Text

OL - TVA Letter to NRC_05-25-11_WBN U2 Samda Remaining RAI Responses
	ML11166A173
Person / Time
Site:	Watts Bar
Issue date:	05/25/2011
From:	; - No Known Affiliation
To:	; Division of Operating Reactor Licensing
References
	Download: ML11166A173 (47)
	v • d • e

1 WBN2Public Resource From:

Boyd, Desiree L [dlboyd@tva.gov]

Sent:

Wednesday, May 25, 2011 1:06 PM To:

Epperson, Dan; Poole, Justin; Raghavan, Rags; Milano, Patrick; Campbell, Stephen Cc:

Crouch, William D; Hamill, Carol L; Boyd, Desiree L

Subject:

TVA letter to NRC_05-25-11_WBN U2 SAMDA Remaining RAI Responses Attachments:

05-25-11_WBN U2 SAMDA Remaining RAI Responses_Final.pdf PleaseseeattachedTVAletterthatwassenttotheNRCtoday.

Replacepreviousemailwithattached.

ThankYou,

~*~*~*~*~*~*~*~*~*~*~*~*~*~*~

Désireé L. Boyd WBN2LicensingSupport SunTechnicalServices dlboyd@tva.gov 4233658764

~*~*~*~*~*~*~*~*~*~*~*~*~*~*~

Hearing Identifier:

Watts_Bar_2_Operating_LA_Public Email Number:

397 Mail Envelope Properties (7AB41F650F76BD44B5BCAB7C0CCABFAF1F35044B)

Subject:

TVA letter to NRC_05-25-11_WBN U2 SAMDA Remaining RAI Responses Sent Date:

5/25/2011 1:06:22 PM Received Date:

5/25/2011 1:07:25 PM From:

Boyd, Desiree L Created By:

dlboyd@tva.gov Recipients:

"Crouch, William D" <wdcrouch@tva.gov>

Tracking Status: None "Hamill, Carol L" <clhamill@tva.gov>

Tracking Status: None "Boyd, Desiree L" <dlboyd@tva.gov>

Tracking Status: None "Epperson, Dan" <Dan.Epperson@nrc.gov>

Tracking Status: None "Poole, Justin" <Justin.Poole@nrc.gov>

Tracking Status: None "Raghavan, Rags" <Rags.Raghavan@nrc.gov>

Tracking Status: None "Milano, Patrick" <Patrick.Milano@nrc.gov>

Tracking Status: None "Campbell, Stephen" <Stephen.Campbell@nrc.gov>

Tracking Status: None Post Office:

TVANUCXVS2.main.tva.gov Files Size Date & Time MESSAGE 383 5/25/2011 1:07:25 PM 05-25-11_WBN U2 SAMDA Remaining RAI Responses_Final.pdf 703283 Options Priority:

Standard Return Notification:

No Reply Requested:

No Sensitivity:

Normal Expiration Date:

Recipients Received:

Tennessee Valley Authority, Post Office Box 2000, Spring City, Tennessee 37381-2000 May 25, 2011 10 CFR 50.4 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, D.C. 20555-0001 Watts Bar Nuclear Plant, Unit 2 Docket No. 50-391

Subject:

WATTS BAR NUCLEAR PLANT (WBN) - UNIT 2 - RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION ITEM NUMBERS 2, 3, 5 AND 15 REGARDING SEVERE ACCIDENT MANAGEMENT DESIGN ALTERNATIVE REVIEW (TAC NO. MD8203)

References:

1. TVA to NRC letter dated May 13, 2011, Watts Bar Nuclear Plant (WBN) - Unit 2 - Response to Request for Additional Information Regarding Severe Accident Management Design Alternative Review (TAC No. MD8203)

2. NRC to TVA letter dated March 30, 2011, Watts Bar Nuclear Plant, Unit 2 - Supplemental Request for Additional Information Regarding Severe Accident Management Design Alternatives Review (TAC No. MD8203)

3. TVA to NRC letter dated January 31, 2011, Watts Bar Nuclear Plant (WBN) Unit 2 - Response to Request for Additional Information Regarding Severe Accident Management Alternative Review (TAC No. MD8205)

4. NRC to TVA letter dated January 11, 2011, Watts Bar Nuclear Plant, Unit 2 - Request for Additional Information Regarding Severe Accident Management Alternative Review (TAC No. MD8203)

The purpose of this letter is to provide a response to the NRC requests for additional information (RAI) Item Nos. 2, 3, 5 and 15 regarding the Severe Accident Management Design Alternatives (SAMDA) analysis discussed in Reference 2.

