Você está na página 1de 10

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257

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

Development of a local three-dimensional numerical


simulation model for the laser forming process of
aluminium components

G.N. Labeas ∗
Laboratory of Technology and Strength of Materials, Department of Mechanical Engineering and Aeronautics,
University of Patras, Rion 26500, Greece

a r t i c l e i n f o a b s t r a c t

Article history: The laser beam forming (LBF) process uses the energy of relatively high-powered lasers
Received 13 July 2007 to cause permanent deformations of components, through the local introduction of ther-
Received in revised form mal stresses. LBF of aluminium material is a process, complex and sensitive, due to the
14 November 2007 complicated physical phenomena taking place during laser processing. Therefore, defini-
Accepted 20 December 2007 tion of optimal process parameters, such as laser power and processing velocity, which
will result to desired bending patterns, as well as investigation of forming limits of various
components require significant experimental effort. Herein, numerical simulation of LBF
Keywords: process is used to provide partial solution to the problem, by developing a local Finite Ele-
Laser beam forming ment simulation model, capable to predict temperature fields and deformation shapes of
Bending laser beam-treated aluminium specimens. The numerical algorithm is based on a non-linear
Aluminium three-dimensional transient thermal–structural analysis, temperature-dependent thermal
Finite Element and mechanical material properties and a laser beam heat flux model. The developed model
Thermal–structural analysis is validated through the comparison of numerically predicted distribution of temperatures
Local three-dimensional model and bending angles to corresponding experimental data of single and multiple laser beam
passes. The validated model is then used to define optimal process parameters for the laser
forming process of aluminium panels.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction for many technological advantages, especially in cases of


forming complex or semi-assembled structures of various
Laser forming recently emerged as a new shaping tech- thicknesses and material types, as well as in rapid prototyping
nique that offers excellent reproducibility, low manufacturing applications.
time and cost, as well as, relatively low thermal influence Although the traditional application of laser forming has
on the material mechanical properties. For those reasons, been to various steel materials, laser forming process of cer-
laser forming is a promising technique with several poten- tain types of aluminium alloys has been recently the object of
tial applications in the automobile, shipbuilding and in considerable attention in the aerospace industry. However, the
particular aerospace industry, where the demand to form low thickness of formed parts, the elevated laser beam light
integrally stiffened structures is high. In comparison to con- reflection and the high heat conduction of aluminium alloys
ventional forming technologies, LBF provides the potential are some of the main reasons, which make the use of laser


Tel.: +30 2610991027; fax: +30 2610997190.
E-mail address: labeas@mech.upatras.gr.
0924-0136/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmatprotec.2007.12.098
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257 249

