ML20128C476
| ML20128C476 | |
| Person / Time | |
|---|---|
| Issue date: | 10/20/1992 |
| From: | Rothberg O NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Baer R NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| NUDOCS 9212040437 | |
| Download: ML20128C476 (22) | |
Text
- s;
.h* gono,of.
v
' UNITED STATES
'W y*
NUCLEAR REGULATORY COMMISSION:
,g 1
r-WASHWGTON, D, C,20555
?.
j s
%....*y
-0CT 2 01992:
O MD40RRELH FORT Robert L. Baer, 011ef Ergineerirg Issues Branch' Division of Safety Issue Resolution Office of Nuclear Regulatory Pesearch
/\\.
'j h Frag Chemy, Section inader B DiROUGH:
/
Engineering Issues Brarh Division of Safety Issue Resolution Office of 11uclear Regulatory Research FRCti:
Owen Rothberg, Task Manager Ergineering Issues Branch _
Division of Safety Issue Resolution office of Nuclear Regulatory Research SURIECT:
REPORI' OF MEEI'DG WIDI IUMARC AND !UCIIAR UIILITY REPRESDEATIVES ON SEPID4BER 29 AND 30,1992 'IO DISCUSS -
!Lt4 ARC'S VERIFICATION AND VALIDATION PROGPR4 EUR 'IHE DIDUS'I2Y IMPLD4DEATION GUIDANCE FOR DIE MAD;TDIANCE IUII (10 CFR 50.65) e On September 29_ ard 30,1992, menbers of the NRC staff atteMed meetings.at ICMARC headquarters with representatives of IRNARC, EPRI, -DTIO, ard several nuclear utilities. We purpose of these meetings was to discuss the verification and validation of !UMARC'S proposed guidance document, ILEARC 01, Revision 2A, July 9,1992, " Industry Guideline for Monitoring the Effectiveness of !bintenance at-Nuclear Power Plants." Please refer to my s
previous meetirg report to you dated August 27, 1992._
l An atteMance-list and meetiry ageMa are enclosed. - Walt Smith of 10 MARC chaired the meetirg. Although a number of program descriptions ard handouts s'
were provided for discussion, nest of this material was returned to NUMARC by the NRC participants because it is preliminary and not ready for distribution.-
Before the. start of-the general meeting, and later on:as well, the NRC representatives attended breakout sessions regarding methods for detemining.
risk-significanoe of SSCs under the scope of the rule. Walt. Smith-. chaired the sessions. Ed Resnick of INIO, Dave Worlege of EERI, Alan Marie of PeCO, Jott e Ramirez of DECD,. Gary Czeschin of_ Union Electric, ard Tom Galloway of Northeast-Utilities also atteMed. We current draft of the NUMARC guidance document (10K\\RC-93-01, Revision 2A, dated July 9,.1992) outlines several(
methods for determinirg risk-significance of SSCs including two based _on
_ probabilistic risk assessment (FRA) methods. % ese are referred to inlthe-text of NUMARC-93-01 (Section 9.3.1). as " Method 1" ard Methed 2."
- In.
-general, the irdustry is concerned that Method -1 will not be; implemented -
efficiently for a number of technical reasons including complexity and :
possible non-uniform application of PRA techniqaes. We particular concerns included:-
u i
9212040437 921020
- h PDR j;
REVGP ERGNUNRC-
-PDR
' t p.
4 007 2 0 1992 1. -
' hhich ' frisk importance measures" are appropriate for maintenance?
2.
Problems arise involvirn truncation and omission of cut sets in PRAs.
3.
At what _ level (system, function, train, major equipment, conponent) is risk measured?
4.
-Is the choice of the " top 10" SSCs for risk achievem nt importance measure an appropriate number?
5.
Should the risk achievement worth include a cutofI at 10'"?
4 6.
What is the " shadow effect" between plants (10 vs. 106) 7 7.
Can PRAs rodel the containment function?
8.
Should risk achievement worth apply to equipnent out of service only, or should it include degradation?-
9.
hhat inconsistencies exist between plants in considering systems, trains, conponents?
10.
How do we balance judgement vs. PRA calculations?
11.
How are failures accounted for? How is the time that the failure occurred estimated?
