ML20213G440
| ML20213G440 | |
| Person / Time | |
|---|---|
| Issue date: | 02/24/1987 |
| From: | Stello V NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | |
| Shared Package | |
| ML20213E956 | List: |
| References | |
| FOIA-87-137, TASK-PII, TASK-SE SECY-87-054, SECY-87-054-01, SECY-87-54, SECY-87-54-1, NUDOCS 8705180288 | |
| Download: ML20213G440 (31) | |
Text
_ - -..::2.._ _
..2-
,~
mm :--- ~ - -.
~..
,S
.a p * "%
s
~
nQ
\\*.v.../
POLICY ISSUE February 24, 1987 (lgf0ffM8tlOM)
SECY-87-54 For:
The Comissioners Fron:
Victor Stello, Jr.
Executive Director for Operations
Subject:
NUCLEAR PLANT RELIABILITY DATA SYSTEM (NPRDS)
Puroose:
To provide the Commissioners with inforr.ation on the status of the MPRDS for the second and third quarters of 1986.
Discussion:
In its affirmation of SECY-81-494, the Comission directed the staff to periodically assess the quality and Quantity of informa-tion produced by NPRDS and to provide the Comission with status reports on the effectiveness of INP0's management o# tha NPRDS and the responsiveness of the FPRDS to NPC needs. The NRC staff reutinely uses NPRDS in its safety activities. At tachnent C provides the staff's assessment of tha NP9CS. Appendix A of this attachment gives a brief history and the Ccerission's involvement in the NPRDS.
The first report was forwarded to the Commission on July 1, 1982 (SECY-82-279).
Subseouently, the staff has prnvided semiannual reports to the Conmission on the stctus of the NPRDS and the staff's program to monitor improvements in the NPRDS data.
The attached report, which is the staff's ter.th semiannual report, describes the results of the staff's prograr to monitor the quality, quantity and usefulness of data available in the NPRDS data base; the current programs involving the NPRDS; and the future directions and goals of our periodic evaluations.
Current Procran Evaluation At the outset of this evaluation report, wa consider it important to highlight the continued commitment of the industry to the improvement of the NPRDS.
In particular, we CONTACT:
Robert L. Dennig, AEOD B705180288 870513 S g -137 492-4490 P
PDR SECY NOTE: This document contains information that should not be released outside the NRC. Appendix B of Attach:nent D should be removed before release is made.
t
{~ h 7
i
=,
,e w
'The Commissioners -
are pleased with INPO's progress in the conversion of the NPRDS to an IBM based system that will significantly inprove the NPRDS capabilities and greatly enhance its usefulness. Addi-ticnal INPO.iritiatives, to be discussed later, are also in progress. Although this evaluation reveals certain problems with the NPRDS, we believe that there has been progress over the period and that there is still substantial momentum behind the further improvement'of this reporting systen.
A summary evaluation of the NPP.DS overall trend through the third quarter of 1986 indicates that the cuantity, quality, and timeliness of reporting, as measured by the staff's current LER based monitoring program, appear to be stabilizing. Approxi-i mately 75% of the component failures reported in Licensee Event Peports (LERs) are reported in the NPRDS within one year (Figures I and 2; Attachments A and B, respectively). These NPRDS reports are of a quality probably adequate for many users (Figure 3; Attachment C). As will be discussed later, the data quality does pres nt some difficulty for engineering analysis of the NPRDS data.
The following specific observations resulted from the staff's activities involving NPRDS:
1.
NPRDS data show that within nine months of the end of the third quarter of 1985 (i.e., by the end of Jur.e 1986) approximately 76% of the NPRDS-reportable failures described in a sample of LERs were in the NPRDS data base.
For the fourth quarter of 1985 the figure is 51% (Figure l
1). Although the 51% ficure for fourth quarter 1985 is a j
decrease from the previous quarters, which were around 75%,-
it is an increase from approximately 40% observed for quarters in 1984 and before.
In addition, this low figure may reflect a tenporary backlog or other administrative data entry delays since the first quarter of 1986 was already at 63%.
2.
The timeliness of NPRDS failure reporting for the third quarter of 1986 has improved compared to the same period in i
1985. Figure 2 shows that the reporting timeliness has been stabilizing over the years to the point where half of the failure event reports are in NPRDS within about six 4
months of the failure. Current NPRDS reporting guidance indicates reports would best be submitted within about 30-60 days.
Thus, improved timeliness is desirable and has been the sub. lect of recent NPRDS Users Group discussions.
3.
In each of the past two quarters, all but one plant sub-mitted at least one failure report. During the first three quarters of 1986, about 20 plants submitted fewer than 30 -
reports. Additional on-site verification of complete reporting appears appropriate for low reportina plants and INPO has taken the initiative to have such verification efforts in progress as part of their program.
9 3.-
l
' The Commissioners 4 The staff continues to assess whether its currant program for measuring the level of NPRDS reportino accurately reflects the overall-level cf reportieg.
The staff's current yardstick for measuring NPRDS reporting utilizes the component failures reported in LERs. The majority of component failures are not reparted in LERs.
In addition, in an effort to improve the NPRDS, INFO has promoted consistency between LER and NPRDS failure raporting.
Therefore, the reporting lavel of ccmponent failures fecn LERs that ara also in NPRDS may be hicher, perhaps signifi-cantly, than the overall reporting level.
During this past evaluation period, we tested two nethods of measuring the overall completeness of NPRDS failure s
reporting.
The results of one test for components in three systems at one plant found 64% of the reportable failures in the NPRDS data base.
The preliminary results of the second test of four key components ef nine plants found an even smaller percentage of the failures in the data base.
In contrast, a 75% level of completeness is indicated by 1
LER based failures. Although no conclusions car be drawn from these two checks as to the overall completeness of the data base, they do indicate that diversified neasuring nethods should be pursued.
5.
From an NRC user perspective, sete difficulties -heve been found in using NPRDS data, such as irccrsistencies in the
~
engineering reports, nissing engineering records, and errors in the data.
