See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/286818787

Effect of Fiber Laser Welding on the Fatigue Properties of Dissimilar Welded

Joints between DP980 and HSLA Steels

Article  in  SAE Technical Papers · April 2014

DOI: 10.4271/2014-01-0990

CITATIONS READS

2 281

9 authors, including:

Dennis Parkes Daniel Westerbaan

ArcelorMittal University of Waterloo
9 PUBLICATIONS   173 CITATIONS    14 PUBLICATIONS   450 CITATIONS   

SEE PROFILE SEE PROFILE

S. S. Nayak Frank E Goodwin

Hitachi Construction Machinery International Zinc Association
44 PUBLICATIONS   1,373 CITATIONS    555 PUBLICATIONS   1,754 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Effect or RSW Electrode Alloy on Electrode Wear View project

Nuclear Waste Containment View project

All content following this page was uploaded by Frank E Goodwin on 18 February 2016.

The user has requested enhancement of the downloaded file.

 2014-01-0990

Effect of Fiber Laser Welding on the Fatigue Properties of Dissimilar

Welded Joints between DP980 and HSLA Steels
D. Parkes, Q.L. Cui, D. Westerbaan, S.S. Nayak, Y. Zhou, F. Goodwin, D. Liu, S.D. Bhole, D.L. Chen

Abstract welding DP and HSLA steels using various welding techniques

[4,5,7,8]. It was determined from this that fiber laser welding
produced the best quality welds at the lowest cost to the
Advanced high strength steels (AHSS) such as dual phase
manufacturer due to its smaller beam divergence, low
(DP) steels are now being extensively used to achieve light
maintenance costs, higher efficiency, high precision and
weighting goals of vehicles because of their attractive
reliability, and compact size [9]. This work is a continuation of
combination of formability and high strength. High strength low
previous work to enhance the understanding of dissimilar
alloy (HSLA) steels are also used in lightweight bodies-in-
welded joints and to optimise their mechanical properties for
white; DP and HSLA steels are therefore often laser butt-
manufacturing purposes, by changing the welding parameters
welded together into tailor welded blanks and to create other
such as weld power and or welding speed to potentially
joints. Among the laser welding processes, fiber laser welding
increase vehicle output.
(FLW) has been shown to provide excellent quality welds,
including superior weld mechanical properties, at higher
speeds than those possible with other laser welding processes. Experimental Procedure
Using dissimilarly welded DP980-HSLA blanks made with
different welding parameters, the tensile properties were found The materials used in this work are HSLA380 and DP980 steel
to not change in spite of the HSLA being weaker than the soft sheets with a sheet thickness of 1.2 mm and have a
zone on the DP980 side of the weld. The high heat input weld galvanized (GI) coating. The chemical compositions of the
was found to have more softening in comparison to its base steels are listed in Table 1. The welded blanks were made
metal (BM) (55 HV versus 46 HV) and less bainite (8% versus using one 100 mm × 200 mm sheet from each material and
15%) in its FZ in comparison to the low heat input weld. The butt welded using an IPG Photonics YLS-6000 fiber laser
fatigue properties of the high heat input weld were also found system attached to a Panasonic robotic arm to make 200 mm
to be better than those of the low heat input weld. × 200 mm sheets as shown in Fig. 1(a). Two different welding
conditions were used in this work and are the conditions
Introduction summarized in Table 2. It should be noted that a little concavity
was observed in all the welds, and the subsequent mechanical
testing was carried on the samples in the as-welded state
In automotive manufacturing, there is a lot of work being done
without any post-weld processing.
in order to increase the fuel economy of vehicles in order to
help preserve the environment [1]. A common method is to Table 1. Chemical composition of the DP980 and HSLA steel in
make the vehicle more lightweight so that it takes less energy present study.
to move the vehicle. By reducing the weight of the car, it could
be possible to compromise the safety of the vehicle, which is Steel C Mn Si Al Cr Ni N Fe
not an acceptable trade-off. Therefore, by greatly increasing DP980 0.15 1.45 0.33 0.05 0.02 0.01 0.009 Balance
the strength of the materials that are being used and reducing HSLA380 0.08 0.83 0.5 0.05 0.03 0.01 0.007 Balance
the size of the individual components to reduce the weight of
the vehicle is a logical means to keep the vehicle safe, while
reducing weight. One of the many solutions is advanced high Table 2. Welding parameters and corresponding heat inputs. [10]
strength steels (AHSS) which have excellent strength
Welding Speed Laser Power Heat Input
properties while maintaining good ductility and formability Steel
(m/min) (kW) (J/mm2)
properties [2,3]. In this particular study, dual phase (DP) steels, Low Heat Input 20.5 4 9.76
were chosen for their wide range of strengths and excellent High Heat Input 16 6 18.75
formability properties. Specifically, DP980 was chosen since it
has an ultimate tensile strength (UTS) of at least 980 MPa and Metallographic samples were cut from the weld cross-section,
good ductility (~14%) [4,5]; since it still has the same density of then mounted, ground, polished, and etched with a 2% Nital
a conventional high strength low alloy (HSLA) steel, solution. The etched samples were observed first using a light
components can be made from thinner sections. In fact it is microscope attached with Clemex image analysis system
being used to replace some structural components such as B- software and then with a scanning electron microscope (JEOL
pillar [6]. JSM-6380) equipped three-dimensional fractographic analysis.
Vickers microhardness was measured on the polished samples
Since HSLA steel is commonly used in automotive across the weld using a computerized Buehler micromet 5103
manufacturing and with the increase in use of DP steels in microhardness tester using a load of 200 g and a dwell time of
lightweight new vehicles, it is inevitable that the metals would 15 s. Care was taken to prevent any potential effect of
be welded together. An increased understanding of how a localized strain hardening caused by adjacent indentations. To
dissimilar welded joint behaves would be highly beneficial. A ensure the accuracy of each test result, two calibration tests
lot of research has gone into the phenomena that occur when
Page 1 of 6
were carried out using a standard reference test block before [4,5,7,8,10,11]. The HSLA BM has a fine grained ferrite matrix
the microhardness tests on the welded joints were conducted. embedded with ultra fine alloyed carbides [4,5].
100 mm

