ML050610249

Supplemental Information Regarding the Request for Exemption from 10 CFR 50, Appendix R, Section III.G.3 for Fire Area Hh

ML050610249
Person / Time
Site:	Davis Besse
Issue date:	02/25/2005
From:	Bezilla M FirstEnergy Nuclear Operating Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
References
3106, TAC MC1833
Download: ML050610249 (24)

Text

FENOC 5501 North State Route 2 FirstEnergy Nuclear Operating Company Oak Harbor, Ohio 43449 Mark B. Bezilla 419-321-7676 Vice President - Nuclear Fax: 419-321-7582 Docket Number 50-346 10 CFR 50.12 License Number NPF-3 Serial Number 3106 February 25, 2005 United States Nuclear Regulatory Commission Document Control Desk Washington, D. C. 20555-0001

Subject:

Supplemental Information Regarding the Request for Exemption from 10 CFR 50, Appendix R, Section III.G.3 for Fire Area HH (TAC No. MC1833)

Ladies and Gentlemen:

On January 20, 2004, the FirstEnergy Nuclear Operating Company (FENOC) submitted a request for an exemption from Title 10 of the Code of Federal Regulations (CFR), Section 50, Appendix R, "Fire Protection Program for Nuclear Power Facilities Operating Prior to January 1, 1979,"Section III.G.3, for the Davis-Besse Nuclear Powver Station, Unit I (DBNPS)

Fire Area HH (DBNPS Serial Number 3003). By letter dated September 3, 2004 (DBNPS Serial Number 3076), FENOC submitted supplemental information regarding the exemption request.

On September 28, 2004, FENOC received an informal request for additional information regarding the exemption request. This request was clarified in a conference call with the NRC staff on November 12, 2004. Enclosure I provides the response to the request for additional information. Enclosure 2 provides the additional evaluation discussed in the November 12th conference call. Enclosure 3 identifies any regulatory commitments contained in this letter.

Should you have any questions or require additional information, please contact Mr. Henry L. Hegrat, Supervisor - Licensing, at (330) 315-6944.

Very truly yours, (5--Aa,,

2?#>4Asd

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 2 Enclosures cc: Regional Administrator, NRC Region III J. B. Hopkins, NRC/NRR Senior Project Manager C. S. Thomas, NRC Region III, DB-1 Senior Resident Inspector Utility Radiological Safety Board

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Enclosure I Page I RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION REGARDING FIRE AREA HH EXEMPTION REQUEST FOR DAVIS-BESSE NUCLEAR POWER STATION UNIT NUMBER I Ouestion 1:

a. You stated during the teleconference on September 27, 2004 that Rooms 603 and 603A are separated by a "substantial" metal-clad door with a metal louver at the bottom. You also stated that the wall is made out of concrete. Please submit descriptions of the wall and door.

b. Please confirm that no combustibles are stored in Room 603A and that the door is normally locked and has a sign on it stating, "No Storage of Combustibles."

DBNPS Response to Question 1:

a. The wall is made of 12-inch thick concrete masonry units. The door that separates the two rooms, Door 603, is a UL Class B fire door with a louvered opening. The louvered opening contains a damper that is held open by a fusible link.

b. A field walkdown by the Fire Protection Engineer and the Fire Marshall confirm the following: Door 603 is locked; signs stating "SEE SHIFT SUPERVISOR FOR ACCESS,"

"STORAGE OF ANY KIND FORBIDDEN," and "DOOR MUST REMAIN LOCKED," are installed; and no combustibles are stored in Room 603A.

Question 2:

a. Describe the location of the CREVS cables with respect to combustibles and possible sources of ignition in Room 603.

b. Verify that both of the redundant CREVS cables will not be affected by a single fire within the area.

c. Verify that the combustibles in the area are inadequate to cause area wide damage, such as, through the development of a hot gas layer.

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 2

d. If questions 2.b and 2.c cannot be verified, provide the technical bases to show that there is some other means by which a fire in the area will be limited such that damage to both redundant cables will be avoided, reduced, or damage will take an extended period to occur, allowing fire brigade to intervene.