Some of these problems are to be expected of any data base systen; however, some of these observed difficul-ties may degrade the usefulness of the results and ray not be obvious to the casual user.
Improvement of the engineer-ing records is also identified by INFO as an area worthy of further work.
We expect improvement in this area in the futuro.
)
6.
To assess the quality of the narrative information provided in NPRDS failure reports, a sample of reports was reviewed to determine if the text described the failure in suffi-cient detail such that system users could understand the failure events.
For the last two quarters evaluated, over 90% of the samples were rated as at least "probably adequate" (i.e., a kncwledgeable person could understand the characteristics of the failure). However, for useful analysis and proper corrective actions following analysis, sufficient information is needed for users to understand the causes of the failures and the corrective action) taken.
The NPRDS Reporting Procedures Manual requires "a narrative stating the cause, or the suspected cause, of the failure" or "if root cause is not determined, a best judgment cause."
It also requires that the narrative describe both ::hort-term and long-torm corrective actions taken..Recent NRC csse studies that attempted to use NPRDS
l t
a j ',-
l The Comissioners, ;
data found that narrative descriptions in the NPRDS failure reports wtre often too brief to adequately determine f&il-ure causes.
For this reason our recent evaluatinns have also measured the percentage of failure reports that pro-vided good root cause information.
Few reports met this
/
test.
In addition, the current reporting guidance of the NPRDS may not support some anticipated end uses (e.g.,
analysis of wearcut and aging due to not capturing instal-part of periodic preventive maintenance) placements or as lation of a new ccmponent for in-kind re Improvement initiatives in this area were discussed recently at INP0'.
within the context of the NPRDS Users Group reeting and an NPRDS subcommittee has been established that will visit such needs.
Current Programs Using and ' Evaluating NPP.DS Data With the i,nplerentation of the LER rule in 1984, the NPRDS became the primary source of. component level failure data. As the system has improved, it has found increasing use by the NRC and the industry. With this increased usage of the system, i
there has been a concurrent increase in feedback from users.
As roted previously, NRC programs using NPRDS data Feve identified a number of ' areas reeding possible improvenent.
Use feedback is used as an integral part of the evaluation pro' cess. Some of the cr-going NRC activities providing feedback cn the use of NPRDS data are:
1.
NRC programs using NPRDS data for' analyzing and evaluating.
the operational experience of plants,. e.g., AEOD case studies, engineering evaluations, tren'ds and patterns studies, and in preparation for plant visits. NPRDS data are also used by NRC's Office of Inspection and Enforcerrent (I&E) in preparing Information Notices :and Pulletins and in evaluations by I&E's. Vendor Program Branch.
In addition, the Office of Nuclear Regulatory Research 'is using NPRDS data in progri.ms such as the Nuclear Plant Aging Research Program.
2.
In addition to using LER information to evaluate the NPRDS data base, new methods are being used to measure the com-pleteness of the overall NPRDS data base such as the two trial activities noted previously. The first test compared
'the maintenance work orders collected by.NRC's Performance Appraisal Team (PAT) on one of their. recent plant apprais-als to the NPRDS records for that plant. The second test compared information on plant maintenance activities with-the NPRDS failure reports for a sample of key components j
at a sample of plants.
i-j 1
The Commissioners Overall Status and Future Direction for the NPRDS Improvement and Evaluation 4
We are pleased to report that our evaluation of the NPRDS data base over the last two-years has shown improvements in the j
number of plants participating, in the number of reports being submitted by the plants, and in the timeliness and quality of the data. However, the data base completeness and quality continue to warrant attention. Otherwise, its usefulness is limited by the resources required to verify and upgrade the data after it is taken from NPRDS.
As noted in the previous section, user's connents are becoming an increasingly important part of the NPRDS evaluation / improve-4 l
ment process.
We have been and will continue to provide INPO with this user feedback.
Thus, future evaluations by the staff will be directed at determining failure reporting completeness and NPRDS appropriateness for the various end uses, including regulatory activities, and at stimulating needed improvements.
In 1987 INPO will be pursuing a number of prngrams to improve the quality and completeness of the'NPRDS data base. To verify failure reporting, INP0 will continue to review maintenance records for its' plant evaluations and to have all incoming failure reports reviewed by experienced engineers. To monitor the accuracy of the er.gineering reports, INPO is planning on reviewing, during plant evaluations, records of components that have been taken out of service.
INPO's conversion to the use of an IBM based system, which will begin implementation in 1987, will provide NPRDS users major improvements such as extensive interactive help, new calculated data, data analysis tools, enhanced report capabilities, and easier use. The conversion project is a major INPO initiative with a goal of stimulating NPRDS participation through increased usage by utilities.
i We have discussed with INPO the possibility of the staff working closely with'INPO to independently review the data collected by INPO to assess'the quality and completeness of the NPRDS data base.
For example, INP0 reviews a significant number of main-
-j tenance records for each plant evaluation to evaluate the NPRDS reporting for that facility.
The staff is discussing with INPO possible approaches to also review this data and the INPO 1
analysis of it as a substitute for the. current analysis based upon failures reported in LERs.
Should cooperative activities using INPO collected' data not prove practical, the staff will explore alternative methods for determining' the overall completeness of NPRDS failure reporting. Applicable methods to be explored include:
comparison of the maintenance work orders. collected by the i
PAT teams, the possibility of comparing the NPRDS engineer--
ing and failure records with 'the component population and
'. s The Commissicrors failures reported in the In-Plant Reliability Data System (IPRDS), a comparison of the engineering records data with information in the Equipment Qualifications Data Base, and checking the completeness and accuracy of the engineerino data through creputer counts and cross tabulations of the NPRDS data.
As we gain more confidence that the NPRDS is fully adequate for NRC usage in safety activities, we will consider recommending to the Connission that we revise cur evaluation frequency from semiannual to annual and perhaps biennial.
r-
$7 L ~
g a h.
o Victor Stello, Jr.
