ML20211N362
| ML20211N362 | |
| Person / Time | |
|---|---|
| Issue date: | 09/02/1999 |
| From: | Spector A NRC (Affiliation Not Assigned) |
| To: | NRC (Affiliation Not Assigned) |
| References | |
| NUDOCS 9909100172 | |
| Download: ML20211N362 (27) | |
Text
-
September 2, 1999 MEMORANDUM TO:
File FROM:
August K. Spector, Communication Task Leader Inspection Program Branch Division ofInspection Program (bri inal sifned by:)
aSagemen Office of Nuclear Reactor Regulation
SUBJECT:
PUBLIC MEETING REACTOR OVERSIGHT PROGRAM FIRE PROTECTION ISSUES MAY 24,1999 On May 24,1999, a public meeting was held between the NRC and the NEl to continue exchanging information on the reactor oversight program fire protection issues. The meeting agenda, a meeting summary, a list of attendees and a copy of written information exchanged at the meeting are attached.
Attachments: As stated
Contact:
August K. Spector 301-415-2140
)
/
Distribution:
h Central Files 0 4 H ~ l I] Q 'TT PUBLIC llPB R/F h
to O Y V\\ -
\\
,rms DOCUMENT NAME: a:\\S24mtgsum
- See previous concurrence.
To rGeeive a copy of this document, indicate in the box: "C" = Copy without enclosures "E" = Copy with enclosures "N" = No copy I
OFFICE IIPB:DIPM V \\llPB:DIPM
}
l l
l NAME AKSpector W ean '
DATE Of/ 1/ /99 07/'M
/
OFFICIAL RECORD COPY l
99o91oo172 99o9o2 ef PDR REVGP ERGNUMRC
l outro
,0 UNITED STATES l
l g
j NUCLEAR REGULATORY COMMISSION 2
WA8HINGTON, D.C. 2055!MX101 o
s,*****/
('
September 2, 1999 l
MEMORANDUM TO:
File FROM:
' August K. Spector, Communication Task Leader inspection Program Branch Division of Inspection Program Management Office of Nuclear Reactor Regulation
SUBJECT:
PUBLIC MEETING REACTOR OVERSIGHT PROGRAM FIRE PROTECTION ISSUES MAY 24,1999 On May 24,1999, a public meeting was held between the NRC and the NEl to continue exchanging information on the reactor oversight program fire protection issues. The meeting agenda, a meeting summary, a list of attendees and a copy of written information exchanged at the meeting are attached.
Attachments: As stated
Contact:
August K. Spector 301-415-2140 1
l
r:
AGENDA l
. MAY 24,1999 l
I 1.
NEl feedback on Baseline procedure I
2.
NRC discuss SDP process related to Fire protection and show examples l
t i
1 I
l
~
f MEETING
SUMMARY
. May 24,1999 The staff presented current drafts of the fire protection baseline inspection procedure and the fire protection risk significance screening methodology (FPRSSM).
~
- NEl provided the NRC with a set of " quicklook" comments on the baseline procedure. Since these " rough, unedited" comments were somewhat unc' ear, the staff agreed to provide clarification questions to NEl. This was accomplished or May 25,1999. On June 9,1999, NEl provided the staff with a set of " clarified" questions on the fire protection baseline procedure.
The staff intends to prepare responses to the second round of clarified industry questions and present them during a June 28,1999 meeting with NEl.
NEl expressed concern that licensees which meet older licensing criteria (older versions of i
NFPA codes for example) could find themselves trying to deal with high FPRSSM numbers in the absence of any actual non-compliances. The NRC response during the meeting was that, beyond the fact that the FPRSSM is just a screening methodology (the results of which are
. subject to both licensee and NRC refinement), the assessment process is separate from the enforcement process, and the Backfit Rule and its processes still apply within the new assessment and ovemight process.
NEl asked, for purposes c.f use in the FPRSSM risk significance determination process, how are past plant specific performance findings used? For example, if licensee corrective actions have removed the expectation of continued poor fire brigade performance, does the FPRSSM still use recent negative fire brigade findings for the screening calculation? The staff 3
expectation was that use of negative information, the cause of which has since been corrected, would depend on the time period for which any given risk calculation applies.
NEl stated that it is clear that the triennial team can usually develop the variety of dehnse-in-depth information required to apply the FPRSSM (i.e.', calculate a CDF value).
- However, NEl asked what a resident inspector is supposed to do with isolated findings regarding, for example, sprinklers, detectors, combustible control or fire barriers? That is, is there a threshold or trigger point for alerting regional management to his perceived need for more extensive followup regarding the licensee's fire protection program? The staff indicated I
-.that it ' ould provide guidance on this matter for the resident inspector within the Significance w
Determination Process.
