G S Booth, R L Jones and P L Threadgill
Paper presented at ISOPE-2006, 16th International Offshore and Polar Engineering Conference, San Francisco, California, USA, May 28 - June 2, 2006
Developments in welding technology applicable to the offshore industry are described in three broad themes; advanced arc welding, friction stir welding and laser processes. In each case, selected advances are reviewed and potentialapplications and benefits in offshore engineering are presented.
Welding is a fundamental technology in the fabrication and repair of virtually all structures in the offshore industry, whether they be above or below sea level, or onshore. It is an enabling technology without which the offshoreindustry (and many other industries) could not operate at its present level of sophistication, and yet welding technology often sits in the background, taken for granted as a mature and established technology. Like most technologies,it is developing steadily over time, allowing new benefits in terms of what can be achieved, and in terms of process economics.
This paper seeks to review some of the recent developments in a selection of welding processes which may have potential for use in the offshore industry. Those chosen are:
Advanced Arc welding
Friction Stir Welding
Advanced arc welding
Arc welding continues to be the preferred welding method for the manufacture of offshore structures, pipelines and process equipment. Different process techniques and welding consumables are available to suit the requirements of arange of offshore applications. The ability to operate on-site is also a major advantage. Whilst no fundamentally new arc welding techniques have emerged in recent years, there have been significant developments focused on improvingprocess efficiency and productivity together with the facilitation of increased welding automation. Notable recent TWI developments of significance to the offshore industry include:
- Low cost activating fluxes for GTA welding.
- Novel keyhole plasma welding procedure.
- Autonomous robotic welding of large structures.
Activating flux for GTA welding
The use of activating fluxes provides a novel method of increasing the penetration capability of the arc in GTA welding. The process is effected through the application of a thin coating of an activating flux material onto the joint surface prior to welding. The effect of the flux is to constrict the arc which increases the current density at the anode root and the arc force action on the weld pool. The constricted arc produces a narrow deep weld compared with the wide shallow weld bead obtained with conventional GTA. Activating fluxes are available commercially for welding a range of materials, including C-Mn steel, Cr-Mo steels, stainless steels and nickel-based alloys. The fluxes are generally available in the form of either an aerosol or as a paste (powdered flux mixed with a suitable solvent, eg acetone or isopropanol) which is applied to the surface with a brush. The activating fluxes can be applied in both manual and mechanised welding, although it is more difficult to control in the former mode of operation.
Despite the productivity benefits of activated flux GTA welding, industry to date has been slow to exploit the process. This is because the relative cost of commercial fluxes is high and flux application is an additional operation.Furthermore, the commercial fluxes tend to produce an inferior surface finish compared with conventional GTA welding and give a surface slag residue, which is required to be removed. In order to mitigate these disadvantages, TWI has developed a low cost activating flux with the following characteristics.
- It comprises a readily available flux ingredient, which can be applied as a water based paste using a brush or spray applicator. It can be procured as a standard chemical reagent. When purchased in this form, the flux cost is estimated to be less than 15 cents per metre of weld length.
- The flux ingredient is non-toxic. It contains no halides or fluorides.
- Flux performance, including depth of weld penetration, is similar to alternative commercial fluxes, ie up to 12mm in stainless steel compared to 3mm for conventional GTA welding.
- It is suitable for welding carbon steel and stainless steel.
- It produces a satisfactory weld deposit surface appearance with minimal slag residue.
The deep penetration action of activating fluxes can be used to achieve significant productivity benefits for orbital welding of stainless steel tube. Tube wall thickness of up to 5.5mm can be welded in a single pass using a square edge closed butt preparation ( Fig.1). Conventional GTA welding procedure requires a bevel preparation together with the deposition of several weld passes to complete the joint.
Fig.1. GTA with activating flux (left hand side) and conventional GTA (right hand side) welds in 48mm OD 4mm WT 304L stainless tube
Keyhole plasma welding
The most significant difference between GTA and plasma welding lies in the ability of plasma welding to operate in the keyhole mode. Compared to GTA welding, the keyhole plasma welding technique has several advantages including deep penetration and high welding speeds. The keyhole plasma welding technique can penetrate stainless steel material thickness up to 10mm. However, despite the apparent advantages of the keyhole plasma technique, it has not been as widely exploited as the microplasma technique. In the past, several factors have discouraged its more widespread use, but these have been largely overcome in recent years as a result of welding equipment improvements, which can be summarised as follows:
- Lower cost plasma welding equipment.
