ML20013H520
| ML20013H520 | |
| Person / Time | |
|---|---|
| Site: | FitzPatrick  |
| Issue date: | 04/08/1998 |
| From: | Cunningham M NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Bajwa S NRC (Affiliation Not Assigned) |
| References | |
| NUDOCS 9804230294 | |
| Download: ML20013H520 (9) | |
Text
_. _
_._ _ _ i April 8, 1998 l
' MEMORANDUM TO:
Singh Bajwa, Project Director Project Directorate I-1 Division of Reactor Projects 1/II i
Office of Nuc! ear Reactor Regulation
\\
FROM:
Mark Gmningham, Chief i
Probabilistic Risk Analysis Branch l
Division of Systems Technology Office of Nuclear Regulatory Research
SUBJECT:
RdQUEST FOR ADDITIONAL INFOR. CATION ON FITZPATRICK l
IPEEE SUBMITTAL.
Based on our engoir g review of the FitzPatrick Individuai Plant Examination of External Events I
(IPEEE) submittal, we have rieveloped the atisched atquests for additional infonnation (RAI).
l The RAls are related to the analyses of seismic; fire; and high winds, flod, and other extemal events (HFO) in the IPEEE. The RAls in the seismic area were developed by our contractor, i
Brookhaven National Laboratory; the RAls in the fire area were developed by our contractor, Sandia National Laborato iss; and the RAls for HFO were developed by RES staff. All of the RAls were reviewed by the " Senior Review Board'~ (SRB). The SRB is comprised of RES and l
NRR staff and RES consultants (Sandia National Laboratories) with probabilistic.isk assessment expertise in extemal events.
i We request that the licensee provic'e its response within 60 days in conformance with our j
review schedule. When the RAls are sent to the licensee, please put Alan Rubin (RES/ DST /PRAB, Mail Stop T-10E50) on distribution. If you have any questions conceming our review, please contact Brad Hardin at 415-6561.
Attachment:
FitzPatrick RAls i
Distribution:
LMarsh SWest EConnell l
GBagchi DJeng RRothman AMurphy NChokshi TChang ABastik JWilliams RKornasiewicz i
SNowien, SNL MBohn, SNL Central File I
i PRAB SublFile Q
DOCUMENT NAME:G:\\FITZRAI.M27 i
'SEE PREVIOUS CONCURRENCE l
Ts seenive e cow of this stocumere inshcote in the bar: *C a Copy without attachment / enclosure
- E" = Copy with attachmentlenclosure
- N" - No copy 0FFICE DST /PRAB l
DST /PRAB l
DSJ/&ffV l
l l
NAME
- BHardin:bgj
- ARubin tid 6hdingham -
DATE 3/30/98 3/30/98 b /q /98 0FFICIAL RECORD COPY 900423o294 9so4og ks-2!Eh (RES File Code) RES 2C-5
)
m 4(
s speg
[
E iga u_
j
MEMORANDUM TO:
Singh Bajwa, Project Director Project Direriorate 1-1 Division of Reactor Projects I/ll
[
Office of Nuclear Reactor Regulation FROM:
Mark Cunningham, Chief Probabilistic Risk Analysis Branch Division of Systems Technology Office of Nuclear Regulatory Research
SUBJECT:
REQUEST FOR ADDITIONAL INFORMATION ON FITZPATRICK iPEEE SUBMITTAL Based on our ongoing teview of the FitzFatrick Individaal Plant Examination of Extema! Events (iPEEE) submittal, we have developed the attacheG request for additionalinformation (RAl). The RAls are related to the analyses of seismic; fire; and high winds, flood, and other extemal events (HFO) in the IPEEE. The RAls in the seismic area were developed by our contractor, Brookhaven National Laboratory; the RAls in the fire area were developed by our contractor, Sandia National Laboratories; and the RAls for HFO were developed by RES staff. All of the RAls were reviewed by the " Senior Review Board"(SRB). The SRB is :omprised of RES and NRR staff and RES consultants (Sanoia.*1ational Laboratories) with probabilistic risk assessment expertise in extemal events.
i We request that the i!censee provide its response within 60 days in conformance with our retiew schedule. When the RAls are sent to the licensee, please put Alan Rubin (RES/ DST /PRAB, Mail Stop T-10E50) on distribution. If you have any questions conceming our review, please contact Brad Hardin at 415-6561.
