Encyclopedie de la recherche sur l'aluminium au Quebec - Edition 2014 | Page 64
62
TRANSFORMATION ET APPLICATIONS // TRANSFORMATION AND APPLICATIONS
Numerical Simulation of Friction Stir Welding
with Smoothed Particle Hydrodyamics
SIMULATION NUMÉRIQUE DU SOUDAGE PAR FRICTION MALAXAGE
AVEC LA MÉTHODE « SMOOTHED PARTICLE HYDRODYNAMICS »
Simulation numérique du soudage par friction
malaxage avec la méthode ‘smoothed particle
NUMERICAL SIMULATION OF FRICTION STIR WELDING WITH
hydrodynamics’
SMOOTHED PARTICLE HYDRODYNAMICS
Kirk Fraser, Lyne St-Georges and László I. Kiss
Université de Québec à Chicoutimi (UQAC)
Joining of aluminum alloys via friction stir welding (FSW) is gaining widespread acceptance. The process’s ability to
create high strength welds in a solid state with few imperfections makes the method very appealing for industrial and
commercial applications. More recently, a double sided FSW process was developed using a bobbin tool. One of the
main advantages of this approach is that hollow core sections can be welded. Also, the vertical force on the machine is
zero.
The results from the simulation are shown below. A comparison of the shape of the finished weld for the simulation and
from experiment is provided. We can see a close agreement between the real and simulated welds. Fraser et al. [6,7]
showed that the welding defects can be simulated for different process parameters. The required torque to sustain
800rpm was measured and compared for the simulation and experiment, we show that an excellent correlation is
obtained between the two.
A coupled thermo-mechanical model has been created using the
smoothed particle hydrodynamics approach in LS-DYNA. The
model simulates advancing phase of the double sided process.
Once fully validated, the model will be used to perform
parametric studies to determine the optimum welding
parameters.
Tail Length (mm)
Simulation 1 - 600 RPM
800 mm/min
50.0
Experiment 3 - 600 RPM
800 mm/min
45.0
Blowout Hole (mm)
20.0
18.0
25.0
15.0
Blowout Protrusion
(Advancing side, mm)
Blowout Protrusion
(Retreating side, mm)
Rate of rotation, forward advancement speed as well as design
of the shape of the FSW tool can be optimized. The model can
qualitatively and quantitatively determine the quality and strength
of the weld.
Max Torque (N m)
Distance to Form Weld
(mm)
-8.0
6.0
20.0
18.9
35.0
20.0
Image from [5]
Numerical simulation of the FSW process is complicated by the large plastic deformation that the aluminum work pieces
undergo. Typical finite element methods (mesh based) are not able to accurately capture this behavior due to the
excessive distortion of elements in the large deformation zone. The SPH method is a mesh-free method, the elements
are not confined to a fixed grid, this allows for accurate simulation of the large deformation process.
One of the major attractions to the SPH method is that the formulation is Lagrangian, thus tracking of the time history of
all the field variables is easily accomplished. The rate of change of the field variables for a given particle “i”, with N “j”
neighbors in the support domain is given by Lui and Lui [4];
There are three major flaws that are visible in the
bobbin tool weld track: 1) start of weld tail, 2) lead in
distance to form the weld and 3) blowout region at tool
exit. These defects are nicely captured by the SPH
simulation
The temperature results are shown in the adjacent
image. The hottest region is found directly under the
tool, the shape of the temperature profile is strongly
influenced by the flow of the aluminum material and
vice-versa.
Wi,j = Smoothing Function
πi,j = Artificial Viscosity
Hi = Artificial Heat
The aluminum work pieces are modeled using SPH elements. The supporting apparatus and the FSW tool are modeled
using rigid bodies meshed with hexahedral finite elements. Here we consider a short section of a hollow core HSS tube.
The aluminum material is modeled using the
Johnson-Cook constitutive material law. The flow
stress is given by [3];
Kirk Fraser
Lyne St-Georges
László I. Kiss
Université du Québec
à Chicoutimi
• The developed numerical model can be used as a powerful process diagnostic tool for friction stir welding. The model
• Welding defects such cavities, lack of penetration, and incomplete welding can be predicted.
An iterative viscoplastic formulation is used such
that;
financial support and GRIPS for their technical support.
will help to determine the optimal process parameters to create high quality welds for the double sided FSW process.
• Residual stresses and strains can be determined.
• The ultimate strength, hardness and resistance of the weld can be evaluated.
• The author would like to acknowledge l’Université de Québec à Chicoutimi (UQAC), NSERC, and CQRDA for their
Tool motion
The relation between pressure and volume is
modeled by the Gruneisen equation of state;
• We are currently developing a coupled computational solid mechanics and heat transfer code using the SPH method.
The code is written to take advantage of the inherent parallel nature of the graphics processing unit (GPU).
A spall failure model is used to effectively simulate the material
after local failure is determined. The spall model sets the
deviatoric stresses to zero when the pressure in the element
surpasses the pressure cutoff. Subsequently only compressive
stresses are allowed. Following failure, the material behaves
as rubble.
Heat is generated by the friction between the tool
and the aluminum as well as the transformation of
plastic work to heat.
Plate
Advance Angular
Thickness
Rate Velocity
1/2
600
800
in
mm/min rpm
Value
Units
Material Constants
Johnson-Cook Parameters [1]
Thermal Parameters
Gruneisen EOS Parameters [2]
ρ0
G
A
B
C
n
m
Cv
K
C0
S1
γ0
a
6061 T6
2700
27.6
324