Executive Director for Operations Attachments:
As stated CISTRIBUTION:
Cormnissioners C 3C (H Street)
C:
e
Attachnent A l
NPRDS Reporting Completeness
[ Based on LERs Only]
oo-so-n-
3.
o so-n n
a.
i
.=
So "
So-o 1-82 3-s2 1-83 3-es 1-84 3-e4 1-es 3-es 2-e2 4-82 2-83 4-83 2-e4 4-e4 2-es 4-se Quarter in which the Event Occurred ri; v i
-s Attachment B NPRDS Timeliness
[ Based on 1985 LERs]
soo e>
eo- -
re- -
E-E
_a So-e-
M so-so -
e 1
2 3
About 1 yr.
6 Time After Event (Approx. Quarters)
- Indicates range of values for 1985 quarters Figure 2
Attachment C
\\
l
\\
NPRDS Narrative Quality
[Most Recent Yr.]
1 I
1 I
a.
I I. u.
8...
I w
,u..
f Q
I-u..
"T a.s --
y s-es 4-es 1-ee n-se a-se Event Quarter
. -I: ;;;WS.O. Probehty.^_f;,__". - E.0 Figure 3
.p yr-w 2
y-v
-g g,
w-3
-y-,---
ww
--n
- + - -
g
---vv
NPRDS Reporting Comploteness
[ Based on LERO Only)
,oo 30 +
eo +
li
- s ro.I 5
i i5 so-n 4c +
2zs **'
.M so-io +
o 1-82 3-e2 1-s3 3-83 1-84 3-84 1-86 3-86 2-82 4-82 2-83 4-83 2-84 4-84 2-85 4-85 Quarter in which the Event Occurred Figure 1 NPRDS Timeliness
[Besed on 1985 LERs]
.co eo
- eo-ro +
EE-l w
i z
s oo-g.40-so-so-I to-o 1
2 3
About 1 yr.
Time Af ter Event (Approx. Quarters)
- Indicates range of values for 1985 quarters w
.-g
,-y-
3 e
Attachment D-i TENTH SEMIANNUAL STAFF EVALUATION Of THE STATUS OF THE NPRD SYSTEM January 1987 Prepared by:
Offica for Analysis and Evaluation of Operational Data i
NOTE - ihe NRC has agreed not to release plant-soecific NPRDS data.
Thus removal of Appendix B is necessary prior to release of this document outside the NRC.
t O
4
'I 3
j l
INTRODUCTION The cresent Nuclear Plant Reliability Data Syster (NPRDS) is a voluntary pro-cra., for the.vaporting of reliability data associated with selected components and systems in nuclear power plants.
Since January 1, 1982,-the MPRDS has been traraged and operated by the Institute of Nuclear Power Operations. Appendix A i
provides a brief his+ory of the NPRDS.
I C N PLETENESS OF NPPDS FAILURE REPORTING
{
Ccroarison of NPRDS Failure Reports ard LERs The evaluations of the completeress of NPRDS failure data have cere.isted of 3
sarpling approximately 100 LERs and attempting to find corresponding NPRDS failure reports for the failures describt.d in the LERs.
Tha first sample (Samole A) involved failures that occurred in January throu Aucust 1981 (before INP0 assumed respensibility for the niinagement of NPRDS)gh 4
Based on the description of the failures contained in the LERs, the analysts identified 97 failures that they believed should hava been reported to NPP.DS.
The analysts then attempted to fird corresponding NPRDS failure reports for -
the NPRDS 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 d
@ scribed in the LERs were found.
Therefore, less than 12',' (i.e.,11 of 971 of the failures that cheuld have been reported to NPRDS had actually teen reportec as of Octcber 1981.
Subsequent sanples of LERs have been analyzed using LERs from the first ouarter i
of CY 1982 (Sample B) through 'the third quarter of CY 1986 (Sample T). These saroles have been compared to the NPRDS data base.
NPRDS data for more recent quarters will continue to be evaluated quarterly and the results described ir 4
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 T, only the preceding threa quarters were analyzed in this way.
For tapes L through 0, tha analysis included the preceding four quarters in order to determine if substantial numbers of failures were being reported as much as four quarters after the event.
The results of these analyses are summarized in Table 1.
The second and third columns of Table 1 identify the sample period and the number of failures reportable to NPRDS from each sample [e.g., the sample from the period October-i l
December 1982 (Sample E) described a total of 110 component failures that should have been reported to NPRDS].
It should be noted that prior to the i
fourth ouarter of CV 1983, each tape reflected the NPRDS data file as it i
existed one month after the close of the quarter to account for the per-missible 30 day delay in submitting NPRDS reports (e.g., Tape E contains the T
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 i
-,--.a
_.m.
_m TABLE 1. SupetARY Of MAIClllNG LER COMPONENT TAULTS AND CPHD-4 FORM 5 Total Matches (1)
Reportable
_ _ _,,,_,,,,_ Tape Component C
c c
Sample Quarter
- Faults A
8 C
D E
F G
H I
J K
L M
N O
P gL, R S
T A
3-8tb 97 11 21 25 25 30 31 8
1-82 104 0
4 8
14 21 C
2-82 99 1
2 5
16 D
3-82 107 1
6 12 13 E
4-8/
110 4
13 16 38 F
1-83 112 6
I?
43 46 G
2 83 99 3
28 37 43 H
3-83 101 21 35 41 41 41 41 42 I
4-83 100 18 38 42 43 44 50 J
1-84 100 19 41 42 42 43 K
2-84 100 14 28 44 49 68 L
3-84 100 7
24 30 58 M
4-84 100 5
27 64 68 Q
1-85 100 5
63 71 75 0
2-85 100 17 49 72 76 P
3-85 100 11 50 68 16 Q
4-85 100
~13 38. 51 51 R
l-86 100 14 48 63 5
2-86 100 9
50 T
3-86 100 30 a.
Quarter in which events in LER sample occured, b.
For Phase A, et-nt dates were spread among the first three quarters of 1981.
