Journal of Constructional Steel Research 131 (2017) 110–121

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Impact of welding on the strength of high performance steel T-stub joints

M.S. Zhao a,⁎, C.K. Lee b, T.C. Fung c, S.P. Chiew a
a
Singapore Institute of Technology, Singapore
b
School of Engineering and Information Technology, University of New South Wales Canberra, Australia
c
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore

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

Article history: In this study, experiments were carried out to investigate the influence of welding on the strength of the
Received 7 September 2016 reheated, quenched and tempered (RQT) steel S690 and the thermo-mechanically controlled processed
Received in revised form 23 December 2016 (TMCP) steel S385. Firstly, a specially designed welding procedure was applied to produce RQT and TMCP steel
Accepted 29 December 2016
plates that were fully affected by the welding heat input for the whole plate thickness. These fully weld-affected
Available online 6 January 2017
plates were then fabricated into specimens for tensile test and metallurgical examination including microstruc-
Keywords:
ture observation and Vickers hardness measurement. The results showed that welding softened the heat affected
High performance steel zone of the RQT-S690 steel significantly but its effect on the TMCP-S385 steel was insignificant. Secondly, T-stub
Impact of welding joints were fabricated by using the RQT and TMCP steel plates and tested in tension until failure. The design plas-
Heat affected zone tic resistances of the joints are then compared with the prediction by EC3 equations. It is found that the EC3 equa-
Microstructure tions could produce unsafe prediction for the RQT joint. Based on the test results, a reduction factor is proposed to
Vickers hardness modify the EC3 equations to produce reasonable conservative prediction of the RQT joint design plastic
Welded T-stub joint resistance.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Due to its high strength, high stiffness and ductile properties, struc-
tural steel is one of the most popular construction materials employed
in structural engineering [1]. Recently, with the rapid advancement in
thermal and mechanical processing procedures for structural steel pro-
duction, the interest for using high strength steels (HSS) with minimum
yield strength N460 MPa has been increasing in the last decade [2]. In
fact, normal strength steels (NSS) with yield strength up to 460 MPa
have been commonly specified for applications in many structural de-
sign codes [3,4]. The most commonly referred performance characteris-
tics of structural steel are their mechanical and chemical properties,
metallurgical structures and weldability [2]. In many structural applica-
tions, traditional focus is the tensile performance (yield stress and ulti-
mate tensile stress) of the steel. However, after many disasters involving
fatigue loading, earthquake, serious fire and terrorist attacks, the em-
phasis of tensile performance characteristics became more frequently
questioned while other performance-improving properties under ex-
treme loading conditions such as deformability and post-fire perfor-
mance received more and more attentions.
Modern high performance structural steels are usually the product
of advanced heat treatment and characterized by increased strength
⁎ Corresponding author.
E-mail addresses: mszhao@ntu.edu.sg (M.S. Zhao), c.lee@adfa.edu.au (C.K. Lee),
ctcfung@ntu.edu.sg (T.C. Fung), SingPing.Chiew@SingaporeTech.edu.sg (S.P. Chiew).
and toughness together with low carbon equivalents [5]. Different
from high alloy steels, the weldability of heat treated steels is improved
significantly in the sense of reduced pre-heating requirements and re-
duced susceptibility to cold cracking [6]. As a result, higher design flex-
ibility and construction productivity can be achieved. For conventional
steel plates, it is suggested that if the width of the soft zone does not ex-
ceed 25% of the plate thickness, local softening would not necessarily
impair the global strength due to the constraints of the stronger weld
metal and unaffected base metal [7,8]. However, such assumption may
not be applied to high performance steels because their main micro-
structure constitutes, such as martensite and bainite, are not stable at
high temperatures [9]. There is a possibility that the enhanced mechan-
ical properties acquired by means of hardening may deteriorate signifi-
cantly after exposure to heat, due to microstructural changes at certain
critical temperatures [10,11]. This undesirable property was first con-
firmed for work hardened cold-formed hollow sections [12], and then
later for the high strength steel S460 and S690 [13] as well as the cold
rolled steel Bisplate 500 [14]. As a result, it is not surprising that con-
cerns are raised about the performance of welded high performance
steel connections, especially when high heat input welding is applied
[15].
The main objective of this paper is to experimentally investigate and
compare the influence of welding on the strength of two relatively new
high performance structural steels, namely the reheated, quenched and
tempered (RQT) high strength steel grade S690 (RQT-S690) and the
thermo-mechanically controlled processed (TMCP) normal strength

