ML110480686

Lr - Draft Seabrook SAMA Supplemental RAIs

ML110480686
Person / Time
Site:	Seabrook
Issue date:	02/17/2011
From:	Office of Nuclear Reactor Regulation
To:
References
Download: ML110480686 (7)

Text

1 SeabrookNPEm Resource From:

Wentzel, Michael Sent:

Thursday, February 17, 2011 1:55 PM To:

Cliche, Richard

Subject:

Draft Seabrook SAMA Supplemental RAIs Attachments:

Draft SAMA Follow-up RAIs.docx

Rick, Attached are the draft supplemental RAIs related to the SAMA review of the Seabrook ER. Please review and provide any comments. Additionally, if you would like to have a phone call to these, please let me know.

Thanks, Mike MichaelWentzel ProjectManager NRR/DLR/RPB1 (301)4156459 michael.wentzel@nrc.gov

Hearing Identifier:

Seabrook_License_Renewal_NonPublic Email Number:

604 Mail Envelope Properties (C0A338EE37A11447B136119705BF9A3F9FF94470AD)

Subject:

Draft Seabrook SAMA Supplemental RAIs Sent Date:

2/17/2011 1:54:52 PM Received Date:

2/17/2011 1:54:56 PM From:

Wentzel, Michael Created By:

Michael.Wentzel@nrc.gov Recipients:

"Cliche, Richard" <Richard.Cliche@fpl.com>

Tracking Status: None Post Office:

HQCLSTR02.nrc.gov Files Size Date & Time MESSAGE 403 2/17/2011 1:54:56 PM Draft SAMA Follow-up RAIs.docx 45789 Options Priority:

Standard Return Notification:

No Reply Requested:

No Sensitivity:

Normal Expiration Date:

Recipients Received:

1. The responses to RAI 1.f and 2.e state that Tables F.3.1.1.1-2 and F.3.2.1-2 consider initiating events. However, Tables F.3.1.1.1-2 and F.3.2.1-2 show only basic events, and so it is not clear how initiating events are explicitly considered in the importance analysis. For example, the most dominant initiating event contributor according to the listing in Table F.3.1.1.1-1 is loss of offsite power due to weather (LOSPW), but there appears to be no corresponding event in Tables F.3.1.1.1-2 or F.3.2.1-2. Additionally, the responses to RAI 1.f and 2.e state that Attachment F.A describes a number of SAMA probabilistic risk assessment (PRA) cases that specifically addressed initiating events.

While this is the case, it is unclear if all of the dominant initiating events in Table F.3.1.1.1-1 are addressed. example, while Attachment F.A describes NOLOSP as the modeling case used to determine the benefit of eliminating all Loss of Off-Site Power (LOSP) events, none of the SAMAs evaluated using this case (i.e., SAMA s 13, 14, 16, 24, and 156) specifically addressed initiating event FLLP, Flood in Turbine Building (causing LOSP). Other examples from Attachment F.A are PRA cases FIRE1 and FIRE2, which are the PRA model cases used to determine the benefit of eliminating control room induced LOCA and fire in Turbine Building causing loss of power to emergency buses, respectively. Neither of these modeling cases addresses the specific scenario of a fire in Switchgear Rooms A or B and no SAMAs were identified or developed to specifically address fire in Switchgear Rooms A and B. Please provide an explanation, including a table, which correlates SAMA candidates to each of the risk-important initiating events and identifies whether there are any potentially lower cost plant-specific SAMA candidates for the specific initiating events identified. Include, as an example, a demonstration in Tables F.3.1.1.1-2 and F.3.2.1-2 of how LOSPW is considered.

2. Concerning the NextEra response to staff RAI 3.b:

a. Relative to the FIVE fire risk analysis methodology, clarify if fire-induced failures of components or human actions credited for mitigating the initiator are assessed, given that the initiator is now fire-induced (including identifying either new components or actions to be modeled or requantifying the random failure probabilities of previously modeled ones). In addition, clarify if hot shorts are considered and, if so, what probabilities are assigned to these. If the assigned probabilities are not bounded by currently accepted values (e.g., best estimates of 0.3 and 0.6, respectively, for intra-cable hot shorts for circuits protected vs. not protected by control power transformers in Tables 10-1 through 10-5 of NUREG/CR-6850 [2005], which are based on Electric Power Research Institute (EPRI) 1006961 [2002]), either provide a technical basis for the values assumed (such as circuit analysis) or perform a sensitivity evaluation using the currently accepted values to determine the effect on the potential cost-beneficiality of the SAMAs.