NEl requested, and the Oversight Task Force endorsed, a blind " reproducibility" trial on the FPRSSM using selected sets of FPFI findings for which CDF values have already been developed. The intent was that, at the next fire protection baseline inspection procedure meeting, industry would present their FPRSSM CDF calculation results for comparison with the staff's CDF results. The staff provided NEl with the selected findings for industry trial application on June 14,1999.
c It was agreed that the next fire protection baseline inspection procedure meeting would be l conducted on June 28,1999. It is expected that shortly after that meeting a staff decision i
would be taken regarding whether to conduct three baseline pilot inspections, probably in the fall. Such a decision would indicate that both the baseline inspection procedure and FPRSSM risk significance determination process are at or near appropriate levels of development for pilot use.
i
-t b
r l
1 ATTENDEES Public Meeting May 24,1999 l
' NEl Fred Emerson Dave Modeen Steve Floyd Tom Houghton NUCLEAR REGULATORY COMMISSION
' Alan Madison August Spector Leon Whitney Peter Koltay Gareth Parry Nathan Siu Moni Dey Mark Rubin J.S. Hyslop Pat Madden Steve Mays OTHERS Kim Green, NUS Info Services Deann E. Raleigh, Bechtel Power M. Callahan, Self Philip Ohson, GAO i
l L
EXAMPLE O
% A s.
f,.m cres,aN *#
mted V
, gudtA
(,,t,g ec)
Ap,c,e,ume(F6)
=
I F = -2, SSD (drt) = -2 (1) AS = High; Fire Brigade = Medium; j
1 Hr. Barrier = High l
Pot. Risk Sign. = IF + SSD + AS + Brigade + FB l
l
= -2
-2 0
-0.5 0
1
= -4.5 (Yellow) l l
(2) Improve 1 hr. Fire Barrier, such that = Medium l
l Pot. Risk Sign. = -2
-2 0
-0.5
-0.5
= -5.0 (White) i l
l l --
e (3) Improve AS, such that = Medium IF SSD AS Brig.
~
Pot. Risk Sign = -2
-2
-0.75
-0.5
-0.5
= -5.75 (still White)
(4) Improve 1 Hr. Barrier = Low Pot. Risk Sign. = -2
-2
-0.75
-0.5
-1 i
=
-6.25 (Green) l
SUMMARY
IF = -2; SSD = -2 (1) AS = High; Fire Brig. = Med. ; Barrier = High Pot. Risk Sign. = Yellow (2) AS = High; Fire Brig. = Med.; Barrier = Med.
I Pot. Risk Sign. = White (3) AS = Med.; Fire Brig. = Med.; Barrier = Med.
Pot. Risk Sign. = White (4) AS = Med.; Fire Brig. = Med.; Barrier = Low Pot. Risk Sign. = Green
, 7, _..., _
f EXAMPLE O
y Asn f,. acw,86)at rn.sel V
AFLI y y y e. d rc \\
- AFW Tc 2-
, gum
(,,t,g ec)
As, e,.umc wa)
IF = -2, SSD (drt) = -2 (1) AS = High; Fire Brigade = Medium; 1 Hr. Barrier = High Pot. Risk Sign. = IF + SSD + AS + Brigade + FB
= -2
-2 0
-0.5 0
= -4.5 (Yellow)
(2) Improve 1 hr. Fire Barrier, such that = Medium Pot. Risk Sign. = -2
-2 0
-0.5
-0.5
= -5.0 (White) t
\\.
...4..
s (3) Improve AS, such that = Medium IF SSD AS Brig.
FB i
Pot. Risk Sign = -2
-2
-0. 7'5
-0.5
-0.5
= -5.75 (still White)
(4) Improve 1 Hr. Barrier = Low Pot. Risk Sign. = -2
-2
-0.75
-0.5
-1
=
-6.25 (Green) l
.. ~
SUMMARY
l IF = -2; SSD = -2 i
(1) AS = High; Fire Brig. = Med. ; Bstrier = High Pot. Risk Sign..= Yellow (2) AS = High; Fire Brig. = Med.; Barrier = Med.
Pot. Risk Sign. = White (3) AS = Med.; Fire Brig. = Med.; Barrier = Med.
Pot. Risk Sign. = White (4) AS = Med.; Fire Brig. = Med.; Barrier = Low Pot. Risk Sign. = Green
~
e, p
f
DRAFT STAKEHOLDER COMMENTS ON DRAFT FIRE PROTECTION INSPECTION MODULE May.24,1999 First Commenter The inspection basis (last sentence) is inconsistent with the objective Section 02.02.b appears to be demanding that we have an unannounced fire drillin a high risk area at a time of NRC's choosing The triennial general guidance should pick a different high risk area than the previous inspection, to be most comprehensive. It should also include a sampling check oflower risk areas. Otherwise the utilities might be tempted to " teach the test" and rigorously keep.the high risk areas at 100% and let other medium and low areas degrade
'Section 03.01/2: Include the referenced information from the FPFI module to allow this to be sufficient Section 03.02a should also request supporting calculations Section 03.02b: The inspection results and non plant specific fire event information have little to do with ranking fire areas according to risk; this should not be the responsibility of the SRA to acquire / address.
Second Commenter The main comment I have on the proposed guidance deals with the source of their criteria. The source of the data for the sprinkler systems, for example, comes from recent versions of NFPA 13. Those of us with older plants will not meet this criteria, even though we will be in full compliance with our committed code of record. This could result in being classified with a high degradation category deficiency, while being completely in compliance with our commitments.
One specific example is as follows: Older sprinkler codes only considered an obstruction below the head to be a problem ifit was greater than 48" wide.
. The guidance provided states that an obstruction below the head of 24" or greater requires a head below the obstruction. A high degradation condition would be identified if two or more heads were obstructed in this manner.
This would not be that uncommon with the presence of 24" wide cable trays and ducts.
