Probabilistic Corrosion Growth Models of Buried Steel Pipelines Using Inspection and Soil Survey Data
Item Type:Conference Paper
Citation:Emadoddin Farahani, Qindan Huang, Probabilistic Corrosion Growth Models of Buried Steel Pipelines Using Inspection and Soil Survey Data, 14th International Conference on Applications of Statistics and Probability in Civil Engineering (ICASP14), Dublin, Ireland, 2023.
submission_276.pdf (PDF) 988.3Kb
Buried steel pipelines are subjected to many possible threats during their service lives, and one of the dynamic threats to the integrity of such pipelines is external corrosion is present. The time-dependent performance evaluation of buried pipelines considering corrosion could help develop pipeline integrity corrosion management strategies, where corrosion evolution needs to be understood. The corrosion-induced damage evolution on pipelines is known to be influenced by many factors such as the physical and mechanical properties of the pipeline as well as the pipelineﾒs surrounding environment. The goal of this paper is to develop a probabilistic predictive corrosion growth model that could be used for time-dependent reliability evaluation of buried steel pipelines based on in-line inspection (ILI) and soil survey data. For both corrosion maximum depth and length, a power-law function of time model formulation is adopted, which considers nonconstant damage growth rate over time. A Poisson process is considered for the occurrence of defects; therefore, the initiation time of each individual defect is considered to follow a Gamma distribution. To incorporate the environmental impact, the model parameters will be modeled through a linear or quadratic function of the field measured physicochemical variables of the soil, including soil moisture, pH, resistivity, redox potential, etc., along the pipeline. The Bayesian updating framework is employed to evaluate the statistics of the unknown model parameters using the available field data obtained from ILI of a 112-km on-shore pipeline near the Gulf of Mexico through Markov Chain Monte Carlo (MCMC) technique. A bi-variant Normal distribution is employed to construct the likelihood function considering the correlation between defect depth and length growth models, which has normally been ignored in the previous studies. Utilizing the developed predictive models, the time-dependent reliability of the aforementioned pipeline is assessed as a case study through predicting the probability of failure in terms of three failure modes: small leak, large leak, and rupture. The proposed methodology for corrosion growth model development can be adopted when either matched or nonmatched defects data are available; it does not assume uniform corrosion initiation time for all the defects, which can then predict the number of newly generated defects since last inspection; and it considers the correlation between the defect depth and length. Most importantly, it reflects the impact of soil properties, which can be used to evaluate the performance of non-piggable buried pipelines.
Other Titles:14th International Conference on Applications of Statistics and Probability in Civil Engineering(ICASP14)
Type of material:Conference Paper
Series/Report no:14th International Conference on Applications of Statistics and Probability in Civil Engineering(ICASP14)
Availability:Full text available