ML20213F123
| ML20213F123 | |
| Person / Time | |
|---|---|
| Issue date: | 02/08/1985 |
| From: | Dircks W NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | |
| Shared Package | |
| ML20213E956 | List: |
| References | |
| FOIA-87-137, TASK-IR, TASK-SE SECY-85-056, SECY-85-56, NUDOCS 8502200475 | |
| Download: ML20213F123 (22) | |
Text
t i
INFORMATION REPORT l
1 February 8, 1985 SECY-85-56 For:
The Commissioners From:
William J. Dircks Executive Director for Operations
Subject:
NUCLEAR PLANT RELIABILITY DATA (NPRD) SYSTEM Purpose :
To provide the Comissioners 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 INPO's manage-ment of the NPRD System and to provide the Commission with semiannual status reports on the effectiveness of INPO management of the NPRD System and the responsiveness of the NPP.D System to NRC needs.
l,
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), and on August 1, 1984 (SECY-84-44A).
These reports addressed the status of the NPRD System and the staff's program to monitor improvements in NPRDS data.
In addition, on May 27, 1983, Dennis Wilkinson, the President of INPO, briefed the Commission on the status of the NPRD System and on the INPO plans for further improvements. to this paper'is the staff's sixth semiannual report. This report describes the results of the staff's program to monitor and evaluate the quality and quantity of data available in the NPRDS data base..
)
i The staff has the following general observations based on the enclosed report.
1.
The data for the most recent quarter shows that the increase in the NPRDS reporting first seen in the third CONTACT:
p Frederick J. Hebdon,_AEOD 492-4480 i
\\
8c4z266475 XA w, /M, J
t
4 The Comissioners quarter 1983, is being maintained.
However, the percentage of failures reported in LERs that have been subsequently reported to NPRDS has remained relatively constant at about 40%.
Thus, although the data base is substantially improved over the past years, there has been no apparent change or further improvements for the past four quarters.
2.
A large percentage of eligible plants submitted NPRDS reports in the second quarter of CY 1984.
For this quarter ( April to June 1984),70 plants submitted at least one failure report. Only 5 plants did not submit any failure reports during the
. quarter. Although there are still many plants that have submitted a rather small number of reports (40 plants submitted from 1 to 20 reports each during the most recent evaluation quarter), the second quarter results provide a continuing indication that more plants are participating, at least at some '
minimal level, in the NPRD System.
3.
The timeliness of NPRDS failure reporting continues to improve, although at a modest rate.
In the second quarter of 1984, 46% of the reports submitted described failures that occurred in 1984, a substantial increase over previous comparable' quarters.
4.
The quality of the NPRDS reports (particularly the Failure Reports) has improved and remains essentially constant at a relatively high level.
Over 90% of the NPRDS failure reports were judged by the analyst to contain sufficient information to permit meaningful analysis.
5.
There is still considerable variation from plant to plant in the number of components included in the respective engineering data files. This will continue to make the use of the NPRDS data (e.g.,
for trends and patterns analysis) difficult because l
the reportable scope varies from plant to' plant.
6.
At the present time, the staff is makino only limited use of NPRDS data.
As confidence in the NPRD System increases, it is expected that use of NPRDS data by the staff will increase.
L J
The Commissioners
Conclusion:
The. improvements in NPRDS reporting first seen in the third quarter of CY 1983 have been maintained through the second quarter of CY 1984. While the overall improve-ments over the past 2-3 years have been substantial, further are still S plants (pected but have been limited. compared There improvements were ex CY 1983) that did not submit any failure reports during the quarter, and 40 plants (compared to 29 in the third quarter of CY 1983) that submitted 20 or fewer reports in the subject quarter.,
The participation in NPRDS continues to improve in terns of the number of participants and the number and tineliness of reporting.
However, for the second assessment period, it is noted that in some measures, little or not additional improvement was noted. Thus, continued INPO nanagement attention and action will be necessary to achieve a fully operational NPRD System.
INPO is currently undertaking an extensive rescoping i>f the program which staff is monitoring.