12.
hhat is the criteria for detemining risk significant systems?
13.
hhat does the phrase, " consistent with PRAs" rean?
1 How are initiating events to be considered?-
After same discussion it was decided that the tems in method 1 are sufficiently vague that tha results would be, at best,- not unifom.
It was y
also noted that the methods of PRA are not-very easy to apply and-that significant data may be unavailable. Dave Worlege-(EPRI) decided to try to devise wortlirg that would address the concerns that were identified in the
,D discussions. His proposal, and tae subject of methods for defining risk significance, were brought up a number of times during the discussions. A
~
copy of his proposed revised wortling for the methods of determining risk significance usirg PRA is erclosed. - All participants will consider the adequacy of the proposed modification to the guidance for discussion at the-next meeting. n e group agreed that system or train. level was the appropriate level to measure risk significance and this should be used for consistency.-
2e general meeting included a discussion of agenda and previous meetirg topics.- Jorge Ramirez (DECO) gave-a presentation describirg use of PRA -
methodology. An outline of the method used by one licensee to detemine the performance criteria-for instrument air systems was die' w ul and alcopy is enclosed.' m is was considered to be an excellent example of the method to-detemine performance criteria.
h e significance of the mode of operation of a plant was discussed a number of times durirg the meeting, with respect to the scope of the rule, perforrance criteria, and availability of data in PRAs for the various modes._ me related topics-of shutdown risk and aqnipmant being out of service for repair or preventive raintenance were also brought up a number of times in the discussions. My general irpression was that licensees are not very confident that their data, or that of the industry in general, is cumprenhesive erough to make the ne-ry conclusions about the relationship of.the effectiveness of raintenance to these topics.
1 4 4,
OCT 2 01992 - Another issue that surfaced a number of tims is that licensees, with varyirq degrecs of exception, do not employ forral procedures for evaluating plant cordition when takirg equipmnt out of service for maintenance.
In mmy.
plants such matters are left to the shift supervisor or other responsible person, based on their knowledge ard experience. The technical specifications are of li:nited help in efforts to assess the overall risk to the plant of various pieces of equiprent being out of sarvice at any particular tire or in a particular rode of operation. mis is a big c A rn because the raintenance rule will require licensees to explicitly consider the overall effect on plant safety of equipnent beirs out of service. Some of the V&V participants irdicated that they would develop proccdures to specifically address this issue for their plants.
We subject of the anount of time that equipment is considered to have been out of service when it is discovered to be inoperable was raised several tires with regard to probabilistic risk assessment. W e usual convention is that a piece of equipment is considered to have been out of service for one-half et the time since the time that it is known with certainty from objective evidence to have been operational. !!o conclusions were reached but the subject will probably be raised again. We problem is that the relative conservativism of the convention is not really known. We imC staff representatives decided that they would be wM to any change that is less conservative than the current convention unless there was objective evidence presented that such a charge is arreptable.
Discussions also included balancirq unavailability ard mintenance, performance criteria for risk-significant systems, ard use of NPRDS. We subject of " inherent reliability" was also discussed although the term was not defined in the discussions.
tuGRC proposed a new ration in their docunent (8.2.1.7) for renoval of SSCs from the scope of the raintenance rule. It was not di&:ussed in detail. A copy of the proposed wordirg is encloscd. We second paragraph, which proposes removal of an SSC after 7 years of " exceptional" service, would appear to be in violation of the mintenance rule. We imc staff participants plan to develop and present a position on the proposed new section at the next V&V meeting.
Several licensees reported their progress to classify SSC within the scope of the maintenance rule. Roy also included some very rough cost estimates.
Same non-safety related, non-risk significant, non-stardby equipnent under the scope of the raintenance rule does not appear to be covered by plant level goals, for exar1ple the steam generator tube leak monitorirg system.
Additional meetings are scheduled for October 20 ard 21, November 17 'and 18, ard December 9 ard 10.
Wese meetings are to be held at tRNARC headquarters in Washington, DC.
1 i
OCT 2 01992
. During the !L' MARC rectings, OGC coments on the staff's regalatory analysis fot-the rninterance rule were delivered to PES. 'Ihec 'crrents were not dircussed at the meetings, hcuever they are included cv (enclosure 2) for your inforration because the coments apply dirc<:tly _
che !L' MARC gaichmce docunent that the staff interds to erderse.