Reference 1 provided TVAs response to the majority of the requests for information

U.S. Nuclear Regulatory Commission Page 3 May 25, 2011 bcc (Enclosures):

Stephen Campbell U.S. Nuclear Regulatory Commission MS 08H4A One White Flint North 11555 Rockville Pike Rockville, MA 20852-2738 Charles Casto, Deputy Regional Administrator for Construction U.S. Nuclear Regulatory Commission Region II Marquis One Tower 245 Peachtree Center Ave., N.E., Suite 1200 Atlanta, GA 30303-1257

ENCLOSURE 1 RESPONSES TO NRC REQUEST FOR ADDITIONAL INFORMATION ITEM NUMBERS 2, 3, 5 AND 15 By letter dated January 31, 2011, the Tennessee Valley Authority (TVA) provided a response to the Nuclear Regulatory Commission (NRC) staff regarding questions related to TVA's Updated Analysis of Severe Accident Mitigation Design Alternatives (SAMDAs) for Watts Bar Nuclear Plant (WBN), Unit 2. In its review of this information, the NRC staff requires further information and clarification on TVA's response. The information listed in the following request for additional information (RAI) refers to the RAI responses in the January 31, 2011, letter.

2. RAI 1.f The response did not address the assumptions concerning the availability of WBN Unit 1 components/systems for both dual-unit and Unit 2 initiating events. Discuss how the availability of Unit 1 systems during Unit 1 outages is accounted for in evaluating WBN Unit 2 CDF and LERF.

TVA Response:

The WBN model is a dual-unit model which incorporates initiating event fault tree logic for a number of support-system initiators. The relevant fault tree logic is input into both unit-specific and dual-unit initiating events. Therefore, the impact of unavailability of Unit 1 components/systems with respect both to mitigation and to initiating events (unit-specific and dual-unit) is incorporated in the model for Unit 2, and the results are included in the calculations for CDF and LERF for Unit 2.

Primary systems for Unit 2 such as ECCS will not be impacted directly during Unit 1 outages, the impact will be seen in the unavailability of support systems such as service water and component cooling. Also, actual plant risk of the operating unit while the other unit is in a refueling outage will be monitored using the WBN EOOS (Equipment out of Service) model. WBN uses EPRIs EOOS program to monitor plant risk in accordance with 10 CFR 50.65a(4). As WBN has not yet completed a refueling outage on one unit while the other unit is operating, assumptions regarding these unavailabilities were made.

The following describes the current modeling approach used in the CAFTA model.

The WBN Unit 1 maintenance rule database was the primary data source used to develop estimates of system train unavailability due to testing and maintenance of modeled components; i.e., TI-119, Maintenance Rule Performance Indicator Monitoring, Trending, and Reporting, 10 CFR 50.65. This database includes component failure and unavailability for periods when a system is required. For systems such as ERCW, Component Cooling and Compressed Air, the systems are required during power operation (Mode 1) and for some periods of time when the unit is in Modes 2-6. For dual unit operation, excessive unavailability from outage downtime would adversely impact the maintenance rule metrics. In addition, dual unit operation with controlling technical specifications from each unit will limit the outage time for common systems compared to single unit operation. As an example, E1-1

two trains of ERCW must be operable during modes 1, 2, 3, and 4 (per LCO 3.7.8),

or within 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months the unit must begin the transfer to mode 5. This LCO does not apply to single unit operation when in mode 5 or 6 already. However, when dual unit operation begins, if one unit is in an outage (i.e., Mode 5 or 6), the LCO will apply to the unit that is in Modes 1, 2, 3, or 4. Therefore, the duration of such maintenance events while one unit is in an outage will be limited by the TS requirements on the non-outage unit. This train maintenance is included in the model Maintenance Rule data is collected on a train basis for most systems. The data for each train was used to calculate the probability of the train being unavailable due to maintenance. Each of the calculated probabilities was given a variable name which was added to the type code table in the WBN PRA CAFTA database. These variables are assigned to the applicable maintenance basic events for each train during fault tree construction.

Where maintenance rule data were not available, generic information from NUREG/CR-6928 was generally used. However, since NUREG/CR-6928 data for unavailability consider only events during critical hours from the Mitigative System Performance Index (MSPI) data, the generic experience data from plant outages are not considered. Also, unavailability was calculated or estimated in some instances where neither appropriate maintenance rule data nor NUREG/CR-6928 data were available. However, based on the maintenance rule unavailability limitations and the technical specification limits, we believe the unavailability times used in the CAFTA model are sufficient for proper decisionmaking with respect to proposed SAMA measures.

3. RAI 1.h

a. The response to the RAI states "The following peer cert findings remain open and are considered documentation related (i.e., are judged to not have the potential for significant change to model results or risk ranking):... Among the findings listed, the following, from the description in the IPE, do not appear to be limited to just documentation:

(1) 1-4 (DG [diesel generator] load sequencer) - states certain failures are missing from the logic model, (2) 2-11 (T-H [thermal-hydraulic] timing for HRA [human reliability analysis] cues, no simulator runs) - does this mean the runs were not performed, or that they were performed but not documented, and (3) 5-1 (Optimistic room heat-up times used).

Provide further justification that each of these items is only a documentation issue and that final resolution is unlikely to impact the SAMA evaluation.

