corrosion and mechanical strength of welded joints of downcomers for rbmk reactors

9
Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors B.T. Timofeev * , G.P. Karzov, A.A. Gorbakony, Yu.K. Nikolaev CRISM ‘‘Prometey’’, 49 Shpalernaja Street, 193015 St.Petersburg, Russia Received 24 October 1998; accepted 12 November 1998 Abstract In the process of operation of RBMK reactors damage took place on welded pipings produced from austenitic stainless steel of the type 08X18H10T. The inspection of damaged sections in pipings has shown that in most cases crack-like defects were of corrosion and mechanical character. This paper considers in detail the reasons for the damage appearance and their development for this type of welded joints of downcomers B325 × 16 mm, which were fabricated from austenitic stainless steel using TIG and MAW welding methods. q 1999 Published by Elsevier Science Ltd. All rights reserved. Keywords: Intercrystalline stress corrosion cracking; Welded joints; Corrosion factor 1. Introduction The type RBMK reactor is a graphite moderated, channel type, direct cycle boiling water nuclear reactor with online refueling. At present there are 15 RBMK reactors in opera- tion: 11 units in Russia, two in Ukraine and two in Lithua- nia. The start-up of these units took place from 1973 (Leningrad 1) to 1990 (Smolensk 3). At the beginning of the 1980s after some operation time a lot of damage was revealed in the primary circuit piping. During normal opera- tion the internal pressure is 8.6 MPa and the temperature of the coolant is 2708C. Austenitic stainless steels are used for the primary circuit piping manufacture. Austenitic stainless Cr–Ni steels of the type 18-10 are widely used in power engineering because of their high resistance to general corrosion. The rate of general corro- sion for these steels does not exceed 0.01 g/m 2 h under conditions of boiling reactor operation. This provided a reli- able, stable operation of reactors, because there was no problem of precipitation and removal of metal oxides from heat exchanging surfaces of reactor channels. In the first place damage appeared on Dy-752 piping welded joints of the main forced-circulation circuit, which were fabricated from low carbon steels of the type 22K (Fig. 1(a)), cladded inside with austenitic metal (thickness of cladded metal is equal to 5 mm). The cause of such a damage was production defects in the weld root which were enlarged during 3–5 years and reached the inner surface. As a result, a contact of a defect with coolant took place so that the defect then developed with a corrosion mechanism [1,2]. Such defects propagate very quickly and it was necessary to solve the problem of their immediate removal and repair. As the cause of the defects’ appearance was related to the welding technology in the process of field welding (formation of martensitic structures in weld root), one succeeded to avoid production defects by a correct sequence of bead fabrication in weld root (at first with pearlitic electrodes YONI-13/45A from the side of the basic weld part and then with austenitic electrodes EA-395/9 (the first layer) and EA-400/10T (the second and third layers) from the cladding side). As it can be seen the repair technology differs in welding materials from the production technology (see Fig. 1(a)). During the operation of welded joints of pipings (B752 × 38 mm) repaired with such tech- nology neither corrosion nor mechanical damage was revealed. At the same time for a number of Russian NPPs numerous cases of crack-like damage were revealed near welded joints in pipings (downcomers B325 × 16 mm), manufactured from the stabilized austenitic steel 08X18H10T (Fig. 1(b)), after their operation for 50 000–100 000 h at various units. As was determined by X-ray inspection, ultrasonic testing and metallographic examination, these cracks initiated from concentrators at the inner surface of tubes in the vicinity of the fusion line of the weld root with base International Journal of Pressure Vessels and Piping 76 (1999) 299–307 IPVP 1911 0308-0161/99/$ - see front matter q 1999 Published by Elsevier Science Ltd. All rights reserved. PII: S0308-0161(99)00005-8 * Corresponding author. Fax: 17-812-274-1707. E-mail address: [email protected] (B.T. Timofeev)

Upload: bt

Post on 03-Jul-2016

218 views

Category:

Documents


4 download

TRANSCRIPT

Page 1: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

Corrosion and mechanical strength of welded joints of downcomersfor RBMK reactors

B.T. Timofeev* , G.P. Karzov, A.A. Gorbakony, Yu.K. Nikolaev

CRISM ‘‘Prometey’’, 49 Shpalernaja Street, 193015 St.Petersburg, Russia

Received 24 October 1998; accepted 12 November 1998

Abstract

In the process of operation of RBMK reactors damage took place on welded pipings produced from austenitic stainless steel of the type08X18H10T. The inspection of damaged sections in pipings has shown that in most cases crack-like defects were of corrosion andmechanical character. This paper considers in detail the reasons for the damage appearance and their development for this type of weldedjointsof downcomers B325 × 16 mm, which were fabricated from austenitic stainless steel using TIG and MAW welding methods. q 1999Published by Elsevier Science Ltd. Al l rights reserved.

