Steve G Shi, Paul Hilton, Steve Mulligan and Geert Verhaeghe
Proceedings of the 23rd International Congress on Applications of Lasers & Electro-Optics 2004
The laser-arc hybrid welding process is characterised by the simultaneous application of a focused laser beam and an arc welding source. The combination of sources can provide several benefits, one of which is to provide an increased gap bridging capability, when compared with that available from the laser source alone. Using on-line monitoring of the joint geometry and adaptive control of the hybrid welding parameters, the tolerance to joint gap and fit-up can be increased and a consistent weld quality and profile can be maintained. This paper describes the results of an exploratory study, to perform adaptive hybrid laser-MAG welding, via on-line control of the process travel speed. The work used feedback from joint gap measurements, acquired from a laser vision system, in an attempt to maintain constant weld quality, particularly in terms of weld profile, when welding samples with varying gaps. When hybrid welding was used to produce square-edge butt welds in 8mm thickness C-Mn steel, a stringent weld quality, as defined for class-B of EN-ISO13919-1, could be achieved for joint gaps as large as 1.2mm, under conditions of adaptive control. Without adaptive control, this stringent quality level could be maintained only when the joint gap was less than 0.8mm.
The laser-arc hybrid welding process is generally characterised by the simultaneous application of a focused laser beam and a MIG or MAG torch. In general, the hybrid laser-arc welding process combines the advantages of both laser welding and arc welding, in terms of high welding speed, good weld profile, improved mechanical properties, high process efficiency and the ability to accommodate relatively large gaps. When hybrid welding is compared to autogenous laser welding, the hybrid process can increase the tolerance to joint fit-up variations and broaden the range of tolerance with regard to edge preparation quality. [1,2] Hybrid laser-arc welding technology has already been investigated for applications in the aerospace, automotive, off-road vehicle, shipbuilding, oil and gas and pressure vessel industry sectors, [3-6] and more recently, has entered production for assembly of stiffened structures for shipbuilding and for the construction of underground oil storage vessels. [5,6]
More recent work by Shi and Hilton  on the gap bridging capability of hybrid laser-MAG welding in 8mm thick C-Mn steel found that the hybrid laser-MAG process had a much larger gap bridging capability than either autogenous laser welding or laser welding with filler wire. With the hybrid process, it was possible to weld joint gaps up to 1.4mm, i.e. five times larger than those tolerated by autogenous laser welding, at a similar welding speed. Larger joint gaps could be accommodated by adjusting the wire feed speed or the travel speed, with other welding parameters remaining unchanged. This indicated, that in hybrid welding, either the wire feed speed of the MAG process or the process travel speed, could be used as a potential adaptive control parameter, to compensate for variations in joint gap, whilst maintaining consistent weld quality.
This paper describes the results of a feasibility study, undertaken at TWI, to perform adaptive hybrid laser-MAG welding, via control of the process travel speed. The work used feedback from joint gap measurements, acquired from a laser vision system, in an attempt to maintain constant weld quality, particularly in terms of weld profile, when welding samples with varying gaps.
In this work, welding trials were carried out on 8mm thickness S275 grade C-Mn steel (BS EN 10025). Samples were machined to 300 x 150mm coupons. The steel plate surfaces and edges were milled and degreased prior to welding. A selection of the coupons had their edges machined so that when placed together, the various gap geometries required were produced, as shown in Fig.1.
Fig.1. Joint fit-up arrangements used for adaptive control welding:
1a) Continuously expanding gap
1b) Stepped joint gap changes
1c) Both variable and constant joint gap
A 1.0mm diameter A18 C-Mn steel filler wire was used for the laser welding with filler wire trials and the hybrid welding trials.
For the hybrid laser-arc welding trials, a commercially available shielding gas (a three part mixture, 55% helium, 43% argon and 2% carbon dioxide) was applied through the MAG torch at a flow rate of 15l/min. No gas shielding was provided coaxially with the laser beam but a high pressure air knife was applied transverse to the beam to protect the laser optics.
