‘The Annals of “Dunarea de Jos" University of Gala
FascicleXIL, Welding Equipment and Technology, Year XVI, 2006
: SSN 1221 4639
EXPERIMENTAL AND COMPUTATIONAL DEVELOPMENTS OF
Fsw
Pedro Vilaga, Luisa Quintino
Instituto Superior Técnico, Technical Univesity of Lisbon, Portugal
edo.vilaca@ is. gt
ABSTRACT
‘For more than one century welding technology plays an important role in mast
ofthe technical development of our society. Friction Sir Welding (FSW) invented
fn 1991 ts unanimously considered one milestone inthe development ofthe welding
lechnology and it i foreseen it will bring important contributions in the
(development of present and future design and consraction of metallic structures.
‘Ths fact has wlready ‘being confirmed by the growing number of industrial
‘pplicatons in many and diferent sectors mainly in the manufacturing of light
(lloys The Group for Welding and Joining Technologies at the Insituo Superior
‘Téenico (SD has sated its activity approximately three decades ago and in 1999
the frst PRD thesis on the fndamentals and modeling of FSW was presented and
‘Since then anew area of RED in solid state welding is being developed.
‘The work performed at IST on FSW started based on literature survey,
participation in dedicated conferences and seminars sideby-side with basic trials
Cabling the development of know-how about the experimental and physics
fundamentals of the process. Later on with the development ofthe computational
‘modeling supported by the participation in cooperate research projects allowed
efitely the consolidation ofthe knowhow about the material flow, thermal flow
‘and metallurgical evolutions
“Although the need for further develop the modeling of the process, both
anclytially and rumericaly, nowadays the efforts are also concentrated in
‘Supporting the Portuguese companies intrested in acquire the FSW technology and
“developing non destrctveteckigues fo access the quality of the joints suramary
(some of the work performed at IST on FSW is presenied with more emphasis 0%
the modelling of the process but also revealing some echnological features.
KEYWORDS: FSW, experimental fundamens, computational modeling.
1, INTRODUCTION
Friction Stir Welding (FSW) is a solid state
welding process, which proceeds below the melting
point ofthe weld material. A rotting non-consumable
fool, with a somewhat complex shoulder and pin
profile, is plunged into the weld joint and forced t0
fraveree along the joint line, hesting the sbuting
components by interfacial and intemal friction, thus
producing @ weld joint by extring, forging and
String the materials from the two workpieces in the
‘iinty of the too 1, 2 31
"The principles behind the technology have
been patented [1], and the development of FSW has
been taken up primarily by large companies including
aerospace, aircraft, naval and automotive sectors [2,
53]. Moreover, research into FSW should not only be
‘eriented to provide solutions for major industial
‘ppliations but also to achieve encugh application
flexibly for is wanston ino small and medium
enterprises.
‘The Group for Welding and Joining
‘Technologies a the Tnsitto Superioe Téenico (ST)
has started is activity approximately three decades
‘ago with the aoqiskion of a power supply for pulsed
‘TAG welding and MSc thesis on TIG welding of
issimilar ‘materials, Since then the group has
cnlarged its activity and has developed a sirategy
based on complementarity of R&D and training
setvtes. In this context the results obtsind in R&D
projects are ured to update training materials and
iceversa topics given for students project lead often
to new ideas for RDT projects. in 1999 a PRD thesis
[2] on the FSW fundamental and modeling was‘The Annals of "Dunaren de Jos" University of Galati
Frasciele XI, Welding Equipment and Technology, Year XVI, 2006
ISSN 1221 ~ 4638
launched and since then a now area of R&D in solid
sate welding is being developed. The experimental
Timitations “Altiough the inexstence of a FSW
machine at IST. many differen issues tave been
ttresed revealing he potential of the process when
Compared with the fusion welding processes in the
‘welding of aluminium alloys [4
FSW. encompasses complex phanamens
elated with plastic flow deformation resting from
the. sirring of the workpicees material. The
‘development of process models conributes towards 8
‘eter understanding of joint formation and resultant
tachanicl properties [5]. Thus the development at
IST of the compustinal modeling of FSW is one
key contrition of IST tothe development of FSW.
