ML20214A152

Partially Withheld SECY-86-216 Providing Commissioners W/ Info on Current Status of Nprd Sys.Ninth Semiannual Staff Evaluation of Status of Sys Encl

ML20214A152
Person / Time
Issue date:	07/24/1986
From:	Stello V NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO)
To:
Shared Package
ML20213E956	List: IA-87-137, Partially Withheld SECY-86-216 Providing Commissioners W/ Info on Current Status of Nprd Sys.Ninth Semiannual Staff Evaluation of Status of Sys Encl ML20213E952 ML20213F123 ML20213G378 ML20213G389 ML20213G440 ML20214A152
References
FOIA-87-137 SECY-86-216, SECY-86-216-01, SECY-86-216-1, NUDOCS 8705190296
Download: ML20214A152 (19)

Text

w

~

a pnapAIETA M ATA ENCL N f *%

?

e 1

g July 24,1986 SECY-86-216 4,,,.

POLICY ISSUE (Information)

For:

The Commissioners From:

Victor Stello, Jr.

Executive Director for Operations

Subject:

NUCLEAR PLANT RELIABILITY DATA (NPRD) SYSTEM Furpose:

To p" ovide the Commissioners with information on the current status of the NPRD System.

Discussion:

In its affirmation of SECY-81-494, the Comission directed the staff to closely monitor the progress of INP0's manage-ment of the NPRD System and to provide the Commission with semiannual status reports on the effectiveness of INP0's management of the NPRD System and the responsiveness of the NPRD System to NRC needs.

The first semiannual report was forwarded to the Commission on July 1, 1982 (SECY-82-279).

Subsequent semiannual reports were forwarded on January 4,1983 (SECY-83-4), on July 5,1983 (SECY-83-4A), on January 27, 1984 (SECY-84-44), on August 1, 1984 (SECY-84-44A), on February 8,1985 (SECY-85-56), on August 20, 1985 (SECY-85-56A) and on January 30, 1986 (SECY-86-35)..

These reports addressed the status of the NPRD System and the staff's program to monitor improvements in NPRDS data.

The enclosure to this peper describes the results of the staff's program to monitor and eva4 ate the quality and quar.tity of data available in the NPRDS data base. The enclosed report is the staff's ninth semi-annual report.

j CONTACT:

Frederick J. Hebdon, AE0D 492-4484 SECY NOTE: This document contains information that should not be released outside the NRC. Attachment A should be removed before release is made.

l i

JROPRIETAR" LATA r m,LusED 8705190276 670513 i

PDR FOIA

) %2 l

JOHNGON87-137 PDR

The Commission The staff has the following general observations based on the enclosed report:

1.

The NPRDS data for the fourth quarter of CY 1985 and the first quarter of CY 1986 show that:

(1) the percentage cf NPRDS-reportable failures that have ultimately reached the data base has increased from approximately 40% to approximately 75%.

Considering that there is some uncertainty in the definition of a reportable failure, this percentage may be as high as 84%. This is a substantial improvement compared to previous quarters.

Further, a total of 4672 failure reports were submitted in the first quarter of CY 1986.

2.

These numbers represent a continuation of the vigorous recovery from depressed failure reporting in the period covering the third quarter of CY 1984 through the first quarter of CY 1985. During that period major effort went into upgrading the component population and engineering-data base (referred to as "rescoping").

3.

However, the possible preferential reporting of failures documented in LERs is of concern for two reasons:

(a)it seriously impairs the ability to judge or. calculate the overall level of NPRDS completeness and (b) it places a high priority on reporting of selected failures which is inappropriate. For example, independent, random reactor trip breaker failures are not reportable as LERs, but are directly reportable to NPRDS. Without an independent source of information on the occurrence of such failures, it is not possible to determine the extent to which such failures are being reported to the NPRD System.

4.

Essentially all of the eligible plants submitted at least one failure report during each quarter of 1985. The INP0 estimate of the expected reporting rate for the median plant if all plants are fully participating (i.e.,100 failure reports per year) has essentially been achieved.

However, the staff feels that additional verification of complete reporting is needed for low reporting plants (e.g., plants with fewer than 40 reports per year).

5.

The timeliness of NPRDS failure reporting has improved compared to the same period in 1985. About one quarter of the reports submitted in the first quarter of 1986 occurred prior to 1985 (i.e., over 12 months prior to the start of the quarter).

I The Commission 6.

Over 85% of a sample of NPRDS failure report narratives continued to be rated at least "probably adequate" (i.e.,

a knowledgeable person could understand the type of failure that occurred).

However, few reports contained good root cause information. Such information is often difficult to determine, but nonetheless must be identified by the licensee in order to assure proper corrective action. Further, its reporting should continue to be a goal of a component reliability data base such as the NPRD System. Thus, the issue of quality warrants continued scrutiny and concern.

In early 1985, INP0 and the industry completed the extensive rescoping and revision of the NPRDS data base.

It is likely

~

that the diversion of resources required for this rescoping depressed routine reporting and, as a result the number of component failures entering the system did not increase above levels fiist achieved in 1cte 1983. While this level was a substantial improvement compared to the level of participation that existed in 1981, it did not constitute full participa-tion.

Since completion of the rescoping the level of participation has increased dramatically. There still may be a few plants that are not fully participating and there are still improvements in the timeliness and quality of the data that should be made.

However, it appears the current levels of participation in the NPRD System warrant increased use of the system as a source of reliability data to meet agency needs.

[

V1'ctor Stello,tJr.

Executive Director for Operations

Enclosure:

Ninth Semiannual Staff Evaluation of the Status of the NPRD System DISTRIBUTION:

Cormnissioners OGC (H Street)

OI OCA ACRS EDO SECY

_~

3 4

I 4

i NINTH i

SEMIANNUAL STAFF EVALUATION

_O_F 4

THE STATUS OF THE NPRD SYSTEM l.

July 1986

]

Prepared by:

Office for Analysis and Evaluation of Operational Data i

T i

i i

j NOTE - The NRC has agreed not to release plant-specific NPRDS data, j

Thus removal of Attachment A is necessary prior to release of j

this document outside the NRC.

t l

Enclosure i

l i

l i

i l

INTRODUCTION The present NPRD System is a voluntary program for the reporting of reliability data associated with selected components and systems in nuclear power plants.

Since January 1,1982, the NPRD System has been managed and operated by the Institute of Nuclear Power Operations.

BACKGROUND Former President Carter's 1977 National Energy Plan recommended that the NRC make mandatory the voluntary reporting of minor mishaps and component failures (i.e.,NPRDS). The plan suggested that mandatory participation would enable the industry and the NRC to develop a more reliable database which is needed to improve the reactor design, construction, operation, safety, and reliability.

Coincident with the NRC's activities directed toward implementing the President's recommendation, the General Accounting Office (GAO) reviewed the NRC's data-gathering activities concerning unscheduled events at commercial nuclear facilities.

In a report issued in late January 1979, the GA0 concluded that it was unlikely that the NRC could justify mandatory NPRDS l

participation when factors such as additional industry costs, limited expected safety benefits, and duplication of the NRC's LER system were considered.

However, the GA0 believed that a full examination of the issue was warranted l

and suggested that the issue be decided using rulemaking procedures.

Following an April 19, 1979 Commission briefing on the collection and analysis of operational safety data, the Commission concurred with the January 1979 GA0 i

recommendation that rulemaking be used to decide the question of whether to make NPRDS reporting mandatory. Accordingly, an Advance Notice of Proposed Rulemaking (ANPRM) was approved by the Commission, and published in the Federal Register on January 30, 1980 (45 FR 6793). Numerous public connent letters were received in response to the ANPRM. The predominant message in the comments was overwhelming opposition to making participation in NPRDS mandatory.

