) 0.0003%
A LERF from current (per year) 3.60E-08 A CCUP from current 0.104%
Changes due to extension from 3 years (baseline) ;
A risk from baseline -
(person-rem/yr) -IA8E-03
% increase from baseline (A risk total risk) - 0.00080/
A LERF from baseline -
(peryear) 1.08E-07 A CCFP from baseline : d 0.312%
RSC 04-02NP E.2 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Based on the analysis and available data the following is stated:
- The person-rem/year increase in risk contribution from extending the ILRT test frequency from the current once-per-ten-year interval to once-per-fifteen years is 0.0005 person-rem/year.
- The risk increase in LERF from extending the ILRT test frequency from the current once-per-10-year interval to once-per-15 years is 3.60E-08/yr.
- The change in conditional containment failure probability (CCFP) from the current once-per-I0-year interval to once-per-15 years is 0.104%.
- The change in Type A test frequency from once-per-ten-years to once-per-fifteen-years increases the risk impact on the total integrated plant risk by only 0.0003%. Also, the change in Type A test frequency from the original three-per-ten-years to once-per-fifteen-years increases the risk only 0.0008%. Therefore, the risk impact when compared to other severe accident risks is negligible.
- Reg. Guide 1.174 provides guidance for determining the risk impact of plant-specific changes to the licensing basis. Reg. Guide 1.174 defines very small changes in risk as resulting in increases of CDF below 10' 6/yr and increases in LERF below 10 7/yr. Since the ELRT does not impact CDF, the relevant criterion is LERF. The increase in LERF resulting from a change in the Type A ILRT test interval from a once-per-ten-years to a once per-fifteen-years is 3.60E-08. Since guidance in Reg. Guide 1.174 defines very small changes in LERF as below 10/7 yr, increasing the ILRT interval from 10 to 15 years is therefore considered non-risk significant. In addition, the change in LERF resulting from a change in the Type A ILRT test interval from a three-per-ten-years to a once per-fifteen-years is 1.08E-07/yr, is only slightly above the guidance. Given the large uncertainty associated with the seismic and fire events that tend to yield somewhat conservative results, this result is believed to be acceptable.
- R.G. 1.174 also encourages the use of risk analysis techniques to help ensure and show that the proposed change is consistent with the defense-in-depth philosophy. Consistency with defense-in-depth philosophy is maintained by demonstrating that the balance is preserved among prevention of core damage, prevention of containment failure, and consequence mitigation. The change in conditional containment failure probability was estimated to be 0.104% for the proposed change and 0.312% for the cumulative change of going from a test interval of 3 in 10 years to I in 15 years. These changes are small and demonstrate that the defense-in-depth philosophy is maintained.
RSC 04-02NP E.3 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension E.3 DESIGN INPUTS The inputs for this calculation, similar to the information documented in the SONGS ILRT main report, come from the SONGS PRA and level 2 update (References 10 and 16).
The SONGS plant damage states are summarized in Table E.2.
Table E.2 SONGS Plant Damage States Plant Damage Representative Sequ ece requn PDS I Transient with loss of secondary heat removal 2.09E-05 Transient with loss of secondary heat removal, and loss of containment spray PDS 2 recirculation 2.82E-07 Transient with loss of secondary heat removal, and loss of containment heat PDS 3 removal 3.06E-07 PDS 4 TransientlSSL with loss of HPSI in recirculation 1.5 1E-09 Transient/SSL with loss of HPSI inrecirculation, and loss of containment heat PDS 5 removal 0.OOE+00 Transient/SSL with loss of HPSI in recirculation, loss of secondary heat PSD 6 removal, and loss of containment spray recirculation 3.00E-07 Transient with loss of HPSI/LPSI injection and lass of containment heat PDS 7 removal 1.45E-06 Transient with loss of HPSI/LPSl injection, loss of secondary heat removal PDS 8 and loss of containment heat removal 1.61E-06 PDS 9 Small LOCA with loss of containment spray recirculation 3.76E-10 PDS 10 Small LOCA with loss of containment heat removal 8.52E-07 PDS 11 Small LOCA with lass of secondary heat removal 2.22E-07 PDS 12 Small LOCA with loss of HPSI recirculation 2.47E-07 RSC 04-02NP F.4 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table E2 (Continued)
SONGS Plant Damage States Plant Damage Rep resentative Sequence -requency State (fyr)
Small LOCA with loss of HPSI recirculation, and loss of containment spray PDS 13 recirculation 5.62E-06 Small LOCA with loss of HPSI recirculation, and loss of containment heat PDS 14 removal 9.07E-08 Small LOCA with loss of HPSI recirculation and loss of secondary heat PDS 15 removal 2.72E-08 Small LOCA with loss of HPSI recirculation, loss of secondary heat removal, PDS 16 and loss of containment spray recirculation 3.08E-08 Small LOCA with loss of HPSI recirculation, loss of secondary heat removal, PDS 17 and loss of containment heat removal 6.22E-07 PDS 18 Small LOCA with loss of HPSI/LPSI injection 4.1 E-07 PDS 19 Large/mndium LOCA with loss of core heat removal 1.39E-06 Large/medium LOCA with loss of core heat removal, and loss of containment PDS 20 spray recirculation 1.09E-09 PDS 21 Large/redium LOCA with loss of core and containment heat removal 4.07E-09 PDS 22 Large/redium LOCA with loss of HPSI recirculation 2.01E-07 Large/nedium LOCA with loss of HPSI recirculation, and loss of PDS 23 containment spray recirculation 1.59E-09 Laige/mrdium LOCA with loss of HPSI recirculation, and loss of PDS 24 containment heat removal 8.40E-09 PDS 25 Large/nedium LOCA with loss of HPSI/LPSI injection 2.01E-09 PDS 26 Transient/LOCA with loss of containment isolation and heat removal 3.OOE-08 PDS 27 Interfacing system LOCA (ISLOCA) initiating event 5.61E-08 Steam generator tube rupture (SGTR) initiating event, no stuck open relief PDS 28 valve (SORV) 1.48E-08 Steam generator tube rupture (SGTR) initiating event, with stuck open relief PDS 29 valve (SORV) 1.25E-08 RSC 04-02NP E-5 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table E.2 (Continued)
SONGS Plant Damage States
- Plant Damage Representative Sequence State (lyr)
TOTAL 3.46E-05 The release category dose information is presented in Table E.3.
