Journal of Materials Processing Tech

Contents lists available at ScienceDirect

Journal of Materials Processing Tech.

journal homepage: www.elsevier.com/locate/jmatprotec

Numerical design of high depth-to-width ratio friction stir welding

Yongxian Huang⁎, Yuming Xie, Xiangchen Meng, Zongliang Lv, Jian CaoState Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, Harbin 150001, PR China

A R T I C L E I N F O

Keywords:High depth-to-width ratioFriction stir weldingAl-Mg-Si alloyHigh-throughput screeningGeometry optimizationMicrohardness

A B S T R A C T

Tools with a high depth-to-width ratio of 0.6 were designed and tested. The thread structure of pin couldpromote material flow but increase fracture risk. The pins with milling facets were more beneficial to achievesound joint. The tapered thread tool with triple facets was the optimal structure for high depth-to-width ratiofriction stir welding at a rotational speed of 800 rpm and a welding speed of 30 mm/min. The measured tem-perature and joint formation agreed with the predicted data well. A narrower average heat affected zone withthe width of 440 μm was obtained, which was far lower than that underwater friction stir welding.

1. Introduction

For conventional friction stir welding (FSW), welding tool consistsof shoulder and pin, in which the ratio of pin to shoulder is about 0.33.The research of Zhang et al. (2011a,b) showed that the shoulder wasresponsible for main part of heat generation. Larger shoulder diameterresulted in that higher heat input concentrated on the surface of joint,which broadened the width of the heat affected zone (HAZ) and re-duced mechanical properties. The rotational pin is the main drivenforce of material flow around the tool. In order to obtain high-qualityjoint, decreasing shoulder diameter and optimizing pin geometricalstructure are of vital importance. However, decreasing heat input bysimply reducing the shoulder diameter easily results in the fracture ofthe pin and formation of defect. Obviously, it is not an acceptablechoice to attempt massive possibilities one by one due to its potentialrisk of pin fracture. Thus, numerical methods were proposed to providea better alternative to welding tool design, which saved time cost andimprove the efficiency of the design process.

Up to present, computer aided engineering (CAE) has potential toelucidate joining mechanism without limitation of objective conditions.The numerical methods CAE are divided into two classical categories:computational solid mechanics (CSM) and computational fluid dy-namics (CFD). Due to its difficulty to guarantee convergence and cal-culate time of CSM methods, researchers utilize CFD methods to makethe numerical process. Feulvarch et al. (2013) proposed a robustmoving mesh technique based on a Eulerian formalism. They dividedthe mesh into two parts: a first one fixed around the stirring zone and asecond one including the base material near the tool and moving with arotational solid motion. Mesh distortions were avoided and computingtime was reduced effectively. Zhu et al. (2016) assumed that the

materials around the tool flow non-uniformly distributed on the inter-face under the driving force of friction. The non-uniform friction forcemodel provided a potential tool for predicting defects. Moreover, Aroraet al. (2011) proposed a numerical criterion for the topology ofshoulder diameter based on the principle of maximum utilization ofsupplied torque for traction, which provided a novel approach to de-veloping tool design. Shi et al. (2015) developed a model to evaluatethe effect of the welding parameters and shoulder size in reverse dual-rotation FSW.

To summarize, conventional computational methods have beenwidely accepted during FSW. However, when designing a new tool fromnonexistence to pass into existence, conventional method is difficult torealize. High-throughput screening method, based on the marriagebetween massively parallel computational methods and existing data-base containing the calculated properties, is capable to explore hy-pothetical candidates. In this study, high depth-to-width ratio FSW wasproposed to reduce the width of HAZ and then increase mechanicalperformances. A high-throughput geometry design method was pro-posed to design high depth-to-width ratio FSW tools. From the aspectsof tool fracture, defect prediction, joint formation and width of HAZ,high depth-to-width ratio FSW was mainly investigated in detail.

