Friction welding of near net shape preforms in Ti-6Al-4V

P L Threadgill and M J Russell
TWI Ltd, Granta Park, Great Abington, Cambridge CB21 6AL, United Kingdom.

Paper presented at 11th World Conference on titanium (Ti-2007), (JIMIC - 5), Kyoto, Japan, 3-7 June, 2007.

This presentation will describe recent developments in the joining of titanium alloys using various friction welding processes such as linear friction welding (LFW), rotary friction welding (RFW) and friction stir welding (FSW). The paper will focus on the use of these processes to produce machining pre-forms for aerospace and other industrial components. Currently, such components are usually machined from solid blocks of titanium alloy, resulting in relatively poor material buy-to-fly ratios. The high cost of titanium alloys, and the relative difficulty in machining titanium, make the use of welded pre-forms a powerful route to significantly reduce production costs for a range of machined titanium parts.

Build up of machining pre-forms by friction processes also provides the opportunity for selection of appropriate titanium alloys in different parts of the structure. This approach allows production of tailored components, resulting in both functional and economic benefits. Examples will be shown of the application of this approach to aerospace components, from simple two-piece friction welded fabrications, to complex multiple-part pre-forms produced by sequential addition of material by friction welding.

In summary this presentation will provide an overview of recent friction welding development work aimed at improving product effectiveness and reducing production costs for a range of titanium alloy components.

1. Introduction

Titanium is a very expensive material to purchase due to the complexity of the refining and metalworking process and the huge amounts of energy required in its manufacture. This is currently compounded by long lead times for delivery. It is also relatively expensive to machine, and therefore there is a strong case to replace the common practice of manufacturing components by machining from solid blocks with fabrication of near net shape components by welding together smaller pieces. This clearly reduces the initial demand for titanium, and also the extent of the machining operation. Although alloys such as Ti-6Al-4V are relatively easy to weld if correct procedures are followed, there has always been resistance to including fusion welds in critical applications. However, solid state welds such as friction welds and diffusion bonds are not subject to the same restrictions, and both are used in critical aerospace components. For example, diffusion bonding is used in superplastically formed fan blades and other parts, and linear friction welding is used to join titanium alloy fan and compressor blades to titanium alloy disks to make integrally bladed disks (blisks). It is therefore a simple extension to this logic to apply these processes to near net shape manufacture, and the current paper gives an overview of the state of the art in this area.

2. Friction welding of titanium alloys

There is a long heritage of friction welding titanium alloys, going back to the late 1960s. Titanium alloys are regarded as relatively straightforward to friction weld, whether by linear friction welding, ^[1-8] or by the more established rotary friction welding. ^[9-14] Material related characteristics of the process and advantages in welding titanium alloys are as follows:

The low thermal conductivity which confines heat generated by the friction welding process to the weld area, where it is needed.
The temperature range in which titanium alloys can be hot worked at high strain rates is relatively narrow. This minimises weld time, and aids in the control of the process.
In a-b alloys such as Ti-6Al-4V, friction welding takes place above the b transus, and therefore the microstructure will be determined by the kinetics of the b to a+b phase transformation. These are such that the microstructures formed will have mechanical properties which equal those of the parent material. This is true of both tensile properties and fatigue. This means that existing mill annealed plate allowables can often be used in component design.
The process window for acceptable quality friction welds is quite large, and this is of obvious benefit to fabricators. This helps make the process very repeatable and reliable, with very few weld failures.
Unlike arc and laser welding, no shielding gas or filler wire is needed for friction welding, as contaminated material is expelled as flash, and surface contamination is minimal, and within the post-weld machining limit.

Friction welding processes have been demonstrated for a wide range of titanium alloys, including developing alloys such as gamma aluminides. It is possible to weld dissimilar titanium alloys to each other, allowing the design and fabrication of tailored structural components. A good example of this would be the use of compressor blades made of one alloy welded to disks made of another, although the extent of other possible examples is limited only by imagination. Joining of titanium alloys to alloys from other systems is difficult, generally due to the formation of intermetallics at the joint line. Titanium forms intermetallics with almost all common structural and functional metals, the only exceptions being refractory metals (W, Nb, Mo etc) and other Group IV metals (Zr and Hf). However, some success has been claimed for joining titanium to steel, aluminium and copper.

