UltraNarrowGroove Laser Welding for Heavy …...In the ultra-narrow groove, the ra-dius of the laser...

Introduction The International ThermonuclearExperimental Reactor (ITER) Projectwas proposed and implemented since1985 (Refs. 1, 2). Many structures inthe vacuum chamber are joined to-gether by several thick-plate compo-nents. Some stainless steel plates areup to 60 mm in thickness (Refs. 3, 4).Their welding has posed a great chal-lenge with current welding techniques. The common welding methods forthick stainless steel include narrow-groove argon arc welding (Ref. 5), narrow-groove submerged arc welding(Refs. 6, 7), and narrow-groove gasmetal arc welding (Refs. 8–10). How-ever, these three welding methods re-quire high heat input and easily giverise to distortion, which greatly affects

the ability to successfully weld thestructures in the ITER vacuum cham-ber. Narrow-groove laser welding witha filler metal that combines the advan-tages of laser welding and narrow-groove welding has increased in popu-larity through its reduction in heat in-put and distortion deformation. Nev-ertheless, various studies (Refs. 11–17) have shown that several problemsstill exist, including incomplete fusionof side walls and keyhole-inducedporosity. Some studies (Refs. 18–22) havealso demonstrated that the effective-ness of multiple light spots or weavingbeams in suppressing the side wall in-complete fusion issues. However, sucha method introduces complexity andmay lead to a need for wider grooves(thus greater distortion) and greatly

increased costs. In addition, Phaoniamet al. (Ref. 23) used the conductionmode for narrow-groove laser weldingwith hot wires to weld 10-mm-thick304 stainless steel plates and acquiredporosity-free welds. This suggests thatthe keyhole mode can be replaced bythe conduction mode to eliminate thekeyhole-induced porosities in narrow-groove laser welding — a major issuein narrow gas welding of thick plates.However, the side wall incomplete fu-sion issues still existed in their study. In summary, a series of problems,such as deformation, keyhole porosity,and incomplete fusion of side walls,exist in the welding of thick plates.The following methods can be adoptedfor overcoming these problems: 1) ultra-narrow-groove welding to re-duce heat input and welding deforma-tion; 2) concurrent heating on the sidewalls and bottom in combination withthe use of a laser beam to avoid the de-fects on the side walls induced by in-complete fusion; and 3) the use of aheat-conduction model to avoid theappearance of keyholes. Comparedwith traditional laser devices, the fiberlaser exhibits a series of comprehen-sive advantages, such as high beamquality, high power, a compact struc-ture, and low operating cost, and thushas attracted extensive attention frommany scholars in recent years (Refs.24–27). Based on the high-qualityfiber laser in combination with weld-ing wire, this paper proposes an ultra-narrow-groove laser heat conductionwelding method in which the laserbeam acts on the bottom and side

WELDING RESEARCH

UltraNarrowGroove Laser Welding forHeavy Sections in ITER

This welding technology can achieve highquality, stable welding of 120mmthick stainless steel plate with a groove width of 4.5 mm and simultaneously reduce distortion,

pores, and incomplete fusion

BY S. K. WU, J. L. ZOU, R. S. XIAO, AND G. W. ZHANG

ABSTRACT Many structures in the vacuum chamber in the International Thermonuclear Experimental Reactor (ITER) were welded thick stainless steel. For thick plates, traditional welding technologies often cause large distortion, pores, or incomplete fusion on theside walls. This paper thus explores ultranarrowgroove laser welding in conductionmode with welding wires using a fiber laser. Results show, while the ultranarrow groovesignificantly reduces the distortion, the concurrent heating from the laser beam on theside walls and bottom of the groove ensures side wall fusion. The conduction mode eliminates porosity. Defectfree buttjoint welds were made on 60mmthick stainless steelplates with a 3mmwide ultranarrow groove in 20 layers. The distortion angle was only0.6 deg. Stainless steel with a thickness and groove width of 120 and 4.5 mm can bewelded using the method and experimental conditions.

KEYWORDS • Thick Plate • UltraNarrow Groove • Fiber Laser • Welding Wire • HeatConduction Welding

S. K. WU, J. L. ZOU ([email protected]), G. W. ZHANG, and R. S. XIAO ([email protected]) are at the HighPower and Ultrafast LaserManufacturing Lab, Institute of Laser Engineering, Beijing University of Technology, Beijing, China.

WELDING JOURNAL / AUGUST 2016, VOL. 95300-s

Zou Supplement August 2015117.qxp_Layout 1 7/14/16 2:46 PM Page 300

walls of the groove simultaneously. In this paper, the transmissioncharacteristics of the fiber laser beamthrough the ultra-narrow groove willbe studied to order to achieve a maxi-mal thickness without defects. Thetheoretical maximum thickness will beanalyzed and the optimal parameterswill be determined. The effectivenessof the proposed method together withthe optimized parameters will beproved by successfully welding 60-mm-thick stainless steel plates.

Theoretical Basis of theUltraNarrowGrooveLaser Thermal ConductionWelding Method

Transmission Characteristics ofthe Laser Beam through theUltraNarrow Groove

Ultra-narrow grooves are defined asgrooves below 6 mm for the welding ofthick plates with a thickness above 30(Refs. 27, 28). The range of the ultra-narrow groove was set between 2 and 5mm owing to the consideration of thewire-feeding nozzle size, which shouldpenetrate into the groove. This studyused the heat conduction weldingmode, which could be achieved by re-ducing the power density via positivedefocusing. The laser spot that reachesthe bottom of the groove was largerthan the groove width as a result of theuse of the positive defocusing mode.Therefore, the bottom and side walls ofthe groove could be simultaneouslyheated by the light beam — Fig. 1.