. c.
Tape N contains data for quarters L(3-84), M(4-84), and N(185).
e b
a u
2-o day of the subject cuarter (e.g., Tape X contains the NPRDS data base as it existed en June 30,1984).
Finally, the renairing columns cf Table 1 certain the results of ef'erts to find correspcoding NPRDS failure reports for failures described in the sarple LERs (e.o., Tape T contained NPRDS failure reports for 51 of the 100 NPRDS-reportable failures in LER Sample Q, fourth ouarter 1985; 63 of the 100 NPRDS-reportable failures in LER Sample R, first quarter 1986, etc.)
Figure 1 shews tha percentages from Table 1 that are in the NPRDS data base aporoximately three quarters after the event. Although the 515 fige e for frurth quarter 1985 is a decrease from the pravious quarters, it is an increase from approximately 40% observed for quarters in 1984 and before.
In addition, this low figure may reflect a temporary backlog or other administrative data entry delay. Data from Table 1 were used to develop Figure 2 which shows the percentages of reports that were in the data base one, two, three and four quarters after the event, based on 1985 LERs. Approximately 75*,of the crrponent failures reported in LERs are reported in the NPRDS within one year.
It is interesting to note that new NPRDS failure reports are still beirg 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 firal quarter of CY 1983 (Sample I).
- n addition, it is important to note that the repcetability of a failure is subject to some interpretation.
Based en INP0's review of one set c' failures ce' ired by us as recortable, we estimate that for apprcximately 105 of the
'aures, the utilities could reasonably conclude that the failure was not recortable to NPRDS because it was incipient in nature, (i.e., a cor.dition v:rere preemptive corrective action was prudent, but no loss of function had yet cccurred).
Thus, the number of reportable failures in each sample as ce: ermined by the reporting utilities, may be as icw as about 90 failures.
Firally, it is also important to note that results for our samples may over-estimate utility reporting against the entire NPRDS scope in the most recent cuarters.
It is our understanding that INP0, in an effort to improve failure reporting, 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 ware subse-cuertly reported to NPRDS may not accurately represent (may be hipher than) the percentage of all NPRDS-reportable failures that are actually reported to NPPDS, because the sample is no longer representativa.
During the past evaluation period the staff has been investigating other scurces of infermation on comporent failures which might be used to verify the coroleteress of the overall NPRCS data base.
Some of these methods of evalua-tion are discussed at the and of this report.
Comoarison of NPRDS Failure Reports and Maintenance Records at One Plant and for a Sample of Key Components at a Sample of Plants 1
For the current evaluation report we have pilot tested two of the proposed i
methods of evaluation.
I'
' 7 3-
~
o The first test compared the maintenance records collected by NPC's Performance Aporaisal Team (PAT) with tha NPRDS failures reported by that plant since Jaruary 1, 1984. The PAT team is a grcup of experienced inspectors who conduct corprehens've inspections of operating reactor facilities' manager.ent contro' systems and related performance to determine their adequacy. The team focuses on such selected areas of plant activities as operations, maintenance, surveil-lance testing, design. change and modification, and training.
This fall the PAT team inspected the HPCI, LPCI and Containment Spray System. Following their review our analysts compared the maintenance work orders for these three systems collected by the PAT team with the NPRDS failure reports. The results of this evaluation are shown in Table 2.
Sixty four percent of.the reportable failure events since January 1,1984 for these three systems were reportad to NPPDS.
The preliminary results of the second test which compared the maintenance activities for four key components at nir.e plants with the NPRDS_ failure.
reports from those plants found even less than 64% of the reportable fail-ures in the NPRDS data base.
In contrast, a 755 level of completeness is indicated by LER based failuras.
Conclusions The NPRDS data for the fourth quarter of CY 1985 and the first three quarters of CY 1986 show that the percentage of component failures from LERs that have ultimately reached the NPRDS data t'ase has increased frcm approximately 40% to app-oximately 755. This is a substantial improvement compared to previous quarters.
Table 2.
Summary Of Matches Ft,r Maintenance Work Orders Since January 1,1984 System Reportable Events Matches Found
% Found HPCI 15 7
47 LPCI 4
4 100 l
3 100 l
1 22 14 64 l
[
j However, the possible preferential reporting of failures documented in LERs is of cencern for two reasons:
(a) it seriousl the overall level of NPRDS completeness and (b)y impairs the ability to judge it places a high priority or d
emphasis on reporting of selec+ed failures which is inapprnpriate. For example, independent, random reactor trip breaker failures are not report-able as LERs, but are directly reportable to NPRDS. The results of~two tests l
to measure the ccmpleteness of overall NPRDS reporting seem to indicate that our concerns over preferential reporting of LER failures may be justified.
_,,m_
_. - _ _ _ _ ~, _ _ _ -,.. _.. - -,,,, _ -., - - _. -., - --.,_._---., _,-,,.,..._.m.
- w 4-MEASURES OF NPRDS PARTICIPATION Table 3 indicates that the cuantity of reporting has increased sharply in recent quarters.
In the third quarter of CY 1986, a tctal of 6009 reports were received; a 130" increase compared to the third quarter of CY 1985.
Figure 3 is a plot of component failure transactions for the NPRDS as a function of calendar quarter. While these numbers do not reflect'the lavel of timely reporting, they do in a sanse 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. This effort, called "rescoping," was necessary to conform the information in the NPRDS data base to revisions to the MPRDS Reportable System and Component Score Manual and the NPRDS Repor~tino Procedure Manual.
Table 3.
NPRDS Reporting Rate Failure Reports Quarter Received 84-1 2307 84-2 2356 84-3 1663 84-4 1095 85-1 2121 85-2 4677 85-3 4625 85-4 4189 86-1 4672 86-2 6094 86-3 6009 One is tempted to divide the quarterly figures by the nurber 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 can 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 com-ponent 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) may not be good measures of NPRDS participation.