I h.
illiam. Dircks Executive Director for Operations
Enclosure:
Sixth Semiannual Staff Evaluation of the Current Status of NPROS i
1 i
L-
/
SIXTH SEMIANNUAL STAFF EVALUATION E
THE STATUS OF NPRDS January 1985 Prepared by:
Office for Analysis and Evaluation of Operational Data e-
INTRODUCTION The present NPRD System is a voluntary program for the reporting of reliability data associcted 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 i
failures (i.e., NPRDS). 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 Office (GA0) reviewed the NRC's data-gathering activities concerning unscheduled events at commercial nuclear facilities.
In a report issued in late January 1979, the GAO 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. Accordir. gly, an Advance Notice of Proposed Rulenaking (ANPRM) was approved by the Commission, and published in the Federal Reaister on January 30, 1980 (45 FR 6793).
Numerous public comment letters were received in response to the ANPRM. The predominant message in i
the comments was an overwhelming opposition to making participation in NPRDS mandatory.
l Despite the opposition to a mandatory system, the staff identified a strong t
need for failure rate 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 developed conceptually a revised reporting system. 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.
Wnile the staff still believes that both types of data are essential to the NRC mission, the possibility arose that the NRC could obtain tha 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, INPO 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 progress of INP0 management of the NPRDS and provide the Commission with semiannual status reports on the effectiveness of INPO management of the NPRDS.
In order to monitor the completeness and quality of NPRDS data, AEOD estab-lished an evaluation program at 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 NPRDS 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), and August 1, 1984 (SECY 84-44A).
In the past, each semiannual report has included analysis of NPRDS data from two calendar quarters.
However, in the third quarter of CY 1984, INPO, working with individual plants, began an extensive quality assurance review of all of the engineering data records (approximately 226,000 records) currently in the data base. The effort, called "rescoping", was necessary to conform the information in the NPRDS data base to recent revisions to the NPRDS Reportable and System Component Manual and the NPRDS Reportino Procedure Manual.
ine rescoping involves reviewing around 4,000 engineering records per nuclear plant and identifying plant components and systems covered by NPRDS.
Along with a change to the NPRDS system codes, the rescoping will add a new retrieval feature, 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 is slated for completion 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 has pro-duced a major short-term disruption in the data base.
In order to facilitate the rescoping the NPRDS data base was broken into three separate pieces; (1) a data base of data from plants that have not yet been rescoped (2) a data base of data from plants that are in the process of being rescoped (this data base cannot be accessed outside INPO), and (3) a data base of data from plants that have completed rescoping.
The staff found that analyzing NPRDS data while the rescoping was in progress was impractical
o because (1) data from some-plants (i.e., those in the process of being rescoped) was unavailable, and (2) data from plants that had been rescoped differed considerably from data for plants that had not yet been rescoped.
Thus, data for the third quarter of CY 1984 was not analyzed. This temporary disruption of the data base may also affect the next semiannual analysis.
As a result, a decision about whetner to analyze data from the fourth quarter of CY 1984 or the first quarter of CY 1985 will be made prior to preparation of the next semiannual report.
Thus, this report contains information using a copy of the NPRDS data base as it existed at the end of the second quarter of CY 1984.
COMPARIS0N 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.
Task A The first sample (Sample A) involved failures that occurred in January-August 1981 (before INP0 assumed responsibility for the management of NPRDS). Based on the description of the failures contained in the LERs, the analysts iden-tified 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 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 that should have been reported to NPRDS had actually been reported as of October 1981.
t h
l l
i
' The analysts also reviewed the entry in Biock 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 match referred to above was for an LER coded as not being reportable to NPRDS, i.e., a "N" in Block 24).
Table 1:
Results of the Task 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)
~
Task B and Beyond Subsequent samples of LERs have been analyzed using LERs from the first quarter of CY 1982 (Sample B) through the second quarter of CY 1984 (Sample K). These samples have been compared to the NRPDS data base.
NPRDS data for more recent quarters continues 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).
Beginning with Tape G, only the preceding three quarters were analyzed in this way.
The results of these analyses are sumarized 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 i
produced.
