Information Notice 2002-27, Recent Fires at Commercial Nuclear Power Plants in the United States

Recent Fires at Commercial Nuclear Power Plants in the United States
	ML022630147
Person / Time
Issue date:	09/20/2002
From:	Beckner W; NRC/NRR/DRIP/RORP
To:	;
	Petrone C , NRC/NRR/RORP, 415-1027
References
	-nr, TAC M3555 IN-02-027
	Download: ML022630147 (18)
	v • d • e

UNITED STATES

NUCLEAR REGULATORY COMMISSION

OFFICE OF NUCLEAR REACTOR REGULATION

WASHINGTON, DC 20555-0001 September 20, 2002 NRC INFORMATION NOTICE 2002-27: RECENT FIRES AT COMMERCIAL NUCLEAR

POWER PLANTS IN THE UNITED STATES

Addressees

All holders of operating licenses for nuclear power reactors, except those who have

permanently ceased operations and have certified that fuel has been permanently removed

from the reactor.

Purpose

The U.S. Nuclear Regulatory Commission (NRC) is issuing this information notice (IN) to inform

addressees of recent fire incidents at commercial nuclear power plants (NPPs) in the United

States. The NRC anticipates that recipients will review the information for applicability to their

facilities and consider taking appropriate actions. However, suggestions contained in this IN do

not constitute NRC requirements and, therefore, no specific action or written response is

required.

San Onofre Nuclear Generating Station (SONGS)

Description of Circumstances

On February 3, 2001, SONGS Unit 3, was operating at 39-percent power following a refueling

outage. While switching offsite power sources for Unit 3, a 4.16-kV breaker (3A0712) faulted

and initiated a fire. This resulted in a loss of power to Unit 3 nonsafety-related systems, a

reactor trip, a turbine/generator trip, and an automatic start of both Unit 3 emergency diesel

generators (EDGs).

The main control room (MCR) received an annunciator fire alarm, along with a visual report of

smoke and flames at the 30-foot elevation switchgear room of the turbine building. The incident

was further complicated when the MCR annunciators were lost as a result of a tripped breaker

approximately 5 minutes into the event.

The San Onofre onsite fire department was dispatched upon receipt of the fire alarm in the

switchgear room and arrived at the scene within 7 minutes. The on-scene fire department

captain requested additional support from an offsite fire department. Firefighters observed that

the room was completely filled with heavy smoke, with essentially zero visibility. The source of

the heavy smoke and heat was determined to be within the closed cubicle 4.16-kV switchgear

cabinetry. The firefighters also noted flames from burning instrument gauges on the front of

cubicle 3B14, which is located directly across from the 4.16-kV breaker cubicle. The onsite fire

department captain established a command post, initiated fire suppression using portable fire

extinguishers, and began ventilating the area. Communication between the onsite fire

department captain and the plant shift manager was through the technical advisor at the scene.

The firefighters discharged portable Halon and dry chemical fire extinguishers through the

cabinet vents in an attempt to extinguish any active fire within the cabinet. The extinguishing

agents had no noticeable effect on the production of smoke. The technical advisor transmitted

to the operations shift manager a report that the fire was out, although the fire department

captain only advised the operations technical advisor that flames were no longer visible.

With the exception of some low-voltage circuits, all power was isolated to the 4.16-kV

switchgear. The firefighters then determined that the cubicle door could be opened safely.

Upon opening the cubicle door, the firefighters observed flames within the cubicle, and

discharged additional dry chemical in another attempt to extinguish the flames. The firefighters

then closed the cubicle door as a containment measure. The cubicle door was subsequently

opened several times, and each time the door was opened, in-rushing air caused the fire to

reflash. Firefighters then used dry chemical each time the fire reflashed.