DBNPS Response to Question 2:

a. The majority of conduits and components for both trains of the Control Room Emergency Ventilation System (CREVS) are located on the north portion of Room 603. The air-cooled condensers are located on the roof above this room. Surrounding the CREVS equipment in Room 603 are conduits and components for the normal ventilation system and to a lesser extent other plant components, i.e., CREVS conduits and components are in close proximity to combustibles and possible sources of ignition.

b. Due to the proximity of Train 1 CREVS cables and components to Train 2 CREVS cables and components, it can not be stated that both will not be affected by a single fire. Both trains are located in the same part of the room. The conduits and components are very close together. Some conduits and components are only inches apart. Some components of the CREVS system are only inches from other plant components, including portions of the non-essential ventilation system.

c. While the total amount of combustibles in Room 603 are low (less than 16,000 BTU/ft 2 ), the room is not void of combustibles. Given the proper ignition source(s) and sufficient time, a hot gas layer can form. See further discussion below.

d. The tools (Excel spreadsheets) distributed with NUREG-1805, "Fire Dynamics Tools (FDTs)

Quantitative Fire Hazard Analysis Methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program," can be used to provide insight to this question. For example, the temperature of the compartment with the doors closed and the room fan in operation would not exceed 250'F after 20 minutes, assuming a 500 kW fire (200 lb of rags, gloves, solvent per NUREG-1805 Table 2-7), a room size of 50.2 ft x 61 fi x 14.5 ft, and a flow rate of 1000 cfm. The 20 minute duration provides adequate time for the brigade to respond, and the limited room temperature is low enough to preclude damage to cables in conduits not in the plume or ceiling jet (reference NUREG-1805 section 2.9). FENOC believes these assumptions to be reasonable, however, the inputs to these or other fire models can yield many possible results depending on what size fire is used, what type of ventilation is used (forced vs. natural), etc.

A probabilistic evaluation of the impact of adding an automatic sprinkler system to Fire Area HH has been performed. See Enclosure 2 to this submittal.

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 1 Determination of Change in Plant Risk Due to a Sprinkler System in Fire Area III

1. Purpose The purpose of this evaluation is to determine the change in plant risk that would result if Room 603, the Control Room Ventilation Room, was equipped with an automatic sprinkler system. This room is the largest of three rooms in Fire Area HH. Room 603 contains both the normal and the emergency equipment used for control room ventilation.

As part of the Davis-Besse (DB) Probabilistic Safety Assessment (PSA), the loss of ventilation in various rooms of the plant was considered. For the main control room a qualitative evaluation arrived at the following conclusion':

The control room is served by a normnal HVAC system, with backup by the control roomn eniergency ventilation system. In the event that there is insifficient cooling available, procedures explicitly describe steps to take to reduce the heat load in tile room and, when necessary, to achieve an orderly shutdowvn. Because of the redundantsystenis and tile presence of operators to detect and correctan abnormalsituation aliost inmnediately, loss of HVA C to this area does not merit treatment as a separateinitiatingevent.

In part as a result of this conclusion, the Individual Plant Examination of External Events (IPEEE) for Davis-Besse determined that Room 603 posed a negligible risk with respect to the potential for internal fires2 .

2. Summary A small reduction in plant risk would be achieved by installing a sprinkler system in Room 603.

The benefit associated with the reduction in core-damage frequency would be too small to justify installation of the sprinkler system.

3. References A. Letters

1. Individual Plant Examination of External Events for Severe Accident Vulnerabilities for the Davis-Besse Nuclear Power Station, Unit 1 (Response to Generic Letter 88-20, Supplement 4), Serial 2412, December 16, 1996 l DB PSA Initiating Event Notebook Volume 1 (Summary) Page 32 2 DB IPEEE, p. 4-69.

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 2

2. Response to Request for Additional Information Pertaining to the IPEEE, Serial 2549, August 21, 1998

3. Response to the Supplemental RAI on the IPEEE, Serial 2621, October 29, 1999

4. Request for Exemption from 10 CFR 50, Appendix R, Section III.G.3 for Fire Area HH, Serial 3003, January 20, 2004

5. Supplemental Information Regarding the Request for Exemption from 10 CFR 50, Appendix R, Section HI.G.3 for Fire Area HH (TAC No. MCI 833), Serial 3076, dated September 3, 2004 B. Drawings

1. E-416, Raceway Aux. Bldg. El. 643' Areas 5 & 7

2. M-450, HVAC: Control Room - Equipment Room Plan at El. 638'-O"

3. OS-32B, Operational Schematic: Control Room Emergency Ventilation System C. Other

1. Fire Hazard Analysis Report

2. C-FP-013.10-006, Combustible Loading Calculation

3. EPRI Methods of Quantitative Fire Hazard Analysis, TR-100443, 1992

4. EPRI Fire PRA Implementation Guide, TR-105928, 1995

5. NRC Fire Protection Significance Determination Process, Appendix F of 0609, issue date 05/28/04

6. DB PSA Initiating Event Notebook Volume 1 (Summary)

7. NRC NUREG-1805, Fire Dynamics Tools

4. Assumptions None

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 3

5. Methodology The methodology used in the evaluation will evaluate the reduction in risk that might be achieved by installation of an automatic suppression system in Room 603. This will be done by estimating the core-damage frequency (CDF) associated with a fire in the room (including explicit consideration of the potential need to evacuate the main control room) based on the current configuration. This estimate wvill then be compared to the reduced CDF that would result if an automatic sprinkler system were available in the room.

