Text

Response to NRC Request for Additional Information for the Review of the Fermi 2 License Renewal Application - Severe Accident Mitigation Alternatives
	ML15009A358
Person / Time
Site:	Fermi
Issue date:	01/09/2015
From:	Kaminskas V; DTE Energy
To:	; Document Control Desk, Office of Nuclear Reactor Regulation
References
	NRC-15-0013
	Download: ML15009A358 (110)
	v • d • e

Vito A. Kaminskas Site Vice President DTE Energy Company 6400 N. Dixie Highway, Newport, MI 48166 Tel: 734.586.6515 Fax: 734.586.4172 Email: kaminskasv@dteenergy.com DTE Energy-10 CFR 54 January 9, 2015 NRC-15-0013 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington D C 20555-0001

References:

1) Fermi 2 NRC Docket No. 50-341 NRC License No. NPF-43

2) DTE Electric Company Letter to NRC, "Fermi 2 License Renewal Application," NRC-14-0028, dated April 24, 2014 (ML14121A554)

3) NRC Letter, "Requests for Additional Information for the Review of the Fermi 2 License Renewal Application - Severe Accident Mitigation Alternatives," dated November 18, 2014 (ML14308A358)

Subject:

Response to NRC Request for Additional Information for the Review of the Fermi 2 License Renewal Application -

Severe Accident Mitigation Alternatives In Reference 2, DTE Electric Company (DTE) submitted the License Renewal Application (LRA) for Fermi 2. In Reference 3, NRC staff requested additional information regarding the Fermi 2 LRA. The Enclosure to this letter provides the DTE response to the request for additional information (RAI).

The following commitment is being made in this submittal:

The potentially cost-beneficial Severe Accident Mitigation Alternative (SAMA) candidates (discussed in the responses to RAIs 5.a.vii, 6.h, and 7.b) will be evaluated.

Should you have any questions or require additional information, please contact Lynne Goodman at 734-586-1205.

USNRC NRC-15-0013 Page 2 I declare under penalty of perjury that the foregoing is true and correct.

Executed on January 9, 2015 Vito A. Kaminskas Site Vice President Nuclear Generation

Enclosure:

DTE Response to NRC Request for Additional Information for the Review of the Fermi 2 License Renewal Application - Severe Accident Mitigation Alternatives cc: NRC Project Manager NRC License Renewal Project Manager NRC License Renewal Environmental Project Manager NRC Resident Office Reactor Projects Chief, Branch 5, Region III Regional Administrator, Region III Michigan Public Service Commission, Regulated Energy Division (kindschl@michigan.gov)

Enclosure to NRC-15-0013 Fermi 2 NRC Docket No. 50-341 Operating License No. NPF-43 DTE Response to NRC Request for Additional Information for the Review of the Fermi 2 License Renewal Application -

Severe Accident Mitigation Alternatives

Enclosure to NRC-15-0013 Page 1 RAI 1.a Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

a. The environmental report (ER) states that No other planned major plant modifications, which could adversely impact the SAMA analysis results, have been identified. Request confirmation that this applies to all planned modifications, major or not, and to changes in operating practices/procedures.

Response

All plant modifications since the FermiV9 model freeze date have been evaluated. Only one plant modification was identified as having an impact on the Fermi 2 internal events model. This modification was to add a third breaker row to the 345 KV switchyard. Although this modification took place after the FermiV9 model freeze date, it was included in the FermiV9 model due to its impact (primarily on Maintenance Rule (a)(4) risk evaluations). Pending Fukushima modifications will also not have an adverse impact on SAMA. In addition, no operating practice/procedure changes have been identified that would have an adverse impact on SAMA. Additional information is provided in the response to RAI 4.k.

Enclosure to NRC-15-0013 Page 2 RAI 1.b Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

b. Describe any credit being taken in the FermiV9 PRA for the mitigating strategies required by 10 CFR 50.54(hh)(2).

Response

There are several instances where B.5.b mitigating strategies are taken credit for in the Fermi 2 Probabilistic Risk Assessment (PRA) model. These strategies are outlined in the Fermi 2 Extreme Damage Mitigation (EDM) procedures. Below is a listing of the instances where B.5.b actions are credited in the model, with accompanying implementing procedure numbers, modeled human action basic events, and associated PRA gate information:

Action Procedure Human Action PRA Gate 1 Refilling Condensate 29.EDM.11 HE1RCSTSCSTEDM FIRE_WTR_MAKEUP Storage Tank (CST) via (OPERATOR FAILS (INADEQUATE FIRE Fire Protection header TO FILL CST PER WATER MAKEUP TO EDMG) CST) 2 Venting the Containment 29.EDM.08 HE1FHVNTACVT-- HVNT-NOAC Locally without AC (Failure to Vent Locally (FAILURE OF AC Power without AC power) POWER AND LOCAL VENTING ACTION) 3 Alignment of portable 29.EDM.15 HE1FSRVSPORTPWR T-LP-DC (FAILURE power to the SRVs (CREW FAILS TO TO ALIGN ALIGN PORTABLE PORTABLE POWER POWER TO SRVs) TO SRVs)

Enclosure to NRC-15-0013 Page 3

Action Procedure Human Action PRA Gate 4 Alignment of the B.5.b 29.EDM.05 HE1FFPROB5BRPV B5BINJFAILS-EX pump to provide ex- 29.EDM.14 (OPERATOR FAILS (B.5.B PUMP FAILS vessel injection via TO ALIGN B5B FOR TO INJECTION TO drywell sprays following INJECTION [Late]) RPV EX-VESSEL INJ) core damage.

Note: Injection to the RPV via the B.5.b pump to prevent core damage is not credited.

5 Replenishment of fuel oil 29.EDM.13 HE2FFPROB5BFO B5BPUMPFAILS supply to the B.5.b 29.EDM.14 (FAILURE TO (B.5.B PUMP FAILS pump. REPLENISH B.5.B TO DELIVER FLOW)

FUEL OIL SUPPLY)

Note: Injection to the RPV via the B.5.b pump to prevent core damage is not credited.

Enclosure to NRC-15-0013 Page 4 RAI 1.c Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

c. Relative to Table D.1-1:

i. Provide a more detailed listing of the initiating events percentage contribution to the internal events CDF. Include the breakdown of loss of coolant accidents (LOCAs) by size and/or cause and the various contributors to the general transients and special initiator groups.

ii. Clarify what is meant by the term, (without LOSP) for the general transients descriptions. Are consequential losses of offsite power (LOSPs) modeled and not included in this value?

iii. Discuss briefly the modeling of the LOSP and station blackout (SBO) scenarios including how the combustion turbine generators are incorporated in the model and how common cause loss of alternating current due to weather is considered.

Response

i. The table below contains a list of initiating events which contribute greater than 1% to the Core Damage Frequency (CDF) solution (Fussel-Vesely > 0.01) at the 1.0E-12/yr truncation limit in the FermiV9 PRA model sorted by Fussel-Vesely (from largest to smallest).

Event Name Probability Fus Ves Description

%LOSP 6.34E-02 1.42E-01 TOTAL LOSS OF OFFSITE POWER TURBINE TRIP WITH BYPASS INITIATING

%TX 8.37E-01 1.30E-01 EVENT

%S1-WA 1.44E-04 1.04E-01 MEDIUM LOCA BELOW TAF (WATER)

%FL-AB-FPRO-RELAY- Nominal rupture in FPS line in AB propagating to 6.70E-06 7.36E-02 N Relay Room

%S1-LP 9.46E-05 6.96E-02 MEDIUM LOCA IN LPCI LINE

%S1-FW 8.72E-05 3.99E-02 MEDIUM LOCA IN FW LINE LOSS OF CONDENSER VACUUM

%LOCV 1.61E-01 3.54E-02 INITIATING EVENT

Enclosure to NRC-15-0013 Page 5 Event Name Probability Fus Ves Description

%FL-TB-MCWS-TBXX- Major rupture in Circulating Water pipe or 1.25E-03 3.44E-02 M expansion joints in Turbine Building.

%ISLOCA-SDC 5.90E-08 3.41E-02 ISLOCA IN RHR SDC SUCTION LINE (X-12)

%TMS 1.31E+00 2.92E-02 MANUAL SHUTDOWN INITIATING EVENT

%PLOOP301 2.11E-02 2.54E-02 PARTIAL LOSP FOR DIV. 2

%BS301 1.17E-02 2.45E-02 LOSS OF BUS #301 INITIATING EVENT

%FL-AB-ECW2-B20XX- Nominal rupture in RBCCW/EECW Div 2 line in 1.79E-06 2.20E-02 N DC Switchgear Room

%FL-AB-FPRO-RELAY- Major rupture in FPS line in AB propagating to 1.46E-06 1.60E-02 M Relay Room

%S2-WA 3.09E-03 1.58E-02 SMALL LOCA BELOW TAF (WATER)

%LOFW 7.30E-02 1.56E-02 LOSS OF FEEDWATER INITIATING EVENT LOSS OF GENERAL SERVICE WATER

%GSW 1.00E+00 1.37E-02 INITIATING EVENT

%TBCCW 1.00E+00 1.28E-02 LOSS OF TBCCW INITIATING EVENT

%FL-AB-FPRO-CCHV2- Nominal rupture of Fire protection piping in Div 2 1.07E-05 1.18E-02 N CCHVAC Room.

%RBCCW 1.00E+00 1.18E-02 LOSS OF RBCCW INITIATING EVENT

%BS101 2.34E-02 1.06E-02 LOSS OF BUS #101 INITIATING EVENT ii. This annotation refers to general transients without a partial loss of offsite power. At Fermi 2, there are two separate switchyards (120kV and 345kV), each of which supply offsite power to a single division of Engineered Safety Function (ESF) and Balance of Plant (BOP) power (these are often referred to as divisional switchyards and the loss of offsite power initiators are referred to as divisional Loss of Offsite Power (LOSP) events). The 120kV switchyard supplies Division 1 power and the 345kV supplies Division 2 power. The partial loss of offsite power events (%PLOOP101, %BS101, %PLOOP301, and %BS301) are processed via the General Transient event tree. Consequential losses of offsite power are modeled as total LOSPs and are processed via the LOSP event tree (not in the General Transient event tree).

iii. The effects of a loss of offsite power are evident in its impact on the BOP systems. The feedwater and condensate systems are rendered initially unavailable due to the LOSP initiating event. Similarly, because of the equipment lost when offsite power is interrupted, the main condenser is also initially unavailable for core decay heat removal.

In addition to the impact on BOP systems, the loss of offsite power event challenges the emergency Alternating Current (AC) power systems to provide AC power. The failure of the emergency AC power systems following a LOSP would result in what is commonly referred to as a station blackout (SBO). These conditions are the complete loss of all AC power to the

Enclosure to NRC-15-0013 Page 6 unit. This requires the failure of the following:

1) Emergency diesel available to the unit

2) Interconnection with the on-site gas turbines The loss of offsite power and station blackout probabilistic evaluations concentrate on the description of possible sequences which may occur during the operator's attempt to successfully maintain the critical safety functions which include core cooling and containment heat removal while attempting to restore AC power and mitigate the challenges of this scenario.

The SBO event tree begins with sequences which are transferred from the LOSP event tree for sequences where a total LOSP (as opposed to a partial LOSP) initiating event occurs and the Emergency Diesel Generators (EDGs) and Combustion Turbine Generators (CTGs) fail to provide onsite AC power to mitigate the event. If at least one EDG or one CTG is available as a source of onsite AC power, the LOSP event tree is utilized (rather than the SBO event tree).

The LOSP event tree is characterized in general terms by evaluating the following mitigating capabilities:

1) The availability of suppression pool cooling early in the sequence.

2) The ability for Direct Current (DC) powered high pressure systems (High Pressure Coolant Injection (HPCI) and/or Reactor Core Isolation Cooling (RCIC)) or Standby Feedwater (SBFW) to provide high pressure Reactor Pressure Vessel (RPV) makeup.

3) Manual depressurization per the Emergency Operating Procedures for scenarios where high pressure injection or early suppression pool cooling is not successful.

4) RPV injection with low pressure systems.

5) Restoration of offsite power (at various time points in the sequence).

6) Long term containment cooling via late suppression pool cooling, shutdown cooling, or venting of the containment.

7) Continued RPV injection following containment failure.

The SBO event tree is characterized in general terms by evaluating the following mitigating capabilities:

1) The ability for DC powered high pressure systems (HPCI and/or RCIC) to provide high pressure RPV makeup.

2) Manual depressurization per the Emergency Operating Procedures.

3) The ability to recover offsite power (at various time points in the sequence).

4) The ability to successfully mitigate core damage in the long term following successful recovery of offsite power. This is accomplished via successful high or low pressure injection followed by either:

(a) Successful containment cooling (as described for the LOSP event tree).

Enclosure to NRC-15-0013 Page 7 (b) Continued RPV injection following containment failure.

The combustion turbine generators are modeled with a support system fault tree. They are modeled for the capability to provide onsite AC power to loads supported by the downstream electrical busses. The CTG system model is characterized by the following:

1) Failure to start and failure to run terms for each CTG.

2) Failure to start and failure to run terms for the standby diesel generator which provides blackstart capability to CTGs 11-2, 11-3, and 11-4.

3) Maintenance unavailability terms for each CTG and the standby diesel generator which provides blackstart capability to CTGs 11-2, 11-3, and 11-4.

4) Common cause failure groups for both failure to start and failure to run composed of CTGs 11-1, 11-2, 11-3, and 11-4.

5) Human Failure Event (HFE) to represent the failure to start the CTG from the control room.

6) HFE to represent the failure to locally blackstart CTGs 11-2, 11-3, or 11-4 via the standby diesel generator.

7) Failure of fuel oil delivery.

8) Failure of CTG 11-1 transformer.

9) Failure of busses and breakers which connect the CTG to the downstream electrical distribution system.

Partial loss of offsite power scenarios are addressed via the General Transient event tree.

Equipment failures (including failures of onsite AC power systems) are addressed via the supporting system fault tree logic and functional fault tree logic. Since, by definition, a Station Blackout is the loss of all offsite and onsite AC power, a partial loss of offsite power (with power from the opposite divisional switchyard available) does not meet that definition.

Common cause loss of AC power due to weather-related events is included in the model in the following manner:

1) Conditional probabilities for a weather centered loss of offsite power event are included in the model as follows:

(a) Basic event LOOP-IE-SW for total LOSP.

(b) Basic event CPFFACP1COND-WE for a Division 1 (120 KV Switchyard) partial LOSP.

(c) Basic event CPFFACP2-COND-WE for a Division 2 (345 KV Switchyard) partial LOSP.

2) The phenomenological failure probability event PHPHCTG1WEATHERLOSP represents the common cause failure of all four CTGs in the event of a weather centered total LOSP.

3) The phenomenological failure probability event PHPHCTG1WEATHER101 represents the common cause failure of all four CTGs in the event of a weather centered loss of the 120kV (Division 1) Switchyard.

Enclosure to NRC-15-0013 Page 8 It should be noted that CTG 11-1 has integral blackstart capability and is the credited Alternate AC (AAC) power source in the Fermi 2 design basis. CTG 11-2, 11-3, or 11-4 may be blackstarted with the aid of a standby diesel generator which is dedicated to the purpose of providing starting power for these CTGs.

