ML20116J629
| ML20116J629 | |
| Person / Time | |
|---|---|
| Issue date: | 09/30/1992 |
| From: | Bell L, Oreilly P NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| References | |
| NUREG-1275, NUREG-1275-V07, NUREG-1275-V7, NUDOCS 9211160262 | |
| Download: ML20116J629 (80) | |
Text
e n
a a
a n..
NUREG-1275 Vol. 7 Oaerating Ex;perience Feechac1 Report - Ex;perience wit a Pump Sea _s Insta::ed in Reactor Coo ant Pum;ps Manu ~acturec 3y Byron Jacxson
~
Commercial Power Reactors U.S. Nuclear Regulatory Commission OITice for Analysis and Evaluation of Operational Data
, L G. Bell, P. D. O'Reilly 4,
- sA"n8!!! **
1275 R PDR
l AVAILABILITY NOTICE Availabihty of Reference Materials Cited in NRC Publications Most documents cited in NRC pubhcations will be available from one of the following sources:
1.
The NRC Public Document Room 2120 L Street NW., Lower Level, Washington, DC 20555 2.
The Superintendent of Documents. U.S. Government Printing Office, P.O. Box 37082, Washington, DC 20013 7082 3.
The National Technical information Service, Springfield, VA - 22161 Although the hsting that follows represents the majority of documents cited in NRC publica.
tions, it is not intended to be exhaustive.
P Referenced documents available for inspection and copying for a fee from the NRC Public r
Document Room loclude NRC correspondence and internal NRC memoranda: NRC bulletins, circulars, information notices, inspection and investigation notices; licensee event reports; vendor reports and correspondence: Commission papers: and applicant and licensee docu-ments and correspondence, i
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 proceed-ings, international agreement reports, grant publications, and NRC booklets and brochures.
Also available are regulatory guides. NRC regulations in the Code of Federal Regulations, and Nuclear Regulatory Commission Issuances.
Documents available from the National Technical information Service include NUREG-series 4
reports and technical reports prepared by other Federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to tho Nuclear Regulatory Commission.
Documents available from public and special technical hbraries include all open hterature l
items, such as books, journal adicles, and transactions. Federal Register notices, Federal j
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, i
Single copics of NRC draft reports are available free, to the extent of supply, upon written l-request to the Office of Administration, Distribution and Mail Services Section, U.S. Nuclear Rogulatory Commission, Washington, DC 20555.
Copies of industry codes and standards used in a substar.tiva manner in the NRC regu!atory l
- process are maintained ai the NRC Library, 7920 Norfolk Avenue, Br.thesda, Marylandi for use by the pubhc. Codes and standards are t.'sually copyrighted and may be purchased..
from the originating organization or, if they are Amencan National Standards, from the L
American National Standards institute,1430 Broadway, New York, NY 10018.-
~,
c l
l-
.. = - - - -.
i i
Vol. 7 i:
I Operating Experience Feedback Report - Experience with Pump Seals Installed in Reactor Coolant Pumps Manufactured by Byron Jackson j
4 1
Commercial l'ower Reactors I
Manuscript Completed: July 1992 1
Date Published: September 1992 i
1 L. G. lictl P. D. O'Reilly i
Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulatory Commission Washington, DC 20555 y"'%,,,
p s...0 i
s 6
i AUSTRACT
'lhis report examines the reactor coolant pump (RCP) maintenance of a clean seal coolmg water system. As seal operating experience through August 1990 at plants morc plants have implemented corrective measures such with Ilyton Jackmn (ll J) RCPs. The operuting experi-as these, the number of 11.J RCP seal failures experi-ence examined in this analysis included a review of the enced has tended to decrease, practice of continuing operation with a degraded scal.
'lhe findings and conclusions documented in this report
'Ihis study included a review of the practice of continued are based on the following: (1) a review of failutes re.
operation with a degraded seal in the case of PWR plants ported to the Nuclear Plant Reliability Data System with Ilyron Jackson reactor coolant pumps. Operating (NPRDS) for the period 01/01/84 thiough 01/01/90;(2) experience reinforces the value of the folh> wing factors to site visits: (3) plant RCP scal experience reports: (4)licen.
safely manage operation of a reactor coolant pump witl*
see event reports (IERs); (5) detailed discussions with indications of a degrading scal:
key plant staff at all plants with Il J pumps:(6) reviews of staff NUREGs:(7)a review of Electric Power Research (1) Plant operators should assure that RCP scals are not Institute (EPRI) studies; (8) reviews of operating proce.
tJteady in a degraded state prior to the loss of the dures for ll-J pumps; (9) discussions with key staff and first seal stage.
industry experts involved with the resolution of Generic Issue (GI) 23, " Reactor Coolant Pump Seal Failures,"
(2) "Ihe operators and engineering support staff should and (10) discussions with a laboratory with extensive be familiar with RCP seal operation and the behav-experience in the area of RCP seal reliability.
ioral characteristics of seals under a spectrum of off-normal conditions.
'the operating experience for Il J RCPs supports (1) lim-(3) Plant operators should be provided with clear, un-iting seal cooling transients, and (2) applying high quality ambiguous procedures covering all aspects of not.
standards to reactor pump seals. Maintenance procc-mal and abnormal RCP se:d ope;ation and have dures appear to be of criticalimportance if reliable seal been thoroughly trained in their use.
operation is to be achieved. Plants with Il J RCPs that i = hn.d relatively good experience with their RCP seals (4) Sufficient instrumentation should be provided to le this success to a combination of different fac-allow plant staff to monitor the performance of the acluding: enhanced seal QA efforts, mcdified/new intact stages so that the operators will be able to take su. designs, improved maintenance procedures and appropriate and timely action to prevent grosc leak-trainmg, attention to detail, improved seal operating pro-age in the event that further (e.g., another seal stage cedures, knowledgeable personnel involved in seal main-fails) degradation of the RCP seal occurs.
tenance and operation, reduction m frequency of tran-sients that stress the seals, seal handling and installation (5) Seal operating limits should be prudent and safe equipment designed to the appropriate precision, and limits that are strictly adhered to, iii NUREG-1275 l
]
1 1
3 CONTENTS AllSTilACT....................................................................................
iii EX EC LT11 VE S U M M A R Y........................................................................ vii 1
IN11tODUCf10N 1
1.1 l l a c k t ro u n d...........................................................................
1 2
REACIOR COOLANT PUMP SEAL OPl! RATING EXPERIENCl! FOR ll Y R ON J A C KSO N R CPS..................................................................
2 2.1 Seal Design Configuration For ll-J Pu mps.................................................
5 2.2 S U S eal Ex pe ri e nce....................................................................
5 2.2.1 Plant. Specific Experience with SU 'Iype Scals........................................
9 2.3 M odified Ilyron J ackson S U Seals......................................................... 12 2.3.1 Plant. Specific Experience with Modified SU Seals...................................... 12 i
2.4 Byron J a ckson N-9000 S cals............................................................. 12 2.4.1 Plant. Specific N-9000 Seal 11xperience............................................... 13 2.5 llingh a m.Willam e t t e Seal............................................................... 13 4
2.5.1 Plant Specific Experience With *llic 11Pham Seal..................................... 13 3
S U M M AR Y O F O PER ATIN G EX PER I ENCE................................................. 14 3.1 Tem perat u re Sensitivity of S U Seals...................................................... 15 3.2 Quality Assurance and Maintenance Considerations......................................... 15 3.2.1 Se n sing IJ n e 12ak s................................................................ 16 3.3 Operat ion Wit h Degrad ed Seals.......................................................... 16 3.3.1 Instrumentation and Operator Guidance.............................................. 19 4
S U M M AR Y O F CO N CLU SI O N S............................................................ 20 5
REFERENCES.............................................................................
21 4
FIGURES 1
Schematic Diagram of Dyron Jackson Reactor Coolant Pump 3
2 Byron Jackson Reactor Coolant Pump Seal Cartridge..........................................
4 3
Schematic Diagram of Hyron Jackson Reactor Coolant Pump Seal Stage.............................
6 4
flow Scheme for Four-Stage Byron Jackson Reactor Pump Seal Installed in a Combustion EnE neering-Designed PWR...................................................
7 i
v NUREO-1275
TAllLES 1
PWRs With Ilyron Jackson lleactor Coolant Pumps...........................................
8 2
Advanced Ilyron Jackson N-90(X) Rr'P Scal Failures............................................ 14 3
Component Cmling Water Transient-Initiated SU Seal Failures................................... 16 4
Continued Operation with Failed Seal Stages................................................... 17 i
APPENDICES A
Summary of itcactor Coalant Pump Scal llistories.......
...................................A-1 11 Seal Failure Data Prior to 1985..............................................................11-1 NUlti!G-1275 vi 1
EXECUTIVE
SUMMARY
'lhis re; ort examines the reactor coolant pump (RCP) seal is constructed of seal stages arranged in series and seat operating experience through August 1990 at plants contained in a cartridge. Gross failure of the P,CP scal can with Dyrc,n jackson (Il J) RCPs. !!yron Jackson RCPs are result in a small break loss-of-coolant accident installed in PWR plants with Combustion Engineering (SHl.OCA).1his type of acciden: can occur during not-(CE) and llabcock and Wilcox (ll&W)-designed nuclear mal operation if the leakage past the seals exceeds the steam supply systems (NSSSs)flhe opetating experience normal reactor wolant makeup capacity. Off normal c::amined in this analysis included a review of the practice conditions that can lead to failure of all seal stages simul-of continuing operation with a degraded seal.1hc find-taneously include loss of offsite po ver, station blackout, ings and conc!usions documented in this report are based loss of component cooling water (CCW), and failures that on the following: (1)a review of failures reported to the cause loss of senice water. Ilouver, the operating expe.
Nuclear Plant Reliability Data System (NPRDS) for the rience data for the period considered in this study did not period 01/01/84 through 01/01/90; (2) site visits to the include any accidents and contained only a few events Calvert Cliffs, Waterford 3 and Arkansas Nuclear One involving off-normal conditions (mostly loss of CCW).
(Arkansas 1 and 2) plants; (3) plant RCP seal experience Ilence these types of events became somewhat periph-reports; (4) licensee event reports (LERs); ($) detailed eral to this study.
discussions with key plant staff at all plants with D-J pumps; (6) reviews of staff NUREGs; (7) a review of 1hc 111 RCP seal designs differ, depending on the NSSS Electric Power Research Institute (EPRI) studks; (8) re-vendor, in the number of seal stages and method of seal views of operating procedures for H J pumps; (9) discus-cooling.1here are four distinct seal designs currently in sions with key NRC staff and industry experts involved in use in U-J RCPs: (1) the original B-J "SU,"(2) the modi-the resolution of Generic issue (GI) 23,
- Reactor Coolant ficd ll-J f% (3) the B J N-9000, and (4) the llingham.
Pump Seal Failures."
Willamette replacement seal. Each of these designs uses three seal stages in a stacked arrangement within a car-Two factors provided the initial motivation for this study.
tridge. With one e7ception, the RCP seal design used by First, reactor coolant pump seals have been identified by CE-designed PC ncorporates three active seal stages the industry as one of several ongoing equipment-related and a fourth " vapor stage," and uses an injectionless sys-concerns. Second, since the NRC Staff is currently pre-tem in which primary coolant is circulated through a heat paring a resolution to GI 23, there was a desire to review exchanger to cool the seals. The lone exception, Maine recent operating experience in that regard. Although Yankee, has an injection system along with a fourth stage RCP seat failure experience was analyzed for all pump seal. The three-stage seal design used in B J RCPs at types in support of GI 23, that analysis was restricted to B&W-designed PWRs has an independent supply of high data prior to 1985.1he narrowing of focus to H.J pump pressure water and an injection system for seal cooling.
seals was the result of a preliminary data review that revealed a concentration of reported failures against seals Most of the operating experience with B-J RCPs to for those particular pumps, mid-1990 has been accumulated with SU type seals, and SU seals are still installed in the majority ot B J pumps.
Although there have been no recent gross RCP seal fail.
Early operating experience with the SU type seals can be urcs at plants using H.J RCPs (the 8/88 seal leakage characterized as unreliabb Operating data show that SU event at Arkansas 2 resulted in a leakage rate of 40 ppm-type seals have frequenti, jerienced single seal stage with'in normal makeup capacity.), the data do show that failures. As a result of the relatively high frequency of seal seals used in B J pumps, particularly the original *SU" stage failures, a variety of corrective actions have been design, have a history of unreliable operation (for the and are being pursued by utilities to improve RCP seal period January 1984-March 1990, NPRDS data show 1.6 reliability. Maintenance and quality assurance improve-degraded or total seal failures per pump for B J pumps, as ments have been made, liowever, as with Westinghouse compared to 0.4 failures per pump for Westinghouse RCP seats, the loss of CCW, even for very short periods, is pumps), where failure of a multi stage sealis defined as sufficient to cause seal damage at plants with B-J RCP any event in which at least one stage of the seal had failed seals which do not have sealinjection systems, since a loss
(
or degraded to the point where continued operation was of CCW at these plants amounts to a loss of all seal impacted. It should be noted that a plant with b J RCPs, cooling. Changes in senice water flow, or rapid small unlike a plant with Westinghouse RCPs, is not required to CCW temperature changes, are st.fficient to cause degra-shut down upon incurring an RCP seat stage fattute, dation of the seals. CCW temperature changes are espe-t cially critical for the SU seal design. Testing performed I
1hc RCP seals limit leakage of the high pressure, high with alternative seal designs (e.g., B J Advanced N-9000 temperature reactor coolant to the ret.ctor containment and Hingham RCP seal designs) indicates that they may l
building from the area along the pump shaft. The RCP be less sensitive to CCW temperature changes. Some vii NUREG-1275 i
_- - - - ~_- - - -
- - - ~ _
?
utilitics have used a combination of maintenance im-ity and communicate via leakage, controlled bicedoff provements and reduction in the incidence of CCW tran-(CilO) flow and pump shaft movement /vibmtion; the sients to achieve improved performance with these scals, stages share the same cooling source; all stages accumu-t at least to the point where a change to a different design late service hours (i.e., wear) at roughly the same rate.
does not appear cost-beneficial.
Operating experience rcilects these vulnerabilities, as the historicalincidents of multiple stage failure have been the lhe alternative RCP seal designs for Il-J RCPs, the Ilin.
result of mechanical failures, in particular C110 flow gham-Willamette (specific problems may not be design.
blockage, or the introduction of contaminants into the related)and the 11 J N-9000, have not been troubic free, scal.
but it is difficult to make definitive performance compari-sons with SU seal performance. 3 to the mtet smaller Determining the exact cause or nature of a given scal amount of operating time accun.
'ed by these seal de.
prob!cm (e.g., whether it involves an independent or a signs in the field. At the time im. report was prepared, common cause mechanism) while online is generally not only two plants had operated with N-9000 seals installed possible, so the emphasis shifts to close monitoring and for more than one year. This stut did not consider the prompt reaction by operating personnel in the control updated operating experience provided by the N-9000 room. Continued operation with a degmded seal requires Seal Users Group in their September 1991 responr to Og personnel with knowledge and understanding of seal op-23.
cration and the signs of progressive failure, liowever, in their assessment of a given situation plant staff have Plants with Il J RCPs that have had relatively good expe-consulted with outside experts to supplement their diag-rience with their RCP seals attribute this success to a n stic capability.
combination of factors, including: enhanced seal QA ci-forts, modified /new scal designs, improved maintenance Thus, operating experience reinforces the value of the procedures and training, attention to detail, improved Mlowing factors in order to safely manage operation of a seal operating procedurcs, knowledgeabic personnelin-reactor coolant purnp with indications of a degrading scal:
wived in seat raaintenance, reduction in frequency of transients that stress the scals, seal handling and insta11a.
g pg g gg
,gg tion equipment designed to the appropriate precision, and maintenance of a clean seal cooling water system. As already in a de8raded state Iirior to the loss of the more plants have implemented corrective measures such first seal stage (e.g., consideration of installed in-as these, the number of B-J RCP seal stage failures expc.
service time, any accumulated time seal has been rienced has tended to decrease.
exposed to loss of coohng conditions)..
(2) The operators and engineering support staff should Although operating experience indicates general im.
be familiar with RCP seal operation and the behav.
provement in seal performance over time, operating data ioral characteristics of seats under a spectrum of also show that plants with these pumps frequently oper-off-normal conditions, ate with one or more failed seal stages. Continued opera-tion with a degrading seal was a factor in the May 1980 (3) Plant operators should be provided with clear, un-gross seal failure which occurred at Arkansas 1. Since this ambiguous procedures covering all aspects of nor-event, no other mstances of gross seal failure have oc-mal and abnormal RCP seal operation and have curred at a plant with Il-J RCPs. Ilowever, the operating been thoroughly trained in their use.
experien:e with Il J RCPs contains a number of cases in which a plant's operations staff was challenged to take (4) Sufficient instrumentation should be provided to expeditious action to prevent or mitigate multiple stage allow plant staff to monitor the performance of the failure, after the first signs of seal degradation appeared.
intact stages so that the operators will be able to take appropriate and timely action to prevent gross seal The redundancy of the ll-J RCP seal design is such that, failure in the event that further (e.g., another seal i
l' with a random failure isolated to one seal stage, there are stage fails) degradation of the RCP seal oxurs.
sufficient intact seal stages remaining that the seal cua -
perform its design function satisfactorily, thereby allow-($) Seal operating limits should be prude'nt and safe ing continued operation of the affected RCP, while still limits that are strictly adhered to. The established maintaining a degree cJ seal redundancy. An increased limits should be comparable to the limits recom-wear rate of the remaining stages would be the only ex-mended by the vendor. Any limits that are signifi-pected result,110 wever, as with most redundant equip-cantly different than those recommended by the ment, there are common cause failure vulnerabilitics; in ve.: dor should be internally justified. 'lhese limits -
general, the stages are constructed of the same materials should consider the experiences related in this re.
. and have the same design; the stages are in close proxim; port and the specific RCP seal design installed.
NUREG-1275 viii
i 1
1 INTRODUCTION tional experience. Following a revicy of operational ex-penence, Section 3 summarires the issues raised by the operating experience and additional information on oper-In this report, the reactor coolant pump (RCP) scal oper-ating procedures and instrumentation. A summary of conclusi ns is contained in Section 4.
ating experience through August 1990 at PWR plants with Byron Jackson (B-J) RCPs is examined. 'lhe conclu-sions documented in this report are based on a review of 1.1 llackgr0HHd failures reported to NPRDS for the period 01/01/84 through 01/01/90, site visits to the Calvert Cliffs, Water
- 1he RCP seals limit leakage of the high pressure, high ford 3 and Arkansas Nuclear One plants, plant RCP seal temperature reactor coolant to the reactor containment experience repurts, licensee event reports (LERs), and building from the area along the pump shaft. The RCP detailed discussions with key plant staff at all PWR plants seal is constructed of stages arranged in series and con-with B.J pumps and with a laboratory (AECL) which has tained in a cartridge.'the number of seal stages and cool-extensive experience with the issue of RCP seal reliabil-ing method depend on the NSSS vendor. Gross failure of ity. Relevant staff NUR3Gs, Electric Power Research the RCP seal can result in a sma!! break loss of-coolant Institute (EPRI) studies, and operating procedures for accident (SBLOCA). This type of accident can occur Il J pumps were also reviewed. 'lhe draft Generic Letter during normal operation if the leakage past the seals l
and Regulatory Guide for the resolution of GenericIssue exceeds the r.ormal reactor coolant makeup capacity. Off-j (GI) 23, " Reactor Coolant Pump Seal Failures," were normal conditions that can lead to failure of all seal stages reviewed in detail. 'this analysis included a review of the simultaneously include loss of off-site power, station practice of continuing power operation with a degraded blackout, loss of component cooling water (CCW), and seal, given uncertainties about the propagation of failure failures that cause loss of service water.110 wever, the and the demonstrated sensitivity of the SU type seal still operating experience data for the period considered in in use at a large number of PWR plants with B J RCPs.
this study did not include any accidents and contained only a few events involving off normal conditions (mostly Two factors provided the initd motivation for this study.
loss of CCW). llence these types of events became some-First, reactor coolant pump seals have been identified by what peripheral to this study.
the industry as one of several ongoing equipment relat,:d concerns. Second, since the NRC staff is currently pre
- While General Design Criterion (GDC) 14 requires that j
paring a resolution to GI 23, there was a desire to review the RCS boundary be designed to have an extrernely low recent operating experience in that regard. Although probability of abnormal leakage, RCP seal performance RCP seal failure experience was analyzed for all pump has been undermining this design premise. As periodi-types in support of GI 23 (Ref.1), that analysis was re-cally replaced parts, RCP seals had been considered ex-stricted to data prior to 1985. The narrowing of focus to empt from coverage by the ASME Code, which is used by li J pump seals 5 as the result of a preliminaty data review the NRC staff to interpret GDC 14. In November 1982, that revealed a concentration of reprted failures of seals the staff assigned a "high" priority to the investigt, tion of for those particular pumps. A recent operating event at RCP seals. In October 1983, Generic Issue 23, " Reactor Fort Calhoun which involved continued operation with a Coolant Pump Seal Failures" was authorized and the multi-stage seal in a failing condition, ultimately attrib-study began.
uted to blocked CBO flow. During preparation of this report, another event involving total loss of CBO flow Work in the intervening years by the NRC staff has fo-occurred at San Onofre 2 (which has B-J RCPs equipped cused extensively on the behavior of Westinghouse reae-with Bingham-Willarrette seals).
tor coolant pumps and their seals under blackout condi-tions. While RCP seal failure experience was analyzed for i
On September 27,1991, the same day that the draft of this all pump types in support of 04 23 [Ref.1], that analysis report waa sent out for Peer Review, the Byron Jackson was restricted to data prior to 1985 Advanced N-9000 Seal Users Group submitted a letter to the NRC staff transmitting their response to Generic Operating data for B-J RCPs show that plants with these issue 23. Among other things, the response contained pumps frequently operate with one or rnore failed seal design data for the N-9000 scal, which is the latest B-J stages. A review of operating pocedures governing RCP seal design, and prodded updated operating experience operation with SU type seals indicated that operator with the N-9000 seal. Since the staff's review of the Own-action to shut down the plant and shut down the affected er's Group submittal is still in progress, this report does RCP is not required until failure of at least two stages, not address it.
