ML20150C178
| ML20150C178 | |
| Person / Time | |
|---|---|
| Issue date: | 03/16/1988 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-2561, NUDOCS 8807120339 | |
| Download: ML20150C178 (85) | |
Text
_
m f f-d$Y 3Q-.
1 uy; ew M/d b
2 pB q
G D{
, N f
DATE ISSUED:
3/16/88 t
OJA9 4 lain
SUMMARY
/ MINUTES OF THE ACRS SUBCOMMITTEE MEETING ON THE AUXILIARY SYSTEMS MARCH 9, 1988 WASHINGTON, D.C.
INTRODUCTION The ACRS Subcomittee on the Auxiliary Systems held a meeting on Wednes-day, March 9, 1988 at 1717 H Street, N.W., Washington, D.C., to discuss primarily the results and conclusions of the Fire Risk Scoping Study that was perfonned by the Sandia National Laboratories (SNL) for the NRC.
The entire meeting was open to public attendance except a portion of the meeting that was closed to discuss proprietary infonnation associated with the fire protection provisions at French Nuclear Power Plants. Mr. Sam Duraiswamy was the cognizant ACRS Staff Engineer for this meetir.g. A list of documents submitted to the Subcomittee is included in Attaciiment A, and a tentative presentation schedule for the meeting is contained in Attachment B.
ATTENDEES ACRS:
C. Michelson (Subcomittee Chairman),
J. C. Ebersole, C. J. Wylie, C. P. Siess (Part Time), and D. W. Moeller (Part Time)
Sam Duraiswamy (Cognizant ACRS Staff Engineer)
Principal Speakers:
NRC:
J. Flack, A. Busiik, T. King, and A. Datta SNL:
M. Bohn, J. Lambright, and S. Nowlen EXECUTIVE SESSION Mr. Michelson, the Subcomittee Chainnan, convened the meeting at 8:35 a.m., and stated that the purpose of the meeting was to hear presenta-tions by and hold discussions with representatives from the Office of Nuclear Regulatory Research (RES) and SNL with respect to the following 8807120339 880316 P
^
2 61 PDR Certified Fy
/
J
1
\\
Auxiliary Systems Meeting Minutes March 9, 1988
- Objectives of the Fire Risk Research
- Overview of the Fire Risk Scoping Study s
- Results and Conclusions from Tasks 1 and 2 of the Scoping Study
- Identification of Fire Risk Issues
- Results and Conclusions from Tasks 4 and 5 of the Scoping Study
- Peer Review Coments on the Results of the Study
- SNL Recommendations for Follow-on Efforts In addition, a closed session will be held to discuss proprietary information associated with the fire protection provisions at French Nuclear Power Plants.
He stated that the Subcommittee had received no written coments from members of the public.
However, it has received a request from Mr. Kaminski, Fire Protection Officer, Wisconsin Electric Power Company, for time to make oral statements on the results and conclusions of the Fire Risk Scoping Study.
FIRE PROTECTION PROVISIONS AT 900 MWE FRENCH NUCLFAR POWER PLANTS (CLOSED) - MR. A. BUSLIK, RES Mr. Buslik stated that the fire protection program at 900 MWE French Nuclear Power Plants consists of the following:
- Prevention and limitation of propagation
- Detection
- Suppression
e 1
i j
Auxiliary Systems Meeting Minutes March 9, 1988 i
He discussed briefly the following:
)
' Cable protection
' Smoke control measures
- Fire detection systems
- Fire suppression systems j
i
- Manual fire suppression FIRE RISK RESEARCH OBJECTIVES - MR. J. FLACK, RES Mr. F' lack stated that the main objectives of the NRC fire risk research are to:
- Determine the need for additional Fire Protection Research j
' Set priorities for future research, if required In order to cbtain additional information for use in fulfilling the j
above objectives, the Fire Risk Scoping Study was initiated at SNL.
The main purpose of the Scoping Study is to assess the risk significance and dominant sources of uncertainty associated with fire risk issues. The results of the Scoping Study focus en:
' Major fire risk contributors
- Root sources of uncertainty
- Completeness of Appendix R requirements Dr. Siess asked how much money has already been spent on the Fire Protection Research. Mr. Datta responded that about $6 million has been spent between FY 1974 and FY 1986.
3 Auxiliary Systems Meeting Minutes March 9, 1988 Dr. Siess asked whether there was any program initially to detemine the need for Fire Protection Research prior to initiating such a research.
Mr. Datta responded that based on the Office of Nuclear Reactor Regu-lation (NRR) needs, a program on Fire Protection Research was developed and documented in NUREG-1148.
Dr. Siess comented that it would be interesting to develop a history of the NRC Fire Protection Research in relation to its need and the amount of money spent, and to find out why it was decided now to reassess the importance of this research.
FIRE RISK SCOPING STUDY Introduction, Overview, Conclusions, and Recomendations of the Tasks Performed - Mr. M. Bohn, SNL Mr. Bohn stated that the main objectives of the Fire Risk Scoping Study are to:
- Review and requantify certain past fire risk scenarios in light of updated data base resulting from the NRC-sponsored Fire Protection Research Program end updated computer fire modeling capabilities.
- Identify potentially significant fire risk issues that have not been addressed in the previous fire PRAs and determine the risk significance of these issues.
- Review current regulations associated with fire protection and plant implementation practices for relevance to the unaddressed fire risk issues.
He stated that the following five Tasks were perfonned under the Fire Risk Scoping Study:
Task 1:
Assess Uncertainties in Four Previous PRAs
1 i
Auxiliary Systems Meeting Minutes March 9, 1988 Task 2:
Requantify Fire Scenarios from Past PRAF J
Task 3:
Identify Potential Fire Risk Issues Task 4:
Assess the Risk Significance of Potential Issues Task 5:
Evaluate Completeness of Appendix R Requirements Mr. Bohn stated that using the new data they h w e.'eviewed and requan-tified the fire PRAs associated with Limerick, Seabrook, Oconee, and Indian Point nuclear plancs.
He mentioned that they did not make any i
changes to the original assumptions used in these PRAs.
The results of the requantified PRAs showed that the fire-induced core-malt frequency is higher than that in the original PRAs (Attachment C, Page 1). Even
'ith Appendix R modifications, it was found that fire is a signif mant w
contribetor to core-melt frequency.
Mr. Bohn stated that based on the assessment, tney found that uncer-tainties in the past four fire PRAs were driven by judgmental factors.
Some of the jwigment-based factors used iv. the prey!ous PRAs are:
' Partitioning of fire occurrence frequency
- Fire severity factor
- Area of rire influence
' Types of initiating events caused by fire
- Conservatism factor for COMPBRN fire code l
i
1 s
Auxiliary Systems Meeting Minutes March 9, 1988 He stated that because of the differing and evolving fire PRA method-ologies, it was difficult to come up with a relative comparison of plant-to-plant risk.
However, they observed that even with differing methodologies, ff re-induced core-melt frequencies were high.
Mr. Bohn p.ovided some recomendations for improving fire risk perspec-tive (Attachment C, Page 2):
' Develop defensible estimates of fire occurrence frequency uncer-tainties reflecting plant-to-plant variations.
- Develop defensible guidelines for partitioning and fire severity factors.
- Validate COMPBRN and more advanced fire codes against actual test
- data, Mr. Bohn stated that the following six issues have not been addressed in the past fire PRAs:
- Manual Fire Fighting Effectiveness
' Control Systems Interactions
- Total Environment Effects (Smoke and Inadvertent actuation of Fire SuppressionSystems)
- Adequacy of Fire Computer Codes
- Barrier Effectiveness
' Seismic / Fire Interactions
3 s
Auxiliary Systems Meeting Minutes March 9, 1988 l
Mr. Bohn discussed briefly the potential impacts of the issues that were not addressed in the pcst PRAs (Attachment C, Page 3).
He said that most of the above issues may contribute significantly to the core-melt frequency.
Since the issues on Smoke Control, and Inadvertent Actuation of Fire Suppression Systems are plant specific, it was difficult to determine the extent to which they would contribute to the core-melt frequency.
Mr. Bohn discussed briefly the conclusions and recommendations resulting from the assessment of the issues that were not addressed in the past PRAs. Some of the conclusions and recomendations associated with the unaddressed issues are given below.
Manual Fire Fighting Effectiveness
==
Conclusions:==
- Must suppress fires in 5-20 minutes to be effective.
' Many areas exist where critical damage is likely prior to the arrival of the fire brigade.
- Knowledge of critical plant systems is very crucial.
' Dense smoke will hampyr fire fighting efforts.
- Fires suppressed by Halon or C0 are likely to re-ignite.
2 Recomendations For Improvement
- Develop guidelines for required response and extinguistraent time for various equipment types.
1 r
Auxiliary Systems Meeting Minutos March 9, 1988
- Evaluate fire brigade effectiveness under actual fire con-ditions including dense and toxic smoke.
- Provide plant-specific training to fight fires in areas where equipment susceptible to spray are located.
Adequacy of Fire Computer Codes
==
Conclusions:==
- Accurate calculation of time to component damage is critical to assess fire suppression response time.
- Results of COMPBRN III code are sometimes conservative and i
sometimes non-conservative.
- No existing fire codes have been benchmarked against actual test data.
Recommendations for Improvement
- Validate correlations in COMPBRN III against test data,
- Benchmark COMPBRN III and CHAM codes against Fire Enclosure Test data to determine their limits of applicability.
- Develop control volume or flow code capable of predicting accurately component damage times.
1
i s
\\
Auxiliary Systems Meeting Minutes 9-March 9, 1988 1
l Control Systems Interactions
==
Conclusions:==
- Even with electrical independence between control room and remote shutdown panel functions, subtle interactions involving control room panel fires and random failures can cause signif-icant core damage scenarios.
- Many remote shutdown panels have limited capability.
Lack of good indication at remote shutdown panel may increase risk.
