Embed Size (px)
Citation preview
doi:10.3723/ut.28.115 International Journal of the Society for Underwater Technology, Vol 28, No 3, pp 115–127, 2009
Tec
hnical
Pap
er
Underwater wet welding made simple: benefits ofHammerhead r© wet-spot welding process
David J KeatsSpeciality Welds Ltd, Cleckheaton, West Yorkshire, UK
AbstractA new method of wet welding was investigated toevaluate potential improvements in weld quality, easeof use, increased welding speed and the eliminationof welding skill. The new welding process, whichhas been called Hammerhead ‘wet-spot’ welding,eliminates the need for skilled welder-divers as wellas traditional cleaning and preparation techniquesnormally associated with conventional manual metalarc (MMA) wet welding. In addition, the process alsoallows welding to be conducted in nil visibility, yetremains a MMA process, using a specially designedFe–Cr–Ni–Mo electrode.
The process utilises a control device, which mustbe pre-set before the diver enters the water. Throughthis device, weld parameters are controlled and qualityis maintained, thus the role of the diver is simplifiedto three steps: make contact with the material, strikethe arc and maintain pressure to the electrode whilewelding.
A series of spot welds were produced both wet anddry on 8.0mm carbon steel plates. The welds wereevaluated with regard to ease of use and setting upof the device, speed and final weld quality. Initially, theperformance of the process was assessed and usagediagrams produced. Work regarding an automatedversion of the system has also been proposed.
Keywords: Hammerhead, welding, wet-spot, under-water, subsea, welder-diver
1. IntroductionUnderwater wet welding has been employed formany years now, but has commercially been re-stricted to conventional manual metal arc (MMA)welding techniques (Keats, 2004, 1990; Associationof Offshore Diving Contractors, 1985; AmericanWelding Society, 1999; Hibshman and Jensen,1933; British Standards Institute et al., 2002).The typical problems associated with under waterMMA welding fall into two categories: mechani-cal/metallurgical quality, and skill and ability. Itwas with both these issues in mind that a newmethodology 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 dueto the rapid cooling (Keats, 2004, 1990; Associationof Offshore Diving Contractors, 1985; AmericanWelding 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 appreciatedthat the skills and abilities necessary to execute highquality, conventional MMA wet welds are extremelyhigh, therefore labour and training costs are sig-nificant factors (Kralj et al., 2003; Gooch, 1983a,b;Grubbs, 1986). This new welding methodology,which has been developed by the author, provides asolution to both of these issues. The process is calledHammerhead wet-spot welding.
This process provides an alternative approachto welding, in which the role of the operator isminimised and is therefore no longer required tohave hand-eye coordination skills. Rather, for thismethod, two materials are joined together by a spotor plug weld by a programmable control device.In this way, the operator simply becomes a meansof making contact with the material and providingmomentum to push the electrode into the materialonce the arc is struck. The process also eliminatesthe need for traditional cleaning, joint preparationsand chipping of weld slag. It utilises one electrodeto produce each spot weld, which is localisedthrough the thickness dimensions of the material.
The author has shown that the final mechanicalweld qualities have been improved, as has theoverall speed of joining when compared to conven-tional wet welding techniques. Unlike conventionalMMA welding, the process provides a methodof controlling the welding current necessary toproduce a weld, without requiring the operator tohave any welding skills or knowledge, because thecurrent is automatically regulated and controlled bythe device on each weld cycle. Thus, the role of thediver is reduced to simply that of an operator.
2. Experimentation and results2.1. Design of apparatusThe Hammerhead MMA wet-spot welding methodemploys 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 controldevice (bottom left)
number of key welding functions to produce a spotweld. 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 functioncables.
2.2. Control functionsThe control device, which is housed within a utilitycase, consists of an on/off switch to power the unit,high/low/auto current control potentiometers, atimer, and amp and volt meters (see Fig 1). Toensure a suitable, safe current is available, thedevice is fitted with a transformer to adapt a 110-voltsupply down to a more suitable and safe 9 volts,which is then rectified to direct current (DC). Areed switch is fitted to trigger a relay, which startsa timer when the arc is first struck. Two currentcontrol potentiometers independently control highand low current settings.
Once these have been set, the device can beswitched into ‘auto’ mode. These potentiometersare adjusted to deliver the appropriate currentto penetrate and fill the materials and thusproduce the spot weld. Once the timer hasbeen triggered (following arc initiation), the highcurrent potentiometer delivers the preset currentfor the set time period. Expiration of this controlthen triggers the low current potentiometer to act,thereby initiating the required low-level currentautomatically. This low current function continuesuntil the arc is broken, after which the deviceautomatically resets and is ready to make the next
spot weld, although a 5-second delay prevents thesystem resetting, should the diver accidentally breakthe arc.
