NL-16-1390, License Amendment Request to Revise Technical Specification Section 5.5.12 for Permanent Extension of Type a and Type C Leak Rate Test Frequencies, Correction to Attachment 3

License Amendment Request to Revise Technical Specification Section 5.5.12 for Permanent Extension of Type a and Type C Leak Rate Test Frequencies, Correction to Attachment 3

ML16238A477
Person / Time
Site:	Hatch
Issue date:	08/24/2016
From:	Wheat J Southern Nuclear Operating Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
NL-16-1390
Download: ML16238A477 (94)

Text

Justin T. Wheat Southern Nuclear Nuclear Licensing Manager Operating Company, Inc.

40 Inverness Center Parkway Post Office Box 1295 Birmingham, AL 35242

& Southern Nuclear Tel 205.992.5998 Fax 205.992.7601

.f\UG 2 4 20l6 Docket Nos.: 50-321 NL-16-1390 50-366 U. S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, D.C. 20555-0001 Southern Nuclear Operating Company Edwin I. Hatch Nuclear Plant Units 1 and 2; License Amendment Request to Revise Technical Specification Section 5.5.12 for Permanent Extension of Type A and Type C Leak Rate Test Frequencies Correction to Attachment 3 Ladies and Gentlemen:

On July 1, 2016, pursuant to 10 CFR 50.90, Southern Nuclear Operating Company (SNC) requested an amendment to the Edwin I. Hatch Nuclear Plant (HNP) Unit 1, Renewed Facility Operating License (DPR-57), and Unit 2, Renewed Facility Operating License (NPF-5), by proposing changes to the Unit 1 and Unit 2 Technical Specifications (TS). (ML#16188A268). Specifically, the proposed changes would revise TS 5.5.12 "Primary Containment Leakage Rate Testing Program." of the July 1, 2016 Licensing Amendment Request (LAR) provided the evaluation of the proposed change and included attachments with mark-ups and clean copies of the TS pages, mark-ups of TS Bases pages, and the risk assessment supporting the proposed amendment. The risk assessment was provided in Attachment

3. Subsequently, SNC discovered that approximately half of the Attachment 3 risk assessment document was inadvertently cut-off in the electronic version transmitted to the NRC.

The purpose of this letter is to submit the corrected Attachment 3.

This letter contains no NRC commitments. If you have any questions, please contact Ken McElroy at (205) 992-7369.

U.S. Nuclear Regulatory Commission NL-16-1390 Page 2 Mr. Justin T. Wheat states he is the Nuclear Licensing Manager for Southern Nuclear Operating Company, is authorized to execute this oath on behalf of Southern Nuclear Operating Company and, to the best of his knowledge and belief, the facts set forth in this letter are true.

Respectfully submitted,

\. *::*-f ',f  ;:'/;

-;.1' Justin T. Wheat Nuclear Licensing Manager ~

jtw/efb/lac

.. 1 I '. I -J

. tJ- A

~ to and subscf!::_ef:e me this~ day of l'l v.~ , 2016.

C2f-~ £ t::;--

Notary Public .

My commission expires: I 0 - ?- cl-tJ 11 , Plant Hatch Units 1 and 2 Risk Assessment to Support ILRT (Type A)

Interval Extension Request cc: Southern Nuclear Operating Company Mr. S. E. Kuczynski, Chairman, President & CEO Mr. D. G. Bost, Executive Vice President & Chief Nuclear Officer Mr. D. R. Vineyard, Vice President - Hatch Mr. M. D. Meier, Vice President - Regulatory Affairs Mr. B. J. Adams, Vice President - Engineering Mr. 0. M. Scott, Principal Engineer, Risk-Informed Engineering Mr. Mitch Etten-Bohm, Senior Engineer Mr. G. L. Johnson, Regulatory Affairs Manager- Hatch RType: Hatch=CHA02.004 U. S. Nuclear Regulatory Commission Ms. C. Haney, Regional Administrator Mr. M. D. Orenak, NRR Senior Project Manager - Hatch Mr. D. H. Hardage, Senior Resident Inspector - Hatch State of Georgia Mr. J. H. Turner, Director - Environmental Protection Division

ATTACHMENT 3 Plant Hatch Units 1 & 2 Risk Assessment to Support ILRT (Type A) Interval Extension Request

Plant Hatch Units 1 & 2 Risk Assessment to Support ILRT (Type A)

Interval Extension Request

TABLE OF CONTENTS Section Page 1.0 PURPOSE OF ANALYSIS ................................................................................. 1-1 1.1 PURPOSE .......................................................................................... 1-1

1.2 BACKGROUND

................................................................................... 1-1 1.3 ACCEPTANCE CRITERIA ....................................................................... 1-3 2.0 METHODOLOGY ............................................................................................ 2-1 3.0 GROUND RULES ........................................................................................... 3-1 4.0 INPUTS ....................................................................................................... 4-1 4.1 GENERAL RESOURCES AVAILABLE ........................................................ 4-1 4.2 PLANT-SPECIFIC INPUTS ..................................................................... 4-6 4.3 IMPACT OF EXTENSION ON DETECTION OF COMPONENT FAILURES THAT LEAD TO LEAKAGE (SMALL AND LARGE) .............................................. 4-17 4.4 IMPACT OF EXTENSION ON DETECTION OF STEEL CORROSION THAT LEADS TO LEAKAGE .................................................................................... 4-19 5.0 RESULTS ..................................................................................................... 5-1 5.1 STEP 1 - QUANTIFY THE BASE-LINE RISK IN TERMS OF FREQUENCY PER REACTOR YEAR .................................................................................. 5-3 5.2 STEP 2 - DEVELOP PLANT-SPECIFIC PERSON-REM DOSE (POPULATION DOSE) PER REACTOR YEAR .................................................................. 5-9 5.3 STEP 3 - EVALUATE RISK IMPACT OF EXTENDING TYPE A TEST INTERVAL FROM 10-T0-15 YEARS .................................................................... 5-12 5.4 STEP 4 - DETERMINE THE CHANGE IN RISK IN TERMS OF LARGE EARLY RELEASE FREQUENCY ....................................................................... 5-16 5.5 STEP 5 - DETERMINE THE IMPACT ON THE CONDITIONAL CONTAINMENT FAILURE PROBABILITY ...................................................................... 5-17 5.6 STEP 6 - DETERMINE THE IMPACT ON THE POPULATION DOSE RISK ...... 5-18 5.7

SUMMARY

OF INTERNAL EVENTS RESULTS .......................................... 5-18 5.8 EXTERNAL EVENTS CONTRIBUTION .................................... ~ ............... 5-20 6.0 SENSITIVITIES ............................................................................................ 6-1 6.1 SENSITIVITY TO CORROSION IMPACT ASSUMPTIONS ............................. 6-1 6.2 EPRI EXPERT ELICITATION SENSITIVITY ............................................... 6-3 6.3 NON-EARLY RELEASE SENSITIVITY ....................................................... 6-6 6.4 ILRT EXTENSION RISK BENEFIT ........................................................... 6-8

7.0 CONCLUSION

S .............................................................................................. 7-1

8.0 REFERENCES

............................................................................................... 8-1 APPENDIX A NUREG/CR-4551 PEACH BOTTOM POPULATION ESTIMATE ....................... A-1 APPENDIX B Hatch PRA Technical Adequacy Evaluation ............................................... B-1

1.0 PURPOSE OF ANALYSIS 1.1 PURPOSE The purpose of this analysis is to provide an assessment of the risk associated with extending the currently allowed containment Type A integrated leak rate test (ILRT) interval to a permanent fifteen years< 1> for Hatch Units 1 & 2. The extension would allow for substantial cost savings as the ILRT could be deferred for additional scheduled refueling outages. The risk assessment follows the guidelines from NEI 94-01 [1], the methodology used in EPRI TR-104285 [2], the NEI "Interim Guidance for Performing Risk Impact Assessments In Support of One-Time Extensions for Containment Integrated Leakage Rate Test Surveillance Intervals"

[3, 21], the NRC regulatory guidance on the use of Probabilistic Risk Assessment (PRA) as stated in Regulatory Guide 1.200 [28] as applied to ILRT interval extensions, and risk insights in support of a request for a plant's licensing basis as outlined in Regulatory Guide (RG) 1.174

[ 4], the methodology used for Calvert Cliffs to estimate the likelihood and risk implications of corrosion-induced leakage of steel liners going undetected during the extended test interval

[19], and the methodology used in EPRI TR-1009325, Revision 2-A [22] for performing a risk impact assessment of extended integrated leak rate testing intervals. The EPRI TR-1009325 Revision 2-A methodology incorporates the specific limitations and conditions outlined in the NRC acceptance of the EPRI TR-1009325 Revision 2 methodology documented in the NRC Final Safety Evaluation [32]. The format of this document is consistent with the intent of the Risk Impact Assessment Template for evaluating extended integrated leak rate testing intervals provided in Appendix H of the EPRI methodology report [22].

1.2 BACKGROUND

Revisions to 10CFR50, Appendix J (Option B) allow individual plants to extend the Integrated Leak Rate Test (ILRT) Type A surveillance testing frequency requirements from three-in-ten years to at least once in ten years. The revised Type A frequency is 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 limiting containment leakage rate of 1.0La (allowable leakage).

<1 l The ILRT risk assessment is to be used to support a request to a 1 in 15 year ILRT test frequency on a permanent basis. The risk assessment methodology and results equally support a request to extend the ILRT test frequency to 1 in 15 years on a one time basis, as has been performed by many utilities.

1-1

The basis for a 10-year test interval is provided in Section 11.0 of NEI 94-01, Revision 0, and was established in 1995 during development of the performance-based Option B to Appendix J.

Section 11.0 of NEI 94-01 states that NUREG-1493 [5], "Performance-Based Containment Leak Test Program," September 1995, provides the technical basis to support rulemaking to revise leakage rate testing requirements contained in Option B to Appendix J. The basis consisted of qualitative and quantitative assessments of the risk impact (in terms of increased public dose) associated with a range of extended leakage rate test intervals. To supplement the NRC's rulemaking basis, NEI undertook a similar study. The results of that study are documented in Electric Power Research Institute (EPRI) Research Project Report TR-104285 [2].

The NRC report on performance-based leak testing, NUREG-1493, analyzed the effects of containment leakage on the health and safety of the public and the benefits realized from the containment leak rate testing. In that analysis, it was determined that for a representative PWR plant (i.e., Surry) containment isolation failures contribute less than 0.1 percent to the latent risks from reactor accidents. Consequently, it is desirable to show that extending the ILRT interval will not lead to a substantial increase in risk from containment isolation failures for the Hatch plants.

Earlier ILRT frequency extension submittals have used the EPRI TR-104285 [2] methodology to perform the risk assessment. In October 2008, EPRI TR-1018243 [22] was issued to update the generic methodology for ILRT extensions to 15 years using current performance data and to incorporate the specific limitations and conditions outlined by the NRC in the final safety evaluation of the methodology [32]. This more recent EPRI document considers additional risk metrics and criteria including the change in population dose, large early release frequency (LERF), and containment conditional failure probability (CCFP), whereas EPRI TR-104285 considered only the change in population dose.

Hatch requested a one-time extension of the ILRT test frequency from 1 in 10 years to 1 in 15 years for Unit 1 [23] and Unit 2 [24]. The NRC approved the one-time extensions for both Unit 1 [33] and Unit 2 [34].

It should be noted that containment leak-tight integrity is also verified through periodic inservice inspections conducted in accordance with the requirements of the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code Section XI. More specifically, 1-2

Subsection IWE provides the rules and requirements for inservice inspection of Class MC pressure-retaining components and their integral attachments, and of metallic shell and penetration liners of Class CC pressure-retaining components and their integral attachments in light-water cooled plants. Furthermore, NRC regulations 10 CFR 50.55a(b)(2)(ix)(E) require licensees to conduct visual inspections of the accessible areas of the interior of the containment. In addition; Appendix J, Type B local leak tests performed to verify the leak-tight integrity of containment penetration bellows, airlocks, seals, and gaskets are also not affected by the change to the Type A test frequency.

1.3 ACCEPTANCE CRITERIA The acceptance guidelines in RG 1.174 are used to assess the acceptability of this permanent extension of the Type A test interval beyond that established during the Option B rulemaking of Appendix J. RG 1.174 defines very small\ changes in the risk-acceptance guidelines as increases in core damage frequency (CDF) less than 10-6 per reactor year and increases in large early release frequency (LERF) less than 10-7 per reactor year. Because the Type A test does not impact CDF for Hatch, the relevant criterion is the change in LERF. RG 1.174 also defines small changes in LERF as below 10-6 per reactor year provided that the total from all contributors (including external events) can be reasonably shown to be less than 10-5 per reactor year. RG 1.174 discusses defense-in-depth and encourages the use of risk analysis techniques to help ensure and show that key principles, such as the defense-in-depth philosophy, are met. Therefore, the increase in the conditional containment failure probability (CCFP) that helps to ensure that the defense-in-depth philosophy is maintained is also calculated.

Regarding CCFP, changes of up to 1.1 % have been accepted by the NRC for the one-time requests for extension of ILRT intervals. In context, it is noted that a CCFP of 1/10 (10%) has been approved for application to evolutionary light water designs. Given these perspectives, a change in the CCFP of up to 1.5% (percentage point) is assumed to be small. This criterion is articulated in the NRC Final Safety Evaluation Report [32] associated with NEI 94-01 and the EPRI ILRT methodology.

1-3

In addition, the total annual risk (person rem/yr population dose) is examined to demonstrate both the relative change and absolute change in this parameter. Examinations of NUREG-1493 and Safety Evaluation Reports (SER) for one-time interval extensions (summarized in Appendix G of EPRI 1018243 [22]) indicate a range of incremental increases in population dose that have been accepted by the NRcC 1l. The range of incremental population dose increases is from <== 0.01 to 0.2 person-rem/yr and/or 0.002 to 0.46% of the total accident dose. The total doses for the spectrum of all accidents (NUREG-1493 [5], Figure 7-2) result in health effects that are at least two orders of magnitude less than the NRC Safety Goal risk.

Given these perspectives, a very small population dose is defined as an increase from the baseline interval (3 tests per 10 years) dose of < == 1.0 person-rem/yr or 1 % of the total baseline dose, whichever is less restrictive for the risk impact assessment of the proposed extended ILRT interval. This criterion is articulated in the NRC Final Safety Evaluation Report

[32] associated with NE! 94-01 and the EPRI ILRT methodology.

(l) The methodology used in the one-time ILRT interval extension requests assumed a large leak magnitude (EPRI class 3b) of 35La, whereas the methodology in this document uses 100La. The dose risk is impacted by this change and will be larger than those of previous submittals.

1-4

2.0 METHODOLOGY A simplified bounding analysis approach consistent with the latest EPRI approach [22] as accepted by the NRC [32] is used for evaluating the change in risk associated with increasing the test interval to fifteen years. The approach is consistent with that presented in EPRI TR-1018243 [22], NUREG-1493 [5] and the Calvert Cliffs liner corrosion analysis [19]. The analysis uses results from a Level 2 analysis of core damage scenarios from the current Hatch Unit 1 PRA model and the subsequent containment responses for the various fission product release categories (including containment intact release). This risk assessment is applicable to Hatch Units 1 & 2 because Unit 2 can be adequately represented by Unit 1 PRA results (see Section 4.2).

The six general steps of this assessment are as follows:

1. Quantify the baseline risk in terms of the frequency of events (per reactor year) for each of the eight containment release scenario types identified in the EPRI report.

2. Develop plant-specific person-rem (population dose) per reactor year for each of the eight containment release scenario types from plant specific consequence analyses.

3. Evaluate the risk impact (i.e. the change in containment release scenario type frequency and population dose) of extending the ILRT interval to fifteen years.

4. Determine the change in risk in terms of Large Early Release Frequency (LERF) in accordance with RG 1.174 [ 4] and compare this change with the acceptance guidelines of RG 1.174.

5. Determine the impact on the Conditional Containment Failure Probability (CCFP)

6. Evaluate the sensitivity of the results to assumptions in the corrosion analysis, external events, and to the probability of undetected leaks from containment (due to corrosion breach) to LERF.

This approach is based on the information and approaches contained in the previously mentioned studies. Furthermore,

• Consistent with the other industry containment leak risk assessments, the Hatch assessment uses LERF and delta LERF in accordance with the risk acceptance guidance of RG 1.174. Changes in population dose and conditional containment failure probability (CCFP) are also considered to show that defense-in-depth and the balance of prevention and mitigation is preserved.

• This evaluation uses ground rules and methods to calculate changes in risk metrics that are consistent with those in the EPRI methodology [22].

2-1

• The EPRI methodology [22] specifies that emergency core cooling system (ECCS) net positive suction head (NPSH) requirements be assessed regarding whether containment over pressure is credited in the design basis ECCS analysis, and if containment over pressure is credited, the potential impacts on the core damage frequency (CDF).

As documented in Section 6.3.3.9 of the Hatch FSAR [36], containment over pressure is not required or credited for Unit 2 for either short term (i.e., < 10 minutes following LOCA initiation) or long term Residual Heat Removal (RHR) pump or Core Spray (CS) pump operation. For Unit 1, the design basis calculations indicate that 3.24 psig (7.5 ft) of containment over pressure is required to ensure adequate NPSH to the RHR pumps, and 3.2 psig (7.4 ft) of containment over pressure is required to ensure adequate NPSH to the CS pumps (at the peak calculated suppression pool temperature of 211.3 °F) for a period from about 2.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> to 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br /> after LOCA initiation. These design basis calculations utilize conservative inputs (e.g., reactor operation at 100.5%) and one RHR heat exchanger.

To provide sufficient margin, the long term NPSH evaluation takes credit for 4.2 psig (10 ft) of containment over pressure for the period of 1.5 to 26.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> following LOCA initiation. No over pressure credit is required for Unit 1 for the short term response (i.e., < 10 minutes following LOCA initiation).

MAAP runs in support of the Hatch PRA demonstrate that if RHR containment heat removal is available, the suppression pool water temperature stays well below 211 °F in the long term for a large LOCA and loss of ECCS NPSH is not a concern. Table 2.0-1 presents the results from four MAAP sensitivity cases performed in support of the ILRT analysis.

The MAAP cases model a large break LOCA (i.e., 28" diameter recirculation line break), core injection via core spray, one train of RHR for containment heat removal, with varied containment leakage values.

MAAP Case 51601 serves as the base case and models maximum allowed Technical Specification containment leakage (i.e., 1 La). The other three cases model increased containment leakage areas to estimate leakage for approximately 100La, 200La, and 400La. The general EPRI methodology is based on assuming a 100La leakage rate. Consistent with the design basis calculations, these MAAP cases utilize an initial torus water temperature of 100 °F and drywell temperature of 150 °F. The MAAP cases demonstrate that a single train of RHR containment heat removal is adequate to keep the suppression pool temperature approximately 181 °F or lower up to a leakage of 400La. With the suppression pool temperature well below 211 °F, loss of ECCS NPSH is not a concern for sequences where one or more trains of RHR containment heat removal operate.

In the event that containment heat removal (i.e., RHR and containment vent) is unavailable, containment pressure will increase to the point of containment failure due to over pressure. In the Hatch PRA, containment failure is assumed to result in a loss of ECCS core injection due to the potential for disruption of injection lines and degraded environmental conditions in plant areas housing injection equipment. While the potential exists for a pre-existing containment failure (as might be detected by the Type A ILRT) to preclude the containment pressure from reaching the point of containment over pressure failure and instead result in loss of adequate NPSH to the ECCS pumps taking suction from the suppression 2-2

pool, the end result would be the same, i.e., loss of ECCS injection leading to core damage. Therefore, there is no change in CDF associated with loss of containment heat removal sequences.

Regarding the consideration of successful containment vent or other containment isolation failures (e.g., random containment isolation valve failures), the CDF associated with such accident sequences is not impacted by the ILRT frequency. The ILRT frequency only impacts risk (i.e., CDF or LERF) associated intact containment configurations, (i.e., as characterized by EPRI Classes 1, 3a, and 3b in the EPRI methodology). Containment configurations which are not intact (e.g., EPRI Class 2 for large containment isolation failures) are not impacted by the ILRT frequency because containment integrity is failed independent of the containment failure mechanisms evaluated by an ILRT.

Based on the above discussion, there is no meaningful change in the CDF associated with the ILRT interval pertaining to the long term containment over pressure credit for Unit 1 ECCS NPSH. Therefore CDF is not quantitatively evaluated in this ILRT risk assessment as a figure of merit.