Enclosure to NRC-15-0013 Page 9 RAI 1.d Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

d. The Fermi 2 internal events CDF is considerably lower than those for other boiling-water reactor (BWR) 3/4 units. A comparison of CDFs was to have been made in the resolution of the PRA Peer Review (Item 4-16). Provide a discussion of the Fermi 2 CDF in comparison with other similar units and the reasons for any significant differences.

Response

The Fermi 2 CDF is lower than other BWR-4s. There are several unique features of the Fermi 2 plant that result in this difference.

Fermi 2 has a Standby Feedwater system that is not found at other BWRs. This external injection system has the ability to inject to the vessel at high and low pressure. It consists of two motor driven pumps, although only one is needed for most scenarios. One pump is powered by Division 2 and the other pump is powered by Division 1 (which has CTG 11 backup). This system is credited early in general transient and loss of offsite power scenarios and is a backup to HPCI and RCIC.

Fermi 2 has a Residual Heat Removal (RHR) complex which contains 2 divisions of the ultimate heat sink. This facility contains the Emergency Diesel Generators, the RHR service water pumps, the diesel generator service water pumps, and emergency equipment service water pumps. This building is protected from design basis tornados.

The ultimate heat sink has adequate inventory for 7 days without makeup.

Fermi 2 has two independent switchyards. Division 1 offsite power is provided by a 120 kV switchyard that is fed from three offsite lines. Division 2 offsite power is provided by a 345 kV switchyard which is fed from two offsite lines. These switchyards are electrically and spatially separated. There is the ability to cross-tie from one division to the other using a maintenance tie breaker.

Fermi 2 has 4 emergency diesel generators (2 support Division 1, 2 support Division 2).

One EDG can provide adequate power to shutdown the plant in general transients and loss of offsite power. Fermi 2 also has 4 combustion turbine generators (CTGs). The CTGs provide power to the Division 1 switchyard (which in turn can provide power to the Standby Feedwater system). CTG 11-1 has blackstart capability from the Control

Enclosure to NRC-15-0013 Page 10 Room for station blackouts and units 11-2, 11-3, and 11-4 can be manually aligned for blackstart using a standby diesel generator.

Fermi 2 has several closed cooling water systems used to cool plant systems. The benefit is that these closed cooling water systems have a finite amount of water which leads to a reduced internal flooding impact.

Enclosure to NRC-15-0013 Page 11 RAI 1.e Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

e. Sections D.1.4.3 and D.1.4.4 of the ER list the same model changes from the FermiV2 model for the FermiV3 and FermiV4 models. The CDF for FermiV3 and FermiV4 models are different (3.3E-06 and 5.8E-06, respectively). Clarify the differences between the FermiV3 and FermiV4 models, including the differences between the Equipment Out of Service (EOOS) software and the Computer-Aided Fault Tree Analysis (CAFTA) software used to develop the FermiV3 and FermiV4 models, respectively. Characterize the unavailabilities due to testing and maintenance included in each model and describe the changes to the FermiV4 model resulting from the cited incorporation of recommendations from the prior peer review in the FermiV4 model.

Response

The FermiV3 and FermiV4 models were similar to one another. The primary motivation for the release of the FermiV3 model was the incorporation of the Level 2 model for the calculation of large early release frequency (LERF) and other Level 2 release terms.

The modeling changes in the FermiV3 model were carried forward to the FermiV4 model; an exception to this is the data analysis methodology employed to calculate the test and maintenance unavailability basic event values. In the FermiV3 model, the test and maintenance unavailability basic event values were calculated based upon data obtained from plant operating experience (this method is consistent with the supporting requirements in the current American Society of Mechanical Engineers (ASME) Standard). Prior to the FermiV4 release, a decision was made by Probabilistic Safety Analysis (PSA) Group supervision to revert to the method of calculating the values for these basic events in the FermiV2 model. In this method the test and maintenance unavailability basic events were set to values based upon the Maintenance Rule (MR) performance criteria (this introduced considerable conservatism, since few of the systems attained unavailability near their MR performance criteria). This utilization of this method introduced considerable conservatism into the FermiV4 model quantification and was the primary reason for the CDF values for the FermiV4 and FermiV3 being different in magnitude.

The practice of using test and maintenance unavailability basic event values derived from MR performance continued until the release of the FermiV9 model. The current modeling practice is

Enclosure to NRC-15-0013 Page 12 to utilize plant data to establish unavailability values; this practice meets the applicable data analysis requirements in the 2009 ASME PRA Standard.

The FermiV3 model being developed for the EOOS software and the FermiV4 model being developed for the CAFTA software had no bearing on the difference in CDF between the two models.

The table below is a summary of the model changes that were incorporated into the FermiV4 PRA model. The addition of the reference leg break initiator was performed specifically to address a 1997 NEI Peer Review item.

ITEM DESCRIPTION Fault Tree Added Linked Level 2 Release Trees Added SBFW short term tree. This takes credit for use of SBFW without CST makeup, which permits more time for the operator to depressurize the RPV.

Initiating Updated IE frequencies utilizing plant specific data Events (IE) Divided BOC IE into individual lines.

Added Instrument line break and Loss of DW Cooling initiators.

T/M Out of Utilized values that were identical or compatible with current MR performance Service criteria. These values were explicitly modeled in the new fault trees.

Demand Utilized the PLG generic data for all failure data. Then a Bayesian update was Failure performed on the generic data when plant specific CPEP data was available.

Data MOV data from MOV Coordinator was also utilized.

Demand Data from CPEP for 5 CP systems were used to update basic event values.

Common Utilized Demand Failure data for updating singletons, then NRC CCF data was Cause used to generate new CCF values Recovery Updated Power Recovery for Total Loss of Offsite Power (LOSP)

Enclosure to NRC-15-0013 Page 13 RAI 1.f Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

f. Section D.1.4.8 of the ER indicates that one of the reasons that the FermiV9 CDF is lower than previous revisions is the update of Level 1 and Level 2 dependent human error probabilities. On ER page D-74, an expert panel review of the human reliability analysis (HRA) dependency was discussed. Specify the latest update and its relationship to the HRA dependency analysis reviewed by the expert panel.

Response

The Expert Panel reviewed the FermiV9 draft model that was reviewed by the peer review team.

The FermiV9 model is the latest update and was used for the SAMA after being finalized. In addition, prior to the peer review another team performed a gap assessment of the draft model to the ASME/ANS requirements. This gap assessment included a review of the human reliability analysis (HRA) dependency analysis.

There was one change to the HRA dependency analysis that was included in the FermiV9 model version (that was used for the SAMA submittal) and was not present in the draft model that was utilized for 2012 Peer Review (on which the expert panel performed its evaluation). This change involved the incorporation of the resolution to Peer Review Finding 3-28.

Finding 3-28 stated that the use of the event HE1FRXPCHSML (a failure to depressurize during a steam loss of coolant accident (LOCA)) to represent cutsets in which HE1FRXPCHWML (a failure to depressurize during a water LOCA) was non-conservative. This finding applied to the use of timing information ascertained from the steam LOCA failure to depressurize basic event (HE1FRXPCHSML) for dependent human error probability (HEP) groupings in which HE1FRXPCHWML (a failure to depressurize water LOCA event ) was involved (as one of the single HEPs that was replaced by the dependent HEP grouping).

The resolution to this finding involved the creation of a new dependent HEP grouping HE1D-D-HPIMLW (Common Failure to Operate High Pressure Injection Systems Given MLOCA Water).

The probability for this dependent HEP was higher than the dependent HEP involving medium steam LOCA. Therefore, this model change contributed (along with other changes discussed in the response to Question 1.g) to an increase in CDF from the draft PRA model used for the 2012

Enclosure to NRC-15-0013 Page 14 Peer Review to the FermiV9 model (from 1.27E-6/yr to 1.50E-6/yr). The contribution to CDF for this basic event is nearly 5% (FV of 4.86E-2). The resolution of this finding did not result in any methodology changes.

Enclosure to NRC-15-0013 Page 15 RAI 1.g Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

g. Page D-73 of the ER states that the Fermi 2 Boiling Water Reactors Owners Group PRA Peer Review was performed during August 2012. The PRA used for the SAMA analysis is FermiV9 issued in March 2013 (ER page D-8). This version is described as a complete upgrade of the previous model (ER page D-72). While the changes made in the FermiV8 model, to produce the FermiV9 model, may or may not be upgrades as defined in the American Society of Mechanical Engineers (ASME) PRA standard, the changes appear to be extensive and cover almost all Level 1 and 2 PRA tasks. Provide further justification for the technical adequacy of the FermiV9 model including the applicability of the peer review to the upgraded model for both Levels 1 and 2. If the peer review was performed on a draft version of the FermiV9 model, provide the results of the draft and identify changes to the draft to produce the model used for the SAMA analysis.

Response

The draft model that was reviewed during the August 2012 Peer Review was the same model as the FermiV9 model (the model utilized for the SAMA application) with the exception of the items listed below. The reasons for making changes to the draft model were:

1) To respond to Peer Review findings.

2) To correct modeling issues that the Fermi 2 PSA Group deemed significant.

In preparing the Quantification and Summary Notebook for the FermiV9 model release, the Fermi PSA team performed a thorough review of the model to evaluate the effects of these model changes on the final quantification.

Because of the small number of modeling changes (many of which were to resolve findings from the peer review team) when compared to the scope of the model, the relatively small impact of the changes in terms of the effect on risk metrics (as defined in the table below), and the fact that the peer review team evaluated each supporting requirement (SR) in the ASME standard associated with initiating events and internal flooding (no carryover from the NEI peer review was credited), the results of the peer review on the draft model are applicable to the FermiV9 model on which the calculations in the SAMA submittal were evaluated.

Enclosure to NRC-15-0013 Page 16 The CDF and LERF for the draft model used for the 2012 Fermi 2 PRA Peer Review were 1.27E-6/yr and 3.86E-7/yr. The table below is a complete listing of the model changes that were made in the time period between these versions. The final CDF and LERF for the FermiV9 model are 1.50E-06/yr and 3.73E-7/yr. The most significant items below are: 2, 4, 6, 10, 17, and 19.

Item Change Description Quantitative Effect 1 Support for the resolution to Finding 1-20. The result of this change is an increase in Basic event ATOPB72SS024 and type code the frequency for sequences involving ATOP were added to the model and associated basic event ATOPB72SS024. The FV for database to represent the auto-transfer of the the basic event is 1.12E-4.

Bus 72S feed from Bus 72M to Bus 72R.

2 Support for the resolution to Finding 1-25. Set There is an increase in CDF associated the probability of the basic event with this change. The FV for this basic HE1FRBCW-FL-ISOL-AB3 (FAIL TO event is 2.25E-2.

TERM. FLOOD FROM EECW IN AN AB3 SWGR ROOM) to 1.0.

3 Support for the resolution to Finding 1-26. This resulted in an increase in CDF The initiating event frequency for the major resulting from cutsets involving basic Circulating Water pipe breaks in the Turbine event %FL-TB-MCWS-TBxx-M; the FV Building (%FL-TB-MCWS-TBxx-M) is for this event is 1.55E-03.

increased by 1E-3/Rx Yr to reflect the calculated maintenance induced failure frequency.

4 Support for the resolution to Finding 3-24. This resulted in an increase in the CDF.

Corrected the calculation for the TPFS HPRI The FV values for the basic events which (HPCI/RCIC System Pumps FAIL TO START) use this type code are:

failure to start type code failure rate. This resulted in the failure rate increasing from a) TPFSHPCIC001A (Failure to start 5.77E-6 to 4.21E-3. of HPCI Pump) - 2.30E-02.

b) TPFSRCICC001 (Failure to start of RCIC Pump) - 9.20E-4.

c) TPFSHPRCCC22_1 (Common Case Failure of the HPCI and RCIC Pumps) - 8.43E-4.

Enclosure to NRC-15-0013 Page 17 Item Change Description Quantitative Effect 5 Support for the resolution to Finding 3-26. This resulted in an increase in the CDF.

Added common cause terms for The FV value associated with the common a) The feedwater check valves cause failure of the feedwater check (B2100F010A and B2100F010B) valves (basic event CVFORXFWCC22_1) failing to close. is 1.42E-4.

b) The 4160 V breakers (B6, C6, E6, and F6) which isolate the grid from the The FV values for basic events associated diesel supported output busses in the with the failure of the 4160 V breakers to event of a loss of power event. open and isolate the grid are:

a) CHFOTF64CC24_1 - 2.08E-6 b) CHFOTF6XCC24_2 - 5.85E-06 c) CHFOTF6XCC24_3 - 3.24E-05 d) CHFOTF6XCC24_4 - 5.97E-06 e) CHFOTF6XCC24_5 - 2.19E-06 f) CHFOTF65CC24_6 - 5.85E-06 g) CHFOTF6XCC34_6 - 4.43E-06 h) CHFOTF6XCC34_7 - 1.51E-05 i) CHFOTF6XCC34_8 - 1.34E-05 j) CHFOTF6XCC34_9 - 1.51E-05 k) CHFOTF6XCC44_1 - 6.63E-04 6 Support for the resolution to Finding 3-28. This resulted in an increase in the CDF The resolution to this finding, i.e. the addition value (basic event HE1D-D-HPIMLW has of dependent HEP HE1D-D-HPIMLW a FV of 4.86E-2).

(Common Failure to Operate High Pressure Injection Systems Given MLOCA Water), is discussed in detail in the response to RAI 1.f.

7 Support for the resolution to Finding 3-35. This change affected the Level 2 model The four (4) main steam lines and isolation of only. Because the containment isolation the MSIVs are added to the model for failure mode is dominated by pre-existing containment isolation (failure to close failure leakage paths and failure of both inboard mode). The main steam line drain valves and outboard containment isolation valves (which are normally closed during at power is a low probability event, no significant operation) were also added to the model for quantitative effects result from the containment isolation completeness (failure to addition of these failure modes to the remain closed failure mode). model.

It should be noted that the failure of the MSIVs to isolate a break outside containment event in the Level 1 PSA model was (and still is) included at the time of the peer review.

Enclosure to NRC-15-0013 Page 18 Item Change Description Quantitative Effect 8 Support for the resolution to Finding 4-5. This The initiating event frequency for the change modified the frequencies for the %TDCAB initiator decreased and the following initiators: initiating event value for the %TDC-BOP increased (by a higher percentage than the a) %TDCAB (Loss of Multiple 130VDC decrease in %TDCAB). This change DC Buses) - From 1.68E-6/rcy to resulted in an increase in CDF based 1.40E-6/rcy. upon:

b) %TDC-BOP (Loss of BOPDC Bus) -

From 1.35E-3/rcy to 2.34E-3/rcy. a) The fact that the initiating event value for %TDCAB was lowered from that in the peer review model but by a lesser magnitude than

%TDC-BOP was increased.

b) The FV value for %TDC-BOP (3.65E-3).

9 Support for the resolution to Finding 4-8. The The model modification resulted in an Turbine Trip (%TX) initiating event frequency increase in CDF (as characterized by the Bayesian update was modified (to incorporate fact that the initiating event frequency a scram that had occurred during low power increased by 0.6%).

operation that may be applicable for at-power operations). The result was an increase in the

%TX initiating event frequency from 8.32E-1/rcy to 8.37E-1/rcy.