LEDs light up against each function so theoperator can monitor the process at any givenmoment. All welding parameters are set prior tothe diver entering the water and involve the devicebeing connected to the weldingmachine via remotecontrol and 110-volt power supply cables. Onceconnected to the device, complete manipulationof the welding machine is provided. Amp andvolt meters are fitted to provide a visual displayof the welding current/voltage, as is a 400-ampsafety switch to isolate the current/voltage to thediver, which is required under Health and SafetyExecutive (HSE) regulations.
The setup of the device is quite straightforward.Prior to entering the water, the diver selects asuitable ‘high’ current (determined by eye) to allowfor adequate penetration of the two materials tobe joined. This high current time is recorded inseconds, and penetration is measured visually byexamining the back of the material for a heat mark,or blister. Providing this is visible on the outsidesurface of the back-face, penetration is adequateand the timer control and high current functionare programmed and set, the operator now selectsthe ‘low’ current control. This function does notrequire the use of the timer and is set simply toprovide a suitable current to complete the weld.
The device may now be set to automatic and isfully programmed to produce welds automatically.The diver can now enter the water and make anysmall adjustments as might be necessary for thegiven water type and working depth. After theseadjustments are made, the device may be reliedupon to give consistent and reproducible weldingparameters, as programmed, for each and everyweld. 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 theoperator does not over penetrate the base mate-rials. Should this occur, weld properties would becompromised by the effects of water pressure, extin-guishing the arc and causing slag entrapment, lackof fusion and/or cracks. As the only opportunity forburst-through is while the ‘high’ current cycle is inoperation, the timer controls this critical function.Excessive penetration is a combined function ofboth high current and arc time. By accuratelycontrolling both functions, penetration control isachieved. It is not possible for the diver to burstthrough the material while the ‘low’ current cycleis functioning, as the current is too low. Thus,the device reduces the role of the diver to that of
116
Vol 28, No 3, 2009
Table 1: Composition of steel plates
Element C Si Mn P S% (max.) 0.2 0.55 1.6 0.035 0.035
Table 2: Carbon equivalent value of steel
CEV 0.35
simply ‘pushing’ the electrode into the materialsand ensuring that contact is maintained.
In operation, this requires no more than 5–10kgfand, provided the operator consistently maintainsthis force, poor visibility conditions will in noway affect the outcome or quality of the weldproduced. The applied force was estimated, basedon experimentation, and became a basis forcalculating the necessary pressure to be applied,using a 3.2mm electrode.
Pressure N/mm2 or (MPa) =Force (kgf)Area (mm2)
1kg = 9.80665 Newtons (1)
Although the core wire of the electrode mea-sured 3.2mm, the outer flux coating also needs to betaken into account, which increases the diameter toapproximately 6.0mm. Therefore, an applied forceof 5–10kgf by the operator will ensure a pressure atthe tip of the electrode of some 1.73–3.49N/mm2
(MPa). The actual applied force each diver usedwhilst welding was clearly onerous. A ‘best estimate’was made by each individual, but the applied forcewas based on the above calculation.
Formuch of the welding operation, the electrodetip is deep within the wall thickness of the material,so no arc is visible, thereby minimising the diver’sinfluence on weld quality.
2.3. Weld samplesAll welding was performed on plate that was 150 ×150× 8.0mm. Material was restricted to low carbonsteel having the composition as shown in Table 1,with CEV (IIW) as shown in Table 2. CEV formulacalculated as:
C+Mn6+(Cr+Mo+ V)
5+(Ni+ Cu)
15(2)
The following weld IDs were assigned for eachtest 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 lapjoints, with one plate overlapping the otherby ∼50%.
2.4. Equipment, facilities and environmentAll welding was conducted onsite in open airconditions, utilising the following equipment andfacilities:
Table 3: Composition of Hammerhead electrodes
Element Cr Ni C Mo Mn Si NbMin. 21 11 0.020 3 0.60 0.70 —Max. 24 14 0.10 5 2 2 —
• A 400-amp Gen-Set diesel welding generator• Piranha welding monitor/safety switch, fittedwith the Hammerhead control system• Underwater welding stinger• Welding leads (50mm2 copper) double-insulated• Brass parallel closing earth clamp.
The diving equipment used was standard com-mercial surface demand, i.e. the diver being fedwith an air supply through an umbilical rather thana scuba bottle. Full radio communications were alsoin place throughout, enabling welding data to besupplied and recorded. Welding was conducted ina freshwater dive tank, and working depth was 3m.The environmental conditions recorded duringthese experiments were as follows: air, −3◦C (±1◦)and water, 0◦C (±1◦).
2.5. ElectrodesThe 3.2mm electrodes were used, having thechemical composition shown in Table 3. Theelectrode was designed specifically to allow forhigh dilutions while maintaining very short-arcconditions. These electrodes have the potential toallow for dilutions up to amaximumof 38%withoutthe risk of martensite formation. To evaluate thispotential fully, the Shaeffler diagram was used(Fig 2) to plot a dilution line based on the meanvalues, as shown in Equation 3.