It seems like we will be expected to backfit our systems to meet the current
n W ~- -
STAKEHOLDER COMMENTS ON FRSSM May 24,1999
+
i
~ ~ ~ ~ ~ ~
~
First Conunenter The nomenclature on degree of degradation has a fully operational and compliant system classed as low degradation. Suggest using high, low and no degradation instead, which also is clearer for the user.
General comment: The examples given in all the categories are more representative of what
)
would be found in an FPFI inspection, not that which would be found by a site resident in routine rounds or reported by a licensee in an LER.
The following comments apply to specific sections:
- 3. The description of the IPEEE analysis bears little relationship to what I remember us doing for it. This is probably even truer for those who did a fire PRA.
- 4. Step 5 refers to Table 4.2, implying that you are to find numbers in it. You find qualitative terms that have some arithmetic associated with them. There appears to be no way to implement this step with that table.
Step 8 last paragraph: It appears that SRT and DRT are reversed in the second and third sentences
- 5. Guidance, first paragraph: Need to explain what recovery equipment is being looked for. Ifit is cold shutdown Appendix R equipment that is repaired or " recovered", call it that. There is a lot of EOP equipment that is unrelated to FSSD.
- 6. General Assumption 3: Why here and elsewhere are compensatory actions not credited, particularly continuous fire watches? These are actions that are acceptable by Tech Specs or TRMs and make you " compliant" in the eyes of the regulation. Compensatory measures are not always equivalent but they should not always be dismissed l
- 7. Evaluation guidance footnote 6: The code edition referred is not necessarily the code edition committed to by licensees. So if we meet our code edition we are degraded ifit doesn't match this one Detectors: Since it doesn't address non - spot detectors nor explains what they are, drop the term. POC detector table: How do you know what growth rate you have in an area?
Footnote 9: What is the source of this statement? I don't think this is necessarily true.
Categories need to reflect the level that you are in if 1 or 2 detectors in an area are bad vs. all of them.
Sprinklers: For head problems, this does not look at what percentage is misplaced or inoperable; i.e.,2 bad heads out of 100 is different than 2 of 10. Also, where they are makes a difference.
V,..
C ~ r _.
g i DRAFT i
- 8. ' Brigade: The time summary says it is almost a 1/2 hour before a fire brigade shows up at a fire at best, at worst almost an hour. This seems a little unreal. The General l
Evaluation Guidance should be in the inspection module, not here. When you get here I
you already know that there is a specific problem with the fire brigade. It states if you can't witness a fire drill the brigade degradation is medium. I thought this method was l
to evaluate'actualimpairment significance.
1 I
- 9. SSD: This is not looking at problems with the FSSD methods, it is looking at the i
number of options you have for FSSD in a given fire area. 'I believe all the conditions listed from extra high to low are fully compliant with the rules. High degradation:
Because you use alternate shutdown in an area like the control room, for which it is specifically called for, high degradation is wrong.
i l
The Case 1 example is degrading items that meet their design without indicating that the design is inadequate, is questioning qualified fire barriers because of the way they are qualified (if this is an issue this is not the docurent to bring it up in; a GI would be more appropriate if they don't trust the qualifications of fire barriers).
The Case 2 example states that given no known issues with sprinklers and detection, you should assume that some may exist undetected and therefore a medium degradation assigned. If there are no issues, there is no degradation.
Second Conunenter l' The technical bases for the applied criteria need to be provided, including the bases for the detector spacing, sprinkler obstructions, and credit / degradation for quick response, ESFR, large drop systems, etc.
- 2. When determining degradation categories, criteria do not appear to be consistent between sprinklers and detection. For example, when discussing detection,25% not
- meeting placement requirements results in "high degradation", while for suppression, 25% or 2 heads not meeting placement requirements results in "high degradation".
1
~ 3. The application of the same degradation category for sprinklers as for detectors appears to overly conservative. Some credit should be allowed for having a fully code compliant sprinkler system even if the detection system does have some code deviations.
- 4. The application of the same degradation category to the sprinkler system as the detection system in essence results in a standard degradation for the purposes of this document, since for most plant areas, at least one, and usually both of these systems will have some level of deviation (s).
1
- 5. The degradation categorizations does not appear to give significant credit for having code compliant detection and suppression systems.
2
[
~
f DRAFT
- 6. The threshold (s) for getting a red and/or yellow categorization do not appear to be of the
.~
magnitude applied in other areas utilizing performance indicators.
Third Commenter Section 1.0, item 1) lists Prevent fires from starting as a DID principle. Suggest replacing this with " Control of combustibles and ignition sources" as this is the wording used in the Baseline document.
Section 2.0 discusses the use of exemptions. No: sure how existing exemption requests will be treated. Ifyou have an exemption, i.e. lack of detection or suppression, will you be penalized for it in Section 7?
{
Section 4.0, Step 2 discusses defining the fire scenario. There appears to be a broad spectrum of possible fire scenarios and resulting fire damage that can be used and will vary from region to region based on the fire protection engineer base of knowledge. See fire scenarios and resulting fire damage as a potential area of disagreement between the utility l
and inspector. Would like to see less importance on human judgment and more emphasis on the use of computer fire modeling.
See engineering evaluations and exemption requeste playing a major role in CDF values.
i Would like to see more discussion on how to treat these items.
How will the use of PRAs as allowed by the proposed NFPA-805 factor into this document?
You may not meet the deterministic requirements that play a major role in establishing the CDF but are acceptable from a PRA point of vievi.
