Probabilistic physical modeling of randomly corroded surface and its use in reliability analysis of corroded pipelines under spatiotemporal vibration

Abstract

Focusing on reliability-informed and risk-informed pipeline integrity management, this paper proposes a degenerated model for naturally corroded surfaces of pipelines using the random field method and in-line inspection (ILI) data, performing statistical sensitivity analysis of physical parameters. The randomly mechanical properties of the corroded pipelines are subsequently examined via simulation of spatiotemporal vibration. The closed-form solution of the nonlinear vibration model is determined based on the equations of mathematical physics. With the probabilistic framework of Bayes inference and Markov Chain Monte Carlo (MCMC), the ILI data is introduced to update the failure probability of corroded pipelines. The numerical example results demonstrate that the synergic effects of random corrosion and random loads are contributed to degraded mechanical properties of pipelines, including uninspected or shallow corrosion-degraded areas. The proposed method is proven to be an effective and practical representation of naturally corroded surfaces, which can accurately evaluate the structural reliability of the corroded pipeline subjected to spatiotemporal excitation.