L-94-104, Forwards Documentation of Util Approach for Resolving Thermo-Lag Fire Barrier Issues at Facilities
| ML17352A574 | |
| Person / Time | |
|---|---|
| Site: | Turkey Point, Saint Lucie  |
| Issue date: | 04/29/1994 |
| From: | Bohlke W FLORIDA POWER & LIGHT CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| L-94-104, NUDOCS 9405100274 | |
| Download: ML17352A574 (56) | |
Text
ACCELERATED DISTRIBUTION DEMONSTPA.TION SYSTEM t
j <<-.'i,'l.
REGULATORY INFORMATION DISTRIBUTION SYSTEM (RIDS)
DOCKET g
05000250 05000251 05000335 05000389 R
ACCESSION NBR:9405100274 DOC.DATE: 94/04/29 NOTARIZED:
NO FACIL:50-250 Turkey Point Plant, Unit 3, Florida Power and Light C 50-251 Turkey Point Plant, Unit 4, Florida Power and Light C 50-335 St. Lucie Plant, Unit 1, Florida Power
& Light Co.
50-389 St. Lucie Plant, Unit 2, Florida Power
& Light Co.
AUTH.NAME AUTHOR AFFILIATION BOHLKE,W.H.
Florida Power
& Light Co.
RECIP.NAME 'ECIPIENT AFFILIATION Document Control Branch (Document Control Desk)
D S
NOTES:
SUBJECT:
Forwards documentation of util approach for resolving Thermo-Lag fire barrier issues at facilities.
DISTRIBUTION CODE A029D COPIES RECEIVED:LTR t ENCL t SIZE: MD
'ITLE: Generic Letter 92-008 Thermal-Log 330 Fz.re Barrier RECIPIENT ID CODE/NAME PD2-2 LA CROTEAU,R INTERNAL: NRR/DE/EELB NRR/DSSA/SPLB RGN2
....FILE EXTERNAL: NRC PDR COPIES LTTR ENCL 1
0 1
1 1
1 2
2 1
1 1
1 RECIPIENT ID CODE/NAME PD2-2 PD NORRIS,J NRR DRPW PD3-1 EG E
01 NSIC COPIES LTTR ENCL 1
1 1
1 1
1 1
1 1
1 D
D D
NOTE TO ALL"RIDS" RECIPIENTS:
D D
PLEASE HELP US TO REDUCE WASTE! CONTACT THE DOCUMENT CONTROL DESK, ROOM Pl-37 (EXT. 20079) TO ELIMINATEYOUR NAMEFROM DISTRIBUTION LISTS FOR DOCUMENTS YOU DON'T NEED!
TOTAL NUMBER OF COPIES REQUIRED:
LTTR 12 ENCL 11
p
~
I>
I 0
P.O. Box 14000, Juno Boa oh, FL 33408-0420 I 94-104 10 CFR 50 APP.R April29, 1994 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, D. C. 20555
Subject:
St. Lucie Units 1 & 2 and Turkey Point Units 3 & 4 Docket Nos, 50-335, 50-389, 50-250, and 50-251 Performance-based A
roach for Resolvin Thermo-La Fire Barrier Issues In the March 10, 1994 meeting, Florida Power & Light (FPL) presented to the NRC a performance-based approach for resolving Thermo-Lag fire barrier issues at the Turkey Point and St. Lucie facilities.
On April 14, 1994, FPL and NRC staff reviewers met to discuss some of the specifics of the FPL fire methodology.
The purpose of this submittal is to provide documentation of the FPL approach for resolving Thermo-Lag fire barrier issues at both plant facilities. Please find attached:
Attachment A: A summary of the FPL Thermo-Lag resolution methodology including performance-based acceptance criteria; Attachment B:
A description of the fire barrier evaluation methodology used for evaluating Thermo-Lag fire barriers, and Attachment C:
A description of the Probabilistic Safety Assessment (PSA) techniques used by FPL as an independent method for examining fire risk in areas that have Thermo-Lag fire barriers.
The methodology described in Attachments A, B, and C meets the objectives of Appendix R by integrating performance-based acceptance methodology forfire barrier evaluations and PSA verification techniques.
This methodology includes conservative criteria and safety factors, conservative fire barrier evaluation methodologies, and PSA verification techniques.
Further, as discussed at the March 10 and April 14 meetings, FPL has used the Thermo-Lag product for outdoor fire barriers at the Turkey Point Nuclear facility. These fire barrier configurations are unique to the Turkey Point site, and to our knowledge, there has been little or no outdoor use by other utilities within the industry. These are low combustible areas which are not subject to fire damage from stratiGed gases or ceiling jet layers that can occur from a fire in an indoor area.
Ar.r p 940S 100274 940429 PDR ADOCK 05000250
.F..
t
't C
ac Xs
'sw'
U.S. Nuclear Regulatory Commission Page L94-104 10 CFR 50 APP.R Attachment D to this document provides the description and bases for meeting the objectives of Appendix R for Turkey Point outdoor areas.
This would allow the use of 30-minute Thermo-Lag fire barriers for shutdown functions that are located 50 or more feet away from major in-situ combustibles.
This use offire barriers outdoors provides equivalent or better protection as compared to that ofradiant energy shields in large indoor areas, such as the reactor containment.
In order to maintain FPL's current Thermo-Lag resolution schedule as described in FPL's letters I 94-24 dated February 7, 1994 and I 94-33 dated February 11, 1994, we ask that you review these submittals at your earliest opportunity and communicate your estimated schedule for informing us ofyour Gndings and comments.
We willcontact the NRC project managers periodically for a status on this request.
Please let us know ifwe can be of any help during the review process.
Very truly yours,
. H. Bo lke Vice President Nuclear Engineering and Licensing WHB:abk Attachments 4 cc: Stewart D. Ebneter, Regional Administrator, Region II, USNRC (2 copies)
Senior Resident Inspector, USNRC, Turkey Point Plant Senior Resident Inspector, USNRC, St. Lucie Plant
)
a f
L-S4-104 Pays 1 of 5 St. Lucie Units 1 A 2 Turkey Point Units 3 R 4 Docket Nos: 50-335, 50-389, 50-250, 50-251 Attachment A SUFI'NARY OF FPL APPROACH FOR RESOLVING THERMO-LAGFERjK BARRIER ISSUES FLORIDAPOWER &LIGHT
C l
'I,
.i f
SUMMARY
OF FPL APPROACH FOR RESOLVING THERMO-LAGFIRE BARRIER ISSUES L94-1%
Attachment A Page20f 5 Introduction 1
As in the case with other licensees, Florida Power & Light (FPL) has used the Thermo-Lag product for fire barriers that protect safe shutdown functions at its nuclear facilities. Now that the industry testing programs are nearing completion, it is clear that the ASTM E119 ratings for this fire barrier product are less than the original design and installation speciflcations.
While investigating methods for resolving Thermo-Lag fire barrier issues, FPL has implemented interim compensatory measures at its nuclear facilities to assure continued compliance with Appendix R and other fire protection requirements, as appropriate.