Recomendations to Reduce Control Systec,s Interactions
- Develop a method and guidelines for detailed review of elec-trical independence between control room panels and remote shutdown panel.
- Perfonn review of remote shutdown panel indications and capabilities and prepare recomendations for enhancement.
Total Environmental Survivability gnclusions-SmokeControl:
- Test results show that total losc of visibility will occur in 5-8 minutes given a typical cabinet fire.
- Normal control room ventilation rate is not effective in preventing loss of visibility.
i i
- Spread of smoke could cause inadvertent actuation of fire suppression systems and hamper fire brigade access.
4 e,--n
,,.-m
.,---n
.-,-.,~--s,
..,-e-
1
\\
Auxiliary Systems Meeting Minutes March 9, 1988
- Potential impacts due to smoke are very plant specific.
Conclusions - Inadvertent Actuation of Fire Suppression Systems
' Past experience shows that inadvertent actuations of fire suppression systems occur with a frequency of about 10-2/ year.
- Without actual fire, inadvertent actuations do not seem to have significant generic safety impact.
With actual fire, the impact seems to be small; however, it is very plant specific.
Recommendations for Total Environmental Survivability Issues
- Conduct plant-specific review of area combinations where spread of smoke could hamper fire fighting or cause inadver-tent actuation.
- Investigate impact of in-cabinet gaseous fire suppression systems on control circuitry.
- Investigate impact of smoke on high-voltage equipment.
Effectiveness of Fire Barriers
==
Conclusions:==
- Aging cracks or barrier test methods may have significant impact on the reliability of barriers than anticipated.
- If penetration barrier reliability is less than 90 percent, it could increase the core-melt frequency significantly.
1 s
Auxiliary Systems Meeting Minu'.es March 9, 1988 Recomendations:
I
- Evaluate reliability of fire barriers based on experience and test data.
- Perform tests under fire conditions on aged penetration seals to evaluate fire barrier reliability.
Mr. Ebersole asked whether they have any feel for the frequency of fires caused by plant personnel. Mr. Lambright responded that roughly abou+
50 percent of the fires were caused by plant personnel.
He mentioned that some of the fires caused by people didn't get reported because they were extinguished imediately.
Stating that the value of the core-melt frequency at the Indian Point plant has been reduced significantly, from 6.5x10-5 to 8.8x10-6, as a result of the implementation of App'endix R modifications, Mr. Wylie asked whether there was one major change that contributed to such a reduction. Mr. Lambright responded that rerouting of a separate power cable for one charging pump in one train of the cooling system contributed to this major reduction.
Mr. Ebersole asked whether they have looked at any plants that use compartmentalization. Mr. Bohn responded that they have not looked at any plants designed that way.
Mr. Michelson asked whether they have adequate test data for use in validating the fire computer codes.
Mr. Bohn responded that they have data from 21 different tests.
However, efforts to reduce data frem these tests had been curtailed owing to budget constraints.
i s
Auxiliary Systems Meeting Minutes March 9, 1988 With reference to a statement made by Mr. Bohn that collateral damage hasonlyasmallimpactoncore-meltfrequency(AttachmentC,Page3),
Mr. Michelson asked how did they arrive at this conclusion. Mr. Bohn responded that in the plants that they have looked at they did not find lot of situations to indicate that collateral damagc would be a signif-icant contributor to core-melt frequency.
Mr. Ebersole asked whether they have looked at the effects of charcoal filter fires.
Mr. Lambright responded that there were several events associated with charcoal filter fires.
However, since they were away from safety-related components, they did not pose any challenge to the plant.
Mr. Michelson asked whether they have looked at the access routes to a room assuming that the lighting system in that room has been lost due to fire. Mr. Bo;1n responded they believe that it is a potential problem and should be looked at.
l Mr. Ebersole asked whether they have looked at the possibility of the occurrence of explosion resulting from the reignition of combustible vapors.
Mr. Bohn responded that they did not investigate this issue.
Mr. Nowlen stated that too much vapor buildup in a small room may cause some major problem.
However, in a large room it may not cause a significant concern.
Mr. Michelson asked whether they found any oil-filled transfonners in the plants that they have looked at. Mr. Lambright responded that they did not perform walk-downs on all the plants that they investigated.
To his knowledge, there are no oil-filled tra.,2fonners in those plants.
1 1
Auxiliary Systems Meeting Minutes March 9, 1988 Mr. Michelson asked wilether they have looked at the effects of heat propagating from a cabinet fire on the equipment and solid state control devices that are mounted on the other side of the cabinet. Mr. Nowlen responded that they did not investigate in detail the propagation of heat from a cabinet fire and its impact on equipment mounted on the other side.
However, results of the cabinet tests conducted so far showed that the temperature in the adjacent cabinets were well below the auto-ignition temperature.
In response to a question from Mr. Michelson, Mr. Nowlen stated that results of the cabinet tests showed that the maximum temperature on the other side of the wall was about 390 F, and the air temperature was about 150*F.
Mr. Michelson commented that the cooling devices may not perform effec-tively under a 150 F environment.
Mr. Michelson asked whether they will be able to put out a cabinet fire with a hand-held CO fire extinguisher.
Mr. Nowlen responded that 2
recently a cabinet fire at the Rancho Seco plaat was put out by a hand-held C0 extinguisher.
Mr. Bohn stated that if the fire is small 2
and detected early, hand-held CO extinguisher would be effective in 2
putting out that fire. Otherwise, it would not be effective.
Mr. Michelson asked whedler someone in the fire brigade has the au bori-ty to make a decision to use water to extinguish the fire in the event that C0 was found to be ineffective.
Mr. Nowlen responded that they 2
did not ask this question specifically when they talked to utility personnel.
The fire fighting practices vary tremendously from plant to plant. Some utilities allow their fire fighters to take whatever measures necessary to extinguish the fire.
1 Auxiliary Systems Meeting Minutes March 9, 1988 Mr. Ebersole asked whether there are any procedures that require that some electrical apparatus should be de-energized prior to putting out the fire. Mr. Nowlen responded that some utilities may require that electrical equipment should be de-energized before putting out the fire and some may not.
Mr. Michelson asked whether they have looked at the effectiveness of fire fighting in total darkness. Mr. Nowlen responded that several utilities train the fire brigades to fight fire under such circumstance.
Mr. Michelson comented that the conclusion that "inadvertent actuation of fire suppression systems associated with fire has only small impact" does riot seem ret.listic.
Since they do not have adequhte data base on this matter, he does not believe that they can arrive at such a conclu-sion.
RESULTS AND CONCLUSIONS FROM TASKS 1 AND 2 - MR. J. LAMBRIGHT, SNL Mr. Lambright stated that under Tasks 1 and 2 specific scenarios from four past PRAs were requantified using updated data base and modeling techniques.
The intended objectives of these tasks are to:
' Identify the remaining sources of uncertainty in the previous analyses.
- Determine whether the overall perspective of fire risk had been altered significantly as a result of a better understanding of the fire problems.
' Detennine the effects of the implementation of the Appendix R requirements on risk estimates.
1 i
Auxiliary Systems Meeting Minutes March 9, 1988 He said that in the requantification process certain limitations were placed.
The fire areas were reevaluated based on the scope of issues considered in the original analyses.
Reevaluation was performed using, as closely as possible, the same analysis techniques used in the original work. Unaddressed issues identified under Task 3 were not incorporated into the analyses of Tasks 1 and 2.
COMPBRN 1 computer fire code was used in the past four PRAs to predict the fire growth and fire damage; in the requantification process an improved version of COMPBRN III was used.
1 Mr. Lambright stated that based on the evaluation of the past PRAs, they have identified major uncertainties in the following areas:
- Initiating event frequencies and partitioning factors
' Fire propagation mcdeling
- Fire suppression modeling
- Fire-induced transients vs LOCAs He discussed briefly the uncertainties associated with the past PRAs (Attachment C, Pages 4-7).
He said that in the Limerick PRA, it was assumed that 87 percent of the core-melt frequency would be from self-ignited panel fires and self-ignited c ble fires.
On the other hand, the other three PRAs performed by Pickard, Lowe, and Garrick (PL&G) assumed that transient oil fires would be he significant contributor to core-melt frequency.
Also, Limerick dia r.ot consider LOCAs to be credible due to cable separation.
On the other hand, PL&G assumed that at Indian Point 2, 95 percent of the core-melt f*equency would be from fire-induced small LOCAs, and at Oconee it would be 52 percent.
1 i
Auxiliary Systems Meeting Minutes March 9, 1988 Mr. Lambright stated that initiating event frequencies for all tour previous PRAs were requantified using the data base developed by Ted Wheelis (Sandia Fire Program).
Based on the work done under Tasks 1 and 2, they made the following observations:
- Fire-induced core-melt frequencies increased for all four previous PRAs (Attachment C, Page 8). This increase alone will result in direct increase in overall fire risk even if all other factors remain constant.
- Even with the modifications resulting ' rom the implementation of Appendix R requirements, fire-induced core-melt frequency is a significant contributor.
- Use of COMPBRN III code for the prediction of fire growth and fire damage resulted in mixed effects.
The propagation times predicted by COMPBRN III are considerably less than those predicted by COMPBRN I in the original PRAs.
However, based on experience, they believe that there are number of inadequacies in COMPBRN III, and therefore the results of COMPBRN III should be viewed as tentative at this time.
- There are large uncertainties still exist. These could be reduced by applying a consistent methodology to all four previous PRAs which is not constrained by the original assumptions.
Mr. Michelton commented that since COMPBRN III code was used in the requantification of past PRAs, some people may question the validity of the requantification results.
Mr. Michelsnn asked what is the total core-melt frequency for Limerick.
Mr. Lambright responded he believes that it is between 60 and 80 per-cent.
1 Auxiliary Systems Meeting Minutes March 9, 1988 Stating that Oconee has a separate shutdown facility as well as remote shutdown panel, Mr. Wylie asked how much impact these systems have on the core-melt frequency. Mr. Lambr yht responded that to his knowledge, Oconee did not seem to give credit to Le remote shutdown panel.
