AECL--9939 CA9200406

OPERATIONS

Operating Performance ofCANDU Pressure Tubesby B. A. Cheadle* and E. G. Price**

presented a!:The lAEATechnicai Committee Meetingon the Exchange in Operational Safety Experienceof Heavy Water ReactorsVienna, Austria, February 20-24,1989

AECL-9939

April1989

A T O M I C E N E R G Y O F C A N A D A L I M I T E D

Atomic Energy £nergle atomiqueof Canada Limited du Canada limiteeCANDU Operations Operations CANDU

Sheridan Park Research CommunityMississauga, Ontario, Canada LSK 1B2

Operating Performance ofCANDU Pressure Tubesby B. A. Cheadle* and E. G. Price**presented at:The IAEA Technical Committee Meetingon the Exchange in Operational Safety Experienceof Heavy Water ReactorsVienna, Austria, February 20-24,1989

AECL-9939

Aprll1989

Atomic Energy of Canada Limited, Chalk River NuclearLaboratories, Chalk River, Ontario KOJ1J0

Atomic Energy of Canada Limited, CANDU OperationsSheridan Park Research Community, 2251 Speakman Avenue,Mississauga, Ontario LSK 1B2

Atomic Energy f-nergie atomiqueof Canada Limited du Canada limiteeCANDU Operations Operations CANDU

Sheridan Park Research CommunityMisslssauga, Ontario, Canada LSK 1B2

Operating Performance ofCANDU Pressure Tubesby B. A. Cheadle* and E. G. Price**presented at:The IAEA Technical Committee Meetingon the Exchange In Operational Safety Experienceof Heavy Water ReactorsVienna, Austria, February 20-24,1989

AECL-9939

April 1989

AbstractThe performance of Zlrcaloy-2 and Zr-2.5 Nb pressure tubes in Candu reactors is reviewed. The acceleratedhydriding of Zircaloy-2 in reducing water chemistries can lower the toughness of this material and it isessential that defect initiating phenomena, such as hydride blister formation from pressure tube to caiandriatube contact, be prevented. Zr-2.5 Nb pressure tubes are performing well with low rates of hydrogen pick-upand good retention of material properties.

Atomic Energy of Canada Limited, Chalk River NuclearLaboratories, Chalk River, Ontario KOJ1J0

Atomic Energy of Canada Limitedv CANDU OperationsSheridan Pant Research Community, 2251 Speakman Avenue,Mississauga, Ontario L5K1B2

Tenue des tubes de force duCANDU en exploitationB. A. Cheadte*. et E. 6. Price**PresentedI'lAEATechnical Committee Meeting on the Exchange inOperational Safety Experience of Heavy Water Reactorstenue du 20 au 24 ffivrier 1989, a Vienne, Autriche

AECL-9939

Avrll 1989

R6sum6

On passe en revue le comporetment des tubes de force en Zircaloy-2 et en Zr-2,5 Nb dans les rfiacteursCandu. L'fiydruration accfile're'e du Zircaloy-2 en milieu aqueux reducteur peut reduire ia resistance de cemateriau et il est essentiel d'Sviter les phgnomenes, lels que la formation de pustules d'hydrure par suite ducontact entre le tube de force et le tube de cuve, favorisant rapparition de dGfauts. Les tubes de force en Zr-2,5Nb ont une bonne tenue et pr6sentent une faible absorption d'hydrogene ainsi qu'un bon maintien despropriety du materiau.

890557

Energie atomique du Canada limitee, Laboratoires nuclgairesde Chalk River, Chalk River, Ontario, KOJ1J0

Energie atomique du Canada limitee, Operations CANDU,Sheridan Park Research Community, 2251 Speakman Avenue,Mississauga, Ontario, L5K1B2

AECL-9939

List of Tables and Figures

Page

Table 1 Fuel Channel Design Data 1

Table 2 Summary of Frac ture Toughness Results and Specimen Propertiesof Zircaloy-2 Pressure Tubes from Pickering Units 9

Table 3 Problems Experienced and Rectication 20Figure 1 Waterside Corrosion of Zircaloy-2 Pressure Tubes as a Function of Service Life 5Figure 2 Profile of Oxidation along Zircaloy-2 Pressure Tubes in Pickering A Units 1 and 2 5Figure 3 Hydrogen Concentration of Zircaloy-2 Pressure Tubes as a Function of Service Life 6Figure 4 Profile of Deuterium Concentration along

Zircaloy-2 Pressure Tubes in Pickering A Units l a n d 2 6Figure 5A Effect of Fluence on the Ultimate Tensile Strength of Zircaloy-2 Pressure Tubes at 280-309°C 7Figure 5B Effect of Fluence on Ductility of Zircaloy-2 Pressure Tubes at 280-300°C 8Figure 6 Estimated Critical Crack Length as a Function of

Temperature for Zircaloy-2 Pressure Tubes fromNPD 8Figure 7 Histogram of Spacer Displacements in P1 and P2 from Site Inspections 10Figure 8 Waterside Corrosion of c. w. Zr-2.5% Nb Pressure Tubes as a Function of Service Life 11Figure 9 Profile of Oxidation along Zr-2.5% Nb Pressure Tubes in Pickering A Units 3 and 4 11Figure 10 Maximum Hydrogen (Deuterium) Pick-up with Service Life

for c. w. Zr-2.5% Nb Pressure Tubes 12Figure 11 Effect of Deuterium Concentration along Pickering A Zr-2.5% Nb Pressure Tubes 12Figure 12 Effect of Fluence on the Tensile Strength of c. w. Zr-2.5% Nb Pressure Tubes at 300°C 13Figure 13 Effect of Fluence on Duciility of c. w. Zr-2.5% Nb Pressure Tubes at 300°C 14Figure 14 Critical Crack Length of Zr-2.5% Nb at 240-300°C Estimated from

Small Specimen Fracture Toughness 15Figure 15 The Velocity of Delayed Hydride Cracking in c. w. Zr-2.5% Nb in