Figures 4 and 5 plot the number of plants vs the number of reports submitted during the quarter (e.g., from Figure A, in the second quarter of 1986, 19 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 ouarter.
From Figures 4 and 5, it can be seen that in each of the last two quarters only one plant did not submit any failure reports during the past two quarters.
a
-r.
-*.L
=+u..
~, -
n4 A
_.g 4
a 4
m 4
i 8500 i
i i
i i
i i
i i
i i
i i
i 1
i 6000 5500 5000 4500 4000 cr 3500 3000 2500 u>
2, 2000 1
3 1500 i
as
!1 1000 i
500 i
O i
i i
i i
i i
i i
i i
i i
i i
i i
i i
i j
1 2
3 4
1 2
3 4
1 2
3 4
1 2
3 -4 1
2 3
4 1
2 1981 1982 1983 1984 1985 1986 Quarter - Year Figure 3. Component failure transactions by calender quarter.
4 j
^
The staff also considered tha level of participation of irdividual units by lecking at the number or reports submitted during the year by the plant.
Figure 6* irdicates the number of reported failures that occurred in CY 1985 arc Figure 7, tha number of reported failures that occur-ed in CY 1986 that had been reported by the end of the third quarter.
The staff is still concerned about the number of plants that reported a rela-tively small number of failures in 1985 (e.g., 27 plants reported fewer thar 40 faflures for the year).
INP0 has stated that during an en-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 occur-ed in 1984. INP0's on-sita review confirmed the completeness of the unit's NPRDS reporting. While the low nurber of reports from this unit signaled INPO to investigate rore closely, INPO concluded that the low number of failures accurately represented the per#ormance of reportable components at this unit.
Conclusions Essentially all of the eligible plants submitted at least one failure report during the second and third ouarters of CY 1986.
TIMELINESS OF NPRDS DATA On January 28, 1983, INFO sent a letter to utilities urgirg that priority in failures reporting be given to recent failures.
The staff has examined the data to determine the degree of respense to this latter ir tarns cf quantity and tineliness of the data.
Table 4 shcws the component failure records sorted by the transaction quarter (i.e., the cuarter that the record was added to the data base) and the year in which the failure occurred.
Table 4 Timeliness of NPROS Failure Reports Year in Which the Failure Transaction Ouarter Occurred Quarter in Which the Report was Added to the Data Base 84-3 84-4 85-1 85-2 85-3 85-4 86-1 86-2 86-3 1986 331 2282 3730 1985 185 1643 2424 2952 3204 2496 1305 1984 1241 1002 1217 2622 1466 970 714 837 489 1983 217 70 322 221 222 109 230 165 129 1982 56 10 153 42 34 26 96 97 86 1981 37 1
66 45 15 29 83 85 27 1980 51 3
135 42 82 24 2
8 53
{
1979 14 0
5 25 158 19 2
17 33 1978 12 4
17 26 44 12 1
28 54 1977 14 5
9 3
27 8
3 49 56 1976 10 0
2 3
114 20 3
23 45 1975 7
0 7
5 33 12 2
7 2
1974 4
0 3
0 6
3 1
0 0
1973 0
0 0
0 0
0 0
0 TOTAL 1663 1095 2121 4677 4625 4184 4672 6094 6009
- The underlying data to support this summary table are provided in Appendix B.
The NRC has agreed not to publicly release plant-specific NPRDS data la n
annanA4v D\\
Number of plants (total = 88) w w
T T
T T
Y M D
E_'
O.
g hx g x u % \\ % \\\\\\\\\\\\ m \\ u g x u xx\\t s c
i-20 k%\\
21 - 4 0 41 - 60 %\\%\\%%%%%%y M f
j ph\\\\%%%%%\\%%%4 g
[
61-80 81-100 h\\\\ % \\ % \\\\ % % % \\\\N $
l 0
l 0.
101-120 \\\\%\\\\N e l
g
{%\\%y e Q
c z
O c
121-140 3
o-141-160 Y
161-180 O
181-200 h%y W E
es o
g O
201-220 N
,o a
E 9
V 221-240 [-
M i
O 3
241-260 a
o
~
261-280 g
E E
h {E O'
281-300 z
C 301-320 g
q l
EI r-l 321 -340 m
3 CD E k 5
h_. 5, 341-360 o
z O
l iii O
c' 3
361-380 h-E V
2 s
O 381-400
-[a y
M
]
401-420 p*
5-421-440 8
O e
c 441-460 O
a a
in 461-480 g
E 481-500 n
l s.
Number of plants (total = 68) u n
u i
?
T T
T TW
[-
g l
0 m
6' 1-20
%%%%%%%$ Ce O
c e
21 -40 h%%%%%%%%%%%%\\%\\%%%\\$ %
4
(
41-60 \\\\\\%%%\\M\\\\\\%\\\\%\\N s
~
~
CO 6 -80 %\\\\\\M\\\\M\\' *
~
~T 81-100 hh\\%%%%%%
[
e Q
101-120 W
O 121-140 %\\\\Y
- a 141-160 W
g-00
'm
["
161-180 j
O Q
[~
E 18!-200 g
O O
O 201-220 3
5 V
221-240 N
S.
O 3
241-260 [-
E O
c.
3 261-280 o
see O
3 D
281-300 E
b, 301-320 g
l 3
g G
j e
l E
g 321-340 5
(
R3 341-360 i
e2 e
[
8 o
[a M
O 541-560 O
Ca o
C w
l3 l
l
j s_Dy(n vOI rm mZO i q-Z3$-
[ _ S :j
- 5. c r.
)5 8
wVI
=
la wO!