TVA Response:

a. (1) This relates to standard Supporting Requirement SY-B10. The peer reviewers comment about this element is as follows: Assessment: Cat 2-3 E1-2

is MET. Basis: Appropriate actuation signals from RPS and ESFAS are modeled. However, the actuation signals from the DG load sequencers are not modeled for each load. This is expected to have minimal impact on the results because the sequencer dependency is modeled for the ERCW system which is required for diesel generator support.

ERCW is the only load with an individual sequencer modeled. The sequencer is modeled with the diesel generator and fails the diesel generator if failed. The supporting requirement is judged as met with respect to this element; therefore, the significance of the issue is believed to be relatively minor. The peer reviewer also specifically indicated that a minimal impact on results was expected. Therefore, any final resolution is unlikely to impact the SAMA evaluation.

(2) This relates to standard Supporting Requirement HR-E4. The requirements for that element are as follows:

Category I No requirement for using simulator observations or talk-throughs with operators to confirm response models.

Category I/II USE simulator observations or talk-throughs with operators to confirm the response models for scenarios modeled.

The peer reviewers comment about this element is as follows: Assessment:

Cat 1 is MET. Basis: The Process for Post initiator actions provided in MDN-000-999-2008-0144 Section 7 does not discuss use of simulators. However, timing data for post initiator actions are provided in MDN-000-999-2008-0153 section 8.0 based only on judgment of the operators, which includes some very optimistic times. Also, the operator interviews should produce some insights related to the risk model, use of procedures, etc. CAT I met for this element.

The WBN Model used walk-throughs rather than simulator runs. The peer review comment indicates that the supporting requirement was judged to be met at least at capability category I, and it is also noted that talk-throughs with operators were in fact performed. The talk-throughs were performed and were well-documented, which would appear to meet the requirements of capability category II/III. Documentation demonstrates that the talk-throughs with members of the operations staff at WBN were conducted over a significant number of hours and with several highly experienced individuals having diverse operations backgrounds. Therefore, the results are believed to be generally reasonable and appropriate. There is no information to indicate that any significant changes might result should simulator runs be performed.

(3) The peer review comment reads as follows:

The mission time used for room heat-up calculations (MDN-000-999-2008-0143, Appendix B, WBNOSG4-242, 200, and 197) was optimistically justified.

E1-3

(This F&O originated from SR SC-A5.)

Basis for Significance. According to WBNOSG4-197, 200 and 242, the mission time for mitigation was verified based on simplified calculations and optimistic engineering judgment. Because the component cooling relies on HVAC, the results of room heat-up calculation affect the ability of components to function without room cooling.

The resolution of this item will be that based on room heat-up calculation results, judge whether the safe and stable condition is met and the basis of the judgment should be presented explicitly.

TVA calculation MDN-000-999-2008-0143 summarized results and identified a number of areas as requiring room cooler/ HVAC support including residual heat removal pump rooms, safety injection pump rooms, containment spray pump rooms, charging pump rooms, the turbine-driven auxiliary feedwater pump room, and the emergency diesel generator rooms. Mission times of 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months were generally assumed. No information was identified in MDN-000-999-2008-0143 to indicate that these results are optimistic with respect to those for other Westinghouse PWR sites. The peer review report indicates that all supporting requirements for Success Criteria were met at Category II (or II/III) or above, therefore success criteria elements are indicated to be adequately met. The proposed resolution indicates that additional review, judgment, and documentation of this process may be sufficient to resolve the question. There is therefore no indication that any changes will be developed which could substantially impact the SAMA evaluation.

b. TVA stated (p. E1-13), with reference to internal flooding, that "the current model is judged to be adequately bounding for this application," which is based on conservatisms in the flooding model described in the response. While this may be true, there are a large number of open internal flood findings from the peer review and an updated flooding analysis is to be included in the next model update. Section 3.7 (p. 70) of the WBN Unit 2 IPE submittal states that two sets of sensitivity studies were performed on the internal flooding analysis, with one set focused on evaluating alternative design/procedural changes that would significantly impact (i.e., reduce) the flood related CDF and LERF while the other was designed to address epistemic uncertainties identified in the WBN internal flooding probabilistic risk assessment (PRA). Provide a description of these studies, their results and conclusions, and how the results and conclusions support the conclusion that the current flooding model is bounding for the SAMA application.

TVA Response:

b. A sensitivity study was performed to evaluate re-routing or sheathing with guard pipes certain raw cooling water piping segments in flood areas 772.0-A8, 772.0-A9, 757.0-A9 and 757.0-A17. A reduction in CDF of approximately 7.32%

E1-4

and a reduction in LERF of approximately 12.98% were estimated for this configuration.

A sensitivity study was performed to examine the flooding risk contributions associated with raw cooling water and high pressure fire protection piping failures in areas 757.0-A2, -A5, -A9, -A17, -A21, and -A24, 772.0-A7, -A8, -A9 and -A10, comparing the risk associated with carbon steel piping vs. stainless steel piping.

The study compared a baseline configuration of stainless steel piping with carbon steel piping. The carbon steel piping configuration CDF was approximately 45.43% greater and LERF was approximately 81.30% greater.