http://dx.doi.org/10.1016/j.jcsr.2016.12.023
0143-974X/© 2017 Elsevier Ltd. All rights reserved.
 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121
steel grade S385 (TMCP-S385). The whole study is divided into two
phases. In the first phase, a specially designed welding procedure is de-
veloped and applied to both the RQT690 and TMCP385 steel plates in
order to induce full welding heat effect in the plate's thickness direction.
These fully weld-affected plates are then fabricated into standard size
specimens for tensile test and metallurgical examination including mi-
crostructure observation and micro-hardness test. In the second
phase, the tensile performance of welded T-stub joints fabricated by
using the RQT-S690 and the TMCP-S385 plates are studied. The load-
displacement curves, failure modes, first yield resistance of the tested
specimens and plastic stage behaviour are investigated. By comparing
the first yield resistances obtained from the test and the design plastic
resistances predicted by the EC3 equations, the effects of welding on
the joint strength of the specimens are evaluated. Furthermore, based
on the test results, a new reduction factor which will take account for
the welding effect is proposed to provide reasonable conservative pre-
dictions of the plastic resistance of HSS RQT-S690 T-stub joints.
2. Phase 1: material property of fully weld-affected specimens
2.1. Base metal
In the first phase, 16 mm thick TMCP-S385 and RQT-S690 steel
plates with minimum yield strengths of 385 MPa and 690 MPa respec-
tively are examined. TMCP is an advanced thermo-mechanical process
to produce low carbon plate steels “microalloyed” with Ti, Nb and V.
The production concept of TMCP is to combine controlled hot rolling
with accelerated cooling to generate a bainite and ferrite dominating
microstructure [16]. By using the merits of grain refinement, precipita-
tion hardening and low amount of transformation hardening, the
TMCP enables the production of as-rolled steels (i.e. no offline transfer
is needed during the production process) with final properties that
are tailored to the requirements and specifications of the target applica-
tion [17]. In addition to strength, hardness and toughness, weldability
and corrosion resistance are usually made features of TMCP. The
TMCP-S385 steel plates tested in this study has minimum yield strength
of 385 MPa and tensile strength between 550 MPa and 670 MPa. The
RQT690 plates used in this study are manufactured by a refined
quenching and tempering technology. Similar to the conventional
quenched and tempered steels, RQT steels also depend on the extremely
fast cooling to produce martensite in the microstructure to enhance
their strengths [10]. However, RQT steel plates exhibit better homoge-
neity in through-thickness mechanical properties compared with tradi-
tional directly quenched and tempered steel plates. The RQT-S690 steel
plates studied in this paper comply with the EN 10025-6 grade S690
specification [18] and have minimum yield strength of 690 MPa and
tensile strength between 790 MPa and 930 MPa.
Fig. 1. Stress-strain curves of RQT-S690 and TMCP-S385 steels.
111
Table 1
Mechanical properties of the TMCP-S385 and RQT-S690 plates tested.
Strengths fy fu (MPa) E Elongation
(MPa) (GPa) (%)
RQT-S690 (16 mm) 745.2 837.8 208.9 14.5
EN 10025-6 S690Q/QL 690 770–940 – 14
(3 mm ≤ t ≤ 50 mm)
TMCP-S385 (16 mm) 443.3 568.0 208.4 37.8
EN 10025–4 S420 M/ML(t ≤ 16 mm) 420 520–680 – 19
t = thickness of plate.
The stress-strain curves and the mechanical properties of the steel
plates obtained by standard coupon tensile test are shown in Fig. 1
and Table 1, respectively. Table 1 also compares the properties of the
RQT and TMCP plates with the respective standards of EN 10025-6
[18] and EN 10025-4 [19]. From Table 1, it can be seen that RQT-S690
steel's yield strength is almost twice the yield strength of S355 but
both of its elongation at fracture and tensile ratio are lower when com-
pared with traditional normal strength steel and the TMCP-S385 steels.
Table 1 also indicates that the TMCP-S385 steel literally fulfilled the me-
chanical property specifications of the slightly higher grade S420M/ML
steel.
2.2. The fully weld-affected plate
Theoretically, the mechanical properties of material within the heat
affected zone (HAZ) can be assessed by direct examination of a very
small size sample from the welded joints. However, this method pre-
sents many difficulties in practice, such as the needs of delicate position-
ing and extraction of a very small size sample from the HAZ within a
narrow zone with high microstructure gradients. Hence, the material
properties within the HAZ zones are often assessed by performing ex-
periments on test samples that undergone simulated thermal treat-
ments closely resemble that experienced in the HAZ [20]. The main
idea of the HAZ material property test in this study is to manufacture
and examine the properties of plates that have been fully affected in
the thickness direction by the heat input during welding. In order to en-
sure that the HAZ is fully developed through the thickness of the speci-
mens, a special welding process was designed: Welding was first carried
out on both sides of the 800 mm × 300 mm × 16 mm specimen plates
(Fig. 2). The welding was carried out in the center area of the plate
along the longitudinal direction and covered a width of 120 mm long.
Shielded Metal Arc Welding (SMAW) and a single electrode LB-70L
were employed to finish the welding at a voltage of 26 V using a current
of 170A and welding speed of 2.2 mm/s. The equivalent heat input is
about 2.0 kJ/mm. Special caution was also paid to the welding sequence
to minimize the deformation and residual stress associated with uneven
heating and cooling. As shown in Fig. 2, every time when two passes
Fig. 2. Welding procedure for the fabrication of fully weld-affected plates.
112 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121
Fig. 3. The fully weld-affected plate.
were finished on one side of the plate, the welding was moved to the
other side. The fully weld-affected plate is shown in Fig. 3.
2.3. Mechanical property tests of the fully weld-affected coupon
The fully weld-affected plates were further cold-grinded (Fig. 4) to
remove the welding materials on both sides. They were eventually cut
and fabricated into standard coupon specimens (Fig. 5). It should be
noted that all grinding and cutting processes following the welding
are accompanied with water cooling to avoid additional heat input
into the specimens. For each type of steel plates, two coupon specimens
were tested by standard tensile test according to EN 10002-1 [21]. The
gauge length of the specimens was 100 mm, while the parallel length
was 120 mm and is fully covered by welding (Fig. 6). The loading rate
was set as 1 mm/min until fracture took place and data points were cap-
tured at a frequency of 1 Hz. During the tensile test, extensometer was
used to monitor and record the stress-strain curves and strain gauge
was employed to verify the results obtained by the extensometer.
Fig. 4. Grinding machine with water cooling system.
Fig. 5. Welding affected coupon specimens after (top) and before (middle and bottom)
grinding.
Fig. 6. Welding affected zone created on the final coupon specimens.
The tensile test results in terms of stress-strain curves of the fully
weld-affected RQT-S690 and TMCP-S385 specimens are shown in
Figs. 7 and 8, respectively. The nominal yield strengths (the 0.2% strain
offset strengths), tensile strengths, the tensile ratios as well as the
strains at fracture obtained are summarized in Table 2. Note that in
Figs. 7 and 8 and Table 2, data obtained from testing the original base
metal (RQT-S690 and TMCP-S385) are also plotted for comparison. It
can be seen from Figs. 7 and 8 and Table 2, the deformation capacities
of all the welded specimens in terms of elongation at fracture decreased
slightly when compared with the respective base metal. The TMCP-
S385 steel lost more deformation capacities than the RQT-S690 steel
but both the nominal and tensile strength were slightly increased. In
general, the effect of welding on the mechanical properties of the
TMCP-S385 specimens is similar to that of hardening effects, i.e. welding
improved the strengths at the expense of ductility. By comparing the
two TMC-S385 specimens (Welded-1 and Welded-2), it can be seen
that the stronger the hardening effects, the worse the ductility results.
On the other hand, besides the slight decrease in ductility, welding
caused serious deterioration in the strengths of the RQT-S690 speci-
mens with nearly 23% reduction of both yield and tensile strengths.
Fig. 7. Stress-strain curves of 16 mm RQT-S690 steel after welded.
Fig. 8. Stress-strain curves of 16 mm TMCP-S385 steel after welded.
 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121 113

Table 2
Mechanical Properties of RQT-S690 and TMCP-S385 after welding.

Material Measured yield strength (MPa) Diff 1 (%) Tensile strength (MPa) Diff 2 (%) Tensile ratio E (GPa) Fracture strain (%)

RQT-S690 Base metal 745.2 100 837.8 100 1.12 208.9 14.7
Welded-1 576.8 77.4 643.3 76.8 1.10 206.9 14.1
Welded-2 584.8 78.5 650.5 77.6 1.09 209.3 13.9
TMCP-S385 Base metal 443.3 100 568.0 100 1.28 208.4 37.8
Welded-1 476.6 107.5 598.4 105.4 1.31 211.0 34.1
Welded-2 481.1 108.5 611.5 107.7 1.27 207.2 32.8

Diff 1 = Yield strength of Welded-1 or 2 / base metal nominal yield strength × 100%.
Diff 2 = Tensile strength of Welded-1 or 2 / base metal tensile strength × 100%.

Table 3
Grinding and polishing methodologies.

