Materials and Design 91 (2016) 398410

Contents lists available at ScienceDirect

Materials and Design

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

Improving bonding quality of underwater friction stitch welds by

selecting appropriate plug material and welding parameters and
optimizing joint design
Xunda Zhang a, Caiyan Deng a, Dongpo Wang a, Zhijiang Wang a,, Jinhu Teng a, Jun Cao b, Wei Xu b, Fan Yang b
a
b

Tianjin Key Laboratory of Advanced Joining Technology, School of Materials Science and Engineering, Tianjin University, Tianjin 300072, China
Offshore Oil Engineering Co., Ltd., Tianjin 300452, China

a r t i c l e

i n f o

Article history:
Received 10 July 2015
Received in revised form 27 November 2015
Accepted 28 November 2015
Available online 30 November 2015
Keywords:
Plug material selection
Joint design
Forging force
Discontinuity
Hardness
Underwater friction stitch welding

a b s t r a c t
Friction stitch welding of S355 steel was conducted under wet conditions to study the bonding quality. The effects
of plug material, joint design, and welding variables on the weld defects, microstructural characteristics, hardness
levels, and tensile properties were investigated. The underwater stitch welding performed with a S355 steel plug
exhibited visible bonding defects at the weld interface, while the weld with LF2 as the plug material yielded a
qualied joint without cracks or discontinuities, provided the plug and hole geometries were well-designed.
The LF2-plug-stitch weld contained more upper bainite and fewer lath martensite grains, which was consistent
with the substantially decreased hardness values in the weld metal. The joint was designed with large initial contact area and gap between the plug and hole, so a higher amount of heat was generated when welding was
started. Further, the welding time was extended from 11 to 16 s, resulting in a larger heat affected zone and
the ow of the plasticized (or squeezed) material was improved too. Stitch welds realized with a 40 kN forging
force exhibited better ultimate tensile strength and elongation than those with a 35 kN forging force, when the
other conditions were kept constant.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Underwater damage in marine structures, undersea oil pipelines,
and platforms needs to be properly addressed, in order to extend the
service life of such infrastructure. The maintenance of such structures
usually requires localized underwater repair operations [1]. Although,
there are several accepted methods for underwater repair operations,
underwater welding is one of the most promising methods.
An increasing number of countries are getting involved in exploiting
deep-sea resources by utilizing marine oil/gas elds, and conditions
under which undersea structures and pipelines are serviced or assembled are becoming increasingly complex. In current underwater arc
welding processes, an unstable arc and bad shaping of the weld may
occur under wet conditions at over 40 m depth. Further, since underwater welding at depths of over 50 m would be even more difcult, to ensure the safety of divers, the welding operation is usually automated in a
dry cabin, and this results in high associated costs. Therefore, developing a new variant of underwater welding and repair technology that is
suitable for operations at underwater depths of over 50 m is of great
value.
Corresponding author.
E-mail address: wangzj@tju.edu.cn (Z. Wang).

http://dx.doi.org/10.1016/j.matdes.2015.11.114
0264-1275/ 2015 Elsevier Ltd. All rights reserved.

Some new technologies including friction based underwater

welding have been proposed to handle the operations over the depth
of 50 m [2,3]. Friction stitch welding, which is illustrated in Fig. 1, is an
innovative friction welding process. It is a practical solid-state repair
method for repairing long cracks in marine structures and undersea
oil pipelines. In this process, when repairing a long crack, a stitch weld
is fabricated by overlapping a number of friction taper plug welds at a
given distance along a pre-dened path [4]. Friction taper plug welding
(FTPW) is a combined drilling and lling process, which involves drilling a blind hole and then forcing a rotating tapered consumable plug
to ll the hole concentrically [57]. Friction stitch welding circumvents
the adverse effects associated with an increase in the ambient pressure
with underwater depth, and the process can be easily automated [8,9].
Therefore, this process can be considered to be a breakthrough in
deep water welding.
So far, several studies on FTPW, particularly those conducted in air,
have been reported: The possible bonding mechanisms, inuence of
process parameters on the bonding quality, and practical applications
of the method have been investigated. Pinheiro [10] focused on the metallurgical and mechanical properties of similar and dissimilar welds fabricated by FTPW. It was found that qualied joints could be produced in
air in a wide operating window. Meyer [11] found that as a result of sequential welding, the former welds were subjected to heat treatment

X. Zhang et al. / Materials and Design 91 (2016) 398410

399

Fig. 1. Schematic illustration of the friction stitch welding process.