forming process more difficult and complex, as compared to al., 2001; Hu et al., 2001; Geiger et al., 2004; Welsink, 2005).
its application to steel forming. In Chan and Liang (2000), Merklein et al. (2001) and Welsink
During laser forming, the irradiated material is formed (2005), experimental investigations concerning bending angle
under the action of local plastic strains induced by laser as a function of the laser parameters, as well as, effects on
heating of the material, instead of the action of mechanical material microstructure are presented. In Hu et al. (2001)
forces and moments applied by the common sheet bending experimental investigation and limited Finite Element sim-
techniques. The local nature of laser irradiation yields high ulations concerning temperature distribution and bending
temperature gradients between the irradiated surface and angle during laser beam forming process of steels and a non-
the neighboring material. The high temperature gradients, specified aluminium alloy were presented. In Geiger et al.
force the material to expand non-uniformly, which results (2004) only thermal FE simulations are performed for investi-
in irregular thermal expansion between the target and lower gation of the temperature distribution through the thickness
surface. As a result the specimen initially bends negatively, of aluminium alloy sheets for automotive applications.
as viewed from the laser beam. The non-uniform expan- In the present investigation a local three-dimensional
sion of the material leads to non-uniform thermal stresses, Finite Element model for simulation of the laser forming
which result to plastic deformation at locations where thermal process of aluminium parts is developed. Non-linear ther-
stresses exceed the material’s yield point. During cooling, the mal and mechanical analyses are consequently performed,
upper material layers shrink more than the bottom, resulting using temperature-dependent thermal and mechanical mate-
in permanent specimen bending towards the laser beam. rial properties. The laser beam is modelled as a step-wise
From the above short description, it is obvious that the moving heat source with Gaussian distribution of heat flux,
laser forming process comprises many simultaneous physical as proposed in Hu et al. (2002). The simulation model pre-
mechanisms and is affected by several process and mate- dictions for different laser beam forming conditions (power
rial parameters. The most important parameters are the laser and speed), material types, specimen thickness and number
scanning path, the heating conditions including laser power of passes (multiple passes) are compared to the corresponding
and type, the scanning velocity, the material thermal param- experimental results of Welsink (2005). Both results correlate
eters emissivity and conductivity, as well as, the material well, indicating the capability of the developed model in the
coefficient of thermal expansion. successful prediction of the transient temperature distribu-
The laser forming technique has been extensively inves- tion and the resulting bending angle. Consequently, the model
tigated during the recent years. Numerical and experimental applicability is demonstrated through the prediction of the
investigations have been carried out to better comprehend the optimal forming parameters, i.e. laser power and speed for
mechanisms and the effects of the control parameters on, for two types of aluminium material panels and three different
example, bending angle and mechanical behaviour. Most of thicknesses.
the investigations are mainly limited to steel applications, e.g.
Thomson and Pridham (1998), Kyrsanidi et al. (2000), Hennige
2. Description of the numerical simulation
(2000), Cheng and Lin (2001), Hu et al. (2002), Shichun and
model development
Zhong (2002), Zhang et al. (2004), Zhang and Michaleris (2004),
Chen et al. (2004) and Hsieh and Lin (2004). More specifically
in Thomson and Pridham (1998), Hennige (2000) and Chen et 2.1. Overview of the simulation approach
al. (2004) deal with process development, process parametric
investigations and optimization, in Kyrsanidi et al. (2000) and The experimental set-up of a typical laser forming process is
Cheng and Lin (2001) refer to simplified analytical solutions, presented in Fig. 1. Fixation plates are used to prevent twisting
while Zhang and Michaleris (2004) focuses to the comparison of the specimen during the forming process. The specimen
of Eulerian and Lagrangian approaches to LBF problem. In Hu is placed directly on ceramic strokes to prevent the warmth
et al. (2002), Shichun and Zhong (2002), Zhang et al. (2004)
and Hsieh and Lin (2004) different thermo-mechanical sim-
ulations are presented, which have some similarities to the
presently developed model. However apart from the difference
in investigated target material, which is steel in Hu et al. (2002),
Shichun and Zhong (2002), Zhang et al. (2004) and Hsieh and
Lin (2004), other key differences also exist; in Hu et al. (2002)
the investigated sheet forming mechanism is buckling and not
bending; in Shichun and Zhong (2002) mechanical results are
presented, based on a handling technique for the transitional
zone lying between the elastic and plastic material regions;
in Zhang et al. (2004) the focus is placed on the definition of
minimum discretization requirements of LBF Finite Element
models; in Hsieh and Lin (2004) the laser source comprises a
single pulse and not a continuous heat source scanning the
metal sheet.
Existing investigations referring to aluminium alloys are Fig. 1 – Overview of Laser Beam Forming experimental
still small in number (i.e., Chan and Liang, 2000; Merklein et set-up.
250 j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257

