ISSN 0957-798X ‘Lop lb, tee ase ca um un ee ge dc= kamWEE war ~~ Fi English transiation of the monthly sAvtormaticheskaya Svarka+ journal published in Russian since 1948 Eaitor-in-Chiet B.E, Paton International Conference «Welded Structures» Papers of Plenary Sessions soo: oF Est dail PATON 8.£ Modern trends toward increase in Strenath cand ite: ientst YuSBorsov vE.Grain | Steed suche: 2 eati- Aa. resskt ALYQSHIN Pana GORNAYA SP. Now apvaach othe technology ext Yuya. | ALYQSYNL HE, and GORA S Ne some 8 - | AYaJlshchenko GARF EF end NETREBSKY M.A. Ancosement of ine strength and sing TE Ronen CaBla feof ties or Chavon canon cata ne 13 sues cunionseraneno | MERLE Sod HOIMODE N! Eminaton of eracram sol atuchwer | titanin vant | goer Yale AY LEONOV UP and AL VENEVENY VA - 5 a0 Victposeor Uttiovanew | Seageslacnsoaies shéngth otveldad ots of GH srtures DVORETSKY V1 Evehaton of residual le of load-carrying weldes Sy | ViMaktinenko AAMazur | Structures and extension of ther service ile me 34 arty LPMojgov VF Moanin | ZUBCHEAKO AS, VASLCHENKO GS and DAAGUNOY VG sf eae 9 ak beta beak cone rowan Sate ths | onezareno vv.roonen | SPREE Sosyriacar omen M8 ne LiPokhodnya LARyabisey | ZUBCHENKOA.S., KHARINA 10 RUNOV AE sad MAKHANEV V.O conf Heenan ang tod ang estore woes ™ V.KSheles OFEBM roacto ee eee 45 YuASterebooen | KAASOUGKY Aa ad ORTINYAEE IY Rapmssment of esa he Gf welds phe ne fea Vase a mcle pee seo The ViLTruakow NMVorenai | Gatteged by hteleqtaling canceion voce Dower one 5A poe K.AYushehenko- KRIVOSHEEY Pj, Problems regarding welded structuses of the «Shelters and t eee Object at the Chomnobyt nuclear power staan 82 = cam KNYSH V.V. Determination of ayclic life of Structure ‘elements | AT Zelnichenko ea ate eases 6 LOBINOV Li. PYTORAK VA and RLVSHINSKY WG Diagnostics | Peeters eta an ate ran vt MORGTSR, and -eleciron spacWle- interferometry and shearagrapty ....~... 72 on WELDED STRUCTURES Pompe: a a Cc) Figure 2. Operational, as well as erica GA and tolerable (ll pressures caleulated with allowance for corrosion dabige 1 {vo sections of ll pipelines negbouring the congss0r plans P~"presute, Length of the pipeline section aiven to two approaches: pressure testing and io- pipe diagnostics, It should be admitted that the use of the first approach does not ensare the desirable performance of a pipeline. This is associated with the fact that long with reveating the most dangerous zones, it contributes to development of the less dangerous defeeis the effect of which on strength can show up fater with time. At the same time, in-pipe di- agnostics enables volumetne coreesion damaye to be detected and wall chickness of a pipe to be mea- sured at a high degree of accuracy and reliability However, detection of plane defects in welded joints involves certain difficulties, This makes it necessary to wse-a more precise method te estimate condition of a pipeline by the data af selective inspection of individual, most critical locations, i.e. hot zones. Liagastics of these anes is a sepa- rate sclentific-and-technical challenge. It involves number of problems associated with substantia- tion of the method for determination of the pitting locations, as well as other external NDT methods Importance of such problems is illustrated in Fi- gure 2, which gives an idea of the quantity and degree of danger of corrosion defects in one of the coil pipelines. Problems associated -with diagnostics of main pipelines are not limited just to results of in-pipe NDT. There are some aspects which are related Table 4. Fracture roughness of pipe metal (, ‘unstable condi both to estimation of permissible defuets and opti mization of repair of main pipelines. It should be noted that existing standards and literature data have substantial differences regarding determina tion af sizes of permissible defects, In addition, conditions causing growth of technological defects with Lime require further investigsstions, In par ticular, apart from the corrosion impact, i is ne= eessary to establish the exact conditions which ‘cause Propagation of fatigue cracks from conven tional technological defects. {fn evaluation of serviceability of pipelines itis necessary to take int-account: probable deygradation of service properties of metal as a result of ttsageins, with time (e.g. general and local clranges in fracture toughness associated with peculiarities of hydraulic and mechanical expansion of pipes, forming, ett.) Ignoring these peculiarities may lead to formation ‘of inadmissible crrors in estimation of crack rests tance of pipelines: ‘As proved by practice, conditions of operation of pipelines differing from the normal ones can cause a fundamental change in mechanical prope ties of metal. In this case the important factors are the stressed-strained state and environment. It should be noted that under the effect of these fac tors embrittlement of metal occurs in local zones, and it should not spread to all metal of the pipeline Propertics of metals are grcatly affected by plas tic deformation. It can occur in zones of desixn stress raisers (zones of welding of T-jointsor branch pipes to pipelines), different types of defects cracks, zones of lack of penetration or lack of fusion, dents, scratches, tool marks) and changes of pipe geometry. Results of investigation of the clfect exerted by the level of plastic deformation of tension e and curapression e2 on fracture tough= frss are shown in Table 1. Plastic deformation ‘causes a substantial decrease in fracture toughness. Thus, at a value of plastic deformation equal to 20 %, resistance of steel of the 17G1S grade to initiation of tough fracture decreases almost to the level of the brittle state. The lowest values of frac- ture toughness were fixed on sampies siunaating typical mechanical damage of metal of the pipes, ice. dents with shallow surface tears (cracks) Since most of welded structures operate under canditions of repeated-alternating loading, thet tigue is a very important problem. It determines at the stage of initiation of tough crack, 8 atthe slage of its transition to a depending upon the level of plastic deformation of tesion &y and compresin & Sie gade sm as — a i oe o . w 1761S 5 0.13 on 0.09 0.08 0.20 ont 0.04 ee ee ee | recast 8 nat asta aw : Roa yan as The ‘ AED ena. ———— * - malt 1012000and opti- should be ture data serminae addition, al defects “In par- it is ne. 8 which eonven= nes fe is ralation ageing fraccure ylraulic g,ctc.), ration ke resis 2a. Ma @ w ° a 2 oe @ WF oF or Figure 3. Faligue curves plotted for tubular connections at asjuiiericl loadigeyele” 1 — cabnertions trented by ule ‘onic peening; 2 — untreated eaanections (a, — load ampli the) the life of welded bridges, railway rolling stock, hoisting machines and apparatuses, off-shore sta~ tionary platforms, tower-type antenna structures, agricultural machines and many other structures This makes it necessary 6 give a serious considera tion to development of methods for increasing fa- tigue resistance of welded joints. An efficient means in this respect is the use of strengthening opera tional treatment of welds, especially ultrasonic peening. This method was elaborated from tech- nologies of surface plastic deformation of welds In addition, this treatment leads to relaxation of residual stresses in the weld and induecs favourable tensile stresses in the surface layers of the wetal Figure 3 illustrates the efficiency of ultrasonic peering of welded joints in terms of increasing fatigue resistance of tubular connections. Tt re- sulted in doubling the Fatigue limit and inercasing the Fatigue life by an order of magnicute. Premature formation of fatigue rvacks was de- tected in welded joints of solid-will superstrur tures of railway bridges, frames of ears und bodte of some types of railway locomotives and wayons, cranes and other structures at an early staxe of their operation. Most often it is the case of those structural members in which formation of such cracks was not expected and which were not de- signed for the effect of alternating loads, Fornvation of