The intent of this document is good, but see major problems with its use. If each region implements this document on their own there will be little consistency in its application.
Each region now varies as to what are their concerns and how they handle new issues.
What avenues are available to a utility if they disagree with the assumptions and resulting CDF? It is similar to the IPEEE, the majority of utilities submitted what they thought was an acceptable document only to have it questioned by the NRC.
Fourth' Commenter General comment: This is a considerable improvement over subjective criteria used previously. Addressing specific comments will make it much more usable.
Section 2.0, second paragraph: The term " findings" should be restricted to conditions which do not meet regulatory requirements, not items which have risk implications.
l Section 3.0,last paragraph: The description of Section 7.0 repeatedly uses phrases like "provides guidance for assessing the effectiveness of...", and thus reads like inspection criteria, not criteria for evaluating the risk significance ofinspection findings.
3
E I r-.
]
DRAFT r.-
Section 4.0 would be enhanced by providing an example for each step. Also in the last paragraph, Step 9: The second through fourth sentences do not explain clearly why it is necessary to " reduce the frequency of the area by a factor of 10."
Table 4.2: Suggest reviewing the consistency of the ratings assigned. For 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> barriers, "high" detection and suppression and " medium" fire barrier provide an " extra high" overall DQR, while "high" detection and suppression and " medium" fire barrier produce only a "high" overall DQR. There are other similar examples in that table and in the 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> barrier table.
Table 4.2: Instead of trying to show 3 and I hour barriers on one table, and have the table run on from page to page, suggest showing the I hour table on one page and the three hour table on as few pages as possible, even ifit means reducing the type size and re formatting.
This would make it a lot more useful.
Section 5.0, first paragraph, last sentence: This strongly implies the need to do fire modeling. Is this correct? If so, which model?
Fifth Conunenter The assumptions require more research to be adequately supported. The assumptions also need to be better connected with the method. However, this is a very good start.
Sixth Conunenter i
General Comments:
There is discussion and guidance on fire scenario considerations, but there is no quantitative end result. I would suggest adding a factor in determining the total CDF.
Let's call this factor, " Fire Effects Factor" (FE), and add it to the existing list of factors.
The result will be a CDF consisting ofIF, FE, SSA, FB, AD/AS and D/MS. The FE factor will be determined by the guidance provided in the document and tabulated as follows:
-0 =
Self propagating fire, w/o suppression of any kind.
-1 =
Plume, w/ ceiling jet layer < 700 oF, self extinguishing, with cable damage, w/o suppression of any kind.
-2 =
Plume, w/ ceiling jet layer < 325 oF, self extinguishing, without cable damage, w/o suppression of any kind.
-3 =
No plume, smoke only, fire confined to box or cabinet, w/o suppression of any kind.
Defense in Depth:
My concept of DID is that there is more than one scenario by which a plant can safely shutdown following an unexpected event. Each of these scenarios, can provide the desired
!~
4
f
~~~m
^
DRAFT s
result independent of the effectiveness or existence of the other. For fire protection these scenarios are-
- 1) Prevention,
- 2) Detection with Suppression, and
- 3) Separation (SSD)
If any one of these were perfectly applied, then safe shutdown would be assured. In the PSA world,1E-06 seems to be the ruler to tell us that we have assurance for safe shutdown.
Additionally, experience tells us that 1E 03 is the minimum achievable in any one segment.
Now let's apply this concept to the three areas of DID.
Starting with Separation, it would seem appropriate that if two trains of SSD equipment l
were available following a fire then the objective would be met. The numbers in table 4.1
= and the description in Table 9.1 do a good job for this section.
- The first line of defense is prevention. For this to be a 1E 03 scenario, an extremely
. reliable and proven method and application must be applied. If all of the following were true then a 1E-03 could be applied:
e.
Low or no in-situ ignition sources e'
Low or no maintenance ignition sources Fire retardant or high ignition temperature in-situ combustibles
. - Low or no transient combustibles Detection with Suppression says that even if a fire starts, it will be detected, suppressed and extinguished before there is any damage to SSD cable or components. Factors involved for this scenario include:
Detection sensitivity, reliability, availability, coverage.
Suppression capable of extinguishing all credible fires Suppression actuated in time to prevent any damage to SSD.
e
. Ability of the suppression system (automatic and/or manual) to respond in time to prevent SSD equipment damage.
e' Fire Brigade performance.
Detection / suppression has two means of being added, the combination for automatic suppression (associated detection system implied) and detection with manual suppression.
The total score should not be less that 3.
In the double room scenario, FB should be assessed at a maximum of-3 when SSA is typically assessed at zero, to be consistent with the DID method above.
This provides a potential minimum assessment of-12 for a perfect room, a room that certainly would have been screened out using previous methods.
5
[ ' *:.:
~
1, 1
DRAFT 4
Scorine each asnect of DID:
- . y.- -.
l This document has varying degrees of detail for assessing the acceptability or score for each L
of the areas in DID. I think that an overall agreement of the scoring should be attained, like the following:
Extra-High Item does not meet any of th'e design or assessed requirements.
)
High-Item functions, but at a very degraded level (ex. Takes a long time or only performs part of the intended function).
I l
Medium Item functions, but in a slightly degraded fashion.
Iow Item functions and meets the design requirements Comments by section:
Section 3.0; I agree that IPEEE analyses should be used with caution. However, it is not fair to say generically that these analyses considered a limited set of small non bounding, type fires.
