Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/281667962

FusionZoneMicrostructureandGeometryinComplete-Joint-PenetrationLaser-ArcHybridWeldingofLow-AlloySteel

ArticleinWeldingJournal·April2015

CITATION

1

READS

173

4authors,including:

Someoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:

DigitaltwinoflaserassistedadditivemanufacturingViewproject

HuiliangWei

PennsylvaniaStateUniversity

17PUBLICATIONS48CITATIONS

SEEPROFILE

ToddAPalmer

PennsylvaniaStateUniversity

104PUBLICATIONS1,138CITATIONS

SEEPROFILE

TarasankarDebroy

PennsylvaniaStateUniversity

286PUBLICATIONS7,782CITATIONS

SEEPROFILE

AllcontentfollowingthispagewasuploadedbyTarasankarDebroyon09January2016.

Theuserhasrequestedenhancementofthedownloadedfile.Allin-textreferencesunderlinedinblue

arelinkedtopublicationsonResearchGate,lettingyouaccessandreadthemimmediately.

WELDING RESEARCH

APRIL 2015 / WELDING JOURNAL 135-s

Introduction Hybrid laser-arc welding is a processthat combines laser beam welding andconventional arc welding in order to in-corporate the benefits of both processes(Refs. 1–4). Hybrid laser-gas metal arc(GMA) welding produces wider weldpools than autogenous laser welding,and deeper weld penetration than GMAwelding with the same parameters(Refs. 5–8). The combination of laserand arc energy sources allows for com-plete-joint-penetration welds to be achieved at sig-nificantly higher welding velocities in asingle pass, while at the same time al-lowing large root openings in weldjoints to be bridged (Refs. 9, 10). As aresult, welding productivity can be

greatly enhanced over that achieved byeither laser or GMA welding alone(Refs. 11–13). In addition, hybrid laser-GMA welding has significant advan-tages in acquiring the desired weld met-al microstructures, since lower coolingrates can be more easily obtained thanin autogenous laser welding. However,martensite, which has very low ductilityand toughness (Ref. 14), can still formin hybrid welding (Ref. 15). Previous work on the hybrid laser-GMA welding of steels has largely fo-cused on the experimental postcharac-terization of weld geometries, mi-crostructures, and mechanical proper-ties (Refs. 15–21). However, thesepostmortem evaluations provide littledetail on the evolution of weld poolgeometries and the cooling rates with-

in the fusion zone. In order to under-stand and predict weld metal mi-crostructural evolution, the thermalcycles experienced during these weld-ing processes must be known. The in-teractions between the heat sourcesand materials during complete-joint-penetration hybrid laser-GMA weldinglead to rapid thermal cycles in theweld pool, which in turn impacts mi-crostructure evolution. Phase transfor-mations during cooling in the weld fu-sion zone have been extensively inves-tigated both experimentally and theo-retically. Bhadeshia et al. developed aphase transformation model (Refs.22–24) based on thermodynamics andphase transformation kinetics. Thismodel can quantitatively predict themicrostructures and properties ofweld deposits for different alloy com-positions, cooling rates, and prioraustenite grain sizes. Direct measurement of temperatureprofiles in the interior of the weld poolstill remains a major challenge. On theother hand, a well-tested three-dimen-sional mathematical model can provideaccurate temperature fields and coolingrates at discrete locations throughoutthe fusion zone (Refs. 25, 26). Severalstudies focused on the numerical mod-eling of the fluid flow and heat transferconditions within the molten weld poolof hybrid laser-arc welding (Refs. 6, 27).Ribic et al. (Ref. 27) numerically studiedthe effect of laser arc separation dis-tance and laser power on heat transferand fluid flow in partial-penetration hy-brid laser-gas tungsten arc (GTA) weld-ing by using a three-dimensional nu-merical model. They found that the dis-tance between the laser and arc signifi-

Fusion Zone Microstructure and Geometry inComplete­Joint­Penetration Laser­Arc Hybrid

Welding of Low­Alloy SteelA process map indicates a martensite­free microstructure can be

maintained over a wide range of welding parameters

BY H. L. WEI, J. J. BLECHER, T. A. PALMER, AND T. DEBROY

ABSTRACT The fusion zone geometry and microstructure in complete­joint­penetration hybridlaser gas metal arc welds of a low­alloy steel are examined experimentally and theoreti­cally. Weld geometry and spatially variable cooling rates are investigated using a three­dimensional heat transfer and fluid flow model. Experimentally measured microstruc­tures are compared with those estimated from a microstructure model based on kinet­ics and thermodynamics of phase transformations, for a range of laser arc separationdistances and heat inputs. Considerable variations in both cooling rates and microstruc­ture were observed for the range of process parameters utilized. In fact, the experimen­tal results and calculations show that for the same heat input, a predominantly ferriticand predominantly martensitic microstructure can be obtained, depending on the laserarc separation distance and resulting cooling rate. A process map is constructed show­ing the effect of welding speed, laser power, and laser arc separation distance on cool­ing rates and microconstituent volume fractions. The map indicates a martensite­freemicrostructure can be maintained over a wide range of welding parameters.

KEYWORDS • Hybrid Welding • Heat Transfer and Fluid Flow • Microstructure • Complete Joint Penetration • Laser • Gas Metal Arc • Cooling Rate

H. L. WEI (huw15@psu.edu), J. J. BLECHER, and T. DEBROY are with Department of Materials Science and Engineering, and T. A. PALMER is with Applied Re­search Laboratory, The Pennsylvania State University, University Park, Pa.

WELDING RESEARCH

WELDING JOURNAL / APRIL 2015, VOL. 94136-s

cantly affected the cooling rates andthat the weld penetration was maxi-mized at an optimal laser arc separationdistance (Ref. 27). Cho et al. (Ref. 6)simulated the molten weld pool geome-try in laser-arc hybrid welding by solv-ing the equations of continuity, mo-mentum, and energy using a commer-cial package. They reported that thewidth of the weld was determinedmainly by the GMA heat source and thepenetration depth was strongly influ-enced by the laser (Ref. 6). However,there are very few systematic studies fo-cused on the numerical simulation of

weld profile evolution, cooling rateswithin the fusion zone, and the corre-sponding influence on the weld metalmicrostructures in complete-joint-pene-tration hybrid laser-GMA welding. Theprevious numerical studies (Refs. 6, 27)have discussed partial-penetration hy-brid welding, where the fluid flow andheat transfer at the bottom of themolten weld pool are significantly dif-ferent than those in complete-joint-penetration welding. In this work, the evolution of macro-and microstructures of complete-joint-penetration laser-GMA hybrid welds inDH 36 steel is analyzed using funda-mental transport phenomena andphase transformation theory. A three-dimensional heat transfer and fluidflow model has been developed to studythe effect of welding velocity and laserarc separation distance on weld geome-tries and cooling rates. Using the calcu-lated cooling rates from the heat trans-fer and fluid flow model, a phase trans-formation model (Refs. 22–24) basedon thermodynamics and phase transfor-mation kinetics is used to provide a

quantitative description of the final mi-crostructures of the weld metal. Thecomputed volume fractions of the weldmetal allotriomorphic ferrite, Wid-manstätten ferrite, acicular ferrite, andmartensite are validated with corre-sponding experimental results for vari-ous welding conditions. The combinedmodels are used to construct processmaps capable of predicting the effect ofwelding parameters on resulting coolingrates and microstructures.