The existing decision making process for continued plant operation is examined further in this study in Section 3.3.
Section 2 of this report discusses the various seal designs Currently, when a seal begins to degrade, the operations now in use in B-J RCPs, along with the related opera-staff, in consultation with the systems engineer and 1
.,,,7_,y.-
~
,y vw..
possibly the RCP vendor, makes the technical recommen-leakage diversion seal only. The number one seal is the J
dation related to continued plant operation, main seal for the pump, and produces a 2000 psi pr essure i
drop.'the major function of the number two sealis to act
'Ihc recent Fort Calhoun RCP seal failure event of as backup in the event of the failure of the number one 8/15-24/90 illustrates the uncertain nature of continued seal. 'lhe number two seal has full operating pressure operation with a degrading seal. It has to be first noted cap bility. Upon failure of the number one seal, opera.
tors shut the number one seal leak-off control valve, that Fort Calhoun has had an excellent operating history with their ILJ SU type seals. On August 15,1900, the which directs all leakage from the number one seal plant detected anomalous pressure readings or, the seal through the number two seal. Once the seal leak-off is for RCP RC-3A. While monitoring the status of the seal diverted to the number two scal, operators shut down the and consulting with the manufacturer, the plant contin-madq following the normal shutdown procedures. If ucd operation for approximately nine days after it was these is indication of reactor coolant leakage mto the determined that an RCP seal was degrading 'the pump containment, the operators are directed to trip the reac.
manufacturer advised the ticensec that there would be no tor and to use the emergency procedures to brmg the serious problems encountered unless the pressure break-plant to a shutdown condition, More importantly, when down orifice (CilO flow path) became completely the reactor reaches hot shutdown either 14 reactor trip or bhicked, in which case there was a possibility of complete normal shutdown, the RCP with the lenkMg or degraded seal failure. Ilowever, there are no reliable means to seal is stopped.
anticipate when or if the breakdown orifices would be-come completely clogged. 'lhe middle stage failed while grigure 1 shows a schematic diagram of a typical llyron the plant was shutting down.'Ihis evperience shows that Jackson reactor coolant pump. IbJ RCP sca' auemblics seal stages can degrade fast. llut, it should also be noted consist of three or four-stage mechanical stals, depend-that the seal did not fait completely.
ing on the nuclear steam supply system (?.SSS) esign:
plants with CIkdesigr.ed NSSSs have a 'ourth seat or A more recent experience with IbJ RCPs occurred on
" vapor stage," and plants with H&W-designed NSSSs 4/8-10/91 at San Onofre 2 (San Onofrc 2 and 3 currently h, rate with three stages. Where provided, the backup have llingham-Willamette seals installed in their RCPs),
V Nr scal located above t he mechanic:d seals fu nctions to when the plant experienced degraded controlled bicedoff reduce or mitigate liquid or gaseous leakage from escap-(CllO) conditions, and the operator eventually manually ing to the containment. Each of the stages within a scal tripped the reactor and shut down the plant when a total assembly has the same design and is fabricated of the loss of CllO flow from one RCP seal was observed. Al.
- ame rnatetials. Futthermore, each stage is designed to 7 hough this event did not involve RCP seal failure per se,
. (thstand full reactor coolant system pressure should it did ir.volve a degradstion of RCP conditions to the point other stages malfunction. Unlike the Westinghouse seal where the operators were required to take appropriate design that uses one seal to contain reactor coolant sys-remedial action in a timely manner to prevent a possib!c tem pressure, IbJ seal designs use three seals to contain threat to RCP seal integrity, reactor coolant system pressure. Using pressure break-down devices (onfices), Ib3 seals normally operate with E*""*8
- ^
E U"
W"
'Ihe operational response to degrading seals at plants seal.1450-1500 psi on the middle scal, and 700-750 psi on i
with lbJ RCPs differs from that for plants with Westing-u m n sea e operating pressure in the cavity be-I house RCPs. At plants with Westinghouse RCPs, shut-tween tk wp wal and tbay wal is ammximately down commences upon determination that the first stage tad penum @g p$
ua unw mn l
of the seal has failed. 'Ihere are design differences be-
"" ##8 " "" " {##M ""wa cangg atmmngthme wsa gum l
tween the seats used in the different pumps that tend to support a less conservative approach for llyron Jackson pumps because of greater redundancy in the scal. 'these differences are briefly summarized below, llowever, our 2
REACTOR COOLANT PUMP review of operating experience indicates that the vulner-ability to temperature transients of "SU" type seals and SEAL OPERATING EXPERIENCE uncertainties about the nature (i.e., whether a common FOR IWRON JACKSON RCPs cause or a random failure mechanism is involved) and progression of seal failwe reinforce the need for appro-priate precau'ionary measures prior to continued plant operation with a degraded RCP seal.
Since January 1985, there have been no gross RCP seal failures at plants using IbJ RCPs (Ihe 8/88 Arkansas 2 Westinghouse RCP seal assemblics consist of three seals, event resulted in a leakage rate of 40 gpm-within normal two of which are full design pressure seals; the third is a makeup capacity.) See Table B-1 in Appendix 11 for a list NUREG-1275 2
4 2
h 4pp - _ -
I 17 17 7
~
3 7
.)
)Ih
. t r
l
']*
[
ii :
\\
t w
l-h
.l i
3 s
\\
\\
l i-Coupling l l
Spacer-,
f t
4 j
]
[
t
{
i{apor$F
' L lh Bifee I
1 P.
l DN n l
I Cooling 16f Water M*7:1 y
Inlet -- g, f a.
'hh*t f 1; HeatIntegral
' -- Drive C
I Closure E
"j
. 4.9 Exchangert' Mount
~.
A i
K Stud i
G l
j S,7 W
,. /_.
Vj,,,//,,/s', Pump Co
/
gM
/,
W.iiiiiiii
,,)..k.
h" Draln y.p-f- -
,/
4 inW h -
Assembly b
. lo
-3 w A
s.
'g 9-.
Rif d.
.r:
R r static N
Impc!r
/
- D'M':
Bearing Auxli!ny/ /
N}
_ j.i
_N l
_i k.2-
.d, Impe ler -
. ' ~
V.
.,\\
,, J, '
p.
~.
}_
.[
. Case-Figure 1. Schematic Diagram of Hyron Jackson Reactor Coolant Pump 3
NUREO-1275
i
+
l i
1 i
~
i C
r l
(<
ll Q'N s vApon l
>t
.scA L o$
d nr/_
UPPER SEAL n
tg*
/
p 1
/
/k; Y
tl t w-I MIDDLE
/
3 CAL e
/f w,h
?
r.
yer w
Mh
%_ad d
1 1
J 4
i Figure 2.. Ilyron Jackson Reactor Coolant Pump Seal Cartridge NUREG-1275 c
of significant seal failure events. Howaver, n.i from the RCP seats frequently exprienced seal stage failures dur-Nuclear Plant Reliability Data System (NPR DS) indicate ing power operation, aw. ni staging problems on start-that seals used in B-J pumps, particularly the original up after an outage. Tr.
1 identifies the plants with "SU" type design, have a history of unreliable operation, installed ll-J RCPs an,
, cal type currently m ne.
i.e., at least one stage of the multi-stage seal had failed or Waterford 3 and Maine represent special cases.
degraded to,he extent that continued operation was im-pacted.
As described in detail in a subsequent section, the Water.
ford 3 plant began operation with Il-J SU RCP seals These data are consistent with results of a recent industry installed, replaced them with N-9000 seals during analysis of main coolani pump performance over the ten.
1987-88, and then re-installed the SU type after the year period 1980-1990. [Ref. 2) 'lhe results of that analy.
N-9000 seals in all four RCPs failed in November 1988.
sis indicated that, for PWRs, plants with Ilyron Jackson pumps had the highest total percentage (by a factor of approximately two over the next highest contributor) of The Maine Yankee plant originally operated with SU seals, then later installed N-9000 seals. Following the capacity factor losses due to reactor coolant pump prob-fallut af all three N-9000 sealsdue to foreign materialin lems than plants with RCPs manufactured by other ven-the reactor coolant, SU seats were teinstalled until the dors. Problems with pump seals were responsible for dates shown in Table 1, when N-9000 seals were again more than 45% of these losses, installed.
At those units that do not have a seal injection system, a As of September 1991, the distribution of installed RCP loss of CCW constitutes a loss of all cooling to the RCP seals according to seal manufacturer is as follows: 23 SU seals, which could result in a catastrophic failure of the seal.'lhe redundancy of the SU type seal design notwith-seals,4 rnodified SU seals,16 Ilingham-Willamette re-placement seals, and 16 N-9000 seals. Of interest is the standing, operating data have shown that upon loss of fact that all four of the PWRs which have Ilyron Jackson C'~ m RCP seal damage can be incurred very quickly for RCPs with seal injection (3 Ilabcock and Wilcox-designed SU type seals. 'lhe data also show that even relatively PWRs, and one Combustion Engineering designed small temperature increases, such as those experienced during normal operation, can damage the seals. The ad-PWR) installed the B J advanced N-9000 seal in incir vanced seal designs for ll J RCPr lave been designed to RCPs during 1988-89. Further discussion of the N-9000 reduce these sensitivities.
seat may 'oc found in Section 2.4.
The following discussions briefly describe some of the 2.1 Seal Des.ign Configurat. ton For IbJ differences among the seal designs and summarize plant-PunipS dy. plant operating experience with each type. The discus.
4, sion of experience is fairly complete for those plants vit.
ited during this study (i.e., Arkansas Nuclear One, There are four distinct seal designs currently in use in B-J Calvert Cliffs, and Waterford 3). NPRDS data war used RCPs: (1) the original B J "SU" scal,(2) the modified B-J as the basis for discussions with the remaining plants; SU seal, (3) the B-J N-9090 seal, and (4) the Bingham-however, hardcopy data obtained during plant visits are Willamette replacement seal. Each of these designs uses considered more complete than information obtained three mechanical seal stages in a stacked arrangement through telephone contacts.
within a cartridge. As mentioned above, the B-J RCP seals installed at CE-designed PWRs incorporate a fourth in discussing the operating experience, the term failure is
" vapor" 8tage; the Bingham Willamette replacement seals installed at CE-designed PWR plants also have a applied to any event where at least one stage of the fourth " vapor" stage, mul;i stage seal had failed or degraded to the point where continued operation was impacted.
A schemati diagram of a typical B-J seal stage arrange-ment is shown in Figure 3. Figure 4 shows the basic flow 2.2 SU Seal Experience scheme for a four-stage B-J RCP seal installed in a CE-designed PWR Appendix A contains flow schemes for other B-J RCP seal configurations, including instrumen-The SU seal design is the original seal installed in B J tation locations.
RCPs. The seal assembly is a self-contained package 2.n-chding seal faces (rings), back-up rings, clastomers, shaf t sleeve, springs, pressure breakdown orifices, and other Site visits and plant surveys revealed a diversity of ap-ancillary parts. [Ref. 4] As noted previously, the seal-proaches to address unsatisfactory operating experience nrrangement and cooling method in a given RCP depend with the original B-J SU type seal design, Plants with SU on the NSSS design.
Byron Jackson Seal Arrangement FUNCTIONAL SCHEMATIC SWT g-7
^
0RDCS BACX tP RDC
/(
s
\\
\\ \\-a"*'
- B" OI4(TER
'A' OIA)CTTA
//
70TATING
/
FACC RING p.0CX RDC-t SuFT ROTATING FACE SLEEW m c y#
g
/
w E
R AI.
p, F
Y mm-U C,9$
///A^"~
/
4
-con. $PRM
'NOT1h The stationary face and backup ring in the llyron Jackson seals installed in the -
i reactor coolant pumps at the Fort Calhoun plant do not contain tne drilled holes shown in this diagram.
Figure 3. Schematic Diagram ofIlyron Jackson Reactor Coolant Pump Seal Stage -
j NURiiG-1275 6
LEMAGE N
4-f rtoW
[
SEAL ruar 4 3 CrH
$ Hart SHArt b ~ ~*[
$1EEEcrr s9') \\
l l mf= %
so PSI L
i I -
LOW PPSSSURE-1 e
J p 3RD STACE
/( u lh o
vso est
} ?m
,~
i l
2ND STACE
[
INTEPMEDIATE SEAL
'f
/
S RE DEVICE
/
p
/
isso est 2d / 7 o
}
_=1 we
, STATIONARY IST STACE j,_,,
'fh HICH PRESSURE SEAL W.
s 90 tai 1Y*-
\\ U E#
5"~ 8'"
- y/
y
(
SucoNDARY also est I
NPI 1 CYM
!w~
RUMP END Figure 4. '"e# dheme for Four-Stage Byron Jackson Reactor Pump Aal Installed in a Combustion Engineering Detigned PWR.
7 NUREG-1275 :
Table 1 PWRs With Byron Jackson Reactor Coolant Pump Plant NSSS Current Seal Type Pump Seal Design Arkansas 1 BAV N-9000 (4 Pumps 11/88)
Injected 3 Stage Davis-Besse B/W N-9000 (1 Pump 6/88)
Injected 3 Stage N. 9000 (3 Pumps 1/90j C a ;ral River 3 BAV N-9000 (4 Putnps 6/90)
Injected 3 Stage Arkansas 2 CE SU Injec.ionless 4 Stage Calvert Cliffs 1 CE llingham-Willamette injectionless 4 Stage (1 Pump 11/89,3 i. mps 3/90)
Calvert Cliffs 2 CE Bingham-Willamette injectionles-4 Stage (4 Pumps 6/90)
Fort Calhoun CE SU Injectionless 4 Stage Maine Yankee Cli N-9000 Injected 4 Stage (1 Pump 11/89,2 Pumps 4/90)
Millstone 2 CE N-9000 (1 Pump 10/89)
Injectionless 4 Stage SU (3 Pumps)
Palisades CE Modified SU (1983)
Injectionless 4 Stage San Onofre 2 CE Bingham-Willamette injectionless 4 Stage (Early 1986)
San Onofre 3 CE Bingham-Willamette Injectionless 4 Stage (10/86)
St. Lucie 1 CE SU Injectionless 4 Stage St. Lucie 2 CE SU Injectionless 4 Stape Waterfora 3 CE SU Injectionless 4 Stage n.
Prict to 1986, the SU seal was the standard seal design flow (controlled bleedoff) line closure / blockage, (7) low installed in all Il-J RCPs. The original SU seal design has pressure starts or excessive differential pressure across a history of generic problems involving high friction char-seals, and (8) improper venting.
act istics of the f.econdary seals (elastomer), and quality cotacol of parts.1 The same failure causes identified in Reference 4 were I!PRI-2965 [Ref.4] documents the results of a shaft seal also identified in the subject study.The review of opernt-operating survey which yielded the following causes of ing data and plant discussions indicated that poor seal seal failures: (1) water contaminant such as abrasive operating experience can be attributed to operating and debris, or " crud," (2) pump transients, such as trips or maintenance errors by plant staffs, inherent SU seal de-othe emergency events, (3) loss of injection flow, (4) sign characteristics, deficient quality assurance measures, temperature and/or pressure transients,(5) loss of cool.
lack of in depth understanding of seal characteristics on ing water and/orI ce temperature excursions,(6)r-ging the part of utility staffs, and inattention to details or lack of in-oepth knowledge about the failure mechanisms and
'Dased on information from " Agenda of Mee'ing between Re resenta.
installation requirements on the part of plant engineering tives of the Nuclear Induistry and the NRC," on October 2,1988.
and maintenance staffs.
1
2.2.1 Plantepecific Experience with SU Type coolant could icak into the containment.ne licensec ielt Seals assured, after discussion with Hyron Jackson that, if the lower seal breakdown device were to become completely This section summarizes operating experience with SU blocked, the plant had ample time to safely shut down.
seals, including experience at chose plants which eventu-ally replaced their SU seals with a different seal type. Fort At approximately 4:00 p.m. on August 24, the plant man-Calhoun, Arkansas 2, Waterford 3. Millstone 2, and St.
ager made the decision to shut down the plant before the Lucie 1 and 2 continue to operate with SU seals installed lower seal breakdown became completely blocked while in thcir RCPs. Summaries of scal-related experience for shutting down damage to the remaming seals occurred.2 Waterford 3, St. Lucie 1 and 2, and Arkansas 1 and 2, as At 4:19 p.m., the licensee ommenced a controlled shut-transcribed from plant records and plant rep vts, are down from 100 percent power. At approximately 600 provided in Appendix A. Reference 2 contains a..imraary p.m., plant instrumentation (pressure readings) indicated of early seal failure experience (prior to 1985) for plants that the middle seal had failed. He system engineer was not visited during this study.
concerned that the on-going pressn i tillations would damage the rerrining seals. The t was made to Detailed discussion of three particular seal failures is reduce power tc. er in order to re m I where RCP 4
provided in the discussions for Fort Calhoun, Arkansas 2 RC-3A could be hutdown.He RW - secured at 7:44 and Calvert Cliffs 2. Rese events were highlighted be-p.m., and the system engineer detecued that the re-cause they either illustrate the typical decision-making maining seal stages were no longer in jeopardy.
process when seal failures occur, or show particular fail-ure modes.
The pump seal cartridge was taken apart on September 5, 1990, under the direct observation of the system engi.
Fort Calhoun neer. There was no obstruction found in the inlet of the pressure breakdown ortfice, but some metal balls were Until a seal failure which occurred in 8/90, Fort Calhoun found at the outlet.The system engineer suspected these had not experienced an RCP seal failure since 1978. At were the lock wires that had been forced through the the Fort Calhoun' plant, all four seals are changed out breakdown orifice, which is only about 1/10-inch in di.
every fuel cycle, and a dedicated individual is responsible ameter, Since this event, the Fort Calhoun licensee has for seal rr intenance. The staff at Fort Calhoun attrib-developed an improved method for safety wiring the first-uted their larBely successful seal operation to the follow-stage pressure breakdown device capscrews. The hcensee ing: (1) dedicated individual responsible for seat mainte-believes that this new wiring technique will ensure that nance, (2) attention to detail by personnel performing the lockwires do not work loose.This new method will be maintenance, (3) procedui es used to operate the seals, (4) used when all four RCP seals at Fort Calhoun are re-the low number of systems transients that tend to stress placed during the 1992 refueling outage.
the ceals, and (5) relatively small RCP shaft diameter resuus in a low seal surface velocity. Thus, the seal "PV" Arkansas Nuclear One, Unit 1 is relatively low. However, the recent v 9 failure event in August 1990 illustrates the potential fa RCP seal failure Arkansas 1 is a PWR with a B&W-designed NSSS using a even at a plant with an excellent sral operating record. It three-stage, seal injected RCP. He plant replaced the also illustrates the uncertainties associated with operat.
SU RCP seal cartridges 81 times between 12/74 and ing with a seal stage in a degraded state. The followir.g
.11/88, at which time the B-J N-9000 seals were installed.
description is adapted from NRC Inspection Report No.
He seal cartridges were replaced following a variety of 50-285/90-35, dated September 28,1990.
seal-related and seal sensing line failures (See Appendix A). A review of operating data indicates that failures of On August 15,1990, the pressure measured between the RCP seal sensing lines were a recurring problem at Ar-lower RCP seal and the middle seal on RCP RC-3A kansas 1.
began to decrease. The cognizant system engineer sus-pected a partial blockage of the pressure breakdown de-The gross RCP failure event which occurred at Arkansas vice betwe'en the lower and middle seals. On August 20, 1 in 1980 illustrates the difficulties mat a plant operations.
the system engineer contacted the pump manufacturer, staff can encounter when attemptmg to anticipate RCP '
Byron Jackson, for assistance. The licensee was informed seat behavior while operating with a seal in a degrading that, if the lower seal stage were to catastrophically fail, condition. On May 10,1980, while the plant was at ap-the middle and upper seals, as well as the vapor barrier, pr ximately 86% power, the SU RCP seal installed in were designed to withstand full reactor coolant system RCP "C" expertenced a gross seal failure that resulted m pressure. He manufacturer stated that the main concern a peak leakage rate of 200 300 gpm. Several differing would be if the pressure breakdown device were to be-detailed discussions of the progression and mechanisms -
come completely blocked. Dis could cause damage to the
%im immoun RCPicahare remaining seals and then, if the lower seal failed, prunary die, upper, and vapor in order moving away froINed lower, mid-cs boundary.
9 NUREG-1275 -
~
d k
i i
j l
1 of the seal stage fadures are available. [Ref. 4-page 44, seal failures, manually tripped the reactor and stopped and Appendix E, page E-2; Refs. 5,6] Ilowever, there RCP 2P32A.
p are three signific:mt points about the event which are relevant to this study; (1) the RCP sesi had been sub.