cracks is attributable ta ignoring the effect of focal stresses and vibrations caused by a groundless choice of calculation diagrams used in design, as well as other factors. This gives rise to the requi ment of development of efficient measures for de- celeration and arresting of cracks in members of the existing structures Figure 4 shows results of comparison of the efficiency of different wethnds used to arrest fa- tigue cracks. The most common and practicable method for arresting fatigue cracks is welding on preliminarily macle grooves in regions of structural members damaued by a crack. In the eases where repair of structural members by welding causes dif- ficulties, the method recommended for erack ar- 9-10/2000 ™p, WEB ava vasrencries € ae | ——— “Ty ote ae eee f 5 sabes esas ire Figure 4. Cyclic fatigue life of samples at R, = 0 and 6, © 155 MPa sbjecled to different methods of arresting offal cracks: 1" intial condiion: 2 — drilling of boles at crack apex, 3 — same plus bsalling high-seength bolts in ther; Satie averloutng to 240 MP3; 5 ~ local explosion tres nent. 6 local heating, 7 — repair welding of eacks resting is drilling of holes at the crack tips and installing into them high-stsenuth bolts, ur it may be some other comparatively easy-to-use method. based on inducing favourable residual compre: sve stresses ahcad of the erack tip by applying seauie overloading, local explosion treatment oF local heating. ‘The fundamental reserve far increasing veliabi lity und fatigue life of welded structures and, at the same time, decreasing metal consumption Ties in improvement of approaches to regulation of de sign loads and permissible stresses. Very topical are investigations of actual loads applied to welded connections. Basic causes of formation of extra stres.es consist in differences between operational oading and actual service conditions of a structure and those used in design calculation diagrams These differences are associated first of all with peculiarities of transfer of loads in structures, spe- cific redistribution of forces in structural connec- tions, mutual effects of members, use of tolerances for deviations of geometrical dimensions and vibra tion of individual structural components. Forexample, in beam superstructures of railway bridges fatigue cracks are formed in webs of main beams near breaks of transverse stiffeners (Fi- gure 3). Such a damage is caused by extra stresses formed as a result of interaction of members of the superstructure, as well as superposition of vibra tions in sections of the beam webs. To increase fatigue life of such connections, itis recomnimended that stiffeners are welded to the beam chord and. web with through penetration and the welds are subjected to ultrasonic pening. Results of testing, the experimental ring at VNIIZKT (Russia) at an ‘operating time equivalent to 20 years of service in 5> ‘WELDED STRUCTURES Figure 5. Attachayent of stiffeners to Inerease fatigue [ile of ‘beam superstructure of allway bridges: a ~ typial truce; 0 Special design ensuring fligue resistance: 1 fate cracks, 2 — transverse stiffeners; 3 ~ fillet weld with through penetration a railway network proved efficiency of this ap roach to upgrading of a structure, ‘Another pressing problem is development of sei entifically grounded approaches to extension of as- signed life tne of critical lnad carrying structures and ensuring their sale operation. Extension of life ‘of such structures leads to a fundamental saving, although it requires revision of specifications in force, Residual life can be increased through a more complete utilization of safcty factors used in design and allowance for differences in calewlated and ac tual operating time, The reserve for extension of, sorvice Life can be found om the basis of analysis of actual service conditions of a structure, fenerali- ation of operational experience, upgrading, repair and reconditioning of the weakest connections, im- provement of the design methods and use of new more substantiated criteria for evaluation of limi- ting states. Such an approach to praviding the yea ranteed residual life, aimed at extension of the total service life of structures, was tried out an super: structures of bridges, tanks for transportation of liquid carbon dioxide, travelling cranes and exea- vating machines. It is used currently for extension of life of load-carrying structures of traction rolling, stock in railways nf Ukraine, Rational utilization of high-strength martonsi- tic-bainitic steels with a yield strength of 600 ~ 1000 MPa plays an important role in solving, the problems associated with decreasing specific metal consumption and increasing operational reliabilicy and life of machines, mechanisins and engineering structures (able 2). The desirable combination of their properties is achieved at a carbon content of 0.10 = 0.17 % by alloying them with manganese, chromium, nickel, molybdenum and other elements (total content of alloying elements — up to 4 — 6%) and applying heat trealment. Main difficulties in welding high-strength ‘marteasitic-haiDitic steels are usually associated with the necessity to eliminate cold cracking of the sweld and HAZ metal and formation of structures that decrease brittle fracture resistance of welded joints. Special filler metals were developed for the hrasic are welding processes to produce welded jommts in such steels with strength equal to that of hase metal and prevent their cold cracking (Ta- ble 3). The absence of stress raisers and tow level of residual stresses in high-strength steel welded pints contribute to a satisfactory resistance te ini- tiation and propagation of fatigue and brittle frac- tures, Experience of manufacture and many-Year successful operation of a mumber of eritical welded stroctures made from high-strengih steels Chigh- capacity mining and oil facilities, super-high ton- nage track beds, building and hoisting machinery, engineering structures, ete.) are a conclusive proof of reliability and efficiency of the developed weld- ing materials and processes. Increase in reliability and extension of life of ‘welded structures depend upon improvement of the diagnostics methods, The method of acoustic emis: sion (AE) shows high potential in this respect Available are specialized procedures and devices based on the effect of AF formed in materials during their defarmatian and fracture Portable equipment. provides a reliable moni- toring of technical state of pressure vessels nai pipelines and other structures. Comparing At sig ‘Table 2. Chetnicsl composition and streasth progestics of high-srensth stels Tap tain su Scere *% & 2 eo Fd ea ae MeO Me ne eh Rae aa as a aa eas ae ranved apnea aces. eset. wid aun 0 won wee DooumnSiectie aizes eapeisn 2 poe me ee fugowsro—niacaat eacat tanetan 1 a we tRacmraveok ies ta-e4 Ia0eise = a8cay ogecew ws omen Wusmonowes —oigeayy Gavek hacia agus Gioia eae cee vm tyson tie ogee Leche sanee wooo wo IRowmenrunayea}-ais o2-es wy tale eo me paar 010850205, tana) ecu oo wo ye wie, 6 — 9-10/2000 RMALh martensi- tof 600 — wlving the cific metal reliability aineering ination of content of langanese, relements up to strength iated ing of the for the welded Figure 6. Interference feinges characterizing quality of a fiction stie welded joiat (material — aluminium alloy 6 ran thick) a= defeet-free region; B = region containing defect nals with actual defects proves the high level of confidence of the information obtained. The method is helpful in location of weakened zones at carly stages of failure. The end task is to estimate the residual life, the AE method is used currently for diagnostics of various industrial objects ‘The AE-based systems for continuous monitor- ing of performance of structures mecting inereascd requirements for safe operation can be ereated in tive future. Such systems are capable of solving the problems of evaluation of a technical state not only of ground objects but also, far example, of space stations by transmitting the information to the Earth. The use of movable diagnostic