Table E.3 Release Category Radionuclide Fraction Release Frequency -
Category - r N/y oble Gas' Iodine -Cesium' Tellurium' Strontium' Total Dose IC-1 (S) 2.94E-05 NA2 NA NA NA NA 2.22E+02 3 B4 1.40E-06 6.50E-02 I.OOE-03 1.OOE-03 6.OOE-09 3.OOE-07 1.98E+05 D1.25E-08 9.74E-01 5.50E-02 3.30E-02 7.OOE-03 2.OOE-06 5.90E+06 G6 2.79E-07 5.48E-01 2.OOE-02 2.OOE-02 2.1OE-02 6.00E-04 2.93E+06 C 2.08E-06 l.OOE+OO 4.OOE-03 7.OOE-03 2.OOE-04 2.00E-06 2.05E+06 5.61E-08 9.98E-01 8.42E-01 8.42E-01 7.70E-02 2.OOE-03 8.61E+07 1.38E-06 l .OOE+OO 1.20E-01 1.21EOl 4.OOE-05 3.001-07 1.36E+07
- 1. Contributing fission product groups are discussed in Appendix B.
- 2. Release fractions not necessary for this calculation.
- 3. Intact containment representing design basis leakage (developed in Appendix C).
- 4. Release category B is defined by containment bypassed with less than 0.1% of volatiles released.
- 5. Release category D is defined by containment bypassed with up to 10% of volatiles released.
- 6. Release category G is defined by early or isolation failure, containment failure prior to or at vessel failure with up to 10% of volatiles released.
- 7. Release category L is defined by late containment failure with up to 1% of volatiles released.
- 8. Release category T is defined by containment bypassed with greater than 10%h of volatiles released.
- 9. Release category W is defined by late containment failure with more than 10% of volatiles released.
E.4 CALCULATIONS Following the methodology presented in the main report, the following calculation steps were performed:
- 1. Map the Level 3 release categories into the 8 release classes defined by the EPRI Report (Reference 7)
RSC 04-02NP E.6 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- 2. Calculate the Type A leakage estimate to define the analysis baseline
- 3. Calculate the Type A leakage estimate to address the current inspection frequency
- 4. Modify the Type A leakage estimates to address extension of the Type A test interval
- 5. Calculate increase in risk due to extending Type A inspection intervals
- 6. Estimate the change in LERF due to the Type A testing.
- 7. Estimate the change in conditional containment failure probability due to the Type A testing.
R-5 SUPPORTING CALCULATIONS Step 1: Map the Level 3 release categoriesinto the 8 release classes defined by the EPRIReport EPRI Report TR-104285 defines eight (8) release classes as presented in Table E.4.
Table E.4 Containment Failure Classifications (from Reference 7)
Failure Classification Descrption -rpretation I fo Assigl ase a egory l Containment remains intact with Intact containment bins containment initially isolated 2 Dependent failure modes or common Isolation faults that are related to a loss of cause failures power or other isolation failure mode that is not a direct failure of an isolation component 3 Independent containment isolation Isolation failures identified by Type A testing failures due to Type A related failures 4 Independent containment isolation Isolation failures identified by Type B testing failures due to Type B relted failures 5 Independent containment isolation Isolation failures identified by Type C testing failures due to Type C related failures 6 Other penetration failures Other faults not previously identified 7 Induced by severe accident phenomena Early containment failure sequences as a result of hydrogen burn or other early phenomena 8 Bypass Bypass sequence or SGTR Table E.5 presents the SONGS release category mapping for these eight accident classes.
Person-rem per year is the product of the frequency and the person-rem.
RSC 04-02NP E.7 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table ES SONGS PRA Release Category Grouping to EPRI Classes (as described in Reference 7)
Person-Class Descrip.i on Release ategory -Frequency, Person-Rem Rem/yr 1 No containment failure IC-i (S) 2.94E-05 2.22E+02 6.53E-03 2 Large containment None :
isolation failures . . -
3a Small isolation failures None Not O.OOE+OO (liner breach) addressed 3b Large isolation failures None Not . - ..
(liner breach) addressed .O-E+-O Small isolation failures -: N E failure to seal (type B)
Small isolation failures- - None failure to seal (type C)
Containment isolation 6 failures (dependent failure, G 2.79E-07 2.93E+06 8.17E-01 personnel errors)
Severe accident 7 phenomena induced failure L, W 3.46E-06 1.36E+07 4.69E+01 (early and late) 8 Containment bypass B, D, T 1.47E-06 8.61E+07 1.26E+02 Total 3.46E-05 1.74142E+02
- 1. E represents a probaillisuically insignnicant value.
Step 2: Calculate the Type A leakage estimate to define the analysis baseline (3 year test interval)
As displayed in Table E.5 the SONGS PRA did mt identify any release categories specifically associated with EPRI Classes 3, 4, or 5. Therefore each of these classes must be evaluated for applicability to this study.
RSC 04-02NP E.8 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Class 3:
Containment failures in this class are due to leaks such as liner breaches that could only be detected by performing a Type A ILRT.
Reference 4 states that a review of experience data finds that Type A testing identified only 4 leakage events of the 144 events identified. Thus about 3% (0.028) of containment leakage events are identified by the ILRT. The remaining events were identified by a local leak rate test (LLRT) (Type B and C testing) and are not included in the analysis. This probability, however, is based on three tests over a 10-year period and not the one per ten-year frequency currently employed at SONGS (References 1 and 2). The probability (0.028) must be adjusted to reflect this difference.
For this estimation, the question on containment isolation was modified consistent with the previously approved methodology (References 3 and 4), to include the probability of a liner breach (due to excessive leakage) at the time of core damage.
Class 3 is divided into two classes using this approach. Class 3a is defined as a small liner breach and Class 3b is defined as a large liner breach.
Calculation of Class 3b probability To calculate the probability that a liner leak will be large (Class 3b), use was made of the data presented in NUREG-1493 (Reference 11). One data set found in NUREG-1493 reviewed 144 ILRTs. The largest reported leak rate from those 144 tests was 21 times the allowable leakage rate (L.). Since 21 xL does not constitute a large release, no large releases have occurred based on the 144 ILRTs reported in NUREG-1493.
To estimate the failure probability given that no failures have occurred, a conservative estimate is obtained from the 95th percentile of the x2 distribution. This is consistent with the Indian Point 3 (Reference 3) and Crystal River 3 (Reference 5) templates. In statistical theory, the x2 distribution can be used for statistical testing such as goodness-of-fit tests (Reference 12). The X2 distribution is really a family of distributions, which range in shape from that of the exponential to that of the normal distribution.