CrEq = %Cr+Mo+ (1.5× Si)+ (0.5×Nb)
NiEq = %Ni+ (30× C)+ (0.5×Mn) (3)
The Hammerhead electrode provides for thefollowing Cr and Ni equivalents, based on thefollowing formula:
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 thiselectrode provides for a maximum dilution of38.4%, without risking the formation of martensitein the body of the weld. It is accepted whenusing MMA welding, a dilution of around 25%can be expected. Under water, this is normallyslightly reduced, due to ambient temperature andrapid cooling, to approximately 20% (Keats, 2004,1990; Gooch, 1983a,b; Masubuchi, 1981; Abson andCooper, 1998).
117
Keats. Underwater wet welding made simple: benefits of Hammerhead r© wet-spot welding process
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
Nic
kel-A
quiv
alen
t = %
Ni +
30x
%C
+ 0
,5x%
Mn
+ 3
0x%
N..
Chrom-Aquivalent = %Cr + 1,4x%Mo + 1,5x%Si + 0,5x%Nb + 2x%Ti..
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
F +M
Austenit
A + M
Tie-Line
118mm
M + F
A + M + F
A + F
Max Dilution
45.35nm (38.4%)
0% F
errit
5%
10%
20%
40%
80%
100%
Ferrit
Martensite
Fig 2: Maximum dilution possible without risk of martensite forming is 38.4%
2.6. Welders and operatorsFour individuals were engaged to carry out weldingand were identified as follows:
• Welder A: skilled welder – conducted welds D1and W1• Welder B: non-welder – conducted welds D2and W2• Welder C: skilled welder – conducted welds D3and W3• Welder D: non-welder – conducted welds D4and W4
Each diver was asked to produce one dry and onewet weld.
2.7. Welding proceduresTo ensure accurate data collection, all weldingoperations were recorded. Applied arc energy wascalculated by use of the standard formula given inEquation 5:
Arc energy =I × V (total power in watts)∗ROL (mm)/time in seconds
(5)
where I is current, V is volts and ∗ROL equalsrun out length of the electrode (in this case ROLreferred to the burn-off rate of the electrode).
No specification exists for wet-spot welding,thus in order to determine how many welds maybe required to bear a given load, the formulashown below was used. The size of any given weld,and therefore the number of welds required, isbased on the required shear stress exerted on thecomponent. Thus, each single spot weld can offerthe following theoretical strength properties.
Max load =πd 2
4× shear strength
(neglecting any bending moment) (6)
In calculating the shear strength for plain carbonsteel, it is common industrial practice to assume thisto be ∼80% of the ultimate tensile strength. TheHammerhead electrode offers a tensile strengthof 650MPa (all-weld ‘dry’ test); based on thisassumption, shear strength becomes 520MPa.
Thus, load carrying area (mm2) is πd2
4 where d isthe spot diameter (mm), so for a 10.0mm spot weld,the area is π1002
4 = 78.54mm2. Max design shearstress for a 10.0mm spot weld is therefore 40.84kN.The number of 10mm diameter welds required canbe calculated using Equation 7:
N =total shear load (kN)
40.84(7)
Alternatively, the shear stress can be calculatedper mm2 of weld. This would produce the total spotweld area required and thus lead to a selection ofspot welds.
Shear Stress (X) kN mm2=
40.84 (kN)78.54 (mm2)
(8)
Thus shear stress equals 0.52kN/mm2. Thetotal spot weld area required for a load of 45kNis therefore 86.54mm2. The actual test resultsobtained for all wet and dry spot welds are shownin Fig 3. The results presented in Fig 4 show thedesired joint strength against a specific numberand/or size of individual spot welds, based on thecalculations discussed.
2.8. Spot weldsThe welding parameters and techniques for allwelds were pre-set and recorded as follows:
• Amps: primary value 250–260, secondary value150–160• Timer: 5–6s for peak (primary) current value
118
Vol 28, No 3, 2009
60
50
40
30
20
10
0
LOA
D k
N
AREA mm2
Actual WET welds Theoretical average WET weld Actual DRY welds Theoretical average DRY weld
HAMMERHEAD SINGLE WET & DRY SPOT-WELDS
63.2 107.16 152.7 97.8 66.62 78.07 87.92 112.36 86.2465.62
Fig 3: Results of the actual wet and dry spot welds produced
Des
ired
Join
t Str
engt
h (k
N)
Number of Spot Welds
Weld Strength Vs Number of Spot Welds
350.00
300.00
250.00
200.00
150.00
100.00
50.00
0.00
294.05
235.24
176.43
117.62
58.81
204.20
163.36
122.52
81.68
40.84
130.69
104.55
78.41
52.28
26.14
8mm spot weld
10mm spot weld
12mm spot weld
1 3 52 4
Fig 4: Number of welds required against desiredweld strength, based on the calculationsdiscussed
• 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 clampedtogether to prevent relativemovement. No cleaningor other joint preparation was used. Welding wasconducted on plates in as-delivered condition. Atthe time of welding for Welder A, visibility wasmoderate at approximately 30–45cm.