b. Relative to the Fire-induced interfacing system loss-of-coolant accident (ISLOCA)

& Containment Impact section. For the only area where isolation valves both inside and outside containment could be affected, clarify how important isolation valves [that] could be controlled locally at the valve would be so controlled for valves inside containment during operation. For the letdown system and its several; fail closed air-operated valves (AOVs), clarify the basis for the statement it is not credible for all three valves to hot short. For isolation failure of one or more valves of a single train, clarify the statement the ability to

remove power from fail closed valves to provide isolation and the case for failed open valves due to hot shorting (which could not be de-powered?).

3. The response to RAI 3.c.a states that [h]owever, the present PRA model has been updated to the more recent EPRI hazards, and that [t]his was done because, while the methodology and experts used in developing the EPRI hazard are essentially the same as the Seabrook Station Probabilistic Safety Assessment (SSPSA), the EPRl hazard is more recent and the EPRl uniform hazard spectrum (UHS) developed for the Seabrook Station site is more realistic than that used in the SSPSA. Since the EPRI hazard curve was assessed in the individual plant examination external events (IPEEE) as a sensitivity case, and the SSPSA curves were considered the baseline, please clarify the basis for considering the SSPSA curves as the baseline and the EPRI curves as the sensitivity case, and why replacing the SSPSA hazard curves with the EPRI hazard curves is now more appropriate as the baseline representation for Seabrook Station.

4. A seismic core damage frequency (CDF) of 2.2E-05 per year, based on the updated 2008 United States Geologic Survey (USGS) seismic hazard curves (as determined from GI-199 information), was used to develop a factor of 2.6 multiplier on the maximum attainable benefit (MAB). The potential effect of this factor is three-fold: (1) for those Phase II SAMAs that were previously screened on high cost or low benefit, the use of this factor could result in some of these SAMAs being reconsidered for a quantitative Phase II evaluation, (2) for those Phase II SAMAs previously evaluated quantitatively and dismissed as non-cost-beneficial, the use of this factor could now render some of these potentially cost-beneficial, and (3) for those Phase II SAMAs previously evaluated and determined to be cost-beneficial, the use of this factor could increase the degree of cost-beneficiality. The revised MAB based on the 2.6 factor was used to re-assess only those Phase II SAMAs that were qualitatively screened based on high cost or low benefit (item 1). Phase II SAMAs that were quantitatively evaluated in the original analysis were not re-assessed in the RAI response (items 2 and 3). Based on a scoping assessment by the NRC staff, applying the 2.6 multiplier to the estimated benefit for these non-re-assessed SAMAs will result in many SAMAs becoming potentially cost-beneficial, based on the current cost estimates. Please provide an assessment of the impact of the higher seismic CDF on all SAMAs, including those identified and evaluated in response to the staffs original RAIs. Note that, as discussed in the Attachment to this letter, this re-evaluation applies to all SAMAs, whether or not they result from internal or external event considerations. Specifically, please discuss whether random failures are included in the seismic analysis and, if not, use the 2.6 factor to assess the impact of the higher seismic CDF on the SAMA evaluation.

5. The response to staffs RAI 5.b states that all top ranked basic events related to large early release frequency (LERF) have been addressed in response to RAI 2.f. This appears to be the case with the exception of basic event FWP161.FS, which has a LERF risk reduction worth (RRW) of 1.0886 (see Environmental Report Table F.3.2.1-2).

Please provide an assessment of this basic event for SAMAs, including an identification of the specific basic events that bound this basic event and the associated SAMAs.

6. The response to staffs RAI 5.b provides the same disposition (not beneficial based on eliminating all supplemental electric power system (SEPS) failures), same associated SAMAs (#9 and #14), and same two pairs of case studies (PRA cases SEPS for #s 8-9 and OSEP1 for #s 10-11) for basic events #8, #9, #10, and #11, yet the estimated benefits are different between the pairs #s 8-9 and #s 10-11. Clarify the differences in

the treatment of these four SAMAs, including the two different case studies and provide additional details on the modeling assumptions for each.

7. Clarify why NextEra believes the uncertainty distribution represents the uncertainty in the fire and seismic portions of the PRA model. Include a discussion of whether probability distributions were assigned for external events (such as for fire ignition frequencies, non-suppression probabilities, hot short probabilities, seismic frequencies, other seismic parameters) and, if not or only done so partially, how this impacts the SAMA analysis, including any specific examples of where probability distributions were applied in the fire and seismic models.