during the subsequent welding processes, which reduced the maximum

hardness values when friction stitch welding was performed in air. In
addition, the welding results for S235 steel in air met the classication
society requirements for repair cracks in ship hulls [11]. Chludzinski
found that the fracture toughness of FTPW joint is much lower than
the base metal (BM) and is hardly inuenced by varying the axial
force in all his cases [12]. However, investigations on underwater friction stitch welding (or FTPW) have rarely been reported.
In our previous study, defects were observed in the bonding line at
the corner when friction stitch welding was performed in underwater
wet conditions, where S355 low-alloy steel was used both as plug and
BM [13]. The hardness values of the joints (up to 490 HV10) were too
high to nish the subsequent welding operations well. Drilling for the
next weld spot of stitch welding was difcult since (a) feeding the
drill into such a hard material was difcult, and (b) the uneven hardness
that the drill encountered on different sides made the drill bend towards the soft sides, thereby reducing the accuracy of the hole shape
and becoming a major cause of defects at the corner.
The objective of this work is to improve the bonding quality of the
friction stitch welds in underwater conditions. In Section 2, the factors
inuencing the bonding quality are analyzed in detail, and some feasible
solutions for improving the bonding quality are proposed. Experiments
designed to prove the solutions suggested in Section 2 are described in
Section 3, and the results are discussed in Section 4.

2. Factors inuencing bonding quality and proposed solutions

The bonding quality is mainly dependent on the proper selection of
materials, joint designs, and welding variables [14].
2.1. Plug material selection and stitch welding sequence
The underwater friction taper plug welds exhibited high hardness
values when S355 was used as both the plug and base materials, because the welds were quenched by the surrounding water. As mentioned previously, high hardness will worsen the weld quality. To
reduce the hardness of the welds, ASTM A350 LF2 steel (LF2 for
short), which has a relatively low hardness and good weldablity, was
employed as the plug material. To balance the uneven hardness on the
different sides of the drill, an optimized stitch welding sequence, in
which every other spot was welded, was selected.
2.2. Joint design
Meyer showed that the hole shape and plug geometry exert a major
inuence on the heat input from friction and material ow, which further inuence the ability to realize bonding all around the joint [11].
Xu found that the plug and hole congurations affect the quality of
the welding product through stress state and temperature distribution

Fig. 2. Friction stitch welding equipment and welding process in the underwater wet conditions: (a) friction stitch welding equipment, and (b) and (c) location of the workpiece in the vice
in the water tank.

400

X. Zhang et al. / Materials and Design 91 (2016) 398410

Fig. 3. The stitch welding sequence: (a) schematic diagram of the welding sequence and (b) friction stitch welding performed in sequence.

Table 1
Measured chemical compositions of the plug materials. All the values are in wt.%.
Material

Mn

Cr

Mo

Ni

Cu

Balance

S355
LF2

0.16
0.11

1.4
1.3

0.034
0.086

0.004
0.004

0.047
0.068

0.01
0.14

0.012
0.013

Fe
Fe

Table 2
Mechanical properties of the plug materials.
Material

HV10

Yield strength
(MPa)

Ultimate tensile strength

(MPa)

Elongation
(%)

S355
LF2

168.9
158.4

362
250

527
485

22
22

[15]. Therefore, the geometry of the plug and hole combination was
modied to increase heat generation and welding time. Improvements
in material ow were also considered.
2.3. Welding variables
The basic welding parameters of the friction stitch welding process
that signicantly inuence the joint properties are rotational speed,
axial force during welding, forging force (which is applied to consolidate
the weld when rotation stops), and burn-off distance (which represents

the consumption of the plug during welding). Unfried suggested that

the welding variables should be chosen such that they are consistent
with the plug diameters and equipment size [16]. According to the
work by Cui on FTPW in air, defects such as lack of bonding and incomplete lling are easily formed when welding is performed with a combination of low rotating speed and low welding force [17]. It was
concluded that a rotational speed of about 7000 rpm and welding
force ranging from 25 to 40 kN were necessary to obtain sound welds
in air. Futamata and Fuji revealed that the application of a forging
force signicantly improved the mechanical properties of the weld
[18]. An appropriate forging force should always be applied in order to
increase the ultimate tensile strength (UTS) and improve ductility
[11]. In the present study, the welding parameters chosen were similar
to those in the work by Cui, but different forging forces were selected to
perform the underwater friction stitch welding experiments. Different
forging forces were maintained for a predetermined time after the cessation of rotation, in order to study the inuence of the forging force on
the bonding quality and mechanical properties.
3. Experimental design
For extensive use in the marine and offshore industries, S355 lowalloy steel plates with dimensions of 200 mm 100 mm 40 mm
were used as the BM. The underwater friction stitch welding experiments were conducted using a hydraulically powered welding equipment developed by Tianjin University, which is shown in Fig. 2a. The
pre-machined base plate was rigidly clamped in a vice, which was
placed in the water tank as shown in Fig. 2b and 2c. A straight shank
twist drill was xed in the weld-head chuck to drill holes with a

Fig. 4. Microstructure of the base metal: (a) S355 steel and (b) LF2 steel.