An integrated hybrid welding system was assembled, consisting of a Trumpf 4kW Nd:YAG laser, a Servo Robot Smart-20 laser vision and joint tracking system, a Kawasaki articulated arm robot and a Lincoln MIG/MAG welding machine.
The laser used was an HL 4006D flashlamp-pumped continuous wave Nd:YAG laser, manufactured by HAAS-LASER GmbH. The laser beam, 1.064µm in wavelength, was delivered via a step-index optical fibre, 600µm in diameter, to an output housing, utilising a 200mm focusing lens to produce a nominal minimum spot size of 0.6mm.
The MIG/MAG torch of the Lincoln PowerWave 450 was attached to the laser output housing, to give an arc travel angle of 30° and a 16mm contact tip to workpiece distance. The output housing and the MIG/MAG torch were mounted onto a Kawasaki JS30 robot, traversing the arrangement over a stationary jig holding the specimens in position during welding. The arrangement can be seen in Fig.2.
Fig.2. Equipment set-up for hybrid Nd:YAG laser-MAG welding with adaptive control
In the integrated system, the sensing and joint tracking device, a 3D laser vision system, mounted with a look ahead distance of 50mm from the centre of the welding laser beam focus, detected both the joint position and the geometry of the gap. The Servo-Robot laser-vision system consists of a Smart-20 laser camera head and a PILOT-LW camera controller and operates using the optical triangulation principle, as illustrated in Fig.3.
Fig.3. Schematic of the operation of the Servo-Robot device
The Smart-20 camera projects a plane of visible light from a 690nm, 10mW laser onto the joint, transverse to the joint line. The reflected laser radiation is imaged onto a photodiode array to provide a joint profile. Signal processing of the image profile is carried out in the PILOT-LW camera controller. The image is divided into segments, with break points defined at features, e.g. inflections in the image, as shown in Fig.4.
Fig.4. Typical joint profile from a square edge butt joint with a fixed gap, measured using the Servo-Robot vision system
For a square edge butt joint, the joint gap is defined as the distance between the two break points across the joint line. The tracking point is at the mid-point of the joint gap. Joint gap and tracking point information is sent to the robot controller with a time stamp which is used by the robot controller to delay the joint gap tracking to account for the look ahead distance of the camera. The tracking point information was used to correct the programmed path during welding, as well as maintaining a constant stand-off distance between the welding head and the workpiece. The joint gap data, on the other hand, was used as source data for adaptive process control using software developed by Kawasaki for the JS30 robot.
All welds were carried out in the flat (PA) position with a workpiece laser power of 4kW, as measured by an OPHIR meter. The samples were tack welded in a jigging arrangement to maintain the required joint fit up and reduce movement of material during welding. The MAG process was operated in the pulsed mode. After welding, all welds produced were inspected visually for the presence of imperfections such as surface breaking porosity, incompletely filled groove, incomplete penetration, excessive undercut, misalignment and cracking.
Initially, hybrid Nd:YAG laser-MAG welding trials were conducted to establish a stable process for welding zero-gap square-edge butt joints resulting in geometrical characteristics acceptable to weld quality class B, further referred to as stringent weld quality, in accordance with BS EN ISO 13919-1:1997.  These conditions were subsequently used on samples with a joint gap that varied between zero at the start and 1.2mm, i.e. slightly larger than the wire diameter used, at the end of the weld, to establish at what joint gap the conditions used resulted in sub-standard welds. From these results, a fixed, parallel joint gap of 1.2mm was then selected and the Nd:YAG laser-MAG welding conditions were re-optimised to produce fully penetrating square-edge butt welds with the same stringent weld quality as achieved for the zero-gap joints. To accomplish the latter, only the welding speed was changed from the initially developed welding parameters.
Information from the above trials was then imported into the Kawasaki JS30 control software and used for adaptively controlling the welding speed, based on joint gap measurements from the SMART-20 sensor. The goal was to achieve fully penetrating welds, of a uniform stringent weld quality, in samples with engineered joint gaps. Adaptive control was carried out in two ways, i.e. stepped or continuous, as shown in Fig.5a and Fig.5b, respectively.