drain of the laser-beamed specimen towards the contacting material properties are introduced in the thermal analysis.
materials. The Nd-Yag laser beam is attached in a movable The results of the thermal analysis, i.e. the transient tem-
frame, capable to follow the desired forming pattern. perature distribution of all loading steps, become the input
During the laser forming process the work-piece is heated to the consequent transient mechanical analysis. The tem-
very locally. The thermal expansion of the heated section is perature dependency of the material mechanical properties
restrained by the surrounding cooler material, leading to gen- and proper mechanical boundary conditions are introduced
eration of compressive stresses. All over the heating period, in the transient mechanical analysis, which results in the cal-
the yield stress of the heated material volumes is reduced; culation of all mechanical magnitudes of interest, such as,
as a result the developed compressive stresses lead to per- residual stresses, plastic strains and distortion distributions
manent compressive plastic strains in the material volumes in the work-piece volume.
that have exceeded the yield stress limit. As the laser beam
moves forward, the material volumes behind the beam start 2.2. Finite Element modelling
cooling down with simultaneous non-uniform lateral shorten-
ing by thermal contraction, which causes bending towards the Three-dimensional (3D) analyses are performed for both the
laser system. The above mechanism involves many interre- thermal and mechanical problems. The reason for a 3D ther-
lated non-linear physical phenomena, such as, high-gradient mal analysis is the need to accurately calculate the high
temperature flow, high stress/large strain development and temperature gradients through the specimen thickness, aris-
abrupt change of material thermal and mechanical proper- ing from the laser beam rapid heating and the consequent
ties, which make the simulation of the laser forming process sudden cooling. The reliable calculation of these temperature
complex and difficult. profiles is very important, as they dominate the local ther-
In the present investigation, a simulation model of the laser mal expansion and contraction of the laser-affected volumes,
beam forming process is developed by means of a coupled which control the amount of plastic deformation and plate
transient thermal–structural analysis. For the numerical anal- bending. For the thermal analysis the element type ‘SOLID 70’
ysis, the Finite Element code ‘ANSYS’ (ANSYS, 2003) is used. An is used, which is an eight-node, three-dimensional thermal
overview of the simulation methodology and its major steps solid with thermal conduction and convection capabilities. For
are presented in the flow chart of Fig. 2. the mechanical analysis the element type ‘SOLID 45’ is used,
A transient thermal analysis of the laser heating/cooling which is an eight-node, three-dimensional structural solid,
process is initially conducted. The geometry of the speci- fully compatible to the ‘SOLID 70’ thermal element, having
men is modelled and a finite element mesh is generated. elastic–plastic non-linear material capability.
Basic parts of the thermal analysis include the specifica- The basic specimen geometry investigated in the present
tion of the heat source characteristics, the development of analysis is a 150 mm × 100 mm rectangular plate of variable
a heat source model, the definition of the irradiation passes thickness. The parametric FE mesh, presented in Fig. 3, is
(thermal load steps) and the introduction of the thermal generated after the identification of the critical component
boundary conditions in the form of heat loss by conduc- volumes; critical volumes are considered those where highly
tion and free convection from the work-piece areas to the increased temperature and strain fields occur during forming,
surrounding environment. Temperature-dependent thermal such as, the material volumes around the laser beam path

Fig. 2 – Flow chart of the laser beam forming process simulation.


j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257 251

Fig. 3 – FE mesh of the entire specimen investigated (left) and FE mesh detail of the half symmetric part where critical
locations are illustrated (right).

and the heat-affected zone. These volumes are empirically ual increase of the element size from the forming line to the
estimated to extend in a zone of width equal to five times boundaries of the component, as shown in Fig. 3.
the laser beam diameter. A dense mesh with average element The investigated specimens are made of 6013-T4 and 2024-
size two thousand times smaller than the laser beam diame- T351 aluminium alloys, as these two alloy types have recently
ter is selected for the above critical areas, in order to enable attracted considerable attention for producing complex-
proper introduction of the laser heat distribution, achieve good shaped aeronautic and automotive components. The material
accuracy of calculated temperature, stress and strain values, surface finish, which is very important as it strongly affects the
as well as to ensure convergence of the solution at reason- laser heat absorption coefficient, is the rolling condition for
able computing times. Concerning the through-the-thickness 2024-T351 material and chemical milling for 6013-T4 material.
discretization, it was found after some FE mesh parametric The temperature-dependent thermal and mechanical mate-
convergence studies, that at least eight elements are required rial properties of 2024-T351 aluminium alloy, which are used
for a proper representation of the high temperature gradients. in the current numerical simulation, are presented in Fig. 4,
The non-critical component volumes are meshed by a grad- as taken from Shi et al. (2003).