- Improved design and reliability of plasma welding torches.
- Improved control of welding parameters including plasma gas flow. This is necessary for closing the keyhole in pipe welds for example.
The deep penetration characteristics achievable with the keyhole plasma welding technique have been exploited to develop novel welding procedures for a range of alternative joint designs in 2.5mm thick austenitic stainless steel.These joint designs could not be easily welded with conventional GTA and GMA welding:
- T-joints can be produced from the external side using a two pass sequence resulting in a fully fused weld with a reinforcement fillet on the internal side ( Fig.2). The level of distortion is also very low.
- Full penetration stake welds in plates of 2.5mm thickness can be produced successfully from one side.
- Similarly, edge lap welds can be made between overlapping 2.5mm thick plates, welding from either side of the joint.
- Full penetration spot welds between overlapping 2.5mm thick plates can be made from one side.
These novel welding procedures offer the possibility of increased flexibility in the design and manufacture of thin sheet components. When using conventional GTA and GMA processes, the design of such components is influenced by the relatively high distortion associated with these processes and the access requirements needed for manufacture. Some of these limitations can be overcome by the use of the laser beam process, particularly with respect to the control of distortion. The present work has demonstrated that similar benefits (but to a lesser extent) can be accrued through the use of the keyhole plasma technique but with modest capital equipment costs and without the need for exacting joint fit-up.
Fig.2. Keyhole plasma welding of a T-joint in thin austenitic stainless steel to give full penetration welds
Courtesy of Rolls Royce
Autonomous robotic welding of large structures
European manufacturing companies are facing increasing competitive pressures which are encouraging a trend to the use of more automated welding methods. Customer preference for customised products, the requirement for shorterdelivery times and a shortage of skilled welders are also contributory factors.
In response to these pressures, NOMAD, the name given to the EU framework project, 'Autonomous Manufacture of Large Steel Structures', was established. The goal of the project was to create a fabrication system capable of weldingsmall production runs and even unique 'one-off' structures, as easily and quickly, as large multiples. This was to be achieved by eliminating as many of the current constraints as possible, for example, eliminating the use of fixturesand dedicated work piece manipulation. The project solution was to develop an autonomous mobile welding robot which moved around a stationary component to complete the required weld joints ( Fig.3).
Fig.3. In-situ welding of an excavator stick using a mobile robot vehicle
In the project, a demonstration cell was developed capable of fabricating steel structures in the 5 to 50 tonne range without the need for special tooling and dedicated handling equipment. Components were loaded into the cell and identified by a vision system. The image of the component was matched to the CAD model and offline programs generated path data to guide a robot arm on a moving platform into position. A laser scanning sensor on the robot arm located the weld start position and welding was carried out under adaptive control.
The project was concluded with a successful public demonstration of the welding system which involved a representative bridge section and a stick, a typical component part of an excavator. The welding system is equally suitable for many applications in offshore engineering.
Friction Stir Welding
Friction stir welding was first developed at TWI in 1991, and the initial patent application was filed in December of that year (Thomas et al, 1991). Initially, it was seen as a process solely for aluminium, and of limited interestto the offshore industry, but this viewpoint has now changed. Much research has been undertaken worldwide to develop the process for steels and other high strength corrosion resistant alloys, for example nickel and titanium alloys.
Fig.4. Principle of friction stir welding
The principles of the process are well documented, and are shown in Fig.4. For butt welds, a rotating cylinder is pushed against the surface of the weld. This cylinder is attached to a pin which penetrates almost the entire depth of the weld. Rotation under pressure causes the development of frictional heat, which softens the workpiece material to the point where it can flow. At this point, the rotating tool is moved along the joint line, and the softened material in front of the tool is extruded between the pin of the tool and the cold material on one side of the pin. During this process, the interface is completely fragmented, and so a solid phase joint is formed behind the tool.