Fig.5. Adaptive control of welding speed based on joint gap:
Fig.5a) Stepped change of travel speed
Fig.5b) Continuous change of travel speed
The first attempt at adaptive control involved a discrete change in the travel speed depending on the range within which the instantaneous measured gap value lay ( Fig.5a). For the range of gaps engineered into the plates used for the adaptive control trials, a control algorithm was developed based on gap width information from the seam tracker. The control software automatically selected one of the pre-established sets of welding parameters (i.e. the travel speed) and applied this to the part of the weld being made, dependant on the gap width seen. This same set of parameters was applied until the seam tracker device measured a gap corresponding to a different parameter set, at which point (in this case) a further change in speed was made.
The second method of adaptive control involved a continuous change of travel speed, as the weld progressed, i.e. with automatic travel speed interpolation between the discrete parameter sets. The travel speed was continuously adjusted based on each instantaneous measurement of joint gap ( Fig.5b).
A schematic drawing showing the principle of the adaptive welding process described above is shown in Fig.6, and the three types of joint preparation used to evaluate the capabilities of the two adaptive control possibilities described above, can be seen in Fig.1.
Fig.6. Principle of hybrid Nd:YAG laser welding with adaptive control
Results and Discussion
Hybrid Welding Conditions for Butts Joints with Constant and Variable Gaps
The following parameters were found to give the best process stability for hybrid Nd:YAG laser MAG welds in 8mm thickness steel plate, using a square edge butt joint, with zero gap, and resulting in a stringent weld quality in accordance with class B of BS EN ISO 131919-1.
- 4kW laser power at work piece.
- 0.6mm diameter spot size.
- 0mm focus position.
- 12 litre/min 55%He-43%Ar-2%CO 2 welding gas through the MAG torch.
- MAG pushing configuration with 1.0mm process separation.
- 1.0mm diameter C-Mn steel welding wire.
- Pulsed MAG mode.
- 4kW MAG power (30V, 168A and 8m/min wire feed speed).
- 0.85m/min welding speed.
A sample with a constant joint gap of 1.2mm was also welded, for which (only) the travel speed was reduced, to produce an acceptable weld profile and quality at this gap. It was found that fully penetrating welds of a stringent weld quality could be made at a speed of 0.72m/min at the1.2mm gap, compared with 0.85m/min for zero gap samples. Figure 7 shows the appearance of a weld made on a sample where the joint gap varied from 0.2mm at the start of the weld to1.2mm at the end of the weld, at a constant travel speed of 0.85m/min.
Fig.7. Appearance of the weld top bead profile produced using welding conditions developed for a zero gap butt joint, but for a sample with a joint gap as depicted in Fig.8a, i.e. increasing linearly from 0.2mm at the weld start to 1.2mm at the weld end
Although it was possible to produce a fully penetrating, square edge butt joint without the weld pool falling through, even at the larger joint gap of 1.2mm, the stringent weld quality could not be maintained over the entire sample length. The geometrical features of the weld profile produced at different positions along the weld, were different. For joint gaps smaller than 0.6mm, transfer of filler material via the arc caused a build-up of metal on top of the weld, leaving the weld with an excess weld metal profile. As the joint gap gradually increased, the amount of excess weld metal was reduced until the excess weld metal became zero, after which the weld top bead exhibited an in completely filled groove. The transition to an incompletely filled groove occurred when the joint gap was around 0.6mm-0.8mm, above which the weld quality fell below that defined as stringent in BS EN ISO 13919-1.
The Adaptive Control Algorithms
Based on the travel speeds obtained earlier for the zero gap butt joints and the butt joints with a constant 1.2mm gap, an adaptive control algorithm was developed for the Kawasaki control software. For the stepped adaptive control, the following relationships were used:
- IF 0<joint gap<1.2mm THEN travel speed = 0.85m/min.
- IF 1.2mm ≤ joint gap THEN travel speed = 0.72m/min.