Despite the cmosition to a mandatory system, the staff identified a strong need for failure rate data and engineering data. The requirements for such data indicated a need to revise and reorient the system, in combination with the existing LER system, and to assure its effective implementation through NRC rulemaking.

In order to obtain the necessary improvements in the LER and NPRDS reporting programs, the staff conceptually ~ developed a revised reporting systems. This Integrated Operational Experience Reporting System (10ERS) would have combined and restructured the NRC LER system and the voluntary NPRD System. The 10ERS concept included two principal features:

(1) the collection of detailed technical description of significant events, and (2) the collection of component reliability data.

While the staff still believed that both types of data were essential to the NRC mission, the possibility arose that the NRC could obtain the needed reliability data without assuming direct responsibility for its collection. On June 8, 1981, the INP0 Board of Directors decided that because of its role as an active user of NPRDS data, INP0 would assume responsibility for management of NPRDS.

Further,

-y

-._-__m

_.-._._,y

INP0 proposed developing criteria to be used in their management audits of member utilities to assess the adequacy of NPRDS participation. Therefore, rather than preempt the INP0 activities by proceeding with the 10ERS rule-making, the staff recommended and the Commission approved (SECY-81-494),

proceeding to modify and codify the existing LER reporting requirements as a separate rulemaking while holding the 10ERS rulemaking in abeyance.

In approving SECY-81-494, the Commission directed that the staff closely monitor the status and rate of improvement of the NPRD System and provide the Commission with semiannual status reports on the effectiveness of INP0 management of the NPRD System.

In order to monitor the completeness and quality of NPRDS data, AE0D estab-lished an evaluation program, with technical support from the Reliability and Statistics Branch, EG&G Idaho.

The initial purpose of this program was to provide a baseline of information concerning the completeness and quality of the NPRDS failure reporting and engineering data files as they existed prior to INP0's assumption of responsibility for the management of NPRDS. Subsequent analyses of the completeness and quality of the files are being conducted periodically in order to provide a measure of the improvements in the NPRD System under INP0 management and technical direction.

The scope and results of this evaluation program have been discussed in reports to the Commission dated July 1, 1982 (SECY-82-279), January 4, 1983 (SECY-83-4),

July 5, 1983 (SECY-83-4A), January 27,1984 (SECY-84-44), August 1,1984 (SECY-84-44A), February 8, 1985 (SECY-85-56), August 20,1985(SECY-85-56A),

and January 30, 1986 (SECY-86-35).

In the third quarter of CY 1984, INP0, working with individual plants began an extensive quality assurance review of all of the engineering data records (approximately 226,000 records) in the data base. The effort, called "rescoping", was necessary to conform the information in the NPRDS data base to revisions to the NPRDS Reportable System and Component Scope Manual and the NPRDS Reporting Procedure Manual. The rescoping involved reviewing approximately 4,000 engineering records per plant, and identifying plant components and systems covered by NPRDS.

In addition, the rescoping added a new retrieval feature, i.e., application codes for key plant comporents. Users will be able to retrieve rapidly, for example, all reports on reactor trip breakers with just one command using these new application codes. The rescoping was completed in April 1985.

Although, in the long-term, the resccping will result in major improvements in the accuracy and consistency of the NPRDS engineering data, it produced a major short-term disruption to the NPRD System. Thus, measures of NPRDS performance (e.g., participation) for the third quarter of CY 1984 through the first quarter of CY 1985 appear to have been depressed due to the extensive effort devoted to rescoping by some utilities and by INP0.

COMPARIS0N OF NPRDS FAILURE REPORTS AND LERs The initial evaluation of the NPRDS data consisted of sa'npling approximately 100 LERs and attempting to find corresponding NPRDS failure reports for the failures described in the LERs.

Sample A The first sample (Sample A) involved failures that occurred in January throug.h August 1981 (before INP0 assumed responsibility for the management of NPRDS).

Based on the description of the failures contained in the LERs, the analysts identified 97 failures that they believed should have been reported to NPRDS.

2 The analysts then attempted to find corresponding NPRDS failure reports for the 97 reportable failures.

In the NPRDS data base as it existed in October 1981 (i.e., Tape A), only 11 failure reports that matched the failures described in the LERs were found.

Therefore, less than 12% (i.e.,11 of 97) of the failures 4

that should have been reported to NPRDS had actually been reported as of October 1981.

The analysts also reviewed the entry in Block 24 of the LER where the licensee is requested to enter a "Y" if an NPRDS failure report had been or would be submitted for one or more of the failures described in the LER. Of the 100 LERs in the sample, 50 LERs (representing 55 failures) indicated that an NPRDS failure report had been submitted. However, only ten matching NPRDS failure reports were found [the eleventh report referred to above was for an LER coded as not being reportable to NPRDS, (i.e., an "N" in Block 24)].

Table 1:

Results of Sample A Event Dates: January 1981-August 1981 NPRDS Data Base as of: October 31, 1981 (Tape A)

LERs Sampled (Sample A):

100 Failures Identified: 121 Failures Reportable to NPRDS: 97 Failure Reports Found in NPRDS File (Tape A):

11 LERs That Indicated an NPRDS Failure Report Submitted:

50 (55 failures)

Sample B and Beyond Subsequent samples of LERs have been analyzed using LERs from the first quarter of CY 1982 (Sample B) through the first quarter of CY 1986 (Sample R). These samples have been compared to the NPRDS data base.

NPRDS data for more recent quarters will continue to be evaluated quarterly and the results described in subsequent reports.

In addition to comparing each sample of LER reported failures to the associated NPRDS tape (e.g., Sample C to Tape C), each version (i.e., data tape) of the data base was used to analyze preceding samples (e.g., Tape C was searched to find reports of failures in Samples B and A).

For tapes G through K and P through R, only the preceding three quarters were analyzed in this way.

For tapes L through 0, the analysis included the preceding six quarters in order to determine if substantial numbers of failures were being reported as much as 18 months after the event.

The results of these analyses are summarized in Tables 2, 3, and 4.

Table 2 identifies the sample period and the number of failures reportable to NPRDS from each sample [e.g., the sample from the period October-December 1982 (Sample E) described a total of 110 component failures that should have been reported to NPRDS]. Table 3 identifies when NPRDS evaluation tapes were produced.

Prior to the fourth quarter of CY 1983, each tape reflected the hPRDS data file as it existed one month after the close of the quarter to account for the permissible 30 day delay in submitting NPRDS reports (e.g.,

Tape E contains the NPRDS data base as it existed on January 31,1983).

In order to be consistent with the analysis procedures used by INP0, the tapes beginning in the fourth quarter of CY 1983 (Tape I) contain the data base as it existed on the last day of the subject quarter (e.g., Tape K contains the NPRDS data base as it existed on June 30,1984).