Each distribution is identified by the degrees of freedom, v. For time-truncated tests (versus failure-truncated tests), an estimate of the probability of a large leak using the x2 distribution can be calculated using the following equation:
p(a)- X 2 (2F+2, a) (Eq. 1)
Where: N is the number of events, F is the number of events (faults) of interest, and a is the percentile distribution (typically assumed to be the 95%/-tile). The result of 2F+2 defines the degree of freedom.
RSC 04-02NP E9 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Given that there have been no large leaks F = 0, therefore v =2) in 144 events (N = 144) the value of X2(2, 0.05) is equal to 5.99. Solving for the 95th percentile estimate of the probability of a large leak yields 0.021 as presented below:
X 2 (2,0.05) 5.99 PC1.3B = 2 144 = 288 = 0.021 (Eq. 2)
Calculation of Class 3a probability The data presented in NUREG-1493 (Reference 11) is also used to calculate the probability that a liner leak will be small (Class 3a). The data found in NUREG-1493 states that 144 ILRTs were conducted. The data reported that 23 of 144 tests had allowable leak rates in excess of l.OxL. However, of the 23 events that exceeded the test requirements, only 4 were found by an ILRT, the others were found by Type B and C testing or errors in test alignments.
Therefore, a best estimate for the probability of leakage is -0.03 (4-of-144). However, the Class 3a probability is estimated using the conservative x2 distribution approach described previously.
This is consistent with the approach taken in References 3, 4 and 5.
The x2 distribution is calculated by F=4 (number of small leaks) and N=144 (number of events) which yields a solution as shown below:
PC1.3A 2 (10,0.05) =18.307 = 0.064 (Eq. 3) 2.144 288 Therefore, the 95th percentile estimate of the probability of a small leak (Class 3a) is calculated as 0.064.
The probability of liner failures must then be multiplied by an appropriate accident frequency to determine the Class 3a and Class 3b frequencies. The IP3 (Reference 3) and CR3 (Reference 5) submittals utilized the entire core damage frequency when developing the contributions for Classes 3a and 3b and then adjusted the Class 1 contribution.
This is somewhat conservative since it does provide the maximum possible contributions due to the extension of the ILRT testing interval. This approach is maintained for the SONGS analysis, in order to be consistent with the approved methodology.
Therefore the frequency of a Class 3b failure is calculated as:
FREQnb - PROBE x CDF = 0.021 x 3.46E-05/yr = 7.20E-07/yr (Eq. 4)
Therefore the frequency of a Class 3a failure is calculated as:
FREQ.,..= PROB...xCDF = 0.064 x3.46E-05 = 2.20E-06/yr (Eq. 5)
RSC 04-02NP E.10 Prnted: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Class 4:
This group consists of all core damage accident accidents for which a failure-to-seal containment isolation failure of Type B test components occurs. By definition, these failures are dependent on Type B testing, and Type A testing will not impact the probability. Therefore this group is not evaluated any further, consistent with the approved methodology.
Class 5:
This group consists of all core damage accident accidents for which a failure-to-seal containment isolation failure of Type C test components occurs. By definition, these failures are dependent on Type C testing, and Type A testing will not impact the probability. Therefore this group is not evaluated any further, consistent with the approved methodology.
Class 6:
The Class 6 group is comprised of isolation faults that occur as a result of the accident sequence progression. The leakage rate is not considered large by the PRA definition and therefore it is placed into Class 6 to represent a small isolation failure and identified in Table E.5 as Class 6.
FREQcas, 6 = 2.79E-07/yr Class 1:
Although the frequency of this class is not directly impacted by Type A testing, the PRA did not model Class 3 failures, and the frequency for Class 1 should be reduced by the estimated frequencies in the new Class 3a and Class 3b in order to preserve the total CDF. The revised Class 1 frequency is therefore:
FREQcjss = FREQciassi - (FREQclass3a + FREQclass3b) (Eq. 6)
FREQcLas.i = 3.46E-05/yr - (2.20E-06/yr +.7.20E-07/yr) = 2.65E-05/yr Class 2:
The SONGS PRA did not identify any contribution to this group above the quantification truncation.
Class 7:
The frequency of Class 7 is the sum of those release categories identified in Table E.5 as Class 7.
FREQc,, 7 = 3.46E-06fyr Class 8:
The frequency of Class 8 is the sum of those release categories identified in Table E.5 as Class 8.
FREQCLSS = 1.47E-06/yr RSC 04-02NP Ell Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension Table E.6 summarizes the above information by the EPRI defined classes. This table also presents dose exposures calculated using the methodology described in Appendix B For Class 1, 3a and 3b, the person-rem is developed based on the design basis assessment of the intact containment as developed in Appendix C. The Class 3a and 3b doses are represented as lOxb1 and 35xLg respectively. Table E.6 also presents the person-rem frequency data determined by multiplying the failure class frequency by the corresponding exposure.
Table E.6 Baseline Risk Profile Class Desc iption - Frequency-
_(lyr), ', Person'-rem alcuated)
Person.rn'c-e m
factrs
'Person-rem I No containment failure 2.65E-05 2.22 E+022 5.88E-03 2 Large containment isolation failures 3a Small isolation failures 2.20E-06 2.22E+034 4.89E-03 (liner breach) 3b Large isolation failures 7.20E-07 7.78E+03 5 5.60E-03 (liner breach)
Small isolation failures -
4 failure to seal (type B)
Small isolation failures -
failure to seal (type C)
Containment isolation 6 failures (dependent failure, 2.79E-07 2.93E+06 ', : . 8.17E-01 personnel errors)
Severe accident 7 phenomena induced failure 3.46E-06 1.36E+07 4.69E+01 (early and late)
Containment bypass 1.47E-06 8.61E+07 1.26E+02 Total 3.46E-05 - 1.74152E+02
- 1. From Table E3 using the method presented in Appendix B.
- 2. lx4 dose value calculated in Appendix C.
- 3. E represents a probabilistically insignificant value.
- 4. 10 times I..
- 5. 35 times L.
- 6. Maximum dose from contributing release categories.
