DCL-10-106, Response to NRC Letter Dated July 6, 2010, Request for Additional Information for Applicants Environmental Report - Operating License Renewal Stage

Response to NRC Letter Dated July 6, 2010, Request for Additional Information for Applicants Environmental Report - Operating License Renewal Stage

ML102440444
Person / Time
Site:	Diablo Canyon
Issue date:	08/27/2010
From:	Becker J Pacific Gas & Electric Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
DCL-10-106
Download: ML102440444 (133)

Text

El Pacific Gas and Electric Company' James R. Becker Diablo Canyon Power Plant Site Vice President Mail Code 104/5/601 P 0. Box 56 Avila Beach, CA 93424 August 27, 2010 805.545.3462 Au ut27 0 0Internal:

691.3462 Fax: 805.545.6445 PG&E Letter DCL-10-106 U.S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, DC 20852 Docket No. 50-275, OL-DPR-80 Docket No. 50-323, OL-DPR-82 Diablo Canyon Units 1 and 2 Response to NRC Letter dated July 6, 2010, Request for Additional Information for the Applicant's Environmental Report - Operating License Renewal Stage

Dear Commissioners and Staff:

By letter dated November 23, 2009, Pacific Gas and Electric Company (PG&E) submitted an application to the U.S. Nuclear Regulatory Commission (NRC) for the renewal of Facility Operating Licenses DPR-80 and DPR-82, for Diablo Canyon Power Plant (DCPP) Units 1 and 2, respectively. The application included the license renewal application (LRA), and Applicant's Environmental Report -

Operating License Renewal Stage.

By letter dated July 6, 2010, the NRC staff requested additional information needed to continue their review of the DCPP LRA. PG&E's response to the request for additional information is included in Enclosure 1.

PG&E makes no regulatory commitments (as defined in NEI 99-04) in this letter.

If you have any questions regarding this response, please contact Mr. Terence L. Grebel, License Renewal Project Manager, at (805) 545-4160.

I declare under penalty of perjury that the foregoing is true and correct.

Executed on August 27, 2010.

Sin erely, James R. Becker A member of the STARS (Strategic Teaming and Resource Sharing) Alliance Callaway

• Comanche Peak

• Diablo Canyon

• Palo Verde

• San Onofre

• South Texas Project

• Wolf Creek

Document Control Desk August 27, 2010 Page 2 pns/50329600 Enclosure PG&E Letter DCL-10-106 cc:

cc/enc:

Diablo Distribution Elmo E. Collins, NRC Region IV Regional Administrator Michael S. Peck, NRC Senior Resident Inspector Andrew L. Stuyvenberg, NRC Environmental Project Manager, License Renewal Alan B. Wang, NRC Project Manager, Office of Nuclear Reactor Regulation A member of the STARS (Strategic Teaming and Resource Sharing) Alliance Callaway

• Comanche Peak e Diablo Canyon o Palo Verde

• San Onofre o South Texas Project o Wolf Creek PG&E Letter DCL-10-106 Page 1 of 131 PG&E Response to NRC Letter dated July 6, 2010, Request for Additional Information for the Applicant's Environmental Report - Operating License Renewal Stage RAI l.a

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

a. Section F.2.1.2 describes changes made in PRA model DCPRA-1991 (the Individual Plant Examination [IPE] model) to include emergency diesel generator fuel transfer system and charging pump back up cooling. The Brookhaven National Laboratory review of PRA model DCPRA-1988 (NUREG/CR-5726) indicates that these revisions were included in the 1988 model. Clarify this discrepancy.

PG&E Response to Severe Accident Mitigation Alternatives RAI l.a The Diablo Canyon Power Plant Model DCPRA-1 988 was the first to implement the emergency diesel generator fuel oil transfer system and charging pump back up cooling. Model DCPRA-1991, the individual plant examination internal model, also contained those model changes as it was the successor to that model.

PG&E Letter DCL-10-106 Page 2 of 131 RAI 1.b

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

b. Section F.2.1.3 describes the changes made to model DCPRA-1988 to create model DCPRA-1993 (the Individual Plant Examination for External Events [IPEEE] model).

Confirm that the changes made to model DCPRA-1988 to obtain model DCPRA-1991 (as described in Section F. 2.1.2) were also made to the IPEEE model.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.b The model changes that were made to DCPRA-1 988 to make Model DCPRA-1 991 were subsumed into the next model revision DCPRA-1993, the individual plant examination for external events model. This is because the development of DCPRA-1993 began with Model DCPRA-1 991.

PG&E Letter DCL-10-106 Page 3 of 131 RAI 1.c

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

c. Provide the internal events core damage frequency (CDF) associated with the IPEEE model. This should be different from the IPE value 'since the 6th emergency diesel generator (EDG) added subsequent to the IPE would be expected to affect the internal event results.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.c The internal events core damage frequency for Model DCPRA-1 993 is reported as 8.8E-05/rx year, which is the same as DCPRA-1 991. We expect the addition of the sixth emergency diesel generator (EDG) to reduce the estimated core damage from internal events; however, at the time of the individual plant examination for external events submittal the addition of the sixth EDG was ongoing and not completed.

Modeling of the sixth EDG first appeared in the DCPRA-1 995 and continues through the current Probabilistic Risk Assessment Model DC01A.

PG&E Letter DCL-10-106 Page 4 of 131 RAI 1.d

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

d. Describe what is meant by "Level 2" in the initiating event "Non-Isolated Steam Generator Tube Rupture (SGTR) for Level 2" on page F-23. Describe the contribution to CDF from isolated SGTR initiating events.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 1.d The DC01A models two kinds of steam generator tube ruptures (SGTR): isolable and nonisolable. The latter are referred to as Level 2 in the initiating event description because they directly contribute to large early release frequency. Nonisolable and isolable SGTRs contribute 14.9 percent (1.2E-06) and 0.2 percent (2.2E-08),

respectively, to internal events core damage.

PG&E Letter DCL-10-106 Page 5 of 131 RAI i.e

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

e. Provide the contribution to the DCOIA internal event CDF due to Anticipated Transient Without Scram (A TWS) and station blackout (SBO) events.

PG&E Response to Severe Accident Mitigation Alternatives RAI I.e Total internal events contribution from anticipated transient without scram is 1.4 percent (1.5E-07) and the contribution from station blackout due to LOOP initiating events is 6 percent (5.1E-7).

PG&E Letter DCL-10-106 Page 6 of 131 RAI 1.f

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

f. PG&E indicates, in Section F.2. 1 that, since the two Diablo Canyon Power Plant (DCPP) units are essentially identical, no separate PRA model for Unit 2 was developed. Describe briefly how the availability or non-availability of Unit 2 impacts Unit I operation and how this is modeled in the Unit I PRA, including availability of shared systems, two-unit initiators, and cross ties between units.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 1.f The probabilistic risk assessment (PRA) model for Diablo Canyon Power Plant Unit 1 takes credit for cross-tying the auxiliary salt water (ASW) system, a shared system.

The detailed ASW model includes, in addition to Unit 1 components, Unit 2 pumps, valves, traveling screens and maintenance on the Unit 2 equipment as well as the cross-tie valves. There are also separate initiating events for loss of ASW due to system faults, and loss of ASW due to flooding that fails ASW for both units. Loss of Unit 2 ASW has no effect on Unit 1 core damage results unless it is needed by Unit 1.

The 4kV vital alternating current (AC) buses can also be cross-tied, which is credited in the PRA model. There are separate models for the Unit 2 vital buses and the breakers needed to cross-tie to the Unit 1 vital buses. Loss of Unit 2 vital buses has no impact on Unit 1 core damage unless it is needed by Unit 1.

PG&E Letter DCL-10-106 Page 7 of 131 RAI l.,q

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

g. Section F.2.3. 1 states that plant specific initiating event frequencies, failure rates, and maintenance unavailabilities are updated approximately every six years with the last update completed in March 2003. Section F.2.1.8 states that the component database in the DCOI model was updated using data through September 30, 2001.

Confirm, via a review of current initiating event frequencies and important component unavailability values, that updating the PRA with these values would not adversely affect the results of the SAMA evaluation.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1._q As part of creating a new fire probabilistic risk assessment (due to be completed at the end of this year) to replace the individual plant examination external events analysis, initiating events and plant equipment reliability data were collected and analyzed. Plant specific initiating events were created with Diablo Canyon Power Plant (DCPP) experience through August 31, 2009, with generic priors taken from NUREG/CR-6928.

The analysis shows that the frequency of all initiators would be reduced except for medium loss-of-coolant-accident (LOCA). This increase for a medium LOCA is due to an industry event where a damaged reactor vessel head was found. This event is not applicable to DCPP because Unit 1 had its vessel head replaced and the vessel head at Unit 2 will be replaced during the next outage. As a result the SAMA evaluation is not likely to be affected by integrating new initiating event frequencies.

Plant specific failure rates, and maintenance unavailability were also analyzed through the same time period. Most failure rates experienced a drop in value, although some increased. Some important component variables that increased in failure rate include the emergency diesel generator (EDG) fail to run during the first hour and the turbine driven pump fail to run. However, the EDG fail to run after the first hour and the EDG fail to start variables fell equally as much. The turbine-driven pump fail to start went up, but the fail to run also went down. As a result the SAMA eval6 ation is not likely to be affected by an update in component failure rates and maintenance.

PG&E Letter DCL-10-106 Page 8 of 131 RAI 1.h

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

h. Section F. 2.1.8 indicates that, except for the charging pump modification described in Section F. 2.1.9, the last update for design or operational changes was made for changes through 2004. Confirm that operational or design changes made since that time will not adversely affect the results of the SAMA'evaluation.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.h As part of preparing an updated fire probabilistic risk assessment (PRA) model, plant design and operational changes since 2004 (through August 31, 2009) were reviewed (PRA Calculation H.4, Revision 4). None of the design or operational changes will have an effect on the Diablo Canyon Power Plant model or the SAMA evaluation.

PG&E Letter DCL-10-106 Page 9 of 131 RAI 1.i

1. Provide the following information regarding the Level 1 Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

i. Identify the PRA model reviewed for each of the reviews described in Section F.2.3.2.

Clarify the model reviewed in the May 2000 Westinghouse Owners Group (WOG) review and the CDF results for and changes to that model as implied in the discussion of the review in Section 1 of the DCPP Risk-Informed In-Service Inspection (RI-ISI) submittal.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.i The following are the Diablo Canyon Power Plant (DCPP) models associated with Section F.2.3.2:

Westinghouse Owners Group (WOG) Peer Review: DCOO (Stage 1 Model)

Scientech/Jacobsen Flooding Review: DC01 ERIN Engineering Self-Assessment: DC01 ABS HRA Review: DC01 Westinghouse Level 2 Review: DC01 Pressurized Water Reactor Owners Group Cross Comparison Study: DC01 I

The DCPP risk-informed inservice inspection (RI-ISI) submittal was based on the DCOO model prior to incorporating the results of the WOG Peer Review. This model was called the Stage 1 model whose results are 5.05E-05/yr and 1.81 E-06/yr for core damage frequency and large early release frequency, respectively (Section 1.2 of RI-ISI Submittal). Note that these totals include internal events, internal fires, and seismic events. The DCOO Stage 1 model was the subject of the WOG Peer Review. No changes were made to the DCOO Stage 1 model as a result of the WOG Peer review.

The WOG findings were discussed in Section 1.2 of the RI-ISI submittal.

PG&E Letter DCL-10-106 Page 10 of 131 RAI 1.i

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

j. Section F. 4 states that the internal events CDF of 8. 44E-06 per year (for PRA model DCOIA at a truncation of IE-14 per year) was used for calculating the Maximum Averted Cost-Risk (MA CR). The CDF reported in Section F.2.1.9 for model DCOIA is 8.13E-06 per year. Address the disparities and the truncation limits on which the internal and external event CDF values given in Section F.2.1 were based.

PG&E Respronse to Severe Accident Mitigation Alternatives RAI 1.j The DC01A quantified core damage frequency (CDF) from internal events is 8.47E-06/yr. Section F.4 states 8.44E-06/yr, which is the sum of the endstate frequencies (used in the SAMA cost benefit calculations). This total is different than the quantified frequency due to rounding error. Section F.2.1.9 should have reported 8.47E-06/yr as the internal events core damage total. The total due to internal events, seismic events and internal fires is as follows:

Internal Events:

8.47E-06 Seismic Events:

3.77E-05 Fire Events:

1.39E-05 Total:

6.01 E-05 The disparity is only with the cited internal events CDF. CDF due to seismic and fire events are consistent with the SAMA calculations and are reported consistently.

PG&E Letter DCL-10-106 Page 11 of 131 RAI 1.k

1. Provide the following information regarding the Level 1 Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

k. The discussion of the assessment of open items/issues resulting from the reviews of the DCPP PRA in Sections F.2.3.4 and F.2.3.5 addresses only the impact on those systems, operator actions, etc., that have been identified in the importance analysis and then only with respect to the SAMA identification process. The assessment did not consider the impact on the evaluation of the benefit associated with any SAMA and the subsequent screening of the SAMAs. While Addendum I to Attachment F to the ER addresses the importance of each open item on the SAMA assessment, it is not clear that this assessment incorporates consideration of the impact on the evaluation of the benefit or subsequent screening of any SAMA. Provide further discussion of the evaluation of open items/issues to support the conclusion that none of the open items/issues will adversely impact the results of the SAMA assessment considering both the benefit evaluation and SAMA identification.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 1.k Addendum 1 contains a list of probabilistic risk assessment (PRA) open items that have been collected by Diablo Canyon Power Plant (DCPP). Each of the open items in this table has been addressed as to whether it impacts the SAMA screening process and list of candidate SAMAs. Given that the PRA impact of each of the open items is small, their resolution is not expected to change the conclusions of the SAMA analysis.

Separately in Sections F.7.2.1 and F.7.2.2, the Phase 1 and Phase 2 lists were re-examined using the 95th percentile of the core damage frequency. DCPP believes that this exercise in re-evaluating the SAMAs more than bounds the impact of PRA open items found in Addendum 1.

PG&E Letter DCL-10-106 Page 12 of 131 RAI 1.1

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

I. Provide definitions of the PRA Issue and Status codes used in Addendum 1. Confirm that all of the unresolved issues from the reviews identified in Section F.2.3.2 are addressed in the addendum.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.1 The following are the definitions of the codes used in Addendum 1:

Status This code is used to prioritize resolving the problem A - Investigate, update, revise model, item must be resolved B - No action required, issue resolved C - No or insignificant impact on the risk results, documentation issue PRA Issue This code is used to define the nature of the problem in terms of which portion of the PRA it belongs.

D - Internal flooding related issue E - Internal fire related issue F - Human reliability analysis related issue G - Level 2 related issue H - Seismic related issue J - General internal events related issue K - Other external events related issue Record #

Unique serial number assigned to each record.

PG&E Letter DCL-10-106 Page 13 of 131 RAI 1.m

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

m. In Addendum 1, Open Item 56 concerns the inclusion of expansion joints in the internal flooding modeling with the comment that expansion joints in the circulating water system would be included in the future. This item was judged to have no impact on the SAMA application because turbine building floods are not significant contributors to risk. Open Item 608 concerns crediting isolation of large turbine building floods to mitigate propagation to areas where equipment failure might be important. This item was judged to have no impact on the SAMA conclusion since SAMAs 4 and 5 are associated with alternative onsite power sources. There is the potential that the.

resolution of both these items could impact the SAMA analysis either by identifying additional SAMAs associated with the flooding events (barriers, automatic isolations,

'etc.) or increasing the benefit associated with the related SAMAs 4 or 5. Provide further support for the disposition of these two open items with respect to the SAMA assessment.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.m Failure of expansion joints has been added to the loss of auxiliary saltwater (LOSW) system initiating event frequency model in the fall of 1999. This initiator, despite the addition of expansion joints to the initiating event model, contributes less than 0.5 percent to core damage frequency (CDF). The flooding equivalent initiating event (Flooding Loss of Auxiliary Saltwater (FLLOSW)) for LOSW contributes less than 0.1 percent to CDF. The total contribution of all flooding initiators to core damage is less than 1 percent. It is unlikely that the addition of a loss of circulating water (CW) flooding initiating event will have an impact on CDF such that new SAMA candidates will arise or SAMA analyses will be affected (see further discussion of CW flooding analysis below). For the reasons listed below, any CW importance to CDF is likely bounded by auxiliary saltwater (ASW). And there are no SAMAs associated with ASW or ASW flooding.

The following is from the Diablo Canyon Power Plant (DCPP) flooding analysis:

A large uncontrolled flood of circulating water in the turbine building has the potential to damage the diesel generators at the 85 ft. elevation. This scenario was addressed in the Final Safety Analysis Report (Reference F.4-1 1) and in a DCPP turbine building flooding analysis (Reference F.4-12). A flood of this nature was considered unlikely for the following reasons:

PG&E Letter DCL-10-106 Page 14 of 131

1.

The CW pump trip switch system employs three float switches seismically mounted to the condenser pit walls. They are placed 3 ft. 9 in. below the turbine building floor elevation (85 ft.). When the contacts of any two switches close, the circulating water pumps are tripped, thereby terminating the leak. An auxiliary contact in this switch system also alarms the control room.

2.

A closed and monitored fire door separates the diesel generator (DG) corridor from the rest of the turbine building to minimize the amount of water that would enter the compartments.

3.

The turbine building sump high level alarm will annunciate in the control room.

4.

When the condensate pump pit fills with water, the resulting condensate pump trip will also alert the control room of a flood condition.

Therefore, the risk of the CW system and, for that matter, any other flooding sources causing flooding above the 85 ft. elevation of the turbine building is judged to be insignificant. For these same reasons, a flood in one unit is not expected to adversely affect the other unit.

PG&E Letter DCL-10-106 Page 15 of 131 RAI l.n

1. Provide the following information regarding the Level 1 Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

n. Section F. 7.2, in presenting the results of the uncertainty analysis for PRA model DCOO, states that the internal events CDF of 1.4 1E-05 per year is a point estimate value. The table of results in Section F. 2.1.6 labels this value as the mean CDF. For all of the CDF values reported in Section F. 2.1, confirm whether these are point estimates or mean values. Also confirm whether the evaluation of base case benefits is based on point estimate or mean values.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.n Section F.7.2 develops an uncertainty distribution for the DC01A model based on the, uncertainty distribution characteristics of the DCOO model' This section presents the DCOO model point estimate core damage frequency (CDF) of 1.41 E-05/yr. Assuming the same distribution characteristics of the DCOO, the table in Section F.7.2 presents the derived mean value of the DC01A model.

The CDF values for the models presented in Section F.2.1 are all point estimate values.

The evaluation of base case benefits was based on point estimate values.

PG&E Letter DCL-10-106 Page 16 of 131 RAI 1.o

1. Provide the following information regarding the Level I Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis in Attachment F to the Environmental Report (ER):

o. Section F.2.1 describes the PRA model changes made since model DCPRA-1998.

For PRA model versions 1995 and later, identify the model changes listed in Sectiom.

E.2. 1 that most impacted the change in CDF.

PG&E Response to Severe Accident Mitigation Alternatives RAI 1.o The following are the Diablo Canyon Power Plant models from 1995 forward and the model changes that had the most impact.

DCPRA-1995

1.

The probability distributions of reactor coolant pump (RCP) seal leakage leading to core uncovery as a function of time, used in the electric power recovery model, (top event RE) were replaced with new distributions, which are based on calculations performed for the qualified O-ring material. Additionally, the electric power recovery model was revised to always select the distributions for core uncovery time (from RCP seal loss-of-coolant-accidents (LOCAs)) for scenarios with no depressurization or cooldown.

2.

The auxiliary saltwater (ASW) system model was modified to: (1) create a new split fraction, ASG, for loss offsite power and all support available, (2) remove demusseling from a number of alignments, (3) use the unavailability variable ZMVU2F/D for the unit-to-unit cross-tie valve (this also effected Top Event Al),

and (4) reflect the train separation of the ASC split fraction. A review of the quantification indicated that split fractions AS4 and AS7 were not being properly selected, so the event tree split fraction rules were modified accordingly.

3.

Addition of the two Backup Battery Chargers 121 and 131 in the model.

DCPRA-1997

1.

Similar electric power recovery factors were added to general transients with offsite power failing independently, as is applied to LOOP (LOOP) initiating events.

2.

For the auxiliary feedwater (AFW) system model, the raw water storage reservoir was added as a backup source of water to the condensate storage tank.

PG&E Letter DCL-10-106 Page 17 of 131

3.

The recovery rules applied when the dedicated fuel oil transfer pumps fail (top event FO fails) were revised to allow recovery of some sequences that are recoverable.

DCOO

1.

Success criteria of the ASW system were changed to be consistent with thermal-hydraulic basis from the "Station Blackout Submittal" and generic letters on service cooling water systems. Several other modeling changes were made, including adding demusseling valves and flow path, and alignment changes to be consistent with current operational practice.

2.

Reactor coolant system (RCS) pressure relief system. Added the third power operated relief valve (PORV) (474) in Top Event PR and included a new Top Event (PRX) in the electric power support system event tree ELECPWR for questioning RCS pressure relief for a specified set of initiators.

3.

LOCA initiating event frequencies were reduced primarily due to the new initiating event frequencies from NUREG/CR-5750.

DCCO

1.

Anticipated transient without scram mitigating system actuation (AMSAC). This system is now credited to actuate the AFW system and trip the main turbine. It also provides a redundant AFW pump start signal when solid state protection system (SSPS) fails. The system model (Top Events AMA and AMB) developed was incorporated into the mechanical support systems event tree MECHSP.

The other event tree models were impacted by the implementation of the AMSAC system: general transient, steam generator tube rupture (SGTR),

anticipated transient without scam, and the interfacing system LOCA event tree model.

2.

The steamline break initiators (steam line break inside contiament (SLBI) and steam line break outside containment (SLBO)) now credit manual SSPS actuation.

3.

The ability to backfeed from the 500kV switchyard and cross-tie the vital buses in accordance with the emergency operating procedures was fully implemented.

DC01

1.

A higher consequential LOOP probability has been implemented.

2.

Modeling of the requirement to depressurize before crediting continuous makeup to the refueling water storage tank (RWST) after loss sump recirculation mode of operation during Small LOCA (SLOCA), and transient induced LOCA scenarios.

PG&E Letter DCL-1 0-106 Page 18 of 131

3.

Longer duration assumed for the emergency diesel generator (EDG) maintenance windows as part of the EDG license amendment request submittal.

DC01A

1.

Revised the charging pump model because of the replacement of the positive displacement pump charging pump (CHG Pumps 1-3 and 2-3) with the centrifugal pumps. One major difference in design between these new pumps and the original centrifugal charging pumps is that they do not require the component cooling water (CCW) for cooling.

Added new top event, CH3PP, representing a~potential RCP seal injection flow path using the Charging Pump 1-3(2-3) regardless the status of the CCW system. This pump is normally in service and remains in operation during a transient, except when manually tripped or upon "a transfer to diesel load shed" signal. It includes the suction path from either the volume control tank or RWST, the charging pump, and the discharge path to the RCP seal.

(I PG&E Letter DCL-1 0-106 Page 19 of 131 RAI 2.a

2. Provide the following information relative to the Level 2 analysis:

a. The discussion in Section F. 2.2.1 through F. 2.2.4, and associated figures and tables, appears to have been taken directly from the DCPP IPE submittal. Discuss the evolution of the current Level 2 model from that of the IPE including changes to the Level 2 analysis to reflect plant design and operation changes, incorporation of severe accident management guidelines, improvement in accident progression knowledge since the IPE (including consequential SGTR events) and changes to the Level 2 and/or Large Early Release Frequency (LERF) model alluded to in Sections F. 2.1.6 and F.2.1.8.

PG&E Response to Severe Accident Mitigation Alternatives RAI 2.a The foundation for the Diablo Canyon Power Plant Level 2 model is the work done for the individual plant examination (IPE), or Model DCPRA-1 991. It was in this model that the containment event tree was developed, containment probabilities assigned, and endstate binning logic built to analyze the core damage effects and assign Level 2 bins.

There has been no major overhaul of that logic since then, except for the significant changes to large early release frequency (LERF) made in the DC01 model. Details of the historical changes are as follows:

The Level 2 analysis was first considered during the development of the individual plant examination external events (IPEEE), which is the DCPRA-1993 model,. The external events impact on containment performance was assessed, which included the evaluation of the containment structure, penetrations, hatches, and isolation valves, and the containment heat removal capability. These structures, system, and components have high seismic capabilities, a focus of the IPEEE. Containment performance for fire initiators was also conservatively evaluated and it was determined that sequences are similar to those of the internal events. The conclusion was that external events do not pose any unique threat to containment performance, and it is not significantly different than that identified in the IPE. Therefore the Level 2 model structure was sufficiently flexible and robust.

