Technical Paper
doi:10.3723/ut.28.115 International Journal of the Society for Underwater Technology, Vol 28, No 3, pp 115–127, 2009
Underwater wet welding made simple: benefits of
Hammerhead r wet-spot welding process
David J Keats
Speciality Welds Ltd, Cleckheaton, West Yorkshire, UK
Abstract
A new method of wet welding was investigated to
evaluate potential improvements in weld quality, ease
of use, increased welding speed and the elimination
of welding skill. The new welding process, which
has been called Hammerhead ‘wet-spot’ welding,
eliminates the need for skilled welder-divers as well
as traditional cleaning and preparation techniques
normally associated with conventional manual metal
arc (MMA) wet welding. In addition, the process also
allows welding to be conducted in nil visibility, yet
remains a MMA process, using a specially designed
Fe–Cr–Ni–Mo electrode.
The process utilises a control device, which must
be pre-set before the diver enters the water. Through
this device, weld parameters are controlled and quality
is maintained, thus the role of the diver is simplified
to three steps: make contact with the material, strike
the arc and maintain pressure to the electrode while
welding.
A series of spot welds were produced both wet and
dry on 8.0mm carbon steel plates. The welds were
evaluated with regard to ease of use and setting up
of the device, speed and final weld quality. Initially, the
performance of the process was assessed and usage
diagrams produced. Work regarding an automated
version of the system has also been proposed.
Keywords: Hammerhead, welding, wet-spot, underwater, subsea, welder-diver
1. Introduction
Underwater wet welding has been employed for
many years now, but has commercially been restricted to conventional manual metal arc (MMA)
welding techniques (Keats, 2004, 1990; Association
of Offshore Diving Contractors, 1985; American
Welding Society, 1999; Hibshman and Jensen,
1933; British Standards Institute et al., 2002).
The typical problems associated with under water
MMA welding fall into two categories: mechanical/metallurgical quality, and skill and ability. It
was with both these issues in mind that a new
methodology of MMA welding was devised.
Underwater wet welding, accepted as a low cost,
practical alternative to dry or hyperbaric welding,
can, however, suffer from quality issues mainly due
to the rapid cooling (Keats, 2004, 1990; Association
of Offshore Diving Contractors, 1985; American
Welding Society, 1999; Hibshman and Jensen, 1933;
Gretskii and Maksimov, 1998; Bailey, 1987, 1991;
Sadowski, 1980; Gooch, 1983a,b; Masubuchi, 1981;
Abson and Cooper, 1998). It is also well appreciated
that the skills and abilities necessary to execute high
quality, conventional MMA wet welds are extremely
high, therefore labour and training costs are significant factors (Kralj et al., 2003; Gooch, 1983a,b;
Grubbs, 1986). This new welding methodology,
which has been developed by the author, provides a
solution to both of these issues. The process is called
Hammerhead wet-spot welding.
This process provides an alternative approach
to welding, in which the role of the operator is
minimised and is therefore no longer required to
have hand-eye coordination skills. Rather, for this
method, two materials are joined together by a spot
or plug weld by a programmable control device.
In this way, the operator simply becomes a means
of making contact with the material and providing
momentum to push the electrode into the material
once the arc is struck. The process also eliminates
the need for traditional cleaning, joint preparations
and chipping of weld slag. It utilises one electrode
to produce each spot weld, which is localised
through the thickness dimensions of the material.
The author has shown that the final mechanical
weld qualities have been improved, as has the
overall speed of joining when compared to conventional wet welding techniques. Unlike conventional
MMA welding, the process provides a method
of controlling the welding current necessary to
produce a weld, without requiring the operator to
have any welding skills or knowledge, because the
current is automatically regulated and controlled by
the device on each weld cycle. Thus, the role of the
diver is reduced to simply that of an operator.
2. Experimentation and results
2.1. Design of apparatus
The Hammerhead MMA wet-spot welding method
employs an electronic control device to control a
115
Keats. Underwater wet welding made simple: benefits of Hammerhead r wet-spot welding process
Fig 1: The control panel with isolation switch,
amp/volt meters and the Hammerhead control
device (bottom left)
number of key welding functions to produce a spot
weld. These functions and features are:
•
•
•
•
•
•
•
•
Main on/off switch
First peak (high) current control
Second background (low) current control
Timer (up to 20s)
High, low and auto current selector
Amp and volt meter
400-amp dual pole isolation switch
110v power supply and remote control function
cables.
2.2. Control functions
The control device, which is housed within a utility
case, consists of an on/off switch to power the unit,
high/low/auto current control potentiometers, a
timer, and amp and volt meters (see Fig 1). To
ensure a suitable, safe current is available, the
device is fitted with a transformer to adapt a 110-volt
supply down to a more suitable and safe 9 volts,
which is then rectified to direct current (DC). A
reed switch is fitted to trigger a relay, which starts
a timer when the arc is first struck. Two current
control potentiometers independently control high
and low current settings.
Once these have been set, the device can be
switched into ‘auto’ mode. These potentiometers
are adjusted to deliver the appropriate current
to penetrate and fill the materials and thus
produce the spot weld. Once the timer has
been triggered (following arc initiation), the high
current potentiometer delivers the preset current
for the set time period. Expiration of this control
then triggers the low current potentiometer to act,
thereby initiating the required low-level current
automatically. This low current function continues
until the arc is broken, after which the device
automatically resets and is ready to make the next
116
spot weld, although a 5-second delay prevents the
system resetting, should the diver accidentally break
the arc.
