C M Allen 1 , C H J Gerritsen 2 , Y Zhang 1 , J Mawella 3
1 TWI Ltd., Granta Park, Gt. Abington, Cambridge CB21 6AL, United Kingdom.
2 Formerly at TWI Ltd., now at OCAS NV (Arcelor), Industry Research Centre, Arcelor Innovation - R&Dmp;D, OCAS NV, John Kennedylaan 3, 9060 Zelzate, Belgium.
3 Defence Procurement Agency, Ministry of Defence, Abbeywood, Bristol BS34 8JH, United Kingdom.
Paper presented at the IIW Commission IV / XII, Intermediate Meeting, Vigo, Spain, 11 - 13 April 2007
Hybrid laser-arc welding, originally proposed in the 1970s, but more recently of renewed interest, combines deep penetration keyhole laser welding with an arc welding process in a single process zone. The hybrid process offers the benefits of the separate processes, and overcomes some of their respective drawbacks, such as the lower tolerance to joint fit up of laser welding, or the higher heat input of arc welding, which increases distortion and subsequent re-work costs. These costs have been estimated to be up to 15-30% of the total labour cost for new construction. Consequently, hybrid welding is already applied in some industries, including shipbuilding, as well as being extensively researched for a variety of materials and industrial applications. For hybrid welding of steel plates, high power output CO2 lasers have typically been used.
This paper describes the development of welding procedures using lower power output CO2 and Nd:YAG lasers, for a selection of DH36, D36 and S275 grade steels. The fit up gaps bridged using a given welding procedure with these two hybrid welding processes have also been determined. Procedures have been developed to make butt welds in flat plate up to 8mm thickness, and, in one case, for a T joint geometry between a 6mm thickness stiffener and an 8mm thickness baseplate. In addition, selected results for weld profile and internal quality, weld hardness, Charpy impact testing, cross-weld static strength, weld fatigue endurance, and weld metal chemical homogeneity are presented.
A wide variety of hybrid processes exist, depending on the laser source used, and the arc welding process with which it is combined. The laser-arc combination used is dictated by the material and application. The hybrid process offers the benefits of the separate processes, and overcomes some of their drawbacks, namely:
- Lower heat input and thereby reduced distortion and rework when compared with conventional arc welding, and even sometimes in comparison with autogenous laser welding or laser welding with filler wire.
- Better tolerance to joint fit-up than with autogenous laser welding.
- Improved penetration and higher joint completion rates than typically achieved with arc welding.
- Lower capital investment and running costs than for autogenous laser welding, as in some applications some of the laser power may be replaced by cheaper arc power.
This paper describes the development of hybrid welding procedures, and illustrates the fit up gaps that can be tolerated when using a given hybrid welding procedure, with a laser source of lower power output than the typical>10kW CO2 lasers used for hybrid welding of steels in industrial applications. Such lower power output lasers are attractive due to reduced capital investment and running costs. Lower power outputs are also more typical of industrial fibre delivered Nd:YAG lasers, which are more readily capable of welding complex geometries than CO2 lasers. This is notwithstanding the recent advent of Yb fibre lasers, which offer both fibre delivery, and high powers comparable to CO2 lasers.
In this work, a 5kW CO2 laser and 4kW Nd:YAG laser were used in hybrid welding procedure development for a selection of DH36, D36 and S275 grade steels. A number of issues were addressed, specifically:
- The capabilities of these lasers to produce welded joints in plate thicknesses up to 8mm, with single sided access and in a single pass.
- The fit up gaps bridged by the welding procedures developed.
- The resulting weld qualities, with levels of imperfections in accordance with the intermediate weld quality class C defined in BS EN ISO 13919-1:1997.
- The resulting weld properties, including hardness determinations in the weld metal and heat affected zone (HAZ), cross-weld static tensile testing, Charpy impact testing and fatigue endurance.
- The chemical homogeneity of the weld metal through thickness.
Three different joint geometries were studied: laser cut, linear butt joints in 4mm and 8mm thickness plate, and a T butt joint between a 6mm thickness roll-formed stiffener and 8mm thickness baseplate.