In Model DCOO, referred to as the combined model, seismic and fire initiators are quantified with the internal event initiators for the first time. Combined core damage frequency (CDF) and LERF results from internal events, fire and, seismic initiators can now be developed. The capability to analyze the entire spectrum of Level 2 endstates was set aside in,.exchange for a simplified LERF result in this or previous revisions of the model.

Model DC01 restored the entire spectrum of Level 2 endstates to the model. Most importantly, the Level 2 model was also reviewed and updated with respect to changes PG&E Letter DCL-10-106 Page 20 of 131 in the state of the art since the early 1990s. The update focused on LERF issues and as a result there were significant changes in containment event tree split fractions that were implemented in the DC01 model. The simplified LERF model, used to quantify seismic LERF was not modified. It is the same as in the previous DCCO probabilistic risk analysis model. The capability to analyze fire LERF was lost upon restoration of the detailed Level 2 model. The model changes necessary to resolve fire LERF were postponed to a future model revision.

There were no Level 2 changes made in the DC01A model, which was developed to support this SAMA submittal.

PG&E Letter DCL-1 0-106 Page 21 of 131 RAI 2.b

2. Provide the following information relative to the Level 2 analysis:

b. The release category frequencies given in Table F. 2-8 appear to be those for the IPE model. Provide updated values for the PRA model DCO IA.

PG&E Response to Severe Accident Mitigation Alternatives RAI 2.b The release category frequencies from the current Model DCO1A are as follows:

Release Category Frequency RC01 2.30E-09 RC01U 5.20E-10 RC02 1.09E-08 RC02U 2.58E-10 RC03 5.02E-10 RC03U 7.99E-09 RC04 1.02E-08 RC04U 3.91 E-08 RC05 6.06E-10 RC05U O.OOE+00 RC06 4.39E-07 RC06U 1.05E-07 RC07 1.10E-10 RC07U 0.00E+00 RC08 4.74E-09 RC08U 3.72E-07 RC09 O.OOE+00 RC09U O.OOE+00 RClO 7.18E-07 RC10U 1.15E-07 RC11 O.OOE+00 RC11U O.OOE+00 RC12 1.70E-08 RC12U 7.29E-07 RC13 6.24E-12 RC13U 1.28E-12 RC14 8.55E-07 RC14U 2.04E-07 RC15 5.84E-13 RC15U 2.34E-11 RC16 5.07E-09 RC16U 5.98E-07 RC17 1.34E-06 RC18 1.82E-07 RC19 2.15E-07 RC20 1.OOE-06 RC21 1.47E-06 PG&E Letter DCL-10-106 Page 22 of 131 RAI 2.c

2. Provide the following information relative to the Level 2 analysis:

c. Provide a discussion of the process used to map the 37 release categories identified in Table F.2-5 to the six source term categories used in the SAMA analysis (e.g.,

Tables F.3-5 through F.3-7). Identify the release categories included in each source term category.

PG&E Response to Severe Accident Mitigation Alternatives RAI 2.c The following paragraphs and table are an extraction from Section 4.8 of the individual plant examination submittal, which discusses consolidation of the 37 source term release categories into 5 significant release category groups. The same steps were taken in the SAMA analysis except for the creation of a sixth category for interfacing system loss of coolant accident (ISLOCA). This sixth category is comprised solely of release category RC18, which means now containment bypass release category group only contains RC17. The identifiers used in the SAMA analysis that correspond to the categories below are as follows:

ST1 - Large, Early Containment Failures ST2 - Small, Early Containment Failures ST3 - Late Containment Failures ST4 - Containment Bypass ST5 - ISLOCA ST6 - Containment Intact "In Section 4.7.1, 37 source term release categories are defined. Quantification of the CET resulted in only 28 of the release categorieshaving significant frequency. In order to gain insights, these release categories were grouped into five release category groups, as shown in the table below. These five groups are discussed in the following sections.

All of the sequences in these release category groups, with the exception of the long-term intact containment release group, involve the failure to arrest core damage before Vessel breach (Top Event AD). As noted in Section 4.5.2, results reported in Reference 4.8-2 for Zion take substantial credit for arresting core damage for a degraded core cooling sequence before vessel breach. However, such credit has not been taken in -this study since it was conservatively assumed that once core damage occurs, no operator action could prevent vessel breach.

PG&E Letter DCL-10-106 Page 23 of 131 As discussed in Section 4.8.2.5, some of the long-term containment intact release group sequences (52 percent of group frequency) are predicted in the CET to be long term overpressurization (greater than 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after event initiation) sequences, which would normally be assigned to the late containment failure release group. It was judged that placing these sequences in the long-term intact group provides a more realistic assessment of containment performance, because of the likelihood of recovering containment heat removal systems using existing procedures."

Release Category Group Definition and Results Description of Release Release categories in Frequency Category Group Group?')

(per year)

Percent Small, Early containment RC13. RC13U. RC14, Failures RC14U, RC15, RC15U, RC16, RC16U 7.61e-06 8.7 Large, Early Containment RCO1, RCO1 U, RC02, Failures RC02U, RC03, RC03U, RC04, RC04U 2.45E-06 2.9 Late Containment Failures RC05 RC05U, RC06 RC06U, RC07, RC07U, RC08, RC08U, RC09, RC09U, RC10, RC10U, RC11, RC11U, RC12, RC12U, RC21 3.97E-05 45.2 Containment Bypass RC17, RC18 1.62E-06 1.8 Long-Term Containment RC19, RC20 I ntact"2I 3.64E-05 41.4 Notes: 1. The release categories are defined In' Table 4.7-1.

2.

Includes 52 percent of the long-term overpressurizatlon sequence frequency (greater than 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after event initiation) (see Section 4.8.2.5).

Frequencies are presented in exponential notation (i.e., le-1 = 1 X 101).

PG&E Letter DCL-10-106 Page 24 of 131 RAI 2.d

2. Provide the following information relative to the Level 2 analysis:

d. From Tables F. 3-5 and F. 3-6 it appears that only the results from six specific Modular Accident Analysis Program (MAAP) cases were used in the source term analysis.

Confirm whether and how results from the parametric code ZISOR were used-in the SAMA analysis. Provide the rationale for using selected MAAP results rather than ZISOR results for the source term release fractions when, in some release category cases, the ZISOR results are higher.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 2.d Results from ZISOR were not utilized for the consequence analysis performed as part of SAMA. Modular Accident Analysis Program (MAAP) results were used to provide the release information for the MACCS2 analysis. The MAAP calculations represent true best estimate calculations specific to Diablo Canyon Power Plant. The representative cases were selected based on their contribution to the total reactor coolant (RC) frequency. For example, for the small early release category, Case RC14 was selected since it contributed 51 percent to the total RC frequency. For large early, Case RC04U represented 54 percent of the total frequency. For containment bypass, two scenarios were selected based on the Level 2 probabilistic risk analysis results. First was a steam generator tube rupture scenario with auxiliary feedwater available. The second bypass event was selected as the residual heat removal pipe break. These individual representative cases are summarized in Table F.3-5.

PG&E Letter DCL-1 0-106 Page 25 of 131 RAI 2.e

2. Provide the following information relative to the Level 2 analysis:

e. Provide more information on the selection of the MAAP case for each source term category, in particular how scenarios of less than dominant frequency but larger potential consequences were considered. Also confirm the version of MAAP used for the source term analysis.

PG&E Response to Severe Accident Mitigation Alternatives RAI 2.e Modular Accident Analysis Program pressurized water reactor Version 4.0.7 was utilized for the source term analysis. The following provides additional discussion of the representative cases.

Large Early - The case selected contributed 54 percent to the total reactor coolant (RC) frequency. The scenario included a loss of all injection, loss of all feedwater, loss of containment sprays and with cooldown starting at 15 minutes into the event. The containment was assumed to be failed with a large area (7 ft2) at the time of vessel breach. This scenario tends to represent the largest potential consequences, since other variations involve operation of containment sprays with and without successful debris coolability.

Small Early - The selected case represents over 50 percent of the total RC frequency.

The case included loss of all injection with a pre-existing containment breach and without containment sprays. Other cases within this release category would either have medium to low RCS pressure and could include water covering the debris outside the vessel. The case selected would tend to represent the highest consequence conditions.

Late Failure - This representative case included a seal loss-of-coolant-accident (LOCA) with successful'operation of auxiliary feedwater (AFW). Core damage occurs at 3.8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> followed by vessel breach at 6.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />. Containment failure occurs at about 38 hours4.398148e-4 days <br />0.0106 hours <br />6.283069e-5 weeks <br />1.4459e-5 months <br /> due to prolonged core-concrete interaction and corresponding pressurization.

This case was judged to be a good representation of a late release. The general characteristics of a late release include substantial time available for passive removal of fission products within the containment and therefore a significantly reduced source term.

Intact - This representative case involves core damage and vessel breach, but without contain~ment breach due to operation of the containment sprays. The source term calculated is limited to that associated with normal design leakage.

PG&E Letter DCL-10-106 Page 26 of 131 Bypass - The bypass results indicated that RC17 represented 89 percent of the total release category frequency. However, because the dominant scenario was a steam generator tube rupture (SGTR) with successful operation of AFW, it was decided to include two different types of bypass events. The higher frequency case, the SGTR with AFW was selected as the first bypass event. The lower frequency case, an interfacing system (ISLOCA) was selected due to expected higher consequences. This is represented as ST5 in Table F.3-5. There was a residual scenario that involved an SGTR without successful AFW and this frequency was added to the ISLOCA frequency.

PG&E Letter DCL-10-106 Page 27 of 131 RAI 2.f

2. Provide the following information relative to the Level 2 analysis:

f, The Westinghouse Level 2 PRA review described in Section F.2.3.2 is-stated to be in support of Risk Informed Technical Specification Test. Frequency (RITSTF) Initiative 5-b and is cited in Reference 56 as an "ASME PRA Peer Review" for LERF. Provide more information on this review including the applicability to the SAMA application, the composition of review team, the criteria used, and the resolution/disposition of review findings for the SAMA application (if not addressed in Addendum 1 to the ER).

PG&E Response to Severe Accident MitiQation Alternatives RAI 2.f Late in 2007, Westinghouse provided a review of the Diablo Canyon Power Plant (DCPP) probabilistic risk a probabilistic risk assessment (PRA) large early release frequency (LERF) model in support of the Risk Informed Technical Specification Test Frequency (RITSTF) Initiative 5-b.

The LERF model was reviewed to determine if there were any modeling limitations that would affect the RITSTF Initiative-5b (I-5b) application. The documentation reviewed for the assessment included the DCPP individual plant examination, DCPP containment fragility assessment, regeneration of the Level 2 Model (part of DCOO model revision),

updated Level 2 split fraction assessments (also part of DC0O model revision), DCPP temperature induced steam generator tube rupture (TI-SGTR) performance assessments, and the most recent LERF quantification results (C.9 Revision 10, June 2006). Results of past peer reviews were also considered. The review was taken with consideration to the high level requirements and supporting requirements in ASME Standard RA-Sb-2005, "Standard for Probabilistic Risk Assessment for Nuclear Power Plant Applications."

The focus of the review was to: (1) assess the status of existing peer review comments, (2) identify areas of excess conservatism, and (3) identify open issues and existing model limitations that may affect risk assessment of components involved in an I-5b Surveillance Test Interval Extension Program. Of particular interest to the LERF model for the I-5b application is the containment isolation valve model. The objective of the I-5b review was to disposition the approach for the expected application, and to recommend an alternative modeling practice.

Findings from the review were placed in the PRA action tracking database. Those findings that were categorized as requiring resolution (with a Status of "A") are listed in Addendum 1. See items 808, 811, and 812.

PG&E Letter DCL-10-106 Page 28 of 131 The review is relevant to SAMA because it provided independent verification of the validity of the DCPP LERF model by industry experts against the ASME PRA standard.

And there are important plant damage endstates that comprise LERF that contribute to the SAMA analysis.

PG&E Letter DCL-10-106 Page 29 of 131 RAI 2.,q

2. Provide the following information relative to the Level 2 analysis:

g. Discuss the binning of plant damage state (PDS) INNNB into key plant damage state (KPDS) INNGB. This appears to violate the fifth binning guideline given in Section F.2.2.3.

PG&E Response to Severe Accident Mitigation Alternatives RAI 2.q The binning of plant damage state INNNB into key plant damage state (KPDS) INNGB took place for several reasons. First the frequency of INNGB is higher than INNNB so INNGB becomes a KPDS, not INNNB. Both plant damage states stem from core damage sequences from steam generator (SG) tube rupture with unaffected SG cooling and some high pressure injection. Their characteristics include medium reactor coolant system (RCS) pressure at vessel breach (250-600 psia) because of some depressurization from SG cooling, but core damage occurs because of failure to inject the refueling water storage tank to balance the RCS inventory losses. For both plant damage states (PDSs) there is a failure of containment spray or containment spray recirculation. In the case of PDS INNGB there is some containment heat removal. The second reason for binning INNNB into INNGB is that both PDSs are characterized by unisolated SGs. Hence, their Level 2 (large early release frequency) characteristics are the same.

PG&E Letter DCL-1 0-106 Page 30 of 131 RAI 2.h

2. Provide the following information relative to the Level 2 analysis:

h. Discuss the extent to which recovery of systems or operator actions following the onset of core damage are credited in the Level 2 assessment, and how the recovery is modeled.

PG&E Response to Severe Accident Mitigqation Alternatives'RAI 2.h There is no recovery with human actions or systems post-core damage. The Diablo Canyon Power Plant containment event tree structure does not determine if systems have been or can be recovered or additional operator actions can be been taken post-core damage.

What may be considered a recovery, because of its direct influence~on large early release frequency, is the operator action to isolate containment when it fails to isolate automatically (see top event 01). This action is credited only for station blackout (SBO) or seismically induced SBO.

I PG&E Letter DCL-1 0-106 Page 31 of 131 RAI 2.i

2. Provide the following information relative to the Level 2 analysis:

i. The DCPP IPE indicates that isolation failure contributes about 80 percent of the small-early release category frequency. The SAMA analysis indicates that this release category accounts for over one third of the offsite consequences. Discuss the importance of this containment failure mode and potential SAMAs to reduce this contribution to offsite consequences.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 2.i The Diablo Canyon Power Plant model includes events for failure to perform a manual containment isolation on failure of the automatic function; however, those actions were below the review threshold for the internal events Level 2 importance list review. For fire and seismic initiators, the Level 2 importance information is not available.

As noted in this request for additional information, containment isolation failure is a significant contributor to the small-early release category, but those contributions are typically related to scenarios in which the support systems for the containment isolation function have failed. The more important support system failures, such as those related to power, would be included on the importance list and the SAMA identification process would address those failures (e.g., via SAMA 5).

PG&E Letter DCL-1 0-106 Page 32 of 131 RAI 2.1

2. Provide the following information relative to the Level 2 analysis:

j. SectionF. 2.2.4.5. 1 discusses the calculation time cutoff at 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br /> and indicates that simulations run for longer times "did not show huge increases" in the release fractions.

Provide a brief but more quantitative discussion of the potential impact to dose-risk and offsite economic cost-risk (OECR) if the simulation time were extended beyond 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br />.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 2.J The fission product release plots for each of the representative Modular Accident Analysis Program cases have been reviewed to evaluate the potential impact of accounting for any fission product releases occurring after the 50-hour mark. In almost all cases, the individual fission product release shows a stable condition at the end of the simulation. The small early release for noble gas (Case RC14 for ST2) does show the potential for continued release beyond the simulation time. This is also the case for the intact representative simulation as well (Case RC20 for ST6). In both of these simulations, the release from containment is through a small opening and has the potential to continue past the end time of 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br /> that was used for the Environmental Report. This seems to be particularly true for the noble gas release, as other release fractions are either stabilized at the end or are of very low magnitude due to natural removal in the containment. It is believed that the current set of sensitivities presented in the SAMA evaluation are adequate to bound the potential impact that a continued noble gas release might have on the offsite consequences. It should also be noted that potential onsite and offsite mitigation measures, beyond the simulation time assumed, could be employed late in the event to reduce the possible consequences of these accidents.

PG&E Letter DCL-10-106 Page 33 of 131 RAI 3.a

3. Provide the following information with regard to the treatment and inclusion of external events in the SAMA analysis:

a. Section F.4.6.2 states that some changes have been made to the external event models, but the models have not been updated to reflect recent plant changes or the full spectrum of current PRA techniques. Provide a discussion of: (1) the model changes that have been made, and (2) the plant design changes that have not been incorporated in the seismic and fire models and their potential impact on the results of the SAMA assessment. In addition, describe the most significant conservatisms and non-conservatisms in the seismic and fire models and their impact on the SAMA analysis.

PG&E Response to Severe Accident Mitigation Alternatives RAI 3.a As part of the process of preparing an updated fire probabilistic risk assessment (PRA) model, PG&E reviewed plant-design and operational changes from 2004 to March 31, 2009. This review was conducted pursuant to existing plant procedures that implement relevant ASME requirements related to the PRA model. The review included all design and operational changes (i.e., the review was not limited to changes related to fire protection; it also included changes that could impact the seismic model). Based on that review, PG&E concluded that none of the design or operational changes will have an effect on the Diablo Canyon Power Plant (DCPP) model or the SAMA evaluation.

The fire model used for the SAMA evaluation has benefited from changes and improvements since the Individual Plant Examination of External Events (IPEEE).

Some of the fire initiators use the internal events model (event trees and top events) for quantification, and the internal events model has undergone several significant revisions since the IPEEE in 1993 as discussed in Section F.2.1. The fire model-was requantified in the DCOO model revision with lower core damage frequency (CDF) contribution due to the internal events model revision. In Model DCCO, fire initiating event FS5 was recalculated and along with internal events model changes a lower revised fire CDF resulted. With Model DC01 the requirement to depressurize the RCS for nonisolable steam generator tube ruptures (SGTRs) caused the fire CDF to increase. It was in this model revision that large early release frequency (LERF) due to fire was not calculated due to the replacement of the simplified LERF model with a detailed Level 2 model. Finally, in Model DCOlA the fire CDF was reduced due to the impact of the revised charging pump model.

The seismic model has also benefited from internal events model revisions since the IPEEE. The seismic analysis was also requantified in the DCOO model revision. In the DCCO model revision, the change to the seismic analysis was the incorporation of the use of the safety injection pumps (and depressurization) for very small loss of coolant PG&E Letter DCL-10-106 Page 34 of 131 accident (VSLOCA) events. The changes to the internal events model resulted in a lower revised seismic CDF. With Model DC01, the requirement to depressurize the reactor coolant system (RCS) for nonisolable SGTRs increased the seismic CDF.

Seismic LERF continued to be quantified using the simplified LERF model. Finally, in Model DC01A there was no change in the seismic CDF. The changes associated with Model DC01 B are discussed in response to request for additional information 3.c.

Other changes taking place after the IPEEE are discussed in Section F.5.1.5 and F.5.1.6 of the Environmental Report.

The external events hazards modeling has not changed since the IPEEE. The conservatisms and nonconservatisms associated with those models are summarized in Tables 3.a-1 and 3.a-2. There are no quantitative means of determining how the issues described in those tables would impact the SAMA analysis. In general, it is believed that the external events results are conservative; regardless, the 95th percentile PRA results sensitivity analysis is considered adequate to account for any net nonconservatism in the external events models.

Table 3.a-1: Summary of Conservatisms and Nonconservatisms in the Fire Model PRA Topic Generic Comment DCPP Specific Example(s)

Initiating Events:

The frequency of fires and their severity Potentially Conservative: The initiating are generally conservatively events frequencies for DCPP are mostly overestimated. A revised NRC fire based on the EPRI Fire Events Database events database indicates the trend and some newer sources indicate the toward lower frequency and less severe EPRI data is conservative.

fires. This trend reflects the improved housekeeping, reduction in transient fire hazards, and other improved fire protection (FP) steps at plants.

System Response:

FP measures such as sprinklers, C0 2, Potentially Conservative: Credit for fire and fire brigades may be given minimal detection and suppression was not taken (conservative) credit in their ability to limit in the model with the exception that main the spread of a fire.

control room fire scenarios credit it implicitly in the geometry and severity factors.

Sequences:

Sequences may subsume a number of Potentially Nonconservative: Fire fire scenarios to reduce the analytic hazards (areas) are screened from burden. The subsuming of initiators and analysis on low probability and are sequences is done to envelope those not included in the total CDF.

sequences included. This results in additional conservatism.

° Potentially Nonconservative: Fire scenarios that contribute less than 10 percent of the total system failure were screened from further review.

PG&E Letter DCL-10-106 Page 35 of 131 Table 3.a-1: Summary of Conservatisms and Nonconservatisms in the Fire Model PRA Topic Generic Comment DCPP Specific Example(s)

Fire Modeling:

HRA:

Fire damage and fire spread are conservatively characterized. Fire modeling presents bounding approaches regarding the immediate effects of a fire (e.g., all cables in a tray are always failed for a cable tray fire) and fire propagation.

There is little industry experience with crew actions under conditions of the types of fires modeled in fire PRAs. This has led to conservative characterization of crew actions in fire PRAs. Because the CDF is strongly correlated with crew

,actions, this conservatism has a profound effect on the calculated fire PRA results.

0 Electric power recovery is not assumed for any fire initiators.

Feedwater and condensate are assumed to not be available.

Neutral: Evaluations of manual operator actions to respond to fire induced failures were examined and credited.

Unknown-: Some operator actions, such as reducing steam generator pressure using the head vent, are not credited atall while others use their nominal values. Precluding credit for the actions may be conservative while using full credit for other actions may be nonconservative. However, treatment of human reliability analysis (HRA) in the current guidance is evolving and while it tends not to credit actions without detailed analysis, it does not imply that the credit taken in the IPEEE is incorrect.

Neutral: The fire analysis is integrated with the PRA system model and any limitations related to the level of detail would be characterized in the "Fire Modeling" and "Sequence" topics.

It is not clear whether the following factors would result in an increase or decrease in the fire CDF given that some changes would likely increase the CDF and others would decrease it:

No fire PRA review comparable to an internal events model peer review was performed on the DCPP fire analysis.

Level of Detail:

Quality of Model:

The fire PRAs may have reduced level of detail in the mitigation of the initiating event and consequential system damage.

The peer review process for fire PRAs is not as developed as internal events PRAs. For example, no industry standard, such as NEI 00-02, exists for the structured peer review of a fire PRA.

This may lead to less assurance of the realism of the model.

The fire modeling has not been updated since the IPEEE (apart from internal events model updates that also impact the fire frequency).

PG&E Letter DCL-10-106 Page 36 of 131 Table 3.a-2: Summary of Conservatisms and Non-Conservatisms in the Seismic Model PRA Topic DCPP Comment Initiating Events:

System Response:

Sequences:

Seismic Modeling:

HRA:

Neutral: The initiating events frequencies for DCPP have been updated to account for the most recent data and to account for the Shoreline Fault.

Potentially Conservative: For seismic events greater than 1.75g, no recovery is assumed.

Neutral: No issues identified. Model structure and plant damage state binning is same as internal events model.

Potentially Conservative: For structure failures, total collapse was assumed to occur at the point where significant strength degradation is estimated to occur. All equipment in the structure was assumed to fail on collapse.

Potentially Conservative: All seismic events are considered to at least cause a VSLOCA and require RCS makeup for success.

Neutral: Both plant specific and generic component fragilities were used.

Unknown: The operator actions originally associated with the internal events model were multiplied by seismic stress factors to account for the impact of the seismic event. HRA guidance is evolving and different approaches may be preferred in a current analysis; however, this does not imply that the credit taken in the IPEEE is incorrect.

Neutral: The seismic analysis is integrated with the PRA system model and any limitations related to the level of detail would be characterized in the "Seismic Modeling" and "Sequence" topics.

It is not clear whether the following factors would result in an increase or decrease in the seismic CDF given that some changes would likely increase the CDF and others would decrease it:

No seismic PRA review comparable to an internal events model peer review was performed on the DCPP seismic analysis.

Apart from the seismic initiating events update performed for the DC01 B model, the seismic model itself has not been updated since the IPEEE. The internal events model updates have impacted the seismic frequency, but those changes are not related to the seismic response model.

Level of Detail:

Quality of Model:

PG&E Letter DCL-10-106 Page 37 of 131 RAI 3.b

3. Provide the following information with regard to the treatment and inclusion of external events in the SAMA analysis:

b. Section 4.6.2 states that "the method chosen to account for external events contributions in the SAMA analysis is to use a multiplier on the internal events results."