.nL Owen Rothterg, Task Manager Dygineerirg Issues Branch Division of Safety Issue Resolution Office of 11uclear Rcqulatory Research
Enclosures:
As Stated
l MY40 # MP &
l 7
p lh V&V Maintenance Meeting September 29 and 30,1992 Revistd Agenda o
NRC/NUMARC Breakout Session 0800-1000 Smith Review of Previous Minutes Review of Overall SS Comparison Resiew of Detail SS Differences o
NSSS Breakout Session 0900 1000 NSSS Partici Review, Disposition and Assessment of Valid SS Differences o
Method #1 Plan Development Breakout 0900-1000 Hall Worledge o
Fermi Presentation on PRA Usage A ( /v ~ #~ ~ /?r /
10001N5 Ramirez o
Report of PRA Subcommittee on Meth #1 INS.1100 Worledge o
Grand Gulf implementation Presentation 1100-1145 Mooney 6 4 ~/ 4 5 7, /+.
Thurman o
Action item review from previous meeting 1145-1230 Smith o
Lunch 1230-1330 All o
Utility Participant Reports 1330-xxxx All
g i
r Maintenance Guideline V&V AHAC Meeting Agenda September 30,1992
\\
j Report on Risk Significant Methodology o
o Discussion on the use of unknown failure start time Reports by utility participants o
o Systems identification to INPO and NRC o
Presious Action item Status o
September lessons learned report
_ =..
.L i
11EETING ATTENDANCE C051511TTEE/SUBC051511TTEE: Verification and Validation AHAC DATE: September 29 and 30,1992 T151E: 9:00 a.m.
LOCATION: NUMARC Main Corderence Room NAAIE ORGAN 12a iION TELEPHONE f
$h$0 bC2-012-l20()
hb V dMA R C-
~
~
WAl/C Af EL[d T/K/C 4MY 82ESCHmi gi.7 a,q g pi g r ?I4-h?$- by w c. c a o.m o G'Cw Q ; " t t ' u t
<c_,,.-
- q a:
S O
[
a p r.S jf ff'~
~ s -hO 9 1 - T d Y - E 2 5 %'~
\\\\ :. L % -
6. ~.. n o e s a m., sm nu -na
~
e, o
i T~E o T~E N ic im c T L(
- fc,,3 7ggg pHtw& ist('bo ac.<j) y V NM ND Ylch0l5 0nn. YRnter Ak< iL;
.oos u.7 scsc-mbn e iGn -
^/~ob wwna um-ow&7 ac t Joi,/
wae sn +n me 6m 9-s mcAm di'4e = 4 L%</iy Lee fPC 96 Y 56 3 - V69 R Sac 2 /%l?annec tc!
FPC 9e g., + s-6'i eb
g h1EETING ATTENDANCE COS1MITTEE/ SUBCOMMITTEE: Verification and Validation AHAC DATE: September 29 and 30,1992 TIME: 9:00 a.m.
LOCATION: NUMARC Main Conference Room NAME ORGANIZATION TELEPHONE 6ARY
$OLL/Ng nmu y - ANo-/
Soi-q,r.gc79 lctm _fWfM kIc Ekt G
z w-6 y c - 6 f*f&
3try D l&RLS&e e-PA E in r-ErK-2h2 to gs.s.ou.A 4:A -h -Aw,
t +4 ARD b.
a 4 l.otr 919 - Sflo-Jf 40 seu-o Sn m n v
% n e.,
enten - asas coi-w - uss
[
t Hn %ea m c, 2,, tc. a v - GG (J5 L-m - 8 m.s m t
,y
,44awa>
th'/rdm6 Ncv Gu.s)-586 W G6AGf,0xx Ec-DEC0 ft5)
S EC-l.5>7 NRc /REGooN zr Cen) 860-8247 lcm Stetka
!%cxRnGr Inc.fR&sma f
( k s) m d6' t h50d-joog, c
rvtta So MEC 2.r2.
t7P
-'T,e) 9 4
r 9
%en-T-
1 gww-p-etus w
W WE-
*VaiT-r'r'Mre7
rm m
M'
'O H /
1 j
duw.azg.