In the ultra-narrow groove, the ra-dius of the laser beam exhibits the fol-lowing variation rules with the propa-gation distance []:

where z denotes the parallel distancebetween the horizontal direction andthe focal point (i.e., the position of thebeam waist); (z) denotes the beamradius at distance z from the focal po-sition; 0 = D/2 denotes the focal ra-dius; ZR = (0

2/M2 denotes theRayleigh length of the laser beam; M2

denotes the quality factor of the lightbeam; and denotes the wavelengthof the laser. By substituting ZR intoEquation 1, z can be written as

Equation 2 shows that the factors af-fecting the transmission distance of thelight through the groove include thegroove width, the quality of the lightbeam M2, the light wavelength , andthe focal radius 0. When the groovewidth is equal to the beam diameter ofthe laser beam at z, i.e., 2(z), the thick-ness of the welded plate achieves a max-imum value. After determining the laserdevice and the groove width, meaningthat , M2, and (z) are known, solvingEquation 2 reveals that z has the maxi-mum (z)2/2M2 when 0 = (z)/√2.̅Thus, when the diameter of the laser fo-

cus D and the groove width 2(z) cansatisfy the relationship D = 2(z)/√2 ,the laser can be transmitted the far-thest in the groove. The transmissiondistance of the laser beam in the groovedecreases if the diameter of the laser fo-cus is too large or too small — Fig. 1. In practical applications, the diame-ter of the laser focus is subject to thebeam transmission system; the focallength of the collimating lens, the focallength of the focusing lens, and the corediameter of the transmission fiber allaffect the diameter of the focus. The ac-tual size of the light beam waist is con-nected with the focal length of the colli-mating lens, the focal length of the fo-cusing lens, and the core diameter ofthe transmission fiber:

where fc denotes the focal length of thecollimating lens; f denotes the focallength of the focusing lens; and c de-notes the core diameter of the transmis-sion fiber. Thus, plates with maximumthickness can only be welded whenthese three factors are matched and thediameter of the laser focus is equal tothe theoretical value. After the laser de-vice and the groove were determined,the positive defocusing mode wasadopted. The laser spot on the bottomshould be equal to or slightly greaterthan the groove width, resulting in thetransmission distance of light in thegroove with the same width to be twicethe distance using the focusing spot. The maximum thickness of the weldedplate can then be written as

� z( )=�0 1+ zzR

��

��

2

1( )

z= ��M2

� z( )2�02 ��0

4 2( )

�0 =ffc�c 3( )

WELDING RESEARCH

AUGUST 2016 / WELDING JOURNAL 301-s

Fig. 1 — Transmission characteristics of the laser beams with different beam waists.

Fig. 2 — Experimental setup of the ultranarrowgroove laserbeam welding.

Groove


where h denotes the maximum thick-ness of the welded plates; and g de-notes the groove width.

Formation of the Heat Conduction Molten Pool

The laser energy that acts on the bot-tom and side walls of the groove shouldbe set between the welded metal’s melt-ing threshold and penetration thresholdto ensure that the laser heat conductionwelding mode is conducted in the ultra-narrow groove. For the semi-infinitewelded metal under the laser’s verticalaction, the temperature at the spot cen-ter on the metal surface can be ex-pressed as (Ref. 29)

where T(0,t) denotes the temperatureat the spot center; T0 = 300 K denotesroom temperature; A denotes thelaser absorptivity by the material;v denotes the welding speed;t = 2(z)/v denotes the characteristic

time of laser action; t denotes thethermal conductivity; at = t/c de-

notes the thermal diffusivity; I de-notes laser power density; denotesthe density of the welding material;and c denotes the heat capacity of thewelding material. In the presentstudy, the laser acted on the bottomof the thick plate after entering intothe ultra-narrow groove. Therefore,the welding material can be consid-ered between the semi-infinite thickplate and the infinite thick plate. Atthis moment, the temperature field atthe center of light spot in the ultra-narrow groove can be modified as

where b denotes the structure coeffi-cient, with the value ranging from 0.5to 1; d denotes the diameter of thelight spot; and P denotes the laserpower. When the temperature at thespot center reaches the melting point,the welded plate begins to melt. Ac-cording to Equation 6, the meltingthreshold of the plate in the ultra-narrow groove can be written as

where Tm denotes the melting temper-ature of welded material; s denotesthe mean thermal conductivity ofwelded material in the solid state; s isthe mean density of solid metal; cs isthe mean heat capacity of solid metal;d denotes the diameter of the laserbeam spot; and As denotes the meanabsorptivity of solid metal. The physi-cal parameters of the metal materialare connected with the temperature.To simplify the estimation, it is as-sumed that the physical parametersare linearly related to temperature inEquation 7, and the average valueswere used in the present work. Based on the self-focusing effect ofthe depression of the molten pool onthe incident laser, the penetrationthreshold, when the flat thick platesare welded using laser, can be charac-terized by P/d (laser power/spot diam-eter) (Ref. 30):