Procedure Abrasive Lubricant/extender Force (lbs) Time (min:s) Platen speed (rpm) Head speed (rpm) Rotation
direction

1 Rigid grinding disc 75 μm diamond Water 6 Until Plane 60 150 NN

2 Rigid grinding disc 9 μm MetaDi supreme diamond MetaDi fluid 6 10:00 60 150 Nb
3 TriDent 3 μm MetaDi supreme diamond MetaDi fluid 6 05:00 60 150 NN
a
4 MicroCloth 0.05 μm MasterPrep alumina 6 02:00 60 150 Nb
a
Refers to washing away using water.

2.4. Metallurgical investigation 2.4.2. Microstructure study and hardness tests

Figs. 10 and 11 show the original microstructures of the RQT-S690
Both RQT-S690 and TMCP-S385 steels are heat treatment hardened and TMCP-S385 steels, respectively. It can be seen from Figs. 10 and
steels in which the strength improvement is due to non-equilibrium mi- 11 that, in accordance with metallurgical theory, the composition of
crostructure [22]. As a member of the QT steel family, RQT steel usually RQT-S690 steel is typical lath-shaped martensite, while the TMCP-
consists of meta-stable transformation microstructure of martensite or S385 steel consists of fine grain mixture of bainite and ferrite. The
bainite, which is tempered after hardening treatment at temperatures HAZs of RQT-S690 and TMCP-S385 steels are shown in Figs. 12 and 13,
below the transformation point A1 (727 °C, for carbon steel with carbon
content 0.15%) [10]. For TMCP steel, the mechanical properties are de-
veloped on the basis of fine grain microstructure which consists of
mainly bainite and ferrite [8]. Theoretically, if the hardened microstruc-
ture is exposed to temperature above A1 and cooled subsequently (as
occurred during welding) the original hardened microstructure would
be transformed into austenite and finally decomposes into microstruc-
ture with reduced strength, such as ferrite or upper bainite [23]. Many
researchers have shown that thermo-mechanically treated microstruc-
ture changes irreversibly if the microstructure is exposed to tempera-
ture above A1 and transformation softening would take place [8,23–
25]. Although tempering softening is less pronounced for TMCP steels
due to beneficial precipitation hardening effect and less transformation
hardening microstructure with compared with RQT steels, the soft zone
is equally inevitable during welding [8]. However, Figs. 7 and 8 and
Table 2 showed that only the strengths of the RQT-S690 was decreased
significantly after welding, while the strengths of TMCP-S385 steel were Fig. 9. Metallurgical examination samples.
even increased slightly. In order to further investigate the effect of
welding on the RQT-S690 and TMCP-S385 steels and search the “miss-
ing” soft zone of the TMCP-S385 steel, metallurgical analysis is carried
out.

2.4.1. Sample preparation

Metallurgical examinations were conducted for the cross sections of
the fully weld-affected coupons (Fig. 3). Two metallurgical specimens
for each type of steel were carefully fabricated by using the precision
cutter, grinder and polisher and finally etched with nital (nitric acid
and alcohol) solution. Table 3 summarized the grinding and polishing
methodologies used and the final samples are shown in Fig. 9. As
shown in Fig. 9, each sample contains two passes of welding and the fu-
sion zones are in circular shapes while the HAZ and base metal zones are
in strip shapes. The metallurgical examination performed consists of
two steps, namely (1) microstructure study by light optical microscopy;
and (2) Vickers hardness measurements according to ISO 6507-1 [26].
The main focus was on the alteration in microstructure and the variation
of hardness within the HAZ. Fig. 10. Original Microstructure of RQT-S690 steel.
114 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121
Fig. 11. Original Microstructure of TMCP-S385 steel.
respectively. It can be seen from Figs. 12 and 13 that the grain sizes of
the both HAZs are greatly increased, especially near the fusion zone. Al-
though the grain sizes gradually decrease with the distance from the fu-
sion zone, the sizes are still larger than those of the base metal (Figs. 10
and 11). Depending on the microstructural changes in the HAZ, normal-
ly three sub-HAZs can be defined [27]: (1) The coarse grained HAZ
(CGHAZ) which is the grain enlargement zone. During welding, temper-
ature in this zone ranges from A3 (about 830 °C for structural steel with
carbon content 0.15%) up to the melting point (about 1500 °C), (2) the
fine grained HAZ (FGHAZ) which is the recrystallized and partially
transformed zone (or inter-critical HAZ zone). This zone has been heat-
ed to a temperature higher than the transformation temperature A1 but
lower than the A3, (3) the tempering zone. The mechanical properties of
this zone are not much different from the base material since no phase
transformation takes place. In this study, at least two of the sub-HAZs,
namely the CGHAZ and FGHAZ, can be observed clearly in Figs. 12 and
13. The tempering zone is the furthest from the fusion zone and is not
within these two figures.
As shown in Figs. 12 and 13, after etching, a local coordinate system
can be defined and established at the boundaries between the fusion
zone and the HAZ. The direction along this curved boundary is defined
as the tangential direction, while that perpendicular to the tangential di-
rection is defined as the radial direction. The Vickers hardness tests
were then carried out along the radial direction to measure the hard-
ness variation from the fusion zone to the base metal. The distance be-
tween two adjacent measurement locations is 0.2 mm. The
measurement covered the area ranging from − 2 mm up to 7 mm
with 0 mm as the boundary between the fusion zone and the HAZ.
Fig. 12. Microstructure in the HAZ of RQT-S690 steel.
Fig. 13. Microstructure in the HAZ of TMCP-S385 steel.
The precision of the measurement positioning was guaranteed by the
digital x-y direction movement control stage mounted on the Vickers
hardness tester (Fig. 14). For all samples, a 500 g force was adopted in
the tests. Typical hardness test results obtained from RQT and TMCP
specimens are shown in Fig. 15. Fig. 15 shows that the hardness distri-
bution patterns of RQT-S690 and TMCP-S385 steels are similar. By
using the hardness of the base metal as an indicator, five distinguishable
zones, namely the fusion zone, the CGHAZ, the FGHAZ, the tempering
zone and the base metal can be seen from left to right. The hardness
in the fusion zones of both specimens is similar, since the same welding
electrode was used. For the RQT-S690 specimens, after a sudden drop of
the hardness near the boundary between the fusion and the HAZ (i.e.
distance = 0), the hardness rapidly rises to a peak in the CGHAZ and
then decreases to a minimum in the FGHAZ which is lower than the
hardness of both the fusion zone and the base metal. The hardness of
the tempering zone is similar to that of the base metal and thus it is dif-
ficult to differentiate this zone with the base metal.
For the TMCP-S385 specimens, the CGHAZ shows obviously higher
hardness with a maximum hardness of 28.0% higher than the base
metal. However, the FGHAZ shows lower hardness with a minimum
value of 15.8% lower than the base metal while the tempering zone
shows comparable hardness. It should also be noted that in this case,
the total hardness contributed by the CGHAZ (area under the hardness
curves) is higher than that contributed by the FGHAZ. As a result, the av-
erage hardness of the HAZ of TMCP-S385 ends up being slightly higher
than that of the base metal. Different from the TMCP-S385 steel, it is
found out that the hardness of the whole RQT-S690 HAZ is lower than
that of the base metal. For the two specimens tested, the highest hard-
ness values in the CGHAZ are 0.7% and 2.7% lower than the base metal,
while the lowest hardness measured in the FGHAZ is about 70% of the
hardness of the base metal. Besides, the hardness of the tempering
zone is also lower than that of the base metal. As a result, the overall
hardness of the RQT-S690 HAZ is much lower than that of the base
metal. It should be noted that these findings from the microstructure
studies and hardness test are consistent with the post-welding
strengths shown in Figs. 7 and 8, while the decrease in ductility after
welding is mainly contributed by the more brittle CGHAZ.
3. Phase 2: the influence of HAZ on the tensile performance of
welded T-stub joints
3.1. Specimen fabrication
In order to investigate the influence of HAZ on the tensile perfor-
mance of weld connections, four welded T-stub joints were fabricated
by using the RQT-S690 and TMCP-S385 plates. Each specimen is fabri-
cated by joining two identical steel plates with dimensions of
 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121 115

