Linear Friction Welding of Ti 6Al 4V for Aerostructure Applications

Bertrand Flipo and Kathryn Beamish
TWI Ltd, Granta Park Great Abington, Cambridge, CB21 6AL, UK

Bryan Humphreys and Martin Wood
CAV Advanced Technologies Ltd, Consett, United Kingdom

Alan Shilton
Ten Solutions Ltd, Wednesbury, United Kingdom

Paper presented at Trends in Welding Research Proceedings of the 10th International Conference, Tokyo, Japan, 11 - 14 October 2016

Abstract

The increasing use of composite materials observed in new aircraft builds leads to a significant demand for titanium alloy structural parts. The rising material costs and popularity, together with their restrictions in supply and processing, is driving the aerospace industry to make increasingly efficient use of available material.

Linear Friction Welding (LFW), a rapid, high integrity, solid-state forging process, has the potential to decrease the buy-to-fly ratio, production time and time to market of aerostructure components. It has already proven its dependability for the production of some of the latest generation aero engines.

The LFW process is not yet used for aerostructures. This is primarily due to the LFW process not being widely known, and not having been developed or proven for aerostructure applications; but also due to a lack of performance data on aerostructures manufactured by LFW being currently available to the aerostructure supply chain and design community.

To address this issue, a large series of Ti-6Al-4V weldments was produced and assessed via metallographic examinations and mechanical testing. The matrix of experiments was able to capture the LFW process window of this titanium alloy, and to measure the impact of the parametric conditions. Metallographic examination revealed a high integrity weld free from contaminants and oxides at the weld interface.

The metallurgy of the welds revealed a characteristic fine-grained equiaxed microstructure. As-welded joints were tested under tensile and alternating fatigue conditions to provide an extended set of joint performance data. Finally, a techno-economic assessment was conducted which demonstrated the viability and potential of the LFW process for a representative structural airframe component.

1. Introduction

Air travel and commercial aviation is booming. Despite uncertainties, air traffic has doubled in the last 15 years, and this  trend  is expected  to  continue over the  next  15  years,  with  world passenger  traffic  and  cargo growing 4.8 % annually ^[1]. The increasing use of composite materials observed in new aircraft builds leads to increased demand for titanium alloy structural parts. Due to limited global manufacturing capacity for titanium structural · parts and the use of inappropriate processing routes, the buy-to-fly, or material utilisation ratio, can be as poor as 20:1 and is rarely as good as 4:1 ^[2]. This is symptomatic of components being machined from oversized ingots, forgings or extrusions. There is a need for a manufacturing route that reduces costs and achieves greater efficiency of titanium utilisation. Linear friction welding (LFW) can help to address this issue by joining smaller workpieces  to produce a preform, which is subsequently machined to size, resulting in significant material and cost savings.

1.1. Linear Friction Welding

The LFW process is not widely known, and it is not yet used by the aerostructure community. However, LFW is a qualified production process for critical titanium alloy aeroengines compressor components called Blisks ^[3] and this is a clear indicator of the capabilities, potential and relevance of LFW to titanium aerostructures. Linear friction welding is a solid-state joining process, meaning that the material is bonded in a plastic state, not molten; resulting in a forged microstructure. LFW works by oscillating two workpieces relative to each other, while under a normal compressive force of sufficient magnitude, to induce expulsion of plasticised material out of the weld interface. The LFW process is self-regulated by the change in material properties as friction heating occurs, and pre-set parameters allow control of a LFW cycle though its phases ^[4] and ^[5], described Fig 1.

1.2. Aims and Objectives

This article aims to demonstrate the advantages of the LFW process for producing Ti-6Al-4V aerostructures. To achieve this, a series of experimental welds were produced so that the impact of the processing conditions on the tensile and fatigue behaviour could be characterised. The experimental welds were also analysed to understand the weld interface microstructure. Furthermore, it is inconceivable to propose a technological innovation to the aerospace industry sector without considering its economic implications. An economic assessment was designed to determine and take account of all arising expenses for producing a demonstrator using conventional and current methods of machining from solid, to the new approach of near-net-shape manufacture using LFW, in order to identify cost savings. 