“The presentation of some of the experimental
tundaments of the FSW developed atthe IST stat
wit a description of the concepts for the tools
Geveloped and follows addessing the diferent
ptems of the plastic flow of the material in FSW.
"The typical hardness feds for aluminium alloys ae
presented pri 1 the mechanical and metallurgical
techniques for evaluation of the FSW joins. Finally
fomasion studies on AASOS3-HIN1 and an
fsuestment of the tallorblanks consiruction wit
FSW, GMAW and GTAW are described. Before the
conclusions the implementation of both thermal
analytical model and coupled fluid. dynamics and
‘Solid mechanics numerical approaches are reviewed.
2. FSW EXPERIMENTS
‘At IST the fundamental conditions for
perfoming experimental iss with FSW were
developed. First the dvelopment of the wools, wih
bath mono-bulk ols and’ modular tools as it
represented in Figare 1. The modular iSTIRtool_¥1
flow the combiraion between different pins and
shoulder geometries ar the continuously variation of
the length of pin coming out of shoulder. Also
‘clamping devices were develope.
frre es
Fig, Modular FSW tool: iSTIRtol_v1 [2,6]
ae
ee lto, [tz me
row (TS Lv Taz a
Classica | ibo_ | SHAZ
Poms ity ittmaz
ns,
Fig2. Classification of he typical material flow patterns (2
Tre mechanical and mtallrgical
ctaraeristics of all tbe main families of wrought
Slminium alloys Rave been investigated from 1.0 10
‘Shout 10mm in differen joint geometries
‘One important esl ffom the sensitivity
analysis of the influence of FSW process parameters
{nthe properties of the jot was the exsbishment of
the main patterns of material low around the tool
(Gig 2) Inthehor conditions the vseoplasic material
‘ow is mere concentrated around the pia andthe heat
affected zone is wider. In opposite under cold
Condit. the themmo-mechanieally est affected
‘ane is wider and the heat affected zones smaller.
Tn figures 3 and 4 the typical hardness fields
obtained forthe two man differen greaps of wrought
‘iulaium alloys are represented. In any of the cases
itis possible 1 conclude about the ear matching of
the hardness eld ofthe welded zone compared with
the base material‘The Annals of “Dunarea de Jos" Univesity of Galati
Fascile XU, Welding |
From the figure 3 showing the behavior of
‘the non heat treatable alloy itis possible to conclude
about the increased hardness of weld bead and heat
affected zones when compared with base material
Because these alloys are very sensitive to sain
hardening the increase is most significant in
‘dynamically recovered zone and thermo-mechancally
‘heat affected zone
Tn figure 4 it is posible to conclude about
the typical location of the global minimum value of
yayguEaeese
Equipment and Technology, Year XVI, 2006,
ISSN 1221 ~ 4659 e
the harmes field located in the itrfae batwoon the
heat affected zone and the thormo-mochanially heat
atfecied zone. Along the heat affected zone there is
typically a local minimum due to over-ageing
Depending on the pattern of material flow around the
tool the FSW onder cold conditions the minimum
value at the flo side (retreating side i smaller than
in the shear side (advancing side. Foe FSW vader hot
‘conditions the minimum values in both sides of the
‘weld bead have moce identical valves.