1 Anachment: FitzPatrick RAls i
Distributiorn LMarsh SWest EConnell GBagchi DJeng RRothman AMurphy NChokshi TChang ABuslik JWilliams RKomasiewicz SNowlen, SNL MBohn, SNL Central File PRAB Sub/ File DOCUMENT NAME:G:\\FI RAl.M27 T, receive a copy of this documeOn - o in tne t.om: *C" = Copy without attachment / enclosure
- E* = Copy with anachmentient Insure
- N" a No copy 0FFICE DST /PRAB V' l6.
DST /PRAB lt DST /PRAB l
l NAME BHardin:bgj ARubin du MCunningham DATE
$ /SO /98 3 / 3a /98
/ /98 l
OFFICIAL RECORD COPY (RES File Code) RES
_2f-5 i
l
1 n
FitzPatrick 1
j Request for Additional Information on IPEEE Submittal f
i l
A.
Seismic f
i i
i 1.
It is stated in the submittal that: "the existing unreinforced concrete block walls were j
j previously evaluated in the IE Bulletin 80-11 program, in which response spectrum analysis was performed using the 1 percent DBE floor spectra and the allowable tensile strengt'. of 35.4 psi. In the iPEEE analysis, the High Confidence Low Probability of
]
Failure (HCLPF) capacities of the block walls were estimatee oy extrapolating the j
fore Jcing linear analysir results by rsiising the damping value to 7 percent arid the l
a!!owable stmss to 42.' psi (1.7 times the design alle,wabie stiess of 25 psi)." in Sec4on j
i 8.3 ri the submittsd, it is stated that the block walls in the emergency diest,i generator l
(EDG) building (EGB-272-6,7,9 and 10) will be -trngthened. The following specific questions relate to the capacity evaluation and E,engthening of the unreinforced concrete block walls:
s.
Provide the technl.;ai basis, such as past test results or detailed
}
analytical evaluation, for the assumed damping value of 7 percent 1
and the allowable stret 2 of 42.5 psi, and for the extrap >Istion from
}
the 80-11linsar analysia results *o chtsin HCLPF estimai s.
\\
b.
Please provide information on the current status of the effort to j
strengthen the block walls in the EDG building.
t' 2.
Regarding the relay rhatter evaluation on pp. 3-31 of the submittal, the submittal states i
that " relays that are either bad-actor relays or are not covered by the generic equipment response spectra "3ERS) were assumed to chatter during an earthquake" (i.e., a failure 4
}
probability of 1.C 6 implied). On the other hand, on pp. 3-81 of the submittal, a failure probability of 0.1 is implied for screening criteria.
Please explain this inconsistency. Aiso, please provide the schedule for resolving the issue of bad-actor relays in the EDG building.
t 3
3.
Regarding the potential rupture of a hydrogen line during an earthquake event, the submittal states that the operating procedure has been changed to close the hydrogen supply line at supply points in the event of an earthquake.
Considering the swift onset of an earthquake and the potentially rapid progress of a hazardous hydrogen explosion, please describe the timing of operatoractions needed to close the hydrogen supplylines andJustify that this simale procedural change is sutNcient to prevent a potential Hre (i.e.,
i explain why this procedural change alone is sufHclent and and how it will i
be implemented in the hazardous environment likely to exist following a seismic event).
I i
w
.-.5 m
._.m
. -. ~ _- - __
i i
B-Fire i
1.
The heat loss factor is defined as the fraction of energy released ' ' a fire that is transferred to the enclosure boundaries. This is a key parameter in the prediction of 3
i component damage, as it determines the amount of heat available to the hot gas layer.
j in Fire-Induced Vulnerability Evaluation (FIVE), the heat loss factor is modeled as being inversely related to the amount of heat required to cause a given temperature rise.