It is additionally noted that design basis NPSH calculations include a level of conservatism. For instance, the manufacturer recommended NPSH limit includes an operational design margin. The amount of margin depends on the specific pump design and the operating condition of the pump. Plant tests at TVA (Browns Ferry) [37] and Monticello have shown that substantial margin exists for ECCS pumps of BWR/3 and BWR/4 plants. Thus, the best estimate NPSH requirements for Hatch ECCS pump successful operation would be expected to be less than either that credited or that calculated by the design basis analysis.

2-3

Table 2.0-1 HATCH MAAP CONTAINMENT OVER PRESSURE SENSITIVITY CASEsCtl MAAP CASE CONTAINMENT MAXIMUM TORUS TIME AT MAXIMUM ID LEAKAGE SIZE TEMP (F) TORUS TEMP 51601 5E-5 ft"-2 181 6.5 to 8.5 hrs (1 La) 51602 5E-3 ft"-2 181 6.9 to 8.0 hrs

("'100 La) 51603 1E-2 Ft"-2 180 5.5 to 9.8 hrs

("'200 La) 51604 2E-2 ft"-2 180 5.5 to 9.8 hrs

("'400 La) 1

' J Cases model LLOCA, core injection via core spray, one train of RHR for containment heat removal, and varied containment leakage sizes.

2-4

3.0 GROUND RULES The following ground rules are used in the analysis:

• The technical adequacy of the Hatch Unit 1 PRA is consistent with the requirements of Regulatory Guide 1.200 [28] as is relevant to this ILRT interval extension. The PRA technical adequacy is documented separately.

• The Hatch Unit 1 Level 1 and Level 2 internal events PRA models provide representative results. (A Unit 2 PRA model is available and the CDF and LERF results are essentially the same as the Unit 1 results. It is judged that the Unit 2 model will not provide any unique or additional insights compared to the results from the Unit 1 model.)

• It is appropriate to use the Hatch Unit 1 internal events PRA model as a gauge to effectively describe the risk change attributable to the ILRT extension. It is reasonable to assume that the impact from the ILRT extension (with respect to percent increases in population dose) will not substantially differ if fire and seismic events were to be included in the calculations; nevertheless fire and seismic events have been accounted for in the analysis based on the available information from the Hatch IPEEE

[18] as described in Section 5.8.

• Dose results for the containment failures modeled in the PRA can be characterized by information provided in the Hatch Severe Accident Mitigation Alternatives (SAMA) analysis and associated responses to Requests for Additional Information (RAis) [9, 29, 30]. Hatch SAMA dose results all represent high magnitude releases. These dose results can be applied to containment failure releases that are lower in magnitude (i.e.,

non-high releases).

• Plant specific dose calculations for containment intact cases are not available from the Hatch SAMA analysis. NUREG-1150 results for such cases are adequately representative for use in the Hatch analysis based on scaling the NUREG-1150 results to account for differences in regional population, power level, and allowed technical specification leakage.

• Accident classes describing radionuclide release end states are defined consistent with the EPRI methodology [22], as summarized in Section 4.2.

• The representative containment leakage for Class 1 sequences is lLa .

Class 3 accounts for increased leakage due to Type A inspection failures.

• The representative containment leakage for Class 3a sequences is lOLa, based on the previously approved methodology performed for Indian Point Unit 3 [6, 7].

• The representative containment leakage for Class 3b sequences is lOOLa, based on the NRC SER [32] and incorporated in the latest EPRI report

[22]. Note that most of the previous one-time ILRT extension requests utilized 35La.

3-1

• The Class 3b can be very conservatively categorized as LERF based on the previously approved methodology [6, 7]. The Class 3b category increase is used as a surr9gate for LERF in this application even though the releases associated with a lOOLa release would not necessarily be consistent with a "Large" release for Hatch.

• The impact on population doses from containment bypass scenarios is not altered by the proposed ILRT extension. Rather it is accounted for in the EPRI methodology as a separate entry for comparison purposes, as accepted in the NRC SER [32]. Because the containment bypass contribution to population dose is fixed, no changes to the conclusions from this analysis will result from this separate categorization.

• The reduction in ILRT frequency does not impact the reliability of containment isolation valves to close in response to a containment isolation signal.

• Consideration of the risk impact of the ILRT on shutdown risk is addressed in Section 6 using the generic results from EPRI TR-105189 [8].

• The ILRT analysis evaluates very small changes in the risk metrics. To facilitate the calculation of these changes and the evaluation of sensitivity cases, the calculations are performed in a spreadsheet. In general, the calculations provided in this report reproduce the calculation results of the spreadsheets. In some cases there may be minor differences in the results between the spreadsheet calculations and hand calculations due to rounding (e.g., a column total in a table may differ). To maintain consistency, results from the spreadsheets are presented in this report.

3-2

4.0 INPUTS This section summarizes the general resources available as input (Section 4.1) and the plant specific resources required (Section 4.2).

4.1 GENERAL RESOURCES AVAILABLE Various industry studies on containment leakage risk assessment are briefly summarized here:

1. NUREG/CR-3539 [10]

2. NUREG/CR-4220 [11]

3. NUREG-1273 [12]

4. NUREG/CR-4330 [13]

5. EPRI TR-105189 [8]

6. NUREG-1493 [5]

7. EPRI TR-104285 [2]

8. NUREG-1150 [14] and NUREG/CR-4551 [26]

9. NEI Interim Guidance [3, 21]

10. Calvert Cliffs liner corrosion analysis [19]

11. NRC SER [32] on EPRI TR-1009325

12. EPRI 1018243 [22]

The first study is applicable because it provides one basis for the threshold that could be used in the Level 2 PRA for the size of containment leakage that is considered significant and to be included in the model. The second study is applicable because it provides a basis of the probability for significant pre-existing containment leakage at the time of a core damage accident. The third study is applicable because it is a subsequent study to NUREG/CR-4220 that undertook a more extensive evaluation of the same database. The fourth study provides an assessment of the impact of different containment leakage rates on plant risk. The fifth study provides an assessment of the impact on shutdown risk from ILRT test interval extension. The sixth study is the NRC's cost-benefit analysis of various alternative approaches regarding extending the test intervals and increasing the allowable leakage rates for containment integrated and local leak rate tests. The seventh study is an EPRI study of the impact of extending ILRT and LLRT test intervals on at-power public risk. The eighth study documents ex-plant consequence results which may be used as surrogate results in the ILRT risk assessment. The ninth study includes the NEI recommended methodology (promulgated in two letters) for evaluating the risk associated with obtaining a one-time extension of the 4-1

ILRT interval. The tenth study addresses the impact of age-related degradation of the containment steel on ILRT evaluations. The eleventh study [32] documents the NRC Final Safety Evaluation of the EPRI 2007 version of ILRT risk assessment guidance (i.e., EPRI TR-1009325, Revision 2). The last study by EPRI complements the previous EPRI report [2],

integrates the NEI interim guidance and NRC SER limitations and conditions, and provides a recommended methodology and template for evaluating the risk associated with a permanent 15-year ILRT interval.

NUREG/CR-3539 [101 Oak Ridge National Laboratory (ORNL) documented a study of the impact of containment leak rates on public risk in NUREG/CR-3539. This study uses information from WASH-1400 [15] as the basis for its risk sensitivity calculations. ORNL concluded that the impact of leakage rates on LWR accident risks is relatively small.

NUREG/CR-4220 [11]

NUREG/CR-4220 is a study performed by Pacific Northwest Laboratories for the NRC in 1985.

The study reviewed over two thousand LERs, ILRT reports and other related records to calculate the unavailability of containment due to leakage. It assessed the "large" containment leak probability to be in the range of lE-3 to lE-2, with 5E-3 identified as the point estimate based on 4 events in 740 reactor years and conservatively assuming a one-year duration for each event.

NUREG-1273 [12]

A subsequent NRC study, NUREG-1273, performed a more extensive evaluation of the NUREG/CR-4220 database. This assessment noted that about one-third of the reported events were leakages that were immediately detected and corrected. In addition, this study noted that local leak rate tests can detect "essentially all potential degradations" of the containment isolation system.

4-2

NUREG/CR-4330 [13]

NUREG/CR-4330 is a study that examined the risk impacts associated with increasing the allowable containment leakage rates. The details* of this report have no direct impact on the modeling approach of the ILRT test interval extension, as NUREG/CR-4330 focuses on leakage rate and the ILRT test interval extension study focuses on the frequency of testing intervals.

However, the general conclusions of NUREG/CR-4330 are consistent with NUREG/CR-3539 and other similar containment leakage risk studies:

" ... the effect of containment leakage on overall accident risk is small since risk is dominated by accident sequences that result in failure or bypass of containment."

EPRI TR-105189 [81 The EPRI study TR-105189 is useful to the ILRT test interval extension risk assessment because this EPRI study provides insight regarding the impact of containment testing on shutdown risk. This study performed a quantitative evaluation (using the EPRI ORAM software) for two reference plants (a BWR-4 and a PWR) of the impact of extending ILRT and LLRT test intervals on shutdown risk. The result of the study concluded that a small but measurable safety benefit (shutdown CDF reduced by lE-8/yr to lE-7/yr) is realized from extending the test intervals from 3 per 10 years to 1 per 10 years.

NUREG-1493 [51 NUREG-1493 is the NRC's cost-benefit analysis for proposed alternatives to reduce containment leakage testing intervals and/or relax allowable leakage rates. The NRC conclusions are consistent with other similar containment leakage risk studies:

• Reduction in ILRT frequency from 3 per 10 years to 1 per 20 years results in an "imperceptible" increase in risk.

• Given the insensitivity of risk to the containment leak rate and the small fraction of leak paths detected solely by Type A testing, increasing the interval between integrated leak rate tests is possible with minimal impact on public risk.

EPRI TR-104285 [21 Extending the risk assessment impact beyond shutdown (the earlier EPRI TR-105189 study),

the EPRI TR-104285 study is a quantitative evaluation of the impact of extending ILRT and LLRT test intervals on at-power public risk. This study combined !PE Level 2 models with 4-3

NUREG-1150 [14] Level 3 population dose models to perform the analysis. The study also used the approach of NUREG-1493 in calculating the increase in pre-existing leakage probability due to extending the ILRT and LLRT test intervals.

EPRI TR-104285 used a simplified Containment Event Tree to subdivide representative core damage sequences into eight categories of containment response to a core damage accident:

1. Containment intact and isolated

2. Containment isolation failures dependent upon the core damage accident

3. Type A (ILRT) related containment isolation failures

4. Type B (LLRT) related containment isolation failures

5. Type C (LLRT) related containment isolation failures

6. Other penetration related containment isolation failures

7. Containment failure due to core damage accident phenomena

8. Containment bypass Consistent with the other containment leakage risk assessment studies, this study concluded:

"These study results show that the proposed CLRT [containment leak rate tests] frequency changes would have a minimal safety impact. The change in risk determined by the analyses is small in both absolute and relative terms.

For example, for the PWR analyzed, the change is about 0.02 person-rem per year ... "

NUREG-1150 [14] and NUREG/CR-4551 [261 NUREG-1150 [14] and the technical basis, NUREG/CR-4551 [26], provide an ex-plant consequence analysis for a spectrum of accidents including a severe accident with the containment remaining intact (i.e., Tech Spec leakage). This ex-plant consequence calculation is calculated for the 50-mile radial area surrounding Peach Bottom. The ex-plant consequence calculation for the containment remaining intact represents a very small contributor to the overall risk spectrum. Because it is a small contributor, this ex-plant calculation (i.e., total person-rem) is considered adequate to represent Hatch if population, reactor power, and the Technical Specification leakage are scaled to represent Hatch. (The meteorology and other site differences are assumed not to play a significant role in this evaluation).

4-4

NE! Interim Guidance [3, 211 NE! "Interim Guidance for Performing Risk Impact Assessments in Support of One-Time Extensions of Containment Integrated Leakage Rate Test Surveillance Intervals" [3] was developed to provide utilities with revised guidance regarding licensing submittals. Additional information from NE! on the "Interim Guidance" was supplied in Reference [21].

A nine step process is defined which includes changes in the following areas of the previous EPRI guidance:

• Impact of extending surveillance intervals on dose

• Method used to calculate the frequencies of leakages detectable only by ILRTs

• Provisions for using NUREG-1150 dose calculations to support the population dose determination.

The guidance provided in this document builds on the EPRI risk impact assessment methodology [2] and the NRC performance-based containment leakage test program [5], and considers approaches utilized in various submittals, including Indian Point 3 (and associated NRC SER) [6,7] and Crystal River [20].

Calvert Cliffs Liner Corrosion Analysis [19]

This submittal to the NRC describes a method for determining the change in likelihood, due to extending the ILRT, of detecting liner corrosion, and the corresponding change in risk. The methodology was developed for Calvert Cliffs in response to a request for additional information regarding how the potential leakage due to age-related degradation mechanisms were factored into the risk assessment for the ILRT one-time extension. The Calvert Cliffs analysis was performed for a concrete cylinder and dome and a concrete basemat, each with a steel liner. Licensees may consider approved LARs for one-time extensions involving containment types similar to their facility. The Hatch assessment has addressed the plant-specific differences from the Calvert Cliffs design, and the Calvert Cliffs methodology was adapted to address the specific design features.

4-5

NRC SER on ILRT Risk Assessment [321 This report documents the NRC review and acceptance of the EPRI ILRT Risk Assessment methodology of EPRI TR-1009325 Revision 2. Based on the NRC review, four conditions and limitations were identified, summarized here as:

1. Licensees must submit documentation supporting appropriate technical adequacy of the PRA.

2. Acceptance criteria for population dose risk and CCFP were revised.

3. Assumed leakage for EPRI Class 3b is revised from 35La to lOOLa.

4. A license amendment request (LAR) is required in instances where containment over pressure is relied upon for ECCS performance.

EPRI TR-1018243 [221 (EPRI TR-1009325 Revision 2-A)

This report presents a generally applicable assessment of risk involved in extension of ILRT test intervals to 15 years on a permanent basis. Appendix H of this document provides guidance for performing plant-specific supplemental risk impact assessments and builds on the previous EPRI risk impact assessment methodology TR-104285 [2], the NEI Interim Guidance

[3,21], and the NRC performance-based containment leakage test program [5], and considers approaches utilized in various submittals, including Indian Point 3 (and associated NRC SER)

[6,7] and Crystal River [20]. The EPRI report codifies minor changes to the ILRT methodology specified by the NRC in the NRC ILRT risk assessment approach SER [32].

The approach included in this EPRI guidance document is used in the Hatch assessment to determine the estimated increase in risk associated with the ILRT extension. This document includes the bases for the values assigned in determining the probability of leakage for the EPRI Class 3a and 3b scenarios in this analysis as described in Section 5.

4.2 PLANT-SPECIFIC INPUTS The Hatch specific information used to perform this ILRT interval extension risk assessment includes the following:

• PRA Level 1 Model results [16]

• PRA Level 2 Model results [17], including release category definitions, and containment failure probability data

• Population Dose within a 50-mile radius [9, 29, 30]

• ILRT results to demonstrate adequacy of the administrative and hardware interfaces 4-6

Hatch Internal Events Level 1 PRA Model The Unit 1 Internal Events Level 1 PRA model [16] is an event tree / linked fault tree model characteristic of the as-built, as-operated plant. This Level 1 PRA model incorporates the resolution of findings associated with the PRA Peer Review of 2009. The total internal events core damage frequency (CDF) used in this analysis is 7.57E-06/yrC 1l (at 1E-12/yr truncation) for Unit 1, as reflected in the combined Unit 1 Level 1 and Level 2 PRA models [17]. (For reference, it is noted that the CDF for the Unit 2 model is 7.42E-06/yr [39], approximately 1.5% less than the Unit 1 CDF. The Unit 1 model is adequately representative of Unit 2 for the purposes of the ILRT risk assessment.)

Hatch Internal Events Level 2 PRA Model The Unit 1 Level 2 PRA model [17] was developed to calculate the LERF contribution as well as the other release categories evaluated in the model. This Level 2 PRA model incorporates the resolution of findings associated with the PRA Peer Review of 2009. Table 4.2-1a summarizes the pertinent Hatch Unit 1 Level 2 results in terms of end states. The total Large Early Release Frequency (LERF) in Table 4.2-1a for Unit 1 is 1.12E-6/yr. The Unit 2 model LERF value is 1.03E-06/yr [31], approximately 8% less than the Unit 1 LERF. The lower Unit 2 LERF value is primarily attributed to a plant design difference. The Unit 2 feedwater injection lines have an additional check valve which lowers the break outside containment (BOC) contribution to LERF for Unit 2. This design difference does not impact the risk assessment because the ILRT interval does not impact the BOC LERF contribution.

(ll The Unit 1 Level 1 CDF value of 7.57E-06/yr used in the Level 2 evaluation [17] is slightly higher than the Level 1 CDF value of 7.53E-06/yr from the latest version of the Hatch Unit 1 Internal Events Level 1 PRA model [16]. To support the Level 2 quantification, Level 1 sequences are binned into accident classes. However, this separate quantification of the individual accident classes may result in duplicate or non-minimal cutsets to be binned into more than one accident class. This may result in the numerical sum of all individual accident classes to be higher than the CDF if all the cutsets were merged together. However, the apparent deviation of the Level 1 CDF quantified for the Level 2 model is less than 1 % and is judged not to significantly alter the results.

4-7

The Level 2 release category end states are defined [38] as follows:

Release Magnitude CsI Release Fraction High > 10%

Moderate/Medium 1% to 10%

Low 0.1% to 1%

Low-Low < 0.1%

Release Timing Time (hrs)

Early < 5 Intermediate 5 to 24 Late > 24 Table 4.2-1b summarizes the core damage frequency contributions by the PRA accident class.

4-8

Table 4.2-1a HATCH LEVEL 2 DETAILED RELEASE CATEGORIESC 1 >

RELEASE FREQUENCY CATEGORY DEFINITION (/YR)

INTACT Containment remains intact. 1.lSE-06 H-E High-early release (i.e., LERF). Dominant accident class l.12E-06 contributors are as follows:

. Class lA (loss of RPV injection, RPV at high pressure):

6 .13 E-0 8/yr

. Class ID (loss of RPV injection, RPV at low pressure):

6.17E-08/yr

. Class 4 (ATWS) : l.93E-07/yr

. Class 5 (BOC): 7.79E-07/yr H-I High-intermediate release. Dominant accident class contributor 2.83E-06 is Class 2A (loss of containment heat removal, CD post-containment failure) at 3. l 7E-06/yr.

M-E Moderate-early release. Dominant accident class contributor is l.19E-06 Class lD (loss of RPV injection, RPV at low pressure) at 8.04E-07/yr.

M-I Moderate-intermediate release. Dominant accident class 9.64E-07 contributor is Class 2A at 8.23E-07/yr.

M-L Moderate-late release. Dominant accident class contributor is 4.64E-08 Class lA at 4.80E-08/yr.

L-E Low-early release. Dominant accident class contributor is Class l.OlE-08 lA at 2.24E-09/yr.

L-I Low-intermediate release. Dominant accident class contributor 9.56E-08 is Class 1D at l.OSE-07/yr.

L-L Low-late release. Dominant accident class contributor is Class 6.94E-09 lA at 7.42E-09/yr.

LL-E Low Low-late release. Dominant accident class contributor is l.33E-07 Class lD at 4.21E-08/yr.

LL-I Low Low-intermediate release. Dominant accident class l.OSE-08 contributor is Class lD at l.02E-08/yr.

LL-L Low Low-late release. Dominant accident class contributor is 4.63E-09 Class lA at 4.86E-09/yr.

Total Total Release Category Frequency (No Intact) 6.40E-06 Total Total CDF 7.SSE-06 (ll From Table 5 of Reference [17] for Unit 1. The High-Late release category had zero frequency and is therefore not listed.

4-9

Table 4.2-1b HATCH CDF CONTRIBUTIONS BY PRA ACCIDENT CLAss< 1 >

PRA FREQUENCY ACCIDENT (/YR)

CLASS DESCRIPTION IA Transients - core melt with vessel at high pressure 1.07E-06 IBE Station blackout - early 1.18E-08 IBL Station blackout - late 4.89E-07 IC ATWS with loss of injection 1.73E-07 ID Transients - core melt with vessel at low pressure 1.35E-06 IIA Core melt after containment failure due to loss of DHR 3.39E-06 Ill Core melt after containment failure due to loss of DHR and 4.llE-10 LOCA IIIB LOCA - core melt with vessel remaining at high pressure 1.SOE-08 me LOCA - core melt with vessel at low pressure 2.75E-09 IV ATWS - containment fails before core damage 3.58E-07 v LOCA outside containment 7.12E-07 Total Total CDF 7.57E-06

<1 > From Table 5 of Reference [17] for Unit 1.