Keywords: Intercrystalline stress corrosion cracking; Welded joints; Corrosion factor

1. Introduction

The type RBMK reactor isagraphite moderated, channeltype, direct cycle boiling water nuclear reactor with onlinerefueling. At present there are 15 RBMK reactors in opera-tion: 11 units in Russia, two in Ukraine and two in Lithua-nia. The start-up of these units took place from 1973(Leningrad 1) to 1990 (Smolensk 3). At the beginning ofthe 1980s after some operation time a lot of damage wasrevealed in theprimary circuit piping. During normal opera-tion the internal pressure is 8.6 MPa and the temperature ofthe coolant is 2708C. Austenitic stainless steels are used forthe primary circuit piping manufacture.

Austenitic stainless Cr–Ni steels of the type 18-10 arewidely used in power engineering because of their highresistance to general corrosion. The rate of general corro-sion for these steels does not exceed 0.01 g/m2 h underconditionsof boiling reactor operation. Thisprovided areli-able, stable operation of reactors, because there was noproblem of precipitation and removal of metal oxidesfrom heat exchanging surfaces of reactor channels.

In the first place damage appeared on Dy-752 pipingwelded joints of the main forced-circulation circuit, whichwerefabricated from low carbon steelsof thetype22K (Fig.1(a)), cladded inside with austenitic metal (thickness ofcladded metal is equal to 5 mm). The cause of such a

damage was production defects in the weld root whichwere enlarged during 3–5 years and reached the innersurface. As a result, a contact of a defect with coolanttook placeso that thedefect then developed with acorrosionmechanism [1,2]. Such defectspropagatevery quickly and itwas necessary to solve the problem of their immediateremoval and repair. As the cause of the defects’ appearancewas related to thewelding technology in theprocessof fieldwelding (formation of martensitic structures in weld root),one succeeded to avoid production defects by a correctsequence of bead fabrication in weld root (at first withpearlitic electrodes YONI-13/45A from the side of the basicweld part and then with austenitic electrodes EA-395/9(the first layer) and EA-400/10T (the second and thirdlayers) from the cladding side). As it can be seen the repairtechnology differs in welding materials from the productiontechnology (see Fig. 1(a)). During the operation of weldedjoints of pipings (B752 × 38 mm) repaired with such tech-nology neither corrosion nor mechanical damage wasrevealed.

At thesametimefor anumber of Russian NPPsnumerouscasesof crack-likedamagewererevealed near welded jointsin pipings (downcomers B325 × 16 mm), manufacturedfrom the stabilized austenitic steel 08X18H10T (Fig.1(b)), after their operation for 50 000–100 000 h at variousunits. As was determined by X-ray inspection, ultrasonictesting and metallographic examination, these cracksinitiated from concentrators at the inner surface of tubesin the vicinity of the fusion line of the weld root with base

International Journal of Pressure Vessels and Piping 76 (1999) 299–307

IPVP 1911

0308-0161/99/$ - see front matter q 1999 Published by Elsevier Science Ltd. Al l rights reserved.PII: S0308-0161(99)00005-8

* Corresponding author. Fax: 17-812-274-1707.E-mail address: [email protected] (B.T. Timofeev)

Page 2: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

metal. Cracks also propagated along the heat affected zone(HAZ), i.e. within the narrow zone (0.3–1.0 mm) near theweld fusion line. Sometimes a crack started in a weld, buthere it did not further develop and again returned to theHAZ. The fracture surface of these cracks is similar tothat of the intergranular corrosion fracture (further definedas intergranular stress corrosion cracking—ISCC) [3–5].The same fracture modes were observed in welded jointsof instabilized austenitic steels applied at NPPs in westerncountries [6–8] and resulted in their replacement. The ISCCof welded joints was found both in steels without stabilizingadditions of titanium and niobium and in steels with thesestabilizing additions. In the latter case ISCC can be revealedonly after a prolonged operation time.