He said that the following scenarios were considered to be dominant con-tributor to the core-melt frequency at Oconee:
' A cable shaft fire leads to either a seal LOCA or a struck-open relief valve. Under this case, high pressure injection must be initiated locally (operator error dominates random system fail-ures).
- A cable shaft fire causes a transient sequence.
(Operator error at the separate safe shutdown facility leads to core damage.)
Considering the above, it seems that the operator error wipes out the effectiveness of the separate shutdown facility.
TASK 3 - IDENTIFICATION OF FIRE RISK ISSUES - MR. S. N0WLEN, SNL Mr. Nowlen stated that based on the results of the research performed in the fire protection area, they have developed a list of fire risk issues that have not been addressed in the previous fire PRAs.
They reviewed LERs, I&E Bulletins and Information Notices to see whether other issues need to be added to the original list.
As a result of this effort, no new issues were identified.
Then, they solicited opinions of experts in various fields associated with fire safety, fire research, fire protec-tion, risk analysis, plant design, etc.
As a result of expert opinion polling, the issue associated with the Adequacy of Fire Barriers was identified.
Based on the results of the requantification of previous j
fire ? ras, they have identified the issue on Adequacy of Analytical
~..
i Auxiliary Systems Meeting Minutes March 9, 1988 Tools for Fire.
He said that the final list of unaddressed fire risk issues consists of the following:
i
- Control Systems Interactions 1
- Seismic / Fire Interactions
- Manual Fire Fighting Effectiveness (Including Smoke Control)
- Total Environment Equipment Survival (Including Spurious actuation of fire suppression systems)
- Adequacy of Fire Barriers
' Adequacy of Analytical Tools for Fire Mr. Michelson asked whether they have obtained any input from utilities on this matter. Mr. Bohn responded that they have consulted fire protectio.' engineers at several utilities to obtain input for use in the fire risk analysis being performed by SNL for NUREG-1150.
Also, several utilities, such as Wisconsin Electric Power Company, and Duke Power Company, participated in the peer-review process associated with the Fire Risk Scoping Study.
They have contacted also the personnel at the Institute for Nuclear Power Operations (INP0); however, they declined to participate in the process.
Mr. Ebersole asked why they did not solicit the opinion of the National Fire Protection Association. Mr. Nowlen responded that this group, which consists of members from vide variety of manufacturing companies,-
is involved primarily in writing standards for different industries.
They do not concentrate specifically on nuclear power plants.
There-fore, SNL did not solicit the opinion of this group.
i l
l
\\
Auxiliary Systems Meeting Minutes March 9, 1988 Mr. Michelson asked how many fire protection experts are with the NRC at the present time.
Mr. Datta responded that there are about seven qualified fire protection experts in the NRC regional offices.
RES has one and NRR has four or six fire protection experts.
RESULTS AND CONCLUSIONS FROM TASKS 4 AND 5 Manual Fire Fighting Effectiveness - Mr. J. Lambright, SNL Mr. Lambright stated that in order to obtain information to quantify the risk significance of manual fire fighting effectiveness, Professional Loss Control (PLC) surveyed 45 nuclear plant sites.
Results of the PLC survey indicated wide variability in the organization, size, training, and equipping of fire brigades (Attachment C, Page 10).
The information obtained from the survey was applied to the previously reviewed fire risk scenarios under Tasks 1 and 2, and core-melt frequencies were i
recalculated by modifying only the factor related to manual fire fight-ing.
The effects of fire brigade response and extinguishment time on core-melt frequency are included in Attachment C, Page 11.
Mr. Lambright stated that the uncertainty as to where to direct sup-pression efforts is a major concern associated with the manual fire fighting issue. There is a potential that misdirected suppression efforts may actually result in damage to equipment not directly involved in the fire. He discussed briefly the approach used in quantifying the risk associated with the misdirected suppression effort and the associated results (Attachment C, Pages 12 and 13).
Stating that systems intended to supply water for core cooling might fail in a fire environment, Mr. Ebersole asked whether they have con-sidered the alternative sources for core cooling in their analysis.
Mr.
Lambright responded that they have used the same scenarios used in the past PRAs.
They did not attempt to come up with additional scenarios or new ways for core cooling.
j
\\
Auxiliary Systems Meeting Minutes March 9, 1988 Mr. Michelson asked whether they always use fog nozzles on the hose and what is the ('fective range of these nozzles.
Mr. Lambright responded that some people use adjustable nozzles and some use fog nozzles.
He believes that people using adjustable nozzles are normally trained to adjust them properly when dealing with electrical cabinet fires.
He is not sure of the effective range of the fog nozzles.
Mr. Bohn stated he believes that the range of fog nozzles is about 10 or 12 feet, i
Mr. Michelson asked whether a 8 ft. high wall with an open ceiling could be rated as a 3-hour fire barrier.
Mr. Nowlen responded that unless the ceiling is closed, they cannot rate that as a 3-hour fire barrier.
Stating that he has seen feedwater pumps separated by 8 or 10 ft. high concrete walls with open ceilings, Mr. Michelson asked what is the rating of these walls. Mr. Lambright responded that they came across a similar situation where charging pumps were separated by 8 or 10 ft.
high walls, but they did not rate the walls as 3-hour fire barrier.
Mr.
Bohn stated that Appendix R states that major structural walls of reinforced concrete or equivalent are considered to be 3-hour fire barriers.
However, it does not make it clear whether this will apply to walls with open ceiling.
Control Systems Interactions - Mr. J. Lambright, SNL Mr. Lambright stated that the approach used to quantify the risk associ-ated with the issue on Control Systems Interactions is as follows:
- NUS Corporation assessed the electrical independence of the remote shutdown panel from the control room.
- Using the Risk Methods Integration and Evaluation Program (RMIEP) system models of LaSalle County Nuclear Power Plant, Unit 2, a multiple spurious actuation or loss of control function analysis
l Auxiliary Systems Meeting Minutes 21-March 9, 1988 was performed for the most critical control cabinets within the control room.
Mr. Lambright stated that based on the results of the remote shutdown q
electrical independence analysis performed by NUS Corporation (Attach-ment C, Page 14), three general types of interaction were identified:
i
' Loss of control power for the operated device as a result of a blown fuse before transferring to remote shutdown.
- Spurious actuation of components either leading to LOCAs or compo-nent damage.
' Total loss of component function due to redundant control equipment failure.
He said that the first two scenarios mentioned above were examined and their impact on fire scenarios developed in Limerick, Indian Point, and Seabrook studies was assessed. The results (Attachment C, Page 15) showed that even with the assumption that one of the safe shutdown systems remain operable, the contribution of the control systems interactions to core-relt frequency is substantial.
Mr. Lambright discussed briefly the analysis performed to quantify the effects of single or multiple spurious actuations due to a control room cabinet fire at LaSalle and its results (Attachment C, Pages 16-20).
Stating that the ECCS'and the electrical distribution panels in the LaSalle control room are partitioned by using approximately one-eighth inch thick steel (Attachment C, Page 18), Mr. Michelson comented that using an one-eighth inch thick steel as a fire barrier dces not seem appropriate; it will get real hot under fire and will be ineffective.
r g
+
ns
\\
Auxiliary Systems Meeting Minutes March 9, 1988 Total Environment Equipment Survival - Mr. S. Nowlen, SNL Mr. Nowlen stated that this item consists of two issues:
' Environment created by the application of fire suppression systems during actual fires.
- Environment created by the spurious actuation of fire suppression systems when there is no fire.
Types of environment of potential concern are included in Attachment C, Page 21. He said that data on the vulnerability of equipment in a fire environment is very sparse.
There are some limited information avail-1 able on certain types of table damage due to themal effects.
- Also, they have some information on smoke deposition.
For components, other i
than cable, there is virtually no data available on themal damage.
Due to lack of sufficient data, they had difficulty in quantifying the risk significance of this issue.
Mr. Nowlen discussed briefly the history of events associated with spurious actuation of fire suppression systems, equipment and systems vulnerable to the environment created by the spurious actuation of fire protection systems (Attachment C, Pages 22-28).
He discussed also the significance of the events related to spurious actuation of fire pro-tection systems (Attachment C, Page 29).
Mr. Nowlen stated that operating experience shows that spurious actua-tion of fire protection systems have been responsible for the degrada-tion of several safety systems such as diesel generators, high pressure coolant injection systems, reactor core isolation cooling systems, etc.
In one incident, spurious suppression system actuation during a refuel-ing outage resulted in the spilling of over 30,000 gallons of contam-inated water within a reactor building.
At least 12 of the 71 spurious suppression events, used in the Scoping Study, have forced an automatic
Auxiliary Systems Meeting Minutes March 9, 1988 trip, manual scram, or normal shutoown of the~ plant itself.
Loss of offsite power, reactor scrams, and control systems interactions result-ing from the spurious actuation of the fire protection systems are considered to be the most serious events.
However, experience shows that the frequencies of such incidents occurring as a result of spurious actuation of suppression systems are 1 to 2 orders of magnitude less than these incidents as a whole.
For these incidents to represent a significant contributor to plant risk, they must involve the degradation of one or more safety systems.
In response to a question from Mr. Michelson, Mr.Nowlen s'ated that for any given fire it should be demonstrated that there is one train available to bring the plant to hot shutdown status.
Mr. Michelson asked whether there is still any doubt that dust can set off smoke detectors.
Mr. Nowlen responded that it has been well recoc-nized that dust will set off smoke detectors.
He believes that using heat sensors to detect fire may preclude the possibility of the occur-rence of this problem.
Mr. Michelson asked whether they have tried to find cut the types of fire protection provided in the turbine area.
Mr. Nowlen responded that they did not perform an extensive plant review on this issue.
Mr. Michelson stated that General Design Criterion (GDC)-3 requires that spurious actuation of fire protection systems should not affect any safety-related components. However, operating experience shows that spurious actuation of fire protection systems had affected safety-related components.