Thus, 'or example, a larger heat loss factor means that a larger amount of heat (due to a more severe fire, a long er buming time, or both; e needed to cause a given temperature rise. It can 'e seen that if the value assumed for the heat loss factor is unrealistically high, fire scenarios can be improperly screened out. Figure R.1 provides l
a representative example of how hot gas layer temperature predictions can changu l
assuming different heat loss factors. Note that: 1) the curves are computed for a 1000 l
kW fire in a 10m x Sm x 4m compartment with a forced ventilation rate of 1130 cfm: 2) the FIVE-recommended damage temperature for qualified cable is 700*F for qualified l
cable and 450*F for unqualified cable; and,3) n e SFPE curve in the figure is generated j
frm a correlation r ovided in the Society for P.e Protection Engineers Handbook [R1].
j i
Basw
' avidence provided by a 1982 paper by Cooper et al. [R2), the EPRI Fire PRA i
implemt,. nation Guide recommends a heat loss factor of 0.94 for fires with durations greater than five minutes and 0.85 for " exposure fires away froa a wall and quickly developing hot gas layers." However, as a general statement, this appears to be a misinterpretation of the results. Reference [R2], which documents the results of multi-
- compartment fire experiments, states that the higher heat loss factors are associated with the movement of the hot gas layer from the buming ccmpartment to adjacent, cooler compartments.- Earlier in the experiments, where the hot gas layer is limited to the buming compartment, P.eference [R2] reports much lower heat loss factors (on the order of 0.51 to 0.74). These lower heat loss factors are more appropriate when analyzing a single compartment fire.
In summary, (a) hot gas layer predictions are very sensitive to the assumed value of tha heat foss factor; and (b) large heat loss factors cannot be justified for single-room scenarios based on these references.
The FittPatrick IPEEE Mre study does not describe the heat loss factor value(s) assumed in assessing damage. What value(s) for the heat loss factor were usedin the FitrPatrick analysis? For each scenario where the hot pas layer temperature was calculated, please specify the heat loss factor value urad in the analysis. In light of the precedng dscussion, please either: a) Justify the value used and dscuss its effect on the identiMcation of Mrs vulnerabilities, or b) repeat the analysis using a anoreJustlRable value and provide the resulting change in scenario contribution to core damage frequency.
2
O Time-Temperature Curves i
900
/
800 SFE j
H.F = 0.70 700.
l + H.F = 0.85 I l
.-~E 600
+ FtF = 0.94
--x-FLF = 0.99 l
g 400 i l
5
/
r i
300 l
200 I
x A x x x x.x x x x.x x.x x x.x x.x u x x.x x.x x.x l
100 R?88,8R E8o o 88R?
88 i.
8e
-a n
e s a n
- m Time (s) l l
Figure R.1 Sensitivity of the hot gas layer temperature predictions to the assumed heat loss factor 2.
The Fitzpatrick submittal assumes propagation time delays of 5 minutes for tray-to-tray fire spread and 15 minutes for fire spread from an electrical panel. The basis upon I
which these propagation time delays have been assumed has not been provided. In particular, these assumptions have been applied in fire zones RR-1, CS-1, CT-3/RR-1, RR-1/AD-4, RR-1/TB-1, and RR-1/RB-1 A. These assumptions pre-set the fire growth i
and propagation time, which in turn impacts the time-dependent rate of heat release and the resulting critical damage times, without appropriate consideration of the specifics of a given fire scenario. In particular, the propagation behavior of a fire will be influenced by a number of case-specific factors including the intensity and duration of the initial source (or exposure) fire, the proximity of the secondary fuels to the initial source (or exposure) fire, the fire behavior properties of the cables er other secondary fuels (e.g.,
ignition temperature, heat of combustion, total fuel mass), the physical configuration of the exposed trays (e.g., vertical versus horizontal, exposed sMace area, and the density of the cable loading), proximity of the fire source and secondary fuels to walls and/or the ceiling, and the room-specific ventilation conditions.