However, the Phase II SAMA evaluations do, in fact, use the IPEEE fire and seismic PRA models rather than the external events multiplier (see also RAI 6.e). Discuss this apparent discrepancy. Justify the use of the fire and seismic PRA models for the SAMA analysis given: (1) the number of strong motion earthquakes worldwide since the Long Term Seismic Program (LTSP) seismic risk assessment that have provided the opportunity for reassessing and updating the methodologies associated with evaluating the impact of ground accelerations on structures and equipment, (2) the significant evolution in fire PRA methodology since the IPEEE as set forth in the fire PRA standard (ANSI/ANS-58.23-2007) and the joint NRC/EPRI fire PRA development guidance (NUREG/CR-6850), and (3) the publication of the external events PRA standard (ANSI/ANS-58.21-2007).

PG&E Response to Severe Accident Mitigation Alternatives RAI 3.b The Diablo Canyon Power Plant (DCPP) quantification method is a hybrid process that employs both a multiplier and the available external events probabilistic risk assessment (PRA) results (fire and seismic).

The baseline maximum averted cost-risk (MACR) is calculated by applying a multiplier to the internal events cost-risk. The multiplier is based on the ratio of the total core damage frequency (CDF) (internal events plus external events) to the internal events CDF. This is described in Section F.4.6.2 of the Environmental Report (ER).

The averted cost-risk for a SAMA is obtained in a multi-step process that includes the use-of the baseline MACR as well as the internal events results, a multiplier for nonfire, nonseismic external events, and the external events PRA results.

The averted cost-risk is the difference between the baseline MACR and the MACR for the configuration in which the SAMA has been implemented (MACRSAMA).

The internal events portion of the MACRSAMA is calculated in the same.

manner as for the baseline MACR using the CDF, Level 2 PRA results, etc., as shown in Sections F.4.1 through F.4.6.1 of the ER.

PG&E Letter DCL-1 0-106 Page 38 of 131 The contribution from the nonseismic, nonfire external events to the MACRSAMA is accounted for by multiplying the internal events MACRSAMA by the ratio of the sum of the baseline internal events CDF and the baseline nonseismic, nonfire external events CDF to the baseline internal events CDF (8.47E-06 + 2.56E-06) / 8.47E-06 = 1.3).

The fire contribution to the MACRSAMA is obtained by multiplying the base fire contribution to the MACR by the ratio of the fire CDF with the SAMA implemented to the base fire CDF.

The base fire contribution to the MACR is developed by multiplying the base external events contribution (Section F.4.6.3) to the MACR by the ratio of the fire CDF to the total external events CDF

($6,400,000

• 1.39E-05 / 5.416E-05 = $1,642,500)

The seismic contribution to the MACRSAMA is obtained by multiplying the base seismic contribution to the MACR by the ratio of the seismic CDF with the SAMA implemented to the base seismic CDF.,

The base seismic contribution to the MACR is developed by multiplying the base external events contribution (Section F.4.6.3) to the MACR by the ratio of the seismic CDF to the total external events CDF ($6,400,000

• 3.77E-05 / 5.416E-05 = $4,455,000)

The validity of the external events model results is an industry wide issue that is not specific to DCPP. The NRC has previously accepted SAMA analyses in license renewal applications that have relied on the results of the individual plant examination of external events risk models when more current external events PRA models are not available. In addition, SAMA is not a risk-informed Regulatory Guide 1.200 process.

For reasons unrelated to this request for additional information (RAI), DCPP has committed resources to updating the seismic and fire PRA models. The results of the updated seismic model are available and the impact on the DCPP SAMA analysis is documented in the response to RAI 3.c. The fire model update is not complete; no insights are available.

PG&E Letter DCL-10-106 Page 39 of 131 RAI 3.c

3. Provide the following information with regard to the treatment and inclusion of external events in the SAMA analysis:

c. The seismic hazard analysis used to determine the seismic CDF is that developed for the LTSP in 1988. Since then there has been a continued investigation of seismicity in the DCPP region. The U.S. Geological Survey, (USGS) has updated its assessment of seismic hazards across the US including California. In addition a new fault, the Shoreline Fault, has been identified. This new fault could impact the seismic non-exceedance frequency used for the seismic analysis. Provide a justification for the continued use of the 1988 seismic hazard curves for the seismic analysis used in the SAMA assessment. Address whether consideration of the Shoreline Fault could result in identification of additional candidate SAMAs specific to failures introduced by the fault, or an increase in SAMA benefits sufficient to change a SAMA disposition from not-cost-beneficial to cost-beneficial.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 3.c The probabilistic seismic hazard analysis (PSHA) for Diablo Canyon Power Plant (DCPP) has been updated from the results in the 1988 Long Term Seismic Program (LTSP) (PG&E, 1988) using current information on the ground motion models and the seismic sources. The most significant revision is to the ground motion model.

Preliminary revisions to the source characterization based on new data and evaluations are also included. The changes to the source characterization and ground motion models are summarized below.

Since the development of this RAI response, the probabilistic seismic hazard has been further refined. This discussion, up through "Probabilistic Risk Assessment" (PRA),

describes the most current seismic hazard analysis. The discussion after "Probabilistic Risk Assessment" and the updated SAMA analysis uses slightly older and slightly different data as shown in the table below:

Seismic Initiating Event Frequency Update Summai Initiating Event DC01A DC01 B Most Current Frequency Frequency Frequency SEIS1 (g levels 2.02E-1 to 1.25) 1.72E-02 7.40E-03

.1.16E-02 SEIS2 (g levels 1.25 to 1.75) 8.69E-04 3.50E-04 3.53E-04 SEIS3 (g levels 1.75 to 2.00) 1.56E-04 7.OOE-05 7.32E-05 SEIS4 (g levels 2.00 to 2.50) 1.24E-04 6.70E-05 6.92E-05 SEIS5 (g levels 2.5 to 3.00) 2.94E-05 2.60E-05 2.64E-05 SEIS6 (g levels 3.00 to 3.99) 7.64E-06 1.74E-05 1.45E-05 New (g levels greater than 4.0)

Note 1 Note 1 2.40E-06 PG&E Letter DCL-10-106 Page 40 of 131 Note 1: The DC01A and DC01 B PRA models did not consider acceleration levels greater than 4.0 g based upon low frequency of occurrence. The most current seismic hazard data has a more significant contribution at acceleration levels greater than 4.0 g and is included here.

As can be seen in~the table, the most current frequencies of occurrence for SEIS1 through SEIS6 are bounded by the DC01A and DC0.1 B model frequencies.

Accelerations at greater than 4.0 g contribute only slightly to the total seismic core damage frequency. The total seismic core damage frequencies for the DC01A, DC01 B and most current seismic hazard are 3.77E-05, 3.15E-05 and 3.44E-05, respectively.

Because the seismic core damage frequency calculated in the DC01A and DC01B models bounds the seismic core damage frequency using the most current seismic hazard, the SAMA analyses using the DC01A and DC01 B frequency data would bound the SAMA analysis if it used the most current seismic frequency data.

Ground Motion Models There has been a significant increase in the number of strong motion recordings from large magnitude (M>7) earthquakes.at short distances (R<10 km) since the 1988 LTSP model was completed. Figure 1 compares the strong motion data set used by the 1988 LTSP and the additional data that were compiled as part of the Pacific Earthquake Engineering Research Center (PEER) Next Generation Attenuation (NGA) project. The NGA project developed five updated ground motion models: Abrahamson and Silva (2008), Boore and Atkinson (2008), Campbell and Bozorgnia (2008), Chiou and Youngs (2008), and Idriss (2008). These models represent the state of the art for ground motion models for earthquakes in California.

An important feature of the NGA data set is that the shear-wave velocity in the top 30 m (VS30) was estimated for each recording. Four of the five models use VS30 as the site parameter. This avoids the inconsistencies in the generic definitions of "rock" and "soil" that were an issue in the development of the LTSP models. For DCPP, the VS30 has been computed from measurements at the power block. The VS30 for DCPP is 1200 m/s.

Based on the increased near fault data, the NGA models give significantly lower values of the high-frequency spectral accelerations close to large earthquakes compared to the 1988 LTSP model. As an example, the median spectral acceleration for 3-8.5 Hz from a magnitude 7 strike-slip earthquake for the DCPP site conditions are compared in Figure 2.

A second important change in the ground motion model is the standard deviation. The NGA models have much larger standard deviations for large magnitude earthquakes than given in the 1988 LTSP model. As an example, the standard deviations for the PG&E Letter DCL-1 0-106 Page 41 of 131 LTSP and NGA models for 3-8.5 Hz are compared in Figure 3. The increase in the standard deviation will offset some of the effects of the reduction in the median for low probability hazard levels.

Source Characterization In the LTSP PSHA, the seismic hazard at DCPP was controlled by three faults: Hosgri, Los Osos, and San Luis Bay. Revisions to the source characterization for these faults based on new data collected since the LTSP was completed and are described below.

In addition, the Shoreline fault is added to the source characterization. The revisions described below are preliminary but represent our current understanding of the faulting in the DCPP region, including the new data from Hardebeck (2010).

Hosgri Fault In the LTSP model, the down-dip geometry and style-of-faulting of the Hosgri fault was uncertain. As a result, three alternative models for the style-of-faulting were used:

strike-slip with 70-90 degree dip (wt=0.65), oblique slip with 45 to 90 degree dip (wt=0.30), and reverse slip with 15 to 60 degree dip (wt=0.05). Recent seismicity data show that the micro seismicity follows a near vertical dip for depth to 12 km (McLaren and Savage, 2001). New geophysical data (magnetic, gravity, shallow seismic) also constrain the Hosgri fault to a near vertical dip. Therefore, the source characterization for the Hosgri fault was modified to remove the oblique slip and reverse slip branches.

The Hosgri fault is modeled as a vertical strike-slip fault in the vicinity of DCPP.

Los Osos Fault In the LTSP model, the down-dip geometry of Los Osos fault was uncertain. Four dip angle were considered: 30 degrees (wt=0.27), 45 degrees (wt=0.02), 60 degrees (wt=0.63), and 75 degrees (wt=0.08). Recent gravity data and geological evaluations suggest that the Los Osos fault has a steeper dip than previously estimated.

Furthermore, the geometry of the low dip (30 degrees) is not consistent with the vertical Shoreline fault extending to a 12 km depth. Therefore, the 30 degree dip for the Los Osos fault is removed and the dips are modeled as 45, 60, and 75 degrees with weights of 0.3, 0.4, and 0.3, respectively.

San Luis Bay Fault For this update, the characterization for the San Luis Bay fault is not changed from the model used in the 1988 LTSP.

Shoreline Fault The Shoreline fault is currently being studied and the characterization will be completed in December 2010. A preliminary model for the Shoreline fault is used in this analysis.

PG&E Letter DCL-10-106 Page 42 Of 131 In this study, the Shoreline fault is segmented into southern, central, and northern segments (Figure 4). Individual and multiple segment ruptures are considered. The source characterization models for the Shoreline fault segments are shown below (Table 1). All three segments are modeled as vertical strike-slip faults based on the seismicity pattern at depth and on the average focal mechanisms given by Hardebeck (2010).

Hazard Results The 1988 LTSP and updated hazard curves for spectral acceleration averaged over 3-8.5Hz are shown in Figure 5. The updated hazard curves are shown with and without the Shoreline fault. At spectral accelerations less than 2.3g, the updated hazard curves including the Shoreline fault are lower than the 1988 LTSP hazard curves. This reduction is due to the lower median ground motion and the revised source characterization of the Hosgri and Los Osos faults. At spectral accelerations greater than 2.5g, the updated hazard is greater than the 1988 hazard. This increase is due to the larger standard deviation of the NGA ground motion models.

Although additional seismic investigations to better characterize the Shoreline fault are ongoing, the results of the PSHA are not sensitive to the outcome of these studies because the hazard is dominated by the Hosgri and Los Osos faults. Even using conservative assumptions about the length of the Shoreline fault (assuming that the Shoreline fault extends to the Hosgri fault for a total length of 25 km) and using conservative assumptions about the slip rate (up to 0.3 mm/yr), the seismic hazard of the Shoreline fault has only a small effect on the total hazard at DCPP.

Table 2 shows the impact of the updated seismic hazard including the Shoreline fault and the 1988 LTSP hazard in terms of yearly core damage frequency (CDF). The total seismic core damage frequency considering the updated hazard with Shoreline is reduced from 3.80 E-05 to 3.44 E-05 per year.

PG&E Letter DCL-10-106 Page 43 of 131 Table 1. Source characterization models for the Shoreline fault segments.

FAULT T

FAULT SHORELINE FAULT TYPE DIP SLIP RATE WIDTH Mchar Southern SS (1.0) 90 (0.8) 0.01 (0.2) 12(1.0) 5.8 (0.2) 80 (0.2) 0.05 (0.3) 6.0 (0.6) 0.2 (0.3) 6.2 (0.2) 0.3 (0.2)

Central SS(1.0) 90(0.8) 0.01 (0.2) 12 (1.0) 5.8 (0.2) 80 (0.2) 0.05 (0.3) 6.0 (0.6) 0.2 (0.3) 6.2 (0.2) 0.3 (0.2)

Northern (Seismicity)

SS (1.0) 90(0.8) 0.01 (0.2) 12(1.0) 5.8 (0.2) 80 (0.2) 0.05 (0.3) 6.0 (0.6) 0.2 (0.3) 6.2 (0.2) 0.3 (0.2)

Northern (Geomorphology)

SS(1.0) 90(0.8) 0.01 (0.2) 12 (1.0) 6.0 (0.2) 80 (0.2) 0.05 (0.3) 6.2 (0.6) 0.2 (0.3) 6.4 (0.2) 0.3 (0.2) 6.05 Southern + Central SS(1.0) 90(0.8) 0.01 (0.2) 12(1.0)

(0.2) 6.25 80(0.2) 0.05 (0.3)

(0.6) 6.45 0.2 (0.3)

(0.2) 0.3 (0.2)

Southern + Central +

Northern (Seismicity)

SS (1.0) 90(0.8) 0.01 (0.2) 12 (1.0) 6.3 (0.2) 80 (0.2) 0.05 (0.3) 6.5 (0.6) 0.2 (0.3) 6.7 (0.2) 0.3 (0.2)

Southern + Central +

Northern (Geomorphology)

SS(1.0) 90(0.8) 0.01 (0.2) 12 (1.0) 6.3 (0.2) 80(0.2) 0.05 (0.3) 6.5,(0.6) 0.2 (0.3) 6,.7 (0.2) 0.3 (0.2)

PG&E Letter DCL-10-106 Page 44 of 131 Table 2. 1988 LTSP and Updated Hazard w/ Shoreline Contribution to Damage Frequency.

DCPP Core LTSP Hazard Curve Seismic CDF - CDF based on Seismic Model CCDPs Initiator IEFrequency CDF SEISI 0.2 to 1.25g 1.72E-02 8.85E-06 SEIS2 1.25 to 1.75g 8.69E-04 1.05E-06 SEIS3 1.75 to 2.Og 1.56E-04 3.16E-06 SEIS4 2.0 to 2.5g 1.24E-04 9.59E-06 SEIS5 2.5 to 3.Og 2.94E-05 9.82E-06 SEIS6 3.0 to 4.Og 7.64E-06 5.27E-06 Total 3.77E-05 Updated Hazard Curve With Shoreline Seismic CDF - CDF based on Seismic Model CCDPs Initiator IEFrequency CDF Delta CDF SEIS1 0.2 to 1.25g 1.16E-02 5.97E-06

-2.88E-06 SEIS2 1.25 to 1.75g 3.53E-04 4.27E-07

-6.25E-07 SEIS3 1.75 to 2.Og 7.32E-05 1.48E-06

-1.68E-06 SEIS4 2.0 to 2.5g 6.92E-05 5.35E-06

-4.24E-06 SEIS5 2.5 to 3.Og 2.64E-05 8.82E-06

-1.00E-06 SEIS6 3.0 to 4.Og 1.45E-05 1.OOE-05 4.73E-06

>4g

> 4g 2.40E-06 2.40E-06 2.40E-06 Note Total 3.44E-05

-3.30 E-06 Note: The DC01A PRA model uses the 1988 LTSP seismic hazard data and does not consider accelerations greater than 4.0 g based upon low frequency of occurrence.

The most current/seismic hazard data has a more significant contribution at acceleration levels greater than 4.0 g and is included here. To determine the CDF due to accelerations > 4.0g, the conditional core damage probability for any seismic event greater than 4.0 g was conservatively set to 1.0.

PG&E Letter DCL-10-106 Page 45 of 131 E)

Magnitude Figure 1. Comparison of the strong motion data sets used for the 1988 LTSP model and the additional data that were included in the 2008 NGA data set. The 2004 data refer to the additional data that were included in the NGA database and were considered in the development of the NGA ground motion models.

PG&E Letter DCL-1 0-106 Page 46 of 131 2\\

m LTSP - SS SAS08 mCBO CY08 In 0.1+-

1 10 50 Rupture Distance (kmn)

Figure 2. Comparison of the median spectral acceleration (3-8.5 Hz) from M7 strike-slip earthquakes for the LTSP (1988) model and the NGA models.

72 a

.0.7-0.6-0.4-0.3 '

• LS

-AS08 0.2-1 BA08

- CB08 0.1-CY08 Magnitude Figure 3. Comparison of the standard deviation for spectral acceleration for 3.0-8.5 Hz from the 1988 LTSP model and NGA models PG&E Letter DCL-1 0-106 Page 47 of 131 Shoreline (Northern Geomorphotogy)

Shoreline (Northern Seismicity)

Shoreline (Centralý,v Shoreline,(Southern)

Figure 4. Shoreline Fault Segments PG&E Letter DCL-1 0-106 Page 48 of 131 0.01

-u-e--Updated WithotA Shoreline

-4m-Updated With Shoreline

--.--LTSP II i.I.

£

• 1 0.001 0.0001 0.00001 0"000001 0ý2 0.7 1.2 1.7 2.2 2.7 3.2 4;2 0.0000001 L------

Spectral Acceleration, 3 to 8.5 hertz (g)

Figure 5. Comparison of the hazard for spectral acceleration averaged over 3-8.5 Hz from the 1988 LTSP and the updated hazard with and without the Shoreline fault.

PG&E Letter DCL- 0-106 Page 49 of 131 Probabilistic Risk Assessment Diablo Canyon Power Plant (DCPP) has performed a seismic probabilistic risk assessment (PRA) using seismic hazard curves that accounts for new seismic data, including the Shoreline Fault and use of an updated methodology. The scope of the update was limited to a change in the seismic initiating event frequencies based on the use of those curves and the resulting model is identified as the DC01 B model. As shown in the table below, all of the seismic initiating event frequencies decreased with the exception of SEIS6, which increased:

Seismic Initiating Event Frequency Update Su mmary Initiating Event DC01A Frequency DC01B Frequency SEIS1 (g levels 2.02E-1 to 1.25) 1.72E-02 7.40E-03 SEIS2 (g levels 1.25 to 1.75) 8.69E-04 3.50E-04 SEIS3 (g levels 1.75 to 2.00) 1.56E-04 7.OOE-05 SEIS4 (g levels 2.00 to 2.50) 1.24E-04 6.70E-05 SEIS5 (g levels 2.5 to 3.00) 2.94E-05 2.60E-05 SEIS6 (g levels 3.00 to 3.99) 7.64E-06 1.74E-05 The SAMA analysis was updated to reflect these changes, which are summarized in this response.

Recalculation of the External Events (EE) Multiplier Based on the changes to the initiating event frequencies, the seismic core damage frequency (CDF) decreased and impacted the EE multiplier. This has been recalculated using the same process described in section F.4.6.2 of the Environmental Report (ER). The following is a summary of the updated external events CDF contributors:

IPEEE Contributor Summary External Event Initiator Group CDF Fire 1.39E-05 Seismic 3.15E-05 High Winds 3.20E-07 Transportation & Nearby Facility - ship impact 1.90E-08 Transportation & Nearby Facility - accidental aircraft impact 7.OOE-07 External Flooding 7.20E-07 Chemical Release 8.010E-07 Total EE CDF 4.80E-05 PG&E Letter DCL-10-106 Page 50 of 131 The EE multiplier is calculated as follows (using the internal events CDF of 8.44E-06 that was used in the ER):

EE Multiplier = (8.44E-06+4.80E-05) / (8.44E-06) = 6.7 Recalculation of the Maximum Averted Cost-Risk (MACR)

Using the updated external event multiplier of 6.7, the updated MACR is $6,700,000:

MACR = Internal Events Cost-risk

• External Event Multiplier MACR = $1,000,000

• 6.7 = $6,700,000 Importance Review Threshold Determination For the seismic analysis, the risk reduction worth (RRW) threshold was derived using the assumption that the seismic CDF is directly proportional to its component of the MACR. Once the seismic component of the MACR is defined for the unit, the factor that would reduce the MACR by $50,000 (it would be $100,000 for the site) is defined to be the RRW review threshold. For DCPP, the seismic contribution to the MACR is defined as follows:

Seismic Contribution to MACR = (Unit MACR - Unit MACR/External Event Multiplier)

• (Unit Seismic CDF/ Unit External Event CDF) or Seismic Contribution to MACR = ($6,700,000 - $6,700,000/6.7)

• (3.15E-05/4.80E-05) = $3,740,625 The RRW threshold is then calculated by dividing the seismic contribution to the MACR by the seismic contribution to the MACR minus $50,000:

RRW Threshold ='$3,740,625 / $3,690,625 = 1.013 After rounding down to 1.01, this is the same value used in the ER. The internal events and fire RRW thresholds also remained the same.

Importance List Review The updated seismic importance list was reviewed down to the 1.01 RRW level and it was determined that essentially the same contributors were included on the list, but in a different order. The result was that the SAMA identification process produced nearly identical results to those documented in the ER. No new SAMAs were identified that PG&E Letter DCL-10-106 Page 51 of 131 are unique to the Shoreline Fault. However, one issue was identified that resulted in a change in the way some of the events were dispositioned.

During the review of the quantification of SAMA 18, it was determined that credit was not being applied to seismic events larger than 1.75g. This is because most of the Class I equipment needed for safe shutdown is only seismically qualified up to that level. In those scenarios, seal loss-of-coolant-accidents (LOCAs) are always assumed to occur. While SAMA 18 could provide makeup for a smaller seal LOCA, injection inventory would eventually deplete and recirculation mode would be required. For large seismic events, no credit is taken for alternating current (AC) power recovery and restoration of residual heat removal operation even if SAMA 18 is initially successful. In addition, other critical failures also may occur, such as reactor coolant system leaks, steam generator ruptures, and excessive LOCAs.

For those large seismic events, only a specialized system would be capable of mitigating the damage. Such a system would include a 4kV power source, a core spray type injection system (with a qualified pressure operated relief valve (PORV) capable of filling the reactor cavity, a connection to a large seismically qualified source of water (wells or seawater), and a heat exchanger system (defined as SAMA 26). A DCPP specific cost has not been developed for this SAMA, but the Limerick Severe Accident Mitigation Design Alternative analysis estimated that the cost of an independent suppression pool cooling system would cost $25,600,000 (Bechtel 1989), which is a large scale system installation similar in scope to work that would be required for DCPP.

The Limerick system used a diesel powered pump instead of relying on AC power and did not have the same level of seismic requirements, but even the cost of that system is greater than the 95th percentile MACR for DCPP ($6,700,000

• 2.36 = $15,812,000).

SAMA 26 is identified as a potential SAMA for the split fractions associated with the large seismic events in the updated importance list review, but it would be screened out by cost in the Phase I analysis and it is not considered further. Table 3.c-1 provides the results of the updated seismic importance list review.

Phase I Re-Evaluation There were no changes to the Phase I screening process based on the reduction in the MACR from $7,400,000 to $6,700,000.

PG&E Letter DCL-10-106 Page 52 of 131 Phase II Re-Evaluation For most cases, the only changes to the quantifications were the updates of the seismic CDF values. Table 3.C-2 provides the updated seismic CDF values for the Phase II SAMAs. These updated CDFs were used to recalculate the net values for the SAMAs.

The exceptions include the following:

The revised quantification method described in the response to RAI 6.o is used to provide the benefits for SAMAs 3, 11, 24, and 25.

For SAMA 9, it was determined that the PRA results used in the ER had not been updated to reflect the final quantification for the SAMA. The values have been updated to reflect the results of the final quantification.

The PRA modeling for SAMA 18 has been modified to address additional failures in the model that were deemed to be recoverable by the SAMA.

The quantification for SAMA 5 was reperformed to address an oversight in the process.