LEDs light up against each function so the
operator can monitor the process at any given
moment. All welding parameters are set prior to
the diver entering the water and involve the device
being connected to the welding machine via remote
control and 110-volt power supply cables. Once
connected to the device, complete manipulation
of the welding machine is provided. Amp and
volt meters are fitted to provide a visual display
of the welding current/voltage, as is a 400-amp
safety switch to isolate the current/voltage to the
diver, which is required under Health and Safety
Executive (HSE) regulations.
The setup of the device is quite straightforward.
Prior to entering the water, the diver selects a
suitable ‘high’ current (determined by eye) to allow
for adequate penetration of the two materials to
be joined. This high current time is recorded in
seconds, and penetration is measured visually by
examining the back of the material for a heat mark,
or blister. Providing this is visible on the outside
surface of the back-face, penetration is adequate
and the timer control and high current function
are programmed and set, the operator now selects
the ‘low’ current control. This function does not
require the use of the timer and is set simply to
provide a suitable current to complete the weld.
The device may now be set to automatic and is
fully programmed to produce welds automatically.
The diver can now enter the water and make any
small adjustments as might be necessary for the
given water type and working depth. After these
adjustments are made, the device may be relied
upon to give consistent and reproducible welding
parameters, as programmed, for each and every
weld. The device may also be set to manual mode,
in which the diver can request either ‘high’ or ‘low’
only current values to be selected.
When under water, it is essential that the
operator does not over penetrate the base materials. Should this occur, weld properties would be
compromised by the effects of water pressure, extinguishing the arc and causing slag entrapment, lack
of fusion and/or cracks. As the only opportunity for
burst-through is while the ‘high’ current cycle is in
operation, the timer controls this critical function.
Excessive penetration is a combined function of
both high current and arc time. By accurately
controlling both functions, penetration control is
achieved. It is not possible for the diver to burst
through the material while the ‘low’ current cycle
is functioning, as the current is too low. Thus,
the device reduces the role of the diver to that of
Vol 28, No 3, 2009
Table 3: Composition of Hammerhead electrodes
Table 1: Composition of steel plates
Element
C
Si
Mn
P
S
Element
Cr
Ni
C
Mo
Mn
Si
Nb
% (max.)
0.2
0.55
1.6
0.035
0.035
Min.
Max.
21
24
11
14
0.020
0.10
3
5
0.60
2
0.70
2
—
—
Table 2: Carbon equivalent value of steel
CEV
0.35
simply ‘pushing’ the electrode into the materials
and ensuring that contact is maintained.
In operation, this requires no more than 5–10kgf
and, provided the operator consistently maintains
this force, poor visibility conditions will in no
way affect the outcome or quality of the weld
produced. The applied force was estimated, based
on experimentation, and became a basis for
calculating the necessary pressure to be applied,
using a 3.2mm electrode.
Pressure N/mm2 or (MPa) =
1kg = 9.80665 Newtons
Force (kgf)
Area (mm2 )
(1)
• A 400-amp Gen-Set diesel welding generator
• Piranha welding monitor/safety switch, fitted
with the Hammerhead control system
• Underwater welding stinger
• Welding leads (50mm2 copper) double-insulated
• Brass parallel closing earth clamp.
The diving equipment used was standard commercial surface demand, i.e. the diver being fed
with an air supply through an umbilical rather than
a scuba bottle. Full radio communications were also
in place throughout, enabling welding data to be
supplied and recorded. Welding was conducted in
a freshwater dive tank, and working depth was 3m.
The environmental conditions recorded during
these experiments were as follows: air, −3◦ C (±1◦ )
and water, 0◦ C (±1◦ ).
Although the core wire of the electrode measured 3.2mm, the outer flux coating also needs to be
taken into account, which increases the diameter to
approximately 6.0mm. Therefore, an applied force
of 5–10kgf by the operator will ensure a pressure at
the tip of the electrode of some 1.73–3.49N/mm2
(MPa). The actual applied force each diver used
whilst welding was clearly onerous. A ‘best estimate’
was made by each individual, but the applied force
was based on the above calculation.
For much of the welding operation, the electrode
tip is deep within the wall thickness of the material,
so no arc is visible, thereby minimising the diver’s
influence on weld quality.
2.5. Electrodes
The 3.2mm electrodes were used, having the
chemical composition shown in Table 3. The
electrode was designed specifically to allow for
high dilutions while maintaining very short-arc
conditions. These electrodes have the potential to
allow for dilutions up to a maximum of 38% without
the risk of martensite formation. To evaluate this
potential fully, the Shaeffler diagram was used
(Fig 2) to plot a dilution line based on the mean
values, as shown in Equation 3.
2.3. Weld samples
All welding was performed on plate that was 150 ×
150 × 8.0mm. Material was restricted to low carbon
steel having the composition as shown in Table 1,
with CEV (IIW) as shown in Table 2. CEV formula
calculated as:
Mn (Cr + Mo + V) (Ni + Cu)
C+
+
+
(2)
6
5
15
The following weld IDs were assigned for each
test plate: dry spot welds – D-1, D-2, D-3 and D-4;
wet spot welds – W-1, W-2, W-3 and W-4.
All spot welds were conducted on simple lap
joints, with one plate overlapping the other
by ∼50%.