Fig. 14. The Vickers hardness tester used in this study.

440 × 150 × 16 mm. The joints are designed as complete penetration
butt weld joints according to the AWS structural steel welding code
[28]. Three bolt holes were drilled at each side of the chord plate in
order to fix the specimens to the test rig. The center-to-center distance
between two rows of bolt holes is 290 mm. The configuration of the
joints is shown in Fig. 16. SMAW is again employed to fabricate the
joints and the employed electrodes matches the corresponding
strengths of each material. However, the energy input is still kept at
2 kJ/mm so that the same welding heat input effect as in the fully
weld-affected plates experiment will be reproduced.
3.2. Tensile test set-up and testing procedure
Tensile tests for the T-stub joints were carried out in a servo-hydrau-
lic universal test machine which has a maximum loading capacity of
2000 kN. To fix the specimen into the test machine, tailor-made
“inverted” support joints made of 50 mm thick S355 steel plates were
manufactured. The configurations of the support joints are the same
as the tested joints (Fig. 16). The specimens are fixed into the support
joints by six M24 high strength hexagon bolts of grade 10.9HR. It should
be noted that in certain situations, the response of T-stub joints is influ-
enced by the type of bolt assembly. Selection of a suitable bolt type ac-
cording to the specific structural usage may avoid premature
reduction of joint strength for higher levels of joint rotation and provide
further rotational capacity [29]. In this study, the blots used were select-
ed to guarantee that they will work elastically and cause little influence
to the load-displacement relationship. This assumption was verified
after testing since no plastic deformation was found in the bolts after
the tests. The full testing set-up is shown in Fig. 17. In order to ensure
that the T-stub joints are loaded vertically, a spirit level was employed
to adjust the position of the specimens during mounting. In order to
capture the load-displacement relationship of the specimens precisely,
LVDT was employed to record the real-time vertical displacement at
the brace end. Since it would be easier to control the testing time,

Fig. 15. Hardness variations in the HAZ.

116 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121
Fig. 16. Configuration of the T-stub joints.
displacement control instead of force control was adopted during the
testing. The loading rate was set as 1 mm/min for all time so that
quasi-static response could be obtained.
3.3. T-stub joints test results
3.3.1. General descriptions
Fig. 18 presents the test results in terms of the load-displacement
curves of the 16 mm RQT-S690 and TMCP-S385 T-stub joints. Although
Fig. 17. The T-stub joint test setup.
Fig. 18. Load-displacement curves of the T-stub joints.
the specimens were failed at different loadings, these curves are of the
same pattern. In general, three different stages in the load-displacement
curves can be distinguished. They are (1) the elastic stage, (2) the plastic
hinge stage and (3) the failure stage [30]. In the elastic stage, the stiff-
ness and the elastic modulus govern the behaviours of the joints until
yielding takes place. When the specimens are further loaded, plastic de-
formation appears and four plastic hinges, two near the weld toe and
two near the bolted area, were formed (Fig. 19). In this stage, the defor-
mation grows more rapidly but the resistance of the joint increases
slowly. If the loads are further increased, the behaviour of the joint
would then enter the large deformation and failure stage. In this stage,
due to large deformation, the original configuration of the joint is
changed from T-shape to Y-shape with the strong presence of catenary
action. The performance of this stage is governed by both the plastic
hinge and the catenary action mechanism. These two mechanisms fur-
ther change the resistance of joint from flexural forces only to combined
flexural and axial forces. In this stage, due to strain hardening, the axial
catenary force could contribute more to the total load carrying capacity
than the plastic hinge moment resistance. As a result, strain concentra-
tion appeared near the plastic zone and the weld toe. Due to the lack of
sufficient ductility that is required to support the development of plastic
hinge and catenary mechanisms, the final failure modes usually happen
in the form of weld toe through thickness fracture (Fig. 20) or bolt hole
necking failure (net section failure under tension, Fig. 21) [30].
3.3.2. Design plastic resistance
For design purpose, the behaviour of the joints under the elastic
stage and the first yield resistance is of the most importance. To quanti-
tatively evaluate the effects of these two parameters, the elastic stage
and plastic hinge stage of the curves are taken out and simplified into
Fig. 19. Plastic hinges formation.
 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121 117

Fig. 20. Failure mode: weld toe through thickness fracture.

a straight line model (Fig. 22). The intersection of the two straight lines
or the turning point of the model is defined as the first yield resistance of
the joint, which is widely accepted as the design plastic resistance of the
joint before large deformation appears [31,32].
In EC3, to estimate the design plastic resistance of T-stub joints, three
failure modes, namely (1) complete yielding of the flange, (2) bolt fail-
ure with yielding of the flange and (3) bolt failure [32] are identified. In
this study, all the specimens were failed by complete yielding of flange
(i.e. plastic hinge formation). To predict the design plastic resistance of a
T-stub joint when it is failed in complete yielding of the flange, two
methods based on the yield line analysis are adopted by EC3 [32].
Method 1:
4Mpl;1;Rd
F¼
m results of the TMCP specimens are predicted conservatively. The aver-
Method 2:
ð8n−2ew ÞMpl;1;Rd
F¼
2mn−ew ðm þ nÞ
with the moment resistance of the section is defined as