4-10

Population Dose Conditional population dose results for containment failure end states are available for Hatch based on the Hatch SAMA evaluation performed for Units 1 & 2 and submitted to the NRC in 2000 [9], and subsequent responses to Requests for Additional Information (RAls) [29, 30].

Conditional population dose results for an intact containment end state (not quantified for the SAMA analysis) are available via ex-plant consequence results for Peach Bottom [26] and can be scaled to represent Hatch. The Hatch specific and Peach Bottom surrogate conditional population dose results may be combined with the most recent Hatch Level 2 analysis results

[17] to develop population dose risk for use in the ILRT assessment.

The SAMA dose analysis utilized the projected population to year 2030 (i.e., 498,834 people in the 50 mile radial region) and a Hatch power level of 2,763 MWth. The population projection is adequately representative for use in the ILRT assessment. The Hatch power level used in the SAMA analysis is slightly less than the current and anticipated Hatch power level in the future, which is 2,804 MWth. The SAMA dose values may be scaled for use in the ILRT analysis by applying a reactor power level scaling factor of 1.015 (i.e., 2,804 MWth / 2,763 MWth).

The Hatch SAMA population dose results are presented in Table 4.2-2. These dose results are based on MACCS2 calculations and accident sequence frequencies applicable at the time.

Included in Table 4.2-2 is a column presenting the ILRT assessment dose values after applying the reactor power level scaling factor. It is noted that the release categories represented in the Hatch SAMA analysis all represent high magnitude releases. Doses associated with non-large releases from containment failure can be conservatively represented by this data.

The population dose associated with an intact containment (Technical Specification leakage) case can be estimated based on scaling the NUREG/CR-4551 dose results for Peach Bottom from Accident Progression Bin (APB) #8 (Core is damaged, Vessel is breached, no containment failure)C 1>. The Peach Bottom dose for APB #8 is not specifically identified in NUREG/CR-4551, but can be back-calculated to be 4,940 person-rem as presented in Table 4.2-3.

(lJ APB #8 is described in more detail in NUREG/CR-4551 [26] Section 2.4.3.

4-11

The APB #8 person-rem result can be used as an approximation of the dose for Hatch if it is scaled for regional population, reactor power level, and allowable containment leakage rate (La). Values for these attributes for Peach Bottom (as evaluated in NUREG/CR-4551) and Hatch are summarized in Table 4.2-4, where the applicable scaling factors are calculated.

Applying the calculated scaling factors, the population dose for Hatch for an intact containment technical specification release is 1,150 pers-rem (i.e, 4,940 pers-rem

• 0.114

• 0.852

• 2.4 =

1,150 pers-rem).

Table 4.2-5 presents the current Hatch Level 2 release frequencies, the assigned dose for the category, and the calculated annual dose risk. The annual dose risk calculated in Table 4.2-5 is not directly used in the ILRT assessment since the EPRI methodology utilizes a different release category scheme, but is presented for completeness.

EPRI Release Category Definitions Table 4.2-6 defines the accident classes used in the ILRT extension evaluation, which are consistent with the EPRI methodology [22]. These containment failure classifications are used in this analysis to determine the risk impact of extending the Containment Type A test interval as described in Section 5 of this report.

Hatch ILRT Results The surveillance frequency for Type A testing in NEI 94-01 under option B criteria is at least once per ten years based on an acceptable performance history (i.e. two consecutive periodic Type A tests at least 24 months apart where the calculated performance leakage rate was less than 1.0 La) and consideration of the performance factors in NEI 94-01, Section 11.3. Based on completion of two successful ILRTs at Hatch Unit 1 and Unit 2, the ILRT interval became once per ten years. Subsequently, a one time ILRT interval frequency of once per fifteen years was approved for both Hatch Unit 1 and Unit 2 [33, 34] based on demonstrating acceptable risk impacts. Each Hatch unit has successfully completed another ILRT (i.e., Unit 1 in March 2008, Unit 2 in March 2009) since these one time ILRT interval extension approvals.

4-12

Table 4.2-2

SUMMARY

OF SAMA MACCS2 CALCULATIONS AND ILRT SCALED VALUES SAMA Annual SAMA Adjusted Dose for Risk Frequency SAMA Dose ILRT Assessment (Person-Rem/Yr)

Level 2 End State Seq# Sequence Description (per yr) cioi (Person-Rem) (Person-Rem) cui [29, 30]

Containment Bypass 5 BOC 1.66E-7c6 > 1.15E+6C2 > 1.17E+6 0.19 Early Cont. Failure 2 SBO 1.79E-6c 6 l 1.06E+6C3 l 1.08E+6 1.90 4 Loss of Cont. 7.43E-7c 6 l 1.02E+6C4 l 1.04E+6 0.76 Heat Removal (CHR) 11 ATWS 7.43E-7C 6 l 7.02E+5C 5 l 7.13E+5 0.52 3.18 total Late Cont. Failure 12 High pressl!re transient 2.0E-7C 1l 5.7E+5 5.8E+5 0.112C*l with loss of CHR 14 SBO with cont. isolation 3.1E-9C 1l 0.0008 failure Intact Cont. (DW Vent) 15 High pressure transient 9 .24E-10C 5 l 1.13E+6C9 l 1.15E+6 0.001 with venting No Containment Failure NA NA NA NAC 7 l NA NAPl NA TOTAL 3.48 C1 l SAMA RAI response to Q#4 [29].

cii SAMA RAI response to Q#14; Sequence #5 [29] clarification provided to NRC by SNC [30].

C3 l SAMA RAI response to Q#l4; Sequence #2 [29].

C4 l SAMA RAI response to Q#14; Sequence #4 [29].

C5 l SAMA RAI response to Q#l4; Sequence #11 [29].

CG) SAMA RAI response to Q#l.b-1 [29].

C7 l Not calculated for SAMA.

C*l SAMA RAI clarification provided by SNC to Question #5 [30].

C9 l SAMA RAI response to Q#l4; Sequence 15 [29].

C10> It is noted that the Hatch PRA model has been updated since the SAMA analysis and the accident sequence frequencies and the associated annual population dose has decreased from that used in the SAMA evaluation.

cui Original SAMA Dose values of previous column, scaled by a factor of 1.015 to account for a higher reactor power level for ILRT assessment.

4-13

Table 4.2-3 PEACH BOTTOM APB #8 50-MILE POPULATION DOSE CALCULATIONC 1 >

ALL APBS APB #8 50-MILE APB #8 50-APB #8 CONTRIBUTION DOSE RISK MILE DOSE APB #8 SO-FREQUENCY TO SO-MILE (PERS- RISK (PERS- MILE DOSE

(/YR) DOSE RISK REM/YR) REM/YR) (PERS-REM) 7.99E-7< 2l 5E-4< 3 l 7,9< 4 l 3.95E-3< 5l 4.94E+3< 5 l 1

<l NUREG/CR-4551 [26] does not document dose results as a function of accident progression bin (APB); as such, the dose result for APB #8 is back calculated from the documented APB frequency and dose risk results.

2

<l From Figure 2.5-6 of NUREG/CR-4551 Vol. 4, Rev. 1, Part 1. Frequency for APB #8 of 7.99E-7/yr is calculated as 0.184 contribution of 4.34E-6/yr CDF.

<3 l From Table 5.2-3 for the mean fractional contribution to risk (MFCR) of NUREG/CR-4551 Vol. 4, Rev. 1, Part 1.

4 C l From Table 5.1-1 for mean value 50-mile population dose of NUREG/CR-4551 Vol. 4, Rev. 1, Part 1.

5 Cl APB dose risk is calculated by multiplying the APB dose risk fractional contribution (column 2) by the total 50-mile radius dose risk of 7.9 person-rem/yr (column 3).

5 C l Calculated by dividing the APB #8 dose risk (column 4) by the APB #8 frequency (column 1)

Table 4.2-4 HATCH APB #8 DOSE SCALING FACTORS Reactor SO-mile Power TS Leakage Plant Population (MWth) (wtO/o/day)

Hatch 498.834(!) 2 804( 2) 1.2% (2)

Peach Bottom 4I 359 I 575< 3 l 3,293< 4 l 0.5%( 5)

Scaling Factor 0.114 0.852 2.4 1

<l Hatch SAMA year 2030 population [9]

2

<l Hatch current and anticipated future value.

3 C l NUREG/CR-4551, Vol. 2, Rev. 1, Part 7, Appendix A.3 (SITE MACCS2 File) for Peach Bottom. Population total for 50-mile radius developed in Appendix A of this report.

4 C l NUREG/CR-4551, Vol. 4, Rev. 1, Part 2, Section A.3.1.

5 C l NUREG/CR-4551, Vol. 4, Rev. 1, Part 2, page B.2-9 for no containment failure.

4-14

Table 4.2-5 HATCH POPULATION DOSE RISK AT 50 MILES RELEASE POPULATION DOSE CATEGORY 2030 POPULATION RISK RELEASE FREQUENCIES ASSIGNED DOSE DOSE ASSIGNMENT (PERSON-REM/YR)

(2)

CATEGORY (PER YEAR) (PERSON-REM)< 1 > BASIS INTACT 1.18E-06 1.15E+03 Peach Bottom 1.35E-03 H-E 1.12E-06 1.17E+06 Hatch SAMA BOC 1.31E+OO H-I 2.83E-06 5.80E+05 Hatch SAMA late CF 1.64E+OO 3

M-E 1.19E-06 5.80E+05 Hatch SAMA late CF ( ) 6.90E-01 3

M-I 9.64E-07 5.80E+05 Hatch SAMA late CF( l 5.59E-01 3

M-L 4.64E-08 5.80E+05 Hatch SAMA late CF( l 2.69E-02 4

L-E 1.01 E-08 5.80E+05 Hatch SAMA late CF ( l 5.86E-03 4

L-I 9.56E-08 5.80E+05 Hatch SAMA late CF< l 5.54E-02 4

L-L 6.94E-09 5.80E+05 Hatch SAMA late CF( l 4.03E-03 4

LL-E 1.33E-07 5.80E+05 Hatch SAMA late CF' ) 7.71E-02 4

LL-I 1.0SE-08 5.80E+05 Hatch SAMA late CF ( ) 6.09E-03 4

LL-L 4.63E-09 5.80E+05 Hatch SAMA late CF( l 2.69E-03 Total 7.58E-06 -- -- 4.37E+OO (1)

Includes a scaling factor of 0.233 for application of the Peach Bottom dose results to the Intact Containment case, and includes a scaling factor of 1.015 for other release categories to account for a reactor power level increase since the Hatch SAMA analysis was performed.

(2)

Obtained by multiplying the release category frequency by the conditional dose.

(3)

All Hatch SAMA dose cases represent high releases. The late containment failure dose is approximately a factor of two less than that for other high magnitude releases and is considered reasonable for use for medium magnitude release cases. This is comparable to SAMA population dose results developed for Quad Cities and Dresden Generating Stations [35] (both Mark I containment designs) where moderate magnitude releases had population dose results approximately one half to nearly equal to high magnitude release population doses.

(4)

All Hatch SAMA dose cases represent high releases. Use of the late containment failure for low and low-low magnitude release cases is acceptable because the associated frequencies for these release categories are low compared to other release categories. The population dose associated with low or low-low releases compose less than 3% of the total as developed in this table.

4-15

Table 4.2-6 EPRI CONTAINMENT FAILURE CLASSIFICATIONS [22]

CLASS DESCRIPTION 1 Containment remains intact including accident sequences that do not lead to containment failure in the long term. The release of fission products (and attendant consequences) is determined by the maximum allowable leakage rate values La, under Appendix J for that plant 2 Containment isolation failures (as reported in the IPEs) include those accidents in which there is a failure to isolate the containment.

3 Independent (or random) isolation failures include those accidents in which the pre-existing isolation failure to seal (i.e. provide a leak-tight containment) is not dependent on the sequence in progress.

4 Independent (or random) isolation failures include those accidents in which the pre-existing isolation failure to seal is not dependent on the sequence in progress. This class is similar to Class 3 isolation failures, but is applicable to sequences involving Type B tests and their potential failures. These are the Type B-tested components that have isolated but exhibit excessive leakage.

5 Independent (or random) isolation failures include those accidents in which the pre-existing isolation failure to seal is not dependent on the sequence in progress. This class is similar to Class 4 isolation failures, but is applicable to sequences involving Type C tests and their potential failures.

6 Containment isolation failures include those leak paths covered in the plant test and maintenance requirements or verified per in service inspection and testing (ISI/IST) program.

7 Accidents involving containment failure induced by severe accident phenomena. Changes in Appendix J testing requirements do not impact these accidents.

8 Accidents in which the containment is bypassed (either as an initial condition or induced by phenomena) are included in Class 8. Changes in Appendix J testing requirements do not impact these accidents.

4-16

4.3 IMPACT OF EXTENSION ON DETECTION OF COMPONENT FAILURES THAT LEAD TO LEAKAGE (SMALL AND LARGE)

The ILRT can detect a number of component failures such as containment breach, failure of certain bellows arrangements and failure of some sealing surfaces, which can lead to leakage.

The proposed ILRT test interval extension may influence the conditional probability of detecting these types of failures. To ensure that this effect is properly accounted for, the EPRI Class 3 accident class as defined in Table 4.2-6 is divided into two sub-classes representing small and large leakage failures. These subclasses are defined as Class 3a and Class 3b, respectively.

The probabilities of the EPRI Class 3a and 3b failures are determined consistent with the EPRI guidance [22]. For Class 3a, the probability is based on the mean failure estimated from the available data (i.e., two "small" failures that could only have been discovered by the ILRT; 2 of 217 tests leads to a 0.0092 mean value). For Class 3b, the Jefferys non-informative prior distribution is assumed for no "large" failures in 217 tests (i.e., 0.5/(217+1) = 0.0023).

In a follow-on letter [21] to their ILRT guidance document [3], NE! issued additional information concerning the potential that the calculated delta LERF values for several plants may fall above the "very small change" guidelines of the NRC Regulatory Guide 1.174. This additional NE! information includes a discussion of conservatisms in the quantitative guidance for delta LERF. NE! describes ways to demonstrate that, using plant-specific calculations, the delta LERF is smaller than that calculated by the simplified method.

The supplemental information states:

"The methodology employed for determining LERF (Class 3b frequency) involves conservatively multiplying the CDF by the failure probability for this class (3b) of accident. This was done for simplicity and to maintain conservatism. However, some plant-specific accident classes leading to core damage are likely to include individual sequences that either may already (independently) cause a LERF or could never cause a LERF, and are thus not associated with a postulated large Type A containment leakage path (LERF).

These contributors can be removed from Class 3b in the evaluation of LERF by multiplying the Class 3b probability by only that portion of CDF that may be impacted by type A leakage."

4-17

The application of this additional guidance to the analysis for Hatch (as detailed in Section 5),

involves the following:

The EPRI Class 2 and Class 8 sequences are subtracted from the CDF that is applied to Class 3b. To be consistent, the same change is made to the Class 3a CDF, even though these events are not considered LERF. Class 2 and Class 8 events refer to sequences with either large pre-existing containment isolation failures or containment bypass events. These sequences are already considered to contribute to LERF in the Hatch Level 2 PRA analysis.

The EPRI guidance and examples also note the potential for accident sequences involving the use of containment sprays or those resulting in late releases due 1

to timing (e.g., long term station blackout, loss of containment heat removal) to be subtracted from the CDF that is applied to Class 3b. This is conservatively not performed for the base case analysis, but is evaluated as a sensitivity case in Section 6.

Consistent with the EPRI guidance [22], the change in the leak detection probability can be estimated by comparing the average time that a leak could exist without detection. For example, the average time that a leak could go undetected with a three-year test interval is 1.5 years (3 yr / 2), and the average time that a leak could exist without detection for a ten-year interval is 5 years (10 yr / 2). This change would lead to a non-detection probability that is a factor of 3.33 (5.0/1.5) higher for the probability of a leak that is detectable only by ILRT testing, given a 10-year versus a 3-yr interval. Correspondingly, an extension of the ILRT interval to fifteen years can be estimated to lead to about a factor of 5.0 (7.5/1.5) increase in the non-detection probability of a leak.

It should be noted that using the methodology discussed above is very conservative compared to previous submittals (e.g., the IP3 request for a one-time ILRT extension that was approved by the NRC [7]) because it does not factor in the possibility that the failures could be detected by other tests (e.g., the Type B local leak rate tests that still occur.) Eliminating this possibility conservatively over-estimates the factor increases attributable to the ILRT extension.

4-18

4.4 IMPACT OF EXTENSION ON DETECTION OF STEEL CORROSION THAT LEADS TO LEAKAGE An estimate of the likelihood and risk implications of corrosion-induced leakage of the steel containment occurring and going undetected during the extended test interval is evaluated using the methodology from the Calvert Cliffs liner corrosion analysis [19]. The Calvert Cliffs analysis was performed for a concrete cylinder and dome and a concrete basemat, each with a steel liner. The analysis approach can be applied to the Hatch Mark I containment design consisting of a steel drywell (floor encased in concrete) and steel torus.

The following approach is used to determine the change in likelihood, due to extending the ILRT, of detecting corrosion of the containment steel. This likelihood is then used to determine the resulting change in risk. Consistent with the Calvert Cliffs analysis, the following issues are addressed:

• Differences between the containment drywell floor and the containment walls and head

• The historical steel flaw likelihood due to concealed corrosion

• The impact of aging

• The corrosion leakage dependency on containment pressure

• The likelihood that visual inspections will be effective at detecting a flaw Assumptions

• Consistent with the Calvert Cliffs analysis, a half failure is assumed for the drywell floor concealed steel corrosion due to the lack of identified failures.

• The two corrosion events over a 5.5 year data period are used to estimate the steel liner flaw probability in the Calvert Cliffs analysis and are assumed to be applicable to the Hatch containment analysis. These events, one at North Anna Unit 2 and one at Brunswick Unit 2 (Mark I containment design), were initiated from the non-visible (backside) portion of the containment liner. It is noted that two additional events have occurred in recent years (based on a data search covering approximately 9 years documented in Reference [27]). In November 2006, the Turkey Point 4 containment building liner developed a hole when a sump pump support plate was moved. In May 2009, a hole approximately 3/8" by 1" in size was identified in the Beaver Valley 1 containment liner. For risk evaluation purposes, these two more recent events occurring over a 9 year period are judged to be adequately represented by the two events in the 5.5 year period of the Calvert Cliffs analysis incorporated in the EPRI guidance.

4-19

• Consistent with the Calvert Cliffs analysis, the estimated historical flaw probability is limited to 5.5 years to reflect the years from September 1996 when 10 CFR 50.55a started requiring visual inspection to when the Calvert Cliffs analysis was submitted. Additional success data was not used to limit the aging impact of this corrosion issue, even though inspections were being performed prior to this date and have been performed since the time frame of the Calvert analysis. (See Table 4.4-1, Step 1).

Consistent with the Calvert Cliffs analysis, the steel flaw likelihood is assumed to double every five years. This is based solely on judgment and is included in this analysis to address the increased likelihood of corrosion as the steel ages. (See Table 4.4-1, Steps 2 and 3.) Sensitivity studies are included that address doubling this rate every two years and every ten years.

• In the Calvert Cliffs analysis the likelihood of the containment atmosphere reaching the outside atmosphere given that a flaw exists was estimated as 1.1% for the cylinder and dome region, and 0.11% (10% of the cylinder failure probability) for the basemat. These values were determined from an assessment of the probability versus containment pressure, and the selected values are consistent with a pressure that corresponds to the ILRT target pressure of 37 psig. For the Hatch Mark I containment, the containment failure probabilities are conservatively assumed to be 1 % for the drywell vertical walls and head along with the wetwell torus, and 0.1 %

for the drywell floor for this analysis. Sensitivity studies are included that increase and decrease the probabilities by an order of magnitude. (See Table 4.4-1, Step 4.)