Experimental Procedure Bead-on-plate complete-joint-penetration hybrid laser-GMA weldswere made on 4.8-mm-thick DH 36steel. An IPG Photonics® YLR-12000-Lytterbium fiber laser with a Precitec®YW50 welding head and a Lincoln Elec-tric® Power Wave 455 M/STT weldingpower source with a Binzel® WH 455Dwater-cooled welding gun were used forhybrid welding. The maximum power ofthe fiber laser is 12 kW, with a wave-length of 1070–1080 nm. The opticssystem utilizes collimating and focusinglenses with 200- and 500-mm focallengths, respectively. The laser is trans-ported to the welding head through a200-mm-diameter process fiber. The fo-cal spot of the laser beam in the absenceof plasma was approximately 0.56 mmin diameter. The laser was focused 8mm above the surface of the plate. Thelaser power used in the welding experi-ments was fixed at 5.0 kW for all thecases. The electrode was 0.045-in.- (1.1-mm-) diameter ER70S-6 wire. Thechemical compositions of DH 36 steeland ER70S-6 welding wire are given inTable 1. The metal transfer mode forthe welding wire was in spray mode.The shielding gas was a mixture of 95%argon and 5% CO2 with a flow rate of 95ft3/h (44.8 L/min). The welding veloci-ties and laser arc separation distanceswere varied to study their effects on theweld profiles and weld metal mi-crostructures. The key welding parame-ters are listed in Table 2. The top andbottom surfaces of the plate were

Fig. 1 — Top surface and symmetry planeof weld pool with temperature contoursand velocity fields for the following:A — Welding speed of 20.0 mm/s, laserarc separation distance of 1 mm; B —welding speed of 30.0 mm/s, laser arcseparation distance of 1 mm. Fig. 2 — Comparison of experimental and

simulated weld cross sections for the hy­brid laser­GMA complete­joint­penetra­tion welding of DH 36 steel for thefollowing: A — Welding speed of 20.0mm/s, laser arc separation distance of 1mm; B — welding speed of 30.0 mm/s,laser arc separation distance of 1 mm.

Fig. 3 — Top surface and symmetryplane of weld pool with temperaturecontours and velocity vectors for the fol­lowing: A — Welding speed of 40.0mm/s, laser arc separation distance of 1mm; B — welding speed of 40.0 mm/s,laser arc separation distance of 5 mm.

Table 1 — Chemical Composition of Base Metal DH 36 Steel (Ref. 10) and Welding Wire ER 70S­6 (wt­%)

C Mn Si Ni Mo Cr V P S Al Nb Ti Cu

Base 0.06 1.39 0.19 0.14 0.03 0.11 0.06 0.011 0.004 0.025 0.01 0.01 0.25 Metal

Welding 0.09 1.63 0.90 0.05 0.05 0.05 0.05 0.007 0.007 0.000 0.00 0.00 0.20Wire

A

A

A

B

B

B

WELDING RESEARCH

APRIL 2015 / WELDING JOURNAL 137-s

ground to remove scale prior to weldingin order to avoid weld root defects dur-ing complete-joint-penetration welding.The sides of the plate were supported,so the welds were made without con-tacting the table below. Selected weldswere sectioned, polished, etched, andphotographed to reveal the weld fusionzone profile and microstructures. Thevolume fractions of selected microcon-stituents in the weld metal are deter-mined by using the point countingmethod following the International In-stitute of Welding (IIW) guidelines (Ref.28).

Mathematical Model A three-dimensional heat transferand fluid flow model for complete-joint-penetration hybrid laser-GMA weldingwas developed by modifying previousnumerical simulation work (Refs. 27,29, 31). Zhao et al. (Ref. 31) developed atransport phenomena-based numericalmodel to predict the keyhole geometryand temperature profiles in the weld-ment during keyhole laser welding. Raiet al. (Ref. 29) developed a convectiveheat transfer model for both partial andcomplete-joint-penetration keyholemode laser welding of a structural steelbased on the work of Zhao et al. (Ref.

31). Ribic et al. (Ref. 27) proposed athree-dimensional heat transfer andfluid flow model for partial-penetrationhybrid laser-GTA welding. In this work,complete-joint-penetration hybridlaser-GMA welding is studied.Marangoni force-driven velocityboundary conditions at the bottomsurface are assumed, which is different

from that of partial-penetration hybridwelding. In addition, the heat transferfrom the metal droplets during GMAWis integrated into the numerical modelfor hybrid laser-GMA welding. The ma-terial properties used in order to com-plete the welding calculations are givenin Table 3. Details of the numerical sim-ulation model are presented below.

Calculation of Keyhole Profile

The keyhole geometry is calculatedusing a model that considers materialproperties, welding process parameters,and specimen geometries. The detailedinformation about the model is avail-able elsewhere (Refs. 30, 31) and onlythe salient features are presented here.The keyhole profile is calculated basedon a point by point energy balance atthe keyhole walls and is determined it-eratively. Multiple reflections of the

laser beam within the keyhole are as-sumed and the number of reflections isdependent on the keyhole geometry.The keyhole wall local temperature istaken as the boiling point of the alloy(Refs. 30, 31). Planar heat conductionfrom the keyhole wall into the work-piece is assumed due to the significantlyhigher temperature gradient in all direc-tions in the horizontal plane comparedto the vertical directions. Once the pro-file calculation is completed, the tem-perature distribution from the keyholemodel is stored in a data file with alltemperatures inside the keyhole as-signed the boiling point temperature.This file is read into the heat transferand fluid flow model, and at each hori-zontal x-y plane, the keyhole boundaryis identified by a minimum and a maxi-mum x value for any y value.

Heat Transfer in Weld Pool andBoundary Conditions

After the calculation of the keyholeprofile, equations of conservation ofmass, momentum, and energy aresolved in three dimensions in the heattransfer and fluid flow model. Detailsabout this model are available in the lit-erature (Refs. 22, 23, 29) and only the

Fig. 4 — Comparison of experimental andsimulated weld cross sections for the hybridlaser­GMA complete­joint­penetrationwelding of DH 36 steel for the following: A— Welding speed of 40.0 mm/s, laser arcseparation distance of 1 mm; B — weldingspeed of 40.0 mm/s, laser arc separationdistance of 5 mm.

Table 2 — Welding Process Parameters for Complete­Joint­Penetration Hybrid Laser GMA Welding ofDH 36 Steel

Weld Laser Power Arc Current Arc Voltage Welding Laser ArcNumber (kW) (A) (V) Speed (mm/s) Separation (mm)

1 5.0 248 31 20.0 1.02 5.0 235 31 30.0 1.03 5.0 232 31 40.0 1.04 5.0 232 31 40.0 5.0

A

B

Fig. 5 — Calculated cooling curves at dif­ferent y locations on the top surface of thefusion zone of weld 1 with welding speedof 20.0 mm/s, laser arc separation dis­tance of 1 mm. The symbol y representsthe distance from the weld centerline.