The time that clapsed from the sensing line failure to jected to at least two significant transients prior to the failure of the middle seal stage was approximately 25 minutes. After the mr :al shutdown of the RCP, the May 10 event-a loss of all a.c. power during which leak, age from the 3rd seal stage doubled, and an intentional leakage rate increased u 40 gpm over the next hour.The j ~
loss of seal injection test (non-nuclear instrumentation increased leak rate after securing the pump was attrib-r l
checkout), (2) once RCP seal degradation had been iden.
uted to continued degradation of the seat components, l
tified on May 10, the plant continued to operate with a resulting in leakage past the vapor seal into the contain-j degrading seal and an increasing leak rate for about one ment building. flecause of experience gained during pre-hour after failure discovery, and (3) once failure of one vious sensing line failures at Arkansas 2, the operating j
seal stage occurred, failure of the remaining seal stages staff was able to quickly diagnose the problem. In cattier occurred very rapidly, sensing line failures, sufficient time was available for per-sonncI to enter the containment building and isolate the leak before any significant seal damage (approximately 1 Arkansas Nuclear One, Unil 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> minimum needed to make a containment building entry) occurred. In this event, the sensing line completely severed, and seal degradation proceeded more quickly.
Arkansas 2 is a CE-designed PWR with a four stage, injectionless RCP (see Appendix A for seal failure experi-his event highlights the potentialimpact of sensing line ence). Although the licensee, Arkansas Power M Light failures.
l Company, is one of the utilities involved in u.
selop-ment of the Il-J N-9000 seal, as of late 1990, there were Calvert Cliffs j
no plans to replace the SU type seals. The utility is con-tinuing to evaluate industry seal operating results, and to Prc.1985 SU type seal data for Calvert Cliffs for the assess the impact of possible upcoming RCP seat require-period 805 through 3/86 are included in Appendix A.
4 ments. Arkansas 2 replaced RCP seal cartridges 48 times (Several additional failures occurred during the period from 12n8 through 11/89 for a variety of seal and sensing from 1986 through 1989.) The two units experienced j.
line failures. In 8/88, the failure of a seal sensing line and numerous seal failures. The SU type seal experience at
{
a related seal stage failure caused a 40 gpm leak. The the plant was such that the utility made a decision to refit licensee attributed 20 gpm of this leakage to the actual their pumps with Bingham-Willamette I ' seals. Al-
}
sensing line leak. The following discussion is adapted though Calvert Cliffs has replaced their St; als with from LER 50-368/88-011, which is included in Appendix Bingham Willamette seals, their SU experience illus-1 A.
trates of the sensitivity of SU seals to temperature.The following discussion is derived from IIR 50-318/85-001.
l I
In August 1988,' Arkansas 2 experienced a sensing line On the morning of April 23, 1985, the Service Water failure that ultimately resulted in a estimated RCP leak (SRW) heat exchanger was removed from service for j
rate of 40 ppm. The sensing line failure initiated an un'
. cleaning of the inlet sea water side tubes. Flow variations usual event and a manual scram of the reactor. The sens-occurred when the system was returned to service. The ing line failed at the junction of a 3/8-inch instrumentline salt water (SW) system cools the SRW, and when an SRW i
and a 3/4-inch piping socket weld. The leak rate at the heat exchanger is removed from service, temperatures 1
sensing line was estimated to be 20 gpm. Failure of the and flows change in the SW cooling system. The compo =
j sensing line initiated RCP seal degradation. Conditions nent cooling water (CCW) system is also cooled by the ;
deteriorated very quickly. Within 10 minutes of the sens-
- SW system.Therefore, changes in th_e SRW line-up affect
{
ing line failure, an alarm was received in the control room the SW flow,'which ta turn affects'the CCW heat ex.
that indicated low controlled bleed-off (CBO) flow.
changers, which in turn cause CCW ;cmperatures to
~
Based on observations of the scaPs parameters, it was change. _
evident that the integrity of the RCP seal was degrading i
with time. Following the sensing line rupture, measured The line-up of the SRW on April 23, 1985 caused a seal parameters became extremely erratic. Flow and tem-temperature transierit of approximately 20 degrees F to-i i
peratura changed in various areas of the seal assembly, be imposed on the CCW heat exchanger. The control
- The ser.l' degraded over the next several minutes due to room opemtors (CROs), monitoring the performance of the loss of normal cooling to the seal stages. Degradation the 21A RCP seal during regulation of SW flow, noted continued until eventually tha last two seal stages (upper large pressure oscillations between the seal stages of the and vapor stages) were indicating RCS pressure. At this seal for RCP 21A. The pump's lower seal temperature point, the operators, it. 2ccordance with guidance con-began to rise at a rate of about three degrees per hour 4
tained in the abnortnal mrating procedures for RCP from a value of approximately 125 degrees F. Pressure 8
NUREG-1275 10-
J-eJ--
4
-s4 w
i oscillations varied from 300-500 psi in Inth the upper and March 1988. The new model N-9000 seal failed during e
middle seal cavities. The seals in the remaining three plant start up in November 1988, and was replaced with RCPs were unaffected.
SU type seals in all four pumps. As noted in Table 1, the i
plant continucs to operate with the SU type seals. The l
Hy early morning of April 24, the lower seal temperature plant incurred several additional failures during 1989, leveled off at approximately 155 degrees F. However, the after re-installing the SU seals, pressure oscillations continued. At 9:25 a.m. on April 25, 1985, CBO lii/Im Flow alarms began to alarm. This sug-St. Lucie gested that CBO had begun to increase markedly, indicat-ing further degradation of the seal stages.nc midJle and There were 42 seal failures expedenced at St. Lucie 1 and 4
i upper cavity pressures increased to 2200 Ir' and the 2 prior to the licensee initiating a study inta the root cause j
excess flow check valve on the CBO line shut due to of seal failures, ne failures at the site were determined 2
cacesive flow. Operators manually tiipped the reactor at to be primarily due to component cooling water (CCW) i 0933 on April 25,1985 because of a rapid degradation of transients. The plant initiated a Plant Change Modifica-the 21 A RCP seal.
- tion that modified the closing logic of the CCW isolation valves.The logic was changed so that the valves would fail 1
This event illustrates the sensitivity to temperature ef-open versus failing in the closed position. Since making fccts, and the rapidity with which cascading seal stage the valve modifications, the plant has not experienced 4
failures can occur. The small (20 degrees F) temperature additional CCW transients.
transient caused a temperature-induced failure of three RCPseal stages.The unpredictable effect of temperature Repeated third stage seal failures at the St. Lucie units on the RCP seal was one of the considerations that the were determined to be initiated by improper seal venting utility factored into their decision to replace their B J procwJures. Since the venting procedures were revised, seals with Bingham-Willamette seals.
failures initiated by this mechanism have declined. Main-1 tenance-related failures were also found to be significant, j
Component cooling water is especially critical in injec-These failures were found to t.c the result of testing and j
tionless RCPs, because CCW is the onh source of seal installation crrors by maintenance personnel. Post-instal-cooling available, and seal damage occurs very quickly lation testing has been climinated and maintenance per.
without cooling. For SU seals, the manufacturer's guide-sonnel have been given additional training to help climi-lines 3 set a maximum operating temperature for CBO at nate installation-related errors.
180 degrees. In a 30-minute loss of component cooling water test highlighted in NUREG/CR-4544 [Ref. 5]. all It was learned from discussions with the licensee that four scal stages exceeded 250 degrees F at 10 minutes into -
there are currently no plans to change from the SU type the test. Since a rapid temperature rise would occur with RCP seal design being used. St. Lucie has had SU type 4
any RCP seal design upon total loss of seal cooling, these seals installed in a RCP for as long as 65 months without results apply to other RCP seal designs as well.
failure (time between failure is measured in calendar hours). De licensee expects RCP seals to last for at least I
Waterford 3 48 months, and would expect the SU seal to be able to last for two fuel cycles (based on a 24-month fuel cycle).
Waterford 3 experienced 23 failures of the SU type RCP seals between the plant's initial startup and the end of Davis-Desse -
1988 (see Appendix A).The SU seals were exhibiting an average lifetime of approximately 4 months. In late 1986, The Davis Besse plant initially started up with SU type Waterford 3 hired Atomic Energy of Canada, Ltd.
seals installed in the RCPs. On the average, these seals (AECL) as a consultant to assist them in defining their experienced a failure about every three months. Because RCP seal problems. AECL identified the following seal-of repeated seal problems, the plant switched to the related problems for Waterford 3: (1) the B-J SU seals modified SU seal. He plant's operating experience with are prone to failure because of component tolerances, (2) the modified SU seals improved to the point where the problems were identified with part specifications [i.e.,
seals would last approximately a full fuel cycle. The plant parts quality assurance, packing, and storage (the carbon nvitched to the N-9000 design, beginning with one pump seal faces were found to change tolerances when stored in June 1988, and installed N-9000 seals in the remaining dry)}, and (3) maintenance-related issues were identified.
three pumps during the refueling outage that began in The plant installed a B-J N-9000 seal in one pump in January 1990.
October 1987, and then in the remaining three pumps in Millstone 2 8" Reactor Coolant Pump Operatin Guidelines: Dyron Jackson Four d
Stage Mechanical seal Canridge Design-Cumbustion Enginetring The Millstone 2 plant also originally operated all of their
$" $ 6;s g n g e NuEn*[ " ""
R cps with SU type seals. Currently, the plant is operating s te n
11 NUREG-1275
with SU type seals in three of the four RCPs. The fourth first installed in 1983. Il-J suggested that the seal be run pump has an N-9000 seal installed to test iM operation in only for one cycle (approximately 12 months). Palisades an injectionless, four stage seal. (A discussion of N-9000 found that the seal lasted up to 20 months, or anproxi-seat experience is contained in Section 2.4.) Prior to 1988, mately two cycles. Four seal failures were experieticed by the plant experienced a number of seat failures; however, Palisades during the period 09/84 through 2/86. Dtscus-these problems were not as serious as those experienced sions with plant staff confirmed that, upon failure of one at other plants. In 1985, an incident occurred at Millstone seal stage, the plant contine to operate until it is con-2 in which component cooling water was isolated from two venient to shut down for repair. An increased level of of the P. cps for several hours. According to the licensee, att entica is given the seal if it is approaching 20 months of the RCP seals exhibited some signs of degradation. Ilow-operating service, ever, the seals were kit in place and the plant operated for several months until the next refueling outage, at The engineering staff at Palisades continues to follow the which time the seals were replaced.
progress of the advanced RCP seal cifort. They have evaluated the B J N-9000 seal, the AECL seal, and the According to the licensee, the better-than-average expe-llingham-Willamette seal, and stated that they will wait rience with SU seals at Millstone 2 from 1988 to date can and see if the advanced seals can meet all the technical be attributed to the following: (1) maintenance personnel requirements associated with station blackout and other have been provided with better training and working con-safety-related issues. Furthermore, they expressed some ditions, (2) maintenance tasks were simplified with the concern about the very limited operating experience asso-upgrading of the hoist used to remove the seal package ciated with all of the advanced seal designs.
(the upgraded hoist allows for better work area vision during seal replacement),(3) improved maintenance pro.
As noted in Section 2.2.1, Davis.Besse reported using the cedures and work procedures hay ( Seen developed, and modified SU type briefly, and the RCP seal life increased (4) an optical flat table has been acquired and installed to to approximately one cycle with the modified seal design.
assure seat face flatness. As a result of a 1985 study of RCP seal experience, the licensee upgraded the following 2*4 11 ron Jackson N-9000 Seals Y
mstrumentation: (1) inlet and outlet CCW temperatures.
(2) Cl30 flow indication, and (3) inlet seal pressure.
The ll-J advanced seal is.aulti-stage seal design, de-signed for three-fuel cycle reliability (50,000 hrs.). Ac-2.3 Modified Ilyron Jackson SU ScalS cording to B-J, the N-9000 seal is an analytically ptedict-able seal whose normal and abnormal operating characteristics are well understood. Furthermore, B-J The modified B J SU seal consists of the SU seal modified st tes that the N-9000 seal's normal operating reliability as follows: (1) slots have been added to the rotating face dunng plant transients, including station blackout, has to increase hydromatic lift at low pressure, (2) the shape been increased signihcantly, of the U-cup has been modified to decrease U-cup secon-Since 1983, B-J, B&W, and a number of utilities 5 whose dary seal friction, (3) the rhape of the rotating face body pl nts have B J pumps installed have been devoting most has been modified to accommodate the modified U-cup, of their efforts to developing the advanced N-9000 seal.
(4) seal sleeve finish at the U-cup has been polished to On October 25,1988, an owners group / vendors group decrease friction and wear of the U-cup and sleeve, (5) briefed the NRC staff to mform them about progress on the staging coil flow rate has been increased from 1 to 1.5 the development of the N-9000 seal. During this briefing, gpm to improve the stability of the seal cavity pressures, the mdustry mformed the staff of actions it was taking (6) the heat exchanger flow path has been modified to mdependently to address coacerns about normal opera-provide a temperature environment consistent with bet-tion of RCPs and abnormal transients, mcluding Genene ter SU seal operating characteristics, and (7) the shaft Issues concerns. At the time this report was prepared, sleeve has been vented to prevent internal pressurization experience with the N-9000 seal was hmited. In addition, of the seal sleeve.4 the amount of available detailed technical information about the N-9000 seat was insufficient for the NRC staff 2.3.1 Plant Spec.fic E,xprience with Modified to assess the reliability of the N-9000 seal design. Only i
SU Seals two plants, Arkansas 1 and Maine Yankee, had operated with N-9000 seals installed in all RCPs for more than Currently only the Palisades plant is operating with the several months. Operating experie,ce at one plant modified SU type RCP seat instalMd. nese seats were (Waterford 3) during startup included an event which
'ne utilities involved in the development and funding of the N-9000
- Based on information frorn " Agenda of Meeting between Representa-type n-J seats are nxida Power Corporation. Maine Yankee, Toledo trves of the Nuclear Industry and the NRC," on October 25.1988.
f&on, and Arkansas Power & tight.
I NUREG-1275 12
involved the premature failure of the N-9000 seals in-Instrumentation for the Bingham replacement seals is stalled in all four RCPs.
effectively the same as that used when the B J SU type seals were installed; i.e., middle and upper seal cavity 2.4.1 Plant Specific N-9000 Seal Experience pressures, lower seal cavity temperature, controlled bleed-off (CBO) flow, CBO temperature, and volume As identified previously in Table 1, the following plants control tank (VCT) back pressure.
currently have advanced N-9000 seals installed: Arkan-sas 1, Davis-Besse, Crystal River 3, Maine Yankee, and 2,5.1 Plant Specific Experience With The Millstone 2.
Bingham Seal Table 2 lists the N-9000 seal failures that ' ve occurred Operating experience with the B-J SU seal prompted to date.
Southern California Edison (SCE) to remove these seals from the San Onofre 2 and 3 RCPs in 1986, and replace The failures of the N-9000 seals at Maine Yankee and at them with Bingham-Willamette seals.The installed Bin-Waterford 3 were in part due to the presence of debris /
gham seals have on occasion experienced staging difficul.
crud in the RCS, ne N-9000 seal design is based on the ties during plant startup, but have had none of the gross premise that the seal will be operated in a clean system.
and unexpected failures experienced with the B-J SU Currently Maine Yankee is operating with N-9000 seals seals originally installed in the B J pumps. The root cause in all three of their RCPs. The plant reinstalled the for the staging problems is still under investigation, but N-9000 seals in pumps 1 and 3 in 11/89, and in pump 2 in indications are that it is not related to the seal design 4SO. As noted above, the experience with the N-9000
- itself, seals at Maine Yankee since their reinstallation has been quite good. Only one injectionless RCP has an N-9000 He O-ring on the lower shaft sleeve has been a problem seal installed (Millstone 2). De version of the N-9000 source at the San Onofre plant. The plant is currently in seal cartridge operating at Millstone 2 has a modified slot the process of making a design change to the lowershaft and I-rir.g which make it more debris resistant.
sleeve to prevent lower shaft seal fretting. Miscellaneous problems have been encountered with the seals; however, The N-9000 seals were installed in all four RCPs at Ar-kansas 1 during refueling outage number 8 (8/88-11/88).
none of the problems have caused fail _;es of severity sufficient to shut down the units.
Problems were encountered with third stage pressure on two of the four pumps following installation of the More recently, San Onofre 2 experienced a degradation N-9000 seals.
of CBO conditions that resulted in a manual reactor trip and shutdown. On April 8,1991, operators noticed that In 1986, the Waterford 3 staff began to look for alterna-CBO conditions relative to one RCP seal were degrading, tive seal designs that would solve their operational prob-in that reduced interstage pressure and increased CBO lems. In 10/87, the N-9000 seal was installed in one RCP, and run for the remainder of the cycle. In March 1988, the discharge temperature were observed. The plant staff began trending scal-related parameters to monitor the N-9000 seal was installed on the remaining 3 pumps. All condition of the seal as the situation developed. Finally, the new N-9000 seals failed on startup (11/88), and the on April 10,-1991, a total loss of CBO flow from the seals in all four pumps were removed and replaced with affected seal was suddenly indicated. In accordance with the SU seals.The plant is continuing to run with the SU the plant's abnormal operating procedures gomrning to-seals installed.
tal loss of CBO flow, the operator manually scrammed the reactor. The licensee brought the plant to cold shut-2.5 Bin 0 am-Willamette Seal d wn, and rem ved the RCP seal cartridge. Preliminary h
indications were that all stx bolts holding the rotatmg Re Bingham seals specifically designed for installation in
. baffle in place had fractured, causing the baffle to shift in B-J RCPs are cartridge assemblies, and like B-J seals and position and rub against adjacent components, resulting are installed and removed from the pump as a unit, he in blockage of CBO flow to the RCP seal,in the Licensee seal cartridge contains four seal stages, three active and Event Report describing this event, {Ref. 7] the hcensee one " vapor seal." The seal operates on the same basic stated that the RCP seal vendor (Bingham-Willamette).
principle as the B-J SU seal discussed abovet the differ-had performed a test involving approximately 500 hours0.00579 days <br />0.139 hours <br />8.267196e-4 weeks <br />1.9025e-4 months <br /> ence between the two seal designs lies primarily in the of continuous RCP operation with no CBO flow, ne design of their individual seal stages and materials.e integrity of the installed Bingham-Willamette RCP seal was maintained during this test. At the time of the April 10 event, the San Onofre 2 procedures addressing a total loss of CBO flow event prescribed that the reactor be NNu EnNnYmNtanu$cYuYgTnEchNp]
immediately tripped and tim affected RCP shut down to 1989 (12mited istribution).
avoid Impending damage to the pump seat. These 13 NUREG-1275 o
i Table 2 Advanced flyron Jackson N-9000 RCP Seal Failures j
Plant Failure Date Comments -
l Waterford 3 11/07/88 N-9000 seals installed in all pumps failed during plant start-up. N-9000 seals in all 4
four pumps were replaced with SU seals.
Maine Yankee 12/22/88 During power operation, newly installed N-9000 seals on RCPs 1,2, and 3 lost two out of three seal stages. Licensee's root cause analysis determined that the rapidly deteriorating condition of the newly installed seals (installed during refueling outage which ended 12/16/88) was caused by silicon oxide (sand) in the RCS due to backflow from the ceramic filters in the seal injection system. Thus the seal failures experienced were not design dependent.
c.
\\
procedures had been developed from operating guide.
SU type seals have frequently experienced degradation lines provided by the original RCP seal vendor (B-J).
due to seal stage failure, and that they are temperature Based on the results of the Bingham loss of CBO flow sensitive. As summarized in Section 3.1, the loss of CCW, test, the San Onofre 2 licensee is revising the plant proce-even for very short periods, is sufficient to cause seal dures to direct the operators to perform a controlled damage. Changes in service water flow, or rapid small plant shutdown rather that scramming the reactor, in the CCW temperature changes, can cause degradation of the event of total loss of CBO flow, seals. Some utilities have used a combination of mainte-nance improvements and reduction in the incidence of Baltimore Gas and Electrie (BG&E), after an analysis of CCW transients to improve performance with these seals.
their RCP seal operating experience, made a decision to retrofit the RCP seals at Calvert Cliffs 1 and 2 with r
Bingham-Willamette seals. Unit l's seals were replaced The alternative RCP seal designs for B.J RCPs, the Hm.-
in late 1989 and early 1990 (approximately 11/89 and gham-Willamette and the B-J N-9000, have not been 3/90), and Unit 2's seals in approximately 6/90. The Cal-trouble-free, but, at the time this report was prepared,it vert Cliffs units have limited operating experience with was difficult to make definitive performance compansons the Bingham seal, with SU seal performance due to the limited amount of operating experience accumulated with these alternative d*#8"
3
SUMMARY
OF OPERATING EXPERIENCE Operating experience points to the necessity for en.
hanced quality assurance and maintenance for successful performance with both old and new designs. Most nota-bly, as discussed in Section 3.2, failure progression to As discussed in the previous section, early operating ex' multiple stages can be accelerated in cases where me-penence with B-J SU type RCP seals can be characterized chanical damage has occurred. Improperly performed as unreliable. As a result, a variety of corrective actions maintenance represents a prime source for such damage.
have been and are being pursued by utilities which have Also, as summarized in Section 3.2.1, instrument sensing i
improved RCP seal reliability Maintenance and quality
- I ne leaks have recurred at several plants with B-J RCP assurance improvements have been made. Also, new and seals. The failure of sensing lines can initiate RCP seal modified designs have been developed and installed at
- failure, various plants.
Regardless of which seal design is installed, it is not con-There are four RCP seal designs currently in use at plants sidered off-normal for a B-J RCP to be operated with a with B J RCPs: the SU type, the modified SU type, the failed seal stage, and planta operate for extended periods advanced N-9000 seal, and the Bingham-Willamette seal in this condition, as summarized in Section 3.3. The oper-design. Most of the operating experience has been accu-ating guidelines issued by the pump vendor (B-J) allow.
mulated with SU type seats, and SU seals are still installed continued operation for indefinite periods with a com-in the majority of B-J pumps. Operating data show that pletely failed seal stage, subject to enhanced monitoring NUREG-1275 14
i of available instrumentation, and thus leave wide discre-3.2 Qtiality Assttrance and tion to the operating utility's engineering judgement.
