ML20056E254

AEOD/T93-01, Primary Sys Integrity,Pressurized Water Reactor Coolant Sys Leaks

ML20056E254
Person / Time
Issue date:	06/30/1993
From:	NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
Shared Package
ML20056E247	List: ML20056E246 ML20056E254
References
AEOD-T93-01, AEOD-T93-1, NUDOCS 9308230072
Download: ML20056E254 (16)

Text

'

O 9

AEOD TECIINICAL REVIEW REPORT UNIT:

Various TR Report No.:

AEOD/T93-01 DOCKET NO.: Various Date:

June 1993 LICENSEE:

Various AEOD CONTACT: J. Kauffman

SUBJECT:

PRIMARY SYSTEM INTEGRITY, PRESSURIZED WATER REACI'OR COOIANT SYSTEM LEAKS

SUMMARY

Reactor coolant system (RCS) leaks at U.S. pressurized water reactors (PWRs) were identified and reviewed for 1980 through May 1992. The events were chosen to identify primary system problems with technical similarities between U.S. and Czech and Slovak nuclear power plants (NPPs). The study excluded steam generato- (SG) tube leaks.

because of design differences between U.S. and Czech and Slovak SGs.

This study was initiated and prepared by Mr. Jan Stuller of the Czechoslovak Atomic Energy Commission (CSKAE) (currently Mr. Stuller is Director of the State Office for Nuclear Safety for the Czech Republic) during an assignment at NRC that resulted from bilateral agreements between NRC and CSKAE. The purpose was to identify issues, problems, or " lessons learned" from U.S. operating experience that might be applicable to Czech and Slovak NPPs.

The study included development of a data file containing information about RCS leaks, categorization of the leaks, and evaluation of RCS leaks, their causes, and corrective measures.

The principal findings of the study were:

Reactor coolant (RC) leaks occurred most frequently in the charging system (20.2 percent), pressurizer and RCS overpressure protection system (19.3 percent),

residual heat removal system (RHRS) (113 percent), other lines connected 'o the RCS loop piping (113 percent), and emergency core cooling systems (ECCs) (9.3 percent).

The most frequent component failures that caused RC leaks were: pressurizer pressure relief valves (PRVs) (10.5 percent), and pressurizer code and spray valves (6.9 percent); charging system piping (10.5 percent) and charging system pumps (7.7 percent); loop piping system valves (8.9 percent); and RHRS valves (6 percent).

The most frequent cause of pressurizer relief valve failures that caused RC leaks was failure of valve internal parts (damaged seats).

930e230072 930615 1

PDR ORG NEXD PDR

O w

Charging system piping failures resulting in leaks were mainly cracks caused by vibrations and other types of cyclic stress. Charging pumps were leaking mainly due to seal failures.

The dominant cause of loop piping system valve leaks was back leakage through ECCS/ loop piping check valves.

The major cause of RHRS valve leaks was seal failure.

ACKNOWLEDGEMENTS The author expresses thanks to the Reactor Operations Analysis Branch staff of AEOD for their help in better understanding certain technical issues during his assignment.

Special thanks to M. George Lanik who supervised this project and Mr. John Kauffman who reviewed and prepared the final version of this report.

The author's assignment at NRC could not have been realized without the exceptional efforts of Mr. Hans Schechter and his colleagues from the Office of International Programs who prepared the assignment and ensured excellent working conditions.

1.

INTRODUCTION RCS integrity at NPPs is one of the foundations of nuclear safety. Each inadvertent and uncontrolled leak of the RC may threaten the ability to remove heat from the core and represents a potential danger of an uncontrolled release of radioactivity, especially for those designs without primary containment.

Ongoing monitoring and evaluation of RCS integrity is essential for continued, safe NPP operation. Similarly, it is important to have a clear understanding of the factors affecting RCS integrity.

Results presented in this report are intended to be helpful in developing a clear understanding of these factors. Detailed information regarding the causes of leaks and the affected components and systems in which the leaks occurred is provided.

2.

METIIODOLOGY AND SEARCII STRATEGIES Details regarding the scope of the study, study methodology, data searches conducted, data analyses and evaluation, data base development, and notes for the result tables are provided in the Appendix.

2

J.

it t.O U is i O

=

.+.1 CATEGORIZATION OF LEAKS HY SYSTENTS AND CON 1PONENTS.

