ML20237K451
| ML20237K451 | |
| Person / Time | |
|---|---|
| Issue date: | 07/31/1987 |
| From: | Dennig R, Oreilly P NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| References | |
| NUREG-1275, NUREG-1275-V01, NUREG-1275-V1, NUDOCS 8709040358 | |
| Download: ML20237K451 (285) | |
Text
,
e =
NUREG-1275 Operating Experience Feedback l
Report - New Plants l
Commercial Power Reactors i
i U.S. Nuclea . Regulatory -
Commission l Office for Analysis and Evaluation of Operational Data R. L. Dennig, P. D. O'Reilly ra asc f oq,*
f.. .. ,\ >
B .- .j
\; ..s ... /
q ,, g ry g t,0 3'/J, U '
pgp NUtd G pop l- 127S h 1
l l
NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in N RC publications will' be available from one of the following sources:
- 1. ' The NRC Public Document Room,1717 H Street, N.W.
Washington, DC 20555
- 2. The Superintendent of Documents, U.S. Government Printing Office, Post Office Box 37082, Washington, DC 20013 7082
! 3. The National Technical Information Service, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited in NRC publications, it is not intended to be exhaustive.
Referenced documents available for inspection and copying for a fee from the NRC Public Docu- l ment Room include NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement bulletins, circulars, inforrr ition notices, inspection and investigation notices; Licensee Event Reports; vendor reports anC .; correspondence; Commission papers; and applicant and licensee documents and correspondence.
l l The following documents in the NUREG series are available for purchase from the GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceedings, and l
NRC booklets and brochures. Also available are Regulatory Guides, NRC regulations in the Code of Federal Regulations, and Nuclear Regulatory Commission Issuances. :l Documents available from the Nationa! Technical Information Service include NUREG series i reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.
I Documents available from public and rpecial technical libraries include all open literature items, such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.
Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the publication cited.
Single copies of NRC draf t reports are available free, to the extent of supply, upon written requee,t to the Division of Information Support Services, Distribution Section, U.S. Nuclear Regulatory Commission, Washington, DC 20555.
Copies of industry codes and standards used in a substantive manner in the NRC regulatory process -
are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually copyrighted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards institute,1430 Broadway, New York, NY 10018.
l l-
[_-___--__-__-______-___-_-________-________-__-_______--__ _ _ _ - _ - _ _ _ _ _ _ _ _ _ _ - _ _ _ __ _. _ _ - _ _ - _ _
NUREG-1275 Operating Experience Feedback Report - New Plants 1
Commercial Power Reactors Manuscript Completed: July 1987 Date Published: July 1987 R. L. Dennig, P. D. O'Reilly Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulatory Commission Washington, DC 20555 I
ABSTRACT This report documents a detailed review of the cause of unplanned events ouring the early months of licensed operation for plants licensed between March 1983 and April 1986. The major lessons and corrective actions that appear to have the greatest potential for improving the effectiveness of plant startups are provided for consideration through the operating experience feedback programs and activities of the industry and the NRC staff.
l l
l 1
l iii
i TABLE OF CONTENTS Page ABSTRACT ..................................................... 111 AC KNOWL E DGME NT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i x EX E C UT I V E S UMMA R Y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x i i
- 1. INTRODUCTION ............................................. 1 l
- 2. BACKGROUND ............................................... 4 2.1 Overview of Sta rtup Tes t Programs . . . . . . . . . . . . . . . . . . . 6 2.2 Sunnary of Relevant Performance Indicator Data . . . . . . 7
- 3. REVIEW OF EXPERIENCE ..................................... 8 3.1 Reacto r Scram Ex pe ri ence . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.1.1 Data Analysis ............................ 10 3.1.2 Cause Analysis ........................... 18 3.1.2.1 Equipment-Related Causes ....... 19 3.1.2.2 Personnel Errors and Procedural Deficiencies . . . . . . . . 25 3.1.3 Findings ................................. 28 3.1.4 Lessons Learned .......................... 29 ,
3.1.5 Perspective From Japanese Experience ..... 31 i 3.2 Engineered Safety Feature Actuation Experience ..... 32 3.2.1 Data Analysis ............................ 33 j 3.2.2 Cause Analysis ........................... 34 !
l.
3.2.2.1 Equipment-Related Causes ....... 35 3.2.2.2 Personnel Errors and Procedural Deficiencies ........ 39 3.2.3 Findings ................................. 39 l 3.2.4 Lessons Learned .......................... 40 l 3.3 Tecnnical Specification Violation Experience ....... 41 3.3.1 Data Analysis ............................ 41 1 3.3.2 Cause Analysis ........................... 44 3.3.3 Findings ................................. 46 3.3.4 Lessons Learned .......................... 47 v
r l
TABLE OF CONTENTS Page 3.4 Unplanned Losses of System Safety Function ........ 48 3.4.1 Data Analysis ............................ 49 3.4.2 Cause Analysis ........................... 50 3.4.3 Findings ................................. 51
- 4. EVENT SIGNIFICANCE AND INSIGHTS ......................... 52 4.1 Scrams With Complications .......................... 52 4.2 Results of Systematic Screening for Significance ... 53 4.2.1 Topics For Operating Reactor Events Meetings ................................. 53 l 4.2.2 AEOD Event Categorization ................ 54 i 4.3 Causes of Significant Events ....................... 56 4.3.1 Accident Sequence Precursor (ASP)
Insights ................................. 56 4.3.2 Augmented Inspection Team (AIT)
Program /Special Inspections .............. 58 4.3.2.1 Catawba 2 AIT .................. 58 4.3.2.2 Hope Creek AIT ................. 59 4.3.2.3 Palo Verde Special Inspection .. 61 4.4 Findings ........................................... 62 4.5 Lessons Learned .................................... 63
- 5. STAFF AND INDUSTRY LESSONS FOR IMPPOVING STAPTUP EFFECTIVENESS ........................................... 63 5.1 Improvement Lessons for Consideration by Licensees . 64 5.1.1 Management lessons ....................... 64 5.1.2 Equipment lessons ........................ 66 5.2 Improvement lessons for Consideration by the NRC Staff .............................................. 67 i
5.2.1 Management Lessons ....................... 67 5.2.2 Equipment Lessons ........................ 68 1
- 6. INDUSTRY IMPROVEMENT LESSONS ............................ 68 6.1 Scram Reductior, Ini ti a ti ves . . . . . . . . . . . . . . . . . . . . . . . . 69
- 7. CONCLUSIONS ............................................. 70 vi
TABLE OF CONTENTS Page APPENDIX A - PERFORMANCE INDICATOR DATA .................... A-1 APPENDIX D - CAUSE CATEGORY DEFINITIONS .................... B-1 APPENDIX C - SCRAM DATA FOR NEW PLANTS ..................... C-1 APPENDIX D - FORMAL STATISTICAL ANALYSIS ................... D-1
- 1. Background ............................... D-2 j
- 2. Overview of Statistical Procedures ....... D-2 {
- 3. Results .................................. D-3 1
- 4. Detailed Output .......................... D-4 i
APPENDIX E - ESF ACTUATION EXPERIENCE ...................... E-1 1
ES F Ac tua tion Da ta . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-2 Detailed Experience with Radiation Monitoring Systems at Selected Plants ......... E-25
- 1. Byron 1 .................................. E-26
- 2. Callaway ................................. E-27 l
- 3. Wol f Cree k . . . . . . . . . . . . . . . .............. E-28 l 4 Palo Verde 1 and 2 ....................... E-28
- 5. WNP-2 .................................... E-28
- 6. Susquehanna 2 ............................ E-29
- 7. Hope Creek ............................... E-29 APPENDIX F - TECHNICAL SPECIFICATION VIOLATION DATA .. . .... . F-1 APPENDIX G - TECHNICAL SPECIFICATION VIOLATIONS BY SYSTEM .. G-1 APPENDIX H - LOSS OF SYSTEM SAFETY FUNCTION DATA ........... H-1 APPENDIX I - MATURE PLANT DATA FOR NRR AND AEOD SIGNIFICANT EVENTS ........................... 1-1 l
l l
Vii
TABLE OF CONTENTS List of Figures
.P, age Figure 1 - Number of Plants in the First Two Years of Operation ..................................... 1 Figure 2 - Milestone Dates for Fach of the 22 Newly Licensed Plants Considered in This Study ...... 6 Figure 3 - New Plants - Total Unplanned Scrams From Critical.(From NRC Performance Indicator Program)...................................... 7 Figure 4 - Unplanned Reactor Scrams: Washington Nuclear 2 .. 9 Figure 5 - Unplanned Reactor Scrams: Callaway .............. 9 Figure 6 - Composite Unplanned Scrams: Newly Licensed Plants Grouped by NSSS Vendor ................. 11 Figure 7 - Unplanned Reactor Scrams: Percentage Distribution by Activity, System, and Cause ... 15 Figure 8 - Normalized Unplanned Scram Rate: Comparison by Activity, System, and Cause ................... 16 Figure 9 - Average Unplannea Scram Rate for Newly Licensed Plants as Function of Power Level Distributed According to Plant Experience ..... 17 Figure 10 - Unplanned ESF Actuations ........................ 33 Figure 11 - Technical Specification Violations .............. 42 Figure 12 - Loss of System Safety Function Events ........... 49 List of Tables Table 1 - NSSS Vendor, NSSS Design Type, Turbine-Generator Supplier, Architect-Engineer, and Milestone Dates for the 22 Newly Licensed Plants ........ 5 Table 2 - Ave rage Unpla nned Scram Ra tes . . . . . . . . . . . . . . . . . . . 12 Table 3 - Average ESF Actuation Rates ..................... 34 Table 4 - Average Technical Specification Violation Rates . 43 Table 5 - Average Rates for Loss of System Safety Function ...................................... 50 Table 6 - Distribution of Events Occurring at the 22 Newly Licensed Plants That Were Discussed at NRR Reactor Events Meetings During 1983, 1984, 1985, and the First Hal f of 1986 . . . . . . . . 54 Table 7 - Distribution of AE0D Category 1 and 2 Events for the 22 Newly Licensed Plants for 1983, 1984, 1985, and the First Half of 1986 ........ 55 Table 8 - 1969-1981 and 1985 Precursors: Frequency of Errors / Failures (PerPrecursor) Principal Systems and Frequency of Failures (Per i Precursor) for Systems Involved ............... 57 viii
1 1
i ACKNOWLEDGMENT The authors express their appreciation to the members of the AE0D technical staff who contributed their time and efforts to the performance of this study, preparing for, conducting, and documenting the visits to plant sites, contributing material to the report as requested, or reviewing the various drafts of this document.
l The authors also acknowledge the technical assistance provided by the AE00 j contractor, EGt<G Idaho, Inc., for data base generation and analysis. ]
1 The assistance of AE0D's technical editor, Sheryl Massaro, is acknowledged.
l l
l l
ix
EXECUTIVE
SUMMARY
I. Background In response to the Commission's Policy and Planning Guidance, the Office for the Analysis and Evaluation of Operational Data (AE0D) conducted a study of the operational experience of commercial power reactors during their first two ,
years of operation. The goals of the study were to: 1) characterize the trends in operational events and their causes for this phase of operation,
- 2) identify correlations between plant attributes and performance, and
- 3) provide feedback to facilitate improvement.
An initial data compilation was published in August 1986 (AE0D/P604), which indicated that additional analysis of new plant trends would be performed.
This report provides that additional analysis, using three additional plants, expanding the time. frame, and analyzing the root causes behind unplanned event frequencies.
The study scope included the operational experience for 22 new plants from January 1983 through June 1986. It consisted of a comprehensive data analysis with plant-specific evaluations and field visits to selected facilities to f pursue potential corrective actions. The analysis examined the reportable events specified in 10 CFR 50.73, including the four major event classes:
reactor scrams, actuations of engineered safety features, technical specifica-tion violations, and loss of system safety function. Plant-specific event frequencies and event cause categories were classified, trended, and evaluated for possible correlations with attributes such as plant type, constructor, or licensee experience. The causes of events were classified by system and nature (e.g.,equipmentorhuman). These cause areas were then pursued through site visits with licensees to identify corrective actions that have proven effective as preventive measures 1/ for plants coming on line _to avoid high event frequencies.
This report has been subjected to the formal AE00 peer review process that involved the plants that were analyzed, industry organizations, and all . major offices of the NRC staff.
II. Results The study concluded that it is possible to achieve significant improvements in early plant performance. An assumption that new plants must experience high
, event frequencies during their first two years of operation need not be l accepted. Learning curves may be dramatically improved and early commercial operation would likely show a corresponding improvement. Indeed, individually, some power reactor licensees already have recognized the need for preventive programs and developed programs that, when implemented effectively, result in improved performance in _many respects.
The analyses also show that, without effective corrective action early in life, the root causes of a high event frequency will likely persist during early commercial operation. At this time of life, the relatively high challenge frequency coupled with the potential of undetected systems problems M NOTE: No requirements are imposed by this report.
xi ;
may present a significant challenge to a relatively inexperienced operating crew. Based on this study, it was determined that a root cause corrective action program, in response to events, is a necessary factor in achieving good performance.
Since October of 1986, both the NRC staff and the industry have taken ,
additional steps to focus on newly licensed plants. This study reinforces l the need for these efforts and ider,tifies improvement lessons that, based on I the operating experience, have proven effective in reducing the high event frequencies. Improvement lessons are identified in the areas of management and equipment, seven lessons in each area. Seven potential lessons are identified for consideration by the NRC staff.
Nuclear safety measures consist of prevention and mitigation items. At the outset, it must be noted that this study identified measures to reduce the frequency of reportable events. Therefore, the improvement lessons are all related to prevention. Mitigation measures, such as operator training for response to transients, continue to be important, but were not the focus of this review.
III. Methodology The basic methodology of the study was to analyze the reporced events (about 2400 events) to determine common root causes among the new plants and search for correlations to focus the pursuit of corrective actions. The root causes were trended across the plants to identify the systems and source of the event (e.g., feedwater control system with human error, procedural error, or '
equipment failure). The degree of correlation was evaluated between the event frequencies and the attributes of the plants (about 60 correlations). The attributes considered included date of operating license (0L) issuance, first or second site unit, architect-engineer (A/E), nuclear steam supply (NSSS) vendor, length of the startup period, and others. The cause categories for events at new plants were contrasted against the root causes for events at mature 1/ plants.
Analysis of the information in the licensee event reports (LERs) was successful in identifying the most prorninent systems / components that were sources of unplanned reportable events. The correlations provided some indication of sources of problems, but no strong correlations (i.e., that held across all types of events studied) were observed. Also, the LERs '
generally did not reveal the circumstances surrounding the failure or human error (such as lack of training or inadequate component testing or checks prior to systems tests). Therefore, to identify effective corrective actions, the most frequent root causes of events frcm the analysis were pursued with licensees. For example, LER analyses that indicated instrumen-tation and controls (18C) technician error caused challenges to safety systems were pursued to identify relevant training, or lack thereof, and the reasons behind any inadequate training, such as late systems turnover or lack of good training equipment, to identify a corrective measure. What resulted was a list of corrective actions that focus on the most frequent 1/ The data for mature plants (those licensed prior to January 1,1983) used in this study were obtained from licensee event reports (LERs) submitted for Calendar Year 1985.
l l xii l
l e____________-__________-____________________ _ _ _ _ _ _ _ _ _ _ _ _ _
causes of reportoble events at new plants. Although numerous corrective actions are possible, based on the analysis, these actions are .iudged to be among the highest payoff to substar.tially reduce high event frequencies at new l plants. ]
No specific correlation, such as NSSS vendor / design, was identified to be the l cause for high event frequencies. The specific improvement lessons provided by this report provide one tool to assist licensee management in assessing and prioritizing any improvement initiatives to reduce the frequcncy of reportable events, thereby improving safety and availability. The specific findings and lessons of the study follow. l IV. Analysis Findinos UNPLANNED REACI'OR SCRAMS is sw The analysis takes explicit $"=.,ng, account of the startup g g ,,,
a oan-n sa period ano the plant status in the data for each a E io plant. For example, the g
illustration on the right h a shows scram counts displayed t -
5 by months since OL issuance. t 5 a y l The months of initial 1*
criticality (IC), full power ! {E licensing (FP) and start of f 9 ""
ccmmercial operation (CO) are indicated, alcng with R ,
$g2 ll ;
the reactor critical hours
- _
j H i PylEEEE E[j , i for each month and the cause ot xvre co o a i4 ie is ao se a4 category for each scram, 4 e e
,iya g ,, , l A formal statistical analysis was perforned using average unplanned event rates for each of the four event types during both the startup program and the first six months after ccmpletion of the startup program.
The results showed no statistically significant correlation of these event rates with NSSS vendor, type of A/E (the utility itself or an outside firm),
the length of the startup program, date of OL issuance, or prior licensee nuclear plant experience. The analysis did indicate a significant correlation ,
between the average event rate during the startup program and the average y i event rate durino the first 180 days of consnercial operation for scrams, ESF "
1.
actuations and losses of system safety function. In the case of violations of ,i technical specifications (TSs), a statistically significant correlative was found between the event rate during early commercial operation and whether or not the plant was a second unit at a site (i.e., lower TS violation rates e prevailed for subsequent units at a site).
Specific findings from the data analysis and review of significant events were as follows:
A. Scrams (1) The scram rate for new plants exceeds that nf mature (those plants which were licensed before January 1, 1983) plants. Viewed as a class, new plants may be characterized by the following:
xiii
1 I
Pre-comercial operation total scram rates are about five times the mature plant rate on the average, and the frequency of scrams with complications (i.e., scrams which had additional fdilures or personnel errors beyond those which initiated the scram) is about a factor of four times that for a mature plant.
I. However, these scrams are from lower we levels.
T s
Post-commercial operation total scram rat - and complicated scram rates are about a factor of two to three higher than the mature plant rate (during the first few months). These scrams are from higher power levels, which is similar to a mature plant. A correlattan did exist between the startup and early post-commercial operation average scram rates. The scram rate ,
during startup was'a statistically significant indicator of the i scram rate that will be experienced during early post-commercial l opereLion. .
(
(2) Analysis for potential correlations of the pre- and post-commercial scram rates with various factors indicated no statistically signifi- )
cant correlation between the NSSS vendor or the type of A/E and the scram rate. Nor were there strong correlations between the scram rates and the length of the startup period, licensee prior experience, or the presence of an older unit at the site found in this analysis.
(3) On a percentage basis, the twses of scrams at new plants are very sin'ilar to those for mature plants. The primary causes are associ-ated with balance-of-plant (B0P) systems, with the feedwater system s dominating. There are some differences in the cause profiles between new plants and mature plants as follows: '
During pre-commercial operation, testing (primarily surveillance testing) contributes more to new plant scrams than in the mature plants by c factor of three. During this period, procedural deficiencies that caused scrams are higher by a factor of two than for mature plants.
B. ESF Actuations .
Previous analysis of operating experience has shown that engineered safety features (ESF) actuations rarely occur in response to an actual event which the specific features were designed to mitigate. Thus, ESF actuation reports do not represent a set of inherently safety-significant events, but rather, unnecessary actuations to which plant staff (operators, technicians, main-tenance personnel) must respond. The possible safety issues associated with unnecessary ESF actuations stem from: unwanted challenges to safety systems; the burden placed en a plant staff to respond to, troubleshoot, and report them; the pctential for an inadequate response to a real event; and possible erosion of a positive attitude toward safety due to the preponderance of false alarms.
Although ESFs hcve a wide diversity in design features, there are opportunities
, for new plants to benefit from ESF operating experience at other facilities in )
l) i.
their efforts to reduce the frequency of unnecessary ESF actuations. Examples
- (
xiv J _ ___
include experience with radiation monitoring, chlorine monitoring, and reactor )
water cleanup (RWCU) systems. {
i (1) New plants showed a wide range of unnecessary ESF actuation rates, with l some plants equal to or lower than the mature plant average. However, the majority of the new plants exhibited ESF actuation rates which averaged about feur and one-half times higher during startup than the average rate for mature plants. The plants with initially high rates i i
generally experienced large decreases beginning between three and eight months post-OL issuance.
(2) Statistical correlations of the pre- and post-commercial ESF actuation rates with various factors indicated that the higher the pre-commercial actuation rate, the higher the actuation rate during the first six months of the post-commercial period. There was a tendency toward lower ESF actuation rates during startup for second units at a site and for units operated by utilities with previous nuclear experience.
(3) As with the scram data, examination of the cause category (design, equipment, human error, procedures, other, and unknown) distribution for the E' actuati ns revealed no dominant category or correlation between a cause category, such as human error, and higher actuation rates. Of the ESF actuations reviewed, about one-third were attributed to human error or inadequate procedures. The majority of the human errors were i committed by 1&C technicians performing surveillance. l (4) High ESF actuation rates at newly licensed plants result from:
Unique or new design features that have not been tested previously (e.g., first of a kind, one of a kind, and state-of-the-art features).
l Actuation setpoints and logics which were too conservative as designed and not suited to the operational environment.
Errors in performing survei) lance testing, calibration, troubleshooting, and maintenance.
C. Technical Specification Violations Technical Specification (TS) violations have a high regulatory visibility and can be important to the NRC perception of licensee competence and safety. The study notes a clear need fer early finalization of the TSs to allow generation and validation of supporting plant procedures, and familiarization of the plant staff with procedural requirements. Discussions with licensee staffs indicate that difficulties with TSs can be traced to having TSs in a state of flux up to licensing (changes in the TSs have been proposed by some licensees right up to OL issuance), and lack of adherence to the TSs due to a lack of 1 training and familiarization (a licensee may wait until the last changes in i the TSs are implemented before the TSs are incorporated into plant procedures I and personnel training or the TSs is started). Analysis results are as i follows:
(1) Almost all TS violations occurring at new plants are the result of personnel errors or problems with procedures. This finding is consistent xv
. , a ,
- l. ,4 r s
g 'j j l
a f g ,a
? N '? j'
- . r g j
$ 4
- with reviews of operational exper ince for all licensed l ants'.
';* Although innan erMr is the d41nant cause of TVviolations, '
equipment problems can generate the c. opportunities.for hum n error that result in violt. dor:s. t -
(2)Over70differentsystems'havebeeninvolvsdintheTShidlations 7, , j occurring at new plants. \ /
'Y 1
j The systems involved in the most violations, the dre detection ar.d the radiation monitoring systems, together accounted for only about .
- - j 25% of the total number of violations, (3) The initial post-OL issuance vi6iition rates for new plants vary widely, but are generally higher than the riature plant average rate of 0.8 violations per n,anth. Significant moderation in violation rate usually peginsaboutsittoeightmonthspost-0Lissuanceinaboutone-halfof tia cases.
(4) Theresul'ts of a statistical analysis inoicated that the rate of TS violatidns was significantly hower during the first six months of commer-cial operation for plants that' are the second unit at a multi-t, nit site.
No statistically significant correlation was found between TS Miolation rate and NSSS vendor, or OL issuance date.
(5) A brief evaluation of the role of TSs indicated that, for the pew plants studied, variations in operating performance are probably not driven by regulatory changes in the scope of TSs, major differences in the quantity of; limiting conditions for operation (LCOs), or related surveil, lance requirements amorjg NSSS vendors.
, 4.
f ,$ cme differences in the plant-to-plant experience result from dif4 rences in designation of plant features as ESFs.
, e D. Loss 'of System Safety function !
(1) The loss of system safety function (LSSF) is not frequent., even at new plants .
In gcneral, plants go for long p 'fods of time with on events interspersed with brief 1-2 monta periods when a few events occur.
7 (2) TN statistical analysis af sLSSF events showed some correlation between the event rate during tht!startup period and the event rate during the e, first six months of commercial operation.
This result tends to reinf0fce the similar correlation found earlier for scram rate and ESF actuation rate.
u.
- . (3) No particular cause category dominates the LSSF events. Due to the low numbers, sporadic occurrence frequency, and variety of causes for LSSF
' events at new plants, no particular new issues or lessons 6 ave Men -
identified in this area.
l' e k f
3 xvi ,
4 >
l t c J
V. Event Significance Since the significance of the reportable events at new plants is dif ficult to measure, the study addressed significance from different points of view.
These included deterministic methods employing engineering judgment, such as event screening by AEOD, NRR, or detailed inspection findings (e.g., augmented ,
inspection teams (AITs)), and probabilistic methods, such as those developed through the NRC Accident Sequence Prscursor (ASP) screening techniques. A summary of this perspective of the 11gnificarce of these events follows.
A. The significance of scrams cccurring during pre-commercial operation may be reduced to some degree by their generally low initial power level, the lower fission prcduct inventory, the onsite presence of additional I personnel, and a sensitivity of the operating staff and management to unexpected problems with an attendant response readiness and capability.
However, the high scram frequency, with its associated percentage of scrams with complications (i.e., additional failures following the scram), still yields a rate of occurrence of scrams with complications that is about four times that of a mature plant.
Early post-commercial operation produces generally fewer scrams, but they occur from higher power. The scram rate can still be two to three times that of a mature plant. The factors that would minimize the event significance, noted above, may not be present.
B. NRC staff methods to screen events (AE00 event categorization or hRR event selection for further review /discussica) reveal a disproportion-ately higher number of events from the 22 new plants studied versus the 76 mature plants.
C. The probabilistic event evaluation methodology of the ASP program was utilized to select significant events at new plants for root cause evaluation. Even with the inclusion of a different set of plants from an earlier period, the sane root causes emerged for the more significant events as produced by the generic analyses of all events. This implies that event reduction measures which reduce the probability for significant ,
events should also prove effe::tive in helping prevent less serious events (
ano failures. The root causes cited by AITs were also reinforced in the I root cause determination process.
In addition, the ASP methodology was utilized to estimate the risk from new plants, on the average, versus mature plants. Although the technique used prcvides only an estimate of risk, it indicated that, considering the higher frequencies for transients and scrams observed for new plants along with a higher number of observed failures for some equipment and the relati/e inexperience of the plant operating staff, the risk associated with plants in their first two years of operation appears higher on the average than that for mature plants.
VI. Lessens Learned The analysis, described previously, pointed the way tc pursue corrective l actions for many of the classes of reportable events. Scram cause profiles.
l with their emphasis on surveillance testing and feedwater control [ Finding A(3)2, ESF actuations due to conservative setpoints or state-of-the-art burn-in l
xvii I
problems [ Finding B(4)], and T.S. violations due to procedural problems
[ Findings C(l) & (2)], were indicators of some more generic common causes among plants. Site visits, utilizing plant-specific data from the analysis along with the generic root cause findings, were made to identify some undarlying factors that contributed to these events. As a result, the study prodJced the following improvement lessons to reduce reportable events.
Although there may be many other improvement lessons that might improve the startup schedule or plant safety, these are directed to reduce the frequency of the events analyzed. The improvement lessons are grouped generally under the categories of management and equipment.
Some of these lessons apply to newly licensed plants only (identified by OL),
while others should be of benefit to any operating reactor experiencing similar problems (identified by Ok).
As part of the Peer Review, each of the 22 plants in this study was requested l' to identify those initiatives presented in the draft report which were judged to have the highest potential benefit. Over one-half of the olants responded to this request. Accordingly, Items VI.A(1) through (6), dealing with manage-ment lessons, and Items VI.B(1) through (7), dealing with equipment lessons, are ordered, highest to lowest, based on the original AE0D priority con-siderations and our evaluation of the comnents received. Item VI.A(7),
shich deals with training, feeds into the other items, and was not independently ranked.
Improvement lessons for Consideration by Licensees A. Management Lessons (1) Establish an operatina plant mentality well prior to initial criticality.
i Ensure that plant operations personnel have the responsibility for operating all equipment as early as possible in the construction completion process. Take early, complete centrol of the transition l frcm construction to operation. (0L)
Have personnel who will be resper.sible for maintenance and testing of plant systems after licensing begin these activities using post licensing precedures before fuel load. This lets procedures get debugged, ano the plant staff gains experience in working under licensed ccnditions. (OL)
Stress the importance of cetails, the need for discipline in follow-ing procedures, the need for awareness of plant conditions and the regulatory requirements associated with these conditions, tight coordination throughout the plant staff, and the need for expedited resolution of problems. (OR)
Minimize continued construction activities after fuel load that may have an adverse impact on plant cperations. Reduce plant staff to operational size, remove constructicr. equipment, and establish housecleaning programs. Bring A/E, NSSS vendor key personnel onsite so that problems can be. resolved promptly when discovered. (0L)
(2) Conduct a deliberate, evenly paced, thorough and well-planned xviii
i prMperational and startup test program.
I Conduct thorough reviews and dry runs for planned testing and allow time for additional testing during either the preoperational or startup testing program. Emphasize planning to reduce the frequency of unplanned scrams and unnecessary ESF actuations. A detailed review of operational experience of similar plants should be a principal guide to the areas needing additional attention. (0L)
Minimize the number of deficiencies and outstanding items carried forward. Establish a policy of complete resolution before proceed-ing. (0L)
(3) Use the finalized Technical Specifications (TSs) to generate and validate (e.g., against the as-built plant) surveillance testing procedures as early as pcssible. In this regard, great discipline should be exercised to restrict the number of last-minute changes in the proposed TSs. Once final draft TSs are issued, the licensee should begin to incorporate TS requirements into plant procedures instead of waiting until tN last few changes have been implemented. In conjunction with this activity, have plant staff (as opposed to NSSS vendor or special startep group) perform all surveillance. It is recognized that development of finalized TSs involves a joint licensee / NRC staff effort. Therefore, this measure relates directly to item A(3) under Improvement Lessons for Consideration by the NRC Staff, which addresses the corresponding staff effort. (0L) l (4) Improve administrative control of surveillance. For example:
Since problems have been experienced when work has been performed in the vicinity of instrument racks during plant operation, licensees should evaluate the location and nature of work activities during operation in terms of adverse effects on plant operation and take appropriate administrative actions. (OR)
Implement schemes to separate channel testing, such as a specific day of the week assigned to work on each channel, and to identify the channel in test, such as posting on control rcom panels. (OR)
I Blend engineering staff into the I&C organization. (OR)
Flag, categorize, and schedule surveillance according to risk of scrams or other ESF actuations. (OR)
Organize the ISC staff to establish accountability for specific equipment. (OR)
(5) Give high visibility to the sources (i.e., organizational element) of unplanned scrams (and other unplanned events) caused by human error and establish performance goals. (OR)
(6) Ensure that operating experience feedback programs: (a)combineinternal events and relevant events from similar plants, (b) communicate them directly to the appropriate first level supervisors and working level staf f at the plant on a periodic basis, including prior to startup, and (c) address preventive measures. For example, segregate the trip and xix
i ESF actuations data involving human errors from recent plant startups j into the specific positions, organizational or functional M ement, I working activity, systems and components, time of day, etc. Feed this information back at the lowest levels so that the experience of others, the complexity of what is being done, and the ramifications of errors can be seen and appreciated. (OR)
(7) A number of improvement lessons are directed at training as follows: l (a) Establish as a major goal an increased commitment to training in performing surveillance testing, calibration, and troubleshooting activities well prior to operations. I&C training initiatives, such as repeated practice for those surveillance testing activities that could cause a transient and which should be cunducted on actual in-plant equipment on live systems prior to operations, should be emphasized. An additional action to improve surveillance testing suggested by licensee staffs was training for ISC personnel in valving instrumentation in and out of service. (OR)
(b) Emphasize training for routine operations involving power level changes and the associated communications among shift personnel (i.e., feed flow and turbine evolutions) that have historically caused trips. Accelerated programs / efforts appear appropriate for newly licensed plants regarding steam generator level control.
Emphasize the need for site specific simulators to include, priar to startup, the best achievable fidelity of the simulator to the plant regarding feedwater effects (lead / lag characteristics of level indication and control methods), and include provisions to continue to improve fidelity as the startup progresses. (0R)
(c) Establish extensive, detailed training for al? segments of the onsite plant staff, including I&C technicians, maintenance mechanics, security staff, operations, and management. (OR)
This training would emphasize: (a) the applicability of the various TSs to the changing plant modes of operation and associated schedules, (b) the relationship of the TSs to the plant procedures, (c) the NRC requirements for deportability of violations, and (d) the basis for the TSs and discussion of LC0 requirements. (OR) ,
B. Equipment lessons (1) Focus on the 80P prior to operation and early in life appears to provide a high return regarding the reduction of unplanned scrams and ESF actuations. Within this area, attention could be given to:
Conducting additional reviews of feedwater and turbine control and bypass systems to identify sensitivities and plant-specific characteristics that could contribute to transients or the ability of the system to cope with expected transients. (OR)
Conducting a systematic review of equipment-protective logics and I setpoints on components such as pumps (suction trip, time delay, I vibration trip) or power supplies to identify areas where a time xx
(
I delay or additional channels for coincidence could reduce the l potential for unnecessary transients or spurious actuations. Give special attention to first-of-a-kind features not incorporated in earlier designs. Additional examples obtained from the plants visited include the main steam reheater drain high level trip and other turbine protective trips. (OR)
(2) Install test jacks and bypass switches at appropriate points in actuation circuitry. (OR)
(3) Implement on a priority basis vendor or licensee trip reduction measures.
Licensee trip reduction programs should focus on safety-related equipment as well as on 80P equipment. (OL) i (4) Pay attention to the design and installation of equipment located in the vicinity of radiation monitors and associated cabling to ensu"a that adequate grounding of equipment, cable shielding, etc., are provided to prevent the occurrence of EMI, which can triager this extremely sensitive instrumentation. (OR)
(5) Thoroughly test new or unique plant features, such as new RPS systems, )
l electrical systems, etc., prior to fuel load to reduce unanticipated failures or unexpected erratic behcvior. Emphasize planning to reduce the frequency of unplanned scrams and unnecessary ESF actuations. (0L)
(6) For future designs or major plant modifications, preference for proven j designs and standardization of design in plant feedwater and turbine I systems appears justified. Conduct further analyses of any first of a 1 kind, one of a kind, and state-of-the-art features, since they have generated a large number of problems during plant startups. (Examples of remedial actions are more extensive preoperaticnal testing, reexamination 4 of actuation logic to better achieve reliable indication and actuation; l for example, reanalysis of actuation on a single input or loss of a single input.) (OR)
(7) Incorporate scram prevention measures such as:
Develop a color coding scheme for single point scram components whose misoperation could cause a scram (for example, pressure sensinglines). (OR)
Install cages or covers over switches or racks that could provide trip signals. (OR)
VII. Lessons fcr Consideration by the NRC Staff A. Management lessons (1) Track new plant licensee progress via periodic (approximately monthly) meetings with the licensee to review the root causes of all reportable events and licensee corrective actions with plant staff present (NRC regions, with NRR and AE0D support) from OL issuance through early months of commercial operation (until mature plant levels are approached).
xxi
(2) Orient readiness reviews to deal with the evaluation of management effectiveness in the oversight of practical aspects of operations.
(3) Ensure that the review of the final plant TSs is scheduled and staffed to allow apm 3 val as early as possible prior to licensing. Since development of finalized TSs involves a joint NRC staff / licensee effort, there is a direct link between this measure and-Item A(3) under Improvement Lessons for Consideration by 1.icensees, which addresses the i corresponding ' licensee measure. ;
(4) Review the progress of individual new plants (those with less than two years of operation) in semiannual senior management meetings to provide visible NRC senior management oversight of their operation.
(5) Develop an operational readiness review element that formally addresses a ;
compendium of items from previous similar plant startups.
B. Equipment lessons (1) Highlight, through discussions with licensees, systems that historically cause, or can cause, a high number of challenges to safety systems.
(2) Provide more focused atter. tion to balance-of-plant operations (e.g., the ability of the operators with the as-built plant to survive feedwater transients and load rejections) to reduce the frequency of these transients.
i l l l
1 xxii
- 1. INTRODUCTION TMI-2 was a relatively new plant NUMBER OF PLANTS IN THE FIRST TWO when the accident occurred in YEARS OF OPERATION March 1979, about 13 months after oisetavto av riscat vtAn the plant received an operating a tectwo license (0L). Among the concerns f2L. .
that arose from the accident was "" M - -- == =
the adequacy of attention paid to C = == = ~~
operational safety by the g* y 1 regulators and plant operators 2
- 7) {f f2 j during the initial phases of i the plant's operation. Both l" the 1985 and 1986 Policy and Planning Guidance of the '
z {
Commission 1/, 2/ directed the ' d I l
I staff to ".T.coiitinue to closely []H f monitor the first two years of .o i .2 .3 4 .....,.....o operation of new plants coming on line, particularly those of Figure 1 licensees who have no prior experience with nuclear plants." The pace of licensing increased in the mid-1980's, and the number of plants in the first two years of operation is peaking in FY 1987, as shown in Figure 1. In August 1986, the NRC's Executive Director for Operations (ED0) requested that the NRC staff make an effort to better understand the reasons behind the usually high unplanned event frequencies at recently licensed plants. 3/ Various NRC program office !
initiatives are underway. In addition, as a result of reconsideration of early operating experience at Fermi 2, the staff recognized the need to
) rethink the process used to assess license performance during initial opera-i tion. 4_/ This report provides input into the overall NRC program.
It is generally acknowledged that newly licensed reactors experience a higher frequency of unplanned operational events during their early years of
, operation when compared to later years. Since the first 10 to 12 months of licensed operation are usually concerned with a detailed startup test program, and some plant systems cannot be fully tested until power operation is achieved, this is not entirely unexpected. For the power conversion systems, this program represents the first opportunity to test the plant systems throughout their full design range.
1/ U.S. Nuclear Regulatory Commission, " Policy and Planning Guidance 1985,"
USNRC Report NUREG-0885, Issue 4. Available for purchase from National j Technical Information Service, Springfield, Virginia 22161. !
2/ U.S. Nuclear Regulatory Commission, " Policy and Planning Guidance 1986,"
USNRC Report NUREG-0885, Issue 5. Available for purchase from National Technical Information Service, Springfield, Virginia 22161.
3/ Memorandum, Denton to Distribution, August 12, 1986.
SI Memorandum, Keppler to Stello, September 23, 1986.
Memurandum, Dentc1 to Stello, September 18, 1986.
As an operating plant, the regulatory climate is dramatically changed from that experienced during construction. At operating license (0L) issuance, the licensee becomes responsible for meeting technical specification system operability requirements for plant operation in various modes, and for per- '
forming surveillance testing in accordance with technical specifications. At the same time, new pressures arise to achieve and maintain full power and begin commercial operation.
Some plants achieve commercial operation in a smooth manner. They have rela-tively few unanticipated events. Similar to these U.S. plants, Japanese plants appear to be relatively event-free in their early years. Other domestic plants have difficult early years, and some may experience over 25 unanticipated scrans, over 100 unplanned actuations of plant safety features, and over 50 violations of their Technical Specifications during their first two years of operation.
Events within the first two years after OL issuance have the potential for increased risk compared to later years of operation. The frequency of significant events within two years of OL issuance, on the average, is double that of the mature 1/ U.S. plants, and decreases with time. Since the per-
~
formance of certain new plants in some areas was equivalent to that of mature plants, the concept of a learning curve should not be accepted without bounds.
Actions are pursued by licensees to reduce the high event frequencies at plants in their early years of operation. One method of identifying such action is '
through the feedback of operating experience from other plants. By -identifying lessons learned during new plant startups and encouraging licensee action in these areas, improvements may be obtained.
During this study, AE0D found that, generally, power reactor licensees recognized the need for action to achieve good early operation. They have developed programs in the areas of treining and operations that cover the spectrum of actions which, if implemented effectively, would result in improvement. However, during the period from the later stages of construction to full power operation, resource priorities may be such that improvement activities not specifically required for licensing or full power operation rray not be sufficiently high to assure implementation. As a result, in some cases, their implementation is delayed or abandoned due to the press of other activity. Late systems turnovers, potential delays for modifications, financial limitations, or the urgent need for personnel (with too little time to complete hands-en training prior to startup) may compromise the priority of some issues known to be important for reliable long-term operation.
Therefore, the primary purposes of this report are to (1) review the opera-tional experience within two years after OL issuance for a number of plants, l (2) assess and characterize the operational trends, for example, against i operation beyond the first two years, and (3) collect and reemphasize some i improvement lessons for feedback to licensees and focus by the NRC staff. The initial efforts of this study are directed toward statistical analyses of these events and the identification of trends for possible correlation to J plant and/or licensee characteristics. 1 4
1/ The data for the mature plants (those licensed prior to January 1,1983) were taken from the licensee event reports (LERs) submitted for Calendar Year 1985.
2
An initial report on new plant experience was published in August 1986, en-titled " Trends and Patterns Analysis of the Operational Experience of Newly Licensed United States Nuclear Power Reactors." For 19 units licensed in the period 1983-85, event counts for Reactor Protection System (RPS) actuations, Engineered Safety Feature (ESF) actuations, security events and miscellaneous events were extracted from the NRC Operations Center computer data base. The cumulative totals of these events as a function of time since issuance of the OL were examined for trends over time and correlation with the nuclear steam supply system (NSSS) vendor, previous nuclear experience at the utility, and whetherornottheutilityservedasarchitect-engineer (AE).
For this new study, three plants licensed in 1986 were added to the scope of the August report, and the plant event data examined were extended through June 1986. The event types covered in the analyses of this report are reactor scrams (a subset of RPS actuations), ESF actuations, violations of technical specifications and losses of system safety function. An analysis of the causes underlying the observed variation in event rates or the high frequencies was performed. In aodition, the inconsistencies in the interpretation of reporting requirements across all licensees were considered by this study.
The use of reactor scrams instead of all RPS actuations had one material con-sequence for the findings of this report versus the August report. In general, the ratio of non-scram RPS actuations to scrams has been found to be much higher for BWRs than for PWRs. Thus, the BWR scram event frequencies are significantly lower than the RPS actuation frequencies for BWRs. In contrast, the respective frequencies for PWRs do not exhibit this behavior.
Also, a further review of the reasons for the security event reporting pattern indicated it is driven by reporting requirements and their interpretation. In dt least one Case (Call 6way), there was a fairly unique reporting requirement with a lower threshold for reporting than for other licensees. As a result, Callaway achieved high counts in the August report. This was not indicative of comparative performance with other plants. Interpretations of security event reporting requirements were found to vary with time and among licensees and regions. Therefore, security events have been eliminated from the scope of this report.
The analysis reflected in the August report to identify characteristics such as NSSS and A/E type that might be related to event frequency was formalized and expanded. (Resu? s are discussed in Section 3.) This statistical analysis also addresses the preliminary finding that early operational experience seems to be puJictive of later experience. In this study, a correlation was found between the startup and early post-commercial operation event rates for scrams, ESF actuations, and loss of system safety function. Thus, rates for these events during startup were statistically significant indicators of the rates experienced during early post-commercial operation.
Further, this study makes use of the more detailed data provided in liccnsee event reports, rather than that of the more prompt 10 CFR 50.72 reports of the i August report. Discussions with licensee plant staffs, NSSS vendors, NRC l regicnal office personnel, and a review of formal startup reports allowed us j to better assess the operational profile of newly licensed plants, and to 4 provide the underlying causes for problems encountered.
t As noted above, this report incorporates three additional plants, later 1 information and data, and a number of changes in the analysis in order to l 3
l I
l improve the consistency and accuracy of the results. Therefore, this . report supersedes the August 1986 report.
Since this study is directed at the causes of events, its recommendations are directed to the prevention of events rather than event mitination.
Section 2, Background, of this report provides: (1)detailedinformationon the plants covered in this study, (2) a discussion of the elements of a typical. startup or power ascension test program, and (3) a summary discussion of what the NRC's Performance Indicator (PI) data show for piants during early operation.
Section 3, Review of Experience, examines early. plant operating histories, and provides.the resuits of detailed analysis of the frequency, trends and causes of unplanned scrams, ESF actuations, violations of technical specifications, and losses of system safety function. General causes (e.g., human error, equipment) are discussed, as well as specific examples based on review of LER text or discussions with licensees. Plant specific corrective actions also are discussed.
Section 4, Event Significance and Insights, provides additional perspective on early operational experience by focusing on events determined to be signifi-cant by various NRC programs. Systematic event screening by NRR and AE00, scrams with complications, accident sequence precursors, and events selected .
for Augmented Inspection Teams or Special Inspections are treated.
Section 5, Staff and Industry Lessons for Improving Startup Effectiveness, summarizes efforts aimed at improving the transition from construction to operation.
Section 6, Industiy Improvement lessons, provides a discussion of industry efforts to improve new plant operation.
Section 7, Conclusions, summarizes the major conclusions of this study on the status of efforts to improve operations at newly licensed plants.
- 2. BACKGROUND This study includes 22 plants that received their initial operating licenses after January 1,1983, i.e., between March 1983 and April 1986, and reviews the reportable operational events which occurred within the first 24 months after OL issuance. Design information and milestone dates are provided in Table 1, which follows.
4
Table I l
I I I Turbine- ! I MILESTONEDATES I I I I 16enerator I I I IFull Power I I i Unit Nase i NS$$ 1 Design i Supplier i A! l OL 1 Critical ! License ICoseercial !