6. Evaluation Existing Condition Fire Area HH consists of:

Air Conditioning (A/C) Equipment Room (Room 603) 3,013 ft 2 Records and Storage Area (Room 603A) 77 ft2 Vestibule (Room 603B) 60 ft2 These rooms are all on the 638' elevation of the Auxiliary Building. Room 603 sits immediately above the Control Room complex. Room 603 houses both trains of normal Control Room ventilation as well as both trains of the Control Room Emergency Ventilation System (CREVS).

The fire area consists of 3,150 square feet of floor area. The evaluation will deal with Room 603 as this room contains the majority of the combustible loading and the safe shutdown circuits within the area. The ceiling of Room 603 is 14.5 ft high. The walls are primarily poured concrete, with some masonry walls.

The in-situ combustible loading in Room 603 consists of cable insulation; neoprene heating, ventilation and air conditioning (HVAC) duct insulation; and small quantities of grease, lube oil, and miscellaneous combustibles. The total combustible loading in the room is less than 20,000 BTU/ft2 , which equates to an equivalent fire severity of 15 minutes in accordance with the National Fire Protection Association "Fire Protection Handbook," Sixteenth Edition. There are charcoal filters in the room, but due to the robust nature of the steel filter housings they are not included in the above values.

Rooms 603A and 603B are assigned a loading of 400 BTU/f12 to account for "incidentals" such as signage and other non-quantifiable loads.

Existing fire protection capability in the area consists of a fire detection system that protects the A/C Equipment Room (Room 603) and manual (not fixed) fire suppression capability consisting of portable fire extinguishers and standpipe hose stations for the protection of the entire area.

Fire Area HH has primarily 3-hour fire barriers on the walls and floors as described in the FHAR.

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 4 Fire damage to the circuits for the Control Room normal and emergency ventilation in Fire Area HH could disable the Control Room HVAC. Loss of the Control Room HVAC is not expected to have an immediate effect on the ability to shutdown the plant from the Control Room. Although procedural guidance to mitigate a temporary loss of HVAC is provided, the Operators may need or choose to abandon the Control Room due to high temperatures.

Evacuating the Control Room and shutting down the plant from the Auxiliary Shutdown Panel can provide alternate shutdown capability. Plant procedures include instructions for these manual operator actions if Control Room cooling is disabled.

For the loss of all Control Room ventilation analysis, it was assumed that eventually all equipment with controls in the Control Room would be affected by a fire in Room 603 due to the temperature in the Control Room envelope eventually rising above the equipment's temperature rating and failing. As discussed in Serial 3076, it would take 30-40 minutes before the Control Room would reach the temperature limits established in plant procedures.

6.1 Pinch Points Based on a review of the ventilation drawings, the electrical raceway drawing and a listing of safe shutdown circuits associated with the CREVS, a number of "pinch points" were identified.

These are places where there were multiple circuits and/or components that are in close proximity and that are more likely to be damaged by a single fire. Based on the drawing review there are only a few places where both trains of CREVS can be lost (since the space of all its equipment takes up only about 25% of the room floor area). Since it is not known exactly which circuits may cause the loss of normal ventilation it is assumed that there may be many places in the room where this could occur.

Areas that could cause total loss of CREVS include the following:

• Area near cabinets C6714 and C6715. These are control cabinets for the two respective trains of CREVS. These cabinets have ventilated front doors. Due to their solid sides and several inches of separation, a fire would not propagate from one cabinet to the other, but there would be the possibility for a single transient fire in front of the cabinets to affect the components in both cabinets.

• Area around cabinet C6705. This non-ventilated control cabinet has the controls for both trains of normal ventilation. A fire originating in C6705 would not affect CREVS because this is non-ventilated. There are also two circuits for train 1 of CREVS located about five to seven feet away in conduit 67002D. These circuits are the following:

MS3311 control circuit (lCCEACIC) and control power to C6714 (lCY104A). This conduit runs along the south wall of the room until it gets half way across then goes directly north. The possibility exists for a single transient fire in front of the cabinet that

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 5 could affect the normal ventilation and Train 1 of CREVS, but Train 2 of CREVS would still be available.

Area of the CREVS controls and compressors (west end of the skid). The two trains of components have little or no separation.

6.2 Fire Scenarios This section discusses the potential fire scenarios in Room 603. Note that manual fire fighting is the primary suppression method in all cases. Fire brigade response (commencement of manual fire fighting) is expected within 15 minutes of an alarm. The fire brigade would use portable extinguishers and/or hose streams to control and extinguish the fire.