Maintenance Considerations Plant operating procedures generally do not require plant l
shutdown until a second stage fails completely, llowever, seat failure progression is not predictable-a point indi.
cated from our review of operating experience and subse-Forced by operational experience of short seal life, utili-quently confirmed through discussions with utility per-ties have placed greater emphasis on quality assurance sonnel. The result of our review and subsequent issues to achieve operational and reliability improve-discussions with licensees indicate the need for licensees ments with both the original and modified SU seal de-to establish strict and safe operating limits for their RCP signs. The improvements reduced plant downtime. The 4
seals, thereby eliminating concerns regarding continued experience at Fort Calhoun with the original SU seal operation with a degraded seal. Limits that are substan-design points out the degree to which quality assurance 4
tially different from the vendor's recommended limits and attentica to detail can control the frequency of seal should be justified internally by the licensee. Plant procc-degradation. The most important ingredients for success dures and available instrumentation are discussed with with the SU seals are the understanding of the seals
- close respect to continued operation in Section 3.3.1.
tolerances and the expertise of the maintenance person charged with the overall responsibility for seal-related maintenance. Also contributing to the success at Fort Calhoun is the attention to re-start of the seal after a 3.1 Teniperature Sensitivity of SU change-out. After seal installation, the seal cavity is al-Seals ways re-f ooded with clean water via the seal flush line, and the seals are carefully vented by procedures. The use of a clean source of water for flooding the seals prevents crud that has settled in the RCS from being introduced i
Temperature transients have been shown to have a sig-into the seal cartridge and causing seal damage on pump nificant impact on SU seals installed in B-J pumps. Minor start-up. Plant staff at Fort Calhoun credit exceptional temperature changes have been sufficient to degrade SU seal experience to the fellowing: (a) changing out e
seals. Damage to the seals has occurred in a very short seals every cycle (b) following proven operating and period, as is illustrated by the event at Calvert Cliffs 1 maintenance procedures, (c) having a knowledgeabic and described in a previous section (L11R 318/85-001). The dedicated maintenance person charged 'vith all scal-re-experience at St. Lucie also highlight.c the sensitivity un-lated tasks, (d) continuous operation without transients der fairly normal operating conditions. Table 3 provides a that stress the seals, (e) attention to details related to summary of reported cases of the impact of CCW flow component tolerances, and (f) as stated previously, the transients on SU seals.
relatively small RCP shaft diameter results it a low seal surface velocity. Thus the seal "PV" is relatively low.
The data regarding the CCW flow transient-initiated i
RC, seal failures shown in Table 3 (and in Table B-2 in In an effort to improve the performance of their SU type Appendix B) is incomplete, since similar data were not seals, several utilities have used Atomic Energy of Can-available for all plants which have B-J RCP seals in-ada, Ltd., (AECL) as consultants. AECL provided seal stalled. Data from sources such as NPRDS contain a maintenance training, parts quality control, technical number of RCP seal failures where the failure causes guidance during seal cartridge rebuilding, and seal car-were identified by the reporting utilities as unknown.
tridge testing. Based on discussions with two different Applying engineering judgement to the event descrip-utilities, this assistance was considered beneficial. For -
tions, some of these seal failures could be attributed to example, many of Waterford 3's routine operating prob--
seal cooling transients. 'Ihe data in Table 3 are based on lems have declined since obtaining AECL assistance.
failure summaries provided in Appendix A.
The operating data show that the rate of progression of Although not all plants with B-J pumps were evaluated, multiple seal stage failure is related to the nature of the based on the limited discussions with plants to date, it initiating event (i.e., mechanical problems, or the intro-4 appears that plants are well aware of the problems associ-duction of foreign materials into the seals). Mechanical ated with perturbation of CCW flow. The B J RCP seal failures and the introduction of contaminants into the operating guidelines appear reasonable for short term scal can lead to failures that start out slovly, and then losses and restoration of CCW (approximately 10 min-t.nexpectedly and unpredictably place in jeopardy addi-utes). However, the guidelines for recovery and seat in-tional se9 stages (See the Fort Calhoun event of spection for losses of CCW for periods longer than 10 8/15-8/24/90 discussed in Section 2.2.1). Other examples minutes provide less specificity.
of unexpected and rapid seal failures are as follows:
15 NUREG-1275
Table 3 Component Cooling Water Transient Initiated SU Seal Failures Since January 1985 =
Number of
_ Plant Date Pumps Affected Comments Arkansas 1 1/85 One ICW transient.
Arkansas 1 2/85 One ICW transient.
Arkansas 1 8/87 Two Dispatcher error led to loss of cooling for 3 minutes, Arkansas 2 4/80 One loss of power (elective seal change).-
Arkansas 2 9/85 One Seal replaced after two loss of cooling events.
Waterford 3 5/85 One Loff of ac power testing during power ascension.
Calvert Cliffs 2 4/85 One
. CCW transient initiated due to service water heat exchanger removal.
(1) Crystal River 3 (1/2/87)-within an hour of the first Calvert Cliffs 1 and 2, althougit data may not be complete stage failure, the second stage began to degrade for plants not visited. Plant specific support and piping (this plant's RCP seal design has only three seal arrangenents probably play a part in the frequency of this stages). The Crystal River failure was initiated by type of problem.
metal filings from lost drive keepers.
The Arkansas 2 event of 8/88 highlights the potential (2) Crystal River 3 (7/2/87)-seal failure initiated by impact of sensing line failure.The " vapor scal" which is
-axial upthrust.
used to contain RCP leakage on failure of the lower seal stages also began to fail, and seal degradation continued (3) Maine Yankee (12/22/88)-an unexpected loss of after the pump had been secured.
two of three stages on all three pumps (N-9000 Sensing line failures have been attributed to vibrations, seals).
thermal stress, improper rigging during seal change-outs, E W"E '## *# "I*"""" "" '* E'"'# " " '
(4) Maine Yankee (9/29/89)-one of the recently in-s nded ahe, @ potendal-ic m e area stalled modified SU seals failed because of metal due to sensing line failure is evident from 8
filings introduced into the seal from a lost keeper ea which had broken off the driver ring on the first seal stage.
3.3 Operation With Degraded Seals.
(5) Palisades (9/16/84)-all three stages failed (vapor All plants with B J RCPs are allowed by procedure to seal remained functional)while the plant was oper-
' perate for an mdetermmate period of time with seal-ating at approximately 60% power due to failure of a shaft coupling bolt.
stages.m a failed condition, long-term operation with -
failed RCP seal stages reduces seal stage redundancy, and -
may increase the tisk of RCS leakage.The operating data (6) St. Lucie 1 (6/24/90)-possible bent shaft or contain numerous events which illustrate uncertainties misalignment of pump and motor, or pump unbal-regarding the behavior of seals under off normal condi :
anced.
tions, such as_ the time to failure for seals previously ex-posed to elevated temperatures. These uncertainties 4
3.2.1 Sensing Line Leaks
- complicate the_ management of continued operation with a degrading scali Leaks or cracks of various instrument sensing lines and -
other seal taps ' injection, vent) have been a recurring -
-Guidelines from the vendor and plant operating proce-:
"i
_ problem at several plants with Il-J RCPs. Appendix A dures allow RCP operation with a seal package in a de-contains tables listing examples found in data made avail-graded state, i.e., at most plants with installed U-J RCP able by plants for this study.The majority of the reported seals, the RCP can be operated with one failed seal stage events appear to have occurred at Arkansas 1 and 2 and and a second seal stage degraded. Table 4 provides a NUREG-1275 16 1
.. _ _ _ ________:____ ___________-____ x ___
1
i compilation of situations where such operation has oc-observed seal degradation, vendor guidance recommends curred.This table is most complete for the plants visited stepped-up seal performance monitoring, which, as dis-during this study, and probably incomplete for others.
cussed below in this section, relics to a great extent on Note that plant-specific data from NPRDS are proprie-engineering judgement. Utilities set criteria related to tary. Since the " Days Operating" column in Table 4, how long the plant will continue to operate with a failed which was calculated from NPRDS data, was not verified seal stage and other stages in a degraded conditioniin with each individual plant, the plant names have been effect, each plant is making decisions based on its own omitted from the first portion of the table. In the event of criteria and experience.
4
. Table 4 Continued Operathi With Failed Seal Slages Seal Discoven Days Plant Type Date Operating Comments From NPRDS Data:
Plant 3 SU 04/25/85 2
Forced outage upper and middle seal failed.
I Plant 4 SU_
06/12/85 60 First stage seal failure-operated for two months.
Plant 5 -
SU 07/29/85 46 Plant operated for approximately 1-}2 months.
I Plant 4 SU 02/19/86 18 Plant operated with failed seal due to misaligned U-cups.-
Plant 4 SU 02/19/86 18
. Plant operated with first stage seal failure. Pump i
50 8.
Plant 5 SU 05/20/86 24 Plant continued to operate with an increasing CBO flow. Pump 2P-328.
Plant 6 SU 06/30/86 149-Plant continued to operate with increasing CB0 y
<lue to degraded seal. Pump 1-B.
nata 1; SU 07/16/86 149 Plant continued to operate with increasing CBO flow due to degraded seal. Pump 2-B.
4 3.
Plant 6 SU-08/24/89 120a Plant continued to operate with degraded seal until refueling.
I
, ant 6 SU 08/24789 120a Plant continued to operate with degraded seal -
I' until refueling. Pump 2-A.
l Plant 7 SU 07/30/85
< 30 Leaking seal forced shutdown.
Plant 7 SU 10/09/85
< 100 Leaking seal forced shutdown.
7 Plant 5 SU 11/22/85 b
Leaking at power.
Plant 7 SU 10/31/88 123 Continued operating with one stage failed.
Plant 7 SU 03/02/89 120 Continued operating until second stage failed.
Plant 8 '
SU
-02/16/90 b
Sealfailure forced a shutdown.
Plant 9
_SU.
12/05/87 b
Plant continued to operate with a degraded seal.
}
17
4 4
Table 4 (continued)
Seal Discovery Days -
l Plant Type Date Operating i
From Plant Reports and Plant Logs:
I Arkansas 1 (Pump P32A)
SU 08/87 22 SU 08/88 42 4
(Pump P32C)
SU 01/85 b
SU 01/85 10 SU 11/85 37 SU 08/88 7
1 Arkansas 2 (Pump 2P32A)
SU 09/85 5
(Pump 2P32C)
SU 09/85 2'-
(Pump 2P32D)
SU 11/87 62 Fort Calhoun SU 08/90 9c l
There were indications for four anonths that the seal was degrading.
a b
Time could not be determined.
4 e
NP.C Inspection Report No. 50-285/90-35.
4 All of the cases in Table 4 involve the SU type seal, for rapid progression appears to arise from a mechanical-re-4 which the most operating experience was available when lated cause. However, at the onset of failure of the first this report was prepared. The sensitivity of the SU seal to stage, the operator monitoring the seal may have little cooling water temperature was discussed previously, idea of the root crase (and whether it is of a random Continued operation with a degraded SU seal could rep-nature or involves a common ca use failure mechnism) of resent an undesirable vulnerability during normal opera-the problem, and hence of the amount of time until fur-tions when additional cooling perturbations could occur, ther damage occurs.
Discussions with key staff members at a number of oper-
- ating plants indicate that small chan :es in CCW flow occur frequently at their plants, rebated as pressure Numerous failure records in the NPRDS database refer-spikes in seal stages. Additional CCW flow adjustments ence continued operation of the RCP for periods ranging are used in attempts to counter pressure oscillations in from two days to several weeks, until the second or third -
seal stages. Published risk analyses, such as those per-stage of a RCP seal failed. The oEerating data show that formed in support of NUREG-1150 for Westinghouse pl nts continued to operate with increas ng CEO flow i
l seals, [Ref. 9] have azumed that all seal stages are fune-(CllO flow plus seal leakage). Operatmg m this manner Saning at the onset of loss of cooling. Analyses of sunival.
can cause the C110 excess flow valve to close, accelerating 4
times for Ilyron Jackson seals with seal degradation pre-total seal failure, if it hasn,t already occurred.
existing the loss of cooling were not identified in this study.
While the operating experience generally reflects suc-cessful management of degrading seals, the uncertair. tics The time period between tbc onset of additional seal concerning the rate of propagationi the nature of the -
stage failures after the initial stage iails is not predictable, failure mechanism (random or common cause), the vul-and thus enhanced monitoring of seal stat us is performed, nerability to cooling transients, and the limited ability to L
Operatirig data show that plants have operated from a few assess the seal condition through instrumentation suggest hours (limited by additional degradation) up to approxi-that operation with a degraded seal is a course of action mately 180 days with a single failed seal stage. As noted that should not be followed unless appropriate precau-above in the discussion of quality assurance, the most tions have been taken.
. NUREG-1275 -
18
~
=
In the case of four stage seals, operators are also depend-temperatute should be maintained below 200 degrees F ing on the failure-limiting performance of the " vapor (temperature reduced from 550 degrees F to 125-150 seal" to limit leakage in the worst case. nis seal under degrees F)J [Refs. 4 ] Specific plan t operating procedures normal operating conditions only sees volume control examined also state that the lower seal cavity temperature tank pressures (40-50 psi). Operating data coritain exam-should be maintained below 200 degrees F. It is our un-plcs of seal stage failures on start up. The sudden expo-derstanding that other plant procedures specify that this sure of the vapor seal to RCS temperature and pressure temperature be maintained below 170 degrees F. B-J may introduce enough stress on the seal to cause damage.
RCPs installed in operating plants do not have direct first stage (lower) seal temperature measurement capability 3.3.1 Instrumentat. ion and Operator (see Appendix, Figure 2). [Refs. 5,8] Although the ad-Guldance vanced B-J N-9000 RCP seal design does have the capa.
bility to measure the first stage seal temperature directly, A review of all the instrumentation available for seal most plants which have N-9000 seals installed in the monitoring currently m place m plants with installed B-J RCPs do not use this feature.
RCPs is beyond the scope of this analysis. He adequacy Plant operating staff indicated they believe the most use-of RCP seal-related instrumentation in responding to ful and dependable instruments are the temperature anticipated seal failures was evaluated in detail in measuring devices. It should be noted, however, that the NUREG-4544 [Ref.4] The purpose of the discussion in temperature measurements in some cases are only ap.
this section is to summarize the observations from plant proximations. Due to physical restraints, locating the visits as they relate to the reliability of indication for probe closer to the stage cavity is difficult (See Figure 2 in i
monitoring seal behavior du ring continu ed operation with Appendix A). Although cavity pressares are considered a degraded seal.
important, discussions with licensees indicate that large pressure oscillatior.s are quite common, and it is not un-The linated cases of instrumentation which were exam-usual for them to start and stop periodically. At one plant, ined during plant visits appeared to be adequate for nor-pressure fluctuations of 200-300 psiwere noted for a seat mal seal operation and for detecting the beginning of seal stage. At this particular plant, the operators place less degradation. Existing instrumentation allows control emphasis on pre *ute and more emphasis on CBO tem-room personnel to detect off-normal seal conditions by perature to determine the status of the seal !!owever, way of alarms or changes in measured parameters. Plant discussions with experts in tiie area of RCP scal 'erform-l operations staff and the system engineer respoasible for ance indicate that such oscillations are reflecting the ac-the RCP are then brought in to analyze the off normal tual seal behavior-namely that it is rapidly opening and -
seal conditions indicated by installed instrumentation.
closing-and may be experiencing excessive wear or dete-Upon detcetion of such a seal related operating condition rioration in the process.
(e.g., failed stage, pressure fluctuations, or increasing temperature trends), decisions for continued operation A flowmeter for indicating CBO flow rate is usually to-are left to the system engineer, operatioris staff, and plant cated in the CBO outlet which is directed to the volume management (if conditions deteriorate-i.e., additional control tank. Because of the range of possible flow rates stages degrade or fail-the opua'. ors are instructed to (normal CBO flow is on the order of 1 GPM) involved, take appropriate actions to secure the affected RCP and flow instrumentation may not be very reliable. CBO flow shut the plant down). However, accurate prediction of metering should encompass CBO flow from normal up to impending seal failure is difficult using existing instru.
a level recommended by the pump and seal manufac-mentation.
turer(s) at which the cooling water heat exchanger's ca-pacity is exceeded. Exceeding the capacity of the heat The parameters ofinterest in assessing seal performance exchangers at some plants causes an isolation of CCW, (and hence for determining seal failure) for B-J pump which may increase the probability of seal damage.
seals are seal cavity pressures (middle, upper, and va-por-the lower seal cavity pressure is assumed to be the The limited cases of instrumentation reviewed appeared RCS pressure, which is measured elsewhere), seal cavity to meet the operating requirements for normal opera-temperatures (increases confirm leakage), and CBO flow tion.1deally, the more information that can be provided rate. [Refs. 3,4] At a plant with B-J RCPs which has seal to the plant staff regarding seal behavior under off nor-injection, normally there should not be an increase in mal ecmditions, the less the uncertainty in assessing the temperature associated with leakage because the water risk of continued operation with a degrading seal. For passing through the seal cartridge is seal mjection water example, some plants have no provisions for measuring and not water from the RCS. However, if the tempera-one or more of the following parameters: upper seal ture were increasing, then this increase would be seen in
" Reactor Coolant Pumpglidg norNrYnnNria E ratin Guideliner n Jacbon Four s;ag,kcgn the CBO outlet temperature. Based on operating guide-scai s m lines from B-J, the desired lower seal cavity operating Pump Division, No e e
mited Disfribunon). '
19
cavity temperatue, middle seal cavity temperature, or age rate of 40 gpm, which was within normal makeup seal !cakage flow. For purposes of seal operation under capacity) have occurred at a plant with B-J RCPs.
normal conditions, this is not considered a problem.1-low-ever, the Arkansas 2 event of 8/88, the Fort Calhoun The redundancy of the B.J RCP seal design is such that, event at 8/90, and the more recent event at San Onofre 2 with a random failure isolated to one seal stage, there are in 4/91 point to the potential for accelerated deteriora-sufficient intact seal stages remaining that the seal can tion of RCP seals.
perform its design function satisfactorily, thereby allow.
ing continued operation of the affected RCP, while main-4
SUMMARY
OF CONCLUSIONS tainin8 a degree of wal adundancy. An inacased wear rate of the remaining stages would be the only expected result. Ilowever, as with most redundant equipment, there are common cause failure vulnerabilities: in gen-The operating experience for B J RCPs supports the cral, the stages are constructed of the same materials and value of (1) limiting exposure to operational transients, have the same design; the stages are in close proximity c.p., cooling watcr transients, and (2) of applying high and communicate via leakage, controlled bleedoff flow quality standards to reactor pump seals. Utilitics have and pump shaft movement / vibration; the stages share the taken corrective action in both areas, in particular, utili-same cooling source; all stages accumulate service hourt ties have adopted tighter and tighter quality controls over (i.e. wear)at roughly the same rate. Operating experience time. Insights gained during the evaluation of B-J seals reflects these vulnerabilities, as the historical incidents of have shown that seal quality assurance (QA) efforts were multiple stage failure have been the result of mechanical not a priority or consideration durinF carlyyears of opera-failures, in particular CBO flow blockage, or the intro-tion. On-gomg operational problems with seals have duction of contaminants into the seal.
forced utilities to consider implementing QA programs for certain seal-related efforts (i.e., scal inspection, main-Determining the exact cause or nature of a given seal tenance personnel training, seal procerement, and seal problem (e.g., whether it involves an independent or a storage),
common cause mechanism) while online is generally not possible, so the emphash shifts to close monitoring and Maintenance procedures appear to be of critical impor.
prompt reaction by operating personnel in the control tance if reliable seal or ration is to be achieved. Plants room. Continued operation with a degraded seal requires with poor maintenance procedures tended to experience personnel with detailed knowledge and understanding of frequent problems with their RCP scals. Most plants have seat operatton and the signs of progressive failure. In recognized that maintenance procedures are important many cases, plant staff consulted with outside experts to and, during plant visits and surveys, n was noted that supplement their diagnostic capability.
plants were in the process of upgrading their maintenance procedures.
Thus, operating' experience reinforces the value of the following factors in order to safely manage operation of a a
pump n a ns of a &gMngwah Plants with B-J RCPs that have had relatively good expe-rience with their RCP seals attribute this success to a combination of different factors, including: enhanced scal (1) Seal operatmg limits should be prudent and safe Q A efforts, modified /new seal designs, improved mainte-limits that are strictly adhered to. The established nance procedures and trainiag, attention to detail, im' limits should be comparable to the litaits recom-proved seal operating procedures, knowledgeable per-sonnel mvolved m seal mamtenance, reduction in mended by the vendor. Any limits that are signifi-cantiv different than those recommended by the frequency of transients that stress the seals, seal handling vend'or should be internally justified. These limits and installation equipment designed to the appropriate should consider the experiences related in this re-precision, and mamtenance of a clean seat coolmg water port and the specific RCP seal design installed system. As more plants have implemented corrective measures such as these, the number of B-J RCP seat (2) Plant operators should assure that the RCP seats are failures experienced has tended to decrease.
not already in a degraded state prior to the loss of the first seal stage (e.g., consideration of installed This study included a review of the practice of continued in-service time, any accumulated time seal has been operation with a degraded sealin the case of PWR plants -
exposed to loss of coeling conditions).
with Byron Jackson reactor coolant pumps. Continued operation with a degrading seal was a facter m the May (3) The operators and engineering support staff should i
1980 gross seal failure which occurred at Arkansas 1.
be familiar with RCP seal operatior. and the benav-Since this event, no other instances of gmss seal failure foral characteristics of Seals under a spectrum of (the August 1983 event at Arkansas 2 resulted in a leak.
off-normal conditions.