laboratories equipped with the basic AE devices is also very efficient, Methods of optical holography have high po tential for diagnostics of welded structures, The PWI developed procedures based on holographic interferometry for quality conteel and evaluation of stressed strained state of welded joints and struc- tures. Available are compact holographic devices, which are placed dicectly onan object investigated and require no special vibration isolation, Otfered are the automated systems for computer processing of holographic interference patterns. Helography is especially helpful for revealing hard-to-control defects in structural members and assemblies of metallic and non-metallic materials, Figure 6 shows results of inspection of a welded joint produced by a new advanced welding method, Le. friction stir ‘welding. This method 1s now used to advantage for welding critical structures of aluminium alloys “Table 3, Properties of welded joints i high strengs ets (0p » = 600 - S00 MPa) Fa, Figen ce Pa ‘Bite tramation ‘Automatic submerged-ate welding Welding w SeOskhn Se-lOKHN Hox AN’ Mechanized gas-sielded welding Manual electec ane welding usin covered electdes 70 and 85 ty) 9-10/2000 Electrodes of ANP-3, ANP-9, ANP- IO and ANP-AT grades 60-70 022-023 70-90 020-024 018-020 ™ Paton,> WELDED STRUCTURES However, the issue of ensuring reliableand efficient control of quality of welded joints i still open with this welding method the probability exists of formation of defects of the «atickings type, which are difficult to detect by X-ray and ultrasonic me thods, Holographic interferometry allows such «le gets to be revealed They are detieted by changes in the interference fringes compared with those registered in a defect free region of a welded joint In conclusion it is appropriate to note that find- ing solutions to important problems of increasing sisength, reliability and life of welded structures requires ‘active interaction of specialists from the ‘world-leading welding centres. NEW APPROACH TO THE TECHNOLOGY OF ULTRASONIC TESTING OF COARSE-GRAINED MATERIALS N.P, ALYOSHIN aad SP, CORNAYA NE. Bauman MSTU, Moscow, Russia ABSTRACT ‘The payer descsibes the solution of the problem of construction af an scousto-solidifieation model of coarse-grained materials, allowing forthe features of structure formation, geometrical and eneray factors of ultrasonic waves sutenustion. Analytial depeodeacies were derived, a program was prepared and some results of ealeulation of the passage and scattering of al types of US waves are given, A conelusion 4s made aa the aced to take into account the actual geometry and energy of the US beam for a cornet suterptetation of the results of UT of the abjects from materials ofthe above ype. Key words: ultrasome testing, weld, course grained materials, ergsalline sructure The main problems arising in performance of ul- trasonic testing (UT) of coarse-grained materials (grain size is commensurate with the US wave ‘ength) are considerable attenuation and high level of structural noise caused by scatteringat US waves in the polycrystalline material. At UY of coarse grained materials with an oriented structure cha: racteristic of austenitic welds, as well as the welds wade by the process of electroslag welding, castings, etc. these problems ate aggravated by deviation and distortion of the US beam, thus leading to incorrect interpretation of the contro! results, and very often also to the impossibility of its perfor- mance. ‘The existing procedures of UT of coarse- grained materials pravide a partial solution of this problem. as the majority of them do not take into account the actual crystalline structure of the ma- terials. Therefore, a procedure suitable for control of certain types of coarse-grained materials, may bbe unfit tor others. NE. Bauman MSTU performed the work al- lowing a new approach to the technology of UY of coarse-grained materials to be presented. Proceed- ing from analysis of the process of metal solidift cation, the weld or casting macrostructure js cal- . WEB ras culated and the solidification model is plotted Then, the US wave passage through a lon-uniform anisotsopie polycrystalline medium is studied and an acoustic model is constructed which permits the main factors of damping to be taken ito account, namely: attenuation, refraction on the fusion boundary and its transparency; deviation of the beam propagation from the wave normal and dis- tortion of the sound beam. The thus constrocted acousto-solidification inodel enables determination by calculation of the ‘optimal parameters and interpretation of the results of UT of coarse-grained materials with different polycrystalline structure. Material stracture, Three kinds of steuctares can be distinguished, namely coarse-grained with ‘equiaxial crystals (HAZ of welded joints, central part of ingots, forgings), coarse-crystalline with columnar crystals (austenitic welds, welds made by electroslag welding, ingots) and fine-crystalline (base metal of the welded joints, outer crust of ingots). Control of materials or zones of welded joints with a coarse-crystalline or coarse-grained struc= ture where energy lesses are found due to US wave scattering on the grain boundaries, deviation and distortion of US beams, is of special interest. Description of the solidification model. An elvetrosiag welded joint can serve as a typical ex- 9-10/2000ferference adefect= hat find- creasing, iroctures from the WELDEO STRUCTURES € S10 __X me S10 Xe Ber de 9 ° 3 1 7 2s as 50 25 50 sa Yom 2mm 0 ORs OS Oe K, o O25 B50 OFS K, a . a 8 Figure 4, Projections of a erystallite axis on the coordinate Figure 2. Chane of the angles of inclination ofthe ersiaite ses NOY (2), X0Z (6) C= 30 mm, p= 10 me, HTS sam, anf tothe coordinate anes dependiag an Ry YP Ca) and wd K.AZ/H (o) a0 ZH 05, 4/L 08, ¥/P 205 ample of production of the coarse-grained struc- other (Figure 3). In this ease, we assume that the ture. When the welding miodel is constructed, the grains are cubes with side a, and erystallites are solidification front, shape of the crystallite axes regular hexahedral prisms of length’ and trans- and change of the crystallite width with its growth, verse dimension a, which are in close contact with ae calculated. each other. The acoustic properties of such media Solidification front. The bcaadest class of weld are characterised by moduli of elasticity produced pools found in practice, are described by the equa- by averaging the moduli of elasticity of the single tion of an ellipsoid with semiaxes ?, p and h crystal by its possible orientations. Let us consider , the passage of plane US waves through such inho- ee ey mogencous media Pipe ‘Attenuation. To the general case, the coefficient of attenuation is made up by the coefficients of whese x, y, 2 are the current coordinate: J, p, atl absorption and scattering. In the metal with a poly- Heare the length, half-width and depth of the so- crystalline structure, the determiaant factor ts scal~ 2 2 lidification front, respectively. Introducing a fieti- tious ellipsoid [1], we obtain che design section of the weld pool of an ellipsoidal shape tering, therefore let'us only take into account the scattering of ultrasound in calculation of attenu- ation. Shape of the crystaitite axes. Calculation of Let as consider an individual grain (scatterer) the spatial position of the crystallites axes is theo which is located in a system of disoriented grains retically possible. Study [1] gives the basic solu- (environment). Let us represent the woduli of elias: tions for ealculation of the projections of the crys- ticity of such an inhomogeneous medium in the tallite axeson the coordinate planes. For weld pools form of of an eltspsoidal shape, using expressions : tive = yu = Chart By at re Vs 6k, 7 iy oye? vy Pow J Con = Bye BCyy at re V, where 8Cyuy = “iia — Chu is the difference in moduli of elasticity in both media; Wis the grain ‘volume; r is the radius-vector uf the point of ob- servation. Tt is known that the erystals of the ferritic-peas- litie and austenitic steels are characterised by where if, P, L are the semiaxes of the fictitious, ellipsoid; x= zy , we obtain the projections of the crystallite axes on planes XO¥ and XOZ (Figare 1), as well as angles c, B and y, made by the tangent to the erystallite axis with the coordi- nate axes (Figure 2) i Change of the crystallite width. Reduction of theerystailite croxs-section with their graweh lends itsolf to analytical assessment and depends o” the solidification schematic. Calculation dependencies are given in (11 Description of the acoustic model, fn orderto eoastruet the acoustic model let us conditionally a b® represent the coarse-grained! structure in the form igure 3, Models for dessn ofthe coarse-crystalline (a) and of a totality of grains disoriented relative to each eaatseained () seartaren S-a0/z000 RED cena.