X. Zhang et al. / Materials and Design 91 (2016) 398410

401

Fig. 5. Geometric parameters of the plug and blind hole for the experiments described in (a) Section 3.1 and (b) Section 3.2.

rotational speed of 400 rpm. After drilling, the twist drill was replaced
with a plug to nish the subsequent welding job.
The optimized stitch welding sequence mentioned above is illustrated in Fig. 3a. The nal stitch weld was fabricated by overlapping ve individual FTPW welds with the given distance of 12 mm. The rst three
welds were placed over a distance of 24 mm along a pre-dened path,
while the remaining two were placed right in the middle of the rst
three welds. This was done to ensure that the drill would not be
deected to one side. The friction stitch weld was carried out in the sequence shown in Fig. 3b, where the remaining sections of the rst three
plugs were machined off.
3.1. Experiments on material selection for the plug

reveals that the material consists of a ferrite matrix with pearlite strips.
On the other hand, LF2 steel (Fig. 4b) consists of ne equiaxed ferrite
and pearlite grains. The plugs were machined according to the geometry
shown in Fig. 5a, and the corresponding hole is shown in Fig. 5a. The
main welding parameters are listed in Table 3.
The as-welded specimens were cut into samples through the predened path mentioned above. The weld sections were examined macroscopically and microscopically using an OLYMPUS GX51 optical microscope. The sections were also inspected for bonding defects. The
microstructure of the weld sections was further examined using a
Hitachi S4800 scanning electron microscope (SEM). Hardness distributions of the different regions were examined using a 432SVD Vickers
hardness tester with a 10 kgf load and 15 s holding time.

To compare the effect of the plug materials on the weld quality, S355
steel and LF2 steel were selected as the plug materials. Their chemical
compositions are listed in Table 1, and the mechanical properties are
shown in Table 2. The microstructure of S355 steel shown in Fig. 4a

3.2. Experiment on joint design

Table 3
Welding parameters used for the experiments on material selection for the plug.

Table 4
Welding parameters used for the experiments on tapered joint design.

Rotational speed
(rpm)

Welding
force (kN)

Forging force
(kN)

Burn-off
distance (mm)

Forging time
(ms)

7000

30

35

16

8000

The tapered joint was modied as illustrated in Fig. 5b, with the goal
of achieving good bonding quality. The diameters of the at bottom and

Rotational speed
(rpm)

Welding
force (kN)

Forging force
(kN)

Burn-off
distance (mm)

Forging time
(ms)

7000

30

35

14

8000

402

X. Zhang et al. / Materials and Design 91 (2016) 398410

Table 5
Welding parameters used for the experiments on forging force.
Specimen no.

Rotational speed (rpm)

Welding force (kN)

Forging force (kN)

Burn-off distance (mm)

Forging time (ms)

A
B

7000
7000

30
30

35
40

14
14

8000
8000

rounded transient were enlarged, and the gap between the plug and
hole was signicantly increased. The depth of the blind hole was
decreased.
LF2 was used as the plug material. The welding parameters used in
the experiment are listed in Table 4. Examinations and analyses similar
to those described in Section 3.1 were performed.

3.3. Experiments on forging force

Different forging forces were applied in the experiments, as shown
in Table 5. The other welding parameters are also listed in Table 5. LF2
was used as the plug material and joints were prepared according to
the design shown in Fig. 5b.
The equipment used is designed with high rotational speed and low
axial force. The applied force is limited by the machine capability, if the
force is applied above 40 kN, it may do harm to the welding machine.
And an extremely high forging force is not benecial for former welds
in the stitch sequence, it might break the bonding of the former welds.
Besides, the forging force should be higher than the axial force. The increased forging force results in an abrupt increase of the burn-off rate
and the friction moment, and at the end of the friction welding process
(when rotation stops) oxide layers in the outer regions of the weld zone
produced by insufcient friction contact can be expelled by an increased
force [11]. Therefore, 35 kN and 40 kN were selected as the forging
forces, which are larger than the welding force and can be executed
by the machine.
Tensile tests were conducted using the CSS-44100 universal testing
machine with a 100 kN load cell and a crosshead speed of 3 mm/min.
The samples were machined to the geometry illustrated in Fig. 6, and
three samples of each specimen were tested.