If a utility selects a method based on PRA to establish risk significance, the process should begin by assembling a panel of individuals, experienced with the plant PRA and with operations and maintenance. This panel will exercise judgrnent in fornung a list of the functions that are important to the plant risk. This process will use insights drawn from the PRA results such as the leading risk contributions, the contribution to core damage frequency from initiating events, and knowledge of important support system dependencies. All PRA software provides information on Fussell Vesely importance or Risk Reduction importance which can also be used in forming the list of important functions.
The list of risk significant functions should be developed into a list of risk significant systems and/or trains that support those functions. Most of the important functions and supporting systems / trains will be identified confidently through this process of expert judgment. Where uncertainties arise or where the risk significance is less obvious and a decision must be mada where to truncate the list of SSCs that are risk significant, the judgments can be calibrated by specific risk importance calculations of the following tvpes.
AfETHOD 1 An SSC would probably be considered risk significant if its Pih Reduction importance Measure with respect to CDF exceeds about 1% of the CDF. Risk s;gnificant containment SSCs may be identified in a similar way.
AfETHOD 2 SSCs would probably be considered risk significant if included in the cut sets that, when ordered in decreasing size, cumulatively account for about 90% of the CDF. Risk significant containment SSCs may be identified in a similar way.
1-
1 Method 1 and Method 2 are similar in concept and differ principally in how the i
cut-off is applied. Judgment should still be employed as to whether these calculations should be used directly and v hether some combination of these methods may be appropriate, such as using method 2 but also identifying individual cut sets larger than 1% of CDF as risk significant.
METHOD 3 wE This method should 1wsps be38estto generate additional insights in addition to 9
either of Methods 1 or 2 because it addresses a difi ant concept of what is important to risk.
When SSCs that can be taken out of service are ranked by their Risk Achievement Importance Measure with respect to CDF an SSC is risk significant if it shows at least a doubling of the CDF. Risk significant containment SSCs may be identified in a similar way.
When exercising any of Methods 1,2, or 3, whenever two or more systems / trains are functionally redundant and none have been individually selected as risk significant, if the function is risk significant, the at least one of the functionally redundant systems / trains should be selected.
The above combined process of expert judgment using PRA input should recognize the limitations implied by the scope of the PRA, the limitations brought about by the PRA structure (i.e., model assumptions, treatment of support systems, level of definitions of cut sets, cut set truncation, and inclusion of repair or restoration of failed equipment) and limitations in'the meaning of the importance measures.
}
}
INSTRUMENT AIR SYSTEM PS3 i
.In determinag the performance criteria for this system, maintenance outage hours will be considered first. It is estimated that any MPFF can be determined by monitoring maintenance outage hours. Figure I shows the history (back to January of 1988) for the maintenance outage hours. The only year with obvious outage hour problems was 1991. The total hours for this year was 387 hours0.00448 days <br />0.108 hours <br />6.398809e-4 weeks <br />1.472535e-4 months <br /> (which is 200 hours0.00231 days <br />0.0556 hours <br />3.306878e-4 weeks <br />7.61e-5 months <br /> greater than any other year with records).
Other means of determining the performance criteria for this system are:
- PRA estimated outage hours
- Surveillance pass / fail history
- Work order history (Completed RCM)
- Any EPRI, NPRDS, or any other industry guidance To better determine the occurrence and duration of outages, a history search was conducted for this system. Figure 2 shows, for 1991, the total work hours for components in the Instrument Air system that would have made the system unavailable. This figure does not include the duration for tag installation and removal. Figure 2 will be used to relate outages with associated work documents in order to determine the cause of the outage.
The assumptions made with the plant PRyAJonservative_.or. restrictive with mgards to the.
hours unavailable. With respect to the evaluation on the Instmment Air System P53, the number of main:enance hours or hours the standby compressor can. be unavailable is approximately 17.5. It is not certain how sensitive the overall plant PRA evaluation is to this number of outage hours for Instrument Air. There is a discrepancy between plant scheduled outages along with the number of required PMs forInstrument' Air and what the PRA assumes.