RSC 04-02NP E 12 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension The percent risk contribution due to Type A testing is as follows:
/oRiskBwsE =[( Class3aBASE + Class3bBAs,) / TotaLA.] x 100 (Eq. 7)
Where:
Class3aaur = Class 3a person-rem/year = 4.89E-03 person-rem/year Class3bwE = Class 3b person-rem/year= 5.60E-03 person-rem/year TotaLsE = total person-ren year for baseline interval = 1.74152E+02 person-rem/year (Table E.6)
%Risk.,E = [(4.89E-03 + 5.60E-03) / 1.74152E+02] x 100 = 0.0060%
Step 3: Calculatethe Type A leakage estimate to address the current inspection interval The current surveillance testing requirements as proposed in NEI 94-01 (Reference 6) for Type A testing and allowed by 10 CFR 50, Appendix J is at least once per 10 years based on an acceptable performance history (defined as two consecutive periodic Type A tests at least 24 months apart in which the calculated performance leakage was less than l.OxL 3 ).
According to NUREG-1493 (Reference 11), extending the Type A ILRT interval from 3-in-10 years to 1-in-10 years will increase the average time that a leak detectable only if an ILRT goes undetected from 18 to 60 months. Multiplying the testing interval by 0.5 and multiplying by 12 to convert from 'years" to "months" calculates the average time for an undetected condition to exist.
Since ILRTs only detect about 3% of leaks (4/144) that are not detected by other local tests, the increase for a 10-yr ILRT interval is the ratio of the average time for a failure to detect for the increased ILRT test interval (60 months) to the baseline average time for a failure to detect of 18 months (i.e., 0.03 x 60/18 = 0.10). References 3 and 5 indicate this is a 10% increase in the likelihood of a Type A leak.
Risk Impact due to 10-year test interval Based on the previously approved methodology (References 3 and 4), the increased probability of not detecting excessive leakage due to Type A tests directly impacts the frequency of the Class 3 sequences. Consistent with Reference 3 and 5 the risk contribution is determined by multiplying the Class 3 accident frequency by the increase in the probability of leakage (1.1 x Class 3 baseline). The results of this calculation are presented in Table E.7 below.
RSC 04-02NP E13 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table E.7 Risk Profile for Once in Ten Year Testing
'Class Description -Frequency(iyr)- "Person-ret (lyr)
I No containment failure 2 2.62E-05 2.22E+02 5.82E-03 2 Large containment isolation failures 3a Small isolation failures (liner 2.42E-06 2.22E+03 5.38E-03 breach) 3b Large isolation failures (liner 7.92E-07 7.78E+03 6.16E-03 breach) 4 Small isolation failures - failure E to seal (type B)
Small isolation failures - failure to seal (typeC)
Containment isolation failures 2.79E-07 2.93E+06 8.17E-01 6 (dependent failure, personnel errors)
Severe accident phenomena 3A6E-06 1.36E+07 4.69E+01 induced failure (early and late) 8 Containment bypass 1.47E-06 8.61E+07 1.26E+02 Total 3.46E-05 1.74153E+02 1 * *U*At *Tohl.. t..U A. rsua-. a Vz; &u. v
- 2. The IPE frequency of Class I has been reduced by the frequency of Class 3a and Class 3b in order to preserve total CDF.
- 3. £ represents a probabilistically insignificant value.
Using the same methods as for the baseline, and the data in Table E7 the percent risk contribution due to Type A testing is as follows:
%Riskio =[(Class3a, 0 + Class3b, 0 ) /Total.] x 100 (Eq. 8)
Where:
Class3aU. = Class 3a person-remlyear = 5.38E-03 person-rem/year Class3b,. = Class 3b person-rem/year = 6.16E-03 person-rem/year Total. = total person-rem year for current 10-year interval = 1.74153E+02 person-rem/year (Table El7)
RSC 04-02NP E.14 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension
%Risk,. = [(5.38E-03 + 6.16E-03) / 1.74153E+02] x 100 = 0.0066%
The percent risk increase (AORisk10 ) due to a ten-year ILRT over the baseline case is as follows:
ARisk10 = [(Total. - TotaLt.) / Tota1>s 0 ] x 100.0 (Eq. 9)
Where:
TotalsE = total person-rem/year for baseline interval = 1.74152E+02 person-rem/year (Table E16)
Total. = total person-rem/year for I 0-year interval = 1.74153E+02 person-rem/year (Table E.7)
A%Risk10 = [(1.74153E+02 - 1.74152E+02) / 1.74152E+02] x 100.0 = 0.0006%
Step 4: Calculate the Type A leakage estimate to address extended inspection intervals If the test interval is extended to 1 in 15 years, the average time that a leak detectable only by an ILRT test goes undetected increases to 90 months (0.5 x 15 x 12). For a 15-yr-test interval, the result is the ratio (0.03 x 90/18) of the exposure times. Thus, increasing the ILRT test interval from 10 years to 15 years results in a proportional increase in the overall probability of leakage.
The approach for developing the risk contribution for a 15-year interval is the same as that for the 10-year interval. References 3 and 5 indicate that the increase is a 50% increase from that for the 10-year interval or a 15% increase from the baseline. Different values are provided for the probability of leakage. In addition, the containment leakage used for the 10-year test interval for Class 3 is used in the 15-year interval evaluation (1.15 x Class 3 baseline). The results for this calculation are presented in Table E.8.
RSC 04-02NP E15 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table E.8 Risk Profile for Once in Fifteen Year Testing Class Description Frequency (yr) Person-rem ' Person-rei (/yr)-
_ No containment failure 2 2.60E-05 2.22E+02 5.79E-03 2 Large containment isolation a3 failures 3a Small isolation failures (liner 2.53E-06 2.22E+03 5.62E-03 3a breach) 3b Large isolation failures (liner 8.28E-07 7.78E+03 6.44E-03 breach)
Small isolation failures -failure E to seal (type B)
Small isolation failures - failure E to seal (type C)
Containment isolation failures 2.79E-07 2.93E+06 8.17E-01 6 (dependent failure, personnel errors) 7 Severe accident phenomena 3.46E-06 1.36E+07 4.69E+01 induced failure (early and late) 8 Containment bypass 1.47E-06 8.61E+07 I.26E+02 Total 3.46E-05 1.74154E+02
- 1. From Table E.6.
- 2. The IPE frequency of Class I has been reduced by the frequency of Class 3a and Class 3b in order to preserve total CDF.
- 3. a represents a probabilistically insignificant value.