The Hammerhead electrode provides for the
following Cr and Ni equivalents, based on the
following formula:
2.4. Equipment, facilities and environment
All welding was conducted onsite in open air
conditions, utilising the following equipment and
facilities:
CrEq = %Cr + Mo + (1.5 × Si) + (0.5 × Nb)
NiEq = %Ni + (30 × C) + (0.5 × Mn)
(3)
CrEq = 22.5%Cr + (3.6) + (1.5 × 1.1)
+ (0.5 × Nb) = 27.75
NiEq = 12.7%Ni + (30 × 0.045)
+ (0.5 × 0.8) = 14.45
(4)
As the Shaeffler diagram shows, the use of this
electrode provides for a maximum dilution of
38.4%, without risking the formation of martensite
in the body of the weld. It is accepted when
using MMA welding, a dilution of around 25%
can be expected. Under water, this is normally
slightly reduced, due to ambient temperature and
rapid cooling, to approximately 20% (Keats, 2004,
1990; Gooch, 1983a,b; Masubuchi, 1981; Abson and
Cooper, 1998).
117
30
28
26
Austenit
24
0%
F
5% err
it
..
Nickel-Aquivalent = %Ni + 30x%C + 0,5x%Mn + 30x%N
Keats. Underwater wet welding made simple: benefits of Hammerhead r wet-spot welding process
22
%
10
20
18
%
tion
Dilu .4%)
8
Max
m (3
n
5
45.3
A+F
A+M
16
14
12
Line
Tie- m
118m
10
8
%
40
%
80
Martensite
100%
A+M+F
6
4
F
2 +
M
0
0 2
20
Ferrit
M+F
4
6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
..
Chrom-Aquivalent = %Cr + 1,4x%Mo + 1,5x%Si + 0,5x%Nb + 2x%Ti
Fig 2: Maximum dilution possible without risk of martensite forming is 38.4%
2.6. Welders and operators
Four individuals were engaged to carry out welding
and were identified as follows:
• Welder A: skilled welder – conducted welds
and W1
• Welder B: non-welder – conducted welds
and W2
• Welder C: skilled welder – conducted welds
and W3
• Welder D: non-welder – conducted welds
and W4
D1
D2
D3
D4
Each diver was asked to produce one dry and one
wet weld.
2.7. Welding procedures
To ensure accurate data collection, all welding
operations were recorded. Applied arc energy was
calculated by use of the standard formula given in
Equation 5:
I × V (total power in watts)
Arc energy = ∗
(5)
ROL (mm)/time in seconds
where I is current, V is volts and ∗ ROL equals
run out length of the electrode (in this case ROL
referred to the burn-off rate of the electrode).
No specification exists for wet-spot welding,
thus in order to determine how many welds may
be required to bear a given load, the formula
shown below was used. The size of any given weld,
and therefore the number of welds required, is
based on the required shear stress exerted on the
component. Thus, each single spot weld can offer
the following theoretical strength properties.
πd2
× shear strength
4
(neglecting any bending moment)
Max load =
118
(6)
In calculating the shear strength for plain carbon
steel, it is common industrial practice to assume this
to be ∼80% of the ultimate tensile strength. The
Hammerhead electrode offers a tensile strength
of 650MPa (all-weld ‘dry’ test); based on this
assumption, shear strength becomes 520MPa.
2
Thus, load carrying area (mm2 ) is π 4d where d is
the spot diameter (mm), so for a 10.0mm spot weld,
2
= 78.54mm2 . Max design shear
the area is π 100
4
stress for a 10.0mm spot weld is therefore 40.84kN.
The number of 10mm diameter welds required can
be calculated using Equation 7:
total shear load (kN)
(7)
40.84
Alternatively, the shear stress can be calculated
per mm2 of weld. This would produce the total spot
weld area required and thus lead to a selection of
spot welds.
N =
Shear Stress (X) kN mm2 =
40.84 (kN)
78.54 (mm2 )
(8)
Thus shear stress equals 0.52kN/mm2 . The
total spot weld area required for a load of 45kN
is therefore 86.54mm2 . The actual test results
obtained for all wet and dry spot welds are shown
in Fig 3. The results presented in Fig 4 show the
desired joint strength against a specific number
and/or size of individual spot welds, based on the
calculations discussed.
2.8. Spot welds
The welding parameters and techniques for all
welds were pre-set and recorded as follows:
• Amps: primary value 250–260, secondary value
150–160
• Timer: 5–6s for peak (primary) current value
Vol 28, No 3, 2009
HAMMERHEAD SINGLE WET & DRY SPOT-WELDS
60
50
LOAD kN
40
30
20
10
0
63.2
65.62
107.16
Actual WET welds
152.7
97.8
66.62
AREA mm2
Theoretical average WET weld
78.07
Actual DRY welds
87.92
112.36
86.24
Theoretical average DRY weld
Fig 3: Results of the actual wet and dry spot welds produced
Weld Strength Vs Number of Spot Welds
350.00
294.05
Desired Joint Strength (kN)
300.00
250.00
235.24
204.20
200.00
8mm spot weld
10mm spot weld
176.43
12mm spot weld
163.36
150.00
130.69
117.62
122.52
104.55
100.00
81.68
78.41
58.81
50.00
52.28
40.84
3. Results
26.14
0.00
1
2
3
4
Number of Spot Welds
5
Fig 4: Number of welds required against desired
weld strength, based on the calculations
discussed
•
•
•
•
•
•
•
•
Welder B was permitted a short practice period
for familiarisation. The welding parameters and
techniques for Welder B were exactly the same as
for Welder A. At the time of under water welding,
visibility was very poor at <25cm.
Welder C was asked to produce his welds after
a brief introduction of the technique. At the time
of under water welding, visibility was very poor
at <25cm.
Welder D was asked to produce his welds
completely unaided and without any opportunity to
practice, in a bid to demonstrate the feasibility of
a no-skill process. At the time of welding, visibility
was completely nil and all welding was carried out
by touch.
