ML051750704

Response to Request for Additional Information Regarding License Amendment Request for a Change to the Fire Protection Program

ML051750704
Person / Time
Site:	Millstone
Issue date:	06/23/2005
From:	Grecheck E Dominion Nuclear Connecticut
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
05-066, TAC MC3100
Download: ML051750704 (38)

Text

b P

Dominion Nuclear Connecticut, Inc.

hI ill\toiic I'owcr Stxion borninion 1 tiopc I-crty R o d Wxr:-lord. <:I 06.385 June 23, 2005 U.S. Nuclear Regulatory Commission Serial No.05-066 Attention: Document Control Desk NSS&UDF RO Washington, D.C. 20555 Docket No. 50-423 License No. NPF-49 DOMINION NUCLEAR CONNECTICUT, INC.

MILLSTONE POWER STATION UNIT 3 RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION REGARDING LICENSE AMENDMENT REQUEST FOR A CHANGE TO THE FIRE PROTECTION PROGRAM (TAC NO. MC3100)

In a letter dated April 15, 2004, Dominion Nuclear Connecticut, Inc. (DNC) submitted a license amendment request related to proposed changes to the Millstone Power Station Unit 3 (MPS3) cable spreading area carbon dioxide fire suppression system. In a facsimile dated September 23, 2004, the NRC transmitted a draft of a request for additional information regarding the proposed changes. On November 17, 2004 a teleconference was held to discuss this information with the NRC. DNC's response to the NRC questions is provided in the attachment to this letter.

The additional information provided in this letter does not affect the conclusions of the safety summary and significant hazards considerations discussion in DNC's April 15, 2004, submittal.

If you should have any questions regarding this submittal, please contact Mr. Paul Willoughby at (804) 273-3572.

Very truly yours, Eugene S. Grecheck Vice President - Nuclear Support Services Attachments (1)

Commitments made in this letter: None

Serial No.05-066 Docket No. 50-423 Page 2 of 3 cc: U.S. Nuclear Regulatory Commission Region I 475 Allendale Road King of Prussia, PA 19406-1415 Mr. G. Wunder Project Manager U.S.Nuclear Regulatory Commission One White Flint North 11555 Rockville Pike Mail Stop 08-B-3A Rockville, MD 20852-2738 Mr. S. M. Schneider NRC Senior Resident Inspector Millstone Power Station

Serial No.05-066 Docket No. 50-423 Page 3 of 3 COMMONWEALTH OF VIRGINIA )

1 COUNTY OF HENRICO )

The foregoing document was acknowledged before me, in and for the County and Commonwealth aforesaid, today by Eugene S. Grecheck, who is Vice President -

Nuclear Support Services of Dominion Nuclear Connecticut, Inc. He has affirmed before me that he is duly authorized to execute and file the foregoing document in behalf of those companies, and that the statements in the document are true to the best of his knowledge and belief.

fl Acknowledged before me this 2 3 3 a y of ( /&4c e ,2005.

My Commission Expires:

Notary Public (SEAL)

Serial No.05-066 Docket No. 50-423 ATTACHMENT RESPONSE TO A REQUEST FOR ADDITIONAL INFORMATION REGARDING LICENSE AMENDMENT REQUEST FOR A CHANGE TO THE FIRE PROTECTION PROGRAM (TAC NO. MC3100)

MILLSTONE POWER STATION UNIT 3 DOMINION NUCLEAR CONNECTICUT, INC.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 1 of 34 BACKGROUND:

In a letter dated April 15, 2004,) Dominion Nuclear Connecticut, Inc. (DNC) submitted a license amendment request related to proposed changes to the Millstone Power Station Unit 3 (MPS3) cable spreading area carbon dioxide fire suppression system. In a facsimile dated September 23, 2004,(2)the U.S. Nuclear Regulatory Commission (NRC) transmitted a draft of a request for additional information. On November 17, 2004, a teleconference was held to discuss this information with the NRC. DNC's response to the NRC questions is provided in the balance of this Attachment.

NRC QUESTION 1 The cover letter of the submittal, states:

"During original plant licensing, MP3 requested a deviation from the requirements of the Branch Technical Position (BTP) CMEB 9.5-1, "Guidelines for Fire Protection for Nuclear Power Plants," Revision 2, July 1981, to allow an automatic C 0 2 fire suppression system to be installed in the Cable Spreading Area (CSA) in lieu of the recommended fixed water suppression system."

The original NRC guidance for fire suppression in the CSA is for a fixed water suppression system. MP3 requested and received approval to change from a fixed water system to an automatic C 0 2 fire extinguishing system. The current requested amendment is requesting to change this automatic C 0 2 fire extinguishing system from automatic to manual. How is MP3 tracking cumulative changes that may affect vulnerabilities? (See R.G. 1.174, Section 3.3.2)

DOMINION RESPONSE Currently, plant changes affecting fire protection are assessed against License Condition 2.H, "Fire Protection," to ensure the changes do not adversely affect the ability to achieve and maintain safe shutdown in the event of a fire. Changes to the plant that impact the fire protection program are reviewed by fire protection engineering

(') DNC letter, "Dominion Nuclear Connecticut, Inc., Millstone Power Station Unit 3, License Amendment Request Regarding a Change to the Fire Protection Program (TAC No.

MB8731 dated April 15,2004.

),I1

(*) V. Nerses (NRC) Facsimile to Mr. D. Dodson (DNC), "Draft Request for Additional Information (RAI) to be discussed in an Upcoming Conference Call (TAC No. MC3100),"

September 23,2004.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 2 of 34 to determine the level of impact to the fire protection program in a documented detailed fire protection review. As a result of this review, changes are incorporated as necessary in the MPS3 Fire Protection Evaluation Report, which is part of the MPS3 Updated Final Safety Analysis Report (FSAR).

Although a fire PRA model and updating process do not currently exist for MPS3, DNC performed the following analysis to provide insights into cumulative effects for the proposed change to the MPS3 fire protection program.

The only NRC approved risk-informed MPS3 license change determined to potentially result in an increase in core damage or large early release was an extension of the emergency diesel generator technical specification allowed outage time (AOT) to 14 days. (Refer to Millstone Power Station Unit No. 3 - Issuance of Amendment RE:

Emergency Diesel Generator Allowed Outage Time TAC No. MB3125, dated August 26, 2002.) The risk analysis of this extension shows the changes to the Core Damage Frequency (CDF) to be 8.OE-O7/yr (delta CDF) and changes to the Large Early Release Frequency (LERF) to be 1.5E-O8/yr (delta LERF).

Based on the risk analysis described in the response to QUESTION 11, the delta CDF for the proposed amendment for a manually initiated C 0 2 system in the CSA is 2.3E-O7/yr and the delta LERF is 7.5E-09/yr1resulting in an MPS3 cumulative change in delta CDF and LERF of 1.OE-G/yr and 2.3E-O8/yr respectively. The individual changes are categorized as very small increases in CDF and LERF according to Regulatory Guide 1.I 74 An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis November 2002 (RG 1.I 74) acceptance guidelines. The cumulative change in CDF is categorized as small per RG 1.174, while the cumulative increase in LERF is categorized as very small. In summary, the individual and cumulative changes are considered acceptable due to the small increases in CDF and LERF.

Table 1. Cumulative Changes Summary.

Risk Informed Change Delta CDF Delta LERF EDG 14 Day AOT Extension 8.OE-O7/yr 1.5E-O8/yr Manual C02 Suppression for CSA 2.3E-O7/yr 7.5 E-OS/yr Total 1.OE-OG/yr 2.3E-O8/yr

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 3 of 34 NRC QUESTION 2 Control Room Evacuation Procedure Page 16 of attachment 1 of the submittal states:

"In the event a control room evacuation becomes necessary due to a fire in the CSA, the operators will need to proceed to areas adjacent to the CSA to perform an alternate plant shutdown."

The submittal does not provide the criteria for when control room staff will go to the alternate shutdown panel. Fire damage may affect control room indications in unexpected ways, perhaps even providing nominal readings while plant systems are in a degraded condition. Is staffing the alternate shutdown panel proceduralized (for example, is the panel staffed upon fire detection in the CSA)? At what point will operators go to the alternate shutdown panel?

