ML20133N845

From kanterella
Jump to navigation Jump to search
Summary of 970109 Meeting W/Nei on Maint Rule Issues. Attendees List & Related Info Encl
ML20133N845
Person / Time
Issue date: 01/17/1997
From: Black S
NRC (Affiliation Not Assigned)
To: Matthews D
NRC (Affiliation Not Assigned)
References
NUDOCS 9701240005
Download: ML20133N845 (5)
v • d • e


Text

'

{k, [gRCEip\\ UNITED STATES Ij NUCLEAR REGULATORY COMMISSION s" C WASHINGTON, D.C. 2066H1001 January 17, 1997 MEMORANDUM TO: David B. Matthews, Chief Generic issues and Environmerc.si Projects Branch Division of Reactor Program Management O,b FROM: Suzanne C. Black, Chief t-ML ) Quality Assurance and Mdinte nce Branch Division of Reactor Controls and Human Factors

SUBJECT:

SUMMARY

OF JANUARY 9,1997 MEETING BETWEEN NRC AND NEl ON MAINTENANCE RULE (MR) ISSUES The NRC staff met with the Nuclear Energy Institute (NEI) and representatives from industry to reach a " common understanding" on the staff's position regarding the j iollowing Mr1 inspection issues: Technical Basis for Using Maintenance Preventible Functional Failures (MPFFs) as a e Reliability Performance Criterion i Use of Questions and Answers (Q&As) from the 1993 MR Workshops e e Scoping Timeliness of Dispositioning Structures, Systems, and Components (SSCs) from (a)(2) to (a)(1) e Technical Guidance for Monitoring Structures e Perceived Prescriptiveness in MR Inspections Technical Basis for Using MPFFs as a Reliability Performance Criterion NEl and the industry were concerned that the NRC staff had established a new interpretation of reliability performance criteria for safety (risk) significant SSCs that had not been previously identified as a generic safety issue during the pilot site visits. NEl and industry proposed to establish a common understanding which defines NRC staff f expectations for establishing reliability performance criteria that could be linked to [ reliability assumptions used in the Probabilistic Risk Assessment ( PRA)/ individual Plant Examinations (IPEs), 3 ey The NRC staff restated its concerns regarding the adequacy of reliability performance criteria established by some licensees and its proper link to the reliability assumptions used in PRA/IPEs as provided in a letter from Frank J. Miraglia, NRC, to Mr. Beedle of NEl, dated c a 0 0 8 + >--9 7 c E uu DW# A $ 7 j g fp d 5 k _ppspppps x w "" W 7

o i i D. Matthews : 1 1 l October 22,1996 (Attachment 1). Individual licensees should provide a technical justification that statistically links reliability performance criteria to assumptions used in the PRA or provide appropriate justification for any significant deviations from assumptions used in the PRA. The staff stated that using generic reliability performance criteria of two l MPFFs por fuel cycle on safety (risk) significant SSCs without an adequate technical basis was not acceptable. The staff reiterated that licensees do not have to count the number of actual demands for SSCs but licensees could estimate the number of expected demands for most SSCs given the Technical Specification surveillance test and post maintenance test frequency for safety (risk) significant systems. NEl and the industry stated that they understood NRC's position on reliability performance criteria and the industry would establish more appropriate reliability performance criteria for safety (risk) significant systems. NEl also stated that the industry would follow the guidance for establishing reliability performance criteria contained in Electric Power Research Institute (EPRI) Technical Bulletin 96-11-01, Monitoring Reliability fer the Maintenance Rule, dated November,1996 (Attachment 2). Additionally, the industry stated that some form of condition monitoring (e.g., channel functional failure rates) would most likely be appropriate for some safety significant SSCs (i.e., Reactor Protection System). The NRC staff stated that the approach to establishing and monitoring reliability performance criteria documented in the EPRI bulletin appeared reasonable; however, the NRC staff could not endorse this document. The NRC staff stated that other approaches to establishing and monitoring reliability performance criteria may also be considered acceptable. Use of Q&As from the 1993 MR Workshops NEl stated that the O&As from the 1993 MR workshops were used to determine NRC's position regarding activities that can cause MPFFs. NEl and the industry determined from these Q&As that the NRC staff did not consider operator errors to be MPFFs. NEl did not understand the NRC staff's current position on operator errors which cause functional failures that could be considered MPFFa. The NRC staff provided NEl and the industry with a letter from Suzanne Black, NRC, to Mr. Ray Ng of NEl, dated June 29,1994, regarding industry's use of the Q&As from the 1993 MR workshops (Attachment 3). The letter states in part that "the staff's responses could change as more experience is gained during the implementation of the rule. Therefore, licensees should understand that these answers represent the staff's current thinking and that information gathered during future site visits, future workshops, or other activities prior to the implementation date of the rule, July 10,1996, may affect these answers." Based on the information provided in this letter, NEl and the industry understand the staff's position on Q&As from the 1993 MR workshops.

J D. Matthews i i 4 l The NRC staff provided NEl with the NRC's policy statement on maintenance of nuclear power plants published in the Federal Register on March 23,1988 (Attachment 4). This j policy statement is referenced in NUMARC 93-01, Rev O, in the definition of maintenance. This document defines the activities that form the basis of a maintenance program. i i The policy statement states that activities that form the basis of a maintenance program - include surveillance, post maintenance testing and return to service activities. Based on this, the staff determined that operator errors that cause equipment failures should be considered MPFFs if the operators are participating in surveillance, post maintenance testing or return to service activities following maintenance. This includes operator errors l in removing a system from service prior to surveillance testing or maintenance. l NEl and the industr/ stated that they needed to explore this area further and make i clarifications to their own MR procedures to take these factors into account when determining MPFFs. j Scoping } i NEl was concerned that NRC inspectors were expanding the scope of the MR to include i non-safety related SSCs utilized in emergency operating procedures (EOPs) such as the emergency lighting and communications systems. i The staff considers the scopmg criterion of paragraph (b)(2)(i) of the MR to include not only those SSCs that are directly used to address an accident or transient or explicitly used in the EOPs, but also those SSCs whose use is implied and that provide significant fraction of the mitigation function. Examples of SSCs that should be considered include i communications and emergency lighting systems, which are a necessity for operators to successfully mitigate accidents, transients and use the EOPs, although they may not directly address the accident or transient or not be explicitly used in the EOPs. 1 i NEl stated that the licenseos and their MR expert panels should be given more flexibility in determining which SSCs add significant value to the mitigation functions of SSCs such as communications and emergency lighdng. However, they understood the staff's position on the matter and would inform all NEl members of the NRC staff's position on these SSCs. 1 Timeliness of Dispositioning SSCs from (a)(2) to (a)(1) NEl stated that the complexity of the cause determination and corrective action process have led to a timeliness issue with regard to dispositioning SSCs from (a)(2) to (a)(1). 1 i

