Pitting Corrosion of Mild Steel under Marine Anaerobic Conditions—Part 2: Statistical Representation of Maximum Pit Depth

Longer-term pitting corrosion of mild steel under field exposure conditions has been shown to occur in the entirely anaerobic environment created by the build-up of corrosion products. There is evidence of lateral pit growth and pit coalescence, both for sub-extremal and extremal depth pits. Statistically this implies the generation of a new population of pits, not directly dependent on earlier pit depth distributions. Moreover, new pits have been shown to initiate in the subsequent anaerobic environment without obvious reference to the pattern of earlier pitting. These observations have implications for the application of an extreme-value statistical theory for modeling maximum pit depth in the anaerobic range. Data from multiple replicate coupons exposed to marine immersion corrosion are reviewed and reinterpreted. Extreme-value distributions are examined for consistency with the data and a new statistical model for extreme depth pitting in the anaerobic range is proposed—the Frechet distribution. It predicts a greater probability of occurrence for a given depth of pitting than does the conventional Gumbel distribution for the same data set. This has important practical implications where long-term pitting is governed by anaerobic conditions.