Extinguishers and hose stations are located both in the room as well as just outside the room.

6.2.1 In-situ Cable Fire - Self-Irnition and Transient Ignition Self-ignited fires are not expected as all cables are IEEE-383 or equivalent design and are either enclosed in conduit or in cabinets. There are no cable trays in the room. Additionally page F3-2 of the NRC SDP indicates that IEEE-383 cables are not ignition sources so long as protective devices are properly sized. See section 5 of the FHAR for further discussion of the overcurrent protection used at DB.

6.2.2 Electrical Cabinets and other Equipment Fires There are a number of electrical cabinets (ventilated and non-ventilated), two non-essential MCCs, ventilation systems, pumps, etc. in the room.

The majority of these "cabinets" are breaker cubicles associated with MCCs E23B and F23B.

They are non-ventilated cubicles and contain qualified cables. They are not related to the operation of CREVS.

MCC E23B is not located near areas of concern and therefore will not be studied further. MCC F23B is located less than three feet from the combustible insulation on the CREVS filter F22-2 ductwork. Per the EPRI Fire PRA implementation guide, page E-8, "Fire cabinets that are not vented do not propagate." The NRC SDP, page F3-2, also indicates that fires within non-vented cabinets will stay within the cabinets. This section also adds that arcing faults in 440VAC cabinets and higher can propagate outside solid metal cabinets. (Subsequent discussion in the SDP, page H-27, indicates that it is the 4160VAC and higher equipment that is subject to the highly energetic faults.) Since all the CREVS circuits are IEEE-383 type cable and are located in conduit and none of these conduits are in the immediate vicinity of the this MCC, nor the section of ductwork facing the MCC, these cabinets are not studied further.

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 6 The remaining 20% of the electrical cabinets are distributed throughout the room. Rather than investigate each of these individually, as a bounding case it is assumed that a fire in any of these might have the potential to threaten operation of the CREVS.

A rapidly developing fire is one in which the combustion stage occurs almost simultaneously with ignition. This type of fire would require that a combustible liquid be available to ignite and rapidly burn.

There are four sources for this: the two normal chillers (S12-1 and S12-2 with 4.25 gallons of oil each) and the two emergency compressors (S33-1 and S33-2 with 1.4 gallons of oil each).

Control Room Chillers (S12-1 and S12-2)

S 12-1 is located immediately above a floor drain and thus any spill should be drained away.

Spills are evaluated should the drain fail. Using the industry guidance3 , a spill will spread until the limiting depth is reached for the assumed liquid or a physical restriction is reached. Thus an unrestricted liquid spill can achieve a large diameter of little depth. Once an ignition source is introduced, the resulting fire results in a very high heat release rate (HRR) but a short duration (lasting seconds). Alternatively the oil is assumed not to spread and results in a smaller heat release rate but longer duration. In either case the fire at S12-1 should not affect both trains of CREVS.

The NRC SDP, page F5-7, defines two different spills. One is the "high confidence spill" (i.e.

100% of the contained oil) that would occur only 10% of the time and the "expected spill" (i.e.

10% of the contained oil) that would occur 90% of the time.

Note: Page F3-2 of the SDP indicates that for combustible liquid fires less than one gallon "damage is postulated only if the target is in the plume or suffers direct flame impingement."

The impact of a fire resulting from an oil spill at control room normal chiller S 12-2 was not calculated, as the chiller is farther from CREVS and located immediately above a floor drain.

Should the floor drain fail, the results would be the bounded by the evaluation for chiller S 12-1.

CREVS Condensers (S33-1 and S33-2)

The "high confidence spill" (i.e. 100% of 1.4 gallons) and the "expected spill" (i.e. 10% of 1.4 gallons, or 0.14 gallon) will be retained within the CREVS skid due to the arrangement of the skid's steel framing. Due to the intimate design of the skid (the two compressors are approximately two feet apart and the filter housings will act as if the source is in a room corner),

both trains of CREVS are assumed lost with either fire.

3 EPRI TR-10043 Section 6.1

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 7 See Attachment 1 for more discussion on the fire frequency due to equipment in the room.

6.2.3 Transient Combustible Fire This scenario considers a transient combustible fire involving a Class A material or flammable liquid. While there are administrative controls that restrict the amounts and types of combustibles brought into Room 603, it is assumed that there may be enough material to result in a transient fire. The scenario evaluated is one in which material is brought into Room 603 and ignites and bums. Due to the neoprene anti-sweat insulation that is used on many of the sections of ductwork, including ones in the overhead, a transient fire could spread.

See Attachment I for more discussion on the fire frequency due to transient combustibles in the room.