Using the same methods as for the baseline, and the data in Table E.10, the percent risk contribution due to Type A testing is as follows:
%Risk,, =[( Class3a,, + Class3b,,) / TotaL 5 ] x 100 (Eq. 10)
)Where:
Class3a15 = Class 3a person-rem/year = 5.62E-03 person-rem/year Class3b,, = Class 3b person-rem/year= 6.44E-03 person-rem/year Total, = total person-rem year for 15-year interval = 1.74154E+02 person-rem/year (Table E.8)
%Risk,, = [(5.62E-03 + 6.44E-03) / 1.74154E+02] x 100 = 0.0069%
RSC 04-02NP E.16 Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension The percent risk increase (ARisk,,) due to a fifteen-year ILRT over the baseline case is as follows:
ARisk,, = [(Total, - TotaltME) / Tota.ASE] x 100.0 (Eq. 11)
Where:
TotaIJME = total person-rem/year for baseline (3 per 10 years) interval = 1.74152E+02 person-rem/year (Table E.6)
Total, = total person-rem/year for 15-year interval = 1.74154E+02 person-rem/year (Table E.8)
ARisk5 , = [(1.74154E+02 - 1.74152E+02) / 1.74152E+02] x 100.0 = 0.0008%
Step 5: Calculateincrease in risk due to extending Type A inspection intervals Based on the previously approved methodology (References 3 and 5), the percent increase in risk (in terms of person-rem/yr) of these associated specific sequences is computed as follows.
%Risk,. ,, = [(PER-REM,, - PER-REM,.) / PER-REM,.] x 100 (Eq. 12)
Where:
PER-REM,. = person-rem/year of ten years interval (see Table E.7, Classes 1, 3a and 3b) =
1.736E-02 person-rem/yr PER-REM,, = person-rem/year of fifteen years interval (see Table E.8, Classes 1, 3a and 3b) =
1.785E-02 person-rem/yr
%Risk,.,, = [(1.785E 1.736E-02) / 1.736E-02] x 100 = 2.83%
The percent increase on the total integrated plant risk for these accident sequences is computed as follows.
%Totaho-1s = [(Total, - Total.) / Total.] x 100 (Eq. 13)
Where:
Totalo = total person-rem/year for 10-year interval = 1.74153E+02 person-rem/year (Table E.7)
TotaI5 = total person-rem/year for 15-year interval = 1.74154E+02 person-rem/year (Table E.8)
% Totaho-is = [(1.74154E+02 - 1.74153E+02)/ 1.74153E+02] x 100 = 0.0003%
Step 6: Calculate the change in risk in terms of large early releasefrequency (LERF)
The risk impact associated with extending the ILRT interval involves the potential that a core damage event that normally would result in only a small radioactive release from containment RSC 04-f02NP E17 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension could in fict result in a larger release due to failure to detect a pre-existing leak during the relaxation period.
From References 3 and 5, the Class 3I dose is assumed to be 10 times the allowable intact containment leakage, L (or 2,220 person-rem) and the Class 3b dose is assumed to be 35 times L (or 7,780 person-rem). The dose equivalent for allowable leakage (Lo) is developed in Appendix C. This compares to a historical observed average of twice La (Reference 11).
Therefore, the estimate is somewhat conservative.
Based on the previously approved methodology (References 3 and 5), only Class 3 sequences have the potential to result in large releases if a pre-existing leak were present. Class 1 sequences are not considered as potential large release pathways because for these sequences the containment remains intact. Therefore, the containment leak rate is expected to be small (less than 2xLe). A larger leak rate would imply an impaired containment, such as classes 2, 3, 6 and 7.
Late releases are excluded regardless of the size of the leak because late releases are, by definition, not a LERF event. At the same time, sequences in the SONGS PRA (Reference 10) that result in large releases, are not impacted because a LERF will occur regardless of the presence of a pre-existing leak. Therefore, the frequency of Class 3b sequences is used as the increase in LERF for SONGS, and the change in LERF can be determined by the differences.
References 3 and 5 identify that Class 3b is considered to be a contributor to LERF. The assumed dose for this class is compared to other LERF sequences to determine if it truly represents an increase in LERF. In order to be a LERF sequence, it must be both early in time and large in population dose. The first condition is met since the failure represents an existing isolation failure. However, the dose is small compared to other early sequences. Table E.9 compares the doses for this and several other cases.
Table E.9 Comparisons of Release Class Doses
.Release"Class.' . .Pop lation Mse (Person-rem)
Class 3b (Table E6) 7,780 Class 8 (Table E.6) 86,100,000 Class 7 (Table E.6) 13,600,000 The table shows that even a conservative estimate for the release (person-rem) is found to be less than 1.0 percent of that obtained from the other early release classes. On a best-estimate basis the average expected leakage would be less than 444 person-rem (using a similar method for the person-rem leakage developed in Appendices A and C for the main report) and would be less than 1.0 percent of the other classes associated with large early release. The conclusion can be drawn from this data that the potential consequence of a Type A leakage event is not large and the proposed change has no impact on LERF. However, conservatively Class 3b is considered to RSC 04-02NP E18 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension be an estimate for the change in LERF to be consistent with the accepted methodology (References 3 and 5). Table E.1O summarizes the results of the LERF evaluation assuming that Type 3b is indicative of a LERF sequence.
Table E.lO Impact on LERF due to Extended Type A Testing Intervals lLRT nspectionlnterval -3 Year aseline)
'1 10 Yars Yars Class 3b (Type A LERF) 7.2OE-07 7.92E-07 8.28E-07 ALERF (3 year baseline) -7.20E-08 .08E-07 ALERF (10 year baseline) 3.60E-08 Reg. Guide 1.174 (Reference 8) provides guidance for determining the risk impact of plant-specific changes to the licensing basis. Reg. Guide 1.174 defines very small changes in risk as resulting in increases of core damage frequency (CDF) below 1.OES06/yr and increases in LERF below 1.OE-07/yr. Since the ILRT does not impact CDF, the relevant metric is LERF.
Calculating the increase in LERF requires determining the impact of the ILRT interval on the leakage probability.
Since guidance in Reg. Guide 1.174 defines very small changes in LERF as below l.OE-07/yr, increasing the ILRT interval to 15 years (3.60E-08/yr) is non-risk significant. It should be noted that if the risk increase is measured from the original 3-in-10-year interval, the increase in LERF is 1.08E-07/yr, which is just above the l.OE-07/yr screening criterion in Reg. Guide 1.174.
Given the large uncertainty associated with the seismic and fire events that tend to yield somewhat conservative results, this result is believed to be acceptable.