In fact typically, if the fire is not analyzed in detail, a large all encompassing fire is assumed.
Section 4.0; Steps 3 and 4 assign degradation qualitative rating (DQR) from table 4.1 and then proceeds to try to determine what this means in table 4.2. This process is extremely subjective and would be vety hard for any two people to agree upon. I would suggest eliminating table 4.2 and using the table at the end of this document.
i An explanation of the reasoning process behind this approach is warranted. Iet me try the following. - Degradation of a particular feature in DID is only bad ifit brings the overall CDF to an unacceptable level. To determine this, we must first know the initial CDF for the area, zone or room being analyzed. This document does a pretty good job of doing that.
Each area, zone or room being analyzed first needs to be assessed as to the lowest CDF (highest negative) which can be attained, then determine the CDF after accounting for the degradation. This will then be the Delta CDF (DCDF). Certainly a decade change in CDF at a beginning value of IE-05 is of more importance than a decade change at a beginning value of 1E-09. This method would then provide that result. Assign a number to each feature of DID depending on the assessed performance of that feature, then determine the total area, zone or room performance by adding up the numbers. Compare the as designed CDF with the as found CDF to determine the relative effect of the field condition. See the table in the example at the end. To calculate the Delta CDF, we first need to raise each of the two numbers to the power of 10, subtract the two numbers, and then take the base 10 Log of the difference. For large differences this will just be the larger (less negative) of the two numbers. Thus, a good approximation of the Delta CDF is the As-found CDF. Now to determine the risk significance of an as found condition, the Delta CDF or As Found CDF can be compared to the green (< -6), white ( 5<x<-6), yellow (-4<x<-5) and red (>-4) bands.
6
n DRAFT 4,
The conclusion is that the degraded CDF is basically the same as the Delta CDF, thus only the AS-found CDF needs to be determined.
Eliminate tables 4.2 and 4.3, and use the table in this section (Step 7) to determine the overall rating of the area, zone or room being analyzed. The right column should be j
Degraded CDF and Delta CDF. Eliminate fudge factors like the one described in Step 5.
Step 6: The way PSA treats ignition frequency, it is the same as Prevention in DID. The i
numbers for a particular area are derived from three basic factors shown in the table at the j
end. Each of these factors has a potential for degradation and should be included in the 4
overall assessment. Table 4.4 is a good starting point, but plant specifics should be utilized.
The time that a degraded condition has existed, certainly is a factor in the analysis.
Something that has been out of service for just a few hours is certainly not as much of a concern as something out of service for a year. I think the approach mentioned needs i
refinement. Try something on the order of a) OOS for less than 3 days = no degradation, b) something OOS for 3-30 days = one decade of numeric degradation and c) something OOS for over 30 days = full as-found numeric degradation.
Step 8: The double room term is a good idea, but will take a lot of reanalysis since the PSA typically assume the barrier to be functional. I would actually recommend expanding this concept to include two barrier scenarios. The first would be the loss of raceway protection within an area; zone or room being analyzed and the second would be the barrier between fire areas as described herein. The case of raceway protection would be more severe since it is already in the room with the fire scenario and a fire does not need to propagate to another room to cause potential damage. The FB term needs to be defined to compensate for these differences. Also, the rules for when a DRT needs to be calculated are too stringent. If a barrier meets present requirements (3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />,1-hour + supp, analyzed as adequate), then a DRT should not be required. Even minor degradation in a barrier (ex.
Crack in wall, small bolt hole in masonry wall with no SSD equipment near the hole, etc.)
should not force a DRT to be generated, only featurcs which would allow the effects of a fire to propagate through the barrier. Also, FB should have a minimum value of-3 to be consistent with the other sections of the analysis.
Step 9: It is a little confusing on how the spurious actuation factor is applied. Is this intended to adjust the SSD factor? Right now the SSD factor assumes the all consuming Appendix R fire with all the failures and spurious actuation which apply. Is this factor intended to relax this number if multiple spurious is already evaluated for the particular area?
Section 5.0: Good section. See general comment for application of a number to this analytical methodology. The assumptions about cabinet fire propagation seem extreme, please provide nuclear power plant examples for these scenarios. This section would be very difficult for a novice to perform.
7 t
.g-7 h
1 DRAFT Section 6.0: Some leeway should be given to minor defects and deficiencies in barriers between fire areas and doors which will allow something less than a perfect assembly to meet the low category.
Section 7.1: This section is an abbreviated version of NFPA 72 Appendix B. These appendices are not part of the code and are only recommendations. Rather than just try to translate a portion of the code in this document, it would be much more appropriate to have the inspector utilize the code for the inspection. If there are specific areas where you want the inspector to concentrate, then mention that in this document. In other words, remove all the biased technicaljargon and reference the NFPA code. Additionally, the assessment for High Degradation item (b) seems a little strict (just one detector short would apply). (b) should be eliminated, item (c) covers this situation appropriately.
Section 7.2.1: This is mostly a regurgitation of specific section of NFPA 13. There is no need to rewrite the code here. Again, if specific characteristics of the system need attention, then state those characteristics and have the inspector use the code for the inspection. The code has a very good section on obstructions; there is no need to try for a different interpretation in this document. After all, the initial premise was that if the system meets the code it is satisfactory. Remove all of the code interpretations. Items b &
c under HIGH Degradation need to be removed, they are properly covered under item (e).