Out of 19 surveillances reviewed (both I.LRTs and stmke times), only one task had unsatisfactory results. Task # 1886 (WO# 67626) on valve P53F006 resulted in unsatisfactory leakage rate. This one failure to meet minimum leauge rate is not significant due to the interface between Service Air system and the offline Instrument Air unit. Instrument Air has a standby function with a backup unit. If the online urJt fails to supply air flow and pressure to plant components, the Service Air system will be the backup system. Therefore, the offline unit is not the primary backup and the maintenance outages hours are not looked upon as a -
significant loss of system function. Also, Unit IIInstrument Air System cannot be started from the Control Room and does not have an auto start function.
1
n.
Presently, the NPRDS program does not monitor or have information on the Instrume a Air System other than containment isolation valves.. This is a lesson leamed that T NPRDS reporting is to be used by the Maintenance Rule, improvements will need to be, mah n broaden the type and amo2nt of infonnation within the scope of NPRDS.
From the attached Figures of ma.:ntenance outage hours, 200 hours0.00231 days <br />0.0556 hours <br />3.306878e-4 weeks <br />7.61e-5 months <br /> will be the chosen goal for this system. This means that for the year 1991 this goal or criteria was not achieved.
Therefore, in following the guidance of the Maintenance Rule for this system, a failurehause detennination will be perfomled for 1991.
4 CAUSE DETERMINATION ON MAINTENANCE OUTAGES The following outages were reviewed to determine cause:
QMlage Date Duration Associated Work Order-1 1/3/91 7 hrs 32462, 32364 2
1/11/91, I hr 29315 3
1/18/91 6 hrs 29007,29004 4
1/21/91
- 45 hrs 30515, 30921, 32827, 32916 5
1/28/91 8 hrs 34597,32902 6
1/29/91 16.5 hrs 13801 7
2/6/91 12 hrs 35803,31600 8
2/8/91 4 hrs 30825 9
2/20/91 9 hrs 35244 10 2/28/91 4.5 hrs 30867, 28254, 36682, 32816, 33027 11 3/6/91 16 hrs 38017, 35520, 36254, 35521, 38007, 35522 12 3/20/91 2 hrs 35519 13 4/4/91 2 hrs 32862, 27907 14 4/7/91 7 hrs 34497 2
9 15 4/16/91 6 hrs 40'324, 40725 16 5/17/91 5.5 hrs 43237 17 6/8/91 12 hrs 44646 18 6/11/91 4 hrs 42896,45288 19 6/14/91 3 hrs 45551 20 7/1/91 8 hrs 46134 21 7/11/91 7 hrs 46788 22 7/19/91 7 hrs 43291, 44409 23 7/23/91 2 hrs 44955 24 8/16/91 13 hrs 48086 25 9/5/91*
I hr 48419 26 9/13/91
- 38 hrs 49381,46289 27 9/18/91 4 hrs 52060 28 9/21/91 4 hrs 52053 29 10/1/91 8 hrs 44995 30 10/3/91 6 hrs 52114 31 10/8/91 3lus 52512, 52513 32 10/15/91 4 hrs 53974 33 10/27/91 2 hrs 54213 34 10/30/91 6 hrs 52656,52657,54616 The total outage hours for 1991 for Instmment Air system'is 283 hours0.00328 days <br />0.0786 hours <br />4.679233e-4 weeks <br />1.076815e-4 months <br />. Some outages included both corrective and preventive maintenance time.
3-
(O
~
e Attention vill be focused initially on outage #4 with 45 hours5.208333e-4 days <br />0.0125 hours <br />7.440476e-5 weeks <br />1.71225e-5 months <br /> and outage #26 with 38 hcurs.
These are two outaFe that accounted for a large percentage of the total outage hours.
%Ltag W This outye took place in January 1991 and the RChi for this system was complete in October 1991. Work order #32827, hich logged 45 outage hours, rebuilt the compressor. Previous work orders replaced damaged vibration probes. An evaluation of the damaged vibration probes led to the disassembly and inspection of the compressor.
During the inspection, thrtist clearances were found to be out of tolerance. This was identified to be cause for the compressor problems and the reason for the damaged probes. The RTM recommended adding steps to existing PM tasks to check thrust cleamnces and coupling lut)ricatt -
This outage could have been prevented by performing the proper mainten:nce. It will be concluded that Ais is an Maintenance Preventable Functional Failure MPFF.