A sensitivity study was performed to evaluate the significance of operator actions to mitigate certain flooding scenarios. A case was run setting the failure probability of all the subject mitigating actions to 0 (certain to succeed) and then another case was run setting the failure probability to 1.0 (certain to fail). Setting the actions to 0 resulted in a reduction in CDF of approximately 2% and a reduction in LERF of approximately 1%. Setting the actions to 1.0 resulted in an increase in CDF of approximately 44% and an increase in LERF of approximately 23%.

A sensitivity study was performed to evaluate the contributions due to maintenance and human-induced flooding events. Suppressing contributions from these causes resulted in a decrease in CDF of approximately 0.3% and a decrease in LERF of approximately 0.4%.

The studies were not performed to demonstrate that the current model is bounding but rather to compare various possible plant changes or to examine the importance of certain inputs to the analysis. However they do demonstrate that the model responds as expected, showing reductions in flood risk associated with improvements, thereby providing additional confidence in its results.

TVA has previously committed to installing flood detection in WBN areas 772.0-A8 and 772.0-A9. Please see commitment 9 of Enclosure 2 of the January 31, 2011 letter from Marie Gillman (TVA) to the USNRC, WATTS BAR NUCLEAR PLANT (WBN) - UNIT 2 - RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION REGARDING SEVERE ACCIDENT MANAGEMENT ALTERNATIVE REVIEW (TAC NO. MD8203) (T02 110131 001).

TVA now numbers this as a new SAMA, 340, to install flood detection in areas 772.0-A8 and 772.0-A9.

5. RAI 2.a.iv

a. The discussion of the determination of release category characteristics (i.e.

source terms) indicates that the SEQSOR methodology was used. This methodology does not calculate release fractions from first principles but uses input from other calculations and has been used in the past for Sequoyah. The WBN Unit 2 lPE submittal discusses the use of Modular Accident Analysis Program (MAAP) 4.0.7 for the LERF analysis. Clarify the origin of the source terms used for the SAMDA analysis. Note that in the July 23, 2010, TVA response to RAI 2.f, TVA took the position that results from MAAP analysis were E1-5

more valid than those from SEQSOR. Clarify this apparent change in TVA's position.

TVA Response:

The WBN Unit 2 IPE submittal did use the Modular Accident Analysis Program (MAAP) 4.0.7 for the LERF analysis. The use of this code, however, was specifically for the thermal hydraulic analyses needed to justify sequence success criteria, not for source term calculations. TVA has not taken the position that release fractions computed by MAAP are more or less valid than the SEQSOR approach for WBN.

In the original IPE for WBN Unit 1 (published in 1992), release fraction results from MAAP 3B were used to satisfy IPE requirements. The latest WBN Unit 2 CAFTA model, used for the SAMA analyses, results in different dominant scenarios for each Release Category than those previously assessed. Since the SEQSOR approach makes use of release fractions computed for WBN Unit 2s sister plant, Sequoyah, it was deemed prudent to use these same inputs for the SAMA evaluation.

The SEQSOR methodology itself uses release fractions originally calculated from first principles for the Sequoyah plant. The origin of the source terms used for the revised SAMA analysis is described in the previously submitted response to RAI 2.a.iv. The approach is to identify dominant scenarios for each release category specific to WBN Unit 2, note their specific release characteristics in terms of accident progression, and then follow the SEQSOR methodology to arrive at mean release fractions for each Release Category.

The SEQSOR emulator developed for use in this analysis is described in the following paragraphs.

The SEQSOR program used to calculate the source terms for Sequoyah in NUREG-1150 was not available for this analysis. SEQSOR is explained in NUREG/CR-4551 Volume 5, Chapter 3 and Appendix B of that volume. Instead, for this analysis, a spreadsheet based SEQSOR Emulator was developed and used to emulate the basic calculation of the SEQSOR program used during the NUREG-1150 analyses.

The SEQSOR code was developed because of the large number of possible sequences a plant could undergo during an accident. The complexity and time of running a phenomenological code, such as MAAP, for each of these sequences would have been impossible. Instead, the SEQSOR code was developed as a relatively simple parametric code to select from a representative set of results of detailed phenomenological codes as probability distributions. This approach allows one to estimate a large number of cases in a short time.

SEQSOR uses blocks of data containing probability distributions, by release class (the nine release classes are groups of elements with similar chemical E1-6

behavior) for a variety of terms in the basic SEQSOR equations, given in equations 3.1 and 3.2 of NUREG/CR-4551, Volume 5. Equation 3.1 gives the behavior in the early phase, before the reactor vessel breach (if any). Equation 3.2 gives the behavior in the late phase, which considers the core-concrete interaction. Each of the data blocks represents a term in one or both of these equations and the data in each is a function of a probability level between 0%

and 100%, and in most cases is also a function of the radionuclide group. During the Monte-Carlo process, a random variable between 0 and 1 is used to select a value (or a set of values for each radioisotope group) for the calculation. The same data blocks were used in the SEQSOR emulator, except where processes or equipment that needed to be considered for this analysis were not included in the NUREG-1150 analyses. The blocks considered in both SEQSOR and the SEQSOR emulator are shown below:

