Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Cracks in submerged-arc welds

A number of articles have been written about the occurrence of cracking in welds. This case study provides an illustration of the causes of cracking.

It involves a panel line in a shipyard where 10mm C-Mn steel plates were submerged-arc welded from one side onto a glass fibre tape backing. The tape lay within a flat groove in a copper backing bar which was maintained in close contact with the plates by air pressure. A square edge preparation was used with a 4mm gap between plates. The weld was 4m long and sound apart from the final 400mm, which contained a centreline crack. Welding conditions were as follows:

Flux: Agglomerated alumina, semi-basic 
Wire: 3mm diameter Sl, 0.5%Mn 
Current: 520A 
Voltage: 36V 
Electrode polarity: -ve 
Electrode extension: 300mm 
Travel speed: 400mm/min 
Width of weave: ± 6mm

One sound and one cracked transverse section of the weld were cut out and prepared for examination. Measurements of both cross section showed the size and shape of the weld to be the same. Therefore, the dilution and chemical composition of the weld should have been similar. Spectrographic analysis of the cracked specimen showed that the composition of the weld was: 0.11%C, 0.020%S, 0.022%P, 0.25%Si, 0.56%Mn, 0.002%Nb.

The factors controlling solidification cracking are: weld metal composition, weld solidification pattern (which depends on the shape of the weld), and the strain on the solidifying weld. A parameter has been developed to calculate cracking susceptibility using the weld metal composition:

UCS (units of cracking susceptibility) = 230C + 190S + 75P + 45Nb - 12.3Si - 5.4Mn - 1.

The UCS of the submerged-arc welds in question was 24. Research at TWI has shown that UCS levels of less than 10 indicate a high resistance to cracking and levels above 30, a low resistance.

In the case of butt welds, trouble should not be expected if the composition gives a UCS of less than 25, provided that the weld has an acceptable shape. The shape of the weld influences the solidification pattern and to minimised cracking the columnar grains of metal should appear in an upward pattern rather than inwards.

In one-sided welds on glass fibre backing the solidification pattern of the weld tends to have an unfavourable grain structure. This tendency to grow inwards rather than upwards can give a more pronounced centreline segregate of impurity elements and also concentrate the contraction strain in the came region. To avoid cracking here consumables should be selected with low carbon and sulphur, and high manganese and silicon contents.

For this case study it was recommended that the wire be replaced by one having higher manganese content so that the weld metal would contain approximately 1.0%Mn. It is possible that the plate, being relatively thin, may have been getting hotter towards the end of the weld, and thus tending to reduce the cooling rate of the weld. A lower cooling rate increases the tendency for cracking and this could explain why the crack was restricted to the last 400mm of weld. Alternatively, the cause may have been the increasing restraint on the weld due to contraction on the weld metal further back in the joint. The exact cause of the problem was not identified, but a change in wire composition provided the solution.

For more information please email contactus@twi.co.uk.  

}