Prior to the two most recent sample quarters, each tape reflected the NPRDS data file as it existed one month after the close of the quarter to account for the pemissible 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 fourth quarter of CY 1983 contain the data base as it existed on the last day of the subject quarter (e.g., Tape K contains the NPRDS data base 4
as it existed on June 30,1984).
c y
5-Table 2: Samples NPRDS Reportable
- Sample Sample Period Failures in Sample A
Jan-Aug 1981 97 B
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 Table 3:
Versions of the NPRDS Data Base Version Date of Version A
Oct 1981 B
Apr 1982 C
Jul 1982 D
Oct 1982 i
E Jan 1983 F
Apr 1983 G
Jul 1983 H
Oct 1983 I
Dec 1983 J
Mar 1984 K
Jun 1984 L
- Data for the third quarter of CY 1984 (Tape L and Sample L) have not been analyzed (See explanation on Page 2).
l
~
. s Finally, Table 4 contains the results of efforts to find corresponding NPRDS failure reports for failures described in the sample LERs (e.g., Tape D contained NPRDS failure reports for 25 of the 97 NPRDS-reportable failures in the LER Sample A).
Table 4: Summary of Matching NPRDS Failure Reports and LER-Reported Failures Version (Tape)
A B
c U
t t
G H
I J
K L
A 11 21 25 25 30 31 B
0 4
8 14 21 C
1 2
5 16 Sample D
1 6
12 13 E
4 13 16 38 F
5 12 43 46
~
G 3
28 37 43 H
21 33 40 40 1
18 38 41 J
19 41 K
12 L
Conclusion The data for the most recent quarter shows that the increase in the NPRDS reporting, first seen in the third quarter 1983, is continuing. However, the initial percentage of matching NPRDS failure reports dropped slightly to about 12%.
However, for earlier samples (failures that occurred in late 1983 and early 1984) the percentage of failures that have been reported to NPRDS by June 1984 has remained relatively constant at about 4D%.
Thus, the data base is substantially improved over the past years, but there has been no apparent change or further improvement for t" past four quarters.
- Data from the third quarter of CY 1984 (i.e., Tape L and Sample L) have not been analyzed (see explanation on Page 2)
6 MEASURES OF NPRDS PARTICIPATION Table 5 indicates that a little improvement in the quantity of reporting was made.
In the second quarter of CY 1984, a total of 2356 reports were received; a 1% improvement over the previous quarter.
However, this is more than four times the number of failure reports added to the data base in the comparable quarter one year earlier.
In addition, Figure 1 is a plot of component failure transactions (additions plus revisions) 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 do show a definite rise in~ the level of activity for the system.
Table 5: TIMELINESS OF NPRDS FAILURE REPORTS Year in Transaction Quarter which the (Quarter in which the report was added to the data base) failure occurred 82-2 82-3 82-4*
83-1 83-2 83-3 83-4 84-1 84-2 1984 371 1072 1983 72 191 1088 1304 1019 602 1982 33 30 273 604 260 639 454 222 185 1981 188 13 264 201 34 146 132 159 139 1980 166 1
192 90 28 77 54 223 168 1979 105 0
149 75 6
19 43 74 113 1978 30 2
187 58 5
6 29 35 32 1977 14 0
233 13 1
3 14 50 25 1976 32 1
109 0
7 4
3 39 6
1975 7
0 86 3
4 3
8 37 8
1974 0
3 24 0
2 1
16 71 6
1973 5
1 11 1
0 0
0 7
0 580 51 1528 1117 538 1986 2057 2307 lEENI
- Significantly affected by one-time mass revisions.
l 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 often reflects the contribution of only a small group of plants reporting to the system.
For example, in the fourth quarter of CY 1982 a total of 1,528 component failure reports were entered into the file.
However, two units accountec for 783 (51%) of the total entries for the quarter, and there were no entries for 44 plants. Thus, summary statistics (e.g., average number of failure reports per plant per year) continue not to be good measures of NPRDS participation.
,, _ -. _ _ -. ~
yy-._..
-..m
Figure 1 2500 l
l
-?c 2250 s
2000
. '5 1750 1500 1
~
g 1250
=
1000 1
750 si 500 -
>s v
\\
250 s
0 1
2 3
il 1
2 3
ft 1
2 3
11 1
2 1981 1982 1983 -
198'l
' YEAR QUARTER e
O 9
g L
_g.