Fig. 6 — Comparison of the calculatedcooling rates of the top center fusionzone for weld 1 with welding speed of20.0 mm/s, laser arc separation distanceof 1 mm, and weld 2 with welding speedof 30.0 mm/s, laser arc separation dis­tance of 1 mm.

WELDING RESEARCH

WELDING JOURNAL / APRIL 2015, VOL. 94138-s

salient features are presented here. Themolten metal is assumed to be an in-compressible, laminar, and Newtonianfluid. The liquid metal flow in the weldpool can be represented by the followingmomentum conservation equation(Refs. 30, 31):

where is the density, t is the time, xi isthe distance along the ith (i = 1, 2, and3) orthogonal direction, uj is the velocitycomponent along the j direction, isthe effective viscosity, and Sj is thesource term for the jth momentum equa-tion and is given as

where p represents pressure, U is thewelding speed, and is the coefficientof volume expansion. The third termrepresents the frictional dissipation inthe mushy zone according to the Car-man-Kozeny equation for flow througha porous media (Refs. 32, 33) where fL isthe liquid fraction, B is a very smallcomputational constant to avoid divi-sion by zero, and C is a constant ac-counting for the mushy zone morpholo-

gy [a value of 1.6 × 104 was used in thepresent study (Ref. 34)]. The fourthterm is the buoyancy source term (Refs.23, 34, 35). The last term accounts forthe relative motion of the workpiece rel-ative to the laser and arc heat sources(Ref. 23). The following continuity equation issolved in conjunction with the momen-tum equation to obtain the pressurefield.

In order to trace the weld pool liq-uid/solid interface, i.e., the phasechange, the total enthalpy H is repre-sented by a sum of sensible heat h andlatent heat content DH, i.e., H = h + DH

u

t

u u

x

x

u

xS

j i j

i

i

j

ij

( )ρ

∂∂

+ ρ∂

∂

= ∂∂

μ∂∂

⎛⎝⎜

⎞⎠⎟

+ (1)

Sp

x x

u

x

Cf

f Bu

g T T Uu

x

jj j

j

j

L

Lj

refj

j( )

( )

= − ∂∂

+ ∂∂

μ∂∂

⎛

⎝⎜

⎞

⎠⎟

−−

+

⎛

⎝⎜⎜

⎞

⎠⎟⎟

+ρ β − − ρ∂∂

1

(2)

2

3

u

xi

i

( )∂ ρ∂

= 0 (3)

Fig. 7 — Comparison of the calculatedcooling rates of the fusion zone located 1mm above the bottom surface for weld 3with welding speed of 40.0 mm/s, laser arcseparation distance of 1 mm, and weld 4with welding speed of 40.0 mm/s, laser arcseparation distance of 5 mm.

Fig. 8 — Comparison of the optical mi­crostructures of the top center of the fu­sion zone by different welding speeds.Magnification 500x. A — Welding speedof 20.0 mm/s, laser arc separation dis­tance of 1 mm; B — welding speed of 30.0mm/s, laser arc separation distance of 1mm. The symbols , W, and a representallotriomorphic, Widmanstätten, and acic­ular ferrite, respectively.

Table 3 — Material Properties Used for the Calculation of Temperature and Velocity Fields (Ref. 29)

Physical Property DH 36 Steel

Boiling point (K) 3133Solidus temperature (K) 1745Liquidus temperature (K) 1785Density (kg/m3) 7200Thermal conductivity (W/m­K) 21Inverse Bremsstrahlung absorption coefficient (1/m) 100Absorption coefficient (flat surface) 0.16Molecular viscosity (Pa­s) 0.0067Coefficient of thermal expansion (l/K) 1.96 x 10­5

Temperature coefficient of surface tension (N/m K) –0.5 x 10­3

Enthalpy of solid at melting point (J/kg) 1.20 x 106

Enthalpy of liquid at melting point (J/kg) 1.26 x 106

Specific heat of solid (J/kg K) 710.6Specific heat of liquid (J/kg K) 836.0

A

B

Fig. 9 — Comparison of the optical mi­crostructures of the fusion zone located 1mm above the bottom surface by differentlaser arc separation distances. Magnifica­tion 500x. A — Welding speed of 40.0mm/s, laser arc separation distance of 1mm; B — welding speed of 40.0 mm/s,laser arc separation distance of 5 mm. Thesymbols , W, a, and M represent allotri­omorphic, Widmanstätten, acicular ferrite,and martensite, respectively.

A

B

WELDING RESEARCH

APRIL 2015 / WELDING JOURNAL 139-s

(Ref. 36). The sensible heat h is ex-pressed as h = CpdT, where Cp is the spe-cific heat and T is the temperature. Thelatent heat content DH is given as DH =fLL, where L is the latent heat of fusion.The liquid fraction fL is assumed to varylinearly with temperature for simplicity(Ref. 23) and is given as

where TL and TS are the liquidus andsolidus temperature, respectively. Thus,the thermal energy transportation inthe weld workpiece can be expressed bythe following modified energy equation:

where k is the thermal conductivity. Thesource term Sh is due to the latent heatcontent and is given as

The heat transfer and fluid flowequations were solved for the completeworkpiece. For the region inside thekeyhole, the coefficients and sourceterms in the equations were adjusted toobtain boiling point temperature andzero fluid velocities.

A 3D Cartesian coordinate system isused in the calculation. Only half of theworkpiece is considered since the weldis symmetrical about the weld center-line. At the bottom of the weld pool,Marangoni force-driven fluid velocityboundary conditions are assumed forcomplete-joint-penetration welding. A187 × 77 × 26 grid system is used in thecalculation and the corresponding calcu-lation domain dimensions are 522 mmin length, 36 mm in half-width, and 4.8mm in depth. The interactions betweenlaser and arc as well as the heat transferand fluid flow within the weld pool areaffected by the separation distance be-tween laser and arc. In the numericalmodel, the effect of arc energy on theformation of the keyhole and the ener-

gy transportation from the keyhole wallto the liquid weld pool are calculated.However, the laser-arc interaction,which has been characterized experi-mentally with optical emission spec-troscopy (Ref. 37), is not rigorously sim-ulated here. During laser-GMA hybrid welding,the rates of heat, mass, and momentumtransport are often enhanced because ofthe presence of fluctuating velocities inthe weld pool. The contribution of thefluctuating velocities is considered bythe incorporation of a turbulence modelthat provides a systematic frameworkfor calculating effective viscosity andthermal conductivity (Refs. 32, 33). Thevalues of these properties vary with thelocation in the weld pool and depend onthe local fluid flow characteristics. Inthis work, a turbulence model based onPrandtl’s mixing length hypothesis (Ref.32) is used to estimate the turbulentviscosity.