He asked whether this means that GDC-3 requirements have not been fully enforced or the utilities have not complied with the requirements of GDC-3. Mr. Nowlen responded that many utilities have
1 i
Auxiliary Systems Meeting Minutes March 9, 1988 gene to great lengths to comply with GDC-3. However, there are still some unidentified vulnerabilities exist, and until these are identified j
and taken care of these things will continue to happen.
Mr. Michelson comented that some of the issues that need to be con-sidered in the evei;t of a loss of offsite power are as follows:
- Inadvertent actuation of fire' protection systems after the restora-tion of power due to the malfunction of deluge valves (in the fire protection system) that are controlled either by water pressure or air pressure, j
' Restarting of ccupressors (that normally use freon) imediately after restoring power.
(The concern here is that after loss of power, the oil in the freon gets cooled.
Once it is cooled, it is difficult to restart the compressors until the oil is reheated to a certain temperature),
i Mr. Michelson commented that there should be a Regulatory Guide or Rule to require that loss of auxiliaries supporting fire protection systems should not result in the inadvertent actuation of these systems.
Mr. Michelson asked whether they know for sure that cardox.,ystems will not get actuated accidentally upon loss of all power.
Mr. Bohn respond-ed that they have dedicated battery backup system to protect the cardox system for a limited period of time.
Mr. Michelson asked if there is an inadvertent actuation of fire pro-tection system in one area while putting out the fire in another area, will there be enough water supply to cope with the actual fire. Mr.
Bohn responded that they did not specifically look into this issue.
Mr.
Michelson suggested that SNL take a look at this matter and the result-ing consequences.
4 r,
-y
i Auxiliary Systems Meeting Minutes March 9, 1988 Mr. Nowlen discussed briefly the adequacy of Appendix R requirements to deal with the issue on Total Environment Equipment Survival (Attachment C,Page30).
With reference to a statement by Mr. Nowlen that Appendix R requires that safety-related equipment should be protected from fire suppression systems, Mr. Michelson comented he believes that Appendix R requires that the functions of The safety-related equip.v.:t should be protected from fire suppression systems.
However, GDC-3 requires that safety-related equipment should be protected from fire suppression systems.
Mr. Bohn stated that he will look into this matter.
Adequacy of Fire Barrier Qualification Standards - Mr. Nowlen, SNL Mr. Nowlen stated that U.S. Standards associated with the qualification of fire barriers do not specify a differential pressure to be maintained across the barrier during fire exposure testing. Testing organizations in the U.S. maintain nonnally a negative pressure on the fire side of the barrier in order to remove efficiently combustion products to the air-handling systems. However, experimental results show that a posi-tive pressure would develop on the fire side of the barrier due to the generation of Feat which results in an expansion of the air in the enclosure.
Standards used in other countries, except U.S. and Canada, specify that fire barriers should be qualified under conditions of J
positive exposure pressure.
Experimental results show that under actual fire conditions a barrier that passes U.S. Standard qualification tests i
may fail sooner than such qualificat'ons may indicate.
Mr. Nowlen stated that fire barriers are very important protective features.
Fire barriers should be about 99 percent reliable in order to minimize risk significance.
If the barriers are only about 90 percent reliable, they will contribute significantly to core-melt frequency.
They believe that adequate data is not available to detennine the
+ - -
y-
,- - - ~ - -. -
.-4 y
T Auxiliary Systems Meeting Minutes March 9, 1988 reliability of fire barriers. A historical review of both nuclear and non-nuclear fire experience base could demonstrate the degree of reli-ability of fire barriers used in the U.S.
Mr. Michelson stated that prior to the fire event at the Browns Ferry plant, most of the plants used flamable material for penetration seals.
After the Browns Ferry event, it was required that non-flamable mate-rials should be used for penetration seals.
He asked whether there are any plants that still use flamable materials for the penetration seals, and is there any NRC requirement that such seals should be replaced.
The Staff and SNL didn't know the answer to this question.
Mr. Ebersole comented that dampers used in the ventilation ducts that are triggered closed by fusible links normally let hot gases and smoke pass through prior to closing, ar.d hot gases passed through these dampers may have significant impact on solid state components.
Mr.
Nowlen responded that utilities are aware of this problem.
In Order to avoid such a problem, they have been using fusible links that operate at low temperature, j
Adequacy of Analytical Fire Tools - Mr. S. Nowien, SNL Mr. Nowlen stated that COMPBRN 1 computer code has been used in all previous fire PRAs to predict the fire growth and the fire-induced damage to equipment.
However, COMPBRN 1 has several inconsistencies; it seems to be overly conservative sometimes.
Although COMPBRN III is better than previous versions, it still has some inconsistencies (At-tachmentC,Page36). The modified version of COMPBRN III has eliminat-ed several problems associated with the previous versions.
Mr. Nowlen discussed briefly the differences in fire damage times calculated by using different versions of the COMPBRN codes (Attachment C, Pages 37 and 38).
i Auxiliary Systems Meeting Minutes March 9, 1988 Sei'smic/ Fire Interactions - Mr. Bohn, SNL Mr. Bohn discussed the potential seismic / fire interaction mechanisms (Attachment C, Pages 39 and 40).
Some of the mechanisms are:
- Earthquakes can cause fire initiators
- Earthquakes can damage fire suppression systems
- Earthquakes can inadvertently actuate fire suppression systems and actuate many fire alarms Mr. Bohn stated that there are some evidences that earthquakes have caused fires in non-nuclear facilities.
However, there is no evidence that fires have been caused by earthquakes in nuclear power plants.
Available data on damage to fire suppression systems due to earthquakes is very limited.
There is virtually no data available on damage caused by earthquakes to Halon or C0 systems.
2 Mr. Bohn discussed briefly the conclusions of the Fire Risk Scoping Study on this issue and the measures that could be taken to reduce the seismic / fire interactions (Attachment C, Pages 41 and 42).
Mr. Ebersole commented that fire protection systems are vulnerable to the vibration caused by seismic activities.
During his recent visit of the Sequoyah plant, he was given to understand that if there is a fire in the diesel generator room, it will be put out by the CO syrtem.
2 They seal the generator room in order to buildup C0 to suppre;s the 2
fire.
Since the diesel generators are air-cooled, once the room is sealed, there will not be enough air to cool the diesel engines.
For a while, the cooling could be achieved by the recirculation of air inside the room.
However, af ter a certain time, inside room air temperature will be so high and will be ineffective to provide proper cooling.
Under such circumstance, the generator may not survive.
In his opinion, 4
Auxiliary Systems Meeting Minutes March 9, 1988 this is a poor design practice.
Mr. Bohn responded that probably this concern of Mr. Ebersole should be pointed out in the results of the Scoping Study.
PEER-REVIEW COMMENTS - Mr. J. FLACK, RES Mr. Flack discussed briefly the peer-review comments on the results of the Fire Risk Scoping Study (Attachment C, Pages 43-49).
Some of the comments came out of the peer-review process are as follows:
- Research should find pitfalls so utilities know what to do to lower risk.
- Because of very limited nuclear experience, it would be difficult to draw realistic conclusions on certain fire risk issues.
- It is questionable whether future research will be able to narrow urcertainties much further.
' Scoping Study has a pessimistic view point which nay be misinter-preted and result in unnecessary backfits.
- Since the fire at Browns Ferry plant, significant improvement has been made in the area of manual fire fighting.
- Fire detection times associated with control room fires do not seem realistic.
- Only a few plants use poorly designed systems for fire suppression; this should not be construed as an industry-wide practice.
' Differential pressure across the barrier is believed to have little impact on barrier integrity.
Auxiliary Systems Meeting Minutes March 9, 1988
' The statement by SNL that "Seismic / Fire interactions has been largely dismissed" is not accurate.
- Electric Power Research Institute (EPRI) is about to issue a request for proposal to study the effects of fire suppression systems on components and systems.
Mr. Michelson asked whether EPRI study will include all types of fire suppression systems, water, C0, and Halon. Mr. Kaminski, Wisconsin 2
Electric Power Company, stated that depending on the budget, EPRI plans to look at all systems.
Stating that it seems that EPRI study is not intended to look at the effectiveness of various fire suppression systems in putting out fires, Mr. Wylie comented that some research on this matter should be per-formed. Mr. Nowlen responded that there has been a lot of work done to determine the effectiveness of various fire suppression systems.
There are still some questions on the effectiveness of these systems to put out cabinet fires.
Other than that, he believes that enough information is available on this matter.
Mr. Wylie comented that since source of fire is an important factor, it should be factored into the Scoping Study.
Mr. Michelson comented that although there is no evidence to show that earthquakes have caused damage to safety-related equipment, the industry should not consider this as a non-problem.
They should take initiatives to show that the design is adequate to cope with this problem.
ORAL STATEMENTS BY MR. KAMINSKI Mr. Kaminski from Wisconsin Electric Power Company stated that fire protection issues have been or are being considered in significant
Auxiliary Systems Meeting Minutes 30-March 9, 1988 detail by utilities.
Some coments have already been provided by industry on the Fire Risk Scoping Study.
They plan to provide addi-tional coments in the near future.
SUBCOMMITTEE REMARKS Stating that at the July 1987 Auxiliary Systems Subcomittee meeting it was mentioned that an integrated report on fire research will be made available by the end of 1987, Dr. Moeller asked what is the status of that report.
Mr. Nowlen responded that the preparation of such a report has been delayed since they have to concentrate on completing the Fire Risk Scoping Study.
Mr. Wylie suggested that the RES Staff provide the Subcomittee with copies of'the additional peer review coments.
Mr. Flack agreed to do so.
Mr. Michelson asked after receiving a final report on the Fire Risk Scoping Study what the Staff plans to do.
Mr. King, RES, responded that after receiving the final report from SNL, the recomendations will be evaluated to determine whether:
' Some issues raised by SNL could be resolved through the Generic Items process.
- Further research is needed to obtain additional information on I
uncertainties associated with certain issues.
