LAUSR

Abstract The quality of life-cycle assessments and civil and structural engineering reliability analyses depends much on the quality of the models on which they are based. Models for loads and material properties are well-established. For structural or material deterioration, including corrosion, good quality models are still being developed. Empirical models exist but often have large uncertainty and extrapolation is problematic. High quality models must be based closely on fundamental scientific principles and observations and be calibrated to real-world data. This approach has been exploited for some years to develop corrosion models for steels in marine environments, including the use of data from field experiments and selected laboratory observations. This work is reviewed and the effect of various influencing factors described, including that of microbiological corrosion. The approach has been extended to cast iron, aluminium and copper-nickels. Two practical application examples are described.