where k is the Boltzmann constant(1.38 × 10–23 J/K); denotes the sur-face tension coefficient of the moltenpool; l denotes the mean thermalconductivity of liquid metal; Lv de-notes the latent heat of evaporation; l ̅ is the mean density of liquid metal;cl is the mean heat capacity of liquidmetal; Al denotes the mean absorptivi-ty of liquid metal; Al denotes the ab-sorptivity of the incident laser by theliquid materials; mv denotes the quali-ty of metal atoms; and Ts denotes thetemperature on the surface of themolten pool, which is equal to the boil-ing temperature when the keyhole isformed. Since the boundary condi-tions of the ultra-narrow groove aredifferent from the flat plate, the struc-ture parameter b should be taken intoaccount. Therefore, the penetrationthreshold in the ultra-narrow groove(Equation 8) can be rewritten as

T 0,t( )�T0 =AI� z( )

�t 2�( )12

tan�1 8att

� z( )2�

��

�

�� 5( )

T 0,t( )�T0 =bAP

2�32�td

tan�1 32�t�cvd

��

��

6( )

pd=

�32 Tm �T0( )�s

b212 As tan

�1 32�s

�scsvd

�

��

�

�

12

7( )

Pd= �Lv

Al

�mvkTs

+2 2� Tv �Tm( )�l

Al exp ��lc lvd4�l

�

��

�

��

+2 2� Tm �T0( )�sAs exp ��scs vd

4�s

��

��

8( )

h=2z= �D2�M2 g2 �D2 4( )

WELDING RESEARCH


Fig. 3 — Theoretical and experimental values of the penetration threshold of the flatplate laser welding as the diameter of laser spot varies.

–

––

–

–

–

Zou Supplement August 2015117.qxp_Layout 1 7/14/16 2:46 PM 02

The input power range of the laser giv-ing rise to the formation of the moltenpool in the ultra-narrow groove can beestimated based on Equations 7 and 9for ensuring the laser heat conductionwelding mode.

Experimental Conditionsand Setup The present study focused on thewelding of thick stainless steel plates

for the ITER project. A fiber laser(YLS-6000, IPG Photonics, USA) wasused. The maximum output power ofthe laser is 6 kW, and the transmissioncore diameter of the fiber is 200 μm. Acollimating lens with the focal lengthof 120 mm, and a focusing lens withthe focal length of 250 mm were used,the diameter of the focusing spot was0.42 mm, and the quality factor of thebeam was 22.3 (M2 = 22.3). The six-axis-joint manipulator (KRC-30,KUKA Systems, Germany) was usedfor the welding operation. The frontwire feeding mode was selected, andthe diameter of the wire feeding noz-zle was 2 mm. The wire for weldingwas kept at a constant distance fromthe point of laser action. The side-blowing shield gas was used to protectthe molten pool, and the shielding gasnozzle with a caliber of 7 mm wasplaced at a 45-deg angle from the di-rection of the laser beam. Ar gas with apurity of 99.995% was adopted as theshielding gas. In the welding process,

the gas flow of Ar was 25 L/min, andthe shielding gas nozzle was pressedagainst the groove so the shielding gascould smoothly enter into the grooveto guarantee the protection effect. Fig-ure 2 illustrates the setup in weldingexperiments. The 304 stainless steel thick plates,which were composed of 71.25% Fe,0.07% C, 2% Mn, 0.045% P, 0.03% S,0.075% Si, 18.28% Cr, 8.15% Ni, and0.1% N, were welded. The ER347 stain-less steel welding wires, which werecomposed of 64.86% Fe, 0.08% C, 2%Mn, 0.03% P, 0.03% S, 0.6% Si, 20.5%Cr, 10.25% Ni, 0.75% Mo, and 0.9% Nb,had a diameter of 1.1 mm. The abovecompositions are presented as massfractions. The molten pool was irradiat-ed by the laser, and the wires were in-serted into the molten pool to melt thegrooves via the heating of the moltenpool. The welding wires can be meltedby the molten pool for improving weld-ing process stability. A total of 4 typesof grooves were used: the U-shaped ul-tra-narrow grooves of 2.5, 3, 3.5, and4.5 mm were processed in the stainlesssteel plates with the thickness of 20, 60,90, and 120 mm, respectively. Beforewelding, the groove was polished andcleaned by acetone. The smoke and dustleft by the last weld bead were cleanedbefore the next welding process. Foreach experimental parameter, the weld-ing process was repeated three times.Twelve cross sections were randomlyobtained from the weld joint in eachwelding sample by cutting. After grind-ing, polishing, and etching, these crosssections were measured with an Olym-pus optical microscope.

Experimental Results andAnalyses

The HeatConduction WeldingMolten Pool Formation

The range of P/d that can lead tothe formation of a heat-conductionmolten pool can be calculated theoreti-cally. The common physical parame-ters of the 304 stainless steel materi-als for theoretical calculations are list-ed in Table 1, in which the surface ten-sion coefficient and the atomic massof iron (Ref. 31) were used, and theother physical parameters are temper-

Pd= �Lv

Al

�mvkTv

+2 2� Tv �Tm( )�lbAl exp ��lcl vd

4�l��

��

+2 2� Tm �T0( )�sbAs exp ��scs vd

4�s

��

��

9( )

WELDING RESEARCH


Fig. 5 — Typical crosssection patterns of the weld joints in the ultranarrow groovewith varying P/d values: A — P/d = 0.96 kW/mm; B — P/d = 1.11 kW/mm; and C —P/d = 1.21 kW/mm.