(a)
200 mm

DP980 HSLA

Figure 2. BM microstructure of (a) DP980 and (b) HSLA. (F: ferrite, M:

martensite)

Rolling direction
L In the FZ of the high heat input welded joint, the structure is
(b) B found to be highly martensitic with very little bainite found in
A the structure (~8%) (Fig. 3(a)). However, the low heat input
C
W
G the weld appeared similar in both heat input conditions with an
R
Position G W R L A B C
Dimension (mm) 50 12.5 12.5 200 57 50 20
Figure 1. (a) Geometry of the welded sheet, (b) Transverse tensile joint
machine from the dotted line in (a).
Tensile and fatigue test samples were machined to form
transverse welds in accordance with ASTM: E8/E8M, the
example of which is indicated by the dashed line in Fig. 1(a)
with the geometry and dimensions of the test coupons shown
in Fig. 1(b). The specimens were machined in such a way that
the weld line was positioned at the center of the gauge length
(Fig. 1(b)). Tensile tests were conducted on a fully
computerized United tensile testing machine at room
temperature and with a strain rate of 1×10-3 s−1. An
extensometer with a gauge length of 50 mm and a strain limit
of 20% was used to measure the strain during the tensile tests.
Load control fatigue tests were performed following ASTM:
E466 on a fully computerized Instron 8801 servo-hydraulic
testing system. Tension-tension cyclic loading was used to
prevent potential buckling of the samples, so a stress ratio of R
= 0.1 was used. The test conditions were conducted at room
temperature with a frequency of 50 Hz and the load was
applied using a sinusoidal waveform. A minimum of two
specimens were tested in the tensile tests and fatigue tests at
each of the cyclic stress amplitudes. The fatigue fracture
surfaces were examined after failure via scanning electron
microscopy (SEM) to view the fatigue fracture mechanisms.
The fusion zones (FZs) of the two different heat inputs show
some interesting differences in the welding behaviour (Fig. 3).
welds were found to contain approximately 15% bainite in the
FZ (Fig. 3(b)). The heat affect zone (HAZ) of the two sides of
upper-critical HAZ, an inter-critical HAZ, and on the DP980
side of the welded joints a sub-critical HAZ [12]. In the upper-
critical HAZ, the temperature was well above the Ac3 line and
thusly mostly transformed into mostly martensite with a small
amount of ferrite (Fig. (c) and (d)) [12]. In the inter-critical HAZ,
the temperature was between the Ac1 and Ac3 lines and,
therefore, had a hardness similar to that of the DP980 BM;
however, due to the difference in initial microstructure, the
inter-critical HAZ of the DP980 (Fig. 4 (b)) looks very similar to
the DP980 BM (Fig. 2(a)) because there was martensite
already present in the structure; however, the HSLA did not
have martensite initially and therefore transformed partly into
martensite and bainite (Fig. 4(e)). In the aforementioned sub-
critical HAZ on the DP980 sides of the weld, the temperature
was below the Ac3 line [4,5,7,8,10,11]. This caused the
martensite that is normally present in the BM to temper (Fig.
4(a)). This tempering process causes the presences of a “soft
zone” in which the DP980 is weakest at these points [10,11]. It
should also be noted that in the low heat input weld, the FZ
and HAZ width are 400 µm and 500 µm respectively, and in
the high heat input welded joints, they are both 500 µm which
would also be related to the amount of extra material that
would have been melted in the FZ of the high heat input weld.