Calculation of Heat Transfer fromGMA Metal Droplets

The hot molten metal droplets pro-duced by the GMAW process impingeinto the weld pool at high velocities andcarry a significant amount of heat intothe liquid weld pool (Refs. 34–36, 38,39). The heat transfer from the metal

droplets was simulated by considering acylindrical heat source with a time-aver-aged uniform power density (Sv). Theuse of a cylindrical volumetric heatsource assumes the spray transfer modeof the droplets, which is consistent withthe welding conditions in the presentstudy. In order to calculate Sv, the radiusof the heat source, its effective height,and the total sensible heat input bythe droplets are required. The radiusof the volumetric heat source is as-sumed to be twice that of the dropletradius, and the effective height, d, iscalculated from the following equation(Refs. 34, 38, 39):

d = hv – xv + Dd (7)

where hv is the estimated height of thecavity caused by the impact of metaldroplets, xv is the distance traveled bythe center of the two successive im-pinging droplets, and Dd is the dropletdiameter. The total sensible heat inputfrom the metal droplets, Qd, is given as(Ref. 34)

Qd = prwrw2wfCp(Td – Tl) (8)

where rw is the density of the electrodewire, rw is the radius of the wire, wf isthe wire feeding rate, Td is the droplettemperature, and Tl is the liquid tem-perature. The values of hv and xv in Equation 7are calculated based on an energy bal-ance as (Ref. 39):

f

T T

T TT T

T T T

T T

L

L

S

L SS L

S

=

>−−

≤ ≤

<

⎧

⎨⎪⎪

⎩⎪⎪

1

0

(4)

ht

u ,h

x xk

Chx

Si

i i p ih (5)

( )ρ∂

∂+ ρ

∂∂

= ∂∂

∂∂

⎛

⎝⎜

⎞

⎠⎟ +

SHt

u H

x

Uhx

UH

x

hi

i

i i

( )( )= −ρ∂ Δ

∂− ρ

∂ Δ∂

−ρ ∂∂

− ρ ∂Δ∂

(6)

hD g D g

D v6gv

d d

d d= − γρ

+ γρ

⎛⎝⎜

⎞⎠⎟

+2 2(9)

2 2

Fig. 10 — Calculated CCT diagrams withthe cooling curves superimposed. Coolingrates are taken at the top center of weldswith 1 mm laser arc separation distanceand 20 and 30 mm/s welding speed(welds 1 and 2, respectively). Coolingrates are taken at 1 mm from the bottomsurface of welds with 40 mm/s weldingspeed and 1 and 5 mm laser arc separa­tion (welds 3 and 4, respectively). Thesymbols , W, and a represent allotri­omorphic, Widmanstätten, and acicularferrite, respectively. Ms is the startingtemperature for martensite formation.

Fig. 11 — Variation of the volume frac­tions of allotriomorphic ferrite, Wid­manstätten ferrite, acicular ferrite, andmartensite with different cooling rates.

WELDING RESEARCH

WELDING JOURNAL / APRIL 2015, VOL. 94140-s

where is the surface tension of themolten metal, r is the density, g is accel-eration due to gravity, vd is the dropletimpingement velocity, and Dt is thetime interval between two successivedrops, which is the inverse of thedroplet transfer frequency. As shown inEquations 9 and 10, calculation of thedimensions of the volumetric heatsource requires the knowledge of thedroplet transfer frequency, radius, andimpingement velocity, which can be de-termined from literature (Refs. 34, 38,39). From the calculated values of Qd,Dd, and d, the time-averaged power den-sity of the volumetric heat source, Sv, iscalculated as follows (Ref .34)

Equation 11 is only valid for grid pointswithin the cylindrical heat source, andthe power density is zero outside thecylinder. Free surface calculation of theweld top and bottom surface is not con-sidered in the current model, and thesimplification of flat pool surfaces is notexpected to affect the calculated weldshape and microstructure significantly(Ref. 40).

Phase TransformationCalculation

In the weld fusion zone of low-alloysteels, allotriomorphic ferrite is thefirst phase to form and it nucleatesheterogeneously at the boundaries ofthe columnar austenite grains duringcooling. It is a reconstructive transfor-mation involving diffusion (Ref. 26).As temperature decreases, diffusionbecomes sluggish and gives way to adisplacive transformation. At relative-ly low undercoolings, plates of Wid-manstätten ferrite form by a dis-placive mechanism. At further under-coolings, bainite nucleates and growsin the form of sheaves of smallplatelets. Acicular ferrite nucleates in-tragranularly around inclusions insidethe austenite (Ref. 26). The diffusion-less martensite transformation maytake place if the cooling rate is highenough. The isothermal time-temperature-transformation (TTT) and continu-ous-cooling-transformation (CCT) dia-grams together with various transfor-mation starting temperatures are cal-culated using the phase transforma-tion model based on thermodynamicsand phase transformation kineticswith weld deposit compositions as in-put variables (Refs. 22–24). The incu-bation times for both reconstructiveand displacive transformations are cal-culated by Russell’s expression:

where t is the incubation time for atransformation, T is the temperature,DGmax is the maximum driving forcefor nucleation, and a, b, c, and d areconstants. The details of calculation ofDGmax and determination of a, b, c, andd are given in the literature (Ref. 22). The CCT diagrams are calculatedfrom the corresponding TTT diagrams

SQ

D dv

d

d2=

π(11)

x hD g

cosgh

t

v vd

v

= + γρ

⎛⎝⎜

⎞⎠⎟

−⎛⎝⎜

⎞⎠⎟

Δ⎡

⎣⎢⎢

⎤

⎦⎥⎥

⎧⎨⎪

⎩⎪

⎫⎬⎪

⎭⎪

2

1 (10)1/2

T G expcT

damax

b( )τ = × Δ × ⎛⎝⎜

⎞⎠⎟ × (12)

Fig. 12 — Comparison of the experimental and calculated microstructure volume frac­tions of the weld metal for weld 1 with welding speed of 20.0 mm/s, laser arc separa­tion distance of 1 mm; weld 2 with welding speed of 30.0 mm/s, laser arc separationdistance of 1 mm; weld 3 with welding speed of 40.0 mm/s, laser arc separation dis­tance of 1 mm; and weld 4 with welding speed of 40.0 mm/s, laser arc separation dis­tance of 5 mm. The symbols , W, and a represent allotriomorphic, Widmanstätten,and acicular ferrite, respectively.

A B

Fig. 13 — Computed volume fractions ofallotriomorphic, Widmanstätten, and acic­ular ferrite, and martensite with corre­sponding cooling rates from 1073 to 773 Kof the fusion zone lower center as a func­tion of laser arc separation distance andwelding speed by different laser powers. A— Laser power = 4.0 kW, arc current = 232A, arc voltage = 31 V; B — laser power =5.0 kW, arc current = 232 A, arc voltage =31 V; C — laser power = 6.0 kW, arc cur­rent = 232 A, arc voltage = 31 V. The sym­bols , W, a, and M representallotriomorphic, Widmanstätten, and acic­ular ferrite, and martensite, respectively.