The process has the advantage of being fully mechanised, therefore weld quality is reliable and repeatable, and no filler metal is required. However, an inert shielding gas is strongly recommended for welding steels. The process is entirely solid state, so hot cracking, porosity etc are not encountered. The process is also very energy efficient, and produces no fume, spatter or radiation. The principal disadvantages are the need for control of fit-up (although perfection is not required), and the need for more substantial fixturing than is common in other processes, due to the high process forces. The exit hole left by the rotating probe can be a problem on closed loop welds (e.g.circumferential welds) but there are methods for dealing with this.
The major challenge with friction stir welding of steels has been the choice of tool material. The tool experiences very high temperatures, often of the order of 1200°C, along with high rubbing stresses and high process stresses. It is no surprise that very few materials have the potential to survive these conditions.
Fig.5. Examples of basic friction stir welding tools for 5mm material a) PCBN
Initial studies concentrated on refractory alloy tools, based on the W-Re system. This alloy maintains its strength to temperatures well in excess of the likely service temperatures, but unfortunately its low temperature strength isnot particularly high. However, there are very few stronger alloys at 1200°C. However, the potential of improving on what is a very simple alloy system is recognised, and much improved alloys are now becoming available. Atpresent, their performance is under investigation. Some typical first generation tools are shown in Fig.5, although the designs are evolving.
An alternative approach has been to use ceramic tools, in particular polycrystalline cubic boron nitride (PCBN). This is the second hardest material known, but it is also very brittle, meaning that it must be treated with greatcare. Friction stir machines which have been specifically designed to eliminate almost all vibration are also needed to allow such tools to operate. Since PCBN is very brittle, the complexity of the tool design is limited, as it isessential to avoid stress concentrating features. The benefit of this material is the very high strength and low wear under operating conditions. Some tools have lasted for over 100m of weld, a remarkable performance by any standard,but this level of performance cannot yet be guaranteed. However, like W-Re tools, the quality and performance of the tools is steadily improving.
Neither tool material is perfect. PCBN has excellent wear properties, but is very fragile. Once worn or broken, the tool cannot be re-dressed, whereas W-Re tools are very tough, and worn tools can be re-dressed several times, givinga substantial cost advantage.
A third option, currently under development, is to use alternative ceramic materials. Although these may have shorter lives than PCBN, they are cheaper, and can be re-dressed when worn. The economic equation to predict the bestsolution is very complicated, and there is at present insufficient data to make informed decisions in most cases.
Fig.6 shows examples of friction stir welds in several steels. The process has been demonstrated for a variety of pipeline steels up to X100, as well as many other high strength steels (HSLA-65, RQT700). In addition, successhas been reported for various CRAs such as 316L, 304L, duplex and super-duplex stainless steels, 9%Ni, 13%Cr and other non-ferrous alloys such as IN600, IN625, Ti-6Al-4V etc.
At the time of writing, friction stir welding is not used in production for any steel fabrication, but its potential advantages and rapid rate of technology development suggest that this will change. Since no filler is required,problems of finding suitable consumables for high strength pipelines will disappear. The process is fully automated, so should give very repeatable results, and the potential for remote operation exists, although at present this hasnot been exploited.
Since the process occurs entirely in the solid state, many of the problems associated with fusion welding (e.g. solidification cracking, porosity, lack of fusion, spatter, fume etc) do not occur. There have been no reports of anyhydrogen related problems with the welding process, which is perhaps not unexpected. In almost all materials, the centre of the weld contains a microstructure consisting of fine equiaxed grains caused by dynamic recrystallisation,although the detail in ferritic steels and many titanium alloys is complicated by the allotriomorphic phase change which occurs on cooling. Further from the weld, the microstructures generally resemble conventional fusion welding HAZs,although the extent of grain growth is generally less due to the lower heat input and peak temperatures.
At present, the volume of data on weld properties is limited. This is understandable, as efforts are concentrated on perfecting the process rather than establishing weld performance. The data available have not yet shown anyunexpectedly poor results. Similarly, techniques for NDE are still under development.
