Metal powder additions in submerged-arc welding

TWI Bulletin, March/April 1988

Part 1 Equipment and process characteristics

by Kevin Middleton

Kevin Middleton, BSc (Hons), is a Research Engineer in the Flux Processes and Surfacing Section of the Arc Welding Department.

This review of techniques and developments in metal powder additions for submerged-arc welding, shows that considerable scope exists for further application of the technique to take advantage of increased productivity and enhanced mechanical properties.

In recent years, metal powder addition in submerged-arc welding (SAW) has been exploited to increase productivity whilst maintaining high degrees of weldment mechanical properties. Higher deposition rates can be achieved by increasing the welding current, but this can only be exploited if the welding speed is also increased to control arc energy and therefore mechanical properties, particularly toughness. For single wire SAW, the maximum speed is limited by the occurrence of undercut, thereby imposing a limit on the maximum usable current level and, therefore, the metal deposition rate achievable. This limitation can be overcome by the use of multiwire techniques which allow much faster welding speeds to be used, ^[1] so that the total welding current, and therefore deposition rate, can be increased.

However, for SAW of thick section steel plate, requiring multiple weld runs to complete the joint, the setting up time between runs significantly influences the overall joint completion rate. For these applications, the higher metal deposition rates achievable with multiwire techniques should result in a reduction in the welding arc time compared with single wire SAW. However, if a similar arc energy is used, a similar number of weld runs will be required to complete the joint so that the setting up time will be unaltered (or even slightly longer because of the increased complexity of setting up a multiwire system). Therefore the welding duty cycle will be decreased. A reduction in the number of runs to complete a given joint would require an increase in arc energy regardless of the number of welding wires used. This may not be possible if critical levels of mechanical properties are to be achieved.

In conventional SAW, only a limited proportion, some 10-20%, of the available arc energy is used in melting filler wire. The remainder is dissipated in melting the flux, superheating the molten weld pool and causing fusion of the base material. This results in the characteristic high degree of penetration and dilution of the base material, which is often in excess of that required to ensure adequate weld fusion, particularly for SA surfacing and the fill runs of a multi-run joint. Whilst melting the flux is an essential process requirement, arc energy expended in superheating the molten pool and causing excessive fusion of the base material is redundant so that considerable excess energy is available for the melting of additional filler metal.

Metal powder additions were initially applied to SA surfacing, ^[2] where the ability to produce low dilution deposits at high metal deposition rates contributed to a reduction in surfacing costs. The first use of a metal powder addition in SA joining applications was to increase the thicknesses of steel which could be welded in a single run using conventional techniques. ^[3] However, the reduction in penetration caused by the powder demanded the use of higher currents and/or wider preparations to ensure good root fusion, considerably reducing the economic effectiveness of using a powder addition.

With the development of consumables tolerant of high dilution welding conditions, the joining of thicker section materials could be achieved using joint preparations of low volume, requiring only one or two deep-penetration SAW runs to complete the weld. This development heralded a decline in the use of metal powder additions. In the late seventies exploitation of North Sea Oil reserves created a demand for a high quality and highly productive technique for multi-run SAW of thick section steel plate. High dilution, high penetration, one and two run welding procedures could not be used because base material property requirements, in the thicknesses used, restricted the arc energy that could be used and demanded the use of multiple runs. Increases in joint completion rate could only be achieved by a technique which completed the joint in as few runs as possible without an increase in arc energy. This requirement rekindled interest in the use of metal powder additions in SAW and sparked the major developments in this area.

Methods of powder addition

Over the years many different methods of adding metal powder have been developed. They may be classified into one of two categories where:

1. The powder is laid ahead of the flux burden and arc;
2. The powder is added directly to the molten pool.

The first category includes techniques where the joint preparation is partially or completely prefilled before welding commences (not practical for circumferential welds or multi-run welds) and the 'forward-feed' technique where the powder is metered into the joint preparation ahead of the flux during welding.

The second category includes the use of metal powder containing tubular wires, fluxes containing metal powder and the magnetic attachment of the metal powder to the current carrying electrode wire.

All the above techniques have been used in the past but those that have found most favour are those in which the powder addition rate can be independently controlled. Two such techniques have found industrial favour, the most popular is the 'forward -feed' technique ( Fig.1) and the other is the 'magnetic attachment' method ( Fig.2).

Equipment

The forward-feed and magnetic attachment systems require additional equipment in the form of a metering and delivery system, which is inexpensive and simple to operate. The forward-feed system needs only one delivery tube to deposit the required volume of metal powder into the joint preparation ahead of the flux burden. The magnetic attachment system usually requires two or more tubes to ensure an even distribution of the powder around the electrode wire.

Two metering systems are currently available for industrial use ^[4,5] and are suitable for both methods of powder addition. The most popular metering system is by means of a bucket wheel driven by a permanent magnet, direct current motor ( Fig.3a) the speed of which is set by a separate control box ( Fig.3b). The electrical power is derived from the mains current supply which, in large welding shops, may be subject to fluctuations in load. The mains fluctuations may cause speed variations in the powder feeding motor, although the extent of these variations has not been measured.