10 Support for the resolution to Finding 4-11. This model modification resulted in an The draft peer review model was blocking increase in CDF due to the increased certain SBO sequences from appearing in the contribution from late SBO sequences.

cutset solution (specifically, sequences which involved equipment and human action failures following offsite power recovery). The functional fault tree logic was modified to address this issue.

Enclosure to NRC-15-0013 Page 19 Item Change Description Quantitative Effect 11 Support for the resolution to Finding 4-12. An The model modification resulted in an operator action HE1FRHRSFILLVENT was increase in CDF (as characterized by the added to the HRA calculator and the fault tree FV for the three affected basic events):

model to account for the operators performing a fill and vent to prevent the water hammer of a) HE1FRHRSFILLVENT - 2.48E-3 the system following the receipt of a pump b) PHPHRHRS-DISCH-J - 3.06E-3 restart following a voiding of the system after c) ZLMMRHRSTC2 - 2.35E-3 a previously occurring system stop (e.g. from a loss of offsite power during torus cooling operation followed by a RHR pump start upon power restoration).

The resolution of this finding also included two other items. First, the basic event PHPHRHRS - DISCH-J (system train disabled by water hammer) was changed from a value of 1.0E-3 to 1.0E-2. The water hammer phenomenology was also divisionalized by explicitly modeling the phenomenology for the water hammer event following Division 2 of RHR being in torus cooling (ZLMMRHRSTC2).

12 Support for the resolution to Finding 7-1. This resulted in an increase in the CDF Added modeling for the failure to isolate a value (as characterized by basic event flow path between the HPCI room and the MVFCSUMPT4500F601 having a FV of Division 2 RHR Room. The implementation 2.12E-05).

in the logic model includes basic event MVFCSUMPT4500F601, which represents the failure to isolate the flow path from the HPCI to the Division 2 RHR Room. Gate B72ENORM_TF65 was added to the model to represent the power dependence of the valve while eliminating circular logic issues; this represents a slight conservative treatment since it only credits offsite power support for Bus 72E. Gate RHR2-HPCI-ISOL, RHR2FLOOD-HPCI, and HPCIFLOOD-RHR2 were added to implement the propagation of flooding from the Division 2 RHR Pump Room to the HPCI Room and from the HPCI Room to the Division 2 RHR Pump Room.

Enclosure to NRC-15-0013 Page 20 Item Change Description Quantitative Effect 13 HEP HE1FMSOVINTLCK (which has a This modification resulted in a decrease in guaranteed probability of failure) was set to ATWS CDF and LERF due to the True in the master quantification flag file. elimination of some non-minimal ATWS combinations.

14 The change removed credit for the torus This modification resulted in a CDF cooling repair basic event. Basic event increase for instances where the repair CPFFTC12-LO-REC-OVRD was introduced was credited (typically long term loss of into the model to guarantee failure in each decay heat removal sequences).

modeled instance of CPFFTC12-LO-REC.

CPFFTC12-LO-REC was also set to True in the master flag file.

15 Created conditional probability basic events This modification resulted in an increase for weather centered PLOOP101, switchyard in the frequency for partial LOSP centered LOSP, and switchyard centered sequences. The FVs for the basic events PLOOP101 in a similar manner as for weather which were added to the model are:

centered LOSP. Basic events PHPHCTG1WEATHER101 and a) PHPHCTG1WEATHER101 -

PHPHCTG1SWYD101 were added to the 4.83E-05 model. Also, a separate equipment failure b) PHPHCTG1SWYD101 - 3.54E-05 related CTG complex failure gate (CTG1EQMT) was created to eliminate non-minimal cutsets for grid centered LOSP.

16 Added dependencies for BOP DC Power to This modification resulted in an increase gates CSTSEPMP and CNDS-QUV (gates in frequency for sequences where BOP which credit the operation of the emergency DC batteries are lost (independent of a hotwell supply pump). loss of power to AC busses which support the emergency hotwell supply pump).

17 Created new gate RWLCHI_BATT, which is This change resulted in a decrease in the not supported by battery chargers to represent frequency for sequences where short term DC power input to the HPCIST and RCICST HPCI and RCIC operation is credited trees. Also changed the titles to HPI-4HR and (typically LOSP and SBO scenarios).

T-HPI-4HR. This corrected a modeling issue where the HPCI and RCIC systems would fail in four hours (or less) due to a failure of the Level 8 trip logic in gate RWLCHI (which contained battery charger and AC power dependencies). The insertion of gate RWLCHI_BATT into these sequences resulted in a dependency on the DC batteries for the short term operation of these systems.

Enclosure to NRC-15-0013 Page 21 Item Change Description Quantitative Effect 18 Changed the calculation method for basic This change resulted in a decrease in events PVTSHPCIF035 and PVTSRCICF015 frequency for cutsets involving these basic from Method 5 (periodic test) to calculation events (since the basic event values method 3 (mission time). decreased). The FV for these basic events are listed below:

a) PVTSHPCIF035 - 4.83E-05 b) PVTSRCICF015 - 0.0 19 Increased the scenario time for HRA event This change resulted in a decrease in the HE1FRXPCHWML (based upon an frequency for medium break LOCA examination of MAAP runs). This resulted in sequences. Note that the FV value for a decrease in the HEP value from 4.6E-2 to HE1FRXPCHWML is 9.30E-2.

2.3E-2.

20 Implementation of EDP 36567. This There is no effect on CDF for this change.

modification inserted the modeling for the This is evidenced by the fact that the basic plant modification that inserted the 3rd row of events are not present in the cutset report breakers to the ring bus in the 345 kV (e.g. have a FV of 0.0) for the baseline switchyard. model CDF calculation at a 1.0E-12/yr truncation limit.

21 This change modified the gate T-HPCI-LOCA This increased the frequency of medium for medium LOCA below TAF to preclude water LOCA sequences due to the credit for SBFW as a successful RPV removal of SBFW credit.

depressurization mechanism.

Enclosure to NRC-15-0013 Page 22 RAI 1.h Request the following information regarding the Probabilistic Risk Assessment (PRA) used for the Severe Accident Mitigation Alternative (SAMA) analysis. Basis: Applicants for license renewal are required by Title 10 of the Code of Federal Regulations (CFR) 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of internal events and calculation of core damage frequency (CDF) in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

h. For the peer review findings discussed in Table D.1-21 of the ER:

i. The resolution of Item 3-26 is a repeat of the finding. Provide the resolution of this item.

ii. For Item 4-23, confirm that there are no floor drains from one flood zone to another that have check valves but no sump pump.

Response

i. Below is the resolution of this finding.

A search was performed to determine instances where common cause failures, specifically those for check valves and circuit breakers, were improperly screened. In general, the screening was found to be appropriate.

Two instances were found where the common cause failures were screened improperly.

These are discussed below.

The feedwater line injection check valve 2/2 common cause failure (B2100F010 A/B) would, in fact, fail HPCI, RCIC, SBFW, and condensate injection (in the event of a loss of feedwater). An exception to the modeling assumption used in the Fermi 2 PRA for subsuming common cause check valve failures into the other common cause terms used for multiple train systems (consistent with the guidance provided in ASME/ANS PRA Standard Supporting Requirement SY-A15a) is the treatment of these two FW injection check valves F010A and F010B. The common cause failure of these two check valves to open would preclude success of the following systems:

1) Main Feedwater

2) Condensate

3) HPCI

4) RCIC

5) SBFW

Enclosure to NRC-15-0013 Page 23 Based on this scope of failures, a common cause failure mode for this particular group of two check valves was added to the PRA model.

There was one instance where it was deemed appropriate to model common cause for breaker failure. The breakers (B6, C6, E6, and F6), which isolate the Engineered Safeguard System (ESS) busses 64B, 64C, 65E, and 65F from the grid when the respective diesel is demanded to support these busses, are required for the EDG to perform its mission. These breakers receive an auto-open signal as part of the response to an undervoltage signal on the supported bus and manual operation is not credited. Common cause events were inserted into the model under the Gates CHFOTF64S001BB6, CHFOTF64S001CC6, CHFOTF65S001EE6, and CHFOTF65S001FF6.

ii. The blanket statement that was made in Item 2 in the peer review finding resolution Table D.1-21 for Finding 4-23 (e.g. The check valves in the drainage system are associated with sump pump discharge) does not properly characterize the resolution to the finding. A search of every drainage system check valve in the plant was not performed to verify this statement when the flooding model was developed. Rather, the check valves which were not screened from the evaluation based upon other considerations were examined to determine the presence of sump pumps in conjunction with the check valves. The considerations that were used to limit the scope of examination were:

1) The potential for backflow was analyzed per Section 3.4.4.4.1 of the Fermi 2 Updated Final Safety Analysis Report (UFSAR); the following attributes were included in the equipment and floor drain system designs to prevent backflow flooding through the equipment and floor drainage systems:

a) Independence of building systems to negate the possibility of abnormal occurrences in one building from affecting normal operation in other buildings b) Check valves and manual isolation valves in each sump pump discharge line to prevent backflow to the sump c) Redundant check valves and a manual isolation valve located in both the equipment drain and floor drain 6-inch transfer lines near the secondary containment boundary to prevent backflow into secondary containment.

2) In turbine building scenarios, no credit is taken for the drainage system; therefore, any postulated backflow returning water to the turbine building would merely result in damage to equipment that has already been postulated as failing in the turbine building flooding scenarios (this statement is found in Item 3 of the Finding 4-23 resolution in Table D.1-21 in the Environmental Report (ER)). It should be noted that the Turbine Building and Radwaste Building are connected at the basement level. The SBFW pumps are located in the Turbine Building and are considered to be failed for major Turbine and Radwaste Building Flooding scenarios.

3) Major or moderate severity flooding scenarios in the Radwaste Building result in flooding of the lower levels of the Radwaste and/or Turbine Building. The Radwaste Building is the location to which the sump pumps in the Turbine and Reactor Buildings discharge their flow. Additional equipment failures do not occur as a result of backflow

Enclosure to NRC-15-0013 Page 24 in the these scenarios, since mitigating equipment on the lower levels of the Turbine Building and Radwaste Building is already assumed to fail as a direct effect of the scenario. The potential for backflow to the Reactor Building from the Radwaste Building is addressed in Item 7 of the Finding 4-23 resolution in Table D.1-21 in the Environmental Report.

4) The Reactor Building consists largely of general areas on the upper floors. These floors are connected to one another and to the basement and sub-basement levels by open stairwells (which direct water to the sub-basement level where water accumulation takes place). The reactor building drainage system is designed to route water from flooding in an unimpeded fashion to the reactor building sumps located on the sub-basement level of the reactor building. Therefore, the search for check valves failing to close with the potential to adversely affect reactor building flooding scenarios was limited to the sub-basement level.

5) Flooding in the Auxiliary Building results in direct propagation and/or propagation via the floor/equipment drains to the Turbine Building and/or Reactor Building sumps in a similar manner as described for the Reactor Building (it should be noted that there is more potential for flood water holdup in the Auxiliary Building due to some of the rooms being enclosed). Backflow through check valves has no impact on equipment in the Auxiliary Building during flooding scenarios.

To ensure that the process for screening check valves was appropriately applied, a search for check valves in the Floor and Equipment Drain system was performed from the Fermi 2 equipment list to ensure that there were no outliers. The search confirms that there were no check valves that were omitted from the assessment of potential for backflow.

Enclosure to NRC-15-0013 Page 25 RAI 2.a Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

a. The ER states that FermiV9 upgrade included the conversion of the Risk Management (RISKMAN)-based CAFTA Level 2 to an upgraded CAFTA Level 2 model (based on first principles). Request more detail on the extent of use of the RISKMAN software based individual plant examination (IPE) Level 2 model in the current model and on subsequent changes to the RISKMAN IPE model.

Response

The original Fermi 2 RISKMAN individual plant examination (IPE) Level 2 PRA model was developed by the same technical lead that developed the Fermi 2 CAFTA Level 2 model. There has been significant evolution in procedures since the IPE was submitted, particularly in the replacement of the EPG Rev. 4 containment flood contingency with the Emergency Procedure Guidelines/Severe Accident Guidelines (EPG/SAGs). This significant improvement in severe accident response coupled with plant modifications and improved understanding of severe accident core melt progression have led to changes in the Fermi 2 Level 2 CAFTA model.

The failure modes, phenomenological effects, and dependencies (including HRA) are treated in a similar manner in both models. There is consistency in the treatment of these model aspects.

Some model changes were incorporated based upon updated severe accident core melt progression based on the use of MAAP 4 deterministic code calculations in lieu of the older MAAP 3B code.

The following similarities between the RISKMAN IPE Level 2 and the latest CAFTA Level 2 are noted as follows:

Event trees have the same basic structure

The Level 1 and Level 2 models remain linked to ensure that dependencies on Level 1 failures and successes are appropriately accounted for

The same end state characterization scheme based on a radionuclide release magnitude and time of release are used

Success criteria are essentially the same

The level of detail is the same

The nodal functional fault trees are similar

Enclosure to NRC-15-0013 Page 26

Documentation of the model remains at a very detailed level Differences between the RISKMAN Level 2 IPE and the updated CAFTA Level 2 include the following:

The RISKMAN "equations and rules" are replaced with fault tree logic that provides the dependency connections in CAFTA (i.e. fault tree linking)

The evacuation time and Emergency Action Levels are updated to the latest available for Fermi 2

Radionuclide releases are based on MAAP 4 instead of MAAP 3B. This affects the following:

Core melt progression times RPV breach time Containment failure time Radionuclide release magnitude

Significant differences in the SAGs relative to EPG Rev. 4 containment flood contingency have been incorporated in the CAFTA Level 2 model. This includes no RPV venting for RPV breach cases, more controlled containment flooding, and increased assurance that drywell (DW) sprays will be operated if RPV breach is anticipated.

The Level 1 model that feeds into the Level 2 was updated to incorporate plant modifications, procedure changes, changes in plant specific data, and common cause data plus the shift to a CAFTA model format

The HRA is updated based on the latest procedures, SAGs, crew interviews, and simulator observations

No credit for Reactor Building decontamination factor for containment failure cases is provided in the CAFTA model compared to modest credit included in the RISKMAN IPE Level 2 Information from the RISKMAN Level 2 model was used for several different purposes:

1) To provide a template for the CAFTA event tree attributes (number of event trees, functional top nodes, etc.).

2) To provide for a starting point for the creation of CAFTA functional fault tree logic from RISKMAN equations.

3) To provide an input (along with other sources) for the development of phenomenological failure events and their associated probabilities.

The course of the Level 2 model development for the PRA upgrade, however, was treated as the development of a completely new model from first principles. In many instances, new model logic was put into place which was not present in the RISKMAN templates. The Level 2 model documentation was completely rewritten.

A full internal review of the newly created Level 2 model and associated documentation was performed by DTE (with no reliance on the attributes of the RISKMAN model). The 2012 PRA

Enclosure to NRC-15-0013 Page 27 Review in 2012 examined each of the Level 2 supporting requirements with respect to the upgraded CAFTA model and associated documentation; none of the SR evaluations were carried over from previous peer reviews.

Since the Level 2 model and associated documentation were completely rewritten from first principals, no comparison to the RISKMAN Level 2 model was made and no change log was maintained.