Welder B was permitted a short practice periodfor familiarisation. The welding parameters andtechniques for Welder B were exactly the same asfor Welder A. At the time of under water welding,visibility was very poor at <25cm.
Welder C was asked to produce his welds aftera brief introduction of the technique. At the timeof under water welding, visibility was very poorat <25cm.
Welder D was asked to produce his weldscompletely unaided and without any opportunity topractice, in a bid to demonstrate the feasibility ofa no-skill process. At the time of welding, visibilitywas completely nil and all welding was carried outby touch.
3. Results3.1. Visual appearanceThe overall quality of welds produced for both wetand dry was surprisingly similar, especially whenone considers the visibility factor. Equally, therewas no substantial difference between welds madeby the skilled welders over non-welders. All weldsshowed adequate fusion between basematerials andweld metal. Although not completely defect free,some wet spot welds did show evidence of minorgas voids/slag inclusions in the weld body.However,none of the recorded defects appeared to make asignificant impact on the overall results of weldsmade wet as compared to welds made dry. Noappreciable defects were observed by the naked eyefor any dry spot welds.
Welds generally had an overall convex circularappearance, but a clear difference did exist betweenwet and dry. Wet spot welds had a somewhat untidyappearance and failed to blend in well with the topplate surface, unlike the dry welds. This appearancewas due to the existence of a more restrictedweld 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 dryspot welds during the final stages of welding, whichdid assist in working/wetting out the weld puddle.This manipulation produced a smoother, moreblended appearance, and as a result, dry welds didnot show the excess ‘flash’ material (which wasevident in all wet spot welds). This flash was dueto excess material being ejected from within themolten nugget, resulting from additional electrodeweld metal, causing still, molten metal to be ejectedas a result of continued pressure applied to theelectrode. Although untidy in appearance, thisflash material was easily removed with a simplehammer blow.
One common feature for both wet and dry spotwelds was the heat mark, or blister, formed onthe back-face of the base material. This provideda very useful indicator as to the success ofpenetration. Although not accurate in terms ofmeasurement or depth, it did provide an excellentvisual method of establishing whether adequatepenetration had occurred. Where no heat markwas present, the depth of penetration into the backmaterial was limited.
The overall diameter of the welds produced inair (measured across the top outside diameter of
the weld) was somewhat larger than welds producedwet, with the average diameter for a dry weld being21.48mm against 14.39mm for that of a wet weld.Wet welds on average were nearly 50% smaller indiameter (49.27%) compared to dry welds madeunder similar current/voltage conditions. Thisincrease in diameter appeared to be mainly dueto operator manipulation of the electrode, despitebeing requested not to do so, andmay also be due inpart to the input energy vaporising the water. Thiscan be seen from studying the weld shapes moreclosely in themacro-photographs shown in Figs 5–8.
3.2. Transverse tension shear testsIn order to establish the load required to failure,both wet and dry spot welds were subjected totransverse shear tensile tests. Tables 4 and 5 showthe individual test results for wet and dry spot welds.
The average failure load of each weld type was45.63kN for dry spot welds and 39.95kN for wetspot welds. A difference of 5.68kN between wet anddry was found. Thus, the average dry spot weldoffered a 14.2% strength improvement over theaverage wet spot weld. The average cross-sectionalarea (CSA) of weld nuggets for all welds was86.24mm2 for dry spot welds and 97.17mm2 for wetspot welds.
120
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 66.62 717 47.8Welder B – D2 78.07 587 45.9Welder C – D3 87.92 594 52.2Welder D – D4 112.36 326 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 63.20 668 42.2Welder B – W2 65.62 607 39.8Welder C – W3 107.16 379 40.6Welder D – W4 152.7 244 37.2
Average 97.17 474.5 39.95
Wet welds, therefore, showed an increase of10.93mm2, thus increasing the CSA of depositedweldmetal by 12.67% (12.7) even though the actualdiameters were far smaller than any of the dry welds.By factoring in this percentage change in the CSA ofwet welds to match the CSA of dry welds, a new loadrequired to failure of 34.88kN (34.9) may be cal-culated. This difference of 10.75kN further reducesthe wet strength results, as compared to the dry, by23.55% (23.6) See Appendix 5 for more details.