Unirradiated and Irradiated Condition 15

AECL-9939

Table of Contents

Page

1.0 Introduction "1

2.0 Monitoring their condition 22.1 Dimensions 22.2 Corrosion and Hydriding 22.3 Mechanical Properties 32.4 Integrity 3

3.0 Equipment and Procedures for Monitoring 33.1 Inspection in the Reactor 33.2 Tube Removal and Testing 3

4.0 Results ot Inspections and Examination 44.1 Zircaloy-2 Pressure Tubes 44.1.1 Corrosion 44.1.2 Hydrogen Ingress 64.1.3 Microstructural Features 74.1.4 Mechanical Properties 74.1.5 Spacer Location 104.2 Cold Worked Zr-2.5 Nb 104.2.1 Corrosion 104.2.2 Hydriding 124.2.3 Metallography 134.2.4 Mechanical Properties 134.3 Heat Treated Zr-2.5 Nb 16

5.0 Operating Instructions 165.1 Leak-Before-Break 165.2 Leak Detection and Location 175.3 Break-Before-Leak 17

6.0 Review of Fuel Channel Problems and Solutions 186.1 Cracks near Rolled Joints 186.2 Crack Initialisation at Blisters 186.3 Manufacturing Flaws 196.4 Channel Elongation 196.5 Gas Annulus Hydriding 19

7.0 Summary 20

8.0 Conclusion 21

9.0 Acknowledgements 21

10.0 References 21

1.0 1NTKO1XJCT1ON

CANUU pressure tubes are performing very well. Over the last 3D years theCANDU reactor fuel channel has been developed to produce much higher powers andthe properties ot the tubes have also had to be developed to accommodate the moresevere operating conditions, Table 1. This required a change in material fromZircaloy-i! to Zr-2.5 Nb which was iortunate because subsequently it. was tound thatthe performance ot Zircalov-'2 deteriorated after long term exposure to the reactorenvironment. Alter halt their predicted service lite, Zr-2.5 Nb tubes haveexcellent corrosion resistance dimensional stability and resistance to failure.

TABLE I

FUEL CHANNEL DESIGN DATA

Power Output,

Mw (electric)

Number of

Channels

Unit

Total

Douglas Pickering Pickering Bruce Pickering Bruce 600(1) Darlington Cernavoda

NPD Point 1 and 2 Kanupp 3 and 1 1-1 5-6 5-8 MW 1-1 1-5

25

132

200

306

510

390

780

125

208

540

390

780

750

180

1920

510

390

1560

750

180

1920

600

390

1520

850

180

1920

600

380

1900

Annul us Gas

Annulus Spacer

Material

Nuaber of Spacers

Air Air N,

1x750 ZrNbCu ZrNbCu

CO, N, CO, CO, CO, CO,

1x750 ZrNbCu ZrNbCu ZrNbCu ZrNbCu ZrNbCu

(5-7)

1x750(8}

2 2 2 (1,2) 1 1 1

1 (3,1)

CO,

1x750

CO,

1x750

Coolant Pressure

MPa

7.6 9.65

Coolant Teaperature 518

K (•axiaui)

Pressure Tube A

Material (2)

572

Wall Thickness 1.2 3.9

9.65

570

5.0

10.9

572

1.0

9.6

570

1.1

10.2

581

1.1

9.6

570

1.1

10.3 11.1

581 585

1.1 1.2

11.1

586

1.2

11.1

585

1.2

Internal Dia>eter 82.5 82.5 103 82.5 103 103 103 103 103 103 103

First Power

froi Reactor

1962 1967 1971 1971 1972/73 1977/79 1982/85 1981/87 1982 1989/92 1992

(1) Gentilly-II, Point Lepreau, Eabalse, Wo1song.

(2) A U Zlrcaloy-2, B is heat truted Zr-2.5 «tX Nb, C Is cold worked Zr-2.5 wtX Nb.•tOJJ»/D?MICE••/oa/i i

AECL-9939Page 2

The excellent performance of the CANDU fuel channels is due to the inherenisafety of the design and the principle of Leak-before-Break. It a delect formsand propagates until it is through wall, the small leakage of PHTS heavy waterwill be detected and the reactor shut down before the defect can grow to thecritical unstable size. Since the pressure tubes operate in a high neutron fluxat high pressure and temperature their properties change during service andconfidence is needed that the properties particularly those that influence leakbefore break behaviour, remain within acceptable limits.

The CAKDU design facilitates non-destructive inspection of the pressure tubesand also the relatively easy removal of individual tubes to monitor the materialcondition. Inspection and sampling devices have been developed to obtain in-situ,detect, dimensional and compositional information. These results can bereferenced to the condition of surveillance tubes, which are removed at intervalsusing a lead reactor criterion and examined and tested in the laboratories. thisensures that the tubes are in a satisfactory condition and anv detect will leak-be tore- break.

2.0 MONITORING THE1H CONDITION

Four principal factors have the potential to affect the service lite of apressure tube:

- changes in dimensions;corrosion and changes in hydrogen concentration;changes in mechanical properties;change in integrity.

2.1 Dimensions

Thermal creep, irradiation creep and irradiation growth change the shape olpressure tubes during service. The tubes eLongate axially, expand in diameter aridsag. The rate of deformation is linearly related to fast neutron flux. Thedimensional changes establish a definite channel life since there are practicallimits that can be accommodated. The diametral expansion limits are establishedby considerations of the coolant flow through the fuel bundles and elongation canonlv be accommodated to the design limit of the bearing travel. The limits in sagare determined by interference with other in-core devices or by contact of thepressure tube with the calandria tube between spacers. Sag is unlikely to limitthe passage of fuel within the expected life of the channels.

2.2 Corrosion and Hydriding

The products of the chemical reaction between the pressure tube and the PHISheavy water are zirconium oxide and deuterium. The loss of metal from corrosionis structurally insignificant but some of the deuterium is absorbed bv the tubes.Deuterium also enters the tubes via the stainless steel end fittings at each endof the channels. It mav also enter from the outside surface of the tube if thereis deuterium present in the annulus between the pressure tube and calandria tuboand the oxide on the pressure tube loses its effectiveness as an adequate barrier.The increase in deuterium concentration increases the susceptibility to delayedhydride cracking (DHC) and may decrease the fracture toughness.

AECL-9939Page 3

2.3 Mechanical Properties

Neutron irradiation introduces damage to the crystal lattice that increasesthe strength and decreases the ductility* but appears to have little ettect on thesusceptibility to DHC. Fracture toughness decreases due to the damage to thecrystal lattice. Hydrogen in solution has very little eftect on mechanicalproperties but when precipitated as hvdrides it can reduce fracture toughness aridthe ductile-brittle transition temperature. The amount ot the reduction dependson the number ot hvdrides and their orientation relative to the stress direction.In the tirsl. few vears ot service, radiation damage causes the mechanicalproperties to change rapidly but subsequent changes are much more gradual.