The categorization of all selected events by systems in which the leakage occurred and type of component involved in with the leakage is shown in TABLE 1.

TABLE 1 CATEGORIZATION OF LEAKS BY SYSTEMS AND C051PONENTS 1

COM PONENTS Piping Valves Pzr Pumns Filters Vessels Ht Rx Other Total PERCENT RV Exch Flan SYSTEhtS I

REACTOR 2

1 2

4 9l 23 1 LOOP PIPING 4

23 1

{28:.1 ill'.3 T PRESSURIZER 2

17 26 3

$8l

^$;1'9.3 l-RCP 10 8

1187 17.32 LETDOWN 4

2 7

2 1151 i6ill w

Cil ARGING 26 4

1 19

$0)

(2d.2 NI AKE-UP 1

3

T4 :

! 1.6 -

SEAL 1

tIf

):0!4 3 INJECTION 110RIC ACID 1

5 W

J2)-

IlORON I

di.

FOAD

'^

g INJECTION ECCS 8

II 2

2 I$i f 9'.f CVCS 2

2 2

1

17.L F2.8 i RilR 4

15 4

5 128J M1:39 UNIDENTIFIED 9

1 Il 3 W.'Of TOTAL 64-

?89

~41:

-39l 1-

!L

< 2 :: '

21 di

'$2'483 11001 ^

PERCENT 25.8 35.9 t

' 16.5 '

'15.7 OAL 1.2:

0'.85 L O'.8 T2!8K Ilod!

~

• See the Appendix for notes Pzr RV = Pressurizer Relief Valves, Ht Exch = lieat Exchangers, Rx Han = leactor Hange (in Tables 1-3).

3.2 CATEGORIZ ATION OF LE AKS IlY C0511'ONENTS AND CAUSES The categorintion of events by type of component and leak causes is shown in TABLES 2 and 3.

TAHLE 2 CATEGORIZATION OF LEAKS IlY CONIPONENTS AND C AUSES CON 1PONENTS Piping Vah es Pzr Pumps Filters Vessels lit Rx Other Total PERCENT RV Exch Flan CAUSES CORROSION 1

2 1

2 2

' 8'-

.i3.2. '

EROSION /

I L1?

$0A:

h CAVITATION w

FATIGUE /STR 38 1

4 U3[

^

3.7 3l:

ESS TIIER51 AL 1

2 I3; f.h2 STRESS TIIERSTAL A

AGING VALVE 24 21

4 5.

j l8.1 :

51ECII ANISN1 SEAL FAILURE 3

39 1

33 1

1 1

79.!

f 318 CONTROL 5

8 J 13 :'

5.2

FAILURE OTIIER 5

2 1

1

' 14 :.:

J 54 4-UNIDENTIFIED 15 13 9

1 1

2 1

'421

1. 16.99 TOTAL 64.

89

41 39-

.: 1 -

3

'2'

.2 c L7 '

2481 PERCENT

25.8 35.9'

16.

15.7

'. ' O.4 -

':1.2 0;8 :

' 0.8 E 12.8.

1100?

5 ::

• See the Appendix for notes

TABLE 3 CATEGORIZATION OF LEAKS BY COMFONENTS AND CAUSES COMPONENTS Piping Valves Pzr Pumps Filters Vessels lit Rx Other Total PERCENT RV Exch Flan OF ALL CAUSES EVENTS IIUMAN 3

3 5

1 2

$15/

2hM FACTOR DESIGN 3

1

^ $4!

0133

'm h

DEFICIENCY INADEQUATE 1

1 1

N$pXii dij2 PROCEDURE INADEQUATE 5 1 1 2 gs93 If33) QUALITY ~ TOTAL 11

10:.