Fig.3. Typical hardness profile forthe non heat treatable aluminium alloys,
(AASO83-II111; thickness: dn) [6]
FFig.A Typical hardness profile forthe heat treatable aluminium alloys
(AAGOGI-T4; thickness: 4.8mm) [2]
‘The ardoess profile enables a rlible
assessment of the static mechanical resistance of the
joims bu the fatigue resistance is moce dependent on
the geometric characteristics of the weld bead at the
top and botiom surfaces and eventual interal defects
both volumetric or layers of socand phase particles or
oxides, eg. root defects (2 7)
“Metallargcal features were also investigated
I, such as, the diffusion of the initial
precipitate, and changes of grain size in the heat
affected zone and thermal-‘mechanically affected
‘one, resulting from the thermal-mechancal eycle of
the process [2, 7,8]. The results allow understanding
the mechanisms that determine the metallurgical
changes. in the weld bead, The metallographic
analysis is compared with hardness ests developed in
19
all zones of the welded joim, enabling conclusions
bout the mechanical resistance efficiency.
Resuks of motllurgy analysis canbe
imerpreted analysing. the grain morphology and
precipitates density and location, eg. in Figure 5 the
‘material analysed was the non-heat treatable
aluminium alloy AASOB3-HI11, where is possible to
emphasise the small coalescence of griin and the
‘eduction of precipitate in the grin boundaries of he
HAZ and TMAZ which it i know 10 play an
important role in the prevention of corosion
evelopment and also the that. TMAZ/Nogget
interface enhances the significant difference between
the stractre of iil gran andthe grain resultant of
the dynamical recrystalisation process of the nugget“The Annals of “Dunares de Jos” University of Galati
Fascicle XIL Welding Equipment and Technology, Year XVII, 2006
ISSN 1221 ~ 4639
The transfer of the FSW ino industrial
pplications, such as shipbuilding, demands for =
detailed investigation aout the infience of eomosion
fn the performance of both parent materials and
‘welded joins. In cooperation with the Mondego
Shipyard, Figueira es For, Ponugal, the IST
fevalusied for the AASOS3HIIT the main
mechanisms of corrosion, sich as: inergrnulit
‘corrosion and exfliaton. In figure 6 iis posibe to
‘Sbicrve some resuitant coroson ia FSW specimens
‘of AASO8S-HIT] after 7 deys of exposition 1 rel
conditions.
rom the fnal results it should be
emphasised thatthe loss mass in the base material
samples i much higher than inthe ficton si welded
amples because icorgranular eorasion mechanism is
fnost suscepible atthe intergranular precipitates and
these are more abundant in bate materia 6,7].
Tn order to access the prtntal of FSW t0
improve the consrucion of talorblanks the IST
esublished a comparison of relevant welding fares,
eg. surface finishing, residual deformation
retallorgical analysis and suength efficiency ofthe
joins performed by FSW, GMAW and GTAW
between plate with diferent thickness of AAIOSO;
AAZO2A-TS snd AASORYHIL. A sample of the
tess forthe AA224-T3, ie presented in Fig.7
Fig.6, Corrosion developed in bse material of AASQSS-H111 afer 7 days Exposition in Tagus River (20g/
NaC) (7+) Intergranular corosion slong the BAI-Mg precipitates) Pitng formation mechanism under the
ship of the weld bead‘The Annals of “Dunarea de Jos” University of Gala
FascicleXIL, Welding Equipment and Technology, Year XVIL, 2006
ISSN 1221 ~ 4639, :
Welding Process
cma
Welding Process
‘maw
GTAW,
(ceatiteetieeati]
a
Fig.7. Assessment ofthe macrographic chraceristies and NDT x-ay test results ofthe A2024T3 dissimilar in
thickness (38mm/4mm) welded by FSW, GMAW and GTAW [4]
3. COMPUTATIONAL MODELING
‘The development of computational models
‘can greatly contribute 10 better understanding any
Industrial process, particularly FSW. A validated
‘model has the "potential to produce reliable
Tnonssaon abu he ueformaon end ising
patiems that are imporant when designing. FSW
tools and thus shouldbe capable of producing welds
{hee of defects and voids.
Further, a model can measure process
characteristics that are difficult to. observe
experimentally such a local stains, strains ates and
stresses [9, 10), These strain and stress fields,
fogether with temperature histories are seen as
critical in predicting mierestracture evolution.