GtlagsAf2 Work order Gn$1 was issued ori the dryi.:g tower which was not drying. The work order took the Unit 2 compressor (2P53C001) out of service in order to replace the 4.way valve, adjusted the valve coupling, installed 3 new mufflers, adjusted timing cams and limit switches. The cause of failure documented (in the work package) was the switch cams and valve were improperly align"'
This failure or misalignment shculd have been prevented - thus a Maintenance Prev le Functional Failure (MPFF).
Previous history (RL..f on P53) identifies problems in the timing circuitry for this system. This was al.o the suspected cause when this work order was initiated. After troubleshooting, the 4-way valve w.ts found to be misaligned
' Iso, the stop limits on the valve were found needing adjustment.
The RCM on P53 initially recommended disassembly and inspection of the 2P53FA04 4 way valve. This recommendation was not implemented since the failure of this va' e is not significant to the plant and failure of this valve is casily detected. Additionally, the vendor provided informati.: that these failures are most likely due to over torquing the valve in the rebuild pnxess. The RCM recommended that the valve be torqued at a lesser value than the recommended torque requirements specified in the vendor manual, if failures continue, then a periodic rebuild she.ld be evaluated.
4
jj CONCLMS10N Considering the total number of >utage hours for the year of 1991 with the causes for the outege
- 4 and #26, Instrument Air system is an A(1) system. In following the Maintenance Rule process, a system goal will be established which will te 200 maintenance outage hours.
j Outage #4 and #26, which totaled 83 hours9.606481e-4 days <br />0.0231 hours <br />1.372354e-4 weeks <br />3.15815e-5 months <br />, have already been addressed in previous cause detennination either through the MNCR process or the RCM that was perfonned on this systein.
Following implementation of the recornmendat.lons to correct the causes, Instmment Air system i
has performed more reliably. Figure 3 shows for the year 1992 the outages and the total number of outage hours. As can be seen by this figure, the present trend predicts less than 200 outage hours for end of year.
Therefore, this system presently is in the A(1) category with irnproving trends for the year 1992, It is expected that the trend in 1992 will continue. At the end of 1992, an evaluation will be performed in order to determine if the 200 maintenance outage hours goal is met and if the system can be retumed to the A(2) category.
F 5
i
.. _.. _. _. ~..,,,
, i-i
- ll;i;L!;
F
-rt
,l,t;t
[>
l!!Lter 29 tc O
32 y-
~~
9
'ya~,
2=
2 s
9
~
,O
- m..
,A r
p
.L 4
'^
- Yey4.'
~.
,:v Qy, s
n 6
-;';w x.
0
+
,r
{
2
- x' a^
l' 1
1 9
9 9
pe
, S 9
1 8
1 9
8
-ra
,M e.4 E
3 1
0 a'y. '
,+3 s
r,.,
I w~y;: y;h.=s. <c%f clyY b
q
.r
- i -
_A
. Mvw2~
&s.g;;p' gg.
.,.#.$gz%@
I i
- g e
f.
f f ~ h" "'
g-
-s.*dW e
0 g
0 n.
hm$: s, ; g g w-- a ,u A fO,$ ,0 t
- 9p h
5 C= M.*pp#p .:nf+$:&,1 I u ai e m%# +g <gpWg iAS!tNhn?Q g 4
- @4v%?h ym^
e ~. O t &y r .i a n ~ Y:r. ? u
- &0'. 4 g~ M ' n%[
D i pC Q. g g ..? g
- j Z,,w 14 I$1 3ha e
"9 i F r J i 7 i 2 A 98 t 9 lu n , J 8 a 1 e 1 m B8 u g -c , e r Q. D 4.jQ p er"q #r ;;. >p 3 t
- ii,
2 p- '- E>$ c s .nny. w,. 8 c,n 1 s' m1 . I,e %.. y-n
- l 8
I 8' .- w 0 .r' m7' + - 8 7 ~) , n - + - u ~ J
- t y0 _x 6
a 0 7 8 -v o N 9 1 0 0 0 0 0 0 0 o 0 5 0 0 5 0 5 4 3 3 2 1 1 L ~ T' ,I N_ 44 :
- 14 j
l
- 1) ;: ; j, 4 }i!, <
(,
- {1,,!