2
t are conservative when predicting the design plastic resistance for the
Mpl;1;Rd ¼ leff fy
2 TMCP joints but its reliability may not be assured for the RQT-S690
In Eq. (3), leff is the effective width of the T-stub flange of the joint
which is equal to 150 mm in this study (Fig. 6.2 of [15]). t and fy are
the thickness and the yield strength. m, n are geometrical parameters
of the T-stub joints (Fig. 23). ew is either equal to 1/4 of the washer
Fig. 22. Definition of design plastic resistance of T-stub joint.
diameter or the width across points of the bolt head of nut [15]. Note
that in Method 2, instead of concentrated at the center line of the bolt,
it is assumed that the force applied to the T-stub flange by a bolt is dis-
tributed uniformly under the washer (or the bolt head/nut). Since the
distance between the center lines of the weld toe plastic hinge and
the bolt area plastic hinge is smaller than m, this assumption leads to
higher but more realistic resistance.
The design plastic resistance of the studied T-stub joints obtained by
Eqs. (1) and (2) and the actual moment resistance obtained from tests,
which were obtained by using the graphic method shown in Fig. 22, are
shown in Table 4. Compared with the EC3 equation predictions, the test
ð1Þ
age test result of the TMCP-S385 specimens is 23.0% and 14.6% higher
than Eqs. (1) and (2), respectively. However, the predicted results
from the EC3 equations for the RQT-S690 specimens are not so conser-
vative when comparing with the test results. The average test results of
ð2Þ
the RQT-S690 specimens is just 4% larger than Eq. (1) and is even −3.1%
lower than Eq. (2) even the actual yield strength (745.2 MPa) of the ma-
terial was already 8% higher than the nominal yield strength used in Eqs.
(1) and (2). Based on these test results, it appears that the EC3 equations
ð3Þ
joints. In fact, the results obtained from the joint tests are consistent
with the results obtained from the fully weld-affected coupon tests
(Figs. 7 and 8) and the Vickers hardness test results (Fig. 15). Therefore,
it can be concluded that while the influence of HAZ on the strength of
the tested TMCP-S385 joints is negligible, welding reduces the strength
of the tested RQT-S690 joints considerably in the HAZ and eventually
impacts on their design plastic resistance. Furthermore, as a higher

Fig. 21. Failure mode: necking failure at the bolt hole area. Fig. 23. Force diagram for design plastic resistance calculation.
118 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121

Table 4
Design plastic resistance of the tested T-stub joints.

Thickness (mm) Test results (kN) EC3 Eqs. (Kn) Nominal (Measured) Diff 3(%) Diff 4 (%)

Test1 Test2 Test average Eq. (1) Eq. (2)

RQT-S690 16 217.8 219.5 218.7 210.3 (227.1) 225.6 (243.6) 4.0 −3.1
TMCP-S385 16 146.5 142.0 144.3 117.3 (135.1) 125.9 (144.8) 23.0 14.6

For EC3 Eqs. results, the nominal series refer to results obtained by using the nominal yield strengths of fy = 690 MPa and 385 MPa for the RQT-S690 steel and the TMCP-S385 steel, re-
spectively; while the measured series refer to results obtained by using the measured yield strengths of fy = 745.2 MPa and 443.3 MPa for the RQT-S690 steel and the TMCP-S385 steel,
respectively.
Diff 3 = ((Test average − Eq. (1) − N) / Eq. (1)) × 100%.
Diff 4 = ((Test average − Eq. (2) − N) / Eq. (2)) × 100%.

energy input rate during welding will inevitable create a larger and where Mwt bh
pl ,1,Rd and Mpl,1,Rd are the moment resistances of plastic hinges
weaker HAZ near the connection, the predictions from the EC3 equa- at the weld toe and the bolt hole area, respectively. The main idea of Eqs.
tions will be less conservative (or even unsafe) if welding methods in- (4) and (5) is to employ Mwt pl,1,Rd to replace Mpl,1,Rd at the weld toe to ac-
volving higher energy input such as Submerged Arc Welding (SAW) count for the reduction the moment resistance due to welding effect.
using multiple electrodes performed by a welding robot is employed Assuming that the moment resistance of the plastic hinge at the bolt
to fabricate the joints. area is the same as Mpl,1,Rd (which is valid for most practice situations),
it is easy to deduce from Eqs. (4) and (5) that
4. Proposed reduction factor for design plastic resistance calculation
2Mwt bh
pl;1;Rd þ 2M pl;1;Rd Mwt
pl;1;Rd
4.1. Derivation of the reduction factor for T-stub joint χHAZ ¼ ¼ þ 0:5 ð6Þ
4Mpl;1;Rd 2Mpl;1;Rd

From the coupon and T-stub joint tensile tests, it is shown that
welding induces localized, large and transient heat input into the fusion
zone and alters the mechanical properties of RQT-S690 steel in the HAZ
significantly. Subsequently, the plastic resistances of T-stub joints made
of RQT-S690 steel are reduced. As a result, Eqs. (1) and (2) from EC3
may no longer always be conservative for predicting the design plastic
resistance of the T-stub joints, especially when welding methods involv-
ing high energy input are employed. Hence, in this section, it is pro-
posed that more reliable and conservative predictions could be
obtained by introducing a reduction factor χHAZ in Eqs. (1) and (2) to re-
duce the predicted design moment resistance. The proposed equation
forms for Methods 1 and 2 are given by
Method 1:
2Mwt bh
pl;1;Rd þ 2M pl;1;Rd 4ðχHAZ ÞM pl;1;Rd
F¼ ¼ ð4Þ
m m
Method 2:
Hence, the problem of determining the value of χHAZ is equivalent to
obtaining the value of Mwt pl, 1, Rd. Such approach is sensible as the reduc-
tion of the joint strength is due to the strength reduction near the
weld toe HAZ. Since the mechanical properties of the HAZ are signifi-
cantly altered (weakened) from the base metal, the original stress
state of the plastic hinge cross section of the steel plate (Fig. 24a) will
be affected. In general, the yield strength of the material inside the
HAZ, fy , HAZ(x) is not constant and it will vary as a complex function
with the distance from the Plastic Neutral Axis (PNA) x (Fig. 24b). In
order to obtain a simple expression of the reduction factor χHAZ for prac-
tical design calculation, as shown in Fig. 25, the complex variation of fy,-
HAZ(x) is idealized as an average constant value of fy,HAZ over the entire
depth of the HAZ, dHAZ. In general, two scenarios are possible. They are
corresponding to the case that dHAZ is either less than (Fig. 25a) or larger
than (Fig. 25b) the depth of the PNA.
From Fig. 25a, when dHAZ is exactly equal to the depth of the PNA, one
has d1 = 0 so based on the equilibrium of stress state

       
ð8n−2ew Þ 2M wt bh f y;HAZ ðdHAZ Þ ¼ f y ðd2 Þ ¼ f y ðt−dHAZ Þ ð7aÞ
pl;1;Rd þ 2M pl;1;Rd =4 ð8n−2ew ÞðχHAZ ÞM pl;1;Rd
F¼ ¼ ð5Þ
2mn−ew ðm þ nÞ 2mn−ew ðm þ nÞ

Fig. 24. Stress states of the plastic hinge at weld toe with and without HAZ.
 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121 119

Fig. 25. Simplified stress states of the plastic hinge at weld toe with constant fy,HAZ.