A

B

C

WELDING RESEARCH

APRIL 2015 / WELDING JOURNAL 141-s

based on the Scheil additive rule (Ref.41):

where ta is the incubation time requiredto reach a specified state on a TTT dia-gram for isothermal reactions, t is thetime to that stage for the nonisother-mal reactions, and dt is the time inter-val at temperature T. In this procedure,the total time to reach a specified stateof transformation for nonisothermal re-actions is obtained by adding the frac-tions of time to reach this stage isother-mally until the sum reaches unity. The inclusion of the weld depositcompositions also leads to the calcula-tion of the appropriate part of thephase diagram needed to obtain para-equilibrium compositions for kineticanalysis. These data are combined withaustenite grain parameters and thecomputed cooling curves from the heattransfer and fluid flow model to calcu-late the volume fractions of allotriomor-phic, Widmanstätten ferrite, acicularferrite, and martensite (Ref. 23). Themodeling results are used to better un-derstand the experimentally observedmicrostructures of the weld metal incomplete-joint-penetration hybridlaser-GMA welding.

Results and Discussion

Calculated Temperature andVelocity Fields

The effects of welding speed andlaser arc separation distance on weldprofiles and temperature and velocityfields are studied separately for com-plete-joint-penetration hybrid laser-GMA welding. Both these parametershave been shown to significantly affectthe weld quality and welding productivi-ty. Figure 1 shows the calculated tem-perature and fluid flow fields in threedimensions when the welding speed in-creases from 20 to 30 mm/s with a con-stant laser arc separation distance of 1mm. The temperature is indicated bycontour lines and the velocity field isrepresented by arrows. There is an in-tense temperature gradient on the topand bottom surfaces of the weld poolbecause the temperature of the liquidmetal at the keyhole wall equals theboiling point, while the liquid metal at

the weld pool boundary remains at thesolidus temperature. The molten metalmoves radially outward for both the topand bottom surfaces because of theMarangoni convection produced by thespatial gradient of surface tension re-sulting from the temperature gradient.The maximum outward flow velocitiesof the top surface are 155.6 and 194.2mm/s for welding speeds of 20 and 30mm/s, respectively. The momentum isthen transferred by viscous dissipationto the inner layers of the weld pool adja-cent to the surface. The comparison between these cal-culated and experimental weld poolgeometries is shown in Fig. 2. The topsurface width decreases from 9.6 to 7.0mm, and the bottom surface width de-creases from 3.6 to 2.0 mm when thewelding speed increases from 20 to 30mm/s. The boiling point contours indi-cate the cross-sectional geometry of thekeyhole, while the solidus temperaturecontours indicate the fusion zoneboundary. It is observed that the widthsof the top and bottom surfaces, as wellas the shape of the fusion zone predict-ed by the heat transfer and fluid flowmodel, agreed well with the correspon-ding experimental results. The weldpool dimensions decrease significantlywith the increasing welding speed as thetop surface widens due to Marangoniconvection. Furthermore, the bottomwidth is also larger than the minimumweld profile width at the middle of theplate thickness, indicating the signifi-cant effect of convective heat transfer. Figure 3 shows the calculated tem-perature and fluid flow fields when thelaser arc separation distance increasesfrom 1 to 5 mm with a welding speed of40 mm/s. When the laser arc separationdistance is changed, the nominal totalheat input of the welding process is con-stant, while the heat input decreasesfrom 0.63 to 0.41 kJ/mm with a weld-ing velocity increase from 20 to 30mm/s, as shown in Fig. 2. The comput-ed length of the weld pool increasesfrom 18 to 23 mm with the increasinglaser arc separation distance. The heatfrom the laser and arc as well as thedroplet is more concentrated near thelaser beam incident point at shorterlaser arc separation distance. The heatdistribution over the top surface of theweld pool significantly changes whenthe arc axis is further separated fromthe laser beam. The top part of the liq-

uid weld pool is stretched and the dis-tance along the welding direction fromthe maximum weld width to the laserbeam incidental point is significantly increased. Figure 4 shows the comparison be-tween these calculated and experimen-tal weld pool geometries. The top sur-face width increases from 5.1 to 6.8mm, and the bottom surface width in-creases from 1.2 to 1.6 mm when thelaser arc separation distance increasesfrom 1 to 5 mm. It can be observedfrom Figs. 1 and 3 that the effects ofwelding speed and laser arc separationdistance on the weld profiles are differ-ent. When the welding speed increasesfrom 20 to 30 mm/s, the weld poolshrinks significantly in all locationsalong the weld depth. However, only thetop part of the weld pool significantlyincreases when the laser arc separationdistance increases from 1 to 5 mm. Thebottom part has a slight change and themiddle part of the weld pool is almostunaffected by the laser arc separationdistance. Changes in these processingconditions impact the resulting weldpool geometries in different ways. Forexample, the increase of the weldingspeed reduces the net energy absorbedby the workpiece. On the other hand,the increase of the laser arc separationdistance mainly influences the energydistribution of the heat sources.

Cooling Rates

The cooling rates over the austenitedecomposition range from 1073 to773 K within the weld fusion zone arecalculated using this same heat trans-fer and fluid flow model. The calculat-ed cooling rates are then used in themodeling of the weld metal mi-crostructures of low-alloy steels. Fig-ure 5 shows the calculated coolingcurves between 773 and 1073 K at dif-ferent locations on the top surface ofthe weld with a welding speed of 20mm/s. It can be seen that the coolingrates are almost independent of posi-tion. The cooling rates within the fu-sion zone and the cooling rate in theheat-affected zone (HAZ) are both at alevel of approximately 20 K/s. Thesesimilarities in cooling rates allow a sin-gle cooling condition to be assumedfor each horizontal plane across theweld depth. The comparison of the calculated

∫ ( ) =dtt Ta

t1 (13)

0

WELDING RESEARCH

WELDING JOURNAL / APRIL 2015, VOL. 94142-s

cooling curves at the top surface of thefusion zone for welding velocities of20 and 30 mm/s and a 1-mm separa-tion distance are shown in Fig. 6. Thecalculated cooling rate between 1073and 773 K increases from 21.3 to 48.4K/s with the increasing weld speed.The corresponding cooling times are14.1 and 6.2 s, respectively. As expect-ed, the cooling rate increases with in-creasing welding speed and resultingdecrease in heat input. The calculated cooling curves be-tween 1073 and 773 K along the center-line 1.0 mm above the bottom surfaceof the weld for laser arc separation dis-tances of 1 and 5 mm at a welding speedof 40 mm/s are shown in Fig. 7. This po-sition is chosen because the width ofthe weld pool is the smallest at thisdepth, which indicates that the weldmetal at this region is more prone tocontain martensite. The calculated cool-ing rates from 1073 to 773 K are 83.3and 66.7 K/s when the laser arc separa-tion distance increases from 1 to 5 mm.The corresponding cooling times are 3.6and 4.5 s, respectively. Ribic et al. (Ref.27) reported that for hybrid laser-GTAwelding with heat input of 0.10 kJ/mm,the cooling time from 1073 to 773 K in-creased from 0.45 to 0.75 s with the in-crease of laser arc separation distancefrom 3.5 to 9.2 mm. As can be seenfrom Fig. 3A, B, the weld pool length in-creases with the increasing laser arc sep-aration distance. The isothermals alsoexpand for regions beyond the liquidweld pool so that the spatial gradient oftemperature decreases for the tempera-ture range from 1073 to 773 K. There-fore, the cooling rate decreases with in-creasing laser arc separation distance,which is consistent with the result re-ported by Ribic et al. (Ref. 27). In order to validate the calculation ofthe cooling rates between 1073 and 773K, the calculated cooling rates for differ-ent welding conditions are comparedwith the results available in the litera-ture that examine the effects of changesin heat input and plate thickness on thecooling rate during arc welding (Ref.42). The literature results show thecooling time from 1073 to 773 K forthe welding speeds of 20, 30, and 40mm/s are approximately 13.1, 6.0, and3.9 s, compared with the values of 14.1,6.2, and 3.6 s calculated by the heattransfer and fluid flow model. There-fore, the cooling rates between 1073

and 773 K obtained from the literatureand the 3D heat transfer and fluid flowmodel show good agreement.