!....................l.........!.... ...l..........l..................l...... ...l.... .....l...........l .........!
1 I I I I I I 4 I I 1 i McGuire 2 1 West 1 4-Loop ! West i Utility 1 03/03/83 1 05/08/83 ! 05/27/831 03/0!/B4 I I Diablo Canyon 1 ! West 1 4-Loep I West i Utility 1 11/08/83 1 04/29/84 l !!/11/84 1 05/07/05 1 ,
I Callanay I West 1 SNUPPS I BE I Bechtel 1 06/11/84 1 10/02/B4 1 10/18/84 1 12/19/84 I I I Catawba 1 I West ! 4-Loop I BE I Utility I 07/10/84 1 01/07/85 1 01/17/B5 1 06/29/B5 1 1 Byron 1 i West i 4-Loop i West i Sargent & Lundy i 10/31/94 1 02/02/85 1 02/14/B51 09/16/B5 !
! Wolf Creek l West i SNUPPS 1 BE I Bechtel/S&L 1 03/11/85 1 05/22/B51 06/04/85 1 09/03/B5 1 1 Diablo Canyon 2 i West i 4-Loop ! West I utility I 04/26/85 1 08/20/B5 I 08/26/B5 1 03/12/86 I t Millstone 3 I West i 4-Loop I 6E I Stone & Webster i 11/25/B5 1 01/23/B6 1 01/31/06 1 04/23/06 I I Catauta2 i West i 4-Loop i BE I Utility 1 02/24/B6 1 05/08/06 1 05/15/861 08/19/06 l l
l St. Lutie 2 l Cosb I 2/4 i West i EBASCO l 04/06/03 1 06/02/83 1 06/10/83 1 08/08/03 I I Waterford 3 1 Coat i 2/4 I West i EBASCO I 12/18/84 1 03/04/65 1 03/16/951 09/24/85 I I Palo Verde 1 1 Coeb ISysten80! 6E I Bechtel I 12/31/84 1 05/25/85 1 06/01/95 1 02/13/96 I I Pale Verde 2 l Cosb ISystes801 6E I Betthel i 12/09/85 1 04/18/B6 1 04/24/B6 1 09/22/06 l
! LaSalle2 1 6E I BWR/5 1 6E I Sargent & Lundy 3 12/16/93 1 03/10/84 1 03/23/841 10/19/84 i Ikashington Nuclear 21 6E I TWR/5 I West i Burns & Roe i 12/20/931 01/19/04 1 04/13/84 l 12/13/84I I Susquehanna 2 1 6E I BWR/4 i BE I Bechtel 1 03/23/84 1 05/08/64 I 06/27/94 1 02/12/95 1
! Liarrick I 6E I BWrt/4 1 6E I Be:htel
- 1 10/26/84 1 12/22/04 1 08/08/85! 02/01/06 I I Shorehas ! 6E I PA/4 1 6E i Stone & Webster i 12/07/B4 1 02/15/85 l I I I Ferst 2 l 6E I PWR/4 IEng. Elet.1 Stone & Webster 1 03/20/65 1 06/21/95 1 07/15/B5 1 1 1 F.iverBend 1 6E I BhR/6 1 6E I Stone & Webster 1 08/29/85'l 10/31/B5 1 11/20/B5 1 06/16/B6 1 i Perry I 6E I BWR/6 i BE I 6ilbert 1 03/1B/861 06/06/B6 1 11/13/86I I I Hope Creek I 6E I BWR/4 i Bechtel BE I I 04/11/06 1 06/28/861 07/25/86 1 12/20/B6 I l
l l
l I
Since only data through June 1986 are used, most plants have fewer than 24 months coverea. A ma,ior reason for the June 1986 cutoff was to allow the analysis to be viewed as historical, thus avoiding misinterpretation as an alternathe PI approach, or as addressing current performance problems. This decision limited the amount of data available for plants such as Perry and Pope Creek. Nonetheless, a review of experience from such cases indicated the causal experience to be more or less typical of that seen for plants with more data in the study, and that, as long as the data were handled appropriately (e.g., event frequencies averaged over the startup program would not be calculated unless the program had been completed), the data were valuable and their use would not skew the results.
5
l i 4 'I l.
2.1 Overview of Startup Test Programs Figure 2 As illustrated in Figure 2, the 24-month period after OL includes, at least partially,
. . . . . . the startup. testing program m,, , .,.., l ' 4 ' ' ' 1 ' ' ' " " g'," " " " " " ;" " " " '"' for each plant. However, the data are displayed such l.7tl'. i
..s.n. ,
l%%
i...,
i #y'x l
. i 4i m'i that this power ascension C.".I.l c.n ,
,* %f'ibe' n.u-
- e l e %- ]
m i ,
r phase can easily be identified and separated if
- x
""*'.' l%% 5.'N e' I e desired. The power ascen-7,,2 ;;* ;x.f l el ld sion program consists of
. . . , i.. . N l e i activities commencing with
"!' 7.* ' ll%:", l" initial fuel' load (which
"+r =~~ 1 xX*J$l
'l~h generally begins within a ll*.";'" ' ll2'I i ve,, l d '
fewdaysofOLissuance)and
, 'lt** 'N y' ' ending with a warranty run T.R. , 1 4.* . x of 100% power and declaration
""',.. l[2 Ix of commercial operation
,t, ,,,, (which, for the plants in this study, were nearly
- ',",',""',',l".".
always concurrent). Within x mn.i wo..ni, each of the three NSSS 1
$ ';".l*"",'"".",, vendor classes represented by the plants in this study I
[ General Electric (GE), !
Westinghouse (WEST), and Combustion Engineering (CE)], the formal startup test program defined by and committed to in Chapter 14 of the Final Safety Analysis Report (FSAR), as amended, is fairly uniform.
For the PWRs, the startup test program consists of the following major segments:
(1) Fuel load.
(2) Pre-critical or post-core load hot functional testing.
(3) Initial criticality and zero/ low power physics testing.
(4) Power ascension testing, including warranty demonstration.
The fourth segment consists of testing at various increasing power level plateaus. While testing occurs at many levels from 5% to 100% power, the major benchmarks for Westinghouse plants are 30%, 50%, 75%, 90%, and 100%, l while for CE plants they are 20%, 50%, 80%, and 100%. The formal startup j program includes from four to five planned scrams, including in all cases J scrams for loss of offsite power, shutdown from outside the control room, and ]
plant trip from full power. inese or other planned scrams were not counted or {
displayed in the analysis. Only unplanned scrams were identified.
i 6
i l
For BWRs the formal startup test program is fairly standard across the GE i product line (BWR-4, -5 and -6), and is comprised of the following major l segments: '
(1) Open vessel testing (fuel loading and low power physics tests). q (2) Critical heatup to rated pressure and temperature.
(3) Power ascension tests.
(4) Warranty demonstration.
The power ascension tests comprise testing in six different zones of the BWR power flow map. While slight variations exist, these zones are defined analogously across the product line. The startup test program includes five to seven planned scrams as part of the testing, again including loss of offsite power, shutdown from outside the control room, and plant trip from full power.
2.2 Sumary of Relevant Performance Indicator Data Currently, the j NRC's perform- "'" PLAmis - UnPLAsusto AuTonATsc '
^"' " " " "'"
ance indicator (PI) program is a LEGEND used to collect '"
- l data to assist
- the NRC staff in ,_
5" "
recognizing poor s DNTm or declining ja~
safety perform-m*' '
ance. Selected $
operational M ...
areas are W monitored to 2 .-- '
identify E Y[ \ ~ , ....
potentially undesirable
\,p\f #\/ '
trends in a a.
timely manner .
for further 07"'
QTR2 QS3 QTR4 Q"5 QTR6 Q"7 QTRS QT"' T"'1 QTR10 QTR12 investigation. oTR AFTER INITIAL CRITICAUTY Curves, such note: new piant average based on 9 - is observations: oid piant average is o.89.
as the one shown in Figure 3 Figure 3 for unplanned automatic scrams from critical, were constructed using 1985 and 1986 data (see Appendix A).
For each operating plant and each indicator, data was taken from the PI database and transposed to the proper quarter following initial criticality.
For example. he first calendar quarter of 1985 might correspond to the second quarter fol ,ing initial criticality for Plant X. The individual values assigned to each quarter (both the number and identities of plants varied from quarter to quarter) were averaged to develop a composite curve for each indicator. The standa,d deviation for ea+ quarter was also plotted to 7
provide perspective on the spread of the individual values about each mean.
Plants which did not contribute to any of the 12 quarters after initial criti-cality (i.e., the first quarter of 1985 was at least the thirteenth quarter after initial criticality) were used to calculate the "old plant" average.
It is important'to note at this point that the standard deviation on the PI data is extremely large. The scram rate is so diverse that it is not well characterized based on such average data as is shown in Appendix A. The scram rate for some newly licensed plants was as low as that of their mature counterparts, while-some mature plants have indicators similar to the averages for the poor performing new plants.
There is no absolute scale with which to measure the new plant statistics. The PI data reveals average scrams per quarter (for unplanned automatic scrams only) of about 3.5 times the mature plant over the first year. For this same time period, the significant events data (based on the PI definition) reveals that the new plants have about two times the number of significant events as the-mature plants. Therefore, both the number of the events and the number of significant events are higher during this initial operating period.
However, new plants, as a class, show remarkable improvement over their first year of cperation. All of the indicators ' initially trend down sharply.
Again, this data is based on averages. On a plant-specific basis, the characteristic learning curve and the general improvement may or may not be the case. Over the next 2 years of operation (note that the PI data show experience over 12 quarters of operation), the scran rate for unplanned .
automatic scrams drops to less than twice the mature plant average. The {
indicators for significant events, equipment-caused forced outages, and safety system failures appear to trend to the mature plant levels, while the safety feature actuations (other than scrams) and the average forced outage rate per quarter actually improve to better than the mature plant average.
The INP0 performance indicator program includes plants beginning January 1 of the second full calendar year of operation following full power licensing. !
(This milestone occurs between 1-1/4 to 2-1/4 years after OL issuance). In this manner, a learning curve should have occurred such that inclusion in the ;
industry average should minimize the new plants' effects. Similarly, the NRC's PI program separates new and mature plants for any averaging of a populatien i group, j
- 3. REVIEW 0F EXPERIENCE For this study we reviewed information on four major classes of reportable un-planned events: reactor scrams 1/, ESF actuations, violations of technical specification,, and losses of system safety function (LSSF). These areas are covered by the reporting requirements established January 1,1984 by 10 CFR 50.73. Sources of 1983 data for plants licensed prior to 1984 are described as needed in the discussions of experience in each area. ;
i 1/ In this study, we have included all unplanned RPS actuations with rod motion after initial criticality as scrams. Both unplanned manual and automatic actuations are included.
8 1 1
1 i
)
I
\
For each area, event counts were generated from LER data for each plant, and cause categories (defined in Appendix 8) were assigned to each event based on the LER description. This information was examined for any general trends or patterns for up to two years of operation for the 22-plant population. The general cause categories were: design, equipment, personnel error, procedures, other and unknown. It should be noted that it is very difficult to determine from LER texts whether equipment failure is random or due to a design deficiency. Therefore, the distinction between design and equipment is often unclear.
Figures, such as those shown in Figures 4 and 5 for scrams, are provided for each of the four trended parameters (Appendices C, E, F, and H). These I figures are constructed such that plant status (operating or prolonged shutdown) is trended along with the number of events and cause categories for the events. In this manner, the plant operating cxperience is clearly displayed and variations in causes among plants are easily observed. In these figures, OL marks license issuance, IC initial criticality, FP full power license, and C0 commercial operation.
UNPLANNED ltEACIVR SCRAMS UNPLANNED REAUIUR SCR AMS WASH. NUCLEAR 2 CALLAWAY
$ Gi".t.,, , 0 Ef='.,,.,
i$Ob:T i$$b:T -
I na
-a7 Iw .
f e e
t I ft i l
f . .
?
,y! h3 1 , ,
dallBBM i8, ua.:.:x,,::,: - - - ua,:.:x,,:: : - - -
Fioure 4 Figure 5 For example, both of the above plants (WNP-2 and Callaway) experienced high scram frequencies. It can be observed that the availability of Callaway (based on critical hoLrs) was good. The cause categories for its scrams car be observed to be dominated by equipment, while those of WNP-2 show a slightly higher proportion of personnel error. Therefore, these figures reveal, for the plants in this study, the frequency of the events, their general causes, trcnds in these cause categories over time, and variations in causes among the new plant population, and also inform the observer of the plant status or availability over the period.
The study of causes through such a simple categorization process too often masks the true nature of the events' root causes. Indeed, the information in LERs, even though subjected to a thorough review by licensees, often does not link the root causes to underlying factors such as the overall management of operations or schedule pressures. Thercfore, more detailed cause analysis was 9
l performed using startup reports, discussions with NRC regional office personnel, site discussions with licensees, discussions with NSSS vendors and other available documentation of early post-licensing operational experience. Site visits were conducted for the following plants: Byron 1, Callaway, Diablo Canyon 1 and 2, Hope Creek, Palo Verde 1 and 2, St. Lucie 2, Susequehanna 2, Washington Nuclear 2 (WNP-2), and Wolf Creek.
During these visits, two related factors underlying good startup performance at a new plant were brought out repeatedly by discussions with plant staffs. Although these factors were cited by licensees as being influential in achieving good performance in each of the specific areas covered in this study, they are stated here as a context for the specifics in the sections which follow.
The first key factor was the need to establish an operating plant mentality in place of a construction mentality prior to fuel load. This involves conveying the importance of details, the need for discipline in following procedures, the need for awareness of plant conditions and the regulatory requirements associated with these conditions, close coordination throughout the staff and the need for expedited resolution of problems. The importance of these points was echoed in the Region III review of lessons learned in licensing Fermi 2.
This review also highlighted the need to ensure that training programs address routine operations as well as abnormal operating conditions.
Secondly, there was a real benefit associated with a deliberate, evenly paced pre-operational and startup program wherein problems are resolved as they come up and the number of open items carried forward is minimized. Generally, proceeding with a startup program with numerous outstanding construction items, i.e., with systems not turned over to the startup organization or turned over with deficiencies, resulted in difficult or prolonged startup programs. This latter case often involves construction personnel remaining onsite, which negatively impacts the transition to an operating mentality, as well as serving as a source of spurious ESF and scram signals and technical specification compliance problems. l 3.1 Reactor Scram Experience As a class, unplanned reactor scrams can be a safety concern, such as unplanned reactor scrams from high power where recovery is complicated by additional equipment failures or operator errors. A new plant would appear more vulnerable than a mature plant in this regard, given an elevated challenge frequency, mitigating systems which have not been thoroughly tested, and personnel with limited operating experience on the given plant. It is recognized that generally during the early months of operation the licensee staff is augmented by experienced NSSS and A/E personnel, plant power is usually low, the core fission product inventory is also low, and all involved are aware of the potential for unusual events in this early phase. Unless stated otherwise, all scram data described in this section is for unplanned scrams, both manual and automatic, that resulted in rod motion. Scrams as part of a planned test were excluded.
3.1.1 Data Analysis Figure 6 displays the cumulative number of unplanned scrams versus the cumulative critical hours for all 22 plants, grouped by NSSS vendor (Appendix C contains a separate curve for each plant.) The slope of the curve at any point represents the average scram rate. The steeper the slope, the higher the scran rate; the rate of change of the slope represents the rate of 10
1 i,
Composito Unpl2nnsd Reactor Scrcms {
40 , i West j 30 - -
O
", a McGulRE 2
> 20 - o MILLSTONE 3 -
'j g a WOLF CREEK
+ BYRON 1 3 .___.a x CALLAWAY E * '
m
- CATAWBA 1 0 / v CATAWBA 2 39 _ _~'
s DIABLO CANYON 1 _
e DIABLO CANYON 2 1945 Average Rate - Mature Plant Slope oL O 5000 10000 15000 Cumulative critical hours 40 i i
, 30 - -
E e
E I 20 -
/ - - -
/ -
I
_=== ~
U o PALO VERD21 gg _ ,
o PALO VERDE 2 l A ST. LUCIE 2 1
+ WATERFORD 3 j 1985 Average Rate . Mature Plant $10pe I 0:
O 6000 10000 15000 Cumulative Critical hours 40 , ,
- O SHOREHAM l
30 - SUSQUEHANNA 2 e a WASH NUCLEAR i E + FERMl 2 m x HOPE CREEK
@
- LASALLE 2
,
- LIMERICK
> 20 -
e PERRY -
"Q
- RIVER BEND A
10 - d ~ ^ ^ ~ -
~
_ ~_ ._ _ 1985 Average mate . Mature Plant slope Oi O 5000 10000 15000 Cumulative critical hours Figure 6 11 I
I t-_______-
improvement (or deterioration). The figure contains a reference line which represents a scram rate of one scram per thousand critical hours, which was approximately the average for 76 mature (i.e., older than 24 months post OL)
LWRs in 1985.
In nearly all cases, the unplanned scram rate greatly exceeds the reference industry rate for the first 2500 critical hours. The fifteen plants which reached this milestone in the time frame of this study, did so within plus or minus 4 months of the coramercial date; nine plants reached 2500 critical hours prior to commercial operation, and six reached it after. A notable exception to high early scram rates is Limerick, a BWR-4. Limerick attributes the very low scram rate during power ascension testing to staff augmentation from another unit by key personnel with sufficient lead time to become familiar with the plant prior to testing. Limerick also performed turbine roll testing while licensed only to 5% power. Completion of the roll testing was considered a critical path time savings, provided assurance of a healthy turbine, and helped build staff morale.
Average scram rates, as shown in Table 2, were calculated for the pre-commer-cial period for those plants which became commercial (i.e., completed the power ascension program) prior to June 1986. In addition, the average post-commercial scram rates were calculated for the first 180 days after the date Table 2 Average Unplanned Scram Rates Pre-Commercial Post-Commercial Scrams Critical Hrs Scrams / Scrams Critical Hrs. Scrams /
Plant 1000 CH 1000 CH McGuire 2 11 3724 2.95 9 3416 2.63 Diablo Canyon 1 12 2745 4.37 5 4160 1.20 Callaway 12 1243 9.65 10 3917 2.55 Catawba 1 9 2266 3.97 3 3483 0.86 Byron 1 22 3746 5.87 5 2691 1.86 Wolf Creek 11 1746 6.30 5, 413$ 1.21 Diablo Canyon 2 17 2474 6.87 N/A, N/A, N/A, Millstone 3 8 924 8.66 N/A N/A N/A St. Lucie 2 6 1101 5.45 7 4024 1.74 Waterford 3 21 1812 11.59 9* 329] 2.73 Palo Verde 1 13 3437 3.78 N/A N/A N/A LaSalle 2 9 3871 2.32 2 3003 0.67 WNP-Z 23 3950 5.82 8 3006 2.66 Susquehanna 2 7 3794 1.85 2, 3741 0.53 Limerick 4 4862 0.82 N/A, N/A, N/A, River Bend 16 3541 4.52 N/A N/A N/A Average 5.30 1.69 Note: These plants had not accumulated 180 days of commercial operation before June 30, 1986. Therefore, their post-commercial experience
.s not considered in the statistical analysis.
12
of commercial operation if the plant had accumulated that amount of commercial time prior to June 1986. This approach was used in order to get comparable and reasonably long time periods over which to everage. As a result, not all plants in the study are used in this part of the analysis. The data in Table 2 (and similar tables for ESF actuations, technical specification violations, and LSSFs in later sections) are grouped by NSSS vendor and then from the earliest to latest (oldest to youngest) OL date within NSSS veador. These data do not provide a Lasis for ranking plant performance.
A formal statistical analysis of these rates looked for correlation between the startup average unplanned scram rate and the post-commercial unplanned scram rate. In addition, each of these rates was tested for correlation with: l l
The length of the startup period (days). j Calendar date of operating license.
First or second unit at a site.
Type of Architect Engineer - utility /non-utility.
First or subsequent nuclear plant for a utility.
NSSS vendor - Westinghouse, CE, GE (BWR-4, -5, -6).
The correlation between the startup and post-commercial average unplanned scram rates was the only statistically significant value determined by computing the Pearson product-moment correlation coefficient, 0.64.1/ No other statistically significant correlations were found; a negative correlation between the length l of the startup period and the pre-commercial average rate came closest to being significant. That is, the unplanned scram rate during the startup program tended to be higher for shorter startup pericas.
Details of the statistical analysis are provided in Appendix D. The appendix provides a discussion of the methods used, the calculated rates, the values of the variables tested for correlation (e.g., the length of the startup period),
the correlation matrices, and results of non-parametric and linear modeling methods.
As shown in Table 2, while unplanned scram rates are generally higher than those for mature plants, the degree to which this is true varies widely from plant to plant. In fact, in the case of Limerick the scram rate during power ascension testing was better than that found at many mature plants, and better than the industry average for mature plants. This is an indication that elevated scram 1/ A correlation coefficient of zero indicates there is no relationship between the variables; when there is perfect correlation and the variables vary in the same direction, then the coefficient is 1.0. When there is perfect correlation but the variables vary in opposite directions, the coefficient is -1.0 (negative correlation). The coefficient can vury between the extremes of 1 and -1 to indicate some intermediate degree of correlation.
l l 13 I
1
frequency is not inevitably linked to early operation.1/ This indication is clearly apparent by an examination of early operation at Japanese reactors.
The first 24 months after fuel load were examined for five Japanese reactors ;
which started commercial operation in 1985: Fukushima 11-3 and Kashiwazaki-Kariwa-1 are both BWRs; Sendai-2, Takahama 3 and Takahama 4 are all PWRs.
While maintaining fairly high critical hours through the period examined, only two of the plants experienced an unplanned scram (one at each plant) during this 24-month period 2/
Scram characteristics (i.e.., cause category, the system that initiated the scram, the activity that contributed to the scram) were aggregated across the 22 plants for comparison with similar data for 76 mature plants in 1985. The statistics are shown in Figure 7. The new plant data is divided into the pre-commercial and post-commercial phases, i.e., the commercial dates for the 22 plants were used to split the data into the two phases. If a plant had not yet gone commen ial, all of its data went in the pre-commercial category. In all three regimes (mature, post-commercial, pre-commercial) complicated scrams with additional failures or personnel errors beyond those which initiated the scram are indicated. These are generally the more significant events. The review highlighted some general results for the pre-commercial and post-commercial new plant e..perience:
Surveillance testing contributes more to new plant scrams in the pre-commercial phase than to mature plants (32.3% versus 11.3%).
New plants showed a slightly greater tendency toward procedural deficiencies during the pre-commercial phase (11.5% versus 6% for mature plants). Howaver, the percentage of procedural deficiencies for new plants during tre post-commercial phase is about the same as that for mature plants.
The initiating system distribution is roughly comparable between mature and new plants: feedwater dominates (slightly higher for new plants).
In the pre-commercial period, the turbine contribution is slightly higher and the electrical system contribution lower than fcr the mature plants or post-commercial.
Figure 8 shows the same scram characteristics and comparisons in terms of scrams per thousand critical hours. This figure combines the frequency of scrams with the relative proportion from a particular source, e.g., from the feedwater system, within each of the three groupings. This figure illustrates the origin of tl.: " excess" scram rate during early operation.
Figure 9 shows the power distribution of unplanned reactor scrams for the date analyzed in this study and data for 76 mature plants in 1985. The figure shows the 1/ We note that several of the nuclear plants included in this study estab-lished records from a production standpoint. For example, Diablo Canyon 1, Callaway and Wolf Creek successively surpassed each other's record for power generation during the first year of operation, and Limerick established a record of 198 days of continuous operation during the first fuel cycle.
2/ Private communication.
14
1 i
UNPLANNED REACTOR SCRAMS Percentage By Activity LEGEND m mmm--- -_____q *
%= 'i.as.d c===nisi inidindd" g ..c~...
. . , , , . . E ,i...
_ gee-
= _
St..*s St.t. ~ f f f f f f f f f f f f f f f f f f ff f fJ g
...... p
. i. = = =
e.~.ai... .e ros.:
UNPLANNED REACTOR SCRAMS Percentage By System LEGEND r .w .71TF77 i277iss~11 i2 lrT T---1, &'""*""'"'"*"'*'
~ " ' " ' ' * " " ' ' '
m ."n a,s ,r 1111 - 7 z 111, .
- n ,------.
5 sm 77Ts> r 1. ,
Y.,
m t.0 - 73
.F
% iw ..o P599B , , , , , , , , ,2 2
~ , . . .
g ;
c ..s. - g I
~ , . ~
h e.i. .ns..i., ., t. .i UNPLANNED REACTOR SCRAMS Percentoge By Cause ,
LEcEND 1 1
,- _?-n_m e~g: ~ =m ,z==:ssa-: a ' g- ....<-n.i
- i. .
I; g ,i . c ,..i
- ,m mm m, m -, . . .
m.-
8 -_
g
.e -
._ y l
. . is is 4~
e.a ,.n....
e
.nv.i.
4 io n Figure 7 15
NORMALIZED UNPLAMdLD SCRAM RATE COMPARISON TY ACT!YlTY ,
LEGEND J erwsummyyn , , _- . ens Plett Pro tamarectal
.t.
ees Flett Pett Comorretet fra.sttte - estere Plants
_ p---
- lteasy State.
.e
.t..
,_t.., p Calik et te .
9.0 0.6 3.0 1,4 2.0 u , s.com./iooo enu. i Howre NORMAllZED UNPLANNED SCRAM RATE COMPARISON BY SYSTEM LEcEND g,,,,,,,,,,,g_,,,,,
W -new Plants. Post teueetial metere Plets Osmer - L E ..
- - we
, l
[ = -P flectrica" p
. it.,
y {
c.e.....
f
.. .t.r p 1 0.0 86 1.0 4 ,6 No. Serems/1000 Cdtical Neuro NORMALIZED UNPLANNED $ CRAM RATE COMPARISON BY CAUSE LEGEND w-c . ..:m : n-e xscu- :e rss s:-zwa. 5mn gm .......t.i "cm.ssvz.9mz:rswz.xmm Reture Plenis ehmap terse .
4
~ ~ . .
f s..
B 0....
- 0. .
p 8.0 0.5 B .0 3.4 3,0 8,8 Pee, serome/tooo cMticot Heur.
Figure 8 16
relative percentage of unplanned scrams originating within each 10% increment of power for the pre-commercial, post-commercial and mature data. It also shows graphically the proportion of scrams in each power level / data class combination where recovery was complicated by additional equipment failures or personnel errors not directly related to the cause of the scram. .
i The profiles shown in Figure 9 reflect the ways in which the plants are a operated in the three different cases rather than tha relative frequency of )
scrams. Plants in the pre-comercial or startup program phase are encountering and eliminating problems as they ascend in power. Thus, this phase is dominated by scrams below 20% power, with 70% of the pre-commercial scrams l occurring below 30% power. 1/ This contrasts with the post-commercial distribution which is very similar to that of the mature plants wherein scrams above 90% power are most frequent. In the latter two cases, difficulties in l
starting up (i.e., at power less than 70%) still exist but are not as i prominent, and the plants basically run at full power as much as possible.
O Figure 9 Average LIB IF E3PSD CERDia Unplanned g Scram Rate i mI h hhh ~1 per thousand sa la is is la 3a ts is is!! Un critical hours 0, J ,
,l 0 C hO hh h ~2 per thousand critical hours om is is em 4m tg 4m Rc in / 4.s n 0 I mana I
1 a
l ]-
- h I hhha ~ 5 per thousand critical hours tm ou la tm s.s is tu tw tu um He ii-a n H e sia n na se lHE me The percentage contribution and distribution by power level of complicated scrams are also very similar for the mature and post-commercial cases. For the mature plants, 21% of all scrams are complicated; the comparable figure for post-commercial experience is 20%. Further, for both mature and post-commercial cases, about 12% of all scrams are both complicated and from greater than 90% power.
To obtain a proper perspective on the relative risk potential of unplanned scrams during early operation, the power and complication profile must be M Adjusting the pre-commercial profile by removing data for plants that had not completed the full startup program had no appreciable effect.
17
combined with the scram rate. As a plant moves through the startup program, the majority of the unplanned scrams occur at low power levels, making up the greatest l contribution to the high average scram rate of close to 5 scrams /1000 critical 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />. Only toward the very end of the startup program is the typical plant spending any appreciable time at higher power levels, e.g., greater than 90%
power. Then the plant completes the startup program and begins to operate at ,
full power in the same My as a mature plant. It is during this period of !
transition from a startup mode to a base load mode that the scram rate begins to decrease; i.e., at around 2500 critical hours or within a few months before or !
after the commercial date. However, as shown in Table 2, the scram rate during the early months of commercial operation can still be a factor of 2 to 3 higher than the mature plant average. At the same time, the test program is over, additional support staff are departing, and the operating staff still has limited experience with the plant. Thus, the period immediately following the end of the startup program could pose a period of increased risk.
3.1.2 Cause Analysis A formal statistical analysis within the 22 plant population was conducted addressing potential correlations of scram frequency with (a) NSSS vendor, (b) length of startup period, (c) first or second unit at the site, 1/ (d) type of A/E (utility /non-utility), and (e) first or subsequent nuclear pTant for utility. The results of this analysis were that no statistically significant correlations were found.
When the scram experience of the 22 new plants is viewed against that for mature plants, the post-commercial plants have approximately the same scram initiating system distribution as mature plants (differences were generally 5 percent or less; see Figure 7). The feedwater system dominated in all cases, although its contribution was slightly higher for the new plants. In the pre-commercial period, the contributions from the electrical system and main generator are less and the contribution from the turbine is slightly greater. The reason for these differences could not be determined from available information.
In terms of the causes of scrams, the distribution among pre-commercial, post-commercial and mature plants was similar, i.e., no marked differences in terms of design, equipment and human error. Procedural deficiencies were a larger source of scrams for pre-commercial plants (as expected). While the proportion or percent contribution from these causes was similar for each experience group, the actual number or frequency of scrams from these causes was not. Since there are, on the average, five times more scrams for the pre-commercial plant than the mature plant, the number of scrams caused by human error, for example, will be five times higher or five times more (
frequent than for the mature plants (See Figures 7 and 8). ]
The pattern of causes on a plant-specific basis was inspected by examining Figures C-1 through C-22 of Appendix C. We looked in particular to see if the earliest months, when the highest number of scrams occur, were consistently dominated by a single cause, such as human error. This did not appear to be the case, as illustrated by the figures for WNP-2 and Callaway shown earlier (Figures 4 and 5, respectively).
There were significant differences in the activity characterizations involved l with the unplanned scrams. Not unexpectedly, testing dominated the activity 1/ This part of the analysis did not compare the first unit with the second unit at the same site. It only characterized a plant as being the first or the second unit at a multi-unit site.
18 l
a
in progress for scrams occurring in pre-commercial _ plants, accounting for about 35 percent of all scrams. The contributions from testing for post-commercial and mature plants were similar, but less than half as much as for pre-commercial plants - about 10 percent of all scrams. In all, almost one-half (actually 45 percent) of all scrams in the pre-commercial phase were associated with testing, calibration, troubleshooting or maintenance - either through personnel errors or procedure problems on one hand, and equipment failures that occurred during testing on the other. (The two situations occur in almost equal proportion across systems, with a slight tilt toward human error and procedural deficiencies for the RPS.)
3.1.2.1 Equipment-Related Causes Unplanned scrams during early operation originate most often in the feedwater system (about 29% of all scrams for pre-commercial plants, slightly less for post-commercial plants). Of the other systems usually classed as B0P or power conversion, i.e., turbine, main steam, main generator, and condensate, the turbine system (about 15% of all scrams for pre-commercial plants,10% for post-commercial plants) causes the greatest difficulty. Within these systems, as well as for the electrical systems, equipment or hardware problems hold a slight edge over procedural and human error problems in causing unplanned scrams. The reverse is true for the reactor protection system and control rod ,
drive system. !
As a whole, the equipment or hardware problems leading to unplanned scrams originate from many different specific hardware features of the plant design.
Recurring failures of subcomponents are the exception rather than the rule. The identification of the equipment problems and their subsequent elimination follow i, j the debugging model that would be expected of a new facility.
As expected (and as is the case for mature plants), equipment problems in the '
feedwater system originate most often with feedwater pumps and the main feed-water control valves and their associated control systems. The analysis revealed the causes to be dominated by equipment problems (52%) and human error (31%). They occur most often while attempting to change plant power level (36%), followed by steady state operation (23%), and testing (19%). The most prevalent feedwater problems involved: (1) feedwater flow control during startup and at low power levels, (2) feedwater regulating valve control systems, and (3) feedwater pump control systems. These areas are discussed below. Discussions with plant staffs provided examples of experience, slightly different in each case, where scrams could be traced to difficulties with operation of the main feedwater control valves, or action was needed to prevent scrams or damage to the plant:
Although Byron 1 benefitted from earlier experience at Zion in selection of the type of valve used for main feedwater regulation, the Byron 1 staff reported that sluggish response of feedwater regulation was identified during preoperational testing. The root cause of this sluggish behavior was identified as improper time response of the feedwater turbine controller.
- At Diablo Canyon, the licensee found that feedwater regulating valve control problems also existed. The valve control design was such that the opening and ciosing rates differed by about a factor of two, resulting in poor dynamic response and instability, i
19
4 Similarly; 'at Palo Verde, the licensee found that the feedwater.
regulating valves would hang up momentarily when changing position, thus inducir.9 perturbations in feed flow.
Although'not a direct cause of re' actor scrams at WNP-2, the licensee s
staff found that when a reactor feedwater pump was operating at minimum speed in a hot standby mode of operation, the differential pressure across the startup level control valve was excessive. The pressure against the valve resulted in excessive piping vibrations that damaged hangers and the valve actuator.
As befits the unique nature af each problem, the corrective action taken was plant specific: "
4 c
Byron 1 eliminated sluggish response of feedwater regulation by )
increasingthe'tjmeresponseoftheturbinecontroller.
At diablo Canyon, management held the plant at about 30% prwer for a month'to fix the feedwater regulating valves. The system was modified so that the valve opening and clos'ng times are the same.
s At WNP-2, they redesigned the nlve and lowered the feedwater pump minimum speed setting.
The following examples are typical of the problems encountered in feedwater pump operation. The example from Waterford 3 illustrates the difficulty that can be associated with root cause determination, and the impact of feedpump protective trip point selection g At both Callaway and Palo Verde, each licensee experienced failures of the feedwater pump speed controller and installed new circuitry (card in the speed controller) to eliminate further reactor trips. At Diablo Canyon, repeated problems with the feedpump Woodward governors prompted their replacement with newer Lovejoy controllers.
At Waterford 3, two scrams occurred because of feedwater pump trips caused by high vibration. After the first event, the licensee determined that the feedwater pump high vibration trip setpoint was lower than desired. However, even though the licensee subsequently raised the setpoint, about 4 months later another feedwater pump trip on high vibration caused a reactor scram on low steam generator level. The
? licensee could not determine the cause of the high vibration feedpump trip after this second event. The licensee decided to remove the feedwater pump high vibration trip after discussions with the pump manufacturer. Three other scrams caused by problems with the feedwater pump involved equipment problems with the feedwater pump speed controller (two scrams), and a feedwater pump oil leak. One of the feedpump overspeed trips occurred during the startup program; the other occurred in early commercial operation. In the latter case, the licensee determined that the feedwater pump governor cup valve seat was undersized and caused the pump to overspeed. Although the licensee did not specifically correltte the two events, a feedwater pump overspeed trip which led to a scram during the start-up program may have also been caused by the improperly sized valve seat.
l 20
^
1 The turbine, main steam and steam dump (or bypass) systems account for roughly 18% of the new plant scrams covered by this analysis. Equipment failure and human error account for 56% and 24%, respectively. Testing contributes'to the scrams in 46% of the cases, followed by 26% during power changes. Turbine stop valve testing and MSIV surveillance testing are largely responsible for the high testing contribution.
Equipment and/or design problems in the turbine and bypass or steam dump systems show the same " learning from experience" characteristic:
The licensee for Byron 1 determined that there were some turbine impulse pressure scaling errors associated with the turbine power signal P-13, which had also contributed to some of the scrams for the turbine system.
These scaling errors were subsequently corrected. Since these corrective measures were implemented, no additional scrams of this type were experienced at Byron 1, either during the remainder of 1985 or during the first six months of 1986.
On an initial attempt to synchronize the generator at Byron 1, the digital electrohydraulic (DEH) turbine control system failed to automatically accept load, and the preliminary settings provided by Westinghouse (the turbine supplier) had to be adjusted.
WNP-2's DEH turbine control system problems were more complex and varied.
Two WNP-2 scrams were caused by failure of electronic circuit cards in the DEH turbine control system. In both cases, the faulty caras were replaced. Two scrams were attributed to hydraulic problems with the system. In one case, pressure fluctuations in the DEH system emergency trip header allowed leakage of DEH fluid past the closed seat of a bypass reset solenoid valve to one of the four main steam bypass valves. The leaked fluid caused the main steam bypass valve to fully open.
Subsequently, the three remaining bypass valves compensated for the initial decrease in steam pressure by cycling to the fully closed l position. This, in turn, caused steam pressure to increase to the reactor high pressure trip setpoint. The licensee's corrective actions included reworking the leaky valves and placing a check valve in the header common to all four solenoids to isolate the valves from any l pressure fluctuations. The other major problem with the hydraulic fluid I system involved inadequate fluid pressure. The plant modified the oil
~
i supply path from the supply pumps and increased the size of the i accumulators from 1 gallon to 2.5 gallons. The other DEH turbine control 1 system problems required modifying the input signals to the system and I instrument calibration.
In regard to the turbine and the turbine control system, the plant staff j believes that many of the initial problems were due to the incompatibility ]
of the turbine system (Westinghouse) to the NSSS (GE). Although a good l number of Westinghouse-designed plants utilize a GE turbine system, very i few GE plants (Cooper and WNP-2) appear to have Westinghouse-supplied turbine systems. Several time consuming modifications were necessary to adjust the turbine system to the NSSS. The selection of the turbine manufacturer was driven by the licensee's desire to use a single source of supply for all its nuclear units regardless of NSSS vendor type.
21
5%
Susquehanna 2 experienced three scrams resulting from turbine trips which l occurred during initial startup testing because of a high level in a i moisture separator drain tank. Susquehanna 1 had not experienced this I problem during the early portion o(its operational history, because the {
piping geometry for its moisture separator drain lines is different from j
. that installed at Unit 2. This problem arose with Unit 2 because of a design change which the vendor made to the moisture separator reheaters
- that was outside the licensee's control. Each drain system contains
, three 8-inch check valves, which are lacated about 10 feet upstream of the drain valves. The original desigr ~of the Unit 2 system had specified a single 12-inch check valve located 1;nmediately downstream of the piping tee junction that branches to the ene"gency dump valve. This location is substantially closer tc the drain tant. The vendor changed the original design f cm one valve to three valve 7 close to the feedwater heaters so that,11/ het feedwater heaters were isolated due to a tube rupture, only one check valve would close due to reverse flow. This would allow the plant to,continus operating with two feedwater strings, and two valves controlli @ tank' level. In the original design, closure of the single chack mlve woolc have isolated all three control valves. In this case, the emsgency dump valve would then be required to control tank level in
> ceder to prevent a high level turbine trip. The purpose of this design modification was to increase the reliability of the moisture separator drain control system and to pro"ide protection for the feedwater heaters.
An investigation and analysis of the root cause of these scrams by the licensee's special task force revealed that the B moisture separator drain tank underwent more severe level transients (including flashing) than the A moisture separator drain tank due to the different piping geometry. In this particular instance, the vendor was aware of the potential for problems which might arise due to the design change.
In order to prevent turbine trips resulting from level control problems in the moisture separator drain tank and the subsequent reactor trips, the licensee for Susquehanna installed a 12-inch check valve in the moisture separator drain tank line, locating the valve close to the drain tank. The licensee also modified the emergency dump valve through the addition of a pneumatic booster. This corrective action resulted in the drain line piping geometry reverting to the design as it had been before the vendor changed it.
As noted above, Byron 1 found that cerbin turbine-relatect settings, based on design predictions, needed adjustment to work properly. Disblo Canyon 2 and Palo Verde 1 and 2 likewise experienced unplanned scrams contrary to the pre-dictions of design analysis. The Diablo Canyon 2 experience was unique in our review in that the unplanned scrams originated in the context of planned startup testing.
The Diablo Canyon 2 plant staff indicated that the plant experienced difficulty during attempts to complete full load rejection testing. They attributed six reactor trips to the attempts to complete this test on Unit 2.
s The test consists of opening the main generator output breakers and l verifying that the generated steam could be successfully routed to the condenser, safety valves, and atmospheric dump valves without causing a reactor trip. The test is intended to demonstrate that a load rejection from 100% power would not necessarily cause a reactor trip. Although f Unit I successfully completed the test with apparent ease, Unit 2 did experience problems due to various reasons.
22
At the Palo Verde units, several reactor trips occurred from power levels between 20% and 40% when a rapid rise in steam generator pressure could not be overcome by modulation of steam bypass valves. An analysis of the system design showed that the failure of one or more of the eight steam bypass valves and procedures (now corrected) prevented the steam bypass control system from handling the transient.
The Diablo Canyon 2 experience discussed above also provided an example of how a change between the first and second units at a site can negate some of the ability to learn from experience even for such closely related plants. The turbine experience at Susquehanna 2 provided a similar example.
The condensate system was not a generic source of difficulty for the plants in this study. However, Waterford 3, with a fairly high scram rate during startup, experienced repetitive equipment difficulties with operation of the condensate system, all during the pre-commercial startup period. Six scrams were initiated in that system. The majority of the problems experienced with the condensate system involved the operation of the condensate polishers.
One isolated scram involved a broken yoke on the condensate minimum recircula-tion valve. One scram occurred when the condensate polisher programmer failed when operators attempted to place a polisher on line. Another scram occurred following a condensate system perturbation which caused a feedwater pump trip on low suction pressure. The licensee suspected that the perturbation was caused by operation of the condensate polishers, and installed a 15-second time delay on the low suction trip logic for the main feedwater pumps. However, about six weeks later, an operator placed a condensate polisher on the line and a low '
pressure feedwater pump trip occurred, resulting in a reactor scram.
The licensee determined that the 15-second time delay in the low pressure feedwater pump trip logic should have prevented the feedwater trip caused by the perturbation, and performed an engineering evaluation of the condensate system. The licensee found that the condensate polisher system bypass valve had a 22-second stroke time. The licensee also fr ad that the condensate polisher flow balance override was not proper 1.y adh.;ted. This prevented the condensate polisher vessel discharge valves from opening during high differential pressure conditions. The itcensee adjusted both the bypass valve stroke time and the flow balance override to prevent further scrams caused by condensate polisher operation.
Equipment failures predominate in the electrical systems and the main generator, and most of ten occur during steady state plant operation.
Equipment commonly involved includes inverters, transformers and motor generator sets located in various support systems in the plant. For example:
To eliminate a source of unplanned scrams, Callaway altered the main and auxiliary transformer low-low level, single element trips to two-out-of-two coincidence logic, and will install relaying to avoid energizing the main generator at standstill.
A problem area related to scrams which was encountered at Diablo Canyon 1 23
consisted of problems with instrument inverters and the associated protec-tion logic. A loss of an inverter caused a trip at power over 35% due to spurious iuss of flow signal. At power over 35%, four loops are needed or a trip occurs. To reduce the number of trips from this source, the licensee replaced the instrument inverters. The instrument logic was also modified to avoid trips.
Some sources of trips are associated with the use of state-of-the-art equip-ment. One source of equipment-caused scrams at Palo Verde 1 consisted of malfunctions in the plant multiplexer system. These malfunctions caused the unwanted actuation of breakers in the plant switchyard, resulting in the loss of offsite power and subsequent RPS actuation (see Section 4.3 for additional discussion of these occurrences). To prevent further RPS actuations of this nature, the licensee hardwired the switchyard breaker control circuits, there-by bypassing the multiplexer system for this function. Palo Verde also experienced problems with the Core Protection Calculator (CPC) System. '
Roughly explained, a drop in pump speed during a fast transfer from the main ,
generator to offsite power caused a low flow trip. This was due to a ;
perceived change in the rate of decrease of pump speed that caused a projected j DNBR trip. Software was needed to fix this problem. Waterford 3 also found ;
some difficulties in using the CPC and related core element assembly ;
calculator (CEAC).
While solid state electronics per se can no longer really be considered state-of-the-art, the lessons that have accompanied introduction of this technology into plant protection systems still have not been consistently applied. AE00 case study C604, issued in December 1986, evaluated the " Effects of Ambient Temperature on Electronic Components in Safety-Related Instrumentation and Control System." One of the events evaluated in the study was an event at McGuire (June 4, 1984) involving the integrated circuit cards in the ,
Westinghouse Process Control System (7300 system). The study found that the McGuire station had experienced printed circuit card failures since 1981 and that the licensee was aware that elevated ambient temperature was one of the contributing causes. A comprehensive plan was initiated at McGuire to address the problem. 4 During its initial startup and early operation, a considerable number of logic card failures occurred at the Callaway plant, some of which led to reactor i trips. The short-tern corrective action taken was to test all appropriate circuit cards and replace any malfunctioning cards. Longer-term action consisted of an independent, in-house assessment of the problem.