For each fire scenario in which fixedpropagation delays were used to estimate the rate and extent of fire propagation, please: (a) indicate if FIVE (or similar) calculations were performed for the scenario and provide the results (equipment 3
i i
or cables damaged) of these calculations; and(b) If experimental data were used, please indicate, which experimental results were used, how they were utilizedin the analysis,
=
Justification of the use of these experimental results to the scenario being analyzed.
i The discussion of results applicability dould compare the geometries, ignition
\\
sources, fuel type and loadings, ventilation characteristics, and compartment characteristics of the experimentalsetup(s) with those of the scenario ofinterest.
4 3.
Fires in the main control room (MCR) are potentially risk-significant because they can catise I&C failures (e.g., loss of signals or spurious signals) for multiple redundant divbions, and because they can force control room abandonment. Although data from two experiments conceming the timing of smoke-induced, forced control room abandonment are available [R3), the data must be carefully interpreted, and the analysis must properly consider the differences in configuration between the experiments and the i
actual control room being evaluated for fire risk. In particular, the experimental configuration included placement of smoke detectors inside the cabinet in which the fire originated, as well as an open cabinet door ivr that cabinet. In one case, failure to account for these configuration differences led to mve than an order of magnitude underestimate in the conditional probability of forced control room fire abandonment
[R4). In addition, another study raises questions about control room habitability due to room air temperature concems [RS).
Please provide the detailed assumptions (including the assumed fire frequency, any frequency reduction factors, and the probability of abandonment) usedin analyzing the MCR andJustifications for these asuumptions. In particular, if the probability of abandonment is based on a probability distribution for the time required to suppress the fire, please Justify the parametric form of the distribution and specify the data used to quantify the distribution parameters. Alternatively, estimate the sensitivity of the fire CDF to the probability of control room abandonment by assuming a value based on the FitzPatrick control room characteristics.
4.
The FitzPatrick submittal utilizes an approach to the analysis of electrical cabinet fires that is the same as the approach recommended by the EPR/ Fire PRA Implementation Guide.
Enclosed ignition sources cannot lead to fire propagation or other damage.
Fire spread to adjacent cabinets cannot occur if the cabinets are separated by a double wall with an air po or if the cabinet in which the fire originates has an open top.
Oil-filled transformers and high-voltage components in cabinets, for example, are susceptible to energetic faults leading to cabinet breach. Switchgear fires at Yankee-Rowe in 1984 and Oconee Unit 1 in 1989 both resulted in fire damage outside the cubicles. Cabinets are also I
susceptible to warping under intense heat loads, which would invalidate any assumption of 4
limited combustion air. Assumptions in the FitzPatrick submittal on fire propagation from enclosed ignition sources should be verified, especially in such typically important areas as the i
relay room and cable spreading room.
Please provida the basis for the assumption and a discussion on how the specific enclosures were analyzed to ascertain that the assumption is applicable to them.
5.
The FitzPatrick sub..T.2al assumes cabinet heat release rates of 65 Btu /s. In contrast, experimental work has developed heat release rates ranging from 23 to 1171 Btu /sec.
Considering the range of heat release rates that have been observed and could be applicable to different control cabinet fires, and to ensure that cabinet fire areas are not prematurely screened out of the analysis, a heat release rate in the mid-range of the r urrently available experimental data (e.g., 550 Btu /sec) should be used for the analysis.
j Please discuss the heat release rates used in your assessment of control cabinet fires. Please provide a discussion of changes in the IPEEE fire assessment rv uits ifit is assumed that the heat release from a cabinet fire is increased to 550 Btu /s.
6.
Tiie submittal notes that non-lEEE383-qualified cabling is present in the plant, but that installations since construction have used only IEEE383-qualified cable. The description i
of the cable parameters. 900 F for ignition,732' F for damage, are more typical of qualified cable.