Enclosure to NRC-15-0013 Page 28 RAI 2.b Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

b. Request more information on the containment event trees (CETs) utilized in the Level 2 analysis including the number of CETs, the sequences handled by each CET, and how LOSP and SBO sequences are addressed.

Response

The containment event trees utilized in the Level 2 analysis are:

Containment Event Tree for Class I and III Sequences (CET1)

Containment initially intact Containment Event Tree for Class II and IV Sequences (CET2)

Containment initially breached LOCA Outside Containment Level 2 Event Tree for Class V Sequences (CET3)

Break outside containment There are several nodes where AC power recovery can be crucial to the assessment of core melt progression and radionuclide releases. These nodes are in-vessel recovery (RX), debris cooling recovery (SI), containment flooding (FC), containment venting (VC), makeup remains available (MU), and containment heat removal maintained (HR).

The time to recover AC power is credited in the SI, FC, VC, MU, and HR node.

There may be an infinite number of times when power recovery can be achieved and accounted for in the model. However, consistent with the approach taken in most published PRAs this continuous spectrum has been discretized into a small number of time phases. Ideally, the number of time phases should be on the order of 3 to 5 depending on the degree of conservatisms that can be accepted. The four principal time phases are:

Recovery of AC power during in-vessel core melt progression, when the recovery of AC power could result in halting the melt progression without RPV breach.

Enclosure to NRC-15-0013 Page 29 Recovery of AC power before RPV breach such that water injection would be available to debris as it exits the RPV (SI Node).

Recovery of AC power before the time when high drywell temperature will result in drywell head failure and potential for direct release to the refueling floor (FC, VC, HR).

Recovery of AC power very late for the purposes of restoring containment heat removal capability (FC, VC, HR, MU).

Recovery of offsite power in RX is subsumed into the SI node. This simplification is possible because of the short time frame available in RX.

The time available for AC power recovery during the in-vessel core melt progression and prior to containment shell debris attack is estimated based upon Fermi 2 specific MAAP calculations.

The probability of AC recovery is based on the use of generic offsite AC power recovery applied to Fermi 2 and segregated by cause of the loss of offsite AC power.

The following information is used to develop the Level 2 fault trees to properly account for the AC recovery credited in the Level 1 PRA and the conditional probability of additional AC recovery in the various nodes/Level 2:

The time credited in Level 1 for offsite AC power recovery

The associated probability of AC offsite non-recovery already credited in Level 1 (for the grid, switchyard, plant, and weather related cases)

The allowed cumulative time for AC recovery in Level 2

The associated probability of offsite AC non-recovery

The conditional probability of offsite AC non-recovery that is to be credited in the appropriate nodes given the Level 1 credit already assigned Assumptions The following assumptions are used in the SBO analysis for the CET:

Class IB means that there is a loss of all power to the 4 safety-related buses and balance of plant (BOP) power Bus failures are always asked in Level 1 The rules only allow recovery of offsite AC if all four emergency buses are without AC power.

Recovery of offsite power in RX is subsumed into the SI node. This simplification is possible because of the short time frame available in RX.

Enclosure to NRC-15-0013 Page 30 Power Recovery AC power recovery actions are important in the assessment of accident sequences involving the loss of AC power support to key frontline systems. This section discusses the AC power recovery attributes as they apply to the CET evaluation in Level 2. As such, there is a strong dependence on the interface of these recoveries that have previously been credited in the Level 1 PRA.

The LOSP/SBO event trees are designed to provide a realistic assessment of the postulated sequences including AC power recovery. As is usual in current PRA technology, there is some binning of similar accident conditions to allow the representation of the infinite set of accident conditions by a finite group of sequences. As a result, the selected sequences subsume a spectrum of conditions. For the timing for AC non-recovery in Level 1, critical times are used:

30 min. for sequences for which no RPV injection is available. This subsumes:

Cases with and without RPV depressurization Cases with HPCI and RCIC run failures 4 hr. for sequences with successful HPCI/RCIC operation but battery depletion at 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months .

This subsumes:

Cases with and without RPV depressurization on the heat capacity limit, pressure suppression pressure Cases with 36 gpm to 120 gpm recirc seal leaks 12 hr. for sequences that have RPV injection but heat removal is required via RHR or venting. This is assumed limited by the time to reach primary containment pressure limit.

Enclosure to NRC-15-0013 Page 31 RAI 2.c Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

c. Relative to the CET functional nodes and descriptions presented in Table D.1-4:

i. Discuss the treatment of containment isolation failure sequences and if subsequent early containment failure is modeled.

ii. Discuss the treatment of, and the nature of, credit taken for containment sprays in the Level 2 model.

Response

i. Containment isolation failure is the first functional node in the containment event trees (with the exception of the containment event trees for core damage accident classes where containment failure has already occurred prior to or in conjunction with core damage).

Failure of containment isolation in conjunction with core damage is assumed to result in a High-Early (LERF) release.

For sequences where containment isolation is successful, several other event tree nodes described in Table D.1-4 of the submittal involve phenomenology that contributes to early containment failure (early for the purpose of response to this question being defined as phenomena that occur prior to or at RPV breach). These event tree nodes are:

1) OP - Depressurization of the RPV

2) RX - Core Melt arrested in-vessel

3) GV - Initiation of Combustible Gas Venting

4) CZ - Containment Intact Before and at RPV Breach ii. The Fermi 2 containment drywell sprays provide a means for both pressure and temperature control using a number of sources with a variety of pumps. When RHR pumps are available, the containment sprays can be manually initiated rapidly from the control room. As part of the Severe Accident Guidelines (SAG) development, additional sources of DW spray are identified (e.g., RHRSW, and B.5.b portable pump).

The drywell sprays have a significantly increased role in accident mitigation with the advent of the SAGs. They provide the following:

1) Radionuclide scrubbing

2) Debris cooling in the drywell

Enclosure to NRC-15-0013 Page 32

3) Potential path for containment flooding

4) Containment pressure and temperature mitigation Torus sprays could be useful to a limited extent. However, the low flow rate and the inability to cool debris make them of relatively low importance for the risk analyses. Drywell spray is considered in the evaluation of the CET nodes as described below.

1) SI (Drywell Shell Remains Intact and Ex-vessel Debris Coolability)

Success at this node requires that water is available (greater than 1000 gpm) to the core debris at the time of vessel failure. Shell failure can occur relatively quickly (i.e.,

minutes) following RPV failure if water is not available to quench the core debris. It is assumed in the model that the core debris will come in contact with and fail the drywell shell if water is not available.

Ex-vessel core debris coolability can be considered to be successful if very high containment temperatures, core concrete ablation, and substantial non-condensable gas generation that can result from poorly cooled debris can be prevented. These are considered preventable if on a best estimate basis a continuous water supply is available to the debris with a flow rate of greater than 1000 gpm. The two methods that may provide adequate coolant injection to the debris bed include continued make-up to the RPV and initiation of drywell sprays.

2) FC (Containment Flooding Initiated)

Drywell Spray mode of RHR is one of the methods available to accomplish flooding of the containment.

3) HR (Post Core Damage Containment Heat Removal)

Drywell spray is considered one method for injection into the containment to establish containment cooling. It should be noted that in addition to an operating RHR pump and a path for RHR to inject into the containment, the following must be present for drywell spray to facilitate containment heat removal:

a) A flow path from the suppression pool must be established b) An RHR heat exchanger must be in service c) RHR service water (RHRSW) must be available to provide cooling to the RHR heat exchanger.

Drywell spray is credited as directed by Severe Accident Guidelines in the Level 2 and associated MAAP analysis. However, the impact of the containment pressure reduction associated with running drywell spray prior to RPV breach is a small contributor to the magnitude of the pressure spike (and associated early [energetic] containment failure

Enclosure to NRC-15-0013 Page 33 phenomenology) compared to other factors such as the RPV pressure at the time of breach. It is noted that the primary functions of Drywell spray in the Level 2 model are to a) put water on the containment floor, b) quench ex-vessel debris following vessel breach, and c) to "scrub" fission products from the containment atmosphere.

Enclosure to NRC-15-0013 Page 34 RAI 2.d Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

d. Relative to the definition of accident classes provided in Table D.1-8 of the ER, the accident subclasses for Class IV appear to be combined and subsequently modeled as a single class having a frequency that is 13.3 percent of the total CDF. Request additional information supporting this treatment and the meaning of (not used) in the class definitions.

Response

Subclass IVT represents sequences involving the failure of adequate shutdown reactivity with the RPV initially intact with core damage induced post high containment pressure; this subclass is designated as not used. This designation is not utilized because it is conservatively assumed in the MAAP analysis and subsequently in the accident sequence analysis that the containment will fail at a torus temperature of 260 °F (before the containment fails on overpressure).

Subclass IVV represents sequences involving the failure of adequate shutdown reactivity with the containment vent operating as designed and the loss of RPV makeup occurring at some time following vent initiation. The suppression pool is saturated but intact in this subclass. This subclass is designated as not used because the containment vent is not credited in ATWS sequences (due to the rapid degradation in containment conditions and the short time available for the operating crew to operate the vent).

Subclass IVA represents sequences involving the failure of adequate shutdown reactivity with the RPV initially intact with core damage induced post containment failure. Subclass IVL represents sequences involving the failure of adequate shutdown reactivity with the RPV initially breached (e.g. LOCA or SORV) with core damage induced post containment failure. These accident subclasses are merged into one class (IV). This merging is performed because the modeling of these two subclasses in the Level 2 model is very similar.

Enclosure to NRC-15-0013 Page 35 RAI 2.e Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

e. Page D-56 of the ER describes a situation in which the release category frequency used in the SAMA analysis is less than that in the Fermi PRA documentation. This is described as due to addressing an issue with under counting of Class II contribution in the PRA. Explain the cause of this undercounting and its potential impact on the SAMA analysis.

Response

The quantification truncation level of 1E-12/yr for the Level 2 sequences resulted in under counting of the total release frequency for Class II sequences when compared to the Class II CDF. When the Level 1 and Level 2 Class 2 sequences were recalculated at a lower truncation of 1E-14/yr, the Class 2 undercounting issue was resolved. The base truncation level of 1E-12 truncates out some of the Class 2 releases causing the calculated releases to be slightly less than the incoming CDF. This was resolved in the quantification notebook Level 2 results by adding the frequency of the truncated cutsets to Moderate/Early release category. As discussed on page D-56 of the ER, the exclusion of these truncated cutsets has a minimal impact on the overall consequences. The impact is further reduced when considering that in the SAMA process, the cost benefit is determined by calculating the Maximum Averted Cost Risk (MACR) of a SAMA and then subtracting it from the base model MACR.

The release category frequencies provided in the MAAP to MACCS2 interface notebook (Reference D.1-31 of the ER) are different from the values provided in the PRA quantification document and those used in the SAMA analysis. The release category frequencies used in the MAAP to MACCS2 interface notebook were based on the draft August 2012 model. However, those frequencies were not utilized in developing the relevant release category information that was used in the SAMA analysis. The SAMA frequencies were based on the FermiV9 model which resolved the findings of the peer review. The relevant information utilized from the MAAP to MACCS2 interface notebook was release timing, energy of the release, and the radionuclide release fractions for each release category.

Enclosure to NRC-15-0013 Page 36 RAI 2.f Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

f. Describe briefly the basis for determining the release category for each of the CET end points without the need for Modular Accident Analysis Program (MAAP) analysis for each of the CET end points.

Response

Level 2 sequence end states were assigned to a release category based on key attributes of the Level 2 sequence (e.g., accident class, mitigating strategies employed, location of release point out of containment) that ultimately impact the timing and magnitude of a release. Based on the body of Fermi 2 specific deterministic calculations available (approximately 57 individual Level 2 MAAP cases) and assignment rules, an understanding of accident phenomenology was inferred to allow the available MAAP 4.0.7 calculations to be used to support the determination of radionuclide release end states for all Level 2 sequences with non-negligible frequency.

Representative MAAP cases were then assigned to the release categories. As noted in ER Section D.1.2.2.5, Appendix D of the Fermi Level 2 PRA Notebook describes which MAAP case(s) support the assignment of each CET end state. Appendix D of the Fermi Level 2 PRA Notebook also lists the individual rules used for assigning timing and release magnitude. See the response to RAI 2.g for representative sequences.

Enclosure to NRC-15-0013 Page 37 RAI 2.g Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

g. Provide a discussion of the representative accident scenarios used for the determination of the release characteristics for each of the release categories including:

i. A description of each scenario ii. Bases for the selection of the representative scenarios iii. Steps taken to ensure that the benefit of a SAMA is not underestimated, particularly for scenarios impacted by the SAMA that may not have the dominant frequency but may have a significantly larger consequence than that for the representative scenario.

See for example, the situation that might occur if a SAMA impacted the High Early (H/E) release category scenario represented by MAAP case EF120521 with a cesium iodine (CsI) release fraction of 0.72 (see ER Reference D.1-31 Fermi 2 Nuclear Power Station - Preparation of Input for Ex-Plant Consequence Analysis MAAP to MACCS2 Interface Notebook), which is modeled in the ER using MAAP case EF120520 with a CsI release fraction of 0.24.

Response

Table 2g-1 provides a summary description of the MAAP scenarios chosen to represent each Fermi 2 Level 2 release category. This table also includes the rationale for choosing the selected MAAP scenario. Note that MAAP scenarios are chosen based on a frequency weighted approach (i.e., the MAAP scenario representing the most dominant sequence(s) or bounding the most dominant sequence is typically chosen). Choosing the scenario with the very most conservative conditions (e.g., highest CsI release fraction) may not best represent the release category.

As part of the SAMA evaluation, an uncertainty sensitivity analysis was performed, as documented in Section D.2.4 of the ER. One of the sensitivity cases evaluated uses a multiplier of 2.5 upon the CDF to examine the impact of uncertainty in the PRA model at a holistic level.

This multiplier can be viewed as addressing a wide range of uncertainties associated with the PRA model, including uncertainty associated with release magnitudes for individual release categories, their underlying accident sequences, and the effectiveness of the SAMA candidates upon those sequences. This uncertainty sensitivity analysis helps insure that the benefit of a particular SAMA candidate is not underestimated.

Enclosure to NRC-15-0013 Page 38 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis H/E-BOC - High EF120524 MAAP case EF120524 represents an H/E release Magnitude / Early following a Main Steam Line break outside of Release containment. This MAAP case adequately represents an (Accident Class V, H/E release with an unisolated LOCA outside of Unisolated LOCA containment (Class V accident). This is the only MAAP Outside scenario representative of this release category.

Containment)

H/E - High/Early EF120520 The H/E bin is dominated by Class IV (ATWS) sequences.

Release (With The H/E frequency evolves primarily from sequences Containment IVA-037 and IVA-012. Sequence IVA-037 represents an Isolation) ATWS event with successful RPV depressurization but failure of the wetwell airspace and eventual drywell shell failure; sequence IVA-012 represents an ATWS event with successful RPV depressurization with the wetwell airspace initially intact leading to a failure of the drywell shell.