A detailed understanding of microstructural
evolution can guide FSW designs by further
Improving, mechanical propane, age strength
‘While considerable experimental work has
been done! 10 improve the knowledge on FSW,
there's yet lot of work needed to create a satisfying
slobal model that can produce consistent results
‘The main difficulties in modeling FSW are
R53)
‘+ Extensive material deformation inthe
ropion containing ful-pasticlzed material
‘+The viscous-plasic flow imposed by the
tool rigid) surface, into the materials
constrained by the interaction wih the cold
base material, with an essentially elastic
a
behaviour, and the rigid plate (ani)
supporting the joint;
‘+ Hout generated de tothe sliding between
the surface ofthe fool and the materials in
the joint depends on an unknown. the
fiition coefice
+ Tieconedt pein uf Ue vscousplastc
flow imposed by the tool rigid surface into
the muterals being welded is also
important because the viscous dissipation
contributes signiticamly to. the heat
‘evelopment during the performance ofthe
weld bead
+ The materials thermomechanical
propenies vary throughout the FSW
process;
‘+The tharmal flow ito the tool and support
plate, needs tobe considered inthe models;
+ FSW process modelling. doss nt allow
‘geometric simplifation Because ic doals
‘with 2 complex 3D material flow around
the pin:
+ Thehighty rating tool pin has typically, 2
complex geomarie profile (eg. threaded),
‘hich is tutor dificult 10 consider for
‘most ofthe numerical methods avalble
‘The challenge is thon fo create a mode able
to fully describe the complex. FSW process as
‘usrated in figure &“The Annals of "Dunarea de os" University of Gala
Fascile Xl, Welding Equipment and Technology, Year XVI, 2006
ISSN 1221 4659,
Materials plaste tow
‘and deformation
peal 2
Heat dissipation by
Inter fetion tscous)
Defective joins
seer ant ro
I
|
Heal Generation
Wtatargie structure
changes (éynaiic
rsrystlization ofthe
et
Fig, Coupled mechanicalhermal/motallurgcal character of FSW process 2, 9,10]
4, ANALYTICAL MODELING
ISTIR is the acronym of a thermal
analytical model for 2D and 3D cases when FSW
Similar and dissimilar materials. The development of
the ASTIR code came about, due to the perceived
fed 10. establish relationships between | FSW
parameters and easy 10 measure geometric features
fndlor mechanical properties via an assessment of.
the Heat Int developed during the steady-state
FSW regime Figure 9).
‘The thermal analytical model, (STIR, is
based on evolutions from thermal flow equations
sstablised by D. Rosenthal [11], fr point heat
Soures with uniform velocity. Tis model also takes
in account the spocificies of the heat generated by
viscous energy dissipation during. plastic flow
(eformation and. interfacial fiction, eg) the
‘symmetic teat goneration mostly atthe periphery
ofthe shoulder of the FSW tool as rest of the
‘imposition of both linear and rational velocities
fd finally the sfferinoes botween hot and cold
SW conditions (2)
2
Due to percived cifficultis in modeling
the FSW process an fterative mehod was employed
‘where the heat rower source is incremented so that
the final value wll minimize the difference betwen
the STIR thermal fled rosulis and the ones
resulting fom the experimental temperatre
“Therefore, ‘STIR code Is imended 10 be
od ina “reverse engincering approach”. based
‘ona thermal field previosly measured, the ress
(ofthe thermal field fom ASTI may be iteratively
Superimposed on results osined from the
experimental thermal meisuremens (ander identical
thermo physical conditions) and the value of the
point power soures, which produces sich thermal
Fields then determined,
"The FSW elficieney coefficient, is then
systemaicaly related wih the FSW parameters and
Characteristics of weld macographs “made in
Sections transverse to the welding direction, eg. the
Thardnes field an the rio Betwocn tho extension of
the thermo-mechanically heat affected zone andthe
fextension ofthe hea affected zone,‘The Annas of “Dunaea de Js" University of Galati
Fascile XI, Welding Equipment and Technology, Year XVI, 2006
ISSN 1221 ~ 4659
i >
Fecerseoctocescchan
Fig 9. — Basic principles ofthe development of STIR [2, 12,13]
iSTIR Code
Inverse Engineering Approach
mon }-(-fes
Fig. 10. Concept of the “revere engineering approach”
s. NUMERICAL MODELING
‘The aumerial modelling, approach of the
[FSW access the material low in the vieiity ofthe
tool and the residual stresses and deformation field
resulting fom the process in the real geometry ofthe
parts being welded,
Because there is no commercial software
able to perform a complete analysis ofthe process,
the strategy adopted was 10 use a coupled process
based on a specially developed code, InegrasD,
‘whieh bridges two commercial software's, namely a
‘uid dynamics (Fluent), used for the viscoplastic
‘ogime steady-state analysis of the material flow in
the vicinity of the tool and structural mechanics
(Abaqus), used for compute in a transient analysis
the residial deformation and sess fields based on
the stress and thermal history resulting from the
FLUENT analysis.