NUREG-1275 20
(4) Plant operators should be provided with. clear, (3) *htain Coolant Pump Shaft Seal Guidelines," EPRI unambiguous procedures covering all aspects of NP-2965, Vols.1-3, Electric Power Research Insti-normal and abnormal RCP seat operation and have tute, Palo Alto, California, March 1983.
been thoroughly trained in their use.
-(4) " Reactor Coolant Pump Seal Related Instrumenta-(5) Sufficient instrumentation should be provided to tion and Operator Response,"NUREG/CR-4544, W. J. Luckas, Jr. et al., Brookhaven National labo-allow plant staff to monitor the performance on the intact stages so that the operators will be able to take ratory, Upton, New York, December 1986.
appropriate and timely action to prevent gross leak-(5) Licensee Event Report No. 80-015/0lX-2, Arkan-age in the event that further (e.g., another seal stage sas Nuclear One, Unit 1, Arkansas Power & Light fails) degradation of the RCP seal occurs-Company, ljttle Rocl Arkansas, April 13, 1981, U.S. Nuclear Regulatory Commission Docket No.
50-313.
5 REFERENCES (6) " Reactor Coolant Pump Seal Information," letter (with attachments) dated July 16,1980, from David C. Trimble, Arkansas Power & Ij ht Company, to g
(1) "The Impact of Mechanical-and Maintenance-In.
R. W. Reid, Chief, Operating Reactors Branch #4, duced Failures of Main Reactor Coolant Pump U.S. Nuclear Regulatory Commission, USNRC.
Seals on Plant Safety," NUREG/CR-4400, M. A.
Docket 50-313.
Azarm et al., Brookhaven National Laboratory, Up-ton, New York, December 1985.
(7) Licensee Event Report No.91-007, San Onofre Nu-clear Generating Station, Unit 2, NRC Docket No.
50-361, Southern California Edison Company, San (2) "An Analysis of the Historical Performance losses Clemente, California, May 10,1991.
Caused by Main Coolant Pumps (MCPS)," Eric A.
J. Olson, Stoller Power Division of RCG /Bailly, Inc.,
(8) " Severe Accident Risks: An Assessment for Five presented at the Fourth International Workshop on U.S. Nuclear Power Reactors," NUREG-1150, Main Coolant Pumps, Phoen'x, Arizona, April Second Draft for Peer Review, U.S. Nuclear Regu-15-19,1991.
latory Commission, Washington, D.C., June 1989.
t I
l l
-l
'1 21 NUREG '?'15 -
.l 1
i 4
4 J
i APPENDIX A i
l Summary of Reactor Coolant Pump Seal Histories 9
i e
i 4
i t-I l
4i 4
k 4
i l
1 4
1 l
?
I i
i 1
l
)
ij l
l 1.
i
,a m
1 a
\\
t i
1-4
+
TAllLE OF CONTENTS Reactor Coolant Pum p Schematics................................................
A-2
{
Figure 1 -Three Stage Reactor Coolant Pump Seal........................................
A-3 Figure 2-Four Stage Reactor Coolant Pump Seal.........................................
A-4 Tables of Reactor Coolant Pump Seal Histories.......................................
A-5 Ar ka n sas U n it 1..................................................................
A-6 Arka n sa s U n i t 2................................................................. A-18 Wa t e rfo rd U nit 3.................................................................. A-22 Calvert Cliffs Units l and 2.......
...............................................A-25 S t. Lu cie U n it 1..................................................................... A-26 S t. Lu ci e U n it 2....................................................................... A-27~
Pump Seal Instrument / Cooling Line leaks.................................................... A-28 LER 368-11: Arkansas Unit 2 Sensing Line Failure............................................. A-31 A-1 NUREG-1275
Reactor Coolant Pump Schematics i
l NUREG-1275 A-2
I 1
i s
)
4 1
,j.
MOTOR JOURNAL-AND THRUST BEARING MOTOR j
r m
MOTOR JOURNAL BEARING y
a RIGID SPACER COUPLING-E EFW
~
~
i
=
j CONTROLLED
- 1
?
BLEEDOFF g --
[-
6P 3
-/:
t
.. STAGING-J.
gp P
g = DIFFERENTIAL ~
STAGING COILS 2
j PRESSURE l
INJECTION gp PUMP MECHANICAL 3
WATER-1 SHAFT SEAL CARTRIDGE d
()-STAGE) i d
~,~I -
R! CIRCULATION PUMP COOLING ATER PUMP BEARING-a-
-4
[
IMPELLER j
L g
- 1. r
/\\
PRIMARY COOLANT J
3 Figure 1.' Three-Stage Reactor Coolant Pump Scal' i '
A-3 NUREG-1275 :
M th I
i=W
.k I, vapor les) testate,
_,;,. f to Radvette (4) gph,,
I' I " '
b Centrollet 81 edoff rf14*tter i
- t. xi n.o, >
lack,rettore I l V9ter $tel Q
150,ite E3 Jl i;';;;;'f'"*"
.i..i. ni v E l
n x,ii.
- t. r 5 5 %
- 1.,,,,to,,e,os, me,.6 Aus118ery lar.11er h
I8
.r i
a i.t v.i..
cm Therme) $4rtl6r C u,v.ie, iv y i._
,ydrostatic learin,
~,VQ 1g f{
met. t,etter M'- "'"4 n T
Figure 2. Four-Stage Reactor Coolant Pump Seal a
NUREG-1275 A-4
Tables of Reactor Coolant Pump Seal Histories 1
3 l.
I A-5 NUREG-1275 I
Q ARKANSAS UNIT 1 FA! LURES / REPLACEMENTS OF SEAT.S OR-PUMPLCOMPONENTS i
P32A 5
Date Description-Source-
'l I
03/22/74 Remove / reinstall seal cartridge J.0, 253 03/15/75 Outer seal destaged, replaced-(ANO -Failure Pro-was operating on this date info.
gress Repor_t-05/09/85 possibly wrong-further invest. )
Dated 05/16/80, 1
-Appendix C 09/06/76 Worn shaft seal cartridge, replaced;-
Progress:
abnormal staging _ press.._ind, oscillations Report of.,
J.O. lo97-05/16/80; NRC:-
'Ietter- 07/16/80 ---
11/23/77 All: four pump cartridges replaced Progress Report 05/16/80 11/26/77 Seal. leakage, replaced cartridge and Progress Report; recirculating impeller lock ring.
NRC letter J. O. -- 4402/77-3 07/16/80 04/07/78 High wear on seal; replaced cartridge
- P.R. 5/16/80;-
Avail. Seminar 10/16/79;' Tech.
Conf.-03/10/81-06/02/79 liigh_ wear - excessive. leakage J.O.
- NRC letter 4403/R3 07/16/_80' 11/03/79 Replaced. seal cartridge;-J.O, 4403/7913
-1 Tech. Conf.--
middle: seal stage failed on-high pressure._
LO3/10/81;NRC start (08/15/79) following loss: of powerL letter 07/16/80i-H to A&C pumps and Rx trip.
Continued ops:
- B&W refurb.:
I until outage 2nd stagescarbon crackedL--
notes; B&W-
_ press, oscillations.-
letter'10/05/79?
'05/10/80
' Abnormal staging; catastrophic, failure
.!NRC letter of P32C - U-cups generally hardened at' 07/16/80;LB&W the _i.d. (working on sleeve. produced letter 06/20/80
' heat and = friction); L1ight heat checking.
B&W letter-on carbides.:
06/16/80. pg. : 9 -
--NUREG-1275
,A-6 i%...
Date Description Source Sign.iicant transients prior to this failure:
l 04/07/80 - Loss of offsite power B&W 1etter 06/20/80' 05/04/80 - Intentional loss of seal Notes from injection for test of NNI B. Garrison (05/04/80) 02/20/81 Elective replacement of cartridge J.0. Attached J.0, 3640
- New seal bindino -
'iis teilation 08/23/81 Replaced seal; J.O. 15908; J.O. 15913 J.0.'s; in 08/81 - a malfunction of the generator B&W notes exciter caused a trip of the Rx and 08/16/81 turbine; lost power to A&C pumps; subseque.nt loss of cooling.
03/31/82 Replaced seal cartridge; installed modified seal.
04/82 Motor overhaul, J.O. 22074 11/82 (IRS)
Replaced seal cartridge; J.0. 35485 Task data
- 1 stage degraded - U-cup fretting all attached 3 stages; rust in upper stage Seal may have been replaced again QC find.- dated prior to S/U (1/20/83) 01/20/83 04/14/83 High vibration on motor-upper guide J.0, 44935/
shoes adjusted; J.O. 44935; motor didn't J.0, 44937 start due to problems with LS-6535A-(J.O. 44937) 12/11/84 Removed cartridge - first' stage hung Task Data (IR6) open on restart 10/05/84 - later restaged; J.0. 74900 01/10/85 Replaced seal cartridge (J.O. 54706),
Task Data; (IR6) due to upper stage failure following MPR Report an ICW transient. _ Inspection-showed of 03/01/85
- 2 stat seal face retaining ring (11-12) was partially out of its slot and was damaged.
Suspect this was the wrong ring (i.e. it was an 11-13, 3rd stage ring).
Date Description Source 02/11/85 Repi m d seal cartridge due to failed Task Data; 3rd stage following elevated press. tast MPR Report (2285 #) and ICV transient.
Modified seal by removing upper shaf t sleeve 0-ring and drilling vent hole through the adj. cap. (J.O. 81176, OCP 85-1010);
inspection showed lower shaft-sleeve 0-ring was out; sleeve expanded to -the point of requiring #3 rot. seal face to be broken for removal.
01/28/86 Some increase in leak 39e during s/u stowed later (lower-o-ring leakage)
Fall 86 Replaced cartridge durir,g IR7 due to RCP shaft inspections.
08/25/87 Unit tripped due to dispatcher error; "A" pump tripped and lost all cooling for 3 minutes. When pump was restarted, Icakage of x 759pm, vibration was 2 mils higher (x 19.'S mils) and upper cavitiy pressure was.80 gpm, 1062 psi upper cavity vibration. On 10/7/87 temp.
transient of approx. S degrees - leakage decreased to.46 gpo; upper and lower cavity pressure increased - On 10/14 upper cavity pressure equal to 1340 psig.
10/16-30/37 Unit was shutdown for midcyle outage.
"A" seal changed out; durino refueling maintenance couldn't cet a proper fit between coupling and motor nachined "A" coupling (this was a'new coupling installed with the new cr tridge i
during IR7 - it was not machined then like the other 3 shafts.) Through subsequent discussions maintenance people said fit was hard at the end of 1R7. Jacked-the motor shaft over and fit was okay
'Maint. procedures are now ::ustomized to make magnetic motor center and geometric center the same). Expect this will reduce vibrations.
j Seal disassembled in late Nov. (24th) -
nothing observed, but no flow test on coils yet.
08/27/88 Installed N-9000 seal l
l (1R8) 01/10/90 3rd stage op decreased to 231 following shutdown from 1 m89 (12/89) leakage is.77 gpm -
Other parameters oksy, trends in c.hron file NUREG-1275
- A-8
P328 Date O_escription Source 08/25/73 Rotor shaf t on motor returned to AP&L letter Allis-Chalmers for repair af ter failure dated 11/19/73 of anti-rotation device.
03/22/74 Replace seal cartridge J.O. 254 08/11/76 Replace cartridri - highwear, debris NRC letter entering seal; failed during S/U on 07/16/J0; Avail.
8/7; J.0. 1575 Seminar 10/16/79; Tech. Conf.
03/10/81 09/05/76 Seal cartridge replaced due to improper NRC letter venting; J.O.1596; lower seal failure 07/16/80 l
12/03/77 Replaced seal cartridge J.0, 4401-77-3; J.O. ; NRC letter l
(1R2) preventive maintenance: motor %spection 07/16/80; VRP l
performed under work request permit 6017-R2 l
6017-R2 473 01/01/79 Replaced seal cartridge - elective NRC letter maintenance J.0. 4403/78-6 07/16/80 11/06/79 Replaced seal cartridge J.O. 4406/79 J.0. 4406/79-3; erratic performance, elective mainterance; B&W inspection oscillations notes 12/79 01/27/80 Replaced seal cartridge - 4406/80-1; NRC letter J.O. 2001; 4408/80-1 07/16/80, J.0. 2001 05/16/80 Remove cartridge following 05/10/80 P32C J. O. ; B&W catastrophic failure J.O. 4409/80-5; letters 6/16/80, inspections showed signs of excessive 6/22/80 heat on U-cups, carbides; significant details provided transients prior to this replacement:
in B&W letters;
- 1) 4/7/80 loss of all A.C., in which and ops, notes 3rd seal leakage doubled (to.59 gpm) from B. Garrison ano 2) intentional loss of seal inj.
test for NNI 09/17/80 Replaced seal cartridge J.O. 4411-80-8 J.0.
02/20/8f Replaced seal cartridge J.O. 3542 J.0.
l (1R4) elective maintenance 07/18/81 Replaced seal cartridge J.O. 6254 J.O.
installed seal cartridge modified in accordance with OCP 81-1051
^
Date Description Source Old seal inspection showed some heat QCSF-81-0761 checking on rotating faces, erosion on o.d. of carbon faces; rub on the shaf t sleeves in the upper stage area (also some i stage U-cup fretting) top of sleeve was black; upper stage more worn than others.
08/16/81 Removed new seal replaced with another Task Data seal (unmodified) - old seal failed due to 3rd seal failure; cavity vent valve leak effectively bypassed #2 stage; cavity press, decrease due to.7 gpm leak in #3 cav. press, sensing line.
Transient proceeding this and inspection details provided in B&W insp. notes 8/16/81 B&D pumps operated ~ 19 minutes w/-C80 isolated.
03/30/82 Overhaul PM-32A - due to higher vibrations J.0, 2480 indicated wiped 'brust bearing J.O. 2489 J.u. 22218 03/31/82 Removed seal cartridge J.0. 22702.,
J.0. 22742 11/17/82 Removed seal cartridge J.0. 35488, Task Data, J.O.
(IRS)
J.O. 35494 Seal cavity pressures indicated #2 and See Task Data
- 3 stage failures following main feed Operating notes pump transient which triggered an RCS transient - max. outflow - 10 gpm 05/25/83 Removed seal cartridge J.O. 50206 -
Task Data seal degraded after flow coastdown test Transient on B&O. #2 seal destaged - rotating Description; parts misaligned on sleeve, carbon PSIP notes fractures #3 stage; staging coil 07/13/83 blockage Residue on carbons analyzed by B&W See analysis No conclusions, results -
11/28/83 12/20/84 Changed out seal cartridge.
Elective J.O. 74901, (1R6) maintenance Task Data NUREG-1275 A-10
I Date Description Source 02/14/85 Changed out seal cartridge to remove-J.O. 81250 (1R6) upper sleeve 0-ring and vent adjusting-J.O. 81252 cap.
OCP-85-1010; rust colored water MPR Report-inside pump to motor spool piece.
03/01/85 1/18/86-Changed seal during shutdown Fall 1986 Changed out seal due to'RCP shaft iilspections during 1R7 (9/4 - 12/24) 2/20/87 Maintenance let the letdown isolation valve shut - x 10 minutes later B press.. lower cavity press decreased 125#'- came back up over 8. hours to normal - appears to have been something passing through P80.
8/17/87-3rd stage on B seal appeared to be gradually-degrading - turned out to be a sensing line-leak - crack was on the spoolpiece -- plant was shutdown - plant restarted - seal-working fine.
11/15/87 Upper cavity pressure decreased, leak increased due to crack in the same location as above entry, the plant *emoved the blank during the mid-cycle outage (10/16-11/2) and-the-pipe rewelded.
The evaluation by engineer said it was okay. The plant reduced power and installed a blank again. Long-term reduction-being pursued. (Such as flex-hose) 8/88
. Seal began leaking on x 8/16
.2 gpm gradually got worse over next 11 weeks.
On 8/27 leakage was approx. 3.5-4 gpm (not includ. CB0) -
reduced power and tried stopping the pump.--
leakage stabilized for approx. 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> then began-rising again - decision made to start refueling outage early and-shutdown unit.
(Unit shutdown was scheduled for 9/2) -
5.5gpm max. outflow - upper cavity pressure sensing line was isolated so little data available - see notebooks 8/27/88 Installed N-9000 seal 1R8 A-11 NUREG-1275
P32C t
Oate Oescription Source 11/19/73 Replaced anti-reverse clutch parts.
AP&L letter
(
11/1','73 2
03/22/74 Replaced seal cartridge J.O. 255 J.O.
07/22/76 Replaced seal cartridge press.
NRC letter oscillations, J.O. 1571 07/16/80 12/03/77 Replaced seal cartridge due to hig out RAC 77-26 leakage (5-6 gpm) J.0. 4406/77-3 01/01/79 Replaced seal cartridge elective NRC letter maintenance J.O. 4406/78-6' 07/16/80 Tech. Sem.
l 03/81 11/06/79 Replaced seal cartridge - erratic Tech. Sem. 03/81 performance-broken 2nd stage face; NRC letter J.0. 8301-79-3 07/16/80 Repair / check TE-6512 upper thrust J.O.
bearing temp. J.0: 8301-79-3 i
05/10/80 Catastrophic failure of seal with ANO 2803 leakage approaching 100 gpm - transient ANO 2226 description and reports are included-RAC 10084 in sources:
J.0. 4412-80-5 J.O. 4412C-80-5 this failure was preceded by a loss J.0. 4418C-80-5 of all AC on 04/17/80 and an intentional-J.O. 4418-80-5 loss of S.I. In 05/80 Fail. Invest.
05/16/80; Memo 05/23/80 J. 0. 1-6009-80-5 AP&L Letter-1-050 05/30; NRC Letter 06/02/80 NRC Letter 06/09/80 B&W Report 06/16/80.
Attach. I to-PSC-80-109; B&W Report 06/20/80 PER 1-80-02 BJ Letter 07/02/80 Memo GTA-185, 07/07/80; LER 80-015 Update, 07/17/80; NUREG-1275 A-12
Date Descriptg Source Failure Event Hotes -
AP&L Letter 1-070-23, 07/17/80;-
HPR Report 10/15/80 Memo ANO-80-5591 11/25/80; AP&L Letter IR-0481-05, 04/13/81; B&W letter 11/20/81 09/20/80 Re iced seal cartridge J.O. - 4408-80-8, J.O.
440;-80-8 RCP coupling to shaft repairs J.O.
J.O. 4407-80-8 01/08/81 Replaced seal cartridge to check J.O.
(IR4) dimensions and clearances J.O. 3544 08/26/81 Replaced seal cartridge af ter loss B&W Notes 08/16/81 of seal inj.
CCV: J.O. 15921; J.0. 15922 - iristalled seal modified in accord. with 81-1051.
08/09/82 Replace seal cartridge due to #3 seal J.0. Task Data failure during H/U - J.O. 31654; LER-82-21-3 AP&L 52 gpm max leakage, J.O. 32865, letter ICAN088214 J.0. 32604 shaft sleeve appeared swollen 06/08/83 Vibration 2.0-2.5 with 6 mil peaks J.0; J.O. 47332 LER 83-19-01; 09/06/83 Crack in lower stage vent line -
AP&L letter blank flange installed.
1.39 gpm leak.
ICAN098307 10/12-12/24/
Changed out seal cartridge J.O. 74896 J.O.
84 (1R6)
PM32C overhauled 01/12/85 Replaced seal cartridge due to improper Task Data, J.0.'s staging.
J.O.'s:
74657, 80037, 54707 01/16/85 Task vibration readings on P32C - J.O.
81062 (Procedure 3005.002)
^~I3 NUREG-1275
Date Description source 01/19/85 Replaced seal cartridge J.O. 80587 -
J 0. Task Data 3rd stage opened during operation; ICW transient J.0. 81067, J.0. 81862 - shaft sleeve expanded - several small pieces of i
RTV between seal faces in 2nd and 3rd stages.
)
l 01/31 -
C-seal cartridge - 3rd stage opened -
Task Data 02/09/85 removed cartridge 02/09/85 PH32C alignment J.0. 81232 (OP 1406.42)
J.0.
02/09/85 Installed seal cartridge - in P32C -
J.O.
J.O. 20006 - made modifications in accordance with DCP 85-1010 - removed upper shaft sleeve 0-ring, vent 5d adjusting cap 11/11/85 5311 - upper cavity and 1262# - lower cavity - no change - began degradit, following power decrease to 43% for MFP work. CB0 temp. on flow; drinking bid outflow.27 gpm - (slowly lost lower stage) disassembly found erosion (heavy) on feedback hold - charged out in Jan. 18, 86 during EfW outage fall 86 charged out seal due to RCP shaft inspections duiing 1R7.
8/1/87 Tripped the pump (not the plant) due to high temperature on the upper thrust bearing on the motor (approx. 310-320 degrees F.;.2 in/sec up to.6 in/sec on brg; coupling 6 mils to 17 mils in about 30 min. - problem due to loss of oil in motor - added 45 gallons of oil on 8/8/87 (15 gallons after low level alarm cleared) restarted thA Dump and temp. and press, were back t r rmal - no evidence of why oil was betrig lost - alarm came back in approx. 6 hrs after restart - continued to run the pump (added 10 more gallons) prior to alarm - 1800 - 8/12/87 tripped C again - run on 3 loop until the outage in Fall.
