Production improved by high power laser welding
TWI Bulletin, September/October 1988
Ian Norris, BMet, is a Senior Research Engineer in the Laser Centre at The Welding Institute.
High power laser welding offers many advantages over conventional welding techniques. While such advantages are recognised and exploited at laser powers up to 5kW, they still need to be established at the higher powers now available from commercial machines. This article describes trials carried out to assess benefits of welding membrane panels for the power generation industry using a 10kW CO 2 laser.
The Welding Institute, with funding from the DTI, has installed, commissioned, and is now operating a 10kW transverse flow CO 2 laser.
The laser, designated CL10, was built by Ferranti to a UKAEA Culham design and is shown in Fig.1. It produces an annular output beam of ~43mm diameter which is transmitted over ~3m to a remote workstation. In the workstation, the beam is directed, via a mirror mounted at 45° to the beam axis, along a 2m single axis gantry on which is mounted a traversing focusing unit. This unit was designed and built at the Institute and comprises two flat mirrors and a spherical mirror of ~680mm radius of curvature ( Fig.2).
Fig.1. 10kW laser and workstation
Fig.2. Laser focusing system
The beam from the laser, when not directed to the workstation, is directed into a water cooled calorimeter where beam power is measured. As some absorption occurs in the mirrors used to deliver the beam, power at the workpiece is generally ~15% less than that recorded at the calorimeter.
The original aim of establishing the laser facility was to allow a series of demonstration projects to be run to highlight advantages offered by high power laser processing in a variety of industrial welding applications.
Discussions with many UK manufacturing companies led to a range of components being suggested for investigation and these were classified into groups, each expected to demonstrate particular benefits from use of the laser. These benefits included deep penetration, fast welding speed, high throughput, low heat input, low distortion and process cleanness. Components from each group were then chosen for laser welding trials on the basis of their suitability for laser processing, material/part availability and individual company interest.
This article summarises trials conducted on one of the components, boiler wall membrane panels, chosen to demonstrate single pass, high speed, low distortion welding. Results of the work and problems and benefits associated with the adoption of laser processing are discussed.
Membrane panel fabrication
Conventional, fossil fuelled power stations and industrial boilers use membrane panels for the walls of various types of furnace. The panels comprise a large number of steel tubes separated by, and welded to, an equal number of flat steel strips.
Membrane panels are built up from smaller sub-panels fabricated by longitudinal welding of tubes and strips. Typically, a sub-panel comprises 4-8 tubes of ~60mm OD and ~5mm wall thickness, welded to a similar number of strips of ~30mm width and ~6mm thickness. In practice, tube and strip dimensions may vary significantly from those quoted to suit particular applications and customer requirements.
Sub-panels can be joined to produce full sized panels with up to 50 tube/strip pairs. A finished membrane panel may be up to 30m long and several metres wide and a complete furnace may use 70 000m of tubing. A large membrane panel is shown in Fig.3.
Fig.3. Large membrane panel ( Courtesy of Babcock Power Ltd)
Sub-panels are manufactured currently using a double sided submerged-arc welding procedure in which strip and tube are traversed under stationary welding heads. In a furnace with 70 000m of tubing, the double sided procedure gives an effective welding rate of up to 1.3 m/min and a total welding time in excess of 1800hr.
As sub-panels are welded from both sides it is possible to join only straight lengths of tube to straight, flat strip. In practice, panels require a number of openings and bends which must be added after fabrication by cutting, rewelding and bending using portable bending equipment.
The current membrane panel production route can also present problems because sub-panels are joined to produce larger panels using manual welding. Time and effort spent in handling and turning sub-panels in preparation for manual welding may be considerable and use of such welding procedures can lead to problems of panel distortion and tube wall burnthrough. Subsequent correction of welding defects is expensive and time consuming.
Laser welding trials
Laser welding permits full penetration welds between tube and strip in a single pass from one side only. Production rate may therefore be expected to increase as a result of elimination of the second side weld and also because of the increased welding speed expected with a laser.
| Sample number | Element, wt% |
|---|
| C | S | P | Si | Mn | Ni | Cr | Mo | V | Cu | Nb | Ti | Al | B | Sn | Co | As | CE | P cm |
| Tube L323 | 0.15 | 0.009 | 0.012 | 0.23 | 1.04 | 0.19 | 0.15 | 0.05 | <0.002 | 0.19 | <0.002 | <0.002 | 0.020 | <0.0003 | 0.03 | 0.01 | 0.017 | 0.39 | 0.233 |
| Strip L324 | 0.08 | 0.011 | 0.008 | 0.22 | 0.22 | 0.03 | 0.05 | 0.005 | <0.002 | 0.03 | <0.002 | <0.002 | 0.064 | <0.0003 | <0.005 | <0.005 | <0.005 | 0.13 | 0.103 |
By welding from one side only, large membrane panels may be built up continuously with the various openings and bends in place and post-weld alterations are minimised. In addition, workpiece handling time and problems currently experienced because of use of manual welding for sub-panel joining are reduced or eliminated.
To benefit from these advantages a system would have to be built on which the panel was laid up on a large flat bed and welded with a moving laser beam manipulated by a gantry system incorporating moving optics.
To assess the suitability of laser welding for membrane panel fabrication, trials were conducted on 300mm lengths of tubes and strip. Effects of laser welding parameters (laser power, welding speed) on weld profile and penetration were determined and trials were carried out to establish tolerance to variations in part fit-up and workpiece alignment with respect to the laser beam focus position.
