Chris Allen, Steve Shi and Paul Hilton
TWI Ltd., Granta Park, Gt. Abington, Cambridge, CB21 6AL, United Kingdom.
Paper published in ILAS Supplement to AILU magazine, Issue 63, July 2011.
Adaptively controlled hybrid laser-arc welding has been demonstrated using a 5kW 6mm.mrad Yb fibre laser. ISO 13919-1 class B (stringent) quality butt welds have been made in both 8mm thickness steel plate and 6mm stainless steel plate. A laser vision sensor has been used to track the joints robotically during welding. In addition, joint fit-up information from the sensor (gap width and mismatch height) has been used during welding to adjust, in real time, the robot speed, or arc welding parameters. This approach increases the tolerance of hybrid welding, particularly to joint gap, producing stringent quality welds over a wider range of joint fit-up cases than with fixed conditions.
The requirements for precise workpiece fit-up and accurate alignment of a laser beam with the resulting joint line, pre-requisites for high quality welds, continue to act as one of the barriers to the wider take-up of laser welding. These requirements put restraints on welding fixtures and edge preparation methods, and lead to tolerances which can prove more difficult to meet in industrial practice. This could particularly be the case when welding large components or fabrications, where the assembly of a number of sub-components, or a build up of distortion, results in cumulative errors in joint fit-up, even with the best edge preparation. The ongoing emergence of high brightness, fibre-delivered, lasers, with their highly focusable beams, promising ever faster, lower heat input welds, makes attention to such requirements even more critical.
Hybrid laser-arc welding is a potential solution to this problem, accommodating greater differences in joint fit-up than laser welding, due to the wire feed addition from the arc. Nevertheless, maintaining weld quality in an industrial setting can still present a challenge, as the requirements of hybrid welding on part positioning, edge preparation and fit-up are still more than those of arc welding, which can increase material preparation and welding fixtures costs, and raise the likelihood of weld defects.
Joint tracking sensors are already used in industry to increase the tolerance of welding processes, including hybrid welding, to otherwise incorrect joint placement. In this article, the use of a laser vision sensor to relay details of joint fit-up (not only position) to the welding equipment, enabling real time adaptive control of the welding parameters to cope with changes, and maintain weld quality, is examined. Broad root-faced butt joints in 8mm thickness S355 steel, or square edged butt joints in 6mm thickness AISI 304 stainless steel, have been adaptively welded as test cases.
Robotic hybrid welding trials were carried out using a Kawasaki FS-060L robot with a D+ controller, an IPG YLS-5000 Yb fibre laser and ESAB synergic MAG arc welding equipment. Suitable shielding gas/wire combinations of Ar/1.2mm A18 (for steel) or Ar-2%O2/1.2mm 308LSi (for stainless) were used. The weld qualities achieved were evaluated with respect to ISO 13919-1:1997. Table 1 summarises selected criteria for Class B (the most stringent class) from this standard.
Table 1 ISO 13919-1 Class B (stringent) weld quality criteria
| ||6mm thickness||8mm thickness|
||Limits for weld quality Class B
|Cracks and crater cracks
||Maximum diameter ≤1.8mm Area fraction ≤0.7%
||Maximum diameter ≤2.0mm Area fraction ≤0.7%
|Excess weld metal
|† In this work, linear misalignments up to 2mm were addressed deliberately as a stretch target.
A Servo-Robot Digi-I/S laser vision sensor, as shown in Figure 1, with a V300 Robonet/Master control unit, was used in the adaptively controlled trials and calibrated with respect to the robot. Welds were made over joints with varying amounts of joint gap and/or joint mismatch, with welding conditions being changed in real time. Following trials making fixed changes to welding parameters, to identify which were most appropriate, real-time changes were made by relaying joint fit-up information from the sensor to response programs (in the form of look-up tables) programmed within the V300 or D+ controllers.
The example in Figure 2 shows how such a table may drive, sending a suitable command voltage signal, an increase in wire feed rate, when a joint gap over a given width is detected.
Figures 3a and 3b show, for two material thicknesses, cross-sections of the Class B hybrid welds made over close fitting, flush, butt joints with the optimum conditions developed in the welding trials.