.} f;:a
- 4-
',i i j: i !-ai: ii
1--m, ~ '~ y. s E 6 8 E N 6 o os 6 1 3-n z a D g +2 60 B ~ g w g U ~ O s i m o E 7 O b ~ ~ p u>II 4 E g e C 2 O 8 g ~ b C s e g2 M 6 CO N g E, H E 5 B E~ 53 be--- "o5 g I J l i I I I I I I I I I '- 7 o-o o o o o o o o o o o o o-o o o m O T N o m O T N o m W 4 N C9 N N N N N e-e- r-r- r-w
2 9 -c ie D 2 2 2 9-v 2 o i N 9 2 1 i 2 9 -tc i s O -3 e 0 g' i a 7' 9 2 t v i t u A -4 O 2 3 2 r - e 9 e f.lu t r i l J-a A D u i 5 g 1 iF i t n 29 n e i uJ-m 5 50 u 2 r 9-r t A p s 6 2 n
- a I
29 ~ ra i M b -7 1 i 29 -be F -60 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O 0 8 6 4 2 0 8 6 4 2 0 8 6 4 2 3 2 2 2 2 2 1 1 1 1 1 I 9~
l6 IS NEW SECTION 8.2.1.7 [ REMOVAL OF SSCs FROM SCOPE OF TIIE MAINTENANCE RULE] V & V ACTION ITEM #17 Both safety and nonsafety related SSCs can be removed from inclusion in the Maintenance Rule if they are deleted as a result of a plant modification. The modification should clearly document the basis for the deletion and also provide a method for feedback to the Maintenance Rule program to delete the SSC(s) from scope. Plant modiGeations which merely replace (changeout) an SSC with a like or r,imilar type will continue to be included in the scope of the Maintenance Rule. For example, a safety related pump is deleted then the pump is removed from the scope of the Maintenance Rule, if the pump was replaced with a different type (centrifugal vs positive displacement) or model (three stage vs two stage) the pump would continue to be included in the Maintenance Rule. Nonsafety related SSCs which have shown exceptional performance over an extended (seven years or longer) period of time can be removed from the scope of the Maintenance Rule. The exceptional performance can be demonstrated by the absence of failures, a high availability or by a documented engineering analysis that all previous known failure mechanisms have been corrected. For example, if traveling screens were a problem and caused trips but the problem was corrected with a modiGeation and performance has been good over seven years then the traveling screens could be removed from within the scope of the Maintenance Rule. The ability to remove SSCs from the scope of the Maintenance Rule will allow utilities to focus and respond to actual problems and not have to maintain a large backlog of past failures which may not be relevant. 'Seven years was selected because it would provide sufficient time to develop historical trends (three or more refuelings) and allow for monitoring SSC performance. In addition the PRA data would be updated to reflect the last five years and therefore would not contain any previous failure history.