Table 2: Samples NPRDS Reportable Sample Sample Period Failures in Sample A

Jan-Aug 1981 97 8

Jan-Mar 1982 104 C

Apr-Jun 1982 99 D

Jul-Sep 1982 107 E

Oct-Dec 1982 110 F

Jan-Mar 1983 112 G

Apr-Jun 1983 99 H

Jul-Sep 1983 101 I

Oct-Dec 1983 100 J

Jan-Mar 1984 100 K

Apr-Jun 1984 100 L

Jul-Sep 1984 100 M

Oct-Dec 1984 100 N

Jan-Mar 1985 100 0

/pr-Jun 1985 100 P

Jul-Sep 1985 100 0

Oct-Dec 1985 100 R

Jan-Mar 1986 100 i

)

Table 3: Versions of the NPRDS Data Base Version Date of Version A

Oct 1981 B

Apr 1982 C

Jul 1982 D

Oct 1982 E

Jan 1983 F

Apr 1983 G

Jul 1983 H

Oct 1983 I

Dec 1983 J

Mar 1984 K

Jun 1984 L

Sep 1984 M

Dec 1984 N

Mar 1985 0

Jun 1985 P

Sep 1985 Q

Dec 1985 R

Mar 1986 Finally, Table 4 contains the results of efforts to find corresponding NPRDS failure reports for failures described in the sample LERs (e.g., Tape P contained NPRDS failure reports for 68 of the 100 NPRDS-reportable failures in LER Sample M).

It is interesting to note that new NPRDS failure reports are still being added to the data base many months after the failure actually occurred.

For example, Tape N (March 1985) contains six new reports of failures that occurred in the final quarter of CY 1983 (Sample I).

In addition, it is important to note that the reportability of a failure is subject to some interpretation.

Based on INP0's review of one set of failures defined by us as reportable, we estimate that for approximately 10% of the failures, the utilities could reasonably conclude that the failure was not reportable to NPRDS because it was incipient in nature, (i.e., a condition where preemptive corrective action was )rudent, but no loss of function had yet occurred). Thus, the number of reporta)le failures in each sample as determined by the reporting utilities, may be as low as about 90 failures.

Finally, it is also important to note that results for our samples may overestimate utility reporting against the entire NPRDS scope in the most recent quarters.

It is our understanding that INPO has prepared a list of the failurescontainedinrecent(i.e.,1984andlater)LERsandhasdistributed the list to all licensees and has encouraged licensees to insure that failures discussed in LERs are reported to NPRDS.

However, failures that are discussed in LERs constitute a relatively small fraction of the total number of failures that are reportable to NPRDS. As a result, the percentage of failures from LERs that were subsequently reported to NRPDS may not accurately represent (may be higher than) the percentage of all NPRDS-reportable failures that are actually reported to NPRDS, because the sample is no longer representative.

.6

l..~.~. - ~ - -. -

. _... - ~ - _. -

-.. _. _ _ -.... _... -. ~

- _ - - - -. - ~ - - ~. -. ~ - _...... _ - _...

)

F $ SIWY OF WLTQt!F.3 LER COWimENT FAtLTS Also NPRD-4 FORMS TAAL Total Matches (1)

J l

Tape l

Reportable C

Sample _ Quar ter*_

Faults A

B C

D E

F G

H I

J K

L' M'

N 0

P L

R_

Cnsponent A

3-81b 97 11

?!

25 25 30 31 8

l-82 104 0

4 8

14 21 E

C 2-82 99 I

2 5

16 D

J-82 10 1 1

6 12 13 4

13 16 38 L

4-82 110 6

12 43 46 F

l-83 II2 l

3 28 31 43 G

2-83 99 21 35 41 41 41 41 42 H

3-83 IJI 18 38 42 43 44 50 I

4-83 100 19 41 42 42 43 J

1 84 100 14 28 44 49 68 e.

t 2-84 100 1

24 30 58 i

L 3-84 100 5 21 64 68 i

n 4-84 800 5

63 11 15 m

I-85 30 0 11 49

/2 16 0

2-85 100

[

11 50 68 P

3-95 100 13 38 Q

4-85 500 l

14 R

l-86 10 0 Quarter in which events in LER segle occured.

l e.

For Phase A, event detes were spread among the first three quarters of 1981.

b.

Tape a contains data for quarters L(3-84). M(8-84), and M(I-85).

c.

l l

1

.r

.,,,.m,_.

,,m-e

-.m.,

_,,.--,._,r3..

,m_

m.

,m.

Conclusions The NPRDS data for the fourth quarter of CY 1985 and the first quarter of CY 1986 show that:

(1) the percentage of NPRDS-reportable failures that have ultimately reached the data base has increased from approximately 40% to approximately 75%. Considering that there is some uncertainty in the definition of a reportable failure, this percentage may be as high as 84%.

This is a substantial improvement compared to previous quarters.

However, the possible preferential reporting of failures documented in LERs is i

of concern for two reasons:

(a) it seriously impairs the ability to judge or calculate the overall level of NPRDS completeness and (b) it places a high priority or emphasis on reporting of selected failures which is inappropriate.

For example, independent, random reactor trip breaker failures are not reportable as LERs, but are directly reportable to NPRDS. Without an independent source of information on the occurrence of such failures, it is not possible to determine the extent to which such failures are being reported to NPRDS. The staff will give increased attention to determining what other possible checks could also be used in order to verify the completeness of the NPRDS data base.

MEASURES OF NPRDS PARTICIPATION Table 5 indicates that the quantity of reporting has increased sharply in recent quarters.

In the first quarter of CY 1986, a total of 4672 reports were received; a 120% increase compared to the first quarter of CY 1985.

Figure 1 is a plot of component failure transactions for the NPRD System as a function of calendar quarter. While these numbers do not reflect the level of timely reporting, they do in a sense measure NPRDS activity or effort. The figures show a sharp increase in NPRDS reporting in recent quarters. This reporting increase follows the completion of the rescoping activities in April 1985.

Table 5.

NPRDS Reporting Rate Failure l

Reports j

Quarter Received

}

84-1 2307 i

84-2 2356 84-3 1663 i

84-4 1095 j

85-1 2121 85-2 4677 l

85-3 4625 4

85-4 4185 86-1 4672 l

One is tempted to divide the quarterly figures by the number of eligible plants i

i and use this figure for tracking NPRDS completeness. Unfortunately, such an 1

average is not always representative since the total number of reports can 1

I i

I

l 5000,

i e

i r

/

4500 -

4000 L 3500 l C

(0 3000 a

CT 2500 ;

cu O-2000 L m

7 2

1500 2

'iB 1000 u_

500 0

1 2

3 4

1 2

3 4

1 2

3 4

1 2

3 4

1 2

3 4

1 1981 1982 1983 1984 1985 Quarter - Year i

Figure 1. Camponent failure transactions by calender quarter.

l l

reflect the contribution of only a small group of plants reporting to the system.

For example, in the first quarter of CY 1985 a total of 2,121 component failure reports were entered into the file. However, one unit accounted for 425 (20%) of the total entries for the quarter, and there were no er.tries for 11 plants. Thus, sumary statistics (e.g., average number of failure reports per plant per year) may not be good measures of NPRDS participation.

Figures 2 and 3 plot the number of plants vs the number of reports submitted during the quarter (e.g., from Figure 2, in the fourth quarter of 1985, 26 plants submitted between 1 and 20 reports). Clearly a key figure to watch is the number of plants that did not submit any reports in a given quarter.

From Figures 2 and 3, it can be seen that only one plant did not submit any failure reports during a quarter.

The staff also considered the level of participation of individual units from a somewhat different perspective.

Figure 4* indicates the number of reported failures that occurred in CY 1985.

In 1983, INPO estimated that the median number of reportable failures (on a per plant basis) in a year should be approximately 100. On this basis, if all of the units are fully participating in NPROS, then half of the plants would submit fewer reports. The median plant submitted 94 reports for failures which occurred in CY 1985.

Thus, this measure of performance indicates that the industry has essentially achieved the INPO expected level of full NPRDS participation for CY 1985.

However, the staff is still concerned about the number of plants that reported a relatively small nurrber of failures in 1985 (e.g., 27 plants reported fewer than 40 failures for the year).