DOMINION RESPONSE Staffing the alternate shutdown panel (ASP) is proceduralized in the MPS3 emergency operating procedures (EOPs) related to a fire emergency. In the event of a fire emergency, the EOPs direct operators to assess control room (CR) habitability and plant control. The CR is evacuated if either of the following conditions is verified to exist: 1) the CR is judged to be an unsuitable environment in which to continue operating, or 2) instrumentation or control for both trains of safety related equipment required for achieving and maintaining hot standby or cold shutdown, is degraded.

Regarding the habitability assessment, operators are directed by the EOP to verify if fire, smoke, hot gasses or fumes, or carbon dioxide conditions exist to a degree that would warrant CR evacuation. (It should be noted that oxygen and C 0 2 level monitoring systems are installed in the CR, the access pathway, and the east and west switchgear rooms.) Regarding plant control issues, the operators are specifically directed by the EOP to verify instrumentation control and equipment has not degraded for the following systems: auxiliary feedwater, RCS boration, charging, residual heat removal pumps, reactor plant closed cooling water pumps, service water pumps, nuclear instrumentation, and controlled secondary atmospheric steam release. The EOP further directs assessment of 1) the ability of any system or equipment necessary for achieving and maintaining hot standby or cold shutdown to function, and 2) the capability to monitor the performance of these systems. These EOP directions also include assessment of spurious operation of both trains of a system or equipment.

Should a fire in the cable spreading area cause CR habitability concerns or degraded plant control issues as described above, then the CR would be evacuated and the ASP would be staffed. It should be noted that if the C02 system in the CSA has been discharged, the EOP directs operators to don self contained breathing apparatus (SCBA) prior to evacuating the CR. If there has not been a CO2 discharge in the CSA,

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 4 of 34 the operators are directed to obtain SCBA without donning to ensure that the SCBA travels with the operators to the ASP.

Operator training for shutdown from the ASP is conducted once per training cycle (two-year cycle) for all operations shifts. The ASP in the west switchgear room is modeled in the MPS3 simulator and provides an effective training aid for switch manipulations required by the EOPs. The operators are requalified on using SCBAs every year to include transferring from the SCBA to the breathing air system. The SCBA training and the training on the transfer from SCBA to the breathing air system in the switchgear rooms are yearly training requirements of the SCBA program.

NRC QUESTION 3 Sensitivity Study for Various Fire Sizes , Page 19, Table 8, includes the following information:

Scenario I Critical HRR ITime to Dam [min]I q i CSA3

-1 Larger fires are less frequent, but may cause damage before fire brigade activities could be effective. The analysis appears to limit its consideration to the smallest potentially damaging fire. Larger fires, although rarer, may challenge the suppression capability and the conditional core damage probability (CCDP) to a greater extent than a smaller, more common fire and thus may result in higher overall risk. Provide a sensitivity analysis to show that the fire size chosen is limiting. For example, what would be the result if CSA3 had a larger fire that caused damage well before plant staff could perform manual suppression? Use the most limiting fire size identified in the sensitivity analysis.

DOMIN1 0N RESPONSE An evaluation of potential fire size in the CSA was performed by DNC fire protection engineering and concluded that large, or fast growing fires could not be reasonably postulated. However, a sensitivity analysis is provided below for the calculated CCDPs.

(See response to QUESTION 11 (RAI 11) for calculated CCDPs.) The sensitivity analysis on the CCDPs reflects fire scenarios involving damaged cables in addition to those identified in DNCs original April 15, 2004 submittal.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 5 of 34 The following assumptions are made:

Increasing the CCDPs from the base values calculated in RAI 11 is equivalent to assuming larger fires affecting more cable trays in the location of the fire scenario.

Any hotwork activity in the CSA would have required disabling the previously configured automatic C 0 2 system. Therefore, hotwork fire scenarios do not present a change in fire risk in the room when evaluating a system change from automatic to manual C02.

General transient fire scenarios may not credit the presence of a qualified fire watch. As such, manual suppression activities may begin at relatively longer times when compared with any hotwork fire scenarios. The sensitivity analysis for fire size in general transient fire scenarios assumes that the fire brigade will respond to the fire at the detection time. Consequently, manual fire suppression activities will begin at the time of fire brigade arrival at the CSA.

DNCs April 15, 2004 submittal documents that a 6.0 MW fire with a duration of approximately 45 minutes is necessary for generating room-wide damage in the CSA. This is equivalent to a CCDP of 1.0. Such intensity will result from a combination of an ignition source and cable fires. That is, given that fixed and transient ignition sources in the CSA are expected to have relatively lower heat release rates, a contribution of cable fires is necessary to reach and maintain an intensity of 6 MW for a duration of tens of minutes.

Sensitivitv Analvsis for Fire Heat Release Rate The fire frequencies calculated in the original April 15, 2004 submittal are applicable for this sensitivity analysis. That is, only the CCDP for each selected scenario will be varied in order to explore the impact on fire risk in the CSA. It is assumed that larger CCDPs would result from fires propagating from the ignition source to cables trays in the vicinity (larger heat release rates). The fire frequencies do not need to be varied in the sensitivity analysis for the following reasons:

1. The apportioning factors applied to the generic frequency (location, ignition source and geometric weighting factors) were specifically calculated for each scenario. The only factor amenable for sensitivity analysis would be the geometric factors. However, this factor was assumed to be 15% of the floor area for each pinch point, which, given the size of the room, is a conservative estimate.

2. Severity Factor: As indicated in the original submittal, the frequencies are based on the lowest heat release rate that is expected to generate target damage or

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 6 of 34 ignition. Therefore, each frequency represents a fire that is fully capable of propagating. The parameter governing how the fire frequency changes with fire size is the severity factor. Larger postulated fires would result in lower severity factors, which would result in lower frequencies and still will be able to propagate.

Therefore, the frequencies listed in the submittal are considered the highest possible values for damaging and propagating fires.

3. Non-suppression Probability: The non-suppression probability parameter in the fire frequency was calculated using an event tree which models different types of fire brigade or hardware failures. In addition, one of the event tree branches models the event of room-wide damage due to the inability of the fire brigade to control the fire by any of the available means of suppression. Therefore, the calculated fire frequencies consider events involving propagating fires that may cause room-wide damage.

The following table presents the calculated CCDPs for the postulated fire scenarios.

Automatic C02 Manual C02 Delta ACDF Fire Frequency Fire Frequency Fire Frequency CCDPs (per year) (per year) (per year) (Per year)

Scenario 2 9.80 E-09 8.90 E-08 7.92 E-08 1.o 7.92E-08 Scenario 3 2.1 OE-08 1.90E-07 1.69E-07 5.82E-01 9.84E-08 Scenario 4 1.20E-09 1.90E-08 1.78E-08 4.83E-05 8.60E-13 Scenario 5 1.60E-09 3 .OO E-08 2.84E-08 1.o 2.84 E-08 Scenario 6 I 6.1OE-11 6.1OE- 10 5.49E-10 1.61E-06 8.84E-16 Scenario 7 I 4.20E-08 8.20 E-07 7.78E-07 3.00E-02 2.33E-08 I Total I 2.29E-07 With these CCDPs, the estimated room CDF for transient fires is 2.3E-7/yr. Notice that for scenarios 2 and 5, the CCDP has a value of 1.0. Therefore, the sensitivity analysis explores the effect of higher CCDPs in scenarios 3, 4, 6, and 7 on the room CDF.