Appendix R criteria for fire barrier rating selection are extremely conservative and as the industry and NRC have learned, are very difficultand expensive to implement.
The selection of one and three hour fire barriers in Appendix R has provided the designer with a method for meeting the defense in depth objectives of Appendix R. These defense in depth objectives are required in order to assure protection of safe shutdown f'unctions during postulated fires. This method of achieving Appendix R compliance does not require the evaluation ofactual fire area combustible loading or combustible configurations. However, when actual combustible loadings are determined and compared to one and three hour fire barriers, it is clear that many fire areas inside nuclear facilities have levels of safety far beyond those needed to assure protection of safe shutdown functions.
Hence, in many fire areas a reduction in fire barrier rating, as determined during recent industry testing, may have no impact on the defense in depth objectives of Appendix R.
FPL has developed a methodology that integrates fire methodologies and Probabilistic Safety Assessment techniques withperformance-based criteria and safety factors. This approach allows FPL fire protection experts to evaluate the fire areas and determine if the defense in depth objectives of Appendix R are still met with reduced credit for Thermo-Lag fire barriers.
The following discussion of FPL's approach is also provided in the logic flow chart format at the end of this summary.
It is noted that since the use of Thermo-Lag in outdoor areas is applicable only to the Turkey Point site and not of generic industry concern, FPL's resolution approach for outdoor areas is provided as Attachment D in this submittal.
Fire Methodologies Fire Hazards Screenin A fire hazards screening evaluation is conducted using methods from the National Fire Protection Association's (NFPA) "Fire Protection Handbook."
The screening method is a traditional fire hazards evaluation which is used to determine if a more detailed analytical approach should be employed. The intent of the fire hazards screening is to filter out areas which contain combustible loadings that are clearly incapable of threatening the flre barriers.
r I
L94-1N A~hment A Pago 3 of S The combustible loading is calculated based on a review ofin-situ combustibles in the fire area.
This review includes drawings and a walkdown of the fire area.
The uniformity or the non-uniformity of the fire load is assessed during the walkdown. The combustible loading is converted to an equivalent fire duration as described in the NFPA's Fire Protection Handbook.
The resulting fire duration is compared to the installed fire barriers, employing appropriate fire barrier ratings as determined through the use of industry tests and the NEI (NUMARC)
Application Guide.
The results of the fire hazards screening analysis are considered acceptable when:
(a) for areas that have detection but are without automatic su ression the as-installed fire barrier ratings must be a factor of two or more greater than the actual Pre loading; or (b) for areas that have detection and automatic su ression for the fire barrier, an as-installed fire barrier rating is credited with a three fold improvement, and the resulting enhanced barrier rating must be a factor of two or more greater than the actual Pre loading.
When the results of the screening analysis are found to be acceptable, no further fire evaluations are necessary and it is concluded that the fire area meets the objectives ofAppendix R.
Fire Modelin A fire analysis is performed in accordance with a fire model based on the Electric Power Research Institute (EPRI) report, "Fire Induced VulnerabilityEvaluation" (FIVE)methodology.
The fire model method willbe used for barriers that do not meet the screening method or for areas that are not appropriate for the screening method.
The flammable transients for a given fire area are assumed to ignite and burn until totally consumed. The fire modeling technique conservatively assumes that the worst case fire barrier in the fire area is directly over or adjacent to the fire plume. The barrier is evaluated against this worst case impingement fire. Additionally, the model evaluates the worst case ceiling jet outside the plume and determines the potential impact on the worst case fire barrier.
The results of the fire modeling analysis are considered acceptable when:
(a) the fire burn time is less than one-half the as-installed fire barrier ratings.
This fire model method provides assurance that worst case fires would burn out in less than one-half the time of the protection provided by the installed fire barriers.
Probabilistic Safety Assessment (PSA) Techniques PSA techniques willbe used to evaluate fire risk for areas containing Thermo-Lag fire barriers after the areas have passed the fire hazard screening or modeling analysis described above. The
I t
L94-IQ Attachment A Page 4 of 5 PSA fire model uses the EPRI "FIVE" methodology to screen or estimate core damage frequency in the fire areas.
PSA Anal sis The fire ignition frequencies are combined with the failure probability of alternative success paths for placing the plant in a safe shutdown condition. It is first assumed that the plant's unprotected cable functions and in-situ equipment located in the fire compartment fail and there is no credit for detection, suppression, or plant fire brigade response. Ifthe resulting core damage failure frequency is less than 1.0 x 10~ per reactor year (or 1 x 10~ per reactor year for compartments with automatic detection and suppression), the fire compartment is screened and no further PSA evaluation is needed. Ifthe fire compartment does not pass the initialscreening process, additional analysis that considers suppression systems, additional operator fire-related recovery actions, fire geometry considerations, etc. may be performed. Ifat anytime during this analysis the core damage frequency is calculated at less than approximately 1.0 x 10~ per reactor year for a fire compartment, the compartment is considered to have acceptably low risk and no further PSA work is conducted.
The PSA model includes an estimate of the failure probability of cables protected by Thermo-Lag.
This unique feature in the PSA model allows FPL to determine if the core damage frequency of a fire compartment is driven by or is sensitive to Thermo-Lag fire barrier failure.
Ifit is determined that Thermo-Lag fire barriers are the cause of a realistic fire vulnerability, appropriate plant modifications and procedure changes willbe made.
Conclusions FPL believes that this comprehensive performance-based approach willdetermine:
(1) ifnuclear plant indoor fire areas are able to meet the defense in depth objectives of Appendix R with Thermo-Lag fire barriers as currently configured, or (2) ifplant modifications are required.
l-94-104 Attachment A Page 5 of5 FPL PERFORMANCE-BASED APPROACH TO THERMO-LAG lSSUE RES OLUTIOM ST. LUCIE UNITS 1 8c 2 aiid TURKEY POINT UNITS 3 &4
,;,;:;ii4chm?;.)"".,)j($?III>&I,;,,NO
,.<g:j$!ITHN%04A'CUED:Ie,,,
- ?ihlIQSIIll,'iORCUIT:,.":,Ii',I'i;:?ii,',"
(C'I'hj%~:I'REDO (REPA'IIIIIIIIP YES UPDATE SSA i,'.4'i?II3imtii'f5 "g'/ah@
'UNS;:ip?I>?f',";, i
- ?(8;?Q?;@/??$ )gc?;;I
.;,:I@!SIItltR...
ANI'.OUTDO O II'IN?'~'.I?i,".NI:,;::,gjRRiER'RATINi.;:!I.:,lI.;tNO
'.?I?m;'I!aKjkpSEPTA@E,:!Imt."N!?
YES
- j-...!;.F?UIEnWPZARPSN)% KIODEI!',,esp YES MAINTAIN EVAlUATION FOR NRC REVIEW I": I",,;'::~i@7ENTIII;".,':-;
L-94-IQ Page 1 of 21 St. Lucie Units 1 & 2 Turkey Point Units 3 & 4 Docket Nos: 50-335, 50-389, 50-250, 50-251 Attachment B PERFORMANCE-BASED APPROACH TO FIRjE MdRjRjXREVALUATION:
DESCRIPTION ANDMETHODOLOGY FLORIDAPOWER & LIGHT
s a
r L94-1Q Amchmcot B Pago 2 of 21 Table of Contents Section Title
~Pa e
Introduct1on II.