INP0 has stated that during an on-site evaluation at one unit they examined that unit's NPRDS program in detail because the unit had reported fewer than 40 component failures that occurred in 1984.

INP0's on-site review confirmed the completeness of the unit's NPRDS reporting. While the low number of reports from this unit signaled INP0 to investigate more closely, INP0 concluded that the low number of failur,es accurately represented the performance of reportable components at this unit.

Conclusions 1.

Essentially all of the eligible plants submitted at least one failure report during the last quarter of CY 1985 and the first quarter of CY 1986.

• The underlying data to support this summary table are provided in Attach-ment A.

The NRC has agreed not to publicly release plant-specific NPRDS data (e.g.,AttachmentA).

l 4

Number of plants (total = 85) w w

u o

u o

u o

4 (p

L ___.. _ L. _. _.. L _

_L

__ J o

7

_l

_g%%%%%%%%%%%%\\1M y

1-20 LM%\\%%MMM\\M\\k 2'-4o khkk G

41 - 60 khk]4 y

61 - 8 0 N

hkk} w 81-100

• T kk}*

Q

,o 101-120 O

C-121-140 u

141-160 u

m e

U1 161-180 Q

3 0

(D 181-200

)

~

(

O O

201-220 o,

3 V

221-240 O

{R ui*

U 241-260 m

CD 3

261-280

~

o CD 281-300 C

g E

z u

e m

o 301-320 3

3 CD r-321-340 O

m n

y 341-360 m

=

=

z e

67 g

C 361-380 a

T3 g3 m

n 381-400 V

,m q

9 3

401-420 z

n 1

a

~

421-440 n

O o

C 441-460 e

a o

V 3

[

P 461-480 481-500

l Number of plants (total = 85) l l

N N

u O

w a

w o

w O

4 L-.

--.1--

I i

1 y

0 e-o.

hkkk\\

T 1-O k\\\\\\\\%\\\\M\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\W a

2'-4o khkkkh 8

a F

41 - 60

\\hm 61 - 8 0 q

\\]

• m 81-100 hku h

j 101-120 c

121-140 u

R 9

3 i

141-160

$5 V

h-o 161-180 g

O 3

(

181-200 o

O 201-220 klu 3

p

,QU 221-240 g

O y

241-260 g

M 3

261-280 n

O

~

3 CD 281-300 C

u

=

E 9

O k

301-320 m

O h-321-340 m

r U_,.

rn (D

U h

O 341-360 i-E z

m O

g 361-380 g

m 5

381-400 Q

m 401-420 g

4 O

421 -440 O

m C

a 3

441-460 oo

-~

C 461-480 3

481-500 g

- II -

t i

' Number of plants (total = 85)

M N

(A O

tp O

LD O

U1 O

+

I L

I 1-

_.} _.. _ 1.

I i

0 f

kkkkha

(

1-20 kkkkkkkkkkkkkk%5 21 -4 0 kkkkkkkk$G 41-60 l

C hkkkkkkkkk]%

l 3

61 - 8 0 kkkk}*

f f

81-100 f

101-120 W

j f

hu c-121-140 o

g f

1 141-160 u

g.

i a

m 161-180 Q*

i O

O 181-200 k\\\\\\\\\\\\} m g

o i

3

.go O

201-220 \\\\\\\\\\%

a o

a 3

$V 221-240 y

O 3

241-260 O

9.

m o

t 261-280 l

m 4

M CD 281-300 z

g E

O 301-320 3

m l

r-o 321-340 m

n CD I

o O

~

O V

341 -360 m

k O

e Z

m si O

r 0-G 361-380 3

l 5'

381-400

}

m g

00 401-420 V

421-440

-e; o

441-460 ca t

461-480 481-500 c

7.

t ha ZI -

541-560

-.n-,

-~~

e, e-e-

-.,.,-w r

t 2.

Based on INP0's estimate of the median level of reporting expected if all i

units are fully participating in NPRDS, industry is approaching full participation in NPRDS.

However, additional verification for low reporting plants is needed.

TIMELINESS OF NPROS DATA l

On January 28,1983, INP0 sent a letter to utilities urging that priority in failures reporting be given to recent failures. The staff has examined the j

data to determine the degree of response to this letter in tenns of quantity l

and timeliness of the data.

t Table 6 shows the component failure records sorted by the transaction quarter l

(i.e., the quarter that the record was added to the data base) and the year in j

which the failure occurred.

J j

TABLE 6 Timeliness of NPRDS Failure Reports f

Year in which I

the failure Transaction Quarter i

occurred Quarter in which the report was added to the data base) i i

84-3 84-4 85-1 85-2 85-3 85-4 86-1 i

1986 331-~

)

1985 185 1643 2424 2952 3204 t

1984 1241 1002 1217 2622 1466 970 714 i

l 1983 217 70 322 221 222 109 230 i

1982 56 10 153 42 34 26 96 1

1981 37 1

66 45 15 29 83 1980 51 3

135 42 82 24 2

i 1979 14 0

5 25 158 19 2

1 1978 12 4

17 26 44 12 1

i 1977 14 5

9 3

27 8

3 i

i 1976 10 0

2 3

114 20 3

i a

1975 7

0 7

5 33 12 2

l 1974 4

0 3

0 6

3 1

i 1973 0

0 0

0 0

0 0

i TOTAL 1663 1095 2121 4677 4625 4185 4672 In terms of the timeliness of reporting, in the first quarter of CY 1986, I

J 3535 component failures were reported that occurred in 1985 or 1986. This j

number is 76% of the total entries for the quarter. By comparison, in the j

first quarter of CY 1985, 1402 reports (66%) of the reports described failures j

that occurred in 1984 or 1985.

3 i

r Conclusion l

The timeliness of NPROS failure reporting has improved compared to the same l

i period in 1985, in the first quarter of 1986, 76% of reports submitted described failures that occurred in 1985 or 1986.

Thus, about 24% of the reports submitted in this quarter reflect component failures that occurred i

more than 12 months ago, i

W h

t l

QUALITY OF THE NPRDS FAILURE REPORT NARRATIVE To assess the quality of the NPRDS Failure Reports, a sample of NPRDS Failure Reports was selected and the narrative descriptions of the failures were reviewed to determine if the text described the failure in sufficient detail that system users could understand the failure, its causes, corrective actions, and the implications for similar equipment. The narratives were graded as desired, adequate, probably adequate, and inadequate.

The results of this anlaysis are presented in Table 7.

The " adequate" category means that the contents of the report meet the letter of the NPRDS reporting requirements as contained in the NPRDS Reporting Procedures Manual. The " desired" category was added for this analysis to primarily identify the percentage of reports containing good root cause information.

Few reports met this test.

The majority of the failure reports examined continued to be at least "probably adequate."

Overall Summary In early 1985, INP0 and the industry completed the extensive rescoping and revision of the NPRDS data base.

It is likely that the diversion resources required for this rescoping depressed routine reporting and, as a result the number of component failures entering the system did not increase above levels first achieved in late 1983. While these levels were a substantial improvement compared to the level of participation that existed in 1981, it did not constitute full participation.

Since completion of the rescoping the level of participation has increased dramatically.

There still may be a few plants that are not fully participating and there are still improvements in the timeliness and quality of the data that should be made. However, it appears the current levels of participation in NPRDS warrant increased use of the system as a source of reliability data to meet agency needs.

t i

I i

.... -..... _. ~. _.. _.