Clearly, the effects of rapid cooling on weldsmade under water should have caused a change inthe mechanical strength of the weld, due to thefaster cooling rates. To better understand these
results, hardness surveys and weld macros/microswere also examined. Unfortunately, however, theseparticular tests were carried out after shear testingand thus may have obscured any minor defects.It was also noted that the dry spot welds hadlarger weld reinforcement, which accounted forthe initial observation that the CSA of dry spotwelds were actually larger, although this is unlikelyto have offered any real advantages in terms offailure strength. The major influence in effectivejoining was adequate penetration of the nugget intothe base materials, rather than the size of weldreinforcement. It should also be appreciated thatthe 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 weldsusing Vickers method at HV-10
Parent HAZ WeldWelder A – D1Traverse 1 (top) 117 183 272Traverse 2 (bottom) 123 182 282
Welder B – D2Traverse 1 (top) 149 168 237Traverse 2 (bottom) 138 172 458
Welder C – D3Traverse 1 (top) 154 197 188Traverse 2 (bottom) 148 176 189
Welder D – D4Traverse 1 (top) 125 165 215Traverse 2 (bottom) 122 173 401
Combined average 135 177 280
Table 7a: Hardness surveys for wet spot weldsusing Vickers method at HV-10
Parent HAZ WeldWelder A – W1Traverse 1 (top) 124 208 134Traverse 2 (bottom) 122 248 217
Welder B – W2Traverse 1 (top) 139 223 276Traverse 2 (bottom) 164 224 284
Welder C – W3Traverse 1 (top) 147 308 274Traverse 2 (bottom) 147 220 281
Welder D – W4Traverse 1 (top) 116 199 132Traverse 2 (bottom) 120 181 262
Combined average 135 226 233
welds – especially for Welders B and C – was poorand completely nil for Welder D.
3.3. Hardness surveyA number of hardness surveys were carried outin accordance with BS EN 1043-1(1996) with twotraverse lines being used and six indentations forparent metal, six for HAZ and three for weld-metalper traverse line for welds D2 and W3. Table 6ashows the average results for all dry spot welds,whilst Table 7a shows the average results for allwet spot welds. An additional hardness survey wasundertaken on welds D2 and W3 running down theweld centre from top to bottom. Its purpose wasto better understand the microstructural changestaking place within the weld body at differentintervals from the interface to the weld cap (seeTables 6b and 7b).
When considering the combined average valuesfor Tables 6a and 7a, it was seen that wet resultswas similar to the dry welds. Somewhat surprisingly,
Table 6b: Individual hardness survey ofdry weld (D2); weld-metal only, running from
top weld centre to weld bottom
Distance from weldcap (mm)
Hardness Hv1kg
0.3 3121.0 2082.0 2054.0 3736.0 2058.0 3649.0 45410.0 48211.0 44912.0 46813.0 472
Table 7b: Individual hardness survey ofwet weld (W3); weld-metal only, running from
top weld centre to weld bottom
Distance from weldcap (mm)
Hardness Hv1kg
0.3 2101.0 2062.0 1943.0 1964.0 2615.0 4196.0 2839.0 21510.0 38611.0 37011.7 378
however, was that the wet welds produced lowerhardness values than the dry welds. This is contraryto what might be expected with conventionalunder water welds cooling more rapidly, thusresulting in harder weld and heat-affected zone(HAZ) metal (Keats, 2004, 1990; Gretskii andMaksimov, 1998; Gooch, 1983a,b; Masubuchi, 1981;Abson and Cooper, 1998; West et al., 1990).
In the case of the dry spot welds (Table 6a), thisappeared due to two anomalously high readingsin Traverse 2 on welds D2 and D4 and wasassumed to be the result of increased dilutionwhilst operating on the ‘high’ current setting. Thissituation produced a hotter, more fluid puddle,thereby diluting more carbon from the plate intothe weld pool. This, combined with the switch overfrom ‘high’ to ‘low’ current, effectively limited anyfurther alloying, which together with the effects ofplate cooling caused the formation of martensite.This is evident from the results shown in Table 6bfor 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
122
Vol 28, No 3, 2009
15% Cr
14% Cr
13% Cr
12% Cr
11% Cr
10% Cr
9% Cr
8% Cr
7% Cr
6% Cr
5%Cr
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 stillacceptable under BS EN ISO 15614-1 (2004) andAWSD3.6 (1999) and did show some improvementsover conventional wet MMA fillet welds (Keats,2004, 1990; Masubuchi, 1981; Abson and Cooper,1998; West et al., 1990).
3.4. Macro/microscopic surveyTo better understand what has actually happenedto weld D2 (highest hardness dry weld) and wetweld W3, a series of microphotographs were takento study the microstructures present. Weld D2was also scrutinised under a Cameca SX50 EPMAelectron microscope to map the weld area (seeFigs 9–11). The results for weld D2 showed reducedCr, Ni and Mo levels present in the root areaof the weld, located just at the point where theswitch-over from high to low current took place.This demonstrates that higher dilution occurredin the root area, resulting in higher carbon levels.This factor may account for the observed elevatedhardness readings, despite a slower cooling ratethan weld W3.
Microphotographs for weld D2 apparently con-firmed this effect and showed that higher carbonmartensite existed, as did numerous sphericalcarbide particles. Martensite and carbides wereevident to some degree throughout the whole weld
2.5% Mo
2.3% Mo
2.1% Mo
1.9% Mo
1.7% Mo
1.5% Mo
1.3% Mo
1.1% Mo
0.9% Mo
0.7% Mo
Fig 11: Quantitative map plotted for Mo
body, as was the occasional isolated globular oxide.Both welds, wet and dry, showed a microstructuralsimilarity, with the existence of delta ferrite in anaustenitic matrix together with isolated globularoxides being present.