- Appendix R requirements are adequately complied with.
If found otherwise, it will be brought to the attention of NRR for action.
l
- - + -
4 y--
w e-w
i l
Auxiliary Systems Meeting Minutes March 9, 1988 Mr. Michelson stated that the Staff and SNL should be prepared to give a presentation to the full Committee during the April or May ACRS meeting.
He said that he would prepare a draft report on this matter and submit to the full Committee for consideration.
Mr. Michelson thanked all participants and adjourned the meeting at 4:10 p.m.
NOTE:
Additional meeting details can be obtained from a transcript of this meeting available in the NRC Public Document Room, 1717 H Street, N.W., Washington, D.C., or can be purchased i
from Heritage Reporting Corporation, 1220 L Street, N.W.,
)
Washington, D.C. 20555, (202) 628-4888.
l
LIST OF DOCUMENTS SUBMITTED TO THE AUXILIARY SYSTEMS SUBCOMMITTEE MARCH 9,1988 1.
Presentation Schedule.
2.
ACRS report to the Commission on the Fire Risk Scoping Study, dated August 10, 1987.
3.
Draft NUREG/CR-5088, "Fire Risk Scoping Study:
Current Perception of Unaddressed Fire Risk Issues," prepared by SNL, dated January 1988.
4.
Minutes of the July 23, 1987 Auxiliary Systems Subcommittee meeting.
5.
Presentation Materials provided during the meeting, i
l l
ATTACHMENT A v,-
e r--,
--w..
,w s,-e
,w y--
,~a n
-,w e
,w-
.,w
~
<~r- - -
e
1 TENTATIVE PRESENTATION SCHEDULE ACRS SUBC0FNITTEE MEETING ON THE AUXILIARY SYSTEMS MAPCH 9, 1988 ROOM 1046, 1717 H ST., N.W.
WASHIt.GTON, D.C.
ACRS CONTACT:
Sam Duraiswamy 202-634-3267 NOTE:
- Presentation Time should not exceed 50% of the Total Time allocated for a specific item.
The remaining 50% of the time is reserved for the Subcomittee questions and d6swers by the Staff or its contractors.
- Number of copies of the presentation materials to be submitted to the Subcomittee:
25 copies.
]
TOTAL PRESENTATICN ITEM PRESENTER TIME ACTUAL. TIME 15 min 8:30 - 8:45 am 1.
EXECUTIVE SESSION 2.
FIRE PROTECTION OF FRENCH Arthur Busiik 30 min 8:45 - 9:15 am 900 MWe N?Ps (CLOSED)
NRC/RES 3.
FIRd RISK RESEARCH John Flack 15 min 9:15 - 9:30 am OBJECTIVES NRC/RES 4.
FIRE RISK SCOPING STUDY Mike Bohn 15 min 9:30 - 9:45 am Sandia National Introduction and Over-Lab (SNL) view of the Tasks Performed 5.
RESULTS AND CONCLUSIONS John Lambright 30 min 9:45 - 10:15 am FROM TASXS 1 & 2 (SNL)l (a)
Evaluation of Fire Risk in Four Re-quantified PRAs (b) Sources of Uncer-tainties
- BREAK ***
15 min 10:15 - 10:30 am A rrA CH8e m-8 B-1
Auxiliary Systems Schedule March 9, 1988
-TOTAL PRESENTATION ITEM PRESENTER TIME ACTUAL TIME 6.
IDENTIFICATION OF FIRE Steve Nowlen 30 min 10:30 - 11:00 am RISK ISSUES - TASK 3 (SNL)
(a) Methods Used to Identify Issues (b) Results 7.
RESULTS AND CONCLUSIONS Mike Bohn 60 min 11:00 - 12:00 pm FROM TASKS 4 & 5 Steve Nowlen (SNL)
(a) Control System Interaction (b) Seismic / Fire Interaction (c) Manual Fire Fight-ing Effectiveness e/
l
- LUNCH ***
60 min 12:00 -
1:00 pm (d) Total Environment 60 min 1:00 - 2:00 pm Equipment Survival 1
(e) Adeo.uacy of Fire Barriers (f) Adeauacy of Analytical Tools 8.
PEER REVIEW COMMENTS John Flack 30 min 2:00 - 2:30 pm (NRC/RES) 9.
RECOMMENDATIONS FOR Mike Bohn 45 min 2:30 - 3:15 pm FOLLOW-ON EFFORTS (SN!. )
10.
SUBCOMMITTEE REMARKS 15 min 3:15 - 3:30 pm
- ADJOURN ****
3:30 pm b -2s
l i
}
I CONCLUSIONS FROM FIRE PRA REQUANTIFICATIONS
)
i
- LEVEL OF FIRE-INDUCED CORE DAMAGE FREQUENCY IS HIGH EVEN WITH APPENDIX R MODIFICATIONS 4
1 i
PRA ORIGINAL REQUANTIFIED WITH APPENDIX R MODS LIMERICKa 2.3E-5 1.6E-4 5.9E-5 i
INDIAN POINTS 6.5E-5 2.0E-4 8.8E-6 OCONEEA 1.3E-5 2.0E-5 ALREADY INCLUDED l
SEABROOKs 2.1E-5 4.6E-5 ALREADY INCLUDED
\\
A POINT ESTIMATE e
MEAN VALUE i
i l
l Arraegr.est C C:- I
\\
l 1
i
RECOMMENDATIONS TO IMPROVE FIRE RISK PERSPECTIVE
~
l
- DEVELOP DEFENSIBLE ESTIMATES OF FIRE OCCURRENCE FREQUENCY l
UNCERTAINTIES REFLECTING PLANT-TO-PLANT VARIATIONS (HOUSEKEEPING, FIRE WATCH PRACTICES, ETC.)
- DEVELOP DEFENSIBLE GUIDELINES FOR PARTITIONING AND FIRE SEVERITY j
FACTORS i
l
- VALIDATE COMPBRN AND MORE ADVANCED FIRE CODE AGAINST ACTUAL FIRE i
l ENCLOSURE TEST DATA l
i
- UPDATE AND MAINTAIN FIRE DATA BASES (OCCURRENCE, SUPPRESSION AND 1
l COMPONENT DAMAGE THRESH 0LD DATA) i
- PERFORM FIRE PRAS AT SEVERAL PLANTS - INCLUDING PLANT WALKDOWN AND THE "UNADDRESSED" ISSUES - TO GET A TRUE PERSPECTIVE OF THE FIRE RISK C-2,
I 1
POTENTIAL IMPACT OF THE "UNADDRESSED" FIRE ISSUES WAS ASSESSED l
l POTENTIAL j
ISSUE INCREASE IN CM APPLICABILITY 1
MANUAL FIRE FIGHTING EFFECTIVENESS 0(10)
GENERIC CONTROL SYSTEMS INTERACTION 0(10)
GENERIC k
TOTAL ENVIRONMENT SURVIVABILITY i
i SMOKE CONTROL
?
PLANT SPECIFIC
- i l
INADVERTENT SUPPRESSION
?
PLANT SPECIFIC
- i j
COLLATERAL DAMAGE SMALL GENERIC FIRE CODE ADEQUACY 0(10)
~
GENERIC
- BARRIER EFFECTIVENESS 0(10)
GENERIC SEISMIC FIRE INTERVENTION SMALL GENERIC-COULD ELIMINATE BY WALKDOWN
- RELATED TO FIRE FIGHTING EFFECTIVENESS C-3 l
l RECOMENDATIONS TO IMPROVE MANUAL FIRE FIGHTING EFFECTIVENESS j.
- TEST CABINET / CABLE ROUTING COMBINATIONS TO ASSESS FIRE SPREAD TIMES (RESEARCH)
]
- DEVELOP GUIDELINES FOR REQUIRED RESPONSE AND EXTINGUISHMENT TIME FOR i
j VARIOUS EQUIPMENT TYPES (RESEARCH) i
- EVALUATE FIRE BRIGADE EFFECTIVENESS VIA FIRE SIMULATIONS AT SEVERAL i
PLANTS TO ASSESS
- OPERATOR VS FIRE FIGHTER SYSTEMS AWARENESS
- TIME TO RESPOND
- ROOM ACCESSIBILITY (KEYS, ETC.)