The following materials were hybrid laser-MAG welded:
- Zinc silicate primed, 4mm thickness DH36 grade shipbuilding steel plate.
- Zinc silicate primed Holland-profile stiffeners of either DH36 or D36 grade steel in 6mm thickness.
- Zinc silicate primed, 8mm thickness DH36 grade shipbuilding steel plate.
- As rolled finish, un-primed 8mm thickness S275 structural steel grade steel plate.
Plates were typically 250-1000mm long, and 125-250mm wide. Stiffeners for T butts were of a matching length.
All plate edges were used as laser cut by the material suppliers. All butt joints had simple square edge joint preparations with no bevel. For butt joints between primed plates, the plates were acetone degreased prior to welding. For butt joints between as-rolled plates without primer, the plate surfaces were ground, and then acetone degreased. For T butts between primed stiffeners and base plates, any primer in the joint region was removed by grinding, again followed by acetone degreasing. Failure to remove the primer from the T butt joint region was demonstrated in early trials to lead to unacceptable levels of porosity in the weld metal.
In order to determine the gap bridging tolerances of the hybrid welding processes being studied, in most experiments fit up gaps of either 0.5mm or 1mm in width were pre-set prior to welding.
Three different consumable wires were used in the MAG process. A C-Mn solid steel wire of 1mm diameter was used for the majority of butt joints. For the T butt welds, wire diameters of both 1mm and 1.2mm were used. In selected butt welds, a 1mm diameter 308L grade stainless steel wire was also used.
The top bead shielding gas mixture used was 55%He-43%Ar-2%CO2 , supplied through the MAG shroud. The top bead shielding gas used in experiments with 308L wire was Ar-5%CO2 -2%O2 . Shielding gas flow rates were in the range 12-15l/min. No under bead shielding was used for any of the welds.
Equipment and Equipment Set-up
For the CO2 laser-MAG hybrid welding experiments, a Laser Ecosse AF8 fast axial flow CO2 laser was used, with a power at work piece of 5kW as measured using a Primes power meter. The beam was focused using a 250mm focal length parabolic mirror, giving an estimated spot size of 0.6mm in diameter atthe focal plane.
For Nd:YAG-MAG hybrid welding experiments, a Trumpf HL4006D flashlamp-pumped Nd:YAG laser was used. Following on from previous hybrid welding experience at TWI using an Nd:YAG laser, a power at work piece of 3.7kW was used. The beam was focused using a 200mm focal length lens, giving an estimated spot size of 0.6mm in diameter at the focal plane.
For hybrid welding experiments, a Fronius TPS450, a Fronius Time Twin, or Lincoln Electric Power wave 455 arc welding power source was used. These MIG/MAG power sources were all used with their respective synergic pulsed metal transfer modes of operation. Standard synergic pulsed metal transfer welding programmes most suited to the filler wires and shielding gas compositions being used were chosen. They were used with a push wire feeder and a straight-necktorch. Fig.1 shows a typical experimental set up.
Fig.1. Example of a hybrid welding head set up, in this case for CO2 laser-MAG hybrid butt welding
In experiments using the CO2 laser source, the work was traversed with respect to the source using an Aerotech x,y manipulation table. In experiments using the Nd:YAG source, the laser focussing head was mounted on a 6-axis Kawasaki JS30robot, and traversed with respect to the work. In both cases, the laser optics were protected by fume and spatter by using a high pressure air-knife.
The main experimental variables were travel speed and wire feed speed. The ranges of travel speed and wire feed speed investigated depended on each of the joint types considered, but were typically 0.7-2m/min, and 3-13m/min,respectively. Other potential experimental variables in the hybrid processes were fixed on the basis of previous experience, such as laser and MAG torch work and travel angles, and the separation between the laser and the MAGtorch.
In addition to acceptable weld profile and internal quality, additional weld procedure qualification tests were carried out on selected welds, including:
- Weld metal and associated heat affected zone (HAZ) Vickers hardness measurements, using a diamond indenter with 5kg load.
- Tensile testing.
- Charpy impact testing.
