ABSTRACT
ABSTRACT Chloride-induced corrosion is a localized phenomenon, where corrosion initiates first at the weakest spot within an exposed steel surface area. This gives rise to a size scale effect, that is, the critical chloride content for corrosion, Ccrit, is size dependent. This paper illustrates the impact of this size effect on probabilistic service life modeling. Examples are calculated using input data from laboratory studies and from engineering structures. The calculations highlight that the size effect greatly influences the outcome of probabilistic service life modeling. To consider the size effect in future model developments, materials science needs to be linked to structural engineering. Physical interpretations of probabilistic modeling incorporating the size effects – such as in terms of predicted “number of corrosion sites per unit concrete surface area” or “average distance between corrosion sites on rebars” – rationalize defining limit states, which is needed for sound durability design.