NUREG-1275 A -14 1
.....,,,-i
. -~.
date Description _
Source 10/16/.' -
Upper guide bearing was wiped, but not 10/30/87 thrust, bearing tightened and a lot of oil joints, no indication of leaks now, during recoupling couldn't get satisfactory TIR - finally using dimensions from post -
1R6 work, jacked the motor shaf t over for fit.
12/1/87 seal acting erratically - cavity prassures dropped to x 6 psi and 38 psi - no increase in CB0 flow or temp. - seal restaged momentarily then destaged again.
8/27/88 "C" seal began leaking over last week -
unit shut down began 8/27 - leakage was.5gpm max (EOL for med. seal) 8/27/88 Install N-9000 seal IR8 7/89 Indications that upper radial bearing wiped - worked during 11/89 1
I A-15 NUREG-1275
P32D Date Description Source 02/11/74 Replaced seal cartridge J.O.100 J.O.
03/22/74 Replaced seal cartridge J.O. 256 J.O.
03/18/75 Replaced cartridge due to excessive NRC letter 07/1G/80 leakage; 2 carbon rings broken Tect. Sem. 03/81 04/10/75 Replaced cartridge due to degraded Tech. Sem. 03/81 performance Fall. Invest.
05/16/80 07/02/76 Replaced seal a rtridge after 1st stage NRC letter 07/16/80 opened J.0. 1572 Tech. Sem. 03/81 08/16/76 Seal cartridge failed - - 25 gpm NRC letter 07/16/80 outleakage due to improper venting Tech. Sem. 03/81 (J.O. 1587)
RAC 76-22 11/25/77 Replaced seal cartridge due to high wear J.O. 4407-77-3 J.O. 4408-77-3 01/01/79 Replaced seal cartridge for elective NRC letter 07/16/80 maintenance J 0. 4404/78-6 01/21/80 Replaced f.eal cartridge J.0. 4403-80-1 for J.0.
elective maintenance NRC letter 07/16/80 05/16/80 Replaced seal cartri&;9 af ter catastrophic J.O.
tailure of
'C' pump.
J.0. 4413-80-5; 4419-80-5; 4419C-80-5 This failure preceded by loss of all A.C.
See B&W notes, in 04/80 and intentional loss of S.I. on notes from running pump in 05/80 Garrison Inspection data indicated high temperature See data '&W in all three stages rates 06/16/80, 06/20/80 12/01/80 Replaced cartridge due to high total Tech. Conf. 03/81 (IR4) outflow and press, oscillations;
-07/16/81 Replaced seal cartridge following loss J.O. 6260; B&W of seal inj. and ICW for ~ 19 min. (may seal insp. notes or may not have replaced this cartridge) 08/10/81 12/15/81 Removed seal cartridge J.O. 13528 J.O.
(actually.
J.O. 3527 J.O.
02/15/81)
NUREG-1275 A-16 l
l
i Date Description Source 03/25/82 Replace seal cartridge J.0. 22703; J.0.
J.O. 22743 installed modified seal per DCP-81-1051 01/h/83 Removed' seal cartridge J.0. 35495; J.0. Task Data (IR5)
J.O. 35498 middle stage gradually (op. & Insp.
degraded during operation.
Info.)
05/21/83 Replaced seal cartridge J.O. 49183 -
J.0.
blocked staging coil Task Data 08/18/83 High temperature on lower motor bearing; J.0. 47406
- v eventually decreased 08/11/84 4 '.n CB0 line weld RAC 1-84-221; RAC
'. 12580; EAR 84-307 J.O.
12/15/84 Replaced seal cartridge due to C80 weld J.O.
(IR6) crack; J.O. 74897 shaft sleeve undersized - Task Data J.O. 81278 02/14/85 Replaced seal cartridge after removing J.O.
(1RG) upper sleeve 0 ring and vent adjusting cap. - J.0. 81251 Modified in accordance w/DCP 85-1010 FCN-1 Fall 86 Replaced seal due to shaft inspections during IR7 8/27/88 -
Installed N-9000 seal 11/88 (IR8) 7/18/89 Vibration sys showed 1 mil peak to peak increase on pump probe - no shift in phase.
7/19/89 "D" pump stopped due to low lube oil levels (added 25 gpm temp. decreased but oil went straight through -
this bearing worked during 1M89_(4/89) 1/10/90 3rd stage gradually (decreased-to'306/ after12/89) - leakage'.62gpm shutdown from IM89 other stages o.k. - trends in chron file.
A-17 NUREG-1275 f
i ARKANSAS UNIT 2 2P32A i
Q11g Description Source 4/23/85 Replaced seal cartridge HPR report J.O. 76712 1
9/13/85 P32A failed - high oscillations, inspection notes i
rising 3 lower stages broken pictures, logs seal removed on 9/18/85 6/28/80 Replaced seal cart. (4th stage)
J.0. 2-4402 80 4 2 4403 80 4 2-4409 80 4 5/15/79 Replace seal, 4th stage J.0. 4416/79 '
sensing line leak RAC 2/1/80 Seal sleeve adjustment only J.O.4403/80-1 2/28/79 Removed spacer pieces for motor
- proced, work 5/14/79 Weld crack on seal sensing line RAC 80-7-0856 sce11/20/80_1tr 3/24/81 Overhaul motor J.0. 06932 4/81 Replaced cartridge J.0. 065153 5/5/82 Seal middle sensing line leak _
82-2-0432 LER50-268/82-
+
017/032-0 5/26/82 Remove seal J.0. 26755 6/22/82 Leak in seal upper cavity sens. line PEAR 82-0753 J.0. 32059 6/6/81 Replaced cartridge J.0. 6153 9/21/78 Replaced cartridge J.0. 4735 8/25/83 Weld break - middle press. sensing LER 83-039 line replaced seali J.0.~054733 l
9/30/81-Replaced cartridge-J.O. 15951' 2/22/79
_ "A" pump couldn't be rotated by-Hr. GEAP 790305-hand -upper-guide bearing wipea 071.
_3g 3
- 2. a-.... _ _, -.
[laig Description Source 7/86 Changed seal cartridge (2RS) 2/88 Changed seal cartridge - decision was (2R6) made to replace' all 4 during refueling as elective maintenance on A & 8, preventable on C & 0.
8/1/88 Lower cavity sensing line weld break -
t.ER caused rapid seal transient - probably broke all 4 seals - max outflow of 35-40 gpm - manually scramed Rx -
tripped pump excess flow - shutdown on vapor stage (scal torn down approx. 8/15) 10/89 Replaced all 4 seals - provided beefter DCR 89-2016 (2R7) stub out arrangement / flex hoses to pipe 2P328 9/10/79 Replaced seal due to excessive wear 4410/79-8 4405/79-8 4/9/80 Replaced seal - elective maint. after J.O.4402/80-2 loss 'of power 6/7/82 Seal repigced I.O.
69526 8/31/84 Replaced seal following Rx trip due 1.0. 63027 to dropped rod - lost lower stage 5/17/85
- B" failed - oscillations see chart and notes on failure in your' notebook 12/7/85 Lower 2 stages opened during a shutdown
' restaged later -
pump seal not changed.
7/86 Changed seal cartridge (2RS) 8/18/86 Changed seal cartridge (2R5) 2/88 Changed seal cartridge (2R6)
A-19 NUREG-1275
Orts Descriotion Equtc1 8/88 Changed seal cartridge during outage after 2P32A sensing line failure led trip - 4th stage on this pump was leaking x40 dpm changed to avoid later problems.
11/89 Charge out seal - beefier stubaits DCP 89-2016 flex hose 2P32C 6/5/82 Replaced cartridge J.0. 29045 7/21/84 Replaced cartridge J.0. 69527 4/19/84 Seal line failure RAC-2-84137 8/8/82 liigh out flow (28 gpm)
RAC-82154 RAC 82155 J.0. 27591 J.0. 27590 2/1/80 Replace seal due to excess leakage (rebuilt 4th stage)
J.0. 4202 9/12/85 Seal cartridge - oscillations causedbyIkC following 2 loss of cooling testing - causing events - replaced 9/14/85 inadvertent SIAS isolated CCW) 9/27/85 Replaced seal due to leak on on bottom 3/4 of(weld cracked pipe connection pipe)Vance has' picture of break.
4/13/87 Seal began oscillating approx. 80 psi overy 30 sec.; CCW adjustments didn't have an effect. d/21/87 - CCW adjustments were made, oscillations quit.
2/88 Replaced seal cartridge (2R6) 1/6/89 U)per seal sensing line was cracked at 11e seal
- shutdown and replaced cartridge 10/89 Charged out seal - beefter stubouts, flex DCP 89-2016 (2R7) hose NURl!G-1275 A-20
r 2P320 DAla Description So m 4/23-25/85 Remove seal cartridge 2/20/79 Replace seal (elective) 10/17/83 Replaced cartridge J.O. note (disassembly was videotaped -
47861 tapes held by D. Moore) 9/23/82 ISI test on RCP studs J.0. 34627 9/17/78 Replaced seal cartridge J.O. 4159 middle stage opened 7/86 Changed seal cartridge (2RS) 11/16/87 Slowly degrading - middle &
upper seal press. oscillations worsening & CB0 flow u) to 1.5 gpm -
tried CCW adjustments )ut to no avail.
2/38 Replaced seal cartridge 2R6 11/89 "D" pump motor maintenance - charged DCP 89-2016 out seal cartridge - beefier stubouts, flex hose A-21 NUREG-1275
WATDtFORD 3 RCP FAILURES IN DATE PUMP EVENT DESCRIPTION NOTES 10/6/82 1A Lower seal had zero delta-p when Changed after cold hydro checked during cold hydro.
4/1/83 18 Seal breakdown pressures dis? layed Changed seal after HFT 2A oscillation during precore li:T.
1/5/85 2A CB0 indicated 2.5 gpm at start of Changed out seal fill and vent.
Imediately 1/11/85 1A Water leakage around shaft at the Reinstalled properly with seal collar area prompted investi-changed bolts gation revealing the shaft holding bolts were improperly installed.
2/20/85 1A Operator error actuated containment Seals were changed IB isolation signal, securing CCW to immediately 2A the seals.
28 5/29/85 1A Plant conducted power ascension testing Seals changed out during 28 thru 80% loss-of-load trip and loss-of-turbine-generator lead AC testing.
Lower seal stages failed carbonate cicanup outage on restart.
7/15/85 2A CB0 and pressure oscillations increased Seal was changed during in frequency over two week period, the outage Readings were about to go offscale
(>3.5 gpm) when T-G vibration problem forced outage.
10/3/85 2B CB0 and staging pressure oscillations Seal was changed out increased over several weeks until immediately excess flow check valve closed.
Plant was forced to shutdown.
12/8/85 1A During recovery from trip, 28 seal 1 mediate seal changcout 10 failed to stage (all staging pressures on all four pumps to 2A caualized) and pressure breakdown provide " fresh start".
2B decreased on 2A, Some shaft leakage Total seal changcouts te occurred on 18.
date equal 171 12/1/86 1A Seals removed during routine refuel 1 Seals were replaced and IB maintenance to inspect and replace for old seal reworked to 2A next cycle. Wear patterns were signi-
" polish" the worn areas 2B ficant indicating tolerance problems and refine the tolerances or poor installation alignments.
to less restrictive dimensions. The tightness NUREG-1275 A-22
of fit of factory parts may be contributing to the failures. Many parts are received out of tolerance.
8/12/87
'28 Rotating baffle bc'.ts failed causing forced outage to inspect /
sli)page ~of the baffle on the heat repair baffle and replace excianger mounting surface.
PRE 87-xx all bolts. Torque values was generated to document and track and positioning of the
- solution, baffle onto the shaft lipare critical to ensure proper engagement of the pieces and a secure bolt tightness. Bolts were stretched and sheared.
10/17/87 2B CB0 and staging pressure: ramped up.
Seal was removed to inspect Eventually pressures approached again abnormal wear equalization and destaged near 1100 lbs.
patterns detected. New CB0 flow rate reached >2.8 gpm before N-900 seal installed (only shutdown.
one) to try out over the remaining months of cycle 2.
10/23/88 SG Noise emanating from the system prompted Inspection af ter shutdown shutdown for inspection to identify and revealed check valve correct the source. Suspected one of gagging pieces loose and the steam valves to the SGil or internal incomplete from contruc-SG loose parts contrcting the vessel tion WA to rework, interior surfaces.
Situation was corrected.
No evidence was found to indicate internal damage or loose parts inside the SG vessel, 11/7/88 1A New N-9000 seal model failed on startup B/J looked at possible mis-three times and all four were removed, alignment, poor axial Apparent CB0 psi and temps were higher play, shaft eccentricities than normal. Staging pressures looked and design tolerance good at low RCS pressures but seals errors. ANO also using destaged as psi increased to NOT/N0P.
with correct tolerances.
All four were swapped out with SU 9250 models.
Vibration data taken.
Inspections to follow.
12/7/88 1A Inspections of removed seals from 11/88 Byron Jackson to perform outage revealed massive damage to the extensive analysis and shaft s19 eve especially at the third return for report by 4/89.
stage were " bluing" burnthru was Redesign of N-scal to observed. Elastomers were destroyed follow with plan for f n places and rotating elements had reintroduction by RF3.
abnormal ' phono" grooving.
A-23 NUREG-1275
1/26/89 1A Valve stuck open and would not seat Suspect valve stem coupling without " banging" the valve stem is slipping with each tse.
closed several times.
possibic seat wear also hanging the valve osen.
Olaphram will also se inspected for malfunction at RF3 and tested for possible causes. Action plan issued by PM Engrg.
e a
N URFG-1275 Aq4
Calvert Cliffs Units 1 and 2
!TA. FCPLACDCNT H!sf 0RY Serviet Seat Replacernent Puno Date in Date out Time C4ttr10;e NtrCet No.
Serv 1Ct gf of Strv1Ce 2/
(Ho.) 3f No.
Replacement Reason 1
118 1/75 0/75 8
0440 Upper & wlasle Seal $tage rallurt 2
124 1/75 a/76 16 0439 Nortle veld railurt 3
114 1/75 12/76 24 0437 uwevn 4
12B 1/75 17/ 76 24 0438 Unano n 3
21e 12/76 gf a/77 5
0442 Unktown 6
118 8/75 1/78 30 0441 Unkno.n 7
120 3/77 1/78 11 0437 Ure/own g
113 4/78 5/78 2
0438 M1031e Seal railure & Veld Cracki*g 9
22A 12/76 gj 1/79.
26 0443 Upper Seal Stage rallurt 10 210 5/ 77 10/ 79 29 7072 Eacess Lcakage rlov 11 118 5/78 1/80 20 0441 Escess Leakage riow 12 12B 4/78 1/80 21 0439 Excess Leakage Flov 13 21A 12/76 4/
1/83 38 2758 Lower Seal rallure 14 1sA 3/ 77 10/80 34 04t.0 Escess Leakage riow 15 21A 1/ 83 1/81 12 0437 011 Damage 16 11B 2/00 7/81 18 0438 Lower Seal railure 17 11A 12/83 4/82 17 2758 Vater Seat railurt 18 12A 4/76 4/82 72 2757 Preventive Maintenance 19 128 2/83 9/82 32 7002 Lo.er Seal rallurt 20 12A 1/79 10/82 46 04a2 Upper Seal rallure 21 118 7/81 9/83 27 0439 Preventive Maintenance 22 21A 3/81 10/83 31 0441 vapor Seal Fallure 23 218 12 / 79 4/ B4 52 0443 Preventive Maintenance 24 22A 1/43 4/84 16 0440 Preventive Maintenance 23 223 12/76 gj 4/84 88 0444 Preventive Walnter.ance 26 2 18 6/84 6/B4 1
7002A Statt up rallurt c
27 21A 10 / 8 3 3/55 18 0442 Seal rallurt 28 11A 7/82 4/85 34 0438 Pttventive Walntenanct 1 29 118 11 / 8 3 4/85 18 2758 2 $tages ralled 33 120 9/82 4/85 32 2757 Unkrown 31 21A 3/85 10/85 8
044A Seal rallure 32 218 7/B4 10/85 16 7002 Prtventive Maintenance 5/
33 21A 12/85 12/85 1
7002A Start Up railurt 34 118 8/85 3/86 8
2757 Seal Stage rallures Notes:
1/ Date in service is cite of pints rtturn from outage where seal was installed.
2/ Cate out of service is date plant entered outage whert seal was replaced.
3/ Servict time 1rcitdes plant outage time if seal was rot worked on ciering outage, y Actual installaticn date is oncert:In. Power ascension date of 12/76 his tseen assuted.
5/ These preventive malrtenance chang outs are ktown to have ten for rottle eeld-related modificaticns.
St. Lucie Unit 1 G
y
.a s
I i.,k t),. {8 I
Es i
r; c
)
3 1'
v 3 a t
X y g' 3
1 C 3
5 i
- t. e t
O. a.
~
I ca 5 k$
h 5
E E
r 3
h 6 9 y
a 1
tb D
e jl.
I,:
i 6
,v *
=J.
11 I
W m
=
5 g
i y
st M
2 b
E E.
I.
E
!.E.
k.
9 E
ti
?
=
a
=
a t
6 1
E T.
s., w
- g. s ri is !
i ti t,a
- r I
~* 1 dg
,L ji
- I t 1
T2 3
(
t y
- 3. *.)
).
88 9
9 u
a
~
tr
}-<
a i ::
E a;
8. II
.] )
1 1 s
(r 3*
j "a IS 4
$g II
-8 S
O
..[
}
[*
11 r a 155 8! jj,IUI Il 3
g E
- ?
a:
a<
a:
LU O
5 E
E E.
i.
!. E.
I 5
~
J I
M
=
g a
a a
a 5
a=
39 31 4*
A A7
-Io 84 !w Iv to
$4 :=
9 g
<g 13 26 it ib to in it it 46 it 46 in 43 in a
i E
E NUREG-1275 A-26
St. Lucie Unit 2 9
lI i
gi 3,i o
v y
i m
3 a
L
\\1 L
l, 5
m 4
e e
e a
a
~
h (D"
E i
5 E
i j
l
=
32.i i
i n
3 e
I I
-t S
[ li x
li L H1
=
O a
et
!L y
m
=-
U e e y
{
A~t a
ig
.ij fl
! ?
u p
55 d
I. I 63 7
- aa
~
mIe Z
n'I R4' t.
.:n
- =.
e N
sg 9
I
!?
J,1 o
E
}
e l
x x
r.
I g
h
.I-
- E 4
e O
b I.
E.
E E
=
2 g
.J etW N
D I4 5"
E k' k7 Io Is kw kV kw Is 5
39 N
58 !$
$8 {b 7,8 ;$
j) j) ($
ag mg jg
{4g gg i
6
=
=
=
=
3 3
3
=
.J (nL
\\1 A NUREGil275.,
,._ e -
Pump Seal Instrument / Cooling Line Leaks NUREG-1275 A-28
l Events Involving Leakage From Instrument / Cooling Lines On Ilyron Jackson RCP Seals PLANT DATE DESCRIPTION ANO-2 8/1/SS Middle seal sensing line failure.
ANO-2 7/18/88 Inlet weld to 2RCP4026H cracked and leaked.
ANO-2 6/11/88 Crack in tubing 3" toward 21rf-6014 from 2RCP-6014B.
ANO-2 1/3/87 Tube fitting PS2RCP-1000A > Igpm leak, repaired 4/28/87.
ANO-2 8/24/86 Field weld #S was cracked near 2RCP-572.
tN O-2 12/27/85 Upper Seal Cavity Sensing Line Irak due to crack in weld down stream of 2RCP-6026A. Isolated leak & repaired 7/21/86.
ANO-2 4/19/84 A seat injection line at weld downstream of 2RCP-57C.
ANO-2 2/15/84 Broken upper seal cavity pressure line.
ANO-2 12/15/83 leak on flange at middle seal vent line.
He flange was tightened to stop the leak.
ANO-2 8/31/83 RCP pressure sensing line pin hole leak 2RCP-6016B. Orind out cracked weld & re-weld.
ANO-2 8/25/83 A leak in a weld of the middle seal for 2P-32A on the pressure sensing line.
ANO-2 5/18/83 A leak in the weld on the upper seal pressure sensing line of 1.61 gpm. He cause was due to vibration.
ANO-2 7/30/82 Seal cavity sensing line.
ANO-2 6/22/82 A leak in the upper seal sensing line. Downstream of 2RCP 60060.
A ANO-2 5/24/82 Sensing line crack.
ANO-2 5/5/d2 A leak in the middle seal pressure sensing line. Cause by a crack in the weld.
ANO-2 11/03/80 A leak in the pump delta-P transmitter sensing line. Caused by vibration induced fatigue due to binding in the helicoil.
ANO-2 9/17D9 Replace middle seal instrument root for 2RCP-6014B.
ANO-2 6/109 Seal flow upper flange leak. Tightened flange.
ANO-2 5/15!79 Found crack in fitting union on CCW flow indica: ion.
ANO-2 5/1409 A leak at the seal cavity pressure sensing line to 21'r-6006. A crack the weld caused stress cyclic fatigue.
ANO-1 1/18/85 Flange leak on seat injection line second flange in line from seal via MU-1008C.
ANO-1 1/15/85 RC-1054 A have leaks on instr. lines & RC-1054-C. Re fittings were tightened & the leak stopped.
ANO-1 9/6/83 A leak in the seal lower stage vent line. A blank flange was installed.
ANO-1 3/20/83 Leak on flange upstream of RC-1057-C.
ANO-1 1/12/82 ne leak was from loose flange bolts on the seal cavity pressure sensing line.
l A-29 NUREG-1275
PLANT DATE DESCRIPTION ANO-1 7/26/81 Repair wcld on downstream side of RilV-6580-11.