The aim of the present research was to clarify the natureof the ISCC, to determine its mechanism, to develop recom-mendation for the prevention of such failures. The presentreport is a review on investigations performed at the insti-tute. Earlier [3–5] it was shown that the ISCC of weldedjoints fabricated from 08X18H10T steel depended on threemain factors:

• level of acting stresses in the structure;• purity of circulating water in the circuit on the oxygen

content in it;• sensitization degree of the HAZ metal of welded joints.

2. Investigated materials

For the aforementioned piping production one usuallyobtains seamless pipes from austenitic steel of the type

08X18H10T. In this case, it is recommended to use tubeswith the carbon content in the metal being not more than0.08% and a titanium content of not less than 0.4%.

The mechanical properties of the tubes’ metal must meetthe definite requirements of yield strength, reduction of areaand grain number. The yield strength at the elevatedtemperature 3508C to exceed 170 MPa, ultimate tensilestrength at the same temperature to exceed 350 MPa, andthe reduction of area at room temperature to exceed 37%.The grain number according to GOST 5639-82 should notbe more than four. All tubes should be heat treated andtested on resistance to ICC according to GOST 6032-89with a provocating heating. Tubes are delivered with elec-trochemical polished surfaces or machined inner surfaceswith a surface finish of 6.3mm. After the tubes (B325 ×16 mm) were bent, they were subjected to heat treatment.The edge preparation (see Fig. 1) according to PNAE G-7-009-89 [9] was used for the welded tubes.

Two welding methods were used for the production ofthese pipe-lines (MAW and TIG). Electric arc manual weld-ing was performed with EA-400/10Y or EA-400/10T elec-trodes (OST 5.9370-81). Manual welding and automaticnon-consumable electrode arc welding (TIG) wereperformed with Sv-04X19H11M3 wire (GOST 2246-70).Tentative piping welding regimes are given in Tables 1and 2.

The investigations of the mechanical properties of down-comer welded joints were carried out both in as-weldedconditions and after a prolonged operation. The test resultsshowed the degradation of welded joints’ properties after120 000 h service did not practically occur. The degradation

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307300

Fig. 1. Sketches of welded joints of multi-forced circulation circuit piping for RBMK reactors: (a) Dy-800 pipe-line; (b) Dy-300 downcomers.

Table 1Tentative regimes of manual electric arc welding of pipingB325× 16 mm

Electrode diameter (mm) Value of current (A)

3 60–904 110–1405 120–160

Table 2Tentative regime of manual and automatic TIG welding for tubesB325×16 mm

Welding type Value of current (A) Argon consumption, min21

Root bead Other beads Into torch Preliminary

Manual 45–90 90–100 8–10 4–5Automatic 50–95 115–135 — —

Page 3: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

is observed only in local areas (directly at the tip of thepropagating crack in the HAZ of the welded joint) andcan be detected using special test methods, for exampleslow strain rate testing (SSRT).

3. Level of acting stresses1

In the process of operation, stresses from internal pressuresupports weight as well as forces from the self-compensa-tion of temperature transport of tubes, drum-separator andsuction manifold are superimposed on residual stresses ofpiping stated (post-welding heat treatment was not carriedout). Residual welding stresses at the inner surface along thefusion line near the weld root in theB325× 16 mm pipingwelded joint are tensile and equal to (0.3–0.4)YS. As it canbe seen from the stress distribution through the wall thick-ness (Fig. 2) the level of residual tensile stresses along theweld fusion line at a depth of 2 mm from the inner surfaceincreases after welding to the value (0.6–0.7)YS in compar-ison with the value of residual stresses on the inner surfacenear the weld root. At a depth of 2–3 mm some reduction inthe residual stress (0.3YS) is observed, and at a depth of 3–4 mm the second peak of the residual stress growth can beseen, which is 1.5–2 times less, than the first one. Beginningwith a 4-mm wall thickness of the tube welded joint (and

further), the reduction in residual stress takes place, andstarting with a 7-mm tube wall thickness the residual tensilestresses change to compressive ones. However the decreasein stress continues till the wall thickness reaches 12 mm. Atsuch a depth the compressive residual stresses are equal to0.3YS. In this case the third peak of the residual stressgrowth (0.5–0.6)YS appears as a result of which the tensileresidual stresses appear again at this depth, and theydecrease practically to zero on moving to the outer surfaceof the tube welded joint. After the power unit start-up thedistribution of summary stresses through the thickness ofthe welded joint remains invariable in its form, but in itsstress value increases 2.0–2.5 times. After the unit isstopped the distribution of the aforementioned summarystresses through the tube wall thickness takes up an inter-mediate place between the curves described.