6.2.4 Hot Work Fire Room 603 is accessible during all plant modes and thus there are no plant conditions that would preclude hot work in the room. The need for hot work is limited but there is nothing that would prohibit hot work in the area. Thus this would fall into the MEDIUM category, "on the order of once per operating year," of the NRCs SDP page F4-6.

See Attachment I for more discussion on the fire frequency due to hot work in the room.

6.2.6 Unmitigated Fire An unmitigated fire, although considered highly unlikely, is postulated to bum for an extended time period without any active fire suppression actions. The consequence of an unmitigated fire is loss of all Control Room normal and emergency ventilation system equipment and cables.

This bounding scenario has already been considered in the safe shutdown analysis contained within the FHAR. Because Room 603 contains redundant trains of safe shutdown equipment that are not separated, alternate safe shutdown has been provided in separate plant fire areas.

Therefore, an unmitigated fire in Room 603 will not prevent the safe shutdown of the plant.

The following provides this assurance:

• Transient combustibles and ignition sources are controlled by administrative procedures in Room 603 as previously discussed.

• The ionization smoke detection system would be available to provide alarms to alert the control room of a fire.

• A fire in Room 603 will propagate very slowly based on the combustible loading of the room and the spatial separation of the combustibles through the room.

• The site fire brigade response to Room 603 upon receiving an alarm from the smoke detection system ensures fire damage would be limited. For this type of fire, the

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 8 suppression method is manual fire fighting, using fire suppression activities commensurate with the type and size of fire encountered.

Finally, the safe shutdown analysis demonstrates that loss of Room 603 will not prevent the safe shutdown of the plant. Alternate shutdown capabilities are provided for a fire in Room 603 using the Auxiliary Shutdown Panel in Room 324 and other manual stations in the plant. These actions would not require entry into Fire Area HH, and actions would not be impaired by the smoke or fire fighting activities in that area.

See Attachment 1 for more discussion on the core damage frequency due to a fire in the room that goes unsuppressed.

7.

Conclusion:

The lack of an automatic suppression system in Fire Area HH does not result in a significant increase in "damaging fires."

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 9 Attachment I Probabilistic Evaluation of Impact of Adding Automatic Suppression to Fire Area IIII An evaluation has been performed of the reduction in core-damage frequency that might be realized if an automatic fire suppression system were available in Fire Area HH. The general scenario that could lead to a total loss of HVAC for the control room due to a fire in Fire Area HH, and a consequential core-damage event, is depicted in the event tree provided as Figure 1. Each of the top events in this event tree is addressed in the sections that follow.

Fire suppressed Core cooling Fire occurs in area Fire incapable of before all Control room succeeds after contr I HH failing all HVAC HVAC is lost remains Inhabitable room evacuation FeHVPAC PNS PEVAc P.. Outcome OK OK OK OK Core damage Figure 1. Event Tree for Core Damage due to Fire-Induced Loss of Control Room IHVAC Initiation of a Firein FireArea HH The frequency of a fire in Fire Area HH was assessed to be approximately 4.0 x 1OE-3 per year for the Davis-Besse IPEEE [Ref. 1, Table 4.2.1.7]. This was the total frequency for fire initiation based on the data in the EPRI fire events database. Because the potential for a fire in this area was assessed to have a negligible impact on core damage frequency, it was not necessary to refine this frequency further.

As noted in Section 4.2.3 of the IPEEE submittal, however, a more careful review of the EPRI fire events data base indicated that a significant fraction of the fires had essentially no potential to cause significant damage. EPRI suggested severity factors that should be applied for each type of fixed source to determine the frequency of fires with the potential to cause damage [Ref. 2].

Similarly, EPRI had found that certain attributes (housekeeping practices, frequency of fire inspections, etc.) could be credited in reducing the effective frequency of fires from transient sources. EPRI suggested factors that could be applied to account for the effectiveness of these

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 10 provisions [Ref. 2]. A review of operating practices at Davis-Besse confirmed that the criteria were met, such that it was appropriate to apply these factors [Ref. 3]. Furthermore, fires due to welding or cutting activities were most often extinguished before there was time for the fire brigade to respond [Ref. 4].

The overall frequency can therefore be reduced by factors according to the type of source that contributes for a particular area. The refined frequency for Fire Area HH is summarized in Table 1. The total frequency of fires in Fire Area HH (including both fixed and transient sources) is therefore as follows:

Fi(HHO1) = 3.97E-4 /yr Table 1. Summary of Fire Initiation Frequency in Fire Area III FIVE/Fire PRA Frequency Adjustment Adjusted Ignition Source Contribution Factor Frequency Severity Initiation from Fixed Sources Factor Control room electrical cabinets 0.2 Electrical cabinets 1.34E-3 0.12 1.61E-4 Fire protection cabinets 7.74E-5 0.12 9.29E-6 Indoor transformers 8.40E-5 0.1 8.4E-6 Diesel generators - 0.4 Motor generator sets 0.14 Pumps 4.75E-4 0.2 9.50E-5 Ventilation subsystems 1.53E-3 0.08 1.22E-4 Total frequency from 3.51E-3 3.96E4 fixed sources Non-Supp.

