C M Allen, P A Hilton and J Blackburn
TWI Ltd, Granta Park, Gt. Abington, Cambridge, CB21 6AL, United Kingdom
Copyright © TWI Ltd 2012
Paper presented at 37th International MATADOR Conference, Manchester, England, 25-27 July 2012.
The original publication is available at: www.springerlink.com
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 4mm Al alloy, 6mm stainless steel and 8mm thickness steel plates. A laser vision sensor, used to track the joints robotically during welding, sends joint fit-up information (gap width and mismatch height) to a controller. This adjusts, in real time, the welding parameters to increase the net tolerance of hybrid welding, particularly to joint gap. Stringent quality welds can be made over a wider range of joint fit-up cases than using fixed conditions.
Laser welding requires precise workpiece fit-up and accurate alignment of the beam with the joint line, if high quality welds are to be made. The emergence of high brightness, fibre-delivered, lasers, with ever tighter focusable beams, makes attention to these requirements even more critical. Nevertheless, the demands that these requirements put on welding fixtures, edge preparation methods and fit-up tolerances can act to dissuade potential users of laser welding technology, particularly if large components or fabrications, assembled from a number of smaller sub-components, are to be welded.
Hybrid laser-arc welding offers a potential solution, being more tolerant to joint fit-up, owing to the wire feed addition from the arc. Nevertheless, notwithstanding the benefit of higher welding speeds with the hybrid process, this tolerance does still not match that of arc welding. In addition, material preparation and welding fixtures costs are higher, if weld defects are to be avoided.
In the current work, the capacity of a laser vision sensor to relay details of joint fit-up to the welding equipment, enabling real time adaptive control of parameters to cope with changes in joint fit-up, has been evaluated. Square edged butt joints in 4mm Al alloy or 6mm stainless steel, as well as 8mm thickness S355 steel (in this case with a broad root-faced butt joint configuration) have been adaptively welded as test cases. Previously, adaptive control of hybrid welding has only been reported for butt joints between steel plates [Shi and Hilton, 2005, Shi et al, 2005] using CO2 or Nd:YAG lasers.
1.2 Experimental method
Robotic welding trials were carried out using a Kawasaki FS-060L robot with an IPG 5kW Yb fibre laser and ESAB synergic MIG/MAG arc welding equipment. Suitable shielding gas/wire combinations of Ar/1.2mm A18 (for steel), Ar-2%O2/1.2mm 308LSi (for stainless) or Ar/1.2mm 5356 were used. The weld qualities achieved were evaluated with respect to ISO 13919-1 or -2. For the highest quality (class B) butt welds, in a plate thickness, t, these standards require:
- That the welds are free of cracks;
- The porosity content is ≤0.7% (steels) or ≤3% (Al), and of a maximum diameter ≤0.3t (max. 2mm);
- Weld root and cap undercut is ≤0.05t (max. 0.5mm (steels) or 1mm (Al);
- Excess penetration (root) and excess weld metal (cap) ≤0.15t+0.2mm (max. 5mm).
Test welds were made along 300mm long joints, with either perfect fit-up, or with varying amounts of joint gap (0-2mm) and/or joint mismatch (0-2mm). Following trials making fixed changes to welding parameters, to identify which gave the greatest fit-up tolerance, real-time adaptive changes in the robot speed or position, or the arc welding parameters, were made automatically, by feeding joint fit-up information, from a Servo-Robot Digi-I/S seam tracking sensor as shown in Fig.1.1, to control programs coded in to the welding hardware.
Fig.1.1. Laser vision sensor, ahead of hybrid welding head. Reproduced courtesy of TWI Ltd.
Figs.1.2, 1.3 and 1.4 show cross-sections of the Class B weld profiles made over close fitting, flush, butt joints, with the optimum conditions developed at the start of the welding trials for each of the three materials. These welds were made at 4.5, 2.5 and 1.6m/min, respectively; using wire feed rates of 9, 8 and 7m/min.
Fig.1.2. Optimum hybrid weld in 4mm Al Alloy. Reproduced courtesy of TWI Ltd.
Fig.1.3. Optimum hybrid weld in 6mm stainless steel. Reproduced courtesy of TWI Ltd.
Fig.1.4. Optimum hybrid weld in 8mm steel. Reproduced courtesy of TWI Ltd.