This result neglects the fact that the overall dose for any other case involving an impaired containment would be bounded by the existing dose rate such that the predicted Type A dose would be inconsequential. This is supported when the predicted doses are compared in Table E.9. As Table E.9 indicates, only the intact containment case would experience an increase in consequence due to Type A leakage.
If only the intact containment data is used in the analysis, rather than the total CDF, increasing the ILRT interval from the current interval to 15 years 3.06E-08/yr) is non-risk significant.
Also, if the risk increase is measured from the original 3-in-1O-year interval, the increase in LERF is 9.17E-08Iyr, which is below the 1.OSE07/yr screening criterion in Reg. Guide 1.174.
Step 7: Calculate the change in conditionalcontainmentfailureprobability(CCFP)
The CCFP is defined as the probability of containment failure given the occurrence of an accident. This probability can be expressed using the following equation:
CCFP= I- f (ni[ ) (Eq. 14)
RSC 04-02NP E19 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Wheref(nci) is the frequency of those sequences which result in no containment failure. This frequency is determined by summing the Class 1 and Class 3a results, and CDF is the total frequency of all core damage sequences.
Therefore the change in CCFP for this analysis is the CCFP using the results for 15 years (CCFP,) minus the CCFP using the results for 10 years (CCFP,0 ). This can be expressed by the following:
ACCFI0-1 5 = CCF, -CCFPI 0 (Eq.15)
Using the data previously developed the change in CCFP from the current testing interval is calculated and presented in Table E. 11.
Table E. 11 Impact on Conditional Containment Failure Probability due to Extended Type A Testing Intervals ILRTolnspecionIterval 3 Years (baseline): 10 as 1:5 Years JAncf) 2.8680E-05 2.8608E-05 2.8572E-05 Xncf)/CDF 0.829 0.827 0.826 CCFP 1.71E-Ol 1.73E-01 1.74E-Ol ACCFP (3 year baseline) 0.208% 0.312%
ACCFP (10 year baseline) 0.104%
E.6 REFERENCES
- 1. San Onofre Nuclear Generating Station Unit 2. Containment Integrated Leak Rate Test Rev. 4, Procedure S02-V-3.12.
- 2. San Onofire Nuclear Generating Station Unit 3. Containment Integrated Leak Rate Test, Rev. 4, Procedure S03-V-3.12.
- 3. Indian Point 3 Nuclear Power Plant. "Supplemental Information Regarding Proposed Change to Section 6.14 of the Administrative Section of the Technical Specification",
Entergy, IPN-01-007, January 18, 2001.
- 4. Indian Point Nuclear Generating Unit No.3 - Issuance of Amendment Re: Frequency of Performance-Based Leakage Rate Testing (TAC NO. MB0178), United States Nuclear Regulatory Commission, April 17, 2001.
- 5. Evaluation of Risk Significance of ILRT Extensior4 Revision 2, Florida Power Corporation, 1-01-0001 June 2001.
RSC 04-02NP E20 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension
- 6. Industry Guideline for Implementing Performance-Based Option of 10 CFR Part 50, Appendix J. Revision 0, Nuclear Energy Institute, NEI 94-0 1, July 26, 1995.
- 7. Gisclon, J. M., et al, Risk Impact Assessment of Revised Containment Leak-Rate Testing Intervals, Electric Power Research Institute, TR- 104285 August 1994.
- 8. An Approach for Using Probabilistic Risk Assessment in Risk-Informed decisions on Plant-Specific Changes to the Licensing Basis, U.S. Nuclear Regulatory Commission (USNRC), Regulatory Guide 1.174 July 1998.
- 10. San Onofre Nuclear Generating Station Living PRA, SONGS 2/3. PRA Level II Analysis Repo IPE-LEVEL2-000.
- 11. Performance-Based Containment Leak-Test Program USNRC, NUREG-1493. July 1995.
- 12. PRA Procedures Guide- A Guide to the Performance of Probabilistic Risk Assessments for Nuclear Power Plants American Nuclear Society and the National Institute of Electrical and Electronic Engineers, NUREG/CR-2300. January 1983.
- 13. Southern California Edison, San Onofre Nuclear Generating Station 2&3 FSAR, Updated, June 2003.
- 14. Summitt, R., Assessment of Safety Benefit for Installation of a Generator Disconnect Switch at Robinson, RSC, Inc., RSC 98-19 June 1998.
- 15. Summitt, R., Comanche Peak Steam Electric Station Probabilistic Safety Assessment.
Evaluation of Risk Significance of ILRT Extension. RSC, Inc., RSC 01-47/R&R-PN-110. November 2001.
- 16. San Onofre Nuclear Generation Station WinNUPRAIWinNUCAP Model Update, January 2004.
RSC 04-02NP E21 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extensbn Appendix F:
Response to USNRC Request for Additional Information for Degradation of the Embedded Side of the Steel Drywell Structure RSC 04-02NP Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension F.0 INTRODUCTION This appendix provides the San Onofre Nuclear Generation Station (SONGS) response to previous United States Nuclear Regulatory Commission (USNRC) request of other licensee (Reference 1) for additional information regarding how potential degradation of the embedded side of the steel drywell containment structure under high pressure during core damage accidents could impact the risk assessment related to the extension of the integrated leak rate test.
F.1 ANALYSIS APPROACH The analysis approach utilizes the Calvert Cliffs Nuclear Plant (CCNP) methodology (Reference 1). This methodology is an acceptable approach to incorporate the liner corrosion issue into the integrated leak rate test (ILRT) extension risk evaluation The results of the analysis, using CCNP methodology, were that increasing the ILRT frequency from three years to fifteen years did not significantly increase plant risk of a large early release.
F.2 ANALYSIS RESULTS Table F.1 summarizes the results obtained from the CCNP methodology (Reference 1) utilizing plant-specific data for SONGS.