• Consistent with the Calvert Cliffs analysis, the likelihood of leakage escape (due to crack formation) in the concrete encased drywell floor region is considered less likely than the containment walls and had region. (See Table 4.4-1, Step 4)

• Consistent with the Calvert Cliffs analysis, a 5% visual inspection detection failure likelihood given the flaw is visible and a total detection failure likelihood of 10% is used. To date, all liner corrosion events have been detected through visual inspection. For Hatch, there is g_enerally 100% accessibility for visual inspection of the interior surfaces of the drywell above the floor elevation, the outside surfaces of the suppression pool, the inside surfaces of the suppression pool (using divers below the water line), and the vent system. (See Table 4.4-1, Step 5.) Sensitivity studies are included that evaluate total detection failure likelihood of 5%

and 15%, respectively.

• Consistent with the Calvert Cliffs analysis, all non-detectable containment failures are assumed to result in early releases. This approach avoids a detailed analysis of containment failure timing and operator recovery actions.

4-20

Table 4.4-1 STEEL CONTAINMENT CORROSION BASE CASE DW WALLS AND HEAD, STEP DESCRIPTION AND TORUS DRYWELL FLOOR 1 Historical Steel Flaw Events: 2 Events: 0 (assume half a Likelihood failure)

Failure Data: Containment 2/(70

• 5.5) = 5.2E-3 0.5/(70

• 5.5) = 1.3E-3 location specific (consistent with Calvert Cliffs analysis).

2 Age Adjusted Steel Flaw Year Failure Rate Year Failure Rate Likelihood 1 2.lE-3 1 5.0E-4 During 15-year interval, avg 5-10 5.2E-3 avg 5-10 1.3E-3 assume failure rate doubles 15 1.4E-2 15 3.SE-3 every five years (14.9%

increase per year). The average for 5th to 10th year is set to the historical failure rate 15 year average = 15 year average =

6.27E-3 1.57E-3 (consistent with Calvert Cliffs analysis).

3 Flaw Likelihood at 3, 10, 0.71% (1 to 3 years) 0.18% (1 to 3 years) and 15 years 4.06% (1to10 years) 1.02% (1 to 10 years)

Uses age adjusted flaw 9.40% (1 to 15 years) 2.35% (1 to 15 years) likelihood (Step 2), assuming (Note that the Calvert Cliffs (Note that the Calvert Cliffs failure rate doubles every five analysis presents the delta analysis presents the delta years (consistent with Calvert between 3 and 15 years of between 3 and 15 years of Cliffs analysis - See Table 6 of

8. 7% to utilize in the 2.2% to utilize in the Reference [19]).

estimation of the delta-LERF estimation of the delta-LERF value. For this analysis the value. For this analysis, values are calculated based however, values are on the 3, 10, and 15 year calculated based on the 3, 10, interva Is.) and 15 year intervals.)

4 Likelihood of Breach in 1°10 0.1°10 Containment Given Steel Flaw The failure probability of the DW walls, head, and torus is assumed to be 1 % (compared to 1.1 % in the Calvert Cliffs analysis). The DW floor failure probability is assumed to be a factor of ten less, 0.1%,

(compared to 0.11% in the Calvert Cliffs analysis).

4-21

Table 4.4-1 STEEL CONTAINMENT CORROSION BASE CASE DW WALLS AND HEAD, STEP DESCRIPTION AND TORUS DRYWELL FLOOR 5 Visual Inspection Detection 10°/o 100%

Failure Likelihood 5% failure to identify visual Cannot be visually inspected.

Utilize assumptions consistent flaws plus 5% likelihood that with Calvert Cliffs analysis the flaw is not visible (not while also accouting for the through-wall but could be unique arrangement of the detected by ILRT).

Hatch containment.

All events have been detected through visual inspection. A 5% visible failure detection is a conservative assumption.

6 Likelihood of Non-Detected 0.00071% (at 3 years) 0.00018% (at 3 years)

Containment Leakage =0.71%

• 1%

• 10% =0.18%

• 0.1%

• 100%

(Steps 3

• 4

• 5) 0.0041% (at 10 years) 0.0010% (at 10 years)

=4.1%

• 1%

• 10% =1.0%

• 0.1%

• 100%

0.0094% (at 15 years) 0.0024% (at 15 years)

=9.4%

• 1%

• 10% =2.4%

• 0.1%

• 100%

4-22

The total likelihood of the corrosion-induced, non-detected containment leakage is the sum of Step 6 for the DW walls, head, and torus, and the drywell floor:

At 3 years : 0.00071 % + 0.00018% = 0.00089% = 8.9E-6 At 10 years: 0.0041% + 0.0010% = 0.0051% = 5.lE-5 At 15 years: 0.0094% + 0.0024% = 0.012% = 1.2E-4 Based on the above, a corrosion impact factor due to undetected corrosion is calculated as follows for the three ILRT cases investigated:

Total Likelihood of non-detected containment (3b Conditional leakage due to corrosion at Corrosion impact factor = Failure Probability + interval) 3b Conditional Failure Probability Case 1: 3 ILRT per 10 years 2.30E-03 + 8.9E-06 = 1.004 2.30E-03 Case 2: 1 ILRT Per 10 years 7.67E-03 + 5.lE-05 = 1.007 7.67E-03 Case 3: 1 ILRT per 15 years 1.15E-02 + 1.2E-04 = 1.01 1.15E-02 These impact factors are used to adjust the EPRI 3b class frequencies to model the impact of undetected corrosion.

4-23

5.0 RESULTS The application of the approach based on the guidance contained in EPRI TR-1018243 [22],

EPRI-TR-104285 [2] and previous risk assessment submittals on this subject [6, 7, 19, 20, 23]

have led to the following results. The results are displayed according to the eight accident classes defined in the EPRI report. Table 5.0-1 lists these accident classes.

The analysis performed examined Hatch specific accident sequences in which the containment remains intact or the containment is impaired. Specifically, the break down of the severe accidents contributing to risk were considered in the following manner:

• Core damage sequences in which the containment remains intact initially and in the long term (EPRI TR-104285 Class 1 sequences).

• Core damage sequences in which containment integrity is impaired due to random isolation failures of plant components other than those associated with Type B or Type C test components. For example, containment breach or bellows leakage. (EPRI TR-104285 Class 3 sequences).

• Core damage sequences in which containment integrity is impaired due to containment isolation failures of pathways left "opened" following a plant post-maintenance test. (For example, a valve failing to close following a valve stroke test.) (EPRI TR-104285 Class 6 sequences). Consistent with the NEI Guidance, this class is not specifically examined since it will not significantly influence the results of this analysis.

Accident sequences involving containment bypassed (EPRI TR-104285 Class 8 sequences), large containment isolation failures (EPRI TR-104285 Class 2 sequences), and small containment isolation "failure-to-seal" events (EPRI TR-104285 Class 4 and 5 sequences) are accounted for in this evaluation as part of the baseline risk profile. However, they are not affected by the ILRT frequency change.

• Class 4 and 5 sequences are impacted by changes in Type B and C test intervals; therefore, changes in the Type A test interval do not impact these sequences.

5-1

Table 5.0-1 EPRI ACCIDENT CLASSES ACCIDENT CLASSES (CONTAINMENT RELEASE TYPE) DESCRIPTION 1 No Containment Failure 2 Large Isolation Failures (Failure to Close) 3a Small Isolation Failures (containment breach) 3b Large Isolation Failures (containment breach) 4 Small Isolation Failures (Failure to seal -Type B) 5 Small Isolation Failures (Failure to seal-Type C) 6 Other Isolation Failures (e.g., dependent failures) 7 Failures Induced by .Phenomena (Early and Late) 8 Bypass (SGTR and Interfacing System LOCA)

CDF All CET End states (including very low and no release)

The steps taken to perform this risk assessment evaluation are as follows:

Step 1 Quantify the base-line risk in terms of frequency per reactor year for each of the eight accident classes presented in Table 5.0-1.

Step 2 Develop plant-specific person-rem dose (population dose) per reactor year for each of the eight accident classes.

Step 3 Evaluate risk impact of extending Type A test interval from 3 to 15 and 10 to 15 years.

Step 4 Determine the change in risk in terms of Large Early Release Frequency (LERF) in accordance with RG 1.174.

Step 5 Determine the impact on the Conditional Containment Failure Probability.

(CCFP)

Step 6 Determine the impact on the 50-mile population dose risk.

5-2

5.1 STEP 1 - QUANTIFY THE BASE-LINE RISK IN TERMS OF FREQUENCY PER REACTOR YEAR As previously described, the extension of the Type A interval does not influence those accident progressions that involve large containment isolation failures, Type B or Type C testing, or containment failure induced by severe accident phenomena.

For the assessment of ILRT impacts on the risk profile, the potential for pre-existing leaks is included in the model. These events are represented by the EPRI Class 3 sequences. Two failure modes were considered for the Class 3 sequences. These are Class 3a (small breach) and Class 3b (large breach).

The frequencies for the EPRI accident classes defined in Table 5.0-1 were developed for Hatch by first determining the frequencies for Classes 1, 2, 7, and 8, then determining the frequencies for Classes 3a and 3b, and finally determining the frequency for Class 1. Classes 4, 5, and 6 are not impacted by the ILRT interval and are therefore not specifically evaluated.

Adjustments are made to the Class 3b frequency and hence Class 1 frequency to account for the impact of undetected corrosion of the steel containment per the methodology described in Section 4.4.

Class 1 Sequences This group represents the frequency when the containment remains intact (modeled as Technical Specification Leakage). The EPRI Class 1 frequency is calculated as the intact containment release frequency from Table 4.2-1a (1.18E-06/yr) minus the EPRI Class 3a and 3b frequencies (6.31E-08/yr and 1.58E-08/yr, respectively) calculated below. For this analysis, the associated maximum containment leakage for this group is 1La, consistent with an intact containment evaluation.

The EPRI Class 1 frequency is 1.10E-06/yr.

Class 2 Sequences This group consists of all core damage accident sequences for which a large containment isolation failure(s) occurs (e.g., valve failure to close).

5-3

The frequency of this EPRI category is estimated by multiplying the conditional probability of containment isolation failure from the Hatch Unit 1 Level 2 PRA by the portion of the severe accident sequences (CDF) that would be challenged. The Level 2 sequences that have containment isolation already failed (via severe accident phenomena) are PRA accident classes

!IA, Ill, IIID, IVA, and V. The following values are used for this calculation:

• Containment isolation system failure probability = 1.6E-4C 1> [17]

• Total CDF = 7.57E-06/yr [17]

• Class !IA sequences = 3.39E-06/yr [17]

• Class Ill sequences = 4.llE-10/yr [17]

• Class IVA sequences = 3.SSE-07/yr [17]

• Class V sequences = 7 .12E-07/yr [17]

The EPRI Class 2 frequency is calculated as follows:

Frequency 2 = (Cont !sol prob) * (CDF - CDF !IA - CDF Ill - CDF IVA - CDF V)

Frequency 2 = 1.6E-4 * (7.57E-06/yr - 3.39E-06/yr - 4.llE-10/yr - 3.SSE-07/yr

- 7.12E-07/yr)

Frequency 2 = 1.6E-4 * (3.llE-06/yr) = 4.98E-10/yr The frequency for EPRI Class 2 is 4.98E-10/yr.

Class 3 Sequences This group represents pre-existing leakage in the containment structure (e.g., drywell and torus). The containment leakage for these sequences can be either small (in excess of design allowable but < lOLa) or large (rvlOOLa). In this analysis, a value of !Ola was used for small pre-existing flaws and lOOLa for relatively large flaws.

The respective frequencies per year are determined as follows:

PROBc1ass_3a = probability of small pre-existing containment leakage

= 0.0092 [see Section 4.3]

PROBc1ass_3b = probability of large pre-existing containment leakage

= 0.0023 [see Section 4.3]

<1 > Containment isolation system failure probability based on nodal quantification of event node 151 (7.86E-3) minus the pre-existing containment failure probability basic event (7. 7E-3). Pre-existing containment failures are evaluated in other EPRI classes.

5-4

As described in Section 4.3, additional consideration is made to not apply these failure probabilities on those cases that are already LERF scenarios (i.e. EPRI Class 2 & Class 8 contributions). (Removing non-early sequences is addressed as a sensitivity case in Section 6.)

Class 3a = 0.0092 * (CDF - Class 2 - Class 8)

= 0.0092 * (7.57E-06/yr - 4.98E-10/yr - 7.12E-07/yr)

= 6.31E-08/yr Class 3b = 0.0023 * (CDF - Class 2 - Class 8)

= 0.0023 * (7.57E-06/yr - 4.98E-10/yr - 7.12E-07/yr)

= 1.58E-08/yr For this analysis, the associated containment leakage for Class 3a is lOLa and for Class 3b is lOOLa. These assignments are consistent with the latest EPRI guidance.

The Class 3b value calculated above does not address the potential impacts associated with containment corrosion. To include corrosion impacts, the Class 3b value can be multiplied by the corrosion impact factor developed in Section 4.4. For the 3-in-10 year ILRT test frequency, the Class 3b frequency accounting for corrosion is 1.58E-08/yr

• 1.004 =

l.59E-08/yr. The increase in the 3-in-10 Class 3b frequency associated with corrosion is approximately 6.32E-11/yr (i.e., 1.58E-08/yr

• 0.004).

Class 4 Sequences This group represents containment isolation failure-to-seal of Type B test components.

Because these failures are detected by Type B tests which are unaffected by the Type A ILRT, this group is not evaluated any further in the analysis.

Class 5 Sequences This group represents containment isolation failure-to-seal of Type C test components.

Because these failures are detected by Type C tests which are unaffected by the Type A ILRT, this group is not evaluated any further in this analysis.

Class 6 Sequences This group is similar to Class 2. These are sequences that involve core damage with a failure-to-seal containment leakage due to failure to isolate the containment. These sequences are dominated by misalignment of containment isolation valves following a test/maintenance 5-5

evolution. Consistent with the EPRI guidance this accident class is not explicitly considered since it has a negligible impact on the results.

Class 7 Sequences This group represents containment failure induced by severe accident phenomena. The frequency and dose of this EPRI accident class is not affected by the ILRT test interval.

However, for the purposes of the population dose calculations, the EPRI Class 7 is divided into 2 subclasses for evaluation, as shown in Table 5.1-3. The total release frequency and total dose are then used to determine a weighted average person-rem for use as the representative EPRI Class 7 dose.

Class 7a represents High Early release sequences (i.e., LERF), except for those associated with PRA accident class V (break outside containment) since these sequences are addressed by EPRI Class 8. The dose associated with these H/E releases is based on the highest adjusted population dose from Table 4.2-2 for early containment failure sequences. The EPRI Class 7a frequency is determined as follows:

Class 7a = CDF H/E - CDF Accident Class V Class 7a = 1.12E-06/yr - 7.12E-07/yr = 4.0SE-07/yr Class 7b represents all other containment failure releases (i.e., non-LERF). The dose associated with these non-LERF releases is based on the population dose from Table 4.2-2 for high magnitude late containment failure. The Hatch specific dose results do not have data for release magnitudes that are in the moderate category or lower. Since EPRI Class 7 is not impacted by the ILRT frequency, however, further evaluation of doses for such releases (e.g.,

using surrogate dose results from another plant) is not warranted. Additional justification for using the late containment failure dose results for non-LERF release categories is provided in the notes to Table 5.1-3. The EPRI Class 7b frequency is determined as follows:

Class 7b = CDF total - CDF H/E (contains BOC) - CDF Intact Class 7b = 7 .57E-06/yr - 1.12E-06/yr - 1. lSE-06/yr = 5.27E-06/yr The EPRI Class 7 frequency is the total of the EPRI Class 7a and 7b frequencies, as shown in Table 5.1-3. The weighted average dose for EPRI Class 7 is also provided in Table 5.1-3.

The EPRI Class 7 fre'quency is 5.68E-06/yr.

5-6

Table 5.1-3 EPRI ACCIDENT CLASS 7 FAILURE FREQUENCIES AND POPULATION DOSES (HATCH UNIT 1 BASE CASE LEVEL 2 MODEL)

POPULATION POPULATION DOSE EPRI ACCIDENT RELEASE DOSE (50 MILES) RISK (50 MILES)

CLASS FREQUENCY/YR PERSON-REM Cl) (PERSON-REM/YR) C2 >

7a (LERF minus BOC) 4.0SE-07 1.08E+06 0.441 7b (non-LERF) 5.27E-06 5.8E+o5c 4> 3.06 3

Class 7 Total 5.68E-06 6.16E+o5< l 3.50 Cl) Population dose obtained from Table 4.2-2 using the adjusted dose values.

<2 l Obtained by multiplying the release frequency value from the second column of this table by the population dose value from the third column of this table.

(3)

The weighted average population dose for Class 7 is obtained by dividing the total population dose risk by the total Class 7 release frequency.

<4 l All Hatch SAMA dose cases represent high releases. EPRI Class 7b is composed of high non-early releases, moderate releases, low and low-low releases. The late containment failure dose is approximately a factor of two less than that for other high magnitude releases and is considered reasonable for use for medium magnitude release cases. This is comparable to SAMA population dose results developed for Quad Cities and Dresden Generating Stations [35] (both Mark I containment designs) where moderate magnitude releases had population dose results approximately one half to nearly equal to high magnitude release population doses. Use of the late containment failure for low and low-low magnitude release cases is acceptable because the associated frequencies for these release categories are very low compared to other release categories, shown previously in Table 4.2-5.

Class 8 Sequences This group represents sequences when containment bypass occurs. For Hatch this frequency is from PRA accident class V with a value of 7.12E-07/yr.

Summary of Accident Class Frequencies In summary, the accident sequence frequencies that can lead to radionuclide release to the public have been derived consistent with the definition of accident classes defined in the EPRI guidance [22], as shown in Table 5.1-4.

5-7

Table 5.1-4 RADIONUCLIDE RELEASE FREQUENCIES AS A *FUNCTION OF EPRI ACCIDENT CLASS (HATCH BASE CASE)

ACCIDENT FREQUENCY FREQUENCY CLASSES. WITHOUT WITH (CONTAINMENT CORROSION CORROSION RELEASE TYPE) DESCRIPTION (PER RX-YR) (PER RX-YR) 1 No Containment Failure 1.lE-06 1.lOE-06 2 Large Isolation Failures (Failure to 4.98E-10 4.98E-10 Close) 3a Small Isolation Failures 6.31E-08 6.31E-08 (containment breach) 3b Large Isolation Failures 1.578E-08 1.584E-08 (containment breach) 4 Small Isolation Failures (Failure to N/A N/A seal - Type B) 5 Small Isolation Failures (Failure to N/A N/A seal-Type C) 6 Other Isolation Failures (e.g.,

N/A N/A dependent failures) 7 Failures Induced by Phenomena 5.68E-06 5.68E-06 8 Bypass (Interfacing System LOCA) 7.12E-07 7.12E-07 CDF All CET End states (including very 7.57E-06 7.57E-06 low and no release) 5-8

5.2 STEP 2 - DEVELOP PLANT-SPECIFIC PERSON-REM DOSE (POPULATION DOSE) PER REACTOR YEAR Plant-specific release analyses were performed to estimate the person-rem doses to the population within a 50-mile radius from the plant. The releases are based on information provided in the Hatch SAMA Analysis and the latest Hatch Level 2 Analysis as described in Section 4.2, and summarized in Table 4.2-2. The results of applying these releases to the EPRI containment failure classification are as follows:

Class 1 = 1.15E+3 person-rem (at 1.0La, as developed in Section 4.2)

Class 2 = 1.17E+6 person-rem (from Table 4.2-2 for containment bypass)

Class 3a = 1.15E+4 person-rem (1.15E+3 x lOLa, per EPRI methodology)

Class 3b 1.15E+5 person-rem (1.15E+3 x lOOLa, per EPRI methodology)

Class 4 = Not analyzed (per EPRI methodology)

Class 5 Not analyzed (per EPRI methodology)

Class 6 = Not analyzed (per EPRI methodology)

Class 7 = 6.16E+5 person-rem (from Table 5.1-3)

Class 8 = 1.17E+6 person-rem (from Table 4.2-2 for containment bypass)

In summary, the population dose estimates derived for use in the risk evaluation per the EPRI methodology [22] are provided in Table 5.2-1.