2.A Integrity

Monitoring must ensure that there are not any conditions in the channelswhich cause localized loss of integrity from unforeseen mechanisms. For example,fretting can cause localized loss of material.

3.0 EQUIPMENT AND PHOCEDUHJES FOJt M0N1T0K1NG

3.1 Inspection in the Reactor

Techniques and equipment have been developed to measure the condition otchannels without their removal from reactor. The selection of the tubes forexamination is based on either the particular properties of the tubes or on theoperating conditions of the channel.

Most inspections are based on the CIGAR (Channel Inspection and GaugingApparatus for Reactors) device [11 which uses ultrasonics to measure diameter anddo volumetric inspection in the body of the tube and the rolled ioint, eddvcurrent to determine the position of the spacers, and inclinometer to measure sag.In addition, the following have been developed:

Eddy current inspection - for delects, spacer location, oxide thickness.

Ultrasonic inspection - tor delects in the rolled ioint region.

Deuterium sampling - Tools remove 50 mg, 0.1 mm thick slivers of materialfrom the inside surface of the tubes which are analyzed tor deuteriumconcentration. These slivers are thinner than the corrosion and wearallowance designed into the tube and leave a shallow, smooth groove.Channel elongation - The elongation ot each channel is measured at intervalsbv specific instrumentation attached to the fuelling machines at each end ofthe channel.

3.2 Tube Rewoval and Testing

The only way to measure the mechanical properties of a pressure tube is toremove a tube and test samples from it. Ontario Hydro and New Brunswick Powerhave put into place surveillance programs in which pressure tubes are removed fromlead reactors tor testing. Various factors such as service time or fluence definea lead reactor.

AECL-9939Page 4

The physical space in the reactor vault and the availability ol suitablysized tlasks determines the number of cuts that are made in a pressure tube priorto removal. Typically, three cuts are made, at the centre and near each endfitting to produce four sections. The removal procedures are described in detailby J.T. Dunn et al. [2\. The evaluation of removed tubes in cells includes 1 hefollowing:

- Visual: The inside and outside surfaces of the pressure tubes are examinedusing optical and video equipment tor surface damage, appearance of theoxide, marks that show where the spacers were located and evidence of anvcontact with the calandria tube.

Dimensions." The length of the pieces, their inside diameter and sag profileare measured.

- Rolled Joints: The as installed residual stresses in the pressure tube inthe rolled ioint. region are calculated from the inside profile of the tubeand the current residual stresses are measured bv strain gauge-slittingtechniques.

- Presence ol Detects: Both ultrasonic and eddy current techniques are used todetermine if any defects are present.

- Metallography: The microstructure, hydride orientation and distribution,oxide thickness and integrity and any detects and damage are assessed usingoptical and electron microscopes.

Mechanical Properties: Tensile strength, susceptibility to DHC, fatigueresistance, fracture toughness and burst strength are determined usingappropriate specimens and equipment. To conserve material and test a widerange of variables, small specimens are used wherever possible. Specimenscan be prepared by spark machining, punching or mechanical machining.

Chenical Composition: Small through wall coupons are punched out ol the tubeat many locations principally for hydrogen and deuterium analyses.

4.0 RESULTS OF INSPECTIONS AND EXAMINATIONS

4.1 Zircalov-2 Pressure Tubes

In Canada, cold worked Zircaloy-2 tubes were installed in NPD, Douglas Pointand Pickering NGS "A" Units 1 and 2, Table 1. These reactors had operated withoutsignificant interruptions until the rupture of a pressure tube in Pickering Unit 2in 1983. Since that incident, an extensive investigation of the behaviour andcondition of Zircaloy-2 pressure tubes from NPD and Pick&ring Units 1 and 2, andcorrelation with the Hanlord-N reactor data has shown that there are seriouslimitations in the performance of the allov in reducing chemistry water.

4.1.1 Corrosion

In tluxes > 1017 n.m-^.s-1 (E > 1 MeV) and temperatures above 250°C there isan apparent initial transient period of oxidation followed by a low "secondary"oxidation rate. In NPD the low secondary rate lasted tor about 12 vears, whereasin Pickering Units 1 and 2, it continued tor only about six years, Kigure 1 \'S\ .The initial low rate of oxidation was followed by a gradually increasing rate upto oxide thicknesses of *" fol) microns when Pickering Units 1 and 2 were shutdown.

AECL-9939Page 5

The start ot the increasing rates occurred when the oxides reached a thickness ot15-2U microns. The tube removed trosn NPL) in 198^ h«d an oxide thicknesses <>l66 microns. Predictions that oxide thicknesses would be 7O-d(J microns in 1WH(>were confirmed by an inspection and subsequent removal ot" a tube in 1(.)H7. Figure2 shows the profile ol the oxide thickness along Pickering Unit 2 pressure tubes|3 | . The oxide was thickest in regions where I lux and tempera Lure are highest.The profile along NPD tubes was less regular with the peak closer to the centre

wUl

o

X

X

oI

100

90

80

60

50

CORROSION OF 2IRCALOY-2 PRESSURE TUBES

PICKERING-1 AND -2

_ O HANFORD NREACTOR

NRU-TUBE 289

- Q NPD

@ PICKERING

ZIRCALOY-2(HIGH FLUX)

1 I

1000 2000 3000 4000 5000

TIME (EFPD)

6000 7000 8000

FIGURE 1 WATERSIDE CORROSION OF ZIRCALOY-2 PRESSURE TUBESAS A FUNCTION OF SERVICE LIFE.

200 400

DISTANCE FROM INLET (cm)

600

FIGURE 2 PROFILE OF OXIDATION ALONG ZIRCALOY-2 PRESSURE TUBESIN PICKERING A UNITS 1 AND 2

AECL-9939Page 6

.1.2 Hydrogen Ingress

Hydrogen (as deuterium) enters the tube from the corrosion reaction with thePHTS -heavy water, from crevice or galvanic reactions with the end fittings at theends of the tubes and sometimes from the gas annulus. The information fromsurveillance tubes removed from the Hanford N-Reactor confirmed that hydrogeningress with time followed a similar pattern to corrosion, Figure 3, and hydridingrates of 12 ppm/EFPY are attained [3]. The profile of hydrogen concentrationalong the Pickering tubes is shown in Figure A. At the highest temperature

o.