Based on the McGuire study, Callaway evaluated the adequacy of ventilation provided for the 7300 cabinets. Temperature profiles were run on selected cabinets, based on heat load, and all were found to be within design-specifications. Consequently, no design changes were recommended. In addition, information provided by Westinghouse revealed that the failure rate at Callaway for 7300 circuit cards reflected the industry average. Callaway did experience unplanned scrams attributable to the failure of circuit cards in the rod control system, which were shown to be temperature-related. A design change that has been implemented to increase cooling to the rod control cabinets appears to have alleviated that problem.
The other Westinghouse plants in this study also experienced problems with this hardware, but not to the same extent as Callaway. Learning from the ,
24
specific experience at Callaway, Byron 1 and Wolf Creek instituted circuit card burn-in testirg early in the plant's test program. This lesson can be extrapolated to all solid state electronics; energizing prior to startup ,
testing should help identify temperature sensitivities and ver.tilation problems before entering actual operation.
3.1.2.2 Personnel Errors and Procedural L9ficiencies Overall, pcrsonnel error and procedural deficiencies resulted in slightly fewer unplanned scrams at new plants than equipment failures. Contributions originate with the plant operators and the plant instrumentation and controls (180)andmaintenancestaff.
Placing equipment in and out of service and/or changing power level (turbine and feedwater evolutions) can result in unplanned scrams. For example: ,
At Waterford 3, personnel error and procedural deficiencies were cited as contributing to the poor experience in placing condensate polishers in operation.
At Byron 1, the turbine steam bypass system has been designed for a 50%
steam bypass capability. However (according to plant staff), because of concerns regarding the reliability of the steam bypass valves, the set-point for reactor trip on turbine trip was set at 10% steam bypass capacity. The prime sensor for this reactor trip is the turbine power signal P-13. Between March 1985 (when initial synchronization with the utility grid was achieved) and September 1985 (when the licensee declared Byron 1 to be in commercial operation), on a number of occasions person-nel errors were made during operation, surveillance, and testing that resulted in the generation of a P-13 signal and a subsequent reactor scram. Specifically, although operators thought that the turbine power was below 10%, and thus a turbine trip would not result in a reactor trip, the indicated power based on first stage turbine impulse pressure was greater than 10%, and a reactor trip followed the turbine trip.
After investigation of these events, the licensee concluded that it was necessary for the operators to acquire additional operational experience with the turbine synchronization procedure. The purpose of this additional hands-on experience was to learn how to balance turbine power correctly with reactor power, while avoiding 15e generation of an RPS trip signal.
At Hope Creek, an unplanned scram occurred while recovery from a feedwater pump trip which occurred during testing at 50% power was underway. The plant staff felt that the operator's training on the plant simulator contributed to the the scram, since the reset scheme on the simulator was different then that on the actual plant. The licensee is l attempting to improve simulator response by encouraging operator feedback and implementing an administrative program to ensure that actual plant design changes and operator feedback are incorporated in the simulator's modeling.
The most prominent and generic evolution with operator error scram potential is probably PWR steam generator level control, especially prior to achieving moderate power. Much experience is necessary to manually control level in a j multi-loop system through control of pump speed and valve position. The i
25
lead / lag control demands accompanied by instrumentation problems present a formidable challenge for a less experienced operator. As a result, either a great deal of operator skill must be demonstrated, or the system needs to be automated in order to mininize level variations that result in mitigating action, i.e. , reactor scrams.
The startup configurations of PWR feedwater control vary from plant to plant.
The various system designs and training are both underlying contributors to the problem. Some plants were built with no feedwater bypass control system, but later added a manual feedwater bypass control system. In some cases, the manual system may be used with the auxiliary feedwater system in automatic control for startup. Others were constructed with automatic feedwater control, via auto bypass control, for startup. Many of these systems require some manual actions during power ascension. Others have instrumentation and control characteristics that demand experience from operators.
At Waterford 3, the licensee staff attributed three scrams during the startup period to operator error while in manual feedwater flow control during low power operations (less than 1S percent power). In the early ,
months of commercial operation, two additional trips were caused by operator error in controlling feedwater flow during low power operations.
Plant staff at Callaway, Wolf Creek, and Palo Verde commonly cited problems in which the simulators did not accurately model the actual plant response during low power and startup operations, a known inadequacy in the simulator regarding routine operation. This hindered the licensee's ability to train the operators to respond appropriately to prevent trips during feedwater flow evolutions under these conditions.
Specific modeling deficiencies at Callaway and Wolf Creek involved the steam generator (SG) shrink and swell phenomena.
Diablo Canyon plant staff also felt that a significant number of reactor scrams at both units were the result of cuitrol of feedwater flow during startup and low power operations.
Staff for St. Lucie 2 also found ihat manual control of steam generator water level at low power levels during startup and low power operations was a major problem that resultM in scrams during initial operation.
Although the plant's steam generators have a large secondary side volume, the narrow operating range at St. Lucie between the steam generator low level reactor trip setpoint and the high level reactor trip setpoint does not necessarily provide a longer time to act before one of the level trip setpoints is reached, relative to the time available in other PWR designs.
Because problems with feedwater flow control are relatively common, new plants have typically had to seek improvements involving hardware changes and operator training. For example:
At Waterford 3, the operator training program has been modified to emphasize control of the steam generator during low power and startup cperation. Further, the operating band at which the water level will be maintained has been enlarged to increase the amount of time the operators have for responding to level changes.
26
Callaway, Wolf Creek and Palo Verde are developing improvements in the simulator response for the feedwater control system, e.g., steam generator shrink and swell characteristics, used in the operator training j program. Further, efforts are underway by the Westinghouse Owners Group {
to determine possible feedwater control system modifications to improve control under low power conditions.
The Diablo Canyon 1 and 2 licensee has installed feedwater bypass valves and is considering automatic ccatrol of the bypass valves to further reduce control problems during startup and low power operations. The plant staff also expressed interest in a Westinghouse-proposed modification of the steam generator low-low level trip setpoint for Westinghouse PWRs from the current value (15 percent) to the 5 percent level. Roughly, they estimated that changing the trip setpoint may have the potential to eliminate about one-third of the trips for a new plant and about one-fifth of the trips for an older plant. However, much more plant-specific evaluation is required.
St. Lucie 2 has requested that the NRC approve a proposed change in the steam generator low level trip setpoint from 39.5 percent to 20.5 percent. Additionally, the licensee plans to install an automatic (digital) system to control feedwater between the 2 percent and 15 percent levels (planned for installation beginning with Unit 1 in March 1987).
At Palo Verde 1 and 2, feedwater is directed into the steam generator economizer at 15 percent and higher power levels. At power levels below 15 percent, the feedwater is directed into the steam generator downcomer.
Because the manual changeover required at 15 percent power has been a source of feedwater transients and reactor trips, a modification is being pursued which would make the changeover automatic and minimize the potential for operator errors.
This previous experience highlights the need for: comprehensive operator training programs, accurate modeling in the simulator of the steam generator and feedwater system characteristics, low level set points providing maximum time available for operator response, and consideration of automatic control systems below 15 percent.
I&C technicians and maintenance mechanics are also responsible for unplanned scrams through surveillance testing, calibration, troubleshooting, and main-tenance. These personnel and the procedures used are responsible for roughly one in every 12 unplanned scrams. Human errors while testing, calibrating, maintaining, and troubleshooting comprise 50% of the spurious scrams initiated by the RPS. Equipment problems account for only 27%, and most of the equipment problems surface while testing is in progress. The problems stem principally from inexperience, not following procedures, or lack of coordination among plant staff.
l All licensees appeared to be concerned about human error-caused reactor trips, l regardless of the magnitude of the contribution to the overall scram rate. As will be brought out further in the discussion of ESF actuations and technical specification (TS) violations, the competence of the I&C technicians is a major factor in many areas of operations. Some licensees have established preemptive or remedial programs which share the following attributes:
27
1 They give high visibility to the source (i.e., organizational element) of unplanned scrams caused by personnel errur-usually within the context of a plant-wide quality improvement program.
They emphasize the need for good communication among the plant staff and an awareness of how testing activities can result in plant scrams, sometimes through group review of unplanned scrams.
They provide for improvement in the testing process'through improving technician skills, improving procedures, providing engineering expertise within the I&C organization, or assigning responsibility for specific equipment.
In a related matter, the BWR Owners Group has proposed that certain testing l requirements be relaxed based on risk analysis cons iderations.1/ One.of the i potential effects of such changes would be a reduction in the number of unplanned scrums and ESF actuations caused by testing activities. However, the NPC staff's review of these proposed changes has not yet been completed.
Therefore, further evaluation of the proposal is necessary before its acceptability can be determined.
Training in the area of surveillance testing is a major difference that seems to underscore good performance in this area. At Diablo Canyon 1, crews were put through simulator testing for familiarity with surveillance testing. I&C technicians were trained for months on a live RPS and surveillance mock-ups, since time was available. The utility resources and, more importantly, available time before operation due to other delays, afforded them the ability to install good training facilities and complete training. Another utility visited had planned for extensive training by I&C technicians. However, schedule delays in systems turnover and personnel availability resulted in less than optimum in-plant training. Planned training faci?ity equipment was not in place in time for training prior to operations. In reviewing the data from this plant, a high percentage of trips due to I&C personnel error was noted.
For new plants, many trips are a simple result of optimization of the control systems for balance-of-plant (B0P) systems such as feedwater control and steam dump control. However, simulator training and improved dry runs, along with more intensive and experienced vendor support, may help. This in-plant familiarization and hands-on training for the optimization of such control systems appeared as a common thread in the plants with a low number of trips involving B0P control system optimization.
3.1.3 Findings The findings stated below are based on the analyses of operational data for new plants regarding unplanned scrams.
(1) The scram rate for new plants greatly exceeds that of mature (those plants which were licensed oefore January 1,1983) plants. Viewed as a class, new may be characterized by the following:
1/ " Tech Spec Improvement Analyses for BWR Reactor Protection System,"
NEDC-30851P, DRF A00-02119-A, Class 3, May 1985, Proprietary report, Limited distribution.
28
Pre-commercial operation total scram rates are about five times the mature plant rate on the average, and, as can be seen in Section 4, the frequency of scrams with complications (i.e., scrams which had additional failures or personnel errors beyond those which initiated the scram) is about a factor of four times that of a mature plant.
However, these scrams are from lower power levels.
Post-commercial total scram rates and complicated scram rates can be a factor of two to three higher than the mature plant rate (during the first few months). These scrams are from higher power levels, which is similar to a mature plant. A positive correlation did exist between the startup and early post-commercial average scram rates.
The scram rate during startup was a statistically significant indicator of the scram rate that will be experienced during early post-commercial operation.
(2) Analysis for potential correlations of the pre- and post-commercial scram rates with various factors indicated no statistically significant corre-lation between the NSSS vendor or the type of A/E and the scram rate.
Nor were there strong correlations between the scram rates and the length of the startup period, licensee prior experience or the presence of an older unit at the site found in this analysis.
(3) On a percentage basis, the causes of scrams at new plants are very similar to those for mature plants. The primary causes are associated with BOP systems, with the feedwater system dominating. There are some differences in the cause profiles between new plants and mature plants as follows:
{
During pre-commercial operation, testing (primarily surveillance testing) contributes more to new plant scrams than in the mature plants by a factor of three (32% vs. 11%). Also during this period, procedural deficiencies that caused scrams are higher by a factor of 2 than for the case of mature plants (11.5% vs. 6%).
No single cause category appears to be driving the scram rates at new plants. For example, oigh scram frequencies do not appear to be driven by personnel error.
3.1.4 Lessons Learned Based on the lessons learned from the experience of new plants that focus on the reduction of unplanned scrams, the following improvement lessons were developed.
(1) Focus on the B0P prior to operation and early in life appears to provide a high return regarding the reduction of unplanned scrams.
Within this area, attention could be given to:
Conducting additional reviews, prior to fuel load, of feedwater and turbine control and bypass systems to identify sensitivities and plant-specific characteristics that cc,ald contribute to transients or the ability of the system to cope with expected transients.
29
Conducting a systematic review of equipment protective logics and setpoints on components such as pumps (suction trip time delay, vibration trip) or power supplies to identify areas where a time delay or additional channels for coincidence could reduce the potential for unnecessary transients or spurious actuations. Give special attention to first-of-a-kind features not incorporated in earlier designs. Additional suggestions obtained from .the plants visited include the main steam reheater drain high level trip and other turbine protective trips.
Emphasize training for routine operations involving power level changes and the associated communications among shift personnel (i.e., feed flow and turbine evolutions) that have historically caused trips. Accelerated programs / efforts appear appropriate for newly licensed PWRs regarding steam generator level control.
Emphasize the need for site-specific simulators to include, prior to startup, the best achievable fidelity of the simulator to the plant ,
regarding feedwater effects (lead / lag characteristics of level i indication and control methods), and include provisions to continue to improve fidelity as the startup progresses.
(2) For new plants on a priority basis, expedite the implementation of vendor or licensee trip reduction measures. Trip reduction programs should focus on safety-related equipment as well as on B0P equipment.
(3) More complete testing of new or unique plant features, such as new RPS systems, electrical systems, etc., prior to fuel load would likely reduce unanticipated failures or unexpected erratic behavior. Emphasize planning for test objectives to reduce the frequency of unplanned scrams and unnecessary ESF actuations.
(4) Establish as a major goal an increased commitment to training in performing surveillance testing, calibration and troubleshooting activities well prior to operations. I&C training '.nitiatives, such as repeated practice for those surveillance activities that could cause a transient and which should be performed on live systems prior to operations, should be emphasized. An action to improve surveillance testing suggested by licensee staffs at the plants visited wa:, training for I&C personnel in valving instrumentation in and out of service.
(5) Some ways in which administrative control of surveillance may be improved are as follows:
Organize the I&C staff to establish accountability for specific equipment.
Blend engineering staff into the I&C organization.
Surveillance should be flagged, categorized, and scheduled according to risk of scrams.
Implement schemes to separate channel testirg, such as a specific day of the week assigned to work on each channel. Also, ;
notification of the channel in test could be posted on control room j panels. q 30 l.
Considering that some new plants have experienced problems when work has been performed in the vicinity of instrument racks during plant operation, licensees should evaluate the impact of work activities during operation that could adversely affect plant operation, and take appropriate administrative actions to minimize the potential for this problem to occur.
Develop a color coding scheme for single point scram components whose misoperation could cause a scram (for example, pressure sensing lines).
Install test jacks and bypass switches at appropriate points in actuation circuitry.
Install cages or covers over switches or racks that could provide trip signals.
Give high visibility to the source (i.e., organizational element) of unplanned scrams caused by human error and establish performance goals.
(6) For future designs or major plant modifications, preference for proven designs and standardization of design in plant feedwater and turbine systems appears justified.
3.1.5 Perspective From Japanese Experience i
A comparison 1/ between the startup programs and regulatory approach indicates that the contrast between Japanese ano U.S. experience in this area can be attributed to the following:
) (1) The Japanese regulatory emphasis placed on producing a safe and reliable I uninterrupted electrical power supply appears to be a significant contri-butor to their minimal scram /high power production record in early opera-tion. Power conversion systems appear to receive the same level of attention to quality / reliability as do the NSSS systems / components.
Single failure philosophy is utilized in some BOP systems. For example, the feedwater system is provided with a third, fast start, electric standby pump that automatically initiates.
(2) The requirements to obtain a detailed design approval prior to starting construction are unique to the Japanese program. This approval may be applied in stages as construction progresses. Approval is then accom-plished utilizing an accumulation of " inspections before use," whereby licensing approval inspections are required upon completion of every work stage. This process ensures that the actual construction stages are carried out as planned during the regulatory planning stage and conform to the Technical Standards set forth by the Ministry of International Trade and Industry. The accumulation of " inspections before use" is then considered as the operating license.
(3) The strong reliance on preventive maintenance, root cause determinations, and adeouate corrective actions appears to be a principal Japanese relia-bility factor (e.g., corrective actions go as far as replacing the failed component and all other like non-failed components with new cnes in that 1/ Private communication.
31
unit prior to continued operations / construction, and in all other units during their next annual inspection).
(4) The practice of not allowing construction (sometimes critical path) to continue until all event corrective actions have been satisfied, along with the resultant schedular delays, appears to provide a strong incentive for utilities to ensure quality construction and operation.
3.2 Engineered Safety Feature Actuation Experience All licensed commercial nuclear power plants in the United States contain systems which are designed to control and mitigate occurrences that might challenge the integrity of the reactor plant or adversely affect plant personnel or the general populace. Generally known as engineered safety features (ESFs), these systems include those designed to control reactor core reactivity, isolate and cool containment, supply emergency cooling to the reactor fuel, remove residual decay heat, provide emergency power, assure habitability of the control room, and control radioactivity releases to the environment. ESFs are listed and described in Chapter 6 of each plant's FSAR.
The following discussion and statistics pertain to all ESFs except the reactor protection system (RPS), which is treated in the context of scram analysis.
ESFs vary in number, complexity, design and nomenclature from plant to plant.
This is especially true for those systems designed for containment isolation, control room habitability and the control of radioactivity releases.
The data on ESF actuations for January 1,1984 through June 30, 1986 are derived from LERs required by 10 CFR 50.73 after January 1,1984. The 1983 data (for Diablo Canyon 1, LaSalle 2, McGuire 2, St. Lucie 2, and WNP-2) were i
taken from computer logs of 4-hour telephone notifications to the NRC Operations Center as required by 10 CFR 50.72. Because of the preliminary nature of the telephone notification, cause information is often not reliable, and thus no cause category assignments were mde for ESF actuations which occurred in 1983.
The data used in this analysis included actuations of any ESF systems, other than the RPS, which occurred between OL issuance and June 30, 1986. It is recognized that the actuation of auxiliary feedwater (AFW) following a reactor trip at a Westinghouse PWR is considered a design basis event consistent with the system design function. For this reason, this study included actuations of AFW at Westinghouse plants only when the licensee specifically stated that the AFW actuation occurred in response to an ESF actuation signal. Therefore, actuations of the Westinghouse AFW system following a reactor scram were not included in the data base.
Previous detailed analysis of industry experience with ESF actuations has l shown that they rarely occur in response to an actual event, for which the feature was designed. Most actuations are thus unnecessary. This is also the case for the new plants in this study. Thus, ESF actuation reports do nct represent a set of inherently safety significant events. Rather, they repre-sent an accumulation of events to which plant staff (operators, technicians, maintenance personnel) must respond. The possible safety issues associated with unnecessary ESF actuations stem from: (1) unwanted challenges to safety !
systems, (2) the burden and distraction placed on a plant staff to respond to, !
troubleshoot, repair, and report them, (3) the potential for an inadequate i response to a real event due to the preponderance of false alarms, and (4) the '
i 32
possible erosion of a positive attitude toward safety requirements and safety equipment due to the preponderance of false alarms.
3.2.1 Data Analysis Figure 10 provides a composite ESF " learning curve" for all of the plants in j the study. It plots cumulative ESF actuations as a function of calendar month j since initial operating license. A line has been added to the plot with a j slope of one per month, which is the average ESF actuation rate for the 76 older plants operating in 1985. Two of the plants in the study (McGuire 2 and St. Lucie 2) had ESF actuation rates throughout their operation that were lower than the average rates for mature plants. A second group of plants had an actuation rate that was almost equal to the average rate for mature plants ie.g., Susquehanna 2, Diablo Canyon 1 and 2, Palo Verde 2). The remainder of the plants had relatively high actuation rates between the initial licensing date and about eight months post-licensing. Most of this latter group showed large decreases in the ESF actuation rate beginning at three to eight months post-OL issuance.
i U N P LAN N ED ESF ACTUATIONS {
l 1985 Average Rate - Mature Plant Slope ,
no . y
- ** + ~'
~. . - ,/ _....+
no- )/ .
,7 .......+~ , _, , _
soo. ,.- / , + -
o CNSt g PNP 1 j g-~
eo. ...- ? ./
_/ ~
- cNs2 Rest g ,. , , ,*. ' g + ocPi . .x . .. si o
,g
,o.
,e oo-
./'
/
/s
/ _
)
lw ,
. cm o sc=2
\
l
.s ,
. .. e#
- f. . . . .'. ~. . _, . _,
4o. '/ / .+
y ---- _ == . NP2 o..o..a
- y - -
n f l -
tly /*f.g,,," f
,o. i o wo*2 G
-[ m -g-g-g- m g g fr- M ^ ""*'
O 2 4 6 S 10 12 14 16 18 20 22 24 Months Since OL lssuance Figure 10 A statistical analysis was performed using the data in Table 3. Average ESF actuation rates were calculated for the pre-comercial period for those plants 33
which became comercial (i.e., completed the power ascension test program) prior to June 1986. In addition, the average post-comercial actuation rates were calculated for the first 180 days (six months) after the date of comercial operation, if the plant had been in operation for that period prior to June 1986. The data shown in Table 3 do not provide a basis for ranking plant performance.
Table 3 Average ESF Actuation Rates Pre-Commercial Post-Commercial Plant Actuations/ Month Actuations/ Month McGuire 2 0.15 0.33 3fablo Canyon 1 1.38 0.99 Callaway 7.00 4.17 Catawba 1 3.30 4.17 Byron 1 6.08 3.67 Wolf Creek 13.30 4.32 Diablo Canyon 2 1.89 N/A, Millstone 3 2.44 N/A St. Lucie 2 0.24 0.00 Waterford 3.21 2.34 l Palo Verde 1 3.25 N/A LaSalle 2 4.59 5.67 WNp-2 7.02 3.51 Susequehanna 2 1.47 2.4S Limerick 1 7.14 N/A, River Bend 6.90 N/A ;
Average 4.33 2.87 Note: These plants had not accumulated 180 days of commercial '
operation before June 30, 1986. Therefore, their post-commercial experience was not considered in the statistical analysis.
The same correlations were tested as for the scram data (see page 12). The analysis indicated with statistical significance that the higher the pre-commercial ESF actuation rate, the higher the actuation rate during the first six mcnths of the post-commercial period. None of the other correlations was statistically significant, including the correlation with NSSS design. The following came closest to being significant:
Lower ESF actuation rates during the pre-comercial period of operation occurred more often at plants owned by utilities with previous nuclear experience.
' Lower ESF actuation rates during startup occurred more often at plants which were the second unit at a multi-unit site.
3.2.2 Cause Analysis As with the scram data, examination of the cause category data shown in 34
Appendix F revealed no dominance of a particular cause category or a corre-lation between a cause category such as human error and higher actuation rates. The underlying reasons for the observed actuation frequencies were pursued through more detailed review of information provided by the licensees, either through LERs or during discussions with plant staff at selected sites.
Recurring causes or causes common to a number of new plants were identified and are summarized below.
3.2.2.1 Equipment-Related Causes Radiation Monitoring Systems / Subsystems. Spurious operation of the radiation monitoring system contributes significantly to elevated ESF actuation fre-quencies. This source appears to be a factor across the range of plants in this study, contributing in proportion to the overall actuation rate for each plant. This system is installed throughout each plant, and its signals lead to realignment of heating, ventilation and air conditioning (HVAC) systems j from their normal operating mode to the emergency (ESF) mode. The highest j actuation rates were associated with first-of-a-kind or state-of-the-art i radiation monitoring systems. Detailed operating experience with this system at Byron 1, Callaway, Wolf Creek, WNP-2, Palo Verde 1 and 2, Hope Creek and Susquehanna 2 is provided in Appendix E. )
The equipment-related reasons for the relatively high ESF actuation rates due to spurious radiation monitoring actuations can be categorized (although in varying proportions from plant-to-plant) as follows: l I
(1) Field application of state of the art technology (high sensitivity).
(2) Ambitious design requirements.
(3) Inadequate vendor support.
(4) Non-coincidence actuation logic.
(5) Vacuum transducer unreliability.
(6) Electronic noise induced during maintenance / construction or normal l operation. !
(7) Radiation pickup from unintended sources.
The first three reasons result in hardware / software incompatibility in the original design, which requires more time to correct than more simple hardware changes.
In response to the equipment problems identified above, licensees have imple-mented or are considering the following corrective actions to reduce the frequency of ESF actuations initiated by the radiation monitoring system:
(1) Modifications to the radiation monitoring system hardware, such as shielding the detector cables from radiation and electromagnetic inter-ference (EMI), replacing the originally installed vacuum transducers with a more reliable design, and replacing faulty electronics modules and photocells.
35
{
(2) Modifications to the radiation monitoring system logic, such as changing l the non-coincidence logic to a coincidence logic, so that activation of )
more than one channel of instrumentation is required to actuate the '
realignment of HVAC systems to the ESF mode.
(3) Modifications to the radiation monitoring system software to remove the l incompatibility between it and the system hardware. l l
(4) Administrative controls placed on any ESF equipment located in the vicinity of work evolutions, such as requiring that ESF monitors be placed in ESF bypass during performance of work activities.
Chlorine Monitoring System. The control room ventilation system shifts from a normal operating mode to an isolated recirculation ESF or emergency mode on detection of hazardous gases in the local atmosphere. Chlorine may be one of the gases monitored if it is used (e.g., as part of the circulating water treatment process) or stored on the site. Like the radiation monitoring system, the chlorine monitoring system design is tailored for each plant.
This system which actuates the control room ventilation isolation system (CRVIS) has been the major contributor of spurious ESF actuations during early operation at some plants. At the Wolf Creek plant, by far the most persistent problem with the chlorine monitoring system was the chlorine-sensitive paper tape which serves as the detection mechanism for the chlorine detectors.
Among the problems experienced with the paper tape are: tape depletion, tape breakage, and tape sticking in the tape transport. The tape transport itself has also malfunctioned, which in turn has caused some actuations of the CRVIS.
Other problems experienced with the chlorine monitoring system have consisted of burned-out light bulbs in the monitor analysis unit and power supply failures. In response to the recurring number of CRVIS actuations initiated by this system, the licensee has (1) increased the frequency of paper tape replacement, (2) replaced failed or malfunctioning system components, and (3) established a surveillance testing program for the chlorine monitoring system.
However, in spite of all the corrective actions taken by the licensee to reduce the frequency of CRVIS actuations due to the chlorine monitoring system, such actuations continue to occur spuriously at Wolf Creek. Currently, they are considering a measure that would result in elimination of the requirement to monitor for the presence of chlorine gas on-site. This action would consist of a design change to switch the circulating water treatment from a gaseous chlorination process to a process of a different type. If the licensee implements this change, all chlorine gas can be removed from the site entirely.
At the WNP-2 plant, the chlorine monitoring system was also the source of a significant number of control room emergency filtration system (CREFS) actua-tions (an ESF at WHP-2). The problems which WNP-2 experienced with the chlorine monitoring system were mostly the same type as those experienced at the Wolf Creek plant. WNP-2 did report one additional type of problem with the chlorine monitoring system: malfunctioning photocells in the chlorine detectors. After several corrective actions taken by the licensee failed to significantly reduce the frequency of spurious CREFS actuations due to the chlorine monitoring system, the licensee cnanged the circulation water treatment from gaseous chlorination to chemical addition of sodium hypo-chlorite. In this manner, all chlorine gas was removed from the WNP-2 site and this source of CREFS actuations was eliminated.
36
In addition to Wolf Creek and WNP-2, Limerick also experienced a significant number of spurious ESF actuations because of problems with the chlorine monitoring system.
Boron Dilution Protection System. Byron 1 experienced its peak ESF actuation frequency in the first two months after receiving its operating license. The l plant was in fuel loading or cold shutdown prior to initial criticality at this time. Of a total of 37 ESF actuations, 22 were HVAC realignments (mcstly involving the control room ventilation system triggered by radiation monitor (
inputs as discussed above) and 14 were actuations of the boron dilution pro- '
tectionsystem(BDPS). The use of the BDPS as an ESF is unique to Byron 1 among the 22 plants in the study.
The BDPS is part of the NSSS design furnished by Westinghouse on recent plants. Its purpose is to limit boron dilution in the reactor core region during periods when the reactor is shut down. It accomplishes this by changing the charging pump suction from the volume control tank (VCT) to the refueling water storage tank. In the case of Byron 1, this system is included as an ESF in the FSAR. A similar system was installed at Callaway, but is not classified as an ESF. Similarly, a study performed for Wolf Creek during its relicensing review indicated that this system was not required to mitigate a boron dilution accident, hence it is not classified as an ESF. While the licensee has not requested the removal of the BDPS from the plant technical specifications, permission to report only valid actuations was requested from NRC and granted in April 1985 (the sixth month post-0L).
The parameter which is used to actuate the BDPS is the neutron flux monitored by the source range monitoring system. The licensee's investigation of the numerous BDPS actuations at Byron 1 revealed that the source range monitors were extremely sensitive to EMI. As a preventive measure to reduce the fre-quency of such actuations, the licensee changed the appropriate source range monitor cabling to a type which is much less sensitive to EMI.
Reactor Water Cleanup (RWCU) System Isolation. Isolation of the BWR reactor iTa~ter cleanup (RWCU) system was a major rource of ESF actuations at LaSalle 2, WNP-2, Limerick, Fermi 2 and River Bend. During discussions at the site, the Hope Creek licensee also cited the RWCU as an early problem area. The RWCU is a filtration and ion exchange system for maintaining the purity of the water in the reactor vessel. While the RWCU serves no safety function, automatic isolation of the system is necessary to maintain primary containment integrity and to ensure proper operation of the standby liquid control system. The isolation valves receive automatic closure signals from parameters monitored within the RWCU system. These signals derive from various leak detection schemes and include differential flow, differential temperature, and high temperature.
One factor which plays a role in the disparity between the frequency of such actuations at LaSalle 2, WNP-2, and River Bend on the one hand, and Susque-hanna 2 on the other, is a change in design of the actuation logic.
In the former case, a spurious momentary signal in a single channel will result in an actuation, whereas in the latter case, a one-out-of-two taken i' twice coincidence logic 'liminates actuation for that reason. !
LaSalle 2 has possibly had the most difficulty with this type of actuation. l 37
)
l The majority of these actuations have been due to differential flow indica-tions. These were due in part to an incorrect lift point setting procedure for the relief valves on the regenerative heat exchanger. Premature lif ting j was causing the delta flow isolations. The licensee has also cited noise from logic relays, overly sensitive temperature sensing / trip modules (Riley) and power supply overheating as sources of spurious isolations. Lastly, a design change which lowered the temperature of the water in the RCWU pump made the ambient and differential temperature actuations ineffective for detecting leakage. However, since this caused the actuation setpoints to approach the normal operating conditions, a number of spurious isolations were generated.
A technical specification change approved January 8,1985 for LaSalle 2 eliminated this source of spurious isolations. RWCU isolations had been almost completely eliminated about 12 months after issuance of the OL.
At WNP-2, recurring RWCU isolations were attributed to the sensitivity of the Riley Podel 86 thermocouple monitor,1/ and happened mostly during surveillance testing of these systems! They occurred when the operator turned the thermocouple monitor switch to the " READ" position. The sensitivity of this instrument is so high that the electrical noise created by the turning of the switch can cause an actuation signal to be generated. WNP-2 has experienced leakage from pump seals. WNP-2 also experienced a number of recurring RWCU isolations due to erroneous high delta-flow signals caused by air intrusion into common instrument sensing lines. The air intrusion was determined to be due to the physical location of the flow instrumentation.
During the 1986 refueling outage, the licensee relocated the flow instru-mentation below the process stream to prevent air entrapment. No further RWCU isolations of this nature have occurred at WNP-2 since the design change was implemented.
The Hope Creek staff stated that the plant has experienced a number of isola-tions of the RWCU system. The licensee cited design (test points difficult to access), overly conservative setpoints on the RWCU density compensator (corrective action consisted of changes to the setpoints) and RWCU pump seal leakage as the major causes for problems. It is interesting to note that one of the oldest BWRs in this study, WNP-2, and the newest BWR, Hope Creek, both have experienced problems with leaking RWCU pump seals. Hope Creek is in the process of procuring sealless RWCU pumps (without mechanical seals).
Steam Line Break Protection System. At both Diablo Canyon 1 and 2, the steam line break protection system has been a source of recurring ESF actuations.
This system is an older Westinghouse design, and actuates on a high steam flow plus low main steam line pressure or low-low T-ave. When actuated, the system isolates the main steam lines, and generates a safety injection signal and an 1/ This problem was common to other BWRs besides LaSalle 2 and WNP-2. The NRC has performed an evaluation of this generic problem. The results of this evaluation have been documented in the report AE00/E604, issued in December 1986. In addition, the NRC's Office of Inspection and Enforce-ment issued Information Notice No. 86-69 (August 18,1986) to address this problem. This information notice describes corrective action taken by one BWR licensee to solve this problem. This corrective action con-sisted of installing a 1-second time delay in the detection circuitrj tc eliminate spurious isolations, while allowing an actual alarm condition to initiate isolation as designed.
38 l
associated reactor trip. The lead / lag characteristics of the system are very sensitive. Currently, they have problems with a suspected oscillating bistable and a flow transmitter drifting high, such that the low steam line pressure signal may occur spuriously. In order to eliminate the ESF actuations caused by equipment problems with this system, the licensee was monitoring the equipment and also considering a license amendment to admini-stratively resolve the problem.
Balance of Plant ESF Actuation System. The licensee identified the balance- I of-plant engineered safety features actuation system (B0P ESFAS) at Palo Verde !
1 and 2 as a continuing source of operational problems and unplanned ESF l actuations. One of the problems encountered with this system was poor conrector pin contact, which resulted from an inadequate design review and pin selection by the vendor. The licensee addressed this problem by replacing all the connector pins in Units 1 and 2.
Another problem experienced with the BOP ESFAS has been circuit card malfunc-tioning due to inadequate ventilation of the ESFAS logic and control circuitry cabinets. This type of problem has caused a number of complex events in-volving the ESFs at Palo Verce 1. In order to prevent recurrence of events arising from cabinet temperatures which exceed the design value, the licensee has provided additional ventilation for these cabinets. In addition, the licensee has installed high temperature alarms to notify plant personnel if the cabinet temperatures reach undesired levels.
3.2.2.2 Personnel Errors and Procedural Deficiencies Of the ESF actuations reviewed in this study, about one-third were attributed to human error or inadequate procedures. The majority of the human errors were committed by I&C technicians performing surveillance. Every licensee contacted cited the need to reduce this source of spurious actuations, and had focused certain actions to bring about such a reduction. Corrective actions have included: setting goals and tracking spurious actuations (usually within the context of a large plant-wide or utility-wide quality improvement program),
group reviewing of unplanned actuations with I&C personnel, reorganizing the I&C function to assign responsibility for specific equipment; and adding engineering staff to the I&C organization.
The ability to perform " dry run" surveillance on live equipment prior to licensing was cited by plants with a low frequency of ESF actuations as a factor in avoiding spurious actuations. Likewise, the inability or lack of opportunity to do so was a factor at plants with a higher frequency. This is one area where schedule pressures translate directly into early post-OL diffi-culties. Pre-licensing testing provides experience under simulated licensed conditions, allows thorough review of surveillance procedures and technical specifications, and instills an operating plant mentality.
The lessons for reducing ESF actuations due to personnel error are generally the same as for scram reduction.
3.2.3 Findings The findings stated below are based on the analyses of operational data for new plants regarding unplanned ESF actuations.
(1) New plants show a wide range of unplanned ESF actuation rates, with some 39
plants equal to or lower than the mature plant average. However the majority of the new plants exhibited ESF actuation rates which averaged about four and one-half times higher during startup than the average rate for mature plants. The plants with initially high rates generally experienced large decreases beginning between three and eight months post-OL issuance.
(2) Statistical correlations of the pre- and post-commercial ESF actuation rates with various factors indicated that the higher the pre-commercial actuation rate, the higher the actuation rate during the first six months of the post-commercial period. There was a tendency toward lower ESF actuation rates during startup for second units at a site and for units operated by utilities with previous nuclear experience.
(3) As with the scram data, exutination of the cause category (design, equipment, human error, procedures, other, and unknown) distribution for the ESF actuations revealed no dominant category or correlation between a cause category, such as human error, and higher actuation rates. Of the ESF actuations reviewed, about one-third were attributed to human error or inadequate procedures. The majority of the human errors were comitted by I&C technicians performing surveillance.
(4) High ESF actuation rates at newly licensed plants result from:
Unique or new design features that have not been tested previously (e.g., first of a kind, one of a kind, and state-of-the-art features).
Actuation setpoints and logics which are too conservative as designed and not suited to the operational environment.
Difficulty in performing surveillance testing, calibration, troubleshooting, and maintenance.
(5) Although ESFs have a wide diversity in design features, there are opportunities for new plants to benefit from ESF operating experience at other facilities. Examples include radiation monitoring, chlorine monitorina, and RWCU systems and emergency ac diesel-generators.
3.2.4 Lessons learned ised on the lessons leerned from the experience of new plants that focus on
.he reduction of unplanned ESF actuations, the following improvement lessons were developed.
(1) Be aware of any first of a kind, one of a kind, and state-of-the-art features, since they generate a large number of problems during plant startup. As a further example, unique MSIV b611 valves at Nine Mile Point ? had to be replaced recently with typical BWR "Y" pattern globe l
40
valves. Anticipate problems with such features and take appropriate measures, such as more extensive preoperational testing. A reexamination of actuation logic to better achieve reliable indication and actuation may also be bene-ficial, for example, eliminate actuation en a single input or loss of a single input.
(2) Ensure that operating experience feedback programs: (a) combine internal events and relevant events from similar plants, (b) communicate them directly to the appropriate first level supervisors and working level staff at the plant prior to startup, and (c) address preventive measures.
For example, segregate the trip and ESF actuation data involving human errors from recent plant startups into the specific positions, organi-zational, or functional element, working activity, systems and components, time of day, etc. Feed it back at the lowest levels so that the experience of others, the complexity of what is being done, and the ramifications of errors can be seen and appreciated.
(3) Pay attention to the design ard installation of equipment located in the vicinity of radiation monitors and associated cabling to ensure that ahquate grounding of equipment, cable shielding, etc., is provided to prevent the occurrence of EMI, which can trigger this extremely sensitive instrumentation.
(4) Highlight, through discussions with licensees, systems that historically cause, or can cause, a high number of challenges to safety systems.
(5) The measures previously listed for scram reduction in Section 3.1.4 should also be beneficial in reducing the number of spurious ESF actuations caused by personnel error.
3.3 Technical Specification Violation Experience While often highly technical and of low individual immediate safety significance, violations of technical specifications (TSs) have a high regulatory visibility. The primary responsibility for safe plant operation rests with the licensee. Violations of the facility TSs are usually due to personnel related issues and may reflect an inability of the licensee to manage the plant safely, and therefore, violations of TSs can be important to the NRC perception of licensee competence.
3.3.1 Data Analysis Data were assembled from three sources: 1983 data came from a review of LERs that separated violations from non-violations in instances of missed surveil-lance, entry into an LCO, or other TS-related situations under pre-1984 reporting requirements as contained in TSs; 1984 data from a review of LERs reporting TS violations 1/ performed for NRR as part of the Technical Specification Improvement Program; and 1985 to June 1985 data from a review of LERs performed by AE00. Cause categories were assigned only for the LERs i
{
reviewed by AE00. While this causal information is incomplete, the nature of i 1/ 10 CFR 50.73 also requires reporting plant shutdowns required by technical specifications. Since these do not constitute violations, they are not included.
41
l TSs is such that one would c>., . i.t most violations to be the result of personnel errors or problems with procedures, and this has been borne out in the general review of 1985 data for all licensed plants; the cause information shown for the plants in this study also supports the predominance of these causal factors (see Appendix Fj.
Figure 'll presents cumulative violations vs. calendar month. As a point of reference, the straicht line represents a rate of 0.6 violations per month, tFe average violation rate for the 76 olcer plants in 1985.
TECHNICAL L SPECIFICATION VIOLATIONS so / ~
' LEGEND 1985 Average Rate Mature Plant Slope /
,.. ,J - . , . ,
/ -, .
/
/ 1 o
- " \ / .,,...f-+-+ o e,o, , m, J a cas2 4 nest h / 4 r . . .i - + DCP, e .x. .SNS1
. oc , . .
/
5+
o b i/
f*/ r +-
~
jso- _y d e , o--e -e>-e- . -, . a
/
i s'. e A"j..j ~~ * ## + ***
20- e- ?
$l / ,....n o asi . .c.,
,_ _2. .. g :.o., o ..o.. o .. o .. o .. o .. o
..- ...4 g - gMT. j .x. . . .. . . . . . .. . ... . ... . .. . . .. . . . . . . a was:
, , , 7 x..-
C
. . x . . . x .. .x a u'as
, _f O 2 4 6 8 10 12 14 16 18 20 22 24 Months Since OLlssuance !
Figure 11 Initial rates very widely, but generally exceed the mature plant average for up to 6 to 8 months after OL issuance. Then rates show a significant decrease in about half of the cases. It can also be observed that one plant (Palo ..
Verde 1) did not display a learning curve during the period considered by this study. High rates can persist beyond 12 months after the OL.
The statistical analysis of the average violation rates shown in Table 4 resulted in only one statistically significant correlation: a plant which is the second unit at a site has a lower violation rate during the 6 months immediately after becoming commercial than a plant which is the first unit at the site. The data show a statistically weaker indication for the same factor 42
during the startup period. These results are based on the data for McGuire 2, St. Lucie 2 and Susquehanna 2. The data shown in Table 4 do not provice a basis for ranking plant performance.
Table 4 Average Technical Specification Violation Rates Pre-Commercial Post-Commercial Plant Events / Month Events /Menth McGuire 2 2.55 0.00 Diablo Canyon 1 1.20 2.16 Callaway 1.57 1.83 Catawba 1 2.51 2.50 Byron 1 4.22 1.17 j Wolf Creek 1.88 2.17 i Diablo Canyon 2 1.31 N/A, Millstone 3 1.63 N/A St. Lucie 2 1.45 1.00 Waterford 1.50 1.8}
Palo Verde 1 4.47 N/A LaSalle 2 1.17 1.67 WNP-2 3.26 2.17 Susquehanna 2 1.10 0.5Q Limerick 1 2.85 N/A, l River Bend 2.37 N/A Average 2.19 1,55 Note: These plants had not accumulated 180 days of commercial operation before June 30, 1986. Therefore, their post-commercial experience was not considered in the statistical analysis.
It is reasonable to assume that the frequency of TS violations, as well as that for unplanned scrams and spurious ESF actuations caused by surveillance, is somehow proportional to the product of the number of LCOs or surveillance tests required by the TSs, and the probability of making an error or suffering an equipment failure during a test or failing to meet the LCO.
The number of LCOs and/or surveillance requirements has generally increased throughout the history of licensing, from the era of custom TSs to the use of ,
standard TSs, which were first implemented in 1974 The sources of growth l have been the institution of standard TSs, resolution of generic issues, '
lessons learned from the TMI-2 accident, and implementation of Appendix R to )
10 CFR 50. However, a review performed in early 19851/ indicated that the ]
growth had leveled off after approximately 1980. Thus! within the 22 plants reviewed in this study, steady growth af TSs thrcugh new requirements pcr se should not be a factor from plant tu plant within NSSS vendor type. ;
4 The referenced review did not address variations between NSSS vendors, so we 1/ Response to ( sngressional question, Walsh/NRR,1/9/85.
43 l I
1
f have no quantitative perspective from that, standpoint. The analysis of TS violation rates showed neither a correlation with NSSS nor one with date of OL issuance--only lower TS violation rates for the second unit at a site.
Based on the foregoing discussion and a qualitative reading of the ESF experi-ence, we tentatively conclude that within the 22 plants studied, variations in operating experience are not driven by regulatory changes in the scope of technical specifications, or major differences in quantity of LCOs or related surveillance among NSSS vendors. Some plant-to-plant differences result from differences in the number of plant features designated as ESFs. However, regardless of the initial plant-specific features of the TSs, nearly all plants show evidence of a learning process, a process that is accelerated for the second unit at a site.
Over 70 different systems are involved in TS violations (see Appendix G).
However, fire detection (90 of 658) and radiation monitoring (72 of 658) contain the greatest concentration of events. Repetitive events at Byron '.,
Linerick and Wolf Creek are most responsible for the prominence of the ff re detection system; likewise, Byron 1 Callaway and Palo Verde 1 drive the radiation monitoring statistics.