2. Methodology

2.1. Matrix of experiments

A large matrix of experiments was carried out on Ti-6Al-4V weldments each made of two workpieces 40x25mm faying area and 75mm body length. These coupons were assessed via in-process quality monitoring. The pressure was first observed under a high rubbing velocity, to determine the minimum forging pressure required for the friction to reach a stable state, as well as for cold deformation that forms the final stage of the process. The amplitude and frequency were studied for energetic performance and the forge duration was finally defined from the thermal behaviour using thermocouples placed up to 2mm away from the interface. Repeat trials were run to assess the impact of welding with and without a protective argon shield. Table I lists some parameters  observed.

Table 1:Envelope of key parametric conditions observed

2.2. Welds assessments

Selected weld coupons were prepared for metallographic  examination. The preparation consisted of mounting,

polishing and etching with Kroll's reagent. Samples were tensile tested according to TPOlc-1, BSEN ISO 6892-1:2009 A, with a nominal specimen diameter of 8.0mm. Fatigue specimens were machined and tested under alternating fatigue loading using methods described in ASTM STP 566EB. These specimens had a nominal neck diameter of 3.2mm and had the weld centreline displaced slightly (0.4mm) from the minimum neck diameter to position the weld at the maximum stress location.

2.3. Techno-economic assessment of a representative structural airframe component

When an established route for component manufacture is challenged by a new process, it is likely that the current part producer would seek ways of making their production process more efficient  and competitive. The  CNC 'from solid' production route was therefore defined with competitive figures in mind, with a  three  shift staff rotation and full machine utilisation, lean solid material block volumes, and a material removal rate based on the latest tool wear management practices. This assessment also took into account recurring and non-recurring costs. Recurring costs included materials, consumable tooling, energy usage, labour, maintenance, and building costs. Non-recurring costs included machines and non-consumable fixture investments. Non-recurring costs and depreciation estimates were amortised over a credible lifespan to provide realistic values for  comparison.

The down-selection process looked at the geometries intricacies, manufacturing complexities and the number of welds required when forming near-net-shape blanks. Stanchion parts were considered suitable target components for this alternative manufacturing solution. A generic demonstrator, shown in Fig. 2a, was designed  for  the purposes of this case study and to aid dissemination. The generic demonstrator had external dimensions of 300x92x60mm. LFW workpiece plate dimensions were extrapolated from the :finished part drawing, Fig. 2b, as was the minimal volume of solid titanium plate required to machine the part from solid, Fig. 2c.

3. Results and discussions

3.1. Effective weld process window

In-process weld monitoring was used across the entire test matrix and the impact of the parametric conditions assessed. Table 2 details the range of the effective weld parameter envelope identified. At high rubbing velocity, Ti-6Al-4V successfully welded at all the pressures applied, and displayed a wide flexibility to process parameters, as demonstrated in Fig. 3a. Initial, transition and burn-off phases combined, the low pressures offered durations up to 6.3s, the medium parametric range were taking typically 1.0s to 1.75s, and the high pressures took as little as 0.7s.

Amplitude and frequency were controlled most effectively during the medium to high rubbing velocity welds. The measure gradients can be seen in Fig. 3b. Thermal gradients as great as 3600°C/s were observed during the 'initial' and 'transition' stages of the welding cycle. This graph also shows that it took around 12s for the joint interface to cool to 300°C [6], the temperature of at which oxidation starts to occur for this alloy, and less than 25s for the interface to drop below 200°C. This results in typical welding cycle durations of less than 30s. The small changes in welding cycle duration observed during this programme allow LFW to be considered as a time dependant process.