Fig. 4 — Several typical surface patterns and crosssection observations of the weldjoints at varying P/d values: A — P/d = 0.55 kW/mm; B — P/d = 0.94 kW/mm; and C— P/d= 1.25 kW/mm.

Table 1 — Physical Parameters of the 304 Stainless Steel

Item Tm(K) Tv(K) σ (N/m) Lv (kJ/kg) mv (kg)

Value 1670 3143 1.5 6856.5 9.3 × 10–26

A B C

A B C


ature dependent. The solid-state phys-ical parameters were set as the meanvalues in this temperature range, andthe liquid-state physical parameterswere also set as the mean values inthis state. By referring to previouswork (Ref. 32), the following parame-ters can be calculated: s = 24 W/Km;cs ≈ 600 j/kg K; s = 7.35 × 103 kg/m3 ;l= 20 W/Km; cl ≈ 784 j/kg K; and l̅ =6.3 × 103 kg/m3. In view of the highabsorptivity of the laser by the stain-less steel materials at a wavelength of1 μm (Ref. 33), and the roughness ofthe plate surface in the solid state,some parameters can be set as follows:As ≈ 0.40; Al ≈ 0.43; and Al ≈ 0.46. Given the structure coefficient b,the melting threshold and the penetra-tion threshold in the ultra-narrowgroove can be calculated according toEquations 7 and 9. The structure coef-ficient b can be determined by com-paring the penetration thresholds us-ing flat-plate welding and ultra-narrow-groove welding. At a weldingspeed of 0.12 m/min, the variationrules of the penetration thresholdwith the diameter of the light spotwhen using fiber laser flat-plate weld-ing were derived in accordance withEquation 8 and the parameters of the

stainless steel, as shown in Fig. 3. Thetheoretical value of the penetrationthreshold is approximately 1 kW/mm,which also increases slightly as thespot diameter rises. A fiber laser with a 2-mm spot di-ameter was adopted for flat-platewelding to verify the calculation re-sults (Fig. 3), and the welding resultsare presented in Fig. 4. When the val-ue of P/d is 0.55 kW/mm, the pat-terns, which are dense, scaly, smooth,and can even be observed on the sur-face of the weld joint, and the depth offusion is comparatively shallow — Fig.4A. This represents a typical heat-conduction welding mode. When thevalue of P/d is 1.25 kW/mm, the roughscaly patterns can be observed, andthe depth of fusion at the center of thelight spot exhibits a remarkable in-crease — Fig. 4C. This represents atypical deep penetration weldingmode. When the value of P/d is 0.94kW/mm, the surface of the weld jointexhibits a combination of the patternsobserved in Fig. 4A and C. The evaporating pressure is nearlynegligible during the heat-conductionwelding process, which causes the sur-face of the molten pool to be compara-tively smooth. This accounts for the

smooth and even surface of the weldjoint — Fig. 4A. During deep penetra-tion welding, the keyhole forms in themolten pool due to the intense evapo-rating recoil pressure. The unstablefluctuation of the keyholes would thengive rise to the relatively coarse anddense scaly patterns — Fig. 4C. It canbe inferred that the welding mode, i.e.,whether heat-conduction welding ordeep penetration welding was used,can be distinguished based on the sur-face morphology of the weld joint. Theobserved penetration threshold is ap-proximately 0.94 kW/mm (Fig. 4B),which fits well with the theoretical val-ue presented in Fig. 3. The penetration threshold can bedetermined based on the alternatingappearance of the weld surface patternas the welding speed or spot diametervaries (i.e., the appearance in Fig. 4B).Figure 3 displays the measured pene-tration threshold of the stainless steelduring the fiber laser welding of flatplates, which shows the measuredpenetration threshold is irrelevant tothe diameter of the light spot and thewelding speed under conditions withlow welding speeds. This is consistentwith the conclusion that the penetra-tion threshold is an inherent attributefor a given material in our previousstudies (Ref. 30). However, the theo-retical value of the penetration thresh-old slightly increases as the diameterof the light spot increases, which ex-hibits a slight deviation from experi-mental values. This is due to the as-sumption in the derivation of the pen-etration threshold; it has been as-sumed the temperature at the marginof the spot is equal to the boiling tem-perature at the moment when thewelding mode varies. Figure 4Cdemonstrates that deep penetrationappears at the center of the light spotwhen using the large-spot laser forwelding, and the heat conductionmode is simultaneously used at themargin of the spot. Thus, at the tran-sition moment of the large-spot weld-ing mode, the self-focusing effect oc-curs at the center of the spot, and thetemperature at the margin of the spotmay not reach the boiling point. Thismay also account for the deviation be-tween the experimental value and thetheoretical value — Fig. 3. The welding was also conducted onthe 2.5-mm groove with no use of

WELDING RESEARCH


– –

–––

––

Fig. 6 — Morphologies of weld joints using ultranarrowgroove laser welding with welding wire with varying spot diameters. Singlelayer welding: A — 1.5 mm; B — 2.5 mm;and C — 3.5 mm; twolayer welding: D — 1.5 mm; E — 2.5 mm; and F — 3.5 mm.