The reference MAAP cases for these sequences are EF120519 (CsI release fraction (RF) of 2.6E-2) and EF120520 (CsI RF of 0.24) for sequences IVA-012 and IVA-037, respectively. MAAP cases EF120521(CsI RF of 0.72) and EF120522 (CsI RF of 0.10) could also represent this release category bin with similar accident sequences, but case EF120520 is chosen as the representative sequence for the H/E release category bin as it represents one of the primary sequences and the CsI release fraction (0.24) falls between that of cases EF120521 and EF120522. Use of the EF120521 MAAP case (CsI RF 0.72) would be overly conservative since it represents a very small portion of the H/E release category frequency (i.e., <1%).

Enclosure to NRC-15-0013 Page 39 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis H/I - EF120532 The H/I bin is driven by Classes IBL and ID. The High/Intermediate dominant sequences leading to the H/I end state are the Release IBL-040 sequence and ID-040 sequence. The IBL-040 and ID-040 sequences are characterized by successful RPV depressurization followed by a failure of the drywell shell due to shell melt-through.

The reference MAAP cases for these sequences are EF120504 (CsI RF of 9.5E-2), EF120504A (CsI RF of 2.8E-2), and EF120504B (CsI RF of 0.12). Cases EF120504 and EF120504A represent moderate magnitude releases and therefore do not provide adequate release fractions to represent a high magnitude release. Case EF120504B does not provide adequate information to develop plume characteristics for this bin due to the lack of a plateau for the CsI and CsOH release fractions over an acceptable run time. However, case EF120532 has a high magnitude CsI release fraction (0.26) in the intermediate time period and adequately represents (conservatively bounds) this release category bin.

It is noted that the reference MAAP cases in the Level 2 analysis are not necessarily exact models of the sequence, but are instead used along with the Level 2 Release Category rules to assign an appropriate end state to the Level 2 sequence.

H/L - High/Late EF120504C The H/L bin is composed entirely of the IIIC-027 Release sequence (100% of the H/L frequency). The sequence represents a LOCA scenario with reactor makeup available but without successful containment venting thereby leading to a containment failure.

The reference MAAP case for sequence IIIC-027 is case EF120504C which models a drywell shell failure following core damage and SBFW injection. This case has a 0.34 CsI release fraction to the environment. This is a direct representation of the only release category sequence.

Enclosure to NRC-15-0013 Page 40 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis M/E - EF120519 The M/E bin is dominated by the Class IIA sequences.

Moderate/Early The dominant sequence, IIA-012, represents a loss of Release decay heat removal with successful RPV depressurization that leads to a failure of the drywell shell.

The reference MAAP case for sequence IIA-012 is EF120514 (CsI RF of 0.48), which is noted in the Level 2 analysis to be a surrogate reference case because it did not include RPV depressurization during the core melt progression. Experience with MAAP has indicated that if RPV depressurization occurs prior to core melt progression then the radionuclide release will be significantly reduced. The end state for sequence IIA-012 is an M/E release because there is successful RPV depressurization prior to core melt progression. This end state for the representative sequence has been confirmed with a MAAP case that included successful RPV depressurization. MAAP cases in the Level 2 analysis are not necessarily exact models of the sequence, but are instead used along with the Level 2 Release Category rules to assign an appropriate end state to the Level 2 sequence.

This reference case results in a release magnitude that is not moderate. A MAAP case representing a moderate release is required.

MAAP case EF120519 represents an ATWS scenario with no SLC injection quickly leading to a wetwell airspace failure and eventual drywell shell failure. This MAAP case results in a moderate release. This MAAP case can adequately represent the IIA-012 as the selected MAAP case has similar but shorter release timing compared to the IIA-012 sequence. The CsI release fraction for this MAAP case is 2.6E-2. This case is selected for the M/E end state because it well represents the magnitude of moderate release.

Enclosure to NRC-15-0013 Page 41 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis M/I - Moderate/ EF120501A The M/I bin is dominated by the Class IIIC and IA Intermediate Release sequences, primarily IIIC-005 and IA-052. Sequence IIIC-005 represents a LOCA event with successful RPV depressurization and containment venting but no inventory makeup. Sequence IA-052 represents a loss of RPV injection with failure to depressurize the RPV (core melt arrested in-vessel) leading to a loss of containment integrity via a small drywell head failure (27 in2).

The reference MAAP case for sequence IIIC-005 is EF120517C (CsI RF of 1.4E-2). However, this MAAP case results in an early release (vent at 4.0 hours0 days 0 hours 0 weeks 0 months ). A MAAP case with an intermediate release is required. It should be noted that the reference MAAP cases in the Level 2 analysis are not necessarily exact models of the sequence, but are instead used along with the Level 2 Release Category rules to assign an appropriate end state to the Level 2 sequence.

No other representative MAAP cases exist for the Class IIIC sequences. The reference MAAP case for sequence IA-052 is EF120502 (CsI RF of 1.1E-2). MAAP case EF120501A (CsI RF of 1.5E-2) is similar to EF120502 except containment failure occurs via a large drywell failure (2 ft2). Both model the use of drywell sprays until containment failure.

The release timing of EF120501A and EF120502 is intermediate. Case EF120501A is chosen as the representative release fraction for the M/I group since it reasonably bounds the most dominant Class IA sequence.

M/L - N/A The M/L bin release frequency was calculated as Moderate/Late negligible in the Fermi 2 Level 2 PRA model. This group Release is subsumed by the M/I end state.

Enclosure to NRC-15-0013 Page 42 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis L/E - Low/Early EF120529 The L/E release frequency is dominated by the Class ID Release and IA sequences. The end states are driven by sequences (primarily IIIB-061 and ID-022) leading to successful drywell venting with concurrent containment flooding.

The reference MAAP cases for sequences IIIB-061 and ID-022 are cases EF120534 (CsI RF of 7.7E-3) and EF120534A (CsI RF of 3.5E-7). However, case EF120534 does not adequately plateau over the MAAP run and as such does not provide adequate information for plume characteristic development. Case EF120534A results in a low-low magnitude release that does not have appropriate release fractions to represent a low magnitude release. It should be noted that the reference MAAP cases in the Level 2 analysis are not necessarily exact models of the sequence, but are instead used along with the Level 2 Release Category rules to assign an appropriate end state to the Level 2 sequence.

MAAP case EF120529 represents a sequence ending with containment flooding and venting and is judged to adequately represent the L/E release category bin. This MAAP case represents a low magnitude release (CsI RF of 1.5E-3) with early timing. This MAAP case is also similar to the dominant sequences of this release category bin with regards to containment flooding and venting. Therefore, case EF120529 is chosen.

L/I - EF120534B The L/I bin is dominated by the Class IIIC and IA Low/Intermediate sequences. The dominant sequence of the Class IIIC Release contribution, IIIC-022, represents a LOCA scenario ending with successful containment venting and flooding. The dominant sequence of the Class IA contribution, IA-004, represents a sequence with successful RPV depressurization followed by a loss of containment integrity with makeup available.

Enclosure to NRC-15-0013 Page 43 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis L/I - The reference MAAP case for sequence IA-004 is Low/Intermediate EF120501B (CsI RF of 1.8E-3). The listed MAAP cases Release for sequence IIIC-022 are EF120534 and EF120534A (CsI (Contd) RF of 3.5E-7). MAAP case EF120501B results in release fractions that do not provide adequate plume characteristic information due to the absence of a release fraction plateaus over the MAAP run time. MAAP case EF120534A results in a low-low magnitude release and is not appropriate for a low magnitude bin.

MAAP case EF120534 represents a Class ID sequence ending with successful RPV depressurization, containment flooding via the RHRSW system, and a controlled containment vent from the drywell. This MAAP case results in a low magnitude release. This MAAP case was modified to properly capture the release fractions from containment to the environment (i.e., avoid non-conservative release fractions when run beyond the typical Level 2 end point). This case is titled EF120534B (CsI RF of 1.0E-2).

It should be noted that the reference MAAP cases in the Level 2 analysis are not necessarily exact models of the sequence, but are instead used along with the Level 2 Release Category rules to assign an appropriate end state to the Level 2 sequence. A more conservative MAAP case is appropriate if it bounds the reference MAAP case and is similar to the dominant sequence(s).

L/L - Low/Late N/A The L/L bin release frequency was calculated as negligible Release in the Fermi 2 Level 2 PRA model. This group is subsumed by the L/I end state.

Enclosure to NRC-15-0013 Page 44 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis LL/E - Low- EF120534A The LL/E frequency is dominated entirely by the Class IA Low/Early Release sequence IA-001 which represents a loss of RPV injection and successful containment combustible gas venting without suppression pool bypass.

The reference MAAP case for sequence IA-001 is EF120507B. However, this MAAP case represents an OK end state with negligible radioisotope release.

MAAP case EF120534A represents a scenario with no RPV injection coupled with containment vent and concurrent containment flooding. This MAAP case adequately represents sequence IA-001 as containment flooding provides radioisotope scrubbing similar to that of the suppression pool during a combustible gas vent without suppression pool bypass. The CsI RF of 3.5E-7 for this case is representative of the LL/E end state and conservatively bounding of the one dominant sequence.

LL/I - Low- EF120507C The LL/I bin is dominated by the Class ID and IA Low/Intermediate sequences ID-029, ID-031, and ID-024. These three Release sequences represent successful depressurization of the RPV ending in either containment failure or containment venting with and without makeup available.

The reference MAAP case for sequences ID-029 and ID-024 is EF120507 which represents failure of RPV injection resulting in containment failure but with drywell sprays available. This MAAP case was rerun for longer than the original run time of 40 hours4.62963e-4 days 0.0111 hours 6.613757e-5 weeks 1.522e-5 months as well as modified to continue operation of the drywell sprays indefinitely to more thoroughly examine release characteristics for plume definition. RHRSW can be credited for makeup for the dominant Class ID sequences (ID-029, ID-031, and ID-024). This case is titled EF120507C (CsI RF of 2.4E-5) and adequately models this release category.

LL/L - Low- N/A Due to the very low frequency of the LL/L frequency, this Low/Late Release end state is subsumed into the LL/I end state bin. There are no representative MAAP cases for the LL/L release category bin.

Enclosure to NRC-15-0013 Page 45 Table 2g-1 RELEASE CATEGORY BIN MAAP CASE SELECTION BASIS MAAP Release Category Scenario Bin Assigned Assignment Basis CI - Containment EF120525 MAAP case EF120525 is chosen as the MAAP case to Intact represent Tech Spec leakage out of an intact containment.

The MAAP case simulates a Tech Spec leakage of 0.5%

from the drywell. This case is chosen over the MAAP case simulating a Tech Spec leakage of 0.5% from the torus airspace as the DW leakage case results in a higher CsI release fraction (CsI RF of 1.5E-6).

The CsI and CsOH release groups are characteristic of key contributors to early and latent health effects and experience has shown they are key contributors to MACCS2 dose and cost results.

Three radionuclide release categories (i.e., H/E-BOC, H/E, and H/I) 1 compose 92% of the MACCS2 offsite population dose risk and cost risk. Each of the other release categories individually contributes only about 3% or less to the dose risk and cost risk. Table 2g-2 identifies the dominant accident sequences for the three significant release categories. The relative contribution of each sequence to the release category along with the sequences representative MAAP scenario (including the CsI and CsOH release fractions) is included in the table. The highlighted rows indicate the sequence/MAAP case selected to represent the release category. Note that the sequence totals for each release category in the table are less than 100%.

Those sequences representing <2% of the release category frequency were not included since they represent non-dominant scenarios that have negligible impact on dose risk and cost risk.

The H/E-BOC release category has one representative MAAP scenario and is therefore adequately represented by that case.

The H/E release category representative MAAP scenario (EF120520) has CsI and CsOH release fractions that bound those of other H/E sequences with frequency contributions of 51% and 78%,

respectively. For comparison, calculation of frequency weighted CsI and CsOH release fractions from the dominant H/E sequences in the table yields values of 0.28 and 0.18, respectively.

Considering this, the calculated frequency weighted CsI release fraction is approximately 17%

greater than the representative H/E MAAP case, but the representative H/E MAAP case CsOH release fraction is approximately 70% greater than the frequency weighted value. Experience has shown that CsOH is a dominant contributor in terms of land contamination that drives cost risk. The representative CsOH release would be expected to lead to a significantly higher cost risk relative to the weighted value since the representative release is significantly larger (~70%).

1 H/E refers to the to the radionuclide release category with High Magnitude (H) and Early Release Timing (E).

I refers to an Intermediate Release Timing (I). H/E-BOC refers to the category of H/E releases that are the result of a break outside containment (BOC) sequence.

Enclosure to NRC-15-0013 Page 46 This cost risk difference would likely bound any impact from the representative CsI release that is slightly lower than the weighted CsI release.

Table 2g-3 provides the representative MAAP case, release start time, and CsI and CsOH release fractions at 48 and 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months for the three (3) dominant Class II sequences that contribute to the H/E release category. The scenario run time for each of these scenarios is 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months . However, per SOARCA assumptions, mitigating equipment would be available to mitigate any fission product releases after 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months . The weighted CsI and CsOH release fractions after 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months for these three cases are 0.45 and 0.20, respectively, and these sequences make up about 17% of the H/E release category frequency. The calculated Class II frequency weighted CsI release fraction is approximately 88% greater than the representative H/E MAAP case but the representative MAAP case CsOH release fraction is approximately 55% greater than the frequency weighted value and therefore significantly offsets the difference in the CsI release fractions.

Table 2g-4 provides an estimation of the total cost and dose risk from all release categories if the Class II H/E sequences are considered as a separate category consisting of 17% of the total H/E frequency. The Class II sequence H/E category is conservatively assumed to have the characteristic results (i.e., MACCS2 conditional population dose and offsite economic costs) of the H/E-BOC scenario. This results in a 15% increase in total dose risk but only a 0.6% increase in cost risk over the base case values of 4.91 person-rem/yr and 15,600 $/yr, respectively. Since the offsite exposure cost is only approximately 34% of the baseline maximum averted cost risk (MACR), this dose risk increase would be much less significant to the total MACR. This increase would not significantly impact the SAMA analysis. It is also noted that inclusion of the Class II accident sequences into an early release timing is judged conservative. For the three Class II sequences, the release start time for each MAAP case is more than 29 hours3.356481e-4 days 0.00806 hours 4.794974e-5 weeks 1.10345e-5 months following accident initiation. This time period between accident initiation and fission product release provides a considerable amount of time (e.g., more than 29 hours3.356481e-4 days 0.00806 hours 4.794974e-5 weeks 1.10345e-5 months ) for consideration of emergency response actions to mitigate release impacts (e.g., declare a general emergency (GE)).

Although the Fermi 2 emergency action levels (EALs) provide flexibility for the Emergency Director to declare a GE earlier than reactor conditions reaching quantitative set points defined in the EALs (e.g., RPV water level and containment pressure), it has not been credited in the Fermi 2 Level 2 PRA model.

Based on reasonably bounding representative CsI and CsOH release fractions for the H/E release category and the relatively minor contribution of the Class II sequences to the H/E release category, MAAP case EF120520 with its associated CsI and CsOH release fractions is judged to adequately represent the H/E release category.

The H/I release category is conservatively represented by MAAP scenario EF120532 that bounds all the dominant H/I sequence CsI and CsOH release fractions.