Volts: 25–35
Polarity: DC-Ve electrode
Electrode angle: 90◦ , ±10◦
Pressure applied: constant 5–10kgf
Material thicknesses: 2 × 8.0mm plates
Electrode: 3.2mm Hammerhead
Position: flat
Weld time: 25–27s
Prior to welding, plates were simply clamped
together to prevent relative movement. No cleaning
or other joint preparation was used. Welding was
conducted on plates in as-delivered condition. At
the time of welding for Welder A, visibility was
moderate at approximately 30–45cm.
3.1. Visual appearance
The overall quality of welds produced for both wet
and dry was surprisingly similar, especially when
one considers the visibility factor. Equally, there
was no substantial difference between welds made
by the skilled welders over non-welders. All welds
showed adequate fusion between base materials and
weld metal. Although not completely defect free,
some wet spot welds did show evidence of minor
gas voids/slag inclusions in the weld body. However,
none of the recorded defects appeared to make a
significant impact on the overall results of welds
made wet as compared to welds made dry. No
appreciable defects were observed by the naked eye
for any dry spot welds.
Welds generally had an overall convex circular
appearance, but a clear difference did exist between
wet and dry. Wet spot welds had a somewhat untidy
appearance and failed to blend in well with the top
plate surface, unlike the dry welds. This appearance
was due to the existence of a more restricted
weld puddle. Also, as the operator was discouraged
119
Keats. Underwater wet welding made simple: benefits of Hammerhead r wet-spot welding process
Fig 5: Macro-photograph for welds D1 (left) and W1 (right) conducted by Welder A
S405490
Fig 6: Macro-photograph for welds D2 (left) and W2 (right) conducted by Welder B
from manipulating the electrode. It was possible,
however, to manipulate the electrode for the dry
spot welds during the final stages of welding, which
did assist in working/wetting out the weld puddle.
This manipulation produced a smoother, more
blended appearance, and as a result, dry welds did
not show the excess ‘flash’ material (which was
evident in all wet spot welds). This flash was due
to excess material being ejected from within the
molten nugget, resulting from additional electrode
weld metal, causing still, molten metal to be ejected
as a result of continued pressure applied to the
electrode. Although untidy in appearance, this
flash material was easily removed with a simple
hammer blow.
One common feature for both wet and dry spot
welds was the heat mark, or blister, formed on
the back-face of the base material. This provided
a very useful indicator as to the success of
penetration. Although not accurate in terms of
measurement or depth, it did provide an excellent
visual method of establishing whether adequate
penetration had occurred. Where no heat mark
was present, the depth of penetration into the back
material was limited.
The overall diameter of the welds produced in
air (measured across the top outside diameter of
120
the weld) was somewhat larger than welds produced
wet, with the average diameter for a dry weld being
21.48mm against 14.39mm for that of a wet weld.
Wet welds on average were nearly 50% smaller in
diameter (49.27%) compared to dry welds made
under similar current/voltage conditions. This
increase in diameter appeared to be mainly due
to operator manipulation of the electrode, despite
being requested not to do so, and may also be due in
part to the input energy vaporising the water. This
can be seen from studying the weld shapes more
closely in the macro-photographs shown in Figs 5–8.
3.2. Transverse tension shear tests
In order to establish the load required to failure,
both wet and dry spot welds were subjected to
transverse shear tensile tests. Tables 4 and 5 show
the individual test results for wet and dry spot welds.
The average failure load of each weld type was
45.63kN for dry spot welds and 39.95kN for wet
spot welds. A difference of 5.68kN between wet and
dry was found. Thus, the average dry spot weld
offered a 14.2% strength improvement over the
average wet spot weld. The average cross-sectional
area (CSA) of weld nuggets for all welds was
86.24mm2 for dry spot welds and 97.17mm2 for wet
spot welds.
Vol 28, No 3, 2009
Fig 7: Macro-photograph for welds D3 (left) and W3 (right) conducted by Welder C
Fig 8: Macro-photograph for welds D4 (left) and W4 (right) conducted by Welder D
Table 4: Tensile test results for dry spot welds
Welder – weld No.
CSA of weld (mm2 )
UTS (MPa)
Failure load (kN)
Welder A – D1
Welder B – D2
Welder C – D3
Welder D – D4
66.62
78.07
87.92
112.36
717
587
594
326
47.8
45.9
52.2
36.6
Average
86.24
548.5
45.63
Table 5: Tensile test results for wet spot welds
Welder – weld No.
CSA of weld (mm2 )
UTS (MPa)
Failure load (kN)
Welder A – W1
Welder B – W2
Welder C – W3
Welder D – W4
63.20
65.62
107.16
152.7
668
607
379
244
42.2
39.8
40.6
37.2
Average
97.17
474.5
39.95
Wet welds, therefore, showed an increase of
10.93mm2 , thus increasing the CSA of deposited
weld metal by 12.67% (12.7) even though the actual
diameters were far smaller than any of the dry welds.
By factoring in this percentage change in the CSA of
wet welds to match the CSA of dry welds, a new load
required to failure of 34.88kN (34.9) may be calculated. This difference of 10.75kN further reduces
the wet strength results, as compared to the dry, by
23.55% (23.6) See Appendix 5 for more details.
Clearly, the effects of rapid cooling on welds
made under water should have caused a change in
the mechanical strength of the weld, due to the
faster cooling rates. To better understand these
results, hardness surveys and weld macros/micros
were also examined. Unfortunately, however, these
particular tests were carried out after shear testing
and thus may have obscured any minor defects.