Chemical compositions of materials used in the trials are shown in the
Table. All materials were supplied by Babcock Power. For welding, the parts were shot blasted, degreased with acetone and held in the jig shown in
Fig.4 which was mounted below the laser focusing unit. The part cleaning procedure, though more rigorous than that currently required for submerged-arc welding, could be automated for production use.
Fig.4. Laser welding jig for holding membrane panel specimens
The influence of welding parameters on weld penetration and profile was investigated with the parts tightly clamped and with the laser beam focused at the strip surface. Welds were made at laser powers of 6-10kW and speeds of 0.75-2.5 m/min.
Prepared sections of the welds made are shown in Fig.5 which indicates that full penetration was achieved over a wide range of parameters. Lack of penetration occurred only at welding speeds of >1.5 m/min though some welds made at speeds of >1.25 m/min exhibited centreline cracking. At a given welding speed, weld width increased with increasing laser power.
Fig.5. Welds made in welding parameter study
From the results of the welding parameter study, the following three welding conditions that produced full penetration welds with wide, medium and narrow fusion zones respectively, were chosen for trials to determine effects of laser beam focus position, part fit-up and workpiece alignment on profile and penetration. These were 0.75, 1.25 and 2.0 m/min, all at 9kW.
Effects of variations in beam focus position with respect to the strip surface are indicated by the sections in Fig.6. Full penetration was achieved at all settings of focus position and only small changes in weld profile were observed when the focus position was varied within the limits investigated.
Fig.6. Welds made to assess the effect of different focus positions
Effects of variations in fit-up between strip and tube are shown in Fig.7 and 8. Figure 7 compares sections taken from welds made with good fit-up and welds made with a 0.25mm joint gap. Welds made with a joint gap exhibited undercut in the top bead and, in some, in the underbead though full melting was achieved across the joint in all.
Fig.7. Welds made to assess the effect of joint gaps
Fig.8. Cross sections from different positions along weld length for assessing effect of variation in alignment of strip with tube centreline
Figure 8 shows sections taken from the low end, the centre and the high end of the specimens set up with vertical mismatch. Full penetration was achieved along the weld and variation in fit-up affected only the details of weld profile.
Welds made to investigate effects of misalignment of the laser beam with the joint line are shown in Fig.9. With the beam axis set towards the tube by 0.5mm, melting of the tube some distance above the strip surface occurred and the gas shielding nozzle melted because of reflection of the beam from the tube surface. This problem was more severe with the beam axis set towards the tube by lmm. With the beam misaligned towards the strip by 1mm, full fusion of the interface was achieved at the lowest welding speed (0.75 m/min) but lack of fusion defects occurred at higher speeds. With a misalignment of 2mm towards the strip most melting occurred in the strip material with little, if any, fusion across the joint interface.
Fig.9. Effects of misalignment of the laser beam with respect to joint line
In general, results of the laser welding trials are encouraging. The fact that full penetration welds were produced at all laser powers investigated at welding speeds of ≥1.25 m/min is important from a production point of view as it means that significant variations in parameters may be tolerated during manufacture.
Tolerances to focus, fit-up and alignment found in current trials provide a useful comparison with equivalent values for submerged-arc welding.
At present, the system used to hold parts beneath the submerged-arc welding head positions the strip surface well within the ± 4mm tolerance to focus position found in the laser welding trials. A similar clamping system would, therefore, be sufficient in a laser welding membrane panel machine.
The tube and strip that are submerged-arc welded to form sub-panels are held in intimate contact during welding. It is encouraging that full fusion across the interface was achieved when laser welding with a 0.25mm gap, though mechanical tests are required to determine the influence of top and underbead undercutting.
A best estimate figure for misalignment of strip and tube centreline allowed with the bridge clamping jig used in the current panel welding machine is ± 0.5mm. The tolerance of half the strip width found in the laser welding trials on misaligned tube and strip is equivalent to ± 3mm and this again indicates that the current clamping arrangement could be used in a laser welding machine.
The laser welding trials showed very little tolerance to misalignment of the beam with the joint line, particularly when the beam axis was misaligned towards the tube. An allowable variation of less than 0.5mm in beam/joint alignment suggests that some form of seam tracking device would be necessary in a production machine. The tube to strip joint lends itself to tracking and a relatively simple mechanical device would probably suffice.
Although the results are encouraging, maximum welding speed for full penetration, (2.0 m/min) is significantly lower than expected from previous trials on laser butt welding of flat, 6mm thickness steel plate. It is possible that further work to optimise angle of approach of the laser beam with respect to the joint and to optimise weld shielding could lead to increased speed and, therefore, significantly greater benefit from adoption of the process.
The profile of the laser welds made in the trials may give rise to concern over joint fatigue properties and further work to investigate weld mechanical properties fully is required.
Conclusions
- Single-sided laser welding shows considerable promise and offers significant potential benefit in this application and should be investigated further as a production technique.
- Visually acceptable welds may be produced between strips and tubes over a range of welding parameters. With 7mm wall thickness tube and 6mm thickness strip, acceptable welds were made at laser powers of 6-10kW and speeds of0.75-2.0 m/min.
- In most instances, fit-up requirements for successful laser welding were within those presently achieved by jigging and clamping arrangements used in submerged-arc welding of membrane panels.
- Clamping currently used is sufficient to hold parts well within tolerances to laser beam focus position.
- Laser welding is intolerant to variations in beam/joint misalignment and a production membrane panel welding machine would probably require a simple seam tracking device.
- The maximum welding speed at which full penetration welds were made was not as high as expected from previous work. There is, however, scope for increasing speed with further process optimisation.
- Further work is required to investigate weld mechanical properties.
Acknowledgements
Funding was provided by the DTI and Research Members of The Welding Institute. Babcock Power supplied materials and information on current production techniques.
