MIAB - a process rediscovered, for joining thin walled non-ferrous tubulars

TWI Bulletin, January/February 2001

Stephan Kallee is Collaborative Project Manager for Friction and Forge Welding Processes. He focuses on leading the EUREKA EuroStir ^TM project on European industrialisation of friction stir welding and the CRAFT LinFric ^TM project on low-cost linear friction welding machines. Before he joined TWI in 1995, he had obtained qualifications as European Welding Engineer and Dipl-Ing for Mechanical Engineering in Munich.

Recent enquiries by Industrial Members of TWI, together with information generated from a market survey of 50 companies, have identified the need for improvements in joining processes for thin wall tubes ( Table 1 ). As Stephan Kallee reports, automotive, aerospace and oil exploration companies are now considering whether the process could be applied to aluminium space frames, titanium tubular structures and stainless steel tubes. This would enable them to produce joints in lightweight or corrosion resistant structures by a one-shot process with an extremely short welding time.

Although magnetically impelled arc butt (MIAB) welding has been successfully applied to carbon steels for industrial applications, little progress has been made with aluminium and stainless steel alloys. Historically, the problems met when welding such materials have been ascribed to inadequate power sources, unsteady arc rotation, low rates of axial displacement, unsatisfactory gas shielding and a lack of reproducibility. Commercial machine producers also saw a conflict in the relatively long gas purging time compared to the short welding cycle. It is now believed that improvements in power source technology, control systems, magnetic coil designs and gas shielding will provide opportunities for significant improvements in the process when welding these materials.

MIAB welding is a forge welding technique in which heat is generated prior to forging by an arc which rapidly moves around the circumference between tubular workpieces ( Fig.1 ). Although melting occurs at the ends of the workpieces, this process produces solid phase joints, as all molten material is expelled into the flash, when the tubes are forged together.

MIAB welding operates with very short weld times. It produces a flash with little spatter. Only hollow workpieces can be joined, including those with non-round shapes ( eg square or hexagonal tubes). MIAB welding is currently only being exploited for carbon steel tubes with a maximum wall thickness of approximately 5-6mm. ^[1] For wall thicknesses of more than 6mm the arc fails to heat the full width of the joining faces, and so a successful weld cannot be achieved. MIAB welding is also known as rotating arc or Magnetarc ^TM welding.

Although the original MIAB patents and publications date from the 1940s, significant industrial exploitation has only really occurred since 1970. ^[2] The process is commercially applied mainly to joining mild steel tubes and/or cast iron parts. The automotive industry uses it, for example, to produce safety critical parts such as axles, benefiting from the very short cycle times and thus high productivity.

How it works

In general the MIAB process works as follows: to start the process, the tubes are squeezed together and a direct current (DC) is applied to the workpieces. The tubes are moved apart to strike the arc and are kept at a small distance (1-3mm) while applying a radial magnetic field to the joint area. This magnetic field can either be produced by permanent magnets or by applying a direct current to electromagnets. The presence of a magnetic field results in a force on any moving charged particles. The direction of the force is perpendicular to the direction of the magnetic field as well as to movement of the charged particles. In the case of MIAB welding this force affects the current in the arc, which is struck between the ends of the workpieces. The magnetic force accelerates the arc in the tangential direction and the arc follows the circumference at a constant speed. After the arc rotating time, when sufficient material has been molten, the tubes are forged together, while switching off the electric current. The forging force is applied until the heated metal has been consolidated.

Research need

TWI has three MIAB welding machines. One has been purpose built for welding of aluminium alloys. ^[3] This vertical machine has not operated for approximately 10 years, but was re-commissioned in 1996 for welding thin walled aluminium tubes.

This report covers the machine modifications and the principles to be applied for welding thin wall tubes made from non-ferrous alloys and stainless steel. Further experiments will be necessary to develop welding parameters and to achieve repeatable welding results. These experiments were initiated in July 1999 and will be reported at a later date.

Objective

• To develop improved systems for MIAB welding thin wall tubulars of non-ferrous materials and stainless steels

State of the art MIAB welding of aluminium

Existing technology

MIAB welding is used in industry only for materials that do not need to be protected against the atmosphere during welding. However, by applying shielding gases, it is also possible to weld non-ferrous materials and stainless steels. Initial work on welding aluminium alloy tubes has been carried out at TWI and the following principles were used as a starting point for this study.