3 to
]Q t
(
Q s
tn
._ V l M
3 a
yQ l
G Q
p Q Q %
fo O1 e
\\
Q h % %
r e
m 1
M h h Q Q
\\ % h b
1 u
A y*
h h $ $
u U1
\\
k $
\\
]%
u u
u y
'N
% h M $ ($
j N ]\\ M($
$ h
\\\\
\\
P h_
_
- P
\\
- h
\\
H S
O O
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
2 4
6 8
0 2
4 6
8 0
2 4
6 8
0 2
4 6
8 0
2 4
6 8
0 6
3 3
4 4
4 4
4 5
5 3_
2 2
2 2
2 3
3 1
1 1
1 1
1 2
4 6
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 i
1 1
1 8
0 2
4 6
8 0
2 4
6 8
0 2
4 6
8 0
2 4
6 8
4 2
2 2
2 2
3 3
3 3
3 4
4 4
4 4
5 1
1 1
1 1
OOg~OOoO
%@,g O 0 O3 0
c3 a' gC3 f x$ 2*
ooJ $~*9 C eM 3o 0agC o0_J "W0W*
0 c
_ _ =. _ - _ _. _
s 1986 1'
30 LEGEND f4 V[/M Number of plants within range
^m m
25-g i
11 l
f 20-C 17 15-i 2
i a
~
i
%o to-5 i
.o i
E i
5-4 3
Z i
2 2
2 2
p C @'
I I
e===ea e
=
=
=
= =1.'
1 1
1 1
,1 T
T T
i
-==
s g
g i
Number of component failure reports Figure 7 Reported component failures ht occured in 1986.
i i
i
In terms of the timeliness of reporting, in the third quarter of CY 1986, j
3730 componant failures were reported that occurred in IPP6.
This number is 62% of the total entries for the quarter.
By conparison, in the third aparter cf CY 1985, 24:4 (52",) raports described failures that occurred in 1985.
Conclusion The timeliness of NPRDS failure reporting has irproved cmpared to the same paried in 1985.
In the third quarter of 1986, 62% of raports submittad described failures that occurred in 1986.
Thus, only about 38% of the reports submitted in this cuarter reflect component failures that occurred more than six months ago.
OUALITY OF THE NPRDS FAILURE REPORT NARRATIVE To assess the cuality 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 taxt 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 ware graded as desired, adequate, probably adequate, and inadequate.
The results of this analysis are presented in Table 5.
The " adequate" category neans that a knowledgeable person could understand the type of failure that occurred.
The " desired" category means the report co-taired good root cause in#crration.
Few reports net this test.
The majority of the failure reports examined continued to be at least "probably adaouate."
The results of scoring the NPRDS narrative quality on a system in which 4=
desired. 3= adequate, 2= probably adequate, and 1= inadecuate are shown in Figure 8.
USECULNESS AND QUALITY OF NPRDS DATA FOR USERS One of the main sources of data for NRC's analysis and evaluation of events and the operational experience of plants is the NPRDS data base. Through our case studies, engineering evaluations, trends and patterns analysis studies, ard follow-up visits to plants we have identified data problems and issues which cculd improve the usefulness and quality of NPRDS data for users. These itens have been providad to INP0 for their consideration.
Fren Case Studies:
As part of a recent case study nn motor operated valves, the 1984 and 1985 failures in the NPPDS data base which involved motor operators were reviewed to evaluate the NPRDS fields: "cause of failure," " failure description," and
" corrective action."
The primary problem identified through this review was that narrative descrip-tions were often too brief and were frequently inadequate or inconsistent concerning how a valve failed and the stated failure cause.
Further, in most
,n
,,,--g,
..., e.,.
instances, the narrative did not include additional discussion to provide information about operating conditions or circumstances of the failure to operate which would suggest the root cause of inoperability.
Ir examining the "cause of failure", the staff felt that rary of the reported caures actually represerted symptoms rather than the underlying or " root" cause of valve assembly inoperability.
In the study we estinatad that nearly 80% of the NPRDS cause data may be indicative of synotems rather than root causes of failures. Nor did the reported information certain sufficient detail about the aperating corditions to permit an in-depth review or independent evaluation to determine the true root cause of failure. The identification of potential sa'ety issues pertaining to either individual events or detection of possible trends or patterns that warrant in-depth evaluation may be a very difficult task because tha available information does not adequately describe operating cenditions which would be needed to evaluate the true root cause of the reported failure.
Another case study of rod cluster control assemblies fcund deficiencies in the plant procedures used to identify NPPDS reportable events; this resulted in events not being reported to NPRDS. At one plant the personnel preparing the LERs noted whether an event was reportable to NPRDS on the LER but the plant procedure for selecting NPRDS reportable events did not include the review of LEPs to identify possible NPRDS reportable situations. At another plant the plant staff prepared the LER and noted whether it was reportable to NPRDS; a ruclear operations group at the corporate location, however, made the final cetermination of whether an event was reportable to NPROS but there was no fe=dback cornunication to the plant personnel who prepared the LER.
Fron Trends and Patterns Analysis Studies:
In 1986 NRC began its NPRDS Trende and Patterns Analysis Program which uses a variety of statistical methods to analyze NPRDS data to determine the factors (i.e., NPRDS fields) which may be statistically related to the failure of the component. An engineering evaluation is conducted to follow-up on the issues identified through the statistical analysis. Since the statistical analysis of this program is based exclusively on NPRDS data, correct and complete information on failures is of utmost importance.
The first system selected for analysis was the main feedwater system; within this systen we have completed the statistical analysis of the main feedwater regulating valves, main feedwater bypass valves and main feedwater turbine driven pumps. During the analysis of these selected components, a significant amount of the data from the engineering records was found to be missing or questionable. Because of the importance of this information to these studies, the NPRDS coordinators for several units were contacted by telephone to verify the data.