4 The dominance of a few plants in each quarter appears to have resulted from one or more plants conducting an extensive review of their maintenance records and submitting a large batch of reports of failures that occurred over an extended period of time. This adds to the difficulty of using any single summary statistic associated with NPRDS reporting.as a measure of the overall industry level of participation in the NPRD System.
i Figures 2 through 6 plot the number of plants vs the number of reports for each plant for the quarter (e.g., from Figure 2, in the second quarter of 1983, 22 plants submitted between 1 and 20 reports). Clearly a key figure to watch is the number of plants submitting zero reports in a given quarter.
By comparing Figures 2 and 6, a substantial decrease from 41 to 5 plants that did" not provide any reports is seen between the second quarter of CY 1983 and the second quarter of CY 1984.
However, this is not a sufficient measure of success since one report from each plant for a period of three months would result in 100% of all plants participating but would not constitute acceptable-participation. For the second quarter of 1984, for example, 30 plants submitted more than 20 reports.
Thus these plants were participating at the average rate projected by INPO (approximately 100 reports per year) for plants that are fully participating in NPRDS.
However, the participation of the majority of plants (45 of 75 plants) fell short of this goal.
Conclusions A large percentage of eligible plants submitted NPRDS reports in the second quarter of CY 1984.
For this quarter (April to June 1984),70 plants submitted at least one failure report.
Only 5 plants did not submit any failure reports during the quarter. Although there are still many plants that have submitted a rather small number of reports (40 plants submitted from 1 to 20 reports each during the most recent evaluation quarter), the second quarter results provide a continuing indication that more plants are participating, at least at some minimal level, in the NPRD System.
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 5 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.
It should be recognized that the 82-4 column (i.e., the fourth quarter of CY 1982) is somewhat distorted due to the mass revisions of older reports by two plants during that quarter.
~
l t
,n--
,---,---.m--
<_----,.,-,-..w--
--,---,,w--,.>
,,--v,nm-_-v, a,----e n-n--.
a----,,-n~----l
s I-
~
009-19t e
09t-ISt (f) c E~
- l2 09t-Itt '
C
-Z
. Ott-13t O
=
U I
029-10t 0
z!2 00t-19e. -
F cez Z
Oj obc _tec O
Mc.
CL
- ~.
e N
092-Itt e
co Ed e
OtC-12C $
O M
2 c.
02t-100 g
~
cd p
9 e
000-193 e A
2 093-192 &
~
<g
~
092-152 *c
,E E
5 P
093-123 g, !
Z c.
N 032-103 7
O 003_19; u.
CL O,
091-191 @'
2 o
O 091-Iti 3,
O z
y I >, Oti-IZI CO t-
\\
CD -
. I s' 021-101
~.
gg 001-19 L
m O'
09-19 09-It Z
NNyov-iz N
v s EE $ \\ % \\\\ % \\ % %
03-1 t>h%%%%N%%%\\%%%NNN O 6
6
+
6 e
e e
a 6
i
a
.AA.
~
009-IBt te
' 089-19)
,O C
09t-itt Z
z D
Ott-12t O
i.:
U OBt-10) o:
7 i
u Z r-00t-ISC NZ i
09 O.
gf-Oec-isc r.
u.
090-1)C.'
n o
ir-'
w n
.g Ott-IEC 5:
5 2
t-a g
Oze-10c g e
~
Occ-isa s 9
1 2
e 033-192 g g
3 093-ItB 'r-z.
L OtE-1EE j g
~ OIE-102 3 m
t E
Z m
Oct-let e cn O
c c
h 021-191 s es 4ss 3
E g.'s' 091-ItI :
%/
- ]
2 n
I!N Oti-1EI
%. J W
CD 1[s 031.10i Z
Z; 2,Q 001-19 y-U
~.
v)(-Q 03-19
~
~
~~~
'.... [.
.h,.....
9[R(Ns ',N's 09-It ',
~
.h.'Nh*!?M\\\\\\
M OHB 91 SE h % Ml M N k\\N'8.'I.) b k M 'N'N 03-1
~
6),hh% % h 0 k
f h
ct i
5 SINY'id 30 229MDhl n
Figure 4,
4TH QTR 83 COMPONENT FAILURE REPORT COUNTS
~
. LEGEND E21 NUMDER OF PLANTS WITillN RANGE a_
n
[f/,
n-pa-o, g
l/
/7 h o n-b-
f
/ /
g g
g bh n s e
.N.
.b.
1 7
7 i
7 7
7 i YT Tg 8
9Pf7 q.8 8
P S
8 8
8 8
o-S o
o' o
8 8'8
?