Enclosure to NRC-15-0013 Page 47 Table 2g-2 Dominant Release Category Sequence Contributions and Representative MAAP Cases(1)

Percent of Sequence Representative CsI Release CsOH Release Sequence Contribution to MAAP Case Fraction Fraction Release Category Release Category H/E-BOC V-003 100% EF120524 0.34 0.24 Release Category H/E IVA-037 27% EF120520 0.24 0.31 IVA-012 18% EF120519 0.03 0.04 IVA-050 9% EF120523 0.58 0.16 IIA-063 7% EF120516 0.60 0.35 IIA-024 6% EF120514 0.49 0.12 IVA-024 6% EF120522 0.10 0.05 IA-041 4% EF120500 0.44 0.19 IIA-037 4% EF120509 0.33 0.17 IIID-008 4% EF120517B 0.25 0.08 (2)

Release Category H/I IBL-040 EF120504 0.11 0.23 55% EF120504A 0.03 0.06 EF120504B 0.12 0.14 ID-040 38% Same as IBL-040 IIIC-007 3% EF120517C 0.02 0.01 (2)

Class IBL sequence EF120532 0.32 0.84 Notes to Table 2g-2:

(1)

Other release categories (i.e., H/L, M/E, M/I, L/E, L/I, LL/E, LL/I, CI) are not included because of their low contribution to dose risk and cost risk (<=3%)

(2)

A MAAP scenario was chosen for the H/I release category that bounds the dominant sequence representative MAAP scenario release fractions. The MAAP scenario, EF120532, represents an IBL scenario which is the dominant accident class for the release category.

Enclosure to NRC-15-0013 Page 48 Table 2g-3 REPRESENTATIVE MAAP SCENARIOS FOR THE DOMINANT CLASS II SEQUENCES OF THE H/E RELEASE CATEGORY Release Representa- Release Fraction at Release Fraction at Start Sequence tive MAAP 48 Hours 72 Hours Time Case (Hours)

CsI CsOH CsI CsOH IIA-063 (7% of EF120516 33 0.58 0.32 0.60 0.35 H/E frequency)

IIA-024 (6% of EF120514 33 0.42 0.10 0.49 0.12 H/E frequency)

IIA-037 (4% of EF120509 29 0.28 0.16 0.33 0.17 H/E frequency)

Enclosure to NRC-15-0013 Page 49 Table 2g-4 FERMI 2 SAMA Dose RISK and Cost Risk with Separate Class II H/E Release Category Offsite Offsite Population Population Characteristics of Release Mode Economic Economic Dose Dose Risk Cost Cost Risk person- person-Release Category yr-1 $ $/yr rem rem/yr H/E-BOC 5.93E-08 2.18E+07 3.03E+10 1.29E+00 1.80E+03 H/E Class II 5.32E-08 2.18E+07 3.03E+10 1.16E+00 1.61E+03 Other 2.60E-07 8.10E+06 2.80E+10 2.11E+00 7.28E+03 H/I 7.20E-08 9.52E+06 5.26E+10 6.86E-01 3.79E+03 H/L 2.46E-10 8.98E+06 1.67E+10 2.21E-03 4.11E+00 M/E 6.17E-08 2.48E+06 8.39E+09 1.53E-01 5.18E+02 M/I 3.71E-08 2.76E+06 6.10E+09 1.03E-01 2.27E+02 L/E 4.36E-08 2.26E+05 2.26E+07 9.85E-03 9.85E-01 L/I 5.46E-08 2.14E+06 8.25E+09 1.17E-01 4.51E+02 LL/E 5.02E-10 1.31E+04 3.81E+05 6.57E-06 1.91E-04 LL/I 7.75E-08 1.29E+05 4.05E+06 1.00E-02 3.14E-01 Cl 7.83E-07 6.46E+01 1.96E+00 5.06E-05 1.54E-06 Totals 5.64E+00 1.57E+04

Enclosure to NRC-15-0013 Page 50 RAI 2.h Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

h. Provide the duration of the MAAP analysis for each release category and provide an assessment of the adequacy of the time to characterize the releases over the full accident duration.

Response

The MAAP scenario run times are a minimum of approximately 36 hours4.166667e-4 days 0.01 hours 5.952381e-5 weeks 1.3698e-5 months following core damage. This time (36 hours4.166667e-4 days 0.01 hours 5.952381e-5 weeks 1.3698e-5 months after core damage) is set based on the clear actions that have been taken within 36 hours4.166667e-4 days 0.01 hours 5.952381e-5 weeks 1.3698e-5 months for disasters such as Chernobyl, Fukushima, and other non-nuclear evacuation events. In addition, the NRC has also adopted a release duration of approximately 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months after accident initiation in NUREG-1935 (SOARCA) as the basis for their radionuclide release duration calculations. The SOARCA analysis assumes adequate mitigating measures could be brought onsite and be connected and functioning within 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months . For Fermi 2 many of the radionuclide release categories have their calculations extend well beyond 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months because no release fraction plateau is reached or the plateau is reached more than 36 hours4.166667e-4 days 0.01 hours 5.952381e-5 weeks 1.3698e-5 months beyond core damage. A time of approximately 36 hours4.166667e-4 days 0.01 hours 5.952381e-5 weeks 1.3698e-5 months following core damage or at least 40 hours4.62963e-4 days 0.0111 hours 6.613757e-5 weeks 1.522e-5 months from accident initiation as used in some of the radionuclide release categories is a reasonable estimate considering the high probability of mitigating equipment to mitigate an active release after this time.

The assumption is also supported by the US nuclear industrys FLEX strategy. Portable site equipment in addition to readily available offsite resources following a severe accident lend credibility to the likelihood that containment or RPV injection can be provided to cool core debris in a reasonable time period.

The adequacy of the MAAP scenario run time is based primarily on the reduction in rate of release associated with the CsI and CsOH release fractions as a function of time (i.e., a plateau is reached). These two release groups are characteristic of key contributors to early and latent health effects and experience has shown they are key contributors to MACCS dose and cost results. In most cases, the release fractions have reached a plateau by the end of the MAAP run.

In those cases where the release fractions have not fully plateaued by the end of the MAAP run, either the release categories are small contributors to the overall results or a considerable amount

Enclosure to NRC-15-0013 Page 51 of time has passed (i.e., much greater than 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months ) and mitigation actions not explicitly modeled in the PRA are expected to reduce the scenario impacts.

Table 2h-1 provides the MAAP scenario run times for the representative MAAP scenarios for those release categories that are risk significant to the SAMA analysis (i.e., >3% of the total dose or cost risk). The three release categories of H/E-BOC, H/E, and H/I compose 92% of the SAMA offsite dose risk and cost risk. An assessment of the adequacy of the time to characterize the release over the accident duration for these three release categories is provided. For the H/I release category the MAAP case duration was 191 hours0.00221 days 0.0531 hours 3.158069e-4 weeks 7.26755e-5 months . This scenario run time is assumed to be adequate per the SOARCA assumptions noted above. For the H/E-BOC and H/E, release categories the MAAP cases were re-run with a duration of 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months to examine the impact to the release fractions. The CsI and CsOH release fractions are reported for the original scenario run time and 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months after accident initiation. These additional run durations demonstrate that release fractions associated with the risk dominant release categories do not exhibit any significant additional release between their original end time and 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months . Therefore, the use of the 40 hour4.62963e-4 days 0.0111 hours 6.613757e-5 weeks 1.522e-5 months case run time is adequate.

Table 2h-1 SAMA MAAP Case Adequacy Assessment(1)

Release MAAP Case Release Fractions Release Fractions Category Cases Run at End of Scenario at 48 Hours Adequacy Assessment Bin Designator Time CsI CsOH CsI CsOH CsI and CsOH release fractions increase slightly between 40 and 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months .

H/E-BOC EF120524 40 hr 0.34 0.24 0.35 0.25 Trends indicate relative plateau is reached by 40 hours4.62963e-4 days 0.0111 hours 6.613757e-5 weeks 1.522e-5 months and run time of 40 hours4.62963e-4 days 0.0111 hours 6.613757e-5 weeks 1.522e-5 months is acceptable.

CsI and CsOH release fractions increase by a negligible amount between 40 and 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months . Trends H/E EF120520 40 hr 0.24 0.31 0.25 0.32 indicate relative plateau is reached by 40 hours4.62963e-4 days 0.0111 hours 6.613757e-5 weeks 1.522e-5 months and run time of 40 hours4.62963e-4 days 0.0111 hours 6.613757e-5 weeks 1.522e-5 months is acceptable.

H/I EF120532 191 hr Not applicable. Runtime is greater than 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months .

Notes to Table 2h-1:

(1)

Other release categories were not recalculated because of their low contribution to offsite dose risk and cost risk (<=3%)

Enclosure to NRC-15-0013 Page 52 RAI 2.i Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

i. Clarify whether plant-specific fission product masses of the relevant fission product elements were used in the MAAP 4.0.7 analyses instead of the isotopic activity of those elements recommended by the MAAP Users Group.

Response

MAAP 4.0.7 allows users to utilize the input of particular fission product element mass only (input FFPIN=0) or a combination of fission product element masses and fission product nuclide activities (input FFPIN=1). The Fermi 2 MAAP 4.0.7 analysis utilizes both element masses and nuclide activities (input FFPIN = 1) in the MAAP calculations as recommended by the MAAP 4.0.7 code. Use of both element masses and nuclide activities is consistent with the guidance in MAAP-FLAASH #68.

Enclosure to NRC-15-0013 Page 53 RAI 2.j Request the following information relative to the Level 2 analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of accident propagation and radionuclide release in the Level 2 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Level 2 PRA model for supporting the SAMA evaluation.

j. Specify the design basis leakage for containment and compare that leakage rate to the release fractions for the containment intact release category.

Response

The maximum allowable technical specification leakage (Tech. Spec. 5.5.12.c) from the Fermi 2 containment is 0.5% of the containment air weight per day. The Containment Intact MAAP scenario (EF120525) assumes the same leakage rate of 0.5% per day. The MAAP scenario selected therefore appropriately models the maximum allowable leakage out of containment (i.e.,

directly represents a maximum containment leakage scenario).

Enclosure to NRC-15-0013 Page 54 RAI 3.a Request the following information with regard to the treatment and inclusion of external events in the SAMA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of external events in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

a. The Fermi 2 license renewal application utilizes a seismic CDF from the Generic Safety Issue (GSI) 199 assessment in developing the external events multiplier used in the SAMA analysis. In response to NRC requests following the accident at the Fukushima Daiichi Nuclear Power Plant, new seismic hazard curves have been developed for each nuclear power plant site. Based on this information, the Electric Power Research Institute has produced updates to the GSI-199 seismic CDFs. Discuss the impact of using the updated Fermi 2 seismic CDF on the Fermi 2 SAMA analysis.

Response

The seismic CDF used in the Fermi 2 SAMA analysis was based on the Generic Issue (GI) 199 assessment and was determined to be 4.2E-06. Since that time EPRI has produced new hazard estimates for Fermi 2. Using the same method used for the GI-199 assessment and the new hazard estimates a CDF of 2.26E-06 was calculated. This would result in a reduction to the external events multiplier. Since the new seismic CDF is lower than the seismic CDF used in the SAMA analysis, the SAMA analysis is conservative.

Enclosure to NRC-15-0013 Page 55 RAI 3.b Request the following information with regard to the treatment and inclusion of external events in the SAMA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants treatment of external events in the Level 1 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the sufficiency of the applicants Level 1 PRA model for supporting the SAMA evaluation.

b. As noted in the NRC staffs evaluation report on the individual plant examination of external events (IPEEE), and as can be seen in the IPEEE (Tables 4-6 and 4-13), there is a 1.5E-05/year CDF from the remaining areas screened (with CDFs less than 1E-06/year) that was subjected to the same detailed analysis as the unscreened areas.

Because this 1.5E-05/year CDF was not included in the 2.15E-05/year CDF from the unscreened fire areas, provide justification for not including it in the total fire CDF used in the SAMA analysis and/or assess the impact on the SAMA cost-benefit evaluation, particularly with respect to determining the external events multiplier.

Response

It was determined that a sensitivity should be performed by using the CDF contribution from the screened fire compartments in the individual plant examination of external events (IPEEE). This was accomplished by adjusting the fire CDF reduction factor that is included in calculation of the External Event multiplier (EEM). Note that for these calculations the new seismic CDF of 2.26E-06 (based on latest EPRI seismic hazard curves) was used.

For the first calculation the fire CDF continued to use the factor of two reduction, this time including the screened areas from the FIVE analysis:

The revised fire CDF is then calculated as follows:

CDFfire = R * (CDFunscreened + CDFscreened) = 0.5 * (2.15E-5/yr + 1.50E-5/yr) = 1.83E-05 The revised EEM is then calculated as follows:

EEM = (Internal Event CDF + Seismic CDF + Fire CDF) / Internal Event CDF EEM = (1.5E-06 + 2.26E-06 + 1.83E-05)/1.5E-06 = 14.7 For the next calculation the scaling factor for the fire event contribution was based on the exact ratio of internal event CDF values from the model used for the SAMA analysis (FermiV9) to the

Enclosure to NRC-15-0013 Page 56 IPE model (compatible with the quantitative evaluation for the IPEEE). Note that internal flood hazards were removed from the FermiV9 CDF as they were not included in the IPEEE model.

Rfire = CDFV9noflood / CDFIPE = 1.27E-06 / 5.7E-06 = 0.223 The revised fire CDF is then calculated as follows:

CDFfire = Rfire * (CDF unscreened + CDF screened) = 0.223 * (2.15E-05 + 1.50 E-05) =

8.1E-06 The revised EEM is then calculated as follows:

EEM = (Internal Event CDF + Seismic CDF + Fire CDF) / Internal Event CDF EEM = (1.5E-06 + 2.26E-06 + 8.1E-06)/1.5E-06 = 7.9 So averaging the two External Event Multipliers results:

EEM = (14.7 + 7.9) / 2 = 11.3 The results of this sensitivity yield a result that is approximately equal to the EEM value of 11 that is used in the submittal. Therefore, the EEM value used in the submittal is deemed to be appropriate.

Enclosure to NRC-15-0013 Page 57 RAI 4.a Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

a. Section D.1.5.2.3 of the ER indicates that a watershed index of 1 (drained by rivers) was used for all spatial elements for conservatism. Explain why drainage by rivers is conservative compared to drainage by large water bodies.

Response

This is conservative as it assumes that all radionuclides that contaminate a specific area remain in that area and therefore increase dose to the population. Larger bodies of water would be more effective at washing off contamination.

Enclosure to NRC-15-0013 Page 58 RAI 4.b Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

b. The offsite economic cost risk calculation includes an assumption that all crops exposed to radiation are destroyed.

i. Specify the distances from the point of release that apply to this assumption.

ii. Clarify if the assumption includes destroying crops grown in contaminated soil in the years following the radioactive release and/or crops not grown in contaminated soils.

Response

i. This applies to all crops within a 50 mile radius of the plant site. Crops outside of a 50 mile radius of the plant site are assumed to be unaffected.

ii. Milk and crop disposal costs are calculated only for the year of the accident. Beginning in the second year after the accident, acceptability of food production is evaluated by comparing the projected individual dose for year 2 with DOSELONG (in Chronic file, 0.005 effective and 0.15 thyroid dose, Sieverts).