Eq. (7a) can be expressed as Similarly, for the stress state scenario shown in Fig. 25b, when dHAZ is
larger than the depth of the PNA, the following three equations can be
dHAZ obtained.
f y;HAZ =f y ¼ t− or εHAZ ¼ ð1−ρHAZ Þ=ρHAZ ð7bÞ
dHAZ
dHAZ þ d2 ¼ t ð13Þ
where
t=2
εHAZ ¼ f y;HAZ =f y ≤1 and ρHAZ ¼ dHAZ =t ≤1 ∫ −t=2 σdy ¼ f y;HAZ d3 − f y;HAZ ðdHAZ −d3 Þ−f y d2 ¼ 0 ð14aÞ

t=2
Furthermore, from Eq. (7b), the condition for dHAZ to be less than the Mwt
pl;1;Rd ¼ ∫ −t=2 σydy
"
 #
depth of the PNA (or d1 N 0 in Fig. 25a) is given by 2 2
d2 ðdHAZ −d3 Þ d
¼ leff f y d2 dHAZ −d3 þ þ f y;HAZ þ f y 3 ð14bÞ
ε HAZ bð1−ρHAZ Þ=ρHAZ ð8aÞ 2 2 2

Obviously, the condition Solving Eqs. (13)–(15) in a similar way will give the following ex-
pression for χHAZ
ε HAZ Nð1−ρHAZ Þ=ρHAZ ð8bÞ
εHAZ ρ2HAZ ð1−ρHAZ Þ2
implies that dHAZ is larger than the depth of the PNA (or dHAZ N d3 in Fig. χHAZ ¼ − −ρHAZ þ 1:5 ð15Þ
2 2εHAZ
25b).
For the stress state scenario shown in Fig. 25a, when dHAZ is less than
the depth of the PNA, the following geometrical relationship exists, In particular, if ρHAZ = 1, that is the whole weld toe thickness is fully
affected by welding, Eq. (15) will be reduced to
dHAZ þ d1 þ d2 ¼ t ð9Þ
ðεHAZ þ 1Þ
χHAZ ¼ ð16Þ
Form the equilibrium of stress state, one can write 2

t=2
∫ −t=2 σdy ¼ f y;HAZ dHAZ þ f y d1 − f y d2 ¼ 0 ð10aÞ From Eq. (16), if εHAZ = 0 (i.e. the weld toe reduces to a pin joint),
χHAZ = 1/2 which simply indicates that moment resistances are only
t=2 provided by the plastic hinge at the two bolt ends.
Mwt
pl;1;Rd ¼ ∫ −t=2 σydy
2  2  3
2
f y d2 þ d1

4 dHAZ 4.2. Range of the reduction factor χHAZ
¼ leff þ f y;HAZ dHAZ þ d1 5 ð10bÞ
2 2
As shown in Eqs. (12) and (15), the reduction factor χHAZ can be
expressed in terms of εHAZ and ρHAZ, only which are both the key param-
By using Eqs. (9) and (10a), one can solve the values of d1 and d2 as
eters that reflect the effects of welding induced on HAZ. In general, both
t t dHAZ and fy, HAZ will vary according to the material, plate thickness and
d1 ¼ ð1−ρHAZ ð1 þ εHAZ ÞÞ; d2 ¼ ð1−ρHAZ ð1−εHAZ ÞÞ ð11Þ the welding method adopted during the joint fabrication. It should be
2 2
noted that while the boundary of HAZ could be established approxi-
Substitute the values of d1 and d2 into Eq. (10b) and using Eqs. (3) mately by the microstructure and hardness test, it is almost impossible
and (6), the reduction factor can eventually be expressed as to determine its exact location since both the microstructure and hard-
ness in the tempering zone and the base metal are highly similar. How-
ρ2HAZ   ever, in general, dHAZ and fy,HAZ will respectively increase and decrease
χHAZ ¼ 1−ε2HAZ −ρHAZ ð1−εHAZ Þ þ 1 ð12Þ
2 as the heat input per unit length during welding increases. For example,
for the same plate thickness, the depth of the HAZ created by welding
Obviously, for the degenerated cases that εHAZ = 1 or ρHAZ = 0, using multi-pass SMAW will be smaller than that generated by high
χHAZ = 1 which is corresponding to the case of no HAZ. speed multiple-electrode single pass SAW performed by a welding
120 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121

machine. While fy,HAZ for multi-pass SMAW will be higher than the cor-
responding value for the single pass SAW.
In this study, since direct tensile tests have been conducted on the
fully weld-affected coupons, the average yield strength of the HAZ can
be estimated from the coupon test results by averaging the nominal
yield strengths shown in Table 2 so that.

576:8 þ 584:8 580:8

f y;HAZ ¼ ¼ 580:8MPa and ε HAZ ¼ ¼ 0:842 ð17aÞ
2 690

For the depth of the HAZ, since no microstructure study and hard-
ness tests was conducted on the T-stub joint specimens, a conservative
value of dHAZ = 8 mm (i.e. half of the thickness of the tested welding-af-
fected coupon specimen) could be used so that
Fig. 26. Low welding heat input case with constant depth of HAZ. Note: Weld passes
8 1−ρHAZ 1−0:5
ρHAZ ¼ ¼ 0:5 and εHAZ ¼ 0:84b ¼ ¼1 ð17bÞ shown in this figure are not the actual weld passes used in the T-joint specimens.
16 ρHAZ 0:5