8 8
8 o
8 I
7T T
T T
T 1.T i
a
.1 u
$8 a
a a
a e,e a,e u#
a e
n
=
- a. m 8
81 N
n n
n n
,e
~
~
m N MDER OF COMPONENT FAILURE REPORTS
m
._a g ;
., e Figure 5 1ST Q'i'R
- 84. COMPONENT FAILURE REPORT COUNTS LEGEND t21 HUMUER OF P1,AN13 WITHIN ItANGE 3
~
g f
n-
/
f a.
1 g
//
3
/
~
f a n.
m
$p,.
g.
r.
~
g illLii
~
z s u,, 4.
. m.
. x e
..y
.se 3,,,,s
.*hEd,1~EsE'EsEE~EE.EEE,~,M55;*55 ss nm,nu u r vuwnn aura
~
00s-tet e
OBt-19)
E
=
Z 09t-1t9 p
5 y
o ii t; Ott-tzt U
E 02t-10t oZ@
6 00t-19C IX Q
o et OBC-19C n
4 C
p 09C-15C p~,,
g g
5 Ott-1BC lt 2
Ci M
2 E
Ozz-1ot
^
Cd 5
6 p.
00C-193.E 4
E
-4 093-193 N
-093-1 taw 2y OtE-153 2 Z
g
'032-103 j
=
[
Z L
O 003-191 t m
C n
f Q09-191f I
I T
i O'
091-Iti 3 O
3 Oti-IEI
~
x a(001-Is
- p g
C h 09-19
'Lgg 09-1t h Q Q Q N OF-12 N
il Ot h h h\\ M \\ M \\ h M M M \\ M k h 02-1 9h%
0 W
S.!RY74 JD ESEKDR
y t
i
' i In terms of the timeliness of reporting, in the second quarter of CY 1984, 1072 component failures were reported that occurred in 1984. This number is 46% of
^
the total entries for the quarter.
By comparison, in the second quarter of CY 1982, only 33 reports (6%) of the reports described failures that occurred in 1982.
Conclusion
]
The timeliness of NPRDS failure reporting continues to improve, although at a i
modest rate.
In the second quarter of 1984, 46% of the reports submitted described failures that occurred in 1984, a substantial increase over previous comparable quarters.
QUALITY OF THE NPRDS FAILURE REPORT NARRATIVE To assess the quality of the NPRDS Failure Reports, a sample of one NPRDS Failure Report per plant was selected for each plant filing a report in the latest quarter.
The narrative description of the failure was reviewed to determine if the text described the failure in sufficient detail that system users could understand the failure, its cause, corrective actions, and the implications for similar equipment. The narratives were graded as adequate,
]
probably adequate, and inadequate. The results of this analysis are presented j
in Table 6:
i Table 6: QUALITY OF NPRDS FAILURE REPORTS Third Quarter Fourth Quarter First Quarter Second Quarter 1983 1983 1984 1984 Adequate 50%
71).
63%
46%
82%
95%
97%
92%
l Probably Adequate 32%
24%
34%
46%
Inadequate 18%
5%
3%
8%
Total Number of j
Plants 60 62 68 72 The percentage of adequate and probably adequate narratives increased from 82%
in the third quarter of CY 1983 to 97% in the first quarter of CY 1984.
Unfortunately, it decreased slightly to 92% in the second quarter of CY 1984.
i a
4 s
.__~..- -__-,..~,
e Conclusions The quality of the NPRDS reports (particularly the failure reports) has improved and remains essentially constant at a relatively high level.
Over 90% of the NPRDS failure reports were judged by the analyst to contain sufficient information to permit meaningful analysis.
ESTIMATED AVERAGE RATE OF NPRDS REPORTING The completeness of the NPRDS data base has been a long standing concern, particularly with respect to failure reporting.
For many. years, most plants were submitting few, if any, failure reports, so it was easy to conclude that the completeness of the data base was unacceptable.
Now, however, most plants are submitting some number of NPRDS failure reports [e.g., in the most recent quarter evaluated, only 5 (7%) of 75 plants did not submit any failure reports during the quarter). However, many plants submitted a rather small number of reports (e.g., 40 plants (537,) submitted from 1 to 20 reports during the quarter].