D. Matthews. The NRC staff's position is that based. on a review of NUMARC 93-01, Rev O, the timeliness of dispositioning SSCs from (a)(2) to (a)(1) can be interpreted in a very broad manner, in NUMARC 93-01, Rev O, Section 9.4, states that "the results of monitoring (including (a)(1) and (a)(2) activities) should be analyzed in a timely manner to assure that appropriate action is taken." The timeliness of cause determinations and corrective actions in most licensee's programs should be linked to the safety (risk) significance of the SSC. The NRC staff believes that licensees' MR processes and procedures for cause analysis, cause determination and corrective action should include MR dispositioning activities to complete the proper link of these activities to maintenance rule required activities. The NRC staff recognizes that some cause analyses and cause determinations may involve difficult tasks to identify the exact cause for functional failures or MPFFs and under these situations licensees should be given flexibility to determine whether SSCs should be dispositioned from (a)(2) to (a)(1) with goals established. However, in cases where cause analyses and causa determinations are clear and straight forward, and the performance criteria has been exceeded, then timely corrective actions, dispositioning SSCs from (a)(2) to (a)(1), and establishment of goals should be completed during the same time period. In all cases, licensees should establish a reasonable schedule to accomplish all these activities. Technical Guidance for Monitoring Structures NEl also asked the NRC staff if the staff will add additional guidance in regulatory guide (RG) 1.160, Rev 2, " Monitoring the Effectiveness of Maintenance at Nuclear Power Plants," on methods licensees should use to disposition structures from (a)(2) to (a)(1). The NRC staff plans to finalize RG 1.160, Rev 2, and issue it by the end of February, 1997. The staff stated that RG 1.160, Rev 2, willinclude guidance which generally states that a structure should be placed in the (a)(1) category if (1) degradation is to the extent that the structure may not meet its design basis, or (2) the structure is degrading such that if the degradation were allowed to continue uncorrected until the next normally scheduled surveillance, the structure may not meet its design basis. The structure should remain in the (a)(1) category until the degradation and its cause have been corrected. NEl stated that they plan on revisi^ 1 their own industry guidance document, NEl 96-03, " Industry Guidance Document for Monitoring Structures," in the near future. Perceived Prescriptiveness in MR Inspections NEl initially thought that MR baseline inspections would only look at performance issues. After several NRC MR baseline inspections, NEl understood that the NRC staff's MR baseline inspection efforts were focused on program and performance issues. NEl has the perception that MR implementation was not going as smoothly as anticipated. NEl stated their desire to make MR implementation an excellent example of a successfully

~ D. Matthews implemented risk-informed, performance based rule which other rules in the future could i follow. They expressed concern that NRC inspectors are questioning expert panel decisions. The NRC staff stated that the MR has both performance based and compliance based aspects. The staff agreed that MR baseline inspections have focused on MR program i issues as was always intended and necessary. Licensees were using the guidance contained in NUMARC 93-01, Rev 0; however, some licensees took certain exceptions to NUMARC 93-01 to implement their program and each of these MR program exceptions nad to be reviewed based on its own merits. The inspectors are appropriately questioning the bases for expert panel decisions and other MR related decisions. The heade,uarters j oversight of MR baseline inspections is meant to ensure inspectors permit liccusees i maximum flexibility in implementing the MR. The NRC staff stated that the MR is one of the first performance based regulations. Industry and the NRC have very little experience whh these type of regulations. The MR baseline inspections have been somewhat focused on program issues because the staff believes that it needs to assess whether licensees have established adequate programs to consistently implement MR requirements. The staff expects that once a licensee has demonstrated that their MR program implementation is adequate, then NRC inspections i would focus on performance issues. This is also the current inspection method employed by NRC site resident inspectors to verify compliance to the MR. Attachments: As stated

5"84 9 i UNITED STATE 4 g g y NUCLEAR RE2ULATORY COMMISSION ~ f WASHINGTON, D.C. aoseMoo1 o Y+9 * ,o October 22, 1996 l Mr. Ralph E. Beedle Senior Vice President and Chief Nuclear Officer Nuclear Generation Nuclear Energy Institute (NEI) 1776 Eye Street, N.W., Suite 300 Washington, D.C. 20006-3706

Dear Mr. Beedle:

I am responding to your letter of September 30,1996, regarding your concem that the maintenance rule baseline inspections have identified a generic industry issue. Your letter described the NRC's position on the use of reliability as a performance " indicator," discussed the industry's choice of reliability performance indicators, and implied the NRC has established "new interpretations of compliance expectations through inspection and enforcement." I, too, am concemed that the industry and regulatory guidance developed over the last several years may not be completely understood by the licensees, even though the nine site pilot visits did not reveal such a problem. My concem is based on the fact that, in four out of the five maintenance rule baseline inspections (MRBis) completed by the NRC through October 4,1996, potential violations of 10 CFR 50.65 (the Maintenance Rule) have been found in the area of goals and performance criteria that the licensees have established for i reliability of systems, structures, trains, and components (SSCs). (A more detailed explanation of my concem is contained in the enclosure.) Paragraph (a)(1) of the maintenance rule requires that " goals shall be established commensurate with safety. " Although not a requirement, quantitative methods - with individual plant examinations (IPEs) or plant-specific probabilistic risk assessments (PRAs), for example, as the basis - hav. s:9en used to establish this required link with safety. Your guidance document, NUMARC 90-01, ties the requirements for goals (10 CFR 50.65 (a)(1)) to similar requirements for performance criteria (10 CFR 50.65 (a)(2)). NUMARC 93-01, Paragraph 9.3.2," Performance Criteria for Evaluating SSCs," states," Performance criteria for risk significant SSCs should be established to assure that reliability and availability assumptions used in the plant-specific PRA, IPE, IPEEE, or other risk determining analysis are maintained or adjusted when determined necessary by the utility." It is the lack of a clear i link to PRA/IPE/IPEEE or other reliability assumptions that is at the root of the NRC's concems. DO k oh- _q R1 tg