Weld W3 showed evidence of a small crack in theroot area, which may have been the result of theshear testing, as no other significant metallurgicalfactors were observed that may have caused such adefect. This may be explained by the pronouncedloss of material that occurred from the weldnugget, which appeared to show a ductile break.In addition, it was clear that the material for weldsD2 and W3 were not the same, despite the materialspecification (shown in Table 1).
Weld D2 clearly had a higher carbon content,as shown by the ferrite and pearlite content, whichmay be as high as 0.25% (see Fig 12). Weld W3showed a lower percentage of pearlite (smaller andless carbide platelet formation). This may moreaccurately reflect a material composition closer to0.15% carbon (see Fig 13).
4. DiscussionIn considering weld strength versus weld size, andthus the number of welds required for any givenload-carrying capacity, the following principle tocalculate 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 was86.24mm2 with the average load to failure beingcalculated at 45.63kN, whilst the average ultimatetensile stress (UTS) was 548.5N/mm2. The wettensile results in Table 5 and Fig 3 show the averageCSA for wet spot welds was 97.17mm2 with theaverage load to failure being calculated at 39.95kN,whilst the average UTS was 474.50N/mm2. Thusby comparison, the average dry spot weld CSAwas 10.93mm2 smaller than the average wet spot
123
Keats. Underwater wet welding made simple: benefits of Hammerhead r© wet-spot welding process
a
b
Fig 12: (a) Weld macro of weld D2 (dry weld)conducted by Welder B; (b) parent materialcomprising predominantly ferrite and pearlite
weld, but offered an increase in shear strength ofsome 5.68kN.
By comparing the wet shear test results tothe theoretical based value (10.0mm-diameternugget), which produced a load to failure of40.8kN with a UTS of 408N/mm2 (as shown inFig 4), the actual weld deposited offered a slightreduction of 0.85kN or 2.1%. When calculatingthe reduction in CSA that equated to 2.83mm2,however, the strength reduction became 2.9%.Thus, the design principle that predicts a givennumber of spot welds for a given load would appearto overvalue wet weld strengths by approximately3%. Nevertheless, this approach demonstrates thatsimple calculation would provide a reliable basemethod for determining the number of spot weldsnecessary to carry a given load (see Appendix 5 formore details).
The overall appearance of wet spot welds, ascompared to dry spot welds, was somewhat untidy,with clearly a more restricted weld puddle inevidence. The size and profile (cap) of the weldsdid not appear to significantly affect the results ofmechanical testing.
The average hardness values shown in Tables 6aand 7b for wet and dry spot welds were acceptable,showing no particular hardness concerns. In fact,considering the average values between wet and
a
b
Fig 13: (a) Weld macro of weld W3 (wet weld)conducted by welder C; (b) parent materialcomprising predominantly ferrite with a smallamount of pearlite (carbon content appears to beapproximately 0.10–0.15% based on thepearlite present)
dry (excluding D2 and D4), the difference wasso minimal as to be irrelevant. It should be ap-preciated, however, differences in material carboncontent for welds D2 and W3 could alone besufficient to show a difference in the hardnessreadings obtained. The hardness results shown inTables 6b and 7b for welds D2 andW3 clearly showsa significant difference in the overall hardness onweld metal, with the dry weld (D2) suffering fromincreased carbon dilution as compared to the wetweld (W3).
The weld macros showed that weld quality wassimilar between wet and dry, though not defect-free, and no significant incidence of defects wereproduced wet, as compared to dry. It should alsobe noted that all welds, both wet and dry, hadbeen mechanically tested prior to macro/microexamination and hardness surveys. This may,therefore, have had some effect on the resultsobtained. Nevertheless, the quality of wet spot weldsproduced showed that this method of welding canbe relied upon to produce under water welds, atthe very least, as effective as those described inthe referenced literature for conventional wet fillet
124
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 wasperformed prior to welding, unlike that of con-ventional wet fillet welding, thus welding efficiencywas significantly increased, with a completed weldbeing produced in less than 30 seconds. The controldevice provided a suitable means to control theessential welding parameters and demonstratedthe means to reduce the role of the diver, evenunder nil visibility conditions. The presence ofthe diver is still essential in the production of asatisfactory weld, due to the need to apply adequatepressure. Nevertheless, this welding method hasdemonstrated a successful means of joining carbonsteels that eliminates the need for skilled welders,as well as all conventional cleaning/preparationmethods. Furthermore, successful wet welds wereproduced under conditions of nil visibility.
5. Conclusions and further workTo demonstrate the commercial advantages ofthis process, the experiments concentrated on thefollowing conditions:
1. Producing spot welds in nil visibility, while stillproviding for an effective weld
2. Elimination of preparation/cleaning of materi-als and increased welding speeds
3. Elimination of welding skills4. Repeatability and consistent weld quality.
The experiments demonstrated that the spotwelding method tested was more than capable ofmaking an effective mechanical fixing under water.It also provided benefits in the way of speed, qualityand repeatability over conventional wet MMA filletwelding, without using skilled welders and workingin poor/nil visibility conditions for both dry and wetspot welds.