‘used inthe STIR code [13]
‘Tis inegration i presented inthe figure
3. The invegration of the to approaches is made
through a FORTRAN code routine named
Integra3D. The concept of integration of the two
‘approaches used in this work is based onthe fct
thatthe FSW process has two regions with different
behaviours. Aft 20ne, near the tol, whore the
‘material as viscous plait behaviour, and a second
zone; that includes the rest ofthe domain, which has
‘rodominantly elastic plastic bchaviou
"The material behaviour used by the fluid
Aynamies analysis fellows the Zener-Hollomon (14,
15] material model which is based on 2 viscosity
parameter, depending of the temperature and sain
rate, The material behaviour wsed by the structural
analysis i the pial elastic-plastic. behaviour
depending onthe temperature“The Annals of “Dunares de Jos” University of Gla
{2 nga Fenian do proces attr por eo
Feu ies eprom wn. THe
[St UnveraeTecnn cs os Senter 203
{3} Lata Quin, Pare Wing, Slt por Pro
Ener o\Qar ite Pogue Praia « Pre
‘Ser Teagan «Guth Revs do sao. de
Stra utd pe, 35, Aas 2.
ete amy ie Gt,
wae tmcnicet” ise Weblog OW). psec
Fay eas 08
(el Pope Ne Vitge P Quin Meni nd
Coron Pons 9 rain Sar Wing f A430
Th Doc RBA hga Ane tthe en
oie of Wel i Cmmison IK Papua 1-8,
{i Pepe: , Corocrttcar Meigles,Conportnen 3
Fresca 3 Create de tin Sia por Fe
Liner Mie Tr 1t-Urierasace Tee de Libs,
‘Seater 205"
WP Dertl Te Ving Pe uation L,Coracerinee
Meta do Pro de Sante por Pe Liner
[Bltrege it Macgie NndrceeYlanas de Poca de
Suda por Pg Leo, MSc These: 1ST Une
‘au de sou Spent
[to Frags Es Vloge Qoinin Aap ou Numeral
oweng ay Pat Bow hots Fw Visas
Sepa GR3E tne Cen 3" hwy
{in| Rota Te hor of Moving Sere of Het nd
Inigo strane, Troma of te ASME,
Fags Pe Quine 1, Dot Santee STIR — Asoc
‘Bera al or Pon sr eng asa of Matas
Frecuing Tcl. taper pliant 15 Desember
bos
1 vaage Fy Quinto Ly Shell. dos. Saas 3
Deion of te Terma Eilny iv FSW vi an
Noting Frmatson Set Aral” Menaly ef e
tnt a of Weng. Commision Rl Bucaes,
yas 208,
[12 zener tn Haman 431 Et of Si Rate Upon
Pini Se tne pid Ph spp
isa.
ita ahepard Tan Wright BS, Detrmiaton of fom
fase Par i souane sston froma al ot
[ocd nercrer Maal Teanibgy, pp. 31523, ne
im