It was also noted that the dry spot welds had
larger weld reinforcement, which accounted for
the initial observation that the CSA of dry spot
welds were actually larger, although this is unlikely
to have offered any real advantages in terms of
failure strength. The major influence in effective
joining was adequate penetration of the nugget into
the base materials, rather than the size of weld
reinforcement. It should also be appreciated that
the visibility conditions for making the wet spot
121
Keats. Underwater wet welding made simple: benefits of Hammerhead r wet-spot welding process
Table 6a: Hardness surveys for dry spot welds
using Vickers method at HV-10
Parent
HAZ
Weld
Welder A – D1
Traverse 1 (top)
Traverse 2 (bottom)
117
123
183
182
272
282
Welder B – D2
Traverse 1 (top)
Traverse 2 (bottom)
149
138
168
172
237
458
Welder C – D3
Traverse 1 (top)
Traverse 2 (bottom)
154
148
197
176
188
189
Welder D – D4
Traverse 1 (top)
Traverse 2 (bottom)
125
122
165
173
215
401
Combined average
135
177
280
Table 7a: Hardness surveys for wet spot welds
using Vickers method at HV-10
Parent
HAZ
Weld
Welder A – W1
Traverse 1 (top)
Traverse 2 (bottom)
124
122
208
248
134
217
Welder B – W2
Traverse 1 (top)
Traverse 2 (bottom)
139
164
223
224
276
284
Welder C – W3
Traverse 1 (top)
Traverse 2 (bottom)
147
147
308
220
274
281
Welder D – W4
Traverse 1 (top)
Traverse 2 (bottom)
116
120
199
181
132
262
Combined average
135
226
233
welds – especially for Welders B and C – was poor
and completely nil for Welder D.
3.3. Hardness survey
A number of hardness surveys were carried out
in accordance with BS EN 1043-1(1996) with two
traverse lines being used and six indentations for
parent metal, six for HAZ and three for weld-metal
per traverse line for welds D2 and W3. Table 6a
shows the average results for all dry spot welds,
whilst Table 7a shows the average results for all
wet spot welds. An additional hardness survey was
undertaken on welds D2 and W3 running down the
weld centre from top to bottom. Its purpose was
to better understand the microstructural changes
taking place within the weld body at different
intervals from the interface to the weld cap (see
Tables 6b and 7b).
When considering the combined average values
for Tables 6a and 7a, it was seen that wet results
was similar to the dry welds. Somewhat surprisingly,
122
Table 6b: Individual hardness survey of
dry weld (D2); weld-metal only, running from
top weld centre to weld bottom
Distance from weld
cap (mm)
Hardness Hv1kg
0.3
1.0
2.0
4.0
6.0
8.0
9.0
10.0
11.0
12.0
13.0
312
208
205
373
205
364
454
482
449
468
472
Table 7b: Individual hardness survey of
wet weld (W3); weld-metal only, running from
top weld centre to weld bottom
Distance from weld
cap (mm)
Hardness Hv1kg
0.3
1.0
2.0
3.0
4.0
5.0
6.0
9.0
10.0
11.0
11.7
210
206
194
196
261
419
283
215
386
370
378
however, was that the wet welds produced lower
hardness values than the dry welds. This is contrary
to what might be expected with conventional
under water welds cooling more rapidly, thus
resulting in harder weld and heat-affected zone
(HAZ) metal (Keats, 2004, 1990; Gretskii and
Maksimov, 1998; Gooch, 1983a,b; Masubuchi, 1981;
Abson and Cooper, 1998; West et al., 1990).
In the case of the dry spot welds (Table 6a), this
appeared due to two anomalously high readings
in Traverse 2 on welds D2 and D4 and was
assumed to be the result of increased dilution
whilst operating on the ‘high’ current setting. This
situation produced a hotter, more fluid puddle,
thereby diluting more carbon from the plate into
the weld pool. This, combined with the switch over
from ‘high’ to ‘low’ current, effectively limited any
further alloying, which together with the effects of
plate cooling caused the formation of martensite.
This is evident from the results shown in Table 6b
for weld D2 (weld metal only).
A similar picture was also evident for weld W3
(Table 7b). As far as HAZ results were concerned,
although wet spot welds did show higher hardness
Vol 28, No 3, 2009
15% Cr
2.5% Mo
14% Cr
2.3% Mo
13% Cr
2.1% Mo
12% Cr
1.9% Mo
1.7% Mo
11% Cr
1.5% Mo
10% Cr
1.3% Mo
9% Cr
1.1% Mo
8% Cr
0.9% Mo
7% Cr
0.7% Mo
6% Cr
5%Cr
Fig 11: Quantitative map plotted for Mo
Fig 9: Quantitative map plotted for Cr in weld D2
8.0 % Ni
7.5 % Ni
7.0 % Ni
6.5 % Ni
6.0 % Ni
5.5 % Ni
5.0 % Ni
4.5 % Ni
4.0 % Ni
3.5 % Ni
3.0 % Ni
Fig 10: Quantitative map plotted for Ni
values than dry spot welds, their values were still
acceptable under BS EN ISO 15614-1 (2004) and
AWS D3.6 (1999) and did show some improvements
over conventional wet MMA fillet welds (Keats,
2004, 1990; Masubuchi, 1981; Abson and Cooper,
1998; West et al., 1990).
3.4. Macro/microscopic survey
To better understand what has actually happened
to weld D2 (highest hardness dry weld) and wet
weld W3, a series of microphotographs were taken
to study the microstructures present. Weld D2
was also scrutinised under a Cameca SX50 EPMA
electron microscope to map the weld area (see
Figs 9–11). The results for weld D2 showed reduced
Cr, Ni and Mo levels present in the root area
of the weld, located just at the point where the
switch-over from high to low current took place.
This demonstrates that higher dilution occurred
in the root area, resulting in higher carbon levels.