A

D

B

E F

C

Table 2 — The Maximum Welded Thickness and Power Range of Thermal ConductivityWelding with Varying Groove Sizes

Groove (mm) Max Thickness (mm) Power (kW)

2.5 60 0.84~3.0 3.5 90 1.33~4.2 4.5 120 1.86~5.4


welding wires. The diameter of thelight spot acting on the bottom of thegroove was set as 2.5 mm, and thewelding speed was set as 0.5 m/min.Figure 5 displays the cross-section ob-servation of the obtained weld joints.The welding mode changes from theheat-conduction welding mode to thedeep-penetration welding mode whenthe value of P/d is within the range of1.1 to 1.2 kW/mm — Fig. 5B and C.Thus, the penetration threshold usingultra-narrow-groove fiber laser weld-ing on stainless steel ranges from 1.1to 1.2 kW/mm. This penetrationthreshold is significantly higher thanthe value achieved, 0.94 kW/mm,when using flat-plate welding, which isdue to the favorable heat conductioncondition in the narrow groove. Ac-cording to Equations 8 and 9, thestructure coefficient of the ultra-narrow groove can be estimated as b ≈0.75 in this case. According to the physical parame-ters of the stainless steel material andEquation 7, the melting threshold ofthe ultra-narrow groove using fiberlaser welding is approximately 0.35kW/mm. Therefore, the heat-conduc-tion welding mode is used in ultra-narrow groove of the stainless steelthick plates when the input P/d of thefiber laser ranges from 0.35 to 1.2kW/mm.

Relationship between theWelding Mode and Weld Defect

A groove with the width of 2.5 mmand depth of 15 mm were created on a304 stainless steel with a thickness of

20 mm. The laserwelding experi-ments were thenconducted, inwhich the spotdiameters actingon the bottom ofthe groove werevaried by chang-ing the laser’s de-focusing amount.The laser powerwas set as 2.5kW, the weldingspeed was set as0.5 m/min, and the wire feeding speedwas set as 2.5 m/min. Figure 6A, B,and C displays the one-pass weld jointsusing ultra-narrow-groove laser weld-ing with welding wire when the diame-ter of the laser spot was 1.5, 2.5, and3.5 mm, respectively; Fig. 6D, E, and Fdisplays the corresponding two-passweld joints.

Results show that with single-layerwelding, the weld depth decreases andthe surface pattern varies from convexto concave as the spot diameter in-creases. The pores appear in the weldjoint when the spot diameter is 1.5mm — Fig. 6A and D. The weld defectsare particularly evident in two-layerwelding, as incomplete fusion of theside walls and pores appear. This oc-curred when the spot diameter wasrelatively small and the value of P/d(1.67 kW/mm) far exceeded the pene-tration threshold of the ultra-narrowgroove — Fig. 6A and D. Accordingly,the deep penetration welding modewas used in the ultra-narrow groovewhen the spot diameter was 1.5 mm,and the unstable closure of keyholesled to the appearance of pores. Due tothe small range of spot action, insuffi-

cient energy was received by the sidewalls, giving rise to the incomplete fu-sion of the side walls.

When the spot diameter was in-creased to 2.5 mm, the P/d value ofthe laser was reduced to 1 kW/mm(Fig. 6B and E), which was lower thanthe penetration threshold of the ultra-narrow groove. Here, the heat-conduc-tion welding mode was used in theultra-narrow groove, and the keyhole-induced pore could be effectivelyavoided in the single-layer welding.However, incomplete fusion of theside walls occurred in two-layer weld-ing. Since the spot diameter was equalto the width of the ultra-narrowgroove, the light spot did not heat theside walls of the ultra-narrow groove.This led to a smooth surface of theweld joint, and incomplete fusion wasthen easily generated at the right-angle corner when the wires werefilled twice.

Different surface morphologies ofthe weld joint would impose differenteffects on the integral molding of the next weld joint. When the ultra-narrow groove was filled by the moltenwires, a three-phase boundary wouldbe formed regardless of the melting

WELDING RESEARCH


Fig. 7 — Schematic of molten pool at different welding modes: A— Deep penetration welding mode; B — thermal conduction welding mode.

Fig. 9 — Tensile and yield test results of the 304 stainless steel joints.

Fig. 8 — Pictures of the samples with a thickness of 60 mm forultranarrowgroove welding, and the morphology of the crosssection of the welding joint.

A A

C

B B


degree of the side walls. Since a key-hole was formed during penetrationwelding, the melt flow in the surfaceof the molten pool would move towardthe two sides from the keyhole mouthowing to the surface tension gradientof the molten pool. After rapid solidifi-cation, the weld joint showed a convexcenter. The spreading wettability ofthe molten pool in the groove was thekey to forming a perfect molding jointduring the filling process of the ultra-narrow groove. However, liquid metalsdid not readily spread over the solid-state surface. Therefore, the weldjoints using ultra-narrow-groove laserwelding usually had straight patternsor raised centers, and the process eas-ily generated incomplete fusion of theside walls after the next filling, asshown in Fig. 7A. When the spot diameter wasgreater than the groove width, the sidewalls melted completely and the weldjoints had smooth surfaces and con-cave patterns — Fig. 6C and F. More-over, neither pores nor incomplete fu-sion of the side walls was observed af-ter the second filling. Here, the valueof P/d was 0.83 kW/mm, and heat-conduction welding was used. Sincethe spot diameter was greater than thegroove width, the defocusing beamacted simultaneously on the side wallsand the bottom of the groove. More-over, the concave shape of the moltenpool could be formed owing to thebeam’s whole heating on the moltenpool and its surroundings. In the nextfilling process, heating of the beam onthe concave weld joint could be used toeffectively avoid the incomplete fusionof the side walls. By using a laser with

a spot diameter larger than the groovewidth, the incomplete fusion of theside walls could be effectively avoided,and the formed surface of the weldjoint is most favorable for the next fill-ing, as shown in Fig. 7B. This processproduces an optimal welding joint.