(RESEARCH) i
- PLANT SPECIFIC MAPPING OF FIRE AREAS VERSUS CRITICAL RESPONSE TIME (UTILITY) l C-f
'i j
j PAST PRA FIRE PROPAGATION MODELING UNCERTAINTIES
- COMPBRN I USED IN ALL FOUR ORIGINAL PRAS l
i
- PARAMETER UNCERTAINTIES i
i
- QUANTIFIED USING LATIN HYPERCUBE SAMPLING FOR INPUT VECTORS
- MODELING UNCERTAINTIES
- PL&G PRAS USED AN ASSUMED CORRECTION FACTOR MULTIPLIER l
j OF 2 ON DAMAGE TIME (ASSUMED LOGNORMAL) l i
j
- REGARDED AS OVERLY CONSERVATIVE BECAUSE THE ORIGINAL
)
VERSION NEGLECTED HEAT LOSSES FROM FUEL ELEMENTS j
C-5 i
l FIRE SUPPRESSION UNCERTAINTIES l
l
- ENGINEERING JUDGEMENT IS USED TO CONSTRUCT SUPPRESSION MODEL BASED ON LIMITED DATA
- ACTUAL PLANT DETECTION AND SUPPRESSION FEATURES ARE CURSORILY l
INCLUDED IN CONSTRUCTION OF SUPPRESSION MODEL i
l l
i
- THE POTENTIAL RISK FROM MISAPPLICATION OF SUPPRESSION AGENTS IS IGNORED l
l c-4 l
1 GENERAL AREAS NOT RE-EVALUATED WHICH COULD HAVE A MAJOR IMPACT ON CORE DAMAGE FREQUENCY i
- INITIATING EVENT FREQUENCY CALCULATION METHOD
- PARTITIONING OF FIRE FREQUENCY FROM A lARGE BUILDING TO A FIRE AREA WITHIN THE BUILDING
- INITIATING EVENT DEFINITION (TRANSIENT VS. LOCAS)
- SEVERITY RATIO
~
- AREA RATIO WITHIN FIRE ZGNE l
- MANUAL VS. AUTOMATIC SUPPRESSION c-7 i
9 l
FIRE INITIATED CORE DAMAGE FREQUENCY (YR-U i
i i
AS REPORTED-REQUANTIFIED 4
l i
LINERICKa 2.3 E-5 1.6 E-4 i
i INDIAN POINT # 2e 6.5 E-5 2.0 E-4 l
OCONEEa 1.3 E-5 2.0 E-5 l
i l
SEABROOKs 2.1 E-5 4.6 E-5 1
1 A POINT ESTIMATE sMBN l
1 I
i i
i l
i C-8 i
I
r - ----- -- - - - - a
REQUANTIFIED FIRE INITIATED CORE DAMAGE FREQUENCY (YR-1) i 1
WITHOUT APPENDIX R WITH APPENDIX R MODIFICATION MODIFICATION l
I LIMERICK 1.6 E-4 5.9 E-5 f'
]
INDIAN POINT # 2 2.0 E-4 8.8 E-6 i
l OCONEEa 2.0 E-5 i
r i
l SEABROOKa 4.6 E-5 i
[
A APPENDIX R HOGIFICATIONS COMPLETED BEFORF C M LETION OF PRA C-7 l
l 1~-
WIDE VARIABILITY IN FIRE BRIGADE TRAINING, ORGANIZATION AND SIZE MINIMUM NORM MAXIMUM FULL-TIME l
ORGANIZATION PART-TIME BRIGADE PART-TIME BRIGADE FIRE DEPARTMENT I
l SIZE 30 78 218 l
l CLAssnoou 6 hrs.
26 hrs.
400 Hus.
l TRAINING LIVE FIRE 1 Hn.
9.4 hrs.
28 Has.
TRAINING l
l C-/o i
!j EFFECT OF FIRE BRIGADE RESPONSE AND EXTINGUISHMENT 1
TIME ON CORE DAMAGE FREQUENCY 3
)
CORE DAMAGE FREQUENCY (YR-1)
]
RESPONSE AND EXTINGUISHMENT TIME i
PLANT / AREA Mm:4uM AVERAGE MAXIMUM OnIGIwat FRA VALUE INDIAN PoInr #2 SWITCHGEAR Room 2.2E-6 3.0E-5 6.5E-5 5.6E-5(MEAN) i LInERICK SAFEGUARDS ACCESS l
AREA 1.4E-6 8.5E-6 3.8E-5 6.0E-6 (POINT j
EST.)
I l
SEABROOK l
CasLE SPnEADING 8.6E-7 4.9E-6 8.9E-6 4.1E-6(MEAN) j Roon i
c-l
MISDIRECTED SUPPRESSION EFFORT QUANTIFICATION APPROACH FOR THE TASK 1 & 2 FIRE SCENARIOS ASSUME COMPONENTS, SUSCEPc i
TIBLE TO SPRAY,
- FLOODING, OR TEMPERATURE WITHIN AN AREA ARE FAILED BY SUPPRESSION EFFORTS REQUANTIFY ANY SCENARIOS WHERE SUFFICIENT INFORMATION EXISTS 1
ON FIRE AREA SPECIFICS l
IF ANY AREAS TURN OUT TO BE PROBABILISTICALLY IMPORTANT, ESTIMATE THE PROBABILITY OF SPRAYING THE WRONG EQUIPMENT AND l
COMBINE THIS WITH FRAGILITY INFORMATION TO DETERMINE THE CONDITIONAL PROBABILITY OF SUPPRESSION-INDUCED FAILURE l
~
l C -/h
RESULTS OF MISDIRECTED SUPPRESSION EFFORTS
- SEABROOK, OCONEE, AND INDIAN POINT PLANTS ELIMINATED FROM l
CONSIDERATION BASED ON NOT ENOUGH DETAILED INFORMATION AVAILABLE SCREENING ANALYSIS FOR LIMERICK DEMONSTRATED THE POTENTIAL j
IMPORTANCE OF THIS ISSUE l
l FOuR LIMERICK FIRE AREAS CHOSEN FOR A MORE DETAILED ANALYSI9 l
- ALL WERE ELIMINATED AFTER FURTHER ANALYSIS BASED ON THE j
FOLLOWING CONSIDERATIONS i
- CONTAINS ONLY CO2 HOSE REELS AND PORTABLE EXTINGUISHER 2 i,
(13KV SWITCHGEAR ROOM) i
- HAS AN OPEN EQUIPMENT HATCHWAY ( 200 FT2) WHICH WOULD i
MITIGATE THE EFFECT OF SMOKE BUILDUP (SAFEGUARD ACCESS
{
- AREA, CRD HYDRAULIC EQUIPMENT AREA, GENERAL EQUIPMENT l
AREAJ i
1
- METHODOLOGY DEVELOPED WHICH CAN BE APPLIED TO PLANTS WHERE MORE DETAILED INFORMATION EXISTS J
c-ns i
REMOTE SHUTDOWN ELECTRICAL INDEPENDENCE ANALYSIS METHODOLOGY
- PERFORM A SCREENING ANALYSIS TO IDENTIFY INTERACTIONS WHICH COMPROMISE INDEPENDENCE
- DETERMINE PLANT CHARACTERISTICS WHICH MAKE THE IDENTIFIED INTERACTIONS PROBABLE AT OTHER PLANTS
- IDENTIFY METHODS TO AVOID INTERACTIONS
- ASSESS EFFECTS ON CORE MELT FREQUENCY IF ANY OF THE IDENTIQ PLANTS HAVE PREVIOUSLY COMPLETED FIRE PRAS c-iq
Evaluation of the Potential Impact of NCR/RSP Dj Interactions of Three Plants C
l CDF due Plant CDF due to to scenario Overall original Beenario original fire with assumed CDF due to Descrip' tion scenario /yr interactions /yr fires /yr Indian Point
- 1. Fire in CSR 1.9 a 10 6 (1) 7.6 a 10 5 2.0 a 10 4 morthern wall (before Appendix R (ii) 3.3 a 10 6 modifications)
- 2. Fire in control 5.7 x 10 7 (1) 2.3 a 10 5 room (ii) 9.9 m 10 7 Baabrook
- 1. Fire in CSR 3.6 a 10 6 (1) 1.6 a 10 5 1eading to less (ii) 3.6 a 10 6 PCC 2.5 a 10 5 (Appendix R status
- 2. Fire in control 9.1 s 10 6 (1) 4.0 a 10 5 unknown) room leading to (ii) 9.1 a 10 6 1ess of Pcc Limerick
- 1. Fire in C$t 2.0 s 10*7 (1) 1.4 x 10 4 spreading to (ii) 1.1 a 10 5 protected cables 4.4 x 10 5 (before Appendix R
- 2. Fire in control 4.3 z 10*8 (1) 4.5 a 10 5 gtgie ggens) room requiring (ii) 3.6 a 10 6 evacuation Bete: Case (1) all equipeent that een be lost due to possible interaction is test.
(ii) all equipment is test with the exception of one method of bot shutdown.
4 e.
G c-is 4
,-my
.--_,,--w..-..-~,----~,--,m-,,ew---.-r--
_LASALLE SPURIOUS ACTUATION ANALYSIS METHODOLOGY _
ASSUME ELECTRICAL INDEPENDENCE OF REMOTE SHUTDOWN PANEL FROM j
THE CONTROL ROOM BASED ON AN IEAL PLANT SPECIFIC ANALYSIS IDENTIFY ALL CONTROL AND ACTUATION FUNCTIONS IN THE CONTROL RO i
i i
)
MAP THOSE FUNCTIONS TO CONTROL ROOM PANELS TO DETERMINE WHICH PANELS ARE MOST CRITICAL TO SAFE SHUTDOWN l
ASSESS THE PROBABILITY OF FIRE SPREAD TO ADJACENT PANELS j
QUANTIFY TWO TRANSIENT SEQUENCIES UTILIZING THE SETS COMPUTER CODE FOR EACH IDENTIFIED PANEL
- ALLOW FOR RECOVERY OF RANDOM FAILURES (I.E.,
FAILURES NOT RELATED TO THE FIRE) IF SUCH ACTIONS TAKE PLACE OUTSIDE THE CONTROL ROOM C - ib i
~
IDENTIFY CONTROL SYSTEMS INTERACTIONS WHICH COULD RENDER A l
COMPONENT OR SYSTEM CGNTROLLED FROM THE. REMOTE SHUTDOWN PANEL i
l
- ALLOW FOR RECOVERY ACTIONS FROM THE REMOTE SHUTDOWN PANEL 1
l
- I7 g
~
RESULTS OF THE LASALLE ANALYSIS
- TWO CONTROL ROOM CABINETS WErtE IDENTIFIED AS THE MOST CRITICAL!
TO SAFE SHUTDOWN l
l j
- EMERGENCY CORE COOLING SYSTEM PANEL (2H13P601)
}
j
- ELECTRICAL DISTRIBUTION PANEL (2PM01J)
- AS A RESULT OF VISUAL INSPECTION BOTH PANELS WERE FURTHER SUB-'
DIVIDED DUE TO STEEL PARTITIONS APPROXIMATELY ONE-EIGHTH THICK!