Again, higher CCDP values represent the effect of a fire propagating to additional cable trays in the CSA. A base case and five additional cases have been postulated in the sensitivity analysis. The cases are:

1. BASE CASE: Room CDF is calculated using the calculated CCDPs in the response to RAI 11.

2. CASE 1: Scenario 3 is assumed to have a CCDP value of 1.O.

3. CASE 2: Scenario 4 is assumed to have a CCDP value of 1.O.

4. CASE 3: Scenario 6 is assumed to have a CCDP value of 1.O.

5. CASE 4: Scenario 7 is assumed to have a CCDP value of 1.O.

6. CASE 5: All scenarios are assumed to have a CCDP value of 1.0. This case represents room-wide fire damage.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 7 of 34 Table 3 presents a summary of the results of the sensitivity analysis. Table 4 lists the calculations for each of the cases analyzed in the study. In the extreme case of room-wide fire damage, where the CCDPs from all scenarios are assumed to be 1.O, the total room CDF for general transient fires is 1.1E-G/yr.

Table 3. Summary of Sensitivity Analysis Results.

Summary Transient Total Room CDF (per year)

Base Case 2.3E-07 Case 1 3.OE-07 Case 2 2.5E-07 Case 3 2.3E-07 Case 4 9.8 E-07 Case 5 1.1E-06 Table 4. Sensitivity Analysis.

Fire Frequency (/yr) Base Case

[ CDF Transient Automatic C02 Manual C02 Delta CCDP (per year) 9.8E-09 8.9 E-08 7.9E-08 1.OE+OO 7.9 E-08 2.1 E-08 1.9E-07 1.7E-07 5.8E-01 9.8E-08 1.2E-09 1.9E-08 1.8E-08 4.8E-05 8.6E-13 1.6E-09 3.0E-08 2.8E-08 1.OE+OO 2.8E-08 6.1E-11 6.1E-10 M E - 10 1.6E-06 8.8E-16 4.2E-08 8.2E-07 7.8E-07 3.OE-02 2.3E-08 Total CDF: 2.3E-07 Fire Frequency (/yr) Case 1 CDF Transient Automatic C02 Manual CO2 Delta CCDP (per year) 9.8E-09 8.9E-08 7.9E-08 1.OE+OO 7.9E-08 2.1 E-08 1.9E-07 1.7E-07 1.OE+OO 1.7E-07 1.2E-09 1.9E-08 1.8E-08 4.8E-05 8.6E-13 1.6E-09 3.OE-08 2.8 E-08 1.OE+OO 2.8E-08 6.1E-11 6.1E-10 5.5E-10 1.6E-06 8.8E 4.2E-08 8.2E-07 7.8E-07 3.0 E-02 2.3E-08 Total CDF: 3.OE-07

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 8 of 34 I 4.2E-08 I 8.2E-07 I 7.8E-07 I 3.OE-02 I 2.3 E-08 I Total CDF: 2.5E-07 I Fire Freauencv Uvr) I Case 3 I CDF Transient Automatic C02 Manual C02 Delta CCDP (per year) 9.8E-09 8.9E-08 7.9 E-08 1.OE+OO 7.9E-08 2.1E-08 1.9E-07 1.7E-07 5.8E-01 9.8E-08 1.2E-09 1.9E-08 1.8E-08 4.8E-05 8.6E-13 1.6E-09 3.OE-08 2.8E-08 1.OE+OO 2.8E-08 6.1 E-11 6.1E-10 5.5E-10 1.OE+OO 5.5E-10 4.2E-08 8.2E-07 7.8E-07 3 .OE-02 2.3E-08 I I I I Total CDF: I 2.3E-07 I Fire Frequency (/yr) Case 4 CDF Transient Automatic C02 Manual C02 Delta CCDP (per year) 9.8E-09 8.9E-08 7.9E-08 1.OE+OO 7.9 E-08 2.1 E-08 1.9E-07 1.7E-07 5.8E-01 9.8 E-08 1.2E-09 1.9E-08 1.8E-08 4.8E-05 8.6E-13 1.6E-09 3 .OE-08 2.8E-08 1.OE+OO 2.8E-08 6.1 E-11 6.1 E-10 5.5E-10 1.6E-06 8.8E-16 4.2E-08 8.2E-07 7.8E-07 1.OE+OO 7.8E-07 Total CDF: 9.8E-07

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 9 of 34 Fire Frequency (/yr) Case 5 I CDFTransient I Automatic C02 Manual CO2 Delta CCDP (per year) 9.8 E-09 8.9E-08 7.9E-08 1.OE+OO 7.9E-08 2.1 E-08 1.9E-07 1.7E-07 1.OE+OO 1.7E-07 1.2E-09 1.9E-08 1.8E-08 1.OE+OO 1.8E-08 I 1.6E-09 I 3.OE-08 I 2.8E-08 I l.OE+OO I 2.8E-08 I 6.1 E-11 6.1 E-10 5.5E-10 1.OE+OO 5.5E-10

.I 4.2 E-08 I

8.2E-07 I

7.8 E-07 I

1.OE+OO Total CDF: I 7.8E-07 1.1E-06 4

I The effects of larger fires in each pinch point, which are those that would propagate to additional cable trays before being controlled, are captured only in the CCDP parameter of the fire risk equation. That is, the fire frequencies calculated in DNCs April 15, 2004 submittal are based on the smallest fire intensity capable of propagating to adjacent intervening combustibles (cable trays) and are evaluated considering different fire protection alternatives, including fixed systems and fire brigade failures leading to room wide damage. Therefore, the sensitivity analysis was conducted by varying the CCDP parameter calculated in RAI 11 for each fire scenario.

The results of the sensitivity analysis indicate that the worst case, in which all the postulated fire scenarios result in a CCDP of 1.O, the total room CDF is 1.1E-G/yr. This worst case CDF change is considered small according to RG 1.174 acceptance guidelines. For the other sensitivities, the change in CDF (between 3E-7/yr and 1E-G/yr) is considered very small according to RG 1.174.

NRC QUESTION 4 CSA Cabinets Sheet 10, of Attachment 3 states:

Therefore, fires in these two isolation panels do not damage additional equipment and no fire scenario related to the isolation panels is required to be postulated.

Describe the isolation cabinets in the CSA, including plant SSCs that could be affected by the panel being fire damaged or spurious actuations in the panel. Provide an explanation of why damaging equipment in one cabinet would not challenge safe shutdown. Provide an analysis that a fire originating in one cabinet could not spread to nearby cabinets or cable trays along the cables connected directly to the isolation panels. Also, provide analysis that a fire in one cabinet could not affect adjacent cabinets.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 10 of 34 DOMINION RESPONSE The circuits associated with the isolation cabinets are generally used for isolation of annunciator signals from safety related circuits. Failure of these circuits would not impact operation of safe shutdown components. The panels are supplied with 125 vdc control power to operate the isolation relays located within the panels. A review of the MPS3 BTP 9.5-1 Compliance Report list of safe shutdown components has determined that only one circuit from each electrical train could be impacted from a fire within the panels. Pressurizer heater 3RCS*H1A has a cable routed into 3CES*PNLBS30 and 3RCS*HlB has a cable routed into 3CES*PNLBS3P. If one train were to be damaged from a fire within one cabinet, the other train would be available. This has no impact on alternate shutdown from outside the control room since the required components will be isolated from the fire area.

The remaining portion of this response describes the growth and propagation analysis for fires in the isolation cabinets. The analysis covers the extent of fire propagation and fire induced thermal damage assuming ignition of the isolation panels. This analysis supports the conclusion of the original submittal, which indicates that isolation panel fires in the CSA will not result in risk significant consequences.

Fire Propaqation Analvsis for Isolation Panels in the CSA As described in the original submittal, there are two isolation panels in the cable spreading area, one associated with the orange train and the other associated with the purple train. The two isolation panels have the following characteristics:

0 Size: 7.5 ft long, 2.5 ft wide, and 7.5 ft high 0 Each cabinet has six closed doors, 3 in front and 3 in the back 0 Cabinets do not have vents 0 Cables connecting to the cabinets are inside metal conduits 0 The two isolation cabinets are 18 ft apart (see Figure 4a).

