JOINING, QUALITY, ASSURANCE, AND INSPECTION...In the case of flash butt welded rings, however, ultrasonics can be successfully employed whilst the proof testing of c~ircular rings

SECTION V

JOINING, QUALITY, ASSURANCE, AND INSPECTION

CRITICAL REVIEW

JOINING, QUALITY ASSURANCE AND INSPECTION

Raynor H. Wedge Rolls-Royce (1971) Limited

Bristol, England

Introduction

The continuing need for ever-lighter aerospace structural assem­blies has encouraged the introduction of more sophisticated joining techniques in high strength titanium alloys. This trend places a greater emphasis on the quality control, both of the basic wrought or cast materials involved and of the relevant joining technique.

Better House-Keeping

Titanium presents special problems in terms of quality control throughout the whole of its processing, right from the ore stage through melting, ingot, billet, casting, manipulation, and fabrica­tion. Quality cannot be inspected into a product, it must be built in at the proper time. To this end, every endeavour must be made to introduce adequate systems to ensure that the inspection and con-trol functions are properly integrated to form an effective overall quality assurance. The particular affinity of titanium for atmospheric gases can cause problems. Segregation and inclusion are additional fea-tures which have to be detected. Of these, hard alpha segregates or high interstitial defects are particularly difficult. Figure 1 shows a typical ·example of a micro section of a hard alpha segregate. Pro­viding such defects break the surface, then liquid penetrant tech-niques in conjunction with a suitable etchant are effective. The detection of inclusions and sub-surface discontinuities, whether brought about by a forging burst or by cracking at hard segregates, is best achieved by ultrasonic means. Also, frequent over-checking throughout the whole manufacturing cycle, together with the

521

522 R.H. WEDGE

elimination of surface defects, will minimise the possibility of defects being present in the finished components. Once again, this is a question of good house-keeping, although every effort must con­tinue in the development of improved non destructive testing tech­niques.

The detection of surface cracking, be it from machining, weld­ing or operation use, is frequently made more difficult by the narrowness of minute cracks. Considerable effort has been expended in improving the sensitivity of liquid penetrants. There are schools of thought that consider it essential to prepare the surface by a suitable etchant prior to crack detection.

Where there is a prevalence of cracking in a localised area under service conditions, the application of a suitable loading cycle during the application of the liquid penetrant mC\}' well accen­tuate any surface defect present.

The use of edd,y current and ultrasonic techniques is showing returns but the efficacy IDC\Y be limited by difficult local geomet­ries associated with the component design.

Fabrication

There is a range of joining processes for titanium alloys that a.re well-established including:

1. Argon arc welding 2. Resistance welding 3. Electron Beam welding 4. Friction/Inertia welding 5. Brazing 6. Diffusion Bonding

Each of the above techniques will be reviewed in terms of the problems associated with the production of joints together with the relevant inspection techniques.

Argon Aro Welding

Argon arc and plasma welding have been successful in welding a number of titanium alloys. Table I shows proof and ultimate strengths together with ductilities before and after weld­ing, for the single phaae alloy Titanium 2.5%Cu and Titanium 6Al 4v.

CRITICAL REVIEW 523

Table I. Mechanical Properties found in Titanium Fabrications

Material Ti-2.5cu Ti-6Al-4V

Unwelded 0.1% Proof 42.5 60 (tons/sq.in.)

U.T.S. 48.5 70 (tons/sq.in.)

El (%) 5.65 J A 18 11

Welded 0.1% Proof 34.5 62.1 (tons/sq.in.)

U.T.S. 44 70 (tons/sq. in.)

El (%) 5.65 IA 15 10

Heat Treatment Solution Treat- 700°C Stress 0 ment 805 C Nuc- relieve

leate 400°c (4 hours) Weld. Age 475°C (4 hours)

Because of titanium's affinity for atmospheric gases their exclusion is essential during welding to ensure good sound joints. Particular attention has to be paid to the cleaning of parts prior to welding. It is generally accepted that abrasive cleaning, fol­lowed by the use of solvents is a reasonable compromise for pro­duction purposes.

The development of automated plasma welding has significantly improved weld bead geometries, particularly when fabricating sheet metal above .080" thick and this has been further enhanced by the introduction of wire feeding. The need now is to improve further the fade in and fade out conditions. Figures 2 and 3 show the profiles associated with top weld beads of both manual argon arc welding and automated plasma welding respectively. It should be noted that the plasma width is narrower, whilst the crown and underbead are more uniform along the weld run.

Satisfactory joints between commercially pure titanium pipes and titanium alloy end fittings have been achieved using an orbital welding technique.

)

524 R.H. WEDGE

Fig. 1. Micro section of a hard alpha segregate.