Microstructures

The comparison of the fusion zonemicrostructures located at the top sur-face of the weld for welding speeds of20 and 30 mm/s at laser arc separationdistance of 1 mm is shown in Fig. 8.For a welding speed of 20 mm/s, themicrostructure contains 35% allotri-omorphic ferrite, 11% Widmanstättenferrite, and 54% acicular ferrite. Whenthe welding speed increases to 30mm/s, the amount of allotriomorphicferrite decreases to 29%, Widmanstät-ten ferrite increases to 12%, and acicu-lar ferrite increases to 59%. The aver-age length of the acicular ferrite de-creases from 13.1 to 7.1 m and thewidth decreases from 3.1 to 1.2 mwhen the welding speed increasesfrom 20 to 30 mm/s. The hardnesstesting results show the microhard-ness value increases from 223 to 248HV. The average microhardness of thebase metal is 174 ± 11 HV. The differ-ences in microstructure contribute tothe higher microhardness of the weldmetal with a higher welding speed. The comparison of the fusion zonemicrostructures located 1.0 mm abovethe bottom surface of the weld for alaser arc separation distance of 1 and 5mm at a welding speed of 40 mm/s isshown in Fig. 9. The volume fraction ofmartensite is about 64% for the laserarc separation distance of 1 mm, com-pared with a separation distance of 5mm when no martensite is present. Theamounts of allotriomorphic ferrite andWidmanstätten ferrite are about 23 and13%, respectively, for 1-mm separationdistance. The volume fraction of acicu-lar ferrite is about 63%, and theamounts of allotriomorphic ferrite andWidmanstätten ferrite are about 24 and13%, respectively, for a laser arc separa-tion distance of 5 mm. The microhard-ness decreases from 283 to 238 HVwhen the laser arc separation distanceincreases from 1 to 5 mm. The highermicrohardness indicates martensite ispresent and signifies a lower toughnessof the weld (Refs. 21, 43, 44), which isdetrimental to the mechanical proper-ties of the joint. Figure 10 shows the CCT diagramscomputed from the TTT diagram based

on the Scheil additive rule, superim-posed with the cooling curves at select-ed locations in the four welds. The cool-ing curves of the welds with weldingspeeds of 20 and 30 mm/s at a laser arcseparation distance of 1 mm both inter-cept with the diffusive and displacivetransformation curves, so allotriomor-phic ferrite, Widmanstätten ferrite, bai-nite, and acicular ferrite are expected.The cooling curve of the weld with alaser arc separation distance of 1 mm atthe welding speed of 40 mm/s inter-cepts with the upper C curve and themartensite transformation line whilethe cooling curve for the weld with alaser arc separation distance of 5 mm atthe welding speed of 40 mm/s inter-cepts with both the upper and lower Ccurves but not the martensite transfor-mation line. As a result, martensite isexpected with the laser arc separationdistance of 1 mm but not for the laserarc separation distance of 5 mm, al-though the net heat input of the weld-ing process is identical. The variation of the calculated vol-ume fractions of allotriomorphic andWidmanstätten ferrite, acicular ferrite,and martensite with cooling rates isshown in Fig. 11. For the compositionof DH 36 steel and austenite grain sizesobserved in the experimental welds, aci-cular ferrite is the predominant phase,comprising nearly 60% of the mi-crostructure up to the critical coolingrate for martensite. Widmanstätten fer-rite increases with cooling rate at the ex-pense of allotriomorphic ferrite. Thesetwo phases combine to make up 40% ofthe microstructure. The calculated re-sults are consistent with the data re-ported in previous work (Ref. 24). Both the experimentally measuredand calculated quantitative volumefractions of different phases of thefour welds are shown in Fig. 12. Goodagreement between the two sets ofdata is observed. Figure 12A showsthe volume fractions of allotriomor-phic ferrite and acicular ferrite slightlydecrease while Widmanstätten ferriteincreases with the increasing weldingspeed. Figure 12B shows the marten-site volume fraction is about 52%when the laser arc separation distanceis 1 mm, but no martensite is ob-served when the laser arc separationdistance increases to 5 mm. A verysmall amount of acicular ferrite is ob-served experimentally for a laser arc

WELDING RESEARCH

APRIL 2015 / WELDING JOURNAL 143-s

separation distance of 1 mm, com-pared to 5-mm separation distancewhen acicular ferrite is the predomi-nant phase. It should be noted the welding condi-tions including the heat input are allidentical when the laser arc separationdistance changes from 1 to 5 mm. Thereason for the large difference of theweld metal microstructures with thechanging laser arc separation distancelies in the fact that the cooling condi-tion, especially the cooling rate between1073 and 773 K, significantly decreaseswhen the laser arc separation distanceincreases from 1 to 5 mm. The coolingrate exceeds the critical cooling rate formartensite transformation with laserarc separation distance of 1 mm while itdecreases to a value lower than the criti-cal cooling rate for martensite transfor-mation when the laser arc separationdistance increases to 5 mm.

Process Map

In order to further understand theinfluence of welding parameters on thecooling rates and corresponding volumefractions of the microstructures of theweld fusion zone, a comprehensiveprocess map is presented in Fig. 13. Themap considers various combinations ofwelding speed, laser arc separation dis-tance, and laser power. The arc currentand arc voltage are 232 A and 31 V, re-spectively, for all the welding condi-tions. The laser powers for Fig. 13A–Care 4, 5, and 6 kW, respectively. Themaps show the combinations of separa-tion distance and welding speed thatproduce a given cooling rate and corre-sponding microstructure in terms ofvolume fractions of different phasesand microconstituents. The criticalcooling rate for martensite transforma-tion is about 71 K/s, which is calculatedby the model for microstructure evolu-tion (Refs. 22–24). Cooling rates below71 K/s are shown in dashed lines as azone without martensite, while coolingrates greater than 71 K/s are shown indotted lines as another zone withmartensite in Fig. 13. The upper boundon the process maps is the partial pene-tration line, above which complete pen-etration is not possible. In general, welding speed has agreater effect on cooling rate than sep-aration distance, and at low weldingspeeds of 20 mm/s, the cooling rate is