If the projected individual dose for the second year exceeds the dose criteria DOSELONG, the loss of crop production for that year is included in the cost. If the projected individual dose for the second year does not exceed the dose criterion, the production of that year's agricultural production and the production from all subsequent years is allowed. Implicit in the model is the assumption that the food doses resulting from successive years of production do not increase with time. Also, just as with the MACCS food-chain model, the long-term interdiction of farmland applies to all crop categories, and there is no provision for long-term interdiction of a subset of the crops.

If the projected individual dose from the second year of agricultural production exceeds the dose criterion, the projected doses from up to eight successive annual periods are each examined in order to determine if production can be resumed within the first 9 years after the accident. If production cannot be resumed, the crop loss is included in the cost. If the projected doses in each of the years 2 through 9 exceed DOSELONG, no further tests are performed and the farmland is treated as if it were condemned. When farmland is condemned, the associated cost is the market value of the farmland with the dollar values for market value reported on the output listing as FARM DEPENDENT CONDEMNATION COST. If the projected doses for one of the years examined satisfy the long-term dose

Enclosure to NRC-15-0013 Page 59 criterion, agricultural production is assumed to resume in that year. In that case, societal ingestion doses are assessed for the period ending with the duration of the ingestion dose period LASTACUM.

For example, if DOSELONG is satisfied in year 3, and LASTACUM has a value of 10, societal doses are accrued for the period denoted as years 3-10, an exposure period with a duration of 8 years. Since agricultural production was not allowed in the first and second years, economic costs for 2 years of interdiction are assessed using the WASH-1400 economic cost model for farmland. The WASH-1400 model calculates the cost of temporary interdiction as the depreciation of the farm's land and improvements (see variables VALWF and FRFIM, from SECPOP), and as specified by the depreciation rates for land and improvements (see input variables DSRATE and DPRATE, from NUREG-1150).

Enclosure to NRC-15-0013 Page 60 RAI 4.c Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

c. Specify the dollar-to-hectare values and data sources used for farm and nonfarm land in the analysis.

Response

The dollar-to-hectare values for farm land range from $5610/ha to $17,934/ha with an average of

$9,335/ha. The dollar-to-hectare values for non-farm land range from $198,181/ha to

$322,884/ha with an average of $223,430/ha.

Farmland and non-farmland data is taken from SECPOP2000 v3.13.1 county level data from 2002. This data is then multiplied by a factor of 2.1384 to account for inflation from 2002 to 2045. The factor was calculated by extrapolating the consumer price index (i.e. inflation) from current data through 2045 assuming linear growth.

Enclosure to NRC-15-0013 Page 61 RAI 4.d Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

d. For each assessed year, indicate the percentage of missing meteorological data replaced with data substitution.

Response

The meteorological data for the assessed years is as follows:

2003 - 96.41% valid data (3.59% replaced with substitution) 2005 - 99.65% valid data (0.35% replaced with substitution) 2007 - 98.12% valid data (1.88% replaced with substitution)

Enclosure to NRC-15-0013 Page 62 RAI 4.e Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

e. Explain how precipitation events were modeled in the analysis. For rainy days, indicate the duration of precipitation events used in the consequence modeling and if precipitation was spatially dependent. Clarify if there were any assumptions for precipitation in outer boundary cells in the calculation and describe the included assumptions.

Response

Rainfall duration is not an input to the MACCS2 code. For Fermi 2, the weather bin sampling method using full years of meteorological data was used (see Section 5.16 of NUREG/CR-6613).

Input of meteorological data into the MACCS2 code is via the ATMOS input file. This file provides the following input data to the MACCS2 codes:

1. Site specific meteorological data file. Three MET files (i.e., 2003, 2005, and 2007) were used in the analysis.

2. Morning and afternoon seasonal mixing heights.

3. Boundary weather parameters:

- Boundary weather mixing layer height = 1000 meters

- Boundary weather stability class index = 4

- Boundary weather rain rate = 5 mm/hr

- Boundary weather wind speed = 5 m/sec

4. Washout Coefficients (median values from NUREG/CR-6244)

- Washout Coefficients Number One, Linear Factor = 1.89E-05

- Washout Coefficients Number Two, Exponential Factor = 0.664

5. Rain Parameters (recommended values taken from NUREG-1150)

- Number of rain distance intervals for binning = 5

- Endpoints of the rain distance intervals (km) = 3.2187, 6.4374, 11.2654, 20.9215, and 32.1869

- Number of rain intensity breakpoints = 3

- Rain intensity breakpoints for weather binning (mm/hr) - 2, 4 and 6 This data is used by the MACCS2 code to internally determine the rain duration in seconds for the current hour and the rain intensity in millimeters per hour. These values are then used to determine the fraction of the release remaining airborne after wet deposition. These internally calculated values are not provided in the MACCS2 output file. The above data is also used by

Enclosure to NRC-15-0013 Page 63 the MACCS2 code to determine the rainfall intensity distribution for the 16 compass directions and the specified rain distance intervals given above. This rainfall distribution is provided in MACCS2 output. No spatial assumptions were made. Therefore the rainfall distribution determined internally by the MACCS2 code is unchanged. For the precipitation boundary cells, no assumptions were made or input into the MACCS2 code. The only input provided relative to the boundary weather cells is as given in item 3, above.

Enclosure to NRC-15-0013 Page 64 RAI 4.f Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

f. Estimate the sensitivity of the Level 3 results to the assumed plume heat value of 10 megawatts that was used in the consequence analysis.

Response

A sensitivity analysis was performed looking at 0, 10, and 20 MW plume energy releases. The results are shown below:

Plume Energy Modified Maximum Difference (MW) Averted Cost Risk ($/yr) (%)

0 $3,395,276 0.76%

10 $3,369,832 0.00%

20 $3,325,351 -1.32%

Enclosure to NRC-15-0013 Page 65 RAI 4.g Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

g. Explain how the network-wide evacuation speed was computed and how it factored into the total time estimated for evacuation. Indicate if this evacuation speed or time reflects radial distance traveled and if population weighting was included.

Response

A DYNEV II analysis was performed in 2010 for the Emergency Response Organization to understand evacuation time estimates. The DYNEV traffic simulation model is a macroscopic model that describes the operations of traffic flow in terms of aggregate variables: vehicles, flow rate, mean speed, volume, density, queue length, on each link, for each turn movement, during each Time Interval (simulation time step). The model generates trips from sources and from Entry Links and introduces them onto the analysis network at rates specified by the analyst based on the mobilization time distributions. The model simulates the movements of all vehicles on all network links over time until the network is empty.

The computational procedure is outlined as follows:

A link-node representation of the highway network is coded. Each link represents a unidirectional length of highway; each node usually represents an intersection or merge point. The capacity of each link is estimated based on the field survey observations and on established traffic engineering procedures.

The evacuation trips are generated at locations called zonal centroids located within the emergency planning zone (EPZ) and Shadow Region. The trip generation rates vary over time reflecting the mobilization process, and from one location (centroid) to another depending on population density and on whether a centroid is within, or outside, the impacted area. The population density by sector and road capabilities are all considered in this evaluation. This confirms that the analysis is spatially dependent. The average evacuation speed identified in the SAMA includes these factors.

The evacuation model computes the routing patterns for evacuating vehicles that are compliant with federal guidelines (outbound relative to the location of the plant), then simulates the traffic flow movements over space and time. This simulation process estimates the rate that traffic flow exits the impacted region.

Enclosure to NRC-15-0013 Page 66 RAI 4.h Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

h. Specify software codes and the versions used for calculating the core inventory.

Response

Core inventory is taken from DC-6120, Fermi-2 Core Inventory of Accident Source Terms based on NEDC-33043P BWROG Generic BWR Source Terms. This used SCALE 4.4 SAS2H methodology [XSDRNPM, ORIGEN-S] to develop bounding BWR fuel design.

Enclosure to NRC-15-0013 Page 67 RAI 4.i Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

i. Provide the basis for the selection of radionuclides listed in the core inventory. For example, radioisotopes for cobalt are not included in Table D.1-23.

Response

Core inventory is taken from the site MAAP parameter file which references DC-6120, Fermi-2 Core Inventory of Accident Source Terms based on NEDC33043P BWROG Generic BWR Source Terms. Source terms developed in this calculation are utilized by Fermi 2 in development of Regulatory Guide 1.183, Alternative Radiological Source Terms for Evaluating Design Basis Accidents at Nuclear Power Reactors.

Cobalt is not included in the core inventory as it is not a fission product or a transuranic.

Radioactive cobalt in the primary system is primarily due to the presence of stellite in valve seats. Small amounts of cobalt can be released from the stellite which can be activated in the core and lead to an increase in operational dose. Therefore, it is not considered part of the core inventory.

Enclosure to NRC-15-0013 Page 68 RAI 4.j Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

j. Confirm if any changes in future fuel management practices or fuel design are planned or being considered that would influence the core inventory.

Response

There are no changes in future fuel management practices or fuel design planned or being considered. The plan is to continue with 18 month fuel cycles and using GE-14 fuel.

Enclosure to NRC-15-0013 Page 69 RAI 4.k Request the following information relative to the Level 3 PRA analysis. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants analysis of consequences in the Level 3 PRA model. The requested information is needed in order for the NRC staff to reach a conclusion on the acceptability of the applicants Level 3 PRA model for supporting the SAMA evaluation.

d. Specify the thermal power level for the core inventory shown in Table D.1-23. Indicate if changes to this thermal power are anticipated and provide any implications on the cost-benefit conclusions for the SAMA analysis due to differences in the thermal power level.

Response

The core inventory used in the SAMA analysis was based on 3430 MWt, which was the licensed power level at the time of the most recent version of the PRA and at the time the SAMA analysis was performed. In 2014, a license amendment was approved authorizing a Measurement Uncertainty Recapture (MUR) uprate to 3486 MWt. There are no plans currently for further power uprate at Fermi 2.

A sensitivity analysis was performed to determine the impact of the increased power level on the Level 3 consequences analysis and thus impacts on the averted cost risk for each SAMA. The revised SAMA analyses based on the increased thermal power level did not result in any SAMA becoming cost beneficial that was previously not identified as such, including the 95th percentile uncertainty.

Enclosure to NRC-15-0013 Page 70 RAI 5.a Request the following information with regard to the selection and screening of Phase I SAMA candidates. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants basis for the selection and screening Phase I SAMA candidates. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Phase I SAMA selection and screening process for the SAMA evaluation.

a. Relative to Table D.1-2 of the ER:

i. SAMA 001, regarding the addition of direct current power supplies, is identified to mitigate event BTTSEDCSCC33_1 (p. D-11) CC GROUP DC BATTERY FAILS DURING OPERATION 2A, 2B, 2C and others. This is not a Phase II SAMA, as it was screened out on the basis of being already implemented per DTE Electric Company (DTE) addressing NRC Order 12-049 requirements with a FLEX portable, direct current generator. Confirm that the generator is large enough to carry direct current loads without relying on batteries. If not, consider other potential cost-beneficial SAMAs, such as increasing the size of the FLEX generator to carry direct current loads or the use of fuel cells.

ii. For event HE1FUHS1AC001 (p. D-13) Operators manually start MDCT fan, consider a SAMA to automate the starting of the mechanical draft cooling tower (MDCT) fan, unless the design already includes automatic starting.

iii. SAMA 129 is cited in the disposition of event %FL-TB-MCWS-TBXX-M (p. D-14)

Major rupture in Circulating Water pipe or expansion joints in Turbine Building and others. This SAMA is apparently addressed through the External Surfaces Monitoring Program for External Degradation and the Internal Surfaces Miscellaneous Piping and Ducting Components Program for internal degradation.

Clarify if this is an existing program and if the benefit of this program is incorporated in the current Fermi 2 PRA. Identify other actions that might be taken to mitigate this flood event.

iv. SAMA 031 is cited in the disposition of event CPFFHPCIMLTSTART (p. D-15)

HPCI fails during subsequent cycles, FW cntl = F, L8 trip =S. In Table D.1-2, SAMA 031 is said to evaluate upgrading high-pressure coolant injection (HPCI) throttling capability to reduce the number of start/stops required. The title or purpose of SAMA 031, as given in ER Section D.2.3 and in NEI 05-01, is to revise procedures to allow intermittent operations of HPCI and reactor core insulation coolant (RCIC). While a SAMA for upgrading HPCI (and RCIC) throttling capability to reduce the number of start/stops required would mitigate this basic event, SAMA 031 would not. Identify and evaluate a SAMA that would improve current design or operation that would mitigate this basic event.

v. SAMA 009 to reduce the DC dependence between high pressure injection and ADS is cited to mitigate event TPFSHPCIC001A for failure of the turbine-driven high

Enclosure to NRC-15-0013 Page 71 pressure coolant injection pump to start (p. D-18). Explain how this SAMA mitigates the event because common cause failure of direct current would not be included in this event. Include other potential SAMAs that might be applicable in the discussion.

vi. SAMA 101 for improving leak detection procedures is cited for a number of events (e.g., %FL-AB-FPRO-RELAY-N) in Tables D.1-2 and D.1-5. This SAMA is not included as a Phase II SAMA in the cost benefit evaluation in Section D.2.3.

Clarify if this treatment is because of the currently in progress implementation of a risk informed in-service inspection program based on ASME Code Case N-716, which explicitly addresses internal flooding initiators for inclusion in the in-service inspection program. Because internal flooding is a significant contributor to the CDF, discuss the impact of the risk-informed in-service inspection program on this contribution and the potential that other mitigating actions might be cost beneficial.

vii. Discuss the potential for a flood barrier to prevent flood propagation to adjacent flood areas through openings and/or failed flood doors for flood events such as

%FL-AB-FPRO-RELAY-N Nominal rupture in FPS line in AB propagating to Relay Room, %FL-TB-MCWS-TBXX-M Major rupture in Circulating Water pipe or expansion joints in Turbine Building, and %FL-AB-ECW2-B20XX-N, Nominal rupture in RBCCW/EECW Div 2 line in DC Switchgear Room.

viii. The DTE response to EA-12-050, which is stated to include measures that would increase the likelihood of successful containment venting to prevent containment overpressure, is cited in the disposition of event CPFFRBLDFAILDUCTL1 COND. PROB. THAT ADVERSE ENVIRONMENT FAILS EQUIPMENT IN RB BASEMENT (LEVEL 1) (p. D-28). Describe the specifics of this response that would impact this basic event.

Response

i. FLEX, uses a 550 kW AC generator to provide power to key loads, including the DC battery chargers, instrument panels, Residual Heat Removal (RHR) complex and Reactor Recirculation system valves. The FLEX strategy includes capability for refueling the diesel generators so that AC power can be provided to the battery chargers throughout Phase 2 until additional offsite resources become available (i.e., at least 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months or longer).

The chargers are sized so that any charger is able to recharge a totally discharged battery in 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months while supplying maximum predicted load.

The 550 kw AC generator is adequate to carry the necessary loads to mitigate the consequences of an accident. A DC generator, therefore, is not necessary and the high cost of fuel cells makes fuel cells not cost effective.

ii. The benefit for automated starting of the mechanical draft cooling tower fans was conservatively evaluated by setting the probability for basic event HE1FUHS1AC001 (Operators fails to manually start MDCT fans) to 0.0 (guaranteed success) for LOSP scenarios.