General Descri tion III.
Fire Hazards Screenin Methodolo IV.
Fire Modelin Methodolo V.
Fire Model Detailed Descri tion A.
Desi n Basis Fire B.
Identification of Fire Scenarios C.
E osure Fire Source Characterization 12 12 12 13 D.
Peak Fire Intensity Transient Combustible...
~Ti~ Pl Convection Radiation..............
Fire Location Factor......
Virtual Origins Plume Temperature Rise...
Tar et in the Ceilin Jet...
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
13 13 14 14 14 15 15 16 16 Unconf'ined Ceiling Jet Confined Ceiling Jet......
F.
Tar et in the Hot Gas La er
~.
~RCh T
H.
Validation...
~
~
~
16 17 17 18 18 VI.
References, 18 VII. ~UFi 19 VIII. ~Glosss 19
I I
1 Ct L94-10t Attachment B Pttt;c 3 of 21 I.
Introduction Historically, compliance with 10CFR50 Appendix R fire barrier requirements has been a relatively simple matter of applying a nominal criteria of "3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />" fire barriers (or "1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />" fire barriers with suppression and detection) for separation of redundant trains of safe shutdown equipment/components/cables.
These general criteria are based on conservative assumptions which allow the criteria to easily bound most actual plant situations; however, since they give no consideration to actual combustible loadings and other location specific characteristics, they are overly conservative in many situations.
Advances in fire modeling techniques have made this conservatism evident.
As a result of recent Thermo-Lag fire barrier adequacy issues, Florida Power and Light (FPL) has re-examined the fire barrier criteria used in the past.
As an alternative, a performance-based approach to fire barrier evaluation can be used to meet the objectives of 10CFR50 Appendix R. The performance-based approach involves a quantitative determination of fire barrier adequacy for specific plant areas through an evaluation of combustible loadings and other location specific characteristics.
The result is a postulated fire hazards analysis, which is based on a technically sound methodology.
This methodology addresses location specific fire barrier adequacy, while maintaining conservative factors of safety.
II.
General Descri tion FPL's performance-based approach to fire barrier evaluations is comprised of a two-step process which determines case-by-case fire barrier acceptability.
The first step involves a conservative fire hazards screening which is based on traditional National Fire Protection Association (NFPA) methodology. A combustible loading derived fire duration is calculated for each compartment or fire area and compared to the industry test derived fire barrier rating using a conservative safety factor.
The second step involves conducting a detailed fire model of areas which do not pass the initial screening.
The fire model is based, in part, on the EPRI "FIVE"(Fire Induced Vulnerability Evaluation) methodology, and includes consideration of flammable transients and transient combustibles, which could potentially ignite bulk in-situ combustibles.
Key characteristics (such as relative position of ignition sources) are evaluated to determine ifthe ignition source could challenge the fire barrier or create conditions which could ignite the bulk in-situ combustibles.
The performance-based approach demonstrates that 1) redundant safe shutdown trains remain free from fire damage, or 2) modifications (including possible fire barrier upgrades) willbe pursued to achieve safe shutdown objectives. A detailed description of the fire hazard screening methodology and fire modeling methodology is provided in the following sections.
Figure 1 provides a flowchart for the general process.
Fi re 1
- Fire Barrier Evaluation Process Flowchart I 94-lQ Attachment B Page 4 of 21 START IN-SITU FIRE DURATIONVS. TESTED BARRIER RATING w FACTOR FOR SUPPRESSION
- 8. DETECTION BARRIER SAFETY FACTOR 2?
YES NO FIRE DURATIONVS.
BARRIER RATING BARRIER SAFETY FACTOR YES DOCUMENT NO PSA MODIFICATION
L94-1Q Attachment B Pago S of 21 III.
Fire Hazards Screenin Methodolo The intent of the fire hazards screening is to filter out areas which contain combustible loadings that are clearly incapable of threatening the existing fire barriers.
The screening is conducted by comparing the fire duration resulting from all in-situ combustibles to the as-tested fire barrier rating, with all calculations performed on a specific case-by-case basis. A safety factor of 2 is used in making the comparison to ensure that the resulting fire duration is actually less than one-half of the fire barrier rating'.
The specific screening steps are:
A)
The fire barriers (Thermo-Lag as a wall or Thermo-Lag as a raceway fire barrier) in the area are fire rated using applicable industry testing. In the case of Thermo-Lag used for raceway protection, this rating willbe assigned using testing as well as the NEI (NUMARC)Application Guide.
B)
The total in-situ combustible loading in Btu/ft's determined for each area
. This value includes all cable except cable which is totally enclosed in a steel raceway.
C)
The total combustible loading is divided by 80,000 Btu/ft2-hr. This yields an equivalent fire duration in hours for the area (Ref. VI.A).
'A safety factor of two is required to pass the screening process.
However, this process uses a simple acceptance criteria (fire barrier rating ) twice the in-situ combustibles necessary to challenge the fire barrier rating).
The factor of safety of two is used to account for unforeseen circumstances such as potential transient combustibles in the area, potentially elevated temperatures at a particular component due to unusual fire geometry, or the possibility of an in-situ combustible which is not captured in the Fire Hazards Analysis (FHA) report.
See Glossary under "fire load",
for the basis of the screening methodology and III.Ffor the guidelines for determining the applicable room geometry and combustible loading for this screening process.
I t
I
D)
The assigned fire barrier rating is divided by the calculated fire duration, to obtain a safety factor (for the fire barrier):
FireBarrierRating(Minutes) =SafetyFactor Room FireLoading(Minutes)
For fire barriers in fire areas with installed suppression and detection, the barrier rating is increased by a factor of 3 prior to calculating the safety factor':
FireBarrierRating(Minutes) x3 S
-SafetyFactor RoomFireLoading(Minutes)
E)
Ifa safety factor of 2 or greater is obtained, the fire barrier is considered acceptable for the hazard in that area; no further evaluation is required. Ifthe area does not pass the screening, the fire modeling technique described below in Section IVis applied.
F)
The applicability of the screening method depends on the type of combustible present in the rooms and the distribution of the combustibles within the room.
The types of combustibles are consistent with this methodology.
The most prevalent combustible is cable.
The, distribution of the combustibles is evaluated as follows: 1) ifthe overall combustible load is small i.e., = 20,000 Btu/ft', (a few cable trays running in the area) then the method is valid, 2) ifthe barrier(s) being analyzed are not near a concentration of combustibles in the area then the method is valid, 3) ifthe concentrated loading is less than twice the average loading (this is appropriate due to less than a 50% combustion efficiency for the most prevalent combustible cable) then the method is valid. The uniformity or the non-uniformity of the fire load is assessed during a fire area walkdown.
Figure 2 provides typical results from the fire hazard screening methodology.
Credit is given for providing suppression and detection for a particular fire barrier.
Temperatures necessary to challenge a fire barrier willalso activate fire suppression and detection systems.