Microstructure
Figure 3. The FZ microstructure of (a) high heat input joint, and (b) low
heat input joint. (M: martensite, B: bainite)
The base metals (BMs) of both DP980 and HSLA steels can
be seen in Fig. 2. The DP980 BM has a microstructure
consisting of a ferrite matrix with islands of martensite. The
DP980 BM was made by inter-critical annealing of and
quenching which alloyed for the dual phase structure to form
Page 2 of 6
 Figure 5. Microhardness curves of HSLA-DP980 welded joints made
with different welding conditions. [5]

Tensile

According to Haiping et al. [13], in a transverse dissimilar

Figure 4. Weld microstructure of (a) DP980 subcritical HAZ, (b) DP980 welded joint, the deformation is dominated by the weakest
inter-critical HAZ, (c) DP980 upper-critical HAZ, (d) HSLA upper-critical material. Since the UTS of the HSLA steel is lower than the
HAZ, and (e) HSLA inter-critical HAZ. (PTM: partially tempered
yield strength (YS) of the DP980, the DP980 will only deform
martensite, F: ferrite, M: martensite, B: bainite)
elastically leaving only the HSLA to yield and deform
plastically. From the tensile curves in Fig. 6 and from Table 3,
Microhardness it can be seen that both the YS and the UTS of the two welding
conditions are similar. This is due to the reason mentioned
As can be seen from the microhardness curve in Fig. 5, the previously that the HSLA is the material that is limiting the
general shape of the hardness curves between the two heat strength of the welded joint; the difference in the elongation
inputs is the same, which shows the soft HSLA BM which between the two welding conditions is due to experimental
increases in hardness as it moves into the inter-critical HAZ scatter. As previously mentioned, any potential deformation
and upper-critical HAZ reaching the peak in the FZ. This is that may have occurred in the soft zone would be negligible
because of the increasing amount of very high strength since it is only ~200 µm across and is still stronger than the
martensite and the bainite that is present in these regions. A HSLA BM [5].
difference in the hardness on both sides of the FZ has been
reported before [5] and represents the difference in carbon
content between the two BMs and the lack of mixing during the
fusion process because of the very high cooling rates. On the
DP980 side, the presence of the soft zones can be clearly
seen in relation to the BM due to the presence of the partially
tempered martensite. However, it should be noted that the
degree of softening that occurred in both the low heat input
and high heat input welds are 46 HV and 55 HV, respectively.
Even though this soft zone in the DP980 soft zone is present,
since it is welded to HSLA, the minimum hardness of the soft
zone is still significantly stronger than that of the HSLA BM;
therefore, it is expected that most if not all of the deformation
should occur on the HSLA side of the welded joints [5].

Figure 6. Tensile curves of HSLA-DP980 welded joints made under

different welding conditions.