Initiation from Transient Sources Probability Transient materials 1.57E4 0.65 1.02E-4 Welding/cutting 3.73E-4 0.15 5.60E-5 Welding/cables l - 0.15 Transient frequency adjusted for suppression 1.58E-4 Exposure factor 0.1 0.1 Inspection frequency factor 0.083 0.083 Total adjusted frequency from transient sources 1.32E-6

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 11 FireIncapable of Affecting All HVA C The frequency summarized in Table 1 accounts for all fires with the potential to cause damage in Fire Area HH. Because of the distribution of fixed sources in the area, not all fires could create conditions that would affect the normal control room ventilation and both trains of CREVS. The potential for fixed ignition sources to cause unavailability of ventilation includes the following considerations:

• With two exceptions, the electrical cabinets accounted for in this area are all sealed cabinets. The heat load is contained, and according to EPRI guidance, there is no potential for propagation to other equipment in the area [Ref. 4]. These exceptions are the control cabinets for the CREVS, which are ventilated. These are freestanding cabinets, and the heat load from a fire in either would not be sufficient for fire effects to propagate to the other cabinet.

• The pumps in the area are two small chilled water pumps (7.5 hp motors). These pumps were included conservatively in the IPEEE evaluation, but they are too small for inclusion according to the EPRI criteria [Ref. 2," .... count large pumps in core heat removal systems and associated support systems"]. They would be incapable of producing a fire of sufficient intensity to threaten CREVS equipment.

• The dry-type transformer is also in a non-ventilated cabinet and it would not affect CREVS availability. The contribution from a fire in the transformer could therefore be neglected.

• The ventilation subsystems include equipment in both the normal ventilation system and the CREVS. The ten subsystems account for the two trains of normal control room ventilation, two trains of CREVS, and the small ventilation units serving Room 603 itself.

Based on the configuration of the systems in the room, a fire in any of the components associated with the normal control room HVAC could not result in a total loss of both normal and emergency ventilation. It is assumed, however, that a fire originating in one train of the CREVS could lead to loss of both trains of CREVS and a loss of normal HVAC. It is assumed that the CREVS accounts for 40% of the HVAC subsystems in Room 603.

• The fire cabinet is in a separate room, and a fire within it would not produce sufficient heat or smoke to threaten CREVS components.

• Because the fire due to transient sources could occur essentially anywhere in the area, the frequency is not adjusted further.

Thus, the conditional probability that a fire left unsuppressed could affect all ventilation in the control room can be estimated as shown in Table 2. The result is as follows:

PHVAC = 0. 2 1

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 12 Table 2. Summary of Fire Initiation Frequency in Area 1111 Fraction Total Adjusted Relevant to Relevant Frequency Ignition Source Frequency HVAC Electrical cabinets 1.61E4 20% 3.22E-5 Fire protection cabinets 9.29E-6 0%

Indoor transformers 8.40E-6 0%

Pumps 9.50E-5 0%

Ventilation subsystems 1.22E14 40% 4.90E-5 All transient sources 1.32E-6 100% 1.32E-6 Total Frequency 3.97E4 0.21* 8.25E-5

• The ratio of the relevant frequency to the total adjusted frequency.

Fire SuppressedBefore All H VA C is Lost Room 603 is equipped with automatic detection, but currently has no provisions for suppression.

For the situation as currently assessed, no credit is given for manual suppression for fixed sources.

If an automatic suppression system were installed, the conditional probability that the fire could be suppressed before affecting all HVAC can be estimated. For purposes of this calculation, it is assumed that such a suppression system can be adequately represented by a wet-pipe system.

Based on EPRI data used in the IPEEE, the unavailability of a wet-pipe suppression system is 0.02. Therefore, the probabilities for failure of suppression can be applied as follows:

PNS = 1.0 (base case)

PNS = 0.02 (case with installation of automatic suppression)

Control Room Remains Inhabitable If HVAC remains available, it is assumed that there will be no need to evacuate the control room.

The fire will have no other direct impact on the ability to maintain core cooling. Therefore, it is assumed that this would lead to a state in which core cooling was successfully maintained.

