Escolar Documentos
Profissional Documentos
Cultura Documentos
ABSTRACT: Numerical investigations are carried out to study the effect of weld parameters on the temperaturetime history in a butt-joint weld steel plate. A mathematical model of transient thermal process in welding is
established to simulate the transient thermal analysis with moving heat source model (Gaussian function) by using
finite element method. Results are compared to numerical and experimental results obtained from a previous
study. Parametric studies based on numerical results are carried out for different weld parameters including
welding speed, plate thickness, heat input, heat source type and finite element mesh.
1
1.1
INTRODUCTION
Introduction to welding
285
286
3.1 Considerations
Generally speaking, severely physical and chemical
reactions occur during the welding process, between
the base metal and the melted pool, which includes
the thermodynamics and heat transfer, the interaction
between the heat source and metal, solidification in the
welding line, the phase transformations in the welding
joint, and so on. To focus on the temperature field of
the welding structure, some factors that have a small
effect on it are supposed to deserve low weight in
consideration or sometimes even be neglected. Hence,
several simplifications have been considered during
the simulation.
First of all, the chemical reactions and agitation
and convection phenomena, and the phase change are
not taken into consideration.
Secondly, the heat transfer between the base metal
and the experiment table are neglected, assuming that
there is only convection heat transfer between the
edges of the work piece and the air, without radiation
heat transfer.
Thirdly, assume that the welding rod is of the same
material as the substrate.
Finally, assuming that the welding process is conducted with the constant heating speed V; the heat
source is complied with the Gaussian model.
3.2 Material properties
The material properties of metal, for instance, the specific heat, thermal conductivity, modulus of elasticity
287
0
51.9
450
0.28
10
200
1.0
100
51.1
499
0.31
11
200
6.5
300
450
46.1
566
631
0.33
12
0.34
13
200
7.5
and yield stress, vary with the change of temperature. The average values can be used in calculations
if the temperature does not vary too much. While
in the welding process, the temperature of weldment
varies shapely. In this case, to neglect the material
properties differences by temperature will result in big
deviations. Consequently, the temperature-dependent
material properties are supposed to be taken into
consideration in welding process simulation.
In this analysis, eight-node three-dimensional brick
thermal element, Solid 70, is used. The Solid 70
element has a three-dimensional thermal conduction
capability. It has eight nodes with a single degree of
freedom, temperature, at each node and is applicable to
a three-dimensional, steady-state or transient thermal
analysis. The element also can compensate for mass
transport heat flow from a constant velocity field. If
the model containing the conducting solid element is
also to be analyzed structurally, the element should be
replaced by an equivalent structural element (such as
Solid 45 or Solid 185).
The C-Mn steel whose temperature-dependent
material properties are listed in Table 1 was used.
3.3
41.1
150
7.3
550
37.5
706
0.36
14
110
7.2
600
35.6
773
720
30.6
1080
800
26.0
931
1450
29.5
438
0.37
14
0.37
14
0.42
14
0.47
15
88
7.2
20
7.1
20
7.1
2
7.0
FE modeling
In this work, the butt welding process has been simulated in Ansys. The dimensions of the work piece
are 300*260*6 mm3 , as shown in Figure 4 in which the
arc center is moving along the x+ direction in line KI.
Furthermore, for the sake of saving computational
time and reducing computer configurations requirement, a half model has been taken in the simulation,
with symmetric boundary condition setting in the
symmetry plane, since the model is symmetric along
the X-Z plane.
Figure 5 displays the element mesh of the geometric
model to be solved. 2 mm element size is used in the
welding zone, while 8 mm element size is used in the
area far away from the heat affected zone (FAZ). Free
mesh is performed in the middle zone.
Meanwhile, a finer mesh was carried on, to study
the effect of the mesh style on the result. In this case,
four different element sizes, 1.25 mm, 2.5 mm, 5 mm,
and 10 mm are modeled. And all the elements are
hexahedrons (see Figure 6).
FE results demonstrates that the temperature distribution of the node located in the center of the welding
line are almost identical since the maximum deviation
in all time steps are only 0.21%, which means the relatively coarser model is satisfied to be utilized in the
simple butt welding process.
288
3.4 Verification
As presented by several researchers, finite element
method (FEM) has been successfully used to evaluate thermal response of the complex welding process.
In the present work, the finite element (FE) package
ANSYShas been used to simulate the arc welding
process.
Quite a few researchers have addressed this topic,
Tsai & Eagar (1983), Guedes Soares et al. (1998) and
Biswas et al. (2007), to name a few.
In Figure 7, a finite element simulation of a line
heating process was verified using the work of Biswas
et al. (2007). To do this, the same geometry and
material properties as Biswass model were using,
the heating process with total heat input 5350 Watts
and heating speed 6 mm/sec was simulated using
ANSYScodes.
From Figure 4, it shows that the temperature history
result of present work is in a good agreement with the
experimental result obtained by Biswas et al. (2007).
3.5
289
Firstly 7 nodes in the welding line with uniform distance are chosen (see Figure 17), and the temperatures
of these 7 points are plotted in Figure 18, which indicates that, after reaching a stable state, the curve of the
temperature of each node is of the same tendency, and
for a certain node, the temperature shapely increases
when it is subject to heat input, and then fall down
with a relatively lower speed after achieving the peak
temperature. The first node and the last one are with
lower/higher temperature than the other nodes, which
is due to the less/more heat affected time.
Another 7 points are selected in Figure 19, whose
temperature distributions are similar to the previous
ones, expect the peak temperatures are much lower
since they are a little far from the heat source, and the
decrements of temperatures are slower (see Figure 20).
In the Z direction, 3 more nodes are taken as
Figure 21. Figure 22 shows their temperatures distributions. The temperature reduces with the increment
of distance from the heating surface but the effect is
not so significant.
In the Y direction, 7 nodes are marked in Figure 23,
with the distances away from the center line of 2, 4,
6, 8, 14 and 26 mm respectively. The more distances
away from the heating line, the lower temperatures
produced (see Figure 24). Figure 25 illustrates the
relationship between the peak temperatures and the
distances from the welding line.
290
3.6
that the lower the speed is, the higher temperature got
in the result.
The heat input in the previous calculation is 5350 W.
Keeping the fixed welding speed 6 mm/s, and changing the heat input to 3000 W, 6500 W and 7950 W,
then the corresponding temperature distributions are
plotted in Figure 27. There is no doubt that the higher
heat input results in higher temperature.
To study the effect of the plate thickness on the temperature field result, 5 cases are conducted in which
the welding speeds are fixed to 6 mm/s, the heat inputs
291
CONCLUSIONS
ACKNOWLEDGEMENTS
This work has been performed in a EU funded
project Network of Excellence in Marine Structures (MARSTRUCT) under contract TNE3-CT2003-506141.
REFERENCES
292