111

ai

25

90

80

70

60

SO

40

30

20

10

0

o03

HYDRIDING OF ZIRCALOY-2 PRESSURE TUBESPICKERING-1 AND -2HANFORD N-REACTORNRU-TUBE 289NPDPICKERING-AND -2

ZIRCALOY-2(HIGH FLUX)

/ / /

VA?^

700

600

500

400

300

200

100

z

1000 2000 3000 4000

TIME (EFPD)

5000 6000 7000 8000

FIGURE 3 HYDROGEN CONCENTRATION OF ZIRCALOY-2 PRESSURE TUBESAS A FUNCTION OF SERVICE LIFE.

320

280

E 240

P 200

w

o2si

160

120

80

40

ZIRCALOY-2 (PI, 2)HIGH POWER CHANNELS.

(3600-3710 EFPD)

200 400

DISTANCE FROM INLET (cm)

600

FIGURE 4 PROFILE OF DEUTERIUM CONCENTRATION ALONG ZIRCALOY-2PRESSURE TUBES IN PICKERING A UNITS 1 AND 2

AECL-9939Page 7

regions of the channel and at the highest flux positions the hydrogen build-up wasgreatest. Out of tlux the hydrogen concentration was much lower. In NPD tubes,the hydrogen concentration agreed with the pick-up expected from the oxidethickness. In Pickering Units 1 and 2, the pick up on some tubes near the inletend was higher than could be accounted for by the corrosion on the inside surtace.

.1.3 Microstructura1 Features

The mode of hydride precipitation was variable. Some tubes from thePickering reactors showed extensive hydride precipitation on the radial-axialplane as the result, of the hydride build-up and the operating stress. Other tubesshowed mostly circumferential-axial orientation. As hvdride concentrationincreased, particulary in the NPD tubes, more hydride assumed a radial-axialorientation |4]. At the positions of contact in the Pickering Units 1 and 2pressure tubes, localized concentrations of hydride ("blisters") had formed f5).These blisters were typically cracked. There was also a noticeable hydridegradient through the pressure tube wall.

The oxide on the inside surface contained many cracks and had a layeredstructure. On the outside surface the oxide both on NP1J tubes and Pickering Atubes had not significantly increased in thickness since installation. On somePickering A tubes near the inlet end, the oxide was less than one micron thickindicating little oxidation since start-up but on the NPD tube the oxide wastypically six to eight microns thick i3,4].

4.1.4 Mechanical Properties

4.1.4.1 Tensile Properties

Transverse rings and longitudinal tensile specimens were used to measure thetransverse and longitudinal tensile strengths of NPD and Pickering A Zircaloy-2

ISQ.

oW

a.

LU

3UU

800

700

600

500

400

\I

300

200

inn

-

-

" /

]

-

•

O D

D

TRANS

NPD D

PICKERING A -

1

AXIAL

O

#

1

rD

FLUENCE (x 1025

FIGURE 5A EFFECT OF FLUENCE ON THE ULTIMATE TENSILE STRENGTHOF ZIRCALOY-2 PRESSURE TUBES AT 280-300°C

AECL-9939Page 8

o

55sIU

30

(

20(

10

-

0

i

NPO

PICKERING A

AXIAL

O

•

1

FLUENCE (x 1025 n/m*)

FIGURE 5B EFFECT OF FLUENCE ON DUCTILITY OF 2IRCALOY-2PRESSURE TUBES AT 280-300°C

150

1003co

oSo

50

FUELCHANNEL

GO7

FOB

BRITTLEFRACTURE

BURST SMALLTEST SPECIMENS

O D

DUCTILEFRACTURE

SMALLSPECIMENS

J

pRT

X

- 8.04- 41.3- 4.19

MPammmm } Oh - 79.2 MPa

*

(INLET)

CONSERVATIVE RESULTS.CCL LIMITED BY SPECIMENLIGAMENT

100 200

TEMPERATURE. »C

300

FIGURE 6 ESTIMATED CRITICAL CRACK LENGTH AS A FUNCTION OFTEMPERATURE FOR ZIRCALOY-2 PRESSURE TUBES FROM NPD

AECL-9939Page 9

pressure tubes. Because the tubes see a varying flux (and temperature) alongtheir length, specimens from varying positions provide an indication of the effectof fluence. There is little increase in strength and decrease in ductility abovea tluence of 2 x 1O25 n/m2, Figure 5 (4,61.

A.1.4.2 Fracture Toughness

Fracture toughness is measured either by testing 17 mm curved compact tensionspecimens or by bursting 450 mm long tube sections containing an axial, through-wall slit. In both specimens the crack plane is the axiai-radial plane andcracking is in the axial direction.

Both the pressure tubes removed from NPD had high deuterium concentrationsand manv of the hydrides were oriented close to the radial orientation. In thetube removed in 1984 the transition from brittle to ductile fracture behaviour wasat about 22O°C, Figure 6, which was below the range of reactor operatingtemperatures, i.e. it a pressure tube in NPD had fractured during reactoroperation then the crack would have propagated in a ductile manner. However, thetransition temperature of the tube removed in 1987 had increased to about 2bO°Cand if a tube had fractured the crack would probably have propagated in a brittlemanner. The change in fracture behaviour of the NPU tubes between 1984 and iyB7was probably due more to the increase in the concentration of radial hydrides thanthe increase in tluence.

TABLE II

SUMMARY OF FRACTURE TOUGHNESS RESULTS AND SPECIMEN PROPERTIES

OF ZIRCALOY-2 PRESSURE TUBES FROM PICKERING UNITS(7)

Specimen

Number

C5

C7C13

cmC16

C31

C33C35C1C6

C11C15C20C24

C36C37

Test

Temperature

°C

280

280

280280280

280

280280230

230280230280

280230230

Tough

orBrittle

TTTTTTTT

BBBBBBBB

Fracture

Toughness

MPa.m12

54.0a

63.6a

69.7a

56.9a

63.6a

67.3a

66.0a

55.4a

48.7

53.6

75.4

50.3

46.0

62.0

40.0

48.5

Yield

Stress

aMPa

662

662662662662

662662662719

719662719662

662719719

[H] EquivalentM9/9

61

59

9542

4558

42...72

63

110• . .