Early releases from fuel to RCS (two oxidation states)

Fraction of core released from RCS before vessel breach (six conditions)

Fraction of containment contents released in the early and late phase (nine containment failure modes)

Containment state (wet or dry)

Sprays during early phase (three conditions)

Number of holes in the RCS if vessel failure (one or two)

DCH release (two conditions)

Core-concrete interaction pool scrubbing (two conditions)

SGTR scrubbing (three conditions)

Ice condenser decontamination factor for early phase (three conditions)

Ice condenser decontamination factor for late phase (four conditions)

Fraction of core mass ejected from high pressure mass ejection (four conditions)

Early phase ice condenser bypass (three conditions)

Late phase ice condenser bypass (three conditions)

All of these data blocks with the exception of the SGTR scrubbing were used in the original SEQSOR analysis. Three additional containment failure modes were considered: intact containment, failure to isolate, and late unfiltered vent.

The SEQSOR Emulator was developed to use the same SEQSOR logic but in a spreadsheet format. The SEQSOR Emulator was independently reviewed prior to use for SAMA analysis. For each Release Category, the fraction of each condition for each set of data blocks is determined from the scenarios included.

Then, a set of 4096 Monte-Carlo samples from the resulting probability space was selected and the resulting release fractions calculated according to the computational rules in equations 3.1 and 3.2 from the SEQSOR approach. This distribution of release fractions was used to determine the mean and other statistical measures for each radionuclide group. Finally, these release fractions and other sequence attributes (e.g., for timing and release points) were used in MACCS2 to determine the consequences for each sequence sub-category.

b. The discussion of the source terms states that the, The source terms for each set of accident characteristics are weighted in accordance with the % contribution for each release type in Table 2.a.iv-3." This process is valid only if the consequences of the releases are linear with respect to the source terms. This is E1-7

not necessarily true. Provide support that this process provides a valid estimate of consequences.

TVA Response:

The quoted statement could have been more clearly worded. The actual approach taken is as stated in the final sentence of the same paragraph describing Table 2.a.iv-3; Each release characteristic type in Table 2.a.iv-3 are calculated separately with MACCS and then weighted by frequency to determine the averaged dose and economic consequences shown in Table 2.a.iv-6. TVA agrees that the doses and consequences used for SAMA evaluations cannot be assumed linear and must instead be calculated separately and then frequency weighted. This is what was done.

In addition to the average weighted dose and economic consequences applied to the SAMA evaluations, as sensitivity, we have here instead applied the worst accident sequence dose and consequences of the sub-categories from each release category to the SAMAs to see if any of the Phase 2 SAMA results would then become cost beneficial. After applying the worst accident doses and consequences, we then reviewed the SAMAs for cost benefit using the CDF 95th percentile sensitivity. The SAMAs remain well below the cost benefit threshold except for one, SAMA 93.

This new sensitivity of cost and benefits was evaluated using the 2.70 multiplying factor for the CDF 95th percentile. Applying the 2.78 multiplying factor for the CDF 95th percentile, as suggested in NRC item 12, would not change these results developed using the worst accident doses and consequences. Again, the one exception is for SAMA 93 which is discussed further below.

SAMA 93 proposes to install an unfiltered hardened containment vent for the prevention of containment failure in late release sequences. SAMA 93 has a cost benefit of 0.96 at the CDF 95th percentile when using the average late release category dose and economic consequences for the baseline. If the worst case doses and consequences were instead used, the cost benefit ratio would exceed 1.0 for the CDF 95th percentile sensitivity. However, if implemented SAMA 93 would reduce the Level 2 release consequences of the late Release Category but would not change the core damage frequency. The mix of sequences contributing to the late release category would also not be changed by implementing SAMA 93. Therefore, we maintain that the baseline risk should use the weighted average accident dose and consequences in the baseline risk evaluation so that the cost benefit of the CDF 95th percentile would remain at 0.96 for SAMA 93. Further, using the 2.78 multiplying factor for the CDF 95th percentile for SAMA 93 and average accident doses and consequences for the baseline would result in a cost benefit of 0.99 which remains below the cost benefit threshold.

Examination of the contributors to the frequency of the Late Release Category, which would be mitigated by implementation of SAMA 93, reveals that 40% of the frequency is attributed to RCP seal LOCAs with leakages of 182 gpm per pump.

E1-8

TVAs response to Item 16 (Reference 1) commits to follow the progress of the seal package design defined in SAMA 58 and to install it if proven reliable.

Implementation of SAMA 58 would greatly reduce the benefits of SAMA 93.

Further, in the Reference 1 response to item 16, TVA further committed to re-evaluate the benefits of SAMAs 215, 226, 50, 55, and 56 for mitigation of RCP seal LOCA scenarios if SAMA 58 is not proven reliable. TVA further commits to add SAMA 93 to this list for re-evaluation should SAMA 58 not prove reliable.