2. Materials and experimental procedure

Al-Mg-Si alloy rolled sheet was used in this high depth-to-widthratio friction stir butt welding, of which the microhardness was about95 HV. A plunge depth of 0.15 mm and a tilt angle of 2.5° were con-stant. Particular tools with features of high depth-to-width ratio, madeof HS6-5-2C high speed steel, were used as illustrated in Fig. 1. Thesizes of the tools are respectively 8 mm and 4.8 mm in the shoulder

http://dx.doi.org/10.1016/j.jmatprotec.2017.09.029Received 13 August 2017; Received in revised form 17 September 2017; Accepted 17 September 2017

⁎ Corresponding author.E-mail address: [email protected] (Y. Huang).

Journal of Materials Processing Tech. 252 (2018) 233–241

Available online 20 September 20170924-0136/ © 2017 Elsevier B.V. All rights reserved.

MARK

http://www.sciencedirect.com/science/journal/09240136

https://www.elsevier.com/locate/jmatprotec

http://dx.doi.org/10.1016/j.jmatprotec.2017.09.029


mailto:[email protected]


http://crossmark.crossref.org/dialog/?doi=10.1016/j.jmatprotec.2017.09.029&domain=pdf

hyxhi

高亮

diameter and the pin length, which have a depth-to-width ratio of 0.6.The ratio of pin length to shoulder diameter used in conventional FSWsituations of aluminum alloy is relatively low than 0.33, as counted inFig. 2. The workpiece temperature field was recorded by four thermo-couples which were fixed in blind holes with 1 mm diameter. Coupleswere named from T1 to T4, and were placed in their coordinates, asshown in Fig. 1. Specimens were cut from the welded workpiecetransverse to the welding direction to carry out microstructural char-acterization. To analyze the microstructural evolution, the section ofjoint was observed by optical microscopy.

3. Numerical modeling

3.1. Mesh and domain framework

A computational domain with dimensions of 5 mm× 55mm × 50 mm was used in the mathematical evaluation model. Fourkinds of tool geometry design, consisting of thread and milling facets,equal to the geometry parameters and plunging depth, were used tosimulate the welding process. As illustrated in Fig. 3, the domain was

divided into two relative zones by sliding mesh model (SMM) with themoving mesh and the fixed mesh. The origin of Cartesian coordinatelocated at center of the pin root, which was at rest relative to the fixedmesh. The moving mesh rotated synchronously with the welding tool.Though the diameter of the moving mesh has no effect on the numericalresults, the interface between two zones could cause slight discontinuityon strain rate and temperature. The diameter was set as 13 mm in orderto exceed the confines of the banded structures in the model.

3.2. Governing equations

The FSW process includes three subsequent procedures with plun-ging, dwelling and pulling out. The dwelling period was mainly con-sidered. The transient model with SMM was proceeded to achieve aquasi-steady state. The solver was set as pressure-based method. The Al-Mg-Si alloy was assumed as an incompressible, non-Newtonian, con-tinuous fluid. The Navier-Stokes equations of mass, momentum andenergy can be written as:

∇ =v· 0 (1)

= −∇ + ∇ρ dvdt

P σ·(2)

= −∇ ∇ + ∇ −ρc dTdt

K T σ v ϕ( ) ( : )p (3)

where ρ is the density of aluminum alloy, t is the flow time, v is flowvelocity, = ∇ + ∇σ μ v v( )t is the deviatoric stress tensor, μ is the non-Newtonian viscosity, P is the relative pressure, T is the absolute tem-perature, ϕ is the sink term which represents the heat loss of convectionand radiation, cP is the specific heat capacity and k is the thermalconductivity.

3.3. Material property

Temperature-dependent thermal conductivity and specific heat ca-pacity of the Al-Mg-Si alloy rolled sheets are given in Fig. 4.

Attested by Zienkiewicz and Cormeau (1974), the non-Newtonianviscosity of metal materials μ is dependent of temperature and effectivestrain rate, as written below:

=με3 e

σ·e

(4)

where σe is the effective stress which is regarded as the driving force tosustain the non-Newtonian plastic deformation, εe

·is the effective strain

rate:

=σ σ σ32e ij ij (5)

Fig. 1. (a) Illustration of FSW process, (b)Conventional tapered pin, (c) Tapered pin with triplefacets, (d) Tapered thread pin and (e) Tapered threadpin with triple facets.