These systems willprovide early warning of a fire and willmitigate the effects of the fire in the room.
L94-1Q ALLchmcnt B Pago7 of 21 Fi re'2 ical Fire Hazards Screenin Results
I
'L L-94-1Q Attachmcttt B Page 8 of2l Fi re 2 Cont'd ical Fire Hazards Screenin Results
L L94-1Q Ausctuncnt B Page 9 of 21 IV.
Fire Modelin Methodolo The intent of the fire model is to permit the detailed evaluation of locations with respect to potential fire hazards.
Based on in-situ combustibles, transient combustibles and flammable transients assumed to be present, the model determines a fire duration (burn time) assuming that the flammable transients willbe the initiator in the fire scenario.
The fire duration is compared to the fire barrier rating assigned from testing.
The duration must be no more than one-half the assigned fire barrier rating.
This model (Fig. 3) also provides a conservative estimate of the room air temperatures that could evolve at a target as a result of the exposure fire scenario.
The air temperature must be below the ignition temperature for the bulk in-situ combustibles in the room. The fire model input includes credible flammable transients, transient combustibles, in-situ combustible targets in actual locations, and room geometry.
Four separate target scenarios are evaluated:
Targets in the plume Targets in the ceiling jet outside the plume Targets in the hot gas layer outside the plume Radiant targets outside the plume More than one scenario can be evaluated for each room.
Combustible targets may be located in any of the locations listed above.
The model and this description are based on EPRI formulations and methodologies (Ref. VI.B and Ref. VI.C).
Fi ure 3
- Fire Model Conce tual Arran ement Jet Corretstton tBased on Testlntr)
HBSKEBTADCorretaeon tor Plume Temperature Gradient (Based on o CONSERVAT)SMS In4ltu Combustible ls Placed ln Plume Transient Burns at Full Intensity Until Fuel Exhaustion
" Influence of Oxygen Depletion ls Neglected Hest Release Rate tEmplrtcatly Based on Tests)
C
~
~
I 94-101 Auachmcnt B Page 10 of 21 IV.
Fire Modelin Methodolo cont'd An effective peak fire intensity is calculated when the target is located in the fire plume and a mean effective peak fire intensity is calculated when the target is located in the ceiling jet region.
The maximum plume and ceiling jet temperature at the location of the target is evaluated based on this intensity. Ifthe plume or ceiling jet temperature exceeds the ignition temperature of the in-situ combustible target placed in that location, then the in-situ combustible willburn and release heat.
For the purpose of modeling ignition temperatures, grouped electrical cables could be considered as semi-infinite solids.
The semi-infinite solid is representative of targets with non-negligible internal resistance to heat flow and sufficient thickness to prevent heat loss from the back face of the solid. Thermal gradients occur within the material, with the surface temperature rising more quickly than internal temperatures under imposed heat flux conditions due to the internal resistance to heat transfer.
However, for this conservative model, the transient thermal response of the target is neglected.
The target is assumed to instantly equilibrate to the environmental conditions.
Since this model provides a comprehensive tally of room temperatures, it could be used to activate suppression systems.
Although the information is available for additional analysis, it is not used to activate any suppression systems, even though they would aid in extinguishing the fire or cooling the effects of a Qre.
A radiant target may be evaluated; however, ifit is not ignited it is not considered to contribute to the fire loading in the room. The parameters provided for the radiant target are the heat flux at the target and the equivalent surrounding air temperature throughout the scenario. Ifit is determined that a radiant target willburn in a room then it must be placed in the plume or ceiling jet to properly emulate heat release.
Typical fire model results are shown in Fig. 4.
L 94-1Q Attacbmcnt B Page 11 of 21 Fi re 4 -
ical Fire Model Results
L94-10t Attachment B Page 12 of 21 VI.
Fire Model Detailed Descri tion A.
Desi n Basis Fire Since the in-situ and potential transient combustibles vary widely in different areas of the plant (Ref. VI.D), the establishment of a specific "design basis fire" for individual areas is a prerequisite to performance of a valid fire hazards analysis (Ref. VI.E).
Accordingly, each fire area is evaluated for all in-situ combustibles and potential transient combustibles and flammable transients.
This includes a review of plant procedures and an area walkdown to assess the transport and use of combustible materials in each fire area. A flammable transient (flash point < 100'F) such as paint, paint thinner, alcohol, etc, is determined for each fire area.
The model sequence initiates the ignition of the flammable transient, independent of the presence of an ignition source.
The location of the flammable transient is selected conservatively such'hat the impact on in-situ combustibles is maximized.
For Turkey Point the combustible controls program requires all flammable transient liquids to be attended at all times, and requires a permit for quantities in excess of an administrative limit. The design basis flammable transient is assumed to be equal to the administrative limit, and the type of flammable transient is determined on an area-by-area basis.
For St. Lucie the combustible controls program manages transient combustibles to an administrative limit, and requires safety cans (not exceeding one-half the administrative limitin capacity) for flammable transient liquids in buildings. The design basis flammable transient is assumed to be equal to the administrative limit, and the type of flammable transient is determined on ati area-by-area basis.
An inventory of the in-situ and transient combustibles is available.
This inventory includes a) BTU content, b) burn rate, c) quantity, d) ignition temperature, and e) location in the fire area (height, distance from walls, etc).
This conservative combination of transient and in-situ combustibles comprises the design basis fire.
B.
Identification of Fire Scenarios This model determines the temperature of the room in the locations indicated in Section IV (above).
These temperatures are tabulated for each time interval until the termination of the calculation. A representative ignition temperature is provided for each combustible target (e.g., IEEE-383 cable is 700'I').
The room is examined to determine the worst location for a floor based flammable transient fire with regard to in-situ combustibles.
Kx osure Fire Source Characterization I 94-1%
Auachncnt B Page 13 of 21 Peak Fire Intensi Combustible target exposure scenarios require evaluation of the exposure peak fire intensity. It is possible to conservatively estimate peak heat release rates for representative exposure fire fuels. 'A conservative assessment of fire source intensities is made by calculating the peak heat release rate as:
Qp~(
) =e(
)'~,(ft )
Btg Btu sec ft2 Where:
q -
the appropriate unit heat release rate associated with fully involved conditions A, -
the total exposed surface area of the fuel.
This method of assessing the peak fire intensities assumes the entire exposed surface area of the fuel spill willbe actively burning. It neglects the periods of fire development and fuel burnout that can diminish the actual fire intensity from this peak burning rate.
This method also uses unit heat release rates associated with fully-involved burning. With these conservative approximations, this approach can be applied directly for fuels with well deflned planar surfaces, such as liquid pools.
Transient Combustible A flammable liquid spill is postulated as the transient fuel exposure fire which initiates the fire scenario.
One of the tasks is to estimate the surface area of the liquid spill. This is a function of the quantity, viscosity and surface tension of the spilled liquid and the cohesion, contour and absorption rate of the floor finish.
On a smooth floor the area is simply equal to the volume of the liquid spilled divided by the thickness of the spill once it stops spreading.
Floors in power plants, however, are not smooth, level or without obstructions. A spill would migrate to the low points and reduce the surface area calculated for a smooth surface.