This factor may account for the observed elevated
hardness readings, despite a slower cooling rate
than weld W3.
Microphotographs for weld D2 apparently confirmed this effect and showed that higher carbon
martensite existed, as did numerous spherical
carbide particles. Martensite and carbides were
evident to some degree throughout the whole weld
body, as was the occasional isolated globular oxide.
Both welds, wet and dry, showed a microstructural
similarity, with the existence of delta ferrite in an
austenitic matrix together with isolated globular
oxides being present.
Weld W3 showed evidence of a small crack in the
root area, which may have been the result of the
shear testing, as no other significant metallurgical
factors were observed that may have caused such a
defect. This may be explained by the pronounced
loss of material that occurred from the weld
nugget, which appeared to show a ductile break.
In addition, it was clear that the material for welds
D2 and W3 were not the same, despite the material
specification (shown in Table 1).
Weld D2 clearly had a higher carbon content,
as shown by the ferrite and pearlite content, which
may be as high as 0.25% (see Fig 12). Weld W3
showed a lower percentage of pearlite (smaller and
less carbide platelet formation). This may more
accurately reflect a material composition closer to
0.15% carbon (see Fig 13).
4. Discussion
In considering weld strength versus weld size, and
thus the number of welds required for any given
load-carrying capacity, the following principle to
calculate overall stress can be employed:
Stress (UTS) =
Force (load)
Area
(9)
The dry results, as shown in Table 4 and Fig 3,
reveal the average CSA for dry spot welds was
86.24mm2 with the average load to failure being
calculated at 45.63kN, whilst the average ultimate
tensile stress (UTS) was 548.5N/mm2 . The wet
tensile results in Table 5 and Fig 3 show the average
CSA for wet spot welds was 97.17mm2 with the
average load to failure being calculated at 39.95kN,
whilst the average UTS was 474.50N/mm2 . Thus
by comparison, the average dry spot weld CSA
was 10.93mm2 smaller than the average wet spot
123
Keats. Underwater wet welding made simple: benefits of Hammerhead r wet-spot welding process
a
a
b
b
Fig 12: (a) Weld macro of weld D2 (dry weld)
conducted by Welder B; (b) parent material
comprising predominantly ferrite and pearlite
weld, but offered an increase in shear strength of
some 5.68kN.
By comparing the wet shear test results to
the theoretical based value (10.0mm-diameter
nugget), which produced a load to failure of
40.8kN with a UTS of 408N/mm2 (as shown in
Fig 4), the actual weld deposited offered a slight
reduction of 0.85kN or 2.1%. When calculating
the reduction in CSA that equated to 2.83mm2 ,
however, the strength reduction became 2.9%.
Thus, the design principle that predicts a given
number of spot welds for a given load would appear
to overvalue wet weld strengths by approximately
3%. Nevertheless, this approach demonstrates that
simple calculation would provide a reliable base
method for determining the number of spot welds
necessary to carry a given load (see Appendix 5 for
more details).
The overall appearance of wet spot welds, as
compared to dry spot welds, was somewhat untidy,
with clearly a more restricted weld puddle in
evidence. The size and profile (cap) of the welds
did not appear to significantly affect the results of
mechanical testing.
The average hardness values shown in Tables 6a
and 7b for wet and dry spot welds were acceptable,
showing no particular hardness concerns. In fact,
considering the average values between wet and
124
Fig 13: (a) Weld macro of weld W3 (wet weld)
conducted by welder C; (b) parent material
comprising predominantly ferrite with a small
amount of pearlite (carbon content appears to be
approximately 0.10–0.15% based on the
pearlite present)
dry (excluding D2 and D4), the difference was
so minimal as to be irrelevant. It should be appreciated, however, differences in material carbon
content for welds D2 and W3 could alone be
sufficient to show a difference in the hardness
readings obtained. The hardness results shown in
Tables 6b and 7b for welds D2 and W3 clearly shows
a significant difference in the overall hardness on
weld metal, with the dry weld (D2) suffering from
increased carbon dilution as compared to the wet
weld (W3).
The weld macros showed that weld quality was
similar between wet and dry, though not defectfree, and no significant incidence of defects were
produced wet, as compared to dry. It should also
be noted that all welds, both wet and dry, had
been mechanically tested prior to macro/micro
examination and hardness surveys. This may,
therefore, have had some effect on the results
obtained. Nevertheless, the quality of wet spot welds
produced showed that this method of welding can
be relied upon to produce under water welds, at
the very least, as effective as those described in
the referenced literature for conventional wet fillet
Vol 28, No 3, 2009
welds (Bailey, 1987, 1991; Sadowski, 1980; Gooch,
1983a,b; Masubuchi, 1981; Abson and Cooper,
1998; West et al., 1990). It is accepted, however,
martensite is likely to exist at the interface (root)
area of any welds produced using this method.
No weld cleaning or joint preparation was
performed prior to welding, unlike that of conventional wet fillet welding, thus welding efficiency
was significantly increased, with a completed weld
being produced in less than 30 seconds. The control
device provided a suitable means to control the
essential welding parameters and demonstrated
the means to reduce the role of the diver, even
under nil visibility conditions. The presence of
the diver is still essential in the production of a
satisfactory weld, due to the need to apply adequate
pressure. Nevertheless, this welding method has
demonstrated a successful means of joining carbon
steels that eliminates the need for skilled welders,
as well as all conventional cleaning/preparation
methods. Furthermore, successful wet welds were
produced under conditions of nil visibility.