- Fatigue endurance. Stress ranges were chosen to give fatigue lives in the range 105 - 2x106 cycles. 'Dog bone' shaped samples were used for the butt joints, and parallel sided samples for the T joints, as shown in Fig.2. Stresses were applied perpendicular to the welding direction in both cases, in the plane of the welded plates (thus load carrying welds) for the butt joints, and of the baseplate (non-load carrying weld) for the T joint.
Fig.2. Details of fatigue testpiece dimensions, for butt weld (left) and T butt weld (right), showing direction of stresses applied
Inefficient mixing of the filler wire addition in the hybrid weld pool is also a concern in hybrid welding. For example, in certain welds inefficient mixing may be one cause of through thickness variations in weld metal microstructure, with adverse effects on weld toughness.  Similarly, in welds made with a deliberately dissimilar wire addition, poor mixing would lead to a variation in weld metal composition through thickness, with potential metallurgical consequences.  Poor mixing may also act to restrict the gap bridging capability of a given welding condition in the root of the weld, by restricting supply of additional molten metal from the filler wire addition. Given these concerns, as an experimental test case, the degree of mixing of the MAG wire in the hybrid weld pool in welds made using the Nd:YAG laser was determined in two specific cases:
- a partial penetration melt run made on 8mm thickness S275 plate using a dissimilar 308L stainless steel wire, and
- a full penetration butt weld made using 8mm thickness S275 plates with a 0.5mm root gap, using the same wire.
The weld metal chemical homogeneity through thickness was determined by analysing variations in estimated elemental concentrations of a deliberately added tracer element (Cr, from the 308L stainless steel wire MAG wire addition),using an energy dispersive X-ray (EDX) probe.
Results and Discussion
Welding Procedure Development
The procedures were developed primarily on the basis of visual appearance, to reduce any instances of undercut or poor weld face and toe blends, followed by sectioning and radiography to confirm that the welds produced were to the intermediate quality class C as defined in BS EN ISO 13919-1:1997. Subsequent additional quality requirements were also specified by the end customer in selected cases. Table 1 summarises the optimum conditions for the welding procedures developed, along with the fit up gaps used as appropriate, with the corresponding laser sources and joint geometries used.
Table 1. Optimum values of experimental variables determined, for the different joint types and fit up gaps considered.
|Joint type||Welding process||Fit up gap, mm||Travel speed, m/min||Wire feed speed, m/min|
|Butt, 4mm, DH36 plate
||CO2 - MAG
|Butt, 4mm, DH36 plate
||CO2 - MAG
|Butt, 8mm, S275 plate
||Nd:YAG - MAG
|Butt, 8mm, DH36 plate
||CO2 - MAG
|T, 6mm, D36 or DH36 stiffener on 8mm DH36 baseplate
||CO2 - MAG
|Notes: a Using a 1mm diameter wire.
b Using a 1.2mm diameter wire.
Examples of cross-sections of butt welds made using these conditions are shown in Figs.3-6, and of a T butt weld made in Fig.7. The influence or otherwise of weld profile features on fatigue endurance are reported below.
Fig.3. Transverse cross-section of a CO2 laser-MAG hybrid butt weld in 4mm thickness DH36 with a joint gap of 0.5mm. Scale in millimetres
Fig.4. Transverse cross-section of a CO2 laser-MAG hybrid butt weld in 4mm thickness DH36 with a joint gap of 1mm. Scale in millimetres
Fig.5. Transverse cross-section of an Nd:YAG laser-MAG hybrid butt weld in 8mm thickness S275 with a joint gap of 0.5mm. Scale in millimetres
Fig.6. Transverse cross-section of a CO2 laser-MAG hybrid butt weld in 8mm thickness DH36 with a joint gap of 0.5mm. Scale in millimetres
Fig.7. Transverse cross-section of a single-sided CO2 laser-MAG hybrid T butt weld between a 6mm thickness D36 vertical stiffener, and a 8mm thickness DH36 base plate, with zero joint gap. Scale in millimetres
Of particular note during procedure development were the findings with regards to fit up gap. Using 3.7kW of Nd:YAG laser power at work piece, it was not possible to produce a consistently fully penetrating weld without using at least a 0.5mm root gap. Conversely, using 5kW of CO2 laser power at work piece, a fit up gap of 1mm was not consistently bridged in 8mm, with lack of sidewall fusion defects being observed in some of the radiographs of welds made. This indicates there exists a range of fit up gaps that can be consistently penetrated and bridged for a given laser power and beam quality, although the full extent of this range, and its variation with laser source used, was not determined in this work. This also suggests that for welding over fit up gaps much in excess of 0.5mm in 8mm, one or more of the following are required:
- a laser power in excess of 5kW is used, or
- seam tracking is used, to ensure exact following of the joint line, or
- sensors are used to adaptively control and modify welding parameters, this having been demonstrated to cope with variable fit up gap  , or
- tight restrictions on edge quality and preparation are specified.