In addition, the postulated spill willbe an artificiallylarge quantity, which would appear to spread over a large smooth floor area.
This would cause the fire to burn intensely for a very short period of time. This would not occur in an actual spill; thus, the model spill area willbe kept small (i.e., = 8 ft'or 10 gallons) to better simulate actual spill and fire conditions while maximizing the burn time which is the acceptance criterion.
L94-IBt Attachncnt B Pago 14 of21 D.) ~Ti~ Pl Heat which is present in the plume, the ceiling jet or the hot gas layer is transmitted via convective and radiant heat transfer.
These methods are described below:
Convection Heat is transfered from the flame to the body (e.g., wall) via standard convective and radiant heat transfer.
For convective heat transfer the typical formula below is used:
q
=h (T-T)
//
Btu c f s I ft2 Where:
h, -
the convective surface heat transfer coefficient (See Section VI.F below), T, the flame temperature T; the surface temperature of the body.
Radiation According to the Stefan-Boltzmann equation, the total energy emitted by a body is proportional to T'See Ref. VI.F), where T is in absolute degrees.
The total emissive power is:
E=efJT4(W/m2)(Btu/hrft2)
Where:
Stefan.-Boltzmann constant (5.67 x 10'/m'K') (1.713 x 10'tu/ft2 hr 'R4) emissivity (dimensionless),
is a measure of the efficiency of the surface as a radiator The intensity of radiant energy (q") falling on a surface remote from the emitter can be found by using the appropriate configuration factor (P), generally assumed to be unity, which takes into account the geometrical relationship between the emitter and the receiver.
q"=pea T4(W/m )(Btu/hrft )
I I
r c'
Fire Location Factor I 94-1M ALachetcot B Page 15 of 21 Fires located along walls or in corners formed by walls can result in higher plume and ceiling jet temperatures than fires located in the center of compartments.
These higher temperatures result from reduced entrainment of ambient air, which is blocked by the solid boundaries adjacent to the flre. The concept of reflection has been applied to calculate the effects of fire location in plume and ceiling jet temperatures.
This concept concludes that a fire burning along a wall is equivalent to a fire with twice the heat release rate with respect to temperature rise at any location, while a fire in a corner is equivalent to a fire with quadrupled heat release rate.
Once the peak fire intensity is estimated, an effective fire intensity (Q,) can be calculated from the product of the peak intensity and the appropriate fire location factor (FLF).
FLF = 4, fires located in corners FLF = 2, fires against walls FLF = 1, fires located in the open Generally, a fire must be located within a few inches of a wall or corner for these location factors to apply.
Virtual Ori ins The fire plume correlations are founded on the theory developed for a weak point source of heat release.
Real fires are strongly buoyant and have area sources.
To account for the effects of these differences, methods to calculate virtual origins have been developed.
A virtual origin is an elevation offset above or below the actual fuel surface that yields conditions, based on the point source theory, equivalent to those produced by the actual fire.
Virtual Origin calculations are feasible only for pool type fires and other fires without substantial in-depth combustion.
For pool fires Heskestad (Ref. VI-G, page 1-112) suggests virtual origin elevation can be calculated as:
Z, - Virtual Origin (m) = -1.02D + 0.083 Q'"
Where:
D - Effective Fire Diameter (m)
Q - Total Heat Release Rate (kw)
L-94-1Q Attachment B Pago 16 of 21 Plume Tem erature Rise The plume temperature rise is calculated using:
2 5 T('F)=340 (z-zy'y)
Where:
z
- Target Height off the floor (ft)
This formula is found in Ref. VI.B. The differential temperature is recalculated for each time interval, and is added to the hot gas room temperature calculated for that time interval to obtain the resultant plume temperature at the specified elevation (z) above the floor.
E.)
Tar et in the Ceilin et The target is considered to be in the ceiling jet ifit is in the upper 20 percent of the room height.
Ceiling jet targets are evaluated at a mean distance from the fire centerline to the nearest wall. For a fire in a corner or against a wall this distance is one-third the width (shortest room dimension) of the room. For a fire in the open, this distance is one-sixth the width of the room.
Unconfined Ceilin Jet The formula (Ref. VI.B) for determining the unconfined differential ceiling jet temperature (hT) is as follows:
0.3xhT,ff g T pl,cell cj
(')'
Where:
~Tpl,ceil L
H W
differential temp in the plume at the ceiling length of the room (ft) height of the room (ft) width of the room (ft)
This formula is used when the ratio of the room length to the room width is less than 0.5 and when the ratio of the room height to the room width is greater than 2.5.
1 a
a
~
L-94-10$
Attachment B Pat,"e 17 nF21 Confined Ceilin Jet The formula (Ref. VI.B) for determining the confined differential ceiling jet temperature (4T) is as follows:
4TJ=0.37x4T,
<<x( xexp[-0.16x()x() ]
Hg L
W3 This formula is utilized when the ratio of the room length to the room width is greater than 0.5 and when the ratio of the room height to the room width is less than 2.5.
F.)
Tar et in the Hot Gas La er The hot gas layer is considered to be the bulk room temperature outside the plume and the ceiling jet regions. A transient heat balance is performed on the room as follows to obtain the transient bulk room temperature:
Heat transfer from the room is evaluated as follows:
A one-dimensional finite difference method is employed for heat transfer through the walls, floors, ceilings, (Including Thermo-Lag walls which have activated the char layer due to the fire).
Conductive heat transfer is utilized for fire doors, hatches and Thermo-Lag walls which have not activated from the Qre.
Normal ventilation systems are in operation and willbe deenergized for associated circuits in the fire room or fire dampers in the ductwork.
Heat transferred to structural steel in the room (cabinets, cable tray supports, conduits, etc.) is evaluated.
Heat transferred into the fire room is evaluated as follows:
The heat from the flammable transient released to the room at the peak burn rate until the transient is consumed.
As a conservatism, oxygen depletion is not considered.
The in-situ combustibles which are exposed to a room temperature in excess of their ignition temperature in the plume, ceiling jet or hot gas layer are burned at their peak burn rate until the combustible is consumed.
Again, oxygen depletion is not considered.
Heat from the energized equipment in the room is released to the room.
L 94-lent Attaehmettt B Page 18 of 21 Radiant Tar et Thermal radiation can be the significant mode of heat transfer for situations where a target is located laterally from the exposure fire source.
This would be the case for a floor-based fire located adjacent to an electrical cabinet, a vertical cable tray or a conduit containing safe shutdown equipment and/or cables.
Evaluation of the thermal radiation requires the evaluation of the incident heat fluxes at a target. A point source estimate of radiant flux is used in this calculation.
The incident radiant heat flux at a target located at a radius (r) from the point source is then given as:
QR XR Qf/
4<<'here:
XR - (0.4 - Radiant fraction)
Q, - Peak Heat Release Rate (kw)
For conservatism, 0.9 willbe used for the absorptivity and emittance of the target.
This yields a net 81% of the incident radiation which willbe retained by the target.
This is incorporated into the total incident radiant heat flux formula. The radiant heat flux at various distances from the edge of the plume and at the designated radiant target are tabulated for the user of the program.