almost independent of separation dis-tance. The effect of power can also beobserved in Fig. 13. As power increas-es, the cooling rates shift to higherwelding speeds for a given separationdistance. For example, at 1 mm sepa-ration distance, the required weldingspeeds to obtain a 70 K/s cooling rateare 32 and 39 mm/s for 4 kW and 6kW laser powers, respectively. It is im-portant to know how cooling rates af-fect the microstructure. As the coolingrate increases from 20 to 150 K/s, thevolume fraction of Widmanstätten fer-rite increases from 0.10 to 0.22 at theexpense of the allotriomorphic ferrite,which decreases from 0.31 to 0.23.The amount of acicular ferrite decreas-es from 0.59 to 0.57 from low coolingrates up to the critical cooling ratewhen acicular ferrite disappears andmartensite forms instead. These maps provide a means for un-derstanding microstructure evolutionduring hybrid laser-arc welding and canbe used to select welding parametersthat optimize the weld microstructureor minimize welding time yet limit theformation of martensite. For example,as stated previously, if the power is in-creased from 4 to 6 kW for a constantseparation distance of 1 mm, then thewelding speed can be increased by 22%from 32 to 39 mm/s without the forma-tion of martensite. Increasing the sepa-ration distance from 1 to 7 mm furtherincreases the possible welding speed to42 mm/s. Compared to the initial condi-tions in this example, the total increasein welding speed is 31% due to a 33%increase in laser power and a 6-mm in-crease in separation distance. The weld-ing speed and the resulting time tomake a weld are important. When com-paring the costs of hybrid laser-GMAwelding with conventional arc weldingfor pipe joining applications, Reutzel etal. (Ref. 45) found that weld time com-prised between 24 and 41% of the totalfabrication time, depending on pipe di-ameter, with other tasks, such as fitup,preparation of the weld, and movementof the part, making up the rest of thetime. Additionally, since the welding isoccurring with the same cooling rate,there would be no change in the mi-crostructure. The utility of the processmap indicates significant promise forunderstanding the evolution of mi-crostructures in the fusion zone of hy-brid laser-GMA complete-joint-penetra-

tion welding by a combination of phasetransformation model and the thermalcycles calculated from the heat transferand fluid flow model.

Summary and Conclusions The effect of laser arc separation dis-tance and welding speed on fusion zonegeometry and microstructure duringcomplete-joint-penetration hybridlaser-gas metal arc welding of low-alloysteel was investigated experimentallyand theoretically. A heat transfer andfluid flow model was used to calculatethe weld pool geometry and coolingrates, which were applied to a mi-crostructure model to compute thephase fractions of selected microcon-stituents. Experimental weld geome-tries and microstructures were com-pared to the calculated values, and thetwo sets of data agreed well. The follow-ing conclusions can be drawn from thiswork: 1. The effect of welding speed andlaser arc separation distance on weldpool geometry was investigated in com-plete-joint-penetration hybrid laser-arcwelding. The weld length and widthboth at the top and bottom of the poolincreased with increasing laser arc sepa-ration distance for the same heat input.The weld pool dimensions decreasedwith increasing welding speed as ex-pected. 2. Cooling rate was also affected bythe hybrid welding parameters. Whenthe welding speed increased, whichchanged the net heat input, the coolingrate increased. When the laser arc sepa-ration distance decreased, which alteredthe heat distribution of the combinedpower sources but not the heat input,the cooling rate increased. At high weld-ing speeds, the decrease in separationdistance can have a significant enougheffect on the cooling rate to formmartensite in the microstructure. 3. The experimental weld mi-crostructures consisted of acicular fer-rite, allotriomorphic ferrite, Wid-manstätten ferrite, and martensite.Martensite existed only in the weld withthe laser arc separation distance of 1mm and a welding speed of 40 mm/s.Acicular ferrite formed at the expenseof martensite when the separation dis-tance increased to 5 mm. Increasing thewelding speed from 20 to 30 mm/s re-sulted in a decrease of allotriomorphic

WELDING RESEARCH

WELDING JOURNAL / APRIL 2015, VOL. 94144-s

ferrite and an increase in Widmanstät-ten and acicular ferrite. 4. Utilizing the combined models, aprocess map has been constructed thatillustrates the effects of welding speed,separation distance, and laser power oncooling rate and microconstituent vol-ume fractions. Identical cooling ratesand microconstituent volume fractionscan be obtained with various combina-tions of process parameters. In general,welding speed and laser power have amore significant effect on cooling ratesthan separation distance; however, athigh welding speeds, the separation dis-tance can be the deciding factor formartensite formation.

The authors would like to thankMr. Jay Tressler for performing thehybrid welding experiments, Mr. EdGood and Dr. Jay Keist for assistingwith the metallography, and Prof.Lars-Erik Svensson for assisting withthe microstructure identification andquantification. This research was per-formed using funding received fromthe DOE Office of Nuclear Energy’sNuclear Energy University Programsunder Grant Number 120327.

1. Steen, W. M. 1980. Arc augmented laserprocessing of materials. Journal of Applied Physics51(11): 5636–5640. 2. Tusek, J., and Suban, M. 1999. Hybrid weld-ing with arc and laser beam. Science and Technologyof Welding and Joining 4(5): 308–311. 3. Page, C. J., Devermann, T., Biffin, J., andBlundell, N. 2002. Plasma augmented laser weld-ing and its applications. Science and Technology ofWelding and Joining 7(1): 1–10. 4. Chen, Y. B., Lei, Z. L., Li, L. Q., and Wu, L.2006. Experimental study on welding characteris-tics of CO2 laser TIG hybrid welding process. Sci-ence and Technology of Welding and Joining 11(4):403–411. 5. Zhou, J., and Tsai, H. L. 2008. Modeling oftransport phenomena in hybrid laser-MIG key-hole welding. International Journal of Heat andMass Transfer 51(17-18): 4353–4366. 6. Cho, J. H., and Na, S. J. 2009. Three-dimen-sional analysis of molten pool in GMA-laser hy-brid welding. Welding Journal 88(2): 35-s to 43-s. 7. Vollertsen, F., Grunenwald, S., Rethmeier,M., Gumenyuk, A., Reisgen, U., and Olschook, S.2010. Welding thick steel plates with fibre lasersand GMAW. Welding in the World 54(3–4):R62–R70. 8. Le Guen, E., Fabbro, R., Carin, M., Coste, F.,and Le Masson, P. 2011. Analysis of hybrid