5. Conclusions and further work
To demonstrate the commercial advantages of
this process, the experiments concentrated on the
following conditions:
1. Producing spot welds in nil visibility, while still
providing for an effective weld
2. Elimination of preparation/cleaning of materials and increased welding speeds
3. Elimination of welding skills
4. Repeatability and consistent weld quality.
The experiments demonstrated that the spot
welding method tested was more than capable of
making an effective mechanical fixing under water.
It also provided benefits in the way of speed, quality
and repeatability over conventional wet MMA fillet
welding, without using skilled welders and working
in poor/nil visibility conditions for both dry and wet
spot welds.
Dry welds were produced as a baseline comparison to compare weld quality and highlight
any differences in mechanical and metallurgical
qualities. This work was limited to welding low
carbon structural steel plate (8.0mm) in fresh water
at a depth of 3m using a specially designed control
device. The experiment had not taken into account
the possible effects of welding in seawater, nor did
it consider other welding positions (Keats, 2004,
1990; AWS, 1999; Hibshman and Jensen, 1933;
BS/EN/ISO, 2002; Gretskii and Maksimov, 1998;
Bailey, 1987, 1991).
Welding was restricted to the use of a single Fe–
Cr–Ni–Mo stainless electrode of 3.2mm diameter,
with all welding being conducted in the flat position. Further work would be required to evaluate
this welding methodology more fully, including
different sizes of electrodes, welding positions and
water type and depth, together with different grades
of structural steel. All under water welding was
conducted in poor and/or nil visibility conditions
using both skilled and non-skilled welder-divers.
Although only a few welds in total were produced,
which were insufficient to provide for a comprehensive outcome, the evidence supports that divers
with little or no welding skills/knowledge were able
to produce acceptable spot-welds just as easily as
the skilled welders. It was shown that visibility had
no effect on performance or weld quality, nor did
the lack of weld preparation or cleaning appear to
affect final weld quality.
Although not a fully automatic welding method,
the control device proved suitable to manage the
welding parameters essential to produce quality
welds repeatedly. It was axiomatic that each
individual diver must ensure suitable pressure is
applied to the electrode for an acceptable weld to
be produced. The wet spot welds provided suitable
weld quality in terms of strength, with properties
closely matching those of dry spot welds. Due to
the metallurgy, however, the process is likely to be
limited under water to welding non-load bearing
joints (e.g. anodes). During testing, it became
evident that the spot welding method provided
for a considerably faster joining method than
conventional wet MMA fillet welding, as the process
did not require any joint preparation or cleaning of
the material/weld, so spot welds (wet and dry) were
produced in a matter of seconds.
The Hammerhead welding process clearly remains a manual operation, despite the control
device, whereas Sadowski’s work (1980) involved
automatic fixed welding heads working in handdeep test tanks only. In contrast, the Hammerhead
process is used by a diver being fully submerged
under water. The welding process is designed as a
one-spot process – i.e. one electrode produces one
spot weld, thereby eliminating the need to make
a second weld over the first, as well as inter-run
cleaning and the need for a second pass (Abson and
Cooper, 1998).
The Hammerhead welding method appears to
lend itself to automation and may well prove to
be of great interest, as presently there remains
a level of control required by the diver to apply
pressure to the electrode during welding. It was
reported by Gooch (1983a,b), Masubuchi (1981),
and West et al. (1990) that the use of austenitic
electrodes to produce conventional fillet and butt
welds under water often produced cracking in the
weld root and hot pass zone. Abson and Cooper
125
Keats. Underwater wet welding made simple: benefits of Hammerhead r wet-spot welding process
(1998) also found when using this type of electrode
cracking so extensive that it prevented any useful
mechanical testing from being undertaken.
With the exception of a small micro-crack
(≤0.3mm) in zone E on weld W3, no other cracking
was observed in the weld or the HAZ. Although
this crack may be metallurgical in nature, it was
more likely to be as a result of the mechanical
testing, as evidence exists that the weld underwent
significant stress with large sections of weld material
missing from the fracture face. The microstructures
reported by Abson and Cooper (1998) also stated
that martensite was severe, particularly in the
root area, where contact with the parent material
produced high dilutions. Nevertheless, martensite
was observed throughout the whole weld body.
Although martensite was present in the weld body
of the Hammerhead spot welds, it was not as severe
as reported by Abson and Cooper (1998).
The hardness values reported also showed a
significant increase over those produced by the
Hammerhead process, whose values (as shown
in Table 7b) are considerably lower and thus
clearly suggest evidence of an improved welding
process. The average shear strength values, as
reported by Gooch (1983a,b), suggest that the
Hammerhead process provided equally effective
mechanical strength properties.
Van der Brink and Boltje (1983) also demonstrated that the moisture content of the electrode flux covering was critical to avoid hydrogen
cracking. The experiments detailed herein were
conducted on a specially developed electrode,
manufactured specifically to minimise moisture
pickup. Furthermore, the number of electrodes
taken into the water at any one time was limited also
to minimise moisture pickup and thus help prevent
hydrogen cracking. Szelagowski (1991) reported
that the type of waterproof coating could also have
a significant effect on chemical composition of the
weld deposit, though no evidence of this was observed on the welds produced in this work. Clearly,
further work in this regard would be beneficial.
The waterproof coating used for the Hammerhead
electrode was a specially formulated vinyl lacquer.
According to Grubbs (1986), conventional wet
fillet welds using ferritic electrodes can be successfully made in accordance with AWS D3.6 (1999)
Class B at depths down to 60m. The opportunity
to test the Hammerhead process at these depths
was not available for these experiments, although
it is accepted that deep welding trials would be
necessary to fully evaluate this process/electrode
and investigate any differences in weld quality from
the shallow water tests undertaken. From the results
obtained for the Hammerhead process, evidence
suggests that further additional alloying of the
126
electrode may be beneficial, in order to help reduce
martensite in the central weld body area. This,
however, is unlikely to offer any benefits at the joint
interface (root) zone, where the highest dilution
was recorded.