3.3.2 Cause Analysis l
Violations associated with the fire detection system stem from not posting a l fire watch within the time limits allowed by plant TSs when a malfunction of l the fire detection system occurs. For example, at the Byron plant, permanent staff personnel are responsible for the proper functioning of the fire detection system itself. However, the duties associated with the performance of fire watch functions have been delegated to contractor security force personnel. This delegation of authority may have been the primary underlying cause of the violations. These problems were attributed to the security force personnel having difficulty in making the transition from plant construction and startup activities and requirements to the much more stringent requirements associated with a licensed operating plant. Also identified by the plant staff as a contributor to the problem was the conflict between maintaining area security and simultaneous fire watches when equipment, especially the security computer, was out of service due to maintenance or problems. For example, if the security computer went down while a security staff member was performing fire watch duties, all security doors failed in their locked position. Access to areas beyond these locked security doors can be attained only by using a standard door key. However, the security personnel on fire watch are not -
permitted to carry the door keys. Therefore, the person on a fire watch had to go to the proper security area to obtain the necessary keys. Because of the time that was required to complete this action, the fire watch TS time limits were frequently violated.
Violations associated with the radiation monitoring system are associated with failing to take radiation samples either with the system or by other means if the system is malfunctioning. The plants with the greatest difficulty were those with problem radiation monitoring systems, as discussed in Section 3.2.
As noted earlier, fire detection and radiation monitoring are only the most prominent of 70 different systems involved, and together account for only about 25%
of the total events identified. To get some perspective on the factors underlying l the wide range of violations, we relied on discussions with plant staffs.
44
1 First, according to plant staffs, the effort required to adapt " standard" TSs to an individual plant is formidable and not to be underestimated. Early and concentrated efforts are needed to finalize TSs well in advance to avoid situations where the TSs are in a state of flux right up to licensing. At Hope Creek, the licensee's staff stated that the final version of the Hope Creek plant TSs was not approved until shortly before OL issuance. Their preparation for operation could have been enhanced by more timely TS finali-z3 tion. Utility personnel utilized the TS surveillance procedures at mode changes to correct numerous problems. Although it was a major effort to I perform the surveillance, in retrospect, it was deemed very worthwhile. The licensee staff also noted problems with tracking the applicable TS requirements for different modes. Improvements were made to the tracking systems to ensure that TS requirements were met for different modes. Other improvements were also noted regarding the positive assignment of TS responsibilities to discrete organizational units to ensure that TS requirements were app *opriately monitored.
An example for WNP-2 provides further 171ustration. There were 60 TS violations reported by WNP-2 during the first 24 months of operation. By far the most significant contribution to this total came from the emergency ac power system, which was involved in 21 violations. Each of these violations consisted of performing the required surveillance on an emergency ac power diesel generator without a prelube/ warmup as prescribed in the plant TSs.
These violations were attributed to the fact that the final version of the WNP-2 TSs, as approved, required a prelube/ warmup of the diesel engines per recommendation of the manufacturer. To do this would require that the engines be run at idle speed. The diesel engines, as installed at WNP-2, although designed and tested according to the regulatory guidance contained in Regula- I tory Guides 1.9 and 1.108 and applicable IEEE standards, do not have the cap-ability to run at idle speed. Consequently, the prelube/ warmup could not be performed as specified in the TSs, and every time the surveillance was performed on a diesel-generator, a violation resulted. This situation existed {
until the licensee was able to obtain approval of a license amendment which allowed starting the diesel generators without a prelube/ warmup period.
In order to help assure timely availability of accurate TSs, one licensee established a local office near NRC staff to expedite and coordinate problem resolution.
Another general factor leading to TS violations is that new plants often have difficulty in establishing an operating plant vs. plent-under-construction mentality. Knowledge of, and strict adherence to, TSs and related procedures become essential after licensing. This is an area in which the second unit at a site has an advantage, although even here vigilance is required to avoid problems. At St. Lucie, for example, subtle differences between Unit 1 and Unit 2 TSs caused problems when experienced Unit 1 staff worked on Unit 2.
Two specific examples of such difficulties were cited. The first consisted of different specifications for the safety injection tanks and associated vents at Unit 2 as opposed to those for Unit 1. This caused some TS violations at Unit 2 when the Unit 1 TSs were mistakenly used for Unit 2. The second case l
45 I
t
involver the TS requirements for water level in the reactor vessel during re-fueling; for Unit 1, the level is 23 feet above the fuel, whereas the corresponding requirement for Unit 2 is 23 feet above the flange, which is about 10 feet higher than the level requirement for Unit 1. Once again, this caused some confusion, especially among the operations personnel participating in the startup of Unit 2 who had Unit 1 experience, since the design of the two units is identical in this respect.
To a large extent at new plants, the training program does not emphasize the TSs as much as other requirements. Plants visited identified a clear need, in retrospect, for formal TS training prior to licensing. Ne such program is enacted in many plants. At Palo Verde and Diablo Canyon, management became aware that personnel were not fully informed of TS requirements. For example, in some cases, I&C personnel may not have even known that missed surveillance were reportable to the NRC. Palo Verde and Diablo Canyon relied on training and administrative improvement programs to remedy the situation. TS training in (1) the various TS requirements of different modes, and (2) the significance of TSs would improve this area.
Measures to reduce the number of various TS violations have been incorporated by licensees in improvement programs, such as the Conduct of Operations Improvement Program at Byron and the Quality Improvement Program at Palo Verde. Further, specific training to ensure that the significance of plant TSs is understood by all levels of the plant's operations staff either was instituted or strongly recommended.
3.3.3 Findings The findings stated below are based on the analyses of operational data for new plants regarding violations of Technical Specifications.
(1) Although often highly technical and of low immediate safety significance, TS violations have a high regulatory visibility, and can be important to the NRC perception of licensee competence.
(2) Almost all TS violations occurring at new plants are the result of personnel errors or problems with procedures. This finding is consistent with reviews of operational experience for all licensed plants.
Although human error is the dominant cause of TS violations, equipment I problems can generate the opportunities for human error that result I
in violations.
(3) Over 70 different systems have been involved in the TS violations occurring at new plants.
The systems involved in the most violations, the fire detection and the radiation monitoring systems, together accounted for only about I
25% of the total number of violations.
(4) The initial post-OL issuance violation rates vary widely but are generally higher than the mature plant average of 0.8 violations per month.
Significant moderation usually begins about six h eight months post-OL issuance in about one-half of the cases.
46
f I
High violation rates persist beyond 12 months after OL issuance for a few plants.
(5) The results of a statistical analysis indicated that the rate of TS violations was significantly lower during the first sir months of commercial operation for plants that are the second unit at a multi-unit ,
site. s Evidence for other correlations was not statistically as strong. J.
However, the data show a tendency toward a lower violation rate during startup for second units at a site.
No statistically significant correlation was found between TS violation rate and NSSS vendor, or OL issuance date. :
(6) A brief evaluation of the role of TSs in overall operational experience indicated that, for the new plants studied, variations in operating experience are probably not driven by regulatory changes in the scope of TSs, or major differences in the quantity of LCOs, or related surveillance among NSSS vendors.
Some plant-to-plant differences result from differences in which plant features are designated ESFs.
Nearly all plants showed evidence of a learning process regardless of the initial plant-specific features of the TSs. This process was accelerated for the second unit at a multi-unit site.
~
(7) The effort required to adapt " standard" TSs to an individual plant is formidable and should not be underestimated.
There is a need for early and significant cooperative efforts j between the NRC and licensees to ensure that TSs are properly tailored to the plant, and finalized to allow generation of supporting procedures and familiarization of the plant staff with requirements.
(8) Discussions with licensee staffs indicate that difficulties in working with TSs can be traced to (a) not having final TSs early enough, i.e.,
having TSs in a state of flux right up to licensing, (b) difficulty in establishing an operating plant versus a plant-under-construction mentality, i.e., lacking appreciation of the necessity for strict ,
adherence to the TSs after OL issuance, and (c) lack of training on TSs.
(9) Measures to reduce the number of various TS violations have been incorporated by some licensees in performance improvement programs (e.g.,
the Conduct of Operations Program at Byron and the Quality Improvement Program at Palo Verde).
3.3.4 Lessons Learned Based on the lessons learned from the experience of new plants that focus on the reduction of TS violations, the following improvement lessons were developed.
47
(1) Establish extensive, detailed training for all segments of the onsite plant staff, including I&C technicians, maintenance mechanics, security '
staff, operations, and marap uent.
This training would emphasize: (a) the applicability of the various TSs to the changing plant modes of operation and associated -
schedules, (b) the relationship of the TSs to the nlant procedures, (c) the NRC requirements for deportability of violations, and (d) the basis for the TSs and discussion of LC0 requirements.
(3) Use the finalized TSs to generate and validate surveillance testing procedures (e.g., against the as-built plant) as early as possible. In this regard, great discipline should be exercised to restrict the number of last-minute changes in the proposed TSs. Once final draft TSs are issued, the licensee can begin to incorporate TS requirements into plant procedures instead of waiting until the last few changes have been implemented. In conjunction with this activity, have plant staff (as opposed to NSSS vendor, special startup group) perform all surveillance.
It is recognized that development of finalized TSs involves a joint licensee /NRC staff effort.
The permanent' operating plant staff (as opposed to special startup group) should perform all initial surveillance.
(4) The NRC staff's review of the finalized TSs should be scheduled and staffed to allow approval as early as possible prior to licensing.
3.4 Unplanned Losses of System Safety Function Sinca January 1,1984,10 CFR 50.73 requires that licensees report any event or ccndition that alone could have prevented the fulfillment of the safety function of structures or systems that are needed to:
A) Shut down the reactor and maintain it in a safe shutdown condition, B) Remove residual heat,
[ C) Control the release of radioactive material, or 3 D)-. Mitigate the consequences of an accident.
Events covered by the requirements may include one or more procedural errors, equipment failures, and/or discovery of design, analysis, fabrication, con-struction, and/or procedural inadequacies. Whether or not the system was required to operate or be in an operational status (standby) when the condition was discovered does not affect deportability. Data for these events 3 was ntracted from LERs, exclusively after January 1,1984, and from 10 CFR 5%.72 do.a for 1983 as necessary, supplemented by a review of 1983 LERs.1/ .
~1/ Prior to implementation of 10 CFR 50.73, all plants had the following ,
reporting requirements in their TSs: (a) component failure or malfunction which prevents or could prevent functional performance in a'ccident situations analyzed in the SAR (Safety Analysis Report), (b) personnel error or procedural inadequacy which prevents or could prevent functinnal performance of systems in accident situations analyzed in the ',
SAR, and (c) failure of the RPS or other systems to initiate the required protective function in the appropriate time.
48
s The loss of system safety function (LSSF) is an infrequent occurrence, and actual system failure when needed is even less frequent. The events reported and analyzed nere involved, for the most part, situations wherein a system might not perform satisfactorily under some postulated circumstances (e.g.,
given a design basis event, a fire, or seismic event), or a failure to function during a surveillance test.
3.4.1 Data Analysis Figure 12 shows cumulative event counts vs. calendar months since OL issuance.
A line with a slope of 0.26 events per month, the average frequency for 76 mature plants in 1985, is provided for comparison. The general pattero is for .
a plant to go for a long period of time with no events, then experience a few events in the space of a couple of months, then have another long period with no events. Rarely are more than 20 events accumulated, even over opera-tion for the full 24-month period. Only WNP-2 showed a fairly long learning period.
LOSS OF SY:-TEM SAFETY Function EVENTS
- LEGEND
- Ff31 + PAY 1 j
24 198s Average Rate - Mature Plant Slope ,, , ,,
20--
f +'-+-+
g
.N.,
CNS2 g RS$1 e ( +...+..+....+..#
+ DCPt . x . a . SNS t , ,
E3 .. s x . DeP o sum *
% nd y ++ /*- g EFP2 O SE32 j /+
fW D C 3 f~ 30 4 f-_./g : Fo * " "' ^ ***'
e2 ep y-
..... y = =,< < e .
-- - =c= . -
.: _4$ o o o asi . .cs,
- .. p#*[M= '.-*e--ao a f*_;M = = = = = = " -*
- o was:_}}.-Q[G--
re / Ic : : 7-L.s s..s .s 0 2 4 6 8 10 12 14 16 18 20 22 24 Months Since OL issuance Figure 12 The formal statistical analysis carried out using the data in Table 5 showed a statistically significant correlation between the average rate during startup and the average post-commercial rate. However, in view of the very low event a 49
j \ rates prevailing in the post-commercial period, this latter result is not of much importance. Of the other correlations, the length of the startup period came close to being significantly correlated (negatively) with the average port-commercial rate. The data shown in Table 5 do not provide a basis for ranking plant performance. Table 5 Average Rates for Loss of System Safety Function
~
Pre-Commercial Post-Commercial Plant Events / Month Events / Month McGuire 2 0.33 0.00 Diablo Canyon 1 0.28 0.00 I Callaway 0.93 0. " Catawba 1 0.69 0.1. Byron 1 0.27 0.16 i Wolf Creek 0.33 0.10 1 Diablo Canyon 2 0.18 N/A, Millstone 3 0.82 N/A St. Lucie 2 0.96 0.50 Waterford 0.21 0.0Q Palo Verde 2 0.44 N/A LaSalle 2 0.58 0.34 Wi.P- 2 1.25 0.33 Susquehanna 2 0.45 0.15 Limerick 1 0.45 N/A, River Bend 0.72 N/A Average 0.57 0.20 Note: These plants had not accumulated 180 days of commercial operation before June 30, 1986. Therefore, their post-commercial experience was not considered in the statistical analysis. 3.4.2 Cause Analysis Since the event frequency is generally very low for each plant, one cannot establish a causal trend for a given plant or make a meaningful statement about cause categories in general (see Appendix H). The analysis of LSSFs for the plants in this study collapses into a detailed examination of the experience at WNP-2 from March 1984 to March 1985. During this period, a total of 16 events were reported. Of these 16 LSSF events, two involved the reactor core isolation cooling (RCIC) system, three involved the emergency onsite ac power system, and two involved the low pressure core spray (LPCS) system. The other nine events each involved a different system. An examination of the LSSF LER dr.t.o for WNP-2 indicated that two sets of four ! events each contained related events. The first set of four LSSF events i SU I
I resulted from a review and update of the 10 CFR 50, Appendix R, safe shutdown analysis. This review found that certain cabling for the following four systems was not appropriately protected from fire: emergency onsite ac power 3 system, the standby service water system, the residual heat removal (RHR) system and the reactor building HVAC system. Upon identification of these deficiencies, the licensee immediately set the prescribed fire watches in place until the cables in question could be protected by the application of appropriate fire protection coating. The failure to provide the required degree of fire protection for these cables, which had been installed after the original Appendix R analysis was completed, was due to human error. The second set of related LSSF events involved the loss of room cooling to the RPS Poom #1, Division I battery and battery charger rooms and Division I emergency bus SM-7 on two separate occasions. The loss of cooling to these rooms required that the reactor core isolation cooling and low pressure care spray systems be declared technically inoperable, although the power sources and system functions were never actually lost. The two occurrences were both caused by vibration problers related to the design of the cooling unit fan motor. The first time that the loss of room cooling occurred, the licensee replaced the fan motor. After the second occurrence, the licensee performed a detailed root cause investigation of the fan motor shaft and associated shaft bearings. The results of this investigation prompted the licensee to replace the shaft bearings with those of a different design. No further occurrences of this type were reported for the remainder of the 24-month period. In the other two cases involving the emergency onsite ac power system, none of the diesel generators in the system (including the diesel generator supplying emergency power for the high pressure core spray system) reached the required voltage for automatic closure of their output breakers following a loss of offsite power event. The voltage adjust potentiometers for the diesel genera-tors were found to have been adjusted to the low voltage limit. The erroneous adjustment was caused by a deficiency in the procedure used during the adjust-ment process. The procedural deficiency and the resulting potentiometer misadjustment had been discovered a few days prior to this event, but, at that time, it was felt that the misadjustment would not affect the automatic breaker closure. In response o the event, the licensee read.iusted the affected potentiometers and revised the appropriate procedures to incorporate input regarding the voltage adjust high and low limits. 1 3.4.3 Findings, The findings stated below are based on the analyses of the operational data for new plants regarding LSSF events. ; i ! (1) The loss of system safety function is not frequent, even at new plants. In general, plants go for long periods of time with no events interspersed with brief I-2 month periods when a few events occur. 51
(2) The statistical analysis of LSSF showed some correlation between the event rate during the startup period and the event rate during the first six months of commercial operation. This result tends to reinforce the similar correlation found earlier for scram rate and ESF actuation rate. (3) No particular cause category dominates the LSSF events. A detailed review of the LSSF events at WNP-2, which showed a learning curve different from the other plants in this study, indicated that design and human error were responsible for conditions which might have prevented the successful performance of equipment had it been called upon to function. (4) Due to the low numbers, sporadic occurrence frequency, and variety of causes for LSSF events at new plants, no particular new issues or recommendations have been identified in this area.
- 4. EVENT SIGNIFICANCE AND INSIGHTS Wnen considering the operational experience of new plants, a question ; rises regarding the safety significance of the reportable events which occur during their early months of operation. Some ma) not consider these events to be significant from a safety standpoint, since they occur during a period involving (1) the shakedown of new systems and equipment, and (2) the familiarization of operations personnel with the new plant. Some also note that if a startup and test program identifies a large number of problems that are subsequently resolved prior to routine operations, then the program has accomplished its mission. . 0n the other hand, experience indicates a likelihood that not all proclems have been identified by testing. This so-called " learning curve" philosophy does not exist in Japan. Their o (e.g., unplanned scrams)perating are not aexperience necessar.v/has shown routine that part ofreportable events a startup and test 4 program. They treat unplanned scrans as very significant events no matter whether they occe during startup or at power. In any event, these early phe:es are periods where carefully thought out, deliberate, and carefully planned ,
activities are needed. In order to provide further insights regarding the operational experience of new plants, this section characterizes the significance of these reportable events from some of the different perspectives used by the NRC staff. It also discusses the root causes for such significant events, as determined from other perspectives, ranging from generic analyses to plant-specific inspections. 4.1 Scrams With Complications One of the mere significant event types for which operational data are avail-able for comparison between new plants and mature plants is reactor scrams which had additional failures or personnel errors beyond those which initiated I 52
the scram. As already discussed in Section 3.1.1, the operational experience for new plants in this comparison had been divided into the pre-commercial or startup phase and the post-commercial phase. The operational experience data for the pre-commercial phase of new plant operation indicate that 16% of the scrams occurring at new plants during this phase involve complications. The corresponding data for the post-commercial phase show that 20% of the scrams involve complications. These percentages are comparable to the mature plant data, which indicate haf. 21% of all scrams at mature plants have associated complications. The significance of scrams during pre-commercial operation may be minimized to some degree due to lower power levels, the onsite presence of additional personnel and a recognition of the need to anticipate problems during a testing program. However, the scram frequency combined with the percentage l of complicated scrams still yields a rate of occurrence of complicated scrams which is about four times that of a mature plant. Early post-commercial operation shows a power level and complication profile like a mature plant, while the scram rate can still be two to three times that of a mature plant. 4.2 Results of Systematic Screening for Significance 4.2.1 Topics for Operating Reactor Events Meetings Over the years, periodic Operating Reactor Events Meetings have been conducted ; by NRR to brief the Office Director, the Division Directors, and their repre- ( sentatives regarding events occurring at operating plants which have been mutually agreed upon by several NP,C offices to be of significance or special interest. These events have recently evolved into the basis for the significant event performance indicator of the NRC PI Program. The PI significant events are those events identified by the detailed screening of operating experience by NRR and the NRC's regional offices, and include degradation of important safety equipment, unexpected plant response to a transient or a major transient, discovery of a major condition not considered in the plant safety analysis, or degradation of fuel integrity, primary coolant pressure boundary, or important associated structures. Events discussed at earlier briefings were not measured against these criteria per se. Rather, those events were selected more informally as of particular interest for their safety significance as potential implications for ongoing programs. Nevertkless, the events do indicate a level of NRC interest and may be used as a basis to reflect on differences in plant age classes. Table 6, on the following page, shows the number of events which occurred at each of the 22 new plants that were discussed at the NRR Reactor Events Meetings during 1983, 1984, 1985, and the first half of 1986. It also shows the average monthly avent frequency for each new plant. Using these results, an overall average monthly significant event frequency of approximately 2 events per year (0.16 events per month) was calculated for new plants. The data shown in Table 6 do not provide a basis for ranking plant performance. 53
Table 6
. .. ; -- ,.. .. g g...,..g. g ;..g. g ; .; g g..g. g ; , ..., ,
90CK111 PLANT 18tiffE8 ff NAA ;8t!!F[3 If nRA ltal[F[$ if NRR l$Rl[F(8 Bf tRA i 1 110. OF llaufMB DF i AvFRAa( so. 1 I i 111 1913 1 19 1904 1 lu ltB1 1 III 1986 I tafAL 1 OPERAfl0N Con 51DERED l (WENil/Ehf# 8
~
35 iklelRE 2 l iI 3I l 0l $ 24 0.200 l 389 l51. LLICil 2 1 01 11 01 01 11 24 1 0.082 1 275 lBIABLO Ct'10m i I O1 0t 21 01 28 24 1 0.04 ' 374 ILA$ALLE 2 1 01 01 11 01 11 24 1 0.042 1 197 hSH.nUCL[AA2 1 01 l1 21 0l 31 24 4 0.125 1 380 llLS9trEHAhhA 2 1 01 21 21 0I $I 24 1 0.20I 1 483 lCALLAsAf 1 i $1 11 II OI 21 24 I 0.083 1 413 lCAfauBA 1 1 01 01 31 21 51 24 1 0.200 1 454 lBfRDR I I OI e1 31 01 2i 2! l 0.143 1 352 ItinERICX 1 1 0I i1 01 01 1I ?! l 8.048 1 528 iPALO VERDE 1 1 01 01 61 41 le i 19 ! 0.4N 8 322 llefMan 1 01 01 01 01 01 19 1 0.000 1 382 ltATERFORD 3 1 01 01 21 01 21 19 1 0.105 8 482 150LF CREIK I OI OI 01 01 01 16 1 0.000 1 341 IFIRR! 2 1 01 01 61 21 0i to 1 0.500 1 323 tilASLO CAlltDN 2 1 01 01 11 0t 1i 15 1 0.067 1 j 458 1RIVEA RENL I OI OI 61 51 5I il 1 0.455 1 ' 423 In!LL$f0NE 3 1 01 01 0. v1 01 51 0.000 1 529 lPALO VERDI 2 1 01 01 01 2l 21 71 0.286 1 414 ICATAllM 2 1 01 01 21 11 31 $1 0.400 t . 440 PERRY ! ! 91 01 01 21 21 4I 0.500 lavt2461 E. IVEuf5/RoufM l 354 In0PI CRiff 5 01 01 01 01 01 31 0.000 i fait als Ptauft TOTAL 1 10 32 18 1 61 1 374 FROM NR4 lRl[FING$a 0.14] Event counts for 1985 are shown in Appendix I for the 76 mature plants operating in 1985. Using this information, an overall average monthly significant event frequency of 6 approximately 1.5 events per year (0.13 events per month) was calculated for mature plants. Comparison of the two sets of 1 results indicate that the frequency of selected events which occur at new ' plants and are discussed at the NRR Reactor Events Meetings is a factor of 1.3 greater than the corresponding event frequency for mature plants. This ratio may well be biased low because more thorough and systematic selection accompanied the evolution to significant event input for the NRC PI Program. Therefore, the new plant 1983/1984 data may represent a low count of significant events when ccmpared to the 1985/1986 standards for determining i significant events. With the 1983/1984 data possibly being on the low side, the above factor may even be greater. 4.2.2 AE0D Event Categorization AE00 reviews every LER and other documents reporting operational experience to identify those events judged to be safety significant. These events are classified according to a set of evaluation criteria as Category 1, 2, 3, and 4 events, depending on the degree of safety significance ascribed to the individual event. Category I and Category 2 repr9sent those events deemed to be the most significant from a safety standpoint. 54
Events cli '
'ied ts Category 1 or Category 2 meet the following criteria:
Cate p y 1 - Those events considered to have such obvious significance that actions should be initiated immediately to ensure plant safety. Category 2 - Those eventt (or combination of events) which appear to have safety significance but do not require immediate action to ensure plant safety. Although general, these criteria have been fixed over the years since 1983. A i comparison of the AE0D significant event screening criteria with the recently j established "significant" NRC PI criteria cited in the previous section reveals J much sinilarity in terms of the coverage, kind and threshold for events of interest. Thus, tc a great extent th NRC PI and AE0D selections are rein-forcing (because they are done independently) rather than complementary. Some l differences do exist. For example, the AE00 criteria would consider an LER to ) be significant if it documented extensive BWR IGSCC (a generic design problem), I or if it documented an event with multiple safety-related equipment failures (a plant-specific operational safety problem). AEOD's criteria also require an event to be categorized as significant if it meets any Abnormal Occurrence (AO) reporting criterion. The NRC PI Program criteria appear to include some events which are not normally identified as significant by the AE0D screening program. For example, an event involving a " scram with complications" would not generally be categorized by AEOD as significant unless the complications were caused by one or more safety system failures or resulted in an important design limit being exceeded. Table 7 shows the distribution of AEOD Category 1 and 2 events for the 22 new plants for 1983, 1984, 1985, and the first six months of 1986. Also shown is the average monthly frequency of Category 1 and 2 events for each new plant. These results were then used to cal (= ' ate an overall average monthly frequency of Category 1 ano 2 Events for new p. ants of about 3-1/2 events per year (0.29 events per month). The data shown in Table 7 do not provide a basis for ranking plant perform &nce. ! Table 7
~ ~
it0'.lIditi3 ";~ aG. [44 6 i ~ E0. EvtRTS I 110. (Vitf3 I I I I pctatt; PLAsi la 1993 I la 1984 I lui i*t5 I til 1986 t i no. OF utl6 DF i A4RAst a0. I
- A;CORDia6 ?0 EDD:A:331u610 AIDO. A tDalllis to atS9:Act0R0!ut TO E30t T01Ai DPitAttoli Coi518Etti I (Esil/MouTM i l pr an- , /> ,; el D E- - & 24 O.41, l
"; 30T!'A.'.i ; i; '! ll l' ll lll ::llil 374 lLAinal 2 01 31 21 01 51 24 1 0.204 1
!;' la:0!"' l ll ll ll ll !! !!! ::0Tl at! :tauase 1 01 5I 01 II 61 24 1 0.250 t 4t3 :CATAsta 1 01 41 61 21 12 1 24 I 0.500 t l 91 11 0.04 1 j 45( BTR0a i ! OI 61 1l 21 1 i 332 alntRI:s ! . 91 0: 2l 01 2I 21 1 0.095 I l 528 :PaLO vitBE 1 e 9I ei 21 II 31 19 l 0.158 1 3:2 !5dOREMA8 91 01 lI i1 21 19 I 4.10$ I
'352 ;eAffar0RD 3 1 01 0I 91 11 Ii 19 1 0.053 4 482 :a0LF Catts ; 9l 91 01 41 01 16 1 0.000 3 341 if!Rnl 2 01 0i ti 51 61 16 1 0.375 1 323 ;0:A9.0 Canton 2 1 81 8I 41 01 01 15 1 0.t30 I 438 :R!v!R ttNS I O1 91 11 21 31 11 1 0.273 1 423 lfilu$fCut 3 1 9I 41 01 0i Bl 81 1.0dv I )
$29 IPA 6014012 1 01 0i 91 3 31 71 0.42' I 414 :CAf aen 2 1 91 01 01 31 31 $1 3.600 1 440 l'IRH I I 91 01 0i t1 01 41 0.000 I 01 01 4I II II 34 9.333 lovinaat no, tvinistMoutil 354 :#0't CRitt '
10 % 3 30 11 to 79 376 TOR Eu Plaitil FRA At00 0ATA= 4.283 55
Corresponding information regarding AE0D Category 1 and 2 events for the 76 mature plants during 1985 is provided in Appendix I. Using these results, an overall average frequency of AE0D Category 1 and 2 events for mature plants of about 1-1/3 events per year (0.11 events per uonth) was calculated. These results indicate that Category 1 and 2 events occur at new plants at a rate which is 2.7 times the occurrence rate for mature plants. 4.3 Causes of Significant Events Two methodologies were employed to approach a discussion of root causes in the context of generally recognized significant events. These are some results from the Accident Sequence Precursor Program, and also some results from special inspections at new plants in response to significant events. 4.3.1 Accident Sequence Precursor (ASP) Insights The ASP Program involves the review of LERs of operational events to identify and categorize precursors to potential severe core damage accidents, such as those associated with inadequate core cooling. Accident sequence precursort are events that are important elements in such sequences. Such precursors could be frequent initiating events or equipment failures that, when theoretically coupled with one or more postulated events, could result in a plant condition in which core cooling would not be adequate. Precursor events are selected from all events based on standard criteria. The precursors are then reviewed to assess their relative significance through PRA methods. Therefore, this methodology was employed to review new plant experience.1/ The methodology includes events that involve one of the following: The failure of at least one system required to mitigate a loss of main feedwater, loss of offsite power, small break LOCA, or steam line break; The degradation of more than one system required to mitigate one of the above initiating events; or An actual initiating event that required safety system response. Because losses of feedwater occur frequently within the reactor population, they were selected as precursors only if other failures also occurred. Precursors from 1969-81 and 1985 (the periods for which analysis has been per-formed) with a higher relative level of significance, were reviewed in detail to identify differences in precursors which occurred within two plant age groups (2 years or less vs. greater). Forty precursors were reviewed, and the frequency of errors, system failures and types of systems involved in each event are documented in Table 8. Although the time frame and threshold of analysis differ from that of Section 3 (thus the set of plants ard events are quite different), the frequencies of the new plant precursor failure mechanisms are evenly split between main-tenance and operator error on the one hand and instrumentation, mechanical AI Letter from J. R. Buchtnan, Director, Nuclear Operations Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee, to Mr. M. H. Willisirs, Program Technology Branch, AE0D, dated October 31,19S6. 56
and electrical component failures on the other. This corroborates the findings in Section 3 for scrams, ESF actuations and LSSF that neither human error nor equipment failure dominates in the first 2 years of operation. The balance tilts toward equipment failure for the older plant precursors. Thus the net trend moving out of the first 2 years is away from personnel error. : The most noticeab'e shift is an increase in electrical failure for older plants. Tabic 8 1969-1981 and 1985 Precursors Frequency of Errors / failures (Per Precursor) for Principal Systems: New* Older ** Maintenance Error .58 .44 Instrumentation Failure .21 .13 Operator Error .38 .25 Mechanical Failure .46 .44 Electrical Failure .25 .56 Frequency of Failures (Fer Precursor) for Systems Involved: Main Electrical (High Voltage) .13 .13 Other Electrical .25 .31 . HPI .08 0 l AFW .38 .25 { HPCI .17 0 1 Instruments .25 .19 i EPS .17 .06 l RHR5 .08 .06 PORV/SRV .17 .06 MFW .25 .19 LPCS .08 .06 {
- Licensed 2 years or less.
** Licensed longer than 2 years. j The increase in electrical failure is mirrored in the breakdown by system contributors. Problems involving electrical systems (other than the diesel l generators) appear to increase with plant age (in conformance with the pre-vious observation in Section 3.1). Improvement with plant age wat observed for PWR AFW and BWR HPCI systems. .
l l Many precursors (in addition to the more serious events addressed above) which j occur in the first 2 years after initial criticality exhibit similar ! characteristics to events which occur later in plant life. However, it appears events occurring early in life included more: (1) Problems in communication between members of the operating crew, i (2) Construction, equipment manufacturing and procedural errors discovered l during anticipated transients [for example, during auxiliary feedwater l 57 l i L- _ - _ - _ _ _ - _ _ _ - _ . _
^
(AFW) system demands and loss of offsite power (LOOP) events] and during environmental conditions different than during startup testing and calibration. One would expect these events to occur more frequently during early plant operation. Thus, we find that when the focus is narrowed to a set of events with relatively greater safety significance (as defined within the ASF dpproach), even including a different set of plants from an earlier time, the same lessons on cause emerge as those from a wider lttk at the plants in this study. In addition, the ASP methodology was utilized to estimate the risk from new p? ants, on the average, versus mature plants. Altnough the technique used provides only an estimate of risk, it indicated that, considering the higher frequencies for transients and scrams observed for new plants along with a higher number of observed failures for some equipment and a perceived increase in human error, the risk due to new plants is nominally higher than that for mature plants. This tends to support what had been previously indicated by other unrelated analyses of the data. The following section provides concrete examples of the types of situations described in (2) above through examination of findings from the NRC Augmented Inspection Team Program and Special Inspections. 4.3.2 Augmented Inspection Team (AIT) Program /Special Inspections Significant events are also selected and evaluated by the NRC in other i prog rams . Among these are the Augmented Inspection Team (AIT) Program and other Special Inspections. A specific feature of these programs is that they l provide a thorou h investigation of each selected event and a determination of I the root cause(s for the event. Two AITs established since the program's inception in mid-1985 and one Special Inspection reviewed specific significant events connected with plants in the startup testing phase. 4.3.2.1 Catawba 2 AIT At 9:41 a.m. on June 27, 1986, a Loss of Control Room Test was initiated on Catawba Unit 2. A crew of five, simulating the minimum shift crew available to Unit 2 operations, started the test by leaving the control room, tripped the reactor from the trip breaker panel, and proceeded to the two remotely located auxiliary shutdown panels (ASPS), and the auxiliary feedwater pump turbine control panel (AFWPTCP). At the ASPS and AFWPTCP, switches were activated transferring control of vital functions from the control room to the auxiliary panels. By design, the transfer blocked automatic initiation of safety injection (SI). By error, the transfer of control of steam generator (S/G) power operated relief valves (PORV) to the AFWPTCP also commanded all four PORVs to open to 75 percent of full stroke. Reactor pressure and pressurizer level, which had been decreasing slowly as a result of the cooldown after the trip, fell rapidly. Within a minute of the transfer, pressurizer level indication was lost, and within two more minutes pressure haa dropped below 1845 psig, generating a SI demand signal. After another 3.5 minutes of unsuccessful attempts to manage the situation from the ASPS and AFWPTCP, control was returned te the contiol room, which was staffed by the 58
)
l
regular operations shift. The transfer, which automatically initiated SI, occurred at a pressure of 702 psig, the lowest pressure recorded. Safety injection was terminated about five minutes later,10:00 a.m., with reactor pressure above 1200 psig, pressurizer level near 34 percent, and about 100 F subcooling. On June 28, 1980, an Augmented Inspection Team (AIT) was dispatched to the site to' conduct an investigation to determine the facts-surrounding the incident. In its report (AIT Report Nos. 50-413/86-25 and 50-414/86-27) documenting the results of its investigation of this event, the AIT identified ; number of Causes. The underlying cause of this event was-the failure to specify in the Design Change Authorization (DCA) or other documents that the mode of control of the S/G PORV controllers at the AFWPTCP had been changed. This in turn led to a failure by station personnel to change procedures and train operators on this modification. The situation was further a cerbated by human engineering deficiencies introduced by the modifications. As a consequence, the staff assigned to perform the test did not understand the function and interaction of controls on the shutdown panels. This lack of understanding led to a pre-test setup of the panels that ensured that the PORVs would open on transfer and that ati.empts to shut them would be futile. Ot:er human engineering l factor failures led to reducing charging pump flow and flow to the reactor coolant pump seals by the vary attemptc to increase flow. Although the main control room PORV controllers were replaced with safety-related controllers, the licensee chose not to replace or modify the AFWPTCP controllers. Also, the licensee failed to revise other design documents to reflect needed scale changes and shutdown panel labeling. Other contributing factors to this event included inadequate training on the shutdown panel instruments and controls, inconsistencies in labeling of instruments and controls, and reluctance by the control room crew to assist the shutdown panel crew for fear of invalidating the test. The Catawba experience suggests that a deeper attention to detail and thoroughness may be needed prior to conducting plcnned testing. Specifically, I I training in the as-built plant for test evolutions and thorough prior engineering review are indicated. 4.3.2.2 Hope Creek AIT l As part of its power ascension test program, Hope Creek Nuclear Generating Station conducted a loss of offsite power test on September 11, 1986 from l approximately 21.5 percent of rated power. ' Twenty-four indivkiual test I observations were identified by Public Service Electric and Gas Company from this test. A second loss of offsite power test at Hope Creek was conducted on September 19 from a reactor shutdown condition (T = 200 F), and 17 (Aditional l test observations were identified. The most significant observations from i these two tests were: (1) emergency diesel, generator "C" outpn breaker i failed to close; (2) power supplies for neutron monitoring source and i intermediate range monitors' detector drives and main steam line acoustic monitors were lost; (3) 17 control rods did not provide'a normal full-in position indication; (a) reactor auxiliary cooling system flow was lost; (5) l emergency diesel generators "A" and "B" governors transferred isochronous i 59 l 4 l
(frequency control) to speed drop (load control) mode without operator action; (6) the "B" safety auxiliary cooling system pump failed to auto-start; (7) the "B" safety auxiliary cooling system loop head tank level indicator failed; (8) one control room emergency ventilation f air recirculation) system fan failed to start; ar.d (9) one drywell fan also fiiled to start. On September 24, an Augmented Inspection Team was formed to investigate and evaluate the test observations. In the report (AIT Repo"t No. 50-354/86-50, dated October 28,1986) documenting the results of its investigation, the AIT stated the following conclusions. Of tht. observations reported from the two loss of offsite power tests, the overall safety significance was relatively minor except for the Bailey solid state logic module failures. Of eight hardware failures identified during this review, six were attributable to various malfunctions within the Bailey logic modules. Three weaknesses found with the Bailey logic modules were: (1) the dependency on common equipment for accomplishment of automatic and manual safety actions for the actuated safety systems equipment; (2) limited test provisions to assure the online operability of the Bailey logic modules after their installation into the equipment cabinets; and (3) the usefulness of the bench test equipment in assuring that the Bailey logic modules are operable. The team was also concerned that the failure rate of the Bailey logic modules appeared high. These weaknesses are especially significant since all of the balance of plant safety-related systems (and a part of one LSSS system) use Bailey modules to develop the safety system logic and actuation functions. A number of minor plant design, construction, and manufacturing problems were also identified. Several specific weaknesses in the scope of various system preoperational tests were revealed since the loss of offsite power tests were the first integrated demonstration of the plant response to this event. Several subtle interactions involving the dependency of various systems on cooling and instrument air supporting systems were revealed. A number of observations resulted because instruments or other equipment lost power during the test. A number of these instances involved the apparent failure to meet FSAR commitments to provide reliable power to specific instruments or equipment. The AIT Report discussed in detail the results of the root cause analysis for each reported observation from the loss of power tests conducted on September 11 and September 19. The root causes were grouped into ten cause categories as follows: Equipment Failures, Preoperational Tesi Deficiency, Observation Error /Non-Problem, Design Flaw, Construction, Operator Error, Security, Training, Procedure Flaw, and Indeterminate. This AIT itself had a significant impact from a few perspectives upon the licensee's s aff. AEOD disc; ssed the event with the licer.see staff in January 1987. Upon rurther reflection the licensee's staff indicated that they believed that the AIT would not have occurred if the Regulatory Guide (R.G.) 1.97 acoustic monitors had been powered from the proper bl.s. They indicated that that AIT, therefore, could have been avoided by more thorough preoperational testing of this R.G.1.97 equipruent. However, no NRC requirement made this necessary. 60
Great efficiency and programmatic streamlining during the period from the later stages of construction to operation are of obvious major concern to utility management. Time is at a premium during this period, Although certain testing would be clearly worthwhile in retrospect, this lesson would be difficult to transfer to another plant in a similar status absent some regulatory cencern. Implementing a more thorough and planned program from ! preoperational testing through full power operation may improve this problem, since schedule pressures (e.g., to complete construction, to get OL, to go j commercial) can influence performance adversely during the transition from i construction completion to routine operation. ! 4.3.2.3 Palo Verde Special Inspection On October 3,1985, Palo Verde Unit I was at 52 percent reactor power, when a reactor trip occurred due to flow-projected, low Departure from Nucleate Boiling Ratio on all four Core Protection Calculators (CPCs). This was due to a loss of offsite power (LOOP), which caused the speed of the reactor coolant pumps to decrease. The CPCs sensed an imminent loss of forced coolant circu- ' lation and tripped the reactor. The loss of offsite power was caused by switchyard breakers opening, due to an apparent malfunction in the plant multiplexer (PMUX).1/ All plant electrical loads were aligned to the offsite power supply via the startup transformers, due to preparation for a subsynchronous resonance test which required this alignment, when the LOP occurred. Loss of the PMUX caused loss of indications and remote control. The LOP resulted in che starting and loading of both emergency dM1 generators which restored power to the Engineered Safety Features buses. % N' loads were sequentially loaded as designed. Offsite power was res* M e '., the plant electrical buses within 24 minutes. Due to the LOP, the Fuel Building Essential Ventilation System, the Control Room Essential Filtration System, and the Containment Purge System (ESFs) all actuated as expected. The reactor coolant system reached a peak pressure of approximately 2290 psia and a low of approximately 2120 psia during the event. ' The auxiliary spray system was utilized twice to control the reactor coolant system pressure. The auxiliary system automatically started, and maintained proper steam generator water level. When the "B" Train diesel generator was shut down, the generator was unloaded too rapidly, which caused the diesel engine to shutdown. Normally the diesel er.gine enters a cooldown cycle prior to shutting down. The control room personnel received a diesel generator high priority alarm signal and declared diesel generator "B" inoperable. -The loss of field relay was found in the tripped position and reset. Subsequently, a surveillance test was I successfully performed on the generator and it was declared operable. The I loss of field relay actuatian was a result of non-licensed operator errors. The operator, who was in training, was readvised regarding proper shutdown of the diesel generator. Repeated attempts to restart the reactor coolant pumps were unsuccessful. A l 1/ This is the same problem discussed previously in Section 3.1.2.1. 61
( design review revealed a lockout on RCP restart (intended to accommodate coastdown of the main generator voltage, although the RCPs were not on main generator power at the time). After reset of these newly discovered lockout features, the RCPs were successfully restat ced and forced coolant circulation restored.To prevent recurrence of the reactor trip, the switchyard breakers that were affected by the PMUX failure were hardw1 red by the licensee. The PMUX breaker control was bypassed as a result of the hardwiring. The procedures governing the restart of the RCP's were changed to specify that the relay must be reset after events which result in their actuation. However, before the licensee could complete the corrective action involving the hardwiring of the switch-yard breakers, another loss of offsite power and subsequent reactor trip due to PMUX failure occurred on October 7,1985. Subsequent to this second event (which was very similar to the October 3, event), the NRC dispatched a Special Inspection Team to the Palo Verde site to review these two events and the licensee's corrective actions in light of concerns regarding the operational readiness of Palo Verde 2, which was nearing the completion of construction at that time. The team found that the licensee's corrective actions were appropriate to prevent recurrence of this type of event. The central issue raised by the Palo Verde experie:e, as well as that previously discussed at Catawba and Hope Creek, is the trade >ff between more extensive preoperational testing, training, and engineering review to eliminate problems versus discovering those problems at a later point !n time. In the cases of Catawba and Hope Creek, integrated tests caught the problems and the tradeoffs involve the nature and the timing of the tests. The Palo Verde events did not occur in the context of a test, and the need to catch such things in some kind of test seem clear. While the optimal approach (the balance of preoperational testing and integrated testing during the startup program) cannot be determined here, it does seem clear that additional preparation for testing and additional testing are called for. These activities, and the time to accomplish them, are not likely to materialize without a regulatory mandate.
4.4 Findings
The findings stated below are based on the analysis of significant events which have occurred at new plants. (1) The significance of scrams during pre-commercial operation may be reduced to some degree by their generally Icw initial power level, the lower fission product inventory, the onsite presence of additional personnel, and a sensitivity of the operating shff and management to unexpected problems with an attendant response readiness and capability. However, the high scram frequency, with its associated percentage *f scrams with complications (i.e., additional failures following the se s ), still yields a rate of occurrence of scrams with complications wnich is about four times that of a mature plant. Early post-commercial operation produces generally fewer scrams, but they occur from higher power. The scram rate can still be two to three times that of a mature plant. The factors that would minimize the event significance, noted above, may not be present. (2) NRC staff methods to screen events (AE0D event categorization or NRR event selection for further review / discussion) reveal a disproportionately higher number of events for the 22 plants covered in this study versus the <6 mature plants. 62
1 1 l (3) When the focus is narrowed to a set of events with relatively greater safety significance (as defined within the ASP approach), even including a different set of plants from an earlier time, the same lessens regarding cause emerge as those from a wider look at the plants in this study. (4) The scram review and the AIT/Special Inspection experience appear to support additional testing in either the preoperational or startup program. 4.5 Lessons Learned Based upon the lessons learned from the experience of new plants that focus on the reduction of significant events, the following improvement initiatives were formulated. (1) Conduct more thorough reviews and dry runs for planned testing and allow time for additional testing during either the preoperational or startup prograns. Emphasize planning for test objectives to reduce the frequency of unplanned scrams and unnecessary ESF actuations. A detailed review of operational experience should be a principal guide to the areas needing additional attention. Additional support for this approach has been provided by the loss of offsite power test recently completed at Fermi 2. In this case, the Fermi 2 plant staff had been preparing for the test for about a month i prior to its execution on March 16, 1987. As part of this preparation, I representatives from Fermi 2 visited the Hope Creek site to review that plant's experience with the same test (the Hope Creek experience has been discussed previously in this section).