Critical fluxes for some targets are available in Ref. VI.B and VI.C.
H.)
Validation This fire modeling technique uses empirical formulas which were derived from testing.
These formulas were derived from pool fire tests on hydrocarbon fires.
Thus, the formulas are based on actual fires and actual fire tests.
The fire model computer program has been verified to ensure that it follows the empirical formulas from which it is comprised.
The conclusion is that the fire model is valid to tested Geld configurations for hydrocarbon pool Gres.
VII.
References A.
B.
C.
D.
E F.
G.
NFPA Handbook, Fifteenth Edition.
EPRI Report TR-100370, "Fire Induced Vulnerability Evaluation Methodology",
Revision 1, dated April 1992.
EPRI Report TR-100443, "Methods of Quantitative Fire Hazards Analysis", May 1992.
Generic Letter 86-10, Section 3.8.2 Appendix R,Section II.B(1) and BTP CMEB 9.5-1, Sections C.b(1) and (2).
Fire Dynamics, D. Drysdale, John Wiley and Sons, 1987.
SFPE Handbook of Fire Protection Engineering, First Edition, dated September, 1988.
L-94-N Attachment B Page 19 of 21 VIII.
1.
2.
3.
4 Fire Barrier Adequacy Process Flowchart Typical Fire Hazards Screening Results Fire Model Conceptual Arrangement Typical Fire Model Results IX.
~Gloss a Ceiling Jet:
A thin layer of hot gases from the fire plume which are located in the upper 20%
of the room height.
Fire Location Factor:
The fire location factor accounts for the reduced entrainment rate associated with fires located adjacent to walls and in corners and the consequent increase in the plume temperature.
The fire location factors are:
4 for fires located in corners 2
for fires located against walls 1
for fires located in the open Fire Load:
Ref. VI.ASection 5/Chapter 9.B, "Fire Load", states the following:
Analysis of tests by the National Bureau of Standards developed an approximate relationship between fire loading and an exposure to a fire severity equivalent to the standard time-temperature curve.
The weight per square foot of ordinary combustibles (wood, paper and similar materials with a heat of combustion of 7,000-8,000 Btu per ib) was related to hourly fire severity as described in Table 5-9B. (Table 5-9B relates multiples of 80,000 Btu per square foot to multiples of an equivalent fire severity of one hour.
The fire load-fire severity relationship was the first method developed to predict the severity of a fire that would be anticipated in various occupancies.
It could be used to determine resistance required of fire barriers as well as structural components.
Although the technique has its limitations, it does provide a conservative estimate of the maximum expected fire in most residential, business, institutional and many commercial occupancies.
It is simple and easy to use, although it is often not considered in some code requirements.
A principal caution that must be exercised is in application of this technique to
o l
L94-IQ Attachment B Page 20 of 21 combustibles having a very high heat release rate.
Such fires may expose the barriers to higher temperatures than those at which they are tested.
Fire load is a measure of the maximum heat that would be released ifall the combustibles in a given fire area burned.
Maximum heat release is the sum of the product of the weight of each combustible times its heat of combustion.
In a normal building, the fire load includes combustible contents, interior finish, floor finish, and structural elements.
Fire load is commonly expressed in terms of the average Gre load, which is the equivalent combustible weight divided by the fire area in square feet.
Equivalent combustible weight is defined as the weight of ordinary combustibles, having a heat of combustion of 8,000 Btu per lb, that would release the same total heat as the combustibles in the space could. For example, the equivalent weight of 10 lbs per sq ft of a plastic with a heat of combustion af 12,000 Btu per lb would be:
10 lbs per stlItx12,000 Btu per lb=120,000 Btu per sqft 120,000 Btu per sing 8,000 Btu lb ordinary combustibles On the conservative side, using the gross fire load for the combustibles does not account for the combustion efficiency of the combustible.
For cable the combustion efficiency is 45/o. Additionally, any oxygen depletion is not accounted for in this screening.
The applicability of this method is discussed in III.F.
Fire Room:
A space bounded by non-combustible barriers where heat and products of combustion from a fire within the enclosure willbe substantially confined.
Flammable Transient:
A transient combustible with a flash point < 100'F (paint, paint thinner, alcohol, etc).
Hot Gas Layer:
The hot gas layer is assumed to extend to the floor for this technique, thus the bulk room temperature is the same as the hot gas layer.
I 94-104 Attachment B Page 21 of 21 In-Situ Combus-tible:
Combustible material which is. permanently located in an area and material which is located in an area for long periods of time.
Plume:
The location directly above a fire source which contains the products of combustion from the Qre and associated high temperatures (flame).
Transient Combus-tible:
Combustible material passing through a place with only a brief stay usually associated (but not limited to) maintenance or modifications involving combustible and flammable liquids, wood and plastic products, waste, scrap, rags, or other combustibles resulting from the work activity.
Virtual Origin:
This is the source point of the fire when using a point source formulation to model a fire plume.
a I
a
~
\\
L94-1%
Page1 of7 St. Lucie Units 1 8g 2 Turkey Point Units 3 &: 4 Docket No's:
50-335, 50-389, 50-250, 50-251 Attachment C PERFORMANCE-BASEB APPROACH TO Fl]RE BAlRlRlKRADEQUACY:
PROBABILISTICSAFETY ASSESSMENT DESCIOPI'ION AI'6)METHODOLOGY FLORIDAPOWER &: LIGHT
l
-e
L94-10t Attachncnt C Pago 2 of 7 Table of Contents Section Title
~Pa e
Introduction The FPL PSA Fire Model Methodolo
~
~ t 3 A
Overview B.
Fire I nition Fre uencies C.
Alternate Success Paths 4
D.
Additional Fire Com artment S ecific Modelin.............
4 E.
Com arison of the FPL Model to the EPRI "FIVE"Methodolo 5
The Modelin of Thermo-La Barriers...........
IV.
Review of PSA Model Results
. 7 V.
References...
~
~
~ 7
..7
L94-N Attachment C Pat;e 3 nf7 I.
Introduction Probabilistic Safety Assessment (PSA) technology willbe used to estimate the level of safety, or fire hazards risk, for fire compartments that contain cables with Thermo-Lag application. The PSA analysis willprovide an additional and diverse review of the plant's capability to cope with postulated fires. The PSA review or screen will be applied to those plant fire compartments that have successfully passed the deterministic or traditional fire barrier evaluation in the FPL performance-based approach to fire barrier adequacy.
II.
The FPL PSA, Fire Model Methodolo A)
Overview The PSA fire model utilizes the EPRI "FIVE" (Fire Induced Vulnerability Evaluation) methodology to screen or estimate the core damage frequency (CDF) for each fire compartment containing cables with Thermo-Lag.
The PSA methodology used in this study to calculate a CDF for each fire compartment can be summarized as follows:
5 CDF
=
II I',
2=1 P, =
The fire ignition frequency in the fire compartment being examined.
P3 =
The failure probability of alternate success paths to put the plant in a safe shutdown condition.
P, =
Probability of critical combustible loading causing damage to the target components and cables.