The fire department captain advised the operations technical advisor that the fire could not be

completely extinguished unless the firefighters applied water to the fire. It appeared that the dry

chemical temporarily removed air from the fire, but did not reduce the heat, and the fire would

reflash once air was reintroduced. The operations technical advisor relayed this request to the

shift manager for permission to use water on the smoldering area inside the cubicle to prevent

reflash. Because he was concerned that the buses were still energized with 125-v dc and low- voltage ac power, the shift manager initially denied the fire department captains request to use

water. However, after the fire department captain spoke directly with the shift manager to

advise him that the deep-seated fire could not be extinguished unless water was applied, the

shift manager granted permission to use water to extinguish the fire. The fire was ultimately

extinguished after firefighters applied water. The deep-seated fire burned for approximately 3 hours3.472222e-5 days 8.333333e-4 hours 4.960317e-6 weeks 1.1415e-6 months before finally being extinguished. The licensee later determined that communication

weaknesses in identifying the actual fire status during the event contributed to the delay in

extinguishment.

Discussion

The extensive damage made it difficult to determine the exact cause of the fault. The licensee

found that the 4.16-kV switchgear phase C arcing contact had completely melted, and

concluded that the phase C circuit breaker failed to close completely during the bus transfer.

The breaker was approximately 25 years old and had its last preventive maintenance performed

in 1997. The licensee also believed that arcing, fire, smoke, and ionized gases in the 4.16-kV

circuit breaker caused multiple faults on a 3A07 bus and the offsite power circuit terminal

connection at circuit breaker 3A0714.

The licensee also determined that the fire event generated a much higher heat release rate

(HRR) than would normally be assumed in typical fire risk modeling to perform probabilistic risk

assessment (PRA). In A Supplement to EPRI Fire PRA Implementation Guide (TR-105928),

report SU-105928, the Electric Power Research Institute (EPRI) provides data for electrical

cabinet fires, indicating an HRR of either 68.60-kW or 200.50-kW (65- or 190-Btu/sec),

depending on the type of cable installed. The EPRI data focus on the HRR contributions of

combustibles in the electrical cabinet (only cable insulation) and neglect the large amounts of

electrical energy that may be released from electrical faults. According to a report by the

NRCs Office of Nuclear Regulatory Research, entitled Operating Experience Assessment Energetic Faults in 4.16-kV to 13.8-kV Switchgear and Bus Ducts That Caused Fires in Nuclear

Power Plants in 1986-2001 (ADAMS Accession #ML021290358, February 2002) for medium- and high-voltage applications, the research indicates that these HRR values [68.60-kW and

200.50-kW (65- and 190-Btu/sec)] may be underpredicted by a factor of 1,000.

This operating experience indicates that equipment rated at 4.16-kV and higher is vulnerable to

particularly energetic electrical faults. This event demonstrates that energetic electrical faults

instantaneously release large amounts of electrical energy and may bypass the normal fire

initiation and growth stages. In the SONGS fire event, the equipment that caught fire was

directly connected to the auxiliary transformer (AT), which is powered from the grid or main

generator. If a circuit breaker is stuck or slow in responding, there is sufficient energy to cause

an explosion and vaporize metal in a few cycles.

Conclusion

This event demonstrates the importance of using water to extinguish deep-seated electrical

cable fires. It is similar to previous fire events (Browns Ferry 1975, Waterford 1995) in which

delayed application of water on electrical fires extended the duration of the fires and delayed

recovery from the events. It is essential that fire brigade and operator training address the

appropriate use of water in firefighting operations in energized electrical equipment. This event

also highlights that the HRR from fires in electrical cabinets may be much greater than

assumed in NPP fire hazard analysis (FHA).

Point Beach Nuclear Plant

Description of Circumstances

On April 24, 2001, while Point Beach Unit 1 was shut down and defueled for refueling outage

U1R26, a fire occurred in the A steam generator (S/G) vault on the access platform to the

primary side manway covers. The fire was believed to originate as the result of a short in a

12-Vdc communication box. The fire consumed a bag of rags and testing equipment debris, and lasted for approximately 23 minutes. After multiple failed attempts in which the fire brigade

discharged approximately 70 pounds of dry chemical (3 portable fire extinguishers), the fire was

finally extinguished using 15-20 gallons of water. The licensee reported that approximately 50

percent of the containment basement floor (8 feet elevation), 50 percent of the A S/G vault, and 30 percent of the A reactor coolant pump (RCP) vault were covered in white dust (dry

chemical fire extinguishing agent). Also, a white dust layer was visible on components on the

main refueling floor (66 feet evaluation). Smoke and soot resulting from the fire left a mark

about 4 feet wide by 25 feet high against the vault wall.