The response to RAI 6.c includes a detailed description of the model changes that were to represent SAMA 18 in the PRA. The updated PRA results for SAMAs 5 and 18 are presented below.

SAMA 5 PRA Results Summary for DC01 B CDF Dose-Risk OECR Fire CDF Seismic CDF Base Value 8.44E-06 8.79

$33,699 1.39E-05 3.15E-05 SAMA Value 7.52E-06 7.31

$27,897 1.25E-05 3.12E-05 Percent Change 10.9%

16.8%

17.2%

10.1%

1.0%

A further breakdown of the Dose-Risk and offsite economic cost-risk (OECR) information is provided in the table below according to release category:

PG&E Letter DCL-10-106 Page 53 of 131 SAMA 5 Internal Events Results By Release Category for DC01 B Release Category STI ST2 ST3 ST4 ST5 ST6 Total FrequencyBASE 7.18E-08 1.66E-06 3.97E-06 1.23E-06 2.88E-07 1.22E-06 8.44E-06 FrequencySAMA 6.63E-08 1.66E-06 3.50E-06 1.23E-06 1.53E-07 9.04E-07 7.52E-06 Dose-RiskBAsE 1.18 2.99 0.13 1.55 2.94 0.01 8.79 Dose-RisksAMA 1.09 2.99 0.12 1.55 1.56 0.00 7.31 OECRBASE

$1,206

$11,919

$35

$8,327

$12,211

$1

$33,699 OECRSAMA

$1,114

$11,919

$31

$8,346

$6,487

$1

$27,897 SAMA 9 PRA Results Summary for DC01 B CDF Dose-Risk OECR Fire CDF Seismic CDF Base Value 8.44E-06 8.79

$33,699 1.39E-05 3.44E-05 SAMA Value 8.41 E-06 8.80

$33,730 4.23E-06 3.44E-05 Percent Change 0.4%

0.0%

-0.1%

69.6%

0.0%

A further breakdown of the Dose-Risk and OECR information is provided in the table below according to release category:

SAMA 9 Internal Events Results By Release Category for DC01 B Release Category ST1 ST2 ST3 ST4 ST5 ST6 Total FrequencyBAsE 7.18E-08 1.66E-06 3.97E-06 1.23E-06 2.88E-07 1.22E-06 8.44E-06 FrequencysAMA 7.15E-08 1.66E-06 3.95E-06 1.23E-06 2.88E-07 1.20E-06 8.41E-06 Dose-RiskBAsE 1.18 2.99 0.13 1.55 2.94 0.01 8.79 Dose-RiskSAMA 1.17 3.00 0.13 1.55 2.93 0.01 8.80 OECRBASE

$1,206

$11,919

$35

$8,327

$12,211

$1

$33,699 OECRSAMA

$1,201

$11,949

$35

$8,345

$12,199

$1

$33,730 SAMA 18 PRA Results Summary for DC01B CDF Dose-Risk OECR Fire CDF Seismic CDF Base Value 8.44E-06 8.79

$33,699 1.39E-05 3.15E-05 SAMA Value 7.52E-06 7.31

$27,897 1.25E-05 1.79E-05 Percent Change 10.9%

16.8%

17.2%

10.1%

43.2%

PG&E Letter DCL-10-106 Page 54 of 131 A further breakdown of the Dose-Risk and OECR information is provided in the table below according to release category:

SAMA 18 Internal Events Results By Release Category for DC01 B Release Category STi ST2 ST3 ST4 ST5 ST6 Total FrequencyBASE 7.18E-08 1.66E-06 3.97E-06 1.23E-06 2.88E-07 1.22E-06 8.44E-06 FrequencySAMA 6.63E-08 1.66E-06 3.50E-06 1.23E-06 1.53E-07 9.04E-07 7.52E-06 Dose-RiskBAsE 1.18 2.99 0.13

-1.55 2.94 0.01 8.79 Dose-RisksAMA 1.09 2.99 0.12 1.55 1.56 0.00 7.31 OECRBASE

$1,206

$11,919

$35

$8,327

$12,211

$1

$33,699 OECRSAMA

$1,114

$11,919

$31

$8,346

$6,487

$1

$27,897 Table 3.C-3 provides the 95th percentile cost benefit analysis results for the ER and the DC01 B model. The results show that use of the updated seismic model does not impact the conclusions of the SAMA analysis; the SAMAs shown to be potentially cost beneficial using the DC01B model are the same as those determined to be cost beneficial in the ER. While SAMA 9 is shown to be potentially cost beneficial in this response while it was not in the ER, the difference is due to the correction of the fire CDF used in the cost benefit analysis rather than any change to the seismic results.

PG&E Letter DCL-10-106 Page 55 of 131 TABLES Table 3.c-1 DCPP Level I Updated Seismic Importance List Review Risk Event Probability Reduction Description Potential SAMAs Worth This SF represents the failure of all vital 4kV AC power given that the turbine building does not fail due to the seismic event. In most cases, the 230kV offsite supply is also failed and power is not available to the site at all. Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified pressure operated relief valve (PORV) 1.4900E-Seismic failure (SF) of AC TB capable of filling the reactor cavity, a connection to a large seismically qualified source of water SACSS5 001 1.1049E+000 struct successful (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the loss of all offsite power and is based on the 230kV switchyard seismic fragility, which is significantly stronger than the 500kV switchyard seismic fragility. Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required.

Such a system would include a 4kV power source, a core spray type injection system (with a 9.9700E-qualified PORV) capable of filling the reactor cavity, a connection to a large seismically qualified SOP6 001 1.0992E+000 SF of 230kV (OG) due to fragility source of water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the failure of all vital 4kV AC power given that the turbine building does not fail due to the seismic event. In most cases, the 230kV offsite supply is also failed and power is not available to the site at all. Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the reactor 3.3000E-Seismic failure of AC TB struct cavity, a connection to a large seismically qualified source of water (wells or seawater), and a heat SACSS6

- 001 1.0871 E+000 successful exchanger system (SAMA 26).

This is an intermediate SF for OBIS and is dominated by an operator error to initiate feed and bleed. The primary contributors containing this SF are low magnitude seismic events (SEIS1) in which auxiliary feedwater (AFW) fails due to operator error to switch to an alternate water source after failure of the condensate storage tank (CST) given the raw water storage reservoir (RWSR) 3.4152E-Loss of Instrument Air - PORV is unavailable. Subsequent to the AFW failure, the operators fail to initiate feed and bleed.

OB1SE 001 1.0866E+000 474 DISABLED Automating feed and bleed initiation is a potential means of mitigating this scenario (SAMA 14).

PG&E Letter DCL-10-106 Page 56 of 131 EetRisk

-1 Event Probability Reduction Description Potential SAMAs Name' euto ecito Worth This SF represents the failure of all vital 4kV AC power given that the turbine building does not fail due to the seismic event. In most cases, the 230kV offsite supply is also failed and power is not available to the site at all. Given that this SF is associated With a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required.- Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the reactor 3.9200E-cavity, a connection to a large seismically qualified source of water (wells or seawater), and a heat SACSS4 002 1.0866E+000 SF of AC TB struct.successful exchanger system (SAMA 26).

The primary contributor containing this SF is a low magnitude seismic event (SEIS1) in which AFW system low power top AFW fails due to operator error to switch to an alternate water source given the CST and RWSR 8.0586E-successful if TDP or 1 MDP are unavailable. Subsequent to the AFW failure, the operators fail to initiate feed and bleed.

AWlS 004 1.0793E+000 supplies water to SG for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> Automating feed and bleed initiation is a potential means of mitigating this scenario (SAMA 14).

This is an intermediate SF for AWS1. The primary contributor containing this SF is a low AFW system low power SEIS 1/2, magnitude seismic event (SEIS1) in which AFW fails due to operator error to switch to an 1/4 SG'S, 1/3 PP'S, FSWT as alternate water source given the CST and RWSR are unavailable. Subsequent to the AFW 8.0586E-supplemental (all support failure, the operators fail to initiate feed and bleed. Automating feed and bleed initiation is a AWS1 004 1.0793E+000 available) potential means of mitigating this scenario (SAMA 14).

This SF is dominated by operator error to initiate feed and bleed. The primary contributors containing this SF are low magnitude seismic events (SEIS1) in which AFW fails due to an operator error to switch to an alternate water source after failure of the CST given the RWSR is 3.4152E-unavailable. Subsequent to the AFW failure, the operators fail to initiate feed and bleed.

OB1 S 001 1.0774E+000 Bleed and feed Automating feed and bleed initiation is a potential means of mitigating this scenario (SAMA 14).

This SF represents the failure of the operators to reset seismic relay chatter given a seismically induced loss of offsite power (LOOP). Without relay reset, onsite AC sources cannot be aligned to required loads. This could be addressed by replacing critical relays with high capacity relays (SAMA 19). Use of portable generators to support AFW and alternate seal injection pumps would not likely provide a large benefit given that the operator dependence issues would limit the credit 2.2522E-Operator resets seismic relay for an additional action to mitigate loss of power; however, SAMA 18 could potentially reduce the OClSC 001 1 1.0505E+000 chatter I risk-of these events.

PG&E Letter DCL-1 0-106 Page 57 of 131 Event Risk Name Probability Reduction Description Potential SAMAs Worth This SF represents the loss of all offsite power and is based on the 230kV switchyard seismic fragility, which is significantly stronger than the 500kV switchyard seismic fragility. Given that this SF is associated with a large scale seismic event (greater thanl.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required.

Such a system would include a 4kV power source, a core spray type injection system (with a 9.8400E-Seismic failure of 230kV (OG) due qualified PORV) capable of filling the reactor cavity, a connection to a large seismically qualified SOP5 001 1.0502E+000 to fragility source of water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the failure of the emergency AC power system due to seismically induced relay chatter. Without relay reset, onsite AC sources cannot be aligned to required loads. Given that this SF is associated with a large scale seismic event (greater than1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system 2.7200E-(witha qualified PORV) capable of filling the reactor cavity, a connection to a large seismically SCT5 001 1.0413E+000 Relay chatter qualified source of water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the seismic failure of all six diesel generators and it is combined in the N sequences with a seismically induced LOOP, resulting in an station blackout (SBO). Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the reactor cavity, a connection to a large seismically 6.6700E-Seismic failure of DEG due to qualified source of water (wells or seawater), and a heat exchanger system SDG5 002 1.0400E+000 fragility (SAMA 26).

This SF represents the seismic failure of all six diesel generators and it is combined in the sequences with a seismically induced LOOP, resulting inan SBO. Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified 1.8300E-Seismic failure of DEG due to PORV) capable of filling the reactor cavity, a connection to a large seismically qualified source of SDG6 001 1.0384E+000 fragility water (wells or seawater), and a heat exchanger system (SAMA 26).

PG&E Letter DCL-10-106 Page 58 of 131 Event Risk Eve Probability Reduction Description Potential SAMAs Worth This SF represents the loss of all offsite power and is based on the 230kV switchyard seismic fragility, which is significantly stronger than the 500kV switchyard seismic fragility. Given that this SF is associated with a large scale seismic event (greater than1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required.

Such a system would include a 4kV power source, a core spray type injection system (with a 9.2700E-Seismic failure of 230kV (OG) due qualified PORV) capable of filling the reactor cavity, a connection to a large seismically qualified SOP4 001 1.0383E+000 to fragility source of water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the failure of the emergency AC power system due to seismically induced relay chatter. Without relay reset, onsite AC sources cannot be-aligned to required loads. Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system 4.5100E-(with a qualified PORV) capable of filling the reactor cavity, a connection to a large seismically SCT6 001 1.0335E+000 Relay chatter qualified source of water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the seismic failure of all six diesel generators and it is combined in the sequences with a seismically induced LOOP, resulting in an SBO. Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified 1.6900E-Seismic failure of DEG DUE to PORV) capable of filling the reactor cavity, a connection to a large seismically qualified source of SDG4 002 1.0322E+000 fragility water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the seismic failure of 125V DC power. This SF is typically combined with LOOP events, which result in SBO scenarios given that DC power is required for onsite power alignment. Given that this SF is associated with a large scale seismic event (greater than1.75g),

a new mitigating system capable of responding after seismic events (potentially up to 4g) is c6nsidered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the reactor cavity, a connection to a 1.2600E-Seismic failure of (direct current) large seismically qualified source of water (wells or seawater), and a heat exchanger system SDC6 001 1.0245E+000 DC due to fragility (SAMA 26).

PG&E Letter DCL-10-106 Page 59 of 131 Event Risk Name Probability Reduction Description Potential SAMAs Worth This SF represents the seismic failure of 125V DC power. This SF is typically combined with LOOP events, which result in SBO scenarios given that DC power is required for onsite power alignment. Given that this SF is associated with a large scale seismic event (greater than 1.75g),

a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the reactor cavity, a connection to a 3.8200E-Seismic failure of DC due to large seismically qualified source of water (wells or seawater), and a heat exchanger system SDC5 002 1.0222E+000 fragility (SAMA 26).

This in an intermediate SF for RT4S, which represents the failure to depower the control rods after a plant trip signal. The dominant contributor is operator failure to de-energize the reactor protection system (RPS) bus. A potential means of reducing the contribution of this SF is to use an alternate signal, such as anticipated transient without scram mitigating system actuation 3.1163E-1/1 train (only one SSPS signal

_(AMSAC), to automate the de-energization of the 480V buses feeding the rod drive motor RT4SE 003 1.0218E+000 generated) - SEIS W/ESAM=30 generator sets (SAMA 20).

This SF represents the failure to depower the control rods after a plant trip signal. The dominant contributor is operator failure to de-energize the RPS bus. A potential means of reducing the 3.1163E-contribution of this SF is to use an alternate signal, such as AMSAC, to automate the RT4S 003 1.0218E+000 Reactor trip de-energization of the 480V buses feeding the rod drive motor generator sets (SAMA 20).

This in an intermediate SF for RT1 S, which represents the failure to depower the control rods after a plant trip signal. The dominant contributor is a failure of the breakers to change state combined with the operator failure to de-energize the RPS bus. A potential means of reducing the 8.5123E-1/2 trains (both SSPS signals contribution of this SF is to use an alternate signal, such as AMSAC, to automate the RT1SE 005 1.0212E+000 generated) - SEIS W/ESAM=30 de-energization of the 480V buses feeding the rod drive motor generator. sets (SAMA 20).

This SF represents the failure to depower the control rods after a plant trip signal. The dominant contributor is a failure of the breakers to change state combined with the operator failure to de-energize the RPS bus. A potential means of reducing the contribution of this SF is to use an 8.5123E-alternate signal, such as AMSAC, to automate the de-energization of the 480V buses feeding the RT1 S 005 1.0211 E+000 Reactor trip rod drive motor generator sets (SAMA 20).

PG&E Letter DCL-1 0-106 Page 60 of 131 Event Risk Nae Probability Reduction Description Potential SAMAs Nm Worth This SF represents the failure of all vital 4kV AC power given that the turbine building does not fail due to the seismic event. In most cases, the 230kV offsite supply is also failed and power is not available to the site at all. Given that this SF is associated with a large scale seismic event (greater than 1:75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the reactor 8.3600E-Seismic failure of AC TB struct cavity, a connection to a large seismically qualified source of water (wells or seawater), and a heat SACSS3 003 1.0187E+000 successful exchanger system (SAMA 26).

This SF is important due to its inclusion in LOOP scenarios with other AC power related failures that lead to SBO evolutions. The largest of these contributors are combinations of other emergency diesel generator (EDG) failures. A potential solution is to use a portable 480V AC generator to supply the battery chargers for long-term AFW support in conjunction with a self-6.5530E-DG 1-3 (BUS F) starts & runs for 6 cooled, 480V AC reactor coolant system (RCS) makeup pump that can be used to mitigate GF1 S 002 1.0186E+000 hr SEIS W/ESAM=30 reactor coolant pump (RCP) seal leakage (SAMA 18).

This SF (intermediate for TGIS and TG3S) represents failure of the Unit 2 DG 2-1 EDG, which appears in conjunction with seismically induced LOOP and relay chatter events that result in a SBO. These scenarios could be addressed by replacing critical relays with high capacity relays (SAMA 19). An additional potential solution is to use a portable 480V AC generator to supply the 8.6426E-1/3 diesels unavailable (BUSG) -

battery chargers for long term AFW support in conjunction with a self-cooled, 480V AC RCS GYGS 002 1.0171E+000 SEIS W/ESAM=30 makeup pump that can be used to mitigate RCP seal leakage (SAMA 18).

This SF represents the seismic failure of 125V DC power. This SF is typically combined with LOOP events, which result in SBO scenarios given that DC power is required for onsite power alignment. Given that this SF is associated with a large scale seismic event (greater than1.75g),

a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the reactor cavity, a connection to a 8.4700E-Seismic failure of DC due to large seismically qualified source of water (wells or seawater), and a heat exchanger system SDC4 003 1.0170E+000 fragility (SAMA 26).

PG&E Letter DCL-10-106 Page 61 of 131 Event Risk Name Probability Reduction Description Potential SAMAs Worth This is an intermediate SF for TF1 S, which represents failure of EDG 2-3. These failures are included in sequences that are essentially those represented by other SFs on the list (seismically induced LOOP with seismically induced relay failure (chatter) or loss of component cooling water (CCW)). These scenarios could be addressed by replacing critical relays with high capacity relays (SAMA 19). An additional potential solution is to use a portable 480V AC generator to supply the 8.6426E-1/3 diesels unavailable (BUS F) -

battery chargers for long-term AFW support in conjunction with a self-cooled, 480V AC RCS GYFS 002 1.0167E+000 SEIS W/ESAM=30 makeup pump that can be used to mitigate RCP seal leakage (SAMA 18).

This is an intermediate SF for THIS, which represents failure of EDG 2-2. These failures are included in sequences that are essentially those represented by other SFs on the list (seismically induced LOOP with seismically-induced relay failure (chatter) or loss of CCW). These scenarios could be addressed by replacing critical relays with high capacity relays (SAMA 19). An additional potential solution is to use a portable 480V AC generator to supply the battery chargers for long-8.6426E-1/3 diesels unavailable (BUS H) -

term AFW support in conjunction with a self-cooled, 480V AC RCS makeup pump that can be GYHS 002 1.0167E+000 SEIS W/ESAM=30 used to mitigate RCP seal leakage (SAMA 18).

This intermediate SF for SB1, which is a failure of the "B" train of the solid state protection system, is important because SB1 is paired with breaker failures that result in Anticipated Transient Without Scram (ATWS) events, which are assumed to result in core damage for seismic initiators. A potential means of reducing the contribution of this SF is to use an alternate 1.4154E-SSPS train B fails (general signal, such as AMSAC, to automate the de-energization of the 480V buses feeding the rod drive S11B 002 1.0164E+000 transient) motor generator sets (SAMA 20).

This intermediate SF for SA1, which is a failure of the "A" train of the solid state protection system, is important because SA1 is often paired with breaker failures that result in ATWS events (75 percent of the SF contribution), which are assumed to result in core damage for seismic initiators. A potential means of reducing the contribution of this SF is to use an alternate signal, 1.4154E-SSPS train A fails (general such as AMSAC, to automate the de-energization of the 480V buses feeding the rod drive motor S11A 002 1.0164E+000 transient) generator sets (SAMA 20).

This SF represents the seismic failure of CCW. Given that this SF is associated with a large scale seismic event (greater than 1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system (with a qualified PORV) capable of filling the 1.5800E-Seismic failure of CCW (CC) due reactor cavity, a connection to a large seismically qualified source of water (wells or seawater),

SCC6 001 1.01 57E+000 to fragility and a heat exchanger system (SAMA 26).

PG&E Letter DCL-10-106 Page 62 of 131 Event 1Risk Eve Probability Reduction Description Potential SAMAs Name JWorth This SF, which is a failure of the "A" train of the solid state protection system, is often paired with breaker failures that result in ATWS events, which are assumed to result in core damage for seismic initiators. A potential means of reducing the contribution of this SF is to use an alternate 1.4154E-signal, such as AMSAC, to automate the de-energization of the 480V buses feeding the rod drive SA1 002 1.0151E+000 SSPS train A motor generator sets (SAMA 20).

This is an intermediate SF for GF1S ("F" EDG failure). GF1S is important primarily due to its inclusion in LOOP scenarios with other AC power related failures that lead~to SBO evolutions.

The largest of these contributors are combinations of other EDG failures. A potential solution is to use a portable 480V AC generator to supply the battery chargers for long term AFW support in 6.5530E-1/3 diesels unavailable (BUS F) -

conjunction with a self-cooled, 480V AC RCS makeup pump that can be used to mitigate RCP GXFS 002 1.0151E+000 SEIS W/ESAM=30 seal leakage (SAMA 18).

This SF represents the failure of the operators to reset seismic relay chatter given a seismically induced LOOP. Without relay reset, onsite AC sources cannot be aligned to required loads. This could be addressed by replacing critical relays with high capacity relays (SAMA 19). Use of portable generators to support AFW and alternate seal injection pumps would not likely provide a large benefit given that the operator dependence issues would limit the credit for an additional 3.7537E-operator resets seismic relay action to mitigate loss of power; however, SAMA 18 could potentially reduce the risk of these OC1SB 002 1.0150E+000 chatter events.

This is an intermediate SF for GH1S ("H" EDG failure). GH1S is important primarily in SBO sequences that include AC failures due to relay chatter (and no operator reset). Without relay reset, onsite AC sources cannot be aligned to required loads. This could be addressed by replacing critical relays with high capacity relays (SAMA 19). Use of portable generators to support AFW and alternate seal injection pumps would not likely provide a large benefit given that 6.5530E-1/3 diesels unavailable (BUS H) -

the operator dependence issues would limit the credit for an additional action to mitigate loss of GXHS 002 1.0137E+000 SEIS W/ESAM=30 power; however, SAMA 18 could potentially reduce the risk of these events.

This is an intermediate SF for GG1S ("G" EDG failure). GG1S is important primarily in SBO sequences that include AC failures due to relay chatter (and no operator reset). Without relay reset, onsite AC sources cannot be aligned to required loads. This could be addressed by replacing critical relays with high capacity relays (SAMA 19). Use of portable generators to support AFW and alternate seal injection pumps would not likely provide a large benefit given that 6.5530E-1/3 diesels unavailable (BUS G) -

the operator dependence issues would limit the credit for an additional action to mitigate loss of GXGS 002 1.0134E+000 SEIS W/ESAM=30 power; however, SAMA 18 could potentially reduce the risk of these events.

PG&E Letter DCL-10-106 Page 63 of 131 Event Risk Name Probability Reduction Description Potential SAMAs Worth This SF represents the loss of all offsite power and is based on the 230kV switchyard seismic fragility, which is significantly stronger than the 500kV switchyard seismic fragility. Given that this SF is associated with a large scale seismic event (greater than1.75g), a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required.

Such a system would include a 4kV power source, a core spray type injection system (with a 8.2700E-seismic failure of 230kV (OG) due qualified PORV) capable of filling the reactor cavity, a connection to a large seismically qualified SOP3 001 1.0117E+000 to fragility source of water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF represents the failure of the emergency AC power system due to seismically induced relay chatter. Without relay reset, onsite AC sources cannot be aligned to required loads. Given that this SF is associated with a large scale seismic event (greater than 1.75g); a new mitigating system capable of responding after seismic events (potentially up to 4g) is considered to be required. Such a system would include a 4kV power source, a core spray type injection system 1.3400E-(with a qualified PORV) capable of filling the reactor cavity, a connection to a large seismically SCT4 001 1.0111 E+000 Relay chatter qualified-source of water (wells or seawater), and a heat exchanger system (SAMA 26).

This SF is a DG 1-2 failure and it is typically combined with seismically induced LOOP and the failure of the other two station EDGs, which results in an SBO. A potential solution is to use a portable 480V AC generator to supply the battery chargers for long-term AFW support in 7.0873E-DG 1-2 (BUS G): GF-F - SEIS conjunction with a self cooled, 480V AC RCS makeup pump that can be used to mitigate RCP GG2S 002 1.0107E+000 W/ESAM=30 seal leakage (SAMA 18).

-This SF, which is a failure of the "B" train of the solid state protection system, is often paired with breaker failures that result in ATWS events, which are assumed to result in core damage for seismic initiators. A potential means of reducing the contribution of this SF is to use an alternate 1.3821 E-signal, such as AMSAC, to automate the de-energization of the 480V buses feeding the rod drive SB1 002 1.0101 E+000 SSPS train B motor generator sets (SAMA 20).

