Journal of Materials Processing Technology 214 (2014) 2777–2783

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

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

A physical simulation of longitudinal seam welding in

micro channel tube extrusion
Ding Tang ∗ , Qingqing Zhang, Dayong Li, Yinghong Peng
School of Mechanical Engineering, Shanghai Jiao Tong University, China

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

Article history: Micro channel tube is a newly developed type of aluminum profiles with sub-millimeter-diameter ports
Received 4 April 2013 in the cross-section designed for heat transfer enhancement. Micro channel tube is formed with porthole
Received in revised form 17 April 2014 extrusion die, and the longitudinal seam welding problem is the key issue related to both the design
Accepted 4 June 2014
of the delicate mandrel in the extrusion die and the pressure bearing capacity of the tube. This paper
Available online 12 June 2014
proposes a novel method to evaluate the seam welding strength of the micro channel tube at the stage of
extrusion die design. First, a finite element (FE) simulation of the tube extrusion process is performed for
Keywords:
the seam welding conditions in the die chamber and effect of the welding chamber height on hydrostatic
Micro channel tube
Extrusion
pressure. Then, a thermo-mechanical experiment is carried out for a quantitative relationship between
Seam welding the welding strength and the weld condition parameters. Combining this relationship to the numerical
results, the welding strength under different die design can be evaluated. Pressure bearing tests on the
tube prove the reliability of this evaluation method. This study quantitatively connects the seam welding
strength of the profile to the extrusion die parameters, which is helpful for optimizing the design of the
extrusion process.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction mandrel failure accounting for approximately 50%. Furthermore,

the length of the mandrel teeth decides the welding strength of the
Multi-port extrusion (MPE) tube is a type of heat exchanger inner wall between channels because the walls are welded between
tube with a row of side-by-side channels used in items such as mandrel teeth. Therefore, mandrel teeth height should be designed
automobile air conditioners, radiators, and gas/oil/fuel coolers. In for sound welding strength.
recent years, a new generation of sub-millimeter-diameter MPE In recent years, seaming welding strength are more and
tube, also called micro channel tube, has been developed for higher more concerned as aluminum hollow extrusions are being more
heat transfer efficiency and light weight. Björn (2010) reported that widely used in high-pressure condition, such as in condensers.
when the heat exchanger channel diameter decreases to less than Oosterkamp et al. (2004) has discussed the defects in seam welding
1 mm (from an original diameter of 2–5 mm), the heat transfer effi- process that have detrimental effects on the fatigue life and load-
cient rises by more than 200%. Micro channel tube is a kind of high bearing capacity. Nanninga et al. (2010) have examined fatigue of
precision extrusion profile formed with porthole extrusion die, as is seam welding line, which shows that defects and structural inho-
shown in Fig. 1 (cut away view). The billet divides into two streams mogeneities at seam welds reduce the ductility and fatigue strength
by the bridge of the die and then solid welded in the welding cham- of the extrusion.
ber. A comparison of the cross-sectional size and the die mandrel Many criteria have been proposed for seam weld quality evalu-
size between micro channel tube and traditional flat tube is shown ation. Early research on this topic was performed by Akeret (1992),
in Fig. 2. who considered using the maximum pressure at the interface as
As shown in Fig. 2, for micro channel tube, mandrel is thin the discriminating parameter in material welding. Plata and Piwnik
and easily worn or broken in the extrusion. Hiramoto et al. (2009) (2000) proposed the Q criterion (pressure–time criterion) after
presented statistics on the failure modes of a porthole extrusion introducing the time factor. Donati and Tomesani have performed
die used for micro channel tube production, which shows that a great deal of research on seam welding. Donati et al. (2004)
improved the evaluation methodology by taking multiple factors of
the welding chamber into consideration. Later, Donati et al. (2007)
∗ Corresponding author. Tel.: +86 21 34206794; fax: +86 21 34206313. refined this evaluation methodology by considering the distribu-
E-mail address: [email protected] (D. Tang). tion of the contact pressure on the welding plane. Bozzi et al. (2009)

http://dx.doi.org/10.1016/j.jmatprotec.2014.06.004
0924-0136/© 2014 Elsevier B.V. All rights reserved.
2778 D. Tang et al. / Journal of Materials Processing Technology 214 (2014) 2777–2783