- 5. STAFF AND INDUSTRY LESSONS FOR IMPROVING STARTUP EFFECTIVENESS The preceding sections have pointed to numerous plant-specific factors that contribute to new plant performance. Even in plants of similar size and NSSS design, at the system level, the design differences and the concomitant tailoring of TSs are enough to negate to some extent the industry-wide learning curve and design maturation which seems to contribute to relatively better experience at Japanese plants. These differences are illustrated in i the discussions of unplanned scrams and ESF actuations. Nonetheless, within i
this study of specific plants and broader industry initiatives is ample evidence that generic lessons can be identified, and the systematic feedback from even such a disparate population has cencrete value. A primary goal of this study was to systematically identify, through an evaluation of operating experience, those improvement lessons that would be of the highest benefit in reducing the frequency of reportable events at new plants. This section contains a listing of those lessons for both licensees and NPC staff that were identified in the review, and that appear to have good potential to improve performance, They have been discussed in detail in previous sections of this report, notably Sections 3.1.3, 3.2.3, and 3.3.3. In order to present this information in a logical manner, they are grouped according to management and equipment. Some of these items apply to l 1 1 63 l
newly licensed plants only (identified by 0L), while most could benefit operating reactors through the first two years of commercial operation (identified by OR). As part of the Peer Review, each of the 22 plants in this study was requested to identify those initiatives presented in the draft report which were judged to have the highest potential benefit for reducing the number of unplanned reportable events and improvino startup performance. Over one-half of the plants responded to this request. Accordingly, Items 5.1.1(1) through (6), dealing with management lessons, and Items 5.1.2(1) through (7), dealing with equipruent lessons, are ordered, highest to lowest, based on the original AE0D priority considerations and our evaluation of the comments received. Item 5.1.1(7), which oeals with training, feeds into the other it9ms, and was not independently ranked. 5.1 Improvement lessons for Consideration by Licensees 5.1.1 Management Lessons (1) Establish an operating plant mentality well prior to initial criticality. Ensure that plant operations personnel have the responsibility for operating all equipment as early as possible in the construction completion process. Take early, complete control of the transition i from construction to operation. (0L) Have personnel who will be responsible for maintenance and testing of plant systems after licensing begin these activities using post-licensing procedures before fuel load. This lets procedures get debugged, and the plant staff gains experience in working under licensed conditions. (0L) Stress the importance of details, the need for discipline in follow-ing procedures, the need for awareness of plant cond,tions and the regulatory requirements associated with these cond' ions, tight coordination throughout the plant staff, and the r.eed for expedited resolution of problems. (OR) Minimize continued construction activities after fuel load that may have an adverse impact on plant operations. Reduce plant staff to operational size, remove construction equipment, and establish housecleaning programs. Bring AE, NSSS vendor key personnel onsite so that problems can be resc'.ved promptly wher. discovered. (OL) (2) Conduct a deliberate, evenly paced, thorough and well-planned preopera-l tional and startup test program. Conduct thorough reviews and dry runs for planned testing and allow time for additional testing during either the preoperational or startup testing program. Emphasize planning for test cbjectives to reduce the frequency of unplanned scrams and unnecessary ESF actuations. A detailed reviw of operational experience of similar 64
plants should be a principal guide to the areas needing additional attention. (0L) Minimize the number of deficiencies and outstanding items carried - forward. Establish a policy of complete resolution before proceeding. (0L) (3) Use the finalized TSs to generate and validate (e.g., against the as-built plant) surveillance testing procedures as early as possible. In i l this regard, great disciplina should be exercised to restrict the number of last-minute changes in the. proposed TSs. Once final draft TSs are issued, the licensee should begin to incorporate TS requirements into plant procedures instead of waiting until the last few changes have been implemented. opposed to NSSS In conjunction with this vendor or speciti activity), startup group have plantallstaff perform (as surveillance. It is recognized that development of finalized TSs involves a joint licensee / NRC staff effort. Therefore, this measure relates directly to Item 5.2.1 (3) under Improvement lessons for Consideration by the NRC Staff, which addresses the corresponding staff effort. (0L) , (4) Improve administrative control of surveillance. For example: Considering that some new plants havo experienced problems when work has been performcd in the vicinity' of instrument racks during plant operation, licensees should evaluate th9 location and nature of work activities during operation in terms of adverse effects on plant operation and take appropriate administrative actions. (OR) ' Implement schemes to separate channel testing, such as a specific day of the week assigned to work on each channel, and to identify the channel in test, such as posting on control room panels. (OR) Blend engineering staff into the I&C organization. (OR) l Flag, categorize, and schedule surveillance according to risk of ) scrams or other ESF actuations. (OR) Organize the I&C staff to establish accountability for specific equipment. (OR) j 1 (5) Give high visibility to the sources (i.e., organizational element) of unplanned scrams (and other unplanned events) caused by human error and establish performance goals. (OR) (6) Ensure that operating experience feedback programs: (a) combine internal i events and relevant events from similar plants, (b) corrmunicate them directly to the appropriate first level supervisors and working level staff at the plant on a periodic basis, and (c) address preventive measures. i l For example, segregate the trip and ESF actuations h ta involving human ? errors from recent plant startups into the spec,fic positions, organiza-tional or functional element, working activity, systems and components, time of day, etc. Feed this information back at the lowest levels so ; l 65 l l
that the experience of others, the complexity of what is being done, and the ramifications of errors can be seen and appreciated. (OR) (7) A number of improvement lessons are directed at training as follows:. (a) Establish as a major goal an increased commitment to training in performing surveillance testing, calibration, and troubleshooting activities well prior to operations. I&C training initiatives, such as repeated practice for those surveillance testing activities that could cause a transient and which should be conducted on actual , in-plant equipment on' live systems prior to operations, should be emphasized. An additional action to improve surveillance testing suggested by licensee staffs was training for ISC personnel in valving - instrumentation in and out of service. (0L/0R)- , (b) Emphasize training for routine operations involving power level changes and the associated communications among shift personnel (i.e., feed flow and turbine evclutions) that have historically caused trips. Accelerated programs / efforts appear appropriate for newly licensed plants regarding steam generator level control. l Emphasize the need for site specific simulators to include, prior to startup, the best achievable fidelity of the simulator to the plant regarding feedwater effects (lead / lag characteristics of level indication and control methods) and include provisions to continue to imr.ive fidelity as the startup progresso. (OR) (c) Establish extensive, detailed training for all segmeats of the onsite plant staff, including I&C technicians, maintenance mechanics, security staff, operations, and management. (OR) This training would emphasize: (a) the applicability of the various TSs to the changing plant modes of operation and associated schedules, (b) the relationship of the TSs to the plant procedures, (c) the NRC requirements for deportability of violations, and (d) the basis for the TSs and discussion of LC0 requirements. 5.1.2. Equipment lessons (1) Focus on the B0P prior to cperation and early in life appears to provide a high return regarding the reduction of unplanned scrams and ESF actuations. Within this-area, attention could be given to: Conducting additional reviews of feedwater and turbine control and bypass systems to identify sensitivities and plant-specific characteristics that could contribute to transients or the adlity i of the system to cope with expected transients. (OR) { l Conducting a systematic review of equipment-protective logics and j 4 setpoints on components such as pumps (suction trip, time delay, vibration trip) or power supplies to identify areas where a time delay or additional channels for coincidence could reduce the potential for unnecessary transients or' spurious actuations. Give ] special attention to first-of-a-kind features not incorporated in .a earlier designs. Additional examples obtained from the plants visited include the main steam reheater. drain high level trip 66
and other turbine protective trips. (OR) (2) Install test jacks and bypass switches at appropriate points in actuation circuitry. (OR) (3) Implement on a priority basis vendor or licensee trip reduction measures. Licensee trip reduction programs should focus on safety-related equipment , as well as on BOP equipment. (0L) ' (4) Pay attention to the design and installation of equipment located in the vicinity of radiation monitors and associated cabling to ensure. ; that adequate grounding of equipment, cable shielding, etc., are provided ' to prevent the occurrence of EMI, which can trigger this extremely sensi-tive instrumentation. (OR) (5) Thoroughly test new or unique plant features, such as new RFS systems, l electrical systems, etc., prior to fuel load to reduce unanticipated i failures or unexpected erratic behavior. Emphasize planning to reduce the frequency of unplanned scrams and unnecessary ESF actuations. (0L) (6) For future designs or major plant modifications, preference for proven designs and standardization of design in plant feedwater and turbine systems appears justified. Conduct further analyses of any first of a kind, one of a kind, and state-of-the-art features, since they have generated a large number of problems during plant startups. (Examples of remedial actions are more extensive preoperational testing, reexamination of actuation logic to better achieve reliable indication and actuation; for example, reanal single input.) (OR)ysis of actuation on a single input or loss of a (7) Incorporate scram prevention measures such as: Develop a color coding scheme for single point scram components whose misoperation could cause a scram (for example, pressure sensing lines). (OR)
]
l Install cages or covers over switches or racks that could provide trip signals. (OR) 5.2 Lessons for Consideration by the NRC Staff 5.2.1 Management lessons l (1) Track new plant licensee progress via periodic (approximately monthly) meetings with the licensee to review the root causes of all reportable events and licensee corrective actions with plant staff present (NRC regions, with NRR and AE0D support) from OL issuance through early months of consercial 1 operation (until mature plant levels are approached). (2) Orient readiness reviews to deal with.the evaluation of management < effectiveness in the oversight of practical aspects of operations. ) (3) Ensure that the review of the final plant TSs is scheduled and staffed to 67
i allow approval as early as possible prior to licensing. Since development of finalized plant TSs involves a joint NRC staff / licensee effort, there f is a direct link between this measure and Itam 5.1.1(3) under Improvement : Lessons for Consideration by Licensees, which addresses the corresponding l licensee measure. (4) Review the progress of individual new plants (those with less than two y years of operation) in semiannual senior management meetings to provid visible NRC senior management oversight of their operation. (5) Develop an operational readiness review element that formally addresses a compendium of items from previous similar plant startups. 5.2.2 Equipment lessons (1) Highlight, through discussions with licensees, systems that historically cause, or can cause, a high number of challenges to safety systems. (2) Provide more focused attention to balance-of-plant operations (e.g., the i ability of the operators with the as-built plant to survive feedwater { transients and load rejections) to reduce the frequency of these transients. ! 1
- 6. INDUSTRY IMPROVEMENT LESSONS At the programmatic level, recent NRC initiatives have focused on the performance of new plants as a class. In the fall of 1986 the PI program, AE00 initiatives and consideration of certain plant specific events resulted fa senior management attention and action. Subsequently, the issue was the subject of a Commission meeting with the staff on November 14, 1986.
Within the industry there are numerous ongoing programs aimed at improving economic and regulatory performance of r.uclear power plants. Enhanced l feedback and use of operational experience, and scram reduction are examples which have relevance to improving newly licensed plant performance through startup and early commercial operation. Until very recently none of those efforts has been focused on newly licensed plants. However, that situation is changing. For example, the review of unplanned scrams and ESF actuations pointed to the contribution from unique or new design features and actuation or equipment protective setpoints and logics which are too conservative as designed cr not suited to the operational environment. Through detailed discussion with staff of recently licensed plants for the same NSSS, plants preparin5 for fuel load should be able to compare / contrast their designs to take advantage of specific experience, such as the time delay on a pump trip that was eventually used. A prototypr for such a meeting was recently held (December 1986) by 13 BWR owners to su re startup experience. A significant outcome of the meeting was a recognition of the value of the exchange and the intention to continue the dialogue through future meetings. In December 1986, the Nuclear Utility Management and Resources Committee (NUMARC) notified the NRC that INP0 had agreed to review and appropriately strengthen its initiatives in the area of early plant operation. To that end in March 1987, INP0 informed its members that they had reviewed their programs 68
and had concluded that additional action could be taken in several areas to assist new plants. Since it is still in the formative stage, the INPJ program is very general. However, it does plan to: (1) emphasize lessons learned from the good startup experience of other plants, (2) provide INDO training for senior plant management, and (3) apply the INP0 evaluation and assistance approach to newly licensed plants. 6.1 Scram Reduction Initiatives While not historically focused on new plant early operation, scram reduction initiatives are in various stages of development within the industry. For example, INP0 has published a compilation of scram reduction practices ba:.ed on data review ar.d plant visits.1/ Also, scram reduction is being addressed by NSSS owner's groups. The Westinghouse Owner's Group (WOG) program appears to be the most developed at Cis point. The Westinghouse Owner's Group developed a Trip Reduction and Assessment Program (TRAP) 2/, 3/ to evaluate reactor trip experience in plants having Westinghouse NSS5s so that generic approaches could be developed to reduce the frequency of such trips. Phase I of the TRAP program began in June 1985 and was in effect until June 1986 when Phase II began. Phase 11 should be completed by June 1987. The TRAP frogram has focused extensively on improving the supply and control of feedwater, the dominant sources of scrams in new as well as mature plants. For example, the WOG has performed studies directed at obtaining satisfactory performance for normal plant operation from existing steam generator level and feedwater control systems. These include: a generic study of the feedwater control system configurations (auto / manual control of the feedwater control valve with feedpumps of varying drive mechanisms) to provide guidance on how to improve performance below 30% power; and convening a panel of expert operators to obtain a consensus opinion on the best methods for manual opera-tion of the feedwater control system. The results were published in WCAP-11126, " Low Power SG Water Level Control System Improvements," (July 1986, Proprietary report) and WCAP-11138, " Improved Skill in Feedwater Control During Startup: Results of An Expert Panel Session," (April 1986, Non-Proprietary report). Further, Phase II will include an investigation of remedial actions to reduce turbine-generator- related reactor trips. The CE Cwners Group, through its Operations Subcommittee, is also conducting a trip reduction program. 4/ Recently expanded in scope, the program supports the five general categorTes of plant-specific scram reduction activities discussed in INP0 85-011: Administrative, System Design / Modification, M Institute of Nuclear Power Operations, " Scram Reduction Practices," INP0-85-011, May 1985. Limited distribution. 2_/ WCAP 10948, "U.S. Westinghouse Inadvertent Plant Trip Experience: A Historical View of Information from January 1930 through September 1985," Proprietary report, Limited distribution. 3_/ l WCAP 11156 "W Inadvertent Plant Trips Experience from October 1985 through l March 1986: Periodic Progress Report No.1," Proprietary report, Limited distribution. N Letter dated March 10, 1987, to Mr. Frank Miraglia, Director Division of PWR Licensing-B, USNRC, from J. K. Gasper, Chairman, CE Owners Group. 69 {
Maintenance, Surveillance Testing, and Operations. Past activities have included: (1) evaluation and sharing of experience regarding B0P turbine / reactor trip initiators, especially that involving the main feedwater system, (2) support for improved site-specific root cause investigations, (3) investigation of possible relaxation of certain Technical Specification surveillance frequencies, and (4) quarterly meetings to review program activities and share plant-specific scram reduction information and related experience. Other activities are in the early stages of development at this time. Expedited application of the results of scram reduction programs to new plants would be beneficial, but has not yet or. curred.
- 7. CONCLUSIONS :
It has long been recognized that newly licensed reactors experience a higher frequency of unplanned operational events during their early years of operation i when compared to later years. Rather than resulting in efforts to improve this situation, this observation has historically been translated into segregating the statistics for this period from general industry figures. While this is quite appropriate from the statistical viewpoint, it tends to l reinforce the view that (1) startup experience is less important and safety l significant than post commercial operation, and (2) the high frequency of j unplanned operational events is an inevitable and hence acceptable part of i early operation. The good experience in different operational areas of the plants reviewed in this study demonstrated that frec,uent difficulties are not inherent in the startup process. For this reason, the tacit assumption that new plants are expected to have high unplanned operational event frequencies should not be accepted without question. During this study, AE0D found that, generally, power reactor licensees have recognized the need for action to achieve good early operation. Some have developed programs in the areas of training and operations that, if implemented effectively, would result in improvement. This study confirmed that it is possible to achieve significant reductions in the frequency of reportable events for new plants. The data analysis of the operating experience provided a pointer to focus discussions with licensees and identify a number of lessons that could be used to achieve improvements. Both the NRC and industry have recently taken additional steps to focus on ' newly licensed plant operation, including the NRC senior management focus on this topic in October 1986 with NUMARC and the recently proposed INP0 program for improving performance during early operation. The findings of this study reinforce the need for and benefit of these efforts. For a new plant, an aggressive root cause determination and corrective action program for responding to failures is a key ingredient to improvement, but may not be sufficient. A reactive posture with respect to responding to failures should be supplemented by an aggressive preventive program that results in avoiding these challenges. The improvement measures identified in Section 5 of this report represent an effective action list te consider in formulating a program to prevent high reportable event frequencies. 70
\
l The analyses of operational data for scrams, ESF actuations and loss of system safety function clearly demonstrate the need to correct the root causes for reportable events early in life. Without correction, these root causes will : likely exist during the early commercial operation of the facility. At this I time of life, the relatively high challenge frequency coupled with the potential of undetected systems problems may present a significant challenge to a new operating crew. Therefore, the analyses indicate that increased dttention to operations is appropriate during the period of early commercial operation. Nuclear safety measures consist of prevention and mitigation items. It must be noted that this study identified measures to reduce the frequency of reportable events. Therefore, the improvement initiatives are all of the prevention type. Mitigation measures such as operator training for response l to plant transients continue to be important, but were not the focus of this review. l l t
.1 l
1 71 1
1 1 APPENDIX A PERFORMANCE INDICATOR DATA l l . i A-1
PERFORMANCE INDICATOR DATA Plots for six NRC Performance Indicators are provided in this Appendix. The indicators are defined as follows: Total Scrams - The number of unplanned automatic scrams that occur while the reactor is critical. (Figure A-1) Forced Outage Rate - The percentage of time planned for electrical generation that the unit was unavailable due to forced events. Forced events are unplanned failures or other conditions that require removing the unit from service before the end of the next weekend. (Figure A-2) Equipment Forced Outage - The inverse of the mean critical time between forced outages caused by equipment failures. (Figure A-3) i Significant Events - Events identified by the detailea screening of operating experience by NRR. The events selected are: degradation of important safety , equipment, unexpected plant response to a transient or a major transient; degra- 1 dation of fuel integrity, primary coolant pressure boundary, or important associated structures; scram with complication. A scram with complication is an RPS actuation, when critical, followed by an equipment failure or malfunc-tion or operator error. (Figure A-4) Safety System Actuations - The sum of the number of unplanned emergency core cooling system (ECCS) actuations that result from reaching an ECCS actuation setpoint or from a spurious or inadvertent ECCS signal, and the number of energency ac power system actuations that result from the loss (deenergiza-tion) of a safeguards bus. (Figure A-5) Safety System Failures - The number of times the plant experienced an event or condition that alone could have prevented the fulfillment of the safety function of structures or systems that are needed to: (1) shut down the reactor and maintain it in a safe shutdown condition, (2) remove residual heat, (3) control the release of radiation material, or (4) mitigate the consequences of an accident. (FigureA-6) The Performance Indicator data base covers 1985 and 1986. For each indicator, data was taken from the PI data base and assigned to the proper quarter following initial criticality. The individual values assigned to each quarter (both the number of values and the plants involved varied from quarter to quarter) were averaged to develop a composite curve for each indicator. The standard deviation for each quarter was also calculated and plotted to provide j perspective on the spread of individual values about each mean. i l A-2
N V T A S D E G . E L C I T A M O T L p U A , A C , I D F E I N R N C A L M P O N R U F 3,
- S M
S A T R N C : A S L P W E f= i N 3 1 z, 4 1 a
' i r
s 4j 7 1 s
- o A :
e r u g - ~ _ - - _ - _ i - - . - - _ - _ F
- e 7 . 5 4 3 2 ,
o r5 6a. $ g $v c$ 5O o iI ! , i I r~ , ,l l
l !)I - ilII , I' D N R O A I T D A N I D G A V V T E N A S D E G o E L 2 1 T 4m 4 R 4 1 Q E 4 4? 1 T R T A Q0( R , s 1 RT S 1 TI TE , 9 Q LA
,[ R NG "* T C AA I
Q 8I T LT RR. . PU T C, Q s O . 7 R La i W . T Am ED i Q I NE C 2 6T-M N 1 I n QI a.s 4 R 1 M 5 R Rno O 1 1 T Em F Q 1 w 4 - w i Rt s o T A.o 4 Q e 3 R1 4 m '- R T T .- 2 Q Quo 1 0 2 _ 3 : R e T s Q a
. 2 u.
o 1 A s : R m T r e u r u Q u _ g r _ i F - w t 0 0 0 0 0 0 0 u _ 7 6 5 4 3 2 1 n _ : . n
%O n c wo4OO awon b O c
a
' I 3 y*
l iI
N I D G A V V T E N A S D E ~ G o E L E 5 G 5 1 A 1 2 T ' R T U Q 7 O 5 S 1 T D 9 R N E 7 0 T AC 1 Q L P R 1 t. 6 O 7 F 3 R W l 4 1 T E T Q N N r E M P I U Q E
~
~
~
~
8 8-1 3
/
0, 1 3 L E. 5 R T Q 3 R
._ T Q
5 \ 2 : 3 3 9
- 0 A 3 :
d e r u g i F ~ - ~ - -
- 5 4 3 2 4 o
%o $o_ (Dxdb 8e Oi t
N A I D G A V V T E N A S D E G o E L 2 1 e' R T s./ - 1 1 Q s R S T T Q 0 Y N s N s. 1 R T I S E TL T V N E e./ 2 9QA R T C I SI T Q A LT 3./ 7 RR PN
- TC Q
A , N1 7 ( 3 R L WC I T Q I
- E F 6Tn
_. N N I RI T I Q G 5 R R
- V :s I ;
S . T E Q 5 4 T 7 \ R F TA Q
. ~
6 _ ~
~
~
- 0. 3 R
~ 1 R T
- ~ T
~ Q Q
_ ~ 3 2
. ~ 5 R
_ ~ 1 5 T
-. 4 1 Q
_ A ( R e r T
- u q~
~
Q
. i F -
~
3 0 - 5 0, s. = - O
# 3 3 2 2 i aEr 3 _
8a gEto . M < O._% 5 _ - ii: I I l' ,i 7"
N I 9 G A V y V T E A S D q o 3 S , N - O I T nr4 A ST U m.s TC NA j 1 A LM . PE . T .- WS 6-EY - NS .
. . s4 Y 3 T 9 E
F A 1, S
.c 8
0 2 _A 1 5 -
- 2
- 3 e -
r u g i F - ~ - - -
- ~ - - -
7 5 # s 2 ' 0 Ng "oLtC O ES 7E m .s uO ok c' I i' -, l l; jii IjNl
N I D G A V E N V T A S D E G o E L S g g)/ E R , U L I 7, t. SA T F 1 N 1' AM s LE s. PT S WY E S N Y
- r 7,
- r F
E.
- A / ew S .
6 A e O r u g i F - - - -
- s. 0 5 0 O 3' 2 2 1 1
$a $a. )WB
/
t n 7Eeb <M 1i i1
?*
APPENDIX B CAUSF CATEGORY DEFINITIONS i l i l l B-1 , I
CAUSE CATEGORY DEFINITIONS Summary Of Definitions Of Cause Categories Used In This Analysis Cause Category Definition Design Causes associated with decisions or events that occur before the plant is operational (e.g., a design /manu-facturing/constactioninadequacy),orwhichare associated with changes made to plant design under 10 CFR 50.59 after OL issuance. This type of cause is usually outside the control of plant personnel. Equipment Any equipment or hardware problem that is not related to design. This type of cause includes such factors as mechanical failure, electrical failure, materials inter-action, chemical reaction, electromagnetic interaction, fire / smoke, impact loads, liquid and gas system pressure problems, abnomal temperatures within the plant, and vibration loads. Personnel Human actions of omission and commission and accidental Error human actions committed during plant operation and mainte-nance. This type of cause includes hazardous or extreme working environment (e.g., high heat, excess noise, steam leakage, or high radiation), poor or improper training, or being unfamiliar with the power plant. It also includes actions outside normal operation of the plant if plant employees are involved (e.g., cause failure doing something other than the performance of their jobs). Other Clearly does not fit in any other cause category. Included in this category are natural phenomena, such as acts of nature. This category also includes moisture intrusion or water vapor within the plant atmosphere. The relatively i few ESF actuations which occurred in response to an actual plant condition, though still generally unneeded, are also included. Procedures Any procedures inadequacy (ambiguous, incomplete, erroneous). i Unknown The cause for the event could not be identified, either as a result of the licensee's investigation, or else it was not provided in the LER. This type of cause includes those ESF actuations described as spurious for which no other cause category could be assigned. l B-2
i APPENDIX C l SCRAM DATA FOR NEW PLANTS l The first 22 figures in this Appendix show cumulative scrams vs. cumulative 4 i critical hours for the 22 plants in this study as of June 1986. The second 22 show scram counts as a function of month since initial licensing. Critical hours for each month are also shown by the line plot. The figures in each set are ordered from earliest to most recent license date within NSSS in the following order: Westinghouse, CE, GE l l l C-1
ilIiI (( l
!lll 0
0 0
- 5 1
S M A - R C S R o 0 0 S R O i I 0 0 - T 0 9 C 1 H A E L R 2 A C I D E T E R N I I N U R A G C L C P M E N V U I T E A V 0 L I i I 0 U T 0 M A 5 U L C U ' M U C 1 C e r u g i F -
- ~
G0 0 0 O . 4 2 EaMsE5e8 oA
!lll!Illlj1 I 0
0 0
- 5 1
S - M A - R C S 0 S 0 R R i 1 0 U . O 0 O - T 1 H C 1 A L E N A R O C Y I D N T - E A I N C R - N C = A O L L E _ P B V N A U I
# I T
D ' A E A~ 0 L V , 0 U I M Y 0 I T 5 U A C L U M o U C 2 C e r u g i F o0 0 0 e 4 2 mI4 Ow 1> J 1
.- ?S pw
ll 0 0 0
- 5 1
S M A R C S t 0 S R
-- 0 R i I 0 U O 0 O T 1 'H C
A L E A R Y C I D W A T E A I N L R N C A L A L C E P V N I U T A E 0 L V , I 0 U I 0 M T 5 U A C L U M U C 3 C e r u g i F - 0O 0 e 0 4 e E=M E !5"
=
m?- I lll l ll lll11l , ll1
l\ 1\ j a 0 0 0
- 5 1
S M A R C S 0 S 0 R R i I 0 U O O T 0 H 1 C A L E A R 1 C I D A T E B I N N R N A C A T O L A E P C V N I U T A E V I T A L
^f -_
I 0 0 0 5 L U M U C U - M -_ U C n 4 C e _ r u g i F _ 0O _ 0 0 0 4 2 eaM $Gix3 = p*
0 0 5 S M A R C S 0 R 0 O I ; 0 T 0 C 1 A E o R D 1 o E N N O N R 0 A Y L B 0 P N U ^ E 0 V I l 0 I T 0 A 5 L , U M U C 5 C e r u g i F c8 0 0 O 4 2 m 3 M m E s J. ,. ! O P' I
0 0 0
- 5 1
S N A R C S 0 R 0 O g I 0 T 0 1 C A E _ R K _ E D E R E C _ N N F A L L O P W - N U E 0 V ; I 0 I 0 T 5 O A L U O M U _ C _ 6 _ C _ e _ r _ u g _ i F ( O 0 0 0 4 2 EaMw>%N8 p"
' i
l:il ll!! !Ili l1\ e e e
- s i
S - M A R C S e S e R R I I e U O O T e H C 2 t _ A L - E N A R O C Y I D N T E A I N C R N C A O L L E P B V N A I U I T D A E e L V i e U I T I e M U A s C L - U M - U n C 7 C e r u g i F
~
ce e e 4
> : )
vZ<QUU 1DZDU O
- - ~ - - __
G l Ln M 2 2 U U) G M G M g D S
~ -
@ 8 l-.- G %
O v4 4 J g U w " Q Z h WS " Z H E Z M V
< J J J W
- a. m Z Z $
D s 4 w G J
> ~ _ @ D
@ Z H LD D
< U J
D 2 D U b T u a E l 00 m o G g e T SHVBDS 3Allyqngn3 l C-9
l 0 0 0
- 5 1
S M A R C S 0 S 0 R R U O j I 0 O T 0 H C 1 A L E A R 2 C I D A T E B I NW R N A C A T L A E P C V N I U T A E 0 L V , 0 U I I T 0 M A 5 U C L U - M U C 9 C e r u g i F LO 0 0 O 4 2 EsM s!=S nb
0 0 0
- 5 1
c c S M A R C S 0 R 0 O l I 0 T 0 C - 1 A E 2 R E c _ D I c E C _ N U N L A c L . P T N S c U
^
_ E 0 _ V ; 0 _ I I T 0 _ A 5 _ L U ' _ M O U _ C _ 0 1 C _ e r _ u _ g i F c0 0 0 O 4 2 sM w> e5$U p:: .
, l
l4 1 0 0 0
- 5 1
S _ M _ A R C S 0 S 0 R R I I 0 U O 0 O T 1 H C A L E 3 A R C D I D R T E O I N F R N R C A E L T E _ P A V NW I _ U T _ A E 0 L V l 0 U I I 0 M T 5 U A C L _ U M U C 1 1 C e r g u - i F L0 __ 0 4 0 2 e y&= >- 3a!o 0g
,[ .1l 1lllL
1iiI ' 0 0 0 _
- 5 _
1 S M A R C S 0 S 0 R R i i 0 U O O T 0 H C 1 A I L E A R E C I D D R T _ E N E I - V R - N C A O L L E P A V N U P I T A E 0 L V , 0 U M I i T 0 A 5 U C L U M U C 2 1 C e r u g i m0 F 0 0 O 4 2 EaM WUiE" = oS
0 0 0 5 1 S M A R C S 0 R 0 O I I 0 T 0 C 1 A 2 E RE D D R E E N V N A S L L P A N P U E 0 V l i 0 I T 0 A 5 L U M U C 3 1 C e r u g i F G0 0 e O 4 2 E MWp5e8 er
0 0 0
- 5 1
S M A R C S 0 R 0 O
; I 0
T 0 C 1 A E R 2
*c D E E L N L c
N A c A S L A c P L N U c E ^ 0 V ; I 0 I T 0 A 5 L U ' M U C 4 1 C e r u g i F c 0 0 0 O 4 2 E Mw>4!5" = p
l4 )Il 0 0 0
- 5 1
S M A R C - S 0 S R O T C 2 f I 0 0 0 1 R U O H A n~ L E R A ._ R A C - D L E n~ I T E N N A C U N "g I R C L . E P H V N S - 1 U A T _ W A E 0 L V ; 0 U M I I T 0 5 U A C L U ' M U C 5 1 C e r u g _ i F cO 0 0 0 4 2
=aYn E - e8 Z=
0 _ 0 0
- 5 1
_ O S M o A R _ C _ S S _ 0 R 0 R O I 0 U T 0 O C 1 H A 2 L E A R A C N I D N T E A I N H R N E C A U L O E P S V N U I U S v T E A V I c I 0 0 L U T A c 0 5 M U C L U M c U ^ C 6 1 C e r u g i F 0 0 0 0 0 4 2 EaM $ 3E8 pt:
0 0 0
- 5 1
S M A R C S 0 0 R l I 0 O 0 T 1 C A E R K C D I E R N E N M O A L I O P L N O U E ^ 0 V l I 0 I 0 T 5 A L U ' M U C 7 1 C v e r 0 u g fO i F 0 0 g 4 2 wI< mow w> i<J.oIaU og
0 0 0
- 5 1
S _ M A _ R _ C _ S S _ 0 R 0 R _ O I 0 U T 0 O C 1 H A L E A R M C D A I E H T N E I N R R O C A H L S P E N V I U T E A V 0 L I i I 0 U T 0 M A 5 U L C U ' M U C 8 1 C e r u g i F cO" 0 0 e 4 2
$aM W6!EU =
o$ i
0 0 0 _ 5 1 S M A R C S 0 S 0 R R I 0 U O l 0 O T 1 H C A L . A E . R C . l. 2 D T I E I _ N M R - N R C - A E L F E P V N I U T _ A E V 0 0 L U I l I 0 M _ T 5 U A C L U ~ M U C 9 1 C e _ r u g i F - cO 0 0 e _ 4 2 EaM w> Qs!U - og
0 0 0 _ 5 1 S M A R C S 0 R 0 O l I 0 T 0 C 1 A E D R N D E E B N R N E A V L I P R N U E 0 V I l I 0 T 0 A 5 L U ' M U C 0 2 C e r u g i F LO 0 0 0 4 2 EaM E%l=50 p*
0 0 0
- 5 1
S M A R C S 0 S - 0 R R i I 0 U O 0 O T 1 H C A L E A _ R C I _ D Y T _ E R I N R R N E C A P L E . P V m N I U T A E 0 L V 0 U I
; I 0 M m T 5 U -.
A C - L U ' M U C 1 2 C e r u g i F - dO 0 0 0 4 2 " E 8 E 5e8 pN
0 0 0
- 5 1
S M A R C S 0 R 0 O I 0 T 0 C 1 A E K R E D E R E C N N E A P _ L O P N H U _ E 0 _ V , 0 _ I I _ T 0 _ A 5 L U ' M U C 2 2 C _ e _ r _ u _ g _ i _ F oCO 0 0 O 4 2 E M 's ! 8 mb
)
$ g3.fU 0
0 0 0 0 - 0 0 0 0 1 8 4 2 0 r , o r m- 4 r 2 te s nl e n 2-ee r 2 n m n douw , 2 n g po ren i s eck S i surho eqetrn 0 M DEPOPU 3
- 5i i 2
A. A_6 v05l2 G mr R i
, 8 1 e C o c S n 6
1 a R u s O2 X 4i s T E K 1 L C AIR 2 O 1 e EU ' c RG - i n C 1 - 0 1 s D E M i1 s h N 8 t n N o _ A 6 M L P , N .
. - 4 U
% I 2
3 2 c e
. 0 r
u , s i r - 5 5 0 1 O$ yZ
?%
sO4 g2. - U 0 0 0 0 0 0 0 0 0 0 0 - 1 8 6 4 2 0 r , o r g 4 r , 2 te s nl e n ee r - n uw , 2 n m n do 2 g po r en S i i s ec surhok eqetrn 0 M DEPOPU 2 A A1 5l 1 V55l22G x I R iL r 2 2 8 1 e C 1 - c _ S n _ o 6 a RN O P 1 u s O~ s T Y 4i CN 1 L AA 1 O 2 E C 1 e - c RO - i n - D E B L - 0 1 s s h NA I 8 t n ND o A 6 M L P N - 4 U i - i - 2 4 2
- m -
C e r i 0 u g , i F 5 0 5 0 1 1 m bm%O b 3 Z Em
r o r r e e n n w n n o g o S i s e k n M D U A 2 Z R ( C S R OY TA CW AA EL RL D C A E N N A L P o N U f 5 2 C e r u g i F 5 0 5 1 1
$:O u Eg2; 2*
1 8 r o r r te s nl ee e n r \ 1
\O k n uw m n do n
g po r en S i i surho s eck eqet rn Di1 M DEPOPU A 25 222DCn 2 E25I D R ) C S N j k h
\
R - O1 \{ TA - C A B E W A - b RT of _ t DA E C O b N O h N b hP A > - L ' P f I - N U 6 _ 2 _ C _ e r u g i F - - i i - 5 0 5 1 1 - ) f$sbm% ( e $ gZ n6
r o r , r e l e n n w , n n r o g n s e S i s e rht k n M D O A 2 Z I G R i l - C , S - o ) R S - O T1 6 h CN - AO h ER - RY , DB .; b E ' N N A - L P
,i N o U
i
}
- 1 7 2 . C e r . u g , i F 5 0 5 g 1 1 h bm obM z ob
1 8
- 4 2 r ,
o r , r t
,e s nl e n ,
ee r n uw , n m n do g po r en s eck S i i surho eqet rn M DEPOPU A 3I 5 2 Z2GD E5iCn R [ l C , S , RK o h OE T E , CR o k A C i l E ]< , RFL b DO EW N N A yN L P b N l q\ U ; 8 2 c
= 3 9 ,
5 0 5 g 1 1 mbMN %o E gz e
r o r r t e s nl e ee rn n uw n m n do g po r en S i i s ec surhok eqet rn M DEPOPU A 235i3 ZE2GsO E5liC R ( C S 2 RN OO T C N Y AA b EC RO D L k E B b NA I ND - A , L P N U 9 2 C e r ug i F 5 0 5 1 1 n bd c % pc)M Z
?
r ' o r ' r te s nl e n ' ee r mnn uw ' n do g i po ren ' S i s surhoeck eqet rn M DEPOPU A 2I 5lD Z2Gi5ECO E E' R ' C ' S ' R3 O E T N ' C ' AO E T S RL ' DI L EM N N A L P o N ' U lO; 0 3 c e r u 9 n - - 5 0 5 1 1 mbobm %O 546yZ c
?O
r o r r t e s nl e ee r n n uw m n do n g po r en i S i s surhoeck eqet rn M DEPOPU A 21Z !G3I 2 EGl iEr-R [ C S R O 2 TA C B A W E R T A DA E C N N A L P N U o - 1 3 c e r - u s - i - r 0 5 1 mbab?oc E4$y2;
?M
1 r ' o r r ( (' te s nl e ee r n n uw n m n do g po ren i s eck S i surho eqet rn M DEPOPU A 335 Z 22ECO 55 lid g\ R ( C S R O2 TE CI AC EU RL . DT E S N N A L . P I N U ;f o
?
2 3 C m g i F - - 5 0 5 1 1 m b* oE gZ oS
0 0 0 0 0 1 8 6 4 2 r , o r , r e e n n w , n n o g o n , S i s e s k n M D U A 2 E R o 3 C o 3 S - 7 . R3 o OD TR l . CO AF E R RE , DT A O E I [C NW N - A ! L ' P N P U F
/
C I 3 3 c L e .b O r u 9 . i r
~ -
5 5 1
- m b?ob yZ
?y l
i mEDO$ qa.fU 0 0 0 0 0 0 0 0 0 0 0 1 8 6 4 2 g r , o r , 4 r 2 te s nl e n , ee r mnn uw , 2 2 n do g i po r en , S i s surhoeck eqetrn 0 M DEPOPU 2 A A5521 52 VE5ZL N i , R r 7 b 8 1 e C io , c - S n 1 _ 6 1 a - R O b u s OE O C 14i s TD - CR b L O AE o 2 1 e EV c R O i n 0 s DL 1 s EA 0 b h NP 8 t n N o o A o b P F 6 M L C P jO , I N _ 4 U 3 3 2 3 c L e r 3O0 a 9 , i r - - 5 0 5 g 1 1
*bd Om%O b C gZ
?m
.l ljl{,
r o r r t e s nl e ee r n n m n douw n g po ran S i i surhe s eck eqet rn M DEPOPU A 2352D ZE5ZCn 22 R C S R 2 OE TD CR AE EV R O DL EA NP N A h L P o 5 N U s a c e r u s i r - 5 0 3 - 1 1 n n cbMo< . O L c) M b Z c PM
r o r r t e s nl e ee r n n m n do uw o n g po r en i s eck S i surho eqet rn M DEPOPU 1, f A 31 5I ZE5lZ2GSO iC D f R C S R O2 T E ? C L AL EA RS DA E L N o N A L P N U 6 3 C e r u g i F 5 5 1 1
$cn %o e gZ Z~
l '
r o r r t e s nl e - ee r n nn uw nrn g po do en S i s ck i su o eq rn M DE PU A 235 2 22 [ 25
]
5n R [ C S 2 M' RR OA TE o C L l AC EU RN . DH ES NA l NW A L o \ P N U 7 3 C e r
^
u g i F - - 5 0 5 1 1
*b O bM D*
M
r o r r te s nl e ee rn m n uw n n do g po ren S i i s ec surhok eqetrn DEPOPU M 3352D Z' ( E 5ZCO 22 R T2 A NN TN EAH MEU Q ES MUS M W U . 8 - 3 C e - r u g i F 5 0 5 1 1 mbmb0 b gZ 2-
r o r r t e s nl e ee r n n uw m n do n g po ren S i i surho s eck e qet rn M DEPOPU A 1 2Z252DI CR5Z0r 2 R C S R o O K T C CI AR EE RM I DL E N N A L P N A U S 9 3 c e r u 9 4 r - - 5 0 5 1 1 ebdb* %O n ( O. b g Z o1o
r o r r t e s nl e ee r n n uw n m n do gipo ren s eck S i surho eqetrn M DEPOPU A Z122 Z 55ZG523n R ( C S R OM TA CA H E ER RO 0 DH ES N N f A L - P N U 0 4 C e r u g i F - ~ 5 0 5 1 1 m b m $ m % O L a> D b 3 z mi~ , f
r o r r t e s nl e ee rn n uw m n do n g po r en S i i s surhoeck eqet rn M DEPOPU A 5I 21Z2DD-(Z E5ECF I R C S R O T2 CI AM E R RE DF E N N o1 A i L P N lF 4 U ; 4 C e r u g i F - 5 0 5 - 1 1 bDZ mbM om %0 L
?e
r o r r t e s nl e n ee r n uw m n do n g po r en i s eck S i surho eqet rn M DEPOPU A Z352DO 22 Z25ZC R C S RD O T N C E B A ER RE _ DI V p - ER N o N A L P N U 7 2 4 C e r u g i F 5 0 5 1 1 mbM$* %o L Z n,0 i
r o r r t e s nl e ee r n n uw m n do n g po ren S i i s eck surho eqet rn M DEPOPU A 3352DO [ 22E5ZC Z R C S R O T CY AR E R R PE D E N N A L P - N U 3 - 4 - c e i w r 5 0 5 1 1 mbebm%O bO.6gZ
?*
r o r r te s nl e ee r n mnn uw n do g i po ren S i s surhoeck eqet rn M DEPOPU A 23523 Z25ZGn 22 R C S R O K E T C E A R E C REP _ DO EH N N _ A - L P N U 4 4 - C - e - r g u i F 5 0 5 1 1 mbd$mu E O. b g z ob I
-l l
l 1 l APPENDIX D FORMAL STATISTICAL ANALYSIS i l l D-1
1 FORMAL STATISTICAL ANALYSIS I
- 1. BACKGROUND For each type of data, i.e., scrams, ESF actuations, technical specification violations, and losses of system safety function (LSSF), formal statistical analyses were performed for the startup (pre-commercial) period for those plants which became commercial (i.e., completed the power ascension test program) prior to June 1986. In addition, analyses were performed for the commercial period consisting of the first 180 days after the date of commercial operation for those plants that reached that milestone prior to July 1986. For these analyses, the units for the event rates were average scrams per 1000 critical hours for the scram data and average number of events per day for the other data types. The per-day rates are also converted to the equivalent per month rates in summary tables at the beginning of the presentation in Section 4 for each data type (Tables D-2 through D-5). For each type of data, three analyses were performed:
(1) Average startup rates versus the following six variables: The length of the startup period (days) Calendar date of operating license First or subsequent nuclear unit at a site Type of Architect Engineer - utility /non-utility First or subsequent nuclear plant for a utility NSSS vendor - Westinghouse, CE, BWR-4, -5, -6 (2) Average commercial rates versus these same six variables, and (3) Startup versus commercial period rates. The values of the six variables assigned to each plant in the analysis are shown in Table D-1. In Table D-1 (and Tables D-2, 0-3, D-4 and D-5) the plants are grouped by NSSS vendor, and then arranged within NSSS vendor from earliest to latest license date. Because they had not completed the startup ) phase (i.e., declared commercial) prior to June 30, 1986, the following i plants were not included in the analysis: Catawba 2, Palo Verde 2, Shoreham, Fermi 2, Perry 1, and Hope Creek. Also, Diablo Canyon 2. Limerick, River Bend, Millstone 3, and Palo Verde 1 were not used in the analysis of post commercial rates since they had not accumulated 180 days of post commercial service prior to June 30, 1986,
- 2. OVERVIEW 0F STATISTICAL PROCEDURES The two rates under study for each data type are quantitative; so is the length of the startup period. OL date can also be considered quantitative; the statistical software treats it as the number of days since January 1,1960.