I 3 l 4 j- - R.E. Beedle October 22, 1996 i i i Maintaining the link between reliability astumptions in the plant-specific risk-determining i analysis and the performance standards under the maintenance rule has been a longstanding j position of the agency. For example, the meeting summary regarding the April 22,1992, j i NRC/NUMARC public meeting for developing maintenance rule implementation guidance i noted: 1 "The industry Guideline allows but does not recommend or require the use of IPE/PRA results for performance criteria or goal setting. The NRC believes the j maintenance (monitoring) results should be used to confirm performance and conditions (including component and train availability and reliability) in available IPE/PRA and other safety analysis results." 1 As indicated above, this concem was resolved prior to the NRC's endorsement of NUMARC 93-01. The maintenance rule is a risk-informed, performance based regulation that requires licensees to provide reasonabie assurance that SSCs remain capable of performing their I intended functions. The NRC does not expect licensees to perform highly sophisticated, rigorous analyses to demonstrate that reliability performance criteria are mathematically equivalent to the values used in PRAs. Rather, our expectation is that licensees provide a i reasonable and appropriate technical basis for selecting performance criteria to meet the regulation. However, it is expected that such approaches would incorporate some consideration of demands for standby systems and service time for normally operating systems. Acceptable approaches exist for linking performance levels to safety (risk). During the nine pilot site visits performed to review early implementation of the maintenance rule, reviews of the licensees' goal-and performance criteria-setting processes were performed. As stated i above, the inspectors found that licensees did understand the issues related to developing performance standards for reliability that were linked to safety. Several of those licensee programs described in significant detail the link to safety (risk) and justified the use of functional failures in the measure of SSC reliability. Therefore, the issue was not raised in the trip reports or meetin with NEl, since none existed. o In short, the NRC's position has been, and is, that performance standards - goals and performance criteria - must be demonstrably linked to safety, and our enforcement decisions will continue to be made based on licensee compliance with 10 CFR 50.65. l

~ ) e R.E. Beedle October 22, 1996 As requested by your letter, a public meeting between the industry and the NRC was arranged and held on Tuesday, October 15,1996, to discuss this issue. During that meeting, the staff and NEl agreed that additional guidance t:, the industry is warranted. On October 16,1996, at an NEl workshop, discussions took place among industry participants to propose approaches to solution of the issue for further consideration. I anticipate that guidance on this issue will be promulgated by NEl at the earliest possible time so as to give those licensees that may not currently have acceptable reliability performance criteria the basis for making the necessary adjustrnents in their programs. Sincerely, Dri inal signed by B Frank J. Miraglia Frank J. Miraglia Acting Director Office of Nuclear Reactor Regulation

Enclosure:

As stated cc: Thomas E. Tipton Vice President, O&E Dept. Nuclear Energy Institute (NEI) 1776 Eye Street, N.W., Suite 300 Washington, D.C. 20006-3706 DISTRIBUTION: Central Files /PDR HQMB R.F. DRCH. R.F. SEE PREVIOUS CONCURRENCE DOCUMENT NAME: G: BEEDLE.LTR Ts recenre e copy of this do*ument, indicate in the boa: "C" = Copy without attachment / enclosure

  • E' = Copy with attachment / enclosure

'N' = No copy 0FFICE HOMB:0RCH l SC:HOMB:DRCH l BC:HOMB:DRCH DD:DRCH A D:DRCH O" NAME WEScott:cct RPCorreia SCBlack RLSpe3stfrd BABoger l> DATE 10/18/96* 10/1 3.0/18/96* 1 /96 g D: AD NAME G@>f ahan " BSher W ACTMdani F#)Waglia DAi~E 10/ s /96 10/fy/96 C40/[J/96 10W /96 10/ /96 OFFICIAL RECORD COPY o

1 l THE RELIABILITY PERFORMANCE STANDARD 2 NUMARC 93-01 defines reliability as "(a) measure of the expectation (assuming that the SSC is available) that the SSC will perform its function upon demand at any future instant in time." Numerically, for normally operating SSCs, reliability is the complement of the ratio of the expected number of failures to a given time of required performance. The ratio of functional i failures to a specified number of operating hours could be shown to describe a reliability level: ) that could be related to the plant-specific PRA/IPE/IPEEE or other risk-determining analysis. i Likewise, for standby SSCs, reliability is the complement of the ratio of the expected number of failures to a given number of start demands ard. once started, run demands. The ratio of functional failures to a specified number of attempted starts and attempted runs could be shown to describe a reliability level that could be related to the plant-specific PRA/IPE/IPEEE i or other risk-determining analysis. In four of the five maintenance rule baseline inspections conducted thus far, the licensees j used maintenance preventable functional failures (MPFFs) over time as their reliability performance standard. For normally operating SSCs, that performance standard could be acceptable if it described a satisfactory relationship to plant-specific PRA/IPE/IPEEE or other risk-determining analysis. The onus is upon the licensee to demonstrate the satisfactory nature of that relationship, and those four licensees had not done so. j More of a problem, however, was their use of MPFFs over time as a performance standard for standby SSCs. As described above, the reliability calculation for a standby SSC must incorporate both failures and demands. All four licensees failed to incorporate demands in their calculations and, therefore, used unacceptable performance standards, clearly not demonstrating a relationship to plant-specific PRA/IPE/IPEEE or other risk-determining analysis. Enclosure

j 1 3 ji l ( I l ,i 6 i MONITORING RELIARUJTY j FOR THE MAINTENANCE RULE i EPRITechnical Bulletin 51141 i i i November 1996 i i 1 l i l .i 2, i f 1 i i j i j Papamd by Applied Reecatch M-

=ent 1

I Canales, Nuw M dm ~ EPRIProW Manager l J.u.cink= l i l t I +e* e emaue edess = ~

t. 1/ [ EPRITECHNICAL BULLETIN 96-11-01 i MONITORING RELTARUJTY FOR THE MAINTENANCE RULE i JohnGiscion EPRI j; Ihvid Worledge ARM ii November,1996 i