that occurs in porthole dies during the manufacture of hollow

aluminum extrusions. Bariani et al. (2006) presented physical-
simulation experiments on longitudinal welds in hot extrusions.
However, the above two studies only obtained qualitative rules of
the influences on the welding strength. So their studies cannot ful-
fill the requirement of weld strength predictions. In practice, an
evaluation method for an extrusion die design is still lacking.
In this work, a new method connecting the extrusion die design
to the welding strength of the extrusion products is introduced.
The method is demonstrated in Fig. 3 with flow chart and is car-
ried out in design of the micro channel tube extrusion die in this
Fig. 1. Schematic of micro channel tube extrusion process.
paper. Firstly, FE simulations of the micro channel tube extrusion
process are performed. From the simulation results, basic weld-
used a novel Lagrangian approach to integrate variables along the ing conditions such as temperature and pressure of the billet in
actual welding paths to get a deeper insight into the seam weld the welding chamber are obtained. And by varying the welding
quality along the tube thickness. Recently, Gagliardi et al. (2012) chamber height, a curve between pressure and welding strength
developed a novel experimental equipment to study the influence is gained. Then, physical simulation experiments are carried out
the bridge shape and width of the porthole extrusion die on welding for welding conditions and provide welded specimen under cer-
quality. Schwane et al. (2013) studied the gas pocket at the bridge of tain conditions. Finally, after tensile tests on the welded specimen,
the porthole die. Welding quality is evaluated by Donati & Tome- quantitative result of welding strength-chamber height curve can
sani criterion. Güley et al. (2013) compared the porthole die and be obtained and welding strength of the micro channel tube is able
normal flat-face die on extrusion of recycled AA6060 chips in order to be evaluated. The evaluation method is validated by hydrostatic
to improve the chip welding quality. And welding quality is evalu- pressure tests.
ated by several criterions. However, these criterions do not provide
a quantitative connection between the welding parameters and the 2. FE simulation of the extrusion process
weld strength.
Some experimental work has been done on the seam weld- In this section, the extrusion process is studied with numerical
ing strength. Jo et al. (2003) studied the relationship between the method. FE simulation was performed using the software package
welding pressures and strength of the extruded tube. Strength of DEFORM 3D. The assembly of the porthole extrusion dies and the
the tube is evaluated with expanding test using a conical punch. FE model is shown in Fig. 4. The FE simulation of the billet and
However, as processing parameters, such as billet temperature and the porthole die is carried out with quarter symmetry boundary
extrusion speed, influence each other, effect of single parameter conditions.
can hardly be evaluated. Recently, physical studies on weld behav- The material of the billet is AA 3003 aluminum alloy with basic
ior have been explored, and parametric studies on weld strength material properties shown in Table 1. The most commonly used
are able to be performed. Edwards et al. (2009) performed physical- constitutive equation for aluminium alloys, Sellars–Tegart hyper-
simulation experiments to study the solid-state bonding process bolic sine law, was applied to the billet material in the FE simulation,
which is
 
1 ε̇ exp(H/GT )
¯ = arcsin
˛ A

where ¯ is the flow stress, ε̇ the mean effective strain rate, ˛

and A the temperature independent constants, H the activation
energy for deformation(141,550 kJ/mol), G the universal gas con-
stant (8.314 J/mol K) and T is the absolute temperature.
The extrusion tools which is composed of die, container and
stem was considered to be thermo-rigid. Heat exchange between
the workpiece and tooling was incorporated in the FE calculation,
while heat transfer between tooling/extrudate and the ambient
surrounding was ignored.
The shear-type friction conditions at the workpiece and tooling
interfaces were imposed as part of the boundary conditions. The

Table 1
Processing parameters of the extrusion process.