> WELDED STRUCTURES 5, d8/rmm 008 noe 002 0 1s Tara Figure 6, Change of he costceat of catering of US wanes, lepending on the angle of encidence on the crystallites (f~ 25Mile, a= 100 um, & = mm): SV, SH wave types moduli of elasticity of the cubie system. The moduli of elasticity of the environment are derived by ave- raging the moduli of elasticity of the single-crystal by the possible crystallographic orientations for the crystallites system (Figure 3, 4 a Ba = 1209 § Cia tod 9. a for the grain system (Figure 3, 6) chu = 170K f Jai dt. aa where is the angle of rotation of the erystallite relative to its axis; a and y are the angles of the grain rotation In onder to determine the scattored fields, let us use the volume integral equation of Lipman- Schwinger type. Let us caleulate the scattered fields U } in Born’s approximation (First approxi- mation of Lipman-Schwinger equation) using Green's function Gig (7, ry Upre uD Joye rn her Oh ar”, where fier", Uf) is the function dependent on the incident wave amplitude and describing the influence of inhomogeneities; r "is the radius-vee- tot of the source, s index denotes the satered eld Solving the problem [2 yielded analytical ex pressions for the coefficients of scattering for the coarse-erystalline structure which is the trans- versely-isotropic medium at the frequencies used in UT, as well as for the coarse-grained strucrure which is the isotropic mediuin. ‘The coefficients of scattering in the HAZ have an especially simple form characteristic of Raleigh scattering (8 10 8, d8/mm os 30 ° rr a Figure $. Dependence of the coefficient of scattering of an $7 type transverse wave on frequency (a™ 100 pa, 6 tim, 0 459 ~a® f 4). The weld metal unlike the metal of the HAZ, is not an equiaxial medium. The scaticrer length exceeds the US wave length. Seattering in this region does not obey the laws characteristic of the cquiaxial polycrystalline media. Calculation ‘has shown that in the range of frequencies for which the ratio of wave length to the transverse sinc of the crystallite exceeds m, the coefficient of seater ing im the weld metal approximately follows the 3 ~ af law (Figure 4) and depends on the direc~ tion of the wave incidence and its polarisation (Pi- ‘gure 5). For all types of the waves, the coefficients Gf scattering become zero in the cases, when the direction of wave propagation or (and) its polari- sation are parallel to the crystallite Jonger axi For the longitudinal and the horizontally-polarised transverse wave, the coefficients of scattering are maximal, if the wave propagates normal to the crystallite long axis. The maximum of scattering of vertically-polarised transverse wave was rexis- tered at the angle of incidence on the crystallites 0-45" Transparency of the fusion boundary. In UT of welded joints the problem of passage through the fusion boundary is of considerable importance as the establishment of the directions and determina- tion of the intensities of the waves transformed at the boundary are necessary for a correct interpre tation of the testing results. Considering that the ‘boundary media differ not only in the wave veloci- ties, but also in their crystalline structure, we have calculated the coefficients of transparency of the boundary of an isotropic (HAZ) and transversely- isotropic (weld metal) media (Figare 6). For the coefficient of transparency Dr (Brae) of SH type wave, the following analytical expression was de rived Di (Bane) = 209 £05 Bine/L1 + 0 (tet) C08 1, where v is the transverse wave velocity in the HAZ; © (aes) Is the velocity of the horizontally- polarised wave in the weld metal; Bixe is the angle WEAR ewan 9-10/2000Ot WELDED STRUCTURES € bros ep 075| re) 2s) a Figure 6. Cocticicats of causparency by ehe energy of ax atenitie weld fasion boundary of wave incidence on the boundary; arts the angle af refraction into the weld. The coefficients of transparency of the fusion boundary were numeri- cally calculated at the passage of longitudinal and vertically-polarised waves, When the coefficients of transparency of the fusion bcundary were calculated, the dependencies of the change of the elastic wave velocities on the direction of their propagation were investigated. ‘The weld material presented in the form ofa totality of disoriented crystallites, is characterised by the elastic constants of the hexagonal system. Solving the Christoffel’s formula for the waves propagating in the plane which passes through tlie sixth order axis of crystals of the hexagonal system, yields the values of velocities of three waves [3]. However, the equations for calevlation of the: velocities of the quasi-longitudinal and quasi transverse waves for engineering applications, are rather complex. Use of a coordinate system referenced to the wave vector, permitted derivation of simple enough cx- pressions for determination of «lie velocities af the. longitudinal 0, (age) and vertically-polarised transverse ty (Ging) waves In the transversely: is0- tropic medium 825 A sin ee Din 1 ve @ Wind = NC O44 A sin” Ga) ye where Gine is the angle of incidence on the crystal- ites: €1- Cr2, Cag nte the elastic constants of the cubic system crystal: = Cty + C12 ~ 2Cg4: p is the density The velacities of the waves calculated by equa- tions (1), (2) are given in Figure 7. Graphs (solid curves) calculated when solving Clwristoffel’s for- mula are given for comparison. Deviation of the direction of the beam propa gation from the wave normal arid beam distortion. In anisotropic materials the direction of eneray ¥-10/2000 WERE OR iva esky 00 ° 9 oa Om T a b Figure 7. Dependenctes of the velocities of the Jongituéina! ‘SHI (a) and transverse SU (5) waves on the aosle oF incidence on ee cele (hexageal yew) soi, shed are — walucsealeulsted by equations (1), (2) respectively 0888 OFT ag propagation does not coincide with the direction ‘of the US wave. This leads to deviation of the US beam from the wave normal, The maximal ampli tudes of particle displacement: in the wave which are the ones detected by the transducers, are ob- served in the direetion of the Umov vector. Procceding from analysis of the directions of the wave energy transfer (31, we derived the reli tionships which correlate the beam deviation with the elastic constants of the single-erystal and the direction of the wave vector: sin 28 (cos? @ ~ 0.75 si 26 — A (Cos! 0+ 2 sin® 8/2 atcla [A sin 20 sin” @ ~ 0.75 cos? ol Ay (8) = { | cy +74 six? 2018 Ay (@) = arctg {4 sin 26/18 Czy + 2A sin? Oh, where @ is the angle between the wave vector and the exystallite axis; r 0, and A indices pertain to the longitudinal, vertically- and. horieantally-po- larised transverse waves, respectively “alculation of angles Ae is given in Figure 8. We see that the greatest deviation of the wave -xcewoe from the direction of energy transfor (up to 33°) is found in the vertically-polarised transverse wave at the angles of incidence un the crystallites of 30 and 0°. The maximal deviation of the lon- 46, 488 ‘35 Figure $8 Deviation A. of the acoustic axis from the wave form!) in the retal af an austenitic weld, depending on the direction of Incidence on the erystalite —> WELDED STRUCTURES Simm ‘J Figure 9. Model of deviation and distortion of