Fig. 4 – Temperature-dependent mechanical (a) and thermal (b) material properties of Al 2024-T351.
252 j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257

2.3. Laser beam source model rolled 2024-T351 material, the coefficient n is around 0.22–0.24,
while for the chemically milled 6013-T4, the coefficient n takes
The heat input to the work-piece essentially depends on the values between 0.10 and 0.13. The shape parameter c of Eq.
laser beam energy supplied and the material surface absorp- (1) takes values between 1 and 3 and controls the maximum
tion capability. In order to quantify the heat flux distribution value of the heat flux at the centre of the laser beam. A para-
towards the material surface, a laser beam source model is metric study for the shape parameter c value has shown a
required. In the present analysis the laser beam is assumed to small effect to the calculated bending angle results for the
introduce a three-dimensional Gaussian heat flux distribution 5 mm diameter laser beam which is currently applied. Fur-
on the plate surface, as proposed in Hu et al. (2002) and pre- thermore, parametric studies investigating the influence of
sented in Fig. 5(a) and (b). The laser beam moves continuously the laser absorption coefficient n on the bending angle have
with velocity u along the forming path, however in the simu- shown a significant influence of the value of n on the temper-
lation, the introduced heat flux is assumed to move step-wise ature history and consequently on the formed shape. These
at a step p, indicated in Fig. 5(a). two observations indicate that most important for the laser
The surface heat flux distribution is computed according forming simulation is the amount of the total thermal energy
to the formula: introduced in the local area of the formed part, which should
be distributed according to a Gaussian function, however its
cnP 2
+(y/r0 )2 ))
Q(x, y) = e(−c((x/r0 ) (1) peak value, which is controlled by the value of the coefficient
r02 c is rather secondary. In all presented results hereafter, c is 3.
The heat flux distribution is implemented into the FE code by
where n is the laser absorption coefficient of the irradiated
means of a FORTRAN subroutine.
surface, P is the laser beam power, r0 is the laser beam radius
equal to 2.5 mm for all investigated cases, c is a shape parame-
2.4. Thermal and structural Finite Element solution
ter of the heat-flux distribution and x, y and z are the Cartesian
coordinates.
During the thermal analysis the calculation of nodal temper-
The laser absorption coefficient n of an irradiated surface
atures and their dependency on time is performed by solving
depends mainly on material type, surface treatment, color
the non-linear heat-transfer Eq. (2):
and roughness. It has been shown in Welsink (2005), that for

[Cp (T)]{Ṫ(t)} + [k(T)]{T(t)} + {Q(t)} = 0 (2)

In Eq. (2), C(T) denotes the temperature-dependent


specific-heat matrix, k(T) denotes the temperature-dependent
conductivity matrix, Q(T) is the heat flux vector and T(t)
and Ṫ(t) are the time-dependent nodal temperature and the
time derivative of the nodal temperature vector, respectively.
The heat conduction equation is solved first in fixed coor-
dinates with the Gaussian heat source and then integrated
with respect to time in moving coordinates to provide a quasi-
steady-state traveling solution.
The heat source moving step value, indicated as p in
Fig. 5(a), by which the heat flux is assumed to step-wise move
along the laser line is defined, after a parametric study, to be
equal to the laser beam radius. Higher values of the heat flux
pitch p, lead to poor results, while lower values penalize the
analysis, without any significant effect on the results accuracy.
A part of heat energy transferred by the laser beam to the
metal is lost by free convection. The heat loss by free convec-
tion follows Newton’s law, where the coefficient of convective
heat transfer is assumed to vary with both temperature and
orientation of the boundary:

kc Nu
qc = (T − Ta ) (3)
Lc

where kc is the thermal conductivity of the material, Lc = V/A2


is the plate characteristic length (V and A are the plate volume
and surface, respectively), Ta the ambient temperature, and Nu
Fig. 5 – (a) Assumed Gaussian heat flux distribution of the
the Nusselt number defined by
laser beam source, (p indicates the heat-flux movement
step during the simulation); (b) graphical representation of
the heat source distribution over the FE mesh. Nu = 5.67 Pr1/3 Gr1/3 (4)
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257 253