Discussion

The dry chemical extinguishing agent is discharged by an inert gas when a fire extinguisher is

used. All forms of dry chemical act as extinguishing agents to suppress the flame of a fire

(Friedman, 1998), but may require extensive cleanup after use, as illustrated by this event. Most chemical extinguishing agents can produce some degree of corrosion or other damage, but of the seven types of dry chemicals, monoammonium phosphate is especially acidic and

tends to corrode more readily than other dry chemicals, which tend to be more neutral or mildly

alkaline. Furthermore, corrosion resulting from the other dry chemicals is stopped in a

moderately dry atmosphere, while phosphoric acid generated by using monoammonium

phosphate has such a strong affinity for water that an exceedingly dry atmosphere would be

needed to stop the corrosion.

Application of dry chemical agents on electrical fires is considered a safe practice from the

viewpoint of electric shock. However, these agents, especially monoammonium phosphate, can damage delicate electrical equipment.

One potential issue with using dry chemical extinguishers results from the sudden release of

the agent and the large area of discharge. Dry chemicals become sticky when heated and, therefore, are not recommended for locations where it may be difficult to remove residue from

equipment. It is important to note that when water is applied to the affected areas, corrosion

will occur because moisture initiates a chemical reaction that accelerates corrosion of

equipment exposed to the dry chemical.

Dry chemicals are generally nontoxic, but can pose a health hazard when used in closed areas.

Persons who breathe concentrations of the dry chemical powder may experience respiratory

irritation and coughing. When dry chemicals are discharged into an enclosed area, impaired

breathing and reduced visibility should be considered.

Conclusion

Although the Point Beach incident lasted approximately 23 minutes, it was not a large fire in

terms of HRR. The dry chemical extinguishing agent did suppress the fire, but failed to

completely extinguish the fire (the fire reflashed twice). The fire brigade unsuccessfully

attempted to extinguish the fire with dry chemical agent three times before easily extinguishing

it with a hose line (water). A more thorough selection of extinguishing media should be

considered in light of the cleanup effort from the small fire. It is important to recognize that the

fire was successfully extinguished with a relatively small quantity of water, which required

minimal post-fire cleanup.

Prairie Island Nuclear Generating Plant

Description of Circumstances

On August 3, 2001, at 8:44 p.m., an operator enroute to the Unit 1 bus 11/12 area observed fire

and smoke, but could not identify the cubicle from which it was originating. The operator

entered the bus 13/14 room and called the main control room (MCR) to report the fire. The

MCR immediately initiated the fire alarm and activated the onsite fire brigade. The MCR also

notified the offsite Red Wing Fire Department (RWFD).

The fire brigade entered the turbine building to assess the extent and exact location of the fire.

They reported flames in the upper and lower compartments of the 12-4 cubicle and along the

left side of the breaker. They also found that the door in both the upper and lower

compartments of the cubicle were blown open. At 8:58 p.m., the fire brigade began initial suppression of the fire using three portable carbon

dioxide (CO2) extinguishers and one Halon extinguisher through the open front door of the

breaker cubicle. The fire was not extinguished, and the fire brigade observed electrical arcing

in cubicle 12-4.

The fire appeared to be localized in one area and not spreading. The initial efforts to

deenergize the bus from the MCR failed. The fire brigade chief reported to the MCR that

breaker 12-4 was still energized, as evidenced by arcing observed in cubicle 12-4. Because of

the uncertainty as to whether bus 12 was deenergized, the Unit 2 shift supervisor decided to

deenergize the 1R transformer. The fire department reported to the MCR that there were small

flames and heavy smoke in breakers 12-1 and 12-4. At 10:13 p.m., approximately 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months

into the event, the fire brigade extinguished the fire with assistance from the RWFD.

Discussion

The fire was extinguished after 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months by using more than 20 portable CO2 fire extinguishers

in the evolution (in addition to the 3 CO2 extinguishers and 1 Halon extinguisher used in the

initial attack). One factor that complicated extinguishing the fire was the decision not to use

water because of energized electrical equipment. This resulted in continued burning and

elevated temperature. Because of the elevated temperature caused by this electrical fire, two

fire brigade members were treated for heat exhaustion at the site, and one of them was

subsequently transported to the hospital for further treatment. In addition, several inches of the

copper feed stabs from the 1M transformer completely vaporized during this fire (providing

additional evidence of high temperature).