The alternative metering system is independent of mains electricity supply. Two dies are rotated into line by a simple lever between two feed tubes and the hopper containing the metal powder ( Fig.4). For the forward feed process the two tubes from the hopper could be connected together to form a single delivery tube. The dies are made of non-magnetic abrasion resistant stainless steel, and the orifices are machined to give the required constant flow rate of powder. The powder falls under its own weight. Different flow rates require different die orifice sizes.

The development of suitable metal powders

The forward-feed process

The high level of weld metal mechanical properties, particularly toughness, required for North Sea offshore structures could not consistently be achieved with the metal powders initially available. Subsequent development work ^[4] identified an increase in weld metal vanadium, derived from the powder, as the major cause of poor weld metal fracture toughness. Metal powder compositions were modified to eliminate this effect and to provide a matching composition with the filler wire, thereby eliminating any compositional variations caused by changes in powder addition rates during welding. The metal powder chemical specification, typically used for offshore fabrications, is shown in Table 1. 

Table 1 Powder composition limits, element wt % ^[6,7]

The magnetic attachment technique

Metal powders to assure weld toughness also had to be developed for this technique although a different pattern of development evolved. Metal powder added via the wire comes into intimate contact with the arc and molten slag, resulting in reduced element transfer efficiency compared with the forward feed technique. In particular, weld metal carbon and manganese content are reduced. ^[5] Powders with increased carbon and manganese contents were tried, but no benefits were accrued. Subsequent additions of up to 2.5%Ni improved Charpy impact values but fracture toughness results were erratic. It was not until molybdenum was used in combination with the nickel that consistently high fracture toughness values could be guaranteed. The metal powder currently recommended for use with the magnetic attachment method of powder addition for joining steels of BS4360 grade 50 contains 2-2.5%Ni and 0.35-0.5%Mo ( Table 1). Because of essential differences in the composition of wire and powder it was then necessary to prove that if the powder supply was interrupted, the resulting weld would still meet the required mechanical properties.

Powder particle size

There has been no in-depth investigation into the effects of powder particle size and current commercial specifications are largely based on experience. ^[6,7] Metal powders for welding are produced by atomisation to produce a specific particle size distribution ( Table 2). This distribution, determined empirically, ensures that the particles are large enough to flow quickly under their own weight, and small enough to give even fusion. Large particles of small surface area to volume ratios ( e.g as in the extreme case of cut wire) may cause localised quenching of the molten pool and may be only partially melted, leading to uneven penetration and lack of fusion.

Table 2 Powder particle distribution specification (both powders) ^[6,7]

Process characteristics

Powder additions in SAW can be used to achieve any of the three following objectives:

1. To decrease the number of runs required to complete a joint whilst using the same arc energy.
2. To decrease the arc energy used to complete a joint in a given number of runs.
3. A combination of 1 and 2.

The first objective is the most commonly sought and it is the approach adopted by offshore fabricators, who carried out much of the process development work in recent years. This is illustrated in the following example. Two joints in 5omm thickness steel plate were welded using a single wire technique and an arc energy of 3 kJ/mm ( Fig.5). ^[8] Welding conditions and joint preparation were exactly the same except that in W2 a powder addition at a rate of 9.0 kg/hr was made from the third run onwards using the forward feed technique. The number of runs required was reduced from 31 to 16 when a metal powder was added. There was also a large reduction in the amount of flux used; without a powder addition 7.2kg of slag was produced for each metre of completed weld, whilst with the powder addition, only 3.7 kg/m was generated. Therefore, in addition to an increase in joint completion rate, significant savings are to be made in consumable costs. Although the metal powder is a more expensive consumable than filler wire, the additional cost incurred for filler metal is outweighed by the reduced flux consumption.

The second objective, i.e. the use of a metal powder addition to decrease the arc energy required to complete a joint in a given number of runs, is rarely required. Generally speaking, a fabricator will maximise productivity by using the highest arc energy possible provided the completed weld possesses the required mechanical properties. However, the ability to reduce arc energy is relevant to single run weldments. Conventional welding procedures for completing a single run weld in 12.5mm thickness plate require an arc energy in excess of 4.5 kJ/mm. Recent work at The Welding Institute ^[9] has shown that the arc energy, required when using a metal powder addition and a copper backing bar can be reduced to 3.0 kJ/mm ( Fig.6). A powder addition therefore decreases the arc energy required to complete a given plate thickness in a single run and, conversely, increases the thickness that can be welded in a single run at a specific arc energy. Besides the productivity benefits already illustrated, metal powder additions have other potential advantages. Some of these process benefits are specific to the method in which the metal powder is added to the weld. It is therefore necessary to describe these benefits separately in relation to the principal powder addition techniques, and this is the subject of Part 2 of this article.

Part 2

References