a US beam of the transverse wave in the eld, Sis the metal thichuess gitudinal wave is 10°, that of the horizontally-po- Tarised wave 12° Beam deviation from the wave normal leads to the sound beam distortion. As a result, focwsing and defocusing of the sound beam can be observed when passing through the weld. This is the most pronouniced in passaye of a transverse wave with vertical polarisation (Figure 9) Noise generation. Scattering of US waves in an austenitic welded joint, in addition to attenu- ation of the intensity of the valid signal, leads to generation of structural noise, which hinders the UT performance. At the ratio of valid signal and noise intensities below 1.5 to 2.0, defect detection is extremely difficult. This results in lowering of the control validity. From the equations of scat- tered fields, we obtain the dependence of the in tensity of structural noise Jy on polar angle 7 be tween the direction of US wave emission and the direction of observation In (W/™ = c08? e) The dependence was derived for the problems of LT of welds when the emitting and the receiving transducers are located from the same side of the item in one point (combined schematic) or ace in poirts close to each other (transmitter-receiver schematic), The angles of introduction of the emit ter and the receiver are the same. The expression ie |) oe 1a 20| 30 Sys am Figure 10. Model of distortion ofthe front ofa transverse US wave it af electrosalg welded jain is valid for any type of plane elastic waves. Analysis, of equation (3) shows tbat the intensity of struc- tural noise is maximal at y= 0 (back seattering) andat y= 180". At the angles of observation of 70 to 110° with the emission direction, the minimum of structural noise intensity Was registered CD / Tt value does not exceed 0.1). ‘The above dependencies were used in prepara- tion of @ program of computation of an acousto- solidification model. The grain size, inelination of the fusion boundary, size and inclination of the crystallites are deterinined with the US beam pas sae, and the coefficients of transparency of the fusion boundary, coefficients of scattering, devia~ tion and distortion of the US beam are calculated. The developed model can be used for calculation of the change in the front of US waves in uicir passage through coarse-grained structuves (Fi- gure 10), this perinitting selection of the optimal UT parameters and correct interpretation of the results REFERENCES 1. Brokhocow, NN. (1968) Physiat process in metas in eiding. Moscow Metallurg 2 Goins, SEs Abo, N'. (1997) Atenuation of wl Shale waves genes welds Nomdeste Tost Fools, 1 188 3, Delon. Rosje, D. (1982) lathe wages on solide Mpiaton for igal provsiog” Mancow Neu, 9-10/2000psterse US Analysis o truce string) on of 70 ininum Ristered lias in aa Tas, lite, substantiated Key words: nuclear and thermal power stations, pipe lines, erosionscovrosion damage, strength, residual life re doa! ‘ell laches, length of thinming zone, chonning Ae One of the main causes for preinature wear under the conditions of high values of pressure, tempera- tore and speed, is the joint action of mechanical (rosion) and electrochemical (corrosion) processes 1, 2}. Especially urgent is the problem ot erosion- corrosion wear (ECW) of pipelines of the therwal and nuclear power stations, 4s a tule, damage duc to ECW is of a local nature. It reduces the pipeline reliability, and this heads to the need to perform expensive repair-re. toration work in individual sections. The currently existent in CIS codes and provisions which specify the conditions for regarding the pipeline sections affected by ECW as unfit for service, Gu not provide an objective estimate of their performance. In par- ticular, the anti-aceident {3] and operational |4 provisions. specify the omlly criterion of admissibi- lity of further service, namely the pipeline wall thickness in ECW zone, determined during the regular inspection, which is limited by 0.85 fynin value (tin is the calculated pipeline wall thie ness). Without denying the importance of this pa rameter influence on the pipeline strength, we be- lieve itis absolutely unjustified to ignare the otiver parameters and, priniarily, the extent of the ECW zone in the axial and circular directions, Taking into account just the wall thickness does not eli nate the fact that in some cases the section of pine line with ECW will not ltave the necessary strensth mangin (which is confirmed by the damage oceur- ring during service), and in other cases this strength margin will be sufficient for extension of Ue ope- rational period between repair. | cermin rom 9-10/2000 ASSESSMENT OF THE STRENGTH AND RESIDUAL LIFE OF PIPELINES WITH EROSION-CORROSION DAMAGE EF, GARF and Mu The E.O, Paton Blsesic Welding Institute, NASU, Kyiv, Ukeaine ABSTRACT “Tho urgeney of the considered problem for nuclear and thormal power generation industry, is noted. Tks shown Us cocies now in force pernit the pipeline wall thinning by 15% ofthe design thickness, which, in a musber of cases leads to 4 sgitieant lowering of the structure reliability sud can cause fractare. Iatraduetion of 3 restriction not only an the wall thinning vafuss, but also on the main dimensions of the thinuing zone, Is Ww PAR OB evar WELDED STRUCTURES € NETREBSKY If the oxtont of the defect #8 considerable, its influence on the pipeline strength cannot be re- garded as a local influence. In this caso, the strength of the section of pipeline with ECW is determined by the minimal value of the wall thickness and with the current requirements, it will be equal to 0.85 of the desig value Not the least of the factors influencing the op rational reliability of the pipeline, is the fact that the current standard procedure does not take into account the prebistory of formation of the defect caused by ECW. Prohably, the sitwatians when the ‘wall thickness has reached limit value during one operational peried between inspections, and those when reaching the limit value was preceded by a long. period of service, should not be treated as ideutical. As the safety of pipelines with ECW should be guaranteed till the sext scheduled in- spection, it is easy to see that im the most unia- vourable case the thickness of the wall with ECW will be reduced to 0.7 tia, and taking into account the non-lincar dependence of the pipe wall thinning ‘on time and the probable initial pipe wall thickness > thn, it can reach even smaller values Therefore, there is no doubt that with an unfa- vourable concurrence of circumstances, occurrence of an emergency situation is inevitable, and this is related to the need for reactor shut-down before time and great financial losses, Note, that in this case the current safety requirements will not be formally violated. This paper gives the results of investigations which allow assessment of the influence of the ins dividual factors on the strength and residual life ‘f pipelines with OW, and describes the procedure cf design evaluation of the strength and residual life of such pipelines based on information about their operation and results of subsequent. inspec tions. 