Figure 1 provides a collection of pictures of these two isolation cabinets. Notice in both cabinets the three closed doors and all the connecting cables inside conduits.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 11 of 34 Figure 1: Purple and orange isolation cabinets:

Orange isolation Cabinet

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 12 of 34 PurDle Isolation Cabinet (3CES

• PNL BS1P, PNL BS2P. PNL BS3P). The two closest cable trays to the purple isolation cabinet run parallel to the cabinet length approximately 4.5 ft from the cabinet. The two trays are stacked. The bottom tray is 13 ft from the floor. The second tray is approximately 1.5 ft above the first one. Finally, the bottom tray is associated with the purple train (3TC417P, and 3TC416P). No orange cables are located near the panel. Figure 2 provides a pictorial representation of the described geometry.

Figure 2: Purple isolation cabinet and nearby cable trays (Drawing not to scale)

(Looking East) 4.5 ft Purple: 3TC417P, 3TC416P isolation 13 ft cabinet Oranae Isolation Cabinet (3CES

• PNL BS10, PNL BS20, PNL BS30). The orange isolation cabinet has two cable tray stacks parallel to its length, one on each side of the cabinet. In addition, the cabinet also has a nearby panel and a purple conduit along a nearby column.

One of the stacks has 8 trays. The bottom tray is around 7 ft from the floor, and trays are around 1 ft apart. The first, and fourth through seventh trays in the stack are associated with the purple train. The second stack consists of three trays. The bottom tray is 10 ft from the floor, and trays are also around 1 ft apart. The second tray in this stack is associated with the orange train.

The panel (3MSS*PNLlO) is located 2.3 ft from the isolation panel. The panel, associated with the orange train, contains no active components. The main steam isolation valve test switches were removed many years ago. This panel contains energized cables, and acts only as a very large junction box, mounted on the floor. The size of this panel is 3.6 ft x 3.6 ft x 4 ft high. Figures 3 and 4 provide a pictorial representation and pictures of the trays and equipment near the isolation cabinet respectively.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 13 of 34 Figure 3: Pictorial representation of trays and equipment nearby orange isolation panel (Drawings not to scale)

SIDEVIEW Emcm m a lS',4'h through 7'h trays are m purple looocl zndtray is orange:

lOOOCX m a 3TC4550, and 25ft 3TCC4540 rn -lcmm 7 ft Dc A Orange isolation 7 ft panel 10 ft Conduit 3CX95OP Looking West Cable tray 1 3MSS*

PNL10 1 Cable tray TOP 1

VIEW 2.3 ft I

2.5 ft I isolation panel 2.5 ft L

Conduit

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 14 of 34 Figure 4: Equipment and trays near orange isolation cabinet. The purple isolation cabinet is also shown in fig. 4a, on the far left.

Figure 4a: Side view Fire in Isolation Panels. A heat release rate of 95 kW has been selected for analyzing the impact of a fire in any of the isolation panels. NUREG/CR 4527, An Experimental Investigation of Internally Ignited Fires in Nuclear Power Plant Control Cabinets: Parts 1

& 2, which document fire intensities from fires inside electrical cabinets list heat release rates for experiments ST7 and ST8 as 93 KW and 95 KW. Experiments ST7 and ST8 were conducted in a cabinet with closed doors and qualified cables. The selected value of 95 kW is the highest heat release rate measured in these two tests. This selection is considered to be conservative because the cabinets in the experiments had top and bottom openings, which is not the case in the isolation panels in the CSA.

Damaae to EauiDment Near the Isolation Panels. A zone of influence (ZOI) for 95 kW fires in these two isolation cabinets is described in Table 5. Any equipment within the distances listed in Table 5 is expected to be damaged by fire. Based on the location of nearby equipment and cables, no targets are located within the ZOI. The resulting ZOI was calculated assuming damage criteria for IEEE qualified cable (damage temperature 625 OF, critical heat flux 11 kW/m2 as specified in NRC Inspection Manual Chapter 0609, Appendix F, Fire Protection Significance Determination Process, (Attachment 6) dated February 28, 2005).

Table 5. ZOI for Isolation Cabinets in the CSA.

Region ZOI Model in Five-Rev1 Library Flames (ft) 2.75 Heskestads flame height correlation Plume fft) 3.4 Heskestads Dlume temDerature correlation Ceiling Jet (ft) 0.02 Alperts ceiling jet correlation Flame Radiation (ft) 1.7 Point source radiation model Smoke Layer (OF) 104 MQH Room temperature correlation

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 15 of 34 The calculated flame height, resulting from Heskestad's flame height correlation, is 2.75 ft. The inputs for the model are listed in Table 6. A fire diameter of around 2 ft was assumed for the analysis. This is nearly the width of the panel. Increasing this diameter will result in lower flame heights.

Table 6. Flame Height Analysis using Heskestad's Correlation.

IEPRl's Five-Rev1 Analvsis Heskdestad's Flame Height Correlation I

Inputs Fire diameter [m] 0.6 HRR [kW] 95 Results Flame height [m] 0.84 Table 7 lists the input parameters used in Heskestad's fire plume correlation. This model was selected for calculating the distance at which a cable target is expected to be exposed to damaging gas temperatures. Based on the analysis, the plume temperature is 315 "C 3.4 ft above the fire source. As in the case of the flame height correlation, a fire diameter of 2 ft was assumed. Increasing the fire diameter will reduce the fire plume temperatures as a function of height above the fire. Therefore, the distance above the fire at which a target will see damaging temperatures will be reduced.

Table 7. Plume temperature Analysis using Heskestad's Correlation.

EPRl's Five-Rev1 Analysis Heskestad's Plume Temperature Correlation Fire location factor 1 HRR rkWl 95 Fire elevation [m] 0 Target Elevation [m] 1.03 Radiation Fraction 0.40 Fire Diameter [m] 0.6 Results IPlume Temp [C] I315 I

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 16 of 34 A ceiling jet analysis provides the critical distance at which a target in the ceiling jet would be exposed to damaging gas temperatures. Table 8 lists the inputs to Alpert's ceiling jet correlation. The ceiling jet analysis indicates that a target would need to be near the fire plume to be exposed to damaging temperatures if the fire is located on top of the cabinet. That is, no target in the ceiling jet is expected to be damaged.

Table 8. Ceiling Jet Analysis using Alpert's Ceiling Jet Correlation.

Five-Rev1 nalvsis Ipert's Ceiling Jet Inputs Ambient temperature [C] 20 Fire location factor 1 Horizontal radial distance

[m] 0.025 ResuIts Ceiling Jet Temp [C] 322 As illustrated in Figure 4b, a purple conduit is located near the floor 2.5 ft away from the orange panel. If an isolation panel fire at floor level is assumed, a cable would need to be 1.7 ft away from the fire to receive a damaging level of flame radiation. Therefore, the cable(s) in the conduit are not expected to be damaged because it is beyond the critical distance, This analysis does not credit any protection the conduit may provide to the cable by shielding incident thermal radiation. The point source radiation model was selected for the analysis. Table 9 lists the inputs to the point source radiation model.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 17 of 34 Table 9. Point Source Thermal Radiation Model.

h EPRl's Five-Rev1 nalvsis IPoint Source Flame Radiation Model I Fire heat release rate IlkWl I 95 0.40 Results Heat flux lkW/m21 11.2 Finally, a 95 kW fire is not expected to produce damaging hot gas layer temperatures.

Temperatures were calculated assuming both a room with an open door and a room with closed doors. The CSA is not mechanically ventilated. These two calculations do not predict a room temperature rise larger than 68 O F . The inputs for the models selected for room temperature analysis are listed in Table 10.

Table 10. Room Temperature Analysis using Several Room Temperature Correlations.

Open Door Closed Door EPRl's Five-Rev1 Beyler Temp Correlation MQH Temp Correlation

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 18 of 34 In the case of the purple isolation panel, the closest cable tray is located more than 4 ft away from the panel, which is outside of the radiation zone of influence. In addition, there are no orange-train targets near the isolation panel.

In the case of the orange isolation panel, the closest equipment is a purple conduit 2.5 ft away, an orange switch panel, 2.3 ft away, and an orange cable tray 2.5 ft away. All of this equipment is outside the zone of influence.

In summary, no equipment or cables are expected to be damaged if a fire occurs in the isolation panels.

