Probabilistic Modeling of Corroded Pipeline Structures

Corrosion is one of the most important damage mechanisms for in-service pipelines, and a significant portion of the maintenance budget is directed toward corrosion-related problems. A major challenge associated with the assessment of the impact of corrosion on the integrity of pipeline structures involves quantification of the amount and severity of corrosion damage present in the structure. Corrosion defects are typically characterized by spatially random distributions and variabilities in size, shape, and morphological characteristics throughout the exposed part of the structure. For pipeline corrosion, such spatial randomness and variability are best modeled using a nonhomogeneous random field approach. A review of some existing random field modeling strategies and their potential for modeling in-service pipeline corrosion data (including their limitations) is presented. A practical random field modeling strategy is developed, which is suitable for in-service pipeline corrosion modeling and circumvents the limitations of existing models. The application of the strategy is demonstrated via example problems, wherein the model is applied to actual pipeline corrosion data. A preliminary application of the corrosion model is also undertaken to assess the residual strength of a pipeline subjected to corrosion damage.