===. Background=== j Most nuclearplant licensees have chosen to monitor the number of Ametional failures or maintenance preventable functional failures for systems, structurcs, or components that i j require reliability to be munkored under paragraph (s)(2) of 10CFR50.6, the mamtenance rulc. In four out of the first five maintenance mic baseline inspections, potential violations of 10CFR50.65 have been found that relate to the way miiability se i being monitoisd. Specifically, the inspection fmdings have found that there is not an j adequale Enk between the performance criteria expressed in terms of failures per 1-operstmg cycle, and IPF/PSA mesmnptions and data. Licensees have expressed concem j that the solution to this conflict will require them to track the number of demands experienced by each standby SSC with a failure performance criterion, and to monitor the probability of failure-on-demand, rather than simply to monitor the number of failwes. This ischnical bulletin describes a process, and its techriical basis, that would address the 3 i NRC cr~rns by establishing a quantitative relationship betwccri the performance i i-crikaa.nd PSA data, without requiring that demands be tracked, and which woukij ) the practice of monitoring failures. The process has been describcd before, m the EPRI repoit TR-106280, Insights Fmm EPRI Maintenance Rule Projects, of May 19%, Section 6. " Technical Basis for Performance Crireria". 1 i i i Outline of the Link to FSA i it will be demonstrated, below, that if is not techniently possible to nwnitor an individual SSC's rollability by monitanng the number of failues mul dernands that it exings over a time as abort as one operating cycle (exceptions: hsy Diesel Generators, Ma= tbsy are tested monthly, and other SSC's with very fmquent test or domsad achedules). The reason is that any ostimate of the pmbability of railure-on-demand horn i such data will be too uncertain to draw useful conclusions ham it as to whethe probability of fshiro-on<lernand is in reasonable consoaance with the value used in ibe iPE/PSA. Renannable estimates can be made but only by increasing the time duratiou to molude a grooter number of tests, mud /or by including data on the performance of other shnitar SSC's from other trans, systems, or other units in the same plant (site). Sneh estimates are very suitable for PSA purposes but will not serve 10CFR50.65 which aquires individual SSC's to be monitored over singic operating cycles. l , qi e y9,hJaq,.eu'h4Mi sp WM M

j.

ii!

De recent response from Mr. Frank J. Miraghs. NRC Acting Director of the Office of

I.

Nuclear Reactor Regulation to Mr. Ralph E. Beedic, Senior Vice President and Chier Nuclear Officer of the Nuclear Energy institute, dated October 22,1996, contams a i smanbcr ofqualifications that indicale that the required consistency with IPE/PSA i assumptions can be achieved by the EPRl-reconunended process for establishing i' perforenance criteria on reliability, despite the imponibility of diroetly anonitoring the i probability of failuso-on-denumd. Mr. Miraglia notes that althoush a clear link to IPE/PSA assumptions is required, this may be achieved by usag the results of monitoring to confirm the perfonnance or oondition in the IPE/PSA. He also notes that Die NRC does not expect highly j sophisticated or rigorous analyses to demonetrase this confirmation, but a reasonable and i appropriate basis with some consideration of ht for standby systems ami service times for norrnally operstmg systems. i Estlanating the Probability of Failure-os-Demand i If r failures arc experienced in n tests. the best estimate of the probability of failurr-on-demand, P, is P=r/n. In a period of two years an SSC that is tested quarterly would j expericacc only 8 test 4-'. For longer periods between tests, the number ofIcsts is l smaller. Mors tests might be included for operational reasons. for preparing the SSC for testing. or as post maintenance functional tests. Consequently, the samber orlogitimate j demands for an SSC that is tested quarterly is not likely to exceed about 20 per cycle. j j i Some SSC's covered by 10 CFR 50.65 are tested much less frequently thr. quarterly so that their estimates ofreliability might need to be based on four tests, or even fewor. 1, Page 4 shows results calculated using the binonnal distnbution for an SSC that j experiences up to 5 failures in 10 tests. Page 5 shows similar results for20 tests. The binomial distribution is universally acknowledged as the correct model for devices dan cxperieuw sandum failures with a constant probability of failure at each demand (e.g. j tossing a coin), as assumed by many IPE/PSA's. The results for hnth 10 and 70 test demands include two charts. The first chart shows the antic of the value "best estim plus two standard deviatious" to the best estimate. De second shows the ratio of the upper 80% and 90% coniidence limits to the best estimate. Dese ratios indicate the degree of precision with which statements can be made about tbs probability of failure-on-demand. Ahbough the tables and charts show results for up to i failures it should be runemberod that the expected nuenber of failures for nucicar plant SSC's over a period of a your or two will be close so zuru. The expected number of falhses isjust the IPE value for the probabihty of failure onAmmmad tunes the..umbercf damads. nis numher will typically he of order 0.1 (~ 0.01x10), or less. His means that, in agrocment with experionoc, tbc actual number of raitura on a specific SSC over one operstmg cycle is i 2 . ~ _ ~... _. - ~ - - -

_,.e,-

-w-*---

4 . 2

{

monly zem, with occasionally one, or possibly two failures occurring. It is the resuks for one failme that are the main focus of attendon here kr.nue two railures w recotananded as being unocceptable and constituting an cxecedance of the perfonnance ailmon. 1 i 1 i i s 1 i i 4 e 2 -l 4 h. i 4 i I i a 4 3

i! Suundard Deviations and ConSdence Bounds for 1 to 5 FaDures le 19 Test i. Destad$ l' l if Testetam) r s Perin r (P+e)* r (P+2e)* 4 Number of failurea = r 1 0.095 0.1 1 1.949 1 2.697 EsGmate of prob. of 2 0.126 0.2 2 1.632 2 2.265 failure on demand = 3 0.145 0.3 3 1.483 3 1.966 i P = r/n 4 0.185 0.4 4 1.347 4 1.775 e is One standerd 6 0.168 0.6 6 1.316 5 1.032 i rimwietion in P 1 4' r L80% IJS0% P=rAn U8t%/P ' U48/L50 000% P=rtn Us4%/P ] I i 1 0.01 0.337 0.1 3.370 33.70 0.394 0.1 3.940 l 2 0.066 0.45 0.2 2.250 8.18 0.507 0? 2.535 i 3 0 114 0.552 0.3 1.840 4.76 0.807 0.3 2.023 j 4 0.188 0.648 0.4 1.015 3.44 0.697 0.4 1.743 5 0.267 0.733 0.3 1.466 2.75 0.778 0.5 1.556 1 I l 1 Ratio of Measi Plus 2. Standers Radio of Upper Conndonee Bounde Deviseens to the tiean Value to the Moen Value " 'y [ 3 l N 1531 p!5' 1 2alo. _. + - - g'u r .tr ~ - m _s_d

