IA-86-127, Eighth Semiannual Staff Evaluation of Status of Nprds
| ML20151Y179 | |
| Person / Time | |
|---|---|
| Issue date: | 01/31/1986 |
| From: | NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| Shared Package | |
| ML19332B522 | List: |
| References | |
| FOIA-86-126, FOIA-86-127, FOIA-86-131, FOIA-86-166, FOIA-86-201, FOIA-86-209, FOIA-86-263, FOIA-86-80, FOIA-86-82, FOIA-86-A-225, FOIA-86-A-226, FOIA-86-A-227, FOIA-86-A-228, FOIA-86-A-229, FOIA-86-A-230, FOIA-86-A-231, FOIA-86-A-232, FOIA-86-A-233 NUDOCS 8805040384 | |
| Download: ML20151Y179 (17) | |
Text
'
.- i 4 4 EIGHTH
- SEMIANNUAL STAFF EVALUATION E
THE STATUS OF fiPROS January 1986 Prepared by:
Office for Analysis and Evaluation of Operational Data -
NOTE - The NRC has agreed not to release plant specific NPROS data.
Thus removal of Attachment A is necessary prior to release of this document outside the NRC.
i l
8805040384 880418 OhDEO b^-225 A PDR Enclosure i u ~S
s .
.. 1 INTRODUCTION The present NPRO System is a voluntary program for the reporting of reliabiIity 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 l
Fortner President Carter's 1977 National Energy Plan recommended that the l NRC make mandatory the voluntary reporting of minor mishaps and component
. f ailures (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.
Coinc' dent 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 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 NPROS mandatory.
Despite the opposition to a mandatory system, the staff identified a strong l 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 HPRDS 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 staf f 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
-2 management of NPRDS. Further, INPO proposed developing criteria to be used in their management audits of member utilities to assess the adequacy of NPRDS participation. Therefore, rather than preeept the INP0 activities by pro-ceeding with the 10ERS rulemaking, the staf f 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 staf f 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 AE00 estab. '
lished an evaluation program, with technical support from the Reliability 1 and Statistics Branch, EG8G Idaho. The initial purpose of this program was i
to provide a baseline of information concerning the completeness an1 quality of the NPRDS failure reporting and engineering data files as they existed prior to INP0's assumption of responsibility for the management of MPRDS.
Subsequent analyses of the completeness and quality of the files are being i 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 thira 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 I data bese. The effort, called "rescoping", was necessary to conform the .
j information in the NPRDS data base to revisions to the NPRDS Reportable System and Component Scope Manual and the NPRDS Reporting Procedure Manual.
Tne rescoping involved reviewing approximately 4,000 engineering records per plant, and identifying plant components and systems covered by NPROS. In addition, the rescoping added a new retrieval feature, i.e., application codes l 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 ef fort devoted to rescoping by some utilities and by INPO, 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 sampilng approximately 100 LERs and attempting to find corresponding NPRDS failure reports for the-f ailures described in the LERs.
Sample A The first sample (Sample A) involved failures that ocenrred in January through 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.
The analysts then attempted to find corresponding NPRDS failure reports for the 97 reportable f ailures. 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 that should have been reported to NPRDS had actually been reported as of October 1981.
The analysts also reviewed the entry in bluck 24 of the i.ER 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 kPRDS f ailure 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 ds 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 I LERs That Indicated an NPRDS Failure Report Submitted: 50 (55 failures)
Sample 8 and Beyond Subsequent samples of LERs have been analyzed using LERs from the first gaarter s 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.
l l
\ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ .
l 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, only the preceding three quarters were analyzed in this way. For tapes L through 0, the analysis included additional preceding quarters in order to detennine if substantial numbers of failures were being reported as much as 18 months after the event. '
l 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 NPROS i
i f rom 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 af ter the close of the quarter to l account for the permissible 30 day delay in submitting NPROS reports (e.g., I Tape E contains the NPRDS data base as it existed on January 31,1983). In I order to be consistent with the analysis procedures used by INPO, the tapes
l 1
1 Table 2: Samples i l
l NPRDS Reportable Sample Sample Period f ailures in Sample
~
A Jan-Aug 1981 97 B Jan-Mar 1982 104 C Apr-Jun 1982 99 D Jul-Sep 1982 107 l E Oct-Dec 1982 110 '
F Jan-Mar 1983 112 G Apr-Jun 1983 99 H Jul-Sep 1983 101 1 Oct-Dec 1983 100 J Jan-Mar 1984 100 K Apr-Jun 1984 100
)
L Jul-Sep 1984 100 M Oct-Dec 1984 100 l N Jan-Mar 1985 100 0 Apr-Jun 1985 100 P Jul-Sep 1985 100 l
l
a Table 3: Versions of the NPROS 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 1 Dec 1983 J Mar 1984 K Jun 1984 L Sep 1984 M Dec 1984 N Mar 1985 0 Jun 1985 P Sep 1985 Finally Table 4 contains the results of ef forts 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 af ter 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 1).