5-9

Table 5.2-1 HATCH POPULATION DOSE ESTIMATES FOR POPULATION WITHIN 50 MILES ACCIDENT CLASSES (CONTAINMENT PERSON-REM RELEASE TYPE) DESCRIPTION (SO MILES) 1 No Containment Failure (1 La) 1.15E+03 2 Large Isolation Failures (Failure to Close) 1.17E+06 3a Small Isolation Failures (containment breach) 1.15E+04 3b Large Isolation Failures (containment breach) 1.15E+05 4 Small Isolation Failures (Failure to seal -Type B) NA 5 Small Isolation Failures (Failure to seal-Type C) NA 6 Other Isolation Failures (e.g., dependent failures) NA 7 Failures Induced by Phenomena 6.16E+05 8 Bypass (Interfacing System LOCA) 1.17E+06 The above dose estimates when multiplied by the frequency results presented in Table 5.1-4 yield the Hatch baseline mean consequence measures for each accident class. These results are presented in Table 5.2-2.

5-10

Table 5.2-2 HATCH ANNUAL DOSE AS A FUNCTION OF EPRI ACCIDENT CLASS CHARACTERISTIC OF CONDITIONS FOR ILRT REQUIRED 3/10 YEARS EPRI METHODOLOGY EPRI METHODOLOGY PLUS CORROSION CHANGE DUE TO EPRI PERSON- PERSON- CORROSION ACCIDENT PERSON-REM FREQUENCY REM/YR FREQUENCY REM/YR PERSON-CLASSES DESCRIPTION (50 MILES) (PER RX-YR) (50 MILES) (PER RX-YR) (50 MILES) REM/YRC 1 >

1 No Containment Failure (2) 1.15E+03 1.lOlE-06 1.268E-03 1.lOlE-06 1.268E-03 -7.03E-08 2 Large Isolation Failures 1.17E+06 4.98E-10 5.82E-04 4.98E-10 5.82E-04 0.00 (Failure to Close) 3a Small Isolation Failures 1.15E+04 6.31E-08 7.27E-04 6.31E-08 7.27E-04 0.00 (containment breach) 3b Large Isolation Failures 1.lSE+OS 1.578E-08 1.817E-03 1.584E-08 1.824E-03 7.03E-06 (containment breach) 4 Small Isolation Failures NA N/A N/A N/A N/A N/A (Failure to seal -Type B) 5 Small Isolation Failures NA N/A N/A N/A N/A N/A (Failure to seal-Type C) 6 Other Isolation Failures NA N/A N/A N/A N/A N/A 7 Failures Induced by 6.16E+OS 5.68E-06 3.SOE+OO 5.68E-06 3.SOE+OO 0.00 Phenomena 8 Bypass (ISLOCA) 1.17E+06 7.12E-07 8.33E-01 7.12E-07 8.33E-01 0.00 CDF All CET end states -- 7.57E-06 4.34E+OO 7.57E-06 4.34E+OO 6.96E-06 1

< > Only release Classes 1 and 3b are affected by the corrosion analysis.

2

<> Characterized as lla release magnitude consistent with the derivation of the ILRT non-detection failure probability for ILRTs. Release classes 3a and 3b include failures of containment to meet the Technical Specification leak rate.

5-11

5.3 STEP 3 - EVALUATE RISK IMPACT OF EXTENDING TYPE A TEST INTERVAL FROM 10-T0-15 YEARS The next step is to evaluate the risk impact of extending the test interval from a ten-year value to fifteen-years. To do this, an evaluation must first be made of the risk associated with the ten-year interval since the base case applies to a 3-year interval (i.e. a simplified representation of a 3-in-10 interval).

Risk Impact Due to 10-year Test Interval As previously stated, Type A tests impact only Class 3 sequences. For Class 3 sequences, the release magnitude is not impacted by the change in test interval (a small or large breach remains the same, even though the probability of not detecting the breach increases). Thus, only the frequency of Class 3a and 3b sequences is impacted. The risk contribution is changed based on the EPRI guidance [22] as described in Section 4.3 by a factor of 3.33 compared to the base case values. The results of the calculation for a 10-year interval are presented in Table 5.3-1.

Risk Impact Due to 15-Year Test Interval The risk contribution for a 15-year test interval is calculated in a manner similar to the 10-year interval. The difference is in the increase in probability of leakage in Classes 3a and 3b. For this case, the value used in the analysis is a factor of 5.0 compared to the 3-year interval value, as described in Section 4.3. The results for this calculation are presented in Table 5.3-2.

As noted in Section 3.0, these calculations were performed in a spreadsheet and columns totals may differ slightly from hand calculations due to rounding.

5-12

Table 5.3-1 HATCH ANNUAL DOSE AS A FUNCTION OF EPRI ACCIDENT CLASS CHARACTERISTIC OF CONDITIONS FOR ILRT REQUIRED 1/10 YEARS EPRI METHODOLOGY EPRI METHODOLOGY PLUS CORROSION CHANGE DUE TO FREQUENCY PERSON- FREQUENCY PERSON-EPRI PERSON- CORROSION (PER RX-YR) REM/YR (PER RX-YR) REM/YR ACCIDENT REM PERSON-(50 MILES) (50 MILES)

CLASSES DESCRIPTION (50 MILES) REM/VR<ll 1 No Containment Failure (2) 1.15E+03 9.171E-07 1.056E-03 9.167E-07 1.056E-03 -4.04E-07 2 Large Isolation Failures 1.17E+06 4.98E-10 5.82E-04 4.98E-10 5.82E-04 0.00 (Failure to Close) 3a Small Isolation Failures 1.15E+04 2.10E-07 2.42E-03 2.10E-07 2.42E-03 0.00 (containment breach) 3b Large Isolation Failures 1.15E+05 5.258E-08 6.055E-03 5.294E-08 6.096E-03 4.04E-O~

(containment breach) 4 Small Isolation Failures (Failure to seal -Type B)

NA NA NA NA NA NA 5 Small Isolation Failures (Failure to seal-Type C)

NA NA NA NA NA NA 6 Other Isolation Failures (e.g., dependent failures)

NA NA NA NA NA NA

\

7 Failures Induced by 6.16E+05 5.68E-06 3.50E+OO 5.68E-06 3.50E+OO 0.00 Phenomena 8 Bypass (ISLOCA) 1.17E+06 7.12E-07 8.33E-01 7.12E-07 8.33E-01 0.00 CDF All CET end states -- 7.57E-06 4.34E+OO 7.57E-06 4.34E+OO 4.00E-05

<1 J Only release classes 1 and 3b are affected by. the corrosion analysis.

2 C J Characterized as 1La release magnitude consistent with the derivation of the ILRT non-detection failure probability for ILRTs. Release Classes 3a and 3b include failures of containment to meet the Technical Specification leak rate.

5-13

Table 5.3-2 HATCH UNIT 1 ANNUAL DOSE AS A FUNCTION OF ACCIDENT CLASS CHARACTERISTIC OF CONDITIONS FOR ILRT REQUIRED 1/15 YEARS EPRI METHODOLOGY EPRI METHODOLOGY PLUS CORROSION CHANGE DUE TO EPRI FREQUENCY PERSON-REM/YR FREQUENCY PERSON- CORROSION ACCIDENT PERSON-REM (PER RX- (SO MILES) (PER RX-YR) REM/YR PERSON-CLASSES DESCRIPTION (SO MILES) YR) (SO MILES) REM/YR< 1 J 1 No Containment Failure (2) 1.15E+03 7.856E-07 9.046E-04 7.848E-07 9.037E-04 -9.28E-07 2 Large Isolation Failures 1.17E+06 4.98E-10 5.82E-04 4.98E-10 5.82E-04 0.00 (Failure to Close) 3a Small Isolation Failures 1.15E+04 3.16E-07 3.63E-03 3.16E-07 3.63E-03 0.00 (containment breach) 3b Large Isolation Failures 1.15E+05 7.888E-08 9.084E-03 7.969E-08 9.177E-03 9.28E-05 (containment breach) 4 Small Isolation Failures (Failure to seal -Type B)

NA NA NA NA NA NA 5 Small Isolation Failures (Failure to seal-Type C)

NA NA NA NA NA NA 6 Other Isolation Failures (e.g., dependent NA NA NA NA NA NA failures) 7 Failures Induced by 6.16E+05 5.68E-06 3.50E+OO 5.68E-06 3.50E+OO 0.00 Phenomena 8 Bypass (ISLOCA) 1.17E+06 7.12E-07 8.33E-01 7.12E-07 li.33E-01 0.00 CDF All CET end states -- 7.57E-06 4.35E+OO 7.57E-06 4.35E+OO 9.19E-05 Ct> Only release classes 1 and 3b are affected by the corrosion analysis.

2 C > Characterized as 1La release magnitude consistent with the derivation of the ILRT non-detection failure probability for ILRTs. Release Classes 3a and 3b include failures of containment to meet the Technical Specification leak rate.

5-14

Table 5.3-1 HATCH ANNUAL DOSE AS A FUNCTION OF EPRI ACCIDENT CLASS CHARACTERISTIC OF CONDITIONS FOR ILRT REQUIRED 1/10 YEARS EPRI METHODOLOGY EPRI METHODOLOGY PLUS CORROSION CHANGE DUE TO FREQUENCY PERSON- FREQUENCY PERSON-REM/YR EPRI CORROSION (PER RX-YR) REM/YR (PER RX- (SO MILES)

ACCIDENT PERSON-REM PERSON-(SO MILES) YR)

CLASSES DESCRIPTION (SO MILES) REM/YRC1J 3b Large Isolation Failures 1.15E+05 7.BBBE-08 9.084E-03 7.969E-08 9.177E-03 9.28E-05 (containment breach) 4 Small Isolation Failures (Failure NA NA NA NA NA NA to seal -Type B) 5 Small Isolation Failures (Failure NA NA NA NA NA NA to seal-Type C) 6 Other Isolation Failures (e.g.,

dependent NA NA NA NA NA NA failures) 7 Failures Induced 6.16E+05 5.68E-06 3.50E+OO 5.68E-06 3.50E+OO 0.00 by Phenomena 8 Bypass (ISLOCA) 1.17E+06 7.12E-07 8.33E-01 7.12E-07 8.33E-01 0.00 CDF All CET end states -- 7.57E-06 4.35E+OO 7.57E-06 4.35E+OO 9.19E-05

<1J Only release classes 1 and 3b are affected by the corrosion analysis.

cii Characterized as lla release magnitude consistent with the derivation of the ILRT non-detection failure probability for ILRTs. Release Classes 3a and 3b include failures of containment to meet the Technical Specification leak rate.

5-15

5.4 STEP 4 - DETERMINE THE CHANGE IN RISK IN TERMS OF LARGE EARLY RELEASE FREQUENCY The risk increase 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 an intact containment could in fact result in a larger release due to the increase in probability of failure to detect a pre-existing leak. With strict adherence to the EPRI guidance, 100% of Class 3b contribution is considered LERF.

Regulatory Guide 1.174 provides guidance for determining the risk impact of plant-specific changes to the licens,ing basis. RG 1.174 defines verv small changes in ris!,<, as resulting in increases of core damage frequency (CDF) below lE-6/yr and increases in LERF below lE-7/yr, and small changes in LERF as below lE-6/yr. Because the ILRT does not impact CDF, the relevant metric is LERF.

For Hatch, 100% of the frequency of Class 3b sequences can be used as a conservative first-order estimate to approximate the potential increase in LERF from the ILRT interval extension (consistent with the EPRI guidance methodology). Based on the original 3/10 year test interval assessment from Table 5.2-2, the Class 3b frequency is 1.584E-08/yr, which includes the corrosion effect of the containment steel. Based on a ten year test interval from Table 5.3-1, the Class 3b frequency is 5.294E-08/yr; and, based on a fifteen year test interval from Table 5.3-2, it is 7.969E-08/yr. Thus, the increase in the overall probability of LERF due to Class 3b sequences that is due to increasing the ILRT test interval from 3.3 years (i.e., 3 tests in 10 years) to 15 years (including corrosion effects) is 6.39E-08/yr. Similarly, the increase due to increasing the interval from 10 to 15 years (including corrosion effects) is 2.68E-08/yr.

As can be seen, even with the conservatisms included in the evaluation (per the EPRI methodology), the estimated change in LERF is below the threshold criteria for a very small change in risk when comparing the 15 year results to the current 10-year requirement, and also to the original 3-in-10 year requirement.

5-16

5.5 STEP 5 - DETERMINE THE IMPACT ON THE CONDITIONAL CONTAINMENT FAILURE PROBABILITY Another parameter that the NRC guidance in RG 1.174 states can provide input into the decision-making process is the change in the conditional containment failure probability (CCFP). The change in CCFP is indicative of the effect of the ILRT on all radionuclide releases, not just LERF. The CCFP can be calculated from the results of this analysis. One of the difficult aspects of this calculation is providing a definition of the "failed containment." In this assessment, the CCFP is defined such that containment failure includes all radionuclide release end states other than the intact state. The conditional part of the definition is conditional given a severe accident (i.e., core damage).

The change in CCFP can be calculated by using the method specified in the EPRI guidance. The NRC has previously accepted similar calculations [7] as the basis for showing that the proposed change is consistent with the defense-in-depth philosophy. Including the impacts of corrosion, the change in CCFP is calculated as follows:

CCFP = [1 - (Class 1 frequency + Class 3a frequency) I CDF]

• 100%

CCFP311o = [1 - (1.lOlE-06/yr + 6.31E-08/yr) / 7.57E-06/yr]

• 100% = 84.63%

CCFP 10 = [1 - (9.167E-07/yr + 2.lOE-07/yr) / 7.57E-06/yr]

• 100% = 85.12%

CCFP 15 = [1 - (7.848E-07/yr + 3.16E-07/yr) / 7.57E-06/yr]

• 100% = 85.47%

CCFP CCFP CCFP LlCCFP1s-3/1D LlCCFP1s-10 3 IN 10 YRS 1 IN 10 YRS 1 IN 15 YRS 84.63°/o 85.12% 85.47% 0.84% 0.35%

The EPRI guidance specifies a criterion that the change in CCFP be less than or equal to 1.5%

for impacts to be considered small. The change in CCFP of 0.84% as a result of extending the test interval to 15 years from the original 3-in-10 year requirement is less than the 1.5%

criterion for small risk impacts.

5-17

5.6 STEP 6 - DETERMINE THE IMPACT ON THE POPULATION DOSE RISK The total annual dose risk (person-rem/yr) within 50 miles is examined to demonstrate the relative change in this parameter. Per the EPRI methodology, a very small population dose is defined as an increase from the baseline interval (3 tests per 10 years) dose of<= 1.0 person-rem/yr or 1 % of the total baseline dose, whichever is less restrictive for the risk impact assessment of the proposed extended ILRT interval.

Table 5. 7-1 summarizes the ILRT results and presents the change in population dose risk of 9.90E-3 person-rem/yr for a change in the test frequency from 3-in-10 years to 1-in-15 years.

This change in population dose risk is approximately two orders of magnitude less than the 1.0 person-rem criterion (the less restrictive criterion) and is considerably less than the 1 % of the baseline dose value of 4.34 person-rem/yr (i.e., 9.90E-3 person-rem/yr < 4.34E-2 person-rem/yr). Therefore, the change in population dose risk for Hatch is deemed to represent a very small impact.

5.7

SUMMARY

OF INTERNAL EVENTS RESULTS The results from this ILRT extension risk assessment for Hatch Unit 1 are summarized in Table 5.7-1.

5-18

Table 5.7-1 HATCH ILRT CASES BASE, 1/10, AND 1/15 YR EXTENSIONS (INCLUDING AGE ADJUSTED STEEL CORROSION LIKELIHOOD)

BASE CASE EXTEND TO EXTEND TO 3 IN 10 YEARS 1 IN 10 YEARS 1 IN 15 YEARS DOSE RISK DOSE RISK DOSE RISK EPRI DOSE FREQ (PER- FREQ (PER- FREQ (PER-CLASS (PER-REM) (/YR) REM/YR) (/YR) REM/YR) (/YR) REM/YR) 1 1.15E+03 1.101 E-06 1.268E-03 9.167E-07 1.056E-03 7.848E-07 9.037E-04 2 1.17E+06 4.98E-10 5.82E-04 4.98E-10 5.82E-04 4.98E-10 5.82E-04 3a 1.15E+04 6.31E-08 7.27E-04 2.10E-07 2.42E-03 3.16E-07 3.63E-03 3b 1.15E+05 1.584E-08 1.824E-03 5.294E-08 6.096E-03 7.969E-08 9.177E-03 7 6.16E+05 5.68E-06 3.50E+OO 5.68E-06 3.50E+OO 5.68E-06 3.50E+OO 8 1.17E+06 7.12E-07 8.33E-01 7.12E-07 8.33E-01 7.12E-07 8.33E-01 Total 7.57E-06 4.34E+OO 7.57E-06 4.34E+OO 7.57E-06 4.35E+OO ILRT Dose Risk from 2.55E-03 8.52E-03 1.28E-02 3a and 3b Delta From 3 yr --- 5.76E-03 9.90E-03 Total Dose From 10 yr

--- --- 4.14E-03 Risk< 1 >

3b Frequency (LERF) 1.58E-08 5.29E-08 7.97E-08 Delta From 3 yr --- 3.71E-08 6.39E-08 LERF From 10 yr --- --- 2.68E-08 CCFP % 84.63% 85.12% 85.47%

Delta From 3 yr --- 0.49% 0.84%

• CCFP %

From 10 yr --- --- 0.35%

Cll The overall difference in total dose risk is less than the difference of only the 3a and 3b categories between two testing intervals. This is because the overall total dose risk includes contributions from other categories that do not change as a function of time, e.g.,

the EPRI Class 2 and 8 categories, and also due to the fact that the Class 1 person-rem/yr decreases when extending the IRLT frequency.

5-19

5.8 EXTERNAL EVENTS CONTRIBUTION Because the LERF risk acceptance guidelines in RG-1.174 are intended for comparison with a full-scope assessment of risk, including internal and external events, an evaluation of the potential impact on LERF from external events is presented.

The method chosen to account for external event contributions is similar to that used in many SAMA analyses where a multiplier on the internal events results is utilized based on the IPEEE results. Although Hatch has completed a SAMA analysis [9], external events were not quantitatively included (i.e., not included in the CDF metric). The NRC, via the RA! process, deemed SNC's consideration of external events for the purposes of SAMA acceptable [25]. For the ILRT assessment, Hatch IPEEE [18] quantitative results are utilized to provide additional perspective on potential risk impacts.

Although Hatch is preparing a license amendment request (LAR) to transition to NFPA 805 the Fire PRA is not yet complete, therefore for this ILRT test interval extension risk assessment the results from the IPEEE will be used. The contributions of the external events from the Hatch IPEEE analysis are summarized in Table 5.8-1. The Unit 1 Internal Events CDF at the time of the IPEEE was 2.1E-05/yr [18]. From Table 5.8-1, the external events multiplier could be calculated as external events CDF divided by the internal events CDF as (8.51E-06/yr) I (2.1E-05/yr) = 0.41. This, however, includes no quantitative consideration of seismic events.

Regarding seismic considerations, the IPEEE identified that the extensive evaluation of the design and location of Plant Hatch found no fundamental weakness or vulnerability to seismic hazards. The IPEEE note~ that no major or minor fault zones are near the site and the region is characterized as one of infrequent seismic activity. While no major plant changes were determined to be necessary, the IPEEE seismic analysis identified modifications of certain Unit 1 and Unit 2 components that were necessary to obtain a high-confidence-low-probability-of-failure capacity of at least 0.3g peak ground acceleration. Modifications of these items (itemized in Appendix I of the IPEEE) were completed in 1995.

For purposes of the ILRT analysis, external event CDF contribution from seismic contributors is initially estimated (for quantitative evaluation purposes) as being equivalent to the contribution from internal fires (i.e., 7.5E-06/yr). Considering the historically low seismic activity in the region of Plant Hatch, this is considered to be a conservative assumption. Seismic PRAs have typically found that the relative proportion of LERF associated with seismic hazards is 5-20

significantly larger than the relative proportion of LERF associated with internal events due to impacts of the seismic event upon containment (i.e., the seismic event fails containment integrity). The seismic evaluation for Peach Bottom [26] (also a Mark I containment design) identifies in Section 2.5.5.9 that the probability of early containment failure for seismically initiated events is high (70% or greater). This is driven by the nature of the seismic event which defeats AC power recovery and the characteristics of the dominant plant damage states where vessel injection or containment heat removal are failed, leading to a high probability of drywell melt-through. For the ILRT risk assessment, early containment failures result in LERF independent of the considerations associated with EPRI Class 3b. Applying the Peach Bottom LERF proportion results to Plant Hatch, the seismic CDF may be reduced by 70% to account for this independent LERF (consistent with subtracting EPRI Class 2 and Class 8 from the internal events CDF when calculating the EPRI Class 3b frequency, discussed in Section 4.3).