Table 1 – Laser beam forming cases used in the


validation of the simulation model
Laser Velocity Specimen Case no.
power (W) (mm/s) thickness
(mm)

2750 0.6 10 1
2024-T351 17.5 5 2
1000
17.5 1.5 3

1500 12.5 3.2 4


3.2 5
6013-T4 2500 7.5
5 6
2000 10 2 7

where Pr is the Prandtl number and Gr the Grashof number,


both of them being functions of ambient air properties and
temperature differences between the surface and the environ-
Fig. 6 – Comparison of experimentally measured and
ment. In the current analysis the material is assumed to be at
numerically calculated temperature histories for 2024-T351
room temperature prior to forming and between subsequent
alloy (case no. 1).
forming passes.
For the structural analysis, the non-linear transient struc-
tural dynamics Eq. (5) is solved:
power P = 1500 W, at forming velocity u = 12.5 mm/s (case no.
[M(T)]{ü(t)}+[C(T)]{u̇(t)} + [K(T)]{u(t)} + {F(t)} + {Fth (t)} = 0 (5) 4).
For the experimental case no. 1, in order to determine the
In Eq. (5), M(T), C(T) and K(T) are the temperature-dependent material temperature under the laser path very close to the
mass, damping and stiffness matrices; F(t) is the external load irradiated surface, a blind hole was drilled from the bottom
vector; Fth (t) is the temperature load vector, while, {u(t)}, {u̇(t)} side of the specimen under the laser path and a thermocou-
and {ü(t)} are the displacement, velocity and acceleration vec- ple was placed in the hole (Welsink, 2005). The blind hole of
tors, respectively. Material non-linearity due to plasticity is 9.5 mm was drilled in the middle of the forming path; there-
accounted for by means of Newton–Raphson procedure. For fore, the temperature measurement is only 0.5 mm under the
the plastic yielding, the von-Mises criterion is used to check heated plate surface. For the simulation, the laser beam diam-
for elements where stress has reached the yield point. The eter is set to 5 mm and the absorption coefficient is assumed
Newmark integration method is used to solve the non-linear to be 12%. In Fig. 6 the experimentally measured transient
problem in the time domain. The boundary conditions con- temperature distribution is compared to the computed results
sidered for the structural analysis are representative of the from the simulation model; the origin of the time-axis in Fig. 6
real mechanical boundary conditions of the formed part (see coincides with the time point when the laser passes the ther-
Fig. 1) and are realized by fixing the vertical displacements of mocouple. The maximum measured temperature was 210 ◦ C
one edge point of the plate, as well as, fixing the in-plane dis- and the predicted 194 ◦ C, i.e. a deviation of 7.6% is observed,
placements of the four corner nodes of the plate to prevent indicating a very good correlation.
in-plane twisting. For the experimental case no. 4, the specimen is formed by
Nd-Yag with laser beam diameter 5 mm. For the simulation,
the absorption coefficient is assumed to be 22%. The tem-
3. Validation of the thermo-mechanical perature distribution on the specimen is measured by two
simulation model thermocouples placed on the bottom surface of the speci-
men in two different locations. The first thermocouple ‘t1’
In order to evaluate the model capability to predict the tem- is placed at the plate corner and the second thermocouple
perature distribution during forming and the final bending ‘t2’ is placed at the middle of the forming path line (Welsink,
angles, a number of characteristic trials on plates of dimen- 2005). The experimentally measured temperature history of
sions 150 mm × 100 mm, with different combinations of laser the two thermocouples is presented in Fig. 7 and compared
power and forming velocity have been performed in Welsink to the numerically calculated results; the origin of the time-
(2005), as shown in Table 1. These trials have been used for the axis in Fig. 7 coincides with the time point when the laser
validation of the thermo-mechanical model. passes each thermocouple. For the thermocouple ‘t2’, which
is placed on the heating line, the temperature rises up rapidly
3.1. Validation of the thermal model when the heat source passes the measuring point and drops
relatively quickly when the heating source is moving away
The verification of the thermal model has been performed on from it, until the specimen temperature is gradually reduced
the forming of an Al 2024-T351 plate of thickness t = 10 mm, to room temperature. Thermocouple ‘t1’ has a delay in reach-
laser power P = 2750 W, at forming velocity u = 10 mm/s (case ing its maximum temperature value due to the time required
no. 1) and an Al 6013-T4 plate of thickness t = 3.2 mm, laser for heat conduction into material volumes away from the laser
254 j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257