Fig. 2. Top bead of a manua 1 argon a re weld, • 080" thick.

CRITICAL REVIEW 525

Despite the continuing development, argon arc welding has marked limitations when dealing with difficult assemblies such as turbine compressor rotors, where distor.tion must be kept to an absolute minimum.

Fabrications are subjected to radiographic and liquid pene­trant examination, both of these techniques being well-tried methods• If extreme resolution is required for detecting very small cracks, micro focus projection radiography has been used with encouraging results. Figure 4 shows a conventional radiograph of a titanium weld, full-size, and a projection enlarged image of a selected portion with very fine crack detail revealed clearly.

Resistance Welding

Spot, seam and flash butt welding of titanium are exceptions where satisfactory welds oan be produced without shielding. It is considered that the success achieved is attributable to the short thermal cycle together with the applied pressure minimisink the possibility of trapped air. It is implicit that care must' be taken with surface cleaning prior to welding. There is no really satisfactory W83 of testing the integrity of spot and seam welds other than the time-honoured methods of taking test coupons period­ically during a production run, which is only justifiable because the process is fully automated. The acoustic techniques that dis­criminate the solidification noise levels of the joint are proving to be effective monitors of joint homogenity.

In the case of flash butt welded rings, however, ultrasonics can be successfully employed whilst the proof testing of c~ircular rings does afford an effective "go - no go" test. Figure'5 shows a ring 8211 in diameter, welded cross sectional area 7.7511 sq., weight 700 lbs., material Titanium 6Al 4V, that has been success­fully flash butt welded. Table II shows properties found in such a component.

~able II. Titanium 6Al 4V Flash Butt Welded Ring - Mechanical Properties

Position .2'fo Proof U .T.S. % Elongation % R of A tons/sq.in. tons/sq.in. 5.65 j A

A 58.3 64.3 13 32

B 57. 1 63.8 12.5 33 -

c 57.4 65 16 36

The above results show consistent mechanical properties across the joint face, the test piece positions being indicated in the sketch on page 528.

526 R. H. WEDGE

Fig. 3. Top bead of an automated plasma weld, .080" thick.

Fig. 4. Conventional radiograph of a titanium weld, full-size, together with a micro focus projection of a portion of the same weld.

CRITICAL REVIEW

Fig. 5. Flash butt welded ring, 82" in diameter, in titanium 6Al-4V.

527

528 R.H. WEDGE

A

B c

Electron Beam Welding

This technique is normally reserved for high integrity applic­ations, particularly on thick sections, where minimum distortion is a pre-requisite. The following areas are considered those that are still requiring attention.

a. It is necessary to ensure that full penetration has been achieved and this can be controlled by a closed loop system utilising the strength of the beam after it has passed through the joint. Figure 6 shows the effect of incomplete penetration with a _narrower beam, porosity occurring at the tip.

b. The beam control system to ensure that the joint line is accur­ately followed and encompassed. At present, it is necessary to use wider beam widths up to .040/.05011 in order to achieve this and the improved ductilities accruing are considered an inciden­tal advantage. Figure 7 illustrates a disadvantage of a narrow beam weld which has not encompassed the joint, although full penetration has been achieved.

c. Both upper and lower bead geometries affect the joint fatigue strength, the degree depending upon the loss of section and/or the bead shape. Figures 8 and 9 show good and bad beads associ­ated with electron beam welds. To solve the upper bead problem is relatively simple, either by providing a lip preparation, subjecting the bead to a cosmetic run or by carrying out post­weld machining. The lower bead geometry is more difficult to control and it is not always possible to carry out post-weld machining or cosmetic runs on the internal surfaces of complic­ated assemblies. Unlike other materials, cosmetics runs on titanium alloys are less effective in that the correcting runs tend to follow the original bead shape discontinuities. This technique would benefit from further attention.

A better understanding of machine parameters and material weld

CRITICAL REVIEW 529

Fig. 6. Electron beam welded narrow beam, porosity at the tip.

Fig. 7. Electron beam welding, full beam penetration, without fully encompassing the joint.

530 R. H. WEDGE

Fig. 8. Electron beam welds showing good bead geometry.

Fig. 9. Electron beam welds showing poor bead geometry.

CRITICAL REVIEW

variables might lead to more consistent top and bottom bead geometries.

531

In addition to three dimensional manipulation of the weld beam to ensure that the weld runs are accurately and consistently followed, the elimination of operator error by numerical control and automation is regarded as essential. This gives added confid­ence in the quality being obtained in the joint.

The additional hazard of spatter in the under-bead area needs attention and the development of more effective and easily applied guards is required.