Table 2: Process window of some of the parametric  conditions observed

3.2. Metallographic observations

The micrographs of the samples studied showed that all joints made with an average rubbing velocity of 250.mm/s or more had successfully welded regardless of the pressure, and that all bore a characteristic fine-grained equiaxed microstructure while displaying discrepancies in their weld centreline and thermo-mechanically affected zones as shown Fig. 4. These helped in determining a suitable process window. High integrity welds were consistently produced in an unprotected atmosphere. Their optical metallographic examinations revealed them free  from  surface contaminants and alpha case ingress within the weld line; with only the outside of the welds and flash showing traces of such indications. Protective gas shielding during LFW production may therefore be unnecessary, assuming the component' s expelled and outer joint  areas are fully machined post weld.

3.3. Tensile test results

Cross-weld tensile samples were prepared from selected weldments run with mid-matrix parametric conditions. The tensile tests all failed in the parent material, and their results, Fig. 5, reflect this behaviour. No significant reductions in tensile performance were observed, with proof stress, maximum stress and elongation returning parent material values. The only slight variation was a minor reduction in the cross-sectional area of the fracture surfaces.

3.4. Fatigue test results

The test carried  out in this programme  revealed  that all the welded  and parent material  low cycle fatigue (LCF)

and high cycle fatigue (HCF) specimens were above their respective minimum fatigue allowable data ^[7]. The LCF specimen showed a comparative drop in performance  of 27.9% on average from the parent material, see Fig.

The HCF specimens showed a 50.4% average comparative drop in performance, albeit while exhibiting less scatter in the population. This level of performance in as-welded condition is very high, and certainly infers more than satisfactory utilisation for static designed components. Furthermore, it is probable that the performance of these joints may be further enhanced through parameter refinement and post-weld heat-treatment investigations.

3.5. Potential savings by the LFW near-net-shape manufacturing route

It became apparent during the demonstrator down-selection process that LFW could be applicable to a large proportion of modem titanium aerostructure components. The generic near-net-shape manufacturing route  of joining gussets and a boss to a baseplate reduced the raw material usage by 66%, compared to the manufacturing from solid solution. This infers a reduced buy-to-fly ratio from  10.9:1 to 3.7:1. This also translates in a reduction of 49.9% in production time and an estimated 23.3% savings in production  costs.

4. Conclusions

The manufacture of linear friction welded near-net-shape aerostructure components is an effective and lean manufacturing approach with a mature readiness level, supported by clear economic benefits.

1. The  LFW  process  window  was  identified  for  Ti-6Al-4V  Grade  5  AMS4911 on  25mm  thick  plate;
2. All weldments produced exhibited forged microstructures which were free from oxides and contaminants;
3. The as-welded tensile performance of these weldments were effectively equal to that of the parent material;
4. The alternating load LCF and HCF performance exceeded the minimum design allowable values for AMS4911;
5. All weldments in this study were evaluated in the as-welded condition. It is likely that the performance of these  joints   could  be  further  enhanced  by  parameter  refinement   and  post-weld  heat-treatment.

5. Acknowledgments

The authors would like to thank Innovate UK, the UK's innovation agency, and the TiFab consortium, for funding the research presented in this paper. Innovate UK reference 101799.

6. References

1. O. Cauquil: "Airbus Procurement Keynote Address" Proc. 2nd ITA Europe Conf., Birmingham, UK, May 2015. 
2. J Allen: "An Investigation into the Comparative Costs of Additive Manufacture vs. Machine from Solid for Aero Engine Parts", Cost Effective Manufacture via Net-Shape Processing RlO-MP-AVT-139 (pp. 17-1 - 17-10). 
3. H. Wilhelm, et al: "Linear Friction Bonding of Titanium Alloys for Aero-Engine Applications", Titanium '95 Science and Technology,pp.620-635. 
4. A. Vairis, M. Frost: "High frequency linear friction welding of titanium alloy'', Wear,Vol. 217 (1998) pp. 117-131. 
5. YouTube video - Slow motion LFW of Titanium: https://www.youtube.com/watch?v=umEuqbgVOR8
6. M. J. Donachie Jr: Titanium A technical Guide,ASM International, (Ohio 2007), pp203-217. 
7. MMPDS: MMPDS-10 handbook, Battelle Memorial Institute, (Ohio, 2015), pp5.57-5.122.