In the straining cycle the level of maximum axial tensilestress (with consideration of residual stresses) exceedsconsiderably the yield strength in the regime’s start–stop.The distribution of normal stress on the fusion line in the as-welded condition, after start and stop (obtained in ourlaboratory by Karzov et al. [4]) is given in Fig. 2. By this,most high stresses were generating in the weld root in theprocess of hydraulic tests (HT), which exceed stresses undernormal operating conditions (NOC) by 1.5–2.0 times (Fig.3). Thus it is seen that the stress level in the welded jointthrough its wall thickness is favourable for intercrystallinecorrosion cracking processes. Besides, because of a greatmargin of the elastic energy in the circuit of multiple forced

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307 301

Fig. 2. Distribution of normal stress component on fusion line in as-welded condition (1), in stationary regime (2) and after stop (3).

1 All calculations of strain–stress state were carried out by Dr. V.Kostylev.

Page 4: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

circulation the reduced distribution of stresses does notchange very much with the generation of the first cracksin piping welded joints.

4. Water chemistry regime

In connection with the fact that dissimilar metallic mate-rials (Zr-based alloys, corrosion resistant steels of the type08X18H10T, chromium alloyed and carbon steels) contacteach other in boiling reactors it was accepted that theyshould use a constant neutro water chemistry regime. Thiswas motivated by the fact that a decreased chloride contentis necessary for a reliable operation of the equipment andthe piping, manufactured from corrosion resistant austeniticsteels. As for Zr-based alloys the alkaline regime is imper-missible because of the corrosion cracking of these alloys[10].

In the process of NPP operation, the control of the follow-ing indexes of the water chemistry regime is carried out:conductivity, the Cl2 content, Fe, Cu, hardness of water,pH. The average index of the water quality in the circuitof multiple forced circulation at NPPs, which were exam-ined by our specialists, were maintained within the norma-tive values. The oxygen content in the circulating water isnot a normalized index, but as the NPP operation experienceshows in the process of water radiolysis, oxygen and hydro-gen are given out in large amounts in a steam phase. Usingsamples of circulating water, taken at various periods of theNPP operation, it was determined that the oxygen content inwater of the circuit of multiple forced circulation in thestationary regime is usually equal to 0.03–0.05 mg/kg.However in the process of planned repairs or reactor stop-page for other reasons, the oxygen content in the circulatingwater increases sufficiently. During a prolonged stop theoxygen content in water can reach 8 mg/kg.

The metallographic studies of welded joints ofB325 ×16 mm tubes showed that the thermal effect of welding onthe HAZ of 08X18H10T steel was associated with solution

and carbide formation as well as with grain growth. As aresult, intercrystalline cracks form and propagate along thefusion line recrystallized zone with a width of 0.2–0.6 mm,until in its absence the crack growth ceased. Therefore, theinfluence of technological factors was estimated in accor-dance with the criteria stated earlier. The sensitization wasdetermined metallographically on the basis of etching abil-ity of boundaries, containing chains of fine-dispersedcarbides, precipitated on them.

Investigations were carried out on specimens of weld-ments before and after their operation and also on specimensof experimental welded joints. On specimens made fromweldments after their operation, crack-like defects arelocated along the weld root at a distance of 0.5 mm andfrom the fusion line at a depth of 5–6 mm. These growfrom the inner surface of tube and look like branched inter-crystalline cracks. Throughout their extent the metal failedon a brittle mode. A supposition was made that cracksgenerate as a result of corrosion cracking in an intercrystal-line mode for which the necessary conditions are as follows:the stress level—above the yield strength, oxygen concen-tration in water—within the range 0.1–8 mg/kg (the latter ischaracteristic of a period of maintenance and units puttinginto operation) and sensitization of grain boundaries (i.e.chromium concentration reduction on grain boundaries asa result of Me23C6 carbide formation).