Welding with a vertical axis

Commercial MIAB machines operate with the tubes in the horizontal position to make loading and unloading easier. Gravity, however, affects the molten material at the tube ends and can cause it to drop to the lower half of the tube circumference. When welding steels, the influence of gravity on the molten steel is relatively low. However, molten aluminium and other non-ferrous materials have the tendency to flow much more to the lowest point of the pipe. In some cases the molten material can even bridge the arc gap and extinguish the arc. Therefore the repeatability of horizontal MIAB welding machines appears to be low when welding aluminium. The use of vertical MIAB machines has been recommended in the conclusions of several studies.

Applying shielding gases

Earlier work on MIAB welding of aluminium alloys ^[4] showed that a shielding gas mixture of argon with 5% hydrogen substantially improves arc initiation, arc rotation speed and arc stability. Furthermore, more heat can be generated and a larger arc gap can be used when adding hydrogen to the argon shielding gas. Although hydrogen dissolves in liquid aluminium and can cause porosity during solidification, ^[5,6] this did not appear to be a major problem for MIAB welding, because the hydrogen saturated material is expelled into the flash. From this work it became apparent that a uniform gas shield is essential and that shielding gas should be supplied from both inside and outside the tube.

Using dual stage current cycles

Industrial MIAB welding machines often use a dual stage arc current. This involves a first stage with a lower current for a relatively long duration and a second stage with a higher current for a short duration. A first stage of 220A/0.7sec combined with a second stage of 740A/0.1sec has been reported by Hone ^[4] as suitable for welding 32 and 38mm diameter 6063 aluminium alloy tubes (1.6mm wall thickness). These parameters were used as a starting point for this study.

Applying high displacement rates

Due to the high thermal conductivity of aluminium, it is recommended that high displacement rates are used when MIAB welding these materials. TWI's vertical machine was able to generate a displacement rate of 710-825 mm/sec when being pneumatically actuated by air pressure from the laboratory ring main (5-20kN force). ^[4] Displacement rates of this magnitude or higher appear necessary to expel molten material and contaminants such as oxides from the weld interface during the forge phase before the molten material solidifies.

Machine modifications

Existing equipment

TWI's MIAB welding machine was originally designed to assess the influence of either a vertical or a horizontal tube axis when welding aluminium alloys. Parts of the machine frame could be tilted to conduct comparison trials.

Prior to the modifications, the MIAB welding machine was operated by a 5.5bar pneumatic system. The maximum forge capacity was nominally 20kN generated by two large diameter diaphragms in tandem position. Although pneumatic systems are known to work faster than hydraulic systems, concerns were raised due to the large moving mass of the actuators of this low-pressure system and owing to a lack of reproducibility.

A flexible blade system was used in the existing machine to allow for axial displacement. However, the resulting movement was non-linear and followed a large diameter arc. The blade spring arrangement ensured that the tube axes remained parallel, but a slight shift in the position of the upper axis was experienced. Previously, Hone made allowance for this offset (normally less than 0.25mm) in the initial set-up of the tube components prior to welding. ^[4]

The machine was originally equipped with a positioning plate to align the lower tube against the upper tube. This plate was actuated by two eccentric levers, which could be secured in position after aligning the tubes. However, it was observed during the initial trials that the set-up was very time consuming due to the superimposition of x and y movements.

Modified equipment

Hydraulically actuated forge system

New control system

More powerful magnetic coils

Improved angular and linear alignment

An insert was used to help with the alignment of the tubes. A stainless steel pin, which was electrically insulated against the tubes by plastic bushes, was inserted into the tubes and used as piloted tooling. Shielding gas was supplied through a centrally located hole in the hollow pin.

New power source for arc current

Improved instrumentation and monitoring

An optical sensor system for monitoring the arc rotation speed has been installed. This system has the advantage of not being affected by the magnetic field of the coils in comparison to a magnetic sensor. The output of the optical sensor system was monitored using an oscilloscope but not saved on the computerised data acquisition system. Enquiries for a very fast computer flashcard, which could store the large amount of data during the short weld cycle, have been made, although this has yet to be purchased. The main parameters to be monitored are voltage, welding current, displacement, coil current, and arc rotation speed.

Better gas shield

A new shroud for applying the gas shield has been designed and manufactured. It was made from electrically insulating materials and has some rubber seals contacting the outer diameter of the tubes. Gas distribution systems to supply the shielding gases at several points from inside and outside the tubes were installed. The gas flow was controlled through a solenoid valve by the PLC satisfactorily.