4. Experimental results and discussion

4.1. Discussions on material selection for the plug
4.1.1. Macrographs and defects characterization
Fig. 7 presents the section morphologies of the friction stitch welds
with different plug materials. These underwater stitch welds were prepared with BM, the joint design, and welding parameters mentioned in
Section 3.1. In Fig. 7a, the section of the S355-steel-plug weld shows visible bonding defects at the weld interface and the defects are marked as
A and B. The optical micrographs of the defects A and B are shown in
Fig. 8a and 8b, respectively. In contrast to that a single FTPW process
in air produced sound welds as reported by Cui [17], the presence of
these defects indicates that the FTPW welding parameters optimized
in air with S355 steel as the plug material and with the plughole geometry shown in Fig. 5a, could not ensure good stitch weld quality in water.
Owing to the heat sink effect provided by the water environment removing a greater amount of heat from the weld region fast, friction
stitch welds formed underwater are more prone to root aws. The
quenching effect of water would cause the hardness of the weld to be
much higher than that of the base material, and a large gap in hardness
would lead to local stress concentration, especially along the bonding
line. Severe plastic deformation of the plug would also result in local
stress concentration. As the maximum tensile values of residual stress
occur close to the bonding line at the bottom as well as the sidewalls
of the hole [19,20], the defects will also occur at these same positions.
As shown in Fig. 7b, the LF2-steel-plug weld section exhibits a significantly improved stitch weld quality, although some lack of bonding can
still be observed at the rounded transient to the tapered sidewalls,
which are marked as C and D in Fig. 7b and shown in Fig. 8c and 8d,
respectively.

Fig. 6. Schematic illustration of the tensile samples.

X. Zhang et al. / Materials and Design 91 (2016) 398410

403

Fig. 7. Section morphologies of the friction stitch welds in underwater conditions and microstructural observation regions of the rounded transient: (a) S355-steel-plug weld and (b) LF2steel-plug weld.

The carbon equivalent (CE) parameter is most commonly used to determine weldability, and is directly related to the cold cracking and hardening capacity. The value of CE is calculated in our study using the
following equation, which is recommended by the American Welding
Society (AWS): CE = %C + (%Mn + %Si) / 6 + (%Cr + %Mo + %V) /

5 + (%Cu + %Ni) / 15. The AWS states that for materials with a CE
value above 0.40%, cracking could occur in the heat affected zone
(HAZ) of the welds. CE values between 0.36% and 0.40% can be dened
as indicating very good weldability [21]. Based on the measured chemical compositions of the plug materials (shown in Table 1), the calculated

Fig. 8. Morphological details of the bonding quality for the areas shown in Fig. 7: (a) and (b) are magnied images of the corresponding areas A and B in Fig. 7a, and (c) and (d) are magnied images of the corresponding areas C and D in Fig. 7b.

404

X. Zhang et al. / Materials and Design 91 (2016) 398410

Fig. 9. Typical zones, stitch welding sequence, and hardness test scheme for underwater friction stitch welding joints with different plug materials.

value of CE for S355 as the base metal is 0.408% and that for LF2 is 0.372%.
In addition, for high-strength low-alloy steels, a lower carbon content
will lead to a better weldability. Based on this and the calculated values
of CE, it may be concluded that LF2 steel has a better weldability than
S355, lower hardenability, and lower potential for cracking in the HAZ.
It should be noted that LF2 steel has a comparatively high Ni content.
Ni is an additive that simultaneously improves the UTS and preserves
toughness, besides decreasing the ductilebrittle transition temperature
and notch sensitivity. Additionally, higher Cr and V contents in the LF2
steel also enhance the mechanical properties as well as resistance to
stress-relief cracking by way of carbide precipitation strengthening
[22]. Owing to above reasons and after comprehensive comparisons,
LF2 steel is nally selected as the plug material to replace S355 steel,
and the bonding quality of underwater friction stitch welds is comparatively improved.

4.1.2. Hardness
From the sections shown in Fig. 7, several typical zones may be distinguished as illustrated in Fig. 9. Hardness tests were conducted on the
stitch welding sections and HV10 values were measured along ve test
lines with 1 mm distance between the indentations. Each stitch weld includes four main regions with different features, namely, the stitch weld
metal (WM), mechanical mixing zone (MZ), HAZ, and BM. The WM
consists of two major zones, namely, the forged zone and stir zone.
The upper parts of the plug material are forged by the applied forging
force at the end of each friction welding process, and this region is dened as the forged zone. Below the forged zone, the stir zone is located
at the bottom of the weld where rotational shearing material ow occurs under axial load. Owing to severe plastic deformation and local
melting by friction, mechanical mixing of the plug material and BM results in the formation of the MZ along the interface between the

Fig. 10. Hardness contour proles of underwater friction stitch welding joints: (a) S355-steel-plug weld and (b) LF2-steel-plug weld.