Page 3 of 6

Fatigue and Fractography
In the dissimilar HSLA-DP980 welded joints of both welding
conditions, the peak stress was above the yield stress of HSLA
BM at high stress amplitudes of 250 MPa causing the yielding
and failure in the BM. as shown in Fig. 7(b). At the lower stress
amplitudes, there was more scatter in the experimental data
due to the fact that the samples were failing in the weld area
(Fig. 7(a)) [5].This was due to the stress concentration
developed by the concavity and slight sheet misalignment
during welding as well as the heterogeneous microstructure
caused during the welding procedure because they are far
more sensitive under fatigue than under tensile load (Fig. 8)
[14,15].
data points in Fig. 9 according to a log-log line of Eq. (1) and
are tabulated in Table 3. The fatigue life of the welded joints
could be estimated based on the values of  f ' and b. From
Table 3, it can be seen that the high heat input welds have a
better b value and a higher fatigue limit which means that they
would have a better fatigue life in comparison to the low heat
input samples.
275
High heat input
Low heat input
225

Stress amplitude, MPa

175

125

75
1E+1 1E+2 1E+3 1E+4 1E+5 1E+6 1E+7 1E+8
Number of reversals to failure, 2Nf

Figure 9. Comparison of S-N curves of the HSLA-DP980 welded joints

with different heat inputs.

Figure 7. Typical fatigue fracture locations at (a) low stress amplitude, Table 3. Summary of fatigue and tensile properties [5].
and (b) high stress amplitude.

Steel
f' b
Fatigue
Limit
YS UTS
(MPa) (MPa)
(MPa) (MPa)
HSLA BM 302 -0.023 200 455 546
DP980 BM 1019 -0.098 250 720 1095
Low Heat Input 455 -0.085 100 478 543
High Heat Input 355 -0.049 <150 467 548

Figure 8. Weld profile of (a) low heat input weld and (b) high heat input
weld.
The fatigue limit of the high and low heat input samples where
determined from the unlimited life approximation when the
samples reached 107 cycles without failure. As previously
mentioned, the primary difference between the fatigue limits
was due to the difference in concavity between the two
samples and because of this caused a lot of experimental
scatter, especially in the high heat input samples.