7 1 52 'M NN > JidAd PERCENT .: 4.4 t: 4.0 ?! -:32;85 O.4 '- r 0.8 :: ^ J.6? L14;1? - Inadequate quality (as cause) includes incompatibility of welding materials and bad weld; B

3.3 CATEGORIZATION OF PRESSURIZER POWER OPERATED RELIEF VALVE (PORV) AND SAFETY RELIEtt VALVE LEAKS BY CAUSES AND REMEDIAL MEASURES The categorization of pre;surizer PORV and safety relief valve leaks by causes and remedial measures is shown in TABLE 4. (spm = gallon per minute - 3.785 liters per minute) 1 4 CATEGORIZATION OF PRESSURIZER PORY AND SAFETY RELIEF VALVE LEAKS L LE AK.- 'NUM-CAUSES L tNUM1 1 MEASURES: sSUMs RATESL ~ > 1 gpm h < 1 gpm 2 valve mechanism i replaced I unidentified I repaired I ? 24 control failure 5 unidentified I replaced I repaired 3 valve mechanism 13 repaired 5 personnel intervention 3 replaced I design modif. 2 and replaced g unidentified 2 unidentified 6 repaired 2 replaced I unidentified 3 4

J.4 LM & LO U h 8 0t% i $U19 U A L A AMnu G A90 0 & 0 t Litt i a t 1190 14Lt% n0 8) 1 LM UOLO M19p hLitS LU1M14 MEASURES The categorization of charging system piping leaks is shown in TABLE 5. (gpm = gallon per minute - 3.785 liters per minute) TABLE 5 CATEGORIZATION OF CIIARGING SYSTEM PIPING LEAKS

LEAO 5 NOM

iCAOSESi

NUM( m IIiEASURES

!#

INdMi !RATESi ~ > 1 gpm < 1 gpm 4 unidentified 3 repaired 3 fatigue / stress I design modif. I comp. replaced ? 22 fatigue / stress 15 replaced 3 design modif. 8 comp. replaced or repaired y repaired 4 erosion I replaced I thermal stress I design modif. I comp. repaired inadequate quality I repaired I unidentified 4 repaired 2 g replaced 2

J.3 LA l tbOM14 A llON Ut ML5 l'H LddUML 19UUNDAM)

• A L 0 t. L tLA hd The categorization of RCS pressure boundary valve leaks is shown in TABLE 6. (gpm = gallon per minute

= 3.785 liters per minute) TABLE 6 CATEGORIZATION OF LOOP PIPING SYSTEh! VALVE LEAKS

LEAIC NUbf

CAUSES.-

^:NUhti -.!MEASdREST lNUht s RATES- > 1 gpm 9 valve mechanism 3 repaired 2 personnel intervention 1 O seal failure 4 repaired 4 unidentified 2 personnel 1 intervention repaired I < 1 gpm 10 valve mechanism 10 repaired I o: personnel intervention 8 unidentified 1 7 4 corrosion I repaired I valve mechanism 2 design modif. I and repaired unidentified 1 human factor 1 personnel 1 intervention - LOOP PIPING SYSTE51 includes main RC piping and the piping which is connected to main RC piping and can not be isolated by an isolation valve such as ECCS, charging system and RHR connections and includes related RC sample piping, instrumentation piping, and primary drain piping. o

A e 4. FINDINGS From TABLE 1, an overall picture of RCS leaks which occurred at U.S. PWR NPPs from 1980 through May 1992 may be drawn: RC leaks occurred most frequently in the charging system (20.2 percent), pressurizer and RCS overpressure protection system (19.3 percent), RHRS (11.3 percent), other lines connected to the RCS loop piping (11.3 percent), and ECCS (9.3 percent). The most frequent component failures that caused RC leaks were: pressurizer PRVs (10.5 percent), and pressurizer code and spray valves (6.9 percent); charging system piping (10.5 percent) and charging system pumps (7.7 percent); loop piping system valves (8.9 percent); and RHRS valves (6 percent). From Tables 2 through 6, the important findings are: The most frequent cause of PRVs failures that caused RC leaks was the failure of valve internal parts (damaged seats). Charging system piping failures resulting in leaks were mainly cracks caused by vibrations and other types of cyclic stress. Charging pump leaks were mainly due to seal failures. The dominant cause of loop piping system valve leaks was back leakage through ECCS/ loop piping check valves. The major cause of RilRS valve leaks was seal failure. 9

p) a q w APPENDIX STUDY SCOPE AND OBJECTIVES Three main objectives were established for the study:

1. creation of a data file with information about RCS leaks,

2. categorization of RCS leaks,

3. evaluation of RCS leaks, their causes and corrective measures.