Parameters Values

Billet material property Yield stress,  s (MPa) 44.6

Ultimate tensile strength,  b (MPa) 115
Elongation, ı (%) 24

Processing parameters Container inner diameter (mm) 40

Extrusion ratio 150
Ram speed (mm/s) 5
Billet temperature (◦ C) 480
Die temperature (◦ C) 450
Friction coefficient (billet–container) 0.7
Fig. 2. Comparison of cross-section and mandrel between micro channel tube and Friction coefficient (workpiece–die) 0.4
normal MPE tube.
 D. Tang et al. / Journal of Materials Processing Technology 214 (2014) 2777–2783 2779

Fig. 3. Flow chart of the welding strength evaluation method.

friction factor at the billet–container and billet–stem interfaces was

assumed to be 0.7 and that at the workpiece–die interface to be 0.4.
As the thickness of the tube wall is only 0.25 mm, the initial mesh
size of the billet is set to be 0.5 mm which is smaller than most of
studies on the aluminum extrusion. And self-adapting remeshing
was applied which is able to solved the problems of local intensive
deformation, for example, near the die bearing. The number of total
elements in the simulation model at the initial step is 150,000. The
model uses computational time steps of 0.01 s and 0.003 s for the
dividing and welding stages, respectively.
Fig. 5 shows the filling process of the porthole extrusion die
derived from the FE model results. The billet is first divided into
two sections through the two portholes in the die. Then, as the
metal flows into the welding chamber, the two sections are rejoined
through the mandrel teeth. After the welding chamber is filled, the
metal is pushed out of the die bearing, and the profile of the tube
is formed. Due to the harmonica-shaped cross section, the micro
channel tubes have a relatively long welding line, with multiple
scattered welding points at each web of the channels.
The bonding process for the billet inside the welding chamber is
enlarged and shown in detail in Fig. 6. The webs of the middle chan-
nels are welded first and then spread to the channel webs at both
sides. This process means that the metal flow speed in the central
part of the die is higher than that at the exterior. This description
agrees with the general speed distribution pattern in an extrusion
forming process. According to the thin wall between the channels,
the width of the welding chamber is less than 1 mm, which is the
result of the low speed of the metal flow in the normal direction of
the welding plane. Fig. 4. Extrusion die for the micro channel tube and the FE model.
2780 D. Tang et al. / Journal of Materials Processing Technology 214 (2014) 2777–2783

Fig. 5. The three metal flow stages during the extrusion of the micro channel tube.

Fig. 6. Illustration of the seam welding process used to create the channel walls.

The height of the welding chamber in the FE model is varied hydrostatic pressure is realized for welded specimens. Secondly,
to examine the influence of the die structure on the welding con- welded specimens are tested for welding strength.
ditions. According to the length of the mandrel teeth used in this Physical simulation experiments of the solid-welding process
study, the height range of the welding chamber is between 2 and by done by pressing two aluminum cylindrical specimens under
4 mm. Fig. 7 shows that as the welding chamber height increases, pre-defined temperature and speed according to previous study of
the hydrostatic pressure increases from 96 MPa to 203 MPa. Edwards et al. (2009). A Gleeble 3500 thermo-mechanical simulator

3. Physical simulation of the seam welding

By FE simulation of the extrusion process, relationship between

the welding chamber height and hydrostatic pressure was
obtained. In order to connect the chamber height to the seam
welding strength directly, effect of hydrostatic pressure on welding
strength need to be known. In this study, two steps of experiments
are put forward for this. Firstly, bonding process under different

Fig. 8. Thermo-mechanical experiment: (a) specimen fixture and (b) definitions of

Fig. 7. Influence of the welding chamber height on the hydrostatic pressure. parameters.
 D. Tang et al. / Journal of Materials Processing Technology 214 (2014) 2777–2783 2781

Fig. 9. Stress–strain curves during the welding process under different pressing
velocity. Fig. 10. Stress–strain curves during the welding process under different strain.