PG&E Letter DCL-10-106 Page 64 of 131 Table 3.C-2: DC01B Seismic CDF Results Summary Case Description DCOIB Seismic CDF Baseline 3.15E-05 SAMA2 3.15E-05 SAMA3 3.15E-05 SAMA5 3.12E-05 SAMA6 3.1OE-05 SAMA7 3.15E-05 SAMA8 3.15E-05 SAMA9 3.15E-05 SAMA10 2.78E-05 SAMA11 3.15E-05 SAMA12 3.15E-05 SAMA13 3.15E-05 SAMA14 3.15E-05 SAMA15 3.15E-05 SAMA16 3.15E-05 SAMA17 3.15E-05 SAMA1 8 1.79E-05 SAMA19 2.87E-05 SAMA20 3.02E-05 SAMA24 3.15E-05 SAMA25 3.15E-05 PG&E Letter DCL-1 0-106 Page 65 of 131 Table 3.C-3: SAMA Results Comparison, ER Vs. DC01B SAMA Implementation Averted Net Value Averted Cost-risk Net Value Change in Cost ID Cost (per unit)

(ER, 95th)

(ER, 95th)

(DC01B, 95th)

(DC01B 95th)

Effectiveness?

2

$6,509,256

$179,687

-$6,329,569

$192,451

-$6,316,805 No 3

$5,863,176

$3,085,978

-$2,777,198

$3,722,728

-$2,140,448 No 5

$6,441,418

$681,000

-$5,760,418

$978,024

-$5,463,394 No 6

$14,475,422

$1,100,015

-$13,375,408

$1,044,453

-$13,430,969 No 7

$2,552,563

$18,457

-$2,534,106

$35,678

-$2,516,885 No 8

$6,376,810

$2,914,522

-$3,462,288

$2,940,916

-$3,435,894 No 9

$1,692,730

$120,800

-$1,571,930

$2,731,160

$1,038,430 Yes 10

$6,234,672

$4,478,898

-$1,755,774

$4,281,583

-$1,953,089 No 11

$4,651,776

$2,776,270

-1,875,506

$1,401,292

-$3,250,484 No 12

$775,296

$1,576,857

$801,561

$1,598,456

$823,160 No 13

$775,296

$1,053,484

$278,188

$1,073,644

$298,348 No 14

$11,435,616

$1,074,348

-$10,361,268

$1,081,078

-$10,354,538 No 15

$9,626,592

$289,494

-$9,337,098

$290,934

-$9,335,658 No 16

$3,944,318

$107,235-

-$3,837,083

$132,750

-$3,811,568 No 17

$5,184,792

$6,733

-$5,178,059

$23,397

-$5,161,395 No 18

$6,441,418

$607,986

-$5,833,431

$4,705,349

-$1,736,069 No 19

$15,312,096

$844,430

-$14,467,666

$806,662

-$14,505,434 No 20

$7,526,832

$868,093

-$6,658,739

$401,863

-$7,124,969 No 24

$50,000

$62,484

$12,484

$423,242

$373,242 No 25

$50,000

$95,596

$45,596

$1,219,764

$1,169,764 No Reference

)

Bechtel 1989 Bechtel, "Cost Estimate for Severe Accident Mitigation Design Alternatives, Limerick Generating Station for Philadelphia Electric Company," June 22, 1989.

PG&E Letter DCL-10-106 Page 66 of 131 References Abrahamson, N. A. and W. J. Silva (2008). Summary of the Abrahamson & Silva NGA Ground-Motion Relations, Earthquake Spectra, Vol. 24, No. 1, 67 - 98.

Boore, D. M. and G. M. Atkinson (2008). Ground-Motion Prediction Equations for the Average Horizontal Component of PGA, PGV, PGD and 5%-Damped PSA at Spectral Periods between 0.01 s and 10.0 s, Earthquake Spectra, Vol. 24, No. 1,67-138.

Campbell, K. W. and Y. Bozorgnia (2008). NGA Ground Motion Model for the Geometric Mean Horizontal Component of PGA, PGV, PGD and 5% Damped Linear Elastic Response Spectra for Periods Ranging from 0.01 to 10 s, Earthquake Spectra, Vol. 24, No. 1, 139 - 172.

Chiou, B. S-J and R. R. Youngs (2008). An NGA Model for the Average Horizontal Component of Peak Ground Motion and Response Spectra, Earthquake Spectra, Vol. 24, No. 1, 173 - 216.

Idriss, I. M. (2008). An NGA Empirical Model for Estimating the Horizontal Spectral Values Generated By Shallow Crustal Earthquakes, Earthquake Spectra, Vol. 24, No. 1,217-242.

McLaren, M. and W. Savage (2001). Seismicity of South-Central Coastal California: October 1987 through January 1997, Bull. Seism. Soc. Am., 91: 1629-1658.

PG&E Letter DCL-10-106 Page 67 of 131 RAI 3.d

3. Provide the following information with regard to the treatment and inclusion of external events in the SAMA analysis:

d. Provide descriptions of the most important seismic CDF sequences including initiator, seismic failures and non-seismic failures and their CDF contributions.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 3.d The top ten seismic sequences for the updated seismic model (DC01 B) are listed below with a description of the seismic and nonseismic failures. Also the associated core damage frequency and contribution to total seismic core damage frequency (CDF) is also listed for each sequence. These top 10 sequences contribute more than 30 percenet of the total seismic CDF.

1.

Seismic Level 4 (2.0 to 2.5g) initiator with failure of 230kV alternating current (AC) grid (SOP) due to fragility. Also turbine building AC power sources are failed (SACSS) and offsite power cannot be restored due to the seismic damage (REZ). The sequence frequency is 2.1E-6 contributing 7 percent to seismic CDF.

2.

Seismic Level 5 (2.5 to 3.0g) with similar effects as sequence.1. The sequence frequency is 1.8E-6 contributing 6 percent to seismic CDF.

3.

Seismic Level 1 (0.2 to 1.25g) with seismic failure of auxiliary feedwater and failure of the operators to initiate feed and bleed during a seismic event. The sequence frequency is 1.8E-6 contributing 6 percent to seismic CDF.

4.

Seismic Level 6 (3.0 to 3.99g) with similar effects as sequence 1. The sequence frequency is 1.1 E-6 contributing 4 percent to seismic CDF.

5.

Seismic Level 5 (2.5 to 3.0g) with similar effects as sequence 1, except that the operators also fail to manually isolate containment during a seismic event to mitigate a large release. The sequence frequency is 1.OE-6 contributing 3 percent to seismic CDF.

6.

Seismic Level 4 (2.0 to 2.5g) initiator with failure of 230kV AC grid (9.27E-01 probability) and emergency diesels due to fragility. The sequence frequency is 7.OE-7 contributing 2 percent to seismic CDF.

PG&E Letter DCL-1 0-106 Page 68 of 131

7.

Seismic Level 6 (3.0 to 3.99g) with similar effects as sequence 5 and the additional seismic failure of the emergency diesels. The sequence frequency is 6.3E-7 contributing 2 percent to seismic CDF.

8.

Seismic Level 1 (0.2 to 1.25g) with seismic failure of the reactor to trip.

The sequence frequency is 5.5E-7 contributing 2 percent to seismic CDF.

9.

Seismic Level 5 (2:5 to 3.0g) with similar effects as sequence 6, except that the emergency diesels also seismically fail. The sequence frequency is 4.8E-7 contributing 2 percent to seismic CDF.

10.

Seismic Level 3 (1.75 to 2.0g) with similar effects as sequence 1. The sequence frequency is 4.3E-7 contributing 1 percent to seismic CDF.

PG&E Letter DCL-1 0-106 Page 69 of 131 RAI 3.e

3. Provide the following information with regard to the treatment and inclusion of external events in the SAMA analysis:

e. Several differences in fire CDF results are apparent when comparing the model DCOIA results in Figure F.2-4 with those in Table 4.6-4 of the IPEEE.

Specifically, initiator VB 1, the IPEEE's largest fire CDF contributor, is no longer ranked in the top 6 contributors and the CDF contribution from initiators FS6, FSI, and VB4 have been reduced in both a relative and an absolute sense,.

particularly when the change in the contribution from FS7 is considered. Discuss the reasons for these differences.

PG&E Response to Severe Accident Mitigation Alternatives RAI 3.e The main reason for a significant change to importance ranking and absolute values among fire scenarios (FS) (types versus CSR1/CSR2) is that FS type go through (use) the same event trees as general transients while CSR1/CSR2 use

'their own event trees. The CSR1I/CSR2 event trees are simplified in contrast.

As the Diablo Canyon Power Plant probabilistic risk assessment (PRA) models are continuously updated (mostly system top events and general transient event trees), it is normal that the impacts/importance of FS type changes as well.

However there has been very little change to the modeling for CSR1/CSR2, which are more of vulnerability type modeling (akin to individual plant examination of external events (IPEEE) models). For the IPEEE assessment, VB1 was the highest ranking initiator, but since then PRA model has been routinely updated resulting in core damage frequency decrease from FS type scenarios while CSR1/CSR2 contribution remained high. The end result is that now CSR1 and CSR2 have become the dominant scenarios.

PG&E Letter DCL-10-106 Page 70 of 131 RAI 3.f

3. Provide the following information with regard to the treatment and inclusion of external events in the SAMA analysis:

f, The description of the model changes resulting in PRA model DCOO (2000) suggest that the fire and seismic PRA models were integrated with the internal events Level I PRA model. Confirm whether this understanding is correct. If integrated, explain why the sequences from these models were not mapped into the CET and release categories.

PG&E Response to Severe Accident Mitigation Alternatives RAI 3.f Model DCOO (the combined model) first provided the capability to run internal events, seismic and fire initiators at the same time with one quantification in RISKMAN. At this point, in time in the evolution of the Diablo Canyon Power Plant model, the detailed containment event tree and its associated endstates was not used in favor of a simplified large early release frequency (LEIRF) model.

Detailed Level 2 quantification did not return until Model DC01 where only internal events were treated in detail. For seismic initiators, the Level 2 treatment remained simplified (i.e., LERF only). Because the more detailed Level 2 event tree and structure was returned to, fire initiating events were no longer analyzed after core damage. This continues with the DC01A model. The effort to model fire and seismic scenarios through the detailed Level 2 event tree has been deferred.

PG&E Letter DCL-1 0-106 Page 71 of 131 RAI 4.a

4. Provide the following information relative to the Level 3 analysis:

a. Section F. 3.1 indicates that the complex mountain terrain would be expected to increase the amount of deposition close to the site. Provide a brief discussion of whether/how the terrain effects were incorporated into the MEL COR Accident Consequence Code System Version 2 (MACCS2) analysis.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 4.a Complex terrain effects are not explicitly incorporated into the MACCS2 analysis.

The MACCS2 code utilizes a Gaussian plume model, which is inherently a flat earth model. As discussed in Section F.3.1, use of such a model is judged acceptable for the purposes of SAMA where the focus is on average (i.e., mean value) results.

The complex terrain was addressed in the SAMA analysis qualitatively.

Section F.3.1 notes that previous site analyses of the complex terrain documented in the Final Safety Analysis Report have found that the net result of the complex terrain and coastal wind regime is a very effective and wide daily dispersal of any pollutants that are present. The SAMA analysis additionally notes that the complex terrain would be expected to increase the amount of deposition close to the site due to the deposition process of impaction (i.e., dispersed material impacting and depositing upon surfaces). The complex terrain near the site is expected to increase the impaction of radiological material near the site where there is a low population density, thereby reducing the amount of radiological material reaching population centers located further from the site. As a result, the MACCS2 calculated SAMA population dose estimates are generally expected to be higher (i.e., dose results more conservative) than those that would be obtained specifically accounting for the complex terrain near the site.

PG&E Letter DCL-10-106 Page 72of 131 RAI 4.b

4. Provide the following information relative to the Level 3 analysis:

b. Section F.3.2 indicates that the transient population was included for the region within 10 miles of the site, but not for the 10-50 mile region. Section 2.1.3.4 of the final safety analysis report (FSAR) identifies that the peak transient population within the 50-mile radius is about 100,000 people. Also, from Table 2.1-4 of the FSAR the total transient visitor-days within the 50-mile radius is about 10,000,000, which would equate to a yearly average population of about 28, 000 people. These values appear to be for the year 2000 time frame. Justify not including a transient population in the 10-50 mile radius. Clarify whether/how the transient population was adjusted for year 2045 as these values are a significant fraction of the year 2000 permanent population. Discuss how the 10-mile radius transient population was developed from the FSAR transient population data, which appears to be based on the 6-mile radius low population zone (LPZ).

PG&E Response to Severe Accident Mitigqation Alternatives RAI 4.b The inclusion of transient population data for only the 0-10 mile region is consistent with the guidance presented in Section 3.4.1 of NEI 05-01, which specifies including transient data included in the site emergency plan. The Diablo Canyon Power Plant (DCPP) evacuation time estimate (ETE) assessment (Reference 67 of Attachment F) provides this transient data in Table 3 for the protective action zones, which encompass the 0-10 mile region. Inclusion of transient data for the 0-10 mile region is relevant for considering the potential impacts on evacuation and emergency phase population dose.

Including transient data in the 10-50 mile radius would increase the conservatism in the MACCS2 cost results. In the MACCS2 internal calculations, transient persons are treated as if they are permanent residents (e.g., property owners).

As a result, transients who are relocated as part of a postulated radiological release will accrue per-diem costs such as temporary housing and food in the early and intermediate phases, and per-capita costs associated with loss of property in the long term phase. Such costs are not generally appropriate for transients who would simply return to their permanent residences outside the affected area. Inclusion of transients also creates the potential for double-counting individuals. For instance, transients who work within the 0-10 mile region of the site but who reside in the 10-20 mile annulus from the site may be evaluated twice in MACCS2 by a postulated release, once due to inclusion in the transient data for 0-10 miles and a second time due to inclusion in the residential population for 10-20 miles. Inclusion of transients in the 10-50 mile region would PG&E Letter DCL-10-106 Page 73 of 131 increase the potential for such double-counting and increase the conservatism in the MACCS2 results.

The transient data for 0-10 miles was adjusted for the year 2045 assuming the transient population experiences the same growth rates as the residential population. Based on the Final Safety Analysis Report (FSAR) 0-10 mile residential population estimates for year 2000 and 2025 (FSAR Figures 2.1-4 and 2.1-6), growth factors were determined for each grid element. The FSAR projected zero growth between year 2000 and 2025 for the region of 0-5 miles, and essentially a uniform growth of 69 percent for all sector directions for the 5-10 mile region. For the MACCS2 analysis, the uniform growth of the 5-10 mile region was also applied to the 0-5 mile region rather than assuming zero growth.

The growth factors were applied to the combined residential and transient population data successively to project from year 2000 to 2025, and then to project from 2025 to 2045. A 0.8 factor was applied to each grid element growth factor for the second projection period to address a projection period of 20 years rather than 25. Essentially all the population data for the 0-5 mile region is due to the inclusion of transients since there are very few permanent residents in the 0-5 mile region.

Regarding transient population data development, transient data was obtained from the DCPP ETE rather than the FSAR as implied by the Section F.3.2.

Transient data from the ETE Table 3 for normal weekday summer daytime was used since this time period resulted in the largest number of transients. From the ETE Table 3, transient data was used for protective action zones 1, 2, 3, 4, and 5 which extend to 10 miles from the DCPP site. The transients were allocated to the individual grid elements assuming uniform distribution for the sector directions. The Table 3 data is for year 2000. This transient data was added to the residential population data for year 2000 from the FSAR for each grid element and projected to year 20451as discussed above.

PG&E Letter DCL-10-106 Page 74 of 131 RAI 4.c

4. Provide the following information relative to the Level 3 analysis:

c. Discuss whether and how the evacuation time was adjusted for the difference in population between year 2045 and the year of the referenced evacuation time estimate study.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 4.c The evacuation time was not specifically adjusted to account for the difference between the population associated with the current evacuation time estimate study for the year 2000 and the population projection to year 2045. Regional transportation infrastructure is assumed to keep pace with population growth over the projection period.

Uncertainties associated with the evacuation speed were addressed in the MACCS2 analysis in two ways. First, the evacuation speed used in the MACCS2 model base case of 0.4 m/s is based on the longest evacuation time estimate for the approximate 10 mile region (i.e., Scenario 5 from Table 16 of the Diablo Canyon Power Plant evacuation time estimate assessment, Reference 67 of Attachment F). This evacuation speed of 0.4 m/s equates to 0.9 mph, considerably less than a typical walking speed of 3 mph and is therefore judged reasonably conservative for use in the base case. Second, the evacuation speed parameter was evaluated in a sensitivity case as documented in Section F.7.3.2. For the sensitivity case, the evacuation speed was decreased from 0.4 m/s to 0.2 m/s. This decrease in evacuation speed resulted in a very small increase in population dose of 3 percent. Thus the MACCS2 population dose results are not significantly impacted by a slower evacuation speed, such as might accompany population growth without infrastructure improvements.

PG&E Letter DCL-10-106 Page 75 of 131 RAI 4.d

4. Provide the following information relative to the Level 3 analysis:

d. Section F. 3.5 identifies the reference for the core inventory (Table F. 3-3) as a "DCPP Informal Calculation." Describe the level of review of this calculation.

Confirm that this core inventory reflects the anticipated fuel management/bumup during the renewal period.

PG&E Response to Severe Accident Mitigation Alternatives RAI 4.d The level of review was limited to a review of the reasonability of the core inventories. The original values from the design-basis inventory calculations were not sufficient for the severe accident mitigation alternative analyses.

Discussions were held with the contractors performing the analyses and a representative evaluation was prepared utilizing the ORIGEN-S module of the SCALE4.4A code package. The inputs and results of the ORIGEN-S runs were then submitted to the contractor for review and acceptance. The contractor reviews concluded that the inventories provided were reasonable.

The core inventory was generated based on current fuel management practices and fuel design. There are no plans for uprating at this time. However, PG&E is considering the possibility of changing to a more standard fuel type. This change will result in an increased fuel inventory in the core and would enable a possible increase in the length of future operating cycles. These changes would result in some changes to the core inventories used in this analysis. The impacts of the changes would be evaluated in conjunction with the licensing of the new fuel type.

PG&E Letter DCL-10-106 Page 76 of 131 RAI 5.a

5. Provide the following information with regard to the SAMA selection and screening process:

a. Section F.5. 1.1 describes the basis for the lower risk reduction worth (RRVV) cutoff associated with the importance analysis review for the Phase 1 SAMA identification process. It is stated that, for the internal events model, an RRW of 1.04 corresponds to a single unit averted cost-risk of $49,965. This however does not include the additional benefits in external events. Using a total MACR of $7.4 million, the internal events cutoff of 1.02 actually used in the importance review corresponds to a minimum averted cost-risk of $145,000 for a single unit.

This is considerably greater than the single unit procedure change cost of

$50,000 used in the SAMA analysis. Provide a review of importance analysis results down to a RRW of 1.007 and provide an assessment of any new SAMAs identified.

PG&E Response to Severe Accident Mitigation Alternatives RAI 5.a Because separate, quantifiable, risk models are available for the Diablo Canyon Power Plant (DCPP) seismic and fire contributors, the approach suggested in this request for additional information (RAI) of directly correlating the maximum averted cost-risk (MACR) to the internal events review threshold is not appropriate. Su~ch an approach implies that the averted cost-risk of a given SAMA is calculated by applying a multiplier to the internal events results, which is not the case for DCPP.

The averted cost-risk for a SAMA is comprised of four separate parts:

Internal events.

Nonfire/nonseismic external events Fire Seismic The availability of explicit model results for fire and -seismic events requires the review threshold to be determined and defined separately. Section F.5.1.1 describes how the review thresholds are determined for the fire and seismic importance lists.

Because quantifiable risk models are not available for the nonfire/nonseismic external events contributors, the internal events review threshold-is established to account for the potential averted cost-risk associated with those types of PG&E Letter DCL-1 0-106 Page 77 of 131 events. In this case, the risk reduction worth (RRW) of 1.04 and the nonfire/nonseismic cost-risk of $1,300,000 correlates to $49,965 for a single unit.

While there may be some fire or seismic benefit associated with an event in the internal events importance list, any SAMA important to the fire or seismic results will be identified independently by means of the explicit importance list reviews that are performed for those models.

In summary, the RRW threshold of 1.007 suggested in this, RAI is not relevant to the DCPP analysis due to the nature of the quantification process and the thresholds established in the Environmental Report for the fire, seismic, and nonfire/nonseismic contributors are appropriate.

PG&E Letter DCL-10-106 Page 78 of 131 RAI 5.b

5. Provide the following information with regard to the SAMA selection and screening process:

b. The identification of potential DCPP SAMAs from importance reviews utilizes the importance at the split fraction level as opposed to the basic event or component level.

The latter might lead to the identification of more specific failures that might be mitigated by a design or procedure change. For example, IPEEE Table 3-10 lists the importance of top events, while Table 3-11 lists the importance at the component level. The latter includes failure of block walls that lead to failure of top event SACSS, "Failure of all vital 4 kV AC power." This indicates that a SAMA related to fixing the block walls might be considered as an alternative to SAMA 18. Provide justification that use of top event importance rather than basic event or component importance provides an adequate identification of potential low cost SAMAs, and additional supporting analyses, as appropriate.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 5.b It is rare for a SAMA to be developed to address the failure of a specific component or event. SAMAs are generally generated to address system or functional failures. For example, in the situation where the failure of a component cooling water heat exchanger valve is important, the resulting SAMA would not likely be to install a parallel valve. This approach does not account for the fact if one valve failure in the heat exchanger path is important, other valve failures in the heat exchanger path would likely be similarly important (none of which would be addressed by the installation of a single parallel valve).

Basic event/component level importance results are available for the Diablo Canyon Power Plant (DCPP) model, but the model structure'does support a practical, systematic means of reviewing how an individual basic event contributes to the important accident sequences. It is important to know the function and under what boundary conditions the function is important. Split fraction importance review is the best choice of those available in RISKMAN.

Split fraction importance provides the state of the boundary conditions and the nature of the sequence, neither of which can be found by solely looking at the components or basic events important to risk. Furthermore, the basic event level results are often reviewed during the sequence analysis to determine what the dominant contributors to a split fraction may be, but this process is not documented. Given these factors, split fraction/sequence review is considered to be the most appropriate means of SAMA identification for DCPP.

PG&E Letter DCL-1 0-106 Page 79 of 131 With regard to the contribution of block wall failures to the SACSS top event, Table 3-11 of the IPEEE indicates that their F-V values are at least an order of magnitude lower than the turbine building shear wall failures and two times lower than failures of the 4kV/480V transformers. Consequently, block wall failures are not a primary concern for the sequences including SACSS. Even if block wall failures could be totally eliminated and other equipment failures are ignored, the averted cost-risk would only be $86,873 (seismic PACR

• block wall F-V =

averted cost-risk: $4,455,031,

• 1.95E-02 = $86,873), which is less than the minimum expected cost of implementation of $100,000. If the 95th percentile probabilistic risk assessment results are considered, the averted cost-risk could be scaled to $205,020.

The scope of the changes that would be required to enhance the block walls, however, would preclude the enhancements from being cost beneficial. The block walls in question, which extend from the floor at elevation 119 ft. to the underside of the floor at elevation 140 ft., were strengthened significantly as part of the blockwall modification program (circa 1993). The anchorages to floor and ceiling were strengthened and steel columns were added on the faces of walls to increase out-of-plane strength and stiffness. There is little room for further changes to the current structure. It may be possible to increase the capacity of these walls through the addition of more steel columns or braces, but limitations on space may require them to be replaced with an integral floor to ceiling, reinforced concrete design. In either case, the cost of changes would be significantly larger than $205,020.

PG&E Letter DCL-10-106 Page 80 of 131 RAI 5.c

5. Provide the following information with regard to the SAMA selection and screening process:

c. The NRC Staff Evaluation Report (SER) on the DCPP IPE identifies three potential plant improvements (p. 15 of the June 30, 1993 SER). Two items were not addressed in Section F. 5.1.4 - modification of reactor coolant drain tank (RCD T) door to flood the reactor cavity and incorporating insights from the SGTR results into severe accident management. Provide the implementation status of these items and a Phase 1/1l evaluation of any not implemented or resolved by other means.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 5.c Insights from the steam generator tube rupture results have been incorporated into the Diablo Canyon Power Plant Severe Accident Management Guidelines (SAMGs) and this issue is considered to be closed out.

The proposed improvement to the reactor coolant drain tank (RCDT) door was implemented during the containment recirculation sump modifications of 2007 (Unit 1) and 2008 (Unit 2). The RCDT compartment doors were modified to include a swing flap at the bottom of the door to let water flow more freely into the cavity under the reactor vessel.