Joints with gaps of up to 2mm in width, and/or mismatches (or hi/lo) of up to 2mm in height, were then welded with these same sets of conditions. This demonstrated that these conditions were only optimum for welding flush, close fitting joints. Figs.1.5, 1.6 and 1.7 (not to same scale) show selected cross sections from these experiments on the Al alloy, stainless steel and steel, respectively. Class B weld profiles were achieved up to a mismatch of 0.3mm or a gap of ~0.3mm, when welding the 4mm Al alloy, 0.5mm or 0.5mm, respectively, when welding the 6mm stainless steel, and 0.6mm or 0.3mm, respectively, when welding the steel.
Fig.1.5. Al Alloy welds over a, b. mismatches and c, d. gaps. Reproduced courtesy of TWI Ltd.
(a) 0.3mm hi/lo: Class B;
(b) 0.5mm hi/lo: lack of sidewall fusion;
(d) 0.8mm gap: excess cap underfill
Fig.1.6. Stainless steel welds over a, b. mismatches and c, d. gaps. Reproduced courtesy of TWI Ltd.
(a) 0.5mm hi/lo: Class B;
(b) 0.8mm hi/lo: re-entrant root;
Fig.1.7. Steel welds over a, b. mismatches and c, d. gaps. Reproduced courtesy of TWI Ltd.
(a) 0.6mm hi/lo: Class B;
(b) 1.0mm hi/lo: re-entrant root;
(d) 0.6mm gap: Class C
As Figs.1.5-1.7 show, the weld cap underfill became unacceptably deep as the gap increased. Conversely, as mismatch increased, the weld root toe blend angles became re-entrant on to the higher plate and/or radiography detected lack of sidewall fusion.
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 1.1-1.3, for the three materials, respectively.
Table 1.1. Optimum changes for 4mm Al alloy
|Reducing head stand-off by 2mm, with increasing mismatch
||Lack of fusion avoided but root still re-entrant on higher plate
|Wire feed rate increased (e.g. to 13m/min), with increasing gap
||Tolerance limit increases from ~0.3 to 1mm
Table 1.2. Optimum changes for 6mm stainless steel. Reproduced courtesy of TWI Ltd.
|Welding speed reduced to 2m/min, with increasing mismatch
||Tolerance limit increases from 0.5 to 0.7mm
|Wire feed rate increased (e.g. to 13m/min), with increasing gap
||Tolerance limit increases from 0.5 to 1mm
Table 1.3. Optimum changes for 8mm steel. Reproduced courtesy of TWI Ltd.
|Welding speed reduced to 1.2m/min, and head stand-off by 4mm, with increasing mismatch
||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), with increasing gap
||Tolerance limit increases from 0.3 to 0.5mm
These results suggested that appropriate changes in welding conditions would double, at least, the tolerance to joint gap, and give rise to more modest improvements in mismatch tolerance when butt welding the steel and stainless steel plates. The mismatch tolerance of the aluminium butt joints did not appear to be improved, but reducing the welding head stand-off did at least avoid lack of sidewall fusion defects.
To confirm the positive effects of these changes, adaptively controlled trials were carried out on butt joints where the joint preparation, in terms of gap and mismatch, increased linearly as the weld progressed. To accomplish this, appropriate control responses were programmed in to the welding equipment controllers connected to the Kawasaki robot and the arc welding equipment. These control programs then implemented appropriate changes in the welding parameters automatically, as a function of the joint gap and mismatch detected by the Servo-Robot seam tracking device.
Some of the most successful results (not to same scale) are summarised in Fig.1.8 (for Class B welds, unless indicated otherwise).
Fig.1.8. Adaptively controlled butt welds in a, b: 4mm Al alloy, c, d: 6mm stainless steel and e, f: 8mm steel. Reproduced courtesy of TWI Ltd.
(a) Re-entrant root over 1mm mismatch but lack of fusion avoided, by reducing head stand-off by 2mm;
(b) 0.8mm gap tolerated (Class C), by increasing wire feed rate to 13m/min;
(c) 1.3mm mismatch tolerated, by reducing speed to 2m/min;
(d) 1.3mm gap, by increasing wire feed rate to 13m/min;
(e) 1.2mm mismatch tolerated, by reducing welding speed to 1.2m/min;
(f) 0.6mm gap tolerated, by reducing welding speed to 1.2m/min, and increasing wire feed rate to 13m/min.