The typical fatigue fracture surface of the dissimilar HSLA-

The S-N plot in Fig. 9 could be expressed using the Basquin
equation [14,15],
𝜎𝑎 = 𝜎𝑓′ (2𝑁𝑓 )𝑏 ,
where σa is the cyclic stress amplitude,  f ' is
strength coefficient defined by the stress intercept at 2Nf = 1, Nf
is the number of cycles to failure (2Nf is the number of
reversals to failure), and b is the fatigue strength exponent.
The values of these parameters were obtained by fitting the
Page 4 of 6
DP980 welded joints that failed at the lowest stress amplitudes
(150 MPa for high heat input and 125 for low heat input) can
be seen in Fig. 7(a). It can also be seen that crack initiation
occurred at the specimen surface due to the stress
(1) concentration arising from the presence of concavity and
misalignment. The fatigue striations can be seen to be
propagating in multiple directions, which is characteristic of this
the fatigue region. In the propagation region, fatigue striations can be
seen in Figs. 10 (c) and (d), where fatigue damage
accumulated as the propagation approached the final fracture.
Final fast fracture area consisted of characteristic dimples Fig.
10 (e) and (f). This is understandable since both HSLA and
DP980 are quite ductile [5]. To enhance the fatigue strength of
Figure 10. Fatigue fracture surface features: initiation region (a) high
heat input, and (b) low heat input. Propagation region: (c) high heat
input and (d) low heat input. Fast fracture: (e) high heat input and (f)
low heat input.
such dissimilar welded joints, further efforts are needed to get
rid of the weld concavity and sheet misalignment.
Conclusions
The major conclusions that can be drawn from this work are as
follows:
1. The tensile properties of the two weld heat inputs
were very similar even though the high heat input
weld had more softening than the low heat input weld.
This is because the strength of the soft zone in both
cases was still much greater than the hardness of the
sub-critical HAZ.
2. The low heat input welded joint had significantly more
bainite in its FZ than the high heat input welded joint.
3. The low heat input weld had a lower fatigue life than
the high heat input welds because of a combination of
stress concentration caused by slight misalignment of
the welded sheets and minor concavity made in the
welding process.
References
1. Murray, J., and King, D., “Oil’s tipping point has passed,”
Nature 481:433-435, 2012.
Page 5 of 6
2. Joost, W.J., “Reducing vehicle weight and improving U.S.
energy efficiency using integrated computational materials
engineering,” Journal of Materials 64:1032-1038, 2012.
3. Kim, H.J., Keoleian, G.A., and Skerlos, S.J., “Economic
assessment of greenhouse gas emissions reduction by
vehicle lightweighting using aluminum and high-strength
steel,” Journal of Industrial Ecology 15:64-80, 2011
doi:10.1111/j.1530-9290.2010.00288.x.
4. Xu, W., Westerbaan, D., Nayak, S.S., Chen, D.L., et al.,
“Tensile and fatigue properties of fiber laser welded high
strength low alloy and DP980 dual-phase steel joints,”
Materials and Design 43:373-383, 2013.
5. Parkes, D., Xu, W., Westerbaan, D., Nayak, S.S., et al.,
“Microstructure and fatigue properties of fiber laser welded
dissimilar joints between high strength low alloy and dual
phase steels,” Materials and Design 51:665-675, 2013.
6. Kwon, O., and Baik, S.C., “Manufacture and application of
advanced high strength steel sheets for auto
manufacture,” Pohang:POSCO;785-790, 2005.
7. Farabi, N., Chen, D.L., and Zhou, Y., “Tensile properties
and work hardening behavior of laser-welded dual-phase
steel joints,” Journal of Materials Engineering and
Performance 21:222-230, 2012.
8. Farabi, N., Chen, D.L., and Zhou, Y., “Microstructure and
mechanical properties of laser welded dissimilar
DP600/DP980 dual-phase steel joints,” Journal of Alloys
and Compounds 509:982-989, 2011.
9. Quinto, L., Costa, A., Miranda, R., Yapp, D., et al.,
“Welding with high power fiber lasers – A preliminary
study,” Materials and Design 28:1231-1237, 2007.
10. Xia, M., Biro, E., Tian, Z., and Zhou, Y., "Effects of Heat
Input and Martensite on HAZ Softening in Laser Welding
of Dual Phase Steels," ISIJ International 48(6):809-814,
2008.
11. Biro, E., McDermid, J., Embury, J., and Zhou Y.,
"Softening Kinectics in the Subcritical Heat-Affected Zone
of Dual-Phase Steel Welds," The Minerals, Metals &
Materials Society 41A:2348-2356, 2010.
12. Kou, S., “Welding Metallurgy,” John Wiley and Sons, Inc.,
New Jersey, 2003.
13. Shao, H., Gould, J., and Albright C., “Laser blank welding
high-strength steels,” Metallurgical and materials
transactions 38B:321-331, 2007.
14. Hertzberg, R.W., “Deformation and fracture mechanics of
engineering materials,” 4th ed., John Wiley & Sons Inc.,
New Jersey, 1996.
15. Dieter, G.E., “Mechanical metallurgy,” 3rd ed., McGraw-
Hill, New York, 1986.
Acknowledgments
The authors would like to thank the Natural Sciences and
Engineering Research Council of Canada (NSERC), AUTO21
Network of Centers of Excellence, the basic scientific research
business expenses project of the Ministry of Education (No.
N110407001), NSFC (No. 51304045), and the China
Scholarship Council for providing financial support. The
financial support from International Zinc Association (IZA) and
ArcelorMittal Dofasco is highly acknowledged. One of the
authors (D.L. Chen) is grateful for the financial support by the
Premier’s Research Excellence Award (PREA), NSERC-
Discovery Accelerator Supplement (DAS) Award, Canada
Foundation for Innovation (CFI), and Ryerson Research Chair
(RRC) program. The authors would like to thank Dr. J. Chen
and Dr. Y.L. He (CANMET-Materials Technology Laboratory, their support and helpful discussion. The assistance of Q. Li, A.
Natural Resources Canada, Hamilton, Canada), Mr. E. Biro Machin, J. Amankrah, and R. Churaman in performing the
(ArcelorMittal Global Research, Hamilton, Canada), and Dr. J. experiments in this work is gratefully acknowledged.
Villafuerte (CenterLine (Windsor) Ltd., Windsor, Canada) for

Page 6 of 6

View publication stats