Abstract
This paper discusses possible origins of preferential weld corrosion of 1% nickel welds in sweet systems. A 1-D current distribution model is used to show that thin film, low conductivity conditions may lead to preferential attack on the weld metal due to lack of galvanic protection of the weld by the parent pipe. In high conductivity environments under the same conditions preferential attack of the weld may be prevented by galvanic coupling to the parent pipe. It is shown that the selection of metallurgy to give a cathodic weld metal does not necessarily mean that the weld metal will corrode slower than the parent metal. However, the magnitude of the corrosion rate differences between weld and parent that are often seen in practice cannot be explained simply by galvanic effects. It is shown that severe selective attack of the weld metal occurs when the other components of the weld are well inhibited, but the weld poorly so. This can occur with underdosing of inhibitor, and is significantly affected by the extent of precorrosion of the weldment before inhibitor addition. Inhibitors can perform differently on parent steel and weld metal, and this needs to be taken account of in inhibitor selection procedures.