The work undertaken by Corus (Thompson,
2005) demonstrated that the welding methodology
to be valid, even for above water applications,
and that it was more than capable of joining a
wide range of material thicknesses. The process
offered a rapid method of joining plates and sheet
steel in both very thin and thick materials (1.6–
15.0mm) and the device-controlled penetration
quite satisfactory. Welding was performed in both
the flat and vertical orientation, and no significant
weld defects were reported, though some voids were
present on thicker materials.
Nevertheless, all joints contained large fused
regions, providing mechanically strong joints. The
welding of thin galvanized steel sheet provided
joints with high mechanical integrity, with the weld
nugget being pulled out from the parent material
with significant plastic deformation. It was noted
that further work was needed to investigate the maximum gap tolerance, as no gap was present or preset during the initial trials. This is recognised to be
an area of extreme interest and significance. (The
full Corus report is available as a separate report.)
Appendix 1
Theoretical value for a spot weld, based on
3.141 × d 2/4 × 520
Load (kN) 40.84
Table 1: Recorded dry-spot weld values (kN)
1
2
3
4
(Average)
Load (kN) 47.8 45.9 52.2 36.6 45.63
Table 2: Reduction in wet-spot weld strength
by 14.2%
1
2
3
4
(Average)
Load (kN) 41.01 39.38 44.79 31.40 39.15
Table 3: Further reduced from table 1 values
by 23.56%
1
2
3
4
(Average)
Load (kN) 36.54 35.09 39.91 27.98 34.88
Table 4: Adjusted for weld strength overvalue
of 40.84kN by 3%
Load (kN) 39.61
References
Abson DJ and Cooper MA. (1998). Wet under water
welding trials with commercial MMA electrodes. In:
Proceedings of the Underwater Wet Welding and
Cutting Conference, The Welding Institute (TWI)/Paton
Institute Middlesbrough, 17–18 April, 50–67.
American Welding Society (AWS). (1999). AWS D3.6M-99
Underwater Welding Specification. Miami: AWS, 129pp.
Vol 28, No 3, 2009
Association of Offshore Diving Contractors (AODC).
(1985). AODC 035: Code of Practice for the safe use of
electricity under water, 135pp.
Bailey N. (1987). Exploratory tests on nickel-based wet
welding electrodes for ferritic steel. The Welding Institute
(TWI) members report 355.
Bailey N. (1991). Welding Underwater – A Metallurgical
Appraisal. Paper presented at the 1st International
Offshore and Polar Engineering Conference (ISOPE-91)
Edinburgh, 11–15 August.
British Standards Institute/European Standard. (1996). BS
EN 1043-1: 1996. Destructive tests on welds in metallic
materials. Hardness testing. Hardness test on arc welded
joints.
British Standards Institute/European Standard/International Organization for Standardization. (2004). BS EN ISO
15614-1: 2004. Specification and qualification of welding
procedures for metallic materials. Welding procedure
test. Arc and gas welding of steels and arc welding of
nickel and nickel alloys.
British Standards Institute/European Standard/International Organization for Standardization. (2002). BS EN
ISO 15618-1: 2002. Qualification testing of welders for
under water welding. Diver-welders for hyperbaric wet
welding.
Gooch TG. (1983a). Properties of under water welds, Part 1:
Procedural trials. Metal Construction 15(3): 164–167.
Gooch TG. (1983b). Properties of under water welds,
Part 2: Mechanical properties. Metal Construction 15(4):
206–215.
Gretskii Y and Maksimov S. (1998). Study of
physico-metallurgical peculiarities of wet arc welding
of structural steels. In: Proceedings of the Underwater
Wet Welding and Cutting Conference, The Welding
Institute (TWI)/Paton Institute Middlesbrough,
17–18 April.
Grubbs CE. (1986). Qualification of under water wet welding
procedures at water depths down to 99M. In: Proceedings
of the 16th International Diving Symposium, Houston
Texas, Session 8, Paper 1, 7pp.
Hibshman NS and Jensen CD. (1933). Underwater Arc
Welding. Welding Journal, 4–9.
Keats DJ. (1990). Professional diver’s manual on wet-welding.
Cambridge, UK: Abington Publishing, 80pp.
Keats DJ. (2004). Underwater Welding: ‘‘A Welder’s Mate’’.
Leicester, UK: Matador/Troubador Publishing, 300pp.
Kralj S, Kozuh Z, Garasic I and Dorn L. (2003). Influence of
the water environment on the parameters in under water
wet welding. Welding and Cutting 55(2): 97–99.
Masubuchi K. (1981). Underwater Welding – Factors Affecting Welding Metallurgy. In: Proceedings of Underwater
Welding of Offshore Platforms and Pipelines, 5–6 November, New Orleans, Louisiana, 81–98.
Sadowski EP. (1980). Underwater wet welding nickel base
and stainless steel electrodes. Welding Journal 59(7):
30–38.
Szelagowski P. (1991). Underwater welding – present state
and development trends. Welding and Cutting, 7–11.
Thompson AM. (2005). Welding developments for Speciality Welds. Corus Research, Development and Technology. Reference source No. 115034.
Van der Brink SH and Boltje GW. (1983). Proceedings,
Conference on Underwater Welding. Oxford: Pergamon Press.
West TC, Mitchell G and Lindberg E. (1990). Wet welding
electrode evaluation for ship repair. Welding Journal
69(8): 46–56.
127