Piping welded joints after operation does not exhibit atendency to ICC on AM GOST 6032-89 [11] method.Thus, close to weld zone sensitization it appears sufficientfor ISCC, but insufficient for ICC proceeding. The compar-ison of sensitization degree after operation with that of theas-welded condition shows that grain boundary sensitizationoccurs by welding (initial) and develops at a low tempera-ture (2908C) in the process of operation. Sensitizationduring welding occurs as a result of being close to theweld zone, heated to temperatures 1200–13008C, as a resultof partial solution of titanium carbides and carbon fixing insolid solution, chromium carbides forming in the process ofcooling after welding and repeated heating within the

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307302

Fig. 3. Straining cycle in the weld root in the regimes: normal operation conditions-(start–stop) and hydrotests.

Page 5: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

temperature range 500–6508C. Repeated heating takesplace by producing usual welding beads. A final sensitiza-tion depends on the initial one. The greater the initial sensi-tization the more rapidly its level is achieved, by whichISCC begins.

On the basis of the Strauss data, the stationary potential ofthe solid solution Fe–Cr in 1 N FeSO4 solution with a Crcontent of 6–8% is equal to20.6 V, and with a Cr contentabove 12% it is10.22 V. In the process of the NPP opera-tion because of carbide growth in the HAZ of welded joints,depletion of Cr occurs by a thickness, which does notexceed 0.1 mm. This narrow zone becomes active but atthe same time the rest of the surface is passive (see Fig. 4).A small surface of the anodic zone near the weld root area(which exceeds the cathodic zone by more than 10 times)promotes an intensive corrosion failure of zones depleted ofCr and the generation of the first surface cracks under theinfluence of high tensile stresses. The stated character ofcrack nucleation is observed when the oxygen content inthe circulating water is not less than 0.2 mg/kg [12].

The sensitization degree of welded joints can be assessedon the width of etching bands near the weld zone [12], or onthe failure depth of grain boundaries on the microsection;hence, it is assumed, that the indicator of ICC resistance in

metallographic studies is damage of metal grain boundariesto a depth of not more than 30mm (GOST 6032-89) [11]. Inmost cases the reason for sensitization is the chromiumcarbide growth in welded joints in the process of pipingoperations is explained by the presence of free carbon ingrain boundaries. Therefore to reduce the carbon content,the addition of stabilizing titanium should not be (5–6) foldless than the carbon content in steel. To combine carbon inoriginal carbides of the type TiC, the cooling rate of weldedjoint metals should be higher than the reaction rate

Ti 1 C� TiC:

During cooling of the welded joint in air, this condition ismet and practically all carbon is combined with originalcarbides. In the provocating temperature range 700–4008C(in the case of carbon being combined with carbides) thecooling rate is insignificant. All these requirements arefulfilled when 08X18H10T steel welding is carried outwith a minimum input of thermal energy.

In intensive regimes and with cooling rate increasingfrom 1300 to 9008C, a considerable amount of free carboncan form; this carbon in the provocating temperature rangecan lead to the formation of nuclei of secondary carbides ofthe type Cr23C6. A great amount of free carbon induces

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307 303

Fig. 4. Variation of stationary potential of Cr–Ni steel in 1 N FeSO4 solution depending on temperature by various carbon content in steel. Temperingduration—10 min.

Page 6: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307304

Fig. 5. ISCC crack view on welded joint ofB325× 16 mm downcomers.

Page 7: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

carbide growth and formation of zones depleted of Cr,which causes metal sensitization in the process of operation.The crack generated at the inner surface of the tube near theweld root zone propagates under a simultaneous effect oftensile stresses and corrosive water environment, enrichedwith oxygen. The crack propagates with a decreasing rateand the presence of crack tip blunting with a considerableincrease of its curvature (rounding off) radius point this fact(Fig. 5). The crack development occurs discretely. Theapplication of high tensile stress during hydrotests givesimpetus to each new crack propagation.

Immediately after hydrotests and using NPP in theregime, the crack development, proceeding with a decreas-ing rate, is induced by a mutual action of tensile stressesand corrosion failure of metal depleted of Cr as a resultof contact corrosion with oxygen depolarization.However, while operating at maximum power the oxygencontent in the circulating water decreases to 0.03–0.05 mg/kg and the corrosion process with oxygen depolar-ization ceases. In this case pH in the crack reduces to three.By this, corrosion at the crack tip proceeds, but withhydrogen depolarization causing the crack tip blunting, asmentioned before. In order to resume the corrosion failureprocess with oxygen depolarization it is necessary to bringoxygen to the crack tip (and it takes place duringhydrotests).