_____,___.._...____..._.._-_..._m-.._._._____._.-

i i

e '

IAal.E7. QUALIM OF NPROS FAlttMf. RtroRTS I

First Qawter second Quarter Third Deerter Foesrth Quarter Ftrst Quarter CY 1985 CV 1985 CY 19c5 CV 1985 CY 19%

i It

}

15,

l

-]

It

}

61

}

Pestred I

I I

I Adequate 801{,

911 451 >

141 311,871 301 > 891 h1F M1 i

I i

I lit l }

781 1

501 l

581 1

4bt l

Probably 4

edequate d

Imedegaete 31 761 181 111 141

[

Total member of 106 14 106 85 84

[

8 IePRO-4 fores evelmeted l

e.

One plant had no failure reports during the first quarter 1986 r

a e

>=*

W 8

r I

k s

E 9

e 4

6 i

t

t>8 UDg e

r.

\\...../

l POLICY ISSUE January 30, 1986 (In Ormation)

SECY-86-35 For:

The Commissioners From:

Victor Stello, Jr.

Acting Executive Director j

for Operations

_Subiect:

NUCLEAR PLANT RELIABILITY DATA (NPRD) SYSTEM

Purpose:

To provide the Commissioners with information on the current status of the NPRD System.

Discussion:

In its affinnation of SECY-81-494, the Commission directed the staff to closely monitor the progress of INP0's manage-ment of the NPRD System and to provide the Cormission with semiannual status reports on the effectiveness of INP0's management of the NPRD System and the responsiveness of the NPRD System to NPC needs.

The first semiannual report was forwarded to the Commission on July 1, 1982 (SECY-8?-279).

Subsecuent semiannual reports I

were forwarded on Januery 4, 1983 (SECY-83-4), on July 5, i

1983 (SECY-83-4A), on January 27, 1984 (SECY-84-44), on l

August 1, 1984 (SECY 84-44/), on February 8, 1985 (SECY-85-56), and on August 20,1985(SECY-85-56A). These reports addressed the status of the NPPD Systen and the staff's program to monitor irrprovements in NPRDS data. The enclosure to this paper describes the results of the staff's proaram to monitor and evaluate the quality and quantity of the data available in the NPRDS data base.

The enclosed report is the staff's eighth semi-annual report.

i i

CONTACT:

Frederick J. Hebdon, AF0D 492-4480 1

7( // /C s d?

i 5TCT1 TOTE: This document contains information that should not be released outside the NRC. Attachment A should be removed before release is made.

,n

The Commissioners The staff has the following general observations based on the enclosed report:

1.

The NPPOS data for the second and third quarters of CY 1985 show that:

(1) the percentage of NPRDS reportable failures that have ultimately reached the data base may be as high as 80% based on failure matching with LER-documented failures, and (2) the failure reporting volume exceeded 4600 reports in each quarter. These numbers represent a vigorous recovery from depressed failure reporting in the period covering the third quarter of 1984 through the first quarter of 1985.

Durino that period major the component population effort went into upgrading (referred to as and engineering data base "rescoping").

2.

Essentially all of the eligible plants submitted at least one failure report during the first three cuarters of 1985.

The INPO estimate of the expected reporting rate for the median plant if all plants are fully participating, i.e.,

100 failure report: per year, has essentially been achieved.

However, the staff feels that additional verification of

)

complete reporting is needed for low reporting plants (e.g.,

plants with fewer than 40 reports per year).

3.

The timeliness of NPROS failure reporting has declined compared to the same period in 1984. This decline may be the results of utilities' reporting failures that were back-logged during the rescoping effort in late 1984 and early 1985. About one half of the reports submitted in the third quarter of 1985 occurred more than six months earlier.

4 While the ma.iority of a sample of NPRDS failure report narratives continued to be rated at least "probably adequate" (i.e., a knowledgeable person could understand the type of failure that occurred) there was a drop-off of quality in the two quarters examined. This may be due again to the volume of reporting generated after the lull for rescoping effort. Also, few reports contained good root cause information.

Such information is often difficult to determine, but nonetheless must be identified by the licensee in' order to assure proper corrective action.

Further, its reporting should continue to be a goal of a component reliability data base such as NPRDS. Thus, the issue of quality warrants continued scrutiny and concern.

l

The Commissioners In general, the NPRDS has improved to the point where (1) it warrants increased use as a source of reliability data to meet agency needs, and (2) the staff needs to reconsider and perhaps revise its evaluation methods.

The staff will be exploring ways to better assure that reporting is complete across the entire NPRDS scope and not just for failures found in LERs, and to verify that low reporting rates for some individual plants are justified.

,. L-V ctor Stello, Jr.

Acting Executive Director for Operations

Enclosure:

Eighth Semiannual Staff Evaluation of the Status of NPRDS f

9 f

l

EIGHILH, SEMIANNUAL STAFF EVALUATION THE STATUS OF NPRDS January 1986 Prepared by:

Office for Analysis and Evaluation of Operational Data 4

l NOTE - The NRC has agreed not to release plant specific NPRDS data.

Thus removal of Attachment A is necessary prior to release of this document outside the NRC.

Enclosure i

m

. INTRODUCTION The present NPRD System is a voluntary program for the reporting of reliability data associated with selected components and systems in nuclear power plants.

Since January 1,1982, the NPRD System has been managed and operated by the Institute of Nuclear Power Operations.

BACKGROUND Fonner President Carter's 1977 National Energy Plan recommended that the NRC make mandatory the voluntary reporting of minor mishaps and component

. failures (i.e., hPRDS).

The plan suggested that mandatory participation would enable the industry and the NRC to develop a more reliable data base which is needed to improve the reactor design, construction, operation, safety, and reliability.

Coincident with the NRC's activities directed toward implementing the President's recommendation, the General Accounting Of fice (GA0) reviewed the NRC's data-gathering activities concerning unscheduled events at commercial nuclear facilities.

In a report issued in late January 1979, the GA0 concluded that it was unlikely that tne NRC could justify mandatory NPRDS participation when factors such as additional industry costs, limited expected safety benefits, and duplication of the NRC's LER system were considered.

However, the GA0 believed that a full examination of the issue was warranted and suggested that the issue be decided using rulemaking procedures.

Following an April 19, 1979 Commission briefing on the collection and analysis of operational safety data, the Commission concurred with the January 1979 GA0 recommendation that rulemaking be used to decide the question of whether to make NPRDS reporting mandatory.

Accordingly, an Advance Notice of Proposed Rulemaking (ANPRM) was approved by the Commission, and published in the Federal Register on January 30, 1980 (45 FR 6793).

Numerous public comment letters were received in response to the ANPRM. The predominant message in the comments was overwhelming opposition to making participation in NPRDS mandatory.

Despite the opposition to a mandatory system, the staff identified a strong need for failure rate data and engineering data. The requirements for such data indicated a need to revise and reorient the system, in combination with the existing LER system, and to assure its effective implementation through NRC rulemaking.

In order to obtain the necessary improvements in the LER and NPRDS reporting programs, the staf f developed conceptually a revised reporting system.

This Integrated Operational E.xperience Reporting System (10ERS) would have combined and restructured the NRC LER system and the voluntary NPRD System.

The 10ERS concept included two principal features:

(1) the collection of detailed technical description of significant events, and (2) the collection of component reliability data.

While the staff still believes that both types of data are essential to the NRC mission, the possibility arose that the NRC could obtain the needed reliability data without assuming direct responsibility for its collection.

On June 8,1981, the INP0 Board of Directors decided that because of its role as an active user of NPRDS data, INP0 would assume responsibility for

. management of NPRDS. Further, INP0 proposed developing criteria to be used in their management audits of member utilities to assess the adequacy of NPRDS participation. Therefore, rather than preempt the INP0 activities by pro-ceeding with the 10ERS rulemaking, the staff recommended and the Commission' approved (SECY-81-494) proceeding to modify and codify the existing LER reporting requirements as a separate rulemaking while holding the 10ERS rulemaking in abeyance.

