Distorting the truth - a novel manufacturing route for aluminium alloy stiffened panels
TWI Bulletin, September/October 2001
Nettem Sekhar, a Project Leader at TWI, has been involved in the development of procedures for welding and cutting ferrous and non-ferrous alloys using Nd:YAG lasers. He has been active in the application of the procedures to industrial context. He is currently pursuing a PhD at the University of Cambridge, UK.
Paul Hilton is Technology Manager - Lasers at TWI where he has responsibility for the strategic development of laser materials processing. As such he has been instrumental in the setting up and management of several European collaborative research projects. Dr Hilton has previously worked in the laser systems industry in the UK and, before that, was a researcher at an international scientific institute in France.
How can the unwanted distortion of laser welded components be overcome? Nettem Sekhar and Paul Hilton investigate a novel laser route for the manufacture of aluminium stiffened panels.
The 2xxx series aluminium alloys is favoured in the manufacture of aerospace structural members because of its superior strength to weight ratio. In airframe construction load bearing members (such as the fuselage and wings) cross stringers are used to add stiffening. The development of a procedure for laser beam welding of AA2014 stringers to an AA2014 base plate to produce a stiffened panel is reported. Distortion analysis has been carried out on laser and TIG welded panels. Suggestions on how to further reduce distortion are presented.
Laser beam welding of aluminium alloys
Because of its lower wavelength radiation from an Nd:YAG laser tends to couple better to aluminium alloys than radiation from a CO 2 laser. This has led to an increase in the use of Nd:YAG lasers in the welding of aluminium alloys. To produce a key-hole there is a required minimum threshold intensity in the focused beam at the workpiece. Typically, threshold intensities of at least 0.04 MW/mm 2 are required for CO 2 laser welding and 0.015 MW/mm 2 for Nd:YAG laser welding of aluminium alloys. Deep penetration laser welding of aluminium alloys requires between five and 25 times greater intensity in comparison to steels. Aluminium is a good reflector of light at the infra-red end of the spectrum, which makes it difficult to weld.
In airframe sections, attaching transverse and longitudinal stringers provides additional stiffness to the complete section. Such an application is considered here: the manufacture of a laser welded stiffened panel.
Experimental work
Materials
The base plate of the panel was machined from AA 2014 (AlCu4SiMg) in the T4 condition (solution treated and naturally aged). The stringers were manufactured from AA 2014 in the T4 condition and were 3.5mm thickness. The filler wire used in the study was ER 2319 (high copper content aluminium welding filler, ~6 wt%Cu), 1.2mm in diameter. One set of base plates was of a uniform thickness of 3mm. The second set had pockets (244mm x 133mm x 1.5mm) machined between the stringers and was similar in form to an actual component ( Fig.1). Six stringers, each 14mm high, 250mm long and 150mm apart, were to be welded on this base plate. A simple finger-clamp mechanism held the stringers in position ( Fig.2).
Fig.1. Stiffened panel with machined pockets (not to scale)
Fig.2. Welding set-up used in the manufacture of the panels. The laser head is at an angle of 38° to the horizontal
Laser and beam delivery
A 4kW Nd:YAG laser was used for the laser processing. The focussed spot size was 0.6mm diameter and continuous wave (CW) power up to 3.5kW could be achieved at the workpiece. The output housing was mounted on a robot arm. Welds were made by traversing the output housing over the stationary workpiece.
Parametric study
A parametric study on smaller samples was carried out to obtain the process variables and procedure to be used during dual sided welding of stringers to the panels. The effects of pre-cleaning (chemical etching, mechanical abrasion), shielding gas type (argon, nitrogen and helium), filler wire feed rate, heat input, weld sequencing (
Fig.3) and laser beam angle of approach were studied.
Fig.3. Sequences used for welding the panels. The welding direction is shown.
Analysis
Visual inspection, radiography and optical metallography were carried out on selected samples. For optical metallography, the samples were etched using Keller's reagent. The standard adopted for evaluation of the welds was: prEN ISO 13919 - 2, Welding - Electron and laser beam welded joints - Guidance on quality levels for imperfections - Part 2: Aluminium and its weldable alloys.
Parametric study - results
Effect of pre-cleaning
Welds made with no surface preparation resulted in high porosity levels. In addition, uneven weld top bead surfaces were obtained. Chemical etching to remove the aluminium oxide layer was effective in producing welds with low porosity levels. Welds prepared using a mechanically abrasive scouring pad to remove the oxide layer resulted in variable levels of porosity.
Effect of shielding gas type
Welds made with argon as the shielding gas exhibited root and centreline porosity. Porosity was distributed in the weld length with a maximum pore diameter of 0.8mm. The levels of porosity were low and met the most stringent quality level (B) of the adopted standard. Slight undercut on one side of the joint could be seen. The top surface of the weld was smooth and slightly concave (as required for good fatigue properties). Radiographs of welds made with helium as the processing gas showed finely distributed porosity with the maximum pore size being 0.6mm in diameter. The top surface of the weld was smooth and slightly concave. Porosity was mainly located at the interface of the two welds, which were made sequentially.
In welds made using nitrogen as the shielding gas, the lowest porosity levels were observed. The maximum pore diameter noted was about 0.5mm. In addition, greater weld penetration depths were obtained in each pass. Smooth top bead profiles, slightly concave in nature, were obtained ( Fig.4).