Because all NPPs constructed and operated in the Czech Republic and Slovakia are PWRs, the scope of the study was limited to primary system integrity problems with certain technical similarities between U.S. and Czech and Slovak NPPs. As such, searches were conducted of U.S.A. PWR events only. Similarly, primary to secondary leaks through SGs were excluded because of design differences. METIIODOLOGY Initial data searches were performed using the Nuclear Documents System / Advanced Design and the Crosstalk communications program. More than 100 single and combined searches were made using the Subject / Field / Boolean Menu Mode search method in the domain of licensee event reports (LERs). The following words and their combinations were used to determine RCS: primary system, RCS, RC pressure boundary, reactor, pressurizer, reactor coolant pump (RCP), loop, sealing water, piping, valves, ECCS, chemical and volume control system, and charging system. The words chosen to characterize leaks and potential leaks were as follows: leak, fission product release, tightness, flaw, integrity, defect, crack, rupture, fracture, failure, indication, attack, corrosion, shock, erosion, loading, stress, fatigue, toughness, thickness,. brittle, pressure, and level. 'Ihe number of LERs potentially relative to RCS leaks found using the above mentioned search method, exceeded 5000. This number of LERs was too high for the intended evaluation. In addition, a brief review of many of these LERs found that they were not in fact relevant to RCS leaks. The need to narrow these searches using additional selective criteria was evident. Accordingly, searches were conducted using the Sequence Coding and Search System (SCSS) which is a nuclear plant operating experience event database system managed by the Nuclear Operations Analysis Center at the Oak Ridge National 12boratory (ORNL). SCSS provides flexible and efficient search capability. A-1

A e The resulting group of LERs found by ORNL using the SCSS included 329 LERs with event dates from 1980 through the beginning of May 1992. The SCSS search commands used by ORNL were as follows: Used commands: FIND PSYS (AB AD AE AF BD BF BK BL BS BT) < EFF>. BEG. B END LOCNTE TYPE PWR ST PSYS (AB AD AE AF BD BF BK BL BS BT) < EFF>. BEG. B <T> (A M G) < D > (F N) END Meaning of the used commands: Locate all PWR LERs reporting leaks in the control rod drive, reactor vessel, primary coolant, pressurizer, emergency boration, low pressure coolant injection, chemical and volume control (high pressure coolant injection), core flooding accumulator, or upper head injection systems. Restricted to those events identified due to an operational abnormality or alarm condition. SEARCil RESULTS RCS LEAK DATA FILE To assist in the categorization and evaluation of events reported in LERs identified in the search done by ORNL, a dBase data file

• LEAK.dbf" was created. All 329 LERs found by ORNL were reprocessed into this data file. The stmeture of this data file allowed characterization of each event by 26 items.

The structure of " LEAK.dbf" was developed to perform categorization of events potentially relevant to RCS leaks. Categorization of events according to the leak rate, type of failed component which caused the leakage, system in which the leakage occurred, cause of leakage (corrosion, erosion, fatigue, failure of internal valve mechanism, control system failure, human factors, etc.), corrective measures and some others was accomplished. It is possible, if desired, to extend the number of " fields" (selective items) and update the number of " records" (leak events) in accordance with dBase Manual instructions. The items in the " LEAK.dbf" are as follows: A-2

LEAK DATA HASE FILE P HILD DESCRIPTION LNO Number of the event (three digits)

2. DOCKLT Tbc three last digits of the ER Docket Number IER number relative to the leak event.

i

3. ER lts form is-XX-YYY-ZZ. XX indicates the year.YYY t

indicates the ER number and 2Z the number of IER revision

4. ANT'cRION DCS number reistrve to the leak event (10 digits) j

5. UNTT NAME Name of the unit where the Icak event occurred

6. UNTT_ TYPE Design type of the unit. The following abbreviations were use for different types of units-

. PWR xif for Westinghouse Dectric PWR units (the x indicates the number of loops) - PWR CE or PWRCESO for Comb. Engineering PWR units - PWR 117 for Babcock & Wilcox PWR units j 7.NSSS Nuclear Steam System Supplier of the unit- . WEST for Westinghouse Dearic - COMB for Combustion Engineering - DW for Babcock & Wilcox

8. CONTAINMENT Containment type of the unit

. DRYAMD for dry, ambient pressure containment t . DRYSUB for dry, subatmospheric containment - ICECND for wet. ice cond 'scr containment ?