1091468166

Critical

Crack

Length-

mm

71

102108{LB)

88

95103(LB)

99(LB)

754953

705047604149

2.5 Y.Q2/O2

mm

..

...

...

...

...

...

...

...11

14147.713

12

1927

a. Calculated from K_ _ ^EJ, where J, is read fro* the first pop-in on the J-resistance curve.

* The values are conservative if the specimen fracture mode was elastic-plastic.ICE'02/17

AECL-9939Page 10

The fracture toughness ot the Pickering Units 1 and 2 pressure tubes wasmeasured in regions ot relatively high hydrogen concentration and it was apparentthat the ductile-brittle transition temperature was between 2.iU°C and ^BO°C, Table2 |7|. Again the structure with the greatest amount of radial-axial hydrideprecipitation exhibited the lowest toughness, but the lowest calculated criticalcrack length still exceeded 4U mm.

4.1.5 Spacer Location

The NPD channels had one spacer separating the pressure tube from thecalandria tube. These spacers were of Inconel X750 and were fabricated as gartersprings and at installation were tight on the tube. Four channels were inspectedand the spacers were in position. The spacers removed were in good condition.

Similar inspections on Pickering Units 1 and 2 showed that displacement ofthe spacers had occurred, Figure 7. These spacers were of Zr-2.5 Nb-0.5 Cu butwere not tight on the pressure tube to allow for diametral creep. Channelvibration during installation and commissioning caused the spacers to move.

m

70

60

50

40 FREQUENCY

30

20

10

1.3 .9 .5 .1 .1 .5 .9 13 17

(m) (m)INBOARD ( - ) OUTBOARD (+)

FIGURE 7 HISTOGRAM OF SPACER DISPLACEMENTS IN PI AND P2FROM SITE INSPECTIONS

4.2 Cold worked Zr-2.5 Kb

4•2.1 Corrosion

Although one cold worked Zr-2.5 Nb tube was installed in NPD, reactorexperience effectively started with the operation of Pickering Units 3 and tt. Themajority ot the corrosion information has been obtained from tubes from PickeringUnits 3 and 4 and Bruce A Units 1, 2 and 3.

AECL-9939Page 11

ons)

.&>E.

$

CK

f

I

UJOIX

O

2O2X

I

ou

70

60

50

40

30

20

10

• PICKERING A

^ BRUCE A

-

i i

1000 2000 3000

TIME (EFPD)

4000 5000

FIGURE 8 WATERSIDE CORROSION OF c.w. Zr-2.5% Nb PRESSURE TUBESAS A FUNCTION OF SERVICE LIFE

8in

gxo

45

30

15

3430-3850 EFPD

200 400

DISTANCE FROM INLET (cm)

600

FIGURE 9 PROFILE OF OXIDATION ALONG Zr-2.5% Nb PRESSURE TUBESIN PICKERING A UNITS 3 AND 4

Oxide thickness as a function ot service life is shown in Figure 8 fH|. Thesecondary oxidation rate appears to have stayed constant similar to the secondaryrates of Zircaloy-2. These results are irom the metallopraphy ot tubes removedfrom the reactors. -The eddy-current technique developed to measure oxidethickness in-situ indicates similar thicknesses, but the accuracy ot the techniqueis less than that of the metallography due to surface unevenness and otheraffects. The profile of oxide thickness along Zr-2.5 Nb pressure tubes shows aslight increase near the outlet end (Figure 9).

AECL-9939Page 12

i. 2.2 Hvdriding

The maximum local deuterium pick-up with operating time is shown in Figure 1Uand along a tube in Figure 11 [81. The maximum concentrations were at positionsnear the outJet end. The pick-up of deuterium from the oxidation reaction isconsiderably less in Zr-2.5 Nb (5X-1OX of theoretical) than in Zr-2 (typically> 40%). A small number of tubes in Pickering Units 3 and 4 exhibited higher

70

60

50

40

30

20

to

-

-

••I

--

PICKERING A

BRUCE A

PICKERINGG AMULTI-TUBE SCRAPE RESULTS

444 • 44 4 . 4 i

•

1000 2000 3000

TIME (EFPD)

4000 5000

FIGURE 10 MAXIMUM HYDROGEN (DEUTERIUM) PICK UP WITH SERVICE LIFEFOR c.w. Zr-2.5% Nb PRESSURE TUBES

200 400

DISTANCE FROM INLET (cm)

600

FIGURE 11 PROFILE OF DEUTERIUM CONCENTRATION ALONG PICKERING AZr-2.5% Nb PRESSURE TUBES

AECL-9939Page 13

deuterium pick-ups near the outlet end than other tubes. This increase was notmatched by an increase in oxide thickness on the inside surface. It is thoughtthis increase may be due to localized ingress from the outside surtace. A tubeIrom channel L09 in Pickering Unit 3 had the highest deuterium increase.

The development of a tool to remove a sliver of metal from the top insidesurtace of an operating pressure tube has enabled a greater amount of data to beobtained f9J. This tool has been used in Pickering Unit 3, Unit 4 (twice) andBruce Unit 3. Deuterium pick-up rates from waterside corrosion are found toincrease with increase in operating temperature.

4.2.3 Meta11ographv

The hoop stress necessary to re-orient hydrides in c.w. Zr-2.5 Nb is in therange 180-220 MPa, and since the operating stress is in the range 130-150 MPa thehydride precipitation mode should remain in the circumferential-axial plane. Thishas been observed except where high residual stresses near rolled joints causedre-orientation in early reactors.

The oxide morphology on Zr-2.5 Nb is somewhat different to Zircaloy-2. Thickoxides exhibit cracks perpendicular to the surtace as well as a terdency to crackparallel and close to the metal-oxide interface.

4.2.4 Mechanical Properties

4.2.4.1 Tensile Properties

Cold-worked Zr-2.5 Nb exhibits an increase in yield ana tensile strength withfluence, Figure 12 and a decrease in ductility, Figure 13 110].

i13HItr

v>111

900

800

700

600

UJ 500

400

c 300

200

100

a TRANSVERSEo AXIAL

FLUENCE (x 1025

FIGURE 12 EFFECT OF FLUENCE ON THE TENSILE STRENGTH OFC.w. Zr-2.5% Nb PRESSURE TUBES AT 300«C

AECL-9939Page 14

80

60

tUJ

Q

wa.