Hence, Eq. (12) should be used to compute the value of χHAZ which
gives a value of 0.957. It should be mentioned that if other steel mate-
rials and other welding conditions are used, different values of fy,HAZ
and dHAZ obtained by tests similar to those described in Section 2 should
be adopted. The predicted design plastic resistances of the joint obtain-
ed by using χHAZ = 0.957 are shown in the second row of Table 5. It can
be seen that now both Eqs. (3) and (4) give conservative predictions.
In practice, it is difficult and expensive to employ the microstructure
study and hardness tests to examine individual joint for estimating the
values of ρHAZ and εHAZ. However, depends on the welding method
employed for the fabrication of the T-stub joint, the follow two scenar-
ios can be identified in practice.
4.2.1. Low welding heat input scenario (LWHI)
In this scenario, it is assumed that during the welding process, the
welding speed is low and relatively low heat input per unit length
(e.g. 2 kJ/mm or less) is applied near the weld toe of the T-stub joint.
For example, during the fabrication of the T-stub joints tested in this
study, manual, multi-pass SMAW using one electrode is employed.
Since multiple passes are used, only a relatively small amount of
welding material is deposited in each welding pass. Furthermore, direct
heating of the parent HSS steel plate is mainly due to those passes that
are in direct contact with the plate (Fig. 26). As a result, dHAZ will almost
be constant and independent of the thickness of the HSS plate. This also
implies that the parameter ρHAZ will decrease as the thickness of the HSS
plate increases. Furthermore, since the heat input per unit length is low
in this scenario, the strength ratio εHAZ will be relatively large. For RQT-
S690 HSS steel plate, it would be reasonable to assume that fy,HAZ should
be generally higher than 450 MPa [15] so that εHAZ could end up ranging
from 0.65 to 0.85 (e.g. εHAZ = 0.842 in the full weld affected specimens
tested in this study).
by a welding robot with multiple electrodes is employed during joint
fabrication, the heat input per unit length can reach 20 kJ/mm or
more. Since only a single or at most two passes are used, at least half
of the welding material is deposited in each welding pass. It is also likely
that all weld passes will directly heat the parent HSS steel plate. As a re-
sult, the thickness of the HAZ will increase as the thickness of the HSS
plate increases. In this case, it is more suitable to assume that the param-
eter ρHAZ will be a constant with respect to the thickness of the HSS
plate. Furthermore, since the heat input per unit length is high, the
ratio εHAZ will be relatively low when comparing with the low welding
heat input scenario. For RQT-S690 HSS steel plate, it would be reason-
able to assume that the value of fy,HAZ could be reduced to a level similar
to the “re-normalized” strength of approximately 400 MPa after extend-
ed post-weld heat treated [15] so that εHAZ could end up ranging from
0.5 to 0.6.
In order to investigate the influence of HAZ with respect to different
plate thickness, a small scale parametric study is carried out based on
the above two welding scenarios for T-stub joints with plate thickness
varying from 8 mm to 40 mm. For the LWHI scenario, three different
cases corresponding to εHAZ = 0.65 , 0.75 , 0.85 and a constant value
dHAZ of 8 mm are considered. For the HWHI scenario, two cases corre-
sponding to εHAZ = 0.5 and 0.6 with ρHAZ = 0.5 are considered. The var-
iations of the reduction factor χHAZ from t = 8 mm to 40 mm are plotted
in Fig. 27. Note that for the HWHI scenario, since the value of ρHAZ is
fixed at 0.5, χHAZ becomes a constant for a given value of εHAZ and Eq.
(12) was employed for all plate thickness t. For LWHI, Eq. (12) was
also employed to calculate χHAZ for most cases except for the cases of
t = 8 mm for all εHAZ values and t = 10 mm for εHAZ = 0.85 and 0.75

4.2.2. High welding heat input scenario (HWHI)

In this scenario, it is assumed that during the welding process, the
welding speed is high and relatively high heat input per unit length
(e.g. 4 kJ/mm or more) is applied near the weld toe of the T-stub joint.
For example, when single or double pass high speed SAW conducted

Table 5
Comparison between test results of RQT-S690 16 mm T-stub joints and modified
equations.

Test average (kN) χHAZ Eq. (3) (kN) Eq. (4) (kN) Diff 5 (%) Diff 6 (%)