Thus, correlations among all these variables were investigated by computing- , the Pearson product-moment correlation coefficients. The significance tests D-2 , l { o
I that are associated with these estimates are equivalent to t-tests of whether j the slope of regression lines for each pair of variables is non-zero (Ref. D-1). j Three other variables under study as possibly being associated with the event rates, type of AE, first or subsequent nuclear plant for a utility, and first or subsequent plant at a site, have two values each. Pearson product-moment correlation coefficients for these variables are also meaningful; here, significance tests are equivalent to two-sample t-tests for differences in the means of the rates. For these tests, eoual variances among the rates for, for example, the set of first nuclear units at sites and those that are later units at the site are assumed. The final variable studied for associations with the rates is the NSSS vendor. Since it is not ordinal and has more than two values, an analysis of variance was performed to see if significant differences exist among rates of each type when they are grouped according to this variable. A set of nonparametric procedures (Ref. D-2) was also applied for these tests. Significance levels for the non-parametric procedures do not assume that the rates are nonnally distributed. The Statistical Analysis System (SAS) software package was used for the statistical analyses (Ref. D-2). For each type of data, all of the analyses using the Pearson product-limit correlation coefficient were accomplished using Procedure CORR. Analyses involving the NSSS variable were accomplished using the following methods: l - General Linear Modeling: SAS Procedure GLM was used instead of j Procedure ANOVA because the number of rate observations for each value of NSSS is not constant.
- Nonparametric tests: Wilcoxon Scores, Median Scores, Van der Waerden l Scores (normal) and Savage Score (exponential).
SAS Procedure NPAR1WAY was used for the nonparametric tests.
- 3. RESULTS The exact level of significance required to declare a variable statistically significant depends on the number of multiple comparisons being made. The probability of observing high test statistics when there are no significant differences increases as the number of tests increases. The Bonferroni correction to maintain the overall level of significance at, say, 0.05, is to use 0.05/n er the significance level of each individual test, where n is the number of nultiple comparisons being made (Ref. D-1).
D-1/ Snedecor, G. W., and Cochran, W. G., Statistical Methods, Second Edition, Iowa State U. Press, p. 185, 1980. D-2/ SAS Institute Inc., 1985, SAS/ STAT Guide for Personal Computers Version, Sixth Edition, i D-3
For this analysis, five sets of tests were performed for each data type, as indicated below: Possibly Required Correlated Significance Varfable Variables n_ Level Startup Period OL date, Length of Startup, 5 0.01 Rate AE Type, First Unit at Site, First Nuclear Unit for Utility Startup Period NSSS 1 0.05 Rate (Proc's GLM & NPAR1WAY) Commercial Period OL date, Length of Startup, 5 0.01 Rate AE Type, First Unit at Site, First Nuclear Unit for Utility Commercial Period NSSS 1 0.05 Rate (Proc'sGLM&NPAR1WAY) Startup Period Commercial Period Rate 1 0.05 Rate (Proc CORR) Detailed results of the statistical calculations are provided following this discussion, grouped by data type. Using the significance criteria stated in the preceding table, the following correlations were statistically significant: Correlation Variable 1 Variable 2 Coefficient Significance Startup Scram Rata Commercial Scram Rate 0.64 0.033 Startup ESF Rate Commercial ESF Rate 0.68 0.022 Commercial Tech Spec Rate Subsequent Unit at Site 0.74 0.009 , Startup LSSF Rate Commercial LSSF Rate 0.81 0.003 The last correlation, that for LSSF, is based on very low event frequencies and for that reason should be discounted. Relaxing the significance level slightly in those cases where the Bonferroni correction was applied, i.e., where a significance level of 0.01 was used, brings up two more correlations as nearly significant (borderline): Correlation , Variable 1 Variable 2 Coefficient Significance Starfup ESF Rate Subsequent Unit at Site 0.60 0.014 l S -etup ESF Rate Subsequent Unit for Utility 0.55 0.026
. DETAILED OUTPUT i The following pages present the output for the statistical codes, graped by data type. Ee:h section begins with a table sunsnarizing the average event rates, followed by the Pearson correlation results, and ending with the GLM and non-parametric results. The variable names used in these calculations are provided in Table D-1.
l D-4
)
E P
- Y 44444444 eee 55446 ST SP wwwwwwww ccc bbbbb S(
N y t i rl ai et l U) ) c E ) 1 urP 0 = NoY = ( fT ( t L s s ss s s stP ooeooeoo oee oeooe riF NNYNNYNN NYY NYNNY i nN FU(
)
E f P o Y ) ) T 1 0 e#E = = p A ( ( YEN TA( UUNUNNUN NNN NNNNN t i ) n T ) UeI )1 tN 0=
=(
1 tiU _ - sSN ( _ D r F sssss ss s ss oeeeeeoo oee oeoee _ itN e NYYYYYNN Fa( NYY NYNYY l b _ a _ T _ e _ t a D r ) 33444555 344 33445 a E _ 88888888 888 88888 d T //////// /// ///// nLA 38181165 681 60369 eOD 00113122 013 12222 l P //////// /// ///// afL 31670341 422 22308 Co( 01001001 011 11010 6
) g s n ,
_ y i5 a _ pr- _ D oe _ f( oe o
) l n4
_ pD i- _ h uP 4 gR ttT grR 44260609 66942724 409 280 89631 05269 nW eEB naT 35133131 124 33342 s etS y unE _ LS( t ooG 1 2 i hi _ l gt= n n 2 yi ns o o 1 tt iu6 y y3 23 a iu tbb _ n k n e n1d l - sm, 2a Cy 1 eae edd 2 n n in eo eCn irr ake to WC5 _ e aa1 r o coe e hcB UN b
==
rowb C ot ufV l ei t n il awn l s Lr l2urr == _ ubl aofbl eo a- qee a e4 _ a Gal trl al .tl SPsmv UN wcb _ l ciaayoii taa aNuii P MDCCBWDM SWP LWSLR i *
?*
Table D-2 Average Scram Rates Pre-Commercial Post-Commercial Scrams Critical Hrs Scrams / Scrams Critical Hrs. Scrams / Plant 1000 CH 1000 CH McGuire 2 11 3724 2.95 9 3416 2.63 Diablo Canyon 1 12 2745 4.37 5 4160 1.20 Callaway 12 1243 9.65 10 3917 2.55 Catawba 1 9 2266 3.97 3 3483 0.86 Byron 1 22 3746 5.87 5 2691 1.86 Wolf Creek 11 1746 6.30 5, 413S 1.21 Diablo Canyon 2 17 2474 6.87 N/A, N/A, N/A, Millstone 3 8 924 8.66 N/A N/A N/A St. Lucie 2 6 1101 5.45 7 4024 1.74 Waterford 3 21 1812 11.59 9, 3291 2.7) Palo Verde 1 13 3437 3.78 N/A N/A N/A LaSalle 2 9 3871 2.32 2 3003 0.67 WNP-2 23 3950 5.82 8 3006 2.66 Susquehanna 2 7 3794 1.85 2, 3741 0.5} Limerick 4 4862 0.82 N/A, N/A, N/A, River Bend 16 3541 4.52 N/A N/A N/A Average 5.30 1.69 Note: These plants had not accumulated 180 days of commercial operation before June 30, 1986. Therefore, their post-commercial experience was not considered in the statistical analysis. I 1 0-6
NEW PLANT ANALYSIS SCMCORR Va ri abl e N Mean Std Dev Sum Hinimum Maximum LPDATE 16 8974 289.76536 143579 8462 9460 STRTPD 13 310.81250 112.43559 4973 124.00000 546.00000 AVGSTSCM 16 5.26236 2.78330 84.19781 0.82264 11.59008 AVGCMSCH 11 1.68266 0.83751 18.50921 0.53456 2.73462 NAETYPE 16 0.25000 0.44721 4.00000 0 1.00000 NFPLTYPE 16 0.37500 0.50000 6.00000 0 1.00000 NFNUNIT 16 0.62500 0.50000 10.00000 0 1.00000 Correlation Coefficient / Prob > lRl under Bo: Rho =0
/ Number of Observations LPDATE STRTPD AVGSTSCM AVGCMSCM NAETYPE NFPLTYPE NFNUNIT LPDATE 1.00000 -0.19481 0.37667 -0.09786 -0.26173 0.30456 0.26648 0.0000 0.4697 . 0.1504 0.7747 0.3275 0.2514 0.3184 16 16 16 11 16 16 16 STRTPD -0.19481 1.00000 -0.52350 -0.12867 0.44117 -0.18722 0.32478 0.4697 0.0000 0.0374 0.7061 0.0872 0.4875 0.2197 16 16 16 11 16 16 16 l AVGSTSCM 0.37667 -0.52350 1.00000 0.64406 -0.14452 0.44540 0.16631 0.1504 0.0374 0.0000 0.0325 0.5933 0.0838 0.5382 16 16 16 11 16 16 16 AVGCMSCM -0.09786 -0.12867 0.64406 1.00000 -0.08956 0.54022 0.27356 0.7747 0.7061 0.0325 0.0000 0.7934 0.0862 0.4157 11 11 11 11 11 11 11 NAETYPE -0.26173 0.44117 -0.14452 -0.08956 1.00000 -0.44721 -0.14907 0.3275 0.0872 0.5933 0.7934 0.0000 0.0824 0.5816 ('
16 16 16 11 16 16 16 NFPLTYPE 0.30456 -0.18722 0.44540 0.54022 -0.44721 1.00000 0.60000 0.2514 0.4875 0.0838 0.0862 0.0824 0.0000 0.0140 16 16 16 11 16 16 16 NFNUNIT 0.26648 0.32478 0.16631 0.27356 -0.14907 0.60000 1.00000 0.3184 0.219; 0.5382 0.4157 0.5816 0.0140 0.0000 16 16 16 11 16 16 16 ( D-7 E _ .
NEW PLANT ANALYSIS SCMCORR General Linear Models Procedure
- mpendent Variable: AVGSTSCM Sum of Mean 3ource DF Squares Square F Value Pr > F Model 4 47.69895608 11.92473902 1.91 0.1781
{rror 11 68.50236392 6.22748783 Corrected Total 15 116.20132000 R-Square C.V. Root MSE AVGSTSCM Mean 0.410485 47.421539 2.4954935 5.26236282 Source DF Type 1 SS Hean Square F Value Pr > F
> TYPE 4 47.69895608 11.92473902 1.91 0.1781 Source DF Type III SS Mean Square F Value Pr > F ' TYPE 4 47.69895608 11.92473902 1.91 0.1781 T for HO: Pr > iT! Std Error of "arameter Estimate Parameter:0 Estimate INTERCEPT 5.968697516 B 6.77 0.0001 0.882290175 3 TYPE b4 -4.634812724 B -2.35 0.0385 1.972860808 b5 -1.895180914 B -0.96 0.3574 1.972860808 b6 -1.450163208 B -0.55 0.5947 2.646870526 ce 1.069598459 B 0.63 0.5396 1.689457062 w4 0.000000000 B . . .
90TE: The X'X matrix has been found to be singular and a generalized inverse was used to solve the normal equations. Estimates followed by the letter 'B' are biased, and are not unique estimators of tJie parameters. D-8
I i NEW PLANT ANALYSIS ) SCMCORR 1 General Linear Models Procedure Dependent Variable: AVGCMSCM Sum of Mean 3ource DF Squares. Square F Value Pr > F Model 3 1.93484929 0.64494976 0.89 0.4923 i Error 7 5.07944458 0.72563494 Corrected Total 10 7.01429387 R-Square C.V. Root MSE AVGCMSCM Mean 0.275844 50.624865 0.85184209 1.68265553 Source DF Type I SS Hean Square F Value Pr > F PTYPE 3 1.93484929 0.64494976 0.89 0.4923 Source DF Type III SS Mean Square F Value Pr > F PTYPE 3 1.93484929 0.64494976 0.89 0.4923 1 l Parameter T for HO: Pr > lTl Std Error of Estimate Parameter =0 Estimate INTERCEPT 3.695592025 B 4.88 0.0018 0.3477630754 I PTYPE b4 -1.161034530 B -1.26 0.2474 0.9200946127 b5 -0.032211612 B -0.05 0.9644 0.6955261508 b6 0.000000000 B . . . ! ce 0.541578165 B 0.78 0.4617 0.6955261508 ' w4 0.000000000 B . . . NOTE: The X'X matrix has been found to be singular and a generalized inverse l was used to solve the normal equations. Estimates followed by the letter 'B' are biased, and are not unique estimators of the parameters. l l l l D-9
NEW PLANT ANALYSI!! SCMCORR NPAR1WAY PROCEDURE Wilcoxon Scores (Rank Suas) for Variable AVGSTSCH Classified by Variable PTYPE Sum of Expected Std..Dev M 3 r.a PTYPE N Scores Under H0 Under H0 Se c re. w4 8 82.0 68.0000000 9.52190457 10. 25000t,r b4 2 3.0 17.0000000 6.29814788 1.5000000 b5 2 13.0 17.0000000 6.29814788- 6.5000000 ce 3 30.0 25.5000000 7.43303437 10.0000000 b6 1 8.0 B.5000000 4.60977223 8.0000000 Kruskal-Wallis Test (Chi-Square Approximation) CHISQ: 6.0662 DF= 4 Prob > CHISQ: 0.1943 4 i . , . . D-10
NEW PLANT ANALYSIS SCHCORR NPAR1WAY PROCEDURE Median Scores (Number of Points above Median) for Variable AVGSTSCM i Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under HO Under H0 Score w4 8 5.0 4.00000000 1.03279556 0.625000000 b4 2 0.0 1.00000000 0.68313005 0.000000000 b5 2 1.0 1.00000000- 0.68313005 0.500000000 ce 3 2.0 1.50000000 0.80622577 0.666666667 b6 1 0.0 0.50000000 0.50000000 0.000000000 Median 1-Way Analysis -(Chi-Fquare Approximation) CRISQ= 3.5937 DF: 4 Prob > CHISQ 0.4638 l l l l l l D-11
NEW PLANT ANALYSIS SCMCORR NPAR1WAY PROCEDURE Van der Waerden Scores (Normal) for Varjable AVGSTSCM Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under B0 Under HO Score w4 8 2.43411818 0.0 1.74830927 0.30426477 b4 2 -2.75155790 0.0 1.15639788 -1.37577895 b5 2 -0.70589166 0.0 1.15639788 -0.35294583 ce 3 1.09712266 0.0 1.36477348 0.36570755 b6 1 -0.07379127 0.0 0.84639658 -0.07379127 Van der Waerden 1-Way (Chi-Square Approxissation) CRISQ= 6.7814 DF: 4 Prob > CRISQ: 0.3479 l l l l l 1 1 0-12 - - - _ _ - _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - 1
]
NEW PLANT ANALYSIS , SCMCORR j NPAR1WAY PROCEDURE Savage Scores (Exponential) for Variable AVGSTSCM Classified by Tariable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under H0 Under H0 Score
- 1 w4 8 1.48468476 0.0 1.83443150 0.185585595 b4 2 .80833333 0.0 1.21336239 .904166667 l b5 2 -0.86867577 0.0 1.21336239 .434337884 '
ce 3 1.52945249 0.0 1.43200263 0.509817497 b6 1 -0.33712815 0.0 0.88809033 .337128150 Savage 1-Way (Chi-Square Approximation) CRISQ: 3.7814 DF= 4 Prob > CHISQ= 0.4364 l i D-13
NEW PLANT ANALYSIS SCMCORR NPAR1WAY PROCEDURE Wilcoxon Scores (Rank Sums) for Variable AVGCMSCM Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under H0 Under H0 Score w4 6 36.0 36.0 5.47722558 6.00000000 b4 1 1.0 6.0 3.16227766 1.00000000 b5 2 12.0 12,0 4.24264069 6.00000000 ce 2 17.0 12.0 4.24264069 8.50000000 Kruskal-Wallis Test (Chi-Square Approximation) CRISQ: 3.4091 DF= 3 Prob > CRISQ 0.3327 I l ! 1
]
D-14
i NEW PLANT ANALYSIS SCMCORR NPAR1WAY PROCEDITRE Median Scores (Number of Points above Median) , for Variable AVGCMSCH j Classified by Variable PTYPE j Sum of Expected Std Dev Mean PTYPE N Scores Under RO Dnder H0 Score : w4 6 3.0 2.72727273 0.862439362 0.500000000 , b4 1 0.0 0.45454545 0.497929598 0.000000000 ' b5 2 1.0 0.90909091 0.668042657 0.500000000 ce 2 1.0 0.90909091 0.668042657 0.500000000 Median 1-Way Analysis (Chi-Square Approximation) CHISQ: 0.83333 DF= 3 Prob > CBISQ: 0.8415 t D-15
NEW PLANT ANALYSIS SCMCORR NPAR1WAY PROCEDURE Van der Waerden Scores (Normal) for Variable AVGCMSCM Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under 80 Under H0 Score w4 6 0.00000000 0.0 1.38825982 0.00000000 b4 1 -1.38299413 0.0 0.80151218 -1.38299413 b5 2 0.00000000 0.0 1.07534143 0.00000000 ce 2 1.38299413 0.0 1.07534143 0.69149706 ! Van der Waerden 1-Way (Chi-Square Approximation) CHISQ= 4.0599. DF: 3 Prob > CHISQ= 0.2551 l 1 (
)
l l 0-16
NEW PLANT ANAI,VEIS SCMCORR NPAR1WAY PROCEDifRE Savage Scores (Exponential) for Variable AVGCMSCM Classified by Variable PTYPE Sun of Expected Std Dev Mean PTYPE N Scoces Under H0 Under H0 Score w4 6 -1.05811688 0.0 1.47526170 .176352814 b4 1 -0.90909091 0.0 0.85174274 .909090909 b5 2 0.21078644 0.0 1.14273280 0.105393218 ce 2 1.75642136 0.0 1.14273280 0.878210678 Savage 1-Way (Chi-Square Approximation) CHISQ= 3.2302 DF= 3 Prob > CHISQ 0.3575 l D-17
l Table D-3 Average ESF Actuation Rates Pre-Commercial Post-Consnercial ; Plant Actuations/ Month Actuations/ Month McGuire 2 0.15 0.33 Diablo Canyon 1 1.38 0.99 Callaway 7.00 4.17 Catawba 1 3.30 4.17 Byron 1 6.08 3.67 Wolf Creek 13.30 4.32 3 Diablo Canyon 2 1.89 N/A, 1 Millstone 3 2.44 N/A St. Lucie 2 0.24 0.00 Waterford 3.21 2.3( Palo Verde 1 3.25 N/A LaSalle 2 4.59 5.67 WNP-2 7.02 3.51 Susequehanna 2 1.47 2.4S Limerick 1 7.14 N/A, River Bend 6.90 N/A l Average 4.33 2.87 Note: These plants had not accumulated 180 days of comercial l operation before June 30, 1986. Therefore, their post-comercial experience was not considered in the statistical analysis. D-18
NEW Pl. ANT ANAI.YSIS ESFCORR Mean Std Dev Sum Minimum Maximum Verfable N 16 8974 289.76536 143579 8462 9460 LPDATE STRTPD 16 310.81250 112.43559 4973 124.00000 546.00000 AVGSTESF 16 0.14653 0.11586 2.34443 0.00549 0.~4 4 318
.31 0.09444 0.05999~ 1.03889 0 0.18889 AVGCMESF 16 0.25000 0.44721 '4.00000 0 1.00000 NAKTYPE 1.00000 16 0.37500 0.b0000 6.00000 0 NFPl. TYPE 1.00000 16 0.62500 0.50000 10.00000 0 NFNUNIT Correlation Coefficients / Prob > lRl under Ho: Rho =0
/ Number of Observations '
1,PDATE STRTPD AVGSTESF AVGCMESF NAETYPE NFPITYPE NFNUNIT i.PDATK 1.00000 -0.19481 0.40270 0.60219 -0.26173 0.30456 0.26648 0.0000 '0.4697 0.1220 0.0499 0. 3 P.7 5 0.2514 0.3164 16 16 16 11 16 16 16 STRTPD -0.19481 1.00000 -0.17747 -0.15179 0.44117 -0.16722 0.32478 0.4697 0.0000 0.5108 0.6559 0.0872 0.4875 0.2197 16 16 16 11 16 16 - 16 AVGSTESF 0.40270 -0.17747 1.00000 0.67808 -0.46640 0.55531 0.59943 0.1220 0.5108 0.0000 0.0218 0.0686 0.0255 0.0141 16 16 16 11 16 , 16 16 l AVGCMESF 0.60219 -0.15179 0.67808 1.00000 -0.35687 0.33040 0.31204 1 0.0499 0.6b59 0.0218 0.0000 0.2813 0.3210 0.3502 11 11 11 11 11 11 11 4AM7YPR -0.26173 0.44117 -0.46640 -0.35687 1.00000 -0.44721 -0.14907 0.3275 0.0872 0.0686 0.2813 0.0000 0.0824 0. 581 fi 16 16 16 11 16 16 16 NFi..T d'E 0.30456 -0.18722 0.55531 0.33040 -0.44721 1.00000 0.60000 0.2514 0.4875 0.0255 0.3210 0.0824 0.0000 0.0140 16 16 16 11 16 16 16 SFl.~.i: . 1 0.26648 0.32478 0.59943 0.31204 -0.14907 0.60000 1.00000 0.3184 0.2197 0.0141 0.3502 0.5816 0.0140 0.0000 16 16 16 11 16 16 16 i i D-19
I NEW PLANT ANALYSIS ESFCORR General Linear Modela Procedure
% pendent Variable: AVGSTESF Sum of Mean iource DF Squares. Square F Value Pr > F Model 4 0.02733245 0.00683311 0.43 0.7829 Error 11 0.17402994 0.01582090 .Correcte>d Total 15 0.20136239 !
R-Square C.V. Root MSE AVGSTESF Mean 0.135738 85.841725 0.12578117 0.14852685 l 3ource DF Type I SS Mean Square F Value Pr > F 1 WE 4 0.02733245 0.00683311 0.43 0.7829 Source D/ Type III SS Mean Square F Value Pr > F PTYPE 4 0.02733245 0.00683311 0.43- 0.7829 . T for B0: Pr > !T! Std Error of Parameter Estimate Parameter =0 Estimate TNTERCEPT 0.1522700881 B 3.42 0.0057 0.0444703599 PTYPE b4 .0089397141 B -0.09 0.9300 0.0994387476 b5 0.0410202566 B 0.41 0.6879 0.0994387476 b6 0.0779704617 B 0.58 0.5707 0.1334110796 cu .0780078097 B -0.92 0.3793 0.0851542560 w4 0.0000000000 B . . . NOTE: The X'X matrix has been found to be singular and a generalized inverse was used to solve the normal equations. Estimates followed by the l letter 'B' are biased, and are not unique estimators of the parameters. . I l
)
I 0-20 l
NEW PLANT ANALYSIS ESFCORR General Linear Models Procedure ependent Variable: AVGCMESF Sum of Mean ource DF Squares Square F Value Pr > F lodel 3 0.01310700 0.00436900 1.34 0.3373
.rror 7 0.02288066 0.00326867
- orrected Total 10 0.03598765 R-Square C.V. Root MSE AVGCMESF Mean 0.364208 60.535317 0.05717224 0.09444444 DF Type I.SS Mean Square F Value Pr > F
.ource TYPE 3 0.01310700 0.00436900 1.34 0.3373
.ource DF Type III SS Mean Square F Value Pr > F -
TYPE 3 0.01310700 0.00436900 1.34 0.3373 T for HO: Pr > lT! Std Error of arameter Estimate Parameter =0 Estimate NTERCEPT 0.0953703704 B 4.09 0.0047 0.0233404709 TYPE b4 .0120370370 B -0.19 0.8510 0.0617530815 b5 0.0574074074 B 1.23 0.2585 0.0466809418 b6 0.0000000000 B . . . ce .0564814815 B -1.21 0.2656 0.0466809418 w4 0.0000000000 B . . . OTE: The X'X matrix has been found to be singular and a generalized inverse was used to solve the normal equations. Estimates followed by the letter 'B' are biased, and are not unique estimators of the parameters. l D-21 l l --
NEW PLANT ANAL,YSIS ESFCORR NPAR1WAY PROCEDURE Wilcoxon Scores (Rank Sums) for Variable AVGSTESF Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under H0 Under 80 Score w4 8 67.0 68.0000000 9.52190457 8.3750000 b4 2 18.0 17.0000000 6.29814788 9.0000000 b5 2 23.0 17.0000000 6.29814788 11.5000000 ce 3 17.0 25.5000000 7.43303437 5.6666667 b6 1 11.0 8.5000000 4.60977223 11.0000000 Kruskal-Wallis Test (Chi-Square Approximation) CHISQ: 2.1599 DF: 4 Prob > CBISQ: 0.7064 D-22 l
NEW PLANT ANALYSIS ESFCORR NPAR1WAY PROCEDURE Median Scores (Number of Points above Median) for Variable AVGSTESF Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under B0 Under H0 Score w4 8 4.0 4.00000000 1.03279556 0.50000000 b4 2 1.0 1.00000000 0.68313005 0.50000000 b5 2 20 1.00000000 0.68313005 1.00000000 ce 3 0.0 1.50000000 0.80622577 0.00000000 b6 1 1.0 0.50000000 0.50000000 1.00000000 Median 3-Way Analysis (Chi-Square Approximation) CHISQ= 5.6250 DF= 4 Prob > CRISQ= 0.2290 ) l i 1 I 1 1 l 1 l I D-23 ; i
NEW PLANT ANALYSIS ESFCORR NPAR1WAY PROCEDURE Van der Waerden Scores (Normal) for Varieble AVGSTESF Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under H0 Under H0 Score w4 8 -0.04566873 0.0 1,74830927 .005708591 b4 2 0.20737721 0.0 1.15639788 0.103688604 b5 2 0.94453011 0.0 1.15639788 0.472265057 ce 3 -1.48363054 0.0 1.36477348 .494543513 b6 1 0.37739194 0.0 0.84639658 0.377391944 0 Van der Waerden 1-Way (Chi-Square Approximation) CHISQ 1.7588 DF= 4 Prob > CHISQ: 0.7800 1 J
)
l l 1 l i . I D-24
NEW PLANT ANALYSIS ESFCORR NPAR1WAY PROCEDURE Savage Scores (Exponential) for Variable AVGSTESF Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under H0 Under H0 Score w4 8 0.92243312 0.0 1.83443150 0.115304140 b4 2 0.15824731 0.0 1.21336239 0.079123654 b5 2 0.47812465 0.0 1.21336239 0.239062327 ca 3 -1.65620074 0.0 1.43200263 .552066915 b6 1 0.09739566 0.0 0.88809033 0.097395660 Savage 1-Way (Chi-Square Approximation) CHISQ: 1.3753 DF: 4 Prob > CHISQ= 0.8485 l D-25
NEW PLANT ANALYSIS. ESFCORR NPAR1WAY PROCEDURE Wilcoxon Scores (Rank Sums) for Variable AVGCMESF Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under B0 Under B0 Score. w4 6 38.0 36.0 5.46476316 6.33333333 b4 1 5.0 6.0 3.15508248 5.00000000 b5 2 18.0 12.0 4.23298734 9.00000000 ce 2 5.0 12.0 4.23298734 2.50000000
, Average Scores were used for Ties Kruskal-Wallis Test (L'hi-Square Approximation)
CHISQ: 4.0335 DF= 3 Prob > CRISQ: 0.2579 + l l D-26
NEW PLANT ANALYSIS ESFCORR NPAR1WAY PROCEDURE Hedian Scores (Number of Points above Median) for Variable AVGCKESF Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under 80 Under 80 Score w4 6 3.0 2.72727273 0.862439362 0.50000000 b4 1 0.0 0.45454545 0.497929598 0.00000000 ; b5 2 2.0 0.90909091 0.668042657 1.00000000 ce 2 0.0 0.90909091 0.668042657 0.00000000 Average Scores were used for Ties Median 1-Way Analysis (Chi-Square Approximation) CHISQ: 4.5000 DF: 3 Prob > CHISQ= 0.2123 1 i i l 4 1 i 1 1 D-27
NEW PLANT ANALYSIS ESFCORR NPAR1WAY PROCEDURE Van der Waerden Scores (Normal) for Variable AVGCMESF Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under H0 Under Bo Score we 6 0.43072730 0.0 1.38533844 0.071787883 b4 1 -0.21042839 0.0 0.79982552 .210428394 b5 2 1.59342252 0.0 1.07307854 0.796711261 ce 2 -1.81372143 0.0 1.07307854 .906860713 Average Scores were used for Ties a Van der Waerden 1-Way (Chi-Square Approxianation) CRISQ: 4.2483 DF= 3 Prob > CRISQ= 0.2359 l 1 I D-28
NEW PLANT ANALYSIS ESFCORR NPAR1WAY PROCEDURE Savage Scores (Exponential) for Variable AVGCMESF Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under HD Under H0 Score we 6 -0.04422799 0.0 1.47011754 .007371332 b4 1 -0.43012266 0.0 0.84877276 .430122655 b5 2 1.95642136 0.0 1.13874815 0.978210678 ce 2 -1.48207071 0.0 1.13874815 .741035354 Average Scores were used for Ties Savage 1-Way (Chi-Square Approximation) CHISQ: 4.0348 DF= 3 Prob > CHISQ: 0.2577 i l l D.29 i
1 l
)
Table D-4 Average Technical Specification Violation Rates l i
>1 Pre-Commercial Post-Commercial '
i Plant Events / Month Events / Month McGuire 2 2.55 0.00 Diablo Canyon 1 1.20 2.16 { Callaway 1.57 1.83 . Catawba 1 2.51 2.50 ; Byron 1 4.22 1.17 Wolf Creek 1.88 2.17 8 Diablo Canyon 2 1.31 N/A, Millstone 3 1.63 N/A St. Lucie 2 1.45 1.00 Waterford 1.50 1.8) Palo Verde 1 4.47 N/A LaSalle 2 1.17 1.67 WNP-2 3.26 2.17 Susquehanna 2 1.10 0.5Q Limerick 1 2.85 N/A, River Bend 2.37 N/A Average 2.19 1.55 j Note: These plants had not accumulated 180 days of commercial operation l before June 30, 1986. Therefore, their post-commercial experience was I not considered in the statistical analysis. I 0-30
NEW PLANT ANALYSIS TSCORR Variable N Mean Std Dev Sum Minimum Maximum LPDATE 16 8974 289.76536 143579 8462 9460 STRTPD 16 310.81250 112.43559 4973 124.00000 .546.00000 AVGSTTS 16 0.07426 0.03661 1.18813 0.03681 -0.15892 AVGCMTS 11 0.05202 0.02585 0.57222 0 0.08333 NAETYPE 16 0.25000 0.44721 4.00000 0 1.00000 NFPLTYPE 16 0.37500 0.50000 6.00000 0 1.00000 NFNUNIT 16 0.62500 0.50000 10.00000 0 1.00000
' Correlation Coefficients / Prob > lRl under Ho: Rho:0
/ Number of Observations LPDATE STRTPD AVGSTTS AVGCMTS NAETYPE NFPLTYPE NFNUNIT PDATE 1.00000 -0.19481 0.12673 0.49565 -0.26173 0.30456 0.26648 0.0000 0.4697 0.6400 0.1210 0.3275 0.2514 0.3184 16 16 16 11 16 16 16 .
4
.TRTPD -0.19481 1.00000 0.28688 0.05985 0.44117 -0.18722 0.32478 0.4697 0.0000 0.2813 0.8612 0.0872 0.4875 0.2197 16 16 16 11 16 16 16 VGSTTS 0.12673 0.28688 1.00000 -0.01851 -0.17918 0.27694 0.50539 0.6400 0.2813 0.0000 0.9569 0.5067 0.2991 0.0458 16 16 16 11 16 16 16
.liGCMTS 0.49565 0.05985 -0.01851 1.00000 -0.00418 0.44917 0.74345 0.1210 0.8612 0.9569 0.0000 0.9903 0.1658 0.0087 11 11 11 11 11 11 11 AETYPE -0.26173 0.44117 -0.17918 -0.00418 1.00000 -0.44721 -0.14907 0.3275 0.0872 0.5067 0.9903 0.0000 0.0824 0.5816 16 16 16 11 16 16 16 I
IFPLTYPE 0.30456 -0.18722 0.27694 0.44917 -0.44721 1.00000 0.60000 0.2514 0.4875 0.2991 0.1658 0.0824 0.0000 0.0140 16 16 16 11 16 16 16 l IFNUNIT 0.26648 0.32478 0.50539 0.74345 -0.14907 0.60000 1.00000 l 0.3184 0.2197 0.0458 0.0087 0.5816 0.0140 0.0000 16 16 16 11 16 16 16 l l I ' i l . 1 I D-31
. _ _ - - _ _ - _ _ - _ _ _ _ _ - _ _ _ . -__A
l NEW PLANT ANALYSIS-TSCORR General Linear Models Procedure lependent Variable: AVGSTTS Sum of Mean iource DF Squares Square F Value Pr > F fodel 4 0.00061888 0.00015472 0.09 0.9846 lrror 11 0.01949041 0.00177186
'orrected Total 15 0.02010929 R-Square C.V. Root MSE AVGSTTS Mean 0.030776 56.685424 0.04209341 0.07425791
;ource DF Type I SS Mean Square F Value Pr > F
' TYPE 4 0.00061888 0.00015472 0.09 0.9846
.ource DF Type III SS Mean Square F Value Pr>F 11TE. 4 0.00061888 0.00015472 0.09 0.984G T for HO: Pr > lTl Std Error of
'arameter Estimate Parameter =0 Estimate NTERCEPT 0.0715423974.B 4.81 0.0005 0.0148822691
' TYPE b4 .0056212908 B -0.17 0.8689 0.0332777654 b5 0.0022556708 B 0.07 0.9472 0.0332777654 b6 0.0074954032 B 0.17 0.8697 0.0446468074 ce 0.0142280366 B 0.50 0.6274 0.0284973758 w4 0.0000000000 B . . .
OTE: The X'X matrix has been found to be singular and a generalir.ed inverse was used to solve the normal equations. Estimates followed by the letter 'B' are biased, and are not unique estimators.of the parameters. D-32
l l NEW PLANT ANALYSIS TSCORR 1 General Linear Models Procedure )ependent Variable: AVGCMTS l Sum of Mean inurce DF Squaren Square F Value Pr > F iodal 3 0.00158062- 0.00052687 0.72 -0.5697 l Error 7 0.00510288 0.00072898 ! j 3crrected Total 10 0.00668350 ] R-Square C.V. Root MSE AVGCMTS Mean 0.236496 51.902306 0.02699968 0.05202020 inurce DF Type I SS Mean Square F Value Pr > F l 9ndPE 3 0.00158062 0.00052687 0.72 0.5697 I isurce DF Type III SS Mean Square F Value Pr > F j i " TYPE 3 0.00158062 0.00052687 0.72 0.5697 1 T for HO: Pr > lT! Std Error of Scrameter Estimate Parameter =0 Estimate INTERCEPT 0.0546296296 B 4.96 0.0016 0.0310225750
> TYPE b4 .0379629630 B -1.30 0.2342 0.0291629921 b5 0.0092592593 B 0.42 0.6871 0.0220451499 b6 0.0000000000 B . . .
ce .0046296296 B -0.21 0.8396 0.0220451499 w4 0.0000000000 B . . . 40TE: The X'I matrix has been found to be singular and a generalized . inverse was used to solve the normal equations. Estimates followed by the letter 'E' are biased, and are not unique estimators of the parameters. 0-33
NEW PLANT ANALYSIS TSCORR NPAR1WAY PROCEDURE Wilcoxon Scores (Rank Sums) for Variable AVGSTTS Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scoren Under H0 Under H0 Score w4 8 69.0 68.0000000 9.51490060 8.6250000 b4 2 14.0 17.0000000 6.29351518 7.0000000 b5 2 16.0 17.0000000 6.29351518 8.0000000 ce 3 27.0 25.5000000 7.42756690 9.0000000 b6 1 10.0 8.5000000 4.60638144 10.0000000 Average Scores were used for Ties Kruskal-Wallis Test (Chi-Square Approximation) CHISQ: 0.35898 DF= 4 Prob > CBISQ= 0.9857 ' D-34
NEW PLANT ANALYSIS ' TSCORR NPAR1WAY PROCEDURE Median Scores (Number of Points above Median) for Variable AVGSTTS Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under H0 Under 80 Score w4 8 3.0 3.50000000 1.02469508 0.37500000 b4 2 1.0 0.87500000 0.67777209 0.50000000 b5 2 1.0 0.87500000 0.67777209 0.50000000 ca 3 1.0 1.31250000 0.79990234 0.33333333 b6 1 1.0 0.43750000 0.49607837 1.00000000 Average Scores were used for Ties Median 3-Way Analysis (Chi-Square Approximation) CHISQ: 1.5079 DF= 4 Prob > CBISQ= 0.8252 l l D-35
NEW PLANT ANALYSIS TSCORR NPAR1WAY PROCEDURE Van der Waerden Scores (Normal) for Variable AVGSTTS Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under H0 Under H0 Score w4 8 0.232188855 0.0 1.74747853 0.029023607 b4 2 .843204187 0.0 1.15584840 .421602094 b5 2 .257931941 0.0 1.15584840 .128965971 ce 3 0.645939442 0.0 1.36412499 0.215313147 b6 1 0.223007831 0.0 0.84599440 0.223007831 Average Scores were used for Ties Van der Waerden 1-Way (Chi-Square Approximation) CHISQ: 0.76539 DF= 4 Prob > CHISQ= 0.9430 i I D-36
i NEW PLANT ANALYSIS TSCORR NPAR1WAY PROCEDURE l Savage Scores (Exponential) for Variable AVGSTTS Cla.ssified by Variable PTYPE i Sum of Expected Std Dev Hean I PTYPE N Scores Under B0 Under B0 Score w4 8 -0.74386169 0.0 1.83386357 .092982712-b4' 2 -0.39010434 0.0 1.21298674 .195052170 b5 2 0.00989566 0.0 1.21298674. 0.004947830 ce 3 1.19334138 0.0 1.43155929 0.397780460 , b6 1 -0.06927101 0.0 0.88781538 .069271007 Average Scores were used for Ties Savage 1-Way (Chi-Square Approximation) CHISQ: 0.74312 DF: 4 Prob > CRISQ= 0.9459
)
i D-37
NEW PLANT ANALYSIS TSCORR hPAR1WAY PROCEDURE Wilcoxon Scores (Rank Sums) for Variable AVGCMTS
' Classified by Variable PTYPE Sum of. Expected Std Dev Mean PTn'E N Scores Under H0 Under H0 Score w4 6 40.0 36.0' 5.40201982 6.66666667 b4 1 2.0 6.0 3.11885760 2.00000000 b5 2 14.0 12.0 4.18438656 7.00000000 l ce 2 10.0 12.0 4.18438656 5.00000000 Average Scores were used for Ties Kruskal-Wallis Test (Chi-Square Approximation)
CHISQ: 2.1184 DF= 3 Prob > CHISQ= 0.5482 l f
]
l D-38
NEW PLANT ANALYSIS TSCORR NPAR1WAY PROCEDURE Median Scores (Namber of Points above Med$an) for Variable AVGCMTS Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE- N Scores Under H0 Under H0 Score w4 6 3.0 2.18181818 0.833195581 0.500000000 b4 1 0.0 0.36363636 0.481045693 0.000000000 b5 2 1.0 0.72727273 0.645390522 0.500000000 e ce 2 0.0 0.72727273 0.645390522 0.000000000 Average Scores were used for Ties Median 1-Way Analysis (Chi-Square Approximation) CHISQ: 2.1429 DF= 3 Prob > CHISQ= 0.5433 1 i D-39
NEW PLANT ANALYSIS TSCORR NPAR1WAY' PROCEDURE Van der Waerden Scores (Normal) for Variable AVGCHTS Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under H0 Under H0 Score w4 6 0.934364749 0.0 1.36884485 0.155727458 b4 1 .967421566 0.0 0.70030294 .967421566 b5 2 0.480451144 0.0 1.06030266 0.240225572 ce 2 .447394328 0.0 1.06030266 .223697164 Average Scores were used for Ties ~ Van der Waerden 1-Way (Chi-Square Approximation) CHISQ: 1.8877 DF: 3 Prob > CHISQ= 0.5960 l D-4 0
NEW PLANT ANALYSIS TSCORR NPAR1WAY PROCEDURE Savage Scores (Exponential) for Varlable AVGCHTS Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under H0 Under HO Score w4 6 1.46271645 0.0 1.43974035 0.243786075 b4 1 -0.80909091 0.0 0.83123448 .809090909 b5 2 0.14531025 0.0 1.11521808 0.072655123 ce 2 -0.79893579 0.0 1.11521808 .399467893 e Average Scores were used for Ties Savage 1-Way (Chi-Square Approximation) CHISQ: 1.7643 DF= 3 Prob > CHISQ= 0.6227 i r D-41 l \
Table D-5 Average Rates for Loss of System Safety Function Pre-Commercial Post-Commercial Plant Events / Month Events / Month McGuire 2 0.33 0.00 Diablo Canyon 1 0.28 0.00 Callaway 0.93 0.33 Catawba 1 0.69 0.17 Byron 1 0.27 0.16 Wolf Creek 0.33 0.10 Diablo Canyon 2 0.18 N/A, Millstone 3 0.82 N/A St. Lucie 2 0.96 0.50 Waterford 0.21 0.0Q Palo Verde 1 0.44 N/A LaSalle 2 0.58 0.34 WNP-2 1.25 0.33 Susquehanna 2 0.45 0.15 Limerick 1 0.45 N/A, River Bend 0.7? N/A Average 0.57 0,20 Note: These plants had not accumulated 180 days of commercial operation before June 30, 1986. Therefore, their post-commercial experience was not considered in the statistical analysis. D-42
NEW PLANT ANALYSIS 43 SUCORR 10:37 Saturday, March 14, 1987 Variable N Mean Std Dev Sum Minimum Maximum LPDATE 16 8974 289.76536 142h'9 8462 9460 STRTPD 16 310.81250 112.43559 /~ 3 124.00000 546.00000 AVGSTSU 16 0.01864 0.01036 0. 2SJ 20 0.00625 0.04178 AVGCMSU 11 0.00657 0.00545 0.07222 0 0.01667 NAETYPE 16 0.25000 0.44721 4.00000 0 1.00000 NFPLTYPE 16 0.37500 0.50000 6.00000 0 1.00000 NFNUNIT 16 0.62500 0.50000 10.00000 0 1.00000 Correlation Coefficients / Prob > lRl under Ho: Rho =0
/ Number of Observations LPDATE STRTFD AVGSTSU AVGCMSU NAETYPE NFPLTYPE NFNUNIT LPDATE 1.00000 -0.19481 -0.17705 -0.23740 -0.26173 0.30456 0.26648 0.0000 0.4697 0.5118 0.4821 0.3275 0.2514 0.3184 16 16 16 11 16 16 16 STRTPD -0.19481 1.00000 -0.38257 -0.61781 0.44117 -0.18722 0.32478 0.4697 0.0000 0.1436 0.0428 0.0872 0.4875 0.2197 16 16 16 11 16 16 16 AVGSTSU -0.17705 -0.38257 1.00000 0.80744 -0.36023 0.23736 0.00854 0.5118 0.1436 0.0000 0.0027 0.1705 0.3761 0.9750 16 16 16 11 16 16 16 AUGCMSU -0.23740 -0.61781 0.80744 1.00000 -0.55514 0.05507 -0.25698 0.4821 0.0428 0.0027 0.0000 0.0763 0.8722 0.4456 11 11 11 11 11 11 11 NAETYPE -0.26173 0.44117 -0.36023 -0.55514 1.00000 -0.44721 -0.14907 0.3275 0.0872 0.1705 0.0763 0.0000 0.0824 0.5816 16 16 16 11 16 16 16 NFPLTYPE 0.30456 -0.18722 0.23736 0.05507 -0.44721 1.00000 0.60000 0.2514 0.4875 0.3761 0.8722 0.0824 0.0000 0.0140 16 16 16 11 16 16 16 ;
NFNUNIT 0.26648 0.32478 0.00854 --0.25698 -0.14907 0.60000 1.00000 0.3184 0.2197 0.9750 0.4456 0.5816 0.0140 0.0000 ' 16 16 16 11 16 16 16 l i D-43 I 1
NEW PLANT ANALYSIS 45 SOCORR 10:37 Saturday, March 14, 1987 General Linear Models Procedure Dependent Variable: AVGSTSU Sum of Mean Source DF Squares Square F Value Pr > F Model 4 0.00039525 0.00009881 0.90 0.4990 Error 11 0.00121424 0.00011039 Corrected Total 15 0.00160950 R-Square C.V. Root MSE AVGSTSU Mean 0.245576 56.373186 0.01050646 0.01863734 Source DF Type I SS Hean Square F Value Pr > F PTYPE 4 0.00039525 0.00009881 0.90 0.4990 rource DF Type III SS Mean Square F Value Pr > F PTYPE 4 0.00039525 0.00009881 0.90 0.4990 T for HO: Pr > lT! Std Error of Parameter Estimate Pa rameter:0 Estimate I 0.0037145958 INTERCEPT 0.0160440227 B 4.32 0.0012 PTYPE b4 .0008159158 B -0.10 0.9235 0.0083060888 b5 0.0145876020 B 1.76 0.1068 0.0083060888 b6 0.0080109601 B 0.72 0.4872 0.0111437875 l ce 0.0019795934 B 0.28 0.7859 0.0071129095 w4 0.0000000000 B . . . NOTE: The X'X matrix has been found to be singular and a generalized inverse was used to solve the normal equations. Estimates followed by the letter 'B' are biased, and are not unique estimators of the parameters. D-44
NEW PLANT ANALYSIS 47 SUCORR 10:37 Saturday, March 14, 1987
]
General I.inear Models Procedure lependent Variable: AVGCMSU Sum of Mean ioerce DF Squares Square F Value Pr > F l Model 3 0.00007108 0.00002369 0.73 0.5646 {rror 7 0.00022634 0.00003233 Corrected Total 10 0.00029742 R-Square C.V. Root MSE AVGCMSU Mean s 0.238994 86.606607 0.00568629 0.00656566 Source DF Type I SS Mean Square F Value Pr>F PTYPE 3 0.00007108 0.00002369 0.73 0.5646 Source DF Type III SS Mean Square F Value Pr > F PTYPE 3 0.00007108 0.00002369 0.73 0.5646 T.for HO: Pr > !T! Std Error of Parameter Estimate Parameter:0 Estimate TNTERCEPT 0.0046296296 B 1.99 0.0863 0.0023214192 PTYPE b4 0.0009259259 B 0.15 0.8844 0.0061418978 b5 0.0064814815 B 1.40 0.2054 0.0046428383 b6 0.0000000000 B . . . ce 0.0037037037 B 0.80 0.4512 0.0046428383 w4 0.0000000000 B . . . MOTE: The X'I matrix has been found to be singular and a generalized inverse was used to solve the normal equations. Estimates followed by the letter 'B' are biased, and are not unique estimators of the parameters. 1 I D-45
NEW PLANT ANALYSIS 49 l SUCORR 10:37 Saturday, March 14, 1987 NPAR1WAY PROCEDURE Wilcoxon Scores (Rank Suns) for Variable AVGSTSU ! Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPF. N Scores Under 80 Under H0 Score w4 8 57.0 68.0000000- 9.52190457 7.1250000 b4 2 17.0 17.0000000 6.29814788 8.5000000 ! b5 2 26.0 17.0000000 6.29814788 13.0000000 ce 3 24.0 25.5000000 7.43303437 8.0000000 b6 1 12.0 8.5000000 4.60977223 12.0000000 Kruskal-Wallis Test (Chi-Square Approximation) CHISQ: 3.0276 DF= 4 Peob > CHISQ: 0.5532 4 i 0-46
l NEW PLANT ANALYSIS 50 SUCORR 10:37 Saturday, March 14, 1987 NPAR1WAY PROCEDURE Hedian Scores (Number of Points above Median). for Variable AVGSTSU Classified by Variable PTYPE Sum of Expected Std Dev. Mean PTYPE N Scores Under B0 . Under H0 Score w4 8 3.0 4.00000000 1.03279556 'O.37500000 b4 2 1.0 1.00000000 0.68313005 0.50000000 b5 2 2.0 1.00000000 0.68313005 1.00000000 ce 3 1.0 1.50000000 0.80622577 0.33333333 b6 1 1.0 0.50000000 0.50000000 1.00000000 Median 1-War Analysis (Chi-Square Approximation) CHISQ: 3.5938 DF= 4 Prob > CHISQ:: 0.4638 j D-47
HEW PLANT ANALYSIS 51 SUCORR 10:37 Saturday, March 14, 1987 NPAR1WAY PROCEDURE Van der Waerden Scores (Normal) for Variable AVGSTSU Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under 80 Under 80 Score w4 8 -2,10612156 0.0 1.74830927 .263265195 b4 2 0.00000000 0.0 1.15639788 0.000000000 b5 2 1.78773430 0.0 1.15639788 0.893867151 ce 3 -0.22300783 0.0 1.36477348 .074335944
, b6 1 0.54139509 0.0 0.84639658 0.541395085 Van der Waerden 1-Way (Chi-Square Approximation)
CRISQ: 3.2221 DF: 4 Prob > CHISQ: 0.5214
/
/
4
.- 1 w
i l l l l J l
}.