The licensee determined that the cause of the event was a poor electrical connection between

the breaker 12-4 C-phase primary disconnect assembly (PDA) and the 1MY bus stab, which

caused the PDA to overheat. The arcing also actuated the protective relaying, which resulted in

an automatic turbine/reactor trip. The arcing event at breaker 12-4 released enough energy to

cause the cubicle to expand and the door to be blown open. The breaker compartment was

heavily oxidized and holes were burned through the cubicle on either side of the breaker. The

arcing event destroyed many of the springs and fingers in the PDAs. A few were found at the

very bottom of the debris, particularly below C phase.

Conclusion

The root cause evaluation of the nonsafety-related breaker fire concluded that maintenance

practices could have contributed to the failure of the PDA by creating a poor connection, which

caused localized over heating of parts of the PDA. This overheating caused the PDA to

disintegrate. At that point, the loose parts of the PDA created a short-to-ground path. Once the

arc was struck, phase-to-phase faulting occurred between the A-B and B-C phases. The initial

arcing to ground quickly interrupted the dc circuit below the breaker pan (located directly below

the PDA).

In this fire event, the use of portable CO2 and Halon fire extinguishers may not have been the

most effective choice of extinguishing agent to use. Operating experience in energized

electrical equipment fires shows that the use of a relatively small quantity of water was effective

in successful fire extinguishment. Fort Calhoun Station

Description of Circumstances

In October 2001, the licensee for Fort Calhoun Station began a surveillance of the Unit 1 containment prestressing system. This surveillance included testing the tension of the

containment concrete tensioning cables. It also involved pumping lubricating grease into the

containment tendon sheathings to replace the grease that had been lost as a result of leakage.

In support of this activity, 55-gallon drums of grease were located in the tension gallery. During

this surveillance activity, the plant personnel discovered that the grease was too cold to pump

and would need to be heated before use. Drum heaters were, therefore, used on the outside of

the drums to heat the grease and facilitate pumping into the containment tension sheathings.

Two drum heaters were used, one powered from a receptacle located in the tension gallery and

the second powered from a receptacle located in room 22. In order to supply power from the

outlet in room 22 to one of the drum heaters, two extension cords were connected in series and

routed through the open door separating room 22 from the tension gallery. At the end of the

day, the drum heater powered from the receptacle in room 22 was left energized to keep the

grease warm overnight so that work could begin the next morning.

Unbeknownst to plant personnel involved in performing the surveillance, the extension cords

used to power the drum heater were not rated for this application. The extension cords were

rated at 15 amperes, and had male connections that would only allow them to be connected to

15-ampere receptacles. However, the 20-ampere male connection on the drum heater had

been inappropriately modified to allow it to be connected to a 15-ampere plug or receptacle.

The licensee later determined that the, 2000-watt drum heater drew a current of 17.39 amperes.

As a result of using underrated extension cords, the extension cords continued to heat up

during the evening. The extension cords eventually overheated and ignited the plastic on the

radiological control point stepoff pad and a rubber air hose.

On December 19, 2001, at 2:48 a.m., the MCR operators received an alarm from an ionization

smoke detector located in room 22. A control room operator dispatched the auxiliary building

operator and a radiation protection technician to investigate the cause of the fire alarms. The

auxiliary building operator arrived at the door to room 22, cracked the door open, and

determined that there was too much smoke to enter the room without using protective

firefighting bunker gear and a self-contained breathing apparatus (SCBA) and informed the

MCR. The fire brigade was activated while operators entered the abnormal operating

procedure for fighting fires. During this event, the MCR received another ionization smoke

detector alarm in corridor 4.

The fire brigade laid out an attack line from the hose cabinet outside room 22 and a backup line

from the cabinet outside room 6 before the attack team prepared to enter room 22. The attack

team entered room 22 and proceeded down the stairs toward the entrance to the containment

tension stressing gallery. The nozzle man described room 22 as being completely filled with

smoke with no visibility. The smoke that traveled from room 22 through the open door caused

the actuation of the water curtain open head deluge system on the auxiliary building stairwell, which resulted in water being sprayed onto safety-related motor control centers (MCCs), which

subsequently caused actuation of the 480-V bus ground alarms in the MCR. These MCCs are safety-related but are not required to function during a safe shutdown event, as defined by Appendix R to Title 10, Part 50, of the Code of Federal Regulations (10 CFR

Part 50). Operators also restarted the room 22 ventilation to remove the smoke.