A trend has been observed—with the increase in crackdepth its growth rate decreases, that can be induced by thedifficulty of bringing oxygen to the crack tip. The averagecrack propagating rate in depth on tubesB325× 16 mm isequal to 1024 mm/h. Initially, the crack growth rate is 20–30% greater; and for the crack depth 5–9 mm it is 15–20%less.

5. Sensitization degree of austenitic welded joints

In connection with corrosion damage of welded joints inpiping made from 08X18H10T steel of NPP with the typeRBMK reactors, the influence of technological factors closeto weld zone microstructure was considered. The effect ofrepeated passes, welding method and intermediate coolingof beads was assessed. The metallographic studies ofwelded joints ofB325 × 16 mm tubes showed that thethermal attack of welding electric arc close to weld zoneof 08X18H10T steel was associated with solution andcarbides formation as well as with crystalline cracks form-ing and propagating adjacent to the fusion line recrystallizedzone having a width of 0.2–0.6 mm, and in the case of itsabsence the crack growth ceased. Therefore, the influence oftechnological factors was estimated in accordance with thecriteria stated earlier. The sensitization was determinedmetallographically on the basis of etching ability of bound-aries containing chains of fine-dispersed carbides, precipi-tated on them.

5.1. Effect of repeated passes

By welding with many passes, a repeated thermalinfluence on the near weld zone from subsequent passes ispossible. The analysis of experimental data [13] of maxi-mum temperature distribution by manual welding showsthat steel heating to the temperature 11508C and above ispossible at a relatively small (to 1.0 mm) distance from thefusion line, and heating to the temperature 6508C andabove—at a distance of 5.0 mm. The published data [14]show, that at these temperatures the most intensiveprocesses of grain growth take place which are associatedwith a collective recrystallization and accordingly with Cr-carbides (of the type Me23C6) formation. In this connectionthe grain growth zone from the fusion line as a result ofsubsequent passes should not be more than 1 mm. Experi-ments showed that it was less than 0.6 mm and the carbidezone formation was spreading through the whole section ofthe previous bead, the width of which did not exceed5.0 mm.

Thus, the thermal influence of subsequent beads close toweld zone of previous beads on the grain growth criterion ispartial or complete through the section of the previous beadfor growth processes of the type Me23C6 carbides. As aresult, in multiple-pass welded joints along the fusion linea solid zone of recrystallization can be formed, which ismost favourable for corrosion cracks growth and interrup-tive near beads butts, which are associated with a repeatedheating. The repeated thermal influence of subsequent beadsalso induces the formation (growth) of the type Me23C6

carbides.

5.2. Effect of welding modes

Welding modes can influence significantly the maximumtemperatures and heating time close to the weld zone.According to Eq. (1) the reduction of relative energy ofwelding arc (heat supply) is achieved by welding current(Iw.c.) and arc voltage (Uar) decrease and also by weldingspeed (Vwel) increase

q� 0:24�Iw:c:Uar�=Vwel: �1�However, an excessive limitation of welding modes is

inadmissible because of the instability of welding arccombustion process and the possibility of production defectsformation. Therefore, the welding current is specified by thenormative documentation and depends on wire electrodediameter; arc voltage varies inconsiderably by manual weld-ing. Welding speed should be limited at the expense ofnarrow beads production.

Thus, the heat supply reduction means of the variation ofwelding conditions is quite limited and consequently itcomes down to the meeting of the normative documentationrequirements. As the experimental results showed, in thecase of welding in increased current regimes, the grain

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307 305

Page 8: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

growth zone width and spreading along the weldment heightincreased which can be sensitized completely.