Item d would certainly make the system less effective but not totally ineffective, this item should be moved to the Medium degradation section. Low Degradation section should also have a line for systems designed, installed and evaluated to meet performance requirements.
Section 8.0: Donning of SCBA masks for all fire drills is not a requirement, and should be removed from the assessment. Some level of assurance that proper communication and physical conditioning is provided is appropriate, but is not required for all drills.
Otherwise this is a good checklist. In the HIGH Degradation section, item (b)is a little over restrictive. Certainly if all of the items were true it should be high. The determination of whether this is a high degradation should be determined on the basis of whether the Brigade is considered totally ineffective, not just weak in one area. In a similar fashion the Medium Degradation category should be for an effective fire brigade which could use improvement.
Table 4.4a: It should be noted here that IF is not dimensionless like the other numbers, it is the probability of a fire per year.
Case #1 in the examples is not an accurate depiction of the plant design or the actual FPFI findings. The following is the proper assessment of the area being analyzed.
I highly recommend that a document of this type utilize hypothetical situations rather than inaccurately depicting actual plant conditions.
l l
1 8
\\
L
{..
F--
y l
f DRAFT l
Case 1: Cable Spread Room and its interface with the "A" and "B" Switchhe'ar Rooms l
Example 1A.
The cable spread room (fire area 57) is directly under the MCR and is located in the Reactor Auxiliary Building on elevation 43'-0". The east wall of the cable spread room (CSR)is shared with the "A" switchgear room (fire area 60) and the "B" switchgear room (fire area
- 56) shares the south wall. These interfacing walls between the CSR and the switchgear rooms are 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> fire rated except for the Thermo-Lag wall segment (which includes a door) in the south wall at column lines RAJ/RA3 and RAI/RA3. This wall segment passed all the requirements of ASTM E119 for a period of I hour and 48 minutes, and passed the hose stream, flame and smoke requirements for the full 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />. Temperatures in contact with l
the barrier were elevated above the ASTM E119 criteria after I hour and 48 minutes.
However, the licensee performed an assessment to ensure that there were no combustibles or safe shutdown equipment near the barrier, which could be affected by the elevated temperatures. Thus, the barrier is determined to perform effectively (meet the intent of ASTM E119) for a period of 3-hors. In addition, there is a door in the east wall, which interfaces with fire area 60, and a door from the CSR into the static inverter enclosure and I
a HVAC air return from the inverter room to the "B" switchgear room. The static inverter room is part of the "B" switchgear room and is ventilated separately from the CSR. All of the barriers and barrier penetrants associated with the static inverter room to CSR are 3 hour3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> fire rated. Located in the "B" switchgear room are the hot shutdown panels which are i
used for post-fire safe shutdown and are required to be used to shutdown thn plant in the event of a significant fire in the cable spread room or the MCR.
The CSR is protected by an automatic Halon suppression system, which is provided with cross-zoned thermal fire detectors that activate the Halon system and provide fire alarm i
indication in the MCR. A separate smoke detection system is provided in the CSR for early
)
warning of a fire in this plant area.
A. General Assumptions The fire load in the cable spread room is high (primarily due to cable). The exposed cable in this room meets the requirements ofIEEE 383 or is coated with a flame retardant. Most of the redundant control circuits associated with the plant's safety functions are located within the cable spread room. For the purpose of this assessment,it is assumed that a fire has occurred in the CSR. A fire in the pressurizer heater bus caused by an electrical fault I
would not propagate beyond the bus. This is a sealed dry type of transformer and the effects of an electrical fault would tend to be confined to the housing (smoke could certainly be generated). Thus, a fire from a transient combustible will be assumed for this room.
The licensee had already performed a fire model evaluation of this room utilizing the maximum allowed transient combustible for this location. The result of this analysis was that without any manual or automatic suppression, the fire would burn out in about 10 i
minutes (using maximum burn rate), damage cable directly in the plume, not generate a ceiling jet layer temperature capable ofigniting additional cables.
9
f7 ll DRAFT it_.
It is assumed that critical post fire safe shutdown functions within the cable spread room J.
will be susceptible to fire damage and loss of function. Therefore, manual operator recovery actions may need to be taken outside the main contrcl room in order to regain plant control and achieve and maintain safe shutdown conditions.
l Due to the short duration of the fire and the 3-hour qualification of the barrier assemblies, it is assumed that the fire does not propagate beyond the CSR.
Since the fire barriers and passive fire resistive devices passed fire endurance testing which includes flame and smoke propagation, only minimal smoke (under fire doors)is assumed to propagate into adjacent plant areas.
The fire initiation frequency (IF) for the cable spread room is 5X10E-3/yr (Table 4.4a).
Due to a fire in the CSR, which results in the loss of a majority of the control room functions, the plant would enter into shutdown procedures and implement its alternative safe shutdown methodology (shutdown from a remote shutdown panel and/or other locations outside of the MCR).
l i
B. Post Fire Safe Shutdown - PotentialImpact and Capability For a fire in the CSR alternative shutdown capability is provided. This capability is electrically and physically independent of the cable spread room. The hot shutdown control panels are located in a room in the "B" Switchgear Room on Reactor Auxiliary Building (RAB) elevation 43'-0". Manual operator recovery actions are necessary to regain control of and operate systems, equipment, and components needed to achieve and maintain post fire safe shutdown conditions. Due to a limited number of air leakage paths (under one fire door) between the cable spread room and the switchgear rooms, a challenging fire in the CSR will most likely result in minor smoke propagation into the adjacent rooms.