Table F.1 SONGS Liner Corrosion Risk Assessment Results Using CCNP Methodology Step :.Description -' Containment Cylinder and Containment Basemat Dolme(85%) ( );
- Historical liner flaw likelihood Events 2 Events: 0 Failure data: containment location specific (Brunswick 2 and North Anna 2) Assume a half failure Success data: based on 70 2 / (70 x 5.5) = 5.19E-03 0.5 / (70 x 5.5) = 1.30E-03 steel-lined containments and 5.5 years since the I OCFR 50.55a requirements of periodic visual inspections of containment surfaces RSC 04-02NP F.1 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension Table F.1 (Continued)
SONGS Liner Corrosion Risk Assessment Results Using CCNP Methodology Step -Description Containment Cylinder and "ContainmentlBaseniat
.D - --. v>.' - -;-Doiei (85%o), - : - -(]5%)
." ,V ;-
2 Aged adjusted liner flaw Year Failure rate Year Failure rate likelihood I 2.05E-03 I 5.13E-04 During the 15-year interval, average 5-10 5.19E-03 average 5-10 1.30E-03 assume failure rate doubles 15 1.43E-02 15 3.57E-03 every five years (14.9%
increase per year). The average for the 5th to 10b year 15 year average = 6.44E-03 15 year average = 1.61 E-03 set to the historical failure rate.
3 Increase in flaw likelihood between 3 and 15 years Uses aged adjusted liner flaw 8.7% 2.2%
likelihood (Step 2), assuming failure rate doubles every five years.
4 Likelihood of breach in Pressure (psia) Likelihood of Pressure (psia) Likelihood of containment given liner flaw liner breach liner breach The upper end pressure is consistent with the current 76 0.10% 76 0.010%
SONGS probabilistic risk 98.6 1.54% 98.6 0.154%
assessment (PRA) level 2 137 16.14% 137 1.614%
analysis (Reference 2). 0.1% 182 57.54% 182 5.754%
is assumed for the lower end. 262 100% 262 10.00%
Intermediate failure likelihoods are determined through logarithmic interpolation as documented in Reference 3. The basemat is assumed to be 1/10 of the cylinder/dome analysis.
RSC 04-02NP F.2 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table F.1 (Continued)
SONGS Liner Corrosion Risk Assessment Results Using CCNP Methodology Step Description Containment Cylinder and Containment I3asernat Donie(85%) (15%)
5 Visual inspection detection 10O% 100%
failure likelihood 5% failure to identify visual flaws Cannot be visually inspected plus 5% likelihood that the flaw is not visible (not through-cylinder but could be detected by ILRT)
All events have been detected through visual inspection. 5%
visible failure detection is a conservative assumption.
6 Likelihood of non-detected 0.0133% 0.00339%
containment leakage (Steps 3 x4 x 5) 8.7% x 1.54% x 10%/a 2.2% x 0.154% x 100%
The total likelihood of the corrosion-induced, non-detected containment leakage is the sum of Step 6 for containment cylinder and dome and the containment basemat.
Total likelihood of non-detected containment leakage = 0.0133% + 0.00339% = 0.01678%
The non-large early release frequency (LERF) containment over-pressurization failures for SONGS are estimated to be 2.22E-06 per year. This is the sum of the frequencies for the release categories L and W in Table 3 of the main report. This value does not hclude those SONGS level 2 release categories where containment sprays are available. If all non-detectable containment leakage events where containment sprays are not available are considered to be LERF, then the increase in LERF associated with the liner corrosion issue is:
Increase in LERF (ILRT 3 to 15 years) = 0.01678% x 2.2213-06 = 3.73E-10 per year Table F.2 shows the changes with the corrosion-induced, non-detected containment leakage for the ILRT extension from the current test interval of 10 years to the proposed test interval of 15 years.
RSC 04-02NP F.3 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Table F.2 Changes Due to Extension from 10 Years (current) to 15 Years Method 'LE RF Increase' Peso,-r'my Percentage Increase in
,Incrase Person -remyr:,
NRC approved 1.87E-081 1.45E-04 2 1.45E-04 / 80.0566' =
method (SONGS 0.0001817%
submittal basis)
NRC approved l.90E-08" 3 1.48E-04" 2 1.48E-04 / 80.05661 method with liner 0.0001846%
corrosion
- 1. Person-rem and LERF increase taken from main report.
- 2. *Assumes all leaks associated with corrosion are large (which is conservative)
Person-rem = LERF x 7.78E+03 (from Table 8 of the main report)
Table F.3 shows the changes with the corrosion-induced, non-detected containment leakage for the ILRT extension from the baseline test interval of 3 years to the proposed test interval of 15 years.
Table F.3 Changes Due to Extension from 3 Years (baseline) to 15 Years Metlhod - -- LERF Increase Person-remyr,- 'ercentage Increase in' Increase Person-remy-,
NRC approved 5.60E-080 4.36E-041 4.36E-04 / 80.05611 =
method (SONGS 0.0005442%
submittal basis)
NRC approved 5.63E08' 3 4.38E_041,2 4.38E04/80.0561=
method with liner 0.0005471%
corrosion
- 1. Person-rem and LERF increase taken from main report.
- 2. Assumes all leaks associated with corrosion are large (which is conservative)
Person-rem = LERF x 7.78E+03.
The results of the analysis, using CCNP methodology, indicate that increasing the ILRT frequency from the current ten years or the baseline three years to fifteen years did not significantly increase plant risk of a large early release. This supports the proposed extension of the ILRT interval.
RSC 04-02NP FA4 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension F.3 REFERENCES
- 1. Letter to NRC from Calvert Cliffs Nuclear Power Plant Unit No. 1: Docket No. 50-317, Response to Request for Additional Information Concerning the License Amendment Request for a One-Time Integrated Leakage Rate Test Extension, dated March 27, 2002.
- 2. San Onofre Nuclear Generating Station Living PRA. SONGS 2/3, PRA Level II Analysis Report IPELEVEL2-000.
- 3. Calculation RSC-CALKNX-2004-0301, Estimation of Containment Overpressure Fragility Curve for SONGS, Revision 0, Ricky Sumrnmitt Consulting, Inc., March 2004.
RSC 04-02NP F.5 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Appendix G:
Evaluation of Relevant SONGS ILRT Experience RSC 04-02NP Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension G.0 INTRODUCTION This appendix documents the relevant historical experience for the San Onofre Nuclear Generating Station integrated leak rate tests performed on both units 2 and 3. From this data, comparisons of the resultant leak size for each integrated leak rate test performed as compared to LI are presented.
G.1 ANALYSIS RESULTS San Onofre Nuclear Generating Station (SONGS) has performed three integrated leak rate tests (ILRT) on unit 2 and unit 3. From the final report for each test, the overall mass point and total time analysis leakage rates (in terms of a percentage of the containment leaked per day) are noted and compared to the leakage rate equivalent to La. Table G.1 presents the rsults of the comparison.
Table G.1 SONGS ILRT Resultant Leak Rates Unit Test Date Mass Point - :.Total Point As-Found L omment-
- Numer ;- - ResultantA- Ru ltant Aag e ake te .