Although this data verification effort concentrated primarily on those fields that had been selected for these particular studies, it can be assured that other fields also contain discrepant data. The analysts have prepared an extensive report documenting the results of this NPRDS engineering record i
verification. A sammary of the problems and issues is as follows-The following fields appear to be inadequately defined (there are l
inconsistencies in the reported data and redundant and ambiguous use i
TABLE 5. QUALITY or NPaos FAttunE REPORTS Third Quarter Fourth Quarter Ftrst Quarter Second Quarter Third Quarter CY 1985 CY 1985 CY 1986 CY 1986 CY 1986 Desired it' It' 6%'
11%'
6%l Adequate 311 82%
- 301, 891 351 86%
381 92%
331,
941 Probably 501 581l 45%
431 551 I
Adeqr,e e Icadequate 181 111 14%
81 6%
TItal Number of 80 86 (Wa 878 878 NPRD-4 Forms Evaluated One plant had no f ailure reports during the first and second quarters 1986.
c.
i
2
.O NPRDS Narrative Quality
[Most Recent Yr.]
u-u..
M-8
= at a
]
e8 "f i
l "I
i "I
.. I 4.
na e-es 4-es, 1-ee E-ee a-se Event Quarter Af:;r:t S.O. %W Adequete-a.O Figu a 8 i
i i
- - -, -. - _ ~.
.-n.--..
n.
y
_g_
of NPRDS codes): internal environment, external environment, valve function / application, valve maximum operating pressure (in some instances systen operating conditions were reported while in others it is the component's inherent design pressure), valve cperating pressure (sane comment as with pressure), valve inlet size (in seme instances line size is reported), and pump rotational speed.
Confifeting NPRDS engineering records; (e.g., valve records that report a specific type of valve operator; however, the associated valve operator record indicates something different. This may be due in part to all associated records not being revised subsecuent tocomponentreplacement.)
3 Component types reported not always in acreement with the component l
manufacturer's model number.
' ~
Component records nissing from NPRDS.
It is more likely the records have been ceded with an incorrect " application code."
l Not all reauired fields are completed.
It appears nirimal effort has been expended tn update NPRDS records.
l From their calls it appears that NPRDS' inportance is not recognized; some plants contacted were surprised that the data are beino used for detailed studies such as the trends and patterns analyses. Almost all plant centacts ir.c4cated a concern for discrepant or etissing data in NFPDS and that the cc~ect data would be input either innediately or in the aear future.
Fec i Plant Visits:
Curing the course of our studies NRC staff often visit particular plants to
'ur-her our investigation, to discuss concerns and issues that our analysis has
~
identified and, coincidentally, to verify the data used in our studies.
In November 1986 nembers of our staff visited a plent to discuss the large number of failures of the feedwater regulating valves and bypass valves.
They found that the NPRDS information on these comporents for " operator type, operator subtype and valve material" were in most cases incorrect.
Incorrect information for these fields was of particular concern as the reported information for these fields had been identified as fairly unique for this particular plant and possibly in some way indicative of potential problems with these valves.
For this same study, personnel from P,egion II reviewed the NPRDS informa-tion on these valves on one of their recent visits. They found that in 1976 the units had replaced the Copes Vulcan valve actuators, which had been frequently falling off, with Fisher Control piston actuatnrs with 4
hydraulic damping and spring closure.
This information had not been updated in NPRDS. The valve type was also incorrectly reported.
Although the above discussion would seem to indicate that users of NPRDS may be facing significant problems, NRC feels that this feedback is in part a positive sign for NPRDS. As the system has improved and been made more accessible to t
i
.s ' ;e g.
users, there has been increased usage which in turn has provided more feedback cr the needs and concerrs of users of NPPDS. AE00 will continue to direct its I
evaluations to users' needs and to provide INP0 with this useful feedback.
i
- l l
PRC3 RAMS AT INPO TO IMPROVE COMPLETENESS AND OUALITY OF NPRDS DATA BASE I
In 1987, INPO will be pursuing a number of prograns to inprove the quality and l
ccrpleteness of the NPPDS data base.
To verify failure reporting INPO will continue to review maintenance records for its plant evaluations.and to have all incoming failure reports reviewed by experienced engineers. To monitor the accuracy of the engineering 'eports, INPO is planning on reviewing, during plant evaluations, records of comoonents that have been taken out service.
I INPO's IBM Conversion Project, which will begin implementation in 1987, will provide NPRDS users major improvements such as extensive interactive help, new calculated data, data analysis tools, enhanced _ report capabilities, and. easier -
use. The conversion project is a maior INPO initiative with a enal of stimula-j ting NPRDS participation through increased usage by utilities.
b FUTURE EVALUATIONS OF THE NPRDS DATA BASE User's comments are becoming an increasingly important part of the NPRDS evaluation / improvement process. We have been and will enntinue to provide i
INPO with this user feedback. We also anticipate an expanded crocerative effort betwaen NRC and INP0 to improve NPRDS and meet user's needs.
Future evaluations by the staff will be directed at determining failure reporting ccrpleteness and NPRDS appropriateress for the various end uses, including regulatory activities, and stimulating needed improvement.
.a l
In forthcoming evaluations we may continue to compare the maintenance records j
collected by the PAT teams with the NPRDS reported failures; we may also select a larger sample of key components and plants to enable ~us to make an estimate of the overall level of NPRDS completeness.
i i
We have also discussed with INPO the possibility of the staff working closely j
with INPO to independently review the data collected by INP0 to assess the i
quality and completeness of the NPRDS data base.
For example, INPO reviews a significant number of maintenance records during each plant evaluation. The 4
staff is discussing with INPO possible approaches to also review this data and 4
the INPO analysis of it as a substitute for the current analysis based u;>on -
j failures reported in LERs.
i 3
Should cooperative activities using INP0 collected data not prove practical, the staff will explore alternative methods for determining the overall com-j pleteness of NPRDS failure reporting. Applicable methods to be explored j
include:
I 1.
Conparison of NPRDS engineering and failure records with component pcpulations and failures reported in the In-Plant Reliability Data Systen (IPRDS)
From 1975 throuch 1984 ' data base personnel visited seven plants to copy maintenance work orders describing corrective maintenance at units on pumps, valves and electrical components to develop a comprehensive and component j
,,,mr--,--,m.
--.---.+--,,-,ee,.n-m,.,..
,e------...,---,y,ng.,,nn-,y--w,.--------n.