Attachment The NRC staff position is that a higher seismic CDF will have the following two impacts: (1) for a SAMA identified to address only seismic events, the Phase II evaluation will result in a higher estimated benefit for the seismic SAMA and (2) for a SAMA identified to address internal events, the Phase II evaluation could result in a higher estimated benefit for the internal event SAMA if both seismically-induced failures and random failures are included in the model and the random failure of the component(s) being addressed by the SAMA show up in the seismic cutsets. This is illustrated by the following:

Internal initiator = I; External initiator = E Non-externally-event-induced (i.e., random) failure basic event for component X = X0 Externally-event-induced failure basic event for component X = X1 Other conditional failures due to I = A B, where A = 1 given an external event initiator (E)

Base case CDF = {I X0 A B} + {E (X0 + X1) B}

(1) Assume SAMA involves component X. If effect is internal-event-related (i.e., non-external SAMA), X0 changes to X0, while there is no change in X1 SAMA-case CDF = {I X0 A B} + {E (X0 + X1) B}

CDF = {I (X0 - X0) A B} + {E (X0 - X0) B} = X0 (I A + E) B (2) Assume SAMA involves component X. If effect is external-event-related (i.e., external SAMA), X1 changes to X1, while there is no change in X0 SAMA-case CDF = {I X0 A B} + {E (X0 + X1) B}

CDF = E (X1-X1) B = X1 E B In both cases, if there is a change in E to E, such as an increase in frequency, the CDF will change, since both equations include a contribution from E:*

(1) (CDF) = {X0 (I A + E) B} - {X0 (I A + E) B} = X0 (E - E) B =

X0 E B (2) (CDF) = X1 (E - E) B = X1 E B Therefore, whether or not the SAMA is external, there will be a change in the CDF due to a change in the external initiator frequency (E).*

• The only situation where there would be no change in CDF [i.e., (CDF) = 0] would be if an external SAMA did not affect X1, i.e., case (2) with X1 = 0. Recall that we have assumed that an internal SAMA can only affect X0 and an external SAMA only X1.

However, X0 can occur whether the initiating event is internal or external (hence the presence of X0 following both I and E), while X1 can only occur following E (hence it only appears following E, never I).

The preceding calculations are illustrated using the fictitious values below, where:

Initiator Frequency - Internal & External = I & E Basic Event Probability -

Base Internal & External = X0 & X1 SAMA Internal & External = X0 & X1 CCDP - Given Internal and External* = (A B) & B Case CDF-Internal CDF-External CDF-Total Delta Base SAMA Base SAMA Base SAMA CDF (1) 8.5E-08 4.6E-08 2.7E-07 1.7E-07 3.5E-07 2.1E-07

-1.4E-07 (2) 8.5E-08 8.5E-08 2.7E-07 2.6E-07 3.5E-07 3.4E-07

-8.2E-09 External Initiator Frequency - Updated (x 2.2) 1.7E-02 Case CDF-Internal CDF-External CDF-Total Delta Base SAMA Base SAMA Base SAMA CDF (1) 8.5E-08 4.6E-08 5.9E-07 3.6E-07 6.8E-07 4.1E-07

-2.6E-07 (2) 8.5E-08 8.5E-08 5.9E-07 5.7E-07 6.8E-07 6.6E-07

-1.8E-08 Case Delta-Delta CDF Check (1)

-1.2E-07

-1.2E-07 As the results indicate, whether or not the SAMA is internal or external, the effect of a change in the external event initiator frequency is manifested as a change in the CDF (i.e.,

Delta-Delta-CDF 0). In this example, since the effect of the SAMA was assumed to reduce to failure probability of the basic events associated with component X, i.e., X0 < X0 and X1 < X1, and the external event initiator frequency was assumed to increase due to the update, i.e., E >

E, a larger decrease (more negative) in CDF (i.e., Delta-CDF) results in both cases, such that the effect of the increase in external event initiator frequency is to increase the magnitude (make more negative) of the CDF reduction that can be realized by either an internal or external SAMA. As a result, the potential for a SAMA to be cost-beneficial increases since the larger reduction in CDF translates into an increase in the potential benefit.

• For simplicity, only the product A B and B itself are shown. It can be deduced that A = (9.5E-4)/(8.8E-3) = 0.105.