In approving SECY-81-494, the Commission directed that the staff closely monitor the status and rate of improvement of the NPRD System and provide the Commission with semiannual status reports on the effectiveness of INP0 management of the NPRD System.

In order to monitor the completeness and quality of NPRDS data, AE0D estab-lished an evaluation program, with technical support from the Reliability and Statistics Branch, EG8G Idaho.

The initial purpose of this program was to provide a baseline of information concerning the completeness and quality of the NPRDS failure reporting and engineering data files as they existed prior to INP0's assumption of responsibility for the management of NPRDS.

Subsequent analyses of the completeness and quality of the files are being conducted periodically in order to provide a measure of the improvements in the NPRD System under INP0 management and technical direction.

The scope and results of this evaluation program have been discussed in reports to the Commission dated July 1,1982 (SECY-82-279), January 4,1983 (SECY-83-4), July 5,1983 (SECY-83-4A), January 27,1984 (SECY-84-44),

August 1,1984 (SECY-84-44A), February 8,1985 (SECY-85-56), and August 20, 1985 (SECY-85-56A).

In the past, each semiannual report has included analysis of NPRDS data from two calendar quarters. However, in the third quarter of CY 1984, INP0, working with individual plants, began an extensive quality assurance review of all of the engineering data records (approximately 226,000 records) in the data base.

The effort, called "rescoping", was necessary to conform the information in the NPRDS data base to revisions to the NPRDS Reportable System and Component Scope Manual and the NPRDS Reporting Procedure Manual.

The rescoping involved reviewing approximately 4,000 engineering records per plant, and identifying plant components and systems covered by NPRDS.

In addition, the rescoping added a new retrieval feature, i.e., application codes for key plant components. Users will be able to retrieve rapidly, for example, all reports on reactor trip breakers with just one command using these new application codes.

The rescoping was completed in April 1985.

Although, in the long-term, the rescoping will result in major improvements in the accuracy and consistency of the NPRDS engineering data, it produced a major short-term disruption to the NPRD System.

Thus, measures of NPRDS performance (e.g., participation) for the third quarter of CY 1984 through the first quarter of CY 1985 may be depressed due to the extensive effort devoted to rescoping by some utilities and by INP0, and may be unusually high in the second and third quarters of CY 1985 due to working off any rescope reporting backlog.

. COMPARISON OF NPRDS FAILURE REPORTS AND LERs The initial evaluation of the NPRDS data consisted of sampling approximately 100 LERs and attempting to find corresponding NPRDS failure reports for the-failures described in the LERs.

. Sample A The first sample (Sample A) involved failures that occurred in January j

through August 1981 (before INP0 assumed responsibility for the management of NPRDS). Based on the description of the failures contained in the LERs, 4

the analysts identified 97 failures that they believed should have been reported to NPRDS.

The analysts then attempted to find corresponding NPRDS failure reports for i

the 97 reportable failures.

In the NPRDS data base as it existed in October 1981 (i.e., Tape A), only 11 failure reports that matched the failures described in the LERs were found. Therefore, less than 12% (i.e.,11 of 97 of the failures l

that should have been reported to NPRDS had actually been repor)ted as of October 1981.

The analysts also reviewed the entry in Block 24 of the LER where the licensee is requested to enter a "Y" if an NPRDS failure report had been or would be submitted for one or more of the failures described in the LER.

Of the 100 LERs in the sample, 50 LERs (representing 55 failures) indicated that an NPRDS failure report had been submitted.

However, only ten matching NPRDS failure reports were found [the eleventh report referred to above was for an LER coded as not being reportable to NPRDS, (i.e., an "N" in Block 24)].

Table 1:

Results of Sample A Event Dates: January 1981-August 1981 NPRDS Data Base as of: October 31,1981 (Tape A)

LERs Sampled (Sample A):

100 i

Failures Identified:

121 Failures Reportable to NPRDS: 97 Failure Reports Found in NPRDS File (Tape A):

11 LERs That Indicated an NPRDS Failure Report Submitted:

50 (55 failures)

Sample B and Beyond i

Subsequent samples of LERs have been analyzed using LERs from the first quarter of CY 1982 (Sample B) through the third quarter of CY 1985 (Sample P). These samples have been compared to the NPRDS data base. NPRDS data for more recent quarters will continue to be evaluated quarterly and the results described in subsequent reports.

4

. In addition to comparing each sample of LER reported failures to the associated NPRDS tape (e.g., Sample C to Tape C), each version (i.e., data tape) of th'e data base was used to analyze preceding samples (e.g., Tape C was searched to find reports of failures in Samples 8 and A).

For tapes G through K and P, only the preceding three quarters were analyzed in this way. For tapes L through 0, the analysis included additional preceding quarters in order to determine if substantial numbers of failures were being reported as much as 18 months after the event.

The results of these analyses are summarized in Tables 2, 3, and 4.

Table 2 identifies the sample period and the number of failures reportable to NPRDS from each sample [e.g., the sample from the period October-December 1982 (Sample E) described a total of 110 component failures that should have been reported to NPRDS).

Table 3 identifies when NPRDS evaluation tapes were produced. Prior to the the fourth quarter of CY 1983, each tape reflected the NPRDS data file as it existed one month after the close of the quarter to account for the permissible 30 day delay in submitting NPRDS reports (e.g.,

Tape E contains the NPRDS data base as it existed on January 31,1983).

In order to be consistent with the analysis procedures used by INPO, the tapes beginning in the fourth quarter of CY 1983 (Tape I) contain the data base as it existed on the last day of the subject quarter (e.g., Tape K contains the NPRDS data base as it existed on June 30,1984).

Table 2: Samples NPRDS Reportable Sample Sample Period Failures in Sample

~

A Jan-Aug 1981 97 R

Jan-Mar 1982 104 C

Apr-Jun 1982 99 D

Jul-Sep 1982 107 E

Oct-Dec 1982 110 F

Jan-Mar 1983 112 G

Apr-Jun 1983 99 H

Jul-Sep 1983 101 I

Oct-Dec 1983 100 J

Jan-Mar 1984 100 K

Apr-Jun 1984 100 L

Jul-Sep 1984 100 M

Oct-Dec 1984 100 N

Jan-Mar 1985 100 0

Apr-Jun 1985 100 P

Jul-Sep 1985 100

r Table 3: Versions of the NPRDS Data Base Version Date of Version A

Oct 1981 B

Apr 1982 C

Jul 1982 D

bct 1982 E

Jan 1983 F

Apr 1983 G

Jul 1983 H

Oct 1983 I

Dec 1983 J

Mar 19E4 K

Jon 1984 L

Sep 1984 M

Dec 1984 N

Mar 1985 0

Jun 1985 P

Sep 1985 Finally, Table 4 contains the results of efforts to find corresponding NPRDS failure reports for failures described in the sample LERs (e.g.,

Tape P contained NPRDS failure reports for 68 of the 100 NPRDS-reportable failures in LER Sample M).

It is interesting to note that new NPRDS failure reports are still being added to the data base many months after the failure actually occurred.