. a2g;3 1

MM., = gdgrg! = mges, w tan L e+3wl 1 2 3 4 A 1 2 3 4 5 Puaises 1 Fatues i l 1 i i 1 1 4 i a J i l ~u ,o

i i i Standard Deviations and Comfideem Bonads for i FR 5 Fash-is b 20 Tm Eggands l i 1 { 28 Testa (un) r s Parin r (P+s)/P r (P+2s)/P Number of fadures : r 1 0.049 0.05 1 1.975 1 2.949 ) Estimate of pmb. of 2 0.067 0.1 2 1.871 2 2.342 dailure on dementt = 3 0.080 0.15 3 1.522 3 2.085 P = r/n 4 0.089 0.2 4 1.447 4 1.894 i s is Die utmeidard 5 0.097 0.25 6 1.387 5 1.tre i denotion in P l 1 r L 80% U 80% Par /n tiEG# U80/L80 USO% Par /n U40%/P 1 0.006 0.182 0.1 1.820 A0.33 0.215 0.1 2.150 2 0.027 0.246 0.2 1.226 0.07 0.203 0.2 1.410 3 0.050 0.304 0.3 1.0 13 5.43 0.344 U.3 1.147 1 4 0.09 0.361 0.4 0.903 4.01 0.401 0.4 1.003 5 0.127 0.415 0.5 0.830 3.27 0.456 0.5 0 #12 i Ratio of 80ean Plue 24tandard. Ratio of Upper Confklence Bounds Deviations to the Mean Value to the Mean Value i ~ atED 2300 g-yrne; '~ 181D -, $ ^. 1810 k.. . M'rJ!! l-lP+2syPl

. 2 -

2000 - F., 1J00 g[- N w- _ g;w-- 2 3 4 6 1 2 3 4 s resume Fasmus J ) 4 1 i

If aero, one, or two failures occur, both the mean-plus-two standant deviation < limit, =< l well as the upper confidence limits, permit estimates of the probability of failure-on. demand that coctend a lhetor of 2 to 4 above the best estimate. Even wider constrants apply below the best esthnste as can be seen fmm the ratio of the upper 80% bound to the lower 80% bound, which is gmeer than 30 if one failure is obacived, and is 8 or 9 if 2 failmes are observed. These results show that the observance of zem, one or two failum in 10 or 20 tests provides a poor capability to conshain the value of the probability of failure-en-demand; the chscrmunating power is so weak that it could not be used sensibly in any monitoring scheme. Ifmore than 20 ist desnands are necumulated the precision improves, so that at least the occurronec of two failurca begins to be a practical predictor of the piubability of fkilute-on-demand. The conclusion is that even if demands wore tracked, they could not be used to provide useful estimates of the probability of fallute on-demand in the monitonng processes of the mainicaance rule. Of c,ourse, after several operating cycles have passed, the precision for any individual SSC will improve if all the data for the whole period since the start of the rule is puuled togedwr. However, this will only provide an estimate of the average performance over the whole period and still will not indicate the perfonnance over the most recent cycle. Expected Number of Failures The situation is not quite hopeless, however, because a quantitative link with the IPE/PSA valuc can be obtained by asking wlmat that value implies about the probability of actually observing a specific manber of failures, rather than asking the question the other way round, as above. The hinomial denalty function gives a simple way to calculate the probability of 0,1, 2, or more failurcs. For r failurcs in a denuunin this function is (note that P is distinct fmm P,(r)): P,(r) = g . P'. (1 - P)*d r!(a r)! So that: P,(0)- (1 - P)" P,(1)- nP(I - P/** 1 2 P,(2) = n(n - 11. P. (1 - P)** 2 wbcrc P is the probability of failure-un-demand used in the PSA. b

  • amass =es-me de e mee e
  • +1e w

. ma mw..-we r =

4 j The fallowing table shows the pmhability of observing zero Iailures when 10 and 20 tests j are perfonned. Nasaber of Probabluty of Zero FaGeres 4 Tests When1PPJPSA Valueis Whan IPE/PSA Valueis P-0.01 P 4.001 10 90.4 % 00.0% 20 om 98.0 % I The following table shows the probability of observing a single failwe when 10 and 20 tests are performed. Number of ProbabiMtyof One FaMare Tests WhenIPE/PSA Yalueis When IPE/PSAValucis P=9.01 P=0.001 i 10 9.1% l.0% 20 16.5 % 2.0% Although the first table above shows that acro failures is by for the rnost likely outcome in each case, it can be seen that there is a 1 % to almost 20 % chance of observing one tatlure. 'Ihc chance of obscrying two failures, however, in much smaller than the chance of j observirig a single failure. The following table shows the pmbebility ofobamving exactly t two failures when 10 and 20 tests are performed. Nussber of Prubabaltyof Two FaNares Tests WhenIPBJPSA Valueis WhenIPE/PSA Valucis i P-9.01 P=0.001 4 10 0.4% 0.0045 % 20 1.6% 0.019 % The results show that for most cases ofinterest a single failure is many times more likely than two failures. 4 We have seen that even when the underlying probability of faihne-on<iemand is in the ranga 0.01 to 0.001, one failure will be experiemd by 1% to 16% of SSC's in one cycle. We have also seen that the best estimatos of the pmbability of failure-on.denaaml frun this experiesco am in the range 0.05 to 0.I, and reasonable upper bounds are 0.2 to 0.4. 7 l -,--4. w

a f f Dis means that a monitormg proces that tried tu us,tiumte P on the basis of those sesults l / can be incorrect by a factor of 20 to 400. This is father evidence that trying to estimate the reliability from the number of failures and demands is an unsuitable way to address j maintenance eflixtivstam Performance Criterion on Failures Frern the analysis it can be seen that single failures can casily ocent given the likely PSA input values and the large number of SSC's tha: are being monitored, but that two failures abould bc quite rare. This conclusion applies for a wide range ofvalues of the number of tests and IPE/PSA values of the probability of failure-on<lemand. De conclusion becornes less valid as the probability of failure-on-demand approaches 0.1 (the chance of two inihase becomes signi6 cant), and as it decreases below 0.001 (the chance of one fhiture becomes loss than 1%). However, the conclusion will remem valid fbr a large fraction of the SSC's in the mantenance rule. he conclusiva supports perTurummco criteria such as "I failme cau uccus,2 failmes is au exccodance", or "2 failurcs can occur,3 failures are an exWaw".1he pha of such criteria should n! ways he checked assinet the actual IPE/PSA value and the member of k# =" desnands to be expected in one operating cycle. His is the vital link with the IPE!PSA assumption. tiowever,it must be stremed that if the critens are set accordag to these requirements they will remain apperpriate mtmafor a wide range ofvahats ofthe number ofdemands (e.g.pom 0 to more than 20 demareds). %csc will bc no addod vsluc in closely monitoring the number of demands unless it exceeds a minimum of at least 20 (the minmuun number daaaak on the IPE/PSA value; for P=0.00I the rninimum would be many hundreds of demands). As shown earlier in this paper, for small numbers of demands there is no way to make uw of the exact number when only 0, I or 2 failures are likely to occur. Summary 11 is not possible to monitor the reliability of most SSC's over a period as short as two years. nis is because, even if the exact number ordsmauds were known, a result ur 0,1, or 2 failums would not pemut meaumgful bounds to be placed on the probability of failuro on demand for the purpose of comparinan with the IPF/PSA input value %is conclusion depends mostly on the low values of the nusabor of faihres involved, and much less on the number of danswin providag this is below about 20. The conclusion is not sensitive to whether standard deviations or con 6dcoce bounds are used it is not nonsitive to the value of con 6dence assumed (two.eided 80%, and 90% bounds in the calculatioas abuse), aud thu in uut sousitive in whether uncHdded or two sided bounds are used. 8