- - - .- --,_- . , , - - - - - . - . . ,w - - -. , , - - -
Table 4: Summary of Matching LCR Component Faults and NPRDS Failure Reports Total Matches (1)
Reportable Component Sample Faults A B C D E F G H I J K L M N O P A 97 Ila 21 25 25 30 31 8 104 0 4 8 14 21 C 99 1 2 5 16 D 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 N 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 CY 1985 (Tape 0), of the f ailures in the sample of failures reported in LERs during the second quarter of CY 1984 (sample K), 6 4 had been, reported to NPRDS.
l l
l l
l
. 1 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 f ailures defined by us as reportable, we estimate that for approximately 10% of the f ailures, 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 f ailures in each sample as determined by the reporting utilities, may be as low as about 90 failures, l
Finally, it is also important to note that results for our samples may 1 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 discussed in LERs constitute a relatively small fraction of the total number of failures that are reportable to WPRDS. 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 NPRDS-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 l the percentage of NPRDS-reportable f ailures 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. - l 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 cre 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 staf f 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. ,
l
- 0-1 MEASURES OF NPRDS PARTICIPATION 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 (
j a sense measure NPRDS activity or effort. The figures show sharp increase 1 in NPRDS reporting in recent quarters. This reporting increase follows the completion of the rescoping activities in April 1985.
Table 5. NPRDS Reporting Rate I failure Quarter Reports Received 84-1 2307 84-2 2356 84-3 1663 84-4 1095 l 85-1 2121 l 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 NPROS completeness. Unf ortunately, .
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 NPROS participation.
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 quarter. By comparing Figures 2 and 4, it can be seen that between the 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 i from 11 to 3.
The staff also considered the level of participation of individual units l from a somewhat dif ferent perspective. Figure 5* indicates the nunber of .
reported failures that occurred in CY 1984.
- The underlying data to support this figure is provided in Attachnent A.
The NRC has agreed not to publicly release plant-specific NPRDS data (e.g.,AttachmentA).
-9 j .. j
- t
_ --~ g 1 -
4 i
i
- se i
_~ -
+
x w
g i i e
_ _~ -
g I w
_i s e
t <
m
=
a 1
a 6
n
_ _ ~. -
.a i .
I i
t' se
_ __~ -
1 I O C O O C C C O O O O O O O C O O O O O O C LA O tr. O in C t.n O LO tn = = m n m m - -
8318VnD 83d S38nlIVJ
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 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 consnercial (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 daia 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 INP0 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). INPO 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 coeponent 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, INPO 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
- 1. Essentially all (98t) of the eligible plants submitted at least one failure report during the first three quarters of 1985.
- 2. Based on INP0's estimate of the median level of reporting expected if all '
units are fully participating 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, INPO 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.
fi-i Figure 2: IST QTR 85 COMPONENT FAILURE REPORT COUNTS 2
LEGEND E23 NUMBER OF PLANTS WITHIN RANGE s-8 .
t F-F8' .
h O
[a- .
sn E
3 R-a ,
o
- a s-8 m -
^
$ 9-1 M,h d,o o o o o o o o o o o o o odo o o B ? S 8 8 8 ? 8 8 $. $. 7S g $. $. $. ?9 g@ 8T .2 S 8 8 g
. 1 1" 1" a T T T T T . . . T T T T -
) .
8 5 NUMDER E - OFECOMPONENT E 555555555555555, FAILURE REPORTS
]
i
- rigure 3: 2ND QTR 85 COMPONENT FAILURE REPORT COUNTS 1
c ca suuncR erYt.4E winiis aAncE
- - g u-M-
Y gg.
S .. .
._ U g- .
3.
P o 32-g ._
?..