9. OL, ISSUED issue date of Operating license

10. COMM_OP Date of the beginning of the Commercial Operation

11. MWT The marimallicensed energy pcmer of the unit (in Megawatts Thermat)

12. EVENT DATI?

Date of the leak event i 13, LEAK _RATIi Irak rate characteriz.ed by the leakage flow (in GPM) or by the totalleakage volume (in gallons).

14. POWEREVEL Reactor power level (in percent) at the moment of the Icakar,e detection /signahzation (max. three digits)

15. OP, MODE Operational mode of NPP at the moment of the leak detection / signalization (max. two digits) i A-3

~ LEAK DATA BASE FILE r 11 ELD DESCRIFI1ON

16. FAIL,CAUSE Cause of leakage:

-IIUMAN FACIDR - DESIGN DEFICIENCY -INADEQUNIE PROCEDURE -INADUQUATE OUAIJIY CORROSION (component failure due to corrosion) - EROSION (component failure due to erosion) - FAllGUE (component failure due to fatigue) - VIBRA 110N (component failu.c due to vibration) - CYCLIC STRESS -IIIERMAL STRESS - WATER HAMMER - OVERPRESSURE - OVERLOAD - CAVTTAT10N -11tERMAL AGING - VALVE MECllANISM (failure to provide the appropriate RCS integrity due to a failure of internal valve mechanism) - SEAL FAILURE (for valves, pumps, and nanges) - CDNTROL FAILURE (failure to provide the appropriate RCS integrity due to an inadequate function of a valve caused by a control system malfunction) - WEAR 17 NOTES 1 Enntual explanatory notes relative to items 13,14,15 and 16 )

18. COMPOSTLNT Type of component failure which or the incorrect personnel operation or manipulation which caused the leakage:

- PIPE TUBE - PIPE WELD - PIPE FLANGE (for pipe flange joint) VALVE (PRVs not included) - PRESSURE REllEF VALVE - RUFIURE DISK -PUMP - ITLTER - VESSEL (tanks, accumulators and pressurizers) - IIEAT EXCilANGER (heat exchanger, coolers) - REACTOR ITANGE (reactor main nange) - CONTROL ROD DRIVE SEAL IIOUSING - SPARE CONTROL ELEMENT DRIVE MECIIANISM IIOUSING - REACIDR SEAL PIATE - SIGiff GIASS l i A-4 I

g LEAK DATA IIASE FILE g 11 ELD DESCRIPTION

19. COMPNO111 Detailed specification of damaged or incorrectly operated component (determined by item 18) causing the leakage

20. SYSTEM The system which the leating mmponent (determined by items 11 and 21) was part of:

- REACIOR (reactor pressure vessel including vessel body, reactor flange, scactor seal plate, reactor cover, control rod housings and venting and instrumentation lines) - IDOP PIPING (RC main piping including the piping which is connected to main pipelines without a possibility to be isolated by an isolation valve) - PRESSURIZER (pressurizer and overpressure protection system mmponents) - RCP (RCT including related parts of RCP auxiliary system such as scaling aster system and control bleed off line) - RCS SAMPIIS (RCS sample system) - PRIMARY DRAINS -INSTRUMENTATION (instrumentation piping) - PRT (pressurizer relief tank) - 11"II)OWN - ClIARGING - MAKE-UP - SEAL INJECITON - DORIC ACID - BORON INJECIlON - ECCS - SAf1?IY INJECI1ON - IIIGil IIEAD LNJECITON - UPPER IIEAD INJECI1ON - 131S1 (now head safety injection) - CORE ITDODLNG - IIPI (high pressure injection) - 1[ PSI (high pressure safety injection) - LPSI (low pressure safety injection) - ACCUMULATUR - CVCS (chernical and volume control sptem) - RIIR (residual heat removal) - DECAY IIEAT REMOVAL - SDC (shutdown cooling) - CCWS (component cooling water sprem) - RWSP (refueling water storage tar.k)

21. MEASURES Corrective measures undertaken by licensee:

- PERSONNT11NTERVENIlON - CREPLACED (component replaced) - CREPAIRED (component repaited) - PMODII1ED (procedure modified) - DMODIf1ED (design modified) - 15MODIITED (technical specifications modified) - SIMCIIECK (checking of similar components) - LVALUATION (planned or origoing investigatiort, inspection, analysis, evaluation) - RESEALED

22. MEASURNOTE Explanatory notes relative to corrective measures for item 21

23. L GPIATER1 Logical parameter indeating whether the leak v1ts greater / equal

(.T.) or smatter (.F.) than 1 GPM.