Dry welds were produced as a baseline com-parison to compare weld quality and highlightany differences in mechanical and metallurgicalqualities. This work was limited to welding lowcarbon structural steel plate (8.0mm) in fresh waterat a depth of 3m using a specially designed controldevice. The experiment had not taken into accountthe possible effects of welding in seawater, nor didit 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 posi-tion. Further work would be required to evaluatethis welding methodology more fully, includingdifferent sizes of electrodes, welding positions andwater type and depth, together with different gradesof structural steel. All under water welding wasconducted in poor and/or nil visibility conditionsusing both skilled and non-skilled welder-divers.Although only a few welds in total were produced,which were insufficient to provide for a compre-hensive outcome, the evidence supports that diverswith little or no welding skills/knowledge were ableto produce acceptable spot-welds just as easily asthe skilled welders. It was shown that visibility hadno effect on performance or weld quality, nor didthe lack of weld preparation or cleaning appear toaffect final weld quality.
Although not a fully automatic welding method,the control device proved suitable to manage thewelding parameters essential to produce qualitywelds repeatedly. It was axiomatic that eachindividual diver must ensure suitable pressure isapplied to the electrode for an acceptable weld tobe produced. The wet spot welds provided suitableweld quality in terms of strength, with propertiesclosely matching those of dry spot welds. Due tothe metallurgy, however, the process is likely to belimited under water to welding non-load bearingjoints (e.g. anodes). During testing, it becameevident that the spot welding method providedfor a considerably faster joining method thanconventional wet MMA fillet welding, as the processdid not require any joint preparation or cleaning ofthe material/weld, so spot welds (wet and dry) wereproduced in a matter of seconds.
The Hammerhead welding process clearly re-mains a manual operation, despite the controldevice, whereas Sadowski’s work (1980) involvedautomatic fixed welding heads working in hand-deep test tanks only. In contrast, the Hammerheadprocess is used by a diver being fully submergedunder water. The welding process is designed as aone-spot process – i.e. one electrode produces onespot weld, thereby eliminating the need to makea second weld over the first, as well as inter-runcleaning and the need for a second pass (Abson andCooper, 1998).
The Hammerhead welding method appears tolend itself to automation and may well prove tobe of great interest, as presently there remainsa level of control required by the diver to applypressure to the electrode during welding. It wasreported by Gooch (1983a,b), Masubuchi (1981),and West et al. (1990) that the use of austeniticelectrodes to produce conventional fillet and buttwelds under water often produced cracking in theweld 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 electrodecracking so extensive that it prevented any usefulmechanical testing from being undertaken.
With the exception of a small micro-crack(≤0.3mm) in zone E on weldW3, no other crackingwas observed in the weld or the HAZ. Althoughthis crack may be metallurgical in nature, it wasmore likely to be as a result of the mechanicaltesting, as evidence exists that the weld underwentsignificant stress with large sections of weldmaterialmissing from the fracture face. Themicrostructuresreported by Abson and Cooper (1998) also statedthat martensite was severe, particularly in theroot area, where contact with the parent materialproduced high dilutions. Nevertheless, martensitewas observed throughout the whole weld body.Although martensite was present in the weld bodyof the Hammerhead spot welds, it was not as severeas reported by Abson and Cooper (1998).
The hardness values reported also showed asignificant increase over those produced by theHammerhead process, whose values (as shownin Table 7b) are considerably lower and thusclearly suggest evidence of an improved weldingprocess. The average shear strength values, asreported by Gooch (1983a,b), suggest that theHammerhead process provided equally effectivemechanical strength properties.
Van der Brink and Boltje (1983) also demon-strated that the moisture content of the elec-trode flux covering was critical to avoid hydrogencracking. The experiments detailed herein wereconducted on a specially developed electrode,manufactured specifically to minimise moisturepickup. Furthermore, the number of electrodestaken into the water at any one time was limited alsoto minimise moisture pickup and thus help preventhydrogen cracking. Szelagowski (1991) reportedthat the type of waterproof coating could also havea significant effect on chemical composition of theweld deposit, though no evidence of this was ob-served on the welds produced in this work. Clearly,further work in this regard would be beneficial.The waterproof coating used for the Hammerheadelectrode was a specially formulated vinyl lacquer.
According to Grubbs (1986), conventional wetfillet welds using ferritic electrodes can be success-fully made in accordance with AWS D3.6 (1999)Class B at depths down to 60m. The opportunityto test the Hammerhead process at these depthswas not available for these experiments, althoughit is accepted that deep welding trials would benecessary to fully evaluate this process/electrodeand investigate any differences in weld quality fromthe shallow water tests undertaken. From the resultsobtained for the Hammerhead process, evidencesuggests that further additional alloying of the
electrodemay be beneficial, in order to help reducemartensite in the central weld body area. This,however, is unlikely to offer any benefits at the jointinterface (root) zone, where the highest dilutionwas recorded.