Fig.6. Macrosections of friction stir welds in various steel grades a) 5mm 316L stainless, made with W-Re tool
b) 12mm X80 pipe steel welded with PCBN tool
c) 8mm EN10025-S355 steel welded with ceramic tool
Equipment for friction stir welding is available commercially, although all machines for production and near-production development have been individually designed and built. This obviously adds to the cost of the equipment. Stateof the art equipment is now very sophisticated compared to what was available just a few years ago. For example, the pin and shoulder can be separate entities, each with its own control system, and this allows immense flexibility inprocedure development, and also minimises the risk of defects appearing in the final weld. Tools are also available with two opposed shoulders, known as bobbin tools or self-reacting tools. The force between the two shoulders iscontrolled, and this means that the equipment does not have to react the high z-axis force. However set-up and breakdown of the weld are more time consuming with this system.
Once the not insignificant challenge of the tool material has been overcome then it is expected that friction stir welding will become attractive for many niche and specialised applications. However, it is thought unlikely that itwill displace more conventional and well established arc welding processes for many applications.
The twentieth century was characterised by the harnessing of electrical energy to create broad economic and societal benefits. Examples of developments enabled by electricity include transport, heating, lighting and the recentexplosive expansion of information and communication technologies. Looking to the future, innovative applications of energy in the form of light have the potential to shape the twenty-first century.
Indeed, the previous century has been described as the century of the electron; the twenty-first century will be the age of the photon.
Over the last forty years or so, lasers have matured from laboratory prototypes into reliable and efficient industrial tools. Typical applications range from cleaning of artwork through automotive manufacture to ship hullconstruction.
Irrespective of the component or structure, all manufacturers are continuously striving to gain a competitive edge over their rivals and the advantages and disadvantages of each manufacturing route are thoroughly evaluated beforeproduction begins. Fabrication or manufacture with lasers offers several potential benefits:
- Increased productivity - more rapid part manufacture at reduced cost.
- Improved quality - for example, minimal distortion and hence reduced levels of rework.
- Enhanced performance - this can be seen by longer lifetimes or greater resistance to corrosion or high temperature. Additionally, laser repair technologies can give rise to significant life extensions of otherwise exhausted components.
- Novel fabrication routes and components - where alternative technologies simply are not available.
The choice of manufacturing route, however, is essentially an economic one. Consequently, there is a continuous battle between manufacturing technologies to gain a competitive edge that is closely analogous to the competition between individual manufacturers.
Due to the narrow, deep penetration weld produced, laser welding offers several advantages:
- high joining rates
- low consumable costs
- a reproducible, machine tool welding process
- low manning levels
- low degrees of distortion, leading to greater precision in assembly and reduced rectification.
Two main types of laser have been used industrially for welding thick section steel - CO 2 gas lasers and Nd:YAG solid state lasers. CO 2 lasers generate light in the infrared regime at a wavelength of 10.6µm and are available at power levels up to about 45kW. The light is transmitted from the laser to the workpiece by a system of mirrors,before being focussed to a small spot for welding. This type of laser is now being used in European shipyards for different aspects of ship fabrication, where the reduced distortion enables the costs associated with rework to be dramatically reduced.
Nd:YAG lasers, in contrast, generate light at a wavelength of 1.06µm, which is also in the infrared regime, but can be transmitted by a fibre-optic cable. This is a major advantage, as it enables the complex set of mirrors used for the beam path of a CO 2 laser to be replaced by a simple fibre and welding head that can be mounted on a welding robot - Fig.7.
Fig.7. Nd:YAG welding head positioned in robot arm, showing fibre delivery capability
Advances in laser technology have led to rapid increases in the power available in Nd:YAG lasers and 4kW Nd:YAG lasers are now extensively used in production.
Recently, low power fibre lasers, which have been used in communications for many years, have been developed to the stage where power levels suitable for welding and cutting are available. The lasing medium is a Yb:YAG fibre and thewavelength of light emitted is 1070nm, very close to the Nd:YAG wavelength. Manufacturers claim that fibre lasers have better beam quality than Nd:YAG lasers of a similar power, enabling greater stand-off distances when welding.Additionally, wall plug efficiency is estimated to be between 20 and 30%, significantly greater than Nd:YAG or CO 2 sources; thus relatively low input power and reduced cooling is required. Fibre lasers are also smaller than Nd:YAG or CO 2 lasers of similar power.
Welding performance data for the fibre laser are now available (Verhaeghe 2005). In very broad terms, the results obtained when using a fibre laser are at least as good as those obtained from a comparable Nd:YAG laser. As greaterpowers and increased electrical efficiency are associated with the fibre, it is clear that this new laser source may have many industrial applications.