WITH NO PENETRATIONS l
l l
l c-;g
/
l i
RESULTS OF THE LASALLE ANALYSIS (CONT'D) 1 i
AFTER THE COMPUTER SCREENING ANALYSIS ONLY ONE TRANSIENT e
SEQUENCE FOR ONE SUBSECTION OF ELECTRICAL DISTRIBUTION PANEL 1
SURVIVED
- SEQUEdCE 8 = NFW RCIC HPC$ lEEI CDS LPCS RHR
- $EQUENCE 8 CORE DAMAGE FREQUENCY 8.1E-6/YR
- TOTAL CORE DAMAGE FREQUEHCY FROM ALL INTERNAL EVENT I
INITIATORS ~ 1.0 E-5/YR HPCS, ADS NOT FAILED BY THIS FIRE
- RCIC, MFW, CDS FAILED BY FIRE INDUCED INTERACTION RCIC FIRE INDUCED FAILURES RECOVERABLE FROM THE REMOTE PANEL C-/9 a
i LASALLE CONTROL SYSTEM INTERACTIONS i
FIRE INDUCED INTERACTION SYSTEM (S)
AFFECTED l
i 1.
UNDERVOLTAGE PROTECTION RELAYS
2.
LOSS OF RCIC ROOM COOLING, RCIC INBOARD STEAM ISOLATION VALVE CLOSES 1
3.
RCIC ISOLATION LOGIC MOMENTARY RCIC i
POWER FAILURE (SNEAK CIRCUIT) 4.
LOSS OF POWER TO A NONSAFETY BUS MAIN FEEDWATER, i
CONDENSATE C-20
t Environments Of Potential Concern:
l With An Actual Fire / Fire Suppression:
Heat (Both High-and Low-Level) j Smoke Deposition l
High Humidity Water Spray l
Hose Stream Impact i
High Static Charges l
Locally Intense Cooling Combined Effects l
Spurious Suppression Events:
Wetting i
Intense Cooling l
High Static Charges j
1 i
i
)
i l
C-2,1 i
j History Of Spurious Suppression System Actuation Events Spurious Actuations Occurring At Frequency Similar To That For i
Actual Fire Events l
l Various Factors Causing Events
\\
Event Reports Typically Very Complete In Identifying Equipment Damaged And Mechanism Of Damage Many Types Of Equipment And Many Systems Have Been l
Compromised Insights Applicable To Fire Environments Available From Spurious Suppression Events Fire Suppression Effects On Components 3
Identification Of Vulnerable Equipment I
Design And Installation Factors Contributing To Problems C-2 L i
f i
2@
u - tbte that 1988 and 1987 data is incomplete l
15 i
l 10 i
i 5
~
O 80 81 82 88 84 85 86 87 Reported Spurious Suppression Events by o
C Calender Year for U.
S.
NPPs c-23 i
)
i 1
~
$ -Power Oper.
@ ~ Pre-op Test. $ ~Refue1ing i
25
~
l 20
)
~
l 15 E
E i
10 5
~
m" O
69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 g
1 h
l Histogram of Reported NPP Fire Events by Calender Year c-29
ru to s
en Q
csa ca e
e e
N liIi!IfII!IIII!Iii!!tiii!IIII!IIII IiIi!IIIi IfIi c+
j c8 Dry Piae Water Deluge (42) 2 h % \\\\ % % % % \\N a
o+
o Wet Pipe Sprinklers (4) on C
u, o
a T
C Wet Pipe Water Curtain (1) co m
0 1
M Ce o
p 3
Water Plumbing Leaks (6) a en a
x a
e ya h\\\\N Cardox (9) a m
~
a h\\M Halon (8) o rn
-<m E_
"+
Cardox, Halon, and Deluge (1) m t3"3 X
h 4
C -2 5
,y,-_,m,,.y-,ym-9,---,y,
--.9 m--g.--
,,,,-,9,,,_,,,,,..%y---
-,.q%ym, _,,y,e,__y
,-,,,,,,,,.f.,_7
-pwy,,
_...gm,.
5 G
E m
w iiiiiiiiiiiiiii m
o h % \\ % % % M ] Maint/ Testing E
n\\\\%%N Personne1 p
h%%\\\\' Leaky Fire Water Systems o
h%%
Steam / Humidity / Dust o
co
@DetectorFaults
@FPSControlFaults
[
@WeldingRetivity ag
[SmokeFromPlantEquip.
g h3 Other Identified Causes p
h g % % % y Unknawn Cause C
M CD 9
s c.u
i l
l l
Eauipment Demonstrated To Be Vulnerable J
l To Suppression Environments l
i l
Transformers Switchgear Batteries Motor Control Centers l
Pump Motors j
Motor Operated Valves i
Air Compressors LowVoltage Power Supplies Switches Relays 1
C-27 a
l 1
3
^
Systems Affected By Spurious Suppression Events I
HPCI l
Core Spray Non-Vital Busses Offsite Power Turbine Trips ESF SBGT CRD RPS Motor Trip l
Generator Lockout Access Control Plant Computer Hydrogen Recombiner
~
Main Feedwater Pump Trip l
LPSI C-28
e Sienificance Of Sourious Sunoression Events l
Most Serious' Events Involve LOOP, Induced Scrams, Induced Control l
Interactions 1
l Risk Significance Expected To Be Small Unless Multiple System
)
Involvement Postulated Frequency of Multiple System Impact low i
1 LOOPS and Scrams Dominated By Other Causes i
For Actuation Of Single Suppression System To Have Significant Risk Impact Requires:
Redundant Trains Housed In Area, Or A Flow Path For Suppressant To Redundant Equipment, Or Control Room Incidents (Water Leakage)
May Get Multiple System Involvement From Actuation Of Multiple Suppression Systems; Potential Causes:
Large Line Breaks Maintenance Activities I
Seismically Induced Actuations l
Fire Induced Actuations In Non-Fire Areas C-2'l
\\
i 4
The Relevance of Appendix R Fire Regulations i
To The Issue of Total Environment Equipment Survivai i
j
)
Appendix R Specifies That Certain Equipment Be Maintained "Free Of Fire Damage" l
Criteria Not Defined Environmental Concerns Not Defined l
Equipment Vulnerability Data Lacking I
Appendix R Specifies That Safety Related Equipment Be Protected l
From Fire Suppression Systems i
Ideal Not Met In Practice As Demonstrated By Experience Base j
Equipment Vulnerability Data Lacking i
i c-so j
i
~
4 l
Fire Induced Pressure is the Unconsidered Failure Mechanism e
i Ceble Penetration Seals Most Likely Point of Concern j
U.S. Standard Barrier Qualification Tests Utilize Conservative l
Standard Time-Temperature Profile (ASTM-E119) l Penetration Seals Fail if Outside Reaches Ignition Temperature i
Exposure Pressure Conditions Not Specified Typical Practice Induces Negative Pressure On Fire Side Enclosure Fires Induce Positive Enclosure Pressures j
l Standards Endorsed In All Countries Other Than The U.S. And l
Canada Specify Positive Pressure Exposure, ISO-3008, ISO-834 L
i Lack of Pressure Specification May Be Non-Conservative NRC/Sandia Testing Of Cable Penetration Seals l
UCB Testing Of Fire Doors (For Private Org.)
Recall That 1975 Brown's Ferry Fire Propagated Though Penetration Seal With Cracks l
A Screening Analysis Demonstrates PotentialImpact If Barriers Are In Fact Vulnerable c_3f l
I l
4 Significance Of Premature Barrier Failure i
Screening Values:For RMIEP/LaSalle,20 Area Pairs Identifi i
i Fire Frequency 1E-3 Barrier Failure 0.1 Required Area Randoms Higher Than IE-3Non-Each Area Pair Contributes At Least 1E-8 i
20 Pairs Contribute at Least 2E-7 To Overall Frequency Of Co Damage 1
Required To Fail Barrier) Could Get 2E-6 Or G i
i I
i1 I
i i
C-3h
4 Significance Of Premature Barrier Failure i
For NUREG-1150 Analysis Of Peach Bottom At Least 20 Adjacency I
Pairs Requiring No Randoms Exist Screen Using:
SE-3 Fire Initiator 0.1 Frequency Of Barrier Failure l
0.1 Area Ratio 0.1 Critical Equipment Damage Frequency I,
-l 1
Core Damage Frequency Contribution SE-6 Per Area Pair With At Least 20 Pairs Contribute On The Order Of IE-4 To Overall J
l Core Damage Frequency i'
If Assume 99% Reliability (Barrier Failure Frequency = 0.01) The All Estimates Down By Factor Of Ten Contribution Per Area Pair 6E-7 i
l
~
Overall Contribution Approximately 1E-5 i
c-33 i
+
Sienificance Of Premature Barrier Failure l
Barrier Element Random Failure Frequencies Have Been Estimated Based On Reports Of:
Fire Doors Left Open Fire Dampers Which Fail During Inspections /I'ests Penetration Seals Found Cracked Or Missing During Inspections Estimated Random Failure Probabilities (Per Reactor Year):
Doors: 7E-3
)
Dampers: 2.4E-3 Penetration Seals: 1E-3 l
i l
Assuming 1 Fire Door, And 10 Penetration Seals Per Interface The Random Barrier Failure Probability = 1.7E-2 l
If One Can Assure On The Order Of 99% Reliability Of Properly Installed Fire Barriers The Randoms Will Dominate l
.l l
j l
C-3/f
i P
Fire Barriers Are Very Important Protective Features l
l
. Separation By Passive Barriers Is Preferred Method Of Protection i.
i Due To Design Constraints Many Impor' ant Area Adjacencies Will l
Exist AC. Virtually All Plants 1
Need To Assure On The Order Of 99% Reliability In Order To l
Minimize Significance Information Notice 88-04 Expresses Concern Regarding Current l
j Implementation Of Barrier Qualification / Documentation Procedures (Somewhat Different But Clearly Related Concern) i Historical Review Of Both Nuclear And Non-Nuclear Fire Experience Base Could 'Macastrate Degree Of Barrier Reliability In U. S.