13> |WELOED STRUCTURES | | | 14] | I ~ \ Figuee 1, Diogramof ECW of apipeinesetion: f,,, ~ nominal vwallthickoess ¢~ cussestwalhUcles: L — @ureent length ‘oF ECW: Ly ~ ECW length relative to desi thickness of the pipe wall: L, — ECW width; Ri, ~ average adios relative to Calculated Wall thickness ‘The following approaches were incorporated into the developed procedure uf evaluation of the strougth and residual life of pipelines with ECW. strength of the pipeline with ECW over the entire period of its service up to taking the decision ‘on repair, should not be below the design strength; strength of pipeline with ECW should not onky be estimated at the moment of a regular inspection but should be also forecast for the period up to the subsequent scheduled mspection; prehistory of development of each concrete a- mage cavsed by ECW should be taken into account ’n design, i.e. duration of service of the section of Pipeline with ECW and kinctics of defect deve- lopment: possible simplifications used in development of the procedure, should yield a conservative result, The above approaches are implemented using a hypothesis according to which the influence of a defect caused by ECW on pipeline strength, will ‘be similar to the influence of an individual through: thickness circular hole, at least, in the area of its diameter, not exceeding a critical value of dg, The influence of ao individual through-thickness hole is well studied, confirmed by numerous experimien- tal studies, and, what is expecially important, is incorporated into the codes [5 ~ 7}. The critical dimensions of the defeets correspond to the diame- ter of the through-thickneys hole not lowering the load-carrying capacity of the cylindrical shel! and are defined by the following expression: dy = 0.25V2Ryatmins aw ‘where Rm is the average radius of the shell. Reduction of the defect caused by ECW to an individual through-thickness hole is rational to be performed through its area determined in the sec- tion along the pipeline axis, as the pipeline load- carrying capacity is determined hy cireunferential stresses, 14 WDB rvs a Figure 2. Limit admissible dimensions of ECW in the pictine, not lowering its Toadeareying caparity. 7 — theoretial; 2 — practical TCHR tom Analysis of erasion-corrosion damage in pipe- lines in service shows that its shape and dimensions can be the most diverse, however, no defects caused by ECW which can be regarded as crack-like are found The area of weakening due to ECW is deter- mined by the results of a regular inspection of the technical condition of the pipeline elements. simplification of the inspection procedure, with preference for deliberately conservative results, it is rational to take this arta to be the area of a rectangle (Figure t) Frye Lion ~ ty a where Ls the total length of ECW in the direction of the pipeline axis; fis the minimal thickness af the pipe wall in the ECW zone determined during 2 regular inspection by the results of thickness ‘sneasarement. On the other hand, itis also possible to more accurately determine the defect area and use the derived values in calculation Proceeding from the above, while allowing for the assumptions made, the admissible dimensions of ECW which do not lower the load-carrying, ca pacity of the pipeline, are given by the following dependencies (1) and (2): Li Gin ~ 4) $0.25 bin VER Ein @) Figure 2 gives dependence (3) in dimensionless co-ordinates, according to. which in the ease of a through-thickness defect (¢ = 0) its length is de- terinined hy diameter do, while at 47 > tym the defect lengtl tends to infinity. In practical work the presence of a through-thickness defect is not allowed, and with the increase of its length, the mechanism of pipeline fracture changes. Therefore, certain Timitations should be introduced. Ut is ra tional to eliminate the possibility of appearance of a through-thickness defect caused by ECW by li- miting the minimal admissible thickness of the pipe wall in the ECW zone by 0.3 tin value, this being in compliance with the US codes 18} “With a rather Jong deiect, the pipe strength will be determined by the cemaining material thick. 9-10/2000AR Ton © ipelne, teal; in pipe. nensions scaused like are + deter n of the ls. For e, with ls, it ra of a ca rection sof urin sess sible wand te for sions Wea wing, @) Bless ofa bde- the ork ot the ee, Results of tetas fragments of pipe with defits simulating eosion-cortosion We WELDED STRUCTURES € Sse ret BRENT eH gn FCW rho EEN th, ry Sate 1 ra 4650 1.000 35 500 500.70 14s 2 seas 5.470 1970 585 20 509 10 153 3 arse 7.200 1628 os 700 308.60 225 4 15808 6.850 1450 07 720 06.70 ao soma aan ams Ha 309 52.99 %7 6 6950 6.575 126 600 a0 181.45 450 7 6830 9000 0497 1100 260 181.46 a8 8 B49 6820 1.565 7s ne 51.46 4 9 609 6.595 0438 1500 m3 481.46 a4 tsar 8.80 10 ton 60a 480.70 82 158.40 5.200 0.590 1600 s00 490.70 235 2 ssg00 ‘8200 109 240 0.0 80,70 em Sea tian ulin oh pyre Pa 1 2.160 3,330 12.58 9.560, 0.962 1415 1.030 2 2eo0 tam 2105 0s) S75.— aT 32002076 1524 14.9% 096 ID Lan 42072190 Miz 1460 ost a9. 54m 30 479 S560 03.91.02 6 2s tao tos 7780952 aaa 3.0 7 50% 4.190 793774 os7 ar 08 8 580 1.270 tose e079. 080, 96s om a6 TT Sore aah 0 27% 1670 146 127% = 0938 224.080 4440 1.860 1461700983 sae 0D sam 660 ses tas 22g as tr ness in the thioning zone. The limit admissible length of ECW at which it can be regarded as a local defect and no Jowering of the load-carrying, capacity of the pipeline will be registered, should be substantiated by experimental studies. Accord- ing to [8], the length of the thinning zone should not exceed BVRmfmin - Then, the area located above the dependence represented by a solid line in 1'i gure 2, determines the ratio of dimensions (length and depth) of the damage dve t ECW at which safe service of the pipeline is guarantecd. According to this dependence, at the length of ECW zone 0< £;£0.5VReatimin t= 0.3 tin and U5VRetnin <£i£4VRytqin, the minimal wall thickness in the damage zone is given by expression (), while a AVRofain SLi 8 8VRudyny the mil smal wall thickness in ECW zone changes bya linear law from 0.91 £0 fy In order to casure reliable operation of pipeline systems with ECW zones, it is necessary to take a decision on the pipeline operability by the results 9-10/2000 WEB va Of the last inspection (at least, for the period till the next planned shut down of the technological system when the next inspection can be conducted). ‘Therefore, itis necessary to guarantee that the limit condition of the pipeline elements with ECW will not sct in before the next planned shut dawn of the system. This requires making a forecast of the condition of the ECW zone by the results of pre- vious inspections. It can be performed by extrapo- laGion, The task is made somewhat more Lifficult by a lack of information on the kinetics of deve- lopment of ECW-related damage in time, Tt is known that the corrosion processes are slowed sown with time because of formation of oxide Films which impede oxygen penetration to the metal sur- face [9|. Erosion-corrosion processes have a some- what different mechanism connectet primarily with the continuous breaking up of oxide films, Therefore, one should not expect slowing down of the process of defect propagation. Apparently, the process of FCW development proceeds more o¢ less uniformly in time. 15> WELDED STRUCTURES —— egal oe, a Figure 3, Samples of pipes with ECW for Lscing by inteona pressure: @ — rectilinear pipes; 0 — pipeline bend The regutatities of the process of damage deve- lopment can only be derived as the inspection re- sults are accumulated on each specific defect and a data bank is formed. At this stage (considering, a lack of information on the kinetics of development of damage related to ECW), it is, probably, justi- fied to introduce a certain correction factor K which would compensate the possible error. It is rational to assume a differentiated K value dependent on the number of inspections of each concrete defect over the period af pipeline service according to the following expression KG. OAi+N, where é is the number of inspections of this defect due to ECW over the period of pipeline service. Practical use of the above approach 9 evalu: ation of the strength of critical pipeline systems with ECW sequises comprehensive experimental verification of its individual statements and the correctness of the accepted hypothesis. The series of experimental studies perforincd at the E.0. Pa- ton Electric Welding Institute had the following, goals: Verification of the hypothesis of adequacy of the influence on pipeline strength of the defect caused by ECW and of the through-thickness cir cular hole: determination of the influence of the width of the defect caused by ECW on strength; deterinination of limit length of EW at which it can be regarded as Jocal damage The laboratory of testing of welded joints cer- tified by Gosstandard of Ukraine in’ the Ukr- SEPRO system, performed pressure testing of pipe samples, the