P92 steels are commonly used in advanced steam generating power plants with welding being a main method of component fabrication. Such welds are typically post-weld heat treated at temperatures up to 780uC, which is designed to be as high as possible without the danger of the formation of any austenite during the heat treatment itself. Moreover, it is recognised that, in practice, there will be differences between the specified heat treatment temperature and that which is actually achieved. There are a variety of weld consumables available that have been designed for joining P92 steels; these consumables are heavily alloyed but have ranges over which the compositions of each element in the consumable are allowed to vary. While reaustenitisation of P92 base metals during post-weld heat treatment in the normal temperature range is not likely, it is known that weld metals have much lower A 1 temperatures than parent metals. As such, concerns have been raised that, although weld consumables may exist within their specifications, certain combinations of elemental variations may lead to a reduction in the A 1 temperature to below that at which the post-weld heat treatment takes place. This paper investigates the effect of variations in alloying element content on the A 1 temperature of P92 steels and appropriate weld consumables using a combination of modelling and experimental observations. An equation has been derived to predict the Ae 1 temperature as a function of chemical composition for weld consumables designed for use with P92 steels, which can be used to identify compositions that may be in danger of reaustenitisation during normal post-weld heat treatment.

To encourage realistic power plant weld heat-affected zone failures within reasonable test times in laboratory creep testing, a welding consumable for P92 steel with the best creep performance was sought, to avoid failures in the weld metal itself. Therefore, a short term creep test programme was undertaken to evaluate weld pads manufactured using three commercially available consumables for P92, the results of which are presented in this paper. The effects of weld consumable composition are discussed both in general and with specific reference to the behaviour observed. The best overall and the poorest performing weld consumables were identified by composition. Microstructural investigations revealed the presence of precipitate free zones, that more readily formed using the poorest performing consumable as the result of retained d -ferrite, since its deposited chemical composition led to the lowest value of d -ferrite formation temperature. A consumable not based on P92 but on the stronger steel alloy FB2 did not perform as well as expected, confirming previous suggestions that the compositional factors which make a parent material creep-strong do not apply on a one-to-one basis to weld metals.

No abstract