40 -

20 -

o TRANSVERSE

o AXIAL0

" *

~ • * / fl

i i

0 a oa

•

1

FLUENCE (x 1025 n/m?)

FIGURE 13 EFFECT OF FLUENCE ON DUCTILITY OF c.w. Zr-2.5% NbPRESSURE TUBES, AT 300°C

4.'2.4.'2 Fracture Toughness

The development of small specimen fracture toughness test procedures haveenabled the trend with fluency to be determined. As with ductility, there is arapid decrease in early life to about 50% of the as-manufactured value and a muchsmaller rate of decrease with increasing fluencet Figure 14. Some tubes showsignificantly higher toughness values than others.

4.2.4.3 Delayed Hydride Cracking (DHC)

The important parameters for DHC are the stress intensity factor Km forcrack initiation, the rate of crack propagation and the temperatures at whichcracking can occur. DHC can only occur when hydrides are present in themicrostructure, that is, when the hydrogen concentration is above the terminalsolid solubility (TSiJ). New pressure tubes contain less than 20 ppm hydrogen;with this amount TSS is about 22U°C and DHC can only occur below this temperature.With the ingress of deuterium, the temperature below which DHL' can occur increasesuntil eventually it can occur while the reactor is operating. DHC can occur alreactor operating temperatures much earlier in Zircaloy-2 tubes than in Zr-'2.5 Nbbecause of the higher rate of deuterium ingress.

It is not yet clear if irradiation influences K^f but it does increase therate of crack propagation in Zr-2.5 Nb, Figure 15 fll|.

too

E.

Ul

2cro

eo|-

60

IOUi

I -Ui

20

o00

0

o

oo

o°

1*8 oo

° o° 8 o° o

oo ° °o cPo o

o

8o

0

I

AECL-9939Page 15

FLUENCE (x 1025 n/m2)

FIGURE 14 CRITICAL CRACK LENGTH OF Zr-2.5% Nb AT 240-300°C ESTIMATEDFROM SMALL SPECIMEN FRACTURE TOUGHNESS TESTS

10-5300C 250C 200C 150C

10-6 3

E>

UI

2o

10-7 -

10-8 -

10-9 1 1

A

•

A

\1

UNIRRADIATEDIRRADIATED INUNIRRADIATEDZr-2.5% Nb

b> n

A ^V

NRU

N

0.0018 0.0020 0.0022

1/T, 1/K

0.0024 0.0026

FIGURE 15 THE VELOCITY OF DELAYED HYDRIDE CRACKING IN c.w. Zr-2.5% NbIN UNIRRADIATED AND IRRADIATED CONDITION

AECL-9939Page 16

<i.J Heat-Treated Zr-2.5 Nb

Heat-treated 2r-2.5 Nb pressure tubes are water quenched from the (0:+ 0)phase, cold worked about 12* and then aged at 500°C. This produces amicrostructure that consists ot about 9U* martensitic c*'and 10* equilibrium0*grains. After the water quench, the martensitic c<'is supersaturated in Nb. Thesubsequent cold work and aging treatment results in a small amount otrecrystaliization and small/^Nb precipitates are formed. The martensitic o<' has arelatively random crystallographic texture and hence the properties of heat-treated tubes are much more isotropic than cold worked tubes.

Compared to cold worked tubes, there is much less information on heat-treatedpressure tubes that have been tested atter service in reactors. In general, theirtensile strength is higher, their ductility, corrosion and hydriding, fracturetoughness and susceptibility to DHC are similar, their axial elongation is muchless but their diametral expansion is slightly higher.

5.0 OPERATING REQUIREMENTS

For satisfactory operation, the material should remain in an adequately toughcondition and any defect generating mechanism should produce a crack whose growthwill result in a detectable leak before it becomes unstable: i.e., leak-betore-break rather than break-before-leak.

5.1 Leak-Betore-Break

For this to occur a number of criteria must be satisfied: the crack mustpenetrate the tube wall and leak, the leak must be detected and the reactor shutdown before the crack grows to the critical length. Thus, the followinginformation is required:

- the maximum length ot the crack when it penetrates the wall;

- the crack velocity;

the critical crack length and its variation with fluence;

the time to detect a leak.

The first three parameters can be estimated from tests on surveillance tubes.The last parameter is a function of the reactor design and operating procedures.

Delayed hydride cracking in Zircaloy-2 has been studied in the laboratory,but the only occurrence in service was the propagation from blister inducedcracking in Pickering Unit 2 channel G-16. A DHC crack in Zircaloy-2 shouldpenetrate the wall with a length about twice the wall thickness. Crack velocityhas an Arrhenius relationship with temperature and at operating temperaturescracks could grow to critical lengths in about 24 hours [4] ., In NPD the criticalcrack length was short because of the presence of radial hydrides and the leakdetection system with an open annul us may not have detected a leak in time toprevent unstable failure. Since any failure would have been in a brittle mode(Figure to), reactor operation was terminated in 1987.

AECL-9939Page 17

In Zr-2.5 Nb pressure tubes, the UHC velocity in the axial direction is twicethat in the radial direction. Therefore a crack should penetrate the wall whenthe crack length is tour times the wall thickness. However, in some instancescracks have grown much longer than this before the rate of PHTS heavy waterleakage was sufficient to be detected. Nevertheless, the crack detection methodsare adequate to detect cracks before they could grow to the critical length.

5. '2 Leak Detection and Location

All mechanisms causing leaks in CAN1XJ pressure tubes have been associatedwith delayed hydride cracking (DHL1). This mechanism of crack propagation inzirconium alloys is relatively fast compared to fatigue and the likelihood otinspection detecting a propagating crack is remote.

In early reactors, the annuli were open and leaks could only be detected bv ahumidity increase in the reactor vault. In later reactors the annulus is sealedand has a gas which is circulated and monitored for moisture content. The systemis sensitive to the ingress of small amounts of moisture and leaks as small as alew grams per hour can be detected within an hour from changes in dewpoint.

Leak-before-break is facilitated by reactor temperature manoeuvring after a leakis detected. It has been found that thermal cycling a reactor to about 3U K belowthe intended hot shutdown temperature, stops crack growth fl'2|. This procedureenables searches tor small leaks to be carried out for a limited time in the hotshutdown condition, where through-wall cracks generally continue to leak. Whenshutdown to cold conditions, cracks may plug making leaks difficult to locate.