If HVAC is not available, the operators would be instructed to shed certain loads to reduce the heat load in the control room. If these measures were not sufficient to keep the temperature in the control room acceptably low, the Shift Manager/ Unit Supervisor could initiate actions to provide alternative ventilation. This would entail opening the door to the control room and, if necessary, using portable fans to provide ventilation. This action is feasible and has been suggested by operations personnel, but it is not explicitly called for by procedure. Therefore, according to the methods employed in the human reliability analysis for the Davis-Besse PSA,

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 13 the failure to take this action has been evaluated as a non-proceduralized recovery action. The human failure event has been assigned the event name ZHACRAVR. The worksheet describing the evaluation is provided as Annex A. The probability of failure for this event, based on the probability of non-recovery, has been evaluated to be as follows:

PEVAC = PNR(ZHACRAVR) = 0.05 Core Cooling Succeeds After Control Room Evacuation Even if the control room must be evacuated, the operators may take local action to maintain core cooling. Aside from equipment associated with control room ventilation, the only equipment whose failure is modeled in the Davis-Besse PSA is control power for one of the main steam isolation valves (MSIVs), valve MS100. With respect to maintaining core cooling, however, this valve is called upon in the PSA only for steam generator tube ruptures (SGTRs). Therefore, its failure is not of concern for the case of a fire in Fire Area HH.

The potential for core damage to occur after evacuation of the control room was evaluated for a variety of control room fires, including for fires that would not directly damage equipment that played a role in core cooling. The conditional probability of core damage under this circumstance can be applied for this assessment as well. The conditional probability was calculated to be as follows:

PCD = 0. 0 7 9 Note that this value should be bounding relative to the assessment of the case in which evacuation is necessitated by the loss of control room ventilation. For a control room fire, the control room might need to be evacuated in as little as about 15 minutes. This would be accomplished while trying to respond to the fire itself. It is likely that there would also be some spurious signals and actuations that could distract the control room operators. For the loss of ventilation, the time available would be substantially longer, and conditions would deteriorate more slowly. The procedure would lead the operators to perform an orderly shutdown before it was necessary to evacuate the control room [Ref. 5]. This value is, however, a reasonable bounding probability for the current assessment.

Estintation of Core-DamlageFrequency The frequency of core damage associated with this scenario can be calculated from the following product:

FCD = F1(HHOI) x P11VAC X PNS X PEVAC X PCD For the base case, this frequency would therefore be as follows:

FCD = (3.97E-4 /yr)(0.21)(1.0)(0.05)(0.079) = 3.3E-7 /yr

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 14 For the sensitivity case, accounting for availability of an automatic suppression system, this frequency would therefore be as follows:

FCD = (3.97E-4 /yr)(0.39)(0.02)(0.05)(0.079) = 6.6E-9 /yr The frequency for the current condition is a small fraction of the overall core-damage frequency for Davis-Besse, which is 1.2E-5 /yr. The addition of suppression system could further reduce this frequency, but would it would constitute a negligible reduction in risk for the plant.

Additional perspective on this result can be gained by considering this situation in the context of a permanent change to the plant. According to Section 2.2.4 of Regulatory Guide 1.174 [Ref. 6],

the acceptability of the effects of a permanent change on CDF are considered in the context of three regions:

• Region 1 (ACDF greater than 10- 5/ year) - Applications that result in changes would not be considered.

• Region 2 (ACDF from 10-6 to1 0 5 / year) - Applications will be considered if it can be shown that the total CDF is less than 104 per reactor year.

• Region 3 (ACDF less than 10-6/ year) - The change will be considered regardless of whether there is a calculation of the total CDF.

Thus, if the increase in CDF is no higher than 10-6/ year, the change is deemed to be acceptable.

Even if the increase were larger than 10-6/ year, the change might still be judged to be acceptable.

The bounding ACDF calculated for this assessment is, 3.2 x 10-7/ year, well below the criterion of 106/ year for which changes are deemed to be acceptable without further justification.

Another perspective is available from the PSA Applications Guide, which provides a criterion for evaluating the impact of permanent changes in CDF [Ref. 7]. The Guide provides the following CDF screening criterion for permanent changes (applicable if the baseline CDF for the plant is less than 104/ year):

ACDF (%) lo[(-O.5Log(CDF))-I]

The baseline CDF for Davis-Besse is 1.22 x 10-/ yr. Therefore, the criterion according to this equation is as follows:

ACDF (%/.) = 10 [(-O.5Log(O.0000122))1]

= 29%

The frequency for the scenario assessed in this evaluation would be equivalent to approximately a 3% increase in the baseline CDF. Therefore, it would be well within the criterion from the Applications Guide.

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 15 It should be noted that this assessment also maximizes the assessed benefit of the incorporation of an automatic suppression system because it reflects the following:

• No credit was given to manual suppression for the base case.