P4 =
Ps =
Failure probability of fire suppression systems.
Failure probability of fire-related recovery actions.
B)
Fire Ignition Frequencies Fire ignition frequencies (P,) for each fire compartment are developed using fire database information contained in the EPRI "FIVE"Report (Ref. V.A).
This information is applied to the FPL site specific configurations by calculating fire ignition frequencies based on actual ignition sources present in the fire compartments.
i, c
L94-10$
AtLtchmcnt C Page 4 of7 C)
Alternate Success Paths The PSA fire model is used to calculate the conditional CDF (P,) or the failure probability of alternate success paths to put the plant in a safe shutdown condition.
The initiating fire events are propagated through the appropriate PSA event and fault tree models in the PSA computer model by failingunprotected cables and equipment in the postulated fire compartment.
Figure 1 provides an example for a postulated fire in the safety related auxiliary feedwater area at the Turkey Point Plant.
The PSA model would estimate the failure probability to maintain th'e plant in a safe condition using "defense in depth" or redundant success paths such as the standby feedwater pumps to remove decay heat.
FEED ANDBLEED SAFETY RELATED APJLCTOR AFWPUMPS MAtNFWPUMPS TlStSNS STHUCAÃtS STANDBYFEED WATERPUMPS Figure 1 Example of Alternate Success Paths lfthe resulting CDF, calculated as (P, x P,), is less than 1.0 x 10~ per reactor year, the fire compartment is screened out from f'urther analysis. Ascreening value of 1.0 x 10 per reactor year willbe used for fire compartments with automatic detection and suppression.
For those fire compartments that did not pass this initial quantitative screening, a more detailed analysis is performed which includes the PP4 and P, terms given above.
The quantitative screening, at this step is very conservative.
This is because the PSA model at this point assumes:
- AllPSA equipment in the fire compartment fails.
- Allunprotected cables in the fire compartment fail.
- No credit is given for detection, suppression systems or fire brigades.
D)
Additional Fire Compartment Specific Modeling Additional detailed modeling is performed for the fire compartments that do not pass the initial quantitative screening (P, x P,).
The additional analysis may include the probability of critical combustible loading damage to target components and cables(P,), the failure probability offire suppression systems (P4), the failure probability of manual suppression and additional recovery
t
~
q 9
~
L94-10t Attachmcttt C PageS of7 actions in the PSA model that are in plant procedures (P,).
E)
Comparison of the FPL Model to the EPRI "FIVE"Methodology.
The FPL PSA assessment of fire barriers uses the EPRI "FIVE"methodology as described in Ref. V.A with differences as identified below.
A cross reference in the terms used above and in FIVE are:
P, =
The fire ignition frequency in the fire compartment being examined is the same as F, in FIVE (p. 6-9 in Ref.
V.A).
The failure probability of alternate success paths to put the plant in a safe shutdown condition is the same as P, in FIVE (p. 6-12 in Ref. V.A).
P, x P4 x P, =
The probability of critical combustible loading damage is the same as P, in FIVE (p. 6-36 in Ref. V.A).
The calculation of the probability of critical combustible loading damage will either use the FIVE tables for fire-target analysis or the FPL plant specific fire hazards model.
The FPL PSA analyses willestimate the failure probability ofprotected cables in the PSA fault trees(P,).
See Section IIIfor additional detail.
The scope of the FPL analysis includes only those fire compartments with Thermo-Lag.
The recovery actions in the P, term refers, to manual fire suppression and additional fire related proceduralized actions.
III.
The Modeling of Thermo-Lag'Barriers A key feature of this analysis is the modeling of protected cables or Thermo-Lag in the PSA model.
For the initial quantitative screening described in Section II.C, failure of all PSA components(i.e.,
- valves, pumps, etc.)
located in the fire compartment is assumed.
In addition, PSA components in areas outside the fire compartment are also failed if they have unprotected cables in the postulated fire compartment.
The Thermo-Lag protected cables, or the associated equipment in the PSA model, are assigned a failure probability based on the Thermo-Lag fire barrier rating determined by industry testing and the anticipated fire duration in the fire compartment.
Finally, the probability of Thermo-Lag failure is assumed to be negligible for fire compartments that pass the initialfire hazards analysis screen since there is high confidence that cables willbe protected due to the low amount of in-situ combustibles in these fire compartments.
I
I 94-10t Atttchmcnt C Pago6of7 Figure 2 is a simplified representation of the modeling of components in the postulated fire compartment for the initial PSA quantitative screen(P, x P,).
firComaettment 1
IJn Cable Protected Cable F A Fire Compttttment 2 Failure Syatom Fauk Tree VaNeA Valve B Cable Valve Failure Power Supply Failure Figure 2 PSA Modeling of Equipment and Cables The probability of failure of the Thermo-Lag fire barriers for cables is primarily a function of its recent industry tested fire endurance rating and of the duration of the specific fire hazard to which it is exposed.
A probability distribution function is estimated based on industry testing of the Thermo-Lag fire endurance rating.
Similarly, a probability distribution f'unction is estimated for fire durations based on the FPL Qre hazards analysis, plant procedures, and discussions with knowledgeable plant personnel concerning expected flammable transients in the plant. A stress-strength interference model is used to estimate the failure probability of the Thermo-Lag based on the probability distribution of the Thermo-Lag failure time derived from industry testing (i.e., "strength") and that of the estimated fire duration (i.e.,
"stress" ). The estimated Thermo-Lag failure probability for the worst case as-tested conduit size(3/4 inch and one inch) willbe used for all currently classified one-hour and three-hour protected cable barriers in the fire compartment for the initial screening effort. The distribution curves willbe submitted with the final reports.
Ifa fire compartment does not screen out due to the initialconservative Thermo-Lag failure probability, a more realistic assessment of the probability of failure of the Thermo-Lag may be performed. This additional analysis corresponds to the P,, P4 and P, above. The detailed assessment may consider factors specific to each fire compartment such as the types and quantities of the combustible and the location of the electrical conduits protected by Thermo-Lag.
The conservative screening assumption of using test data for the worst case conduit size for all conduits in a fire compartment is reexamined at this point.
The fire duration used in the stress-strength interference model is generally an upper bound compared to that which a fire barrier willexperience due to the fact that the combustibles and the plume are not always located directly under the Thermo-Lag in an actual fire. Ifthe Thermo-Lag protected cable is not in the fire plume in outdoor fire compartments, the Thermo-Lag failure probability is assumed to be negligible due to the absence of a ceiling affect to trap hot gases.
I 94-104 Attachment C Pat;c7 of7 IV.
Review of PS@ Model Results The PSA analysis estimates a level of plant safety or risk for each fire compartment through the PSA computer model of the probability that a fire willstart combined with the ability of the plant equipment and plant operators to cope with the fire.
The fire compartment willbe screened out from any further evaluation during the PSA analysis if the CDF for a fire compartment is determined to be less than 1.0 x 10'er reactor year. A level of safety of 1.0 x 10'er reactor year for a fire compartment has been accepted in the industry consistent with the FIVE methodology. A fire compartment willalso be screened from further evaluation at 1.0 x 10~ per reactor year ifP, x P, alone is less than 1.0x 10 per reactor year and the Qre compartment has automatic detection and suppression.