'5 I D-48
NFAl PLANT ANALYSIS 52 SUCORR 10:37 Saturday, March 14, 1987 NPAR1WAY PROCEDURE Savage Scores (Exponential) for Variable AVGSTSU Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under HO Under HD Score w4 8 -2.12125375 0.0 1.83443150 -0.26515672 b4 2 -0.54925630 0.0 1.21336239 -0.27462815 b5 2 2.31145799 0.0 1.21336239 1.15572899 cc 3 0.06165640 0.0 1.43200263 0.02055213 b6 1 0.29739566 0.0 0.88809033 0.29739566 Savage 1-Way (Chi-Square Approximation) CHISQ= 4.1299 DF= 4 Prob > CHISQ: 0.3887 l I i l i D-49
NEW PLANT ANALYSIS 54 SDCORR 10:37 Saturday, March 14, 1987 NPAR1WAY PROCEDURE Wilcoxon Scores (Rank Sums) for Variable AVGCMSU Classified by Variable PTYPE Sum of Expected Std Dev Mean PTYPE N Scores Under 80 Under RO Score w4 6 29.5000000 36.0 5.24837637 4.91666667 b4 1 5.5000000 6.0 3.03015151 5.50000000 b5 2 18.0000000 12.0 4.06537486 9.00000000 ce 2 13.0000000 12.0 4.06537486 6.50000000 o Average Scores were used for Ties Kruskal-Wallis Test (Chi-Square Approximation) CHISQ= 2.5536 DF= 3 Prob > CHISQ= 0.4657 l i i I 0-50 l
NEW PLANT ANALYSIS 55 SUCORR 10:37 Saturday, March 14, 1987 NPAR1WAY PROCEDURE Median Scores (Number of Points above Median) for Variable AVGCMSU Classified by Variable PTYPE Sum of Expected Std Dev Mean l PTYPE N Scores Under H0 Under DO Score w4 6 1.0 2.18181818 0.833195581 0.16666667 b4 1 0.0 0.36363636 0.481045693 0.00000000 , b5 2 2.0 0.72727273 'O.645390522 1.00000000 l ce 2 1.0 0.72727273 0.645390522 0.50000000 Average Scores were used for Ties Median 1-Way Analysis (Chi-Square Approximation) . { CHISQ= 4.7619 DF= 3 Prob'> CHISQ: 0.1901 l l 1 . l l l l a 0-51
l 1 ! NEW Pl. ANT ANALYSIS 56 SUCORR 10:37 Saturday, March 14, 1987 ( NPAR1WAY PROCEDURE Van der Waerden Scores (Normal) for Variable AVGCMSU Clansified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under H0 Under 80 Score w4 6 -1.64876956 0.0 1.32538319 .274794927 b4 1 -0.10768182 0.0 0.76521034 .107681825 b5 2 1.38175908 0.0 1.02663740 0.690879539 ce 2 0 37469231 0.0 1.02663740 0.187346156 o Average Scores were used for Ties Van der Waerden 1-Way (Chi-Square Approximation) CHISQ: 2.3126 DF: 3 Prob > CHISQ: 0.5101 D-52
NEW PLANT ANALYS]S 57 SDCORR 10:37 Saturday, March 14, 1987 NPAR1WAY PROCEDURE Savage Scores (Exponential) for Variable AVGCMSU Classified by Variable PTYPE Sum of Expected Std Dev Hean PTYPE N Scores Under HO Under 80 Score w4 6 -2.03285233 0.0 1.42651391 .338808722 b4 1 -0.33250361 0.0 0.82359819 .332503608~ b5 2 1.15086580 0.0 1.10497293 0.575432900 ce 2 1.21449014 0.0 1.10497293 0.607245070 Average Scores were used for Ties Savage 1-Way (Chi-Square Approximation) CHISQ= 2.9472 DF= 3 Prob > CHISQ: 0.3998 l D-53 E_
APPENDIX E ESF ACTUATION EXPERIENCE 1 l E-1 ,
M i ESF ACTUATION DATA l l E-2
EDoA 0 L 0 0 0 0 0 0 0 0 0 0 0 1 8 6 4 2 O r ' o r 4 te r s
? 2 nl e ee r n ?
uw mnn do '
' 2 2
n gipo r en ' s surhoeck i eqetrn 0 DEPOPU U 2 i A3521 ' v25ZL x 22 r
' 8 1 e o c S n N 6 a -
N 1 u O2 I 4i s s _ TE 1 AR UI O L O C1 2e TU c CG C n AC d 0 1 i s - F M ' s h S ' 8 t E n o 1 6 M t ' e r u c 4 9 i r P F
% U /2 C I
6 L 9 OO 0 0 0 0 0 4 3 2 1 m '" DaId % 0 b o. 6 a Z T" l
r ' o r r o ' t e s nl e ee rn n uw n m n do ' g po r en i s Q surhoeck i eqet rn DEPOPU 2352DO 22 2'E5ZC U ( O C 1 - MN OO I Y MNA # MCO P F ELB F A BID - o 2 E C I e r u g i F e M L
$OO 0 0 0 0 0 4 3 2 1 WC9.%]8 %0 upgaZ I
l
r o r r e n N n w T n n r o f g n i s e k s rh eht D O 2 i2 2 # Z ( 5Z S N O f TK I AW UA TL CL AAC g - F S - E I 3 E e r o u g i - - F - 0 0 - 0 4 2 1
- s7
r ' o r r ' t e s nl ee re n 0 mnn do uw \ n g po r en i s surhoeck i e qet r n DEPOPU U I 235lG 2 ZREGCr D- - ' S h M N o O1 I - C TA -' !/ /// AB # U T W A Y O C T C AAC f l F o S o E o *PF o /6 4 C I e e
-s ,
i F ? e 9L
$OO 0 0 0 0 0 4 3 2 1 . wC3.%2O b# aZ min l' '
r o r r t e s
# enl e n e r mn n uw n do g po ren i
i s eck surho e qet rn DEPOPU # 2i52D1 # [ Z2 2 E52CF
/
S o x N O I 1 T AN b UO l TR CY /{ AB _ F 4 - S E 5 E e r u g I i I F J
/
4
/y$
/ '
, gp 0 0 0 4 3 1 oU o b#baZ T
r o r r e e n n w
.n n o o n s
_.g i s e e k D U H(i S N - O I E K o TE o AR k UC f { o T C F AL O FW S E 6 y E e r u g i - F f f < [ 0 0 0 0 4 3 2 1 O5EO yO be6 min
ED ga $ 0 0 0 0 0 0 0 0 0 0 0 1 8 6 4 2 0 r , o r , 4 r 2 te s _ nl e n , ee r 2 n1 4 uw , 2 n g rn po do r en , surhoeck s i i eqetrn , 0 _ DEPOPU 2 A35I 22DMI v25I Lr 8 t 1 e
, c -
2 n S - NN OO 6 1 a u s - 4i s T Y I 3 1 - A UA N m 2 L O 1 e T C O C c n O EL g% m 0 1 i s s F B m h A EID 3 8 t n o_ 6 M 7 5 E P - F e r u T2 /4 C - g I i F
.- 2 L
. OO 0 0 0 0 0 4 3 2 1 mCo$D b#EDZ ma
r o r r te s nl e n e e ru w, n n m n do g po r en i su i s eck eqetrho rn DEPOPU 3352D1 22 - Z25ECF S N O 3 I f E KN U O T T CL S AL I FM S E 8 e c e [ig r u s i r j 4
/
0 0 0 0 4 3 2 1 o tCo 3 %O b#EaZ Eo
8 r o r r t e s nl e n ee r n m n do uw n g po ren i s surhoi e qet reck n DEPOPU 335I ZE5l22GDO iC S N O2 I TA AB UW T CA T AA FC S E 9 E e r u g i I F _f 2 0 0 0 0 4 3 2 1 n fco5g8 %O mpEaZ r~
r o r r te s nl e ee r n n uw m n do n g po ren i i s eck surho eqet rn DEPOPU / 2i523I 5-Z Z2 I E5Z[r S N O2 I TE AI UC T C L U AT F S S E . 0 1 E o e r o u / g i / F
? !i 0 0 0 0 4 3 2 1 oCo5DY t
5 o % O u uc o, g c Z c m'."
r , o r , r t e s nl e , ee r n n uw , n m n do g po r en , surhoeck i i s eqetrn , DEPOPU 235I [ 22GDO 2 E5liC o r 3 S N3 O I D go , TR AO UF 9 o 3 TR 7 1 CE r - AT 7 F A SW 3 E o 1 1 e e b i w kP r - 1 1 rO - 0 0 0 0 0 - 4 3 2 1 nbs]8 %O $#baZ ( - i e0
r o r r te s nl e ee rn n uw n m n do g i po ren i surho s eck eqetrn DEPOPU 21 5 il D 2 22G [ E5l iCO ? S _ N1 - O E I o\E TD o. 6 AR o UE o TV C AO L FA o k SP E - 2 - 1 E e r
;o/
u g i F T5 0 0 0 0 4 3 2 1 go$30 y eaZ mi
L' Do4 $c LU - - 0 0 0 0 0 0 0 0 0 0 0 1 8 6 4 2 O r ' o r 4 r ' 2 te s nl e n ' ee r n uw 2 n m n do ' 2 g po r en i s ' surhoeck i e qet r n 0 DEPOPU 2 I Aa521 2 vto5ZLN r
' 8 1 e
' c S n -
N2 ' 6 1 a u - OE I s s TD ' 4i 1 AR - L O UE 2 TV 1 e c C ' n AO L ' 0 1 i s s FA SP h t 8 E n o - M _ - 1 3
# 6 E 0 e U P r F u /4 i
g C I F 9
- 2 L _
OO 0 - 0 0 0 0 - 4 3 2 1 - W C o 5 3 " 8 % 0 a c) Q 6 s Z TM L
0]e go $ 0 0 0 0 0 0 0 0 0 0 0 1 8 6 4 2 g r , o - r , 4 - r 2 te s nl e _ ee r n n uw 2 m 2 g i pon r en n do s surhoeck i eqetrn , 0 DEPOPU 2 i 1 A32252 x VE5Zt r _ 8 1 e e c S n 6 a N 1 u O2 I 4i s s TE o 1 AL L O UL 2 1 e TA CS - c n AA O 0 i m C1 s p s F L h S 8 t E n o 4 1 6 M E e r i u g 9 P 4 F F / C I 2 3 L 3OO 0 0 0 0 g 4 3 2 1 n C ]8 yO b#EDZ rg
1 8 6 4 2 O r ' o r 4 r 2 t e s nl e ee rn n uw m n do
' 2 2
n g po ren i i s eck E surho eqet rn 0 DEPOPU M 2 23523 ' E252EO 22 8 r 1 e 2 c S n 6 a NR 1 u OA I s s T E o 4i 1 AL UC L O O 2e T U C1 c CN n A H . 0 1 i s s FS T h SA t [f 8 n EW M o s a 6 i i t ' e
'FP r
i
- y 4 r
i' 2 C I
' L tOO 0 0 0 0 0 4 3 2 1
{o 3. O g r6
r o r r t e s nl e ee r n n uw m n do n g po r en i i surho s eck eqet rn DEPOPU - 3 3i5IZ2GiGn 2E5l S2 N A O I N TN AA - U T H C E AU Q FS SU ES 6 1 E e r u g i F 0 0 0 0 2 1 4 3 nC9.%Dta ( b o. 2 D Z y
6 4 r o r r ' 4 e 2 e n n 2 o n n 2 g o n i s s k ' e e n 0 D U 2 2 5 2 E 5 O ' 8 1 e M c S n N O C 16 a O " u s T K I 4i s C 1 AI L UR 2 O TE 1 e c CM AI 7 P 0 i n L F 1 s F s S h t E 8 n o 1 7 6 M E e , r u g i 4 F r C 2 I L eOO 0 0 0 0 0 4 3 2 1 gb 1 j$sZ eG
r o r r t e s nl e ee r n n uw n m n do g po r en i s i surhoeck eqet rn DEPOPU 2552Dn 252E52C ( S N OM I TA AH UE T R CO AH FS 7 S E o 8 1 E e r u g i F - 0 0 0 0 2 1 4 3 D 0coND#S yO aCDEDZ 9; Eo
1 8 6 4 2 O r ' o r ' 4 r 2 t e s nl e n ' ee r n uw ' 2 2 n mn g po r en do i i s eck surho eqet rn ' 0 DEPOPU 2 21Z252D-1 2 E52CF (
' 8 1
e
' c S n 6 a N
1 u O I
@ s s
T2 ? 4i 1 L AI O UM 2 1 e TR o c n CE AF i 0 9 1 s s F 9 h S 8 t E n F o o 6 M 9 1 > h h E e P r u F 4 C g i I r 2 L OO 0 0 0 0 0 4 3 2 1 {o5D E b#E3Z
,g
r , o r r , t e s nl e ee r n n m n do uw , ng po r en i s eck , i su eqetrhorn DEPOPU I 1 21Z GD-- , ER2SCF , S N , ODN I T E , AB U TR CE , AV I 2C FR S H E @ 0 2 E e - r u g i F ' i _ 6-0 0 0 0 0 3 2 4 1 nCo$D om f b E oZ. rN
0 0 0 0 0 1 8 6 4 2 O r ' o r 4 r ' 2 t e s nl e ' ee rn n uw m n do
' 2 2
n g po r en i s eck i surho eqet rn ' 0 DEPOPU 2 2352D Z'E5252Cn f
' 8 1 e
' c S n 6 a N
1 us O I s 4i T ' 1 AY ' L O UR 2 T C E R ' 1 e c n A P ' 0 i s 1 s F ' h S ' 8 t E n o 1 ' 6 M 2 e ' e r e 4 9 4 e
- h. I a'C
$O0 L
2 0 0 0 0 0 4 3 2 1 W C o 5 D " [d % 0 m a) O. 6 3 Z 73 l'
1 8 6 4 2 r _ o r r te s nl e ee rn n uw n m n do g po r en i i surhoks ec eqet rn DEPOPU 3I 1 21 2 [ Z2G E5EEF 3-S N OKE I TE AR U T C CE APO F H S E 2 2 E e r u g i F e' // y 0 0 0 0 4 3 2 1 n fCoND"8 yo aopg0Z r2 f
4 i DETAILED EXPERIENCE WITH RADIATION , f MONITORING SYSTEMS AT SELECTED PLANTS l l l l E-25
DETAILED EXPERIENCE WITH RADIATION MONITORING SYSTEMS AT SELECTED PLANTS
- 1. BYRON 1 At Byron I, the numerous realignments of HVAC equipment from the normal operating mode to the emergency (ESF) mode were due almost entirely to the activation of radiation monitors. Several different but interrelated reasons I have combined to yield an extremely sensitive radiation monitoring system that frequently causes unwanted and unnecessary HVAC ESF actuations. These reasons are identified below:
(1) New technology - The radiation monitoring system at Byron 1 repre-sents the state-of-the-art in radiation detection equipment. The use of beta scintillator to detect abnormal levels of iodine, noble gases, and particulate is a new application of an extremely sensitive instru-ment in an operating nuclear plant. (2) Inexperienced specification writers - Original system specifications were written by people who, though technically competent, lacked practical radiation monitoring experience in a nuclear plant. As a result, some of the equipment specifications were inappropriate for this particular application. (3) Conservative design requirements - The licensee's specifications for the radiation monitoring system included two levels of radiation warning: an alert level, and an alarm level. However, the licensee has set the ESF actuation setpoint of this system to be the lower, radiation alert level, which was specified in the original specifications to be less than a decade above background. As a result, the system reacts to small changes in radiation level by generating an ESF actuation signal, instead of merely providing a warning, as had originally been intended. (4) Acceptance testing - The radiation monitoring syste. was installed just before OL issuance, without any acceptance testing being performed prior to actual operational use. Consequently, this new system was being tested and shakedown problems encountered with little preoperational experience. Although the overa'.1 equipment quality appears to be good, some mis-application of equipment has been discovered. (5) Vendor support - The principal system vendnr did not provide adequate field support to Byron 1 after the equipment had been installed at the plant. Remedial actions which the licensee has taken to solve this problem and all of its nuances include completely redoing the original software, and developing a comprehensive onsite parts inventory and repair program to permit self-sufficient maintenance and repair of the system. E-26 L___________________________-_-______
(6) Coincident logic - The radiation monitoring system, as originally designed and installed, did not require coincident logic to generate an ESF actuation signal. Consequently, any single sensor reaching its setpoint for any reason would cause an ESF actuation. The licensee has modified the actuation logic to a two-out-of-two coincidence logic system. (7) Voltage sensitivity - The radiation monitoring system is extremely sensi-tive to minor voltage supply perturbations, As supplied, any voltage drop in excess of 105 V ac would trigger a sensor. The licensee has modified this system, raainly by changing monitor setpoints, to respond only if the voltage drops below 90 V ac. (8) Deportability - As originally installed, the radiation monitoring system , would detect noble gases, iodine, and particulate. According to the 1 intent of Criterion 19 of the NRC's General Design Criteria, only noble gases would have to be monitored as input to the ESF actuation system. This need is reflected in the Byron 1 technical specifications. However, the Byron FSAR commits to monitoring all three types of radiation. The licensee has determined that Byron 1 is unique in this commitment. Therefore, the licensee has begun modifications of the radiation monitor-ing system to remove the iodine and particulate monitoring functions from J the ESF actuation logic. j ( (9) Allowable outage times - The Byron 1 Technical Specifications reflected ] the original non-coincidence logic design of the radiation monitoring i system, in that they required the appropriate HVAC system to be switched over to its ESF mode of the operation within I hour of the discovery of a . system failure. Combined with the inability to obtain spare parts for ! the system in a timely manner, the original logic led to many unnecessary actuations of the HVAC in the ESF mode. The licensee's modifications to the actuation logic and the accompanying appropriate changes to the technical specifications should avoid this situation. t
- 2. CALLAWAY In response to the large number of ESF actuations triggered by the radiation monitoring system, the licensee created a special task force to determine why l this system was causing so many ESF actuations and to recommend appropriate !
action to be taken to prevent further actuations of this type. As a result of its investigation, the task force identified five types of ESF actuations ! initiated by the radiation monitoring system, and recommended specific l corrective actions for each type as follows: (1) Downscale trip due to loss of power (blown fuses, breaker trips, bus ! transfers, etc.) - Responsible for 14 actuations reported in 11 LERs. i The licensee eliminated the downscale trip function, j (2) Spurious high radiation alarm and ESF actuation due to vacuum transducer failure - Responsible for 10 actuations reported in three LERs. The j licensee replaced the original vacuum transducers with an improved design. l E-27 l l
(3) Spurious high radiation alarm and ESF actuation due to "100 Times Overrun" software incompatibility - Responsible for 10 actuations reported in two LERs. The licensee eliminated the overrange software, which was discovered to be incompatible with the Callaway radiation monitoring system hardware. (4) Spurious high radiation alarm and ESF actuation due to electronic noise induced during work evolutions - Responsible for three actuations reported in two LERs. The licensee prevented this cause of ESF actuations by requiring that all ESF monitors be placed in ESF bypass during performance of all work activities. i (5) High radiation alarm generated by containment atmosphere monitor resulting l in ESF actuation - Responsible for three actuations reported in four LERs. 1 This is apparently a unique requirement for Callaway. The licensee is preparing a proposed change to the technical specifications to remove the requirement that the ESF trip setpoint be set at a noble gas concentration 1 l of 9mR/hr during normal operation, and to less than two times the activity I I during purging.
- 3. WOLF CREEK Like the Callaway and Byron 1 plants, the Wolf Creek plant has also experi-enced a significant number of ESF actuations due to equipment problems with ;
the radiation monitoring system. The most persistent problem manifested at Wolf Creek was an incompatibility between the hardware and the software in the j RM-80 microprocessing unit for the radiation monitors. The licensee worked in i conjunction with other affected licensees, including those at the Byron and i Callaway plants, to develop a solution to this incompatibility problem, which has been discussed previously. Once this solution had been implemented, CRVIS actuations caused by the radiation monitoring system were significantly reduced. One other source of ESF actuations involving the radiation monitoring system, which was caused by radio frequency interference, was addressed by placing strict administrative controls on the use of hand-held radios in the vicinity of ESF equipment.
- 4. PALO VERDE 1 AND 2 The radiation monitoring system has been a major source of problems with the B0P ESFAS. The radiation monitors, manufactured by Kaman Sciences, are a one-of-a-kind, state-of-the-art design and are extremely sensitive. Problems with the monitors have caused actuations of the control room ventilation system and the fuel building ventilation system. Causes for the problems have included loss of power supply, spikes in power supply, personnel testing errors, electrical noise, wrong filter changes, grounding problems, inadequate seals, and conversion factors. A number of actuations were caused by incompatibilities between the system EPROM and its software.
- 5. WNP-2 At WNP-2, the licensee reported that the radiation monitoring system was a major scurce of spurious ECF actuations. Corrective actions which the licensee took to reduce the number of ESF actuations caused by electrical noise E-28 l
i
include (1) changing tb detector grounding scheme, (2) reducing the internal temperature of instrument panels by the addition of ventilation paths, and (3) suppressing relay actuttion-induced electrical spikes by adding thyristors or triacs to the actuation circuitry.
- 6. SUSQUEHANNA 2 A number of ESF actuations involving the standby gas treatment system (SGTS) and the central control complex environmental control system (CREOSS) at Susquehanna 2 (isolation of these two systems is coupled) were caused by radiation " shine" during outages. To prevent further occurrences of ESF actuations of this r,ature, the licensee provided shielding for the detectors.
- 7. HOPE CREEK Although the Hope Creek plant experienced a number of ESF actuations (namely, control room emergency filtration system actuations) that were initiated by the radiation monitoring system, the problems which caused the actuations were somewhat different from the experience at other new plants.
The radiation monitoring system currently installed at Hope Creek was not the original monitoring system. Initially, the Hope Creek radiation monitoring i system represented an advanced state monitoring design. However, the vendor l of the originally installed system went into bankruptcy, and could not provide the licensee any technical support or replacement equipment to resolve the i problems encountered with the installed system. During the early months of l operation, the plant experienced a number of spurious actuations of the control l roon emergency filtration system (CREFS) that were due to equipment problems (particularly electronic drift) with the high voltage power supplies for the l radiation detectors. In response to the first two occurrences of this type, the licensee recalibrates the malfunctioning power supplies and established a program to monitor and trend system performance data. As problems with these power supplies continued, the licensee determined that the installed system was extremely susceptible to problems durino periods of high humidity. Because i of the recurring problems with the originally installed equipment and the i original system vendor's economic plight, the licensee elected to replace the radiation monitoring system with a more widely used, state-of-the-art desic". The decision to change the radiation monitoring system design created a numas of operational problems because it meant that new equipment installation anc testing were being performed during the startup of the plant, a situation whict, would not normally occur. Ordinarily, all of this work would have been com-pleted during the pre-operational testing performed before licensing and startup. The replacement radiation monitoring system design now installed at Hope Creek has had its share of problems, as evidenced by the experience with ESF actuations initiated by this same system at other new plants (e.g., Callaway, Byron 1, Wolf Creek). As a result of their experience, those licensees (including Hope Creek) which use this particular radiation monitoring system at their plants have formed an owner's group to pool their efforts to resolve the problems encountered with this system. E-29
l APPENDIX F TECHNICAL SPECIFICATION VIOLATION DATA l } l l 4 l l F-1
- r o r r - t e s S nl e ee r n ix N n uw m n do m I O n g po r en i s surhoeck i T eqet rn A DEPOPU L 2i52Dn 2 Z2 E5ZC O I ( t V o o N O2 I TE AR CI U I FG I CC 7 EM P , r S L . r' A r' C I r N ) H <1 C E T 1 jI 4 F e r u - - - g j i F 0 0 0 0 4 3 2 1 mC9.M o5 g ajEoZ
,a 1 ,'
1 8 6 4 2 O r ' o r a 4 r # 2 t e s S nl ee e n r N n m n n uw do 2 2 O I i g po r en i s eck surho T eqet rn 0 A DEPOPU 21 9 23 2 L 2 Z2E52En O 8 O ( C 1 I - V1 a 6 a NN 1 OO I Y 4i _ TN 1 L AA C I C ' PF 2 1 e O F ' O I i CL 0 1 EB P SI A 8 h t L D M A o ' 6 C I C I N 4 H t C E 9 2 T L OO 2 F m g 0 0 0 0 0 H 4 3 2 1 a yj bu)Z c C o 3 ~ 9. > I
,.l
r o r - r t e s S nl ee r n e N n m n douw O I i n g po r en s T surhoi eqetrn eck A DEPOPU 21 5I W L 22DDn 2 E5E C O I [ V N OY I TA AW CA I F LL I CA g EC P S L A i V C I N o! 2 H C E T 9 3 F e r u - - - g i F 0 0 0 0 4 3 2 1 m - % aj$DZ
?^
1 8 6 4 2 r ' o r ' 4 r 2 t e s S nl e ee rn to N N m n uw n do t
.{
' 2 2
O n g po r en ' surhoeck i I i s T e q et rn 0 A DEPOPU 222352D U 2 L 2E5ECO O I ( N 8 1 e V - 6 a N \ m 1 O1 I 4i TA 1 L AB O CW I FAT O 2 1 I o! '
#C CA i
i( 0 1 E C P o { h t S O 8 L o s P F M A Y /6 C C I @I N T1 M 4 H 9 C E - 2 T 7L 4
- OO F '
e r - - - u s i r 0 0 0 0 0 4 3 2 1 n rC2.2~9.> g Ly$e2 7-
r o r r t e s S nl ee e rn N nrn nn uw do O I g po r en i s surhoeck i T eqet rn DEPOPU 2 A L 2S 52 5 2D-1 EE52CF a O I V m o N O I O m T 1 o AN f-CO FR I N I Y C B E
}
{ P . E S > L A - C I N H f _ C E T ~ / )- c 91 F - e - r - - u g i 0 0 0 0 F 4 3 2 1 n cCo3 3> - ayhsZ
1 8 6 4 2 r o r r te s S nl e ee r n N n n m n do uw O I i g po r en i s eck surho T eqet rn DEPOPU A 21Z2 51 D L ZE50CO O I V N K O I E 0 o D TE AR o CC I o d FF I C L EO PW S L d A 2 C I / N / H 6 C E T 6 F e r - - - u g F i 0 0 0 0 4 3 2 1 mCo3~3. > $faZ a
039u ~d2. c -U 0 0 0 0 0 0 0 0 0 0 0 1 8 6 4 2 g r , o r 4 r , 2 te s S nl ee rn e , N m n do n uw , 2 O I i n g po r en i s eck , 2 T surho eqet rn 0 A DEPOPU 2 A3521 I L 5 v2EZL N O I r
, 8 1 e V2 , c n
NN , 6 1 a u OO , s
~
I Y 4i s T AN 1 L CA O I C .
, 2 1 e F o O c I O p C n C L 3 0 i
s E P B \ m 1 h s SA I 8 t n L D - m M o A 6 - C I I a P N o F
/4 H C I
C E e 2 T m L a e OO r , e r - - - o 9 i 0 0 0 0 g F 4 3 2 1 o cCo5 6~>%Oaj EaZ Llllll ll' llll[ ll1
r ' o r ' r t e s S e n ' nl ee r N n mn n uw do O I i g po ren i s eck surho T eqet rn ' A DEPOPU 21Z252D-1 ' L 2 E52Cr O [ I V N O 3 ' I E TN AO C T I FS I L CL ' EI PM- _ S ' - L Q A - U C I i1 ig - N H i C - E T 2 8 e - e - - r - u s 0 0 0 0 0 i r 4 3 2 1 n r
~
e~3.> g uj$cZ c - i
r o r r t e s _ nl e n ee r E n m n n uw do D i g po r e i. i s ec surhok T eqet rn A DEPOPU 1 Z52 2552DF_ [ 55ZC 2 A B W A T A C L o o 9 F e r u - - - g i F 0 0 0 0 4 3 2 1 n cC ~3.>%Opy g3Z Io
r o r r t e s S nl ee e n r N n n uw m n do O I g po r en i s surhoeck i T eqet rn < A DEPOPU 3I 5I L 225lZ2GDOiC O I ( 1 V N O2 I TE AIC C I U FL I C S E T P S L A C I N o H o C E ? T j I 0 1
]
F e r - - - u g - i 0 0 0 0 - F 4 3 2 1 - n cd o3e~ c ' > g aj6DZ E~
r - o r r t e s - S nl ee e n r N n n uw m n do O I i g po r en i s eck T surho eqet rn A DEPOPU Z1Z2523 L ZE5ZGn O I o-V o N3 h OD I TR AO CF I o 9 F ER I CT EA g E PW S L - A 9 O C I N $ R H / O C E T n - F e r a 0 0 0 0 9 i 4 3 2 1 r n c@5 O.> %O 'y]E)Z Z
r o r r t e s S nl ee e r n N n n uw m n do O I i g po r en i s eck surho T eqet rn A DEPOPU 222352D L 2E5ZCn O I
?'
V N1 OE I TD AR C E b I F V I CO L E PA S P O L o A o ' C I j N H h C E T m 2 1 F e r - - - u g i F 0 0 0 0 4 3 2 1 m r$ g jeaZ 7t
r o r r t e s nl e ee r n E n n uw m n do g po r en O i i surho s eck eqet rn M DEPOPU a3523 ZE52Gn 22 1 W N2 E
}- D R
E V O L pA gP g o { 0 g
]_
3 1 F e r - - - u g i F 0 0 0 0 4 3 2 1 md 9. %~OI a]65Z
,h
r o r r t e s S nl ee e r n t N nrnnn uw do o m O I g po i r en s eck
/
i surho T eqet rn A DEPOPU 6 m L 23523n 22 ZES2G O I V % N 3 O2 I TE AL m CL F I SA I m CA o E P L S O [, - L r, A m C I N b H f C E o/ T 4 - 1 - F e - - i w 0 0 0 g - r 4 3 2 i n cbs~ - a]E)7-I*
r o r r t e s S nl e o ee r n N n n uw m n do 9 O I g po r en i s surhoeck i T eqet rn E A DEPOPU 3i5GD-l L Z2 ER5liCT O s I ba V2 Rh NR OA T E I o 6 L _ A C C U I FN I C . C EH PS C SA 7//, L W o A < / C I N f H \ C E T 5 1 rl F e r - - - u g i F 0 0 0 0 4 3 2 1 nC9.Y o5 g a]g c 5'
,l r
o r r t e s S nl ee e r n N n m n do uw O I n g po r en i s surhoeck i T eqet rn A DEPOPU 23 L 232E6O 2 ( R 2 O I V 2 NA ON I TN m AA a s s CH I FE m I C U 5 E SQ a P r SUS . 3 L A m C I a r N a r H C , E lI T 6 1 3 F , e r - - - u g i 0 0 0 0 g F 4 3 2 1 n cCo5g5 g ajEaZ 7'0
r o r r te s S nl e ee r n N n uw n m n do O I i g po ren i s eck surho T eqet rn A DEPOPU 31 I 3 S L ZEEl2BGGO O A i I V N o 9 O I K o' TC AI j CR I F E I CI M ' EL 5 P S L A l C I N H C E T 7
)
1 F e - - - r u g 0 0 0 0 i F 4 3 2 1 n %jgeZ cC94"oI 1 I"
r ' o r r ' t e s S nl ee e n r N n m n uw n do O I g po r en i s ' surhoeck i T eqet rn A DEPOPU 2 Z252D 3I L E52Cn O I [ Y V N 9 OM I TA AH C E I e FR O _ I C E H P S S o ' L 9 A C I N 9 H C - E T 8 1 F ' e r u - g i F 0 0 0 0 0 4 3 2 1 n: ccoae~9.> c g aj $zZ 76 I
lfIIj E $ o 4 - 50 2. {. 0 0 0 0 0 0 0 0 0 0 0 0 _ 1 8 6 4 2 g r , o r , 4 r 2 te s _ S nl ee e r n N n n uw m n do
, 2 2
O I g po r en i s surhoeck i T eqetrn DEPOPU 0 2 A Ad521 I , L Vbo 25Ztx r O I
, 8 1 e c
V n 6 a _ N e 1 u s O I 4i s e - T 2 1 L AI m O CM I o 2 1 e c F I E R r P i n CF 3 0 1 s E e s P h t S 8 n 0 O o L M A 6 C I o P N - 4 F 4 H C I C 2 E _ _ T L 9 1 eO0 F , e r - - u g i 0 0 0 0 g F 4 3 2 1 n:co5 o.> %o aj63Z r ms i - 1 , ;' it
r ' o r r ' te s S nl ee e r n N n mn n uw do O I g po r en i s eck ' i su T eqetrhorn A DEPOPU L 2lZ252Dn [ Z' E5ZC O I V ' N OD I N TE AB C R I F I E ' CV ' _ EI PR S L A U C I T N H oI - C 9 E s T 0 2 e '
=
i 9 0 0 0 0 0 - F 4 3 2 1 nCo c g.> g ajEDZ
r o r r t e s S nl e n ee r N n mn n uw do O I i g i po r en s eck surho T eqet rn DEPOPU A Z1 523 L Z2 ZE5ZEO O I V N O I T AR Y C I R FE C P I E P S L A C I N H C E T 1 2 F e - r u g - i 0 0 0 0 F 2 1 4 3 ndo5g3 c y]6aZ I"
r , o r , r t e s S nl ee e n r N mn n uw do O n g po ren surhoeck i I i s T eqet rn , A DEPOPU 21Z252D L EE52Cn O I V , N O I EK TE AR CC I F E I C P , E O PH , S , L , A , C I N , H , C E T /I 2 2 DO F , e r - - - u g i F 0 0 0 0 o 4 3 2 1 n - c C o D' e* 9. > % 0 a ] 6 D Z c TU
l l l j APPENDIX G TECHNICAL SPECIFICATION VIOLATIONS BY SYSTEM This appendix provides the number of violations of technical specifications for each of the 22 plants in the study according to the plant system involved. The EIIS system codes are used; a key for the codes is also provided. l l l l l G-1
b eh g C, C C, C, - O O O OOO O C4 O O OO O C. O C, C, - O o- in O n n P4 C e . e, o O nO +-* r. - N .-e C e-e - e-.J e e4 n O.s P' C.3 e< OO O OOO O O O ** C2 O C, O O O O Cs O ON
.,OOCOO OOO -O O O -O OOO O O O O Cs N
*-e (f S P4 ep e C 9 O V8 O e4 OO OO O o' 4 e-o *4 CJ O O e8 NO n 40 .-e m
( e3 N e4 & O. O - e4 N O O M,w CN N O O NN C) M N O W 0-O e4
- h. W4 OO O OOO O O O O O O O O te oO O OO O -
La. N .e C= t(') Ca NO e4 O O F3 e-e ce e4 O Cr ee O OJ O ==e *4 ea N La's N
- ew Cn OO O Cs OO O O Cs Cs O ee . -e O O O Ca O O O O N La i
- eo Oe4 O O OO C> O O O O Cs O O O O O O O "O Yb 4AJ as O eeO Ca O OO Ca oO Cs O N C.* O O O C, O O C.* - *
- LsJ
<3 ' O C* O OO C* O O O O *-e O ** O O Ce O C O O G'A ce N .
14J La . O C* O O e-e oO Cs O O O O O O O Cs O O Ca O ** ON Gs LAJ O C.s O OO O e-* O O O O O O O O OO O O O NOM
- s. a LJ O O Cp O e4 O O C# C O Cs O O O O OO O C8 O O O we Ut O en C "J Ltd OOO O O OO O O f4 O C# C Ca *-e N O O r4 OO O f ==
e-y..o u c s:La N ee O O O OO O Cs O O O 04 O O Cs O O O NO .e GJ
..J LJ C3 5:a
>4 LaJ G O O ** OO O O OO O O O r4 O O O O OO O O O Vb
> > LJ t
2 C:2 >-- O OO O O O O O Cu e-o O Cs O Ca O Cs Cs O ** Co NO
- LJ > CR 4
*ad Z e-- =~e Us O OO O O OO O O O O Ca e-o OO Cs O C. O +* O Ce PJ
.3 als L .e M.
*~e LAJ O: OOO O Gs O O e4 O eo O Cn O O C' e-e O O Ca O O O Va Le W- sns e*4 e--e U's
& L5 >- O De +4 =-* C.) O O O OO O e-e O F3 O OO C"S Ce OO O e4 ===e La s Ua saa -
O Q. us o ee v e .-, we .e C. C. C, O O L. >> en- O O C, C. .4 O O e4 en g u oO ** O C, O O cO ct e4 c.a O CD CA O N O O NO N O C* 'p--'* O aze e4 C 'M :r O OO Cs O O r4 O
- O C> O O O Ca c. O O Ca O . - . ON
>- .1 L1 Ela L48 **3 O Cr* O O C' O * *
- O O (" 4 C# O O O O O Ce C4 8%4 O O C4 hf1
>- sla ee 4" a C> - O O O N C's e- o O O e-e Ca O C2 Ca O O *'e O Cw Cs 4J ee $3 8
.4 c.e f O OO OO O +* ** --e O O O O O O Cs O O N C2 ** O +4.1 s als b--
t/ D LD C8 OO O C3 OO Ca ea C'e O O ("4 Ca O O O O O O O O *-. 53 e
&# La # D 4** O O 04 O O O C2 C'A O *e 04 Cte Q O O O O e4 O O F'=.
u se
. I gs O C2 C3 e' e O -e O C4 CA O O O V1 Ce O O w4 C3 N O C2 Cb O O Ga **
A tsJ Lb C* C2 O OO Ce O C* O O Cs t4 O O O O O C= O ("% O ON ar ska saa eo O eg oO N t4 O O O C8 CJ O O w-1 CD O O O Vs v4 O g=e as .-e C.3. C. C, esp e. e O f4 O O O O C, Cas O O O O Ce C O O C.m ON 04 CD O C4 O O D O ** O e"e VS C3 O 4) O e1 O O CD CD Q OJ O GJ G LJ O e-* O Cs O e4 O C .* O O O Cr O O Cr co O O OO O O CJ G G4 4J Ce Ci Ce td Cs O O O O C. (g og e-e C,3 gsJ C2 yp O e.e Cs O g(-) CI- N 4 ed O ter O O C/ N O O- .- e O Ps r4 N e4 O O C, CD C Vs O O) EI N r-e t0 4 lir N L' ") La ,e e-e r4 4 e4
> >- Ys N 4 YJ Le e +< e Y F N' e.4 La J Lai gi a e4 E V 3
ee NG &" L44 04 Lak N a s 8' La.3 .s. CJ LEJ
<J LB LJ La J h' e"a s'r (r~1 E L LJ 2: ** G n" La a
# e4 ct G & 04 Of La A L .J O La # La a L4 A e- o GE e e3" LJ IL C.J Cr las Ela ne E-3 LJ A a- a b- o > ;>> ;3 *3 LA f LA 4.J
> F CL X 3e - a ..J **
L.if A p- U3 *e >- O' La d J D O' P4 .3 Y. C. ) fG G 6t a ti e S- 11J G LaJ _a r i C1 Ca LY La ) fr O L4.J t &A. <I. cJ G .A>- >-- s.L m3 La : C4 UJ 3: J E.ft 8 . .J O'. > C-'8 e U2 F- O J >-
.. A D+ G G G H We LAS ( aC >4 *--e 4A G. Et La # b-e '1. 6- D G liiF" C1 C.3 Ca e3 s L J E s ( e 5'.* C1 LA. 'J : e.A E: 3. as . e (a, O- Us U3 Us 32 18 3 6-G-2
7 l l 1 a w s E3J N EP VD N Vs tr3 pC6 @5 ' s-- 41 er ** (J =d ) er O
- N N EP- e4 er C3 tr3 er O Vs e-o N Vm es e1 gw'pN ** N *** e=e e4 s4 W3 FD tr) ad3 H
M OOO O O O O Cs OO O O O OO Cs O O O O N <e V4 N W O Cs D O N D D O C3 *9 O N $#1 *4 OM Ce che *9 O C3 O $4es na
>=e O OOOO OO O O O C.7 OO O O O *'d O O O O .* N
==
M M C OOO O DO OOO O 'O O O O O C=3 O D Cs2 OW M
*-4 3
O N O O CJ O OO OOO O e9 O O O C
- O O O O O F3
- h. OOO O C3 OO ONO O O O Ce O N **e Cs O Cs O Ca tr3 3
Ltd 3 N N O O e'4 O OO OOO O er OO O O OOOOO O* r3 A O O O O ** O O OO O Ca O O Cs O e4 OO Cs OOON .
)
es O OO OO O OOOOO O OO oOMOO OO O es e* N e4 O O N O .-4 O O se eo o e-o e
- O OOOO O e9 e4 edi **
U1 3 OOOOO O O OOO OONO O OO O OOOON U) OO O ** O OOOOOO O ** O O OO OOO OOON 2 C3 fTt O O OO OO O OO O O O O O O O C8 OOO ed O **
- =>
G G OOO OO O NOOO C3 O OO O O O O O OO ON
.A O **O e-e % O O O +-e F3 O O OOOO O OOO *-e O O O O ad3 > us O
w esh 6xn O O e4 mO O O O e4 ==e O N O e=e O N Co O O OO O CF C3 > til w~e .- J
>- CD SC O OO O ** O O O O O OO e --e O O *e COOO O **
C **
- d'L.