UltraNarrowGroove LaserButtJoint Welding Technologyon a 60mmThick Plate

To further verify the feasibility ofthe proposed welding method, ultra-narrow-groove laser heat-conductionwelding with welding wires was con-ducted on two 304 stainless steelplates with thicknesses of 60 mm,since the plate used in the ITER proj-ect can be up to a maximum thicknessof 60 mm. Thick stainless steel with asize of 250 × 200 × 60 mm was used,and the groove was set at 3 mm. Thediameter of the light spot on the bot-tom of the groove was 3.2 mm, theoutput power of the laser was 3.5 kW(i.e., P/d = 1.09 kW/mm), the weldingspeed was 0.42 m/min, and the weld-ing wire speed was 3.4 m/min. Weld-ing was then conducted with 20 weld-ing wires, each of which had a heightof 3 mm. Figure 8 displays the crosssection of the weld joint from the sam-ple. The width of the weld joint wasapproximately 3.8 mm, and it had reg-ular patterns and no defects (ISO13919-1:1996), and the welding defor-mation was approximately 0.6 deg. Tensile and yield tests were thenconducted on the 60-mm-thick weld

joint to examine the mechanical per-formances of the joint. Samples 3 mmthick and 25 mm wide were selectedfrom different positions at the upper,middle, and lower layers of the joint,and the interval between these layerswas 20 mm. The average value of thetensile and yield properties of the weldjoint and base material were also in-vestigated at varying temperatures,since the protective containment usedin the ITER project may operate in alow-temperature environment. The re-sults are shown in Fig. 9. The tensile strengths of the weldjoint at 273, 243, 213, and 183 K were651, 1180, 1220, and 1260 MPa,which were 87%, 96.7%, 94.5%, and92.6% of those of the base material,respectively. At lower temperatures,the tensile strength of the weld jointwas comparatively greater. The joint’stensile strength increased with the in-creasing temperature, which may berelated to the limitations on the dislo-cation slip at low temperatures. Theyield strength of the joint variedslightly with the temperature; it wasroughly equal to that of the base mate-rial within the temperature range of275 to 320 MPa. The relatively weakphase-transformation strengtheningeffect of the austenitic stainless steeland the increase in strength of thebase material under the as-rolled con-dition may be two primary reasons forthe weld joint having a lower strengththan the base material. The slight difference between theweld joint using the ultra-narrow-groove laser welding and the base ma-

WELDING RESEARCH


Fig. 10 — Different samples for laser heatconduction welding with welding wires: A — Sample with width and depth of 3.5 and 90 mm,respectively; and B — sample with width and depth of 4.5 and 120 mm, respectively.

A B


terials also suggests that the weldingstability and quality can be ensuredby adopting the proposed method.Since the welding wires and the basematerial vary in chemical compo-nents, the content of ferrite is higherin the ER347 welding wires than inthe base material, which would de-crease the strength of the weld jointto a certain degree. However, the de-cline in tensile strength was mainlyrelated to the fact that the weld jointwas the as-cast structure that had notbeen strengthened through hotrolling. Additionally, the weld jointhad undergone multiple thermal cy-cles in the multilayer welding of thethick plates. This can be referred to asthe tempering process, which wouldlead to an uneven distribution andoverall decline of strength.

Theoretical Calculation of theMaximum Thickness of theWelded Plate and the RelatedExperimental Verifications

The maximum thickness of thewelded plate in the ultra-narrowgroove using fiber laser welding andthe laser power range that can giverise to the molten pool were calculated(Table 2) based on Equation 4 and theconditions for the formation of themolten pool in the ultra-narrowgroove (Equations 7 and 9). This weld-ing mode produced the following re-sults: when the ultra-narrow groovewas 2.5 mm in width, the maximumthickness of the welded plate was 60mm, and the laser power ranged be-tween 0.84 and 3 kW; when the ultra-narrow groove was 4.5 mm in width,the maximum thickness was up to 120mm, and the laser power ranged be-tween 1.86 and 5.4 kW. The ultra-narrow grooves listed inTable 2 were selected for welding ex-periments to verify the correctnessand feasibility of the proposed calcula-tion method. In the experiments, thewelding speed and the wire feedingspeed were set as 0.36 and 0.32m/min, respectively. An ultra-narrowgroove was processed in the stainlesssteel plate with a thickness of 20 mm,a width of 3.5 mm, and a depth of 90mm. The laser power was set at 4 kW,and the spot diameter, acting on thebottom of the groove, was set at 3.6

mm by controlling the defocusingamount (Table 2). No weld defects ap-peared on the cross section of the weldjoint after completion of two weldpasses — Fig. 10A. This suggests thatno-defect ultra-narrow-groove fiberlaser welding of the stainless steelplate with a depth and width of 90 and3.5 mm, respectively, can be achieved. Another ultra-narrow groove with awidth and depth of 4.5 and 120 mm,respectively, was processed similarly,and a 5-kW laser power and 4.6-mmspot diameter were adopted for con-ducting the welding experiments —Fig. 10B. After twice welding withwelding wire, the weld joints had regu-lar shapes, and defects, such as incom-plete fusion and pores, did not appear.Ultra-narrow grooves with widths of120 mm can also be welded by fiberlaser with welding wires. The resultsindicate that the calculation is accu-rate; this welding technology can beused to achieve high-quality, stablewelding of 120-mm-thick plate.