Note also that scenarios in which there is a loss of control air (e.g., during a station blackout) and failure of the operators to locally control the AFW LCVs to allow the turbine-driven AFW pump to continue steam generator cooling contribute 10% to the Late Release Category frequency. TVA has committed to implement SAMA 339 (in place of the new air accumulators called for in SAMA

70) to provide a new capability to allow the operators to transfer from the normal compressed air supply to the station nitrogen system for control of the AFW LCVs for these sequences. See the response to item 11a for RAI 5.b.

Implementation of SAMA 339 will further decrease the evaluated benefit of SAMA 93, further affirming that it is not cost beneficial even for the CDF at 95th%

sensitivity assessment.

c. The discussion of release category definitions and contributors in response to RAI 2.a.ii and iii indicates that early steam generator tube ruptures (SGTRs) are assigned to the BYPASS release category under the contributor LERF-SGTR (SLERF) corresponding to the SLERF CET end state. The WBN Unit 2 IPE indicates that thermally induced SGTRs make up 32 percent of the WBN Unit 2 LERF. The SLERF end state is not included in the RAI revised model dominant CET end states listed in Table 2.a.iv-2. It is noted that plant damage state (PDS) bin 4B1, which, according to Table 2.a.i-2, is made up of large SGTR sequences, is not represented in the dominant CET end states. Clarify the reason for this and describe the SGTR contribution to WBN Unit 2 consequences.

TVA Response:

There are 2 types of SGTR sequences that contribute to LERF releases in the Containment Event Tree (CET):

1) SLERF Sequences - SGTR Initiating Events that lead to Core Damage (PDS Bin 4B1 or 4B2) and release.

2) BLERF Sequences - Non-SGTR Initiating Events that lead to Core Damage (PDS Bin 2) and Thermal-Induced SGTR (TI-SGTR) or Pressure-Induced SGTR (PI-SGTR) in the CET.

In the WBN Unit 2 IPE model, both BLERF and SLERF sequences contributed to LERF frequency. In the WBN Unit 2 SAMA model, the SGTR IE sequences are separated from the Non-SGTR IE sequences; however, both types of sequences still contribute to LERF frequency.

The sequences that contribute to LERF are shown in Table 5c-1 below. The SGTR sequences contribute negligibly. In the WBN Unit 2 SAMA Model, the Non-SGTR IE (BLERF) sequences that lead to TI-SGTR contribute 3.41E-07 E1-9

(22%) of the overall LERF frequency of 1.61E-06. The difference from the WBN Unit 2 IPE Model is due to the changes made in the WBN Unit 2 SAMA model as documented in Attachment B.4 of the October 2010 SAMA submittal.

Table 5c-1 CET Sequences that Lead to large Early Release Frequency (LERF)

CET End Sate PDS Bin SBO Non-SBO CET Sequence Events CET Sequence Frequency BLERF-01 2

Non-SBO Not PI-SGTR, Not RCS Depressurization, TI-SGTR 3.97E-11 BLERF-02 2

Non-SBO PI-SGTR 6.32E-09 BLERF-04 2

SBO Not PI-SGTR, Not RCS Depressurization, TI-SGTR 2.84E-07 BLERF-05 2

SBO PI-SGTR 5.10E-08 (Non-SGTR Sequences) BLERF Subtotal 3.41E-07 SLERF-01 4B1 Non-SBO (Containment Bypassed - No CET Events) 4.95E-10 SLERF-02 4B2 Non-SBO (Containment Bypassed - No CET Events) 5.13E-10 SLERF-03 4B1 SBO (Containment Bypassed - No CET Events)

<1E-14 SLERF-04 4B2 SBO (Containment Bypassed - No CET Events) 7.69E-13 (SGTR Sequences) SLERF Subtotal 1.01E-09 BLERF and SLERF Total 3.42E-07

d. The development of the RAI revised source term characteristics given in Table 2.a.iv-3 includes four contributors to the late release category, whereas Table 2.a.iv-2 identifies six dominant CET end states from three different PDS bins. Explain the development of the four late release category contributors and their weighting.

TVA Response There are actually 12 CET sequences rather than the 6 CET sequences listed in Table 2.a.iv-2 that contribute to the release characteristics in Table 2.a.iv-3.

All 12 sequences are shown in Table 5d-1 below. The 12 CET sequences were combined into 6 for development of the source terms. Each sequence characteristic contributes a percentage to the release characteristics as shown in the table. The percentages are rounded off to the values shown in Table 2.a.iv-3.