X. Zhang et al. / Materials and Design 91 (2016) 398410

405

Fig. 11. Typical microstructures of S355-steel-plug weld in different regions.

plasticized material of the plug and BM. Further, the HAZ can be divided
into primary HAZ and secondary HAZ.
The hardness contour proles of the two samples of the S355-steelplug weld and LF2-steel-plug are shown in Fig. 10a and 10b,
respectively. Hardness values were transformed to contours proles
by using the OriginPro 9 software. The hardness levels of LF2 are
much lower than those of S355, in general.
In the weld with the S355 steel plug, the maximum hardness value
of 493.7 HV10 is observed for the third weld, which is at a distance of
2 mm below the surface. The hardness values in the WM are signicantly higher than those in the BM (Table 2). Except for the hardness values
in the area around the central line in the second weld, the majority of
the hardness values are in the range of 370493 HV10 in the WM for
the S355-steel-plug weld and 290393 HV10 for the LF2-steel-plug
weld. Both of the stitch welds exhibit heterogeneous hardness distributions, especially in the upper and middle regions. The hardness values
measured in the region around the central line in the second weld are
substantially reduced, because of the reheating treatment experienced
during subsequent welds. This indicates that homogenization of the former welds occurred during the following welding operations.
The hardness values of the primary HAZ of the S355-steel-plug weld
range from 227 HV10 to 246 HV10, while the corresponding values for
the LF2-steel-plug weld range from 295 HV10 to 327 HV10. In addition,
the hardness values of the secondary HAZ are higher than those of the

primary HAZ, but lower than those of the WM. The secondary HAZ hardness values of S355 and LF2 are in the range of 262 to 332 HV10 and 242
to 312 HV10, respectively.
As there have been no specialized criteria for friction stitch welding,
AWS D3.6 (Specication for Underwater Welding approved by
American National Standard) was used as a reference to determine
whether the hardness levels in our study are acceptable. According to
AWS D3.6 the maximum hardness for 485 MPa of tensile strength is limited to 375 HV10. It is important to note that the maximum hardness
value of nearly 493 HV10 in the WM for the S355-steel-plug weld is
too high according to AWS D3.6. As evident from Fig. 10b, apart from
a few areas in which the hardness values are greater than 375 HV10,
most of the hardness values for the LF2-steel-plug weld are acceptable.
Comparison with the hardness results reported by Cui in air [17] indicates that welding in water caused the hardness values to increase
signicantly, owing to the rapid cooling rate. Hardness values up to
493.7 HV10 are observed in the WM of the S355-steel-plug weld, rendering the subsequent operations difcult and leading to a poorer
weld quality. While the hardness values in the WM of the LF2-steelplug weld are substantially reduced, the hardness values around the
bonding line, however, are high, which could lead to high local stress
concentration in this area and nally lead to defects at the corner.
Although the combination of the plughole geometry in Fig. 5a and
the LF2 steel plug results in a signicantly improved quality than

406

X. Zhang et al. / Materials and Design 91 (2016) 398410

Fig. 12. Typical microstructures observed by optical microscope (ab) and typical SEM morphologies (cd) of the LF2-steel-plug weld.

when the S355 steel plug is used, some root defects are still found. The
results suggest that the bonding quality may be improved if a low-alloy
steel plug material with a lower hardenability and better weldability is
selected. Since the local stress concentration is considered to be the
main reason for defects, the plastic ow characteristics corresponding
to the combination of plug and hole geometries will be considered in
Section 4.2 with the goal of eliminating defects.
4.1.3. Microstructures
The typical microstructures of the stitch weld with S355 steel plug,
as shown in Fig. 11, were observed using an optical microscope. The microstructure of BM is shown in Fig. 11g, which is similar as the microstructure in Section 3.1. The observations reveal different grain
structures, mainly due to the microstructural changes induced by the
heat treatment of the subsequent welds. The last weld (the fth weld
in Fig. 3) contains coarse grains of lath martensite in the WM
(Fig. 11c) as a result of the high cooling rates in underwater conditions,
which is consistent with the corresponding hardness values (370 HV10
to 493 HV10). The second weld (Fig. 11b) contains tempered martensite

in the WM, since the second weld undergoes two reheating processes,
i.e., during the fourth and fth welding steps. Tempered features can
also be observed in the rst weld (Fig. 11a).
The mixing zone along the bonding interface in Fig. 11d shows a
band of ne equiaxed ferrite grains with a few pearlite grains. This
ne-grained area located in the gap between the plug and hole was
lled with a plasticized layer of the shear-off plug material.
In the primary HAZ in the bottom area (Fig. 11h), ne granular bainite grains are dominant with a few martensite islands. Fig. 11e shows
the microstructure of secondary HAZ. As the area around the weld interface was heated above Ac1 [23] and underwent rapid cooling in the underwater conditions, the secondary HAZ transformed into the acicular
ferrite microstructure, which results in a signicant decrease in
hardness.
Fig. 11f shows the microstructure of the mechanically sheared material in the stir zone adjacent to the weld interface.
The microstructures of the stitch weld formed with the LF2 steel
plug were observed and compared with those of the S355 stitch weld.
The overall microstructure of WM is presented in Fig. 12b. The WM

Fig. 13. Section morphologies of the friction stitch welds formed with the geometry in Fig. 5b and regions to be observed at the rounded transient.