A graphical representation of the computed temperature


distributions on the formed specimen at four different times-
intervals is presented in Fig. 8, for the experimental case no. 4.
The high temperature gradients at the area close to the laser
beam, as well as, the considerable lower temperatures pre-
dicted away from the heat source, are clearly illustrated in
Fig. 8.
In general, the thermal analysis results of the developed
simulation algorithm demonstrate a good correlation to the
experimental results, as the temperature distributions dur-
ing the heating and cooling stage are predicted with a good
precision.
Fig. 7 – Comparison of experimentally measured and
numerically calculated temperature histories for 6013-T4 3.2. Validation of the mechanical model
alloy (case no. 4).
From the experiments in Table 1, forming case nos. 2, 3 and
5–7 are used in the validation of the mechanical part of
line. The maximum temperatures of thermocouples ‘t1’ and the simulation model, i.e. experimentally measured bending
‘t2’ are 31.9 ◦ C and 95 ◦ C, respectively. The computed maxi- angles from single and multiple laser passes are compared
mum temperatures at the positions of thermocouples are 30 ◦ C to the respective simulation predictions. Concerning the
and 105 ◦ C, respectively. As illustrated in Fig. 7, the calculated multi-pass experiments, the specimen is irradiated repeat-
temperature ‘t2’ tends to drop more rapidly compared to the edly, with the laser line distance between the successive
experimentally measured curve. The same trend may be also passes to be 5 mm, as shown in Fig. 9; in the upper left
observed in Fig. 6. This deviation may be explained by the fact corner of Fig. 9, an illustration of the bending angle defini-
that the working specimen during the experiments is freely tion is also provided. After each pass the specimen is cooled
placed on an almost adiabatic ceramic plate, but it is not forced down until room temperature Ta = 20 ◦ C, before the successive
to a continuous contact on it. Therefore, contact loss occurs pass is initiated. For this reason a high delay time of 300 s
due to initial negative and later positive specimen bending between successive passes is also considered in the simula-
taking place during the forming process. In the simulations, a tions.
continuous free convection to air from the beginning until the The bending angles are experimentally measured by plac-
end of the forming process is considered, as it is not possible to ing a caliper on the free edge of the specimen in order to
estimate the contact surface change due to forming, without measure the displacement of the edge. For the multi-pass
using thermal and mechanical contact elements. This means forming cases the bending angle is measured after the end of
that the real thermal boundary conditions of the bottom plate each pass. The vertical displacement of the deformed panel
surface change locally and continuously between free con- during LBF is presented in Fig. 10, as predicted from the sim-
vection to the air and adiabatic, introducing some deviation ulation of case no. 3. The experimental bending angle values
between experimental and numerical results. from Welsink (2005) are compared in Table 2 to the correspond-

Fig. 8 – Predicted temperature distribution plots at four time intervals during LBF simulation (case no. 4).
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257 255

Fig. 9 – Sequence of the successive LBF passes for the multi-pass tests (dimensions in mm) and illustration of the bending
angle definition (upper left corner).

ing numerical results, as calculated by the current simulation values. Taking into account the remark of an almost linear
model. relation between number of passes and total bending angle,
It may be observed from Table 2, that the bending angle for each investigated case a theoretical single pass value is
increases with decreasing thickness, or by an increase in laser extracted as the average of the five pass experimental value
power or scanning speed. For the cases of multiple-pass LBF, and included in Table 2 (in parenthesis); in such a way, a much
it can be concluded mainly from the simulations, but also better correlation between the experimentally measured and
from the experimental results, that the subsequent passes numerically calculated single pass bending angle values can
lead to an almost linear increase of the bending angle. The be reached. The good correlation of the multi-pass results
maximum deviations between experimentally measured and indicates the capability of the simulation algorithm in the pre-
numerically calculated bending angle values does not exceed diction of the bending angle for different materials, various
28%, 14% and 10% for a single laser pass, three passes and five forming powers and forming velocities.
passes, respectively. The higher deviations refer to the sin-
gle laser beam pass treatments; these deviations arise from
difficulties in the precise experimental definition of the bend- 4. Definition of optimal forming
ing angle, which, due to their low values (even close to zero parameters
for high thickness and low laser power) are significantly influ-
enced by secondary effects, such as non-uniform bending and Optimal forming parameters are those which lead to the
warping, resulting to errors of the experimental single pass desired bending angle at the lower laser power and higher