The overall benefits (in terms of increased tolerance) gained by using adaptive control, particularly in terms of increased gap tolerance, can be seen by comparing Fig.1.8 with Figs.1.5-1.7.
Figs.1.9-1.11 show the extents to which tolerances have been increased for making either Class B or C welds, including those over a combination of mismatch and gap, for these three different butt joint test cases. In these Figures, 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 with adaptive control. Other welds were left unclassified ('X').
Fig.1.9. Increase in fit-up tolerance possible when hybrid welding butt welds in 4mm Al alloy plate with adaptive control. Reproduced courtesy of TWI Ltd.
Fig.1.10. Increase in fit-up tolerance possible when hybrid welding butt welds in 6mm stainless steel plate with adaptive control. Reproduced courtesy of TWI Ltd.
Fig.1.11. Increase in fit-up tolerance possible when hybrid welding butt welds in 8mm steel plate with adaptive control. Reproduced courtesy of TWI Ltd.
These current results appear broadly in agreement with previous data using other laser types. The maximum gap tolerance reported in other work is up to 1.5-1.6mm for CO2 laser-based hybrid welding [Kim et al, 2003, Shi et al, 2005], by increasing the wire feed rate by ~55-60%, or to 1.2mm for Nd:YAG laser-based hybrid welding [Shi and Hilton, 2005], by reducing the welding speed by ~15%.
In terms of tolerance to mismatch, [Thomy et al, 2006] have reported full fusion through a butt joint between 11.2mm thickness plates with a 1.4mm mismatch using a higher power 15kW fibre laser. However, in this case the weld profile quality class achieved was not reported.
Adaptively controlled flat position hybrid laser-MIG/MAG butt welding of Al alloy, stainless steel and steel plates has been carried out, for plate thicknesses in the range 4-8mm, using a high brightness 5kW Yb-fibre laser. ISO 13919-1/2 Class B and Class C welds have been made, over joints with a number of different fit-up conditions.
The main conclusions of these welding trials are:
- Class B weld profiles can be produced, at welding speeds of up to 4.5m/min (depending on plate thickness), over perfectly fitting joints;
- Welding without adaptive control can be used to tolerate joint gaps and mismatches to a limited extent (depending on plate material and thickness);
- Off-line trials can identify those parameters which, if changed, can increase these tolerances;
- Using joint fit-up data from a seam tracking sensor, subsequent adaptive control of welding parameters is then useful in combatting unacceptable levels of weld cap underfill in joints over wider gaps, sometimes in excess of 1mm (depending on plate material and thickness);
- Adaptive control also increase give modest tolerance increases to mismatch, when welding steel or stainless steel butt joints, avoiding re-entrant weld root toe blend angles;
- Adaptive control is more useful in avoiding lack of fusion defects in aluminium butt joints with mismatch, but does not improve the weld root profiles achieved in those cases.
BS EN ISO 13919-1:1997: 'Welding - Electron and laser beam welded joints - Guidance on quality levels for imperfections - Part 1. Steel'.
BS EN ISO 13919-2:2001: 'Welding - Electron and laser beam welded joints - Guidance on quality levels for imperfections - Part 1. Aluminium and its weldable alloys'.
Kim H S, Lee Y S, Park Y S, Kim J K and Shin J H, 2003: 'Study on the welding variables according to gap tolerance of butt joint in laser hybrid arc welding of carbon steel', LIM Proc pp165-169, 24-26 June, Munich, Publ. D-70331 Stuttgart, Germany.
Shi S G and Hilton P A, 2005: 'A comparison of the gap-bridging capability of CO2 laser and hybrid CO2 laser MAG welding on 8mm thickness C-Mn steel plate', Welding in the World, 49, pp75-87.
Shi S G, Hilton P A, Mulligan S J and Verhaeghe G, 2005: 'Hybrid Nd:YAG laser-MAG welding of thick section steel with adaptive control', Welding and Cutting, 4, 6, pp345-350.
Thomy C, Seefeld T, Vollertsen F, Vietz E, 2006: 'Application of fibre lasers to pipeline girth welding', Welding J., 85, 7, pp30-33.