ANO-1 7/15/81 1.cak on l'F--6581 between RilV4580F & RilV-6580ll. Replace valve.
AFO-1 3/5/81 RilV-6580G fitting leak.
ANO-1 8/1808
'the leak was frorn the high pressure seal cavity sensing linc. 'lhe leak was isolated by valve R11V-6580ll.
ANO-1 2/3n8
'Ihe Icak was upstream of I'F-6578 at a weld.
ANO-1 1/2807
'Ihe leak was a pinhole leak on upstream pipe to valve wcld for valve R11V-7711.
ANO-1 6/2005 A crack in a socket wcld on the upstream side of valve 6510F.
ANO-1 12/28D4 A pin hole leak on the upstream pipe to valve wcld on valve RilV-7711.
Calvert Cliffs-1 7/86 Flex line fitting leaking (loose connection) replaced fitting.
Calvert Cliffs-1 5/85 Leak-off line-flex hose damaged and leaking-replaced ficx hose.
Calvert Cliffs-2 7/85 Flex line fitting leaking-replaced fitting.
Calvert Cliffs-1 10/85 111ced-off line cracked-welded.
Calvert Cliffs-1 6/85 1211 RCP-Scal leak-off line leaking ficx hose-replaced ficx hose.
Calvert Cliffs-1 11/84 Sensing line-pin hole leak Calvert Cliffs-2 7/84 111ced-off line (at scal housing) cracked.
Calvert Cliffs-2 11/83 Sensing line.
Calvert Cliffs 4/80 C110 line cracked.
Calvert Cliffs-2 109 Sensing line crack.
Calvert Cliffs-2 11/78-Sensing line cracks.
Calvert Cliffs-1 408 Sensind line cracks Calvert Cliffs-2 10/77 Sensing line cracks Calvert Cliffs-2 507 Sensing line cracks Fort Calhoun-1 209 CBO line crack Saint Lucie-1 10/87 Sensing line crack MUREG-1275 A-30
LER 368-11: Arkansas Unit 2 Ser.ing Line Failure
. A-31 NUREG-1275
kRC fors 366 U.S. Nuclear Regulatory Coevilssion (9 2,3)
Approved OH8 No. 3150 010,4 Capires: 8/31/85 LICEHSEE E V E H f R E P*O'A T (L E R)
TCILIIY NAME (1)
Arkansas Nuclear One sunitalvo IDOCALIhuh6ER(2)lP'ACl~(7) 1015101010131 bl 8tl10fl019 IIILE (4) Unisolatable Etactor Coolint,5ystes, Leak Caused By ylbration Induced Fallgue Failure of A Reactor Coolant Pum Seal Instrumentation L~l6s EVIhl DATE (5)
LT
~stR (61 EE FORi s DAR s (7) -
OIHER F ACIllIIES INVOLVED (8) i i
i 15equential Kevision Monthi Day l Year Year Number
'Ntimbe r Nonth 10ay_ tear faellity Names Docket Number (sl-i i
N/A O'T T T U Of 0 10 __910 N/A 0
5 0 D 0 ITRTl@0llI 15 5UE.HIIRD FUR 5UANLt 0,g)HbE[$] 8 iBL 8 0 81 0 1
~
81 8 81 81-H 01 11 1
' (kITlh
~
1RERENIF0F 10 CFR b M30[ (9)
I 11 (Check one or rore of the 'following f(ll)'-
M6DIl l _1 20~3M(b) i-25 U5(c.
ex 50.73 (iv) l _I 73.71(b)
~
50.'36cl1) 50.'36 'c'
$0.73 v) 1 73.71(c)
LEVill l I 20.405(a)(1)(1) 1 1
~ l 20.405(a (1)(11 l~;
(10) 1110101 g
.50.73 vil) l_l Other (Specify in i~i20.405(a(1(11))
1 7
50273 a 50.73 vill)(A)
Abstract below and w.
l l 20.405(a (1 (iv) 1 '50.'73ah
' ))
~
50.73(a)
(vill)(B) in Test, NRC form
~~
l l~l 20.405(a)(1 (v) 50 73 sh 1
450.73(a)
(x)
~
3C6=)
LICER5T. ' 1 lCT.f 0R.IHIS LER (12) haQe
,i Telephone Number
' Area l Larry A. Taylor. Nuclear Saf ety and Licensing'.5pecial.lst l Code i
~~
COMPLEIE OhE LlhE FOR EALH COMP 0 KENT f AILUEE DESCRIBED IN IHIS EEFORI (1)jl191614l-l3111010 15j0 l
I i
Reportable m
l lEeportablel i
Causel5ysteel tecnonent Manufacturer to NPR058 Cause System
. Component Hanufacturer to NPR05 I i
i i
l I I I I l i i 1 1
I I I I I I I
I I
i S&j l l i
l l I l l l
~l l
i i l i I l i
lEMi hi EEF0ETlXFLclkD (14)
LXPlGILD l Month Day Year I $UBH15510N I gi ves (if yes, comolete Espected Sut?m'I'ssion Date)WlflTNo s 1 OATE (15) l l
1.
I 4 I AsifiACT (Limit to 1400 spaces, i.e., approxim4tely.f t ' teen. single-space typewritten lines) (16)
On 8/1/88. a Notification of Unusual' Event',Wa's%Clared'and imanual reactor trip and plant cooldown performed following the f ailure of 4' reactor 3oola'nt!p p',(RCP) seal cavity pressure sensing instrument llee. The sensing Ilne failure, located at{thelfuslon ine'of,4 3/8 inch instrument tubing to 3/4 loch piping reducer socket weld, resulted infakeactor] Cool nt system (RCS) leak of approximately 20 gallons
~
4 per minute (gpe).
Thepumpshaftsealialso'Drate'.during'the'eventwasapproximately40spe.
degradedifollowing'the sensing line failure and contributed to addittenal RC5 leakage.
The maxlauniltal A high pressure safety injection pump was used,ln'additionitojnoretalicharging pump operations to maintain pressurlier level durIng the rapld,' Controlled *CooldownlT}he cause of the sensing Ilne failure was low
~
stress high cycle f atigue f ailure of the ' veld' tubing'systes configuraties ev8 Induced by vibration associated w v *tet8ve scticas included DCP seal piping and e-t
- W h *es9 ted in r
modifications Jon weld betterments to reduce susceptability of the systems to vibration induced stress.
tung term actions include evaluation of an improved RCP seal design incorporating the use of flealble aoses for process piping connect,lons and other features 'to alnimite the potential for f ailures of this t>pe and tne resulting leaks.
NUREG-1275 A-32
Fore 1062.018 NftC Foro 3G&A U.S. Euclear Regulatory Coualaston (9 83)
A proved OH8 No. 3150 0104 4
Empirest 8/31/85 t!CEN5d Ivrkt'kritRTitlFR)lTf1TICONTINUAi!0N FACILI1Y hAXL (1)
MCALIgbuB5Eil;(2),
~1TINWeill (6) l Fact (3) 15equeritiall liievisionl Arkansas Euclear One. Unit two t
Year Number Number 1 0l$1010[0 316f8E61._6-oT TT 1 ci olot?loriots it U (if more space is required, use additional,hiWtforst3b
,5 4. 17) 1.
Description of (vent A.
Plant Status At the tiee of occurrence of,this' event'.tArkan's'a's' Nuclear One Unit 2 (ANO 2) was in Mode 1 (Power Operation) with reactor /powerlat[ app'rcilsately'100 percent.
Reactor coolant system (RCS) temperature was 580 degrees Fahrenheit?jappRCS pressure was apprentaately 2250 psla.
and The unit had been operatir.g at full!powerjfor roxlattely 36 hours4.166667e-4 days <br />0.01 hours <br />5.952381e-5 weeks <br />1.3698e-5 months <br /> following startup af ter a saintenann outage unrelated to;th s' event B.
Component Identification This
- vent involves the failure"6 ~ eactor'
^
pressure sensing instrument line.' Theyal} Coolant pump (nP) 2P3?a middle seal cavityure* occurr
[18G] at the fusion line of a tub ng'to. piping. reducer sectet weld. The tubing / reducer assembly allows connection of.a pressureitransaltter (Pi] to c. 3/4 loch A5ME 5ection 111 Class !!! pipe connected to the alddle sealfcavity:of the shaf t seat for the RCP. the v
transmitter is used to provide remote; indication *of'aiddle seal cavity pressure in the control room. The transaltter.also'Drovides?aniinput. signal to the plant computer.
C.
Sequence of Events At approstaately 1635 on' 08/01/CS.'.a'socketNeld ~on'a pressure sensing line connected to the elddle seal cavity of the shaf t sea 13for'RCP'12P32A' developed a leat resulting in a small RC5
[ A81 leak to the contatreent atmosphefe'.ifAt 1'd5.9the* controlled bleedof f (680) low flow alare for 2P32A annunciated in the,contro1lrooe' alerting the control room operators to the problem. In response to the alarsatha~ operators monitored the control room indications of the RCP seal parameters and observed that;CBO. flow free the 2P32A pump seal had decreased f rom a normal valve of approximately)'one7callon'per, minute (gpe) to an indicated value of appronlaately.5 gpe.
Also, the. sea icavity' iagnosed as being a leak on one of the seal were starting to oscillate..The'probleelwas, pressures were noted to be abnormally low and d
cavity pressure sensing Ilnes % Actions!verel}lemediately inttlated to prepare for personnel to enter the containment building and isolate the'af fected line if possible. At approaleately 1654 the 2P32A CB0 hlch temperature'alars'annunctated indicating encessive temperature cf the water in the CB0 line..The 2P32A middle? seal cavity pressure had decreased to 0 psia and CB0 flow was indicating'approximatelyi'0 'gpe.V8ased on these indications and continued observation of the, seal parametersilt.was evidentithat the integrity of the RCP shaf t seal was degrading with time. Actlons hereltnitiated to perfore a controlled plant shutdo n.
Over the next few minutes the seal continued to' degrade and by 1700, pressure in the upper two shaf t seal cavities had increased to approximately RCS pressure. At 1702, the control rcos operat:rs manually tripped the reactor and main turbir.e in preparatten f:r n:Lrtr.;
the RCP. At 1703, 2P32A was stopped to lerevent further damage to the shaf t seal.
The emergency feedwater systea generator (5/C) water level resp [BA] actuated automatically on normal post trip steam onse and was Other plant systems responded normally to the trip,used to provide feedwater to the 5/G's.
and the plant was stabilized in hot standbg (Mode 3).
The RCS leak rate at this Llae was estimated to be appronlaately 20 goa. as a result, at 4725. a Notification of Unusal Event (NUE) was declared due to RCS leakage being greater than the Technical Specification allowable leakage of 10 gpe.
The htC =as notified of the manual reactor trip and declaration of a HUE.
An RC5 cooldown f rom the hot standby temperature of $45' Fahrenheit to a temperature of appr) leately $30' Fahrenheit was initiated to allow depressuritation of the RC5 in an attempt to decrease the leakage from the seal. - At 1754, a licensed operator and health physics teChniClan entered the Containment building to attempt to locate and isolate the leak. Af ter eatering the containment sthese personnel determined that access to the area of 2P32A was not possible due to the 3 mount of steaa present in the area from the leat, at 1800, the personnel enited the contaltwent building. At 1850, an RC5 cooldown to cold shutdown was initiated at the maalaus rate allowed by Technica Speelficatices, ie.
depronleately 100 degrees Fahrenheit per hour, lhe RC5 leak rate at this time had increased to a saitsua value of approntattely 40 spa A-33 NUREG-1275
fore 1062.018 (f.$. Nuclear Regulatory Coavelssion Approved OMS he. 3150 0104 (spires: 8/31/85
- m..,_
C tlCEN5(E EVERTTILP0MMER$EYTi omkm!ON TRILIIY KAME (1) rar T T. WHP.I R ( 2 ).
LER HUMBER (f0 l'EI'{
i 5equential i Eevislonl Arkansas Nuclear One Unit Two tear Number Humber 1 0l$10}00F3f41't ~llTTr7 FI Mr71- 0101310fl019 Ital (If more space is required. use additicnalW4 fora;,3 Ws)s(17).
At 1912, in order to c'odn'shMfo theiCoscined ?af fects on RCS inventory of the cooldown and sass loss free the RCS leak Ma;hig pressur safety;lnjection (HP51) pump (8J] was manually started and one motor operated HPs coldll njectlen valve was throttled open. The HP51 pump, with suction being supplied. rom th Irefue g.wateritank and injecting into an RCS cold leg.
was used in conjunction with three.c;eratiduring the reaalnder of the ooldovn.'tilhe' lant'ging pumps ts, maintain pr char cooldown was ctatinued and at 2141, the plant entered Mode 4 (Hot Shutdown)'with RC5] temperature tiest than 300' Fahrenheit ano RCS pressure at approtisately 330 psla.u At'2304
- personnel entered the contalraent building and were able to access the area around 2PJ2A.%lbess' personnel discovered that the middle seal cavity pressure sensing ilne for the pump,wasnevered downstreas of a manual isolation valve in the line. The o.anual isolation valve wasilosed toilsolate the leak path, however, at this time it was observed that RCS leakage.'alsoiexisted.fros'the upper seal area around the pump shaf t.
Following isolation of the broken"siddle?sealicavity' pressure sensing line, the RC5 leak rate decreased but still remained 'greaterjthan';10 'gpe.7.the RCS cooldown and depressurization was continued and at approximately 2348,Tthe' shutdown cooling system (BP] was placed in service for decay heat removal. ; Cold shutdown'(Kade.5) was reached at 0055 on 8/2/88. At 0144, with RCS pressure at 50 psia the RC5(leak'. rate'was' estimated to be approximately 12 cps. By 0535, the RC5 leak rate had decreasedsto',less'.than:10 gpa and the AVE was terminated. The RCS was depressurited to atmospheric corditicnsland' drained down in preparation for seal replacement and repair of the damaged senQgge) ~~
- 11. Event Cause A.
(vent Analysis The four ANO-2 Reactor Coolarit rundsTre"Dyren* Jackson, vertical shaf t, single suction, single
~
stage centrifugal pumps.. lhe punos are: located in a two loop RCS piping arrangement containing two pumps per loop, to prevent RCS traMge from around the pump shaf t, the pump is equipped with four series e.echanical seals,%I;e.Ma'. lower,' middle, upper and vapor seal (see figure 1),
three of the seals are used to contaln] reactor coolant pressurel the fourth, while rapable of eithstanding full systes pressureM, devices,as"a backup vapor seal.
s\\std Located in parallel with the RCP seals are pressure reducing
.'glhe is to divide RCS pressure equallyf a.nongIthe seals. purpose of these pressure reducing devices Q lhe mechanical seals are lubricated and cooled by a controlled one gpa reactor coolant' leakage, f.e.Inth located in the hydrostatic controlled bleedof f (CEO) flow.
Reactor coolant leakage enters'.the'sealJarea;through a labyr tearing area. Also located in this' area'.is a heat exchanger which cools the labyrinth area.
This beat enchanger is cooled by component cooling water. Once in the seal area, the reactor coolant leakage is picked up by,the sealiarea recirculation pump ar.d It discharged to the inner tubes of a concentric coil heat exchanger.'T<lhe seal area recirculattun pump has a flow rate of appronlaately 40 spa. The outer tutas of the concentric heat enchanger are supplied with co*punent cooling water. Once cooled, the seal recirculat!on water is used to cool the outer shell of the se ls and is returned to the seal recirculation pump. Approalmately 39 gallons of the 40 com goes through the recirculation system, as described above, and the other gallon is passed to the seals. Of the one spa that passes into the seal area, less than 11 passes through the seal faces for lubrication and more than 99% passes through the pressure reducing devices. This one see C60 flow is coIIected in an area above the third seal and routed to the Cheeical and Volume Control System [CB) (CYCS) volume control tank (VCT). The water it'then returned to tne RCS by the charging systes as part of normal RCS makeup. A small amount of leatage passes through the vapor seal stage to the reactor drain tank. CB0 flow f rom each CCP is displayed and recorded on chart recorders in the control room. Additional?y. other RCP seal parameters such as CB0 water temperature and seal cavity pressures are disp 1med and recorded. Control room alares (vlsual and audible annunication) are provided f or (80 low /high flow and high temperature conditions. The' seal parameter indications are obtained f rom flow, temnerature, and pressure instruments connected to the RCP shaf t seal by piping and tubing systems. The following piping systems are attached to each RCP.
Cavity Pressure Sensing Lines (upper and alddle cavity)
Controll:1 Bleed cf f Line (CBC) e 5eal Injer.tlon Line e
Component Cooling Vater (CCV) e Inner Casket teak off Detection Linr e
5eal vapor $tage leak of f Line NUREG-1275 A-34 g,
P foro 10C2.018 HRC fore 366A U.S. Nuclear Regulatory Co.witssion (9 33)
Approved OMO No. 3150*0104 Empires; 8/31/85 LIC[N$(Ei[YENilREPOR$fR$i[X6 CONTINUATION t t R huRH R (6)
FACILIIY hAHL (1)
D9QKLb HUMBLA.(2). _
i 1 5ecuentiall i Eevistor Arkansas Nuclear One. Unit Two
}
Year Number
_Nember 7
4, 1 8 mie Clife[0T0F31 oi-oi ii i TEU (if more space is required, use addtuona_) hGi[ ors;3o6Ar,pg7)
Each pump has separete pretsdreise'E's' hMAesgConnected to e.he upper seal cavity, el, cavity and CB0 line. The purposetof category I) is to allow the! operator thesellinesT(ASME section.!! class !!! piping, tQmonitor'RCP, seal perf ormance, ty reading the g
pressure in the two cavitiesand'CC0 IfneMthe. operators are able to assess the seal staging performance ty detersining how such' pressure'.ls, dropped across the seal stages. Downstream of normally open manual isolation! valves *.'I3/4;toch piping ettached directly to the seal assembly is welded to 3/8 inch 03 instrumentation' tubing which is then routed to the pressure sensing Instruments. If the pressure' sensing; Capability;is;1ost for either cf the cavities or bleedoff line. CEO flow and temperature!!ndications are'available to allow the cperator to alluresiin'the Cavity pressure using Ilnes resulting assess intearity of the sealistages%ow,and hlch'CB0 temperature (due to flow diversion f ree in leakage Is detected by low CB0ifl pressure sensing lines tap the{ng'.otheranstalled RCS leakage etetection systees. seal [Caylties*vla a the normal CB0 path) and by us Each of the breakdown devices. L'nder normalioperationWith RCS pressure at a noelnal value of 2250 psla, the elddle seal sensing Ifne 'should ste$8pproXlmately 147$ psla, while the Vpper cavity would be at approntmately 750 psla..: Temperature of. the reactor coolant in these seal areas is normally approxleately 130 degrees f ahrenhalt.'.
A review and analysis of the seal' performance data for 2P32A obtained from the control room chart recorder of the seal parametershln'e'approximately)10 minutes prior indlCates'that a small leak developed in the 2P32A middle seal cavity pressure sensing'1 the first centrol roca annunClator alare*(CBO. flow low which alerted the operators that a probles taisted. The leak apparently-started as 4'partist feljure of the 3/8 inch tubing to 3/4 loch pipe weld and eventually prepacated to'a' complete severance of the line approstaately 10 minutes af ter the init{al: leak.4/Several small RCP seal sensing ilne leaks had previously Cccurred at ANO*2 during' plant operation and the control room operators were f amiliar with the indications produced byfsuchileaks.. This past experience led to the preept and accurate diagnosis of the problem bysthe foperations staf f.
However, during prevleus occurrences of this type,'suf fictsntittee hat been available for personnel to enter the t
contaltwent building and isolate'theJeak prioriti any signiflcant seal damage occurring. A minleve time period of approximately.cne~hoursts.usually necessary eo prepare for and perform a containment entry with the plantLoperating'at power. Additionally, unilke this event,
~
previous sensing Ilne leaks have not resulted in complete Clrcumferential f ailures of the associated piping or tubing or prevented an order,1y, Controlled plant shutdown.
Following the cceplete severance of the alddle seal Cavity sensing line, the indicated seal parameters tecree entremely erratic,as pressures,' flow and temperature changed in verlous areas of the seal assembly. The shaf t' seal Capability degraded over the neat several minutes due to loss of normal cooling to the seal l stages until eventually the last two seal stages (upper and vapor seal) were Indicating approximately RCS pressure. Vith these Indications present, the operators followed the guldance Contained in the abnormal onerating troredure f or RCP seal f ailures and manually tripped the reactor and stopped the af facted RCP. The total 0C5 leak rate at this.tlee was estimated to be approximately 20 Spa and increasing.
1he leak rate subsequently increased to a samtaus value of appreslaately 40 gpa during the neat heur after securing the pump. The cause of this incteased leak rate af ter securing the pump was most likely due to continued degradation of the seal components and increased,
leakage past the vapor seal to the Containment building 4tsosphere.
Response of plant systees to the reactor trip were normal. An RCS cooldown and depressuritation to approslaately 530' Fahrenheit was laaediately initiated by dumping steam to the main concenter. The, so;down was momentarily terminated at this point to allow for boration of the RC5 to the required cold shutdown boron Concentration to ensure adequate reactor snutdown margin was maintained. Additionally, personnel were cosoleting preparations to enter the contatreent building to detersloe if the leak could be isolated. tipon completing the RC5 boration and determining that the area around 2P32A was inaccessible tne RC5 cooldown was recommenced.