Fig.4. Cross section of a two pass weld made with nitrogen as the shielding gas
Effect of filler addition
In welds made without any filler addition, longitudinal cracking extending along the weld length was observed. Using the ER 2319 wire at low filler feed rates (1.5 and 2.0 m/min), small centreline cracks were observed in the weld cross sections. In addition, weld top beads were narrow and uneven. At welding speeds of 2 m/min and wire feed rates of 3 m/min, weld cracking was avoided. Smooth weld beads were also obtained. Increasing the wire feed rate to over 4 m/min only resulted in excessive deposit and shallower weld penetrations.
Effect of heat input
At very low welding speeds (0.75 or 1.0 m/min), wide weld beads with penetration up to 66% of the stringer thickness were obtained. As the welding speeds were raised (3.0 and 4.0 m/min), smoother and narrower weld beads were obtained. Increasing the filler feed rate but keeping the welding speed constant led to a decrease in the weld penetration and in wider top beads. At higher welding speeds (> 3 m/min), weld beads with an uneven top surface were obtained, probably due to the difficulties encountered in feeding the wire into the weld pool at this speed. This was also the case at high wire feed rates (3.5 m/min).
Welding procedure
The parameters chosen based on the results of the parametric study are shown in Table 1. Before welding all stringers were placed in position and clamped in the jig. The direction of welding for both passes of the fillet was the same. The two weld sequences, labelled A and B ( Fig.3) were used. Laser welded panels can be seen in Fig.5. Figure 4 (produced in the parametric study) should be indicative of the fillet weld cross section obtained in the panels.
Table 1: Parameters to manufacture the complete panel/stringer assemblies
| Welding speed | 2.0 m/min |
| Filler wire feed rate | 3.0 m/min |
| Angle of approach | 38° |
| Laser focus | On the surface at the apex of the joint |
| Gas type | Nitrogen |
| Gas shielding | Coaxial |
| Shielding gas flow rate | 30 litre/min |
| Cleaning | Chemical etching |
Fig.5. Completed panels containing machined pockets (weld sequence shown in Fig.3)
Observed distortion
In fillet welding, two types of distortion are possible. The shrinkage force exerted by the cooling weld is the basic cause for these two types of distortion. Angular distortion is a result of contraction of the weld in a direction perpendicular to the weld line and bowing is due to contraction along the length of the weld. In practice, the distortion exhibited by a structure is a combination of the two, ie angular distortion and bowing. The stringers tend to expand more than the base plate along their length with progression of welding. This is due to accumulation of heat in the stringer. On completion of welding, as the panel cools down, the stringer tries to return to its original dimensions by contracting. This results in high compressive stresses along the length of the stringer, leading to the form of bowing distortion.
Effect of sequencing
When unclamped, the laser welded panels displayed two modes of distortion: a vertical movement on one corner of the panel (Mode 1) and a movement in the central portion of the panel, at the weld start and finish (Mode 2).
Table 2 gives the distortion measured in the panels.
Similar panels were manufactured by the TIG welding route, as can be seen in Fig.6, and the distortion observed was higher. This is because the TIG weld beads were much larger and the total heat input into the structure was also higher than for the laser welding technique.
Table 2 Distortion measured in the panels
| Panel No. | Process | Sequence | Machined pocket | Mode 1 | Mode |
| Start | End |
| 1 | Laser | 1 | No | 42 | 23 | 6 |
| 2 | Laser | 1 | Yes | - | 14 | - |
| 3 | Laser | 2 | Yes | 26 | 3 | 10 |
| 4 | TIG | 1 | Yes | - | 60 | 60 |
Fig.6. A TIG welded panel (welded to sequence No.1 as shown in Fig.3a).
Effect of laser beam angle of approach
Using the optimised parameters a fillet weld was also prepared reducing the angle of approach of the laser beam to 20°. This was made possible by placing the stringer at one end of the base plate and by holding this assembly at the edge of a processing table. The result showed that for the same joint configuration a lower angle of approach allowed the welding speed to be increased by 20%, or the laser power to be reduced by 15% when compared to the 38° angle used for the panels. In addition, the amount of base plate melting could be reduced. It is thus probable that the laser welded panels could have been produced with even less distortion if the 20° angle or even lower had been accessible.
The levels of distortion and weld quality observed in the laser welded panels are, as yet, generally unacceptable in the aerospace industry. To reduce distortion further, heat input should be reduced and/or improved welding techniques used. Some of the techniques which could be further evaluated to either reduce the levels of distortion or improve the weld quality are:
- Smaller laser spot size
- Simultaneous welding from both sides
- Reducing the angle of approach of the laser beam
- Twin-spot laser optics
- Hybrid laser-arc processing
- Low stress nil distortion (LSND) techniques
- Pre-bending of the base plate
- Fluxes that improve weld penetration
Conclusions
A laser welding procedure for the welding of aluminium stringers has been developed. Filler wire is required to minimise the risk of solidification cracking. Nitrogen as the processing gas has given the least levels of porosity. The levels of distortion in the full panels are much lower than those seen with TIG welding, but are as yet unacceptable. Further means of reducing distortion have been outlined.