hC.W%
- h OMNND 2
i PROPOSED REGULATORY GUIDANCE - CONCERNS AND QU!)3TIONS CLASSIFICATION OF (a) (2) SSCs The guidance does not appear to contain a clear explanation of those requirements which must be met in order to place a structure, system, or component (SSC) under a plant's preventive maintenance program as provided by (a) (2) of the Maintenance Rule. In order to avoid the monitoring requirements of (a)(1), a licensee must demonstrate that the portormance or condition of an SSC is being ef fectively controlled through preventive maintenance. Without this affirmativo demonstration, the SSC in quest'en has not i satisfied the requirements of (a)(2) a7d must be monitorei under the (a)(1) program. In several sections, language in the guidance implies that an (a) (2) classification occurs automatically, without a licensee demonstration that the SSC is currently being maintained such that there is effective performance. In addition, some other sections are unclear or incomplete as to the requirements of (a) (2) even though their language does not, Der se, create an incorrect characterization of the Rule's requirements. The sections which require particular attention include portions of the Executive Summary (p.iii, 1. 20-23), Figure 1 (p. vi), S 9.2 (p.12, 1. 29-32), 59.3.3 (p. 17, 1. 17-31; p.18, 1. 5-12), and S 10. 2 (p. 26, 1. 16-39). The above sections do not clearly reflect the Rule's requirement that an affirmative demonstration must be made before an SSC qualifies under (a)(2). In addition, the document as a whole does not provide clear guidance as to the requirements of (a) (2). The Maintenance Rule must be implemented by July 10, 1996 pursuant to 50.65(c). Until then, there are no tests or requirements that a licensee must meet under the Rule. The language of the proposed guidance anticipates that licensees will use the period between now and 1996 to identify those SSCs covered by the Rule at their f acili ty and place those SSCs into either the (a) (1) or (a)(2) category. On and after July 10, 1996, however, the licensees must. comply with the dictates of the Ru3e including the test for the classifi. cation of SSCs into the (a) (2) category. If a licensee can generate data which demonstrates the acceptable performance of a particular SSC between now and the 1996 date, that licensee can use that data to support an (a) (2) classification.at the date of implementation in 1996. The proposed regulatory guidance does not make clear the difference between what is permitted up to July 10, 1996 and what is required by 10 CFR 50.65 after that date. The provisions which address (a) (1) and (a) (2) classification do not offer a clear explanation of the requirements of the Maintenance Rule for those licensees who will use the document on and after July 10, 1996. e 4-ir 9 gr =y-1 y -.w9 y .m
.s z. 1 INCORVECT UJE OF THE WORD "DilALL" A portion of 5 9.3.2 of the proposed regulatory guidance states that: " Failures resulting from surveillance and actual system demands phall be corupared to-at 2 aast the previous two f ailures of the same surveillance or previous two demand failures to assure that a repetition of a previous failure has not occurred." (emphasis added) (p.16, 1. 40-41 p.17, 1. 1-2). Regulatcry i guides cannot contain language which creates a requirement or command (such as use of the word "chall") except where such language communicates specific requirements established by the applicable rule or regulation itself. THE MEANING OF INHERENT RELIABILITY The proposed regulatory guide refers to the concept of inherent reliability in several places including 5 10.2 (p. 25, 1. 33-35) and S 9.3.3 (p.17, 1. 33-37). While the regulatory guide provides examples and some ph rases to desc.'be the concept of inherent reliability, the guidance f ails to art..culate an objective standard 4 against which inherent reliability may be deterttined. THE MEANING OF CRITICAL BAFETY FUNCTIONS CYSTEM PERFORMANCE REVIEW. " Critical Safety functions System Performance Re"lew" is a term contained in a list of mothods to be used to develop risk significant criteria in S
- 9. 3.1 (p.13, 1.
11-12).- The proposed regulatory guide does not contain a clear description of this term. +
I cc C. Ider !U 324 !!. Alderran P 315 P. Ihrarowsky M!BB 9112 E. Dockjoni ! U 007 K. Connaughton 0 01 16 H3 R. Correia G mi 10 D18 M. Dey 1 6 169 E. Doolittle Gml 16 H3 T. Toley O mi 10 A19 G. Grant G m1 17 G21 C. lleitemos ! U 007 D. Hic}mn MiBB 9112 S. liudson I U 007 C. Johnson IU 316 T. King tilS 007 M. Falsch G mi 15 B18 E. McKenna Gml 16 G15 W. Minners llLS 007 G. Mizuno Omi 15 B18 T. Marley G 01 12 G18 T. Ilovak MiBB 3701 P. O'Reilly M mB 9112 J. Page I U 302 C. Petzune G m1 10 A19 B. Richter ITIS 129 M. Ring Region III J. Roo GifH 10 H5 W. Russell Gm1 12 G18 B. St Paty MIBB 7714
- 11. Saltos Gmi 10 E4 J. Scutcrough G mi 16 H3 A. Serkiz NLS 314 L. Shao
! E 007 B. Shelton MiBB 7714 D. Smith 11UMARC J. Sniezek Gifti 17 G21 T. Stetka Region IV M. Taylor G ml 17 G21 S. Treby GiFN 15 B18 D. Trirble G m1 16 H3 A. Vietti-Cook CMFN 16 G15 J. Vora NLS 217B G. Zoch OWTN 10 A19 Public Document Room > file = numeet.276 300000 ..}}