Applications l
Simple Plant Inspection And Documentation Procedures Could Assure
{
l MinimalImpact
\\
l s
l c 33 l
\\
4 k
ib m
w a
r w
e m
a d
E a
.o..
a z
zz m
w owe
-rz e
r e
e a
sm-cc - z a
w r
a r
umo o
e o
m
-za z
w u
w woa o
u
=
-vo z
o a
m w
rza m
o m
z-m e
w e
m u
z r
z u
m z
omz o
r u
w u
o ows a
w w>
r r
w m
w r
mme m
u m
=
o
=
c wz w
o e
a E
u mm-2 e
oma m
o m
w z
w woo m
a r
m a
m a
a su e
o m
z a
o
~uz m
m r
e m
w w
w u
o a
r
-w-
-m a
w e
m m
z m
m u
z e
m
=
r z
z z
a o
z zom e
muo er e
c w
z
>z
=
m r
o m
m co>
mr r
z x
2 r
rre w~
r u
r w
o I
mm s
s z
e e
_J uww zo e
e o
m m
z r
m om oa w
M o
o a
w ud z
r e
w o
we m
z m
d w
=
wra a
a m
o w
w a
w m
mu zu m
a r
oz-m
~
c z
a m
w rom ao r
m o
e m
a z-cc cc r
- e m
mer me a
r
=
o 0
uma m
a a
e u
w m
ruz mw e
e r
a z
r E
x a-o ou
=
=
m m
m m
ruz gw M
r zo o
=
S g
Wga r
u z
W wr o
soE
(-
u cr O
O O
O O
O O
4
j i
F)
()
Table 2 COMPBRN Fire Damage Times (minutes)
Postulated Modified Pool Fire COMPBRN COMPBRN COMPBRN Scenario Diameter (m)
I*
III III Indian Point 2 Switchgear Room 0.30 24.2
=
=
0.91 NA 14 10 i
Electrical Tunnel Aisle Fire 0.61 15.8
=
13 Tray Fire **
14.4 12 3-6 Cable Spreading Room 0,30 22.6
=
=
0.91 NA 50 7
oconee Cable Shaft 0.30 6.7
=
3 Auxiliary Bldg Aisle Fire 0.30 9.3
=
=
0.61 NA 24 3
Tray Fire 5.4 28 3
Seabrook FCC Pump area Aisle Fire 0.61 16.1
=
=
0.91 NA 6
=
Tray Fire NA 16 3
Cable Spreading Room 0.30 16.1 35 7
0.61 48.4 35 6
Limerick All Locations Aisle Fira 0.61 10
=
3 Tray. Fire 10
=
=
- -Mean Time multiplied by error factor, with the exception of Limerick.
Taken from utility PRA.
- -Tray fires represent self-ignited cable tray fires.
NA-Not analyzed in utility PRA.
j 62-3 7 n-to
.~'
l RISK SIGNIFICANCE
~
(
j FIRE DAMAGE TIMES FROM COMPBRN CAN BE COUPLED TO l
EXPERIENTIAL FIRE SUPPRESSION DATR TO YIELD PROBABILITIES OR FREQUENCIES OF SUPPRESSING R FIRE BEFORE DAMAGE OCCURS MODIFIED COMPBRN III COMPBRN III lL l
LARGEST OBSERVED DIFFERENCE IN INFINITE TIME 3 MINUTES EFFECT i
DAMAGE TIME FOR ONE SCENARIO ON RISK i
FACTOR CORRESPONDING PROBABILITY OF 1.0
.05 0F 20 l
SUPPRESSION BEFORE DRMRGE OCCURS i
LARGE UNCERTRINTIES AND SENSITIVITIES IN RLL VERSIONS 0F COMPBRN 1
f 1
h c-38
~
POTENTIAL SEISMIC / FIRE INTERACTION MECHANISMS EARTHQUAKES CAN CAUSE FIRE INITIATORS A)
CABLES POLLED CAUSE SHORTS to DCCUR B)
FLAMMABLE LIQUIDS SPILLED OR GASES RELEASED CAN BE IGNITED BY SPARKS (SHORTS OR BROKEN BULBS)
EARTHQUAKES CAN DAMAGE FIRE SUPPRESSION SYSTEMS A)
CAST IRON FIRE MAINS ARE VULNERABLE IN OLDER PLANTS B)
SPRINKLER HEADS CAN BE DAMAGED BY IMPACT C)
FIRE PUMPS (ESPECIALLY ON SHOCK ABSORBER MOUNTINGS)
CAN BE DAMAGED BY SHAKING D)
HALON OR CO2 TANKS OFTEN SKID MOUNTED AND UNANCHORED c-57
e 1
POTENTIAL SEISMIC / FIRE INTERACTION MECHANISMS (CONT'D) i EARTHQUAKES CAN INADVERTENTLY ACTUATE FIRE SUPPRESSION l
4 l
SYSTEMS AND ACTIVATE MANY FIRE ALARMS i
l A)
SMOKE DETECTORS CAN BE ACTIVATED BY DUST l
B)
SUPPRESSION SYSTEMS CAN BE ACTIVATED BY RELAY CHATTER THE POST-EARlHOUAKE ENVIRONMENT CAN HAMPER FIRE RECOGNITION e
i i
AND FIRE FIGHTING EFFECTIVENESS i
A)
MANY SPURIOUS FIRE ALARMS WILL REQUIRE OPERATOR ATTENTION i
B)
AFTERSHOCKS, LOSP AND EMERGENCY LIGHTING WILL HAMPER MANUAL FIREFIGHTING 4
l c-4o
}
THE BNL STUDY CONCLUDED:
i i
"AUTOMATIC SPRINKLER SYSTEMS (IF INSTALLED IN ACCORDANCE l
)
WITH NATIONALLY RECOGNIZED STANDARDS...) WILL NOT SUFFER j
SIGNIFICANT IMPAIRMENT UNLESS... MAJOR STRUCTURAL COLLAPSE I
OR FALLING WALLS.
i j
j FIRE MAINS WITHSTAND EARTHOUAKE FORCES REASONABLY WELL, BUT...
BREAKAGE OF CAST IRON FIRE MAINS... AND PULLING APART OF SLIP l
JOINTS CAN BE ANTICIPATED.
i i
TEMPORARY IMPAIRMENTS OF FIRE ALARM SERVICE DUE TO i
j ACTUATION OF MANY TRANSMITTERS SIMULTANEOUSLY...
4 I
CAN BE ANTICIPATED."
O l
l C-4/
R
i SEISMIC / FIRE INTERACTIONS EASIER TO FIX THAN RECOMMEND A PLANT WALKDOWN LOOKING FOR:
CO, HALON TANKS ON SKIDS, UNANCHORED 2
l t
ACTUATION SYSTEMS WITH RELAYS AND LOCKING C l
i SM0KE (IONIZATION, PHOTOELECTRIC) ALARMS WITH NO HEA OR FLAME ALARM BACKUP i
FIRE PUMPS ON VIBRATION ISOLATION MOUN i
CAST IRON FIRE MAINS OR SEISMICALLY WEA I
SPRINKLER HEADS DROPPED THROUGH SUSPENDED C INADEQUATE SLACK IN CABLES ENTERING CA ANCHORED CABINETS
.UNANCHORED 02 BOTTLES OR H2 TANKS POTENTIAL IMPACT BETWEEN SPRINKLER HEADS PIPES, ETC.
4 i
i c-4zi
A
~
GENERAL COMMENT
S l
l i
Research should find the pitf alls so utilities know what to do in order to lower risk.
4 i
- Managernent is concerned with what they are getting for their i
money (cost effective sa f ety).
I
- Will we see a real reduction in risk or just a fine tuning of the u nc ertaintles?
Difficult to extrapolate test data to real world situations; actual experience is a beffer Indicator.
i 5
Questionable whether limited nuclear experience is sufficient to draw conclusions.
c-43
GENERAL COMMENTS (Contin u ed) i Co n cern over la rg e u n certainties in bottom
- line, especially when co n sid e ring the bounding a n a lysis.
l Questionable whether f uture resea rc h will be a ble to na rrow u n certain ties much f u rthe r.
Study has a
pessimistic viewpoint which may be misinterpreted and result in u nnecessary backfits.
4 1
~
MANUAL FIRE FIGHTING COMMENTS l
- Since the Brow ns Ferry fire t h ere has been significani i
im prove m en t s in training and equipment.
Disag reed with dediccied vs.
non-dedicated fire briga d e conclusions.
Response times used in the report we re rea son a ble.
Regarding training de ficie n cies, need to expand on what was actually being pe rfo rm ed in the field.
4 l
- Proper smoke tower training can address the types of problem s which the brigad e may enco u nf er.
a C-45 l
i
a CONTROL SYSTEM INTERACTION COMMENTS l
- Operators should be gF/en more credit f or recovery.
- Systems interaction analyses are detailed and very tho rou g h.
- Cheaper to go in and physically separate systems rat h e r f
than trace down in te ra ctio n s.
4 Disagreed on the time to detect control room panel fires,
- ventilation fans would blow smoke out
)
- cabinei sm o ke d etecto rs would trip
- corridors a re f requently travelled.
i i
c-+6 i
COMPONENT SURVIVABILITY COMMENTS
- Should not conclude that some p oo rly d esig ned systems,
represent industry wide p ro ciice.
Important to d if f e re ntiate between types of sytems when modeling suppression.
EPRI is about to issue a "Request For Proposal" to study the effects of fire suppression.
C-47
)
i
t o
9 BARRIER SURVIVABILITY COMMENTS 1
Based on p revious (non-nuclea r) experie nce, ba rrie rs
{
have dem o n strated excellent p e rf o rm a n c e.
Wall f u el loading could be im po rta n t.
4 l
Dif f eren tia:
p ress u re across the ba rrier is believed to have little impact on ba rrie r integ rity.
1 i
l Analysis is co nservative because the amount of l
co m b ustible material available is limited.
l Hose strea m iests and tim e-tem pe ratu re c u rves a re m o re conservative than some European tests.
3 C-4 8
I ET SEISMIC FI RE CO M M ENTS Results we re helpf ul to utilities, gave them something to look for at their plant.
The statement th a t "seismic / fire inte ra ctio n s has largely been dismissed" is n ot o cc u ra te.
Data is insufficient to identify this issue as sig nifica n t.
C-49
- -