~.- ----- --...-. -.-._--. 1 Instead, the chance orhing 0,1, or 2 failures can be r@ W'-J using the IPE/PSA inputvalue, and the expected number oflegitimate demands. A failme criterion should be 1 - selected that acknowledges that possibly 1, or in some instances 2 faGures mighi occur, consistet with the IPE/PSA input value, but that the chance of additional failures should be very mucli less. One failure can occur rendomly within an operstag cycle even when I pseventive maintenance is perfarned effectively, because many factam concerning nervice conditions and rates of degradation can not be known with censinty. However, if two failures occur, such pertonnance enteria would indicate that these failures are very unlikely to he random events, and probably represent a trend toward poor performance requiring appropnate cause analysis and corrocrive action. In most cases the estesand number of demands is quite su0icient for this calculation because the above conclusions will romain true for a wide range of numbers of dem==le. No added value is provided to the maintemoce rule monitoring process by deriving a detailed knowledae of the number of demands. \\ 9 m mp se 4 e-******&a

    • we' e' me w -.

.d +msee ses

  • = -

w ter-h*

==--*=edhsN ueeh =W emmWW84DSame at +dibe e6 + *

  • i..

am / UNITED STATES j NUCLEAR REGULATORY COMMISSION l *r]j e WASHINGTON. D.C. 21M55-0001 l ()g%,*****f June 29, 1994-2 1 Mr. Ray Ng, Manager Licensing and Performance Based Regulation Nuclear Energy Institute SUITE 300 1776 EYE ST NW WASHINGTON DC 20006-3706

SUBJECT:

FINAL NRC-STAFF REVIEW OF QUESTIONS AND ANSWERS FROM THE AUGUST 1993 NUMARC MAINTENANCE WORKSHOPS-The NRC staff has completed its review of the questions and answers documented by NEI from the Maintenance Workshops held in Atlanta and St. Louis in August 1993. Based on the staff's attendance at the workshops and subsequent discussions with NEI representatives, the staff agrees that the written answers prepared by NEI are appropriate responses to the questions from the workshops. The NRC staff believes that making these questions and answers available to the industry will promote a better understanding of the maintenance rule. However, licensees using these questions and answers as guidance should understand that because some of the questions were very specific in nature the answers to those questions may be very limited in their applicability to other licensees with different plant or equipment configurations. Licensees are cautioned to use the entire set of questions and answers as an aid in understanding the intent of the maintenance rule and not rely on individual answers to provide the final determination of acceptability. In commenting on the answers to the questions, the staff used the best information available at the time of its review. The staff's responses could change as more experience is gained during the implementation of the rule. Therefore, licensees should understand that these answers represent the staff's current thinking and that information gathered during future site visits, future workshops, or other activities prior to the implementation date of the rule, July 10, 1996, may affect these answers. Licensees who need guidance should refer to the rule,10 CFR 50.65, and the Regulatory Guide 1.160, which represent official NRC positions and to NUMARC 93-01, which was endorsed by the staff in Regulatory Guide 1.160. Licensees should note that some questions have been revised or combined with other questions from the workshop to clarify or illustrate an isst.e. Therefore the questions may differ slightly from those asked at the workshops. +m I

f C_9 O dCOf2 ROSE 6 ides. & v Cdmc w.sn.uus ,.isg.. 188.234 as Stat.gE 83 Stat.444 as amended gypsCTws matthis FinalPahey "ActivitiesWind Forts the Basis of o 142 UAC 2238,23s21; sec. ans as Stat.124s Statement is effectwo Marts 23.1988. Maintemenee Progreta." 1 (42 UAC SM6L Sections 2.500-2.00s also pon Mnut aposusanon costrace. as.t.a===== Progreens leeued under sec.102. Pub. L 91-190. as Stat -Jad W. Roe. Director.Dtvielen of ~ } 853 as amended (42 U.SA 41721. Sections Doensee Performance and Quality Each constreroual nuclear power plant e l 2.7cos. 2.719 also issued under 5 UAC. En i Sections 2.754. 2.7eo. 2.no eleo seemed undee 5 Evaluation. Office of Neclear Reector should develop and implement a well-dasa=d and e5ective program to assure l UAC ss7. Section 2.7so also ieseed under Regulation. U.S. Nucieer Regedetary sec.103. es Stat. sas. as amended (42 UAC Commlaaion. Washington.DC sesss, that maintenance activities are conducted to preserve er restore the 2133) and 5 UAC 552. Sectione 2.400 and telephone (301) 402-1004. availability, performance and reliability 2.aos also issued under 5 UAC 563. Section

2. sos also issued under 5 USC 563 and sec.

Pokey of plant structures, systems, and compements. W program should clearly as. Pub. L 85-25s. 71 Stat. 579 as amended Backsmand (42 UAC 3osel. subpart K also taeved & components and aMties i sec.1es es Stat. ess (42 UAC 223e); sec.134. W Conunission has a program 1o "*"*8"'"! 4 puh. L 97-425. se Stat. 2230 (42 UAC 101Hl. continually evaluate the operational systems used to control those activities. { Appendix A sho issued under sec. a. Pub. L puformatice Mauckar poww planta. Further. the program should include 91-8e0, se Stat.1473 (42 UAC 21351'. pub. L Analysis doPemMonal even& feedback of specific results to ensure i Appendix 8 also issued under sec.10 shown &at.In some cases, na corrective actions, provision's for overall ( SEL240. Se Stat.1842 (42 UAC 2021b et seg). power plant equipenent is not being Pmgrain evaluauon. and ee 4

2. Section V.F. of Appendix C is maintained at a level which ensures.

scation dposs% componenw i revised to read as follows: with a high degree of reliahility, that the Pm Pmh equipment will perform its intended APP **** N 8*"""' *I P"Ec7 function when required. A lisaited NRC Activities Which form the Basis of a

  • "d ***d** I*',MC Enforcement Assions examination of nuclear power plant Molntenone Program maintenance programs has found a wide An ad unte pmpam should variation in the efectiveness of these ide' l

V. Enforcement Actions * * *

  • TechnologyJa the areas of E Aeopening ClosedEnforcement Actions entshas a

an --Cwmetin snanntenance, l If significant new information is received or contributor to plant reliability kW eins -Preventive maintenance. obtsmed by NRC which indicates that en end. hence,is d safety concern.h --Predictive maintenance. enforcement sanction was incorrectly Conunission believes safety can be --Surveillance; applied, consideration may be given. enhanced b ' ro ' the

  • Engineering support and plant NOPening effectivent ma enance programs f,

",I (th' j;t modifications; throughout the nuclearindustry.%e correct the record. Reopening decisions wit! C"""i"*" 18 Proceedmg with

  • Quality assursace and quality decrease the seventy of a sanction or to be made on a case.by. case basis. are rulemaking consistent with this beliel contml; expected to occur rarely. and require the his policy Statement is being loseed to
  • Equipment history and trending:

specific approval of the Deputy Executive provide guidance to the industry while

  • Maintenance records:

Director for Regional Operetions. the rulemaking proceeds.