Accounting for this seismically induced independent LERF results in a remaining estimated seismic CDF of 2.25E-6/yr (i.e., 7.5E-6/yr

• 0.3).

Including this 2.25E-6/yr estimate for seismic contributors, along with the IPEEE maximum CDF estimate for high winds and external floods, provides a multiplier value of 0.51 (i.e., 1.08-05/yr / 2.lE-05/yr).

The EPRI Category 3b frequencies based on internal events for the 3-per-10 year, 1-per-10 year, and 1-per-15 year ILRT intervals (including corrosion impacts) are shown in Table 5.8-2.

The change in the LERF risk measure due to extending the ILRT from 3-per-10 years to 1-per-15 years, including both internal and external hazard risk, is estimated as shown in Table 5.8-2. The total increase in LERF (measured from the original 3 per 10 years required to the proposed 1 per 15 years performance of the ILRT) due to the combined internal and external events contribution is estimated as 9.5E-08/yr. Once again, it is noted that this quantitative estimate is judged conservative due to the assumption that seismic CDF contributors are equal to fire CDF contributors.

5-21

TABLE 5.8-1 IPEEE CONTRIBUTOR

SUMMARY

EXTERNAL EVENT TYPE CDFC 1 > (/YR)

Seismic Not calculatedC 2 >

Internal Fire 7.SE-06 (Unit 1) <3 >

High Winds < 1.0E-06 External Floods < 1.0E-08 Transportation and Nearby Facility Not calculated< 4 >

Accidents Total (for initiators with calculated CDF) < 8.51E-06 Cl) From Section 1.4 of the Hatch IPEEE [18]

C2> Seismic margins analysis utilized.

C3 l Internal fire Unit 1 CDF is approximately 36% of the Internal Events CDF value of 2.lE-05/yr at the time of the IPEEE. Fire CDF for Unit 2 was lower, at 5.4E-06/yr, approximately 25% of the Internal Events CDF value of 2.2E-05/yr.

C4 l Hatch conforms to the Standard Review Plan for transportation and nearby facility accidents. No potential vulnerabilities attributed to these causes were identified.

Table 5.8-2 HATCH CLASS 3b (LERF) AS A FUNCTION OF ILRT FREQUENCY FOR INTERNAL AND EXTERNAL EVENTS (INCLUDING AGE ADJUSTED STEEL CORROSION LIKELIHOOD) 38 38 38 FREQUENCY FREQUENCY FREQUENCY (3-PER-10 YR (1-PER-10 YEAR (1-PER-15 ILRT) ILRT) YEAR ILRT) LERF INCREASEC 1 >

Internal Events ContributionC 2 > 1.584E-08 5.294E-08 7.969E-08 6.39E-08 External Events Contribution 7.70E-09 2.57E-08 3.87E-08 3.lOE-08 (Internal Events x 0.51)

Combined (Internal + 2.28E-08 7.61E-08 1.15E-07 9.5E-08 External)

Cll Associated with the change from the 3-per-10 year frequency to the proposed 1-per-15 year frequency.

C2 > Values from Table 5.7-1.

5-22

NRC Regulatory Guide 1.174 provides NRC recommendations for using risk information in support of applications requesting changes to the license basis of the plant. As discussed in Section 2 of this risk assessment, the risk acceptance criteria of RG 1.174 are used here to assess the ILRT interval extension.

The 9.5E-08/yr increase in LERF due to the combined internal and external events from extending the Hatch ILRT frequency from 3-per-10 years to 1-per-15 years remains in RG 1.174 Region III, below lE-7 per reactor year (i.e., "very small" change in risk). Thus the inclusion of external events does not change the conclusion that extending the ILRT test interval to 15 years results in a very small change in risk in regards to LERF criterion.

5-23

6.0 SENSITIVITIES A number of sensitivity cases are performed to examine the sensitivity of the results to specific assumptions and parameters. These include an examination of the following:

Assumptions in the corrosion analysis Likelihood of undetected leaks within containment (expert elicitation)

Impact of removing late releases (i.e., non-early releases) from EPRI Class 3b Consideration of the risk benefits of an extended ILRT interval 6.1 SENSITIVITY TO CORROSION IMPACT ASSUMPTIONS The results in Tables 5.2-2, 5.3-1, and 5.3-2 show that including corrosion effects calculated using the assumptions described in Section 4.4 does not significantly affect the results of the ILRT interval extension risk assessment.

Consistent with the EPRI guidance, sensitivity cases were developed to gain an understanding of the sensitivity of the results to the key parameters in the corrosion risk analysis. The time for the flaw *likelihood to double was adjusted from every five years to every two and every ten years. The failure probabilities for the were increased and decreased by an order of magnitude.

The total detection failure likelihood was adjusted from 10% to 15% and 5%. The results are presented in Table 6.1-1. In every case, the impact from including the corrosion effects is very small.

6-1

Table 6.1-1 STEEL CORROSION SENSITIVITY CASESC 1 >

VISUAL INCREASE IN CLASS 3B FREQ (LERF)

INSPECTION FOR ILRT EXTENSION

& NON- FROM 3/10 TO 1/15 YEARS VISUAL (PER YR)

- AGE (STEP 3 CONTAINMENT FLAWS IN THE BREACH (STEP 5 IN CORROSION (STEP 4 IN THE THE ANALYSIS) CORROSION CORROSION INCREASE DUE TO ANALYSIS) ANALYSIS) TOTAL INCREASE CORROSION Base Case (1.0% Base Case (10%

Base Case DW walls, head, DW walls, head, Doubles every &WW; &WW; 6.39E-08 7.45E-10 5 yrs 0.1% Floor) 100% Floor)

Doubles every Base Base 6.44E-08 1.34E-09 2 yrs Doubles every Base Base 6.36E-08 4.87E-10 10 yrs 15% DW walls, Base Base 6.39E-08 8.38E-10 head, & WW 5% DW walls, Base Base 6.35E-08 3.59E-10 head, & WW 10% DW walls, Base head, & WW; Base 6.91E-08 5.99E-09 1% Floor 0.1% DW walls, Base head, & WW; Base 6.32E-08 5.99E-11 0.01% Floor LOWER BOUND 0.1 % DW walls, 5% DW walls, Doubles every head, & WW; head, & WW; 6.31E-08 2.92E-11 10 yrs 0.01% Floor 100% Floor UPPER BOUND 10% DW walls, 15% DW walls, Doubles every head, & WW; head, & WW; 8.18E-08 1.87E-08 2 yrs 1% Floor 100% Floor (l) The containment corrosion sensitivities are performed in the same manner as the base case

'(as shown in Table 4.4-1), but the detailed calculations for these sensitivities are not shown here.

6-2

6.2 EPRI EXPERT ELICITATION SENSITIVITY EPRI facilitated an expert elicitation to reduce excess conservatisms in the data associated with the probability of undetected leaks within containment [22]. Because the risk impact assessment of the extensions to the ILRT interval is sensitive to both the probability of the leakage as well as the magnitude, it was decided to perform the expert elicitation in a manner to solicit the probability of leakage as a function of leakage magnitude. In addition, the elicitation was performed for a range of failure modes which allowed experts to account for the range of mechanisms of failure, the potential for undiscovered mechanisms, un-inspectable areas of the containment as well as the potential for detection by alternate means. The expert elicitation process has the advantage of considering the available data for small leakage events, which have occurred in the data, and extrapolate those events and probabilities of occurrence to the potential for large magnitude leakage events.

The basic difference in the application of the ILRT interval methodology using the expert elicitation is a change in the probability of pre-existing leakage in the containment. The basic methodology uses the Jeffrey's non-informative prior for the large leak size and the expert elicitation sensitivity study uses the results of the expert elicitation. In addition, given the relationship between leakage magnitude and probability, larger leakage that is more representative of large early release frequency, can be reflected. For the purposes of this sensitivity, the same leakage magnitudes that are used in the basic methodology (i.e. 10 La for small and 100 La for large) are used here. Table 6.2-1 illustrates the magnitudes and probabilities of a pre-existing leak in containment associated with the base case and the expert elicitation statistical treatments. These values are used in the ILRT interval extension for the base methodology and in this sensitivity case. Details of the expert elicitation process, and the input to expert elicitation as well as the results of the expert elicitation, are available in the various appendices of the EPRI report [22].

6-3

Table 6.2-1 EPRI EXPERT ELICITATION RESULTsC 1 >

EXPERT ELICITATION MEAN PROBABILITY PERCENT LEAKAGE SIZE (LA) BASE CASE OF OCCURRENCE REDUCTION 10 9.2E-03 3.88E-03 58%

100 2.3E-03 2.47E-04 89%

1

<J Data taken from Table D-1 of the EPRI ILRT Guidance [22]

. ~ . .

A summary of the results using the expert elicitation values for the probability of containment leakage is provided in Table 6.2-2. As mentioned previously, probability values are those associated with the magnitude of the leakage used in the base case evaluation (10La for small and 100La for large). The expert elicitation process produces a probability versus leakage magnitude relationship and it is possible to assess higher leakage magnitudes more reflective of large early releases but these evaluations are not performed in this study.

The net effect is that the reduction in the multipliers shown above has a similar impact on the calculated increases in the LERF values. The increase in the overall probability of LERF due to Class 3b sequences that is due to increasing the ILRT test interval from 3 to 15 years is 7.52E-09/yr (about 88% reduction relative to the base model assumptions accepted by the NRC SER[32]). Similarly, the increase due to increasing the interval from 10 to 15 years is 3.28E-09/yr (about 88% reduction relative to the base model assumptions accepted by the NRC SER [32]). As such, if the expert elicitation mean probabilities of occurrence are used instead of the non-informative prior estimates, the change in LERF for Hatch is nearly an order of magnitude lower. The results of this sensitivity study are judged to be more indicative of the actual risk associated with the ILRT extension than the results from the assessment as dictated by the EPRI methodology values, and yet are still conservative given the assumption that all of the Class 3b contribution is considered to be LERF.

6-4

Table 6.2-2 HATCH EPRI SENSITIVITY ILRT CASES BASE, 1/10, AND 1/15 YR EXTENSIONS (BASED ON EPRI EXPERT ELICITATION LEAKAGE PROBABILITIES)

BASE CASE EXTEND TO EXTEND TO 3 IN 10 YEARS 1IN10 YEARS 1 IN 15 YEARS EPRI DOSE PER- PER- PER-CLASS PER-REM CDF/YR REM/YR CDF/YR REM/YR CDF/YR REM/YR 1.15E-06 1.25E-03 1.04E-06 1.19E-03 1

1.15E+03 1.33E-03 7.03E-06 4.98E-10 5.82E-04 4.98E-10 5.82E-04 2

1.17E+06 4.98E-10 5.82E-04 8.87E-08 1.02E-03 1.33E-07 1.53E-03 3a 1.15E+04 2.66E-08 3.06E-04 6.00E-09 6.91E-04 9.28E-09 1.07E-03 3b 1.15E+05 1.76E-09 2.02E-04 6.16E+05 3.50E+OO 3.50E+OO 7

5.68E-06 3.50E+OO 5.68E-06 5.68E-06 7.12E-07 7.12E-07 8

1.17E+06 7.12E-07 8.33E-01 8.33E-01 8.33E-01 7.57E-06 7.57E-06 4.34E+OO Total 7.57E-06 4.33E+OO 4.33E+OO ILRT Dose Risk from 3a and 3b 5.09E-04 1.71E-03 2.60E-03 Delta From 3 yr --- 1.13E-03 1.96E-03 Total Dose From 10 yr

--- --- 8.34E-04 Risk< 1 >

3b Frequency (LERF) 1.76E-09 6.00E-09 9.28E-09 Delta From 3 yr --- 4.24E-09 7.52E-09 LERF From 10 yr --- --- 3.28E-09 CCFP % 84.44% 84.50% 84.54%

Delta From 3 yr --- 0.06% 0.10%

- CCFP %

From 10 yr --- --- 0.04%

(lJ The overall difference in total dose risk is less than the difference of only the 3a and 3b categories between two testing intervals. This is because the overall total dose risk includes contributions from other categories that do not change as a function of time, e.g., the EPRI Class 2 and Class 8 categories, and also due to the fact that the Class 1 person-rem/yr decreases when extending the ILRT frequency.

6-5

6.3 NON-EARLY RELEASE SENSITIVITY As noted in Section 4.3, the EPRI guidance and examples indicate that accident sequences resulting in late releases may also be subtracted from the CDF applied to EPRI Class 3b (in addition to subtracting EPRI Classes 2 and 8 which always result in LERF). Removing late release contributors was conservatively not incorporated in the base case analysis, but is examined in this sensitivity case.

In the Hatch Level 2 analysis, accident sequences associated with loss of containment heat removal (i.e., accident classes IIA, IIL), and long-term station blackout (i.e., accident class IBL) are determined* to be non-early releases. The CDF associated* with* these non-early releases is identified in Table 4.2-lb.

The Class 3 EPRI frequencies may be recalculated as follows:

Class 3a = 0.0092 * (CDF - Class 2 - Class 8 - Class IBL - Class IIA - Class IIL)

= 0.0092 * (7.57E-06/yr - 4.98E-10/yr - 7.12E-07/yr - 4.89E-07/yr -

3.39E-06/yr - 4.llE-10/yr)

= 2.74E-08/yr Class 3b = 0.0023 * (CDF - Class 2 - Class 8 - Class IBL - Class IIA - Class IIL))

= 0.0023 * (7.57E-06/yr - 4.98E-10/yr - 7.12E-07/yr - 4.89E-07/yr -

3.39E-06/yr - 4.llE-10/yr)

= 6.85E-09/yr Using these new EPRI Class 3a and 3b frequencies and repeating the calculations using the EPRI methodology [2_2] provide the results summarized in Table 6.3-1.

Removing those accident sequences from EPRI Class 3a and 3b that are non-early contributors results in a reduction, by a factor of approximately 2.3, in the change in internal events LERF (i.e., from 6.39E-8/yr (see Table 5. 7-1) to 2. 77E-8/yr), in the change in population dose risk (i.e., from 9.90E-3 per-rem/yr to 4.30E-3 per-rem/yr), and in the change in CCFP (i.e., from 0.84% to 0.37%) for a change in the ILRT interval from 3 years to 15 years. This sensitivity case further substantiates the conclusion that the change in the ILRT interval results in a very small risk impact.

6-6

Table 6.3-1 HATCH NON-EARLY SENSITIVITY ILRT CASES BASE, 1/10, AND 1/15 YR EXTENSIONS (INCLUDING AGE ADJUSTED STEEL CORROSION LIKELIHOOD)

BASE CASE EXTEND TO EXTEND TO 3 IN 10 YEARS 1IN10 YEARS 1IN15 YEARS EPRI DOSE PER- PER- PER-CLASS PER-REM CDF/YR REM/YR CDF/YR REM/YR CDF/YR REM/YR 1.15E-06 1.32E-03 1.23E-03 1.01 E-06 1.16E-03 1

1.15E+03 3.0SE-06 4.98E-10 5.82E-04 4.98E-10 5.82E-04 4.98E-10 5.82E-04 2

1.17E+06 3.16E-04 9.14E-08 1.0SE-03 1.37E-07 1.58E-03 3a 1.15E+04 2.74E-08 6.88E-09 7.92E-04 2.30E-08 2.65E-03 3.46E-08 3.99E-03 3b 1.15E+05 6.16E+05 5.68E-06 3.SOE+OO 5.68E-06 3.SOE+OO 5.68E-06 3.SOE+OO 7

8 1.17E+06 7.12E-07 8.33E-01 7.12E-07 8.33E-01 7.12E-07 8.33E-01 4.33E+OO 4.34E+OO 4.34E+OO Total

-- 7.57E-06 7.57E-06 7.57E-06 ILRT Dose Risk from 3a and 3b 1.11E-03 3.70E-03 5.57E-03 Delta From 3 yr --- 2.SOE-03 4.30E-03 Total Dose From 10 yr

--- --- 1.80E-03 RiskC 1 >

3b Frequency (LERF) 6.88E-09 2.30E-08 3.46E-08 Delta From 3 yr --- 1.61E-08 2.77E-08 LERF From 10 yr --- --- 1.16E-08 CCFP % 84.51% 84.72% 84.87%

Delta From 3 yr --- 0.21% 0.37%

- CCFP %

From 10 yr --- --- 0.15%

Cll The overall difference in total dose risk is less than the difference of only the 3a and 3b categories between two testing intervals. This is because the overall total dose risk includes contributions from other categories that do not change as a function of time, e.g., the EPRI Class 2 and Class 8 categories, and also due to the fact that the Class 1 person-rem/yr decreases when extending the ILRT frequency.

6-7

6.4 ILRT EXTENSION RISK BENEFIT The performance of an ILRT occurs during plant shutdown and introduces some small risk for the shutdown configuration. As identified previously, an EPRI study [8] of operating experience events quantified the impact of extending the ILRT and LLRT test intervals on shutdown risk. The study concluded that a small but measurable safety benefit is realized for extending the test intervals from 3 per 10 years to 1 per 10 years. The study showed that shutdown CDF was reduced by lE-8/yr to lE-7/yr.

This CDF safety benefit is not amenable to direct inclusion in the EPRI ILRT quantitative methodology for the at-power plant configuration which focuses on the LERF metric, but the safety benefit is noted here for completeness.

6-8

7 .0 CONCLUSIONS Based on the results from Section 5 and the sensitivity calculations presented in Section 6, the following conclusions regarding the assessment of the plant risk are associated with extending the Type A ILRT test frequency to fifteen years for Hatch Units 1 and 2C 1):

• Regulatory Guide (RG) 1.174 [4] provides guidance for determining the risk impact of plant-specific changes to the licensing basis. RG 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. Because the ILRT does not impact CDF, the relevant criterion is LERF. The increase in internal events LERF resulting from a change in the Type A ILRT test frequency from three in ten years to one in fifteen years is conservatively estimated as 6.39E-08/yr using the EPRI guidance as written and including potential corrosion impacts. The LERF increase using the EPRI Expert Elicitation values is substantially less (i.e., 7.52E-09/yr). Using both approaches, the estimated change in LERF is determined to be "very small" using the acceptance guidelines of Reg. Guide 1.174.

• The change in Type A test frequency from three in ten years to one in fifteen years, measured as an increase in the total integrated plant dose risk for those accident sequences influenced by Type A testing, is 9.90E-03 person-rem/yr using the EPRI guidance values, and drops to 1.96E-03 person-rem/yr using the EPRI Expert Elicitation values. The EPRI guidance states that a very small population dose is defined as an increase of <1.0 person-rem/yr or <1% of the total population dose, whichever is less restrictive for the risk impact assessment of the extended ILRT intervals. The Hatch dose increase results are significantly less than 1.0 person-rem/yr. Moreover, the risk impact when compared to other severe accident risks is negligible.

• The increase in the conditional containment failure probability (CCFP) when comparing the three in ten year frequency to one in fifteen year frequency is about 0.84% using the EPRI guidance values, and drops to about 0.10% using the EPRI Expert Elicitation values. The EPRI guidance states that increases in CCFP < 1.5 percentage points are very small.

Therefore the increase for Hatch is determined to be very small.

• The potential impact on LERF from external events was quantitatively assessed utilizing information from the IPEEE. The total increase in LERF due to internal and external events using the EPRI guidance values is estimated to be 9.5E-08/yr, which remains in the "very small" change region (i.e., < lE-7/yr) of the RG 1.174 acceptance guidelines.

Therefore, increasing the ILRT interval to 15 years is considered to represent an insignificant change in risk for Plant Hatch.

Cl) The ILRT risk assessment specifically quantifies the risk impacts associated with Hatch Unit 1.

Due to the similarity of Hatch Units 1 and 2, however, the risk impacts and therefore the conclusions are considered equally applicable to Unit 2.

7-1

Previous Assessments The NRC in NUREG-1493 [5] has previously concluded that:

• Reducing the frequency of Type A tests (ILRTs) from three per 10 years to one per 20 years was found to lead to an imperceptible increase in risk.

The estimated increase in risk is very small because ILRTs identify only a few potential containment leakage paths that cannot be identified by Type B and C testing, and the leaks that have been found by Type A tests have been only marginally above existing requirements.