PG&E Letter DCL-1 0-106 Page 81 of 131 RAI 5.d

5. Provide the following information with regard to the SAMA selection and screening process:

d. IPE Section 6.1 and IPEEE Section 7.1 identify a number of plant improvements that, although not PRA-related, were considered to have a beneficial impact on the PRA results. Provide the implementation status of these items and a Phase 1/1l evaluation of any not implemented or resolved by other means.

PG&E Response to Severe Accident Mitigation Alternatives RAI 5.d All of the plant improvements from Section 6.1 of the individual plant examination and 7.1 of the Individual Plant Examination of External Events that were identified as "not probabilistic risk assessment-related" have been implemented at Diablo Canyon Power Plant and no further evaluation is required: The improvements include:

Installation of a sixth emergency diesel generator Installation of the anticipated transient without scram mitigating system actuation system Digital feedwater control upgrade Elimination of the boron injection tank 480V switchgear room ventilation enhancement

'Component cooling water abnormal operating procedure enhancements Eagle 21 process protection system upgrade and elimination of the resistance temperature detector, Instrument inverter replacement PG&E Letter DCL-1 0-106 Page 82 of 131 RAI 5.e

5. Provide the following information with regard to the SAMA selection and screening process:

e. Table F.5-3 identifies SAMA 21, "Provide a Portable Air Compressor to Pressurize IA Header," as a Phase I SAMA and dispositions it as not used. This SAMA is not discussed elsewhere in the ER. Furthermore, SAMA 21 appears to be an alternative to SAMA 9 that would mitigate a broader set of basic events.

Clarify-how SAMA 21 was identified and justify why it should not be further evaluated.

PG&E Response to Severe Accident Mitigation Alternatives RAI 5.e SAMA 21 was originally developed as an alternative to SAMA 9 for addressing seismically induced loss of Instrument Air (IA) to the Pressure Operated Relief Valves (PORVs). However, it was determined that most of the IA system piping is not seismically qualified and cannot be assumed to be available during a seismic event. This would preclude using a portable compressor to pressurize the IA'header to support PORV operation. Because the SAMA 9 design supports PORV operation through the portions of the IA lines that are seismically qualified, it was considered to be a more appropriate means of addressing seismically induced IA failures.

Consideration was given to retaining SAMA 21 for nonseismic events; however, the cost of implementation for SAMA 21 was estimated to be about the same as that for SAMA 9. Given the reduced capability of SAMA 21,, relative to SAMA 9, there was no need to retain SAMA 21 and it was eliminated from further consideration.

PG&E Letter DCL-10-106 Page 83 of 131 RAI 6.a

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

a. Section F. 6 states that plant personnel developed DCPP-specific implementation cost estimates for each of the SAMAs. Provide a description of:

the process PG&E used to develop the SAMA implementation costs, the level of detail used to develop the cost estimates (e.g., general cost categories such as hardware design, procurement, installation, and testing, as well as procedure development, quality assurance and licensing support, etc.), and how the calculations are documented. Provide the details of the cost estimates for SAMAs 3, 8, 10, 11, 12, and 14.

,PG&E Response to Severe Accident Mitigation Alternatives RAI 6.a A cost estimate was developed in order to determine the implementation cost for each potential SAMA on the Phase I SAMA list. The process for developing the

-cost estimate is detailed below.

The SAMA cost estimates were created by an independent consultant who previously worked with PG&E at Diablo Canyon Canyon Power Plant (DCPP) as a director of engineering, operations, and maintenance. The consultant previously held a senior reactor operator license at DCPP. The consultant determined the scope of the SAMA including identifying the changes and modifications needed to mitigate the accident as described later in the report.

A cost estimator for DCPP Strategic Projects developed the cost for each potential SAMA based on any procedural changes, hardware, engineering, training, or simulator modifications.

Cost input was received from the civil and electrical planning supervisor. The cost estimates were reviewed by the project management office supervisor and the maintenance planning manager.

Upon receiving a description of each potential SAMA on the Phase I list from ERIN Engineering and Research Incorporated, DCPP evaluated each SAMA to determine the changes and modifications needed to mitigate the accident. As part of understanding each SAMA, DCPP challenged ERIN's logic and attempted to find alternative methods of mitigating the accident.

Once the SAMA was fully understood, DCPP determined if procedures and training alone would resolve the SAMA issue. If this was the case, costs were estimated for procedure revisions and/or emergency operating procedure (EOP)

PG&E Letter DCL-1 0-106 Page 84 of 131 revisions, new specific training material preparation, and incremental training costs. Effective p rocedure changes would have clearly established existing entry conditions (EOP E-O and subsequent EOPs, annunciator response procedures, or abnormal operating procedures (AOPs)). Examples of SAMAs that could be addressed with only procedures and training are SAMA 24, Prevent Clearing of Reactor Coolant System (RCS) Cold Leg Water Seals; and SAMA 25, Fill or Maintain Filled The Steam Generators to Scrub Fission Products.

If procedures changes alone did not resolve the SAMA issues, then hardware changes were evaluated. The first step in evaluating the potential hardware changes was to determine if passive modifications would resolve the issue. If passive modifications would resolve the issue then costs were estimated for those hardware changes. No procedures or training would be part of the cost estimate. Examples of SAMAs that could be resolved with only passive modifications included SAMA 12, Improve Fire Barriers for Auxiliary Salt Water (ASW) and Component Cooling Water (CCW) Equipment in the Cable Spreading Room; SAMA 13, Improve Cable Wrap for the Pressure Operated Relief Valves (PORVs) in the Cable Spreading Room; and-SAMA 23, Reinforce steam generator (SG) and Associated RCS Piping Supports.

If only passive hardware changes did not resolve the issue, the next step was to determine if passive and active modifications (meaning no specific operator actions are required but the equipment is active) would resolve the issue. If passive and active modifications would resolve the issue, costs were estimated for changes. Costs would include Engineering, hardware, hardware testing for qualification, installation, and revised surveillance tests/maintenance procedures.

No costs would be needed for EOPs/AOPs, simulator modifications, or training since the plant configuration would not change for these programs. Examples of SAMAs that could be resolved with passive and active modifications included SAMA 7, Replace or Modify the Battery Chargers to Operate Without the Batteries; SAMA 8, Install an Additional Train of Switchgear Room HVAC; SAMA 10, Install High Temperature RCP Seals; SAMA 19, Replace Critical Relays with High Seismic Capacity Relays; and SAMA 9, Backup Air System for PORV PCV 474.

If passive and active modifications alone would not resolve the issue, the next step was to determine if fully active equipment capability (operator actions are needed to implement or monitor the new equipment) needs to be installed to achieve the SAMA goal. Two versions of this exist: fully automatic and manual, implementation of the alternate equipment. Response time of operators and inclusion of operator error reduction needs are used to determine the extent of automation.

PG&E Letter DCL-1 0-106 Page 85 of 131 Fully automatic: Assume input instrumentation will be upgraded to multichannel to preclude single failure, inadvertent actuation. Consider improved accuracy of the automated input signal sensors to prevent early or late actuation. All logic control boxes are redundant, Class 1, seismic.,,

Class 1 power supplies are generally needed. Costs will include Engineering, hardware purchase, proof testing, installation, modifications to the plant simulator, new surveillance tests, ongoing testing for the automation hardware, revised emergency procedures, and incremental operator training.

Examples of SAMAs requiring fully automatic active equipment include SAMA 2, Automate Swap to Recirculation; SAMA 14, Fully Automate Feed and Bleed Initiation; SAMA 16, Install Automatic Suppression In Vertical Board 4 of the Main Control Room (MCR), and SAMA 11, Install Containment Combustible Gas Igniters.

2.

Manual initiation of alternate equipment: Assume operator action is needed in 30 minutes or less. This time response is a critical input specification that precludes extensive alignment of portable equipment and often may require actuation from the control room. Local operations need to be minimized since response time and reduction of operator errors is important to be successful in the SAMA. Cables, hoses, piping, valves, manual switches, and fuel supplies all need to be permanently installed to minimize errors and installation time. Fuel and piping systems need to be "wet" up to the equipment interface to reduce priming time and miss-operation due to air ingestion. Operator actions in the field should be limited to opening or closing a few manual valves, starting a diesel generator or pump, and throwing a few manual switches. Considerations would include seismic capacity, Class 1/Class 2 interfaces between the alternate source and existing equipment, station blackout conditions, local manual or remote manual actuation. Costs would include engineering, hardware, installation, procedures (surveillance tests and operator procedures), simulator modifications, operator training, and on-going testing/maintenance of the alternate systems.

Examples of SAMAs requiring manual initiation of fully automatic equipment are SAMA 1, Install Primary Side Steam Generator (SG)

Isolation Valves; SAMA 3, Alternate Direct Current (DC) Generator; SAMA 4, Install a Self-Contained Swing Emergency Diesel Generator (EDG); SAMA 5, Use an Alternate EDG to Support Long Term Auxiliary Feedwater (AFW) Operation and a 480V Alternating Current (AC) Self-Cooled Positive Displacement Pump (PDP) for Primary Side Makeup; SAMA 6, Use Alternate Engine-Driven High Pressure Pump for Secondary Side Makeup; SAMA 15, Provide Hard Piped Connection between Fire PG&E Letter DCL-1 0-106 Page 86 of 131 Water and the Charging Pump Lube Oil Coolers and Remotely-Operate'd Motor Operated Valves MOVs); SAMA 17 Install Alternate Power Connections to Centrifugal Charging Pump (CCP) 12; and SAMA 18,'

Seismically Qualified Alternate 480V AC EDG to Support Long Term AFW Operation and a Seismically Qualified 480V AC Self-Cooled PDP for RCS Makeup.

Once the scope of each SAMA was determined, a cost estimate for each potential SAMA was developed based on any procedural changes, hardware, engineering, training, or simulator modifications. The cost estimates were prepared by the plant cost estimating group with input from the civil & electrical planning supervisor and reviewed by the project management office supervisor and maintenance planning manager. The cost estimates were done in 2009 dollars and included contingency costs and capital overhead. Cost estimates did not include escalation or allowance for funds used during construction. Cost estimates from past projects were used when applicable.

DCPP has recently completed many projects and the actual costs of these projects were used to develop the SAMA cost estimates. Recent DCPP projects include upgrading the battery chargers, PDP replacement with a CCP, SG replacement, digital-upgrades to turbines, feedwater (FW) pumps, and FW controls. Recent project costs were used to determine the cost for SAMA 4, Install a Self-Contained Swing EDG; and SAMA 7, Replace or Modify the Battery Chargers to Operate Without the Batteries.

For cost estimates that were not based directly off of past projects, itemized cost estimates were developed to include hardware, procedure changes, engineering, training, or simulator modifications. Hardware costs were estimated, based on the cost estimate for the hardware identified during the scope determination phase for each SAMA. Specific hardware costs from recent projects such as piping, valves, electrical cable, and -switchgear were used when applicable.

Procedure change costs were developed based on man-hours for all the processing needed to update the procedures. Engineering estimates were based on typical man-hours costs for design changes. Training costs were developed based on the man-hours needed to prepare operator training materials. Simulator modification costs were based on the recent costs to modify the simulator.

Costs estimates for the requested SAMAs are provided on the following pages.

PG&E Letter DCL-10-106 Page 87 of 131 SAMA 3, Failure of DC Control Power

==

Description:==

Provide a portable 120VDC diesel generator set to power 120VDC loads 0

S S

0 100kW DG set + fuel tank for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> Seismic/weather proof location Run on 85 ft elevation Supply cables and manual connection switches at each DC distribution panel(3 per unit)

Description Qty.

Unit Unit $

Total Engineering 1

LS

$ 1,000,000 1,000,000 Procedure Revisions 1

LS 100,000 100,000 100kW DG Set + Fuel Tank for 24 1

Hours LS 450,000 450,000 Conduit & Wire to DC Busses 1,000 LF 1,000 1,000,000 Manual Connection Switches 3

Ea 75,000 225,000 Switches to DC Panel Installations 3

Ea 25,000 75,000 Per Weatherproof Enclosure 0.50 Unit

$ 1,500,000 750,000 Indicators in the Control Room 1

LS 600,000 600,000 New Test Procedures 1

LS 175,000 175,000 Per Simulator Modifications 0.50 Unit 250,000 125,000 Per Revise Operator Training Program 0.50 Unit 75,000 37,500 Sub-Total 4,537,500 Sub-Total per Unit X 2 Units 2

Units

$ 4,537,500 9,075,000 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) +

(70% Lab X 40% PGE Lab X 14.3%

Capital A&G) 7.68%

696,960 20.00 Contingency 1,954,392 Total

$ 11,726,352 All costs in 2009 dollars. No escalation or AFUDC has been included PG&E Letter DCL-10-106 Page 88 of 131 SAMA 8, Loss of 480V switchgear ventilation

==

Description:==

Adds a new set of ventilation supply and exhaust fans in the 480V vital bus rooms Add 1 supply and 1 exhaust fan per rm. Class 1, seismic, weatherproof Locate on 165 ft. elevation above control room Vital 480 power. New breakers and wiring. New On/Off circuits Automate system to start on high temperature Temperature sensor in each room (3/unit)

Logic control box in switchgear room Description Qty.

Unit Unit $

Total Engineering 1

LS

$ 450,000 450,000 5,000 CFM Supply Fan 1

Ea 75,000 75,000 5,000 CFM Return Fan 1

Ea 75,000 75,000 New Ductwork 1

LS 4

250,000 250,000 Seismic Bracing 1

LS 75,000

'75,000 Electric Room Penetrations 2

Ea 20,000 40,000 Tie-Ins to Existing Ductwork 2

Ea 30,000 60,000 Automatic Controls 1

LS 70,000 70,000 Power and Control Wiring 1

LS 550,000 550,000 Miscellaneous Louvers/Dampers/Etc.

1 LS 50,000 50,000 Sub-Total

$ 1,645,000 Sub-Total per Room X 3 Rooms 3

Rooms

$ 1,645,000

$ 4,935,000 Sub-Total

$ 4,935,000 Sub-Total per Unit X 2 Units 2

Units

$ 4,935,000

$ 9,870,000 PG&E Letter DCL-10-106 I

Page 89 of 131 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) + (70%

Lab X 40% PGE Lab X 14.3% Capital A&G) 7.68%

758,016 Contingency 20.00%

$ 2,125,603 Total

$ 12,753,619 All costs in 2009 dollars. No escalation or AFUDC has been included.

PG&E Letter DCL-10-106 Page 90 of 131 SAMA 10, RCP Seal Cooling

==

Description:==

Install new emergency reactor coolant pump (RCP) seal that seals off the RCP shaft above 300°F.

Westinghouse currently tested a new high temperature seal the fits into a modified Number I seal housing.

Description Qty.

Unit Unit $

Total Engineering - Including Qualification and QA Approval 1

LS 750,000 750,000 Procedure Revisions 1

LS 125,000 125,000 Modified Number 1 Seal Housing 4

Ea 50,000 200,000 New RCP Seal 4

Ea 300,000

$ 1,200,000 Spare RCP Seal + Housing 1

Ea 350,000 350,000 Installation Complete (No Maintenance Outage Window (MOW))

4 Ea 550,000

$ 2,200,000 Sub-Total

$ 4,825,000 Sub-Total per Unit X 2 Units 2

Units

$ 4,825,000

$ 9,650,000 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) +

(70% Lab X 40% PGE Lab X 14.3%

Capital A&G) 7.68%

741,120 Contingency 20.00%

$ 2,078,224 Total

$ 12,469,344 All costs in 2009 dollars. No escalation or AFUDC has been included.

Based on NRC Approval of Westinghouse RCP Seals.

PG&E Letter DCL-10-106 Page 91 of 131 SAMA 10, Alt A 0

0 0

0 0

0 0

Size: 200 hp for charging pump, 75 hp for battery charger New 480VAC charging pump Run on 85 ft. elevation Cables to all 6 battery chargers Seismic 24-hour fuel supply (estimate for 400-500 gallons)

Swithches to connect EDGto any battery charger.

Description Qty.

Unit Unit $

Total Engineering 1

LS

$ 1,500,000

$ 1,500,000 Procedure Revisions 1

LS 200,000 200,000 100kW DG Set + Fuel Tank for 24 Hours 1

LS

$ 450,000 450,000 Conduit & Wire to DC Busses 1,000 LF 1,000

$ 1,000,000 Manual Connection Switches 3

Ea 75,000 225,000 Switches to DC Panel Installations 3

Ea 25,000 75,000 300kW DG Set + Fuel Tank for 24 Hours 2

LS

$ 650,000

$ 1,300,000 Conduit & Wire to'DC Busses 1,200 LF 1,000

$ 1,200,000 Manual Connection Switches 3

Ea 75,000 225,000 Switches to DC Panel Installations 3

Ea 25,000 75,000 Weatherproof Enclosure 1

Ea

$1,500,000

$ 1,500,000 200 HP 480V Positive Displacement Charging Pumps 2

Ea

$ 450,000 900,000 Suction Piping 3,000 LF 600

$ 1,800,000 Discharge Piping 600 LF 600 360,000 Indicators in the Control Room 1

LS

$ 300,000 300,000 PG&E Letter DCL-10-106 Page 92 of 131 New Test Trocedures 1

LS

$ 250,000 250,000 Rev Operator Training Program 1

Ea 75,000 75,000 Sub-Total

$ 11,435,000 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) +

(70% Lab X 40% PGE Lab X 14.3%

Capital A&G) 7.68%

878,208 Contingency 20.00%

$ 2,462,642 Total

$ 14,775,850 All costs in 2009 collars. No escalation or AFUDC has been included.

PG&E Letter DCL-10-106 Page 93 of 131 SAMA 11, Containment Failure

==

Description:==

Install automatic DC powered igniter system. Costs similar to hydrogen recombiners installed in 1981.

Inerting containment is not practical.

Description Qty.

Unit Unit $

Total Engineering 1

LS

$ 1,000,000

$ 1,000,000 Procedure Revisions 1

LS 100,000 100,000 Hydrogen Igniters 5

Ea 50,000 250,000 Hydrogen Sensors 30 Ea 5,000 150,000 Power Supply Hardware 1

LS 100,000 100,000

.2 Controls 1

LS 250,000 250,000 Conduit & Wire To Containment Dome Class 1 5

Ea 200,000

$ 1,000,000 Scaffolding to Containment Dome N

(Up & Down) Incl. Mat'l Handling in and out of Containment I

LS 750,000 750,000 Sub-Total

$ 3,600,000 Sub-Total per Unit X 2 Units 2

Units

$ 3,600,000

$ 7,200,000 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) +

(70% Lab X 40% PGE Lab X 14.3%

Capital A&G) 7.68%

552,960 Contingency 20.00%

$ 1,550,592 Total

$ 9,303,552 All costs in 2009 dollars. No escalation or AFUDC has been included.

PG&E Letter DCL-1 0-106 Page 94 of 131 SAMA 12, Cable Spreading Room failure affects ASW/CCW Pump Controls

==

Description:==

This project improves fire protection cabling to ASW pump and CCW pump controls for each unit.

2 ASW Pumps per unit 3 CCW Pumps per unit Description Qty.

Unit Unit $

Total Engineering 1

LS 500,000 500,000 Remove & Reinstall Interferences 1

Ea 50,000 50,000 Wrap Vital Conduits in Cable Spreading Room With Fire Protection Wrap 200 LF 250 50,000 Sub-Total 600,000 Sub-Total per Unit X 2 Units 2

Units 600,000

$ 1,200,000 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) +

(70% Lab X 40% PGE Lab X 14.3%

Capital A&G) 7.68%

92,160 Contingency 20.00%

258,432 Total

$ 1,550,592 All Costs in 2009 Dollars, No Escalation or AFUDC has been included.

PG&E Letter DCL-1 0-106 Page 95 of 131 SAMA 14, Fully Automate Feed and Bleed

==

Description:==

Automate Steps 11, 12, 13, 16, and 17 in EOP FR-H.1 Install redundant, seismic, Class 1 logic boxes in cable spreading room.

Run new signal wires for:

S/G level, wide range RCS pressure/PZR level AFW flow PORV position switches Charging flow 0

Run output wiring to control:

Charging pumps Charging flow control PORV open/close New procedure, tests, etc.

Description Qty.

Unit Unit $

Total Engineering 1

LS

$ 2,000,000

$ 2,000,000 Procedure Revisions 1

LS 200,000 200,000 New PLC's 2

Ea 500,000 1,000,000 Controls - Conduit & Cable 1

LS

$ 2,000,000

$ 2,000,000 Terminations 400 Ea 1,000 400,000 Programming 1

LS

$ 2,000,000

$ 2,000,000 Testing 1

LS

$ 1,000,000

$ 1,000,000 Per Revise Operator Training Program 0.50 Unit 500,000 250,000 Sub-Total

$ 8,850,000 Sub-Total per Unit X 2 Units 2

Units

$ 8,850,000

$ 17,700,000 PG&E Letter DCL-10-106 Page 96 of 131 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) + (70%

Lab X 40% PGE Lab X 14.3% Capital A&G) 7.68%

1,359,360 Contingency 20.00%

3,811,872 Total

$ 22,871,232 All costs in 2009 dollars. No escalation or AFUDC has been included.

PG&E Letter DCL-1 0-106 Page 97 of 131 RAI 6.b

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

b. The discussion in Section F. 6 indicates that the Phase 2 benefit analysis utilized the sum of the release category frequencies (8.44E-06 per year) instead of the Level I CDF value of 8.4 7E-06 per year, and that the release category results did not include the contribution from the negligible (intact) release category. The Level 1 CDF is given in Section F.2.1.9 as 8.13E-06 per year and a negligible release category is not discussed elsewhere. Clarify this discrepancy.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.b The language in Section F.6 related to the "negligible release category" is related to Release Category ST4A (BYPASS w AFW - 5 hr GE), which always has a frequency of 0.0 for DCPP. To avoid any confusion associated with reporting a release category with a frequency of 0.0, an attempt was made to remove all references to this release category from the SAMA documentation; however, the following sentence in Section F.6 was overlooked during this effort:

"Furthermore, the release category results provided for each SAMA do not include contributions from the negligible release category." The decision to exclude Release Category ST4A from the results is not related to the difference in the core damage frequency (CDF) and the sum of the release category frequencies.

The difference between the sum of the release category frequencies for Model DC01A (8.44E-06/yr) and the DC01A CDF (8.47E-06/yr) is due to rounding.

The CDF of 8.13E-06/yr that is provided in Section F.2.1.9 of the Environmental Report should have been reported as 8.47E-06/yr, as indicated in the response to RAI 1.j.

PG&E Letter DCL-10-106 Page 98 of 131 RAI 6.c

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

c. In each of the subsections in Section F. 6, there is a discussion of assumptions made in evaluating the respective SAMA. Most of these are described in terms of the split fractions of the PRA and can be confusing without recourse to the details of the split fraction definitions and usage. Provide a brief description of the model changes for each SAMA in layman's terms without recourse to PRA jargon. Furthermore, the discussion describes the development of the split fractions used for the SAMA analysis but does not provide a corresponding description of the baseline split fractions. Provide a description of the baseline split fractions that are being changed in each SAMA evaluation.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 6.c The following is a brief description of the model changes made to the model to quantify the SAMA impact. Also included is the base model description of the split fraction, if any, that was manipulated to help quantify the SAMA impact.

SAMA 2 Description of model changes - This SAMA models replacing the operator action to switchover to sump recirculation to an automated system. To model this, the operator action failure probability (in top event RF) was replaced with a system failure probability. The system failure probability is based on a 2-train Solid State Protection System (SSPS) approximated from a Diablo Canyon Power Plant (DCPP) model split fraction. The following are the base model split fractions for top event RF that were modified to affect the hardware change.

RE1 - Recirculation switchover after small loss-of-coolant-accident (LOCA) or bleed and feed with core spray failed RF3 - Recirculation switchover after medium or large LOCA RF4 - Recirculation switchover to recirculation after core melt RFF - Recirculation switchover is guaranteed to fail due to boundary conditions SAMA 3 Description of model changes - This SAMA models the addition of another train of DC power. This change is made through a direct current (DC) bus recovery human action (split fraction). The screening Human Error Probability (HEP) value (in split fraction REBAT) was replaced with a new human action value plus a hardware term in order to model the manual alignment to this proposed backup train. The probabilities are based on current DCPP split fractions.

PG&E Letter DCL-1 0-106 Page 99 of 131 REBAT - Recovery from loss of all DC buses with offsite power and vital alternating current (AC) buses available SAMA 5 Description of model changes - This SAMA models the addition of an emergency diesel generator (EDG) to supply AFW level control and to supply 480V AC power for an additional charging pump. The auxiliary feedwater (AFW) and charging pump split fractions were altered to model the impact of these changes. For AFW during blackout conditions the turbine driven pump is lost if level indication is not available. Hence, the split fractions associated with these conditions were changed. As for the charging pump, the associated split fraction values were reduced to mimic the presence of an additional train.