Fig. 2. Pin length and shoulder diameter from relevant references for FSW of aluminumalloy. The green, red and yellow shape means that these were counted from journal pa-pers between 2012 to 2017, journal papers before 2012 and conference papers, respec-tively. The points meant single data while the squares and lines meant range data. (Forinterpretation of the references to colour in this figure legend, the reader is referred to theweb version of this article.)

Y. Huang et al. Journal of Materials Processing Tech. 252 (2018) 233–241

234

=ε ε ε23e ij ij

· · ·

(6)

Here, σij and εij·are the components of the deviatoric stress tensor and

the strain rate tensor, respectively. The strain rate tensor is a symmetricsecond-order tensor, which represents deformation parts of Helmholtzvelocity decomposing theorem:

= + =ε v v i j12

( ) , 1,2,3ij i j j i·

, , (7)

The Sellars-Tegart constitutive model used by Sheppard andJackson (1997) and ameliorated by Su et al. (2015) determines thestress to sustain the plastic deformation:

⎜ ⎟= ⎛⎝

− −−

⎞⎠

⎡

⎣⎢⎛

⎝⎞⎠

⎤

⎦⎥ +−σ T

T αZA

σ1 273.15273.15

1 sinhemelt

nmelt

11

(8)

Here, α, A and n are material constants, Tmelt is the melt temperature ofAl-Mg-Si alloy, σmelt is the yield stress above the melting point. Z is theZener-Hollomon parameter, which is defined as:

=Z ε eeQ

RT·

(9)

where Q is the activation energy, R is the gas constant. The constantsused above were listed in Table 1.

3.4. Velocity near-tool-surface

On the premise that the surface contact exists between tool surface

and welding metals, the viscoplastic materials will be driven by thetools. Slip rate η is used to define the speed difference between two wallsurfaces, which is defined by Nandan et al. (2007):

= − −δ e1ωR

δ ω Rijk

s0 0 (10)

where δ0 is a constant taken as 0.4, ω0 is a constant to non-dimensionalize the rotational speed, Rs is the radius of the shoulder.Fig. 5 limns the velocity distributions on longitudinal and transversesections. The horizontal velocity is proportion to rotational speed ω, sothe velocity on any infinitesimal body can be defined as:

⎧

⎨

⎪

⎩⎪

====

v ηωRv v cosθ

v v sinθv v sinξ

horizon ijk

shear horizon ijk

normal horizon ijk

vertical horizon ijk (11)

The other boundaries except inlet and outlet boundaries were con-sidered as stationary wall with zero slip shear resistance. The velocitiesof inlet and outlet boundaries were in accordance with the weldingspeed, while the injections were released from the section of the inletboundary.

3.5. Heat generation equations

Friction and plastic deformation heats were synchronous con-sidered. The friction heat source term F1 was surficial heat source andrepresented the friction between tool and workpiece. F1 equaled to zero

Fig. 3. Schematic view of high depth-to-width ratio FSW numerical model.

Fig. 4. Temperature-dependent thermal conductivityand specific heat capacity of Al-Mg-Si alloy.


235

when the surfaces were separated. Related numerical results likeTutunchilar et al. (2012) attested that there was cavitation phenom-enon occurring due to negative pressure, as shown in Fig. 6. The frictionheat was governed as follows:

⎧

⎨⎩

= ⎡⎣ + − ⎤⎦ >

= ≤

F β R ω δ μP ξ cosθ P

F P

· (1 ) sin · 0

0 0ijk

δijk ijk ijk ijk

ijk

1 1 3

1

e

(12)

The plastic deformation heat source term F2 is volumetric heatsource, which is given as:

=F β σ ε· · e2 2 e·

(13)

where β1 and β2 are the empirical coefficient which represents thermalconversion rate, μ is the friction coefficient which is taken as 0.4.

The tool is made of high speed steel, and heat will be conducted tothe tool and dissipate in the end. It is of vital importance to calculatethe percentage of heat transported into the tool, which can be definedby:

=+

λ JJ J

w

w T (14)

= kρCJ p (15)

where Jw and JT denote the effusivity of the workpiece and the tool,respectively. k, ρ and Cp are the thermal conductivity, the density andthe heat capacity, respectively.

