A phenomenological model for general corrosion of mild, low-alloy steel under near-surface, at-sea immersion conditions was proposed earlier. The model was set in a probabilistic framework. The present paper deals with estimation of the various uncertainties associated with corrosion assessment from field data and the estimation of model bias. Uncertainties arise from inherent variability in corrosion processes, from variability in material properties, from corrosion response to environmental conditions, and in the environmental conditions themselves. Model uncertainty is also involved, as is model bias. In general, there is an acute shortage of appropriate data in the literature since individual coupon results have only seldom been reported. Results from recent field tests carried out on the Eastern Australian seaboard provide possibly the most complete variability information to date and are applied herein. An example is given to illustrate the application of the derived uncertainty information to a corrosion-time relationship and the associated uncertainty probability bounds.

You do not currently have access to this content.