Experimental programme

Materials

• Aluminium alloy - AA 6061-T4 50mm diameter, 0.9mm wall thickness
• Superduplex stainless steel, 10mm diameter, 1mm wall thickness
• Titanium alloy - Ti-3Al-2.5V, 19mm diameter, 1mm wall thickness

Pilot trials with the pneumatically actuated machine

Pilot trials using the pneumatic actuation of this machine were carried out on seamless 6061-T4 aluminium tubes (50mm diameter, 0.9mm wall thickness, Figs 2-4 ). The results of these trials showed that machine improvements regarding displacement rate, control systems and gas shielding were necessary for welding such materials. Parameter development studies, as well as optimisation and reproducibility trials would need to be conducted after modifying the machine.

When using the hydraulic system, high forces were applied to the machine frame on the initial contact of the tubes. This resulted in bending moments in several parts of the frame. The machine frame, which was originally designed for pneumatic actuation, appeared not to be stiff enough for the higher hydraulic forces. It was necessary to increase the rigidity of the frame by welding several (originally mechanically fastened) cross members to the frame. This improved the angular and linear alignment of the welded tubes. It was also necessary to use end stops to avoid the tubes slipping in the collets.

The linear bearings were difficult to install in the existing frame. Within the confines of the existing equipment it was not possible to precision grind the surfaces onto which the linear bearings were mounted because they were not accessible. Consequently, the tolerances in the linear movement were larger than those achievable with bearings installed on ground surfaces. Further improvements might be necessary and there might be a need to replace parts of the existing machine frame by a new framework, onto which the linear bearings could be mounted with greater precision.

The hydraulic power pack has been evaluated by installing several sensors and the pressure distribution appeared satisfactory. Concepts for a further increase of the flow rate have been developed.

The magnetic coil system and the field distribution appeared to work in a satisfactory manner. The PLC worked very well from the very first trials and minor modifications to the programme were easy to conduct.

Experimental results

Commissioning trials were carried out on super duplex stainless steel (19mm diameter, 1.5mm wall thickness, Fig 7 ). These trials showed that it was necessary to install a smaller diameter hydraulic actuator to reduce the forging force and flow rate. The diameter of the cylinder bore was reduced from 50 to 25mm. The new 25mm diameter cylinder worked well for the small diameter tubes, but further improvements appear necessary for achieving higher displacement rates especially for welding aluminium alloys.

Commissioning trials were also made with Ti-3Al-2.5V tubes (19mm diameter, 1mm wall thickness, Fig 8 ). The titanium experiments are documented on a TWI videotape, which can be ordered as an addition to this report. ^[7] Macrosections of a typical titanium weld of the commissioning study are shown in Figs 9 and 10. A weld parameter record of this weld is presented in Fig 11 showing non-optimised parameters.

Discussion

The machine is a substantial improvement over the original. Some trials carried out on stainless steel and titanium tubes showed improvements in reproducibility, but highlighted areas where further improvements might be needed. These are particularly related to displacement rate, machine stiffness and precise angular and lateral alignment of the actuating mechanism. Components such as the new control system (PLC), the new power source, the gas shield and the coils work very well.

Further trials are underway to assess fully the capability of the machine, and to identify where any further improvements might be needed. The results of these trials will be reported at a later stage.

It is expected that this work will provide a unique resource, a fully operational and flexible MIAB welding machine, capable of welding a wide range of non-ferrous and stainless materials.

Conclusions

• TWI's vertical MIAB welding machine has been modified and commissioned for welding thin section non-ferrous and stainless steel tubes
• Successful test welds were produced with titanium and stainless steel tubes
• Design information for equipment manufacturers has been developed

Recommendations for future work

• Further machine modifications appear necessary to increase the displacement rate for welding aluminium tubes and to improve the alignment of the tubes
• Studies should be conducted to generate data on welding parameters and weld quality
• The process reliability and repeatability of MIAB welding non-ferrous tubes should be assessed

Acknowledgements

Most of this work was funded by the Industrial Members of TWI, as part of the 1998-2000 Core Research Programme. The Boeing St Louis Phantom Works funded some of the experimental assessment and released the results. The author wishes to thank Kevin Colligan who advised on the re-design and commissioning of the machine. His thanks go also to the colleagues at TWI, including Edward Watts, Chris Wing, Mike Moore, David Staines, Roger Wise and many others.

References