3.6. Thermal boundary conditions

The boundary heat exchange including convection and radiation ofthe surface is given as follows:

− ∂∂

= − + −∞ ∞k Tz

h T T σε T T( ) ( )4 4(16)

where σ is the Stefan-Boltzmann constant, ε is the emissivity which istaken as 0.4, h is the heat transfer coefficient which is taken as 15 W/(m−2·K−1), 15 W/(m−2·K−1) and 200 W/(m−2·K−1) on upper surface,lateral surface and lower surface, respectively.

3.7. Fluid-solid interface method for tool fracture analyses

As aforementioned, high depth-to-width ratio FSW tools face risk ofpin fracture. It is necessary to find out when the pin fracture will occur.Fluid-solid interface method was proposed to investigate the tool stressstate. The pressure and temperature of the workpiece were transferredto the tool to achieve coupled fluid-solid surface, as shown in Fig. 7.

4. Results and discussion

4.1. Temperature distribution and validation

To validate the reliability of numerical model, a high depth-to-width ratio FSW joint was welded by a tapered thread pin with triplefacets at a welding speed of 50 mm/min and a rotational speed of1200 rpm. The simulated temperature distributions along the long-itudinal, transverse section around the tool and on the upper surface ofthe fluid domain are shown in Fig. 8a, b and c, respectively. Thecomputed peak temperature is about 719 K. The temperature on ad-vancing side (AS) is slightly higher than that on retreating side (RS),which agrees with Mishra and Ma (2005). Fig. 9 shows the historycurves of temperature. The thermal curve corresponds well with that ofexperimental data measured by thermocouples. The average relative

Table 1Material constants and other constants.

Constant α ln(A) n Tmelt σmelt Q R

Value 1.7 × 10−8 Pa−1 27.418 6.66 908 K 20 MPa 181.53 kJ/mol 8.314 J/(mol·K)

Fig. 5. Velocity distributions near-tool-surface: (a)horizontal section and (b) transverse section.

Fig. 6. Illustration of negative pressure cavitation during FSW.


236

deviation of the prediction temperature is 1.7%.

4.2. Fracture criteria

Fig. 10 shows evaluation criteria of fracture potential possibility byfluid-solid interface method. To examine the tool fracture due to thelack of the strength, curves of tensile strength versus temperature andthe computed maximal effective stress are plotted. The green shadesindicate that the tool can work without risk of fracture, while theyellow shades indicate that the tool is likely to crack. The red zone ontools means the distribution position of maximal effective stress, whichis the potential crack initiation position. For example, the taperedthread pin with triple facets at a welding speed of 30 mm/min and arotational speed of 800 rpm locates on the position of the blue penta-gram, which means it will work regularly. The tapered thread pin withtriple facets at a welding speed of 100 mm/min and a rotational speedof 800 rpm locates on the position of the red hexagon, which is the sign

of fracture.Table 2 shows the potential possibility of fracture. At a certain ro-

tational speed, lower welding speed reduces the maximal effectivestress and the potential risk of fracture. The tapered thread pin withtriple facets is the only geometrical structure to remain intact whenwelding speed reaches to 30 mm/min. The riveting discovery is that themaximal effective stress of 587.6 MPa by tapered pin with triple facetsis significantly lower than those of thread pins at a welding speed of10 mm/min, while the stress of thread pins is relatively lower than

Fig. 7. Schematic illustration of Fluid-solid interface method during FSW.

Fig. 8. The computed temperature contour (K): (a)longitudinal section, (b) transverse section and (c)upper surface.

Fig. 9. History thermal curves of experiment and numerical simulation.

Fig. 10. Evaluation criteria of fracture potential possibility.


237

conventional tapered pin. It is conjectured that the thread structure canpromote the material flow but increase the fracture risk.