Conclusions Based on fiber laser transmissioncharacteristics, an ultra-narrow-groove laser heat-conduction weldingmethod with filler metal was proposedfor thick stainless steel. Theoreticalcalculations and experimental verifica-tion were also conducted, and the pri-mary conclusions are as follows. When the beam waist diameter Dand groove width g meet D = g/√2, thelaser beam can spread farthest in theultra-narrow groove, i.e., the weldablethickness of the heavy section for ultra-narrow-groove laser heat-conduction welding is maximum value. Using the ultra-narrow-groove fiberlaser heat-conduction weldingmethod, the penetration threshold inthe groove with a width of 2.5 mm isapproximately 1.2 kW/mm, and thewelding mode can be discriminated ac-cording to the surface morphology ofthe welded plates. Welding deformation can be re-duced by adopting the ultra-narrowgroove. Incomplete fusion of the sidewalls can effectively be avoided by thelaser beam acting on the side walls andthe bottom of the groove simultane-ously. The formation of keyhole-induced porosity can be completelyavoided by adopting the heat-conduc-

tion welding mode. The absence of defects with theproposed method of welding stainlesssteel plates with thicknesses of 60 mmthat are used in ITER project can beachieved by using the ultra-narrowgroove with a groove of 3 mm widthand 20 weld passes. Results demon-strate that the welding joint has regu-lar patterns and little deformation. For the fiber laser and focusing sys-tem adopted in the present article, thestainless steel thick plates with thethickness and groove width of 120 and4.5 mm can be welded using ultra-narrow-groove fiber laser heat-conduction welding with welding wire.

The research was financially sup-ported by the National Natural ScienceFoundation of China (Grants No.51505011, 51475011, and 51275013).

1. Onozuka, M., Alfile, J. P., Aubert, P.,Dagenais, J.-F., Grebennikov, D., Ioki, K.,Jones, L., Koizumi, K., Krylov, V.,Maslakowski, J., Nakahira, M., Nelson, B.,Punshon, C., Roy, O., and Schreck, G. 2001.Manufacturing and maintenance technolo-gies developed for a thick-wall structure ofthe ITER vacuum vessel. Fusion Eng. Des.5(4): 397–410. 2. Coste, F., Sabatier, L., and Dubet, O.2001. Nd:Yag laser welding of 60 mmthickness 316L parts using multiple pass-es. Laser Inst. America 502–509. 3. Jokinen, T., Karhu, M., and Kujan-paa, V. 2003. Welding of thick austeniticstainless steel using Nd:Yttrium-alu-minum-garnet laser with filler wire and hy-brid process. J. Laser Appl. 15(4): 220–224. 4. Coste, F., Janin, F., Hamadou, M.,and Fabbro, R. 2003. Deep penetrationlaser welding with Nd:Yag lasers combina-tion up to 11 kW laser power. Proc. of SPIE4831: 422–427. 5. Khodakov, V. D., Danilov, A. I., Kho-dakov, D. V., Praliev, D. A., Abrosin, A. A.,and Gutorov, D. A. 2015. Investigation anddevelopment of technology for automaticargon-arc narrow gap welding of Du-850pipes for reactor cooling plant to water-wa-ter reactors. Weld. Int. 29(5): 372–378. 6. Qu, Y. B., Cao, B., Wang, L., Zhang, B.Q., Wu, J. D., Cai, Z. P., and Pan, J. L. 2014.Weakening of fusion zones in narrow gapsubmerged arc welding. J. Tsinghua Univ.

WELDING RESEARCH


Acknowledgments

References

–


WELDING RESEARCH


54(3): 305–308. 7. Zhang, W. M., Li, X. X., and Wang, B.2014. Design on cleaning device for slag ofnarrow gap submerged arc welding. Appl.Mech. Mater. 470: 404–407. 8. Kimura, S., Ichihara, I., and Nagai, Y.1979. Narrow-gap, gas metal arc weldingprocess in flat position. Welding Journal59(7): 44–52. 9. Wang, J. Y., Zhu, J., Fu, P., Su, R. J.,Han, W., and Yang, F. 2012. A swing arcsystem for narrow gap GMA welding. ISIJInt. 52(1): 110–114. 10. Wang, J. Y., Zhu, J., Zhang, C.,Wang, N., Su, R. J., and Yang, F. 2015. De-velopment of swing arc narrow gap verticalwelding process. ISIJ Int. 55(5): 1076–1082. 11. Srinivasan, G., Arivazhagan, B., Al-bert, S. K., and Bhaduri, A. K. 2011. Devel-opment of filler wires for welding of re-duced activation ferritic martenstic steelfor India’s test blanket module of ITER. Fu-sion Eng. Des. 86: 446–451. 12. Jokinen, T., Viherva, T., Riikonen,H., and Kujanpaa, V. 2000. Welding of shipstructural steel A36 using a Nd:Yag laserand gas-metal arc welding. J. Laser Appl.12(5): 185–188. 13. Salminen, A. S., and Kujanpaa, V. P.2003. Effect of wire feed position on laserwelding with filler wire. J. Laser Appl.15(1): 2–10. 14. Elmesalamy, A. S., Li, L., Francis, J.A., and Sezer, H. K. 2013. Understandingthe process parameter interactions in mul-tiple-pass ultra-narrow-gap laser weldingof thick-section stainless steels. Int. J. Adv.Manuf. Technol. 68: 1–17. 15. Yu, Y. C., Huang, W., Wang, G. Z.,Wang, J., Meng, X. X., Wang, C. M., Yan, F.,Hu, X. Y., and Yu, S. F. 2013. Investigationof melting dynamics of filler wire during