Please identify the fire zones and scenarios where non-IEEE383-qualified cable is present, the ignition and damage temperatures appropriate to that cable, and the impact on the results of the fire studyif those parameters are assumed. Identify and estimate the i
fire CDF contribution from any scenarios screened inappropriately by assigning qualified cable properties to non-qualified cable.
7.
Section 4.7.3.1 cites the EPRI Fire PRA /mplementatiw Guide methodology for the treatment of automatic suppression system unavailability in the multi-rone analysis.
This reference treats manual recovery of automatic suppression systems as being independent of subsequent manual efforts to suppress the fire. This assumption is optimistic, as the fire conditions (e.g., heat, smoke) that lead to the failure of recovery efforts can also influence the effectiveness of later suppression efforts. Such an approach, therefore, can overlook plant-specific vulnerabilities.
It is important that all relevant factors be considered in an evaluation of the effectiveness of fire suppression. These factors include: (a) the delay between ignition ar.d detector /suppre ssion system actuation (which is specific to the configuration being analyzed); (b) the time-to-damage for the critical component (s) (which is specific to the fuel type and loading as well as to the configuration being modeled); (c) the response time of the fire brigade (which is plant-specific and fire-location-specific); (d) the time required by the fire brigade to diagnose that automatic suppression has failed and to 5
1 i
take manual action to recover the automatic suppression system, and (e) performance shaping factors (PSFs) affecting fire brijde actions. These PSFs could include factors 4
such as perseverance (persistent efforts made to recover a failed automatic suppression system), smoke obscuration, and impaired communications [R4].
Finally, it should be noted that the NRC staff's evaluatica of the FIVE methodology specifically stated that licensees need to assess the effectiveness of manual fire-fighting teams by using plant-specific data from fire brigade training to determine the response time of the fire fighters.
Please identify those scenarios for which credit is taken for both manual recovery of automatic suppression systems and manualsuppression of the fires (if manual recovery efforts are unsuccessful), andplease indicate the plant equipment that may be affected by the fires. In the analysis of these scenarios, how are dependencies between manual actions treated? Please Justify the treatment, considering the expeuted fire environment, the recovery actions required, and the manual fire suppression actions required.
8.
The failure probability for automatic suppression applied values compatible with the FIVE methodology. This data is acceptable for systems that have been des gned, I
insta!Ied, and maintained in accordance with appropriate industry standards, such as those published by the Natienal Fire Protection Association (NFPA).
Please verify that automatic fire suppression systems at FitzPatrick meet NFPA standads.
References for (Big R1.
P.J. DiNenno, et al, eds., "SFPE Handbook of Fire Protection Engineering,"
2nd Edition, National Fire Protection Association, p. 3-140,1995.
i R2.
L. Y. Cooper, M. Harkleroad, J. Quintiere, W. Rinkinen, "An Experimental Study of Upper Hot Layer Stratification in Full-Scale Multiroom Fire Scmarios," ASME Joumal of Heat Transfer,1Q4,741-749, November 1982.
R3.
J. Chavez, et al., "An Experimental Investigation of Intemally ignited Fires in Nuc! ear Power Plant Cabinets, Part Il-Room Effects Tests," NUREG/CR-4527N2, October 1988.
R4.
J. Lambright, et al., "A Review of Fire PRA Requantification Studies Reported in NSAC/181," prepared for the United States Nuclear Regulatory Commission, April 1994.
R5.
J. Usher and J. Boccio, " Fire Environment Determination in the LaSalle Nuclear Power Plant Control Room," NUREG/CR-5037, prepared for the United States Nuclear Regulatory Commission, October 1987.
6 1
C.
High winds, flood, and other external events 1.
It is stated in the submittal that an analysis has indicated that, for the condition of probable maximum precipitation (PMP), the load capacity of the reactor building c
2 roof (50 lb/ft ) will cn'y be exceeded if two of three roof drains on one side of the roof are blocked and, as a result, the water depth exceeds 9.6 inches.
Please provide information on any plant procedures that address the prevention of blockage of tt e reactor building roof orains, orif there are no such procedures, provideJustification for their omission.
t s
7
_____