Bayesian Approach to Coating Life Modelling for Probabilistic Corrosion Growth Rate Analysis

Abstract Selecting the correct pipeline coating, ensuring its proper application, and maintaining it are essential for the long-term protection of pipelines against external corrosion. While industry guidance documents recommend quality assurance and quality control procedures for the maintenance and installation of coatings, the practical service life of pipeline coatings remains a challenge to estimate. As newer generation coating technology matures, anecdotal evidence suggests that coatings can remain reliable for decades under optimal environmental conditions. The authors developed a model based on general industry knowledge of coatings across time and findings from integrity digs. A Bayesian model was used to provide a robust framework to integrate prior knowledge and observed data. In situations where empirical evidence is sparse or unavailable, subject matter expertise acts as an informative prior. This paper introduces a Bayesian method that quantifies the practical service life of coatings and predicts the probability of the coating being intact for a given pipeline segment over time. The model incorporates subject matter expertise with observations from pipeline excavations that reveal the coating to be intact or not. The model is updated to account for the results from integrity digs. Using generalized industry benchmarks with specific inspection findings produces a more accurate estimation of coating integrity. The results are used to estimate the effective corrosion growth rates on a pipeline in a probabilistic pipeline assessment.