ML20248E200
| ML20248E200 | |
| Person / Time | |
|---|---|
| Issue date: | 09/27/1989 |
| From: | Israel S NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| Shared Package | |
| ML20248E198 | List: |
| References | |
| TASK-AE, TASK-T914 AEOD-T914, NUDOCS 8910050099 | |
| Download: ML20248E200 (4) | |
Text
-
l.' -
AEOD TECHNICAL REVIEW REPORT UNIT: .
N/A -
DOCKET NO.: N/A TR REPORT NO.: AE0D/T914 LICENSEE: N/A DATE: September 27, 1989 NSSS/AE: N/A EVALUATOR / CONTACT: S. Israel
SUBJECT:
OVERVIEW 0F DESIGN / INSTALLATION / FABRICATION ERROR
SUMMARY
A review of about 300 LERs from 1988, which were determined to be important in accordance with established AE00 screening processes, indicates that about half (D/1/F) of the reports are attributable to design / installation / fabrication errors.
be related to previous generic communications.About 60 percent of the LER qualification reports. deficiencies each occurred in about 10 percent of the D/1/FSei deficiencies. About 10 percent of these reports concerned emergency power Focused activity by the licensees appears to be the dominant method for identifying these flaws.. Typical discovery modes include review of plant modifications and repair af failed components.
plantdeficiencies the operation. were discovered because of some aberration during normalOnly provide ongoing initiatives needed to capture these types of flaws. E DISCUSSION A survey of important LERs from 1986 and 1987 (Refs. I and 2) indicated that about 50 percent of the reports were attributable to D/1/F flaws. These important LERs are determined by established AE0D screening procedures to identify LERs especially important to safety, and are labeled category 2 LERs.
These D/I/F deficiencies generally go undetected for considerable lengths of time because the problems that escape initial reviews and tests and are usually not identifiable by routine surveillance tests. The latent aspect of these flaws poses a challenge to the licensees to discover these deficiencies before they contribute to a serious event at a plant.
A similar survey was made of category 2 LERs from 1988 and the results indicate that about 50 percent are attributable to D/1/F errors also. Although the results may be biased because of the classification process, it is significant that latent flaws are still being discovered in relatively mature plants in a relatively mature industry having extensive operational experience feedback programs.
may illuminate This the group of LERs was reviewed to identify salient features that situation.
An evaluation was made of the pertinent issue associated with each LER and a generic communication data base was searched to identify documents that may have been issued previously on the issue based on the title of the document.
The letters. generic documents included information notices, bulletins, and generic earlier generic About 60 percent of the selected LERs appear to be associated with communications.
'8910050099 890927 PDR ORG NEXD PNV
_4 l l
l The topics of the generic communications were developed from a much larger survey and cover about 80 problem areas. Only about 50 topics are related to l the.LERs in this study. Seismic design problems are the most frequently This was the subject of an unresolved safety issue that w
. generic letters issued in 1987.
A major requirement of these generic. letters l is a systematic plant walkdown by knowledgeable personnel to identify ;
potential seismic problems in significant systems. i
(
Based on the discussion in the LER itself, it appears only about one-third of k the seismic flaws were discovered by plant walkdowns.
About one-third were discovered or system. during a review of a plant modification of the degraded component plants or from other systems at the same plant.Another large fraction of participation in the resolution of the seismic issue, this heightenedBecause of e explicitly stated in the LER. awareness may be the. underlying factor in all these disc Inadequate of equipment the D/I/F, category 2 LERs. qualification is the principal topic for about 10 percent This area has been the subject of much discussion and generic communications over the last 10 years. Many of the reports are quite narrow, addressing a specific component and its associated problems.
relays, and These sensors.are mainly environmental problems with splices, wires, Identification of these deficiencies may be the result of a review of specific plant modifications or a sensitivity to certain problem areas, such as splices, during routine activities.