218.7 0.957 201.3 215.9 8.6 1.3

0.842 177.2 190.0 23.4 15.1

For Eqs. (3) and (4), the nominal yield strengths of fy = 690 MPa and 385 MPa for the RQT-
S690 steel and the TMCP-S385 steel are used, respectively.
Diff 5 = ((Test average − Eq. (4)) / Eq. (5)) × 100%. Fig. 27. Variation of reduction factor χHAZ with different welding parameters (LWHI: Low
Diff 6 = ((Test average − Eq. (5)) / Eq. (5)) × 100%. welding heat input, HWHI: High welding heat input).
 M.S. Zhao et al. / Journal of Constructional Steel Research 131 (2017) 110–121
where Eq. (15) was used. Fig. 27 clearly shows that the HWHI scenario
always lead to a lower reduction factor than the LWHI scenario for all
plate thicknesses. Furthermore, the LWHI scenario curves indicate that
the effect of HAZ decreases as the thickness of the joint increases. For
the conservative HWHI scenario where ρHAZ = 0.5 and εHAZ = 0.5 so
that fy , HAZ = 345 MPa, which is equivalent to the total loss of quench
and tempering hardening effects of the RQT-S690 steel in the HAZ, a
value of χHAZ = 0.842 is obtained. Finally, the last row of Table 5 show
that if this reduction factor is adopted, the average test results will be re-
spectively 23.4% and 15.1% higher than predictions from Eqs. (3) and
(4). These values give similar level of conservative estimations pro-
duced by Eqs. (1) and (2) for the TMPC-S385 joints (the last row of
Table 4).
5. Conclusions
In this study, the influence of welding on the strength of the
reheated, quenched and tempered high strength steel in grade S690
(RQT-S690) and the thermo-mechanically controlled processed normal
strength steel in grade S385 (TMCP-S385) are investigated experimen-
tally in two phases.
Phase I examined the stress-strain behaviour of the fully weld-af-
fected coupons, the microstructure and hardness variation within the
heat affected zone (HAZ). It is found that welding may cause serious de-
terioration in both strength and ductility of the RQT-S690 steel, while
the strength of TMCP-S385 steel actually was increased slightly. Fur-
thermore, from the microstructure study, it is found that three sub-
HAZs can be distinguished from the fusion zone to the base metal. The
coarse grained HAZ shows relatively high hardness and large-size
grains; the fine grained HAZ shows relatively low hardness and small-
size grains; and the tempering zone has comparable hardness and
grain size to the base metal. It should be noted that average hardness
of the HAZ of RQT-S690 steel is much lower than the base metal while
that of the TMCP-S385 steel is comparable to the base metal.
Phase II investigated the tensile performance of welded T-stub joints
made of those two materials. The first yield resistances from the test re-
sults and the design plastic resistances obtained by EC3 equations are
compared. It is found that the EC3 equations give reasonable conserva-
tive prediction for the TMCP-S385 joints. For the RQT-S690 joints, de-
spite that the actual yield strength of the RQT-S690 steel is already 8%
higher than the nominal yield strength, the EC3 equations only provides
marginal overestimation. Such results reconfirm the conclusion from
the Phase I study that welding heat input weakened the material prop-
erties of RQT-S690 steel near the weld toe. To compensate such adverse
effect, a reduction factor is proposed to modify the EC3 equations. Fur-
thermore, a small scale parametric study was conducted to compute
values of the reduction factor corresponding to the high and low heat
input welding scenarios. The parametric study results shown that rea-
sonable conservative predictions for the design plastic resistance of
RQT-S690 joints could be obtained by combining the reduction factor
corresponding to the high heat input welding scenario with the EC3
equations.
Acknowledgement
This research is supported by the Singapore Ministry of National De-
velopment (MND) Research Fund on Sustainable Urban Living (Grant
No. SUL2013-4). Any opinions, findings and conclusions expressed in
this paper are those of the writers and do not necessarily reflect the
view of MND, Singapore.
121
References
[1] R. Bjorhovde, M. Engestrom, L. Griffis, L. Kloiber, J. Malley, Structural Steel Selection
Considerations: Reston (VA) and Chicago (IL): American Society of Civil Engineers
(ASCE) and American Institute of Steel Construction (AISC), 2001.
[2] R. Bjorhovde, Development and use of high performance steel, J. Constr. Steel Res.
60 (2004) 393–400.
[3] AISC, Specification for Structural Steel Buldings. ANSI/AISC 306–05, Ed, Americal In-
stitute of Steel Construction, Chicago, 2005.
[4] BSI, Eurocode 3: Design of Steel Structures: Part 1-1 General Rules and Rules for
Buildings, BS EN 1993-1-1, British Standard Institution, London, 2005.
[5] W. Yu, Y.-j. Qian, H.-b. Wu, Y.-h. Yang, Effect of heat treatment process on properties
of 1000 MPa ultra-high strength steel, J. Iron Steel Res. Int. 18 (2011) 64–69.
[6] Welding Handbook, Volume 1: Welding Technology, eighth ed. American Welding
Society, Miami, 1991.
[7] B. de Meester, The weldability of modern structural TMCP steels, ISIJ Int. 37 (1997)
537–551.
[8] F. Hochhauser, W. Ernst, R. Rauch, R. Vallant, N. Enzinger, Influence of the soft zone
on the strength of welded modern HSLA steels, Weld. World 56 (2012) 77–85.
[9] J.D. Verhoeven, Steel Metallurgy for the Non-metallurgist, ASM International, Ohio,
USA, 2007.
[10] H.K.D.H. Bhadeshia, R.W.K. Honeycombe, Steels: Microstructure and Properties, 3th
ed. Elsevier Science & Technology, Oxford, United Kingdom, 2006.
[11] A.V. Sverdlin, A.R. Ness, Chapter 3: Fundamental Concepts in Steel Heat Treatment
(Steel Heat Treatment Handbook: Metallurgy and Technology), Taylor & Francis
Group, 2007 121–164.
[12] J. Outinen, Mechanical properties of structural steel at elevated temperatures and
after cooling down, 10th International Conference on Fire and Materials 2007, 29–
31 January, 2007 (San Francisco, CA, United states).
[13] X. Qiang, F.S.K. Bijlaard, H. Kolstein, Post-fire mechanical properties of high strength
structural steels S460 and S690, Eng. Struct. 35 (2012) 1–10.
[14] J.C. Golondrino, G.A. MacRae, A. Abu, J.G. Chase, G.W. Rodgers, B. Fu, et al., Post-Fire
Hysteretic Behaviour of Asymmetrical Friction Connections, Presented at the Pro-
ceedings of the 10th Pacific Structural Steel Conference, Singapore Structural Steel
Society and National University of Singapore, Singapore, October 2013 8–11.
[15] M.S. Zhao, S.P. Chiew, C.K. Lee, Post weld heat treatment for high strength steel
welded connections, J. Constr. Steel Res. 122 (2016) 167–177.
[16] S. Machida, H. Kitada, H. Yajima, A. Kawamura, Extensive application of TMCP steel
plates to ship hulls: 40 kgf/mm2 class high yield stress steel, Mar. Struct. 1 (1988)
219–243.
[17] M. Militzer, S. Hashmi, G.F. Batalha, C.J. Van Tyne, B. Yilbas, 1.10 - thermomechanical
processed steels, Comprehensive Materials Processing, Elsevier, Oxford 2014,
pp. 191–216.
[18] BSI, Hot Rolled Products of Structural Steels: Part 6 Technical Delivery Conditions for
Flat Products of High Yield Strength Structural Steels in the Quenched and Tem-
pered Condition, BS EN 10025-6, British Standards Institution, London, 2004.
[19] BSI, BS EN 10025–4: Hot Rolled Products of Structural Steels: Part 4 Technical Deliv-
ery Conditions for Thermalmechanical Rolled Weldable Fine Grain Structural Steels,
British Standards Institution, London, 2004.
[20] D. Kaplan, G. Murry, Thermal, Metallurgical and Mechanical Phenomena in the Heat
Affected Zone, John Wiley & Sons, Kokoken, USA, 2008.
[21] BSI, BS EN 10002–1: Tensile Testing of Metallic Materials: part1 Method of Test at
Ambient Temperature, British Standards Institution, London, 2001.
[22] G.E. Linnert, Welding Metallurgy. Volume 1, Fundamentals, fourth ed. American
Welding Society, Miami, Florida, USA, 1994.
[23] J.M. Unfried, C.M. Garzon, J.E. Giraldo, Numerical and experimental analysis of mi-
crostructure evolution during arc welding in armor plate steels, J. Mater. Process.
Technol. 2009 (2009) 1688–1700.
[24] A. Takahashi, H. Ogawa, Influence of softened heat affected zone on stress oriented
hydrogen induced cracking of a high strength line pipe steel, ISIJ Int. 35 (1995).
[25] B.K. Choudhary, J. Christopher, D.P. Rao Palaparti, E. Isaac Samuel, M.D. Mathew, In-
fluence of temperature and post weld heat treatment on tensile stress–strain and
work hardening behaviour of modified 9Cr–1Mo steel, Mater. Des. 52 (2013) 58–66.
[26] I.S.O. BSI, 6507-1: Metallic Materials – Vickers Hardness Test – Part 1: Test Method,
British Standards Institution, London, 2005.
[27] Z. Boumerzoug, C. Derfouf, T. Baudin, Effect of welding on microstructure and me-
chanical properties of an industrial low carbon steel, Engineering 2 (2010) 502–506.
[28] AWS, Structural Welding Code, in Steel, American National Standards Institute,
Miami, 2008.
[29] M. D'Aniello, D. Cassiano, R. Landolfo, Monotonic and cyclic inelastic tensile re-
sponse of European preloadable GR10.9 bolt assemblies, J. Constr. Steel Res. 124
(2016) 77–90.
[30] M.S. Zhao, C.K. Lee, S.P. Chiew, Tensile behaviour of high performance structural
steel T-stub joints, J. Constr. Steel Res. 122 (2016) 316–325.
[31] A.M. Girão Coelho, F.S.K. Bijlaard, Experimental behaviour of high strength steel
end-plate connections, J. Constr. Steel Res. 63 (2007) 1228–1240.
[32] BSI, Eurocode 3: Design of Steel Structures: Part 1–8 Design of Joints, BS EN 1993-1-
8, British Standard Institution, London, 2005.