Fire Propaaation from One Isolation Panel to the Other. NUREG/CR-6850 or EPRl 1008239, EPRI/NRC-RES, Fire PRA Methodology for Nuclear Power Facilities. Vol. 2:

Detailed Methodology, April 2005, page G-8, indicates the following: Electrical cabinets that are not vented do not propagate a fire. Penetrations listed above are not considered vents. It is assumed that in the absence of other ventilation, penetrations will not allow sufficient air exchange to replace oxygen consumed by the fire, and an incipient fire will self-extinguish when there is no longer enough oxygen to support combust io n. I As such, a fire in both purple and orange isolation panels are not expected to propagate for two specific reasons: 1) the fire is not going to be sustained inside a closed cabinet with no vents, and 2) cables connecting to the panels are inside metal conduits, which are not expected to propagate flames.

Sensitivitv Analvsis. A sensitivity analysis, which postulates additional damages to the ones identified in the base case, was conducted in the fire modeling study for the orange isolation panel because purple conduit is located 2.5 ft from the isolation cabinet. As illustrated in Figure 4b, the purple conduit is within the height of the isolation panel, which may subject it to flame radiation, if the fire breaches the integrity of the panel walls.

Using the point source flame radiation model, a fire intensity of 200 kW would be required to generate an 11 kW/m2 heat flux. This fire intensity is twice the heat release rate measured in ST7 and ST8 described in NUREG/CR 4527, An Experimental Investigation of Internally Ignited Fires in Nuclear Power Plant Control Cabinets: Parts 1

& 2. This is considered a large margin given that a fire in a closed cabinet with no vents cannot be sustained. Table 11 lists the input for the point source radiation model analysis.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 19 of 34 Table 11. Point Source Flame Radiation Calculation using FIVE-Rev1.

IEPRls Five-Rev1 Analvsis IPoint Source Flame Radiation Model Fire heat release rate [kW] 200 Radiation fraction 0.40 Distance from flames [m] 0.76 Results Heat flux TkW/m21 I 11.03 In summary, a fire propagation analysis was conducted for the two isolation panels in the CSA. The propagation analysis consisted of: 1) a detailed description of fire scenarios involving either one of the isolation panels as fixed ignition sources, and 2) a determination of the extent of fire damage assuming ignition of either of the isolation panels.

The detailed description of the fire scenarios included pictures and drawings describing the cable trays and conduits nearby the isolation panels. The determination of the extent of fire damage consisted of evaluating fire consequences using engineering calculations from the FIVE-Rev1 library of fire models (EPRI 1002981, Fire Modeling Guide for Nuclear Power Plant Applications, August 2002.)

Based on the results of the analysis, the following conclusions are provided:

1. The 18 ft distance between the orange and purple isolation panels is relatively large. A fire in one of the panels is not expected to generate direct damage to the other panel. Furthermore, cables connected to the isolation panels are inside metal conduits. It is assumed that a fire will not propagate through metal conduits.

2. All the cables and conduits associated with one safety division are outside the zone of influence of a fire in a panel associated with the other safety division.

Consequently, it is expected that a fire in either of the isolation panels will not affect both safety divisions.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 20 of 34 NRC QUESTION 5 Smoke Effects on Fire Brigade Operations , Page 5, states:

The following fire fighting equipment is installed in the CSA or nearby areas. Page 29, states:

Dry hose stations and continuous flow hose reels are provided in sufficient numbers and locations within the CSA such that all trays can be reached.

With the fire protection equipment located within the fire area, how has the licensee evaluated fire brigade access to the equipment, considering that there may be smoke within the room? For some room fires the phenomena of smoke stratification may occur, where smoke collects at an elevation below the ceiling of the room. This stratified smoke may obscure fire fighter vision. How has smoke stratification been considered in the fire brigades effectiveness?

DOMINION RESPONSE The new fire fighting equipment installed within the CSA was added in conjunction with the recently installed incipient fire detection (IFD) system. Because the IFD system is designed to alarm during the incipient or pre-combustion stage of a fire, no visible smoke is anticipated when the fire brigade responds to an IFD alarm in the CSA. For that reason, the fire brigade is expected to be able to access the fire equipment locker in the CSA and use the thermal imaging camera stored inside the locker to find hot spots (e.g., in the CSA cable trays). The new booster hose reels located in the CSA augment the original fire fighting capability provided both inside and outside the CSA, by providing additional means to attack high cable trays which might be smoldering or beginning to ignite from a short circuit.

Regardless, should the fire brigade respond to an alarm in the CSA and find a high smoke content in the room, the fire brigade has the option to use fire fighting equipment staged outside the CSA versus inside the CSA. A second thermal imaging camera is also located outside the CSA and available to fire fighting personnel. As fire extinguishment and ventilation are performed, the booster hose reels and hose stations inside the CSA may be used for fire fighting or accessed for salvage and overhaul operations.

The fire brigade regularly trains to the fire fighting strategies specified for the CSA. The fire brigade training includes training to reduced visibility or zero visibility conditions both at the fire training facility at Millstone and in the in-plant CSA fire drills. In the

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 21 of 34 training facility, a high smoke environment is simulated using mineral oil (boiled), and brigade members are required to locate the source of the fire. Standard training for the fire brigade includes zero visibility at the fire scene and use of ventilation to improve visibility. During actual in-plant drills in the CSA, brigade members are shown pictures of the fire condition being encountered in the drill scenario, such as high smoke and fire conditions. The brigade members are trained to react with appropriate fire fighting techniques in response to the drill conditions (i.e. low approach for low smoke layer).

Hydrants, hose houses, and additional hose are readily accessible to the fire brigade from outside the CSA, to bring fire fighting water through either or both access points into the area.

Smoke stratification in the CSA is a concern based on the relatively high ceiling in that area. It was for that reason the installed IFD system was designed with two elevations of sensors and eight zone alarms at the control panel. This provides initial input to the brigade members as to the potential fire location.

NRC QUESTION 6 Gaseous Suppressant Propagation Attachment 1, Page 3, states:

Based on past experience (1999), a C02 discharge in the CSA has the potential to increase C02 levels in areas adjacent to the CSA.

Modifying the system to make it manually actuated will reduce the likelihood of spurious actuation. Also, by making the system a backup suppression system, the fire brigade will be less likely to actuate it. If the fire brigade actuates the system, how has the enclosure been tested to assure that the fire suppressant gas will not propagate to other fire areas and interfere with operator access (travel) to alternate shutdown panel?

DOMIN10N RESPONSE It is DNCs position that fire brigade use of the C02 suppression system will be based on an informed decision made by the brigade captain in consultation with the shift manager, following an assessment of the fire scene. DNC is confident that the C02 system will be actuated by the fire brigade in those fire scenarios where manual suppression will not be effective. Use of the C02 system is reflected in the CSA fire fighting strategies and in fire brigade drills and ongoing training.

Integrity testing of the CSA, as described in DNCs April 15, 2004 submittal, indicated that C02 leakage was observed into areas that are required to be traversed by operators should a CSA C02 discharge occur and a control room (CR) evacuation be necessary. It was for that reason, the plant modifications for barrier penetration enhancement, C02/02 monitors, the breathing air stations, and SCBA voice amplifiers were performed. Additionally, the route least likely to be impacted by CO2 migration

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 22 of 34 and fire fighting activities was written into the emergency operating procedure (EOP) related to a CSA fire emergency requiring CR evacuation. Specifically, this EOP directs the operators to don SCBA prior to evacuating the CR and traveling to the ASP if a C02 discharge occurred in the CSA. Also, if a C02 discharge has not occurred in the CSA, the EOP directed action is to obtain the SCBA prior to evacuating the control room and ensures operators are equipped with the SCBA should it be needed once they get to the ASP. The operators are directed by the EOP to exit through the CR east door when proceeding to the ASP. The described route in the EOP to the ASP is then via the service building west stairwell, into the east switchgear room and then into the west switchgear room where the panel is located. Based on these directions, operators avoid using the alternate route to the ASP (where C02 levels may be higher),

Additional details describing when operators are allowed to remove their SCBA are provided in Attachment 1, Section 4.1.4, of the April 15, 2004 submittal. Plant modifications that have been installed to manage the C02 migration are also described in that section of the submittal.