An inspection team concern was room and electrical equipment cooling and control of smoke at remote and local control stations during a control room fire. The Hot Shutdown Control Panel (HSCP) room is a small (about 8' x 10') concrete room built on the far west outside wall of the "B" switchgear room. The HSCP room has dedicated supply and exhaust fans.
The "B" switchgear room has dedicated supply and exhaust fans as well. Specific problems include: there was an inadequate smoke management plan associated with the fire brigade pre-fire plans for alternative shutdown. Additionally, existing EOP's included appropriate actions to restore HVAC, however these actions were not specified in the ASD procedure.
However, calculations indicate that it will take many hours for the rooms to exceed temperature requirements, thus giving the operators the opportunity to turn on the exhaust fan manually.
The ASD panelis on the opposite end of the "B" switchgear room from the door to the CSR (over 40' away). Additionally the main exhaust fan for the "B" switchgear room is directly above the ASD room and would provide some ventilation.
j i
10 I
1 p
l t
DRAFT C. Conditions Affecting Fire Mitigation Effectiveness ij-r r
n During the inspection, conditions were identified with the design of the CSR Halon fire suppression system that could affect its ability to suppress a fire. Specifically, the system is designed to provide a concentration of 5 to 7 percent for a minimum of 10 minutes. Review ofinitial testing records for this system indicates that the system delivered and held a room Halon concentration more than 6 percent for 4 to 5 minutes and more than 5 percent for 11 to 12 minutes. These concentrations fully meet the criteria of NFPA 12A (1980) and the l
design basis requirements of the system as delineated by the licensee. Halon is typically l
not utilized for deep seated cab!e fires. Deep seated cable fires are not postulated in the CSR since there are no high energy :murces or concentrated highly flammable substances in this location.
l The inspection noted fire brigade effectiveness weaknesses associated with pre-fire plans (smoke removal / control). NRR representatives had a preference on fire brigade dress, however the equipment used to protect fire brigade members from the hazards associated with fire fighting met all OSHA requirements. Based on the observations made during the witnessing of a drill, the inspector had a preference to donning the breathing apparatus worn by the fire brigade during the drill. There were no requirements to perform this activity in this fashion during a drill.
D. Assessment Based on the conditions noted during the inspection, a qualitative screening assessment of the potential risk significance of the fire protection DID principles weaknesses was performed.
Assuming that a fire occurs in the CSR, the fire should be detected rapidly by the smoke detectors in the room. Since, the Halon suppression system is actuated by cross zoned thermal detectors (2 detectors on two separate detector loops have to sense the fire) it is inherent that a time delay in actuation will occur. The time delay associated with having to actuate a cross zoned thermal detection system would not provide the conditions for a deep seated cable insulation fire to become established. In addition,if a fire doo'r or damper in the system does not fully close or an HVAC fan does not stop; the system will not actuate. Considering that the Halon system meets all of the licensees' design basis requirements and meets the NFPA code, the degradations in AD/AS effectiveness would be low. However, use of thermalinstead or smoke detectors would delay the actuation of the system, and the effectiveness is determined to be medium.
The fire loading conditions associated with the CSR are high, however a reasonable fire scenario, based on in situ conditions and allowed operational practices failed to propagate a fire beyond the initiating event, and does not have sufficient potential energy to exceed the Thermo Lag wall qualification fire test conditions. Therefore, the affects of a fire initiating in the CSR will be confined to the CSR and a DRT condition does not exist.
11 1
y
~~
a DRAFT
- It should be noted that the smoke detection system provid.es notificatio[6f a potential fire condition and performs no active function to mitigate the consequences of the fire.
)
Detection system design and installation is a critical factor and can delay operator and fire brigade response to the potential condition. There were no noted deficiencies in the CSR detection system during the inspection. Therefore, the degradations associated with detection effectiveness are low. During the inspection, the fire brigade's ability to effectively implement fire fighting operations in a timely and efficient manner was identified as a strength. However, there was a lack of a detailed smoke removal plan. This coupled with the fast response to the installed smoke detection system moves the DSIS performance to low to medium.
Therefore, an assumed challenging fire in CSR is assumed to cause fire damage to redundant trains of shutdown functions in the CSR. Thus, the post fire safe shutdown methodology requires the manual recovery of plant safe shutdown equipment and functions that have been affected by fire damaged to the cables in the CSR. This is accomplished by abandoning the main control room, isolating' fire affected circuits and re-aligning power to needed equipment from alternative power sources, manually re-aligning valves in reactor water make-up systems, monitoring reactor shutdown and core cooling performance parameters and controlling them from HSCP.