A-35 NUREG-1275
fore 1062.01B hAC form 3t6A t!.5. Nuclear Regulatory Comission (9 83)
Approved OMS No. 3150 0104 Dpires: 6/31/65 L I C [N $[F.' [IEh1][f,0 R,Tj LE A))T[X f /CO NTI NUA T 10N IXEILITY HAME (1)
DMLhhuMSlit (2)
LER AUHBER d IFAGt(3I" lequential Kevisiolil Arkansas Nuclear One, Unit Two Year Number Number i Ol5'l'Ol010p31'6"8 ~ET I 01 11 1 I
ol01510F1019
~
IETT (If more space is required, use additional hkfors.36IA37[7)
Ove to contraction of the reactor. Coolant}du'[Ng the. rapid Cooldown of the RC5 in conjunction with the RC5 mass loss from the11eakhthe.use'ofja,hlgh pressure safety injection (HPSI) pump in combination with the three operating!charolog pur:ps became necessary to adequately salntain pressurizer level. One HPSI pump'vasjmanually' started and a motor operated RC5 cold leg injectionvalvewasthrottledcpenjto.pr the RC5 leak rate decreased as RCS press, ovide,mateup flow to the.RC5. During the cooldown, uretvas1 reduced. By 2300, with RC5 pressure at approximately 320 psla, the leak r' ate"vas*estlaated,to be approximately 20 gps. Following the containment building entry and lsolatton of,the severed alddle seal cavity pressure
(
sensing line, the RC5 leak rate decreased 'however,' personnel performing the evolution noted a significant amount of leakare existed past the pumps' vapor seal to atmosphere. The leak rate continued to decrease as RCS pressure was reduced and was eventually terminated when system pressure reached atmosphorlc and the RC5 vas dralned down below the seal area.
8.
Safety Significance This went represented a challenge to plant' safety equipment, operating procedures and especially to the Operations staf f of,the' unit /rsNo major malfunctions of necessary plant equipment occurred during or following the reactor trip from full power or during the subsequent rapid p! ant cooldown. 4The guidance provided by the procedures, both abnormal op; rating and esergency operating,' used during the event, was adequate to alnimite the consequences of the sensing line failure',and subsequent RCP seal degradation. Thejudgements and decisions made by the Operations personne1'during the event were appropriate and timely considering the circumstances. - Overa11,9these* factors cuntributed to successful miti,ation l
of the event without any threat to,the.bealth and safety of the public, it is also leportant to note that the seal senstng Ilne f ailure 'and subsequent seal degradation did not result in a loss of coolant accident (LOCA)iping and shaf t seals are addressed below.
as defined in 10CFR50.46. The potential safety concerns related to failures of RCP seal p A review of the design functions'of'the' RCP~shaf t' seal system as well as piping attached to the RCP5 (other than reactor coolant piping) Indicates that the most critical challenge to safet upon f ailure of a seal or piping is the potential for a toss of Coolant Accident (LOCA.
Artantas Nuclear One, Unit 7, has been~ fully analyzed for both small and large break LOCAs in accordance with 10CFR requirements and analysis shows that a large break LOCA (greater than 0.5 square feet) is the most 11alting break at ANO-2. Postulated breaks in the RCP seal related piping and tubing fall vell within the range of small break LOO.s. For example, if the two cavity pressure sensing'llnes,'the seal injection line, and the CB0 line on all four pumps were considered to f all catastrophically at once, the resulting total break area would be less than 0.01 square feet.
the Combustion Engineering Plant Analysis Code (CEPAC) was used to analyte the etfect of seal piping systems falling simultaneously on all four pumps (16 total lines),
these tour the ensuing transient resulted in a reactor trip on low margin to saturation temperature and actuation of safety injection flow. Following the initial depressurization, conditions stablilledafterapproalaate1{f(controlledbysteamgeneratorpressureatthe1000psiaset 20 minutes with pressurizer pressure near 1400 psia, RC5.
temperatura approalmately 545 point) and pressurlier level riear 25% and rising; Subsequent to this quasi equilibrium condition the operators would take manual control of the secondary steam bypass control system and initiate a controlled cool down to bring the plant to cold shutdown conditions.
Uo to three of the 16 Ilnes considered could f all with *he resulting leakage reaalaing within the mate *up capacity of the three coolant charging pumps normally available for RC5 eakeup, teakage within the capacity of the coolant charging pumps will not necessarily result in an automatic reactor trip, but is governed by Technical Specification leakage llelts. Overall the transient analysis results indlCate that the fallure of RCP seal related lines coes not result in conditions outside those previously analyzed for ANO-2.
NUlWG 4275 A-36
fore 1062.01B NRC form 3t&A U.$. Nuclear Regulatory Corvalssion (9 B3)
Approved OMS No. 3150 0104
[ spires: 8/31/05 LICLNSEE EVENT.RErop (Lt,R)llEqCONilNUATION IACILITY HAME (1)
DOMET$lkiBLR (2) 1 LER huMBLR (6) l FAGt (3)-
5equential Revision l i
Arkansas Nuctear One. Unit Two TUi (If more space is required. use addition 41shROform)66{3])_6]h.
__ Year Number.
Number 1 015101010 3 8
61 8 01 Il 1 01 0101610Fl019 (l
4 As a result of the seal sens'IEg'line'fallure*at'ANO 2 on 8/1/88, a concern also emerged that relates to the possibility,that a been introduced by the postulated fallu, decrease.in the targin to failure of a RCP seal has re?of'.the" sensing lines. To answer this concern, a discussion of those features which* provide {".this,sargin is needed.the seal cartridge consists of fou As previously stated.
RC5 pressure equally. The foun th7the vapor' stage, normally carries only the dif ferential pressure created by VCT backpressure.i![ach'of. these four stages is designed tn operate with t
full system pressure for a period of:"tise.J1'Past e3pertence has proven the seals capable of performing under these conditionsj Curing occurrenets of seal degradation,*ilfione of ths seals degrades, the :tr.er two stages wiil share the resulting loads.HIf.two,stals~ degrade, the third seal will pick up the dropped pressure. Only if all three lower. seats degrade to the point of being fully open or CB0 flow is blocked, will full syntes pressure be applied to the vapor stage. Additional seal features which aid in the sargin to' safety.to gross seal leakage are retalning rings, which tend to hold any damaged Carbon'5eallfaces,together.
Broken seal f aces will hold some pressure because the retalning rings tends to capture these pieces and keep tha free jassaing other, seal., faces or blocking the pressure breakdown devices.
During past events evees when seal facef.have'expertenced significant degradations the esulting seal leakage has been low. Additionally, the'1/4 loch orifice holes for the process piping connections provide a choke point for flow in'the event of a shear of an instrumentation line. Such as occurred during this' event.
.A choked flow calculation performed following the event concluded that the maalaum flow rate from a 1/4 inch ortflee would be approalmately 37.5 gpe.
In addition to the rugged design of-the=sealsreacn reactor coolant pump and its associated piping contains suf ficient instrumeetation for the operators to monitor performance and diagnose problems rapidly. The instrumentation,'plus other installed RC5 leakage detection systems are provided to alert the opera, tor.to seal salfunctions so that appropriate actions can be taken.
The corclusion reached Is that although the fallt$re of a sensing line or scae other attached ilne may cause some seal degradation, it does not present a significant reduction to the margin of safety. A degraded seal still performs its main function by limiting reactor coolant outflow.
C.
Root Cause f ollowing the event, a detailed Investigation ar'd review of the clpino assnClated =lth RCP seals was conducted with the objective of establishing the root cause of the f ailure associatto with this event and also to review the collective history of previous R(P seal piping tallures.
Based on these reviews it was concluded that both the piping to tubing weld f ailure of this event and previous piping failures were attributable to vibrational effects. Additional information relate. to these areas is provided below.
2P32A middle seal cavity pressure ser. sing line weld f allure on 8/1/88.
lhe Cause of the failure of the 2P32A siddle seal cavity pressure sensing line was determined to be icw stress, high cycle, fatigue failure of the weld materialJ induced by vibration associated with the operating reactor coolant pump. Investigations of the failure also showed that during a modiflCation Bade to this line in 1984. field personnel had installed a tubing to pipe connection with an elbow which deviated free standard details and was an unapproved method of routing instrumentation tubing for this application, This wts conside-ed to be a significant contributing factor to the failure.
A-37 h
=
Foro 1062.018 NRC Foro 366A U.S. Nuclear Rsgulatory Comelssion (9 83)
Approved OH8 No. 3150 0104 LIC[N5(E [YENT"RO0itiMOf TE(CONTINUAi!ON TICitI n hAME (1)
. DOCKET.', E ER.(2) I L[R huMMR (6) l FAGE (37'.
i Sequentiall lRevisioni 8
Arkansas Nuclear One. Unit Two
_Yeari I Number Ntst er l l'
1015J0]OTO 13 '61 er el al- -
01 H I 01 0101710FD H TLAI (If more space is required, use sooltional hRC. fora.h
.s ~(17)
Collective history of RCP seal,il@%l. lures,' ncluding occurrence of 8/1/88, d
the cause of RCP seal piping'fd1'uTesat'AN0 2'vas ceteritned to be that the initial desigi of the piping system and subsequent design modifications to the system following repeated sensing line f allures, relled heavily.on dead vflght, thermal and seismic considerettons and were inadequate to elf alnate vibratlon effects'en the piping system caused by the operating RCP. Although vibration ef fects on the piping' system were considered during evaluations of previous f ailures, the lack of rigorous In-depth vibrational design analyses which would in the design phases of a modiflCation." allow consideration of the combined ef fects of the piping, the supports and anchor points'was c.onsidered to be significant. It should be noted that regular, consistent vihration data has not been available in the past.
O.
Reportability This event is being reported per"the"re'quirements of 10CFR50.73(a)(2)(1)(A) as the completion of a plant shutdown required by the Technical Speelfications and also per 10CFR50.73(a)(2)(lv) as a condition that resulted in the manual actuation of an Engineered Safety features System (HPSI pump) and the Reactor Protection System (RPS). The NRC was nottfled of the manual RP5 actuation and declaration of Notification of Unusual Event per 10CFR50.72(b) (2)(11) and 10CFR50.72(a)(1)(1) at 1725 on 8/1/88.'
!!!. Corrective Actions A.
Imediate Following coepletion of the plant cooldown and cepressurustion, the RC5 was drained to the required level for RCP seal replacement. J A team of personnel was assembled to perform a detailed evaluation of the event. The primary objectives of this group were to:
' y.).a,y.
Perform a detailed engineering evaluation ~of the sensing Ilne failure to estabilsh root cause of the failure.
(.C. ? -
Review previous problems with RCP piping and tubing systems.
Develop any necessary modifications identitled as being required or desirable based on the results of the evaluations and Identificatlon of any necessary long term correctlve actions.
4 B.
Subsequent the 2PJ2A shaft seal was removed and replaced. wore..ctivities undertaken on the seal piping and tubing systems for all four RCPs consisted of system configuration reviews and restoration, modificatlons and weld betterments.
Configuration h v! w s and Restoration (ach piping system was walked down to validate that all applicable portions of the piping were configured as Indicated on approved design drawing and stress calculations. Additionally. any non standard tubing installations were reviewed and evaluated as ACCeCtable or modifled to reflect standard details. Traumatized piping was either corrected or replaced. Vibration isolators were reinstalled or refurbished as needed. Certeln welded fittings were replaced with mechanical connectors which have demanstrated improved performance in this application.
Vibration isolatton shlening and pipe support gaps were erified. Support bolts and clamps 9
were inspected and restored to specifications as necessary, l
Modifications J
The evaluations resulted In modif(cations being required on two 3/4-inch piping systems associated with seal injection lines to 2P32C. These modifications were completed prior to plant heatup.
Veld Betterment Hany of the f ailures of RCP seal piping and tubing have occurred at or near welded connectlons, NUREG-1275 A-38
Form 1062.018 NRC Form 366A U.S. Nuclear Regulatory Cosalssica (9 83)
Approved OHD No. 3150 0104 tapires: 8/31/85
. ~.,
LICENSEE [Y[NT REPOR,T).......lCXT, CONTINVATION LER)
EACitlIY hE(T)
DOCMT NUKiiEA (2),
liR NUHBER (6) h FAGE ( W trkansas Nuclear One. Unit Two
. Year
, Sequential l i
l 1 Revision Number Number
' 0151010 pl:3l_61 8 ~ 61 6 01 0101810F1019 of 11 1 ILxI (If more space is required, use additional hi4 form 3 bA's) (ll)
All piping systems were inspected for ' Weld failures. HDE was f,erformed on welds with f ailure histories er with configurattens known or suspected to be candidates for cracking. All welds with identified problems were restored to speelfication. A number of welds were improved uy use of weld overlay techniques. Particular attention was glven to weld surf ace finish and contours in an attempt to lower the stress intensification factor associated with these welds.
Since f ailures of welds have been observed to begin at the pipe 00, surface conditioning was improved.
In addition to these improvements, a post-asdlfication testing program was developed to Instrument selected lines with strain gages to determine system responses during heatup.
Yibration measurementi and visual coservations were made of certain piping systees to gather both quantitative and qualitative data for-valication of acceptable systee response f ollowing the work performed on the piping systems during the outage. This InforLation will also serve as design input data for any future modifications to the systes.
C.
Future As a result of the detailede ln depth evaluctions and corrective actions taken following the occurrence of this event, the root cause was identified ar.d corrected. Arkansas Power and tight Company is continuing to participate in an industry program established prior to this event which is oriented towards improvements in RCP seal design for use in Byron Jackson pumps. As part of the new seal design, the use of flexible hoses for connecting process sensing ilnes to the teal cartridge is being pursued. Reviews of design information and vibration test results indicate that connection fittinos using flexible hoses should be equal to or better than socket welded piping joints for this' application. ANO-1 is currently Installing the new seal design in the RCPs on that unit during the refueling outage in progress at this time. It is anticipated that the laproved seal will be installed on ANO 2 RCPs following evaluation of their use at AH0 1.
IV.
Additional Information A.
Similar (vents
$1milar Jvents at ANO related to RC$ leaks associated with RCP seal piping and tubing systems were previously reported in the following Licensee Event Reports (LERs).
ANO 2 ANO-1 368776~033 31377FD14 368/80 086 313/75-003 368/82-017 313/77-005 368/83 083 313/78 003 3be/63-039 313/78 021 313/82 001 313/83 019 8.
Supplemental Information
[nergy industry identification System (El!$) codes are identified in the text as (xx).
)
A-39 NUREG-1275
D NRC Foro 366A U.S. Nuclear Regulatory Comissico (9 83)
Approved OM8 No. 3150 0104 Lapires: 6/31/85 LIC(NSEE EVENT REPor* -(LER) TEXT CONTINUATION FACILITY AAME (1) 100CLET HUMBER (2) l LER N M ER (6) l PAGE (3) l l
l l5equentiall lRevisioni Arbantes Nuclear One, Unit Two l
l_Yearl Number l Nyber l 1015101010l'31 61 81 81 81--
of 11 1
--I of 0101910fl019 TUI (if more space is required, use additional hAC f ora.166A's) (17)
FIGURE 1 REACTOR COOLNH PUMP 2P32A rw
".. +.......
p...........
- N-N40
- 7 g
u.w s s :
i; 6
> g:
p gj he.,,
gx a.2 e egii a,
e_ e A
- - ~ ~ - >
1 I
g U
l
-( Aa',
[m
...e....1
- =
.g a m bt-F { @ > M
- *-C-9
{~
,g g g
g
~
_m
-H-H n - >
. H. -H H
I gg;
_pgl
=
I 1
A= (H-mm
- *->4
>4
' 'y,;l;D 2
O N
c,/
s_
s
" = - - - ' j'
\\f I
NUREG-1275 A-40
APPENDIX B Seal Failure Data Prior to 1985
Table 11-1 lists a number of the more significnt RCP seal failures at operating U.S. pressurized wat.:r reactors (PWRs) prior to January 1985.
Tabic B-1 Significant Seal Failure Events Event Estimated Volume Plant Date ofleak/ leakage Rate Oconee 2 0104 50,000 gal /90 gpm Robinscm 2 0505 200,000 gal /300 - 500 gpm Indian Point 2 07n7 90,000 galO5 gpm lladdam Neck 08D7 4,200 gal /N.A.
Salem 1 10n8 15,000 gal /35 gpm Arkansas 1 05/S0 60,000 gal /(200 - 300) gpn.
Arkansas 2
~,.-,.-
' '8 8 N.A./40 com c
4 Table B-2 lists component cooling water transient-initiated SU seal failures.
Table 11-2 Component Cooling Water Transient-Intiated SU Seal Failures Prior to January 1985 No. of Pumps Plant Date Affected Comments St. Lucie 1 407 Three Ixss of instrument air led to loss of CCW.
St. Lucie 1 -
6/80 Four DC ground caused loss of CCW.
St. Lucie 2 12/84 Two Power loss to CCW isolation valve.
St, Lucie 2 8/85 Four Incorrect fuse size led to loss of CCW.
Arkansas 1 1169 Two less of ac power.
Arkansas 1 5/80 Four A loss of offiste power occurred on 4D/80.
Arkansas 1 7/81 Two Imss of Injection and ICW for 19 minutes.
Arkansas 1 8/81 Two Imss of ac power.
11 - 1 NUREG-1275
l Table B-3 identifies instances reported prior to January 1985 in which plants continued to operate with failed seal stages.
l Table 11-3 Continued Operation With Failed Seal Stages 4
5
Discovery Days Plant Typ.
Date Operating Comments i
From NPRDS Data:
Plant 1 oU 01/14/84 ~
2 Upper Seal tracked for several days.
Plant 2 SU 07/25/84 17 Continued operating with sensing line leakage.
Calvert Cliffs SU 1206 21 SU 04U7 7
4 SU 0908 1555 SU 10n9 11 SU 01/81 180 Arkansas 1 (Pump P32A) SU 0809 19 SU 04/80 33 SU 08/80 2
SU 10/84 37 1
1 NUREG-1275 11 - 2
NRC FORM 335 U.S. NUCLEAR REGULATCR COMMISSION
- v. REPORT NUMBER 12-69)
(Assigned t>y NRC. Add Vol..
NRCM 1102 Suppm. Rev., and Addendum Num-32cn ma BIBLIOGRAPHIC DATA SHEET t* - " *"Y 1 (see instructions on in. re.orse'
- 2. TITLE AND Sut3 TITLE
- NUREG-1275, Vol. 7
- 3. DATE ftEPORT PUBUSHED Operating Experience Feedback Report - Experience with Pump Seals Installed in l
uONm vEAR Reactor Coolant Pumps Manufactured by Byron Jackson September 1992 Commercial Power Reactors
- 6. TYPE OF REPORT
- 1. G. Uc!!, P. D. O'Reilly Technical
- 7. PERIOD COVERED (inctuOo Ortes) 1/1/84 - 1/1/90 s
- 0. PERFORMING ORGANIZATION - NAME AND ADORESS (if NRC, provide DMalon, Ott6ce or Region. U.S. Nuclear Regulatory Commission, and mamng address; 11 contractor, provide rarne and malling address.)
Division of Safety Programs Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulatory Commission Washington, DC 20555 9, SPONSORING ORGANIZ ATION - NAM. AND ADORESS (11 NRC, type 'Same as above"; 11 contractor, provide NRC Dtvlsion Office or Hegion, U.S. Nuclear Regulatory Commission, and rnalling ad&oss.)
.r.
rts 2
. is or sess)
- l nu,,c, art examines tne reactor coolani pump (RCP) seal operating expericnce through August 1990 at plants with Byron Jackson (B-J) RCPs. The operating experience examined in this analysis included a review of the practice of continuing operation with a degraded seal.
Plants with B.J RCPs that have had relative y good experience with their RCP seals attribute this success to a combi-nation of different factors, including: enh:.nced seal QA efforts, modified /new seal designs, improved maintenance procedures and training, attention to detail, improved seal operating procedures, knowledgeable personnel involved in seal maintenance and operation, reduction in frequency of transients that stress the seals, seal handling and installa-tion equipment designed to the appropriate precision, and maintenance of a clean seal cooling water system. As more plants have implemented corrective measures such as these, the number of B-J RCP seal failures experienced has tended to decrease.
This study includrd a review of the practice of continued operation with a degraded seal in the case of PWR plants with Byron Jackson reactor coolant pumps. Specific factors were identified which should be addressed in order to safety manage operation of a reactor coolant pump with indications of a degrading scal.
- 12. KEY WOROS/DESCRIPTORS (List words or phrases that will assist researchers in locating the report.)
- 13. AVAA.ABluTY JTATEMENT Unlimited Reactor Cool". a Pumps
- 14. SECURlW CLASSFICANN Reactor Cociant Pump Seals (This Page)
Pressurir.: Water Reactors UnclasMfbd (This Report)
Unclassified
- 15. NUMBER OF PAGES
' S. PRICE NRC FORM 335 (2-69)
s Printed
- on recycled paper Federal Recycling Program
7 ~NUREG-1275, Vol if -
~
OPERATING EXPERIENCE FEEDBACK REPORT-INTGWO[EKarumtUtMuiPSUa13-~~ TRFiTTEEMEliEM"'
INSTALLED IN REACTOR COOLANT PUMPS MANUFACTURED BY BYRON JACKSON..
UNITED STATES SPECIAL FOURW-CLASS RAM NUCt. EAR REGULATORY COMMISSION '
POSTAGE AND FEES PAID
. WASHINGTON, D.C. 20555-0001 usuRe PERMIT NO. G 6?
OFFICIAL BUSINESS PENALTY FOR PRIVATE LF" $300 120555119531 5
14.IllCllF US N9C-of,gy
-phf ghp[rJRLICATIONS SVCS
- -211 WASHI w ny a
i-2
'I
+=
v
-i -,.,,
,--r 3
-m,
,,.,...y
]
'