4.3. Defect prediction

The particle tracing technique is applied to predict the formation ofhigh depth-to-width ratio FSW influenced by material flow and relatedfactors comprehensively. The injections are set as massless particles,which are superimposed on the inlet surface with a random radius of0.3 mm. The boundary conditions for injections are identical to thematrix metal. As shown in Fig. 11, the particles appear grey and thegrey mass means dense particle distribution. The region where the red

background exposes, means the occurrence of potential defects. Redbackground indicates the infeasibilities of this geometry with certainparameter. Defects mainly distribute in the welding nugget zone (WNZ)of AS and the root of joint. The defects at AS can be attributed to theinsufficient flow, and the defects in the root of joint results from ex-treme low heat input. It is necessary to improve the effect of stir andraise root heat input appropriately.

To describe the joint formation, an index to evaluate the distribu-tion evenness of the particles was defined:

= ×CV SDdX

100% (17)

where SDd is the color standard deviation within 5 mm from theweldline, X is the color average value within 5 mm from the weldline,CV is the evaluation coefficient. The formation comes better when CVdecreases. According the function, tapered thread pin with triple facetscomes the best results of which CV value is 4.512%, as shown in Fig. 11.Both tapered pin with triple facets and tapered thread pin with triplefacets which have the geometrical structure of milling facets are morelikely to perform high-quality joint for the lower CV value.

4.4. High-throughput screening

Both tool fracture analyses and joint formation are of vital im-portance to realize sound joints. These two influence factors into anevaluation system were taken which is processed by a massively par-allel way. Fracture evaluation and defect prediction high-throughputmaps comprising all 1600 systems considered are shown in Fig. 12. Allthose systems for which CV value is less than 5% without fracture areincluded in the list of potential candidates. The parameters are ar-ranged linear scale in increasing order. Each circle, pentagram, hexa-gram and square represent a parameter combination. A grey squareindicates that there are risks of fracture, which should be avoided. Acolored circle or a colored pentagram indicates that it can be welded

Table 2Fracture state evaluated criteria.

Geometrical structure A B C

Tapered pin Parameter (rpm,mm/min)

800, 50 800, 30 800, 10

Maximal EffectiveStress (MPa)

2050 1940 1446

State Fracture Fracture FractureTapered pin with triple

facetsParameter (rpm,mm/min)

800, 50 800, 30 800, 10


1225 830 588

State Fracture Fracture SafeTapered thread pin Parameter (rpm,

mm/min)800, 50 800, 30 800, 10


1535 1386 900

State Fracture Fracture SafeTapered thread pin with

triple facetsParameter (rpm,mm/min)

800, 50 800, 30 800, 10


1179 750 656

State Fracture Safe Safe

Fig. 11. Particle distribution at outlet plane and itsCV value: (a) tapered pin, (b) tapered thread pin, (c)tapered pin with triple facets and (d) tapered threadpin with triple facets.


238

without tool fracture. The color contrast from yellow to red indicatesincrease of CV value and deterioration of the joint formation. Thoseyellow pentagrams are the systems of which CV value is less than 5%.There are 0, 0, 1 and 6 candidates in the geometry assembly of taperedpin, tapered thread pin, tapered pin with triple facets, tapered threadpin with triple facets, respectively. A lower CV value means a sounderjoint. The system of tapered thread pin with triple facets at a weldingspeed of 30 mm/min and a rotational speed of 800 rpm is regarded asthe optimized geometry and parameters.

Tapered thread pin with triple facets has a geometry of threadstructure and milling facets. In spite of potential weakening of the tools,the thread structure effectively promotes the material flow. Under theassistance of milling facets to promote the flow and soften the materialsby increasing deformation heat, the structure is beneficial to realize thejoint formation successfully without crack possibilities. These para-meters give a feasible environment to ensure the realization of thedense and sound joint.