wire feed laser welding. J. Mech. Sci. Tech-nol. 27(4): 1097–1108. 16. Kong, F. R., Liu, W., Junjie, M., Lev-ert, E., and Kovacevic, R. 2013. Feasibilitystudy of laser welding assisted by fillerwire for narrow-gap butt-jointed plates ofhigh-strength steel. Weld. World 57: 693–699. 17. Zhou, Z., Wu, W., Wei, J., Du, S.,Han, S., Liu, L., Yu, X., Li, H., Foussat, A.,and Libeyre, P. 2012. Research on manu-facture and enclosure welding of ITER cor-rection coils cases. IEEE Trans. Appl. Super-cond. 22(3): 4202603. 18. Arata, Y., Maruo, H., Miyamoto, I.,and Nishio, R. 1986. High power CO2 laserwelding of thick plate-multipass weldingwith filler wire. Trans. JWRI 15(2): 27–34. 19. Coste, F., Janin, F., and Hamadou,M. 2003. 20 mm thickness Nd:Yag laserwelding of 316L stainless steel with longfocal length. Proc. of SPIE 4831: 422–427. 20. Zhang, X. D., and Ashida, E. 2011.Welding of thick stainless steel plates up to50 mm with high brightness lasers. J. LaserAppl. 23(2): 022002. 21. Tsukamoto, T., Kawanaka, H., andMaeda, Y. 2011. Laser narrow gap weldingof thick carbon steels using high bright-ness laser with beam oscillation. Proceed-ings of ICALEO, United States 141–146. 22. Phaoniam, R., Shinozaki, K., Ya-mamoto, M., Kadoi, K., Tsuchiya, S., andNishijima, A. 2013. Development of ahighly efficient hot-wire laser hybridprocess for narrow-gap welding — Weldingphenomena and their adequate conditions.Weld. World 57: 607–613. 23. Fanton, L., Abdalla, A. J., and Fer-nandes de Lima, M. S. 2014. Heat treat-ment and Yb-fiber laser welding of amaraging steel. Welding Journal 93(9):362-s to 368-s.

24. Blecher, J. J., Palmer, T. A., and De-bRoy, T. 2015. Mitigation of root defect inlaser and hybrid laser-arc welding. WeldingJournal 94(3): 73-s to 78-s. 25. Zou, J. L., Wu, S. K., Xiao, R. S., andLi, F. 2015. Effects of a paraxial TIG arc onhigh-power fiber laser welding. Mater. De-sign 86: 21–327. 26. Zou, J. L., Wu, S. K., He, Y., Yang, W.X., Xu, J. J., and Xiao, R. S. 2015. Distinctmorphology of the keyhole wall duringhigh-power fiber laser deep penetrationwelding. Sci. Technol. Weld. Joi. 20(8): 655–658. 27. Coste, F., Sabatier, L., Dubet, O.,Aubert, P., and Jones, L. B. 2001. Nd:Yaglaser welding of 60 mm thickness 316Lparts using multiple passes. Laser Inst.America 502–509. 28. Phillips, R. H., and Metzbower, E.A. 1992. Laser-beam welding of Hy80 andHy100 steels using hot welding wire addi-tion. Welding Journal 71(6): 201-s to 208-s. 29. Duley, W. W. 1998. Laser Welding.John Wiley & Sons, Inc. 30. Zou, J. L., He, Y., Wu, S. K., Huang,T., and Xiao, R. S. 2015. Experimental andtheoretical characterization of deep pene-tration welding threshold induced by 1-mlaser. Appl. Surf. Sci. 357: 1522–1527. 31. Iida, T., and Guthrie, R. 1993. ThePhysical Properties of Liquid Metals, OxfordClarendon. 32. Agelaridou, A. 2002. Thermal andSolidification Modeling of Welding: A DesignTool Approach. Tufts University. PhD the-sis, pp. 60–568. 33. Hirano, K., Fabbro, R., and Muller,M. 2011. Experimental determination oftemperature threshold for melt surface de-formation during laser interaction on ironat atmospheric pressure. J. Phys. D: Appl.Phys. 44: 435402.


UltraNarrowGroove Laser Welding for Heavy …...In the ultra-narrow groove, the ra-dius of the laser...

Documents

Transcript of UltraNarrowGroove Laser Welding for Heavy …...In the ultra-narrow groove, the ra-dius of the laser...