consideration the same way it does with the necking in the tensile
is used to provide a vacuum environment and precise controlling of test. The contact pressure is expressed with non-dimensionalized
temperature, pressure and strain rate. The installation of the spec- form p/ s , in which P is the contact press and  s is the yield stress
imen and the schematic drawing of the experiment are shown in of the billet material.
Fig. 8. The specimen dimensions are 10 mm in diameter and 45 mm The aim of the thermo-mechanical simulation is to provide
in length. specimen under different hydro pressure according to Fig. 7. This
During the test, the halves of the specimen are axially aligned pressing process is taken as free upsetting on cylinder, so the mid-
and contact each other. Then, the platform is depressurized to dle part of the welding plane can be taken as under the triaxial
10−1 Torr, and the specimen is heated at a rate of 2.5 ◦ C/s by the compression stress state. The stress at the welding plane is approx-
direct resistance heating system. When the desired temperature is imately treated as hydrostatic pressure.
reached, the conditions are held constant for 30 s, after which the In this study, strain (displacement of the clamp) and velocity
specimens are pressed against each other at a pre-defined speed. are varied to generate different hydrostatic pressure at the welding
The stress and strain are recorded during the test. plane. Stress–strain curves with specimens under different press-
In the test, the cross-sectional area and length of the specimen ing velocity and strain is shown Figs. 9 and 10, respectively. It is
changed substantially, so the true stain is used toexpress the short-
 clear that velocity (strain rate) has more obvious effect on pressure
ened displacement and is calculated from ε = ln L0 − L/L0 , in of the welding plane. The pressure stress at the strain of 0.2 is taken
which L0 and L are the initial working length and the shortened as the hydrostatic pressure. With the velocity growth from 1 mm/s
distance, respectively. True stress and strain measures account for to 500 mm/s, the pressure increased from 65.2 Mpa to 121.5 Mpa.
changes in cross-sectional area by using the instantaneous values The tensile tests on the bonded specimens are carried out
for area. Therefore, it partly takes the bulge neat contact surface into using Zwick tensile testing machine. The weld strength is
2782 D. Tang et al. / Journal of Materials Processing Technology 214 (2014) 2777–2783

Fig. 11. Relationship between pressure and welding strength.

non-dimensionalized with the specific strength  by comparing the

weld strength to the tensile strength of the basic AA 3003 material
(44.6 Mpa).
Fig. 11 shows the influence of the pressure on the welding
strength. Increasing the pressure enhances the welding effect.
When the pressure level is low, the strength increases fast with the
increasing of pressure. However, after the pressure is over 87 Mpa,
the welding strength reaches a plateau which is 0.8–0.9 times of
the base material strength. Therefore, the pressure 87 Mpa can be
taken as the critical value of sound welding pressure. Fig. 13. Extrusion die: (a) welding chamber and (b) assembled die.
Further, the interfaces of the welding plane are observed along
the weld axis. Fig. 12 shows the exposed weld line under a micro- temperature and 3 mm in welding chamber height. By simulation
scope. From illustration (a) to (h), the pressure from 63.0 Mpa to the welding pressure is 142.7 Mpa which will result in 0.82 in spe-
216.5 Mpa, respectively, and the interface gradually becomes indis- cific strength.
tinguishable. When the pressure is over 90 Mpa, the micro voids The welding chamber part and the extrusion die are shown in
scatted along the bonding surface is completely closed and the weld Fig. 13. A hydrostatic pressure test is done to examine the tube’s
line is invisible. This result agrees well with the strength tests of welding strength. Fig. 14 shows the testing equipment and the tube
welded specimens described in Fig. 11. Integrate the experiment after failure. The failure pressure of the tube is 35.6 Mpa. According
result in Fig. 11 to the simulation results in Fig. 7, conclusion can be to the Vamadevan and Kraft (2007)’s study on pressure bearing of
made that welding chamber height larger than 2 mm under such micro-channel
√ tubes, the stress on the web can be calculated with
die design can reach the requirement that seam welding strength  = 3/2(Pi (1 + (D/t))), where Pi is the internal pressure, D the
larger than 80% of the basic material strength. diameter of circular channel and t is the thickness of the web. In
this study, D is substituted with the length of the square channel.
4. Experimental validation The ultimate strength of the welding strength after is 92.6 Mpa
which is 0.805 in specific strength. Compare to the calculation by
In order to validate the conclusion obtained in Section 3, form- evaluation method proposed (0.82), it can be taken as an agreement
ing experiments is done according to the FE model: 450 ◦ C in for engineering application.

Fig. 12. Microstructure of the welding line at different strains.

 D. Tang et al. / Journal of Materials Processing Technology 214 (2014) 2777–2783 2783

(3) By combining the configuration of the welding parameters with

the numerical results, the welding strength can be evaluated.

Acknowledgments

The authors would like to acknowledge the supported of NSFC

(Project 51105249), National Science & Technology Pillar Program
(Project 2012BAF01B06), Shanghai Municipal Science and Tech-
nology Commission Project (11dz1202102), China Postdoctoral
Science Foundation (20110490725), Shanghai Postdoctoral Sus-
tentation Fund (10R21414300). The authors also appreciate the
supported of the Research Fund of State Key Lab of MSV, China
(Grant No. MSV-MS-2010-05).