Discussion

Licensee personnel performed unauthorized modifications to the male connections of two drum

heaters, allowing them to be inserted into underrated outlets and extension cords, which

ultimately caused the fire. The licensee concluded that the root cause of the fire was the

modification of the male connection on a 2,000-watt drum heater. The plug on a second

2,000-watt drum heater was also found to be modified. This unauthorized modification

defeated a manufactured safety device (electrical connector standards), thereby allowing the

heaters to be energized using undersized extension cords and electrical outlets. On one of the

plugs, a prong was twisted 90 degrees to make it similar to a 15-ampere plug. On the other

heater, the plug was completely removed and replaced with a 15-ampere plug (see illustrations

below).

Standard Standard Altered

15 Amp Receptacle 20 Amp Plug 20 Amp Plug

G G G

W W W

Conclusion

The fire was a result of modified plugs on two drum heaters, which defeated the intent of the

design of electrical outlets. The licensee failed to with comply the procedural requirements for

a temporary modification. The heavy smoke from the fire caused a deluge sprinkler system to

actuate in a different fire area which sprayed on safety-related electrical equipment. 10 CFR

50.48(a) requires licensees to have a fire protection plan that meets General Design Criterion (GDC) 3 of Appendix A to 10 CFR Part 50. GDC 3 requires that structures, systems, and

components that are important to safety shall be designed and located to minimize, the

probability and effect of fires and explosions, consistent with other safety requirements.

Summary of Recent Fire Events

This IN discusses four separate, noteworthy fire events. These events indicate that addressees

should give special consideration to the means and associated effects of fire extinguishment.

The use of water on electrical cable fires is discussed in Section 9.5.1 (page 9.5.1-15) of the

NRCs Standard Review Plan (SRP), NUREG-0800, dated April 1996, as follows: Experience with major electrical cable fires shows that water will promptly extinguish

such fires. Since prompt extinguishing of the fire is vital to reactor safety, fire and water

damage to safety systems is reduced by the more efficient application of water from

fixed systems spraying directly on the fire, rather than by manual application with fire

hoses. Appropriate firefighting procedures and fire training should provide the

techniques, equipment, and skills for the use of water in fighting electrical cable fires in

nuclear plants, particularly in areas containing a high concentration of electric cables

with plastic insulation. This is not to say that fixed water systems should be installed

everywhere. Equipment that may be damaged by water should be shielded or relocated

away from the fire hazard and the water. Drains should be provided to remove any

water used for fire suppression and extinguishment to ensure that water accumulation

does not incapacitate safety-related equipment."

The affected and nearby equipment may need to be deenergized to eliminate the potential for a

personnel shock hazard. Fire brigade training and pre-fire plans should address fighting fires in

energized electrical equipment, as well as the responsibility to deenergize and open cabinets

for safe access to concealed fires (such as in electrical cabinets). All of these activities

contribute to the duration of and recovery from the event.

The NRC expects addressees to evaluate the above information for applicability to licensed

activities. However, this IN does not require any specific action or written response. If you

have any questions about this notice, please contact one of the technical contacts listed below

or the appropriate project manager in the NRCs Office of Nuclear Reactor Regulation (NRR).

/RA/

William D. Beckner, Program Director

Operating Reactor Improvements Program

Division of Regulatory Improvement Programs

Office of Nuclear Reactor Regulation

Technical Contacts: Naeem Iqbal, NRR Mark Henry Salley, NRR

(301) 415-3346 (301) 415-2840

Email: nxi@nrc.gov Email: mxs3@nrc.gov

Phillip Michael Qualls, NRR

(301) 415-1849 Email: pmq@nrc.gov

Attachments:

1. List of Recently Issued NRC Information Notices

2. References, List of NRC Generic Communications, and

List of Recently Issued NRC Information Notices Related to Fire Protection Experience with major electrical cable fires shows that water will promptly extinguish