For example, Tape N (March 1985) contains six new reports of failures that occurred in the final quarter of CY 1983 (Sample I).

e 4

2 r

-e

=,4--

. Table 4: Summary of Matching LER Component Faults and NPRDS Failure Reports Total Matches (%)_

Tape Reportable Component Sample Faults A

B C

D E

F G

H I

J K

L M

N O

P A

97 11a 21 25 25 30 31 B

104 0

4 8

14 21 C

99 1

2 5

16 O

107 1

6 12 13 E

110 4

13 16 38 F

112 6

12 43 46 G

99 3

28 37 43 H

101 21 33 40 40 41 41 42 1

100 18 38 41 43 44 50 J

100 19 41 42 42 43 K

100 14 28 44 49 68b L

100 7

24 30 58 M

100 5

27 64 68 h

100 5

63 70 0

100 17 48 P

100 11 a.

For Sample A, event dates were spread among the first three quarters of 1981.

b.

This number means that by the second quarter of C'Y 1985 (Tape 0), of the failures in the sample of failures reported in LERs during the second quarter of CY 1984 (sample K), 68%

had been, reported to NPRDS.

G e

G v

. Ir addition, it is important to note that the reportability of a failure is subject to some interpretation.

Based on INP0's review of one set of,

failures defined by us as reportable, we estimate that for approximately 10% of the failures, the utilities could reasonably conclude that the failure was not reportable to NPRDS because it was incipient in nature, (i.e., a condition where preemptive Corrective action was prudent, but no loss of function had yet occurred).

Thus, the number of reportable failures in each sample as determined by the reporting utilities, may be as low as about 90 failures.

Finally, it is also important to note that results for our samples may overestimate utility reporting against the entire NPRDS scope in the most recent quarters.

It is our understanding that INP0 has prepared a list of the failures contained in recent (i.e.,1984 and later) LERs and has distributed the list to all licensees and has encouraged licensees to insure that failures discussed in LERs are reported to NPRDS. However, failures that are discur, sed in LERs constitute a relatively small fraction of the total number of failures that are reportable to NPRDS. As a result the percentage of failures from LERs that were subsequently reported to NPRDS may not accurately represent (may be higher than) the percentage of all hPRDS-reportable failures that are actually reported to NPRDS, because the sample is no longer representive.

Conclusion The NPRDS data for the second and third quarters of CY 1985 show that.,

the percentage of NPRDS-reportable failures documented in sample LERs that have ultimately reached the data base has increased from approxi-mately 40% to approximately 70%. Considering that there is some uncertainty in the definition of a reportable failure, this percentage may be as high as 80%. This is a substantial improvement compared to previous quarters.

However, the possible preferential reporting of failures documented in LERs is of concern for two reasons:

(a) it seriously impairs the ability to judge or calculate the overall level of NPRDS completeness and (b) it places a high priority or emphasis on reporting of selected failures which is inappropriate.

For example, independent, random reactor trip breaker failures are not reportable as LERs, but are directly reportable to NPRDS.

Without an independent source of infor-mation on the occurrence of such failures, it is not possible to determine the extent to which such failures are being reported to NPRDS.

The staff will give increased attention to determining what other possible checks could also be used in order to verify the completeness of the NPRDS data base.

3 1

. MEASURES OF NPRDS PARTICIPATION l

Table 5 indicates that the quantity of reporting has increased sharply in recent quarters.

In the third quarter of CY 1985, a total of 4625 reports were received; a 200% increase compared to the average of the first two quarters of CY 1984. Figure 1 is a plot of component failure transactions for the NPRD System as a function of calendar quarter. While these numbers do not reflect the level of timely reporting, they do in-a sense measure NPRDS activity or effort. The figures show sharp increase in NPRDS reporting in recent quarters. This reporting increase follows the completion of the rescoping activities in April 1985.

Table 5. NPROS Reporting Rate Failure Quarter Reports Received 84-1 2307 84-2 2356 84-3 1663 84-4 1095 85-1 2121 85-2 4677 85-3 4625 One is tempted to divide the quarterly figures by the number of eligible plants and use this figure for tracking NPRDS completeness.

Unfortunately, such an average is not always representative since the total number of ' reports of ten reflects the contribution of only a small group of plants reporting to the system. For example, in the first quarter of CY 1985 a total of 2,121 component failure reports were entered into the file.

However, one unit accounted for 425 (20%) of the total entries for the quarter, and there were no entries for 11 plants. Thus, summary statistics (e.g., average number of-failure reports per plant per year) continue not to be good measures of NPRDS participation, t

Figures 2 through 4 plot the number of plants vs the number of reports sub-mitted during the quarter (e.g., from Figure 2, in the first quarter of 1985, 38 plants submitted between 1 and 20 reports). Clearly a key figure to

{

watch is the number of plants that did not submit any reports in a given I

quarter. By comparing Figures 2 and 4, it can be seen that between the 1

first quarter of CY 1985 and the third quarter of CY 1985 the number of plants that did not submit any failure reports during the quarter decreased from 11 to 3.

The staff also considered the level of participation of individual units from a somewhat different perspective.

Figure 5* indicates the number of.

reported failures that occurred in CY 1984.

• The underlying data to support this figure is provided in Attachment A.

The NRC has agreed not to publicly release plant-specific NPRDS data (e.g., Attachment A).

-p

---

r-

-m--

FAILURES PER QUARTER m

m w

w c

c u,

m O

U1 O

U1 O

U1 O

U1 O

o O

O O

O O

O O

O O

o O

O O

O O

O O

O O

O

~

@b w

M D

N w

O a

C c'

4

=

x

~

Hm I

m e

l w

Nm>

s

~

b u

b_

I l

esud

~

N m mme w

I I

I I

I I

m.

e

. In 1983, INP0 estimated that the median number of reportable failures (on a -

per plant basis) in a year should be approximately 100. On this basis, if all of the units are fully participating in NPRDS, then half of the plants would submit more than approximately 100 reports for a year and one-half of the plants would submit fewer reports. The median plant submitted 93 reports for failures which occurred in CY 1984.

If plants not yet declared commercial (which generally precedes NPRDS reporting) are dropped, the median is 105.

Data for CY 1985 was not evaluated in this manner because only three quarters of data are available and it is difficult to extrapolate this data to a full year because of the time lag associated with reporting.

Thus, this measure of performance indicates that the industry has essentially achieved the INPO expected level of full NPRDS participation for CY 1984.

However, the staff is still concerned about the number of plants that reported a relatively small number of failures in 1984 (e.g., 24 plants reported fewer than 40 failures for the year).

INP0 has stated that during a recent on-site evaluation at one unit they examined that unit's NPRDS program in detail because the unit had reported fewer than 40 component failures that occurred in 1984.

INP0's on-site review confirmed the completeness of the unit's NPRDS reporting. While the low number of reports from this unit signaled INP0 to investigate more closely, INP0 concluded that the low number of failures accurately represented the performance of reportable components at this unit.

The staff will continue to pursue possible approaches in order to verify this conclusion.

Conclusion I

1.

Essentially all (98%) of the eligible plants submitted at least one failure report during the first three quarters of 1985.

2.

Based on INP0's e3timate of the median level of reporting expected if all

~

units are fully p'trticipating in NPRDS, industry is approaching full participation in NPRDS. However, additional verification for low reporting plents is needed.

TIMELINESS OF NPRDS DATA On January 28, 1983, INP0 sent a letter to utilities urging that priority in failure reporting be given to recent failures. The staff has examined the data to determine the degree of response to this letter in terms of quantity and timeliness of the data.

Table 6 shows the component failure records sorted by the transaction quarter (i.e., the quarter that the record was added to the data base) and the year in which the failure occurred.