5.3. Effect of welded joints types

One-sided welded joints of tubes (without backing rings)were fabricated using glass-like edge preparation; the weld-ing of the first pass was carried out by argon arc method withnon-consumable electrode (TIG) and without filling wire.At present in case of repair, V-type edge preparation is used(according to PNAE G-7-009-89 ‘‘Basic Welding Regula-tions’’ [9]). By this, tubes are assembled with a 1.0–2.0 mmclearance and the first pass is performed with filling wire. Inour opinion, the advantage of this technology is that electricarc heat is consumed for the melting-down of not only edgesbut also welding wire. The wire application permits one toprovide the ferritic phase present in deposited metal andexclude the possibility of hot crack generation. Proceedingfrom geometrical dimensions the height of the first three tosix beads by filling of V-type edge preparation is greater incomparison with the U-type preparation and reduces thesensitization close to the weld root zone depending onsubsequent beads. The mechanical treatment of tube edgesis also easier. Therefore, in the future such type of weldedjoints can be used not only under assembly but also underplant conditions. However, the standard metallographicstudies of specimens showed that the application of theearlier stated edge preparation does not exclude a possibilityof sensitization close to the weld root zone.

5.4. Effect of the welding method

The piping welding was carried out by two methods—manual TIG and a combined method (root—TIG, filling—covered electrodes). Sometimes automatic TIG is used. Thecomparison of manual TIG with covered electrode weldingshows that with the use of welding wires and electrodeshaving the same diameter, these welding methods areequivalent in heat supply (q), but differ in the depositioncoefficient:

acl � Qcl=Iw:c:t; �2�whereQcl is the deposited metal mass, andt the welding

duration.By covered electrode weldinga cl � 12 g/A h and is

greater by 2–4 times than that by manual TIG. In thisconnection with the values of welding current, covered elec-trode welding provides a greater amount of deposited metaland a smaller number of beads and, consequently, a lowerheat influence close to weld zone. By submerged arc weld-ing the coefficienta cl � 15 g/A h. However, this weldingmethod is characterized by a higher welding regime (Iw.c. .300 A) and can induce sensitization close to the weld zonedirectly by the production of each bead. The semi-automaticand automatic MIG havinga cl � 16 g/A h and permitting

the increase of welding output with minimum heat applica-tion are also to be considered.

5.5. Effect of metal cooling after each pass

In order to increase the resistance to intercrystallinecorrosion by the production of multi-pass welds in austeniticsteels, it is necessary to cool deposited metal after each passto a temperature not higher than 1008C. Experiments showthat specimen welding without cooling (with ‘‘overheat-ing’’) induces an increased sensitization close to the weldzone, especially the weld root.

A forced cooling of specimens from the inner and outersurfaces of tubes with water which influences favourably thestructure close to the weld zone was not established andgrain growth was observed only by the production of thefirst pass, the welding of which was carried out withoutwater cooling.

However, the sensitization degree depends not only onwelding technology but also on 08X18H10T steel tendencyto the formation of Cr carbides. Thus, in the case of thepresence of coarse carbides precipitation on grain bound-aries, a forced cooling with water and welding in minimumregimes (Iw.c.� 60–80 A) does not exclude sensitization andgrain growth (close to weld zone) of specimens welded byTIG. Therefore the corrosion resistance of such weldedjoints to intercrystalline corrosion can be increased byone-pass cladding on the inner surface of 30–40 mm widetubes from ends, using corrosion resistant welding materials(the type EA-898/21B electrodes and Sv-04X20H10G2Bwire).

It was established that steel, neighbouring a crack,becomes sensitized. The sensitization degree at the innersurface of a tube, close to the weld root is equal to 60%.The maximum sensitization is observed at a distance of notmore than 1.0 mm from the inner surface and was equal to80% for one specimen tested. For two other welded jointsthe sensitization at the inner surface was about 70%. As wemove away from the inner surface, the sensitizationdecreases. In all the cases a crack occurs in the sensitizationzone and ends with the sensitization degree of 30–40%before the metal layer, where it is absent.

6. Conclusion

1. Formation and development of ISCC of welded joints ofpiping B325 × 16 mm of the circuit of multiple forcedcirculation is a result of damage accumulation from theeffect of three basic factors: mechanical, corrosion andmicrostructural.

2. The fraction of the mechanical factor is determined bythe value of elastic energy margin, presence of tensilestresses, exceeding the yield strength; the fraction ofcorrosion factor—oxygen content in circulating water

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307306

Page 9: Corrosion and mechanical strength of welded joints of downcomers for RBMK reactors

with a concentration of not less than 0.2 mg/kg; the sensi-tization fraction of HAZ welded joints is determined bythe correlation of grain areas and depleted of Cr zones ofits near-boundary areas.