A promising development of electron beam welding is that of joining titanium end fittings to pipes using a glow discharge tech­nique, the advantage being that environmental conditions are more readily monitored and controlled, together with the effect-that the distortion is minimised as the weld joint is produced simultane­ously and hence no fade in and out problems.

At the present time, there is renewed interest in the smaller, lower voltage electron beam weld guns, due to their maneouvre­ability• It is recognised that where welds can be located such that gravity effects are minimised, better bottom weld beads result.

The post-weld inspection of electron beam welded components includes etching followed by binocular examination, ultrasonic testing where access is readily possible (although delta scan tech­niques give greater flexibility in the more difficult cases) .radio­graphic and liquid penetrant techniques. While the use of liquid penetrant is very effective for picking up surface imperfections, radiography for finding porosity, the ultrasonic methods are proving to be the most effective for detecting fusion defects and cracks.

Friction/Inertia Welding

Initially, the process was restricted to circular cross sections, but later developments are enabling non circular geometries to be joined. The process also has the potential of enabling the. joining of dissimilar alloys. High axial forces together with high torques are exerted during the joining operation and this demands high rigidity of the basic machines, fixtures and even components.

When considering the relative merits of friction and inertia welding it has been found easier to monitor and control the welding parameters associated with inertia welding and this must result in a higher degree of quality.

The tensile properties of joints are equivalent to those of

532 R.H. WEDGE

the parent materials when joints are produced by either technique, although the fatigue strength is more consistent with inertia weld­ing. Again, the development of an automated process is essential together with a read-out of meaningful process variables to ensure adequate quality control and record. Of the many para.meters being recorded it still remains to be solved which are the most pertinent from the quality control aspect. Figure 10 shows a typical read­out obtained during an inertia welding cycle. Part of the inspec­tion procedure should demand the examination of such a read-out. There are implications that the degree of upset has a bearing on the final joint strength and it is here that further work is required to confirm these early findings. The other areas that require attention are:

a. more precise control of axial dimensions during the joining process.

b. further work on the control of angular positioning of one part relative to another.

The upset created during joining necessitates the machining of the outer and (where relevant) inner surfaces in order to minimise the lowering of fatigue strength. On the credit side-, the lack of fusion minimises the degree of micro cracking.

Figure 11 shows a section through an inertia weld of Titanium 6Al 4v. Figures 12, 13 and 14 show the structures at positions A, B and C respectively, as indicated on Figure 11. Figure 12 shows an over-heated structure in the weld area, which is normally unacceptable due to a drop in fatigue. It has been proved that this particular structure, when occurring in an inertia weld zone does not result in loss of fatigue properties. It is deduced that the amount of residual mechanical work at the weld interface is sufficient to compensate for the unsatisfactory structure, in spite of post-weld heat treatment. Table III shows the mechanical properties associated with a Titanium 6Al 4V inertia welded joint.

Table III. Room Temperature Tensile Properties

0.2% Proof Stress U.T.S. % El. % R of A Fracture (tons/sq.in.) (tons/sq. in.) 5.65 IA Position

61.5 64 10 35 Parent Metal

Table III is continued on page 536.

CRITICAL REVIEW 533

'I tt '-~· c'

b • ... ~ ~.ft • ,., ~~; ';iiwR~ '(~ i T

'' I

I

Fig. 10. A typical read-out obtained during inertia welding cycle.

Fig. 11. Section through an inertia weld of titanium 6Al-4V.

534 R.H. WEDGE

Fig. 12. Structure at position A as indicated on Figure 11.

Fig. 13. Structure at position B as indicated on Figure 11.

CRITICAL REVIEW 535

Fig. 14. Structure at position C as indicated on Figure 11.

536 R.H. WEDGE

Charpy Impaot Properties (Plain - 0.211 dia. x 2" long)

Specimen Energy absorbed (ft. lbs.)

Welded 33 UB

Parent Material 32 UB

The area that has to be satisfactorily resolved is that of non destruotive inspection of joints. Methods used so far have included radiography, penetrants, eddy currents and ultrasonic techniques. Limited success has been obtained with ultrasonic techniques. Both delta scan and longitudinal wave pulse echo methods have been successful in the detection of partially fused or weak joints, whilst eddy currents are mainly useful for s·howing up surface imperfections. Good house-keeping together with the maximum of aut.omation in the process become almost essential requirements.

Diffusion Bonding

Efforts have been made to exploit the significant advantage offered by diffusion bonding, whereby the final bonded joints require a minimµm of subsequent machining, although this is wholly dependent on the ultimate geometry of the component.