~~
N 7
o d -
FZ1 o Ro o o FZ1 o o o o o o o o o 9???8aiaaaBEaaMHRRR!$$$$a
~*~'hkkkkkkk$kkkkkkkkkikk NUMDER OF COMPONENT FAILURE REPORTS O
% _ _____m - . ____ _______m _ _ - - ._. _ m _
1 000-19t (f) $
P g
=
If 00 > -13t 0 099-Itt D b O A 0 O k t-12>
U E Cr O 02 >-10 t h bz, 03 O 00&-19C O P b 5 0 00C-Isc
[2] ,
% u 0 09C-14 5 .
m =
0 0 >C-12C y 0 E -
02C-10C y O E g 0 000-102E
[2L. 0 OR2-193 n
g 0 092-l>2g u
0 0>2-132 0
O 022-10B{
CL O 002-191 L o
2 O 2 001-191 6 U 0 g
091-161 = .
to z e t k% 001-121 9hh\kN 021-101 E :
9hk\k% 001-10 l
- k. 00-19 6h%%\h% 09-It liik\\h\kk\MNkh \ OF-!?.
\
62 N\\\\\\\NNhhh\\\\\\\\\\\\\\\\\\\\\\N 02-I
- eg 0
\
i r
u.
& 4 = 6 & 4 .
(m=1vmt) stnv1.1.40 >nnnns !
- l
- 14 -
t 0 009-19t I w ;
y 0 00t-196
- I 09t-itt
.z.
1 Ot t- L2 > ,
t .
c" zi2 0 02 >-10 >
wz o got.19C 95 l #w6 0 09C-19C o
e 0 09C-ItC w
=
. 2 0 Ott-120 ,
o z
O 02C-10C i e 6 .
I u 1 000-192 8 l gi ! . z m 8t --
tQOst-tsa!c 6
5 v
=
m 1 093-It2 500 $
Ci%E s .c cn Ek%Y 0tz-Ita z i i i b >= m ge i 2, 022-I02 3 ,
ei -
. 9 hNSSSSSSx h 1.we: 91 e
} 9% %\kN001-191
! thhkh%, 091-1t1 5 i a
Lkkkhkh% 0tt-121
,e3 ot hAN h h M % %NS88 Oei-i0t i 001-W ;
ec 5 9k\\\%\\\%%%% 09 -19 l 98NNSSSSSS88SA%\\%N *-i t cih\\MANN%Nh%%%NN888 06-ta SSSSSSN % %
To H O Cl 9 9 9 3 0 .
(W 1v101) SINV1d J0 B3GKM j l
l l
I
l l
l l !
1 1
TABLE 6 Timeliness of NPRDS Failure Reports Year in which I the failure Transaction Quarter occurred (Quarte
- in which the report was added to the data base) 84-1 84-2 84-3 84-4 85 1 85-2 55-3 i
1985 -- -- -- --
185 1643 24 ?4 1984 371 1072 1241 1002 1217 2622 1466 1983 1019 602 217 70 322 221 222 l 1982 222 185 56 10 153 42 34 l
1981 159 139 37 1 66 45 15 1980 223 168 51 3 135 42 82 l 1979 74 113 14 0 5 25 158 i 1978 35 32 12 4 17 26 44 1977 50 25 14 5 9 3 27 1976 39 6 10 0 2 3 114 i 1975 37 8 7 0 7 5 33 1974 71 6 4 0 3 0 6 i 1973 7 0 0 0 0 0 0 TOTAL 2307 2356 1663 1095 2121 ~46f7 4625 l
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 NPRDS 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 ef fort 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 f act component f ailures from 1974 continue to be reported. l OVALITY OF THE NPRDS FAILURE REPORT NARRATIVE I J
To assess the quality of the NPROS 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 suf ficient 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.
1 Table 7. Quality of NPRDS Failure Reports Fourth Quarter First Quarter Second Quarter First Quarter Second Quarter CY 1983 CY 1984 CY 1984 Third Quarter CY 1985 CY 1985 CY 1985 Desired 1% 1%
Adequate 71% 63% 46% RO% 46% 31%
Probably 24% 34% 46% 17%
Adequate 28% 501 i j
Inadequate 5% 3% 8% 3% 25% 18%
Titc) Nweber of 62 68 72 106 MPRDS-4 Forms 75 80 Ev31uated t
1 1
1 -
s h
4 4
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 i
analysis to primarily identify the percentage of reports containing good root cause information. Few reports met this test.
While the majority of the failure reports exanined 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 af ter the lull for rescoping effort. The issue of quality warrants continued scrutiny and concern.
Overall Summary In early 1985, INPO 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 deprest.ec 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 are still some 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 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 te preferential reporting of failures discussed in LERs, and 2) low report counts at some plants are an accurate reflection of their experi-ence.
{
l
_ _ _ - - - _ _ -