results of which are given in the Table: Sample No. 5 is a Section of a pipeline hend cut out of the pipeline after 20 years of service, The other samples were made of rectilinear pipes with defocts simulating ECW made in the middle part of the sample on the internal surface (Figure 3) The diameter of a circular hole de, equal aver the area of the section along the pipe axis, was deter- ‘ined proceeding from the detect dimensions. The limit diameter of a circular hole not lowering the WERE DIG rca 0 "0 a0| ; 8 of | { + of, | oe 10 CR Figure, Distibution ofthe values of the ratio of experinental suid design beeaking pressure in the piges with ECW (N — probability, ° ~ steed deviation) sample strength, is given for each sample for com- Darison purposes. ‘The design strength of the san ples was determined from expressions known from {6,71 Se tin Rn where @ = 2/(deN2Rmelmin + 1.75). The last column of the Table gives the ratio of breaking pressure at testing and breaking pressure derived from calculations. The ratios of these values are close to the nortnal law of distribution with the sample mean equal to 1.013 and staucard de- viation of 0.023 (Figure 4) The results of experimental investigations pro- vide convincing evidence of the validity of the pro- posed hypothesis not only in the region of de * do values, but also in the case when the diameter of an adequate circular hole leads to decrease of the pipe strength ‘The samples for experimental investigations, as can be seen from the Table, had different width of ECW woe. Lm/VRmlmin. ratio changed in. the range of 1.12 to 3.33. The dependence given in Figure 5 shows that the width of ECW zone does ‘ot influence the ratios of experimental and design breaking pressuces. Considering that the width of ECW 2one was ignored at the design breaking pres sure, if it had an iafluence on strength, it would have been visible in the Figure. ‘The presented results give grounds to state that inevaluation of the strength of pipelines with ECW, in the case af pressure loading, the width of the damage zone can be ignored. Note, however, that the US codes take into account the influence of the 9-10/2000 Pa>xr) wiv oF com- he san ya from atio of essure values 1 with nd de- $ pro- epro- Kady ter of the is, as of the m in does sign hot ies wld that Ww the a ee od Figure 5 [fluence af ECW ro0e width on pipe strength width of ECW zone [8]. If it is smaller than Ravin curve 1 is recomended (Figure 6). With defoot width > 2.65VRmmfnin» admissible values of t/taiy are given by curve 2. The dependence as- sumed in this study which determines the dimen: sions of ECW at whieh the design strength of the pipeline is ensured (Figure 2), is close to curve 2 (igure 6). However, lowering of limit values of ECW length and depth with tts width reduction from 2.65.Rntnin tO VRintmin» is, probably, anjus- tified. ‘The data of experimental investigations permit assesment of the degree of thetr correlation with the assumed dependence of limit admissible dinic sions of ECWY not leading to a decrease of the pipeline sirength. With this purpose, only those af the results presented in the Table are consideved, in which the influence of the coefficient is minor or equal to 1. From the compacison given in Fi- gure 7 one can see that the testing results quite satisfactorily correlate with the proposed «lepen- dence of limit admissible dimensions. of whereas the assumption of the limit length at which the defect ean be considered as a local one, eyxral to BYR. Is guite justified Proceeding irom the above, the procedure of taking a decision on the possibility of further ope- ration of the pipeline with ECW, consists of the following steps. The initial parametersof the considered pipedine section are determined. Among them are the pipe ‘outer diameter D, nominal wall thickness thom working pressure P in the pipeline, yrade of steel of which the pipeline is made, and admissible stresses for it lo], data of the start of pipeline operation, design thickness of the wall tin. The latter can be determined from calculations a a sinensis nf 1,2 265K, 9-10/2000 -W according to 18}. $= La RE 2 Figure 6. Dependencies (or determination of limit admissible WARD rca WELDED STRUCTURES € or 2 3 4 5 6 7 Figue 7, Comparison of limit dimensions of ECW zone by the resales of testing pipes in which no los of strength was fount: 1 experiment: 2" design The length of ECW zone tn the direction af the axis of pipeline Lj, minimal thickness of pipe wall tin ECW zone, daration af the pipeline operation T;, are determined proceeding from the results of the regular scheduled inspection of the pipeline section wit ‘The period of pipeline operation till the next: scheduled inspection Piet ts determined. ‘The forecast thickness of the pipe wall in ECW zone at the moment of the scheduled examination is determined from the following expression: |) from ra Length La of ECW zone on which the wall thickness is below fn for the period of scheduled inspection, is calculated from the following expres- Lys KL tom —& ‘ined, where Rip = (D ~ co-ordinates isi/tmin and La/VRn bia marked in Figure 2. If this point falls on the de- pendence which determines the limit admissible ‘atias of the dimensions of ECW zone, ac is above it, further oporation of the pipeline section with ECW till the next scheduled examination, is pos sible, Now, if it is lower than the mentioned de- pendence, pipeline repair is required. The safe residual life of the pipeline section with ECW dimensions established during a regular Cr a er oe Figure 8. Depondence of limit admissible dimnsions of FCW, ‘oun! for assesment of the residual life ofthe pipeline 7> WELDED STRUCTURES inspection, is determined by the following proce- Sate, The dependence shown in Figure 8 which is a mirror reflection af the dependence given in Fi- gure 2, as well as the data of the last inspection of ECW Zone, are used to determine the limit admis sible values of length Lim and depth fre ~ tim of the damage. With this aim, the point with en-or- dinates (oom ~4)/trom and Li/VRofiom is marked in Figure8. A straight line isdrawa through the origin of coordinates and this point up to the point of intersection with the curve of limit adi sible dimensions of ECW and (aia = fly) / thm and Liim/*Rmtgom parameters are determined by the point of intersection. Dimensions f and £5 of ECW sone derived tn the lase inspection and the timi¢ admissible dimen- sions flim and Liig of ECW zane derived from Fi- ure 8, as well as the period T; of operation vf this pipeline section before the last inspection, are used to determine the safe life of the pipeline with ECW. according to the following expression: fon Ts ak ‘The safe residual life AT of the pipetine section with ECW is determined from the following, ex- pression ar= ‘The values of residual service life of the seetion of pipeline with ECW can be used for assigning, the schedule of subsequent inspections and pla ning the repair operations ‘CONCLUSIONS 1. The currently accepted in power generation in dustry procedure of processing, from ane side the pipcline sections with ECW, does not provide the required level of safety in their service, and on the other hand, in a number of eases does not permit full use of the actual residual life af the pipeline. 2. & procedure of evaluation of fitness for ser vice and residual life of the pipelines witt ECW. zones in individual sections, has heen developed. which is based on the principle of provision of the design strength during the entire period of the sys tem service. 