References

Akeret, R., 1992. Extrusion welds-quality aspects are now center stage. In:
Proceedings of the Fifth International Aluminum Extrusion Technology Seminar,
pp. 319–336.
Bariani, P.F., Bruschi, S., Ghiotti, A., 2006. Physical simulation of longitudinal welding
in porthole-die extrusion. Ann. CIRP 55, 11–14.
Björn, P., 2010. Heat transfer in microchannels. Nanoscale Microscale Thermophys.
Eng. 5, 155–175.
Bozzi, S., Vedani, M., Lotti, D., Passoni, G., 2009. Extrusion of aluminium hollow pipes:
seam weld quality assessment via numerical simulation. Metal. Sci. Technol. 27,
20–29.
Donati, L., Tomesani, L., Giacomelli, V., 2004. Evaluation of a new FEM crite-
rion for seam welds quality prediction in aluminum extruded profiles. In:
Proceedings of Eighth International Aluminum Extrusion Technology Seminar,
pp. 221–235.
Donati, L., Tomesani, L., Minak, G., 2007. Characterization of seam weld quality in
AA6082 extruded profiles. J. Mater. Process. Technol. 191, 127–131.
Edwards, S.P., Den Bakker, A.J., Zhou, J., Katgerman, L., 2009. Physical simulation
Fig. 14. Evaluation of the welding quality: (a) test apparatus and (b) specimen after of longitudinal weld seam formation during extrusion to produce hollow alu-
failure. minum profiles. Mater. Manuf. Process. 24, 409–421.
Gagliardi, F., Ambrogio, G., Filice, L., 2012. On the die design in AA6082 porthole
extrusion. CIRP Ann. Manuf. Technol. 61, 231–234.
5. Conclusions Güley, V., Güzel, A., Jäger, A., Ben Khalifa, N., Tekkaya, A.E., Misiolek, W.Z., 2013. Effect
of die design on the welding quality during solid state recycling of AA6060 chips
by hot extrusion. Mater. Sci. Eng. A 574, 163–175.
In this work, seam welding strength is studied by physical sim- Hiramoto, K., Minami, S., Sakihama, H., 2009. Development of the latest heat
ulation. A novel method is proposed and tried in determination of exchanger tube technology. In: Proceedings of the 9th International.
the welding chamber height of the micro channel tube extrusion Jo, H.H., Jeong, C.S., Lee, S.K., Kim, B.M., 2003. Determination of welding pressure in
the non-steady-state porthole die extrusion of improved Al7003 hollow section
die. The welding conditions for the extrusion process are obtained tubes. J. Mater. Process. Technol. 139 (1), 428–433.
using FE analysis. Then, a series of physical simulation experiments Nanninga, N., White, C., Mills, O., Lukowski, J., 2010. Effect of specimen orientation
are performed to for a quantitative investigation. Based on this and extrusion welds on the fatigue life of an AA6063 alloy. Int. J. Fatigue 32,
238–246.
investigation, the following conclusions can be drawn: Oosterkamp, A., Djapic Oosterkamp, L., Nordeide, A., 2004. Kissing bond phenomena
in solid-statewelds of aluminum alloys. Weld. J., 225–231.
(1) The micro channel tubes have scattered multiple welding Plata, M., Piwnik, J., 2000. Theoretical and experimental analysis of seam weld for-
mation in hot extrusion of aluminum alloys. In: Proceedings of the Seventh
points, and the welding conditions of the channel walls at the
International Aluminum Extrusion Technology Seminar, pp. 205–211.
ends are much larger than those of the channels walls in the Schwane, M., Jäger, A., Gagliardi, F., Ben Khalifa, N., Tekkaya, A.E., 2013. Model-
middle. ing approach for the determination of material flow and welding conditions
in porthole die extrusion with gas pocket formation. Key Eng. Mater. 554,
(2) When the pressure is over critical value (87 Mpa), welding
787–793.
strength reaches a plateau of 0.8–0.9 times of the base material Vamadevan, G., Kraft, F.F., 2007. Processing effects in aluminum micro-channel tube
strength and the micro voids scatted along the bonding surface for brazed R744 heat exchangers. J. Mater. Process. Technol. 191 (1), 30–33.
is completely closed and no welding line is visible.