The HSCP is located in 'the "B" switchgear room, which is adjacent to the CSR. In addition,
' the independent ventilation system provided for the HSCP room, which is located in the "B" switahgear room, could be rendered inoperable as result of a CSR fire. Smoke from the CSR fire would have little if any affect on the ability of the operators to safely shutdown the plant from the HSCP. The room and electrical equipment cooling for the "B switchgear
. room, and indirectly for the HSCP room, could be achieved through manual operation of HVAC equipment whose control circuits and components were protected from fire.
j Therefore, a low degradation exists for the CSR fire scenano.
j 1
The potential change in CDF of these identified weaknesses in the fire protection DID principles can be expressed as follows:
)
PRS = IF + FB + SSD + AD/AS +D/MS
~ Therefore, if a challenging fire were assumed to occur in the CSR, the following is an approximation of the potential risk significance (PRS) of the fire protection DID findings:
IF= -2.3 l-FB = -2 SSD = -0 AD/AS = -0.75 i
D/MS = -1.0
[
PRS = -6.05 (green)
This term is the double room ' term (term when fire barrier has a probability of failure).
u l-12
L
~. -
p m
j!
DRAFT
'T o
Example 1B.
(_
The FPRSSM indicates that the single room term (if barrier succeeds) must also be used to evaluate cases with no fire barrier degradation. Therefore the single room term will use SSD =.1 if full credit is given to the SSD and not deducted for smoke. Note that the single room term utilizes FB=high=0 since no barrier exists in the "B" switchgear room to mitigate the fire damage. The single room term is as follows.
IF = -2.3 FB = 0 SSD =.1 AD/AS =.0.75 DafS =.1.0
. PRS (SRT) =.5.05 (white)
The total of the DRT and SRT is -5.01 (white).
Remember that for medium degradation offire barrier, the double room term is adequate for this screening approach since the single room term is no larger than the double room term.
Example 10 l
Fix artomatic suppression: AD/AS = Low degradation IF =.2.3 I
SSD = 0 FB = -2 AD/AS =.1.5 Dais =.1.0 PRS (DRT) = -6.8 (green), this is the minimum achievable PRS for this room (DRT).
~
For the single room term, i
l SSD =.1 FB = 0 l'
Therefore total change in PRS is.5.8 OVhite), minimum achievable PRS for this room (SRT).
Taking into consideration the probability of having a fire, which could challenge the fire protection features, would add another decade (.1.0) to the previous calculations.
13
. 4 4 l
I.
i['
jj; f
i1 T
F m
t tm5 G
j]
i e
f M
A h
r e
t o
f n
5 t
o M & 2 e
F k
1 R
i a
D M g l l
i "n
5 5
t C
a 2
7 5
8 n
t D'.
o 2
?
T 2 p F $m7 a
6 6
6 e
t hdl e
t o d oD M1 ng e aah
,t s s r un, D
eho S
b i
mTt 1
2 3
1 1
S 0
i
.d
- s u sn e
i nr o
r o e c a
s wbmd u
l t
t n
a u.
- 5 5
e u u 5N7 p
5 n
e p
2 5 7. g 0 7
5 h no f
a 0
l t
f s u
0 0
o d
d MS 0
0 0
cws n t
- v -
a e
at a
d re nb t
t h a) n 6l b tf4 e
e D
uf o
,o e (>
m s
- 5. M 0 )e n d op c
m t p 5
5 uu t
1 e
v a r o
AS 0
- (
fi 0
1 1
1 1
1 c
c ot e
afid r gi n
@7 r
e e n w n g )a e
no i
h o s s5 it psk<
t c
5 5
5 e e f
e 5
5 s x l
o t
2 5
7 7
7 h(ir<
e t
0 0
0 r
4 0
0 0
e n
D 0
oe e o
m t gh (
i rt t
s w
M1 o
c r a e o s
e l
e nl t
b ei l 4
g e
e mh my 1
n D
t r
u e e ),
i z
nbt 6 E
i e -
1 2
3 l
n F
0 1
1 otwsd<
i o
t u o <x i
u s
t jt s
r el 5
t e
o m
hl w-r f
i n
t o(
wNt p
)
p e
p AT f
e m
u B N R p/A A
A os
/
i
/
/
i
=
S F(S 0
- ;.N N
N hh
.h 1
2 3
ct Pw s
m a s D ),
M2 e
o e e s
o ecC6 s
R y
nd<
s r
a i
e r
t a r n(
e a
o
. m 5
5 r eu n b
f 5
5 5
of o e
e r
n mo 2
5 7
f e 2
2 t i
- e a
o oe 0
0 0
d d sr 0
0 0
Z CG 0
eeAg e
)
eg e e r
a r
nr i
e /y s
ahh l t t b
4 f
r ts e
A o
s r rs o oit h
r r
.l FFd i
c f
t o P o
e.
e e
f
(
b rt 5
5 Dr mn 2
5 7. e. 5 5
5 w.Ca s
P
)
e F
,e d
D oo p
0 0
0 0
0 0
Pc I
CC 0
t a
nt C (n Del m i
v e o f
Cr M1 o
o o
afDc e
r i
f e
t t
s p
n n e
eb t i ledh n e
e c
w m
t r v
t 5
5 5
De i
e n u 5
5 f a o
s r
go 0
1 1
ho c 1
1 et l
s P I S 0
hf nF e
e b
s t ooD s
i 4
egt C e
A o
h taoa l
i g
d d
md t
b 1
l l i
f a a
4 od H
m g;
e e
n u
n n
n u
c0ixu l
4 t
a a
t si ig o a
a u
b l 1 i g oo e
l i
e r r h L$BPD F.
t c erF e
l d
w s s s
- s b
vg t
g e
l s p h
o eo e s e s
e oaps a e e x
i T
T LDn E H M AD A D Tb aA l
.