These conditions were also used to weld joints with a mismatch up to 2mm high, or a gap up to 2mm wide. Selected cross-sections are shown in Figure 4. Class B weld profiles were achieved up to a mismatch of 0.6mm (Figure 4a) or a gap of 0.3mm (Figure 4e) when welding the steel, or 0.5mm (Figure 4c) and 0.5mm (Figure 4g), respectively, when welding the stainless steel.
a) 0.6mm mismatch: Class B,
b) 1.0mm mismatch: re-entrant root,
e) 0.3mm gap: Class B,
f) 0.6mm gap: Class C
c) 0.5mm mismatch: Class B,
d) 0.8mm mismatch: re-entrant root,
g) 0.5mm gap: Class B,
h) 0.6mm gap: Class C
The root toe blend angles became re-entrant on the higher plate if the mismatch increased (Figures 4b and d). Conversely, the cap underfill became unacceptably deep if the gap became greater (Figures 4f and h).
One or more fixed changes in the welding conditions were then tried, to improve the weld profile along joints with mismatch or gap, including:
- Reducing the welding speed
- Increasing the wire feed rate and/or arc voltage trim
- Changing the stand-off height of the welding head
- Deliberately off-setting the laser beam off of the joint line (by off-setting the welding head)
The optimum results from these trials are summarised in Tables 2 and 3, for the 8mm steel and 6mm stainless steel welds, respectively.
Table 2 Optimum changes in welding parameters to increase fit-up tolerances suggested from fixed change trials on 8mm steel
|Joint fit-up parameter||Optimum change(s)||Effect|
||Welding speed reduced to 1.2m/min, and head stand-off by 4mm
||Tolerance limit increases from 0.6 to 0.8mm
||Welding speed reduced to 1.2m/min, and wire feed rate increased (e.g. to 9m/min)
||Tolerance limit increases from 0.3 to 0.5mm
Table 3 Optimum changes in welding parameters to increase fit-up tolerances suggested from fixed change trials on 6mm stainless steel
|Joint fit-up parameter||Optimum change(s)||Effect|
||Welding speed reduced to 2m/min
||Tolerance limit increases from 0.5 to 0.7mm
||Wire feed rate increased (e.g. to 13m/min)
||Tolerance limit increases from 0.5 to 1mm
As Tables 2 and 3 summarise, it appeared that changing the welding conditions would give rise to modest improvements in mismatch tolerance and would at least double the tolerance to joint gap. To confirm this, appropriate control algorithms were programmed in to the process controllers, and adaptively controlled trials carried out on butt joints, where the joint preparation, in terms of gap and mismatch, increased linearly as the weld progressed. As the seam tracking device detected changes to the weld profile, an automatic change to a selection of the appropriate welding parameters was implemented. The most successful results are summarised in Figure 5. The benefits (in terms of increased tolerance) gained by using adaptive control can be seen by comparing Figures 4 and 5.
5a) 1.2mm mismatch tolerated, by reducing welding speed to 1.2m/min
5b) 0.6mm gap tolerated, by reducing welding speed to 1.2m/min, and increasing wire feed rate to 13m/min
5c) 1.3mm mismatch tolerated, by reducing speed to 2m/min
5d) 1.3mm gapm by increasing wire feed rate to 13m/min
Figure 6 shows the extent to which tolerances have been increased for making either Class B or C welds, including those over a combination of mismatch and gap, in the case of the 8mm thickness steel. In Figure 6, the red and green boxes indicate the estimated tolerances of autogenous and hybrid welding without adaptive control, respectively. Bold symbols are from validation experiments carried out in this work. The dotted lines represent the limits for Class B and Class C welds.
- The tolerance of high brightness hybrid laser-arc butt welding to joint gap and mismatch in 6-8mm plate, whilst still producing the most stringent class of welds to ISO 13919-1:1997, is ~0.3-0.5mm and ~0.5-0.6mm, respectively, if fixed conditions are used.
- However, off-line trials can identify those parameters which, if changed, can increase these tolerances.
- A laser vision sensor can then successfully relay the fit-up details to process controllers, pre-programmed to respond with these changes, to maintain weld quality over a wider range of gaps and mismatches. Using this approach, in this work, these tolerances have been doubled for the production of Class B welds.
The research leading to these results has received funding from a combination of the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 222289, and Industrial Members of TWI, as part of its Core Research Programme.