Stochastic modeling of corrosion growth

Abstract

Corrosion is a major threat to the structural integrity and safe operation of infrastructures throughout the world. Corrosion growth modeling is important in structure maintenance planning. Existing models focus on the maximum corrosion pit depth growth which may lead to inaccurate life predictions since the corrosion volume growth may result in failures before the depth reaches its failure threshold. We develop a stochastic model that characterizes both corrosion volume and depth growth. The distribution of volume growth increments with time and the reliability estimate based on the corrosion depth, the corrosion volume growth and their combined effect are obtained. The influence of stresses, which include the relative humidity, pH level and temperature is incorporated into the model based on the physics of the corrosion reaction mechanisms. Field corrosion data are used for the model's validation. The proposed model results in more accurate predictions of the remaining lives compared with the existing models.