X. Zhang et al. / Materials and Design 91 (2016) 398410

407

Fig. 14. Morphological details at locations A and B shown in Fig. 13.

consists of a volume of upper bainite and a small amount of lath martensite grains. Magnied SEM micrographs of this region are presented
in Fig. 12c and 12d. In these images, some ne and parallel martensite
laths are found grouped side by side in one packet. The observed packets
appear to be oriented in different variations. This typical lathy structure
is the low-carbon martensite. Fig. 12d shows the typical SEM morphology of the upper bainite. A number of bainite laths nucleated at the prior
austenite boundary and grew parallel within the coarse prior austenite
grain. Adjacent to the prior austenite grains, several lath bainite sheaves
are observed with different crystal orientations.
The MZ, as shown in Fig. 12a, exhibits signicant metallurgical bonding characteristics as some acicular ferrite nucleated along the internal
wall of the hole and grew towards the WM. Consequently, the initial interface becomes indistinguishable.
Compared to the welds where S355 is used as the plug material, in
the LF2 welds, more number of upper bainite and fewer martensitic
structures are observed in the WM, and more metallurgical bonding
characteristics are observed along the bonding line, which results in
lower hardness and better bonding quality.

In the welding experiments described in Section 3.2, the welding

times are longer than those in the experiments described in
Section 3.1. The welding time for the experiments discussed in this section is recorded to be 16 s from initial contact to the end of relative rotation, which is about 5 s longer than the welding times in the former
experiments. The section morphology for the weld formed according
to the geometry design in Fig. 5b is shown in Fig. 13. Fig. 14a and b present the morphological details at locations A and B indicated in Fig. 13,
where cracks may form easily. No cracks or discontinuities are observed
in Fig. 14, suggesting that the modied joint design (Fig. 5b) is advantageous over the design shown in Fig. 5a in terms of bonding quality.
In the joint design shown in Fig. 5b, the initial contact area is enlarged and the gap between the plug and hole increases. The enlarged
contact area generates more heat and increases the welding duration,
resulting in a larger HAZ and lower hardness values around the bonding
line. The larger gap between the plug and hole promotes the ow of
plasticized (or squeezed) material, which is expected to further improve the bonding quality. In summary, based on the results described
in Sections 4.1 and 4.2.1, with proper selection of the plug material
and improved joint design, sound underwater friction stitch welds can
be obtained.

4.2. Discussions on joint design

4.2.1. Macrographs and defect characterization
In order to produce high-quality stitch welds in water without apparent discontinuities along the bonding line, a modied geometry
with increased diameters for the at bottom as well as the rounded
transient (as shown in Fig. 5), is chosen. Here, LF2 is used as the plug
material.

4.2.2. Hardness
Hardness tests were conducted on a section of the stitch weld and
HV10 values were examined along ve test lines with 1 mm distance between the indentations, as illustrated in Fig. 15. The hardness prole obtained is shown in Fig. 16. In the WM, except for the area around the
central line, the hardness values mainly range from 350 to 404 HV10,
with the maximum hardness value of 404.9 HV10 locates at a distance

Fig. 15. Different zones, stitch welding sequence, and hardness testing scheme for underwater friction stitch welding joints formed with the geometric parameters shown in Fig. 5b.

408

X. Zhang et al. / Materials and Design 91 (2016) 398410

Fig. 16. Hardness contour prole of underwater friction stitch weld formed with the geometric parameters shown in Fig. 5b.

of 1.5 mm from the surface. The hardness values of the primary HAZ
range from 208 to 241 HV10, while those for the secondary HAZ are in
the range of 268325 HV10. Comparison with Fig. 10b shows that the
hardness values in the WM increase slightly although it is still acceptable, while the hardness values in the HAZ decrease.
4.2.3. Microstructures
The microstructures examined by optical microscopy are summarized in Fig. 17. The microstructure of BM is shown in Fig. 17g. From

Fig. 17c, it is evident that the WM of the last weld mainly consists of
upper bainite with a small amount of lath martensite and it is similar
as the microstructure of the fourth weld (Fig. 17a). On the other hand,
owing to the heat treatment that occurred during subsequent welding,
the microstructure in the WM of the second weld consists of tempered
martensite with cementite nely dispersed in the ferrite matrix, as
shown in Fig. 17b. Compared to the WM shown in Fig. 11ac, the
amount of martensite decreases, as a result of which the hardness also
decreases, as shown in Fig. 16.

Fig. 17. Typical microstructures in different regions of the underwater stitch weld formed with the geometric parameters shown in Fig. 5b.