3. The rate of HAZ failure corrosion processes at the innersurface of welded joints ofB325× 16 mm downcomersachieves its saturation already with an oxygen content of0.2–0.3 mg/kg and a further increase in oxygen contentin the circulating water influences slightly the duration ofthe incubation period of ISCC.

4. The process of ISCC crack development is discrete. Eachinitiation of crack growth takes place in the process ofhydraulic tests. A trend of crack growth rate decrease isobserved with the increase of its depth.

5. It is possible to suppress the nucleation and growth ofISCC cracks completely or partially by means of elim-inating the influence of one of these factors. As it ispractically impossible to suppress the influence of eachfactor, it is necessary to attain the optimisation of allfactors and that it also can increase the incubation periodof ISCC and reduces the rate of its development.

References

[1] Timofeev BT, Vinogradov RP, Generalova SP, Chernaenko TA. Frac-ture resistance for piping materials of BWR nuclear power plants ofthe RBMK type. International Journal of Pressure Vessel and Piping1992;52:303–311.

[2] Timofeev BT, Fedorova VA. Corrosion and mechanical strength ofNPP material welded joints. International Journal of Pressure Vesseland Piping 1995;64:25–42.

[3] Karzov GP, Petrov VA, Timofeev BT et al. Damage mechanism ofpiping welded joints made from corrosion resistant Cr–Ni steel in theprocess of a prolonged operation of the type RBMK reactor. Proceed-ings of the Fifth International Conference. St. Petersburg, vol. 3. 7–14June 1998. pp. 179–190.

[4] Karzov GP, Margolin BZ, Markov VG, Tereshchenko AG, YakovlevVA. The damages nature of tempered piping of RBMK reactor circuitof multiple forced circulation during operation and means to preventthem. Proceedings of the Fifth International Conference. St. Peters-burg, vol. 2. 7–14 June 1998. pp. 27–43.

[5] Nikolaev YK, Vinogradov RP, Zelenin YV. The influence of techno-logical factors on sensibilization degree of welded joints in NPPpiping manufactured from austenitic steel of the type 08X18H10T.Proceedings of the Fifth International Conference. St. Petersburg, vol.3. 7–14 June 1998. pp. 201–205.

[6] Smith E. Crack growth resistance methodologies for austenitic stain-less steels. International Journal of Pressure Vessel and Piping1996;65:327–333.

[7] Ishihara T, Matsushima S, Ohashi S. Effects of environmental factorsand stress on progress of stress corrosion cracks of type 304 stainlesssteel in high temperature water. Met. Corros. Proc. 8th Int. Congr.Mainz, vol. 7. 6–11 September 1981. pp. 1718–1723.

[8] Thomas RG, Keown SR. Stress relief of austenitic weldments—microstructure and mechanical property effects,mechanical behaviourand nuclear applications of stainless steel at elevated temperature.Proceedings of the International Conference, London, May 1981.pp. 30–38.

[9] PNAE G-7-009-89. Welding and Cladding. Moscow: Energoatomiz-dat, 1991.

[10] Gorbakony AA, Timofeev BT, Petrov VA. Damage mechanism ofpiping welded joints made from corrosion resistant Cr–Ni steel inthe process of a prolonged operation of the type RBMK reactor.The Fifth Meeting of Joint Research Programme between RIFTof Tohoku University and CRISM ‘‘Prometey’’ within Japan–Russia Collaboration for 1997–1999. St. Petersburg, 15 May 1998.p. 12.

[11] GOST 6032-89. Steels and Alloys, Corrosion Resistant. Test Methodson Intercrystalline Corrosion. Moscow: Publ. House of Standards,1991. 43 pp.

[12] Matteson E. Electrochemical Corrosion. Moscow: Publ. House Metal-lurgy, 1991. 197 pp.

[13] Anikovsky VV, Zabolotzky VM, Ivanova TI. Temperature and resi-dual strains distribution in HAZ by welding of technological samplesfrom austenitic steel. Svarka 1969;12:00.

[14] Parshin A.M. Microstructure, strength and ductility of stainless andhigh temperature steels and alloys applied in shipbuilding. Leningrad:Sudostroenie, 1972.

B.T. Timofeev et al. / International Journal of Pressure Vessels and Piping 76 (1999) 299–307 307