5.3 Break-Before-Leak

This occurs if a crack grows to the critical length before any PHI'S leakageis detected, as in Pickering Unit 2 channel G-16 when DHC cracks initiated atblisters. The characteristics of blister formation are substantially similar inZr-2 and Zr-2.5 Nb. If a region of the pressure tube wall touches the calandriatube it is cooled. If the temperature is below the terminal solid solubilityhydride precipitation will occur and the thermal gradients will allow furtherhydrogen to migrate to the relatively cool zone and build up a mass ot hydride.The controlling conditions for blister formation are the temperature of theregion, which is determined by the temperatures of each tube and the thermalresistance between them, and the hydrogen concentration in the pressure tube.

The steps in blister formation and cracking are:

- Sag of the pressure tube resulting in contact with the calandria tube. Thisusually involves limited fretting before loaded contact. The frettinggenerates a loose oxide.

Decrease in temperature of the pressure tube at the point ot contact.

•-- Precipitation of hydride if the hydrogen concentration is above the terminalsolid solubility (TSS) for- the conditions. (If the concentration is belowTSS, no precipitation occurs.)

AECL-9939Page 18

- build-up ot solid hydride to a "blister" shape or layered structure dependingon the nature ot contact.

- Growth of the blister with time controlled by the total hydrogenconcentration.

- Cracking ot the blister from volumetric expansion and the hoop stress on thetube during the growth process. The probability of blister fracture and thedepth of the crack in the blister increase with blister depth.

Propagation of the crack outside the blister when a critical depth determinedby the stress, is reached.

Growth ot the defect by DHC preferentially in the axial direction due tothermal gradients in the tube wall causing the hydrogen to concentrate nearthe outside surface.

The higher the hydrogen concentration, the greater the size ot blister thaican be developed. Thus Zircaloy-2, with an accelerating hydrogen ingressphenomenon, is at risk much earlier than Zr-2.5 Nb.

The timeliness of action to prevent blister formation requires:

- a knowledge ot the hydrogen concentration;

- a knowledge of the possibility of contact from the position of the spacersand calculations ot sag, or by inspection.

Since determining the position of the spacers is a slow process, warning ofpossible blister formation is best achieved by measuring the deuterium ingress oitubes that have high initial hydrogen concentrations.

6.0 REVIEW OF FUEL CHANNEL PROBLEMS AND SOLUTIONS

While CANIXJ fuel channels have operated very effectively a number of problemshave had to be resolved. These problems were due to design shortcomings, assemblyerrors or fabrication flaws.

6.1 Cracks Near Rolled Joints

Incorrectly rolled joints in Pickering Units 3 and 4 resulted in many Zr-2.5Nb tubes having very high residual stresses in the rolled .joint region, and sometubes developed cracks by DHC. The solution was to stress relieve joints alreadyfabricated at the Bruce A reactors and to develop a low stress rolled ioint forreactors under construction. However, two tubes at Bruce Unit 2 leaked becausecracking had started before the stress relief was done. Stress relaxation duringservice was sufficient to prevent extensive further crack initiation at Pickeringand subsequent to 1975 only one additional tube has developed a leak (in PickeringUnit 3).

6.2 Crack Initiation at Blisters

A Zircaloy-2 pressure tube ruptured in Pickering Unit 2 (G-16). In thisinstance more than one factor was involved in producing unstable rupture.

AFCL-9939Page 19

Firstly, one of the two spacers had moved out ot position allowing 1hp pressuretube to sag and contact the cool calandria tube near the outlet end. Sermul Iv,the Zircaloy-2 pressure tube had built up a high hydrogen content near the out leiend and hydride blisters formed. The blisters cracked and after thev had grown toa threshold size the cracks propagated axially and grew to an unstable length.Radial hydrogen gradients in the tube prevented the crack penetrating to theinside of the wall and leaking before the unstable length was reached

A number of reactors have the zirconium alloy spacer design which can m m eduring assembly and commissioning. Displacement ot the spacers increases thelength of the unsupported span and allows pressure tubes to touch the calandriatubes prematurely. Contact of the pressure tubes with the calandria tubes is nota serious concern until hydrogen builds up in the tubes beyond a threshold value.In Zr-2.5 Nb tubes this wiJl not occur until the latter halt ot the design life.The prudent move will therefore be to reposition the spacers during the first hnltot the design lite. An inspection and correction system, (Spacer Location andkepositioning (SLAK)) has been developed to correct this situation. This type otoperation will need to be carried out on about eight CANDU units. However,greater time margins are available on these units because thev use /r-2.5 Nbpressure tubes.

However, when only two spacers are present in a channel, the flux enhancedsag in 6 m long channels will allow some pressure tubes to contact calandria tubesbefore the end of design life. Therefore relocating spacers is ot limited benefitin these reactors and they will eventually be retubed and tour spacers will beinstalled.

6.3 Manufacturing Flaws

Two pressure tubes in the Bruce Unit 2 reactor contained defects which hadescaped detection during manufacture. Cracks initiated at the delects by DHC.Both leaked and the leakage was detected but one tube failed unstably when it waspressurized cold after shutdown. Improved process control and inspectionprocedures will eliminate these detects in new pressure tubes.

6.4 Channel Elongation

The design of the Pickering A and earl> Bruce A fuel channels did not allowtor sufficient pressure tube elongation and the bearing travel would be used upbefore 30 years. When the bearing travel is used up, the channels could besupported at one end outside the normal bearings to allow continued operation.However, since these reactors are also the ones that contain onlv two spacers,thev will be retubed and the elongation allowance will be increased.

6.5 Gas Annulus Hydriding

Surveillance tests on tubes from the Pickering Units 3 and 4 reactors haveindicated that some tubes have picked up deuterium near the outlet end faster thanexpected. The circumstantial evidence favours deuterium ingress from the nitrogengas annulus due to the degradation of the normally protective oxide. Carbondioxide is now used in all other units and is expected to prevent degradation.Increased oxygen concentration in the carbon dioxide has now been recommended toensure that a protective oxide barrier is maintained. These design problems andtheir solutions are summarized in Table 3.