-ound eakaei FounidLeakage. Limt, Rate Rate 21 02/1985 0.064 wtO / day 0.053 wtO / day 0.1 wt% / day Test result leakage is less than Le 22 10/1991 0.0485 wt% / day 0.0713 wt% / day 0.1 wt% / day Test result leakage is less than L.
23 03/1995 0.0427 wt% / day 0.0676 wt% / day 0.1 wt% / day Test result leakage is less 34 11/1985 0.054 wt% / day 0.054 wt% / day 0.1 wt% /day Test result leakage is less than L 35 03/1992 0.0536 wt% / day 0.0681 wt% / day 0.1 wt% / day Test result leakage is less than L, 36 09/1995 0.0546 wt% / day 0.0732 wt% / day 0.1 wt% / day Test result leakage is less thanLa RSC 04-02NP G.1 Printed: 05/1212004
Evaluation of Risk Significance of ILRT Extension
- 1. Information for this test taken from Reference I.
- 2. Information for this test taken from Reference 2.
- 3. Information for this test taken from Reference 3.
- 4. Information for this test taken from Reference 4.
- 5. Information for this test taken from Reference 5.
- 6. Information for this test taken from Reference 6.
From Table G.1, it can be seen that there have been no reported leakage rates equal to or greater than L<. From this, the SONGS data is not atypical (or unique) from the industry data as a whole for ILRT results. This supports the proposed ILRT extension.
G.2 REFERENCES
- 1. San Onofre Nuclear Generating Station. Unit 2. Reactor Containment Building Integrated Leak Rate Test, Final Report, February 1985.
- 2. San Onofre Nuclear Generating Station. Unit 2, Reactor Containment Building Integrated Leak Rate Test, Final Report, October 1991.
- 3. San Onofre Nuclear Generating Station, Unit 2. Reactor Containment Building Integrated Leak Rate Test, Final Report, March 1995.
- 4. San Onofre Nuclear Generating Station. Unit 3, Reactor Containment Building Integrated Leak Rate Test, Final Report, November 1985.
- 5. San Onofre Nuclear Generating Station, Unit 3. Reactor Containment Building Integrated Leak Rate Test. Final Report, March1992.
- 6. San Onofre Nuclear Generating Station, Unit 33.Reactor Containment Building Integrated Leak Rate Test. Final Report, September 1995 RSC 04-02NP G.2 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Appendix H:
SONGS PRA Quality Discussion RSC 04-02NP Printed: 05112/2004
Evaluation of Risk Significance of ILRT Extension H.0 INTRODUCTION The purpose of this appendix is to provide a statement of the quality of the probabilistic risk assessment (PRA) at the San Onofre Nuclear Generating Station (SONGS). The information contained in this appendix was provided by the SONGS PRA staff and incorporated into the report format.
H.1 SONGS PRA QUALITY STATEMENT The SONGS 2/3 PRA has a broad scope characterized as a Level 1 and Level 2, internal and external events, all modes PRA. The external events mainly cover seismic and fire events.
The technical quality/adequacy of the SONGS 2/3 PRA has been assessed through a number of cumulative quality PRA reviews over recent years. Most recently, the SONGS 2/3 PRA was reviewed against the ASME PRA standard (Reference 1). This peer review provides insight into the current quality status of the SONGS 2/3 PRA.
A comprehensive independent peer review of the SONGS 2/3 Level 1 and Level 2 internal events Living PRA for full power and shutdown operations was conducted between August 1996 and April 1997 by an outside consultant (Scientech, Inc.). During this rview, documents, procedures, and supporting calculations and analyses were examined. The review was based primarily on the guidance provided in the PRA procedure guides such as NUREG/CR-2300, "PRA Procedures Guide: A Guide to the Performance of PRAs for Nuclear Power Plants,"
(Reference 2) and NUREG/CR-4550, Volume 1, "Analysis of Core Damage Frequency: Internal Events Methodology," (Reference 3) as well as PRA applications documents such as EPRI TR-105396, "PSA Applications Guide," (Reference 4) and NUREG-1489, "Review of NRC Staff Uses of PRA" (Reference 5). The results of all independent review activities performed by internal and external reviewers were documented in the SCE PRA Change Package process and tracked in the PRA Punch List Database. In June 2003, a pilot application of the ASME PRA Standard peer review process for the SONGS 2/3 Living PRA was performed (Reference 6).
The results of this pilot application are documented in WCAP-16165 Rev. 0 (Reference 1).
The ASME peer review team provided a list of comments (known as 'facts and observations' or F & O's). These F & O's were identified based on a review of the SONGS 2/3 Living PRA against the high level and supporting requirements of the ASME PRA Standard. Each F & 0 was graded based on the type of finding (i.e., technical adequacy or correctness, editorial, suggestion, or complementary).
All F & O's were reviewed for possible impact on the results and conclusions of this application.
Three F & O's were determined to potentially have an impact on the results. Two of these F &
O's were related to the assessment of the ISLOCA initiating event frequency (i.e., IE-C12-02 and IE-C12-03) and the third was related to the B-hour mission time for diesel generators during internal loss of offsite power events (i.e., SY-All-01). The SONGS 2/3 Level 1 and Level 2 model used in this application included the resolution of these three F & O's. The remaining F
& O's were determined to not impact the results or conclusions.
RSC 04-02NP H.1 Printed: 05/12/2004
Evaluation of Risk Significance of ILRT Extension Several measures have also been implemented in the development of the SONGS 2/3 Living PRA to ensure quality. Changes in the model that impact assumptions, success criteria, basic event probabilities, and system and plant models formally undergo several levels of review, and depending on the complexity of the change, may also include peer and/or technical expert panel review.
H.2 REFERENCES
- 1. Pilot Application of ASME PRA Standard Peer Review Process For the San Onofre Nuclear Generating Station Units 2 and 3 PRA, WCAP- 16165. CEOG Task 1037, November 2003.
- 2. PRA Procedures Guide: A Guide to the Performance of PRAs for Nuclear Power Plants NUREG/CR-2300. January 1983.
- 3. Analysis of Core Damage Frequency: Internal Events Methodology NUREG/CR-4550, Volume 1, January 1990.
- 5. Review of NRC Staff Uses of PRA, NUREG-1489, March 1994.
- 6. Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications, ASME RA-S-2002, ASME, April 2002.
RSC 04-02NP H.2 Printed: 05112/2004