• Given the insensitivity of risk to containment leakage rate and the small fraction of leakage paths detected solely by Type A testing, increasing the interval between integrated leakage-rate tests is possible with minimal impact on public risk. The impact of relaxing the ILRT frequency beyond one in 20 years has not been evaluated. Beyond testing the performance of containment penetrations, ILRTs also test the integrity of the containment structure.

The findings for Plant Hatch confirm these general findings on a plant specific basis considering the severe accidents evaluated for Hatch, the Hatch containment failure modes, and the local population surrounding Hatch.

7-2

8.0 REFERENCES

[1] Nuclear Energy Institute, Industry Guideline for Implementing Performance-Based Option of 10 CFR Part 50, Appendix J, NEI 94-01, Rev 3-A, July 2012.

[2] Electric Power Research Institute, Risk Impact Assessment of Revised Containment Leak Rate Testing Intervals, EPRI TR-104285, August 1994.

[3] Letter from A. Pietrangelo (NEI) to NEI Administrative Points of Contact, Interim Guidance for Performing Risk Impact Assessments in Support of One-Time Extensions for Containment Integrated Leak Rate Test Surveillance Intervals, November 13, 2001.

[4] U.S. Nuclear Regulatory Commission, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis, Regulatory Guide 1.174, Revision 2, May 2011.

[5] U.S. Nuclear Regulatory Commission, Performance-Based Containment Leak-Test Program, NUREG-1493, September 1995.

[6] Letter from R.J. Barrett (Entergy) to U.S. Nuclear Regulatory Commission, IPN 007, dated January 18, 2001.

[7] United States Nuclear Regulatory Commission, Indian Point Nuclear Generating Unit No. 3 - Issuance of Amendment Re: Frequency of Performance-Based Leakage Rate Testing (TAC No. MB0178), April 17, 2001.

[8] ERIN Engineering and Research, Shutdown Risk Impact Assessment for Extended Containment Leakage Testing Intervals Utilizing ORAMTM, EPRI TR-105189, Final Report, May 1995.

[9] Hatch SAMA Evaluation in Support of Application for License Renewal, February 2000.

[10] Oak Ridge National Laboratory, Impact of Containment Building Leakage on LWR Accident Risk, NUREG/CR-3539, ORNL/TM-8964, April 1984.

[11] Pacific Northwest Laboratory, Reliability Analysis of Containment Isolation Systems, NUREG/CR-4220, PNL-5432, June 1985.

[12] U.S. Nuclear Regulatory Commission, Technical Findings and Regulatory Analysis for Generic Safety Issue II.E.4.3 'Containment Integrity Check', NUREG-1273, April 1988. ~ ..

[13] Pacific Northwest Laboratory, Review of Light Water Reactor Regulatory Requirements, NUREG/CR-4330, PNL-5809, Vol. 2, June 1986.

[14] U.S. Nuclear Regulatory Commission, Severe Accident Risks: An Assessment for Five U.S. Nuclear Power Plants, NUREG -1150, December 1990.

[15] U.S. Nuclear Regulatory Commission, Reactor Safety Study, WASH-1400, October 1975.

[16] Southern Nuclear Calculation Change Notice PRA-CN-H-13-003, Hatch Unit 1 Internal Events PRA Model Revision 4.1, Version 1, October 2013 ..

[17] Southern Nuclear Calculation H-RIE-IEIF-UOl-010, Hatch Unit 1 Level 2 PRA Model, Version 2, November 2015, documenting ERIN Engineering Report, "Hatch 2013 Unit 1 Level 2 Model Integration and Documentation, Rev 0, December 2013.".

8-1

[18] Hatch Individual Plant Examination for External Events, Southern Nuclear Operating Company, January 1996.

[19] Response to Request for Additional Information Concerning ttie License Amendment Request for a One-Time Integrated Leakage Rate Test Extension, Letter from Mr. C.

H. Cruse (Calvert Cliffs Nuclear Power Plant) to NRC Document Control Desk, Docket No. 50-317, March 27, 2002.

[20] Letter from D.E. Young (Florida Power) to U.S. Nuclear Regulatory Commission, 3F0401-11, dated April 25, 2001.

[21] Letter from A. Pietrangelo (NEI) to NEI Administrative Points of Contact, One-Time Extension of Containment Integrated Leak Rate Test Interval - Additional Information, November 30, 2001.

[22] Risk Impact Assessment of Extended Integrated Leak Rate Testing Intervals:

Revision 2-A of 1009325, EPRI TR-1018243, October 2008.

[23] Letter from SNC (H.L. Summer, Jr.) to US NRC,

Subject:

Edwin I. Hatch Nuclear Plant Unit 1 Request to Revise Technical Specifications: Deferral of Type A Containment Integrated Leak Rate Test (ILRT), August 31, 2001.

[24] Letter from SNC (H.L. Summer, Jr.) to US NRC,

Subject:

Edwin I. Hatch Nuclear Plant 2 Technical Specification Revision Request Integrated Leakage Rate Testing Interval Extension Request, April 26, 2004.

[25] U.S. Nuclear Regulatory Commission, Generic Environmental Impact Statement for License Renewal of Nuclear Plants: Regarding Edwin I. Hatch Nuclear Plant, Units 1 and 2 - Final Report, NUREG-1437, Supplement 4, May 2001.

[26] U.S. Nuclear Regulatory Commission, Evaluation of Severe Accident Risks: Peach Bottom, Unit 2, NUREG/CR-4551, Vol. 4, Rev. 1, Part 1. December 1990.

[27] Response to Request for Additional Information Regarding License Amendment Request for One-Time Extension of the Type A Test Interval, Letter from Mr. Carl J.

Fricker (Salem Generating Station) to U.S. Nuclear Regulatory Commission, February 24, 2010.

[28] Regulatory Guide 1.200, An Approach for Determining the Technical Adequacy of Probabilistic Risk Assessment Results for Risk-Informed Activities, Revision 2, March 2009.

[29] Letter from SNC (H.L. Summer, Jr.) to US NRC,

Subject:

Edwin I. Hatch Nuclear Plant, Additional Information Related to the Staff's Review of Severe Accident Mitigation Alternatives (TAC Nos. MA8096 and MA 8098), July 26, 2000.

[30] Letter from SNC (H.L. Summer, Jr.) to US NRC,

Subject:

Edwin I. Hatch Nuclear Plant, Additional Information Related to the Staff's Review of Severe Accident Mitigation Alternatives (TAC Nos. MA8096 and MA 8098), August 31, 2000.

[31] Southern Nuclear Calculation H-RIE-IEIF-U02-011, Hatch Unit 2 Level 2 PRA Model, Version 2, November 2015, documenting ERIN Engineering Report, "Hatch 2013 Unit 2 Level 2 Model Integration and Documentation, Rev 0, December 2013.".

[32] U.S. Nuclear Regulatory Commission, Final Safety Evaluation For Nuclear Energy Institute (NEI) Topical Report (TR) 94-01, Revision 2, "Industry Guideline For Implementing Performance-Based Option of 10 CFR Part 50, Appendix J" and Electric Power Research Institute (EPRI) Report No. 1009325, Revision 2, August 2007, "Risk Impact Assessment of Extended Integrated Leak Rate Testing Intervals" (TAC No.

MC9663), June 25, 2008.

8-2

[33] U.S. Nuclear Regulatory Commission, Letter

Subject:

Edwin I. Hatch Nuclear Plant, Unit 1 Re: Issuance of Amendment (TAC No. MB2842), February 20, 2002.

[34] U.S. Nuclear Regulatory Commission, Letter

Subject:

Edwin I. Hatch Nuclear Plant, Unit 2 Re: Issuance of Amendment Revising the Technical Specifications for the Primary Containment Leakage Rate Testing Program (TAC No. MC2761), February 1, 2005.

[35] Exelon Generation Company, License Renew Application for Dresden Nuclear Power Station and Quad Cities Nuclear Power Station, January 2003.

[36] Plant Hatch Final Safety Analysis Report, Revision 33, September 2015.

[37] Oak Ridge National Laboratory, Loss of DHR Sequences at Browns Ferry Unit One -

Accident Sequence Analysis, NUR~G/CR-2973, May 1983.

[38] Southern Nuclear Calculation H-RIE-IEIF-UOl-010, Hatch Unit 1 Level 2 PRA Model, Version 1, November 2015, documenting ERIN Engineering Report, "Hatch Detailed Unit 1 Evaluation and Level 2 Notebook, Rev 5, November 2010."

[39] Southern Nuclear Calculation Change Notice PRA-CN-H-13-002, Hatch Unit 2 Internal Events PRA Model Revision 4.1, Version 1, October 2013.

8-3

APPENDIX A NUREG/CR-4551 PEACH BOTTOM POPULATION ESTIMATE

A.1 PURPOSE This appendix serves to estimate the 50-mile radius population surrounding the Peach Bottom plant utilized in the NUREG-1150 [Al] study conducted by the NRC for use in the Hatch ILRT risk assessment.

A.2 BACKGROUND The EPRI ILRT risk assessment methodology [A2] includes the option to use ex-plant consequences from NUREG-1150 if a plant does not have a plant specific Level 3 PRA. These consequence results are detailed in the NUREG/CR-4551 series of documents for the different plants analyzed.

Hatch has plant specific consequence results developed in support of the Severe Accident Management Alternatives (SAMA) analysis performed in support of license renewal. These plant specific consequence results, however, do not include results for the condition of an intact containment (i.e., technical specification leakage) required by the EPRI ILRT methodology. Therefore, surrogate results from the NUREG-1150 study for Peach Bottom, as documented in NUREG/CR-4551 [A3], are utilized in the Hatch ILRT risk assessment.

In order to utilize the Peach Bottom results for the Hatch assessment, the results must be scaled to address differences between Peach Bottom and Hatch reactor power levels, allowed technical specification leakage, and regional (i.e., 50-mile radius) populations. This scaling is discussed in Section 4.2 of the main report.

NUREG/CR-4551 [A3] does not specifically identify the 50-mile radius population for Peach Bottom. Therefore, a Peach Bottom 50-mile population estimate must be developed.

A.3 POPULATION ESTIMATE Table 4.2-2 of NUREG/CR-4551 [A3] provides the population within certain distances of the Peach Bottom plant, as reproduced in Table A.3-1. Previous ILRT submittals have typically interpolated the data to develop a 50-mile radius population.

A-1

For the Hatch ILRT assessment, the NUREG/CR-4551 series of documents was further reviewed and the Peach Bottom site regional data was found in the Peach Bottom MACCS SITE data file reproduced in NUREG/CR-4551 Part 7 [A4] in Appendix A. The MACCS SITE population data, to a distance of 50 miles, is reproduced in Table A.3-2. For the select distances of 1, 3, 10, and 30 miles, the cumulative population totals match between the two NUREG/CR-4551 volumes.

Based upon the population data from the Peach Bottom MACCS2 SITE file, the 50-mile radius population is 4,359,675. This is the value used for Peach Bottom in the Hatch ILRT assessment.

Table A.3-1 NUREG/CR-4551 POPULATION FOR PEACH BOTTOM KILOMETERS MILES POPULATION 1.6 1 . - 118 4.8 3 I 1,822 16.1 10 I 28,647 48.3 30 I 989,356 160.9 100 I 14,849,112 563.3 350 68,008,584 1609.3 1000 I 154,828, 144 A-2

TABLE A.3-2 NUREG/CR-4551 PEACH BOTTOM MACCS SITE FILE POPULATION DATA Radial Distances'" '

Miles 0.25 0.51 0.75 1.00 1.51 2.00 2.50 3.00 3.50 5.00 7.00 10.00 13.00 16.00 20.00 25.00 30.00 40.00 50.00 Kilometers Sector 0.4 0.82 1.21

'Z 1.61

• r' ,

2.43 3.22 4.02 4.83 5.63 8.05 11.27 16.09

. 20.92 25.75

', ) ,

32.19 40.23

• ,i.r 48.28 64.37

, ~.;** , ;-

80.47

,J,>

'A "<

1 0 0 0 0 0 0 20 26 31 122 406 867 2 212 5 557 68 185 42 610 36 746 25 232 43 251 2 0 0 1 1 3 7 19 26 30 123 455 977 5 385 1,870 9 336 12 012 13 500 29 877 188 747 3 0 0 1 2 4 8 24 32 37 151 413 889 1 602 2 235 3 231 10 573 14 456 22 521 91 474 4 0 0 0 1 2 4 31 39 48 187 325 700 2,912 1 710 3 207 3 038 26 164 90,241 164 108 5 0 0 6 5 14 21 32 41 49 195 323 698 1 871 4 260 5 207 9,780 37 964 241 423 217 733 6 0 0 0 0 0 0 32 39 46 183 354 755 4 439 760 4 691 11 312 40 619 35 254 18 393 7 0 0 0 0 0 0 28 36 43 169 406 870 3 309 3 109 5 776 1130 3 899 10 750 14 771 8 0 0 1 2 4 8 37 46 56 219 702 1 498 1 005 4 019 24 424 5 722 856 8 539 9 246 9 0 0 4 4 11 17 14 20 23 93 1 048 2 238 4 653 5 262 9 826 24 736 5 216 3 830 3 556 10 0 0 8 6 20 28 16 23 27 109 476 1 024 2 497 11148 9 730 20 245 77 714 765 820 262 612 11 0 0 5 5 13 19 178 219 259 1 019 510 1 090 2 522 2 254 7 890 10 533 48 267 370 900 155 161 12 0 0 4 4 11 16 30 39 46 185 317 683 1,074 2 423 1 931 3,281 4 610 26 679 45 421 13 0 0 9 7 21 30 49 62 73 291 475 1 019 3 353 0 2 586 9 756 4 223 44 975 27 904 14 0 0 6 6 16 22 31 38 46 183 322 691 1 940 1,080 6 848 31 792 83 510 44 075 21,768 15 0 0 10 8 23 34 25 34 40 162 277 601 0 1 973 2 991 15 791 23 371 47 987 222 390 16 0 0 6 6 16 22 10 14 17 69 344 741 925 3,322 15,194 35,881 19,668 33,274 82,407 Radial Total 0 0 61 57 158 236 576 734 871 3,460 7,153 15,341 39,699 50,982 181,053 248,192 440,783 1,801,377 1,568,942 Cum. Total -- -- -- 118 -- -- -- 1,822 -- -- -- 28,647 -- -- -- -- 989,356 -- 4,359,675 A-3

A.4 REFERENCES

[A1] U.S. Nuclear Regulatory Commission, Severe Accident Risks: An Assessment for Five U.S. Nuclear Power Plants, NUREG -1150, December 1990.

[A2] Risk Impact Assessment of Extended Integrated Leak Rate Testing Intervals:

Revision 2-A of 1009325, EPRI TR-1018243, October 2008.

[A3] U.S. Nuclear Regulatory Commission, Evaluation of Severe Accident Risks: Peach Bottom, Unit 2, NUREG/CR-4551, Vol. 4, Rev. 1, Part 1. December 1990.

[A4] U.S. Nuclear Regulatory Commission, Evaluation of Severe Accident Risks:

Quantification of Major Input Parameters: MACCS Input, NUREG/CR-4551, Vol. 2, Rev. 1, Part 7. December 1990.

A-4

Appendix B Hatch PRA Technical Adequacy Evaluation In Support of ILRT Interval Extension Risk Assessment Bl

B.1 OVERVIEW A Probabilistic Risk Assessment (PRA) analysis is utilized in the containment Type A Integrated Leak Rate Test (ILRT) risk assessment to support an extension of the Hatch Unit 1 and Unit 2 test interval from ten year to fifteen years.

The guidance provided in Regulatory Guide 1.200 [Bl] (Section 4.2) indicates that the following items be addressed in documentation submitted to the NRC to demonstrate the technical adequacy of the PRA:

1. Identification of permanent plant changes (such as design or operational practices) that have an impact on the PRA but have not been incorporated in the PRA.

2. The parts of the PRA used to produce the results are performed consistently with the PRA Standard as endorsed by RG 1.200.

3. A summary of the risk assessment methodology used to assess the risk of the application, including how the PRA model was modified to appropriately model the risk impact of the application.

4. Identifications of key assumptions and approximations in the PRA relevant to the results used in the decision making process.

5. A discussion of the resolution of peer review or self-assessment findings and observations that are applicable to the parts of the PRA required for the application.

6. Identification of parts of the PRA used in the analysis that were assessed to have capability categories less than that required for the application.

The purpose of this appendix is to address the requirements identified above.

B.2 TECHNICAL ADEQUACY OF THE PRA MODEL The PRA model version used for the ILRT extension assessment is the Hatch Unit 1 Internal Events PRA model version 4.1, change notice PRA-CN-H-13-003, which was completed in October 2013, updated with the Level 2 model documented in H-RIE-IEIF-UOl-010. This is a maintenance update of the Hatch Unit 1 Internal Events PRA, Revision 4, which incorporated the resolution of findings and observations associated with the November 2009 Peer Review of the Hatch Unit 1 Internal Events PRA model.

Revision 4.1 of the Hatch PRA model is the most recent evaluation of the risk profile at Hatch for internal event challenges. The PRA modeling is highly detailed, including a wide variety of initiating events, modeled systems, operator actions, and common cause events. The PRA B2

model quantification process used for the Hatch PRA is based on the event tree / fault tree methodology, which is a well-known methodology in the industry.

The Hatch PRA model is controlled in accordance with SNC procedure RIE-001, "Generation and Maintenance of Probabilistic Risk Assessment Models and Associated Updates," and associated guidelines. This procedure defines the process for implementing regularly scheduled and interim PRA model updates, for tracking issues identified as potentially affecting the PRA models (e.g., due to changes in the plant, errors or limitations identified in the model, industry operating experience, etc.), and for controlling the model and associated computer files. To ensure that the current PRA model remains an accurate reflection of the as-built, as-operated plants, RIE-001 requires the following activities outlined in the procedure are routinely performed:

Design changes and procedure changes are reviewed for their impact on the PRA model on an on-going basis.

• Reliability data, unavailability data, initiating events frequency data, human reliability data, and other such PRA inputs shall be reviewed approximately every two fuel cycles and updated as necessary to maintain the PRA consistent with the as-operated plant.

As indicated previously, RG-1.200 also requires that additional information be provided as part of the submittal to demonstrate the technical adequacy of the PRA model used for the risk assessment. Each of these items are addressed in the following sections.

B.2.1 Plant Changes Not Yet Incorporated into the PRA Model As part of PRA model configuration control, SNC maintains a PRA model maintenance database that tracks all issues that have been identified that could impact the Hatch PRA model. Per SNC procedure RIE-001 the significance of the pending items in the database is evaluated to determine the impact on model results. Each pending item is prioritized for future model updates according to its significance to model results. A review of the current open items in the database for Hatch identified no permanent plant design or operational changes that would significantly impact the results of the risk assessment performed for the ILRT interval extension evaluation.

B.2.2 Parts of PRA Used The ILRT risk assessment utilizes the overall Level 1 and Level 2 results from the Revision 4.1 PRA Level 1 model and Level 2 model documented in H-RIE-IEIF-UOl-010, as noted in 83

the main report of the ILRT risk assessment. Section 3.2.4.1 of the NRC final safety evaluation [B2] of the EPRI ILRT risk assessment methodology documents that Capability Category I is appropriate since approximate values of Core Damage Frequency (CDF) and Large Early Release Frequency (LERF) and their distribution among release categories are sufficient for use in the EPRI methodology.

A summary of the 2009 Hatch PRA Peer review results is provided in section B.2.5. As explained in section B.2.5, with the completion of Revision 4 of the PRA model, all PRA Standard supporting requirements (SRs) are now assessed to meet Capability Category II, as applicable. The Peer Review Findings are also discussed in section B.2.5.

B.2.3 Risk Assessment Methodology Summary The ILRT risk assessment methodology is based on EPRI TR-1018243 [B3] which has been evaluated and endorsed by the NRC in a Safety Evaluation Report (SER) [B2]. The methodology as applied for Hatch is fully described in the main report of the ILRT risk assessment.

B.2.4 PRA Key Assumptions and Approximations For this application, the EPRI methodology [B3] involves a bounding approach to estimate the change in LERF for extending the ILRT interval. Rather than exercising the PRA model itself, the methodology involves the establishment of separate calculations that are linearly related to the pl~nt CDF co_ntribution that is not aJready LERF. The ILRT, risk assessment methodology [B3] incorporates various assumptions and approximations identified in the main report of the ILRT risk assessment. Key EPRI methodology assumptions and approximations are addressed via sensitivity studies. Any assumptions and approximations utilized in the PRA are judged to have negligible impacts compared to those utilized in the EPRI methodology for the purposes of this application.