AECL-9939Page 20

Table III

Problems Experienced and Rectification

990116/07PRICEt9/02/17

PROBLEM

1. Zircaloy-2 excessivecorrosion and hydriding

2. Displacement ofrelatively looseZr-2.5% Nb - 0.5 Cuspacers

3. Two-spacers. Contactpossible after 17 years

4. Bearing travel onlysuitable for 15-yearservice.

High residual stressrolled joints

a) Open annulus.Activation ofA«,. (NPD, DouglasPoint).

b) Nitrogen Annulus Gasproduction of Ci i, -

c) Annulus Gas - lowoxidizing potential.

d) Slow response toleakage.

High growth rate of c.w.Zr-2.5%Nb

RECTIFICATION

1. Change to c.w. Zr-2.5%Nb.

2. a) change to Inconel"snug" spacers in newreactors.

b) spacer relocationtechniques developed.

3. Four snug spacers preventcontact for 60 years.

4. a) Bearing length andfeeder arrangementmodified to meet30-40 yr. service.

b) Possible channelbearing relocationoff bearing supportsat one end ofchannel.

5. Low stress jointdeveloped.

6. a) Closed annulus sealedwith Inconel bellows.N2 or C02 gas.

b) Use of C02 only

c) Increased 02

concentration.

d) Sealed annulus withrecirculating gas.

7. Modifications tofabrication process.

7.0 SUMMARY

Inspection and measurement of properties or lr-2.5%Nb pressure tubes showthat the tubes will operate tor their design life in good condition. To datethere is no indication that Zr-'2.5*Nb could suffer the deterioration seen inZircaloy-i! material. Channels in out-ot-design condition need to be restored tothe design condition to ensure that hydride blister induced crack initiationmechanisms cannot be activated.

AECL-9939Page 21

8.0 CONCLUSION

(a) Zircaloy-2 in reducing environments can suffer accelerated hvdriding withinsix years of start up.

(b) The loose spacer design can move during installation and commissioning andproduce larger span lengths than designed. Creep sag can cause contactbetween the pressure tube and the calandria tube.

(c) Contact can cause hydride blister formation and the conditions lor crackgrowth by DHC which may not. result in leak-belore-break.

(d) Contact between pressure tubes and calandria tubes must not occur when thetotal hydrogen concentration exceeds TSS of the cooled contact zone.

(e) Cold-worked Zr-2.5 Nb pressure tubes hydride at an acceptable rate andmaintain good properties. However, continued contact between these tubes andcalandria tubes is undesirable after a hydrogen concentration beyond athreshold level is present.

(t) A sensitive leak detection system is essential to limit the consequences olpressure tube cracking.

9.0 ACKNOWLEDGEMENTS

The authors wish to record their thanks tor use of information generatedunder CANDU Owners Group development programs and from Ontario Hydro surveillanceand inspection programs.

10.0 KEJb'EKENCES

II] M.P. Dolbey, "CIGAK - An Automated Inspection System tor CANlrtJ Keactor FuelChannels" ASM 8th International Conference on Non-Destructive Evaluation inthe Nuclear Industry, Orlando, Florida, 1986 November.

(21 J.T. Dunn, B.K. Kakaria, J. Graham, A.J. Jackman, "CANDU-PHW fuel ChannelReplacement Experience", IAEA Specialists Meeting, Kiso, Denmark, 1982September (AECL-7538).

[3] V.F. Urbanic and B. Cox, "Long Term Corrosion and Deuteriding Behaviour ofZircaioy-2 Under Irradiation", Can. Metal 1. Quart. V. 24, No. 3, 1985, p.189-196 (AECL-8874).

|4) C.E. Coleman et al "Evaluation of Zircaloy-2 Pressure Tubes from NPD", EighthInternational Symposium on Zirconium in the Nuclear Industry", San Diegn,1988 June, 19-23.

[5| G.J. Field, J.T. Dunn, B.A. Cheadle, "Analysis of the Pressure 'lube Failureat Pickering NGS 'A' Unit-2, Nuclear Systems Department", Can. Metal I.Quart., Vol. 24, No. 3, p. 181-188, 1985.

((>] W.J. Laugtord, "Metallurgical Properties of Cold Worked Zircaloy-2 PressureTubes Irradiated Under CANUU-PHW Power Keactor Conditions", ASTM STP1971, p. 259 (AECL-3833).

AECL-9939Page 22

|7| U.K. Chow, L.A. Simpson, "Analysis oi the Unstable Fracture of a Head orPressure Tube Using Fracture Toughness Mapping", ASTM, SiiJ 91H, ISlUb, p. Vli-101.

[B] V. Urbanic, A. Manolescu, ti. Warr and O.K. Chow, "Oxidation and DeuteriumPickup ot 7.r-2.5 Nb Pressure Tubes in CANDU-PHW Weactors. EighthInternational Symposium on Zirconium in the Nuclear Industry, San Diego, \9titiJune, 19-23.

|9] W.E. Huth, C.A. Cox and R. Joynes, "Tooling to Obtain Samples ot PressureTube Material In-Reactor tor Deuterium Analysis", CANDU MaintenanceConference, Toronto, Ontario 1987 September, paper 4B-7.

(101 E.F. Ibrahim, "Mechanical Properties ot Cold Drawn Zr-2.5Nb Pressure TubesAfter up to 12 Years in CANDU Reactors", Conference on Materials for NuclearCore Applications, BNES, London, 19b7, Paper 12 (AKCL-9480).

111 I S. Sagat, proposed paper for presentation at Specialists Seminar on Leak-Bel ore-break, Toronto, Ontario, 19Hy November.

1121 C.f. Coleman et al "Minimizing Hydride Cracking in Zirconium Alloys", Can.Metall. Ouart., V. 24, No. 3, 19H!i, p. 2

The International Standard Serial Number

ISSN 0067-0367

has been assigned to this series of reports.

To identity individual documents in the serieswe have assigned an AECL-number.

Please refer to the AECL-number when requesting additionalcopies of this document from:

Scientific Document Distribution OfficeAtomic Energy of Canada LimitedChalk River, Ontario, CanadaK0J1J0

Le numero de s6rie standard international

ISSN 0067-0367

a 6t6 attribue" a cette sSrie de rapports.

Pour identifier les rapports individuels faisant partie de cettess6rienous leur avons attribue' un nume"ro AECL-.

Veuillez faire mention du numfiro AECL- si vous demandezd'autres exemplaires de ce rapport au:

Service de Distribution des Documents OfficielsL'Energie Atomique du Canada, LlmiteeChalk River, Ontario, CanadaK0J1J0

Price Code A C0C3 de prix A