While it is clear that one report per quarter is not indicative of full participation, it is not clear how many failure reports are expected from a
" typical" plant in a " typical" quarter.
The rate of reporting is expected to vary considerably from plant to plant and from quarter to quarter based on a number of factors including maintenance practices, operating cycle, etc. Thus any effort to quantify a " typical" reporting rate should be viewed only as an overall measure of system implementation and not as a reporting quota.
With this caution in mind, work is underway to quantify how many failure reports should be expected from a " typical" plant in a " typical" quarter.
The first phase of this activity was to sort the engineering data for 8 plants for which the rescoping (discussed on page 2) has been completed.
The data were sorted by plant and then by generic component class to obtain counts of the number of each type of component in the engineering data of each plant.
This data reflected considerable variation from plant to plant in both the total number of components and the number of components in each type (e.g.,
the total count ranged from 1556 components for one unit to 3592 components for another unit).
INP0 has estimated that the current reportable scope should ' include approximately 5000 components per unit.
Thus, it appears that much work remains to obtain a consistent reportable scope for all plants.
If there is not a uniform reporting scope for similar plants, this would be a major concern because of the uncertainties introduced in the analysis and use of the data.
.I
i j
i o
- 17 This assessment will continue and be expanded to develop a measure of the level of reporting that may be expected from a typical plant in a typical quarter. As a minimum, this work should form the basis for discussions of measures that could be used to differentiate between plants that are fully 1
participating in NPRDS and plants that are providing only a token number of reports. Such measures are essential to determining the completeness of
)
reporting and the validity of the resulting data.
Conclusions i
There is still considerable variation from plant to plant in the number j
of components included in the respective engineering data files.
This will continue to make the use of the NPRDS data (e.g., for trends and patterns analysis) difficult because the reportable scope varies from plant to plant.
Current Uses of NPRDS Data by the Staff In order to assess how widely and in what ways the staff was using NPRDS data, we requested that appropriate program offices and the regions provide a brief discussion of their recent (i.e., during CY 1984) experiences in l
retrieving and using NPRDS data. Specifically, we requested that they describe when and how they used the NPRD System, including a description of the need to be met, the data requested, the source of the data (e.g., direct on-line access, direct contact with INPO, request for data from NPRDS via 4
AEOD), the data received, and the usefulness of the data in meeting the original need. In addition, we requested a description of any specific situations where they identified a need for NPRDS data but their efforts to i
obtain the required data were not successful.
i The response to this request indicates that the use of NPRDS data by NRC i
offices is rather limited, yet on occasion the data has been useful.
In general, such use consists of a search of the NPRDS data to identify failures 3
with specific ch:racteristics (e.g., steam binding of auxiliary feedwater pur.ps) in order to supplement lists of such failures obtained from other sources (e.g., LERs).
As the NPRD System continues to expand and improve, and confidence in the system increases, it is expected that NRC staff use of the 1
system will expand.
i In addition, AE0D is developing methodology to permit the systematic analysis of NPRDS data for trends and patterns. This analysis will supplement existing j
trends and patterns analyses of LER data.
It is expected that the methodology 1
for this program will be completed in mid-1985 and will be foi Swed by a phased implementation schedule.
l 4
l
.w.. -,.
- Conclusion i
At the present time, the staff is making only limited use of NPRDS data as a supplement to other data sources.
As confidence in the NPRD System increases, it is expected that use of NPRDS data by the staff will increase, j
OVERALL
SUMMARY
The improvements in the NPRD System reporting first seen in the third quarter of CY 1983 have continued through the second quarter of CY 1984.
While the overall improvements have been substantial, further improvements were expected, but have been limited. There are still 5 plants (compared to 9 in the third quarter of CY 1983) that Jid not submit any failure reports during the quarter, and 40 plants (compared to 29 in the third quarter of CY 1983) that submitted 20 or fewer reports in the subject quarter.
The participation in NPRDS continues to improve in terms of the number of participants and the number and timeliness of reporting.
However, for the second assessment period, it is noted that in some measures, the rate of improvement has slowed and little or no additional improvement was noted.
Thus, continued INP0 management attention and action will be necessary to achieve a fully operational NPRD System.
1 l
m----.-.
,p
-._,e.3m,,-...,m-,
_