)

~

t p

e 1-0 00G-10k w

(f) 0 00k-10P E-Z 0 OOP-Itt z

b E

I Ott-tat O

t U

[#

0 02P-IOP m

Ze b

Id 2 0

00t-IOC Z

@3 0

O 0

00C-19C 1

o

[I.l 0

09C-IFC e

W

' 0FC-12C $

D 2

0 M

2 rv 0

02C-10C y 0

00C-102 $

4 D

k 0

002-192 Ez 0 092-102 F h

b 0

0F2-122 g Z

[:.]

022'102 0

7 O

o ooa-ggi a 1

0 2

0 001-191 $

O 091-It! g O

O 3z tf) 1 Oti-121 a

l{

021--101 xp 2

001-10 F

00-19 h) 09-It

.s hh\\

ti 0 9-12 "cL%N%%%%%%%NEM e-i y

7

.v c

(elva.o.t) SLNV1d.iO H30lmN

O

'rigure 3:

2ND QTR 85 COMPONENT FAILURE REPORT COUNTS LEGEND a

EZ NUMDER OF PLANTS WITHIN RANGE

~ ~l J3 m-M-

I E

ps.

s-8 e

n N

z

$ u-o

" s-g

e

'g e-5.-

z e-N

/

P7)

%o F7) o o

o o o o e o e o e o o J. Ii1

~ ~ - = = asa=massass aase: ss NUMBER OF COMPONENT FAILURE REPORTS e

4

Figure 4:

3RD QTR 85 COMPONENT FAILURE REPORT COUNTS LEGEND E2 NUMBER OF PLANTS WITHIN RANCE F. -

g n-3 7

l g-a b

U n

E

]S Ao bo 5

2 M

N V,

o e o o o e o o o o o o o o o PZW7) 8 ? 8 8 8 8 ? 8 8

$, $, v? $, y8 8 @ e? @Q8 8 ? 8 8 @

1 1 1 2 T T T T T

?

T T T T E E E E E N 5 5 5 5 5 5 5 5 5 5 5 5 5 5.

NUMDER OF COMPONENT FAILURE REPORTS

0 009-10>

w 0

0 00t-19F z<*

~

1 09t-itt z

E Ih-Ot t-12t 2*

0 02P-10t czie WZ 0

OOP-!BC 05

c r 1~

0 00C-19C c

0( 09C-ItC O

ew Cm 0

OtC-12C a

2 0

02C-100 6

e t] OOC-18212

~

~

SE s g t-t@- 092-192 h 8"5 c3 6

g,u o e;

t{.

t a

092-!r2 s.

'C y " g 2kk] 0>2-122 y

&.C On O

{~

2Q

,022-102 I ai j

9AggggggY002-ioi!

C 9 k % Q k k k k N 001-191 tAgNh3 091-It! 5 O

Lkkkkkkkk 3 OPI-121 e, 2

'h % NSSS8& h \\ M M \\\\\\\\N " '-' '

Eg SANggM 00i-io s

e e 9xNN g N g g y a

00-19 SSSSSSg%%%NS a

a-it cihhhMMhhMMN 06-13 9th h h M N a-i

~

A%%%N 9

o

(

r i

V g

g y

H U

01 8

9 9

3 0

(90-1 viol) SINVB JO H3GM3N I

+

l a

l t

,,,~-r-

-,,,,.., - - -,.. _,, ~ ~ -,,,,,

,e---,,,-----,.----,w,.---

-~, -. -, - ~ - -, - -

r--,-----,.-,c, e

r-.

--,,,e,

TABLE 6 Timeliness of NPRDS Failure Reports Year in which the failure Transaction Quarter occurred (Quarter in which the report was added to the data base) 84-1 84-2 84-3 84-4 85-1 85-2 85-3 1985 185 1643 2424 1984 371 1072 1241 1002 1217 2622 1466 1983 1019 602 217 70 322 221 222 1982 222 185 56 10 153 42 34 1981 159 139 37 1

66 45 15 1980 223 168 51 3

135 42 82 1979 74 113 14 0

5 25 158 1978 35 32 12 4

17 26 44 1977 50 25 14 5

9 3

27 1976 39 6

10 0

2 3

114 1975 37 8

7 0

7 5

33 1974 71 6

4 0

3 0

6 1973 7

0 0

0 0

0 0

TOTAL 2307 2356 1663 1095 2121 46/7 4625 In terms of the timeliness of reporting, in the third quarter of CY 1985 2424 component failures were reported that occurred in 1985. This number is 52% of the total entries for the quarter.

By comparison, in the third quarter of CY 1984, 1241 reports (75%) of the reports described failures that occurred in 1984.

Conclusion

~

The timeliness of NPROS failure reporting has declined compared to the same period in 1984. This decline may be a result of utilities' reporting failures that were not reported during the rescoping effort in late 1984.

In the third quarter of 1985, 52% of reports submitted described failures that occurred in 1985.

Thus, about one half of the reports submitted in this quarter reflect component failures that occurred more than six months ago, and in fact component failures from 1974 continue to be reported.

QUALITY OF THE NPRDS FAILURE REPORT NARRATIVE To assess the quality of the NPRDS Failure Reports, a sample of NPROS Failure Reports was selected and the narrative descriptions of the failures were reviewed to determine if the text described the failure in sufficient detail that system users could understand the failure, its causes, corrective actions, and the implications for similar equipment. The narratives were graded as desired, adequate, probably adequate, and inadequate.

The results of this analysis are presented in Table 7.

Table 7. Quality of NPRDS Failure Reports Fourth Quarter First Quarter Second Quarter First Quarter Second Quarter Third Quarter CY 1960 CY 1984 CY 1984 CY 1985 CY 1985 CY 1985 Desired 1%

1; Adequate 71%

63%

46%

80%

46%

31%

Probably 24%

34%

46%

17%

28%

50%

Adequate Inadequate 5%

3%

8%

3%

25%

18%

T;tal Number of 62 68 72 106 75 80 NPRDS-4 Foms Evaluated O

O s

C

The " adequate" category means that the contents of the report meet the letter of the NPRDS reporting requirements as contained in the NPRDS Reporting Procedures Manual.

The " desired" category was added for this analysis to primarily identify the percentage of reports containing good 4

i root cause information.

Few reports met this test.

1 While th9 majority of the failure reports examined continued to be at least "probably adequate", there was a dropoff in the two quarters just examined. This may be due to the volume of reporting generated after the

.I lull for rescoping effort.

The issue of quality warrants continued scrutiny and concern.

t j

Overall Summary In early 1985, INP0 and the industry completed the extensive rescoping and revision of the NPRDS data base.

It is likely that the diversion i

resources required for this rescoping depressed routine reporting and, as a result the number of component failures entering the system did not increase i

above levels first achieved in late 1983.

While these levels were a substantial

]

improvement compared to the level of participation that existed in 1981 it did not constitute full participation.

Since completion of the rescoping the level of participation has increased dramatically.

There are still some i

plants that do not appear to be fully participating and there are still improvements in the timeliness and quality of the data that should be made. However, it appears the current levels of participation in NPRDS I

warrant increased use of the system as a source of reliability data to meet agency needs.

The staff will continue to pursue possible approaches in order to verify that 1) the levels of NPRDS reporting indicated by the LER audit are, in fact, representative of the overall NPRDS completeness and are not due to preferential reporting of failures discussed in LERs, and 2) low l

report counts at some plants are an accurate reflection of their experi-ence.

i I

4

..e4e A4 4

4 4

4aoe a

v44 4ae g.

d 4

..a-.-

-m..

.m._

9 g

e g

9.

e i

l.

i 1

I f

I l

l I

J.

4 N

I J

J h

l

't II t

0

,1 6

i 1,

i 11 I

i I

4 4

4