4.5. Joint formation

Based on the optimized system evaluated by high-throughputscreening, the joint produced by tapered thread pin with triple facets ata welding speed of 30 mm/min and a rotational speed of 800 rpm wasobtained. As shown in Fig. 13, the surface presents smooth whichmeans lower heat input. Three distinct zones of heat-affected zone

(HAZ), thermo-mechanically affected zone (TMAZ) and WNZ areidentified, as shown in Fig. 14. The macrostructure displays lanky shapedue to the high depth-to-width geometry of the welding tool. The in-terface between TMAZ and WNZ on the AS is obviously clear thanconventional FSW joint obtained by Cabibbo et al. (2007). However,the interface on the RS becomes obscure. It is inferred that only a few ofgrains in the TMAZ on the RS flows into the region behind the pin onthe AS, while metal is driven by the shear force of the rotational tool tofill the gap behind on the RS resulting in a clear interface betweenequiaxed WNZ grains and elongated TMAZ grains. Interestingly, themultiple onion ring patterns are observed in the WNZ stacked vertically

Fig. 12. The parameters are arranged linear scale inincreasing order. Each circle, pentagram, hexagramand square represents a parameter combination. Agrey square indicates that there are risks of fracture,and this combination will be ignored. A coloredcircle, a hexagram or a colored pentagram indicatesthat it can be welded without fracture. The colorcontrast from yellow to red indicates increasing CVvalue and deterioration of the joint formation. Thehexagram indicates the optimized parameter combi-nation: (a) tapered pin, (b) tapered pin with triplefacets, (c) tapered thread pin and (d) tapered threadpin with triple facets. (For interpretation of the re-ferences to colour in this figure legend, the reader isreferred to the web version of this article.)

Fig. 13. Optical surface macrograph of high depth-to-width ratio FSW joint for 5 mm Al-Mg-Si alloy rolled sheet.


239

through the thickness and close to the AS. Onion rings have alternatebright and dark rings. Since the formation of onion rings is believed tobe the results of maximal effective plastic strain rate by Tongne et al.(2017), the multiple onion rings found in the WNZ are a direct evidenceof thread-driven material transport phenomena during high depth-to-width ratio FSW. The flow velocity near-tool-surface can be defined as:

=v ηωRhorizon ijk (18)

There is a great difference of flow velocity between the top and thebottom of the thread, which results in periodic drastic changes of strainrate tensor. The multiple onion rings pattern stackes vertically throughthe thickness emerge. For the Al-Mg-Si alloys, the zigzag line easilyappears at the WNZ, which is detrimental to tensile property as

described by Dai et al. (2015). In this study, the optimized taperedthread pin with triple facets can produce the multiple onion rings,which results in the elimination of zigzag line and mechanical inter-locking, improving mechanical properties.

The temperature distribution of transverse section is shown inFig. 15. The attenuation of the peak temperature occurs rapidly fromwelding center to sides, which generates less heat effect to the joints. InFig. 14, the width of the HAZ is lower than conventional FSW notably,which is attributed to the lower heat input. Microhardness distributionof the high depth-to-width ratio FSW joint is shown in Fig. 16. Thehardness gap of the WNZ between upper line and others implies that theshoulder affected zone is softened by more heat accumulation than thepin affected zone. The hardness of the TMAZ at the middle is ap-proximately equivalent to that at the upper due to the similar accu-mulated strain and thermal cycling, as shown in Fig. 15. The width ofHAZ ranges from 270 μm to 660 μm, while the average value is about440 μm. As described by Mishra and Ma (2005), HAZ is usually to bethe weakness of FSW joint. This technique can reduce the width of HAZsignificantly, which effectively diminishes the disadvantages of widerHAZ. Compared with underwater FSW by Zhang et al. (2011a,b), themechanism of narrowed HAZ induced by high depth-to-width ratio FSWis reducing heat input significantly while that of underwater FSW isincreasing heat dissipation rate. High depth-to-width ratio FSW reducesthe width of HAZ further, which means the potential promotion ofmechanical performance.

5. Conclusions

(1) The numerical evaluation model was proved to be accurate andpractical for high-throughput screening.

(2) The thread structure could promote the material flow but increasethe fracture risk. The CV value revealed that pins with milling facetswere more beneficial to achieve sound joint.

(3) Optimized geometry of tapered thread pin with triple facets at awelding speed of 30 mm/min and a rotational speed of 800 rpmwas obtained. Sound joint without obvious thickness reduction andflashed was attained based on the optimal geometrical structure.