For the continuous adaptive control, the actual travel speed corresponding to a given joint gap was automatically calculated by linear interpolation between the two travel speeds given in the above relationships. For example, the travel speed for a joint gap of 0.6mm, was 0.79m/min, as shown graphically in Fig.5b.
Stepped Adaptive Control for Hybrid Welding
To test the stepped adaptive control algorithm, a square edge butt sample was machined to provide gaps of 0.2mm, 1.2mm and 0.2mm, as depicted in Fig.1b. The welding parameters developed for the zero gap and the 1.2mm joint gap samples were used as a basis for welding this sample.
Fig.8. The typical weld appearance of a weld produced with stepped changes in travel speed on butt joints with stepped changes in joint gap (see Fig.1b for joint configuration):
8a) Adaptive control parameters
8b) Weld top bead appearance
During welding, travel speed was changed by use of adaptive control. A speed of 0.85m/min was automatically selected for joint gaps measured to be less than 1.2mm and a speed of 0.72m/min was automatically selected for joint gaps measured to be equal to or larger than 1.2mm. For this particular joint configuration, two speed changes i.e. from 0.85m/min to 0.72m/min and from 0.72m/min to 0.85m/min, were made over the length of the weld, as can be seen in Fig.8a. The top bead appearance of the weld produced is shown in Fig.8b. Although it was possible to accommodate the discontinuous joint gaps using this control mode without losing penetration or without the weld falling though, poor weld profiles were observed in the transition region, a scan be seen in Fig.8b, because the hybrid process did not react quickly enough to the stepped change in welding speed initiated by the control system.
Continuous Adaptive Control for Hybrid Welding
Figures 9a and 9b show the appearance, top and underbead respectively, of a weld produced with continuous adaptive control of travel speed, for a sample with a linearly variable joint gap, increasing from 0.2mm at the start of the weld, to1.2mm at the end of the weld.
During welding, the travel speed was decreased uniformly over the 300mm weld length, from 0.85m/min to 0.72m/min, corresponding to the measured joint gap, as shown schematically in Fig.9c. The weld exhibited full penetration and a smooth top and underbead appearance, even at the largest joint gap. The gradual reduction in travel speed as the joint gap increased, increased the amount of filler deposited and as a result, the weld did not exhibit the sunken topbead profile as witnessed in a similar sample welded without adaptive control ( Fig.7).
Fig.9. Appearance of the weld produced using hybrid Nd:YAG laser-MAG welding with continuous adaptive control of welding travel speed to accommodate a continuously varying joint gap of 0.2mm at the start and 1.2mm at the end of a square-edge butt weld in 8mm thickness C-Mn steel:
Fig.9a) Top bead appearance
Fig.9b) Underbead appearance
Fig.9c) Adaptive control parameters
Figure 10 summarises the performance of the continuous adaptive control method with regard to weld quality, in terms of incompletely filled groove and excess weld metal, demonstrating that when using continuous adaptive control of welding speed, stringent weld quality could be achieved, even at the largest joint gap investigated. Without adaptive control, this was not possible for joint gaps larger than 0.8mm. This illustrates the potential of adaptive control in a production environment, where close to zero-gap butt joints, as usually required for laser welding, are often not cost effective or practical to produce.
It is believed that even larger joint gaps than the 1.2mm tested here are weldable using additional adaptive control algorithms. It is also apparent that the application of adaptive control increases the process tolerance to joint gap, when compared to welding without adaptive control.
Fig.10. The measured excess weld metal and incompletely filled groove of welds produced with and without adaptive control of the travel speed. The measured values are compared to values defined in BS EN ISO13919-1:1997.
Figure 11 shows the top and underbead appearance of a weld produced on a sample where the joint gap increased initially, then remained constant for a while, before reducing again to zero at the end of the sample, as shown in Fig.1c. The weld was made using the continuous adaptive control method.