There are no 'perfect' welding processes, and all of the options available for joining Ti alloys have their disadvantages. For friction welding, the equipment cost is high, especially where a direct drive machine is used. Partly for this reason, larger rotary friction welding machines tend to use the inertia principle, where the energy to make the weld is stored in a rotating flywheel.

Linear friction welding machines also require energy transfer at a high rate, and the use of stored energy, for example by the use of hydraulic accumulators, has been shown to result in a very significant reduction in the cost of a welding machine. ^[5,8]

Both rotary and linear friction welding processes produce flash, but this can be removed relatively easily during post-weld machining operations.

3. Near net shape approaches

It is self-evident that each near net shape design is different, and therefore will require a unique set of components to fabricate it. Friction processes may be a significant contributor to the arsenal of processes available, but there will of course be situations where other methods are superior. Linear and rotary friction processes are best suited to adding small to medium size pieces to larger structures. Current linear friction welding machines are limited to weld areas of about 2000mm ² although rotary machines can handle much larger components. For example, a medium size 100 tonne rotary friction welding machine could probably cope with an area of about 12-13,000mm ² . However, rotary methods are inherently less flexible, as the rotating part should ideally have rotational symmetry, although with modern machines specific rotational alignments can be accurately achieved. With either method, complex added parts (or complex parts being added to) will require complex fixturing systems to withstand the high process forces involved, and this will inevitably limit the economic advantages, especially in short production runs. Very small parts may be better manufactured by processes such as direct metal deposition using laser or electron beam processes. Similarly, more traditional processes may also be used to weld additional parts to the preform.

The ideal approach for additive manufacture by friction welding reduces to the use of simple building blocks to fabricate a larger structure which can be finally machined to shape. A concept has been proposed by Boeing (Pat. No. US2005127140), in which a number of prototypical structures can be assembled from simple shapes using friction based processes, primarily linear friction welding. Schematics of the concept are shown in Figure 1. Feasibility studies on the addition of material by LFW have shown that excellent weld properties can be achieved. ^[15] Weld tensile and fatigue properties are commonly comparable to or an improvement over the parent material, and existing mill annealed plate allowables can be used in component design. In addition, the processing window for linear friction welding of Ti alloys is generally large, which allows the welding of a wide range of different size/shape features.

Fig.1. Concept for near net shape component manufacture using linear friction welding
	a) initial stage,
	b) final bridging weld,
	c) machined to shape

Whilst additive manufacture by linear friction welding is rapidly developing into a production process, the use of other friction processes for fabrication of parts is also being considered. Work is ongoing on the use of both rotary friction welding, and friction stir welding, for the additive manufacture of Ti alloy parts.

Rotary friction welding offers the potential for rapid fabrication of relatively large scale parts, and is suitable for the joining of a range of components such as actuators, undercarriage parts, fixings, and engine structures. A schematic of a relatively simple application of rotary friction welding for an aerospace component is shown in Figure 2.

Fig.2. Possible structure fabricated using rotary friction welding

Friction Stir Welding (FSW) is the latest addition to the range of friction welding processes that can be used for additive manufacture. The FSW of Ti alloys has proved to be challenging, however recent development work at TWI has lead to significantly improved capabilities in this area. FSW is now under investigation as a method of adding material to larger structures, and also as a method for rapid prototyping of full components, as illustrated in Figure 3

Fig.3. Conceptual designs for complex structures fabricated by multipass friction stir welding
( Courtesy W M Thomas)

The use of FSW for additive manufacture offers the potential for the creation of high quality solid phase material, at a relatively high material deposition rate. FSW also offers significantly more flexibility in the form of the deposited material when compared to the other more established friction processes. Considerable success has been achieved in the use of this approach in several aluminium alloys.