(4) High depth-to-width ratio FSW has potential to reduce HAZ andimprove joint quality. The average width of HAZ was 440 μm,which was attributed to the lower heat input generated by highdepth-to-width structure.

Acknowledgements

The work was jointly supported by the National Natural ScienceFoundation of China (No. 51575132) and the Fund of NationalEngineering and Research Center for Commercial AircraftManufacturing (No. COMAC-SFGS-2016-33214).

References

Arora, A., De, A., Debroy, T., 2011. Toward optimum friction stir welding tool shoulderdiameter. Scr. Mater. 64, 9–12.

Cabibbo, M., McQueen, H.J., Evangelista, E., Spigarelli, S., Paola, M.D., Falchero, A.,2007. Microstructure and mechanical property studies of AA6056 friction stir weldedplate. Mater. Sci. Eng. A 460–461, 86–94.

Dai, Q.L., Wang, X.Y., Hou, Z.G., Wu, J.J., Shi, Q.Y., 2015. Microcavities accompanying azigzag line in friction stir welded A6082-T6 alloy joint. Sci. Technol. Weld. Join. 20,68–74.

Feulvarch, E., Roux, J.C., Bergheau, J.M., 2013. A simple and robust moving meshtechnique for the finite element simulation of Friction Stir Welding. J. Comput. Appl.Math. 246, 269–277.

Mishra, R.S., Ma, Z.Y., 2005. Friction stir welding and processing. Mater. Sci. Eng. R. Rep.50, 1–78.

Nandan, R., Roy, G.G., Lienert, T.J., Debroy, T., 2007. Three-dimensional heat and ma-terial flow during friction stir welding of mild steel. Acta Mater. 55, 883–895.

Sheppard, T., Jackson, A., 1997. Constitutive equations for use in prediction of flow stressduring extrusion of aluminium alloys. Mater. Sci. Technol. 13, 203–209.

Shi, L., Wu, C.S., Liu, H.J., 2015. The effect of the welding parameters and tool size on thethermal process and tool torque in reverse dual-rotation friction stir welding. Int. J.

Fig. 14. Microstructure of transverse section.

Fig. 15. Temperature distribution on transverse section of high depth-to-width ratio FSWjoint.

Fig. 16. Microhardness profiles of the high depth-to-width ratio FSW joint.


240

http://refhub.elsevier.com/S0924-0136(17)30428-4/sbref0005



















hyxhi

高亮

hyxhi

高亮

Mach. Tools Manuf. 91, 1–11.Su, H., Wu, C.S., Bachmann, M., Rethmeier, M., 2015. Numerical modeling for the effect

of pin profiles on thermal and material flow characteristics in friction stir welding.Mater. Des. 77, 114–125.

Tongne, A., Desrayaud, C., Jahazi, M., Feulvarch, E., 2017. On material flow in frictionstir welded Al alloys. J. Mater. Process. Technol. 239, 284–296.

Tutunchilar, S., Haghpanahi, M., Givi, M.K.B., Asadi, P., Bahemmat, P., 2012. Simulationof material flow in friction stir processing of a cast Al-Si alloy. Mater. Des. 40,415–426.

Zhang, H.J., Liu, H.J., Yu, L., 2011a. Microstructure and mechanical properties as a

function of rotation speed in underwater friction stir welded aluminum alloy joints.Mater. Des. 32, 4402–4407.

Zhang, X.X., Xiao, B.L., Ma, Z.Y., 2011b. A transient thermal model for friction stir weld:part I: the model. Metall. Mater. Trans. A 42, 3218–3228.

Zhu, Y., Chen, G., Chen, Q., Zhang, G., Shi, Q., 2016. Simulation of material plastic flowdriven by non-uniform friction force during friction stir welding and related defectprediction. Mater. Des. 108, 400–410.

Zienkiewicz, O.C., Cormeau, I.C., 1974. Visco-plasticity-plasticity and creep in elasticsolids-a unified numerical solution approach. Int. J. Numer. Methods Eng. 8,821–845.


241





















Documents

Journal of Materials Processing Tech