B.2.5 Peer Review Findings B.2.5.1 Previous Peer Review and Self Assessment of the Hatch PRA Model The Hatch PRA model was reviewed extensively during development and undergoes independent internal review during each update. The Hatch PRA was reviewed twice prior to issuance of the ASME PRA Standard for peer review.

84

The initial peer review was conducted by the BWR Owners Group in April 2001.

The review team used Revision A-3 NEI draft "Probabilistic Risk Assessment (PRA) Peer Review Process Guidance" dated June 2, 2000 as the basis for review.

This review was observed by a team from the NRC.

In 2006, a gap analysis was performed against the available versions of the ASME PRA Standard and Regulatory Guide 1.200, Revision 0 (2003 trial version).

B.2.S.2 Regulatory Guide 1.200 PRA Review of the Hatch PRA Model against the ASME PRA Standard Requirements A PRA Peer Review of all elements of the Hatch Internal Events PRA PRA model including Internal Flooding against Regulatory Guide 1.200, Revision 2 clarifications [Bl], the ASME/ANS PRA Standard [B4], and NEI OS-04 was performed in November 2009.

A summary of the results of the PRA Peer Review [BS] previously provided to the NRC as part of the NEI Risk Informed Tech Spec Initiative (RITS) Sb License Amendment Request (LAR) submittal (NRC Record Locator Number ML103140S10) for which SNC received a NRC safety evaluation report (SER) as discussed in section B.2.7, is shown below.

Based on the results of the Peer Review, 9S% of the supporting requirements (SRs) evaluated met Category II or better. There were 10 SRs that were noted as "Not Met" and S that were noted "Met" at Category I only. All of the "Not Met" findings were resolved as part of the update of the Hatch Internal Events PRA Model, Revision 4.0, to Revision 4.1 as noted in Tables B.2-1 and B.2-2.

Peer Review Summary Previously Provided to NRC as Part of the NEI Risk Informed Tech Spec Initiative (RITS) Sb License Amendment Request (LAR) Submittal [B6]:

The ASME/ANS PRA Standard contains a total of 326 numbered supporting requirements (SRs) in nine technical elements and one configuration control element. There were five not applicable requirements for the Hatch review: AS-B4, IFEV-A8, LE-OS, LE-06, and MU-01.

Among 321 applicable SRs, 9S% of SRs met Capability Category II or higher as follows:

Caoabilitv Cateqorv Met No. of SRs O/o of total aaalicable SRs CC I/II/III (or SR Met) 225 70%

CC I 5 1%

CC II 27 8%

CC III 18 6%

CC I/II 10 3%

CC II/III 26 8%

SR Not Met 10 3%

SR Not Aoolicable 5 1%

Total 326 1000/o 10 SRs were judged to be not met.

1. IE-A6 was not met because the additional initiating events identified from the FME analysis were not included in the model under review.

BS

2. IE-B2 was not met because the IE notebook did not discuss the system for grouping initiating events.

3. IE-B3 was not met because IE-B2 was not met.

4. AS-B3 was not met because the Accident Sequence notebook did not provide a discussion of the phenomenological conditions for each accident sequence.

5. SC-BS was not met because a comparison of the thermal hydraulic calculations with those of other plants was not performed.

6. SC-Cl was not met because, although the success criteria were very detailed, it was difficult to compare criteria to a specific initiating event.

7. IFSN-A10 was not met because credit was not given to operator isolation of a certain flood scenario. This could cause a potential for flooding in other areas.

8. IFQU-AS is not met because IFSN-A10 is not met.

9. IFQU-A6 cites an over conservatis_m by failing operator actions outside the main control room in flooding areas.

10. MU-Al is not met because the SNC process used does not provide a more organized method of documentation.

In addition to the not met SRs, there were five SRs that were met as Category 1.

1. IE-A9 was met as Cat 1 because plant specific experience for initiating event precursors was not provided.

2. LE-C3 was met as Cat 1 only because there was no address as to the possibility of equipment repair. HNP did not credit repair during LERF conditions.

3. LE-C10 was met as Cat 1 because there was no documentation regarding continued equipment operations or personnel actions in a LERF environment. HNP did not credit such actions.

4. LE-C12 was met as Cat 1 because there is no documentation showing that the PRA can credit post containment failure operation of equipment or personnel actions. HNP does not credit such items.

5. LE-C13 was met as Cat 1 because there was not an engineering basis for the decontamination factor used for scrubbing.

Resolution of 10 Not Met SRs and 5 SRs Met at Category 1:

Table B.2-1 shows details of the 10 "SR Not Met" findings and resolutions after the peer review. In addition, Table B.2-2 shows how the five Category 1 SRs are addressed.

86

Table 8.2-1: Resolution of the Hatch PRA Peer Review F&Os Associated with the 10 Not Met SRs Review F&O# Level Resolution Resolution Status Element IE-A6-1-2 IE-A6 Finding Include the additional IEs in the analysis. It The IEs identified in the notebook have been added (SR not met would appear panel failures would be very to the model as well as the common cause failure CC-II) unlikely except for specific types of events, considerations. Two of the buses were not modeled especially spatially related events. as IEs because they serve only to feed two that were added as IEs. The two buses not modeled as IEs are modeled as support for systems; one of the supported systems is in itself modeled as a special initiator. The special initiator modeling for the HNP PRA not only considers what causes a transient/scram but also what would require a shutdown per Technical Specifications IE-82-1-7 IE-82 Finding Describe the process for systematically The IE notebook has been revised to include a full (SR not met grouping the IEs to ensure: discussion of IE grouping. This comment has been CC-11111111) (a) Events can be considered similar in terms addressed.

of plant response, success criteria, timing, and the effect on the operability and performance of operators and relevant mitigating systems: or (b) events can be subsumed into a group and bounded by the worst case impacts within the new group.

IE-83-1-7 IE-83 Finding Describe the process for systematically This comment has been added to the IE notebook.

(SR not met grouping IEs to ensure: There are no subsumed events in the HNP PRA CC-II) (a) Events can be considered similar in terms model.

of plant response, success criteria, timing, and the effect on the operability and performance of operators and relevant mitigating systems; or {b) Events can be subsumed into a group and bounded by the worst case impacts within the new group.

AS-83-1-9 AS-83 Finding Include additional detail for each accident The detail required by this finding has been added (SR not met sequence. Particularly, there was no mention to the accident sequence notebook. The sequence CC-11111111) of the generation of harsh environments descriptions have a discussion of Environmental B7

Table 8.2-1: Resolution of the Hatch PRA Peer Review F&Os Associated with the 10 Not Met SRs Review F&O# Level Resolution Resolution Status Element affecting temperature, pressure, debris, Conditions. The HNP PRA does not typically rely water levels or humidity that could impact the on equipment or operator actions in an area where success of the svstem a severe environment is encountered.

SC-85-3-1 SC-85 Finding Check the reasonableness and acceptability A comparison table for the HNP PRA was (SR not met of the results of the thermal/hydraulic and developed for Success Criteria based on input from CC-11111111) any other analysis used to support the other 8WR facilities (i.e. Pilgrim, Cooper, LaSalle, success criteria. Document in the SC and Nine Mile Point).

Notebook how this reasonableness was performed.

SC-Cl-5-4 SC-C3 Finding The success criteria should be captured in A success criteria summary table has been added (SR not met tabular from in the SC document. Having this to the SC notebook. This table in addition to the CC-1/111111) information in one place will alleviate extreme detail already provided makes the confusion when performing PRA applications Success Criteria notebook extremely informative.

and upgrades. Additionally, it will facilitate peer reviews. Provide a summary of success criteria for available mitigating systems and human actions for each initiating grouo.

IFSN-A10-4-5 IFSN-A10 Finding Operator actions should be developed and The PRA model used in the peer review contained (SR not met added to the scenario development and the over 100 flood initiators. No screening was done CC-1111/111) PRA model to reflect how the plant would be based on operator action input. This finding was operated in the event of this scenario. It may addressed by screening the initiators to 24 and be beneficial to consider use of mitigation applying HRA for these scenarios to mitigate the event trees to assure that all mitigation results.

issues are considered.

IFQU-A5-4-5 IFQU-A5 Finding Operator actions should be developed and The PRA model used in the peer review contained (SR not met added to the scenario development and the over 100 flood initiators. No screening was done CC-1111/111) PRA model to reflect how the plant would be based on operator action input. This finding was operated in the event of this scenario. It may addressed by screening the initiators to 24 and be beneficial to consider use of mitigation applying HRA for these scenarios to mitigate the event trees to assure that all mitigation results.

issues are considered.

IFQU-A6-2-7 IFQU-A6 Finding Consider more realistic operator actions for This item has been addressed for IFQU-A5-4-5.

SR not met floods when action has to be taken outside CC-1111/111) the main control room.

BS

Table 8.2-1: Resolution of the Hatch PRA Peer Review F&Os Associated with the 1O Not Met SRs Review F&O# Level Resolution Resolution Status Element MU-Cl-5-8 MU-C1 Finding Recommend developing a database for use This comment refers to the model update process.

SR not met by PRA. Such a database should have These procedures are under revision. Presently, CC-1/111111) prioritization clearly delineated. This would model change requirements tend to be governed allow a dynamic assessment of the by the use of the corrective action program.

cumulative impact of pending changes.

Additionally, this will allow the more Note: A rigorous Configuration management significant changes to be incorporated before program has been developed after the 5b LAR was less significant ones submitted to the NRC in 2010.

89

Table B.2-2: Resolution of the Hatch PRA Peer Review F&Os Associated with the5 Cat I met only SRs Review F&O# Level Resolution Resolution Status Element LE-C3-7-2 LE-C3 Finding Review significant progression sequences to Statements have been added to L2 NB Section 7 of (SR Cat I met) support evaluations required in applicable notebooks, as well as Appendix K to note that SR Capability Category. significant accident sequences resulting in LERF were reviewed and credit for continued operation or repair (beyond LOOP recovery) was not judged to be credible. Cat 1/111111 is considered met for this SR.

LE-C10-7-2 LE-C10 Finding Review significant progression sequences to Statements have been added to L2 NB Section 7 of (SR Cat I met) support evaluations required in applicable notebooks, as well as Appendix K to note that SR Capability Category significant accident sequences resulting in LERF were reviewed and credit for continued operation or repair (beyond LOOP recovery) was not judged to be credible. Cat II is considered met for this SR.

LE-C12-7--2 LE-C12 Finding Review significant progression sequences to Statements have been added to L2 NB Section 7 of (SR Cat I met) support evaluations required in applicable notebooks, as well as Appendix K to note that SR Capability Category significant accident sequences resulting in LERF were reviewed and credit for continued operation or repair (beyond LOOP recovery) was not judged to be credible. Cat II is considered met for this SR.

LE-C13-7-4 LE-C13 Finding Perform a containment bypass analysis as Text enhanced in Appendix D of notebooks (Footnote 6 (SR Cat I met) described in SR LE-C13 added to Table HA5) and similarly in Section 7.7. The analysis regarding scrubbing is approached with engineering judgment. Only scrubbing for low pressure sequences is considered and then the value is low.

This basis is acceptable and the use of scrubbing is considered in the uncertainty analysis. Cat 111111 is considered met for this SR.

IE-A9-1-4 IE-A9 Finding Include other sources of OE in the search for Several sources are available to determine plant (SR Cat I met) IE precursors. specific initiating events. When preparing the IE notebook, SNC reviewed the following sources. For each source, appropriate section has also been identified.

BlO

Table B.2-2: Resolution of the Hatch PRA Peer Review F&Os Associated with the5 Cat I met only SRs Review F&O# Level Resolution Resolution Status Element

1. Plant Specific Events: [Table 3-4; Appendix C (LERs)]

2. Plant Systems: [Appendix B - FMEA]

3. LOCA inside Containment [3.2.3]

4. LOCA outside Containment [3.2.3]

5. Multiple Failures [3.1.6]

6. Interview [3.1.9; Appendix E]

As a result of these reviews, additional special initiating events were identified and have been modeled.

Bll

B.2.6 PRA Portions With Inadequate Technical Adequacy As previously noted, the NRC safety evaluation [B2] of the EPRI ILRT methodology specifies that Capability Category I is appropriate for the applicable PRA Standard supporting requirements. Based on the update to the Hatch Internal Events PRA model to Revision 4.1, following the 2009 PRA Peer Review, all PRA Standard supporting requirements are met at Capability Category II or higher, as applicable.

B.2.7 NRC Safety Evaluation Report (SER) for Hatch Units 1 and 2 NEI Ris.k Informed Tech Spec Initiative (RITS) Sb Submittal SNC submitted a license amendment request (LAR) and received a SER for NEI RITS Initiative Sb to implement the Surveillance Frequency Control Program (SFCP) at Hatch. The NEI RITS Initiative Sb application is similar to this ILRT interval extension submittal in that they both deal with surveillance frequency extensions. Excerpts from the NRC SER for Hatch pertaining to the Hatch PRA quality are provided in section B.4 and the overall conclusions are considered applicable to this submittal.

B.3

SUMMARY

A PRA technical adequacy evaluation was performed consistent with the requirements of RG-1.200, Revision 2. This evaluation combined with the details of the results of this analysis demonstrates with reasonable assurance that the proposed extension to the ILRT interval for Hatch Units 1 & 2 to fifteen years on a permanent basis satisfies the risk acceptance guidelines in RG 1.174.

812

B.4 EXCERPTS FROM NRC SAFETY EVALUATION REPORT (SER) FOR HATCH UNITS 1 AND 2 RISK INFORMED TECH SPEC (RITS) INITIATIVE SB APPLICATION SUBMITTAL From Sb SER (ADAMS Accession No. ML11108A129):

The requested change would adopt the NRC-approved Technical Specifications Task Force (TSTF) Traveler TSTF-42S, Revision 3, "Relocate Surveillance Frequencies to Licensee Control-RITSTF Initiative Sb" (Reference 1). When implemented, TSTF-42S relocates most periodic frequencies of TS surveillances to a 1icensee-controlled program, the Surveillance Frequency Control Program (SFCP), and provides requirements for the new program in the Administrative Controls section of the TSs.

3.1.4.1 Quality of the PRA The quality of the HNP PRA is compatible with the safety implications of the proposed TS change and the role the PRA plays in justifying the change. That is, the more the potential change in risk or the greater the uncertainty in that risk from the requested TS change, or both, the more rigor that must go into ensuring the quality of the PRA.

The licensee used RG 1.200, Revision 1, to address the HNP PRA technical adequacy. RG 1.200 is NRC's developed regulatory guidance which, endorsed with comments and qualifications the use of the American Society of Mechanical Engineers (ASME) RA-Sb-200S, "Addenda to ASME-RA-S-2002 'Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications'," (Reference 6),

NEI 00-02, "Probabilistic Risk Assessment (PRA) Peer Review Process Guidance,"

Revision 1, (Reference 7), and NEI OS-04, "Process for Performing Follow-On PRA Peer Reviews Using the ASME/ANS PRA Standard," Revision 2 (Reference 8). The licensee has performed an assessment of the PRA models used to support the SFCP against the requirements of RG 1.200 to assure that the PRA models are capable of determining the change in risk due to changes to surveillance frequencies of SSCs, using plant-specific data and models. Capability Category II of ASME RA-Sb-200S was applied as the standard, and any identified deficiencies to those requirements are assessed further to determine any impacts to proposed decreases to surveillance frequencies, including by the use of sensitivity studies where appropriate.

The NRC staff notes that in Revision 2, RG 1.200 endorsed with comments and qualifications an updated combined standard, which includes requirements for fire, seismic, and other external events PRA models. The existing internal events standard was subsumed into the combined standard, but the technical requirements are essentially unchanged. Since NEI 04-10 specifically identified the use of RG 1.200 Revision 1 to assess the internal events standard, the licensee's approach is reasonable and consistent with the approved methodology.

In November 2009, a licensee peer review of the HNP internal events PRA model was conducted using the ASME/American Nuclear Society (ANS) Combined PRA B13

Standard (Reference 11). The NRC staff evaluated the licensee's response to the deficiencies (Table 2 of Enclosure 2 of the license amendment request (ADAMS Accession No. ML103140510)) identified by the licensee peer review. The NRC staff's assessment of the resolution or disposition of the identified deficiencies is provided below:

IE-A6-1-2: Initiating events involving failure of electrical panels were identified but not included in the PRA model. These events have been added to the PRA model, and so the finding has been addressed.

IE-B2-1-7 and IE-B3-1-7: The grouping process for initiating events was not provided in the documentation. The documentation has been revised to include a full discussion of initiating event grouping, and so the finding has been addressed.

AS-B3-1-9: Enhancements to the documentation of accident sequence modeling were identified, in particular for accident environment impacts on plant equipment.

The documentation has been revised to include this information, and so the finding has been addressed.

SC-B5-3-1: A reasonableness check of the success criteria by comparison to the criteria for similar plant designs was not performed. The licensee identified that a comparison table has been developed for success criteria based on four other similar facilities, so the finding has been addressed.

SC-Cl-5-4: The documentation of success criteria could better describe which systems are credited for initiating events, and which initiating events impact systems. A success criteria summary table has now been included in the documentation, so the finding has been addressed.

IFSN-Al0-4-5 and IFOU-A5-4-5: Credit for operator action to isolate a flood source is not addressed. The flood initiator screening did not consider operator actions, but the licensee has now provided a screening to initiators and developed human reliability analyses for 24 scenarios. Therefore, this finding has been addressed.

IFOU-A6-2-7: The PRA conservatively assumes ex-control room actions are failed for systems impacted by the flood. The finding has been addressed in the response to item IFQU-A5-4-5.

MU-Cl-5-8: The process for tracking model changes is identified as cumbersome and inefficient. The procedures governing the process are under revision and model changes are tracked by the corrective action program.

LE-C3-7-2, LE-Cl0-7-2, and LE-C12-7-2: No documentation of review of significant accident progression sequences resulting in a large early release were found. The licensee has updated the documentation, and concluded that credit for continued operation or repairs (other than recovery of offsite power) were not credible.

814

LE-C13-7-4: No evaluation of scrubbing of releases has been performed for class V sequences (containment bypass). The licensee has updated the documentation, and bases limited scrubbing credit for low pressure sequences only on engineering judgment.

Further consideration of scrubbing effects is performed in the uncertainty analyses.

IE-A9-1-4: A more thorough review of available plant operating experience beyond interviews is required. Additional reviews were conducted and documented, and additional special initiating events were added to the model.

Based on the licensee's assessment using the applicable PRA standard and RG 1.200, the level of PRA quality, combined with the proposed evaluation and disposition of the identified deficiencies, is sufficient to support the evaluation of changes proposed to surveillance frequencies within the SFCP, and is consistent with Regulatory Position 2.3.1 of RG 1.177 B.5 REFERENCES

[Bl] Regulatory Guide 1.200, An Approach for Determining the Technical Adequacy of Probabilistic Risk Assessment Results for Risk-Informed Activities, Revision 2, March 2009.

[B2] U.S. Nuclear Regulatory Commission, Final Safety Evaluation For Nuclear Energy Institute (NEI) Topical Report (TR) 94-01, Revision 2, "Industry Guideline For Implementing Performance-Based Option of 10 CFR Part 50, Appendix J" and Electric Power Research Institute (EPRI) Report No. 1009325, Revision 2, August 2007, "Risk Impact Assessment of Extended Integrated Leak Rate Testing Intervals" (TAC No.

MC9663), June 25, 2008.

[B3] Risk Impact Assessment of Extended Integrated Leak Rate Testing Intervals:

Revision 2-A of 1009325, EPRI TR-1018243, October 2008.

[B4] American Society of Mechanical Engineers and American Nuclear Society, Standard for Level 1/Large Early Release Fr~quency Probabilistic Risk Assessment for Nuclear Power Plant Applications, Addenda A, ASME/ANS RA-Sa-2009, 2009

[BS] Hatch Unit 1 Peer Review Report (2009), February 2010

[B6] SNC Letter NL-10-0271, October 29, 2010, Edwin I. Hatch Nuclear Plant - License Amendment Request for Adoption of TSTF-425-A, Rev. 3, Risk-Informed Justification for the Relocation of Specific Surveillance Frequency Requirements to a Licensee Controlled Program Using the Consolidated Line Item (NRC Record Locator Number ML103140510)

B15