The description of the AFW base model split fractions changed for this SAMA are already listed in Section F.6.3. These-are the charging pump split fractions:

CH1

- TWO TRAINS ALL SUPPORT AVAILABLE CH1A-PUMP 12 UNAVAILABLE CH1B - PUMP 11 UNAVAILABLE CH2 - SUPPORT TO CCP1/TRAIN A UNAVAILABLE CH3 - SUPPORT TO CCP2/TRAIN B UNAVAILABLE CH4 - 480V F UNAVAILABLE / Rest Support Available CH4A - PUMP 12 (TRAIN B) AND 480V F (TRAIN A) UNAVAILABLE CH5 - 480V G UNAVAILABLE / Rest Support Available CH5A - PUMPA 11 (TRAIN A) AND 480V G (TRAIN B) UNAVAIL CH6 - LOSS OF OFFSITE POWER (LOOP)

CH6A - PUMP 12 UNAVAIL - LOOP CH6B - PUMP 11 UNAVAIL - LOOP CH7 - SUPPORT TO CCP1/TRAIN A UNAVAILABLE - LOOP CH8 - SUPPORT TO CCP2/TRAIN B UNAVAILABLE - LOOP CH9 - 480V F UNAVAILABLE / Rest Support Available - LOOP CH9A - PUMP 12 (TRAIN B) AND 480V F (TRAIN A) UNAVAIL - LOOP CHA - 480V G UNAVAILABLE / Rest Support Available - LOOP CHAA - PUMP 11 (TRAIN A) AND 480V G (TRAIN B) UNAVAIL - LOOP SAMA 7 Description of model changes - This SAMA models an improved battery charger that can carry the DC bus loads without the battery. It is affected by removing the dependency on the batteries in the model (in the split fraction rules logic).

No split fractions are affected.

PG&E Letter DCL-10-106 Page 100 of 131 SAMA 8 Description of modei changes - This SAMA models the addition of a redundant train of switchgear HVAC. The new redundant train probability is substituted for the HVAC human action recovery term (in top event RE) in the model. The redundant train probability is taken from a DCPP ventilation system analysis.

RE6 - Recovery of Switchgear Ventilation for Loss of Switchgear Initiator RE6A - Recovery of Switchgear Ventilation for System Failures SAMA 9 Description of model changes - This SAMA models the addition of Backup N2 to bolster the performance of the Power Operated Relief Valve (PORV) PCV-474 during feed and bleed scenarios. The failure probabilities associated with feed and bleed are dominated by the operator action to initiate feed and bleed, but there is a hardware component modeled. The hardware component was improved and the associated total failure probabilities reduced.

The effected base model split fraction descriptions are already listed in Section F.6.6.

SAMA 10 Description of model changes - This SAMA addresses the installation of high temperature reactor coolant pump (RCP) seals. The RCP seal LOCA failure probabilities associated (found in top event SE) with loss of all seal cooling were lowered to reflect new high temperature seals. Engineering judgment was used to assign seal LOCA probabilities for the new high temperature seals.

The effected base model split fraction descriptions are already listed in Section F.6.7.

SAMA 11 Description of model changes - This SAMA addresses the important large early release frequency (LERF) impact of a containment hydrogen burn. Containment failure due to hydrogen burn could be mitigated by the addition of a battery powered hydrogen igniter. The containment failure probability and the likelihood of a hydrogen burn were reduced to reflect the failure probability of a proposed hydrogen igniter system.

CECET1 - Probability of containment failure after hydrogen burn HECET1 - Hydrogen burn within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> of vessel breach HECET2 - Hydrogen burn within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> of vessel breach PG&E Letter DCL-10-106 Page 101 of 131 SAMA 12 Description of model changes - This SAMA improves the fire retardancy of the cable spreading room cable wrap. It is implemented by reducing the frequency of a cable spreading room fire initiating event (initiating event CSR1).

Engineering judgment is used to arrive at'a factor of 100 reduction in initiating event frequency for this SAMA.

CSR1 - Cable spreading room fire, loss of auxiliary salt water, and component cooling water SAMA 13 Description of model changes - This SAMA improves the fire retardancy of the PORV cable wrap in the cable spreading room. It is implemented by reducing the frequency of a cable spreading room fire initiating event (initiating event CSR2) that results in a PORV LOCA. Engineering judgment is used to arrive at a factor of 100 reduction in initiating event frequency for this SAMA.

CSR2 - Cable spreading room fire, PORV induced LOCA SAMA 16 Description of model changes - This SAMA models the addition of a fire suppression system in Control Room Vertical Board 4. It is implemented by reducing the failure to extinguish a Vertical Board 4 fire (see split fraction FEF2) by a factor of 100 (see EPRI report 1019189, Fire PRA Method Enhancements Additions Clarifications, and Refinements to EPRI 1019189, Section 3.7 for probability source).

FEF2 - Control Room vertical board 4 geometry / severity factor SAMA 17 Description of model changes - This SAMA models the addition of alternate AC power connections to Charging Pump 1-2 so that 2 redundant power trains are available. It is implemented by reducing the failure probability of charging (top event CH). The reduction factor is based on the comparison of 2 charging pump system analysis results (split fractions CH1 and CH1A).

The affected split fractions and base model descriptions are listed in Section F.6.12.

SAMA 18 Description of model changes - This SAMA represents installing arseismically robust, stand alone 480V AC power source that could provide power to the AFW level instrumentation as well as powering an additional high pressure injection (positive displacement) pump. System failure rates for AFW (top event TD and PG&E Letter DCL-10-106 Page 102 of 131 AW) and charging (top event CH) were modified to account for these new features. The redundant positive displacement pump (PDP) failure rate probability was set at 1.OE-02 and for AFW the existing system failure rates were reassigned. The modified split fractions are as follows:

AW7 - SUPPORT FOR ALL 10 PERCENT STM DMPS AND THE TDP UNAVAILABLE AW7B - SAME AS AW7 - STEAM LINE BREAKS AW7L - SAME AS AW7 - RCP'S TRIPPED AND NATURAL CIRCULATION MODE AWAA- 0 SUPPORT FOR 10 PERCENT STM DMPS/TDP/MDP 1-2 AWAAB -

SAME AS AWAA - STEAM LINE BREAKS AWAAL - SAME AS AWAA - RCP'S TRIPPED AND NATURAL CIRCULATION MODE AWAB - NO SUPPORT FOR 10 PERCENT STM DMPS/TDP/MDP 1-3 AWABB - SAME AS AWAB - STEAM LINE BREAKS AWABL - SAME AS AWAB - RCP'S TRIPPED AND NATURAL CIRCULATION MODE AWD - SUPPORT FOR THE TDP UNAVAILABLE AWDB -

SAME AS AWD - STEAM LINE BREAKS AWDL - SAME AS AWD - RCP'S TRIPPED AND NATURAL CIRCULATION MODE AWF - GUARANTEED FAILURE CH1

-TWO TRAINS ALL SUPPORT AVAILABLE CH1A - PUMP 12 UNAVAILABLE CH1B - PUMP 11 UNAVAILABLE CH2 - SUPPORT TO CCP1/TRAIN A UNAVAILABLE CH3 - SUPPORT TO CCP2/TRAIN B UNAVAILABLE CH4 - 480V F UNAVAILABLE / Rest Support Available CH4A - PUMP 12 (TRAIN B) AND 480 F (TRAIN A) UNAVAILABLE CH5 - 480V G UNAVAILABLE / Rest Support Available CH5A - PUMPA 11 (TRAIN A) AND 480 G (TRAIN B) UNAVAIL CH6 - LOOP (LOOP)

CH6A-PUMP 12 UNAVAIL - LOOP CH6B-PUMP 11 UNAVAIL-LOOP CH7 - SUPPORT TO CCP1/TRAIN A UNAVAILABLE - LOOP CH8 - SUPPORT TO CCP2/TRAIN B UNAVAILABLE - LOOP CH9 - 480V F UNAVAILABLE / Rest Support Available - LOOP CH9A - PUMP 12 (TRAIN B) AND 480V F (TRAIN A) UNAVAIL - LOOP CHA - 480V G UNAVAILABLE / Rest Support Available - LOOP CHAA - PUMP 11 (TRAIN A) AND 480V G (TRAIN B) UNAVAIL - LOOP TD1

- SEISMIC TD AFW PUMP - SCT=F TD2

- SEISMIC TD AFW PUMP - SCT=F & AN I.C.=F TDF - SEISMIC TD AFW PUMP - SCT=F*DG=F SAMA 22 Description of model changes - This SAMA models the inclusion of additional AFW start signals that are not credited in the PRA model. Currently, SSPS is credited as the main start signal source with backup coming from anticipated transient without scram mitigating system actuation (AMSAC). An additional PG&E Letter DCL-10-106 Page 103 of 131 AMSAC-like signal source, with equivalent failure probability was added to the model to credit these signals. The credit was added to the existing SSPS system.

The affected split fractions and their base model description are listed below Split Fraction

Description:

SB1 - SA is successful, general transient SB2 - SA fails, general transient SB3 - SA unavailable due to support, general transient SB4 - SA is successful, large loss-of-coolant-accident (LLOCA)

SB6 - SA fails, LLOCA SB9 - SA is successful, steam generator tube rupture (SGTR)

SBA - SA fails, SGTR SBF - Guaranteed failure SBI - SA is successful, steam line break outside containment (SLBO)

SBL - SA is successful, small loss-of-coolant-accident (SLOCA)

SBM - SA fails, SLOCA SAMA 24 Description of model changes - This SAMA models the procedural change that would preclude operators from clearing the reactor coolant system (RCS) cold legs after core damage occurs. This is addressed in the model by reducing the probability of thermally induced steam generator tube rupture (SGTR)by a factor of 10.

ISCET1 - Induced tube rupture probability, no induced PORV failure, with RCP cooling ISCET2 - Induced tube rupture probability, no induced PORV failure, with no RCP cooling ISCET3 - Induced tube rupture probability, no induced PORV failure, with no RCP cooling SAMA 25 Description of model changes - No model changes were' made. This SAMA suggests a procedure change that directs operators to fill or maintain filled the steam generators (SGs) just prior to core damage to provide mechanical scrubbing of fission products. To evaluate this, the LERF model sequences (cutsets) were recalculated to account for a new human action crediting the filling of the SGs. If the human action was successful, the scenario frequency was rebinned to small early release frequency (SERF). If the human action failed it remained a LERF sequence.

No split fractions were changed.

PG&E Letter DCL-10-106 Page 104 of 131 RAI 6.d

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

d. In each of the subsections in Section F. 6, split fractions and top events are listed under a heading "PRA Model Changes to Model SAMA," followed by a

-discussion of model changes. In most cases the discussion describes the changes made to the items listed. In some cases the description doesn't include all the split fractions listed or includes a discussion of other spit fractions without an explanation. See for example the discussions in Sections F. 6.2 and F. 6.3.

Without a detailed knowledge of the composition and relationship of the split fractions in the DCPP PRA model it is difficult to fully understand the changes made to the model. Explain the intent and source of the split fractions and top events listed in each section and clarify the discussions as to the changes made and why changes were or were not made to those items listed.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 6.d In general, the split fractions listed in Section F.6 under the section called PRA Model Chanqes to Model SAMA are the split fractions taken from the model importance reports. The most important (risk-reduction worth (RRW)) split fractions to core damage frequency (CDF) from internal initiating events, fire and seismic initiators were considered. These split fractions where then grouped by similar functions and a SAMA was developed to address the failed function.

That is where the list of split fractions for this section came from.

Once the SAMA function was determined, only certain top events and split fractions were logical choices to affect those changes. The following provides more details on the specific split fractions chosen and those left out of the analysis:

SAMA 2 Split fractions RF1, RF3, and RF4 are split fractions directly modeling switchover to sump recirculation. Other RF split fractions have negligible RRW and were not changed for this SAMA evaluation. Top event OR and its split fractions were not addressed because this split fraction addresses cooldown and depressurization mainly for steam generator tube rupture (SGTR), and not sump recirculation directly.

SAMA 3 Other split fractions DF1 and DG3 model loss of a specific direct current (DC).

bus and help show the importance of DC power. This SAMA proposes the addition of another DC bus and the most direct way to model that is through an PG&E Letter DCL-10-106 Page 105 of 131 existing DC bus recovery split fraction. It could have been modeled through these associated top events but it is a more complex model change and altering RE was chosen instead.

SAMA 5 All the split fractions listed here are for loss of 4kV power either through the loss of a diesel generator(s) or LOOP. They show the importance of alternating current (AC) power highlighting the opportunity for an additional AC source to help mitigate accidents. An additional power source could be modeled in these associated top events, but the SAMA is more specific in that it addresses the functions that need the power and those functions are targeted: charging and turbine driven auxiliary feedwater (AFW). Hence, the charging and AFWsplit fractions are specifically addressed.

SAMA 7 These listed split fractions are associated with the DC batteries and DC buses highlighting the opportunity for an additional DC train or more robust charger.

The clearest way to model this SAMA is to remove the dependency on the batteries by removing the logical dependency found in the split fraction rules.

The base model rules fail the associated DC bus when the battery fails. This link was removed for this SAMA.

SAMA 8 These split fractions are for HVAC recovery (split fractions RE6A and RE6), loss of 4kV bus (top events AH and AA) due to ventilation loss, and loss of 480V switchgear ventilation (top event SV). They all point to the opportunity to bolster the switchgear room ventilation. Switchgear room ventilation is modeled in top event SV. The most direct way to model a new redundant train is through HVAC recovery split fractions RE6 and RE6A.

SAMA 9 The split fraction listed here is AWS4, which is a seismic specific split fraction where both motor driven AFW pumps are lost. The scenarios in which this split fraction is associated are where feed and bleed fails because of loss of air to one PORV and the other PORV loses power. This scenario is then addressed by modeling Backup N2 to PORV-474. The best way to address this feature is in the feed and bleed split fractions (top event OB) where loss of instrument air is a split fraction boundary condition that no longer would apply.

SAMA 10 Split fraction RF1 models switchover to containment sump after a seal loss-of-coolant-accident (LOCA). If the likelihood of a seal LOCA could be reduced by a SAMA installing new high temperature reactor coolant pump (RCP) seals, then importance of containment sump switchover would also be reduced. Since it is PG&E Letter DCL-10-106 Page 106 of 131 the likelihood of a seal LOCA that is being changed, the seal LOCA top event was altered to incorporate the proposed SAMA.

SAMA 11 Both split fractions that are listed are addressed by the model changes. See Section F.6.7.

SAMA 12 Split fraction FEF6 is associated with the cable spreading room geometry and FRE3 is the fire specific human action probability. Both of these split fractions are part of the cable spreading room fire analysis. To address the impact of improved wrap on the cable spreading room cables, the fire initiating event frequency (CSR1) was reduced as it is most likely to change if the cables are rewrapped.

SAMA 13 The split fractions listed in Section F.6.10 all help to develop the final frequency of a cable spreading room fire that results in a power operated relief vavle (PORV) LOCA. Each split fraction is one facet of the scenario probability (See Section F.6.10 for full list of split fractions and descriptions). The SAMA could be implemented by reducing the probability of one of these contributing split fractions, but since the PORV cable wrap is being improved the induced PORV LOCA initiating event frequency was changed.

SAMA 16 The split fraction that is listed is discussed and is addressed by the model changes. See Section F.6.11.

SAMA 17 This SAMA addresses split fraction SE9F, which is the LOCA probability where seal injection and thermal barrier cooling are lost. If charging were to remain available this split fraction would become less important. Thus for this SAMA the reliability of charging is improved by addressing top event CH split fractions that have been affected by the loss of one train of power as a boundary condition.

SAMA 18 These split fractions represent seismically induced loss of AC power through loss of 4kV power either through the loss of a diesel generator(s) or LOOP. They show the importance of AC power highlighting the opportunity for an additional seismically qualified AC source to help mitigate accidents. An additional power source could be modeled in these associated top events, but the SAMA is more specific in that it addresses the functions that need the power and those functions are targeted: charging and turbine driven AFW. Hence, the charging PG&E Letter DCL-10-106 Page 107 of 131 and AFW split fractions (and those for seismic scenarios) are specifically addressed. This is similar to SAMA 5, except the seismic capability is added.

SAMA 22 Solid state protection is modeled via top event SB. No split fraction or top events are listed in Section F.6.14, but the most important (RRW) top event SB split fractions are addressed by this SAMA. Since the start of AFW is required for secondary side heat removal post trip, this SAMA addresses that important feature of this SAMA in the solid state protection system (SSPS) split fractions.

SAMA 24 Both split fractions that are listed are addressed by the model changes. See Section F.6.15 SAMA 25 Top events AW and OR are discussed already in Section F.6.16.

PG&E Letter DCL-10-106 Page 108 of 131 RAI 6.e

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

e. The discussion in Section F. 6.2 of modeling the impact of SAMA 3 notes that no credit is taken for seismic or fire initiators. However, the results indicate a large reduction in fire CDF. Explain this result. Provide an assessment of the additional risk reduction from this SAMA in seismic scenarios.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.e The note in Section F.6.2, "There is no credit applied for fire or seismic initiators," was describing the recovery action REBAT was only credited in some internal initiators.

The statement that fire initiators are not affected by the model changes for SAMA 3 is not correct. There should be an effect from the proposed SAMA as seen in the results for fire.

The seismic impact of SAMA 3 was evaluated using the same model changes as used for internal events. The difference is making the changes in the seismic portion of the model event trees. The equivalent location of split fraction REBAT in the seismic portion of the event trees is in event tree RECSEIS (the seismic equivalent of internal events RECV for recovery). A new rule was inserted in this event tree taking credit for split fraction REBAT and using similar logic to internal events to assign it. With credit now given for this SAMA, the new core damage frequency (CDF) is 3.7E-05 or a 2 percent reduction in seismic CDF. The cost benefit analysis results using the DC01B model are provided in the response to RAI 3.c.

PG&E Letter DCL-10-106 Page 109 of 131 RAI 6.f

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

f. The discussion in Section F. 6.4 of modeling the impact of SAMA 7 notes that the assumptions result in an underestimate of the risk benefit. Characterize the magnitude of this underestimate.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.f The discussion of the development of SAMA 7 mentions that because the model's alternating current (AC) power macros were not altered for the analysis that the results may underestimate the risk impact. The underestimate should be small and only have a second order effect. This is because in RISKMAN the macro gets evaluated (is assigned logically True or False)after the top event gets evaluated. Since the changes to the model were made in the top event, which is prior to when any macros get evaluated, the effect of not changing AC power macros should contribute, at most, an order of magnitude less than the SAMA impact itself or less than about 0.5 percent on internal events core damage frequency (CDF). This should have no effect on the conclusions of the SAMA analysis.

PG&E Letter DCL-10-106 Page 110 of 131 RAI 6.,q

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

i

g. The evaluation of SAMA 11, "Install Containment Combustible Gas Igniters,"

in Section F. 6.8 indicates a 68.3 percent reduction in the fire CDF. Explain this significant reduction and why the associated split fractions involved do not appear in the fire CDF importance list.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 6.,q There was a computational error that did not make it to the final report. There is no change in fire core damage frequency (CDF) due to SAMA 11. The

'associated split fractions do not and should not appear in the fire importance list.

PG&E Letter DCL-10-106 Page 111 of 131 RAI 6.h

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

h. For SAMA 13, "Improve Cable Wrap for the Power Operated Relief Valves (PORVs) in the Cable Spreading Room," provide the baseline initiating event frequency for initiator CSR2.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.h The base frequency for initiating event CSR2 is 6.7E-03/yr.

PG&E Letter DCL-10-106 Page 112 of 131 RAI 6.i

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

i. The evaluation of SAMA 17, "Install Alternate Power Connections to Centrifugal Charging Pump (CCP) 1-2, " in Section F. 6.12 indicates a very small reduction in internal events CDF and no reduction in fire CDF, even though this SAMA was identified based on the fire analysis importance review. Split fraction SE9F, representing the probability of a seal loss-of-coolant accident (LOCA) given reactor coolant pumps are tripped and all seal injection and thermal barrier cooling is lost, is indicated in Table F. 5-1b to have a RRW of 1.03 for fires.

Based on this non-negligible contribution, some reduction in fire CDF would be expected. Discuss this evaluation.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.i Split fraction SE9F is assigned for cases where the reactor coolant pumps (RCPs) are tripped and there is a loss of all seal injection and thermal barrier cooling. More specifically, the split fraction is assigned where 4.16kV alternating current (AC) Buses F and G are failed (both buses fail or LOOP and failed emergency diesel generator (EDG)), as these are the conditions where seal injection and thermal barrier cooling would be unavailable. The 4.16kV AC Buses F and G are also the power supplies for the charging pumps. Even though this SAMA provides an alternate power source to CCP 12, split fraction SE9F is assigned in cases where both trains of power are lost. Hence, having an alternate feed from the other train of power will have no effect for the loss of AC power cases where SE9F gets assigned. Although this SAMA was chosen based on split fraction importance, the results show that this SAMA has no effect on these fire related core damage sequences. Even if an independent alternate power source was available to power CCP 12, the 1.03 risk reduction worth (RRW) implies an additional risk reduction of only $47,840 ($1,642,500 -

$1,642,500 1 1.03 = $47,840). At the 95th percentile, the total averted cost-risk for SAMA would be $119,635 (($2,853 + $47,840)

• 2.36 = $119,635), which is far less than the $5.2 million cost of implementation.

PG&E Letter DCL-10-106 Page 113 of 131 RAI 6.i

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

j. SAMAs 5 and 18 are essentially identical and involve use of an alternate EDG to support auxiliary feedwater (AFW) operation and a 480V AC self-cooled positive displacement pump for primary makeup. The only difference is that SAMA 18 is stated to be seismically qualified. However, NRC staff notes the following: (1) the costs of implementation are the same, (2) no seismic failure is incorporated in the SAMA 5 evaluation, (3) the model changes described for each SAMA are different with respect to the impact on AFW availability, and (4) the change in internal events CDF, dose-risk, and OECR are different with the SAMA 5 benefit being larger than the SAMA 18 benefit. It would be expected that SAMA 18 would have the same benefit in internal events but a higher benefit in seismic events, and a higher implementation cost due to the higher seismic capacity for SAMA 18. Clarify the differences between SAMAs 5 and 18 and discuss the differences in modeling and in results.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 6.0 The intention is that SAMA 5 be like SAMA 18, except SAMA 18 has a seismic capability for alternating current (AC) power. In light of the new USGS seismic hazard assessment, SAMA 18 has been reassessed. See RAI 3.c response.

The implementation cost used in the analysis of SAMA 18 is the same as SAMA 5. It is expected that the cost of seismic qualified power for SAMA 18 would be more, but the implementation cost from SAMA 5 was used as a lower bound with no effect on the analysis conclusions.

PG&E Letter DCL-10-106 Page 114 of 131 RAI 6.k

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

k. The estimated per unit cost of $1.7M for SAMA 9, "Backup Air System for PORV Pressure Control Valve (PCV) 474," seems high for what is described as installation of nitrogen bottles to support the PORV only. Justify the cost estimate for SAMA 9.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.k The cost estimate for SAMA 9 was developed per the process described in RAI 6.a. This is a new permanent Class 1 system being added to the plant. The cost estimate is below.

SAMA 9, Loss of Air to Power Operated Relief Valve (PORV) 474

==

Description:==

Provide backup air (nitrogen tanks) to PORV PCV-474. Design similar to systems on PORVs PCV 455C and PCV-456.

Description Qty.

Unit Unit $

Total Engineering 1

LS 500,000 500,000 Procedure Revisions 1

LS 70,000 70,000 800# N2 Bottles 3

Ea 5,000 15,000 N2 Regulator / Manifold - Class 1 1

Ea 100,000 100,000 Solenoid Valves 1

Ea 20,000 20,000 Actuator 1

Ea 20,000 20,000 Relays 1

Ea 20,000 20,000 S/S Tubing 1

LS 250,000 250,000 Check Valve 1

Ea 35,000 35,000 Tie-In to Air System 1

Ea 20,000 20,000 PG&E Letter DCL-10-106 Page 115 of 131 Seismic Bottle Support/Rack 1

Ea 150,000 150,000 Power and Control Wiring 1

LS 60,000 60,000 Surveilance Testing 1

LS 50,000 50,000 Sub-Total

$ 1,310,000 Sub-Total per Unit X 2 Units 2

Units

$ 1,310,000

$ 2,620,000 Capital Overheads (Mat'l 30% X I2.25% Mat'l Burden) + (70% Lab X 40% PGE Lab X 14.3% Capital A&G) 7.68%

201,216 Contingency 20.00%

564,243 Total

$ 3,385,459 All costs in 2009 dollars. No escalation or AFUDC has been included.