Laser welding using any source, however, is a very demanding process and tight control of fit-up and steel composition is required to produce welds of satisfactory quality and performance.
Hybrid laser-arc welding for girth welds in land pipelines.
An enhanced range of applicability is available through hybrid laser-arc welding, in which the two welding processes are coupled in a single process zone, as shown in Fig.8. Options available include combining either a CO 2 , Nd:YAG, diode or fibre laser with an arc welding process such as gas metal arc welding, gas tungsten arc welding or plasma arc welding, or even combining two different laser sources.
Fig.8. Typical hybrid laser-arc welding arrangement
Compared with the use of laser power alone, hybrid laser-arc welding offers:
- Increased travel speed (x2) or increased penetration (x1.3).
- Improved tolerance to fit-up gap.
- Ability to add filler material to improve weld metal microstructure, joint quality and joint properties.
- Potentially improved energy coupling.
- Increased heat input and reduced hardness.
There are, however, some drawbacks which include increased complexity ('more things go wrong'); the need to define additional welding parameters; and the requirement to establish the process parameters anew, as these cannot simply be determined from the optimum procedures for the two separate processes.
Nevertheless, hybrid laser-arc welding is now a production process in both the automotive and shipbuilding industries and has been shown to be a candidate process for girth welding gas transmission pipelines.
One approach (Howse 2005) to making girth welds in land pipelines is to deposit an internal GMAW root run using conventional techniques. A hybrid laser-arc weld is then made from the outside and a final GMAW capping pass is deposited. In this way, a single hybrid laser-arc pass may replace a large number of GMAW fill passes, thus reducing pipeline costs significantly by reducing the number of welding stations on the right of way.
Results showed that this approach was suitable for producing deep penetration welds that met the requirements of API 1104. Acceptable hardness levels were achieved and toughness requirements of 40J at -10°C were satisfied. Fig.9 illustrates a typical macrosection of weld, prior to the deposition of the final capping pass.
Fig.9. Macrograph of a hybrid Nd:YAG laser/MAG weld in X60 pipeline steel showing an acceptable weld with 11mm full penetration
Repair using laser direct metal deposition
For high value applications such as drill bits it may be economically attractive to repair worn or locally damaged parts rather than replace them. A candidate repair technology is laser direct metal deposition in which a laser is used to melt powder onto a substrate material. In contrast to approaches in which the powder is placed on the substrate, the powder is fed via a nozzle into the laser spot on the substrate, as illustrated in Fig.10. This results in the formation of a molten pool which, on cooling, leaves a solid deposit of the powder material on the substrate.
Fig.10. A schematic of laser-based deposition using a single, off axis powder feeder
Laser deposition has two main advantages over other powder deposition techniques. Firstly, the operation is a low heat input process which reduces the likelihood of liquation cracking. Secondly, the use of very small spot sizes enables highly accurate and reproducible deposits to be made. Additionally, adaptive control systems have been developed that monitor the characteristics of the molten pool and adjust the process as necessary to maintain deposit quality.
In addition to repair, however, direct laser metal deposition is suitable for original part build. By varying powder composition as a part is built up, a functionally graded component can be developed with particular performance characteristics directly related to position. The process can also be used for rapid prototyping.
For the process to gain widespread acceptability, however, development is required to increase deposition rates, deposition efficiency and to establish appropriate processing parameters for high accuracy and quality.
This paper has summarised significant recent developments in both traditional and emerging welding processes. It is clear that progress is being made, and this should lead to improved quality, greater reliability and lower costs for the offshore industry. The importance of welding technology to the offshore sector is well recognised, but there is still great scope for improvement.
Howse D S, Scudamore RJ, Booth GS 'Yb fibre laser/MAG hybrid processing for welding of pipelines'. IIW document IV-880-05, 2005. Thomas WM, et al 'Friction Stir Welding' International patent application GB9125978.9, Dec 1991. Verhaeghe G 'The fibre laser - a newcomer for material welding and cutting' Welding Journal vol 84 No 8, Aug 2005, P56-60.
The authors wish to thank their many colleagues at TWI who have contributed to the advances described here.