The welding of aluminium and its alloys

Friction welding

Unlike the other processes covered in this book friction welding is a solid phase pressure welding process where no actual melting of the parent metal takes place. The earliest version of the process utilised equipment similar to a lathe where one component was held stationary and the other held in a rotating chuck (Fig. 8.11). Rubbing the two faces together produces suf­ficient heat that local plastic zones are formed and an end load applied to the components causes this plasticised metal to be extruded from the joint, carrying with it any contaminants, oxides, etc. Thus two atomically clean metal surfaces are brought together under pressure and an intermetallic bond is formed. The heat generated is confined to the interface, heat input is low and the hot work applied to the weld area results in grain refinement. This rapid, easily controlled and easily mechanised process has been used extensively in the automotive industry for items such as differential casings, half shafts and bi-metallic valves. Since the introduction of this conventional rotating method of friction welding many developments have taken place such as stud welding, friction surfacing, linear and radial friction welding, taper plug welding and friction stir welding.

One very important characteristic of friction welding is its ability to weld alloys and combinations of alloys previously regarded as unweldable. It is possible to make dissimilar metal joints, joining steel, copper and aluminium to themselves and to each other and to successfully weld alloys such as the 2.5% copper-Al 2618 and the AlZnMgCu alloy 7075 without hot cracking. The primary reason for this is that no melting takes place and thus no brittle intermetallic phases are formed.

8.5.1 Rotary/relative motion friction welding

The rotary/relative motion friction welding process (Fig. 8.11) is suited to the joining of fairly regular shaped components, one of them ideally being circular in cross-section. Equal diameter tubes or bars are the best example since equal heating can take place over the whole contact area. There are a couple of disadvantages to this process. The first is that one of the com­ponents must be rotated and this places a restriction on the shape and size of the items to be welded, the second is that items to be welded cannot be presented to the mating part at an angle.

The welding parameters comprise the rotational speed which determines the peripheral speed, the pressure applied during the welding process and the duration of the weld cycle. The metal extruded from the joint forms a flash on the outside of the weld and this is generally machined off to give a flat surface.

8.5.2 Friction stir welding

The most significant process for the welding of aluminium to be developed within the last decade of the twentieth century was the friction stir process, an adaptation of the friction welding process. This process was invented at TWI in the UK in 1991 and, unlike the conventional rotary or linear motion processes, is capable of welding longitudinal seams in flat plate. Despite being such a new process friction stir welds have already been launched into space in 1999 in the form of seams in the fuel tanks of a Boeing Delta

II rocket (Fig. 8.12). It will soon be used for non-structural components in conventional commercial aircraft and is being actively considered for struc­tural use. Friction stir welding has also been introduced into shipyards with great success and is being actively investigated for applications in the railway rolling stock and automotive industries.

The process utilises a bar-like tool in a wear-resistant material, for alu­minium generally tool steel, a tool lasting in the region of 1-2km of welding before requiring replacement. The end of the bar is machined to form a central probe and a shoulder, the probe length being slightly less than the depth of the weld required. The bar is rotated and the probe plunged into the weld line until the shoulder contacts the surface. The rotating probe within the workpiece heats and plasticises the surrounding metal. Moving the tool along the joint line results in the metal flowing from the front to the back of the probe, being prevented from extruding from the joint by the shoulders (Fig. 8.13). This also applies a substantial forging force which consolidates the plasticised metal to form a high-quality weld.

To provide support and to prevent the plasticised metal extruding from the underside of the weld a non-fusible backing bar must be used. A groove in the backing bar may be used to form a positive root bead - a simple method of determining that full penetration has been achieved (Fig. 8.14).

The technique enables long lengths of weld to be made without any melting taking place (Fig. 8.15). This provides some important metallurgi­cal advantages compared with fusion welding. Firstly, no melting means that solidification and liquation cracking are eliminated; secondly, dissimilar and

incompatible alloys that cannot be fusion welded together can be success­fully joined; thirdly, the stirring and forging action produces a fine-grain structure with properties better than can be achieved in a fusion weld and, lastly, low boiling point alloying elements are not lost by evaporation. Other advantages are low distortion, no edge preparation, no porosity, no weld consumables such as shield gas or filler metal and some tolerance to the presence of an oxide layer.

One disadvantage to the process is that the ‘keyhole’ remains when the tool is retracted at the end of the joint. While this may not be a problem with longitudinal seams where the weld may be ended in a run-off tab that can be removed, it restricts the use of the process for circumferential seams. This disadvantage has been overcome by the use of friction taper plug welding. Tools with a retractable pin are also being investigated and have given some promising results.

Alloys that have been welded include the easily weldable alloys 5083, 5454, 6061 and 6082 and the less weldable alloys 2014, 2219 and 7075. In the case of alloys in the ‘O’ condition tensile failures occur in the parent material away from the weld. As far as the effect on the HAZ is concerned heat input is less than that of a conventional arc fusion weld. This results in narrow heat affected zones and a smaller loss of strength in those alloys that have been hardened by cold-working or ageing. Table 8.3 lists the

results of mechanical tests carried out by TWI Ltd as part of the investiga­tory programme. The results show that the ‘softening factor’, the ratio between the parent metal strength and that of the weld, in both the cold - worked and age-hardened alloys, is close to 1, implying that there is a limited loss of strength.

The softening factors of 0.83 for the 6082-T6 alloy can be compared with the softening factor of 0.50 in Table 4.5 of BS 8118 for an arc weld in the same alloy and condition. The design benefits once this reduction in strength loss can be taken advantage of in the design specifications are obvious.

Plate of 75 mm thickness has been welded using a double sided technique at a welding speed of 60 mm per minute. Plates in the thickness range 1.2-50 mm have been welded in a single pass and at speeds of up to 1800mm/min. The process is completely mechanical and can be carried out with simple machine tool equipment that requires very little maintenance. The conventional non-destructive testing techniques of radiography and ultrasonic examination do not lend themselves to the interrogation of friction welds. However, the welding parameters are machine tool settings and can be easily monitored and used to determine weld quality, any deviation from the required settings being cause for rejection.

Although this development is relatively recent it has been enthusiasti­cally adopted by the rail rolling stock manufacturers and a number of ship­yards in addition to its use in the aerospace industry.