• In the sensitivity case, automatic suppression was credited as being effective for all fires, if the system functioned. No penalty was taken based on the actual effectiveness of the suppression system.

• The potential that the automatic suppression system could cause damage due to spray effects, either when responding to an actual fire or as a consequence of spurious actuation, was neglected.

• The conditional probability of core damage given the need to evacuate the control room was calculated for the situation in which evacuation would be necessitated by a fire in the control room. This probability would be bounding for the situation in which evacuation resulted from a loss of control room ventilation, since there would be substantially more time to achieve a stable state and the conditions would be less challenging for the operators.

References

1. Individual PlantExanination of ExternalEvents for the Davis-Besse Nuclear Power Station. The Toledo Edison Company, December 1996.

2. FirePRA Implementation Guide. Electric Power Research Institute Report TR-105928, December 1995.

3. "Control of Transient Combustibles". Davis-Besse Nuclear Power Station Administrative Procedure DB-FP-00007, Rev. 5.

4. Fire-InducedVulnerability Evaluation (FIVE) Methodology. Electric Power Research Institute Report TR-1 00370, April 1992.

5. "Control Room Evacuation". Davis-Besse Nuclear Power Station Abnormal Procedure DB-OP-02508, Rev. 3.

6. "An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant Specific Changes to the Licensing Basis. U.S. Nuclear Regulatory Commission Regulatory Guide 1.174, July 1998.

7. PSA Applications Guide. Electric Power Research Institute Report TR-1 05396, August 1995.

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 16 Room Layout Sketch NOT TO SCALE

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 17 CREVS Skid (front viewv/ looking East)

I

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 18 CREVS Control Cabinets (front view/ looking South)

', , .,1

Docket Number 50-346 License Number NPF-3 Serial Number 3106 Page 19 Annex A Davis-Besse PSA Event ZHACRAVR (Non-Recovery Event)

Definition Failure to provide alternative ventilation after loss of control room ventilation Situation In the event of a total loss of ventilation for the control room, the operators would be instructed to take measures to reduce the heat load in the control room [Ref. 1]. If they were unable to reduce the heat load sufficiently to maintain an acceptable temperature in the control room, It could become necessary to shut down the plant and evacuate the control room. Core cooling would then be maintained by use of local panels.

Prior to the evacuation of the control room, it should be possible to establish adequate ventilation by opening the control room door and.

if necessary, to use portable fans to facilitate the exchange of air with the auxiliary building. Although this action is not called for explicitly by procedure, It has been identified as a logical response by operators.

The time available to take action would depend on the effectiveness of actions to reduce the heat load in the control room and on ambient conditions. It is reasonable to assume, however, that the time available would be on the order of 30 - 40 min. Therefore, it is characterized as 'short".

There is training related to similar actions for other areas in the plant, and there is procedural guidance for Initial response to the loss of ventilation. In particular, all licensed operators receive fire brigade training. This training Includes actions to provide temporary ventilation where necessary.

The actions that would be required would be relatively simple. The actions would only be successful If they were accomplished before conditions became harsh. Therefore, the environmental conditions are characterized as 'good'.

Influence-Based Assessment Time available short Training/practice, procedural guidance yes Qualitative likelihood of failure: moderately high Complexity of decision and actions simple Working environment good Mean probability: 5.OE-2 Alternative Assessment Use alternative non-recovery probability? l no l Alternative value:

Critical Assumptions

1. Shift Supervisor would be likely to initiate alternative cooling of the control room if necessary.

2. Time available to provide alternative cooling would be on the order of 30-40 min.

References Procedures Number Revision

1. Abnormal Procedure for Control Room Emergency Ventilation System Load Shedding DB-OP-02533 5 Other References Result I. . . . . . . . _ . . . . . . . : . . . . . . . .. . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 5.OE.2 5 I. .. . .. .. .

. - . ........ . . I. . - 1. . .. - - . ....... .... .. . .... ....... .

Docket Number 50-346 License Number NPF-3 Serial Number 3106 COMMITMENT LIST THE FOLLOWING LIST IDENTIFIES THOSE ACTIONS COMMITTED TO BY THE DAVIS-BESSE NUCLEAR POWER STATION (DBNPS) IN THIS DOCUMENT. ANY OTHER ACTIONS DISCUSSED IN THE SUBMITTAL REPRESENT INTENDED OR PLANNED ACTIONS BY THE DBNPS. THEY ARE DESCRIBED ONLY FOR INFORMATION AND ARE NOT REGULATORY COMMITMENTS. PLEASE NOTIFY HENRY L. HEGRAT, SUPERVISOR - LICENSING, AT (330-315-6944) OF ANY QUESTIONS REGARDING THIS DOCUMENT OR ANY ASSOCIATED REGULATORY COMMITMENTS.