The results of the analysis willprovide a relative comparison of the level of safety for the various fire compartments that contain cables with the Thermo-Lag application.
Procedural or physical modifications for fire compartments would be recommended if PSA sensitivity analyses demonstrate that Thermo-Lag protected cables with reduced fire endurance ratings are the cause of a realistic fire vulnerability.
V.
References A.
EPRI Report TR-100370, "Fire Induced Vulnerability Evaluation Methodology", Revision 1, dated April 1992.
VI.
List of Figures 1.
Examples of Alternate Success Paths.
2.
PSA Modeling of Equipment and Cables.
I
~
o cs Ir94-IQ Pago I of S Turkey Point Units 3 & 4 Docket Nos: 50-250, 50-251 Attachment D Y POINT OUjH)OOR FmRE BARRIERS:
DESCRIPTION AI%3 METHODOLOGY FLORIDAPOWER & LIGHT
~
'C
L 94-1%
Page 2 of S Table of Contents Section Title
~Pa e
I.
Introduction
~
~
~
\\
~
3 II.
Turbine Lube Oil S stem...........................................
3 III.
Combustible Control Pro ams, '.....,,...........,,....,......,,,...
4 IV.
30-Minute Outdoor Fire Barriers
p1 I 94-1(A Pago 2 of S Table of Contents Section Title
~Pa e
I.
Introduction II.
Turbine Lube Oil S stem.....
~
~
~
~
0 0
~
~
0 3
III.
Combustible Control Pro ams 4
IV.
30-Minute Outdoor Fire Barriers
~
~
~
~
4
I,
L94-IQ Attactuncnt D Pnt;o3 of S TURKEYPOINT OUTDOOR FIRE BARRIERS I.
Introduction Florida Power & Light (FPL) has used the Thermo-Lag product for outdoor fire barriers at the Turkey Point Nuclear facility. These fire barrier configurations are unique to the Turkey Point site and, to our knowledge, there has been little or no outdoor use by other utilities within the industry. FPL believes that equivalent protection to subsections III.G.2.a and c to Appendix R can be provided for outdoor areas using Thermo-Lag as a fire barrier material where the following conditions are met:
(1)
As-installed fire barriers have a minimum rating of30 minutes, based on test results and the NEI (NUMARC)Application Guide; and (2)
The fire barriers are not within 50 feet of a major in-situ combustible.
Outdoor areas present special characteristics that mitigate the effects of fires.
Fire damage from stratified hot gases or ceiling jet layers is not a concern in outdoor areas or other large enclosures such as the reactor containment.
The fire energy is not localized by physical boundaries and dissipates quickly with the large outdoor heat sink. Major in-situ combustibles in outdoor areas such as transformers, oil reservoirs, etc, are contained and have suppression systems.
A 50-foot separation provides additional assurance that a fire from one of these sources would not challenge a 30-minute fire barrier.
II.
Turbine Lube Oil S stem The major combustibles in the Turbine Lube Oil (TLO) system are the TLO reserviors, filters, and the hydrogen seal oil units. No 30-minute Thermo-Lag fire barriers willbe allowed within 50 feet of these major sources.
The balance of the TLO system consists of a turbine control oil system, turbine auto stop oil system, and associated piping.
Both are supplied by the shaft driven main turbine lube oil pump. This pump operates at about 380 psig. Oil is supplied to the shaft bearings at about 15 psig. The TLO system has annunciators that willalert operators to a high pressure lube oilpipe break, or other pressure loss related problems.
These annunciators include:
Turbine Bearing lube oil press. (E 2/1)
Emergency Bearing oil pump running (E 2/3)
Turbine Aux oil pump running (E 4/1)
Guarded oil actuation (E 5/6)
Turbine bearing oil low pressure trip (E 6/2)
Set Point 8
psig 7
pslg 10 psig 2
pslg 5.5 psig
~
~ Q t
~
o
P L-94-1(A AMachmcnt D E
Pcgc4 of 5
~l
'here are no valves (potentially leak locations) in the TLO piping under the turbine deck. The high pressure supply piping is encased in a low pressure (atmospheric) guard pipe which drains to the TLO sump.
In the event of a leak in a high pressure TLO line, the leakage will be diverted to the TLO sump via the guard piping. A leak in the high pressure portion of the TLO system will cause turbine trip and/or turbine valve closures and/or control room annunciation.
Only a low pressure (atmospheric) TLO leak could occur in the turbine building. This would not be a large release of oil, and would be discovered by operators before any significant accumulation would occur.
Awet pipe suppression system is provided for the TLO piping in the turbine building under the turbine deck.
A leak in the low pressure lube oil piping could result in:
impingement on a hot steam pipe or other ignition source which would ignite the oil.
This would be under the suppression system and would be quickly detected by operators.
no impingement on any ignition sources and end up in a storm drain or sump system.
This would not likely be of any significant quantity and would be discovered by plant personnel.
This low leakage TLO piping design will assure that the fire hazard is low in the TLO piping areas.
III.
Combustible Control Pro ams The Turkey Point Combustible Control Program does not allow storage of combustibles in outdoor areas that contain safe shutdown equipment or cables.
Procedures require that liquid flammable combustibles must be attended at all times and a special permit is required for quantities greater than 5 gallons. Hence, transient combustible controls assure that a worst case transient fire caused by a spill would be far below a hazard level that could challenge a 30-minute fire barrier. As an example, a fire from a 20 gallon flammable transient spill (over =20 ft'rea) would generate a fire with a duration less than 15 minutes. Initiation of a fire of this size would require multiple breakdowns of the plant transient combustible control processes.
However, were these breakdowns to occur, there is still at least a factor of safety of two when compared to a 30-minute fire barrier located directly in the fire plume. Additionally, the actual fire brigade response time for outdoor fires is well below a 30 minute fire barrier fire endurance rating.
L 94-1Q Auachmcnt D Page S of S IV.
30-Minute Outdoor Fire Barriers Subsection III.G.2.fof Appendix R allows for the protection of safe shutdown equipment and cables in containment using radiant energy shields. This protective feature is appropriate given the fact that the containment is a large open area and extreme temperature build-up around a postulated fire is not possible.
For outdoor open areas, this approach provides equivalent protection for safe shutdown equipment and cables as compared to that allowed by subsection III.G.2.fofAppendix R for the reactor containment. Itis also noted that the detection and fire brigade response to a fire in an outdoor area is better than for inside reactor containment.
Additionally, 30-minute Thermo-Lag fire barriers can perform the functions of radiant energy shields. The NRC has provided concurrence with this use offire barriers in Generic Letter 86-10.
Given the identified fire hazards, this method for outdoor areas provides equivalent protection to that of subsections III.G.2.a and c of Appendix R, and is comparable to the provisions of subsection III.G.2.fof Appendix R for protecting safe shutdown equipment and cables.
This method willbe used initiallyto evaluate the Turkey Point outdoor open fire areas/zones described in the Final Safety Analysis Report.
rE Qt '
el