Nd:YAG laser-MAG arc welding processes. Opticsand Laser Technology 43(7): 1155–1166. 9. Cao, X., Wanjara, P., Huang, J., Munro, C.,and Nolting, A. 2011. Hybrid fiber laser-arc weld-ing of thick section high strength low alloy steel.Materials and Design 32(6): 3399–3413. 10. Roepke, C., Liu, S., Kelly, S., and Mar-tukanitz, R. 2010. Hybrid laser arc weldingprocess evaluation on DH36 and EH36 steel.Welding Journal 89(7): 140-s to 150-s. 11. Bagger, C., and Olsen, F. O. 2005. Reviewof laser hybrid welding. Journal of Laser Applica-tions 17(1): 2–14. 12. Gao, M., Zeng, X. Y., and Hu, Q. W. 2006.Effects of welding parameter on melting energy ofCO2 laser-GMA hybrid welding. Science and Tech-nology of Welding and Joining 11(5): 517–522. 13. Rao, Z. H., Liao, S. M., and Tsai, H. L.2011. Modeling of hybrid laser-GMA welding: Re-view and challenges. Science and Technology ofWelding and Joining 16(4): 300–305. 14. Callister, W. D., and Rethwisch, D. G. 2009.Materials Science and Engineering. Wiley, U.S.A. 15. Roepke, C., and Liu, S. 2009. Hybrid laserarc welding of HY-80 steel. Welding Journal 88(8):159-s to 167-s. 16. Gao, M., Zeng, X. Y., Hu, Q. W., and Yan, J.2008. Weld microstructure and shape of laser-archybrid welding. Science and Technology of Weldingand Joining 13(2): 106–113. 17. Moore, P. L., Howse, D. S., and Wallach,E.R. 2004. Microstructures and properties oflaser-arc hybrid welds and autogenous laser weldsin pipeline steels. Science and Technology of Weldingand Joining 9(4): 314–322. 18. Swanson, P. T., Page, C. J., Read, E., andWu, H. Z. 2007. Plasma augmented laser weldingof 6 mm steel plate. Science and Technology of Weld-ing and Joining 12(2): 153–160. 19. Chen, Y. B., Feng, J. C., Li, L.Q., Chang, S.,and Ma, G. L. 2013. Microstructure and mechani-cal properties of a thick-section high-strengthsteel welded joint by novel double-sided hybrid fi-bre laser-arc welding. Materials Science and Engi-neering A – Structural Materials Properties Mi-crostructure and Processing 582: 284–293. 20. Gao, M., Zeng, X. Y., Yan, J., and Hu, Q. W.2008. Microstructure characteristics of laser-MIGhybrid welded mild steel. Applied Surface Science254(18): 5715–5721. 21. McPherson, N. A., Suarez-Fernandez, N.,Moon, D. W., Tan, C. P. H., Lee, C. K., and Baker, T.N. 2005. Laser and laser-assisted arc weldingprocess for DH 36 microalloyed steel ship plate.Science and Technology of Welding and Joining 10(4):460–467. 22. Bhadeshia, H. K. D. H. 1982. Thermody-namic analysis of isothermal transformation dia-grams. Metal Science 16(3): 159–165. 23. Bhadeshia, H. K. D. H., Svensson, L. E.,and Gretoft, B. 1985. A model for the develop-ment of microstructure in low-alloy steel (Fe-Mn-Si-C) weld deposits. Acta Metallurgica 33(7):1271–1283. 24. Bhadeshia, H. K. D. H., and Svensson, L. E.1993. Modeling the evolution of microstructure insteel weld metal. Mathematical Modeling of WeldPhenomena. H. Cerjak and K. E. Eastering, eds. In-stitute of Materials, London, pp. 109–180. 25. Mundra, K., and DebRoy, T. 1996. Numer-ical prediction of fluid flow and heat transfer inwelding with a moving heat source. Numerical heattransfer, Part A 29: 115–129. 26. Mundra, K., and DebRoy, T. 1993. Calcula-tion of weld metal composition change in high-

power conduction mode carbon-dioxide laserwelded stainless steels. Metallurgical TransactionsB – Process Metallurgy 24(1): 145–155. 27. Ribic, B., Rai, R., and DebRoy, T. 2008. Nu-merical simulation of heat transfer and fluid flowin GTA/Laser hybrid welding. Science and Technolo-gy of Welding and Joining 13(8): 683–693. 28. International Institute of Welding (IIW),Guide to the light microscope examination of fer-ritic steel weld metals. 1991. Welding in the World29: 160–176. 29. Rai, R., Kelly, S. M., Martukanitz, R. P.,and DebRoy, T. 2008. A convective heat transfermodel for partial and full penetration keyholemode laser welding of a structural steel. Metallur-gical and Materials Transactions A 39A: 98–112. 30. Kaplan, A. 1994. A model of deep penetra-tion laser welding based on calculation of the key-hole profile. Journal of Physics D: Applied Physics27: 1805–1814. 31. Zhao, H., and DebRoy, T. 2003. Macrop-orosity free aluminum alloy weldments throughnumerical simulation of keyhole mode laser weld-ing. Journal of Applied Physics 93: 10089–10097. 32. Wilcox, D. C. 1993. Turbulence Modeling forCFD (California: DCW Industries). 33. Launder, B. E., and Spalding, D. B. 1972.Lectures in Mathematical Models of Turbulence. Lon-don, UK: Academic Press. 34. Zhang, W., Kim, C. H., and DebRoy, T.2004. Heat and fluid flow in complex joints dur-ing gas metal arc welding — Part II: Application tofillet welding of mild steel. Journal of AppliedPhysics 95(9): 5220–5229. 35. Essers, W. G., and Walter, R. 1981. Heattransfer and penetration mechanisms with GMAand plasma-GMA welding. Welding Journal 60(2):37-s to 42-s. 36. Lancaster, J. F. 1984. The Physics of Weld-ing. New York, N.Y.: Pergamon Press. 37. Ribic, B., Burgardt, P., and DebRoy, T.2011. Optical emission spectroscopy of metal va-por dominated laser-arc hybrid welding plasma.Journal of Applied Physics 109(8): 083301. 38. Kumar, S., and Bhaduri, S. C. 1994. Three-dimensional finite element modeling of gas metal-arc welding. Metallurgical and Materials Transac-tions B 25B: 435–441. 39. Yang, Z., and DebRoy, T. 1999. Modelingmacro- and microstructures of gas-metal-arc weld-ed HSLA-100 steel. Metallurgical and MaterialsTransactions B — Process Metallurgy and MaterialsProcessing Science 30(3): 483–493. 40. Kou, S. 2012. Fluid flow and solidificationin welding: Three decades of fundamental re-search at the university of Wisconsin. WeldingJournal 91(11): 287-s to 302-s. 41. Takahashi, M., and Bhadeshia, H. K. D. H.1991. A model for the microstructure of some ad-vanced bainitic steels. Materials Transactions32(8): 689–696. 42. Masubuchi, K. 1980. Analysis of WeldedStructures. New York, N.Y.: Pergamon Press. 43. Ribic, B., Palmer, T. A., and DebRoy, T.2009. Problems and issues in laser-arc hybridwelding. International Materials Reviews 54(4):223–244. 44. Sevim, I. 2006. Effect of hardness to frac-ture toughness for spot welded steel sheets. Mate-rials and Design 27(1): 21–30. 45. Reutzel, E. W., Sullivan, M. J., and Mikesic,D. A. 2006. Joining pipe with the hybrid laser-GMAW process: Weld test results and cost analy-sis. Welding Journal 85(6): 66–71.

References

Acknowledgments