L) M e-a ** *1: O Vs *-* O O ** N O ==e O OO e=* v4 O O O O O ** e-4 ene V3 e* GA E N .-e s-l E.3 uJ (D Ca ooO C, C) +-e O OO OO OOO O OOOOO C, e
- kJ >~ E O a as un n- u. O OO O- s O O OO O w~ O O Cs O Cs OOO O Ca O e-4 y V2 x e- E <r C O O O O O O O O OOOO O OOOO OO O .-e D L2 2 (O E* w
- s. 2: u N s s Vs O N O Ca O O ee O N ars O O N O O N trl V8 94 41, r Le La) C"3 Fi OOO O Ca OO O e* O O e4 O OOOOO O O O V8
>-- W Q. OONO ee O O O O N O N ee O O O O e4 O O M Ner *-e he:
- >o.
C9 2' OO O O O OOO O OO ee QOO OOO OOO O **
- 3 lan:
up 2.* O ON OO OOOO O O OOO OO OOOO ** OM M b-E Laa O O O O OO -e OOO O OO O O *4 OO OOO ON r.3- M J A. K OOOO OOO OO OO C. O O O O ** O O O O O *-4 W M LA E ee F1 tr3 w O N ** O W *4 O e-4 ** e4 e4 O O e.e O O 04 e4 *4J d 2 e s' N j _J COOOO O O O O O O e4 OO O O O O O O O O ** es
% *'O OO OOO O O CA O O OOO OOO OOO **4 ON La4 O .-e O O O C3 e-o C O OO NNO O OO OOO OO =E.8 es ca O O (* 8 .-4 N O OOO ONO O OO OO OO O OON es L.3 ** O ** O eg C O O O e4 O ya O e O . og ee O e.4 ==e -e e N **
SI's O O * '* O O N OOO O OO O O O OOOO O O O VS em D O e=e O OO O O O O OO O OOO O O O Cs O w Ca s'4 -
>-e e OOO
- O O Cn OO O ONO O O OOO+*OOW
- N at E N C2 C3 o-o ee N e-o e-e
- Wa >- VJ N E 9* ) i 8.43 e4 3 Y ** 4.AJ 84 8 C:3 u-4 E M- ;
N: *
- N 4" C hm !
LaJ N LaA ed Ca sts 2 tsJ R C3 uJ
=
4 d N n Laf L iJ *a - H O kJ sr -:= > La5 b ** 4
>-- x a se a" O m masO _J O_ a G G G4 584 F L&4 E
-e _.a sr 64J un "
A . .J
>- a CY. Ls'4 di O-a. JM > Nw. C3 8 E J
e:I' sa". JA b-- M E a W. Ca. V8 X- J 8.f3 OJ C. J3 Ca ' De a ts a Q. >- J P- G G G w *-e ta l C.9 4 e- e Q. De L a LJ L.3 CD dlll3 4A. 1'
- J .J
*--a ts 4 F. E n. 4 EL La Qd.sk:
>e llE.: P. - '3
(#1 is3 44 8 3 m
@E C3
>=
G-3
Table G-3 Ells SYSTEM CODE DEFIN!TIONS CODE SYSTEM AA CONTROL R00 DRIVE SYSTEM AB REACTOR COOLANT SYSTEM (PWR) AC REACTOR CORE SYSTEM AD REACTOR RECIRCULAi!0N SYSTEM (BWR) BA AUI!LIARY/EMER6ENCY FEEDWATER SYSTEM (PWR) BB CONTA!NMENTCOMBUSilSLE6ASCONTROLSYSTEM BC CONTAINMENT ICE CONDENSER /REFRIGERAi!ON SYSTEM (PWR) BD CONTAINMENT LEAKA6E CONTROL SYSTEM BE CONTAINMENT SPRAY SYSTEM (PWR) - 96 H16H PRESSURE CORE SPRAY SYSTEM (BWR) BH EMER6ENCY/ STANDBY BAS TREATMENT SYSTEM BI ESSENTIAL SERVICE WATER SYSTEM BJ HISH PRESSURE COOLANT INJECTION SYSTEM (BWR) BN REACTOR CORE ISOLATION C00LIN6 SYSTEM (BWR) B0 LOW PRES 5URE COOLANT INJECTION SYSTEM (BWR) BP LOW PRESSURE SAFETY INJECTION SYSTEM (PWR) BQ H16H PRESSURE SAFETY INJECTION SYSTEM (PWR) BR STANDBY LIQUID CONTROL SYS!EM (BWR) BS ULTIMATE HEAT $1NX SYSTEM BT SUPPRES$10N POOL MAKEUP SYSTEM (BWR) CA BORON PECYCLE SYSTEM (PWR) CB CHEMICAL AND VOLUME CONTROL / MAKEUP AND PURIFICATION SYSTEM (PWR) CE REACT 0h WATER CLEANUP SYSTEM (BWR) D0 DIESEL F:JL O!L SYSTE9 DE FUEL O!L RUE! VINS, STORASE, AND TRANSFER SYSTEM DF NJCLEAR FUEL TRANSFEF SYSTEM EA MEDIUM-VOLTA 1E POWER SYSTEM (601V THROUGH 35 (V) EB MEDIUM-VOLTAIE POWER SYSTEM - CLASS lE El DC POWER SYS1EM EJ DC POWER SYS'EM - CLASS lE EK E9ERGENCY OUlTE POWER SUPPLY SYSTEM i' FK SWITCHYARD !ISTEM
!C FIRE DETECT!3N SY! TEM 'i IB INCORE/EIC0FI NEUTRON M0hlTORING SYSTEM
!J LEAK MONITLRING SYSTEM IK CONTAINMENT ENVIRONMENTAL MONITORING SYSTEM IL RADIATION MONITOR!N6 SYSTEM IM TEMPERATURE MONITORING SYSTEM IN SE! Smit MONITORING SYSTEM IP POST-ACCIDENT MONITORIN6 SYSTEM IV VIBRAT!DN MON 110RINS SYSTEM JB FEEDWATER/ STEAM 6ENERATOR WATER LEVEL CONTROL SYSTEM JC PLANT PROTECTION SYSTEM JD REACTOR POWER CONTROL SYSTEM JE ENSINEERED SAFETY FEATURES ACTUAi!0N SYSTEM JJ TURBINE SUPERV!SORY CONTROL SYSTEM JL PANELS SYSTEM JM CONTAINMENT ISOLAi!0N CONTROL SYSTEM KA CONDENSATE STORA6E AND TRANSFER SYSTEM KE HEAT REJECTION SYSTEM KM CHILLED WATER SYSTEM G-4
Table G-3 (cont'd.) Ells SYSTEM CODE DEFINITIONS CODE SYSTEM KX SAMPLING AND WATER QUALITY SYSTEM KP FIRE PROTECTION SYSTEM (WATER) KG FIRE PROTECTION SYSTEM (CHEMICAL) KR FIRE PROTECTION SYSTEM (PAS $1VE) ISPECIAL NON-Ells CODE)
. NA CONTROLBUILDINS/CONTROLCOMPLEI NF AU11L!ARYBUILDINS N6 REACTOR BUILDINS (BWR)
NH REACTORCONTAINMENTBUILDINS PM PLANTMANASEMENTISPECIALNON-E!!SCODE) SB KAIN/REHEATSTEAMSYSTEM SJ FEEDWATER SYSTEM VA REACTOR BUILDINS ENVIRONMENTAL CONTROL SYSTEM VI CRYWELL ENVIRONMENTAL CONTROL SYSTEM IBWR) VC SHIELD ANNULUS RETURN AND EIHAUST SYSTEM V6 FUEL BUILD!NS ENVIRONMENTAL CONTROL SYSTEM VI CONTROLBUIL0lNS/CONTROLCOMPLEIENVIRONMENTALCONTROLSYSTEM VL PLANTEIHAUSTSYSTEM WD L10VID WASTE MANAGEMENT SYSTEM WE 6ASEOUS WASTE MANAGEMENT SYSTEM (PWR) We CFFSAS SYSTEM (BWR) W! STEAM SENERATOR BLCWDOWN SYSTEM (PWR) UK EQUIPMENTANDFLOC1CRA!NSYSTEM II OTHER KN0 H SYSTEM - SEE COMMENT FIELD ISPECIAL NON-El!S CODE) ZY MULTIPLE KNOWN SYSTER - SEE COMMENT FlELO ISPECIAL NON-E!!S CODE) ZZ UNXNOWN SISTER ISPECIAL NON-E!!S CODE) G-5
S 4 ob APPENDIX H LOSS OF SYSTEM SAFETY FUNCTION DATA H-1
. 44
0 0 0 0 0 1 8 6 4 2 O r ' o r o 4 r 2 t e s I nl e n ee r <' uw i1-n 2 N 1 n m n do 2 g po r en O I i i surho s eck eqet rn 0 T DEPOPU 2 C ' N 2552]O 25-EAZE [ 5 8 U e 1
) c F o '
n 6 a Y 1 u T2 E
' 4i s
s F E 1 L AR I O 2 O e S U r
' C 1 c
MG E C s 0 0 1 i n s TM l I s S h t Y 8 n S ' o M F ' 6 O ' S - 4 S O ' P L %I I'
' / 2 F
C f I 1 L OO 9 0 0 0 0 0 4 3 2 1
?cD>c ) %O Lc) 6,Z C
fm l'
0 0 0 0 0 0 1 8 6 4 2 o r , o r 4 r , 2 t e s nl e , ee r n n uw 2 N n m n do g po r en 2 O I i i surho s eck T eqet rn DEPOPU 0 2 C 3352D , N Z25ZCn 22 O 8 U , C 1 e c F1 _ n 6 a YN o , 1 u TO 0 , s EY o6 4i s FN 1 L AA SC , P 2 e F 1 O c MO / i n EL 0 1 s TB s SA YI A_ no 8 h t SD _ M F O , 6 O bI C S / 4 S e O L e 2 e 2 9 L H
&O O e ,
r u - - - g i F 0 0 0 0 o 4 3 2 1
>o % O L c) )2;
="
l l
0 0 0 0 0 0 0 0 0 0 1 8 6 4 2 o r , o r o r , t e s nl e , ee r n n uw a N n m n do g po r en o r O I i i surho s eck _ eqet rn T DEPOPU C 2S51S2 ] N ER565n FE U F , n Y a r T Y , E A m F N , AK S , L ML , EA g , TC , S Y zr S if{ m O F o , C 6 O o , P F S I
,/4 S aI C
O r L m 3 5 n L e
.OO r
u . g i F 0 - 0 0 g 4 2 1 0>o uO a aZ F^ l'
0 0 0 0 0 0 0 0 0 - 0 1 8 6 4 2 g r , o r 4 r , 2 te s nl e o r n ee N nn uw oI 2 n g rn do 2 O i po s eck r en , A i I su rn eqetrho - T DEPOPU m 0 2 C 2552]7 a N Z5 25 Z [X r I E 8 U m 1 e c F , n Y , 6 1 a u T1 , s E s FA , 4 1 i AB SW A m 2 L O 1 e A O c MT EA
, C 0 i s
n _ 1 TC t o s S m h Y o , 8 t n S o dP F o F M o a/6 C O i h I S . 4 S O a c L m 2 4 H
.L OO e
r , u g - i F 0 0 0 o 4 2 1 m*b>o %O o cZ x& ll( l' l
i] r o r r t e s nl e
" ee r n m n duw n
N~ n r. O I i g po r en i surhok s ec eqet rn T DEPOPU C 25523 N 52 E55ZGO U F Y a T lf E1 f l F AO S N R MY EB T , S ) Y S F O S 7t S . [ O 9 L s H e w - - a' 0 0 0 4 3 2 o>DwOue j sz
1 8 6 4 2 O r ' o r ' r t e s nl e ' ee r n n uw N n m n do g po r en O I i i surho s eck eqet rn T DEPOPU C 21Z2 1 2] N 2 [ R52[5O U F ' Y T E E K o o ' F E , f AR S C oI
\
f o MF L a ' E O o_ T S W Y i S F ' O / S / ' S O ' L 6 H e r ' u g i F 0 0 0 0 0 4 3 2 1
) >
$.Co>G "O a s2; zL
1 8 6 4 2 r o r r t e s nl e ee r n n uw N n m n do g po ren O I i i s surhoeck eqet rn T DEPOPU C 2352D N 2 ERLZCOG U F2 YN TO EY FN AA S C MO EL \ TB q SA YI SD F O S S O L 7 H e r u g i F 0 0 0 0 4 3 2 1 e> ; 6c)#ba2
0 4 2 1 8 6 r o r r te s nl e n ee r nn uw N nrn g po ren do O I i s i s eck u rhrno eqet T DEPOPU C 3iZ N ZE2lG3 iGn U F Y T 3 EE FN AO ST S ML EL I T SM Y - S - F O { ig _ S ' - S i O L 8 n e r u _ i 9 - r 0 0 0 0 _ 4 3 2 1 _ ) ) a>06OLcAbDZ Fe
I!l]I n euDo4 eo c LO 0 0 0 0 0 0 0 0 0 0 0 - 1 8 6 4 2 O r o r r ' 4 2 - te s nl e ee rn N nt uw 2 nrn do 2 O I i g po r en i s ec surhok T eqetrn DEPOPU ' 0 2 C i N a223521 L25ZL x r U ' 8 1 e F ' c n Y ' 6 1 a u T2 ' s E s FA ' 4i 1 AB ' L O SW ' 2 e MT A ' 1 c n EA ' 0 i s TC S 1 s h Y ' 8 t n S ' o F ' 6 M O ' S 4 S C O O ' I L c 2 e 9
- L H
$ O0 e
r ' u g i - F 0 0 0 0 O 4 3 2 1 3Ca>* vO ae#6 Z
=E i
sanoy IEo!)!JD 8 0 0 o o S 8 8 8 o l
~
8 t 2
~
$$ Ec Z
~
c ES 55G / o u.e o b gx / -
~ mac o c) CT c O k C QWACAD N.N o -
O N O ~ Z s@SD2] e k o - em y
> ~s bN
~
+ .s
~
w&
% A 4O O wm -
% c) o 2 [.
~
.E m
m9 m Pw - U
- h o w -
2 g - e O h w O g 8+ w O ~t
& O / -
N l r .- a 9= 9 52 $o 1 e p , , , l ? 8 8 S squaAO JO JOQLUnN H-11
r o r r t e s nl e ee r n N nn uw nrn do O I i g po r en i s surhoeck T eqet rn DEPOPU C 31 Z252D N ZR5ZCn U F N Y3 T E D i FR AOF S o M. E R ET o TA SW Y S o F 0 O S S O 6' L 1 1 n er u i s - - - r 0 0 0 0 4 3 2 1 mCe>eWOLeM63Z [ 'u
r o r r t e s nl e n ee r n uw N n mn g po r en do O I i i surho s eck e qet rn T DEPOPU C 21Z252O N 2E5 zen U F ; Y1 o\ TE ED F m R 6 AE S m V MO [ E L T A o S P { Y S o F O jo / S S O L 2 1 H e r u g i - - - F 0 0 0 0 4 3 2 1 B ;c e > e 6 O L )0 M E c Z bw
r o r r ' te s nl e ' ee r n n uw N n m n do ' O I i g po r en i surho s eck ' T eqet rn DEPOPU C 2352D1 ' N 2 2ELZCF
' Q U
F ' Y2 ' T E ' E F D A ~R S E V M O E L T S A ' Y P ' S ' F 0 ' O 0 ' S S e O L e 3 e 1 H e e ' i v - F ' 0 0 0 0 O 4 3 2 1
) )
?Co>G %O L0 sZ
- xg
r o r r U te s nl e n ee r N n n m n douw o
- g po r en O
I i i s eck surho T eqet rn DEPOPU t C t j I N 23122GD I 25ECO Z U U F Y T2 E ' F E A S L L A MS # EA TL # S o Y S F O
^ -
P S S O L 4 1 H e r u g - - - i - F - 0 0 0 0 4 3 2 1
?ceg -
t+ O 0#bsZ xh
r ' o r r ' t e s nl e ' ee r n n uw N n m n do g po r en O I i i s surhoeck eqet rn T DEPOPU C Z2G D 2i5I ' N EE5EiCO b-U FP o ' YR TA /r ' EE ' FL AC SU MN E T H S S Y A y SW " F ' O S S O L s t
=
9 - - - 0 0 0 0 0 - 4 3 2 1
?@>0 6O b#ba2; f
1 8 6 r ' o r r # t e s ' nl e n ee r n uw N n m n do g po r en O I i surhoeck i s eqet rn T DEPOPU C 2 2DI ZE2ZCr N I ' U F 2 ' Y T A E N F N AA SH E ' MU E T S Q d S U Y S S F O ' S C S # O L N d 6 1 a e e ' r u 9 - - i - r 0 0 0 0 0 4 3 2 1
) )
?Ce>Q 6 y0 bsZ xg
r o r r t e s nl e ee r n n uw N n m n do O i g po r en i surhos eck R eqet rn DEPOPU o 21 51 ] 7 E 2 Z2 [ E50Cn5 1 W Y T ECK AIR S E MM EI L H W o T b E n W d 7 h 1 H e r u g - - - i F 0 0 0 0 4 3 2 1
$.co>*LO LoMh y2; c.
.1'
r ' o r ' r t e s ' nl e ee r n n uw ' N n m n do g po ren O I i surhoeck i s eqet rn T DEPOPU C 25523-5 2 1 ' N E55ZEF e U e F i Y e TM E e F A AH SE e t R MO EH TS S M Y S F O S 9 S O L 8 W 1 H e r u - - g - i F 0 0 0 0 4 3 2 1 Bce>0 %O eA$aZ xg l
r o r r te s nl ee re n uw N n mnn do g po r en O I i i surho s eck eqet rn T DEPOPU C 3lZ252D N 2'5ZCn [ E U F Y T E F 2 AI S M R ME EF T S Y S o F O S S O L 9 1 n e r u s - - - i r 0 0 0 0 4 3 2 1
)
- 3Ca>0 u aMEdZ mg
r ' o r ' r t e s nl e ' ee r n n uw N n mn g po r en do O I i i s s u rnrheck o eqet T DEPOPU C 335I 3 ' N 22G Z25EEO U F ' Y T D ' EN ' FE AB S R ME EV TI R - S Y j S F O S A S O U L o 2 m 0 0 0 0 4 3 2 1
?ce>O%OLO#$oZ 3
x$
1 8 6 4 2 r o r r te s nl e r n ee nn uw N nrn do g po ren O I i i s eck surho eqet rn T DEPOPU C 335I N 22G D ER5ECn U F Y T E FY AR SR E M E P T S Y S F O S S O L 2 0 0 0 0 4 3 2 1
$.C e > o % O L a # 6 D Z .
?N
0 6 4 2 1 8 r o r r t e s nl e n ee r uw nn N ng rn po do ren O I i i surho s eck eqet rn T DEPOPU C ii2D N 2Z22E5ZCO U F Y T K EE FE AR SC ME E P O _ T S H Y S F O S S O L 2 2 H - e r u - g i F 0 0 0 0 2 1 4 3 S.Ce>e uO aOM6sZ {m
APPENDIX I MATURE PLANT DATA FOR NRR AND AE0D SIGNIFICANT EVENTS l l l l 1 i l I-1 l
c
.i
/ .Y Table I-1 -
1
. i r N:. incs : e. ceci : hD. enci i hD. inm . l 3: r[T PLAuf ;8R:Efft SY BRA :IA1!/10 AY FA :lt:Er0 P whR llRIEFID If t'R I I h3. Or n3*HS OF ; 44AABE e. I I i ih 983 I IP !*C4 I th 1985 I IN 1986 : TCTAL I OPERAf!Du CDNSIDitLD ' fvEr5/M;h?H :
......l. ,,I
....I_ I. I .I
?!! ' AO
. N/A i h/A i 1! N/A i !1 12 1 0.083 1 368 l AN3-2 I h/A I h!A 1 :I h/A i 11 12 1 0.083 1 334 :stadt vnaty i I h/A 1 h/A i CI h/A 1 01 12 1 0.000 1
!!! :3:8 A00t P:lti i N'A : N/A i 01 11 / 4 1 0I 12 1 0.000 l 251 lB80s45 FERRY l I h/A 1 h/41 2I t/A I 21 12 1 0.167 i 296 :lA3s45 FE9RY 3 1 6'A 1 h/A 1 31 k'A i 31 12 1 0.250 1 325 :8Ry4Ssl:K1 '
hs A l t'A 1 0I h/A i 0i 12 1 0.000 1 324 :ltyN$u:32 i k!A ! h! A I 1I II/A i 11 12 ! 9.083 1 317 04;vfRf CLIFFl 1 i h/A I t/4 l 2I h/41 21 12 1 0.167 ! 3:1 : 4.v!4f CLIFFI 2 : t'A I h/A 1 21 h/A 1 21 12 1 0.167 1 2tB ::00'il I N'A I h/A I eI h/41 01 12 l 0.000 1 302 l3Y5'AL RIVE 8 3 . h/A i b!A I 2I h/A I 21 12 1 0.167 1 346 :DAs:$-BE5SE h/A l h/A I 3l t/A I 3t 12 1 0.250 1 23? :DA!!:En 2 I n't i N/A i 1l h/A I II 12 1 0.003 1 249 :DRE50!4 3 1 h/A : N/A i iI h!A i !I 12 I O.083 1 33: :0JA4f AAhM . N'A l s/A I 2I t'A I 21 12 l 0.167 I 3:5 :C.C. C:Dr I i n/a : t /4 1 0l M/A I 01 12 1 0.0M i 316 :C.C. COCF 2 I h/A I N/A I 3I N/A I 31 12 0.250 l 348 lFALEY I I h/A ! h/A I 01 #!A I 01 12 1 0.000 1 364 lFAR.tv 2 l h/A I h/ A i 1: h/A i 11 12 : 0.083 :
!33 IFit!PA?R!:t I h/4 I t. A I 2l R/A I 21 12 ! 9.167 1 285 ;FOAT CALN3?h l N/A l n'A t 0I s/A i 0l 62 1 0.000 l 244 :l;h4A I hrA I t/A l 1I h!A i 11 12 1 0.063 !
4:6 lBR At: ELP.F 1 1 h/A 1 h/A I 41 h/A 1 41 12 1 0.333 ! 213 NA2:48 NE t i N'A 1 h/A ! 0t k/A I OI 12 1 0.000 1 321 INA':M i 1 h/4 I h/A I 41 t/A I 41 12 1 0.3!3 : 346 lHM:N2 i NfA I N/ A 1 1I N/A 1 Ii 12 1 0.083 1 261 M.I. R l #!3N 2 1 h!A I h/A i 1! h/4 l 1l 12 1 0.013 1 247 lle!As P::ht 2 I aAI n/A i 1I h/A t ll 12 ! 6.083 1 29a :!bC:An PO:nf 3 i t'A i h/h l l1 t'A I Ii 12 4.083 1 305 'rts4Jatt i N'A ! h/A l 2I t'A I 2l 12 1 0.167 1 409 ?LACR !$t f t'A l h/A 1 21 h/4 l 2! 12 : 0.167 3'3 lLASAat i . h/A : N'A i 0I N'A l 01 12 : 0.000 1 309 :#4;h! 14Ntt! ! N/A 1 h/A 1 3l h/A I 3 12 ; 4.250 l 369 .M:6.!RE I I n!A i n/A I 2I h/A I 21 12 : 0.167 1 245 'P:LL5'041 1 I h/A i b/A I 0: N/A : 0l 12 1 0.000 l 336 ;#:45':#E 2 I b'A I h!A f 11 krA l ll 12 1 0.083 1 263 :en':: tao I n/A l li/ A I 0l t'A 1 01 12 1 0.000 1 220 .h!4E n:.E P0lW' 1 t/A I N'A i 1I a'A 11 12 ; 0.083 : 338 1h3R'N ANh41 i hth 1 t/A I 3i N/A 1 31 12 ; 0.250 1 339 :h?'N ANNA 2 I b'A i h/A 1 3I t/A I 3l 12 l 0.250 1 269 :C:04E! 1 a N/A 1 t'A ! ! h/A I jl 12 ! 0.083 : 270 '0 :#Et 2 i h/A I h/A i 1I h/4 : 1l 12 ; 0.083 1 287 :3:04tf 3 n'A i h41 0I t/A l 0: 12 : 0.000 l 23 O'5'It CREEE 4/A : 5'A I 1I n'41 11 12 1 0.083 ; 255 9CSA (5 1 a/A : n/A I 1I t/A 1 11 12 1 0.083 l 2?7 'P(4:* 90't2a 2 I k'A 1 tra i j! N'A I J: 12 : 0.083 1 218 : PEA:M B:?f n 3 ! n'A i n/A l 4: t<A I &i 12 ; 0.333 : 2'? :Pl.6'!' I b'A : N' A l 2I hAI 2 12 t 0.167 266 :' 10 l[4:M1 1 a'A f k'A 1 21 t'A : 21 !! : 0.167 : 301 :P !C I!A> 2 l n'A 1 t'A : I: n/A : 1l 12 0.083 ' 212 at'p:[ !s,44: I: a. A I b'A 1 0 h/A t 0: !! l 0.000 326 :'R$;4:! !$.A4: 2: t'A : aa: 0: h'A l 01 12 l 0.000 1 254 'g.;A: (ltlEl i h/ A i t'A l 1l h'A l 1l 12 l 0.063 ; 261 :9sA: 01':El 2 l'A : t/A : 1 n/A : 1i 12 : 0.033 , 312 :844M 51:0 i tat t'A : 2; hA: 71 12 1 0.553 2'2 :SA.[' i N/A l 4!A t 0 n'A I O: 12 1 0.000 1 3:1 !$4.18 2 h'A k'A I 2l h/A : 2l l21 0.}gt t
' 3I 2^e '5A4 CGP! I t '4 l h>A l t'A : 3: 12 : p,2$0 3ei :Sch Ch78E 2 h3 : n/A : C: l'a : 0 12 0.000 362 '5Aa OCRI 3 t'A : N'A 1 0; k'A l 01 12 0.000 I ?P !!!Cben 1 : N/A l h/A : 0; tAl 0l 12 ! 0.000 l !?! .fD'.!C' A= 2 t'A 1 h'A : 2i h/A I 21 12 : 0.167 1 !!! :l'. if!! I i n'A ! N'A : 1I n'A i 11 12 i 0.083 :
39: :5J 4ts . k'A 1 N'A 1 4: t/A I 4 .. 12 : 0.333 1 280 :53'87 1 krA : n/A ! 0: h41 0: 12 l 4.000 1 28: ilJ'h 2
- k'A l h!4 t t: t/A : 11 12 ; 0.083 38' :1 '5: E*4==A l n't : n 'a : 3l h!A I 31 12 i 0.250 :
28' 0*8EE P!it 15.1: n'A : h'A ; 3: ble i 31 12 1 0.250 344 'ROJ44 t '4 : n!A : 11 N'A 1 Ii 12 l 0.053 1 250 MJRr[v P::a' 3 1 n's i t' A 1 4l b'A l 61 12 0.500 : 251 !?.e g, p; h* 4 h/A 1 h/A : 51 h'A ! $I 12 ; 0.4:7 : 271 lVth0C 'Aut'[I N'A l n'A 1 1l N 'A 1: 12 ! 9.083 4 029 !?anr(( RON! , hA: hA: 3; hag 3l g; ; g,og; ; 29 :!!D% 1 . D'A l t'A : 1: h>A : 1 12 . 0.01: 30s :llon 2 ! 4'A : n't ! 2; k'A : 2l 12 ; 0.167 AvtRASE 40. [4r5!Or* f 'A; ! 116 1
! . 116 912 1 1 F3R AJ Cil PLAC$
F8;" IS# BB:EF! A65* 4.!!? I-2 .
1 Table I-2 l
. I NO. [Vt4?S I ht. [44'S I ko.14N'l t 40. IVEhil i I I I I l Crit: pia 4' ; 14 1983 f IN 1994 t IN 1995 ' la 1986 NO. OF CNTH$ DF t 1 1 i A4AAGE ND.
- A :3Cl45 70 AEDO A : DAD!si TO Af 00:A;t0RDINS t0 AEOD'AOCORDING TO Af3D: TOTAL I OPIRATIOh CONS!0f AID I IVEll'$/CN'N I I I l-- I (
.....;&C l3 1 I n' A i 4/41 0t h/A 1 01 12 1 0.000 I j 38 l AC 2 N/A 1 t/A l 21 h/A I 21 12 1 0.167 l l l4 :tER t# VA. LEY 1 i N/A : t/A : 21 N/A I 21 12 1 0.167 1 55 :f:S RC t P :ht t h/A l a/A I 0I h/A i 01 12 1 0.000 1 59 :5R s4! Ff tet 1 1 h/A 1 N'A I 4: h/A I 41 12 1 0.333 1 4 llP]s4! FitRY 3 I h/A 1 t'A I 0I t/A 1 01 12 1 0.000 1 2! B4'ASel:r ! I t/A I # 'A t 4I N/A I 41 12 1 0.333 1 74 fptuNSu::K2 N/A : h/A i j 1
!I h/A I ll 12 1 0.083 1 4
!? t:A.Vitt OL rF51 l t'A ! n '4 1 0I h/41 01 12 1 0.000 t i B :ta,vtPt CL!FFS 2 I t/A l 4/41 2I s/A l 21 12 1 0.167 1 10 : CO'!R I n'A 8 N/A 1 3i t/A 1 31 12 1 0.250 l l2 ttav!?AL n:qa 3 I n'A I 4/4 I 2i 4/41 21 12 1 0.167 1 44 :D845 #E53! N'a : h/A 3I h/A I 31 12 1 0.250 1 37 :Dat$:(4 2 I n/A I hA1 2I N/A 1 21 12 1 0.10 i I 4' :DRE5:In 3 a'A I t/A 1 2i N/A I 21 12 1 0.1H I
) h TA%E A8C.D n'a : N/A I 31 t/A I 31 12 1 0.250 1
- 5 2.0. 000r ! , a>A 1 N/A : 0i N/A 1 01 12 1 0.000 1
< le D.C. : Or 2 I n'A : n/A I 2. h/A 1 21 12 1 0.10 I f 4R :FA'L!v 1 ! t/A : N'A 1 0i R/A 1 01 12 1 0.000 1 64 lFARLEY 2 . h/4 i N/A i 1i N/A i 11 12 1 0.083 1 13 l Fit:PATR::F I n'A ! N'A I 2I k/A I 21 12 1 0.167 1 15 1F:4' lAJL4 I n/A t h/A 6 2I n/A I 21 12 1 0.161 1 44 6:Ha . 4/A 1 n'A I 01 N/A i 0l 12 1 0.000 l 18 '8AAC EAF 1 4'A 1 h/A I O1 h!A i 01 12 I 0.000 1
;3 '#A::A" N!:r n'A ! h'A i 21 R/A I 21 12 1 0.167 1
- ! WA?> ! ? t'& I h'A i 2i N/A 1 2l 12 1 0.lp i 26 'NA? H 2 h/A I h'A l 01 3'A ! OI 12 1 0.000 1 1l lH.I. R IM!04 2 I h/A n'A i 0I h/41 0l 12 1 0.000 t l' !! CIA 4 PO:R' 2 . k'A I n'A ! OI 4/41 01 12 1 0.000 I h :: CIA 4 PC:n? 3 i R'A : n's 1 0t n!A ! OI 12 ! 0.000 !
?5 :EtaAch!! , t/A : n/A I 3i #'A I 31 12 1 0.250 1 39 lLA:ROSSE N'A : N/A i 0: N/A I OI 12 1 0.000 t j
'3 lL A34.LI 1 f 4/A : k'A I 3I h/A I 31 12 1 0.250 1 19 MA!N! VAvi! 4/A ! N'A : 2: h/A I 21 12 1 0.!M l 9 'C6.llt! ! n'4 ; k'A I O! Nin 1 01 12 1 0.000 1 45 :'!LLf'h[ l I N' A l h/A I 4l A/4 I 41 12 l 0.333 :
IA 'R: LS':4! 2 N'A I N/A I 0I h/A I 01 12 1 0.000 1 63 18:4'!:fLLO I n'A l n!A i 0: h/A 1 01 12 1 0.000 1 20 :hM! Mh! P 14? 1 h!A : 618 ! l1 g/A i 1! 12 1 0.083 1 38 :OfMAMAj i N/A I N/A 1 2l R/A l 2l 12 1 0,10 j h 39 'CP?w AnA 2 I g!A ; geg I gl g'A I 11 12 1 0.013 I I j F :S:S!! 1 N'A I h/A : 0l N/A I OI 12 1 0.000 t i
'O :: MEE 2 I t'A l n'A 1 0I h/A t 01 !! t 0.000 t
- 0 ChEt 3 +
ha: n 's i { l' 1I n!A I 1! 12 1 0.083 l l
- 9 lD'$iE' ORE!f htA l 6/41 4: n;n I 41 12 1 0.333 I I 35 :P A. Sa:E5
- h/A : N/A : 01 h/A 1
" 0l 12 1 0.000 1 I
- PEO 8:?ft# 2 i k'A I n't ! 0i h/A I OI 12 1 0.000 I
'I 3f A> l!'!OP 3 I h/A A'A : n/a 1 3I !! l 0.250 1
)
13 :P LS8M i a'A : N'A I n'A i I 11 12 1 0,003 i s6 M14i illu 1 . h'A ; h/A : 0. N' A 1 03 12 1 0.000 1 t tY I!A> 2 a/A ; h/A I a, n'4 1 0l 12 1 0.000 l 12 > RAM:t!5.8C !: h'A : h:A i ii n't i 11 12 1 0.083 1 it :PRA!'!! I$.AC 2 I k'A 1 N'A l 0: h/A i 01 !! l 0.000 1 54 'EW : ' !$ 1 h/A 1 h/41 01 n!A I 01 12 1 0.0K 1 i! l2J' Ci?:E! 2 4'A : t'A 1 2l #/A I 21
- 12 l 0.10 l
'AO: !!:0 . a/4 : N'A 1 4I h/A ! 41 12 1 0,333 I
'2 SA E' 1 h!A : h/A I 2: N'A 1 21 12 1 0.10 I
' 54.t* nrA 1 n'A l !I 4/A i 1! 12 1 0.083 I t :Sa* S:"I I t'A 1 k'A I 3I wAI 3: 12 l L250 t I
~ '54 009!2 4'A 1 4'A : 0: N/A I Oi 22 :54 DCrgg 3 12 1 0.000 1 h/A I #'A ! OI h/41 01 12 1 0.000 1
' :$! ,,bA* ! I n '4 ' 4/A l 1I h/A 1 )I 12 1 0.083 I
- 8 : TID:VA* 2 htA ! t/4 : Il h/A t ti 12 1 0.093 l
!! :5?. L!:::t 1 . n't I n/A : 0l h/A 1 01 12 1 0.000 i
'! T.**E 8 . k/4 I t'A I 3I g/A I 31 12 1 0.250 l 10 :5s8 A' l I h/A ! h'A I OI h/A I OI il 12 1 0.000 I
- 52H' 2 h/A l h/A i 0i t/A I 01 12 l 0.000 I l' :l;50!
- 441 h>A I h/A 1 0I h/A t 0l 12 1 0.000 1 P :hRE! *:i! IS.,1: k'A i n't : 01 8/A I Ol 12 l 0.000 1 44 17A0'A4 1 t'A : N 'A ! 2i grA I
' 21 12 1 0.10 i L :TWIf PCN 3 N'A I n 'A 1 4I 6/A l Ai 12 1 0,3;} I 51 :f te'D PO:N? 4 . n'a : n'A ' 3: h/A t 31 12 1 0.250 1 71 'V!8"Y VAWE! n't ! 4'A ! 01 5'A i 01 P : WEE O! 12 1 0.0c01 hA: h?A ! OI 4'A I 01 M ;UUh l 12 1 0.000 i NA: hAl 31 t/A t Il 12 1 28 0.250 1 n'A ; k'A
- n/A l 2,O, A. 2 2i 12 1 NO [V[h'l/Cu's I ; og : 1 98 1 012 I 0.10 i144445!,
ra Ait 0;0 pgAnto FADP AE3D DATAs 0.10 1-3
/ i
=xc tow us u n muctsaa neout.voav commission i.at oarauweea u mr ,er r,oc. , v, u, p w,, J n em j
L"o'"38 BIBLIOGRAPHY ATA SHEET j4 su i~staucno~ o~ 1 vios. NUREG-1275 g 2 f eiLE .No SUS 1f rLE 3 Lt.Vi BL.NE 5-OPERATING EXPERIENCE FEEDBA REPORT: New Plants (Commercial Power Reactors) , ,,, , ,,,((,,,,,, ,, s oo~i gl . .a 6 .ui oms > g 3 July [ E D.TE REPORT 155uEO 1987 MoNip Yt.R
'i Jud 1987 7 PtReoHM NG oAG.N:2 ATsoN N.Mt .No MAILING .Qoht $$ f,wypp le 1 8 PRoJECTjfk&KMroRn v~if NuM8(R
\ jf T,. yon ca.~1 ~vo..a Office for Analysis and Evaluation l fOperational Data p U. S. Nuclear Regulatory Commission A F Washington, D. C. 20555 T , [c io 3,0~ ion,~o 0 o ~a.1,0~ ~~ .~o u iu~o ooai ss ,,,,. i. c , n. rvPe or aiP*'
Same as 7., above 5, Technical (
-\ /
4[ ,ene,ooeav m oesw .. o
,, ve.a .~1. ~or , } f 01/01/83-06/30/86
%[.c I
........c,--.o-
,/ i V \
This report documents a detailed review (bf t 'e cause of unplanned events during the I early months of licensed operation for icensed between March 1983 and April 198 5. Themajor.lessonsandcorrectiveactio/ns plantthat' ppear to have the greatest potential for improving the effectiveness of plant startyps are provided for consideration through the operating experience feedback prog 'ms and activities of the industry and the NRC staff. / '; h j Q
/
a j
- i. oocum ~ r .~.< vi,s . . em ti m icm. ions New Plant Ex
/ (\ g i . v,. r.
rience, Unplanned Reportable Events, Reacton Scrams, Engineered fetyFeaturesActuations,TechnicalSpecificakionViolation s, Unlimited Loss of Sy em Safety Function, Significant Events Operatf al , sicumt<cussi ic.no~ Experienc' Feedback
& < ra ,
to iDENTeslEms OPEN E EOTEAMS 9 gg}gggj{jgj tram r.oorrt Unclassified 17 huween on P. Gas M 10 PR ct
*U.S.00vtah*Lhi PR thf inc cr r tct :1987.ig t.69 :50193
&f.-
^
fl .ws g;, s. p, e ;r,p > O:. 97
,s
>> 4 ,. f 3 g' t ;
f '?
. .g' ,,g, ,h. , f,s,c'.w.'s.,.'v p ,V
.s, y x f .. p@. 4. f- '4'
~' ',',mgs.. s
. w.~ f f h' h h> f ,aw L.w n.%r;jy, v ,
j e
- s -,. ;,w r '_, p-" t :,'$
.3
,m. , a -
w n. y 7
.e s- c p ~ ; ;;u; j q , m. .w + mw>
- " m' u%. , m
"~=,.
., s c 2 yn . > ,
~, -
, n.- y. . - + , : 2 ,
i- t a. c .tr a
+
k ,n5.N (C 'y ., _ i ' j,; ) [I. L
,, - ~myt: ", . R &q.
4
; .5 , {.( 4 z..' 4'
, ; -. }; .
g /( ,, g [V/: s ,:A eQ+ ;'it y;
. ' & ' .q' ::! ( h Tt; '+ '
. ?
5 t . g;.. ; hg 4 'b ',. , 4 i er << p '. i * .<& * .->~- - yjg - 4 > eq eQ . .,9,.., -
- ::;)
*tI'; + J_py. . ' _,y
') .
r gy , If G '.
.;: 9 13 F ';
T b, q.:_ ,..,y 3
; ; sa c y
'.} M ' ' *
' '. 0"3 f ' tl , _
'_ , y 7, .
W'f nW . f M . f UNITED'STATESs 3 ,,,@# ~
' ~ ' "
""Y T,_, G
- JNUCLEAR REGULATORY COMMISSIONS < >; a [' "
S$N/EdNNS9 %e V'< M y@ j C JWASHINGTO.NLD;Ci20555 ' 1, '
,G' m
' _?wME4 *
~
e wr , 1 '...p
,- M:A p - . J OFFICIAL BUSINESS : - -e S 1 ; W# M b*' N,. *
( PENALTY, FOR PRIVATE USE,' $300. 3 -'O 3' ' M? " ' "" r m. . r.'
, A , t g ; , *);^
q' . . 1 l t 4 I D g] ;' ,,[ h jE J ! i gs y .) t , e I Y ,j., ,'#- O-, r:> , O!f g. h,.,. , 4
, s I . ,A.,1I
^'t 7> -
p 1 1
<w.-< :
'jp;, ,
sl y. a.3 4 l v ,., , j d a,- ,) 'L, 7 g
.g'r. ;' .
J ,i
,. 1.,u
,.c'S '
r ; d., , f L
> ').Q. ' ' e -. y 'j'7
" ' ' l' { -' (g\
- j. s h. .. t
"<Q:7>
i 3 , a s - ,q, mw" s
,7 j
- Y'k, A-yv-a.>4 i
ra r
.vlg ;
;y_ '. ,
.,r.
x',,, s t. , , s e.ts
.n 4
.r. J.f .t.. .;
3 t'.o.u s i[ t
,.b.,,. ,/ f
,.,y :
3-Er ,, ,
,e,L ,,
s: 1 x', j g q ' b'
.% v 1
;, '-{ > :y,j <. t
- h,. f ,f
, , (
,e g.)...,e v.,
A: r.,:
- e. .
.j 4o' , s
",,, J 3 I
+
! r v., , ,
s ,
;% r r
4
- tr. . t
', 7- I.
y ., g. i I
.k . i n', $ ". 9 1 r v .-->
g g
; k_ - y -E '
4' r l'_ i, F
' )yl .
g. e_ gwp ? Ox:n s:M
-[
/ y4(i*
j;, jg' (-.g g- 1
.). [,.
gi
$g d_ .g'
. 4y*.,
,'j R y ,L' y
'(. yr ; :. ]~ 3.{ - L g,y ,
i ' ' ,P
. , :; e 4
~
1 r .s y2) 7 s
) ii p i
1 ':, ' 1 # I ' # q E I , [.1 e lgf; ( L., .,j , , , ,.%v
.s s. .!
i t
..te ;^2 ei {- f, .
[g+ g i ) f , '
$y I . , 'y n . o i r
' -,J.M:
s A Aq a , I" P s,v
, 4 ' ! ' ,$ 'p" ' .' 1 i
a 5 * .ji. y l .,g 3 'e l
- I
\ ,
1l
;.q c, 4 /* 'J >
i
- 1,%:, <
.y,-
t c s
, ; /-,y ty ._v -e t % yL, I
y r. .,,,OI,- ,
-k ' #
.[ h s
c.n' _.l 4 g
.,J't
t .t
'i f , i fr ,,..
P I
%i g
f y.,' 'l 6 1
-; . g - 4+.
(_m
_' , , 4 g _ . . h),,. . t i
,I i _'.a 4l.
4.. . ,
'n,
- 2 g !, . ,4,
- s ." , ,k5
' E i h q q r
) ,,t . . ,
!L
- _ 7 g
\ <
1 y7 , h, (- }:\ Ir y' /] r j., c .
,, 3 -. g.
, 1 Igg >
..t
. 4' si t >
4 t i: .l. , ; i 5 I y f. i
-C.'
,)
g
, (,': j
- 2. J -
, T 4
EI. \ ' I > i b g..- E
$ ..; .lI -
y* j
~ ;
'l,, ,'t' 1
-{'.,
l,
'j,,\' g'*
1 3 --'d . O,
,e d Q .J i > \
. ,;jh y ;y._' & f ;
y 4 g ( . I ;:( ,'. , [q ; . -
-;L.> -.
. '.' {. ., n ' :p W-F. .. s
! ,,. 'd , l .'<9" I' . g l> l
'4', _,
g \ ,t , 4 Y ' y'r., 1 >
\
-4, ,5
+%-
g ,m. A g i i (p i 3 1 4
,c-L i r f ,
y
; t a. l. \ .1
; t'{ ,,
,E ,Q k
j
- f. C. i '+
1,.,
, ;$ .' '1 -
g. i
+1 Lp , ,
, s s a
,I' b
.,'.( r.
, 34.
.?
?
k ,"ec ' ;t
~3.-,._:[ c. . 4
-: q lu L
'F.M d 4'j' e ' '
i1 ,
- A r s
)
+ ^
xM: n,- u;, w >
': . y ' '- -
+
$_ V@ 4 }g * '
l3f i _..
, jt 4 N;- -
; , 5 &, '}}