E1-10

Table 5d-1 CET Sequences that Contribute to the Late Release Category PDS Bin CET Sequence Frequency Level I Core Damage Sequence Release Characteristics RCS Pressure (Hi or LO)

Hi Lo Hi Lo ARFS Success or Failed (S or F)

F F

S S

Level 2 Release Sequence 6.5%

28.8%

2.4%

60.6 1

10%

LATE_CA-05 1.35E-06 Steam Break Outside Containment, SG isolated, SI terminated, PORV fails to reclose, SI restarted, HPR fails RCS Depressurization, ARFS available, No containment failure early, No CHR 8.9%

1 LATE_CA-01 3.62E-07 Steam Break Outside Containment, SG isolated, SI terminated, PORV fails to reclose, SI restarted, HPR fails No RCS Depressurization, ARFS available, No DCH, No containment failure early, No CHR 2.4%

2 60%

LATE_CNA-28 3.64E-06 Flood IEs leading to SBO and CD Not Bypassed, High RCS Pressure and SG Dry RCS Depressurization, LPME, ARFS failed, No containment failure early, No CHR 24.0%

2 LATE_CNA-09 2.71E-06 Loss of ERCW or CCS leading to CD Not Bypassed Low RCS Pressure and SG Dry RCS Depressurization, ARFS available, No containment failure early, No CHR 17.8 2

LATE_CNA-24 2.95E-07 LOSP leading to SBO and CD Not Bypassed, High RCS Pressure and SG Dry RCS Depressurization, ARFS failed, No containment failure early, No CHR 1.9%

2 LATE_CNA-20 9.15E-07 Flood IEs leading to SBO and CD Not Bypassed, High RCS Pressure and SG Dry No RCS Depressurization, HPME, ARFS failed, No containment failure early, No CHR 6.0%

2 LATE_CNA-11 1.41E-07 LOSP or Loss of ERCW or CCS leading to CD Not Bypassed Low RCS Pressure and SG Dry RCS Depressurization, ARFS failed, No containment failure early, No CHR 0.9%

2 LATE_CNA-12 1.06E-07 LOSP or Loss of ERCW or CCS leading to CD Not Bypassed Low RCS Pressure and SG Dry RCS Depressurization, ARFS failed, No containment failure early, No CHR 0.7%

2 LATE_CNA-16 8.34E-08 LOSP leading to SBO and CD Not Bypassed, High RCS Pressure and SG Dry No RCS Depressurization, ARFS failed, No containment failure early, No CHR 0.5%

3A, 3B 30%

LATE_CA-09 4.92E-06 Loss of ERCW or CCS leading to CD Not Bypassed Low RCS Pressure and SG Wet (3A) or Dry (3B)

ARFS available, No containment failure early, No CHR 32.4 3A, 3B LATE_CA-36 1.98E-07 Flood IEs leading to SBO and CD Not Bypassed, High RCS Pressure and SG Dry LPME, ARFS failed, No containment failure early, No CHR 1.3%

3A, 3B LATE_CA-10 2.22E-07 Loss of ERCW or CCS leading to CD Not Bypassed Low RCS Pressure and SG Wet (3A) or Dry (3B)

ARFS available, No containment failure early, No CHR 1.5%

e. Table 2.a.iv-6 gives the October 2010 RAI offsite population dose for release category III as 8.19E06 person-rem. This is a significant increase over the original October 2010 result of 1.13E06 person-rem, and when multiplied by the release category frequency of 1.3E-05 per year gives an annual population dose of 107 person-rem. This is a factor of 10 higher than that given in Table 5.c-1.

Confirm that the value in Table 2.a.iv-6 should be 8.19E05 person-rem.

TVA Response The value should be 8.19E+05 person-rem.

E1-11

f.

Provide the revised Off-Site Exposure Cost and Off-Site Economic Cost used to develop the maximum averted cost risk (MACR) as given in Section 5.1 and 5.2 of the October 14, 2010, RAI response submittal (and On-Site Exposure Cost in Section 5.3 and On-Site Cleanup and Decontamination Cost and Replacement Power Cost in Section 5.4 if these have changed for any reason) of the October 14, 2010, submittal.

TVA Response Below are the revised costs for the SAMA MACR. The table shows the October 2010 costs as compared to the Revised RAI MACR costs.

Cost Category October 2010, SAMA Report Revised (RAI)

SAMA Results Off-Site Exposure Cost $

$514,379

$369,019 Off-Site Economic Cost $

$466,032

$718,959 On-Site Exposure Cost $

$8,153

$8,153 On-Site Economic Cost $

$666,023

$666,023 Total Base Cost $

$1,654,587

$1,762,154 Base Cost with External Event Multiplier 2.0

$3,309,174

$3,524,309 Base Cost with External Event Multiplier 2.28

$3,772,461

$4,017,712

15. RAI 6 This RAI requested an assessment of the impact of uncertainty on the Phase I screening of SAMDAs due to either excessive cost or very low benefit similar to that given in response to the original RAI 7.a. The response to this new RAI does not provide this assessment. While Table 19 of the October 14, 2010, SAMDA submittal is cited, this table addresses the impact of uncertainty on the Phase II cost benefit analysis not the Phase 1 screening. The current marginal MACR is a factor of 2.6 times that on which the original screening was performed, while the risk profile of the current PRA is considerably different from that used in the original screening. These changes could impact the judgments made in the Phase I screening without requiring a Phase II cost evaluation. Provide the requested assessment.