The work undertaken by Corus (Thompson,2005) demonstrated that the welding methodologyto be valid, even for above water applications,and that it was more than capable of joining awide range of material thicknesses. The processoffered a rapid method of joining plates and sheetsteel in both very thin and thick materials (1.6–15.0mm) and the device-controlled penetrationquite satisfactory. Welding was performed in boththe flat and vertical orientation, and no significantweld defects were reported, though some voids werepresent on thicker materials.
Nevertheless, all joints contained large fusedregions, providing mechanically strong joints. Thewelding of thin galvanized steel sheet providedjoints with high mechanical integrity, with the weldnugget being pulled out from the parent materialwith significant plastic deformation. It was notedthat further work was needed to investigate themax-imum gap tolerance, as no gap was present or pre-set during the initial trials. This is recognised to bean area of extreme interest and significance. (Thefull Corus report is available as a separate report.)
Appendix 1Theoretical value for a spot weld, based on3.141× d2/4× 520
Load (kN) 40.84Table 1: Recorded dry-spot weld values (kN)
1 2 3 4 (Average)Load (kN) 47.8 45.9 52.2 36.6 45.63Table 2: Reduction in wet-spot weld strengthby 14.2%
1 2 3 4 (Average)Load (kN) 41.01 39.38 44.79 31.40 39.15Table 3: Further reduced from table 1 valuesby 23.56%
1 2 3 4 (Average)Load (kN) 36.54 35.09 39.91 27.98 34.88Table 4: Adjusted for weld strength overvalueof 40.84kN by 3%
Load (kN) 39.61
ReferencesAbson DJ and Cooper MA. (1998). Wet under water
welding trials with commercial MMA electrodes. In:Proceedings of the Underwater Wet Welding andCutting Conference, TheWelding Institute (TWI)/PatonInstitute Middlesbrough, 17–18 April, 50–67.
American Welding Society (AWS). (1999). AWS D3.6M-99Underwater Welding Specification. Miami: AWS, 129pp.
126
Vol 28, No 3, 2009
Association of Offshore Diving Contractors (AODC).(1985). AODC 035: Code of Practice for the safe use ofelectricity under water, 135pp.
Bailey N. (1987). Exploratory tests on nickel-based wetwelding electrodes for ferritic steel. TheWelding Institute(TWI) members report 355.
Bailey N. (1991). Welding Underwater – A MetallurgicalAppraisal. Paper presented at the 1st InternationalOffshore and Polar Engineering Conference (ISOPE-91)Edinburgh, 11–15 August.
British Standards Institute/European Standard. (1996). BSEN 1043-1: 1996. Destructive tests on welds in metallicmaterials. Hardness testing. Hardness test on arc weldedjoints.
British Standards Institute/European Standard/Internatio-nal Organization for Standardization. (2004). BS EN ISO15614-1: 2004. Specification and qualification of weldingprocedures for metallic materials. Welding proceduretest. Arc and gas welding of steels and arc welding ofnickel and nickel alloys.
British Standards Institute/European Standard/Internatio-nal Organization for Standardization. (2002). BS ENISO 15618-1: 2002. Qualification testing of welders forunder water welding. Diver-welders for hyperbaric wetwelding.
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 ofphysico-metallurgical peculiarities of wet arc weldingof structural steels. In: Proceedings of the Underwater
Wet Welding and Cutting Conference, The WeldingInstitute (TWI)/Paton Institute Middlesbrough,17–18 April.
Grubbs CE. (1986).Qualification of under water wet weldingprocedures at water depths down to 99M. In: Proceedingsof the 16th International Diving Symposium, HoustonTexas, Session 8, Paper 1, 7pp.
Hibshman NS and Jensen CD. (1933). Underwater ArcWelding. 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 ofthe water environment on the parameters in under waterwet welding. Welding and Cutting 55(2): 97–99.
Masubuchi K. (1981). Underwater Welding – Factors Affect-ing Welding Metallurgy. In: Proceedings of UnderwaterWelding ofOffshore Platforms and Pipelines, 5–6Novem-ber, New Orleans, Louisiana, 81–98.
Sadowski EP. (1980). Underwater wet welding nickel baseand stainless steel electrodes. Welding Journal 59(7):30–38.
Szelagowski P. (1991). Underwater welding – present stateand development trends. Welding and Cutting, 7–11.
Thompson AM. (2005). Welding developments for Special-ity Welds. Corus Research, Development and Technol-ogy. Reference source No. 115034.
Van der Brink SH and Boltje GW. (1983). Proceedings,Conference on UnderwaterWelding. Oxford: PergamonPress.
West TC, Mitchell G and Lindberg E. (1990). Wet weldingelectrode evaluation for ship repair. Welding Journal69(8): 46–56.
127