3. Limit admissible combinations of the depth and length of ECW zone in the direction of the pipeline axis have been established, which da cot lower its load-carrying capacity 4, The main postulates of the procedure have been experimentally contiemed. LL is shows that the width of ECW zone does not influence the strength and residual life of pipelines with ECW. REFERENCES: 1, Torronen, K., Aaltonet, P., Hanninen, H, (1994) Water Shemsury id tutenals degradsion ta LWRS lng Prow of Specelses Merting on Working Material Evetin thd Corosion of NEP organised by the TAEA, Kye. Sop 19" 22, 1984 Viena, 2. Korhonen, R., Hietanen, ©. (994) Erosion- €2) in the direction of which Figure 7, Closed zie-iled crack between the exsspiece an web plate ——— iP FOB eri086, Figure 8. Variants of piece: 1 — web plat ztal and point of # cross piece @ ~ height of the fillet weld; I, and fy ~ distances from be ai the dubleT beam web tothe eae of gravy cangeneut of the welds nese the eros 2 "centro of gravity of the deponted lfect exerted by shrinkage Fores: 3 — the longitudinal shrinkage forces act. But even in this case, ata cross-piece thickness of 45 mm, the shrinkage in the transverse direction is harapered, this leading to extremely high values of residual welding stresses. ‘Cross pieces were welded to double-T beams without slots at the corners. it i at these points, which are the points of crossing of the welds con= verging in three different directions, that the mul- tiaxed stressed state is formed due to'a high rigidity of the structure, which increases the risk of brittle fracture It should be noted that despite the requicement of ensuring a good quality of gas cutting and of the welds in compliance with assessment group C {DIN EN 25 817), the quality left much to be lesire Tn immersing cold thick-walled steel parts into liquid zine (Figure 9), the non-stationary tempera ture fields lead to formation in them of high thermal ———— RAR eae stresses. Superposition of these stresses on residual welding stresses say create conditions for liquid zinte to penetrate into steel. Liquid zine, while pene: trating into steel along the grain boundaries. makes these boundaries weaker, which leads to brittle fracture. The following three factors are required for this effect to show up: * it is necessary that liquid and solid metals have low mutual solubility, which is the case of the zine-iron pair; ‘© no intermetallic phases should be formed in the liquid-solid metal pair, which is also the case vf the zine-iron pair; * solid metal in contact with Jiquid metal has 1 possibility of being plastically deformed, which was fixed in the region of welding cross-pleces to double-T beams. Sensitivity of metal to embrittlement as a result id of penetration of the liquid metel into the ‘one lepends upon effective tensile stresses and de sree of alloying of steel, Thus, the effect exerted by liquid zine on steel containing 0.22 % silicon is, higher than on steel containing no silicon [13] “The presence of numerous defective welds made repair oF welded span structures inexpedient, All the span structures in the workshop ad to be re- placed by the new ones. ‘Measures for elimination of eracks in hot gal- vanazing. With a wall thickness of the steel struc- ‘eares equal to more than 30 mm, even slow immer- sion into hot liquid zine leads to high thermal stresses which cannot be compensated for by strength of the material. In this case the presence ‘of obstacles to strain of the metal and high residual welding stresses complicate elimination of cracks. Preheating may decrease the risk of their formation, but it is hardly suitable for economical ceasons. Therefore, while considering the problem of corro- sion protection of steel structures by hot galvaniz ing, one should take into account the following aspects: ‘= need to develop a part with a design optimal for galvanizing: + rational selection of materials to minimize the risk of penetration of liquid zine along the grain boundaries; * keeping to the welding technology; «# coordinating fabrication of welded structures with companies performing galvanazing. The following requirements should be satisfied in hot galvanizing of steel parts ‘* difference in thickness of the parts welded, tnax/tmins Should he not in excess oF 2.5; * it Is necessary to limit thickness of cross sec~ tions of the plates and profiles (uae $20 - 30 min), as well as round sections tn combined structures (dmx $ 30 mm); ‘flanged edges of the parts should have a large radius: 9-10/2000edad jor liquid hile pene- ies, makes to britle Fequiced id metals e case of ormed in the ease ictal has d, which pices to ‘aresult he solid and de- ts made mit. ALL 9 be re- tot gal Isirue- inmer- beamal for by sence sual racks, lation, axons caren: vaniz owing imal tthe grain tures WELDED STRUCTURES ¢ Rigid structure with volumetric ——| Fa ars tele aetses Hot galvanizing Zinc bath Temperature of the zine bath ‘=> (440 - 480°C) decreases strength => Effect of residual stresses and. obstacles to strains u ¥ Crack initiation (penetration of quid => zine depending upon residual, local and thermal stresses) Figure 9. Flow diagram of the proces af galvanizing and crack formation ‘Figure 10, Drum with 9-10/2000 Figure 11, Cicumferential conaecting welds between the shaft and the drum base RED nia iL 3Figure 12, Connertng weld wth cracks + hollows should have openings to let zine flow in and out; * it is necessary to make stots in paris at the points where cross-picees join a double-T beam; # where possible, it is necessary to avoid fillet welds of lange sizes and replace them by butt welds; * it is necessary to avoid intermittent welds in design of the parts patches with openings to remove air should be welded entirely on their perimeter: # it is necessary to cnsure the low Jovel of re sidual welding, srrsses by optimal arrangement of the welds and keeping to the required sequence af their makings # it is necessary to remove all welding slags prior to hot galvanizing’ + the speed of immersion in galvanizing should be 1 = 8 m/min, depending upon the thickness of a cross section; * galvanizing of complicated parts should be done alter preliminary experiments; * it is necessary to follow the transportation and storage regulations specified for parts in the hot state (suspension, supports along the entire length, ete.) ‘Shrinkage eracks on drums of helt conveyers. It was thought initially that cracks in cireumfer: ential connecting welds (Figures 10 and 11) be- tween the steel $995J2G3 (St.53-3) shaft 65 or 70 mm in diameter and stool S235JR (St.37) hase 15 mm thick were caused by fatigne feacture which took place within only a few months of operation of the drums (Figures 12 and 13). However, as found during further investigations, itis probable that cracks were formed immediately after manu- —— | _] unferentia\ shrinkage erack ater fracture ofthe joint wHPARRR, Figure 14. Transverse section of the jlnt containing eracks between the shaft (@) and the base (2) facture of the drums. The said cracks propagated in the material of the base in parallel with the HAZ, of the welds (Figure 14). Structural elements did vot fracture only because the welds had the V- groove preparation, As the welds were not seb: jected toNDT, the cracks were not detected during the manufacture. It is likely that manufacturers did not consider welding to be a critical process [nour opinion cracking was caused by formation of quenching structures in the HAZ of the fillet welds, which is proved by an extremely high mi- crohardness of this zone, CO2 welding was con- ducted without preheating, High tensile stresses favoured cracking and led to complete separation df the weld from the base. The situation was ag- gravated by the fact that groove preparation was done with a small angle, which caused defects in the weld root (Figure 14, @). In addition, parts ‘were machined for welding and after assembly there ‘was no gap between them, which prevented strains during welding end increased the level of residual welding stresses. This assumption was confiemed by analysis of distribution of residual siresses in a welded joint between the shaft and the drum base during cooling (Figure 15). Distribution of residual stresses was, investigated in cooling from 500 °C to room tem- perature. The presence of the zones with increased residual stresses was noted in the FLAZ of the welds. Cracks were initiated particularly in these zones. ‘Cracks between the shaft and the drum base ‘could have been avoided in the ease of ensuring the 9-10/2000