NRC QUESTION 7 Operator Preparation for Mock Evacuation , Page 17, States:

"A mock evacuation of the control room after a simulated discharge of CSA C 0 2 was accomplished with operators wearing SCBA."

How many operators participated in the mock evacuation? Under what conditions was the evacuation conducted, e.g., were the operators aware of the exercise and "prepared" for it or, was it done without operators' prior knowledge? Was some type of "emergency scenario" staged to cause the evacuation? How extensive was the evacuation, i.e., did the operators continue the exercise to the switchgear rooms and simulate alternate shutdown actions using SCBAs or other fixed breathing equipment?

Were they required to communicate with the SCBAs on and, if so, were they able to do so without undue difficulty? If communication was not required, why not? Did the mock evacuation consider delayed access or alternate means to access the switchgear rooms and alternate shutdown areas considering fire fighting activities, smoke spread, and other environmental and plant condition/activities that may challenge the operators to successfully perform alternate safe shutdown actions?

What is the basis for being able to generalize the performance results from the mock evacuation to remaining operators (crews) that were not tested?

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 23 of 34 DOMINION RESPONSE As a point of clarification, the mock evacuation described in the DNC submittal letter , page 17, and the initial validation (walk-through) of the EOP for a CSA fire subsequent to a control room evacuation, described on page 21 (see NRC question 8), are referring to the same evolution. Two off-shift operators and two fire brigade members knowingly prepared for and performed this walk-through while two observers recorded data. The walk-through was not conducted as part of any simulated emergency senario. All four participants of the walk-through were staged in the MPS3 control room with the initial conditions being that a fire had occurred in the CSA and C 02 had been discharged in the CSA. No physical staging of other personnel or equipment indicating an emergency plant condition (fire) was included as part of the walk-through. The four individuals donned SCBA and proceeded from the control room to the alternate shutdown panel (ASP) using the service building west stairwell and traversing through the east switchgear room as directed by the EOP. The operators then simulated EOP actions at the ASP including switch manipulations, providing directions to other operators, and observing instrumentation. The fixed breathing air station had not yet been installed and there was no simulation of hooking the SCBA into the fixed breathing air station at the ASP. (As discussed in DNCs April 15, 2004 submittal, training has occurred on the breathing air station for the operating crews since it has been installed.) It was determined that the fixed breathing air station would not need to be used by the operators until after actions were performed at the ASP to bring the plant to a stable condition (maintaining pressurizer level within the indicating band on the ASP). Procedures have been established for the breathing air station and use is directed in the EOP. The individuals participating in the walk-through were required to communicate with SCBAs using new voice amplifiers at the ASP and it was determined that the communication was not degraded.

It should be noted that typically four operators would be available during a fire emergency to perform an actual plant shutdown at the ASP and the timed scenario confirmed by the walk-through is considered conservative. Concurrent actions as well as pre-staging of steps would occur if four operators performed a plant shutdown versus two operators used during the walk-through.

The walk-through did not include a simulation of delayed access or alternate access to the switchgear rooms and ASP. An alternate path to the ASP exists and is well known by the operating crews, however, the EOP directed pathway does not result in the operators traversing stairweIIs/hallways adjacent to the CSA (i.e, those areas most likely to induce delays due to fire fighting activities). In addition, fire fighting strategies for the CSA direct activities to be performed away from the EOP directed pathway the operators use to access the ASP.

The basis for being able to generalize that all crews would be able to perform the shutdown scenario at the ASP using SCBA is that the operating crews receive the same

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 24 of 34 hands-on training at the simulator ASP. Additionally all licensed operators are trained in use of SCBA and need to pass yearly qualification testing. Training on the breathing air stations in the east and west switchgear rooms has been conducted for all operations shifts and is included in the periodic training requirements for licensed operators.

NRC QUESTION 8 Walkdown Information , Page 21, States:

"An initial validation (walk-through) of the revised EOP for a CSA fire and subsequent control room evacuation was performed in January 2003. , Page 21, specifically states in Bullets 4 and 5:

"Verifying the power operated relief valves (PORVs) were closed at the auxiliary shutdown panel in the west switchgear room was completed in 14 minutes and 16 seconds which was within the acceptable 15 minute time frame. Establishing reactor head vent letdown was accomplished in 24 minutes and 46 seconds which was within the 30-minute acceptance criteria."

(Part 1)

In addition to the five items (Attachment 1, page 21) described as accomplished successfully at the auxiliary shutdown panel, what other manual actions are required at the panels to achieve safe shutdown? Were these actions also performed as part of the walkdown to assure successful completion? If so, what were the times associated with the completion of these actions? Were required manual actions performed within the time limits to assure safe shutdown without collateral damage to equipment?

(Part 2)

Describe the access path to the alternate shutdown panel and any other locations that may require access for a fire in the CSA. How many shutdown locations must be accessed? How many operators are required?

(Part 3)

The difference between the time to complete the actions described in bullets 4 and 5, are 44 seconds and 5 minutes and 14 seconds, respectively. How do the alternate shutdown activities compare with the manual actions criteria in IP 71111.05, Enclosure 2? Given this was a onetime demonstration by one particular crew under non-hazardous conditions, what confidence is there that any crew could perform the

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 25 of 34 required activities during a fire within the time required? How is the reliability of crew performance assured for possible fire conditions that might be encountered? Was a time margin between the performance time and the minimal required time addressed that would provide confidence that any crew could reliably perform the actions under realistic fire conditions? If such a margin was considered, please describe the analysis.

(Partl)

The EOP directs numerous actions be performed at the ASP. The walk-through was intended to validate the timing of two actions considered to be critical to achieving stable shutdown conditions as well as the overall safe performance of operators using SCBA at the ASP (communications, movement, etc.). No additional actions are time-constrained and therefore were not part of the walk-through.

The required completion time for the critical manual actions specified in DNCs April 15, 2004 submittal, are based on the performance goals set by the Branch Technical Postion (BTP) 9.5-1 and generally do not correspond to an equipment damage state.

For example, the time specified for those manual actions associated with establishing RCS inventory control are based on maintaining pressurizer level within the indicating range, rather than the time to reach an RCS voiding or core uncovery condition. For the purpose of the walkthrough conducted, the more limiting times associated with the BTP performance goals were used to measure success.

(Part 2)

The access path to the ASP is via the service building west stairwell and was chosen and proceduralized to protect the control room operators from potential leakage of C02 from the cable spreading area east side door. The areas that are required to be accessed by the control room operators for shutdown outside the control room are the east and west switchgear rooms. The strategy supporting shutdown from outside the control room is designed to be accomplished with the minimum crew composition described in the plant Technical Specifications. The emergency operating procedure directs the shift manager, the unit supervisor, and the two control room operators to proceed to the ASP. From the ASP, other plant equipment operators are dispatched via radio or telephone to remote areas of the plant. It should be noted however that additional personnel would become available within approximately 90 minutes of event initiation given that the abandonment of the control room results in a declared emergency, ALERT level classification which requires activation of the emergency response organization.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 26 of 34 (Part 3) to NRC Inspection Procedure (IP) 71111.05T describes standards to be used during the NRCs Triennial Fire Protection Inspection for assessing manual actions to be taken outside the control room as adopted by a licensee since original licensing in support of achieving safe shutdown in the event of a fire. Specifically, Enclosure 2 applies to licensees committed to 10CFR50, Appendix R, Section lll.G.2. (As a point of clarification, the MPS3 CSA is similar to an Appendix R, lll.G.3 area.) The strategy described in EOP 3509.1, including manual actions to be taken to achieve safe shutdown from outside the control room were reviewed by the NRC under the original licensing basis of the facility and again in conjunction with the NRCs 50.54(f) facility assessment which completed in 1998. While this procedure has undergone some change, all changes are subject to review against the standards established for verification and validation of changes to plant specific EOPs. Those standards include an engineering review for changes that impact the BTP 9.5-1 fire protection program. A documented fire safe shutdown review of EOP changes determined to impact the fire protection program include consideration of basic elements like those outlined in Enclosure 2 to IP 71111.05T. Additionally, EOP changes are subject to review and approval by the Station Operations Review Committee.

