Predictive analyses of the integrity of corroded pipelines based on concepts of structural reliability and Bayesian inference

• Structural reliability are used to evaluate the probability of failure of pipelines.
• Model considers material properties, geometric and operational parameters.
• Bayesian reliability concepts are used to estimate the evolution of corrosion defects.
• The operator can schedule defect repairs using this methodology.
• The operator can establish inspection intervals using this methodology.

In this paper, structural reliability concepts are used in conjunction with limit state functions proposed in the Recommended Practice DNV-RP-F101 (2010) to evaluate the probability of failure of corroded pipelines during their lifetimes. The model takes into account the natural spread of material properties, geometric and operational parameters, and the uncertainties associated with the sizing of eventual corrosion defects. Bayesian reliability concepts are used to estimate the evolution of a pre-defined distribution of defects obtained, for instance, from an inspection campaign. By comparing the predicted probability of failure with the reliability acceptance criteria, the operator can schedule defect repairs and establish inspection intervals with more confidence. This proposed methodology can provide the basis to develop a risk based maintenance strategy of pipeline systems.