X. Zhang et al. / Materials and Design 91 (2016) 398410

409

cannot reach the values of the BM. Welds subjected to a forging force
of 40 kN show better elongation than those with a forging force of
35 kN. Similar results were also obtained by Ambroziak [14], suggesting
that an appropriate forging force should be applied to increase the UTS.
According to the hardness measurements described above, hardening and strengthening behaviors commonly occur in the WM or HAZ
during the underwater welding process. The large difference in hardness between the WM and BM can cause high levels of local stress concentration at the bonding line. Under tensile stress, yield and plastic
deformations are more prone to occur at the bonding line, rather than
in the WM or HAZ.
5. Conclusions

Fig. 18. Tensile properties of the welds subjected to different forging forces.

In Fig. 17d, a thin layer of the recrystallized ferrite matrix containing

pearlite strips is detected in the MZ. In addition, a large volume of acicular ferrite grains formed and grew towards the WM side, indicating
that sufcient metallurgical bonding occurred along the bonding line.
The microstructure of the primary HAZ (Fig. 17h) is similar to that in
the primary HAZ of the S355-steel-plug weld in previous experiments,
except that there are more number of martensitic structures. The secondary HAZ (in Fig. 17e) has a signicant amount of polygonal ferrite
formed on the austenite grain boundaries during cooling [24]. Further,
the width of the primary HAZ is signicantly larger than those of the
stitch welds in the former experiment, particularly in the bottom area.
In the work by Perrett and Sketchley [25], a layer of mechanically
sheared material was found adjacent to the weld interface. They also
found that mixing had occurred between the plug and BM, yielding a
range of microstructures. As can be seen from the microstructure of
the stir zone (Fig. 17f), mixing occurred in several areas and this region
was subjected to more severe deformation compared to the stir zone
described in Section 4.1.3 (Fig. 11f). As heat is generated owing to the
friction between the plug and substrate, this region is subjected to
higher process energies.
Geometry modications did contribute to changes in the microstructure, owing to changes in the heat input and welding times. According to the work by Pinheiro [10], the welding time not only
determines the microstructure at the interface, but also affects the
depth of heating in the BM as well as the width of the HAZ. Longer
welding times generate higher peak temperatures, which can be associated with the heat input into the joint [10]. Therefore, the microstructures presented in Fig. 17 are formed.

4.3. Discussions on forging force

Fig. 18 shows the UTS and elongation of tensile specimens obtained
from the experiments on forging force (Table 5). The three samples subjected to a forging force of 35 kN are named A1, A2, and A3, whereas
three other samples subjected to a forging force of 40 kN are named
B1, B2, and B3. The UTS of sample sequence A, which are subjected to
the forging force of 35 kN, are in the range of 412 to 451 MPa, while
those of sample sequence B, which are subjected to the forging force
of 40 kN, are in the range of 435 to 465 MPa. The mean value of UTS
for sample sequence A is 430 MPa with an error of 9 MPa, whereas
that for sample sequence B is 451 MPa with an error of 7 MPa. In
other words, the mean value of UTS for sample sequence A is indeed
lower than that for sample sequence B. Additionally, the UTS of the
specimens is generally lower than that of BM (in Table 2), and all the
samples failed at the bonding line. The elongations of the samples also

In this paper, the bonding quality of the friction stitch weld formed
in wet conditions was examined. The effects of plug material, joint design, and welding variables on the weld defects, microstructural characteristics, hardness levels, and tensile properties have been investigated.
The following conclusions can be drawn:
(1) Underwater friction stitch welds can be fabricated without cracks
or discontinuities using LF2 as the plug material and with an optimized geometry design, with a rotational speed of 7000 rpm,
burn-off distance of 14 mm, and welding force of 30 kN. Welds
with a forging force of 40 kN show better UTS and elongation
than those with a forging force of 35 kN.
(2) By using a plug material with better weldability and lower
hardenability such as LF2, the underwater friction stitch weld
quality can be improved. A lower carbon content/carbon equivalent in the plug material will reduce the amount of martensite,
besides substantially decreasing the hardness values in the WM.
(3) The geometries of the plug and hole should be designed properly
to achieve higher heat generation, longer welding duration, and
improved ow of the plasticized (or squeezed) material. The enlarged initial contact area and increased gap between the plug
and hole result in a better joint.