Of these requirements, the last is the least attractive from point of view of industrial implication, and the first from point of view of cost.
Weld procedure qualification tests
The results of selected weld procedure qualification tests performed are summarised in Table 2.
Table 2. Selected weld procedure qualification tests results, for the different joint types considered.
|Joint type||Welding process||Maximum hardness, HV5, and location||Ultimate tensile strength, MPa (mean of three), and failure location||Charpy impact energy, J, at -20°C|
|Butt, 4mm, DH36 plate, 0.5mm gap
||CO2 - MAG
|Butt, 4mm, DH36 plate, 1mm gap
||CO2 - MAG
|Butt, 8mm, S275 plate, 0.5mm gap
||Nd:YAG - MAG
|Butt, 8mm, DH36 plate, 0.5mm gap
||CO2 - MAG
||388 a , HAZ
||51J b mean in weld metal
54J c mean in HAZ
|T, 6mm D36 or DH36 stiffener on 8mm DH36 baseplate
||CO2 - MAG
|Notes: - indicates not determined.
a Lloyds Guidelines indicate that only one value may be in excess of 380HV5, which was the case in these welds.
b These samples did not exhibit fracture path deviation.
c These samples exhibited fracture path deviation.
On the basis of the results presented in Table 2, the welding procedures were considered acceptable by the end customer and additional fatigue testing was carried out on selected CO2 -MAG welded butt and T joints. Fatigue test results are shown in Figs.8-10. As is standard practice in the reporting of fatigue data, minor corrections to the raw data were applied, to correct for specimen misalignment and bending during testing. For reference, the design life and mean life of relevant full penetration arc welded butt joint fatigue classes as specified in BS 7608 are also shown. Class D welds are shop welds made in flat position using specific arc welding processes. Class E welds are out of position welds made by other processes, including submerged arc. Class F welds are welds made on a permanent backing strip, and certain classes of fillet welds. Therefore, a high quality welded butt joint would be expected to be to class D, and a high quality T butt joint to class F or better.
Fig.8. Fatigue test results for the hybrid CO2 -MAG butt welds in 4mm thickness DH36 plate, for fit up gaps of both 0.5mm and 1mm, plotted against the local stress range calculated after corrections for specimen misalignment and bending distortion
Fig.9. Fatigue test results for the hybrid CO2 -MAG butt welds in 8mm thickness DH36 plate, for a fit up gap of 0.5mm, plotted against the local stress range calculated after corrections for specimen misalignment and bending distortion
Fig.10. Fatigue test results for the hybrid CO2 -MAG T butt welds between a 6mm thickness vertical stiffener and an 8mm thickness baseplate, with zero fit up gap, plotted against the local stress range calculated after corrections for specimen misalignment and bending distortion
As Fig.8 shows, the hybrid welded simple butt joints in 4mm thickness plate are equivalent to, or better than, class D arc welds. As Fig.9 shows, the hybrid welded simple butt joints in 8mm plate are of similar fatigue endurance class at low stress ranges. A slight deterioration in performance was observed at higher stress ranges, yielding consequently shorter fatigue lives. As Fig.10 shows, the hybrid welded T joints are equivalent to, or better than, class F arc welds. In fact, these T butt welds approach class E performance. There is again a slight deterioration in performance at high stress range.
