TWI Industrial member Report Summary 950/2010
By C J Miller and J M Nicholas
Background
Large liquid natural gas storage (LNG) tanks operating at ~164°C have for many years been utilising 9% Ni steel for the inner shell and tank bottom and Ni-based filler metals continue to be the preferred choice for welding.
Whilst the resistance of Ni-based filler metals to brittle fracture has meant a long and satisfactory safety record for LNG tanks, they have a number of disadvantages. One of these is that the Ni-based fillers that have suitable toughness are unable to match the strength of 9% Ni steel. This is a fact that is taken account of in design codes for LNG tanks by basing the maximum allowable design stresses on the strength of an under-matching Ni-based filler metal.
In the UK, when the capacity of LNG storage tanks increased and favoured the use of 9% Ni steel rather than aluminium much work was undertaken to develop suitable MMA electrodes and welding procedures to maximise weld metal strength.
One of the discoveries of this early work was that MMA welding in the horizontal-vertical (HV) position (shell circumferential seams) tended to give all-weld tensile strength that was lower than all-weld tensile strength in joints welded in the vertical-up (V) position despite the latter being made using generally higher heat input welding.
The reason for this strength difference was postulated as being related to differences in the preferred orientation of the weld metal grain structure. Welds made in the HV position were found to have a grain orientation that favoured the preferred slip direction whereas V welds did not. The preferred grain orientation hypothesis has remained the perceived wisdom for explaining weld metal strength variations in 9% Ni steel joints welded with Ni-based fillers and this is still referenced in one of the TWI frequently asked questions (FAQs).
However, the authors of this report undertook some investigatory work some years ago that produced data to indicate that there was a simpler explanation for how the strength of Ni-based weld metal in 9% Ni steel butt joints is influenced.
These findings were only published privately but an alternative hypothesis was proposed to explain the observations about variable strength according to welding position. The new hypothesis was that it was principally dilution of the filler metal from the base metal that influences the strength of the weld metal. Moreover, it was shown that dilution was related to welding current and this is influenced by welding position. Welding in the HV position is generally done at higher current than V welding so that the dilution level in HV welds tends to be higher and the strength lower.
Objective
The objectives of this work were to investigate the relationship between weld metal strength and weld metal dilution and the influence of welding current on weld metal dilution.