Enclosure to NRC-15-0013 Page 72 The resulting CDF from the above model change is 1.441E-06/yr. The averted risk cost relative to the base case for this change is $272,772. The averted cost when considering uncertainty is $681,930. Implementation of this modification will require a significant engineering effort for revision of the load shedding and sequencing logic as well as re-examination of various electrical calculations to confirm the adequacy of the EDG to support the additional load during the load sequencing evolution in addition to the required field modifications. The logic also needs to ensure the fans would not automatically start when ambient temperature is low to prevent damage to the cooling towers due to icing.

Based on this effort this modification is estimated to be approximately $2.4 million. As such, the SAMA is considered not cost beneficial.

iii. The Circulating Water pipe and expansion joints will be subject to inspection in accordance with the requirements of the Internal Surfaces Miscellaneous Piping and Ducting Components Program and the External Surfaces Monitoring Program. These programs will be fully implemented prior to the Period of Extended Operation. Neither monitoring program is credited in the Fermi 2 PRA model.

Using motor-operated valves in place of fast-acting hydraulic or pneumatic positioners has minimized the potential for water hammer in the circulating water piping and the associated rupture of expansion joints.

Also, if the failure were to occur in a circulating water line because of a pressure surge, that same surge would probably trip off the circulating water pumps by means of the pressure switches that protect the system. Flooding would thus not occur.

Another consideration is that part of the response to SOER 85-005, Internal Flooding of Power Plant Buildings (DER 88-0069) included creation of preventive maintenance (PM) events AD70 and AD71 to visually examine the Condenser circulating water inlet and outlet Expansion joints. These events are performed every other outage (2R). Acceptance criteria for these inspections were established by Engineering in memorandum NEPJ-90-0113.

Finally, any of three Turbine Building sump alarms (5D11-RWC, 5D13-RWC, 5D3-RWC) or one Radwaste Building sump alarm (5D19-RWC) would activate in the Radwaste Control Room indicating a leak. Response procedures for each of these alarms direct the operator to investigate the source of the leak. In addition, plant operating procedure 23.702.01, Plant Systems Leak Check Procedure, provides a systematized approach for identifying abnormal inputs into the Radwaste Systems.

Therefore, in light of the design features currently installed to minimize or mitigate the consequences of an expansion joint failure together with an established monitoring program, additional programs to be added, and multiple alarms to alert the Operations staff

Enclosure to NRC-15-0013 Page 73 of an abnormal leakage condition, the additional costs to modify the expansion joint to suppress leakage is not warranted.

iv. A cost-benefit evaluation was performed that addressed a site specific vulnerability, specifically basic events CPFFHPCIMLTSTART and PFFRCICMLTSTART. These basic events are HPCI and Reactor Core Isolation Cooling (RCIC) fail during subsequent cycles, respectively.

Current operating procedures allow for cycling of the HPCI/RCIC (on at Level 2 and off at Level 8) before operators take manual control of the pumps to throttle flow and maintain RPV water at a constant level, thus preventing additional cycling. The proposed SAMA to address these site specific basic events is to revise procedures and training to allow operators to take manual control of HPCI/RCIC earlier in the event in order to prevent cycling on and off of the pumps.

Evaluation of the proposed SAMA in this manner is specifically important to Fermi 2 as failure of HPCI during subsequent cycles (CPFFHPCIMLTSTART) appears in the list of Level 1 Risk Significant Terms. The benefit of the proposed SAMA, performed as stated above to prevent subsequent failures of HPCI/RCIC during subsequent cycles, resulted in an averted cost risk of $15,700 ($39,250 including 95th percentile uncertainty). Since the minimum cost of implementation is assumed to be $50,000 for procedure change, this SAMA is not cost-beneficial.

v. DTE agrees that this basic event is not addressed by SAMA 009. However, another SAMA (SAMA 196) performed an analysis that can be used to determine the benefit of an alternative SAMA which would mitigate this basis event. The analysis of SAMA 196 evaluated an improvement in the HPCI pumps starting performance by a reduction of the HPCI pump failure probability of approximately 40%. SAMA 196 resulted in an averted cost risk of less than $5,000 ($12,500 including 95th percentile uncertainty). There are no credible SAMAs that could be implemented for less than $50,000, and thus no cost-beneficial SAMA is possible to mitigate basic event TPFSHPCIC001A.

vi. SAMA 101 is not evaluated in Phase II because Fermi 2 is in the process of implementing an ASME Code Case N-716 risk informed in-service inspection program. N-716 includes risk informed criteria to provide a method of ensuring that any plant-specific piping locations that are important to safety are identified. Therefore, even though ASME Section XI does not include or require any non-destructive examination (NDE) requirements for Safety Class 3 and non-nuclear safety class piping, N-716 would add such piping if it were determined to be high safety significant based on the results of an internal flooding PRA.

However, based on the Fermi 2 PRA results that it is not likely that the important internal flooding initiators will meet the criteria for inclusion into the N-716 program, a new SAMA has been evaluated.

Enclosure to NRC-15-0013 Page 74 The SAMA would be the implementation of an inspection program for the piping associated with the risk significant internal flooding initiators. This change would be implemented by adding visual inspection of these pipes to the regular shift rounds procedure. The SAMA was evaluated by assuming the inspections would result in a 25%

reduction in the initiating event frequency for these initiators. The resulting CDF from this model change is 1.454E-06/yr. The calculated averted risk cost relative to the base case is

$41,651. There averted risk cost for the uncertainty case is $104,128. The cost of implementation is estimated considering both the development of the procedure change and the actual manpower required to perform the visual inspections on a regular and continuing basis. Use of visual inspections takes credit for the leak before break concept and would involve walking down the risk significant internal flooding piping segments to visually identify any leakage from these piping segments. Identification of leakage would allow repair prior to further degradation that results in more significant leakage and an internal flooding initiator. Assuming that this activity requires 15 minutes per day for one individual to perform the inspection at a cost of $150 per hour, the yearly additional cost of these inspections would be $13,687. The present value of this yearly expense using a discount rate of 7% is $158,600. Assuming a minimum procedure cost of $50,000, the total cost of implementation is $208,600. Therefore, this SAMA is not cost beneficial when considering the base case and the uncertainty case.

vii. The potential for use of flood barriers to prevent flood propagation to adjacent flood areas for these initiators was assessed as described below.

Event %FL-AB-FPRO-RELAY-N - Nominal rupture in FPS line in AB propagating to Relay Room This event is used to model propagation of a flooding from a fire protection system line rupture in the Auxiliary Building to the Relay Room. Event %FL-AB-FPRO-RELAY-M, Major rupture in FPS line in AB propagating to Relay Room, models a similar but larger flooding event. The use of a flood/watertight door between the AB and Relay Room was modeled by setting the two events to false in the fault tree model. This action removes Relay Room failures due to flooding propagating from the AB in the results.

Quantification of the revised model resulted in a CDF reduction of 1.35E-07/yr. The averted risk cost of this change was determined to be $44,267 ($110,667 when including 95th percentile uncertainty). While the averted cost risk (including uncertainty) for this SAMA is greater than the minimum hardware cost ($100,000), a detailed cost estimate would likely drive the implementation cost for this SAMA above $110,000 such that this SAMA would remain not cost beneficial.

Event %FL-TB-MCWS-TBXX-M - Major rupture in Circulating Water pipe or expansion joints in Turbine Building This event is used to model a major rupture of circulating water pipe in the Turbine Building. The event would flood the basement of the turbine building up to an elevation of

Enclosure to NRC-15-0013 Page 75 583 ft at which point the water would run out of the building. A watertight door in the AB provides protection from flooding up to Elevation 588 ft. This door is modeled in the PRA as event CPFFTBLDDOORFAIL, Turbine Building to Auxiliary Building Isolation Door Fails, and has a failure probability of 1E-04. The model was revised to simulate no failure of this door by setting the basic event to false. Quantification of the revised model resulted in a CDF reduction of 4.9E-08/yr. The averted risk cost of this change was determined to be $22,721 ($56,802 when including 95th percentile uncertainty). The averted cost risk of this SAMA is less than the minimum hardware costs of $100,000 and would therefore not be cost beneficial.

Event %FL-AB-ECW2-B20XX-N - Nominal rupture in RBCCW/EECW Div 2 line in DC Switchgear Room A conservative assumption is made in the current PRA model of record (to simplify the modeling) regarding the ability of the operator to terminate a flood emanating from RBCCW/EECW piping in the scenarios of interest before the equipment is flooded in the room to which the flood propagates. The human action HE1FRBCW-FL-ISOL-AB3 is set to 1.0 (guaranteed failure) to simplify modeling in the scenarios for which it could potentially be credited, specifically:

1) Floods which emanate in the DC switchgear and propagate through a breached door to the Division 2 AC Switchgear Room (terminating the flood would prevent damage to the equipment in the AC Switchgear Room following the equalization of flooding level in the combined rooms following door breach).

2) Floods which emanate from the Division 2 AC Switchgear Room and propagate to the DC Switchgear Room (terminating the flood will prevent the breach of the door to the DC Switchgear Room).

Of the two scenarios listed above, Scenario #1 is the most bounding with respect to HRA timing, since the timing to irrecoverable failure is shorter (62 minutes).

To examine the quantitative significance of this event, a more realistic formulation for the HEP for event HE1FRBCW-FL-ISOL-AB3 is derived to establish a new baseline for the SAMA evaluation since the event is set to 1.0 to simplify modeling in the base PRA. A screening value of 0.2 is assigned for this purpose based upon the following:

1) The value is deemed conservative in lieu of the long scenario time and the close proximity of the area to the control room.

2) The cue to the operator (low head tank level) occurs quickly.

3) There are secondary operator cues from the loss of electrical equipment in the room where the flooding begins.

4) The value accounts for the generalized process that would lead to the isolation of the flood (the search for system leakage without specific guidance for targeting

Enclosure to NRC-15-0013 Page 76 areas where the impact of flooding is significant and without explicit isolation instructions).

5) The elevated value implicitly addresses any HRA dependencies associated with the event.

The benefits for physical modifications are examined with respect to the revised baseline (HE1FRBCW-FL-ISOL-AB3 = 0.2) by assuming that the propagation between the switchgear rooms is always prevented (HE1FRBCW-FL-ISOL-AB3 = 0.0).

The result of the SAMA analysis with the above changes is an averted cost of $89,872 dollars. The averted cost for the uncertainty sensitivity is $224,680. While the averted cost is less than the minimum estimated cost of a hardware change ($100,000), the averted cost with uncertainty is greater than the minimum hardware cost. Given that the addition of a curb could conceivably be designed and installed with a cost range of $100,000 to

$200,000, this SAMA would be potentially cost beneficial based on the uncertainty sensitivity.

viii. Event CPFFRBLDFAILDUCTL1 is a conditional failure that is applied when valve T4600-F407, RBHVAC to SGTS isolation valve, fails to close. The failure of this valve to close will provide a pathway for steam and other releases to the third floor Reactor Building which results in the potential for damage to important mitigation equipment.

The NRC has rescinded Order EA-12-050 and replaced it with Order EA-13-109, Order to Modify Licenses with Regard to Reliable Hardened Containment Vents Capable of Operation under Severe Accident Conditions. The Order establishes numerous functional requirements for a hardened containment venting system (HCVS) including the following:

Performance objectives

- Minimize the reliance on operator actions

- Minimize plant operators exposure to occupation hazards

- Account for radiological conditions that would impede personnel actions for event response

- Controls and indications shall be accessible and functional during severe accident conditions, extended loss of AC power (ELAP) and inadequate containment cooling conditions

Design features

- Vent 1% rated power, restore and maintain containment pressure below the design pressure

- Discharge the effluent to a release point above the main plant structures

- Minimize unintended cross flow of vented fluids within a unit and between units

Enclosure to NRC-15-0013 Page 77

- Sustained manual operation from a control panel located in the main control room or a remote but readily accessible location

- Capable of manual operation (e.g., reach-rod or manual operation of pneumatic supply valves from shielded location), accessible to operators during sustained operations

- Capable of operating with dedicated and permanently installed equipment for at least 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months following loss of normal power or loss of pneumatic supplies for air operated components during an ELAP

- Include means to prevent inadvertent actuation

- Include means to monitor the status of the vent system (e.g., valve position indication) from the control panel during ELAP

- Include means to monitor the effluent discharge for radioactivity with indication at the control panel during an ELAP

- Withstand and remain functional during severe accident conditions

- Designed and operated to ensure the flammability limits of gases in the system are not reached or designed to withstand dynamic loading from deflagration and detonation

- Minimize the potential for hydrogen gas migration and ingress into the reactor building or other buildings

- Includes features and provisions for operation, testing, inspection and maintenance adequate to ensure that reliable function and capability are maintained Development of a HCVS system, whether by modification of the existing system or installation of new system, which meets the above requirements will result in a robust and reliable venting system and will also mitigate event CPFFRBLDFAILDUCTL1.

Specifically, compliance with the requirements to minimize unintended cross flow of vented fluids within a unit and to minimize the potential for hydrogen gas migration and ingress into the reactor building will prevent or significantly reduce the likelihood of this event since the flow path from the vent to the reactor building HVAC system will be reduced.

Enclosure to NRC-15-0013 Page 78 RAI 5.b Request the following information with regard to the selection and screening of Phase I SAMA candidates. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants basis for the selection and screening Phase I SAMA candidates. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Phase I SAMA selection and screening process for the SAMA evaluation.

b. ER Section D.2.1 indicates that documentation for 11 previous industry SAMA analyses were reviewed to identify potential SAMA candidates. Discuss how SAMAs from these sources were selected for incorporation into the Fermi 2 Phase I SAMA identification.

Response

The review of industry SAMA analyses was focused on identifying SAMA candidates that proved to be potentially cost-beneficial for other plants. These SAMA candidates were then screened based on their applicability to the Fermi 2 plant design, if they had already been implemented, or if they were covered by a SAMA candidate already retained for a cost-benefit analysis. Potentially cost-beneficial SAMA candidates for other sites were not screened from analysis based on implementation costs.

Enclosure to NRC-15-0013 Page 79 RAI 5.c Request the following information with regard to the selection and screening of Phase I SAMA candidates. Basis: Applicants for license renewal are required by 10 CFR 51.53(c)(3)(ii)(L) to consider SAMAs if not previously considered in an environmental impact assessment, related supplement, or environmental assessment for the plant. As part of its review of the Fermi 2 SAMA analysis, NRC staff evaluates the applicants basis for the selection and screening Phase I SAMA candidates. The requested information is needed in order for the NRC staff to reach a conclusion on the adequacy of the applicants Phase I SAMA selection and screening process for the SAMA evaluation.

c. Documentation on the identification and disposition of all Phase I potential candidate SAMAs does not address vulnerabilities or enhancements from the IPE. While no vulnerabilities were found in the IPE, several opportunities for enhancements were identified. The status of these enhancements is not addressed in the license renewal application ER nor elsewhere in the cited supporting documents. The NRC staff Safety Evaluation Report on the IPE notes that the hard-piped containment venting was installed but indicates that several potential plant improvements were identified for further consideration. Confirm whether or not all IPE identified enhancements have been implemented and further consider those that may not have been implemented.

Response

Actions were taken to implement potential improvements identified for further consideration in the Fermi 2 Individual Plant Examination (IPE). The identified enhancements and actions taken are described below: