Atomic Energy of Canada Limited STEAM GENERATOR … · r "r, ;.1~- Atomic Energy of Canada Limited STEAM GENERATOR TUBE FAILURES: WORLD EXPERIENCE IN WATER-COOLED NUCLEAR POWER REACTORS

r " r , ;.1~-

Atomic Energy of Canada Limited

STEAM GENERATOR TUBE FAILURES:

WORLD EXPERIENCE IN WATER-COOLED NUCLEAR

POWER REACTORS DURING 1972

by

P.D. STEVENS-GUILLE

Chalk River Nuclear Laboratories

Chalk River,.Ontario

March' 1^74 - l<'

AECL-4753

r>."'f i

STEAM GENERATOR TUBE FAILURES:

WORLD EXPERIENCE IN WATER-COOLED NUCLEARPOWER REACTORS DURING 1972

by

P.D.Stevens-Guille

ABSTRACT

During 1972, approximately 1 in 3 operating reactors with szeam generators incurredtube failures. They predominated near the tube sheet and in the bend region. Variousforms of corrosion were the most frequent cause of failure. Eddy current inspection wasthe preferred method for locating and investigating the cause cf failure. Extensive use wasmade of both mechanical and explosive plugs for lepair. As a class,steam generators wilhMonel-400 tubes had the lowest failure rates and those of Inconel-600, the highest.

Chalk River Nuclear Laboratories,Chalk River, Ontario.

March 1974

AECL-4753

Rupture de tube Jans les ge"ne"rateurs de vapeur:

Experience mondiale dans les re"aeteurs de puissance

refroidis par eau en 1972

par

P.D. Stevens-Guille

Resume

En 1972, environ 1 sur 3 r-?ac!"urs en service avec des generateurs de vapeur a subi desruptures de tube. Les ruptures se sont produites pr5s de la plaque tubulaire et dans lapartie cintre'e. Diverses formes de corrosion ont ete la cause la plus frequente de cesruptures. L'inspection par courants de Foucault a ete la methode preferee pour localiserles ruptures et determiner leur cause. Pour les reparations, on a souvent employe deshouchons mecaniques et explosifs. Les generatuiu-? de vapeur ayant des lubes enMonel-400 avaient los plus has taux de rupture tandis q\ic ccux en lnconel-f-OO avaienl lesplus hauls taux.

L'Hnergie Atomique du Canada, LimiteeLaboratories Nucleaires de Chalk River

Chalk River, Ontario

Mars 1974

AECL4753

TABLE OF CONTENTS

1. Growth of Reactor Operating Experience 1

2. Survey of Tube Failures in 1972 1

3. Location and Cause of Failure s

4. Leak Lontion, Inspection and Repair d

5. Tube Failure Rates d

b. Summary 10

Acknowledgement 10

References II

TABLES

1. Reactors with Steam Generator Tube Failures in ll>72 2

2. Cause of Steam Generator Failure in ll>72 5

3. Leak Location, Inspection and Repair of Steam Generators in ll»72 7

4. Stainless Steel Tubed Si cam Generator Experience in Jan. I . ll>7.' s

5. Monel-400 Tubed Si cam Generator Experience lo Jan. I . ll»7.i N

6. lneonel-600 Tubed Sieum Generator Experience lo Jan. 1, I ')73 l )

7. Occurrence of lneonel-600 Tube Failures 10

FIGURES

1. Growth of Nuclear Reactors

2. Failure Regions in Ineonel-600 Tubed Steam Generators Dining l')72

STEAM GENERATOR TUBE FAILURES:

WORLD EXPERIENCE IN WATER-COOLED NUCLEAR

POWER REACTORS DURING 1972

1. GROWTH OF REACTOR OPERATINGEXPERIENCE

At the end of 1972 there were approximately 130nuclear power reactors in operation in the world. Theinstallation of power reactors is expected to growrapidly in the next few decades. One current listshows that at the beginning of 1973 over 300 reactorswere either under construction or in the design stage[ l l . Figure 1 taken from an International AtomicEnergy Agency directory [2| shows the predictednumber of reactors installed in the 24-year period,1954 to 19 7 8 . An exponential curve fitted to thesedata has a doubling time of only 3.7 years.

The area under this curve is the product ofreactors and y jars, i.e., operating time. A property ofthe exponential curve is that the area under it in anydoubling period is equal to the area from minusinfinity to the beginning of the period. Thus in thenext four years new operating experience will exceedthat accumulated since the first power reactorentered operation.

Already a formidable amount of operating ex-perience exists. Lessons, from it can lead to sub-stantial improvements in capital and operating costsof future plants poviding that useful information canbe extracted and nade available to reactor designers.This report is one attempt to summarize operatingexperience in a specialized but important area, that ofsteam generator tube failures in water-cooled nuclearreactors. It follows a previous report by the author onexperience to the end of 1971 [3] , and shouldpreferably be read in context with it.

Steam generator tube failures have continued tocause outages and lost production in 1972 as well ascausing high radiation doses to inspection and repaircrews. This report lists operating reactor statistics bytube material to highlight significant trends in failurerates.

2. SURVEY OF TUBE FAILURES IN 1972

By the end of 1972,4! water-cooled reactors withsteam generators were in operation. Table 1 sum-marizes a search of the published literature on tube

failures and a survey of reactor operators made by theauthor. Experimental and shutdown reactors areexcluded from the Table as their design and operationmay differ significantly from current practice. Afailed tubed is defined as one that ceases tu performits function because of either leaks or defects whichmay cause leaks in a short lime.

Stainless Steel Tube Failures

Failures occured in 5 of l> reactors with slainlcsssteel lubes. The largest numbers of failures occurredin Indian Point-1, Dresden-1 and K.WL.

Indian Point-1, in the USA, has had a long his'oryof stress corrosion failures. Approximately 36 lubesfailed in 1972 bringing the total failures to about i 29[4] . The reactor can operate on 3 of 4 steamgenerators, and one is frequently valved out awaitinga repair shutdown. Extensive eddy current inspection

400

<ITUJ

300r-

CO

rr200

LLJDC

§a.

sDO

100

INSTALLATION OFPOWER REACTORS

EXPONENTIAL CURVEFIT, DOUBLING TIME \3.7 YEARS

54 58 62 66 70YEARS

Figure 1 - Growth of Nuclear Reactors

78

___ 1

REACTORS WITH STEAM GENERATOR TUBE FAILURES IN 1972

Reactor, Type"Power |MW(e)j

STEAM GENERATORS (SG)

No./RcactorManufacturer

Type, Size, TubeOperating Conditions

Tube Size,Material Tube Failuies Cause of Defects, Remarks

N-Reanor (NFK)LWC'.R800II.S.A.

12CombustionEngineering Co.

Horizontal, U-tube in shell1116 tubes 1486m2

10.3 MN/m , 277°C(1500 psig, S39°F)

15.9 mmOD(5/8 in.)1.5 mm thick(0.06 in.)2 SG's tubedS. Steel - 30410 SG's tubedlnf:onel-600

I Cause of failure unknown. No failuresin S. Steel SG occured in any of the 10 . nconel-fiOO

tubed SG's.

Indian Poinl - 1 4 Horizontal, U-tubeJn shell 25.4 m m ODI'WR Babcocki 811 tubes, 1280m* (I in.)2(.S Wikox 10.3 MN/m ,270 C S. Steel - 304U.S.A. (1500 psis, 519 V)

Stress corrosion cracking in ail '' SG's.Sensitization which occurred during stressrelief or shell welds probably accentuatedfailures. Some leaks occured at U-bendregion. Total shutdown for repairs 'V 2months.

Dresden - 1IIW200II.S.A.

KWL LingrnBWR256l-ed. R. Germany

ti;;rii:liutioHVVH150Italy

Foster Wheeler

Atlas-WerkeAC

KnninglijkeMaehinefabriekStork (NL)

Vertical, U-tube in shell1801 tubeij. 605 m7.0 MN/m",285 c.(1020 psia, 54b i )

Vertical, tube in shell5000 tubesm 2 360 m2

7.3 MN/m , 261°C(IOSC psig, 5O2°F)

Verlical, U-tube in uU1785 lubes7.0 MN/m , 277°C(1015 psig, 531°1)

15.9 mm OD(5/B in.)S. Steel - .104 I

23 mm ODS. Steel

1 '> mm 1)1)Monel-400

12 inSG"B"

'V'lin 1972t^tal 1 1 540inSG "A"75 inSG "U"

Cause of failure unknown.

Cracks in 8 condensate tubes by ovcrstress-ing. Repaired by hotting and seal welding;^ 100 tubes plugged as they would other-wise become inaccessible after repairs.Repairs took 7 months.

Caused failure not known. Failures occurredin centre of bundle on both inlet and outletsides.

Be/nan - 1PWH Weftinjjhouse

t . e i t n : Corp.

Vertical, U-tuhe in shell2604 tubes, 6195m15.5 MN/m2. .IJ5°C(2250 psig. 5">9°F)

22.2 mm OO 'V'S.lOin 1972 Second hatch of repairs started in Dec. 1971,(7/H in.) Total 952 and completed in May 1972. Failures from1.2 rnm thiel* secondary side; tntergranular caustic cracking(0.050 in.) occurred mainly at inlet end in tube sheetInconel-600 crevices in both SG's. During explosive plug-

ging hi£h chloride concentrations detected >nSG safe ends from primer cord. Dye penetrantinspection revealed I crack in shell, repairedby welding. Power limited by SG's to 75';f.

M«n:mv.i lI'WK.11(1

Japan

1 nmlFnri-

2lltsllivori

nn

nisC...

Vertical. U-442(> tubes! 5.4 MN/m(22.1.1 p ig .

lube in si

2 . 322°C<.io"i)

Inconel-nOO 1 10 plugged in 1972:1 leaker9° with delects10 luhes removed^ 2 0 0 0 pluggedin 197.1

Defect* utcurrtd in bend regions, SOPIC indicat-ed severe wall thinning 'Miner from corrosion«r erosion, hut tm iritcrgranular attack.Repairs took 5 months.

I'oint Hen. l> 1I'WR

U.S.A.

KWO (Ohricjicim)I'WR12Sl-'d. K. Germany

Shippinpport -(Core 2)I'WR«0I VS. A.

11.11. HohitisonI'WK-110r.s.-v.

(tilehofrnlitig-'.hulteSlerkradf AC.:inil llulck AG

2 Hnhcockand Wilcox2 - FoslerWheeler Co.

.!VSfstinphouseIlectric Corp,

Verlical I,1-tube in shell-12i.o lubes.412Km3

15.4 MN/m2. .1I»°C

Vertical, ll-lube in shell2M)7 lubes. 27S0m2

14.7 MN/m , 310°l'(2 130 psig. 590 ° l )

Band W: U-tube in shell921 tubes, 845 m3

F.W.: straight lube20*10 tubes, 771 ni2

I3.R MN/m2,2RI°l"(3000 psiu, 538°F)

Verlical. U-lube in shell32M) luhes, 41 2H m15.4 MN/m2. .1I7°C(223.1 p«ig.(>04°F)

22.2 mm Ol)(7/8 m.)1.2 mm thicktO.OS'iin.)Inconel-600

22 mm OH(7/8 in.)1.2 mm thickInconel-600

S. Slwl - 304

22.2 mm OPO/Rin.)1.1 mm thick(0.050 in.)lncnnt.*[-600

During 1972:10! in SU "A"92 in SG "B",and plugged in 1973Tolal 193

67 in 1972:50 in SC. 117 in SCJ 2

.12 in 1472:2t.inSli "A"4 in SG "U"2 in SG "C"

Average primary to secondary leal1: rate i 001/dav in 1972. All defects 25 in 125 mmahovu lube shell in -entre of hoc 'eg. Detectsranged from 65 to 95'^ nf wall thickness.Caustic stress corrosion crac king suspected inlube failures. Tube sheet cladding and tuherepaired in 5 month shutdown.

Failures in SG I occured ^ SO mm ahnvetube sheet, caused by intergrunular crackingfrom secondary side. Failures in SG 2 occur-ed in region of smallest bend radius, cause nutknown. Primary side cracks observed in bothSG's where lubes rolled into tube sheet.

Failures all occured in SG M1A". Cause offailure not available. SG continued to leak6 l/h after return to service.

Intergranulur corrosion in all 3 SG'sprnably caused by caustic concentrationson tuhe sheet. Leak discovered 0 - 1 0 0 mmabove tube sheet. One explosive plug failed[o bond and was repaired by seal welding.

(Gon tin tied)

Table 1 (Continued)

Reactor, TypePower I MVV(e) |

San OnolreI'WK4.1UU.S.A.

Haddam NeckI'WR575U.S.A.

STEAM GENERATORS 1SG)

No./ReactorManufaclvner

Type, Size, TubeOperating Conditions

Tube Size,Material Tube Failures

WeslinglluuseElectric Carp.

WeslinghouseKiectric Corp.

Vertical, U-lube in shell.1794 tubes, 3 5 7 J i !

14.5 MN/rn ,304°C(2100 psia, 571°l )

Vertical, ll-lube in shell3714 tubes 257.1m2

IJ.H MN/m , M6°l'(2000 psig. 56S°I)

I'I mm I HICiin.)Incnnel-600

l'» mm OL)('a in.)Inconel nun

!» in )l>7215 in I 'm

14 in 1*17 2

in SC: " . I "

Cause of Defects, Remarks

(."JIUKI* nut known, hailuru* at firsi hnfflcElate und U-hvnil region. Kvpairs lunk 1 Sdays tola] down lime. E->.icnsivc cvUWcuircnf inspcchnti rcvfak'ii snme tollingin all Stl's at ;mti-vjhr;ition bars and tlrsltuliL" supports.

lnlL-rt;raiiularHlri.'ss nirrosiun stisprt ii'if

:ll WGK I.iixht-water-cooled, graphite-modenMe.l reactorI'WH Pressurued'lignl-water-mptierated and cooled reactorHWP Hoiliuc-light-watcr-cooled and moderated reactor1'HVVK Pressuri /ed-heavy-water-rnoderated and cooled reactor

has been done; in fact, some steam generators havebeen inspected more than once. A remotely operatedinspection machine is used so that inspectors canwithdraw to a low radiation field area for the dura-tion of the tests.

Dresden-1, a US boiling water reactor with 4 steamgenerators had 12 tube failures in 1972. Althoughthis type of reactor is no longer built, operatingconditions in both the primary and secondarysystems are comparable with other plants of the samevintage, and i! is included here for this reason.

During a pressure test of the steam generators inKWL (Lingen) in the Federal Republic of Germany,cracks were found near the weld between the bottomplate and condensate tubes, caused by overstressing.Special machining and grinding to repair 8 condensatetubes and the attachment of new tubes bolted to thebottom plate caused a 400 mar.-rem dose to therepair crew. This is one of the largest man-rem doseson record in a single steam generator repair [5,6].

Monel-400* Tube Failures

Tube failuie occurred in only one of 7 reactorswith Monel-400 tubes. Garigliano, an Italian boilingwater reactor, had 7 tube failures in 1972. The causeof failure is not known. The remaining reactors withMonel-400 tubes are alt of the C ANDU** type. Theyhave an outstanding record of failure-free operation;only one tube failure has occurred in nearly 6reactor-years of operation.

*Monel-400, Inconel-600 and lncoloy-800 are trademarks of the International Nickel Co.

**CANDU - CANada Deuterium Uranium

Inconel-600* Tube Failures

Eight of 21 reactors with inconel-hOO aibes failedin 1972. The largest numbers of failures occurred inBeznau-1 in Switzerland, Mihania-1 in Japan andPoint Beach-1 in the USA.

Failures in Beznau-1 were due to corrosion. Thereactor was operated during 1470-71 with v.cro solidschemistry treatment | 7 ] , The first hatch of failuresoccurred in 1971 and was caused by intergranularcracking at the tube sheet crevices. Investigationsshowed that the zero solids treatment provided nobuffeiing of hardness impurities brought into thesteam generator from condensor in-leakagev During1971 phosphate was added intermittently, bu( it wasnot sufficient to neutralize the continued in-leakagcof river water via the condensor and high levels offree caustic existed which resulted in over 500 tubefailures in 1972 due to caustic cracking just above thetube sheet mainly at the primary inlet end. By theend of 1972, 18% or the total number of tubes hadbeen plugged and the reactor power limited to 75% toreduce the load on the steam generators.

The Mihama-i failures were of a different naturefrom those in other reactors. Instead of intergranularcracks, Mihama-1 tubes exhibit localized wallthinning at the intersection of tube bends and anti-vibration straps. Failures were thought to be due tofaulty design and chemical control of the secondarysystem which probably resulted in a concentration offree caustic at a semi-stagnant steam and waterinterface. Alternate wetting and drying due to lighttube spacing brought about local wall thinning- Thelocation of the defects showed that mechanicalfretting with the anti-vibration strap was not a causeof failure [7] . Although 110 tubes were plugged in

- 3 -

1972 [8] , failures persisted in 1973 and it has beenreported tlict over 2000 tubes (approximately 25%)have been plugged [9].

Failures at Point Beach-1 followed the patternestablished in Beznau-1 in 1971-72. During the earlyoperation of this reactor, the sodium to phosphateratio in the secondary system was excessively high,indicating the presence of free caustic. Condensortube leakage was also experienced. Starting inJanuary, 1972, regular phosphate addition andperiodic blowdown kept corrosion in check, althoughthe average primary to secondary leak rate was 100!/day until a repair shutdown in October, 1972.During this shutdown the tube sheet cladding wasrepaired. This operation was to correct a fault thathad occurred in a large number of Westinghousesteam generators |10 | . Large-scale eddy currentinspection also revealed intergranular cracking oftubes in the centre of the primary inlet side from 25to 125 mm above the tube sheet. A total of 193 lubeswere plugged. A probe was installed in steam genera-tor 'A' to permit sampling of secondary water andmeasurement of temperature profiles across the tubesheet 111,121.

Less severe tube failures occurred in KWO(Obrigheim) in the Federal Republic of Germany,Shippingport-2, and H.B. Robinson-2, San Onofre-1and Haddam Neck all in the USA.

KWO experienced 67 failures in three locations;cracks on the primary side within the tube sheet, onthe secondary side about 50 mm above the tube sheet,and in the U-bend region in tubes with the smallest(55 mm) bend radius. All the failures in steamgenerator No. 1 exhibit intergranular cracking. Thosein No. 2 have not been investigated. The primary sidecracks in the tube sheet are thought to be due to thetube rolling procedure adopted in manufacture. Thesecracks are the first failures to be reported on theprimary side of any steam generator. The remainingfailures are thought to be due to water chemistryproblems and poor flow distribution in the centre ofthe tube bundle. Replacement steam generators havebeen ordered [13].

Shippingport-1, later modified to Shippingport-2,is one of the oldest PWRs* in the USA. It has steamgenerators of different types made by two manu-facturers. They were originally tubed with stainlesssteel, but were retubeel m 1964 with lnconel-600atterextensive failures. Details of the 37 failures whichoccurred in 1972 are not available [14]; but leakagehas continued in 1973 indicating further failures.

PWR Pressurized-light-water-moderatedand cooled reactor.

The H.B. Robinson-2 reactor had 32 tube failuresin all 3 steam generators. They occurred in a regionup to 100 mm above the tube sheet on the primaryinlet side and were caused by intergranular cracking[15]. A slug feeding method was originally used foraddition of sodium phosphate to control the pH ofthe secondary circuit which may have promotederratic chemical control. After repairs in 1971 tocorrect tube sheet cladding separation, a smallprimary to secondary leak developed which remainedconstant for 7 months. Blowdown of the steamgenerators was stopped duiing the last 4 months ofthis period to keep radial ion releases as low aspossible, but this may have allowed the formation olfree caustic above the tube sheet. The reactor wasshut down for tube repairs in May 1972, after theleakage rate increased abruptly. Large-scale eddycurrent testing was done with the aid of a remotelyoperated prube positioning machine. Three tubesections removed from the steam generators revealedstrong concentrations of sodium and potassiumindicating failure due to caustic cracking. Oneexplosive plug failed to bond during tube pluggingbecause of moisture in the tubes. The plug wassubsequently welded to the tube sheet. Repairsinvolved 148 workers and 145 man-rem dose [16].

San Onofre-I had 19 failures at the first baffleplate above the tube sheet and in the U-bend region.All failed tubes were near the outer periphery of thetube bundle [7, 17, 18]. These failures may berelated to mechanical damage that occurred duringmanufacture which deformed the outer tubes. Large-scale eddy current testing of over 3000 tubes in June1973 indicates some fretting wear in all 3 steamgenerators around the anti-vibration bars in the U-bend and at the first baffle plate [19, 20 ] . SanOnofre-1 is one of the older PWR's with Inconel-600tubes. If fretting wear induced by tube vibration is aphenomenon that increases with operating time, thesefailures may portend other vibration problems insimilar steam generators. On the other hand, thefretting may only relate to the manufacturing history.

Haddam Neck (Connecticut Yankee) experienced13 failures above the tube sheet on the primary inletside. They were thought to be due to intergranularstress corrosion, the cause of two earlier failures in1971 [211.

In the Palisades plant in the USA, approximately1400 tubes have been plugged. While these failuresoccurred in January 1973 and not in 1972, they arementioned as the failure mechanism is similar to thatin Miharna-1; a form of corrosion or erosion at the

— 4 —

intersection of the tubes and anti-vibration straps.The steam generators of both reactors were manu-factured by Combustion Engineering [22, 23] .

3. LOCATION AND CAUSE OF FAILURE

Figure 2 shows a typical nuclear steam generatorwith Inconel-600 tubes. Regions where failure occurredin 1972 are identified with the name of the reactorand the number of failed tubes. Stainless steel andMonel-400 tube failures are nol shown cither be-cause the steam generators construction is differentor because the location of the failures is nol known.

It is significant that failures predominate in the hotleg in the vicinity of the tube sheet and in the bendregion. Only one reactor (San Onofre) has reportedfailures occuring at tube supports in the straight tubesection.

The spatial distribution of failures in cross sectionis not well defined. Failures have occurred in anapparent random pattern; the only characteristic isthat the majority occur on the inlet side ut theprimary coolant.

Table 2 lists the cause and percentage of eachmode of failure. Various types of corrosion wereresponsible for about one half of all failures, li ma\also be the cause of some of the unknown failures.No attempt was made to subdivide corrosion failuresby mechanism because of the complexity of the sub-ject. Fretting induced by vibration occurred in onereactor and tube sheet cladding failure in another. Afurther reactor suffered extensive mechanical damage.

TABLE CAUSE OF STEAM GENERATOR FAILURE IN ;<>73

Vihutum

Tube siiccl it.tiluiili!

Dnknnwn

UCHIIII

Indian I'oirn[k'/tiau IMih.im.i |I'.mil Dc'a.hKWOII II lt.ir.ins,

San Dm. in - . 1

I ' ,ml llr.uh I

San (lrinlre IN - ICcu'lurDresden 1

Sluppurepiirl .

Number

ANT1V1BEAT1ON BAB REGIONSAN ONOFRE-I. [17)

A N I I V I B S A T I O N STRAP REGIONMIHAMA 1. 1—3000]PALISADES |1N 19731.1-14001

fIRST SUPFQRT REGIONSAN ONOF1E.I. |-6|

REGION iUSI ABOVE TUBESHEET/ POINT BtACH-1. 11931' BEZNAU-I. 1-300]

HADDAM HECK. |l'l

rcEvire REGIONBEZNAU-). PNIV7IM-

D1VIOEH PLATE » E G I O HPOINT BfACH.1.

Figure 2 - Failure Regions in Inconel-600 Tubed SteamGenerators During 1972

Conclusions and Possible Remedies

Initial inspection during all phases ol steamgenerator construction is of prime importance toensure acceptable operation. But the task of the lubeinspection alone is formidable; the total length oftubing in some new steam generators exceeds 200km. Even with on-line eddy current and ultrasonicinspection used in tube fabrication there is a smallbut finite probability that tubes used in steamgenerator construction may contain defects.

The optimum choice of tube material is a questionthat utilities and manufacturers have not resolved.Current favoured materials are Inconel-600 and ln-coloy-800. Further operating experience is necessarybefore clear trends are evident.

In operation, the choice and control of thechemistry of the secondary circuit demands carefulconsideration. At present there is no universal watertreatment method in use; in fact there is considerabledebate on the merits of the various methods. Coupledwith this problem is the use of blowdown to removeimpurities which tend to settle just above the tubesheet. Figure 2 shows the large number of lubefailures that have occurred in this region.

- 5 -

Condenser tube failures are another operationalhazard. They should be located and plugged promptlyto prevent impurities entering the feed water andaggravating water chemistry conditions.

Corrosion in the tube sheet crevice region has longbeen an area of concern to designers. It is possiblethat this problem has received more attention than iswarranted; only one reactor has experienced failuresin this region.

Tube vibration caused by cross flow from recircu-lating water in the region above the tube sheet hasnot been a problem. Only one reactor hasexperienced some tube fretting at the first supportplate. Vibration failures at the bend region may bemore serious. Although the massive tube failures inthe Palisades and Mihama-1 steam generators werenot attributed to vibration, they indicule that addi-tional information is required on the steam-waterHow in this region.

The cost of plugging failed lubes is largo in termsol lost product ion alone without considering therepair crew costs or the cost of man-rem close.Typically, steam generators cost about $10 perK\V(e). The cost of alternate electric supply resultingfrom a forced shutdown costs SO.OlO/KWh or more.On this basis, a single 1000 h shutdown costs as muchas the first cost of all the steam generators in thereactor. In ll)72, 4 of 8 reactors were in this situ-ation, and spent an average of over 3000 hours (4months) repairing failed steam generators.

This is not a case tor steam generator replacementon economic grounds, but it does place the high costof repair in perspective and show that money spenton improving reliability is justified.

4. LEAK LOCATION, INSPECTION AND REPAIR

Table 3 summarizes operator response to steamgenerator tube leakage. Methods for locating andidentifying leaks are listed with inspection and repairtechniques. Radiation fields, dose rates and remarksare included.

The majority of operators identify leakage byradioactivity in the secondary circuit - in steam,boiler blowdown or condenser air exhaust. Oncedetected, leaks are nearly always located visually afterflooding and pressurizing the secondary side. (InK.W0 a TV camera was used for remote observation.)

It is usually essential to investigate both thelocation and cause of failure so that corrective actioncan be taken for the future. The preferred method,used in 9 of 13 reactors with failures (and in many

with no failures) is eddy current inspection from theinside of the tube. This nondestructive method issensitive to small geometric changes such as cracks,wall thinning and through holes. At present mosteddy current inspection is performed manually. Thisexposes workers to high radiation fields and does notlend itself to use in newer steam generators whichhave several thousand tubes. Remote eddy currentinspection has been used in a few reactors, and islikely to gain acceptance in the future.

A complementary, though difficult, technique is insitu radiography. It has been used to detect wallthinning in the bend regions in Mihama-1 and Pali-sades. Details have been given at a conference [24 | .

Both ends of defective tubes are blocked, eitherwith mechanical plugs rolled and welded to the tubesheet or by explosive plugs. Although only 1 case ison record of an explosive plug which failed to bondto the tube (in H.B. Robinson-2), mechanical plugsappear to be the most popular. Over 3000 wereinstalled in Mihama-1 and Palisades in 1972-73.

Radiation fields in steam generators are high andoften hamper repair work. For example, 148 menwere required to inspect 4280 tubes and plug 32 inH.B. Robinson-2. The total dose was 145 rnan-rem.Repairs in 1971 to this reactor involved an additional345 man-rem. In KWL, extensive machining andwelding inside the steam generators to repair 8 con-densate tubes involved a total dose of 400 man-rem.

It appears that the complete cycle of leak location,inspection and lube plugging is well within thecapability of present remote systems technology. Iffailures continue to occur, reactor operators mayhave to consider this approach to reduce large ra-diation doses to their staff.

5. TUBE FAILURE RATES

Tables 4 to 6 list all the water-cooled powerreactors in the world with steam generators inoperation at the end of 1972. The purpose of theTables is to provide a brief summary of operatingexperience for each tube material. Six reactors areexcluded. Five are in the USSR and East Germanyfrom which no operating experience is available. Thesixth, KKS in the Federal Republic of Germany hassteam generator tubes of Incoloy-800, an alloydifferent from that used in other steam generators.

Reactors with stainless steel tubed steam gene-rators are listed in Table 4. Operating time in this andsubsequent tables is defined as the time during whichthe reactor generates electricity; down time is not

- 6 -

TA1ILI- * I IAK LOCATION, 1NSPI-CTION AND REl'Alk Ol- ST1AM GiNlRATORS IN 1472

leak Dcieiiion \ n . lulu-1* RadiationReactor and Location3 Inspection Repair Repaired I iekls Rem.uk->

N-Rt-actor (NPRJ I-'R activity in steam. Visual only Mechanical pluming 1 A.- \ K/h I <-ak in s. Sl<'«-I j(i» tuht-sVisual ,w , , v . „. i m , , I K l h m i n i b l . s

lnd'in Point - 1 Radioactivity in steam. Kddy current inspection Rolled in nipples Vl(> ' ^ 4 0 K/b at Ki-m><u- i-Jds , m M , |Visual of 1841 lulus -56'5 total tube sheet insn^ Urni us.-d.' tube

'> l<> IS K/li :K s.-.l r r n i - ^ . a

Dresden - 1 l-uss cf flmv. N>i>ni- S. Steel plug* I ; 0.15 („ u,7 K'hVisual in.stalk-d nianualh -' m,n< rein lutai

KW| lingcn Radioactivity in steam . Lu|i>id penetrant on Knifed and weld-d I I H 0.:$ K/h general I ;l,!ure ->l , • •ii.lens.itf iubp%Visual repair plugs pluip 400 .null n-ni i JUMJ lu^li JUM-

tula] dosi-

tiariglinno ltadioeriemicul analysis None S.SIeeMO.l plugs 7 < t,. h K h .itt.-t H,in-t m.-lb..,! .» ,i,-i.-, in.t

lie/nan - I Condenser air-ejei-tot. Kddy current inspi'dion )• xplosive pi up, v IHD S W li M tul>es retn>>M'dRadioactivity monilorinc, of SMS tubes •_ ̂ HHl'7Visual total

Miliama - 1 Visual Kddy current inspection Mechanical plugs I 10 Not a\;ut:ible in tubes reumveii.of HH52 lubes _ i no'?. and seal v%eld ^20Uiitotal in 197.»

Point Beach - 1 Kadioactivity in steam. Kddy current inspection Kxplosive plups !'».( 0.S I<I 0.7 K/h lube samples removed.Visual of 241\ tuhcs - ,l«"i lotal (plugged general Tield Special sample probe

in 147.M msiiillfd I<I Mi "<\"

K w t l N ' f t activity in steam. Kddy current inspection Kxplosive plugs (.7 0.4 K/h outside 4 mbr-, r<-m'>vt-.!.Radioactivity in air of 'Vyoo lubes - 17'T i : t o i s K hexhaust from condenser. tolnl insulcTV camera

H.H. Robinson 2 Kadioactivity in SCI Kddy current inspection Kxplosive plugs. \2 10 K/h general One i>iploMv* plu^Uaked,blowdovvn. of 4 2 80 tuhes - 44^! Orifice plates welded inside repaired by wddinn. •» '"I" 'V'EU"1 t**tnl in SU "A" to reduce 145 man-rem seclM-ns removed.

moisture carryover dose received1

by I4K men(M) men . • \ :S

San Onofre - I Radioactivity in Stl Kddy current inspection I-xplosive plugs I') 'i to I I K/b 1(10'" eddy current of SCi "A"hlowdown. of 1S47 tuhes = 141? and tube removal planned )r.rVisual total l')7.l.

Shippingport - 2 Radioactivity in steum. Some eddy current Welded plu£>. and M Not availableinspection mechanical plugs or

a new type

Haddam Neck Kadioactivity in steam. Kddy current inspection Kxplosive plugs 14 15 K/h ai midHydrostatic pressure of 2000 tubes = U ^ plane nl waiet l«»fctest. total

Visual location after flooding and pressurizing secondary system.

IABII: I STAINLESS STEEL riJIII-O STEAM CiMRATOR EXPERIENCE TO JAN. 1. 1173

Reactor TypeTube Defects Cumiitativc

in 1972 Tu1 DefectsNo. 1 uhes

ReactorNo. SG's /Rcaclor

OperatingTime

(years)

MTBF1000 Tubes

(years)

MTBIst;

(years)

N-lleatiui

T;irupur - l

la.a|,u . 2

SliippiiiBPorl 1

Indian Point 1

llrusden 1

Yankee Howe

KW1 l.ingetl

Ardennes {( IKKI/l

MZHt Karlsruhe

KKPGiimJrcmmign

IT Vcrcellcse

1'WR

I'WR

IIWK

I'VVU

IIWR

I'WIt

IMIWIt

11W It

„,„„

-V16

12

(1

-\,o

c

(1

t

c

> 274

> n

1 '11 60

.1200

.1200

1.0.14

.1124

72(14

64K0

10000

10

a

2

4

4

4

4

1.5

Prior iiservice

I'rior t!service

4.7

6 . 7

8.8

10.1

.1.0

Kpitlemic hpitleniif

S7K7

f.MH

4.2

4.6

4..1

- 0.2

1.4

1.6

0.K

1.0

Kxcluded frum totals

MONI I 4(1(1 Timi l ) STEAM cnNl-RAIOR EXPERIENCE K> JAN. I. U17J

Type

1IWK

I'HWR

I'HWR

I'HWR

I'HWR

I'HWR

HHWR

lube Uelectin l'>72

• V 7

(1

0

0

0

0

0

( umulative Nu. Tubes-I'uho Uufects Reactur

No. S(I*sReaclnr

Opu ratingTime

(years!

MTBF/J 000 Tubes

MTBI-

(years)

I l i i u p l i i s I ' l iu i l

V'x kerillg • 1

l'ifkerin|i - 2

1'ickeriiiB- J

KANUI'P

RAPP-I

i S h l l f ]

11: oo

11 200

11 200

HIM)

151100

-

H

12

12

12

6 . 1

t .J

1.0

0.8

0 . J

fl.-l

" ^ 0 . 0

lotals

TABI.L 6 INCONI-.L - ftOll TUBED STEAM CENKKATOR EXPERIENCE TO JAN. I, I ' m

Reactor TypeTube Defects Cumulative

in 1972 Tube DefectsNil. Tubes

ReactorNil. S<;\Rea.-lor

Operatinglime

(years)

Ml 111-IIIIHI lubes

[years)

M I III

Ueznuu - I

Miliama lJ

I'riint Heath - I

KWO (OhrigheinO

Shippingpnrl - 2

H.ll. Hr,binson . 2

San Onofre - I

lladdam Netk

NI'll

lanipur• I

Tarapur• 2

Mihamii • 2

Surrey - I

I'oinl Heath - 2

N-Rc actor

K.I (.inn., I

1'uiisades

Jose Cahrera

Malm- Yankee

Be/nnu - 2

Turkey Point - 3

I'WU

I'WK

I'WU

I'WR

I'WK

I'WR

I'WR

I'WH

I'HWH

HWIi

HWR

I'WH

I'WR

I'WR

I.WGR

PW'l

I'WR

I'WR

I'WR

I'WR

I'WR

S 30

I 10

I ' l l

67

37

32

19

14

452

160

11.1

SI

141

32

23

Id

I

5208

8H52

6520

S2I4

00.14

07 80

11.182

15176

200'I

.1200

3200

6520

10164

6520

I •) 160

6520

17038

2604

1710">

5208

9780

4.0

4.7

1.4

U.5

2.3

0.5

3.2

0.1

l-l'iili-mi.

I piilemn

• tVI

0.2

0..1

0.4

1.11

5.1

5.0

I I ,

I P

Tntals: 21

i xcluded from Totals

included. Mean times between failures (MTBF) per1000 tubes and per steam generator are calculated bydividing the product of either tube operating years orsteam generator operating years by total tube failures.It is used as a relative merit factor; its absolute valuemay not have practical significance as more than onetube may be judged defective and plugged at onetime. Extremely high failures early in the life of aplant are regarded as epidemics and are not includedin the totals. The N-Reactor falls into this category.Steam generators from this reactor, Tarapur-1 and 2and Shippingport-1 were retubed in 1967, 1968 and1964 respectively with Inconel-600 and are also listedin Table 6. The remaining 8 reactors have a total ofapproximately 50,000 tubes in 27 steam generatorsand have accumulated 45 reactor years of operation.The MTBF per 1000 tubes is 1.2 years and the MTBFper steam generator 0.7 years.

The overall record of reactors with MoneI-400tubed steam generators, shown in Table 5, is muchthe same as stainless steel. The MTBF per 1000 tubes

is 1.2 years and the MTBf per steam generator ().(>years. However, witli the exception of Garigliano.they are all of the CANDU type, and have a highlysatisfactory record oi only I tube failure in 133,000tubes in 5.7 reactor vears.

Table 6 lists reactors with Inconel-600 luhedsteam generators. Nine reactors have had failures,some for the second or third lime. Eight failuresoccurred in 1972. Beznau-1 and Mihama-I sufferedepidemic failure and are excluded from the totals.The remaining 19 reactors have over 163,000 tubes in57 steam generators and have accumulated 40 reactoryears operation. The MTBF per 1000 tubes is 0.6years while the MTBF per steam generator only 0.2years.

Inconel-600 is to be used in a large number of newreaUors being commissioned or under construction.The overall failure rates are poor and cast doubt onthe long-term performance of this material in steamgenerator service. Only one reactor (Jose Cabrera) has

TAIil.l. 7 (KTL'RRliNCi: OF INCONEL-600 TUBE FAILURES

YearNo. Reactorswith Failures

Approximate Total Failed Tuhes/1000No. Failed Tubes/y in Operation

No. Reactors inOperation withInconel Tubes

52 0.8

1 1 0.1 12

I1J71 8 6 0 8.1

1072 1000 5.6

4650 2 8

exceeded .? vears of oDcialion without a lube failure.Table 7 shows !hc occurrence of lneoncl-600 tubelailures. There has been an alarming increase in thenumber of lubes failing annually which is notmatched by the growth of new tubes entering service.While a large number of tube lailures have occurred ina lew reactors such as Be/nau-l and Mihama-1. thenumber of reactors which have failures is alsoincreasing.

6. SUMMARY

liming \{>12. I.? of 41 operating reactors incurredsteam generator lube failures. At least 1 reactorshad epidemic failures that resulted in a permanentloss of steam raising capacity.

Various forms of corrosion were the most frequentcause of failure. Others included vibration, tube-sheet cladding and mechanical damage.

Failures predominate in the hot leg in the tubesheet vicinity and in the bend region. No tubefailures have occurred above the first tube suppori

and below the bend region.

Hddy current inspection was the preferred methodfor detecting and evaluating defects in steamgenerator tubes.

High radiation doses to repair crews have beenexperienced in a number of reactors. The completecycle of tube defect location, inspection andplugging is wiihin the scope of remote systemstechnology, and such use could lead to drastic-dose reductions.

As a class. Moncl-400 tubed steam generators hadthe lowest failure rales in 1972; however, totaloperating experience is limited to about I 2 reactoryears. Inconcl-600 tubed steam generators ex-hibited the highest failure rates with a meantimebetween failure of less than 1 year per 1000 tubes.Operating experience for this class exceeds 40reactor years.

ACKNOWLEDGEMENT

The author thanks the manywho made this survey possible.

reactor operators

10

REFERENCES

1. "Power Reactors, 1973, World Directory."Nuclear Engineering International. Vol. 18, No.203, April, 1973.

2. "Power and Research Reactors in MemberStates." 1973 Edition, International AtomicEnergy Agency, Vienna, 1973.

3. P.D. Stevens-Guille, "Steam Generator TubeFailures: A World Survey of Water CooledNuclear Power Reactors to the end of 1971."Atomic Energy of Canada Limited Report,AECL4449, April, 1973.

4. Indian Point Station, Unit 1, Semi-annual Opera-tion Report No. 20, April 1, 1972-September30, 1972 (Consolidated Edison Co. of New York,Inc., New York) 15 Dec. 1972, DOCKET-50003-148.

5. "Operating Experience with Nuclear PowerStations in Member States." International AtomicEnergy Agency Report, IAEA-155, Vienna, 1973.

6. O. Deublein, "Repairing the Primary Heat Ex-changers and a Circulating Pump at KWL."Symposium on Operating and Fuelling of NuclearPlants. IAEA/SM-]78/5, Oct., 1973.

7. "The Westinghouse Steam Generator Sympo-sium." Westinghouse Nuclear Energy Systems,Pittsburg, Pa., April 1973.

8. Nuclear Engineering International, Vol. 18, No.200, p 11, 1973.

9. Nucleonics Week, Vol. 14, No. 33, Aug. 16, 1973.

10. "A Technical Report on Steam Generator TubeSheet Cladding." Westinghouse Electric Corp.WNET-102, March 1972, DOCKET 50247-69,1972.

11. Point Beach Nuclear Plant, Unit 1. Steam Gene-rator Tube Defects. (Wisconsin Electric PowerCo., Milwaukee) 20 Feb. 1973, DOCKET50266-101.

12. Point Beach Nuclear Plant, Units 1 and 2.Operating Report 5, July 1, 1972 - December3 1 , 1972 (Wisconsin Electric Power Co.,Milwaukee) 1 March, 197i, DOCKET 50266-103.

13. A. Mayr, "Operating Performance of Ste=,mGenerators and Circulating Pumps in the Primary

Circuit of KWO." Symposium on Kxperieucefrom Operaling and Fuelling of" Nuclear Plains.IAEA/SM-178/16, Oct.. 1973.

14. Shippingporl Atomic Power Slaiion. QuarterlyOperating Reports. (Duqucsne Light Co.,Shippingport Pa). D L L S 5000172, 5000272,5000372, 5000472.

15. D.B. Waters, "Operational and Licencing Aspectsof Steam Generator Tube Leakage in the Il.B.Robinson Unit No. 2 Plant." Carolina Power andLight Co., November 1973.

16. H.B. Robinson, Unit 2 Routine Operating ReportNo. 4. April I September 30, ll)72 (CarolinaPower and Light Co. Raleigh, N.C) 30 November1972. DOCKET 50261-130.

17. San Onolre Nuclear Generating Station. Unit 1,Semi-Annual Operating Report No. 11, July1 December 31, 1972. (Southern CaliforniaEdison Co., Rosemead, Calif.) 15 Feb., ll>73.DOCKET 50206-183.

18. San Onofre Nuclear Generating Station, Unit 1Operators Report for the month of February,1972. (Southern California Edison Co.. Rose-mead, Calif.) Feb. 1972. DOCKET 50206- Hi 1.

19. San Onofre Nuclear Generating Station, Unit I.Semi-Annual Operating Report No. 12, January1 —June 30, 1973. (Southern California lidisonCo., Rosemead,Calif.) 28 August, Il)73.DOCKI:T50206.

20. H.L. Ottoson, V/.R. Gould, "Operating t-.x-perience with San Onofre Nuclear GeneratingStation." Symposium on Experience fromOperating and Fuelling Nuclear Plants. 1AEA/SM-178/49, Oct., 1973.

21.Haddam Neck Piant, Unit 1. Operation ReportNu. 72-6 for the month of June 1972. (Connecli-cut Yankee Atomic Power Co., Haddam) 27 July1972. DOCKET 50213-154.

22. Nucleonics V; * . Volume 14, No. 43.

23. Palisades Plant, Steam Generator Tube Leakage,Additional Information. (Consumers Power Co.,Jackson, Mich.) 6 March 1973. DOCK T50255-139.

24. R. Stone, "Inspection of Tube Thinning in SteamGenerators." Topical Meeting, Radiography andComplementary Nondestructive Testing in theNuclear Industry. Portland, Oregan, Aug. 1972.

- 11 -

Addi t ional copies of this documentmay be obtained from

Scientific Document Distr ibut ion OfficeAtomic Energy of Canada Limiied

Chalk River, Ontar io, CanadaKOJ UO

Price - 50$ per copy