Fig. 10 – Predicted vertical displacement plots at four time intervals of the LBF simulation (case no. 3).
256 j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257

Table 2 – Comparison between experimentally measured


and numerically calculated bending angle values
Case Bending angle (◦ ) Deviation (%)

Experimental Simulation

1 0.08 (0.06) 0.06 −25


2 Pass 3 0.21 0.2 −5
5 0.33 0.32 −3
1 0.96 (0.99) 0.93 −3
3 Pass 3 3.21 2.94 −8
5 4.99 4.89 −2
1 0.19 (0.22) 0.21 11
5 Pass 3 0.57 0.65 14
5 1.11 1.09 −2
1 0.0 (0.03) 0.02 −
6 Pass 3 0.08 0.07 −13
5 0.19 0.21 10
1 0.5 (0.73) 0.64 28
7 Pass 3 2.1 2.03 −3
5 3.66 3.47 −5

forming speed such that high productivity rates succeed at low


cost and low risk of affecting material properties locally. In the
current study, the developed numerical simulation algorithm
is used to calculate the single pass bending angle of a 6013-
T4 and a 2024-T351 plate, of dimensions 150 mm × 100 mm,
having three different thicknesses, i.e. 2 mm, 3.2 mm and
6 mm. The parametric study includes 210 simulation runs (105

Fig. 12 – Final bending angle of 2024-T351 panels as


function of forming velocity for varying laser power and
thickness t = 2 mm (top), t = 3.2 mm (mid) and t = 6 mm
(bottom).

runs per material type), i.e. three different thicknesses, five


laser power values (from 1000 W to 3000 W with 500 W step)
and seven laser velocities (from 5 mm/min to 20 mm/min at
2.5 mm/min step). The predicted bending angles are presented
in Figs. 11 and 12, for 6013-T4 and 2024-T351 material type,
respectively. Optimum forming parameters may be calculated
from the bending angle results of Figs. 10 and 11, after setting
a desired bending angle value.
It may be observed from Figs. 10 and 11 that in all cases
increase of laser power or a decrease of forming speed lead
to higher bending angles, as expected. It is also observed
that increase of plate thickness results in lower final bend-
ing angle. Some data points are missing from Figs. 11 and 12,
e.g. all data points for forming velocity lower than 0.9 mm/s,
for t = 2 mm and P = 3000 W; the missing bending angles corre-
spond to cases in which the processing temperature reaches
the material melting point. Hence, these cases can be consid-
ered as non-feasible. An overall comparison of the predicted
bending angles between 6013-T4 and 2024-T351 aluminium
panels shows that the final bending angle is always higher for
2024-T351 material type. This observation can be explained by
the higher thermal conductivity values of Al6013-T4 material,
Fig. 11 – Final bending angle of 6013-T4 panels as function as compared to Al2024-T351, resulting in a faster heat transfer
of forming velocity for varying laser power and thickness and smoother temperature gradients through-the-thickness
t = 2 mm (top), t = 3.2 mm (mid) and t = 6 mm (bottom). of the plate, therefore lower final bending deformation.
j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 7 ( 2 0 0 8 ) 248–257 257