Acknowledgments
This research is supported by the National 863 High-Tech Project of
China (2011AA090302).
References
[1] L. Stephen, L.O. David, S. Charles, et al., Underwater Welding of Marine Structures,
International Workshop, NEW Orleans, Louisiana, USA, 1994.
[2] J.Q. Zhang, Y.F. Shen, X. Yao, et al., Investigation on Dissimilar Underwater Friction
Stir Lap Welding of 6061-T6 Aluminum Alloy to Pure Copper, Mater. Des. 64
(2014) 7480.
[3] Y. Zhao, Z.P. Lu, K. Yan, et al., Microstructural Characterizations and Mechanical
Properties in Underwater Friction Stir Welding of Aluminum and Magnesium Dissimilar Alloys, Mater. Des. 65 (2015) 675681.
[4] R.E. Andrews, J.S. Mitchell, Underwater Repair by Friction Stitch Welding, Met.
Mater.-Mar. Technol. 796-797 (1990).
[5] E.D. Nicholas, Friction Hydro Pillar Processing. 11th Annual North American
Welding Research Conference, 111995 79.
[6] D. Pauly, Process Development on an Experimental Friction Hydro Pillar Processing
System, GKSS, International report, 1999.
[7] W. Thomas, D. Nicholas, The Need for Gas Shielding Positive Advantages for Two
Friction Processes, TWI Bull. 84-88 (1997).
[8] F.W.T. Yeh, P.H.C. da Cunha, C.R.L. Lessa, et al., Evaluation of Discontinuities in A36
Steel Repairs with Friction Hydro Pillar Processing Using Different Axial Forces, ISIJ
Int. 53 (2013) 22692271.
[9] Y.Y. Yin, X.Q. Yang, L. Cui, et al., Material Flow Inuence on the Weld Formation and
Mechanical Performance in Underwater Friction Taper Plug Welds for Pipeline Steel,
Mater. Des. 88 (2015) 990998.
[10] G.A. Pinheiro, Local Reinforcement of Magnesium Components by Friction Processing: Determination of Bonding Mechanisms and Assessment of Joint Properties,
GKSS, Forschungszentrum Geesthacht GmbH, Geesthacht, 2008.
[11] A. Meyer, Friction Hydro Pillar Processing-Bonding Mechanism and Properties,
Fakultt fr Maschinenbau und Eletrotechnik der Technischen Universitt CaroloWilhelmina zu Braunschweig, 2003.

410

X. Zhang et al. / Materials and Design 91 (2016) 398410

[12] M. Chludzinski, M.P. Paes, F.L. Bastian, et al., Fracture Toughness of Friction HydroPillar Processing Welding in CMn Steel, Mater. Des. 33 (2012) 340344.
[13] J. He, Hardening and Crack Mechanism of Friction Stitch Welded Joints for DH36.
Master Dissertation, Tianjin University, 2014 (in Chinese).
[14] A. Ambroziak, B. Gul, Investigations of Underwater FHPP for Welding Steel Overlap
Joints, Arch. Civil Mech. Eng. 7 (2007) 6776.
[15] Y.C. Xu, H.Y. Jing, Y.D. Han, et al., Numerical Simulation of the Effects of Various Stud
and Hole Congurations on Friction Hydro-Pillar Processing, IJMS. 90 (2015) 4452.
[16] S.J. Unfried, M.T.P. Paes, T.F.C. Hermenegildo, et al., Study of Microstructural Evolution of Friction Taper Plug Welded Joints of CMn Steels, Sci. Tech. Welding Joining.
ens. 15 (6) (2010) 506513.
[17] L. Cui, X.Q. Yang, D.P. Wang, et al., Experimental Study of Friction Taper Plug
Welding for Low Alloy Structure Steel: Welding Process, Defects, Microstructures
and Mechanical Properties, Mater. Des. 62 (2014) 271281.
[18] M. Futamata, A. Fuji, Friction Welding of Titanium and SUS 304 L Austenitic Stainless
Steel, Weld. Int. 4 (10) (1990) 768774.
[19] D.G. Hattingh, D.L.H. Bulbring, A. Els-Botes, et al., Process Parameter Inuence on
Performance of Friction Taper Stud Welds in AISI 4140 Steel, Mater. Des. 32
(2011) 34213430.

[20] D.G. Hattingh, A. Steuwer, M.N. James, et al., Residual Stresses in Overlapping Friction Taper Stud Welds. Proceedings of MECA-SENS 5, fth international conference
on mechanical stress evaluation by neutrons and synchrotron radiation, Mito, Japan,
1012 November, 2009, Mater. Sci. Forum 652 (2010) 111115.
[21] SA-6/SA-6M Specication for General Requirements for Rolled Structural Steel
Bars, Plates, Shapes, and Sheet Piling. ASME BPVC Section II: ASME. 2001
[22] J.G. Nawrocki, J.N. Dupont, C.V. Robino, et al., The Stress-relief Cracking Susceptibility
of a New Ferritic Steel Part 1: Single-pass Heat Affected Zone Simulations, Weld. J.
79 (2000) 355362.
[23] D.G. Hattingh, C.V. Zyl, Temperature Distribution for a Friction Taper Stud Weld in
Thick Wall 10CrMo910 Steel, R&D J. South Afr. Inst. Mech. Eng. 28 (2012) 3745.
[24] G. Thewlis, Classication and Quantication of Microstructures in Steels, Mater. Sci.
Technol. 20 (2004) 143160.
[25] J.C. Perrett, P.D. Sketchley, Hole Repair in Steel Plate Using Friction Taper Plug
Welding Initial Trials in Air and Underwater, TWI 15015-3 (2007).