No.matter which technique is ad.opted, whether it be high or low pressure,, high or lo'A'. temperatures, air or vacuum, all at present are economically disappointing, essentially due to the degree of precision required at the joint faces. There is one exception, that of continuous seam, diffusion bonding where constant section geometries typical .of those found within the airframe industry appear to be a viable proposition. The process utilises the simple rolling technique. Figure 15 shows a continuous seam diffusion bonded joint in Titanium 6Al 4v. The micrograph shows the uniformity and the extent of the heat affected zone together with the good geometric proportions of the blend radii. Here again, although there appears to be an over-heated transformed beta structure in the heat affected zone, the mechanical properties attained those of the parent material, presumably because the metallographic structure is relatively unimportant for this par­ticular joint configuration.

The ability to detect stuck joints presents serious problems. The infra-red technique (suitable for egg-box structures and honey­combs) and the ultrasonic C-scan methods are being used to deter­mine the soundness of joints.

CRITICAL REVIEW 537

Fig. 15. Continuous seam diffusion bonded joint in titanium 6Al-4V.

538 R. H. WEDGE

Fig. 16. Typical TiCuNi diffusion brazed joint using a .004" thick foil.

Fig. 17. Micrograph of an activated diffusion bonded joint between titanitm1 6Al-4V.

CRITICAL REVIEW 539 .·

However, none of these techniques will detect a planar grain boundary condition. It is, therefore, most desirable that.research activities should continue.to establish a satisfactory and practical non destructive technique.

Unless the economics are improved and better joint-N.D.T. techniques are developed, diffusion bonding may have difficulty in finding a place in future joining methods.

Brazing

Experience has been centred around torch and furnace brazing, the latter accounting for the major part of the experience. The inherent metallurgical properties of titanium are such as to cause the fillers to form alloys with the base materials, these generally being brittle inter-metallic compounds. This problem was shown up in the early developments and disappointments ~ere experienced with the silver based alloy fillers. The ductility problems have been overcome to some extent by the development of alloy '.fillers of the AlTiZrBe and TiCuNi types. Using the filler in the powder, foil or even electro-plated form, brazed joints can be obtained. Figure 16 shows a typical TiCuNi .diffusion brazed joint using a .004" thick foil.

Diffusion brazing of titanium alloys using TiCuNi is only satisfactory in cases where the filler is sandwiched in the joint and for the simplest capillary joints. As yet, a satisfactory capillary braze alloy suitable for complex assemblies wh~re gaps have to be filled is awaited. Although work is proceeding on such brazing alloys no commercially viable solution t:ias b_een launched.

A further development known as activated diffusion bonding or brazing has shown that under very low pressures, 1~20 lbs/sq.in. sandwich joints can be achieved where it is difficult. to see a joint line. Figure 17 shows a micrograph of such. a joint between Titanium 6Al 4V. This appears to be very encouraging as there is a noticeable absence of planar grain boundaries at ,the interface and the absence of residual inter layers.

Energy methods such as ultrasonics are being utilised as methods of checking the integrity of joints since radiographic techniques will only indicate that the braze alloy is present.

Challenges

1. Is triple melting the answer to segregation problems?

2. No mention has been made of sophisticated assessment and

540 R.H. WEDGE

reduction of residual internal stress in finished componEnts, arising from manipulation, machining, heat treatment, joining and operation stress/temperature environment. I look forward to comment in this direction today.

3. Im pl icH in ·Q.uali ty Assurance is the continuous maintenance of rigid process control at each and every stage of manufacture of the me.terial and component.

4. The joining of titanium to other conventional me"als has still to'be effectively demonstrated by other than classical bolting mettods.

5. The aerospace industry awa:!.ts further developments of high­strength, high-temperature, readily joined titanium alloys.

6. It is hoped "hat titanium castings will show economic advan­tages when incorporated into fabrications.

7. From the galaxy of talent present a useful contribution is anticipated in the field of N.D.:1

• anC:. Quality Assurance of titanium allo,ys.

Acknowledgements

The author wishes to acl;nowledge the help and advice given by his colleagues at Rolls-Royce in the preparation of this paper and to thank the Directors of the Bristol Engine Division of Rolls­Royce for their permisston to publish this paper. However, the opinions expressed are his own and do not necessarily represent the views or policy of Rolls-Royce (1971) Limited.

Also; due acknowledgement is m1ide to the followini:;~ companies:

1. Reynolds Tube Company (U.K.) Limited, for information on flash butt welded rings;

2. Solar, ·Division of International Harvester Co., San Diego, California, U.S~A., for information on continuous seam dif­fusion bondin§;;

3. The Non Destructive Testinc Centre at the Atortic Energy Research Establishment, Harwell, England, for information on micro focus projection radiography and glow discharge electron bearr; welding.