Although the walk-through performed in support of this change was a one-time demonstration, Millstone is confident that other operators could perform the required activities during an actual fire in the same time period. Operator training on the fire emergency EOPs (shutdown actions at the ASP) is conducted for licensed operators once per training cycle (two-year cycle). The alternate shutdown panel in the west switchgear room is modeled in the MPS3 simulator and provides an effective training aid for switch manipulations required by the EOPs. Licensed operators are trained in SCBA use and have to pass an annual qualification test. Training on the breathing air stations in the east and west switchgear rooms has been conducted for all operations shifts and is included in the periodic training requirements for licensed operators.

Consideration of a time margin was not included in the development of the original performance times upon which operator success is currently determined, nor should the data presented in the application for this amendment be construed to represent the margins available. The current performance times are conservatively based on meeting the performance standards identified in BTP 9.5-1 (i.e., maintaining pressurizer level in the indicating range, reactor heat removal function capable of removing decay heat),

while the performance margins discussed in the NRCs draft position on this subject are based on the point at which equipment damage occurs.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 27 of 34 NRC QUESTION 9 Full Crew Complement , Page 21, States:

"An initial validation (walk-through) of the revised EOP for a CSA fire and subsequent control room evacuation was performed in January 2003. Two operators and two fire brigade members simulated a control room evacuation after a simulated CSA C02 discharge."

The walkdown was conducted with two operators; is this essentially half the maximum number of operators that are needed to perform shutdown actions (see Attachment 1, Page 16)? With a "full crew complement," how much time will the crew take to accomplish the required manual actions to safely shutdown the plant from the auxiliary shutdown panel and how was the time determined? Provide a list of operator actions that impact the application, their error probabilities, and how they were estimated.

DOMINION RESPONSE Two operators were used to verify that adequate communication would occur while wearing SCBA at the ASP. The two fire brigade members were used to simulate the presence of four people at the ASP. This was done to determine the ability of four people to negotiate the area around the ASP. The fire brigade members also ensured the physical movement of the operators in SCBA was safe and could be accomplished without jeopardizing plant equipment since the walk-through was performed at 100%

power operation. The evolution was accomplished satisfactorily with four personnel at the ASP wearing SCBA. Generally, the time to shut down the plant from the ASP with two operators is considered to be very close to the time needed for four operators to perform the same function. This conclusion is based on the reasoning that four operators would result in better pre-staging of steps and performance of concurrent steps, and therefore result in less time to establish control and shutdown from the ASP.

The use of two operators during the scenario was considered conservative with regards to the timing of the scenario. The walk-through was judged by the observers of the walk-through to have been conducted orderly, with tasks performed in a focused and methodical manner. This type of performance is consistent with operator training philosophy.

It is important to note that the actions of controlling the plant from the ASP should a CR evacuation be necessary as the result of a CSA fire are not dependent on the initiation feature of the C02 system in the CSA (e.9. automatic versus manual).

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 28 of 34 The human error probability (HEP) calculation below has been provided for the one manual action considered to be critical for maintaining core cooling, which is to align letdown and charging.

HEP of the Manual Action at the ASP durina the Fire Caused MCR Evacuation Most of the critical actions to achieve safe shutdown during the Main Control Room (MCR) evacuation caused by cable spreading area fires are performed before the operators leave the Main Control Room. For example, the operators need to close the PORV block valve before they leave the MCR. Hence, the action to close the pressurizer PORVs from the auxiliary shutdown panel (ASP) would not be a critical action.

According to the emergency operating procedure (EOP) for a cable spreading room fire, (EOP3509.1) the step to align the letdown and charging suction path is identified as the only critical action to achieve safe shutdown. The HEP of this specific action is evaluated as follows:

Event

Description:

Operator failed to align letdown and charging suction path.

System: Charging System Component: 3CHS*AV8149AJB, C, 3RCS*LCV459,46OJ3CHS*LCV112D, 112E, 112B.

Component Location: Auxiliary Shutdown Panel Procedure: EOP 3509.1 Control Room, Cable Spreading Area or Instrument Rack Room Fire Notes:

EOP 3509.1 Step to Align Letdown and Charging Suction Path states:

a) CLOSE letdown orifice isolation valves (ASP A) 3CHS*AV8149A 3CHS*AV8149B 3CHS*AV8149C b) CLOSE letdown isolation valves (ASP A) 3RCS*LCV459 3RCS*LCV460 c) OPEN RWST TO CHG isolation valve 3CHS*LCV112D (ASP A) d) Check a Train B charging pump - RUNNING e) OPEN RWST TO CHG isolation valve 3CHS*LCV112E (ASP B) f) CLOSE VCT TO CHG isolation valve 3CHS*LCV112B (ASP A)

Based on the procedure walk-through, from the time of the simulated reactor trip, the total elapsed time to don and check out SCBA staged in the control room was approximately one minute.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 29 of 34 For fires resulting in both a total loss of charging and the need to evacuate the control room, the action to isolate letdown is modeled to occur within a 15 minute time limit (DNC Engineering Record of Correspondence 25212-ER-04-0011, Appendix R Inputs to Fire Safe Shutdown Thermo-hydraulic Calculation, February 2004). However, the flow rate of the letdown path is around 120 gpm.

Failure to isolate the letdown path in an hour should not have significant impact to the reactor core since 120 gpm leakage from RCS does not impact natural circulation within the first hour.

According to the computer simulation of EOP 3509.1 performed May 24, 2004, the validation time in this case is 12 minutes to isolate letdown.

At least two licensed operators perform the EOP 3509.1 at the ASP.

During the MCR evacuation, the stress level is justified as extremely high.

Operators need to close six valves and to open two valves from the ASP. If any one of the eight valves does not operate correctly, the action to align the letdown will be justified as failure.

The HEP is estimated based on the THERP methodology in accordance with NUREGICR-1278, Handbook of HRA with Emphasis on Nuclear Power Plant Applications, August 1983.

PCCalculation bv THERP Model This is pure rule-based behavior. Operators are trained to follow EOP3509.1 step by step. The cognitive error to prevent the operator from performing the step to align the letdown and charging suction path in EOP 3509.1 is the omission error. Either operator omitting one item of this step, or omitting the entire step, may cause the action failure.

According to item 43 in Table 20-7 of NUREGICR-1278, the failure rate of the omission of the step to align the letdown and charging suction path is:

HEPmedi, (omit entire step) = 0.01 with error factor 3 According to the item 34 of Table 20-7 of NUREGICR-1278, the failure rate of the omission of one item of the step to align the letdown and charging suction path is:

When procedures without checkoff provisions are used, or when checkoff provisions are incorrectly used for a long list (>lo items) 4 When procedures without checkoff provisions are used, or when checkoff provisions are incorrectly used for a long list ( > l o items)

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 30 of 34 HEPmedia, (omit one item of step) = 0.001 with error factor 3 PC(rnedian) - HEPmedian (Omit entire Step) + HEPmedian (Omit one item Of step)

- 0.01 + 0.001

- 0.01 1 with error factor 3 PF Calculation bv THERP Model To close or open these valves is a simple pushbutton action on the ASP. The likely execution error is to select a wrong button. According to item l 5 of Table 20-13 of NUREGCR-1278, the selection error of each valve will be 1.OE-3 with error factor 3.

For all 8 valves the selection error will be:

PE(rnedian) - 0.008 with error factor 3 If the operator selected a wrong control button, this button may be one of the eight buttons. A recovery factor 0.5 is applied to the selection error.