Recent work in the area of additive manufacture of Ti parts by FSW has produced encouraging results, as shown in Figures 4 and 5. Figure 4 shows an initial concept structure in which four layers of 3mm Ti-6Al-4V have been deposited using static shoulder friction stir welding to make a vertical ligament about 9mm wide. A macro section can be seen in Figure 5. The quality of the weld in terms of freedom from defects was very good, although further refinement of the process and assessment of the component is underway. Current work is assessing the scaling up of this approach to fabricate larger structures. Initial evaluation has shown no obvious differences in material microstructures between the various layers when examined optically, as shown in Figures 6(a) and (b). It will also be necessary to assess the level of distortion, and investigate methods of controlling or correcting this. Although friction stir welding is normally considered to be a low distortion process, this is not always the case in irregular shaped components.

Fig.4. Photographs of SSFSW preform after machining
a)	b)

Fig.5. Macro section of 9mm wide four layer friction stir build-up in Ti-6Al-4V

Fig.6. Microstructures from (a) first and (b) fourth layer deposits, showing no obvious microstructural changes
a)	b)

A further potential advantage to this approach, and indeed to any near net shape approach based on friction welding, is the ability to place different titanium alloys in different parts of the structure, and the possibility exists, albeit with much more research, to include non-titanium parts. It is of course also feasible to integrate more than one friction process, or also use additional processes.

Finally, it is likely that friction stir based methods will result in preforms which require more machining than might be necessary with linear or rotary, as there will always be some waste material at the side of each layer, as can be seen clearly in Figure 3. However, the process is very adaptable, and the complexity of shapes which can be made will offset this disadvantage.

4. Summary and conclusions

The use of friction welding process for the additive manufacture of Ti alloy parts is attractive for a number of reasons including:

The high quality of the friction welded joint and/or deposited material.
The relatively high material deposition rate that can be achieved compared to conventional rapid prototyping technologies.
The relatively low material wastage compared to conventional machine-from-solid approaches.
The flexibility offered by the choice of an appropriate friction welding approach.
The ability to vary the alloys used in different areas of the fabrication, thus producing a tailored product.

Factors that must be considered in the selection of friction welding process for fabrication of parts include:

The capital cost of the equipment required and the overall financial case based on the volume of parts to be produced.
The varying levels of technology maturity for the three main friction welding options, rotary, linear and friction stir welding.
The need for correct process selection, optimisation and procedure development/validation for each desired application.

References

P. Wanjara and M. Jahazi: Met & Mat Trans A, 36A, (2005) pp2149-2164.
A.W.E. Nentwig and L. Appel: Schweissen und Schneiden 47 (1995) pp648-653
W.A. Baeslack, P.L. Threadgill, E.D. Nicholas and T.F. Broderick: Ti-1995, Birmingham, 1, pp424-431
H. Wilhelm, R. Furlan and K.C. Moloney: Ti-1995, Birmingham, 1, pp620-627
S.W. Kallee, E.D. Nicholas and M.J. Russell: Ti-2003, Hamburg, 5, pp2867-2874.
D. Schneefeld, D. Helm and H. Wilhelm: DVS Berichte 208, (2000) pp42-47.
M. Corzo, O. Casals, J. Alcalá, A. Mateo and M. Anglada: Rev. Metal. Madrid 41 (2005) pp403-409
M.E. Nunn, 1st Int. Conf. On Innovation and Integration in Aerospace Sciences, 4-5 Aug 2005, Belfast, N.I.
O. Roder, D. Helm and G. Lutjering: Ti-2003, Hamburg, 5, pp2875-2882
L.S. Smith, M.F. Gittos and P.L. Threadgill; Proc. Seminar on Titanium Risers and Flowlines, Trondheim, 17 Feb 1999.
P. Adam: DVS Berichte 139, 5-6 Dec 1991, pp36-39
A.S. Wadleigh: Ti-1986, San Francisco, 2, pp640-652
P.L. Threadgill; Proc. Conf. on Right Use of Ti, Stavanger, 4-5 Nov 1997.
A. Wisbey, I.C. Wallis, H.S. Uhbi, P.D. Sketchley, C.M. Ward-Close and P.L. Threadgill: Ti-1999, St Petersburg.
M.J. Russell, R.R. Boyer: 'Recent Developments in Linear Friction Welding of Ti for Aerospace Applications', Aeromat 2006, May 2006, Seattle, Washington, USA