PG&E Letter DCL-10-106 Page 116 of 131 RAI 6.1

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

L SAMA 5 provides for a small, alternate EDG to mitigate loss of level instrumentation and AFW control and a redundant positive displacement pump (PDP), powered by the alternate EDG, to mitigate RCP seal LOCAs. Previous SAMA analyses (i.e., Susquehanna and Point Beach) have shown procurement of a portable EDG to be significantly less costly than procurement and installation of an additional injection pump. Provide an evaluation of an alternative SAMA involving procuring just a small portable EDG to supply critical plant systems such as AFW.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.1 The original SAMA design is considered to be the most appropriate for Diablo Canyon Power Plant (DCPP) given that the positive displacement pump (PDP) provides a means of.maintaining the plant in a stable state. Without primary side makeup capability in a long-term station blackout (SBO), the plant would potentially be in a configuration in which it is losing primary side inventory without a means of makeup until 4kV AC power is restored.

However, even if it is assumed that installation of a 480V AC generator alone could deliver the entire 95th percentile averted cost-risk for SAMA 5 (which does credit for the PDP); this modification would not be cost beneficial.

From Section F.7.2.3 of the Environmental Report (ER), the 95th percentile averted cost-risk for SAMA 5 is $68.1,000. The updated 9 5 th percentile averted cost-risk for SAMA 5 is presented in the response to RAI 3.c, is $978,024.

The cost of installing a 480V AC generator that can support steam generator level instrumentation after battery depletion has been developed by DCPP. The following is a summary of the estimated costs:

SAMA 5, Revision 1 Emergency Diesel Generator (EDG) for Steam Generator (SG) level indication only (no 480V 200 HP PDP Al for Reactor Coolant System (RCS) makeup) 480VAC DG set, cart mounted Weather proof storage on 85 ft.

Cables from 85 ft. storage loaction to two 480VAC/120VAC backup (BU) regulating; transformers on the 115 ft. elevation of the auxiliary building Seismic holddowns for diesel generator set 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> fuel supply on board diesel generator cart PG&E Letter DCL-10-106 Page 117 of 131 Disconnect switches for each BU 480V/120VAC BU regulator Description Qty.

Unit Unit $

Total Engineering 1

LS 750,000 750,000 Procedure Revisions 1

LS 150,000 150,000 40kW DG + Fuel Tank for 12 Hours 1

LS 90,000 90,000 Weatherproof Enclosure 1

LS 500,000 500,000 Conduit & Wire to BU Regulators 1,000 LF 1,000

$ 1,000,000 Miscellaneous Switches & Equipment 4

Ea 75,000 300,000 Revise Operator Training Program 1

LS 75,000 75,000 Sub-Total

$ 2,865,000 Capital Overheads (Mat'l 30% X 12.25% Mat'l Burden) +

(70% Lab X 40% PGE Lab X 14.3%

Capital A&G) 7.68%

220,032 Contingency 20.00%

617,006 Total 3,702,038 All costs in 2009 dollars. No escallation or AFUDC has been included.

Even using the updated averted cost-risk for SAMA 5, the 95th percentile net value is negative as shown below:

$978,024 - $3,702,038 = -$-2,724,014 PG&E Letter DCL-10-106 Page 118 of 131 RAI 6.m

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

/

m. Provide the modeling assumptions used to evaluate SAMA 6 in the uncertainty analysis (Section F. 7.2.1.1).

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.m Split Fractions: AWl L, AW4 Associated Split Fraction Top Event: AW Discussion: Auxiliary feedwater (AFW) system split fraction AW1 L is used for LOOP, loss of primary flow and loss of component cooling water (CCW). AW4 is used in more cases such as Anticipated Transient Without Scram (ATWS), general transients and steam generator tube rupture (SGTR). Ninety two percent of split fraction AW1 L's (the split fraction other than AWF with the highest fractional importance, AWl L represents the all support available case) probability is derived from the human action to align supplemental and backup water sources to AFW (HEP AWZTE1 L = 6.53E-04).'

Assumption(s):

The purpose of this SAMA is to~utilize the B.5.b pump, or an equivalent high-pressure secondary side pump to provide steam generator cooling as a backup to the motor and turbine driven AFW pumps. Should this pump be made available, and assuming it would share no support with the AFW pumps, it too would require a human action. That human action would likely depend (joint dependency) on the human action that is part of the AW1 L split fraction development. In consideration of that potential dependency the failure probability of the high-pressure secondary side pump, as a bounding estimate, is assumed to be 1.OE-02 and AWl L and AW4 can be multiplied by this value to arrive at a final value that represents the inclusion of this pump. Furthermore, the guaranteed failed split fraction, AWF, which represents the loss of all support to AFW, can be replaced by 1.OE-02 to.

approximate the case where the high-pressure secondary side pump is the only source of secondary cooling. It is assumed this pump is self contained and requires no support except for operator action. Thus a guaranteed failed split fraction (1.OOE+00) is not needed.

PG&E Letter DCL-10-106 Page 119 of 131 Model Change(s):

AWF - Guaranteed failed due to support dependencies = 1.OE-02 AW1 L - AFW with RCP'S TRIPPED AND NATURAL CIRCULATION MODE =

7.15E-04

• 1.OE-02 = 7.15E-06 AW4 - SUPPORT FOR BOTH MDP'S UNAVAILABLE = 3.66E-02

• 1.OE-02 =

3.66E-04 PG&E Letter DCL-10-106 Page 120 of 131 RAI 6.n

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

n. For SAMA 15 (Section F. 7.2.1.3), provide the implementation cost and the 9 5th percentile averted cost-risk value.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 6.n The 95th percentile averted cost-risk value is $289,485. The cost of implementation, as documented in Table F.5-3 of the Environmental Report, is

$9,626,592.

PG&E Letter DCL-10-106 Page 121 of 131 RAI 6.0

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

o. In estimating the benefit of the reduction in risk from external events, PG&E provides a separate analysis to estimate the benefit of the reduction in fire and seismic risk. Since a Level 2 PRA model was not developed as part of the fire and seismic models, the benefit from the reduction in fire and seismic risk is based on the reduction in fire and seismic CDF. However, this approach may underestimate the benefit for SAMAs in which the benefit is dominated by the reduction in population dose risk or OECR and not CDF. This is the case for SAMAs 3, 11, 24, and 25. Provide an assessment of the impact on the results for SAMAs 3, 11, 24, and 25 that accounts for the potentially higher reduction in dose-risk and OECR than CDF for these SAMAs.

PG&E Response to Severe Accident Mitigation Alternatives RAI 6.o As indicated in this RAI, the fire and seismic probabilistic risk assessment (PRA) models are not integrated with the current Level 2 model and similar insights related to the Level 3 impacts of a SAMA's implementation are not available.

Because these insights are not available, the Diablo Canyon Power Plant (DCPP) Environmental Report (ER) assumed that the fire and seismic cost-risk contributions were equal to the ratio of the fire and seismic CDFs to the internal events core damage frequency (CDF) multiplied by the internal events cost-risk.

The implicit assumption is that the fire and seismic CDFs are distributed among the internal events release categories in the same proportion as the internal events CDF. It is known that this is not necessarily the case, but it is used as an approximation.

Apart from the issues related to the distribution of CDF to release categories, correlating the averted cost-risk to CDF presents the additional complication that some SAMAs may impact the Level 2 and 3 results without impacting the CDF at all. In order to compensate for these issues in the quantifications of SAMAs 3, 11, 24, and 25, the fire and seismic averted cost-risks will be recalculated using information from the internal events Level 3 results. The fire and seismic risk reductions will be estimated by increasing any fire and seismic Level 1 reductions by the percentage the internal events Level 3 reductions exceed the Level 1 reductions. Specifically, the "percent change" reported for the fire and seismic models reported in Section F.6 of the ER will be increased bythe larger of the following two percentages:

• Percent Change Internal Events Dose-Risk - Percent Change Internal Events CDF PG&E Letter DCL-10-106 Page 122 of 131 Percent Change Internal Events.OECR - Percent Change Internal Events CDF For example, in a situation where the internal events CDF decreased by 10 percent while the internal events Level 3 results decreased by 15 percent, the fire and seismic "percent reduction" values would be increased by 5 percent each.

The ER did not explicitly provide the baseline fire and seismic contributions to, the maximum averted cost-risk (MACR) or a detailed explanation of the quantification process. The response to RAI 3.b documents this information.

The same quantification process is used here with the exception that the percent reduction in the fire and seismic results are calculated as described above.

Because the updated seismic PRA results yield a smaller CDF, this response is conservatively based on the external events results reported in the ER. The response to RAI 3.c reports the results using the updated seismic model.

SAMA 3 Section F.6.2 of the ER provides the following information for SAMA 3:

SAMA 3 PRA Results CDF Dose-Risk OECR Fire CDF Seismic CDF Base Value 8.44E-06 8.79

$33,699 1.39E-05 3.77E-05 SAMA Value 7.56E-06 7.46

$28,618 4.40E-06 3.77E-05 Percent Change 10.4%

15.2%

15.1%

68.3%

0.0%

Given that the Dose-Risk reduction is larger than the offsite economic cost-risk (OECR) reduction, it will be used as the basis for the external events change modifier:

Change Modifier = Percent Change Dose-Risk - Percent Change Internal Events CDF Change Modifier = 15.2% - 10.4% = 4.8%

The revised Percent Change values are as follows:

Revised SAMA 3 Percent Change Results Internal CDF Dose-Risk OECR Fire CDF Seismic CDF Percent Change 10.4%

15.2%

15.1%

73.1%

4.8%

Applying the process described in Section F.4 of the ER yields an internal events, cost-risk of $858,968. After accounting for "round-up" of the base internal events PG&E Letter DCL-1 0-106 Page 123 of 131 cost-risk, this value is $859,600. The nonfire/nonseismic external events contributions are accounted for by multiplying this value by 1.3:

$859,600

• 1.3 = $1,117,480 The fire cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base fire cost-risk of $1,642,500:

Fire Cost-RisksAMA3 = $1,642,500 * (1 - 0.731) = $441,833 The seismic cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base seismic cost-risk of $4,455,000:

Seismic Cost-RisksAMA3 = $4,455,000 * (1 - 0.048) = $4,241,160 The total cost-risk for SAMA 3 is the sum. of the fire, seismic, and nonfire/nonseismic cost-risk values:

SAMA 3 Cost-Risk = $1,117,480 + $441,833 + $4,241,160 = $5,800,473 The averted cost-risk is the difference between this value and the baseline MACR of $7,400,000:

SAMA 3 Averted Cost-Risk = $7,400,000 - $5,800,473 = $1,599,527 th In order to account for the 95 percentile PRA results, a multiplier of 2.36 (refer to Section F.7.2) is applied:

$1,599,527,* 2.36 = $3,774,884 Given that the cost of implementation -is $5,863,176, the net value is negative, which is consistent with the conclusion in the ER:

$3,774,884 - $5,863,176 = -$2,088;292 SAMA 11 Section F.6.8 of the ER provides the following information for SAMA 11 (note that the fire contribution has been updated to reflect the response to RAI 6.g):

SAMA 11 PRA Results Internal CDF Dose-Risk OECR Fire CDF Seismic CDF Base Value 8.44E-06 8.79

$33,699 1.39E-05 3.77E-05 SAMA Value 8.44E-06 7.92

$32,820 1.39-06 3.77E-05 PG&E Letter DCL-10-106 Page 124 of 131 Percent Change 0.0%

9.9%

2.6%

0.0%

0.0%

Given that the Dose-Risk reduction is larger than the OECR reduction, it will be used as the basis for the external events change modifier:

Change Modifier = Percent Change Dose-Risk - Percent Change Internal Events CDF Change Modifier = 9.9% - 0.0% = 9.9%

The revised Percent Change values are as follows:

Revised SAMA 11 Percent Change Results CDF Dose-Risk OECR Fire CDF Seismic CDF Percent Change 0.0%

9.9%

2.6%

9.9%

9.9%

Applying the process described in Section F.4 of the ER yields an internal events cost-risk of $959,919. After accounting for "round up" of the base internal events cost-risk, this value is $960,551,. The nonfire/nonseismic external events contributions are accounted for by multiplying this value by 1.3:

$859,600

• 1.3 = $1,248,716 The fire cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base fire cost-risk of $1,642,500:

Fire Cost-RiskSAMA11 = $1,642,500 * (1 - 0.099) = $1,479,893 The seismic cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base seismic cost-risk of $4,455,000:

Seismic Cost-RisksAMA11 = $4,455,000 * (1 - 0.099) = $4,013,955 The total cost-risk for SAMA 11 is the sum of the fire, seismic, and nonfire/nonseismic cost-risk values:

SAMA 11 Cost-Risk = $1,248,716 + $1,479,893 + $4,013,955 = $6,742,564 The averted cost-risk is the difference between this value and the baseline MACR of $7,400,000:

SAMA 11 Averted Cost-Risk = $7,400,000 - $6,742,564 = $657,436 PG&E Letter DCL-10-106 Page 125 of 131 In order to account for the 95th percentile PRA results, a multiplier of 2.36 (refer to Section F.7.2) is applied:

$657,436

• 2.36 = $1,551,549 Given that the cost of implementation is $4,651,776, the net value is negative, which is consistent with the conclusion in the ER:

$1,551,549 - $4,651,776 = -$3,100,227 SAMA 24 Section F.6.15 of the ER provides the following information for SAMA 24:

SAMA 24 PRA Results Internal CDF Dose-Risk OECR Fire CDF Seismic CDF Base Value 8.44E-06 8.79

$33,699 1.39E-05 3.77E-05 SAMA Value 8.43E-06 8.63

$32,758 1.39E-05 3.77E-05 Percent Change 0.1%

1.9%

2.8%

0.0%

0.0%

Given that the OECR reduction is larger than the Dose-Risk reduction, it will be used as the basis for the external events change modifier:

Change Modifier = Percent Change OECR - Percent Change Internal Events CDF Change Modifier = 2.8% - 0.1% = 2.7%

The revised Percent Change values are as follows:

Revised SAMA 24 Percent Change Results CDF Dose-Risk OECR Fire CDF Seismic CDF Percent Change 0.1%

1.9%

2.8%

2.7%

2.7%

Applying the process described in Section F.4 of the ER yields an internal events cost-risk of $980,117. After accounting for "round up" of the base internal events cost-risk, this value is $980,749. The nonfire/nonseismic external events contributions are accounted for by multiplying this value by 1.3:

$980,749

• 1.3 = $1,274,974 The fire cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base fire cost-risk of $1,642,500:

PG&E Letter DCL-10-106 Page 126 of 131 Fire Cost-RisksAMA24 = $1,642,500 * (1 - 0.027) = $1,598,153 The seismic cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base seismic cost-risk of $4,455,000:

Seismic Cost-RiskSAMA24 = $4,455,000 * (1 - 0.027) = $4,334,715 The total cost-risk for SAMA 24 is the sum of the fire, seismic, and nonfire/nonseismic cost-risk values:

SAMA 24 Cost-Risk = $1,274,974 + $1,598,153 + $4,334,715 = $7,207,842 The averted cost-risk is the difference between this value and the baseline MACR of $7,400,000:

SAMA 24 Averted Cost-Risk = $7,400,000- $7,207,842 = $192,158 In order to account for the 95th percentile PRA results, a multiplier of 2.36 (refer to Section F.7.2) is applied:

$192,158

• 2.36 = $453,493 Given that the cost of implementation is $50,000, the net value is positive, which is consistent with the conclusion in the ER:

$$453,493 - $50,000 = $403,493 SAMA 25 Section F.6.16 of the ER provides the following information for SAMA 25:

SAMA 25 PRA Results Internal CDF Dose-Risk OECR Fire CDF Seismic CDF Base Value 8.44E-06 8.79

$33,699 1.39E-05 3.77E-05 SAMA Value 8.44E-06 8.063

$33,220 1.39E-05 3.77E-05 Percent Change 0.0%

8.7%

1.4%

0.0%

0.0%

Given that the Dose-Risk reduction is larger than the OECR reduction, it will be used as the basis for the external events change modifier:

Change Modifier = Percent Change Dose-Risk - Percent Change Internal Events CDF Change Modifier = 8.7% - 0.0% = 8.7%

PG&E Letter DCL-10-106 Page 127 of 131 The revised Percent Change values are as follows:

Revised SAMA 25 Percent Change Results CDF Dose-Risk OECR Fire CDF Seismic CDF Percent Change 0.0%

8.7%

1.4%

8.7%

8.7%

Applying the process described in Section F.4 of the ER yields an internal events cost-risk of $969,321. After accounting for "round-up" of the base internal events cost-risk, this value is $969,953. The nonfire/nonseismic external events contributions are accounted for by multiplying this value by 1.3:

$969,953

• 1.3 = $1,260,939 The fire cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base fire cost-risk of $1,642,500:

Fire Cost-RisksAMA25 = $1,642,500 * (1 - 0.087) = $1,499,603 The seismic cost-risk can be obtained by multiplying the complement of the revised "percent change" fraction by the base seismic cost-risk of $4,455,000:

Seismic Cost-RisksAMA25 = $4,455,000 * (1 - 0.087) = $4,067,415 The total cost-risk for SAMA 25 is the sum of the fire, seismic, and nonfire/nonseismic cost-risk values:

SAMA 25 Cost-Risk = $1,260,939 + $1,499,603 + $4,067,415 = $6,827,957 The averted cost-risk is the difference between this value and the baseline MACR of $7,400,000:

SAMA 25 Averted Cost-Risk = $7,400,000 - $6,827,957 = $572,043 In order to account for the 95th percentile PRA results, a multiplier of 2.36 (refer to Section F.7.2) is applied:

$572,043

• 2.36 = $1,350,021 Given that the cost of implementation is $50,000, the net value is positive, which is consistent with the conclusion in the ER:

$1,350,021 - $50,000 = $1,300,021 PG&E Letter DCL-10-106 Page 128 of 131 RAI 6.p

6. Provide the following information with regard to the Phase II cost-benefit evaluations:

p. PG&E's review of the Prairie Island SAMAs in Section F.5.1.3.5 did not address Prairie Island SAMA 3, "Provide Alternate Flow Path from RWST to Charging Pump Station" or SAMA 19a, "Upgrade Equipment and Procedures for Replenishing RWST Inventory from a Large Water Source," which were determined to be potentially cost-beneficial in response to NRC staff, RAIs.

Review these SAMAs for applicability to DCPP and provide a cost-benefit evaluation, if applicable.

PG&E Response to Severe Accident Mitiqation Alternatives RAI 6.p The following table provides a disposition of the Prairie Island SAMAs 3 and 19a for Diablo Canyon Power Plant (DCPP):

REVIEW OF ADDITIONAL PRAIRIE ISLAND COST BENEFICIAL SAMAS FROM NUREG-1437 INDUSTRY SAMA DISCUSSION FOR DCPP DISPOSITION SITE SAMA ID DESCRIPTION FOR DCPP SAMA LIST 3

Provide alternate flow path from RWST to charging pump suction The refueling water stroage tank (RWST) suction path already includes a pair of redundant valves (Motor Operated Valve (MOV) 8805A/B) for the normal centrifugal charging pumps and the Class 2, third charging pump. The risk-reduction worth (RRW) values for all of the Diablo Canyon Power Plant (DCPP) split fractions (CH*)

related to the RWST suction path to the charging purfips are below 1.002 and are not meaningful risk contributors.

Already implemented; no further evaluation required.

PG&E Letter DCL-10-106 Page 129 of 131 REVIEW OF ADDITIONAL PRAIRIE ISLAND COST BENEFICIAL SAMAS FROM NUREG-1437 INDUSTRY SAMA DISCUSSION FOR DCPP DISPOSITION SITE SAMA ID DESCRIPTION FOR DCPP SAMA LIST 19a Provide a reliable For Prairie Island, the installation of the Already backup water RWST refill source is credited primarily for implemented; but source for increasing the time that is available to an alternate replenishing the perform the RCS cooldown in an steam SAMA evaluation RWST generator tube rupture (SGTR). Cooldown is required.

would equalize primary and secondary side pressures and effectively terminate the inventory loss to the secondary side. DCPP already has the capability to provide makeup to the RWST from the Chemical and. Volume 9

Control System (CVCS) blend tanks as well as from the spent fuel pool. The RWST refill action is directed in the plant's emergency proceduresýand credited in the probabilistic risk assessment (PRA) model. While RWST refill is a reliable action, failure of the makeup action is included in the Level 2 imporance list for interfacing system loss-of-coolant-accidents (LOCAs) (split fraction MUV).

Review of Table F.5-2a of the Environmental Report (ER) indicates that SAMA 2 was identified as a means of mitigating the failure of RWST refill; however, recirculation is. not available for the dominant interfacing system LOCA events. Automating RWST makeup is a more relevant means of addressing RWST makeup failures. This SAMA is evaluated further below.

Evaluation of Automating RWST Makeup for DCPP Automating refueling water storage tank (RWST) makeup is considered to require the installation of tank level sensors/transmitters with logic that will start the spent fuel pool transfer pumps, open the valves in the flowpath, and return the system to standby when the tank is refilled. Three Mile Island estimated the cost of this type of change to be $3,800,000 (Exelon 2008). This estimate is considered to reflect the scope and nature of changes required for implementing RWST makeup at DCPP and it is used for this SAMA's cost of implementation in this evaluation.

A bounding averted cost-risk estimate can be developed to show that this SAMA could not be cost effective for DCPP. The low proportion of the maximum PG&E Letter DCL-10-106 Page 130 of 131 averted cost-risk (MACR) associated with internal events the small role that interfacing system loss-of-collant-accident (ISLOCA) and steam generator tube rupture (SGTR) play in combination with fire and seismic initiators provide a basis for the estimate.

For DCPP, the nonfire/nonseismic contribution to the MACR is:

Internal Events MACR * ((nonfire/nonseismic core damage frequency (CDF) +

internal events CDF) / internal events CDF)

$1,000,000 * ((2.56E-6 + 8.44E-06) / 8.44E-06) = $1,300,000 Even if the entire nonfire/nonseismic cost-risk of $1,300,000 is assumed to be mitigated by the RWST refill SAMA, it would not be cost effective even when the 95 percentile probabilistic risk assessment (PRA) results are considered

($1,300,000

• 2.36 = $3,068,000).

ISLOCA and SGTR events are excluded from quantification in combination with other initiating events in the risk models due to low contribution, so no additional impact would be expected from the fire and seismic scenarios. Even if the relevant RWST makeup split fractions were assumed to exist just below the review thresholds, the contributions would be limited to $50,000 foreach initiating event type. If the $100,000 sum is scaled up to account for the 9 5th percentile PRA results, the total would only be $236,000 (($50,000 + $50,000)

• 2.36 = $236,000). Adding this contribution to the nonfire/nonseismic contribution of $3,068,000 yields $3,304,000, which is still less than the cost of implementation.

No further review of automating RWST makeup is required for DCPP.

REFERENCES Exelon 2008 EXELON (Exelon Corporation). 2008 Applicant's Environmental Report; Operating License Renewal Stage; Three Mile Island Unit 1. Attachment E -

Severe Accident Mitigation Alternatives Analysis. January.

PG&E Letter DCL-1 0-106 Page 131 of 131

RAI 7

7. For certain SAMAs considered in the Environmental Report, there may be lower-cost alternatives that could achieve much of the risk reduction at a lower cost. In this regard, provide an evaluation of the following SAMA: Purchase or manufacture of a "gagging device" that could be used to close a stuck-open steam generator safety valve for a SG TR event prior to core damage.

PG&E Response to Severe Accident Mitiqiation Alternatives RAI 7 Diablo Canyon Power Plant (DCPP) already has steam generator safety valve (SV) gagging devices, but these devices are not designed to close an open SV.

They are designed to block the opening of an SV and to assist in closing a "simmering" or leaking SV. The thin rods that are used to apply force to the tops of the SVs for these applications would fail if an attempt was made to use them to close a stuck open SV.

Even if the gagging devices were physically capable of closing an open SV, local operation of such devices at DCPP would not likely be feasible during a steam

,generator tube rupture (SGTR) event. Unlike Beaver Valley, where this SAMA was initially identified as potentially cost effective, DCPP does not have primary loop isolation valves. At a minimum, SGTR conditions would present an extremely challenging working environment at the safety valve (i.e., heat, noise, radiation, timing issues). Remote operation of a gagging system would be required to realistically ensure reclosure of a safety valve during an SGTR event, which is not a low cost alternative.