24. PIPE _2tNCll logical parameter indicating whether the diameter of the failed equipment was greater / equal (.T.) or smaller than 2-inch %

25. PRESBOUNDR logical parameter indicating whether the leak occurred (.T.) or not (f.) within the RC pressure boundary.

26. AGE Parameter indicating the period between the beginning of commercial operation and the date of Icak event (months)

.A-5

O O i NOTES: -IF THE VALUES FOR LEAK _ RATE, POWERLEVEL OR OP MODE HAD NOT BEEN KNOWN THE FOLLOWING NUMERICAL SIGNS i WERE USED IN APPROPRIATE COLUMNS AND LINES: 99999.999 FOR UNKNOWN LEAK RATE 999 FOR UNKNOWN POWERLEVEL 99 FOR UNKNOWN OP_ MODE - IF ANY OTHER OF THE 26 ITEMS HAD NOT BEEN KNOWN THE QUESTION MARK SIGN ? WAS USED IN APPROPRIATE COLUMNS i AND LINES. I EVALUATION OF LERS IDENTIFIED BY SCSS The evaluation of the SCSS documentation and the corresponding LERs showed that not [ all of the 329 selected LERs were relevant to RCS leak events. 87 LERs did not concern RCS leak events, i.e. electric system failures, overfilling of the pressurizer relief tank with primary make-up water, or reactor trip due to a turbine trip. Such LERs were 1 excluded from further evaluation. 1 Events reported in the remaining 242 LERs were categorized by systems in which the leak occurred, components that caused the leak, causes of failures, remedial measures and leak rate, if provided. Many LERs did not report leak rates, especially in those i cases where Technical Specifications requirements other than leak rate limits were exceeded (e.g. loss of one or two charging pumps, high radiation signals within the containment, high level and dilution of boric acid in safety injection tanks, etc.). Results are provided in the tables in the body of the report. 1 NOTES: (for TABLE 1) - PIPING (as component) includes tubes, tube welded joints and flange joints; - OTHER (as component) includes rupture disk, sight glass and particular reactor vessel and reactor upper head parts such as: control rod drive seal housing, spare j control element drive mechanism housing and reactor seal plate. - REACTOR (as system) includes reactor pressure vessel, reactor flange, reactor t seal plate, reactor cover, control rod housings and venting and instrumentation lines; - LOOP PIPING (as system) includes main RC piping and the piping which is j connected to main RC piping upstream of any isolation valve; - PRESSURIZER (as system) includes pressurizer and overpressure protection system components; A-6

w - ECCS (as system) includes safety injection subsystems and core flooding systems; - RHR (as system) includes RHRS, Decay Heat Removal system and Shut Down Cooling system; - Events indicated as RCP system " PIPING" leaks represent events relative to failures of pipes, welds or fittings in RCP auxiliary systems containing RC such as: sealing water system or RCP bleed off lines; - Events indicated as RCP system PUMP" leaks represent failures of RCP sealing or RCP main flange leaks. - UNIDENTIFIED (as system) represents leak occurrences for which the appropriate system was not specified in the LERs. Some LERs described events with more than one leaking components. If several components were reported leaking, each component was evaluated separately and included in the TABLE 1 as a separate leak event. Therefore, the total number of leaks (248) in TABLE 1 is greater than the total number of evaluated LERs (242). NOTES: (for TABLE 2) - FATIGUE / STRESS (as cause) includes also, water hammer and overpressure; - OTHER (as cause) includes inadequate quality (e.g. chemical incompatibility of welded materials), human factors (e.g insufficiently tightened fittings), inadequate procedure, design deficiency or wear. In some cases, there were several causes for one leak for example as a combination of *OTHER" and contml failum.. For such leaks, only the control railure was indicated as the cause of the leak in the TABLE 2; - UNIDENTIFIED (as cause) includes those leak events for which the causes were not specified in the appropriate LER. 1 A-7 j l}}