TVA Response:

This response provides an assessment of the impact of uncertainty on the Phase I screening of SAMAs due to either excessive cost or very low benefit. A comparison of the Maximum Averted Cost of Risk (MACR) for the Revised RAI SAMA results are shown in Table 15-1. The Revised RAI sensitivity results show the 95% uncertainty (multiplier of 2.7) MACR for an External Event multiplier of 2.28 is $10,847,822. This $10,847,822 MACR value can be used to screen the Phase I SAMAs for either Excessive Cost or Very Low Benefit.

E1-12

Table 15 Revised RAI SAMA Maximum Averted Cost of Risk (MACR) Results Cost Description Revised (RAI) SAMA Results Total MACR MACR Onsite Costs (Function of Core Damage without Release Costs)

MACR Offsite Costs (Function of Core Damage and Release Costs)

Base Cost with External Event Multiplier 2.0

$3,524,309 $1,348,352 38%

$2,175,957 62%

Base Cost with External Event Multiplier 2.28

$4,017,712 $1,526,731 38%

$2,490,981 62%

95% Cost with External Multiplier 2.0 (95%

Multiplier 2.70)

$9,515,634 $3,615,941 38%

$5,899,693 62%

95% Cost with External Multiplier 2.28 (95%

Multiplier 2.70)

$10,847,822 $4,122,173 38%

$6,725,650 62%

If we assume 100% Core Damage Frequency (CDF) reduction is worth

$10,847,822, we can estimate the potential change in MACR of various impacts from proposed SAMAs. Table 15-2 shows the potential change of MACR of 7 cases of risk reduction according to what aspects of the CDF and Release category frequencies the proposed change impacts.

Table 15-2. 95% MACR Risk Reduction Case Types SAMA Case CDF LERF (Early &

Bypass)

LATE SERF Contribution to MACR Potential Change in MACR 1

Changed Linear Linear Linear 100.0% $10,847,823 2

Fixed Changed Fixed Fixed 10.6%

$1,150,720 3

Fixed Fixed Changed Fixed 47.9%

$5,201,087 4

Fixed Fixed Fixed Changed 3.2%

$345,785 5

Changed Changed Fixed Fixed 14.2%

$1,538,656 6

Changed Fixed Changed Fixed 77.0%

$8,348,905 7

Changed Fixed Fixed Changed 11.7%

$1,272,159 Case 1 is calculated as a linear function of the total MACR based on the fractional change in CDF. For Cases 2 through 4, the values are calculated by zeroing the release category frequency contribution and subtracting the result from the total MACR. For Cases 5 through 7, the values are calculated by zeroing the release category frequency contribution, subtracting the release frequency from the CDF, and then subtracting the changes from the total MACR.

For SAMA Case 1 in Table 15-2, if the risk reduction changed the CDF and the LERF, LATE and SERF Release categories are changed linearly, the MACR available for reduction would be the full $10,847,823. A 1% reduction in CDF would be worth $108,478, a 10% reduction in CDF would be worth $1,084,782 and so forth.

E1-13

E1-14 For SAMA Cases 2 through 7, if the risk reduction change affected the LERF, LATE, SERF or combination of CDF and LERF, CDF and LATE, or CDF and SERF, and the other contributors to MACR were fixed, the table shows the potential change in MACR available for reduction. Based on comparing the SAMA change to one of these SAMA cases, we can screen for excessive costs based on the maximum potential change in MACR.

In addition to the potential maximum MACR available for reduction

($10,847,823), we can screen based on the potential benefit using the 7 case values in Table 15-2. Costs would have to be equal to or less than the corresponding maximum potential change in MACR to be cost beneficial.

Using the above cost benefit screening guidelines we have reviewed the Phase I SAMAs for Excessive Costs, Very Low Benefit, or low cost benefit (cost greater than potential benefit). All of the Phase I SAMAs that were screened for Excessive Cost or Very Low Benefit were rescreened and the results provided in Table 15-3. In estimating the potential benefits of each SAMA, often the estimated percentage reduction in MACR was based on observations of RRWs for basic events in the model that would be affected by implementing the associated SAMA. Table 15-3 is sorted first by disposition and then by SAMA number. For others, knowledge of the implementation cost was used to qualitatively bound the anticipated benefit. All of the Phase I SAMAs remain screened out. This conclusion is unchanged if instead of using frequency weighted average consequences for each Release Category to compute MACR, the worst sub-category doses and consequences are conservatively used.

Some of the Phase I SAMAs pertaining to mitigation of Seal LOCA events may be retained for consideration along with the other Seal LOCA SAMAs being considered in Phase II. See the TVA response to item 16 (Reference 1).

Table 15-3. Phase I SAMA Candidates SAMA Number SAMA Title SAMA Discussion Source Phase I Comments Disposition Excessive Implementation Cost. (Table 15-2 Case 1) 2 Replace lead-acid batteries with fuel cells.

Extended DC power availability during an SBO.

ML11166A173

Text

Subject:

Subject:

Subject:

References:

Navigation menu

Search