Overloaded diesel was the subject of four LERs and control room ventilation problems only one oroccurredtwo LERs.in three other reports. The rest of the topics occurred in simple approach to capturing similar flaws in the future.This About 10 percent of diverse group of the D/1/F LERs each dealt with emergency power problems or fire issues. Both of theseactivities.
making areas have received considerabic attention because of previous rule This suggests that some level of ongoing industry discussion specific areas. / activity in an area may sharpen licensee interest and awareness in About 60 percent of the operating plants had category 2 LERs concerned with the D/1/F issue. The actual impact may be higher because of dual plant sites where only one LER is issued if there is a common problem. Most of the plants had only one, two, or three LERs; however, one plant had 9 LERs, one had seven, and three had six.
Thus, these types of discrepancies are fairly common for most plants and a few have a large number of reports. The number cf plants with multiple reports may depend on the licensee's ongoing activities related to design reconstitution or similar programmatic initiatives.
The manner in which these design / installation / fabrication errors were discovered was examined. Six categories considered are:
Task Related Involved specific action related to the component cr system such as a review of a modification, but not concerned with a component failure or plant occurrence.
(w
(, ;.- -
4 Routine Activity Captured by a non-specific plant activity.
Operational Occurrence Discovered during normal plant operation as a component failure or unexpected plant response.
1 Test Discovered during a surveillance test.
! Related Activity Discovered at another plant or another system at the same plant.
Unknown Indeterminate from the LER.
Task related activity was the most frequent means for discovering these problems and'it accounted fer about 40 percent of the reports. Most of the reports in this category were initiated because of a review of a plant modification. ;
change satisfies all the design criteria.This' activity requires making a separate ~
This effort is similar to the original design justification and results in examining much of the original design package 'to identify and understand all the criteria. This review process ends up highlighting deficiencies in the original design if there were any.
Existing NRC requirements such as'10CFR50.59 reports justifying any plant changes and quality assurance criteria in 10CFR50, Appendix B, which covers maintenance and modification of equipment, already direct the licensee to perform these types of detailed analyses and inspections for any plant modification.
Routine about 16activities percent ofandthetests each was responsible for identifying flaws in events.
Operational occurrences (about 13%) and tests could be considered task related activities of a slightly different form.
Thus, combining task related, operational occurrences, and test categories, licensee focused attention accounted for about 70 percent of the reports.
Purposeful
-D/1/F events. activity appears to be the most effective means for capturing these 10CFR50, Appendix RootB. cause analyses of component failures is required by Surveillance tests are required by the technical specifications and post-maintenance testing is required by Generic Letter 83-28, " Required Actions Based on Generic Implications of Salem ATWS Events."
A review of the LERs indicates that 14 of the discoveries were prompted by some NRC activity. Nine of the LERs were the direct result of feedback from an information notice although no specific action is required for these generic communications.
safety system functional inspection, and bulletins.Other prompting came from the r The licensee's feedback program, required by the technical specifications, was also instrumental in plant. information from other facilities to identify design flaws at their using Thus, latent flaws have been identified by many diverse ways.
CONCLUSIONS Design / installation / fabrication errors continue to be a major cause for important events reported each year that affect a wide array of plants and systems.
addressed About in generic60 percent of the reports appear to cover topics previously communications.
A breakdown of the manner of discovery
3-
.. g . i f
indicates that about 40 percent were the result of a directly related task such as a review of a plant modification involving the flawed component / system.
Discovery during general'or routine plant activities or tests accounted for about 16 percent each. About 13 percent were. discovered as a result of a component plant failure response while during the plant was operating or because of an unexpected operation.
. The diverse ways that led to the discovery I of these~ flaws does not point to a simple programmatic approach to improve the effectiveness of capturing similar latent problems.
Seismic design errors and inadequate equipment qualification each accounted for about 10 percent of the LERs. Flaws in the emergency power system were discovered in approximately 10 percent of the reports. The seismic and emergency power. issues probably reflect heightened interest in these topics in the recent past.
This study does not suggest any additional action at this time. Existing licensee and NRC review and inspection processes continue to be effective in uncovering these types of issues as noted by the small fraction (13%)
discovered because of some aberration during plant operation.
REFERENCES 1.
" Overview of Significant Events in 1986," Memo from S. Israel (NRC) to P. Lam (NRC), January 28, 1988.
2.
S. Israel, Insights from Significant Events in 1987, NRC report AE00/T805, May 5, 1988.
-__.______m.___________-__ _ _