PE(median) - 0.004 with error factor 3 HEP (median) - PC(rnedian) + PE(median) = 0.015 Error Factor - 3 Modifv the HEP for the Effects of Stress and ExDerience Levels During the MCR evacuation, the stress level is justified as extremely high. Furthermore, the operators need to don SCBA to perform their job on ASP. These activities are not trained frequently. According to item 6 (b)6of Table 20-16 of NUREG/CR-1278, a factor of 10 should be applied to the HEP. Hence, HEP (median) - 0.015

• 10 = 0.15 Error Factor - 3 Modifv the HEP for the Effects of Recoverv Factor Based on the following reasons, a recovery factor of 0.1 is suitable to apply to the iEP evaluation :

a. The time window of this action is at least one hour (in average, operator may finist this task in 12 minutes). Operators have enough time to find their mistakes.

5 Clearly and unambiguously labeled, set apart from valves that are similar in all of the following:

size and shape, state, and presence tags.

6 Step-by-step actions with extremely high stress and novice.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 31 of 34

b. At least two licensed operators perform EOP3509.1 together. They may find the mistake done by each other.

HEP (median) - 0.15

• 0.1 = 0.015 Err0r Factor -

- 3 HEP (mean) - HEP (median)

• EXP{[ln(EF)/1.645] /2}

- 0.015

• 1.25

- 1.88E-2 In conclusion, the error probability for the actions at the ASP is 1.9E-2. DNC considers this contribution to CDFto be small.

NRC QUESTION 10 Reactor Coolant Pump (RCP)

Fire Protection Evaluation Report, Section 7.4, States; "Failure of the three charging pumps coincident with the failure of the thr e mponent cooling water pumps due to fire Is considered an incredible event."

Based on the statement in the Fire Protection Evaluation Report, it is unclear how the licensee protects the RCP seals if there were a CSA fire. Provide technical analysis of how operators will mitigate coincident loss of component cooling water pumps and charging pumps due to CSA fire, When concluding the coincidence of the two triple failures was incredible, was the possibility of two concurrent triple common-cause failures (i.e., one among the three charging pumps and one among the three CCW pumps) addressed, or were failures assumed to be independent? Please provide any evaluation of the "incredibility" that you performed.

DOMINION RESPONSE Section 7.4 of the Fire Protection Evaluation Report is referring to a fire in the Auxiliary Building where both sets of pumps are in the same fire area (AB-1). There is a water curtain installed between the component cooling pumps (CCP) and charging pumps, and between support cables within the fire area. This was the subject of deviation letter B11852, dated 11/4/85, and has been approved in SSER Supplement 4 page 9-5. The fires in the CR, instrument rack room, or CSA which affects both sets of pumps is one of the reasons to abandon the CR and take local control at the ASP. For this fire, the A Train charging pump will be electrically isolated from the fire area and operated locally at the switchgear once electrical power is restored and valve alignments have been completed. The current shutdown strategy does not restore cooling to the RCP seals if cooling has been interrupted either by loss of power or valve misalignment due to

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 32 of 34 spurious operation. For this event, makeup will be provided through the normal charging path. CCP will not be required in this scenario until later in the event when needed for cool down in support of the RHR system.

NRC QUESTION 11 CDF and LERF Summary Provide an assessment of the change to CDF and LERF, including a description of the significant contributors to the change.

DOMINION RESPONSE An assessment of the change to CDF and LERF is provided below.

1. At each of the seven pinch-points referred to as Case 1 through 7, identified in DNCs April 15, 2004 submittal, all cables that could be damaged by fire were identified. A walk down was conducted to identify the list of cables at each pinch-point of the different fire scenarios.

2. All the equipment powered by the damaged cables were identified from the Fire PRA Cable and Equipment Database and matched to the appropriate Basic Events. Some of the raceways identified during the walk down did not have associated components in the database and were therefore excluded.

3. With the associated basic events failed for each case, the current released Equipment-Out-Of-Service (EOOS) Risk Monitor EPRl Code, Version 3.3a, MPS3 Online, 2003 was used to calculate CDF and LERF for each of the seven cases. The average maintenance model M3EOOS.CAF was used in EOOS and modified the Activity Table in MPS3online.mdb file to insert case mappings.

Also, the Basic Event database file M3Rev2.BE was modified to set all the initiators to 0.0 except for %GPT, which was set to 1.O.

4. To calculate the change in CDF and LERF, the CCDP and CLERP are multiplied by the difference in Fire Frequencies for the Automatic and Manual CO2 system configurations. These fire frequencies are calculated in DNCs April 15, 2004 submittal (summarized in Table 15, Attachment 3 of that submittal). Table 12 below contains the fire frequencies, associated CCDP, CLERP and delta CDF and LERF for each case.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 33 of 34 Assumptions:

1. A fire in the CSA is assumed to result in General Plant Transient (GPT).

Therefore all the Initiating Events (starting with %) in the Basic Event File were set to zero except %GPT which was set to 1.O.

2. The CDF calculation does not account for the additional risk of spurious actuation of the C 0 2 system while it is in automatic mode. Spurious actuation of C 0 2 represents both a nuclear safety and a personnel hazard that can be avoided by allowing manual initiation only. If a spurious actuation occurs it will increase the risk of a plant transient since the control room may need to be evacuated and shutdown actions have to be performed at the auxiliary shutdown panel. Although this risk is not anticipated to be significant, there is a clear risk benefit of avoiding spurious actuations of the C 0 2 system by operating it in a manual mode only.

3. The CDF and LERF calculations do not consider the benefit of the Incipient Fire Detection (IFD) system for detecting fires in the pre-combustion stage. This fact should be reflected in the PRA model as a reduction in the fire frequency.

However, the calculation conservatively does not credit the IFD system with eliminating fires. The risk analysis assumes the IFD system only detects the fire after it progresses to the stage in which a fire has developed. In other words, the IFD system is conservatively treated only as a second detection system. This is particularly conservative for the CSA, which has no significant combustible material other than IEEE-383 qualified cable insulation.

Furthermore, there is only a small likelihood of transient combustibles or ignition sources being in the room (no welding work, not a traveled area etc).

DNC believes the IFD is designed to provide the additional benefit of notifying operators, during the pre-combustion stage, when procedures would direct staff to enter the area and eliminate the hot spot. Although not quantified, this reduction would reduce the CDF/LERF values due to the lower fire frequencies for the manual C 0 2 scenario and the automatic and manual risks would be roughly equal.

Serial No.05-066 Docket No. 50-423 Response to RAI Attachment Page 34 of 34 Automatic Manual C02 C02 System System (per year) (per year Scenario Ascenario Ascenario CCDP ACDF CLERP ALERF Notes:

1. Previously, when the C 0 2 system was operated in the automatic mode, it would be disabled during hotwork activities. Thus, scenarios involving hotwork are omitted from this analysis (i.e. scenarios b & c in Table 15 of DNCs April 15, 2004 letter).

2. For Case l a , the frequency is equal to zero and therefore ACDF and ALERF equal zero.

3. A truncation of 3 to 4 orders of magnitude below the conditional core damage frequency (CCDF) is used for the analysis. As a result a truncation of 1E-05 was used for Case 2a, 3a and 7a and truncation of 1E-09 was used for Case 4a and 6a. Due to a large number of equipment failing in Case 5a, a truncation of 1E-02 was used to reduce the number of cutsets and computational time.

In conclusion, as shown in Table 12, the total ACDF between Automatic and Manual C02 system actuation of the seven cases is 2.3E-O7/yr. Cases 2 and 3 are the dominant contributors primarily because the CCDP for these two fire locations is 1.0 (or nearly 1.O). All of the fires initiating frequencies are very small. Case 7 has the highest fire initiating frequency, however, the CCDP offsets this, which results in a relatively small ACDF.

For Large Early Release Frequency, the total ALERF is 7.5E-O9/yr. In nearly all cases, core damage resulted from loss of A and B electrical buses.

The calculated change in CDF is less than 1E-06 /yr which is very small (Region Ill) according to RG 1.174 acceptance guidelines. In the same manner, the calculated change in LERF is less than 1E-O7/yr which is very small (Region 111) according to RG 1.174 acceptance guidelines. The calculations are conservative since as noted in the assumptions, no credit is taken for the IFD system detecting fires in the pre-combustion stage. Therefore, the change from an automatically actuated C 0 2 system to a manually actuated system in the CSA does not result in a significant increase in CDF and LERF.