Fig.11. Appearance of the weld produced using hybrid Nd:YAG laser-MAG welding with real time continuous change in travel speed to accommodate continuously variable joint gaps on square-edge butt joints with 0.2-1.2mm and 1.2-0.2mm continuously variable joint gaps in 8mm thickness C-Mn steel:
Fig.11a) Top bead appearance;
Fig.11b) Underbead appearance;
Fig.11c) Adaptive control parameters
This sample was produced to demonstrate the effectiveness of continuous adaptive control to compensate for slowly changing joint gap conditions, as often experienced in a production environment, and to ascertain whether the control algorithm could also be applied to closing joint gaps (as it was developed based on trials on samples with opening joint gaps). For this weld, the travel speed was automatically decreased, from 0.85m/min to 0.72m/min, for the first part of the weld, kept constant at 0.72m/min in the central section and then automatically increased from 0.72m/min to 0.85m/min, at the end of the weld. The resulting fully penetrating weld showed a smooth top and underbead. This proved that the earlier developed continuous control algorithm could also be applied to closing gaps, maintaining a predetermined joint quality.
A hybrid Nd:YAG laser-MAG welding system, employing adaptive control of the process travel speed has been implemented. A Servo Robot vision system was used for on-line measurement of the joint gap which provided feedback to the robot controller for changes to the process travel speed. A series of welds with a range of pre-engineered gaps, larger than those normally tolerated by laser welding, was made using this set-up and two approaches for adaptive control of travel speed were investigated. It was demonstrated that the system could automatically adjust the travel speed in either a stepped or a continuous mode, using on-line measurement of the joint gap. In terms of weld profile, however,better results were obtained by applying the continuous adaptive control method. When hybrid Nd:YAG laser-MAG welding was used to produce square-edge butt welds in 8mm thickness C-Mn steel, a stringent weld quality, as defined for class B of BS EN ISO 13919-1, could be achieved for joint gaps as large as 1.2mm, when applying the continuous adaptive control method. Without adaptive control, this stringent control could be achieved only when the joint gap was less than 0.8mm.
This work was carried out using TWI discretionary research funding. The authors would thank Ashley Spencer for his technical expertise and his assistance in carrying out the hybrid welds.
- Dilthey, U. & Wieschemann, A: (2000) Advantages in laser-arc hybrid welding, in Proceedings of the SOJOM Conference.
- Biffin, J. & Walduck, R. P. (1994) Plasma arc augmented laser welding (PALW), in Proceedings of the Second European Conference on Joining Technology, Florence, Italy, 295-304.
- Biffin, J.(1997) Plasma augmented lasers shine ahead in auto body application, Welding and Metal Fabrication, 4, 19-21.
- Ishide, T., Nayama, M., Nagashima, T. & Takano, G. (1999) Hybrid laser welding technique (GTA-YAG coaxial welding) for nuclear power plant, Proceedings for IIW Conference, Osaka, Japan, IIW Doc. XII-1567-49.
- Kaierle, S., Dahmen, M. & Bongard, K. (2000) Hybrid welding of oil tanks- a new joining process, in Proceedings of Welding and Cutting Conference, Nurnberg, Germany 207-209.
- Thomy, C., Sepold, G., Seefeld, T., Walz, C., Hoffmann, R., Stiebe-Springer, I., Kocak, M. & Segar, R (2002) Industry implementation of laser/GMA welding and mechanical properties of the weld, in Proceedings for. Supermartensitic Stainless Steel Conference, Brussels, Belgium, 147-155.
- Shi, G. & Hilton, P. (2004) A comparison of the gap bridging capability of CO 2 laser and hybrid CO2 laser MAG welding on 8mm thickness C-Mn steel plate, TWI Members Report 798/2004.
- BS EN ISO 13919-1:1997 (1997) Welding - Electron and laser beam welded joints - Guidance on quality levels for imperfection, British Standards Institution.
Meet the Author
Gongqi Shi joined TWI in 2001 with a PhD in Engineering Materials and is now a senior project leader in Laser & Sheet Processes Group of TWI. E-mail: firstname.lastname@example.org. Paul Hilton is the Technology Manager and Geert Verheaghe a senior project leader in the Laser & Sheet Processes Group of TWI. Steve Mulligan is a project leader in the Arcs & Surfacing Group of TWI.