  • Management of parts, tools, and -

farItifime

  • Procedures:

Dated at Washington DC. this 17th day of It is the objective of the Commiasi n

  • Past-maintenance testing and March 19ss.

that all components. systems and retur.t-to-service activities; For the Nuclear Regulatory Cosamission. structures of nuclear power plants be . P4easstres of overall program 3,c,, ,p %.. maintained so that plant equipament will eNectiveness, s,

3.c uik, i$*in the a

[FRDoc seh8333 Filed -as. s:45 aml req Ta p h thi objective, org o

    • "8 8888 F"6**

each licensee should develop and pg.,,g,,, implement a snaintenance program fe m which provides for the penodec ,Jr,,,n" evaluation, and prompt repair of plant -Shift c vwage, M CFR Pat 2 components, mteses, andstructwas to --Resoura allocation: ensure their avallabihty.

  • Control of contracted maintenance pinal Commission poncy Statement on Donaltion d h services:

RAalntenance of Nucteer Power planta

  • Radiologicalexposurecontrol h r ommi..iant defines mala aamar=*

n meescv: Nuclear Regulatory as the aagregate of those 8==r* (A1.ARAk required to preserve or restore safety. e personnel qualification and training:

  • latesmal comununications between Commission.

reliability, and availability of plant Actioet Final policy etstement. structures, systems, and -: = =ts, the mainte===a= organization and plant sueessamv: W Commission beheves Maintenance lachsdes not caly activities operations and support youps, safety can be enhanced by improving traditionally manaelated with identifying

  • Casamunications between plant and the effectiveness of maintenance and wskiactualer potential corporate management and the programs throughout the==r4==

degraded conditions,14, emper, mainteamaamargamination. Industry.W Commission is pmceeding surveillanca t=P exassinations. Maintenance =a=====,t.eia== or with rulemaking consistent with this and preventive measures but extends to requirements ofindividual vendors behf.This Policy Statement is being all og Aq functicos for the conduct shound receive appropriate attention m issued to provide guidance to the of these activities.%ees activities and the devalarn===e of the anaintenanos industry while the rulemaking proceeds. functions are listed below modar program. Attactuaent 4 L

Federal Reg'ister / Vcl. 53. Ns. 56 / Wednesday Varch 23, 1988 1 Ildes and Regnbtlana 9431 Future

=====8==1= Actica erroneous PT speed signal during ground one enWne as a failure of the other r De Commissionintends this Policy operation wie the bottoming gownor engme. Dis AD supersedes AD e6-W Statement to provide guidance to the (BG) enabled. St. Amendment 3D-6473 (51 FR 44439;. DAfsa: Effective-44ay 9.19es. December e.1986). industry in improving maintenance programs for their power reactor Compliance.cMule-As prescribed laterested persons have been afforded facilities. Th'e Commission will continue in the body << tlw AD. an opportunity to participate in the to enforce existing requirements 18corPorstion by Aeferenos-making of this amendment.No Approve.s by the Director of the Federal comunents were received. Accordingly, including those that address maintenance practices and will take Regh*a as of May 9.1988. the proposal is adopted without change. whatever action that may be necessary ADtanssesa:The applicable service AD as-10-61. Amendment 30-6473 (51 to protect health and safety, bulletins (SB's) may be obtained from FR 44430). Issued November 18.19e8. The Commission expects to publish a Dowty Rotol Limited. Cheltenham Road requires that the engine BG be disabled Notice of Proposed Rulemaking in the East. Gloucester. England GL2 eQH: when the aircraft powerlever is r naattioned in the beta range (below near future that will establish basic GeneralElectric %=p==y.1000 requirements for plant maintenance Western Avenus.Lynn, Massachusetts liight idle).%e AD was needed to programs. We believe that the contents p1910; and Saab Scania AB, Last as-prevent FT overspeed and resulting and bounds of the proposed rule will fall "*"""- Sweden. uncontained failure caused by reaction within the general framework described A copy of each SB is contained in of the fuel control to an erroneous PT Rules Docket Number as-ANE-21. in the speed signal dunas groimd operation in this Policy Statement. Consideration will also be given to Office of the Regional Counsel. Federal wie the BG enaWW. industry wide efforts that already have Aviation Adaninistration.New England AD e6-10-81 pmvides interim been initiated. We encourage intereste.i Region.12 New England Executive Park. g,% g p vat FT ovmped parties to provide their views on this Burlington. Massachusetts 01808, and a d m m W i h. Si m h important subject to the Commission. may be exammed between the hours of imuuc6cm @ spuid airush ad a.m. and 30 M y through engine opereung proedures which even at this early stage of the rulemaking process. Any notice of y,excep ys. increase crew workload and invalidate proposed rulemaking that is published Pon PustTvenst mercanaAftoes coorfact: the constant torque on takeoff functMn will provide, of course, a period for Barbara Garian. Engine Certafict. tion the FAA has deteramed that a second public comment on its contents. Branch. ANE-141. Engine Certification overspeed protection system with an Sce. Airush Ce uncahn Mon. improved level of safety precludes the Deted at Washington.DC this 17th day of Federal Aviation Administration.New need for these interim instructions and March.1ses. og an Repon.12 New England noturns the aircraft and engme to pre-For the Nuclear llegulatory Commission. Executive Park. Burlington. AD as-10-51 operation. i Samuel J. CWut. Massachusetts 01803: telephone (617) f%"d"=3a= Secretary of the Commissiott 273-7006, p Doc. m N A M tesa N Suerl.tansarTAsty sospoemeATeost A De FAA has determined t'nat this euse caos **** proposal to amend Part 30 of the Federal regulation affects 107 aucraft all of Aviation Regulations (FAR) to include a which are in liance wie els AD. mw D maluiring the instaHation of a g,,go,,,3 - that this action (1)is DEpARTIGENT OF TRAftSpORTATl000 second overspeed protection systees on not a " major rde" under Executive certain GE CT7-.5A series ; t.,,, -"~ Order 13331:(2)is not a "signincant Federal Aviation Administration ensmes as installed in Saab-Fairchild rde" under DCyr Regdatory Policus SF340A aircraft was p=hi'=had in the and Procedures (44 FR 11034; February 14 CFR Port 39 Federal Regleter on October 14.1987. (52 as.13F9):(3) does not warrant