Chen, D.J., Wu, S.C., Li, M.Q., 2004. Studies on laser forming of
5. Conclusions Ti–6Al–4V alloy sheet. J. Mater. Process. Technol. 152, 62–65.
Cheng, P.J., Lin, S.C., 2001. An analytical model to estimate angle
A three-dimensional Finite Element model capable of sim- formed by laser. J. Mater. Process. Technol. 108, 314–319.
ulating single and multi-pass laser beam forming process, Geiger, M., Merklein, M., Pitz, M., 2004. Laser and forming
as well as, predicting temperature distributions and bending technology—an idea and the way of implementation. J. Mater.
Process. Technol. 151, 3–11.
shape of LBF aluminium components has been presented. All
Hennige, Th., 2000. Development of irradiation strategies for
the major physical phenomena of the heating and cooling 3D-laser forming. J. Mater. Process. Technol. 103, 102–108.
stages of the LBF process are simulated through a non-linear Hsieh, H.-S., Lin, J., 2004. Thermal–mechanical analysis on the
thermal–structural analysis. The developed model consid- transient deformation during pulsed laser forming. Int. J.
ers temperature-dependent thermal and mechanical material Mach. Tools Manuf. 44, 191–199.
properties and a Gaussian moving heat flux source applied on Hu, Z., Labudovic, M., Wang, H., Kovacevic, R., 2001. Computer
simulation and experimental investigation of sheet metal
the specimen external surface.
bending using laser beam scanning. Int. J. Mach. Tools. Manuf.
The model is capable of predicting temperature distri-
41, 589–607.
butions and final bending angles in good agreement to Hu, Z., Kovacevic, R., Labudovic, M., 2002. Experimental and
experimental data. The simulation algorithm is applied in numerical modeling of buckling instability of laser sheet
the definition of optimum forming parameters for 6013-T4 forming. Int. J. Mach. Tools Manuf. 42, 1427–1439.
and 2024-T351 components. It is shown that increasing laser Kyrsanidi, A.K., Kermanidis, T.B., Pantelakis, S.G., 2000. An
power or decreasing forming speed leads to higher bending analytical model for the prediction of distortions caused by
the laser forming process. J. Mater. Process. Technol. 104,
angles, as well as that increasing plate thickness results in
94–102.
lower achieved bending angles.
Merklein, M., Hennige, T., Geiger, M., 2001. Laser forming of
aluminium and aluminium alloys-microstructural
investigation. J. Mater. Process. Technol. 115, 159–165.
Acknowledgements
Shi, Q.-Y., Dickerson, T., Shercliff, H.R., 2003. Thermo-mechanical
FE modeling of friction stir welding of Al 2024 including tool
Part of this work was performed in the frame of the European loads. In: Proceedings of the International Conference on
Research Programme “Economic Advanced Shaping Processes Friction Stir Welding, Park city, Utah, 14–16 May.
for Integral Structures” (EcoShape). The financial support of Shichun, W., Zhong, J, 2002. FEM simulation of the deformation
the European Union under contract AST3-CT-2003-502884 is field during the laser forming of sheet metal. J. Mater. Process.
Technol. 121, 269–272.
gratefully acknowledged. The forming experiments were con-
Thomson, G., Pridham, M., 1998. Improvements to laser forming
ducted by EADS Corporate Research Center Germany (EADS-G)
through process control refinements. Opt. Laser Technol. 30,
and the Institute for Machine Tools and Industrial Manage- 141–146.
ment (IWB) of the Technical University of Munich, in the frame Welsink, M., 2005. Planning, realizing and analyzing the
of research activities of the above referred project. experiences of Laser Beam Forming. Diploma thesis.
Inholland University, School of Technology, Department of
Aerospace Technology, Haarlem, The Netherlands.
references
Zhang, L., Michaleris, P., 2004. Investigation of Lagrangian and
Eulerian finite element methods for modeling the laser
forming process. Finite Elem. Anal. Des. 40, 383–405.
ANSYS User’s manual, 2003. Swanson Analysis Systems Inc. Zhang, L., Reutzel, E.W., Michaleris, P., 2004. Finite element
Chan, K.C., Liang, J., 2000. Laser bending of an Al6013/SiCp modeling discretization requirements for the laser forming
aluminium matrix composite sheet. J. Mater. Process. Technol. process. Int. J. Mech. Sci. 46, 623–637.
100, 214–218.

Você também pode gostar