,,n--n-,--,-,-~w--.pn----,~y.-.ww,
- n _ g.,
wn.~,,.nv..._.v-m,
.t a p
. In _
specific data base for PRA and other comporant reliability-related statistical ar.alysis. They also gathered plant equipmert lists and plant drawings and in sc-e cases interviewed plant personnel for inforration on component popelations anc duty cyclas.
The data from these components were encoded inte a computer-ized data base.
In 1985 work was completed on expanding the data base to 4
iaclude two additional nuclear units and plant specific data for selected corponents from three other plants. AE0D is investigating the possibility of using the 1984 and 1985 data to neasure the completeress of NPRDS reporting frnn these plants durino that time period.
IPRDS could possibly be used to ccupare NPRDS information on failure date, failure mode, faHure severity, cause and description.
2.
Evaluation of Engireering Records Completeness and Quality by Computer Counts and Cross Tabulations For this possible new evaluation task, one. system for each NSSS vender would.
i be selected at random.
(Over a period of time all systers would eventually be evaluated.) Counts of engineering records by plant for each kay component within the selected systems. could be tabulated.
Based on each plant's number of loops, possible deficiencies in the completeness for NPRDS engineering records could readily be determined.
In addition for each of the key components in the system, tabulations for each engineering field could be generated. These tabulations would highlight missing data and data incorrectly coded.
Also, some indications of templeteness of failure reporting ray be assessed by cenparisons of the nurber of failures for the selected ccFponents anono the l
pla nts.
Plants with significantly fewer reported failures than the average could de identified.
3.
Comparison of Quality of NPRDS Engireering Records Data with Information in the Equiprent Qualifications Data Base Presently the Equipment Qualifications (EQ) data base contains data from the eouipment qualifications reports on electrical ccmponents through 1985.
For newer plants there are hard copy equipment qualifications reports for pumps and valves which might be coded for the data base.
Information from the EQ data base and NPRDS records would be matched on plant identification number, utility identification number, and system. NPRDS engineering records information could then be compared for such fields as in-service date, component code, manufacturer, model number, terial number, and Juch operating characteristics as temperature, pressure, and humidity. The question of whether the reported maxinum operating temperature and pressure is " actual or design" could also be ascertained.
OVERALL CONCLUSION We are pleased to report that our evaluation of the NPRDS data base shows substantial improvements in the number of plants participating, in tha number of reports being submitted by the plants, in the timeliness of reporting and in the completeness of reporting NPRDS failures which are documented in LERs.
As noted in our last report, we feel that these levels of participation in NPRDS warrant incressed use of the system as a source of reliability data to meet agency needs. We compliment INPO in its success in developing the NPRDS i
W e,e Appendix A NPRDS HISTORY h
The present NPRDS 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 NPRDS has been managed and operated by the Institute of Fuclear Power Operatiens (INP0).
BACXGROUND Former President Carter's 1977 National Energy Plan recomended that the NRC
~
make mandatory the voluntiry reporting of minor mishaps and component failures (i.e.,NPRDS). The plan suggested that mandatory participation wodd enable l
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.
I I
Coincident with the NRC's activities directed toward implementing the President's recomendation, the General Accounting Office (GAO) reviewed the NRC's data-gathering activities concerning unscheduled events at coamercial nuclear facilities.
In a report issued in late January 1979, the GAO concluded that it was unlikely trat the NRC could justify mandatory NFRDS participation when factors such as additional industry cests, limited expected sa*?ty 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 Comission briefing on the collecticn and analysis.
cf operational safety data, the Comission concurred with the January 1979 GA0 recomendation that rulemaking be used to decide the question of whether to nake NPRDS reporting mandatory. Accordingly, an Advance Notice of Prcposed
{
Rulemaking (ANPRM) was approved by the Comission, and published in the Federal Register on January 30, 1980 (45 FR 6793). Numerous public coment letters were received in response to the ANPRM. The predominant message in the con ents was overwhelming opposition to make 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 reouirements for such cata 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 system. This Integrated Operational Experience Reporting. System (10ERS) would have combined and restructured the NRC LER system and the voluntary NPRDS. The 10ERS concept included two principal features:
(1) the collection of detailed technical description of significant events, and (2) the collection of' component I
reliability data.
'l q
.s: *.y.
I o,
' 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 reeded j
reliability data without assuming direct responsibility for its collection.
On June 8, 1981, the INP0 Board of Directors decided that because of its role es an active user of NPRDS data, INP0 would assume responsibility for monace-ment 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 recomended and the Comission approved (SECY-81-494) proceeding to modify and codify the existing LER 1
reporting requirements as a separate rulemaking while holding the IOERS rule-making in abeyance.
In approving SECY-81494, the Commission directed that the staff clcsely monitor the status and rate of improvement of the.NPRDS and provide the Com-mission with semiannual status reports on the effectiveness of INPO managen,ent
~
of the NPRDS.
In crder to monitor the completeness and quality of NPRDS data, AEOD 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 previde a baseline of information concerning the completeness and ouality of the NPRDS failure reporting anc engineerine data files as they existed prior to INP0's assumptien of responsibility for the management of NPRDS.
Subsequent analyses of the completeness and qualf ty of the files are being conducted reriodically in order to prcvide a measure of the improvements in the NPFDS
]
under INP0 management and technical direction.
i The scope and results of this evaluatien 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), Auoust 20, 1985 (SECY-85-56A), January 30, IS86 (SECY-86-35) and July 24,1986(SECY-86-216).
With the implementation of the LER rule in 1984, the NPRDS became the primary source of component and train level failure data. The rule essentially elimi-nated the reporting of component failures or failures of a single train of a multi-train safety system. Only a complete system failure (all trains) is reportable. Therefore, the primary national source of failure data has become the NPRDS.
Since the NRC and the industry are so dependent on this source of data, the i
need continues for a formal and continuing program to provide confidence that NPRDS data is of good quality (accurate and ccmplete engineering and failure records), and is suitable.for the various NRC component and train failure data needs.
l k
f l
.-.