All fracture faces were examined, to determine the fatigue failure initiation site. Butt welds made in 4mm plate with a 0.5mm gap predominantly failed from the toes of the weld cap. Butt welds made in 4mm plate with a 1mm gap failed either from the toes of the weld cap or weld root, in approximately equal numbers. Butt welds in 8mm plate with a 0.5mm gap again failed predominantly from the toes of the weld cap. T butt welds failed exclusively from the toes of the weld root. This last result was in line with the steeper toe blend angles on the weld root side compared to the weld cap, as shown in Fig.7.
Weld metal analyses
To address the concerns previously referred to regarding weld metal chemical homogeneity, selected Nd:YAG-MAG hybrid welds were made using a dissimilar wire combination. Fig.11 and Fig.12 show cross-sections of a partial penetration melt run and a full penetration butt weld respectively, and the corresponding variations through thickness in Cr. In these experiments Cr effectively served as a tracer element, as it was only present in the 308L wire, and was absent in any significant quantity from the S275 plate. As Fig.11 and Fig.12 show, no significant through thickness variations in the weld metal composition were detected. This suggests that full mixing of the filler addition was taking place during hybrid welding.
Fig.11. Transverse cross-section of an Nd:YAG laser-MAG hybrid melt run on plate in 8mm thickness S275 (scale in millimetres), made using 308L stainless steel wire, and variation in ZAF corrected estimations of Cr content in wt%, at various positions through the thickness of the weld metal. These measurements are repeatable with a relative error of ±4%
Fig.12. Transverse cross-section of an Nd:YAG laser-MAG hybrid butt weld in 8mm thickness S275 (scale in millimetres), made using 308L stainless steel wire, and variation in ZAF corrected estimations of Cr content in wt%, at various positions through the thickness of the weld metal. These measurements are repeatable with a relative error of ±4%
Hybrid laser-MAG welding procedures have been developed successfully for single sided butt and T butt welding of C-Mn steels up to 8mm in thickness. In this work, a given set of welding parameters has been shown to bridge a fit upgap of up to 1mm in 4mm plate, and up to 0.5mm in 8mm plate. However, separate work with adaptive control  has shown that fit up gaps of up to 1.2mm in 8mm plate can be bridged successfully. The procedures reported in this paper achieve high strength welds, with a weld quality acceptable to at least the intermediate class C of BS ENISO 13919-1:1997, and, where appropriate, hardness and Charpy impact toughness acceptable to Lloyd's Register Guidelines for the approval of CO2 laser welding.
The fatigue performance of selected welded joint types is to, or better than, the equivalent arc welded fatigue endurance class, albeit with a slight deterioration in performance at higher stress ranges. In the butt joints tested inthis work, fatigue failures initiated from weld cap toes, or a mixture of weld cap and weld root toes. The T butt joints all failed from the toes of the weld root, emphasising the need to achieve smooth toe blends in the weld root for maximum fatigue endurance.
The weld metal composition through thickness of selected welds made using a dissimilar stainless steel wire was uniform. This result is of particular importance when hybrid welding steels that rely on a MAG wire addition to achieve a specific weld metal microstructure or composition, for example, to meet toughness requirements.
- P L Moore, E R Wallach and D S Howse, 'Development of laser, and laser/arc hybrid welding for land pipeline applications', paper presented at 2nd Materials Research Conference organised by the Younger Members Committee of the IOM3, London, 17-18 February 2003.
- C M Allen, G Verhaeghe, P A Hilton, C P Heason, P B Prangnell, 'Laser and hybrid laser-MIG welding of 6.35 and 12.7mm thick aluminium aerospace alloy', paper presented at International Conference on Aluminium Alloys (ICAA 10), Vancouver, 9-13 July 2006.
- S G Shi, P A Hilton, S J Mulligan and G Verhaeghe, 'Hybrid Nd:YAG laser-MAG welding of thick section steel with adaptive control', Proceedings of the 23rd International Congress on Applications of Lasers & Electro-Optics, 2004.