  • AE84 A'#8t 88-FR 3e458).

preparation of a regulatory evaluation j I,, g De proposal was prompted by en as en an6cipendimpact is minimal; engine Pr omspeed and resulung and (4) will not have a significant Airwortfeness DirecWves;Generet uncontalmd failum. caused by reaction .,ana=&c impact. positive or tive. Electric (GE)CT7-5A -5A1,and-5Ag of the fuel control to an erroneous PT on a substantial number of entities Turtsoproponer Engines as insteRed in speed signal during operation under the criteria of the Regulatory SeatWalrchtd W340A AirersR with the BG '" PkxlWBty Act. Aamecv: Federal Aviation Since this emdition is likely to exist Administration (FAA). DOT. or develop on other engines of the some IJet of Subjectsla14 CFR Part se type design, a mw AD is being leased Engines. Air transporation. Aircraft. AcTeose Final rule. that requires installation of a second Aviation safety. Incorporation by guesasAsty:his annendment adopts a overspeed protection system on GE ,,g,,,,,,* new airworthiness directive (AD) which C17-5A series ; E 1. "' engines as requires the installation of a second instaBed in Saab-Fairchild 3F3enA Adoption of the Amendment overspeed protection system on certain aircraft.his AD also requires ' " parenset 2 de au&ority ^ g,jegated to,,s's's, th,e Federal A tion of engine BG descevadon GE CIT GA series ; ', 7" nd eneus as imiand m Saab Fairch .wii in es,ower iever,marmt = A,m,,i,,,, ,(, A m to SF340A aircre't.his AD also prevent an aberse yaw condition in te amend Part as of ee Fderal Aviadon supersedes AD es-to-81. Annandment aircraft that could ocent due to a Regula6am (FAR) as h 3D-6473 (51 FR 44430; December s.1988). mismatched aircrah power conditica his AD is needed to prevent engme resulting from an==ca==aad=d PART 3D -(anssaansrij power turbine (PT) overspeed and lacrease of one engine.his also resulting uncontained feuere caused by prevent the crew froan misinterpreting 1.The authority citation for Part as reaction of the inel control to en the une====nded power increase of continues to read as fouows-

ATTENDANCE SHEET DATE: 1/9/97 SUBTECT: Maintenance Rule Implementation ORGANIZATION: h@l LOCATION: Rockville. MD ORGANIZATION NAME (prin0 TrrLE (print) TELEPHONE AFFILIATION (prin0 ye a-s Pes.sAtay remcrr >Recu qis -iye y Alcue / b y n %2Anae blAdL %u c ther-w r.,oi, puti3,w+ //mu JK 5t44cucmua m.ta. a n a s w /m, Ne/t;sa /wS Soc]t0orrei SacErm O. e V 4i C-tso 9 A)kl'utcYfuri5 ~T5ha Pie.hugeto Jir., Licatsi, 202. at s/ VE7 ih);.(4kreis z.oz 73%d75 Ajr1

  1. Awrfp1A'cE

[f/ art > NAVA' EAU hs.uArity SW 42.8-2S/ M122 7VW M RM / o1 -39 '5-6 'iPo8 ft/sVC % 8kMw 60rwom et+J Ibsues m < e t i n s~ 4..c.. w. %m Ny u - m.., A~ ~ sm# d/lc2T e5 7-6 7C:2 dv/v'G.5 M+'<' $r

  • K Ya ' ele #

N23) rs!' N'v' WA e' r vLILu-G np PN. r.4 as M 0~ .o J / / % c..e w-m--ows u gr s h?e. a 3. 5 a m L %L%%. (sa)ser-Iter k;56~c.. hay muud em. eu'n w *w-ast uos S *"*c' *"h* de tta PtLH &2M zur-sst ts w r..t ir-<. cet.~er soi-r/s - 7770 2 NE /._ _Th n//v L.ma M. LluewaAt';t2%sW #2-Nr-nnr tvi"j'b D.... u T.s.,. t w. e.,,, 7<%5+n A sc /N. i.. s I l

,..o A'ITENDANCE--CONTINUED SuarEcr: Maintenance Rule Implementation Page No. 2 ORGANIZATION NAME (print) TrrLE TELEPHONE AmuADON i (print) 3 D w.16t co>L 32. SE. cps ira l (So'NIS-12/.:.I A.Tc_/C9e hen /m R A?E levuME sure. M.t h t. ses 3,79s'/9 b>PPSS klkk 6cle. L.'cean En:aue nob 4n-+1Ge aanuwh]6#dk l j U U P.Sie U N tL rek N v b ie Se tw ~. e Ier// 3c'/ Lfiy.32]Y lt.s pt. cm /M g /Ofsr9/M g e SMsAme SK. / uo o, r 4 "Ad4 ~ e2 - Ts *? 9 09us to urre J Ahm x ens Eur- /se W r4NT wea Du LvNg SLsLt.l f5jqui nd &c /c c fou n ws % dex G49f ?6/o)7/s-37s2 Peco idexa p.4 u n v. 7Atr307 on E g a._ (24 5 -3%4 f&ll E$/('& [VWL DMb /2/C uui A//,h k/t' #ff//Yct H- ~ T m L t,,, & s Reh<.L;hr Em,. w -179-y&i Wes +,. e L., <.<_ Gyy\\/ifj/_gsorn k41v7 R u1* (mt. ((sic. 3 'f(/M&32 insp. Ec Cw 4/.E n,6 v s m -rw << E w butw<v N yrec 4 % N TSu B>C,6 errJcn t o eI < SfG- % 6-}b W tA W & lt 2 B6LC. f;/hr //R 0.JsseIn 'ird ni-3sw M % C.. b e 7> u e x. Li c.. Au,t 3 r-se-vz9r a're w 4 'h w Gaasd' t<><resevse, tusracn (><d w -iiir "M*: 1:st* .sw muu. %w mua (,Sw-am asc/sa a 1}}

Retrieved from "https://kanterella.com/w/index.php?title=ML20133N845&oldid=4236163"