Prediction of Corroded Pipeline Performance Based on Dynamic Reliability Models

This paper focusses on developing an initial model for dynamic reliability analysis to predict the aging pipeline performance due to corrosion. The corrosion failure mechanism and the associated data requirements are identified by combining outputs from the literature and project experiences. Bayesian networks (BN) are developed to manage and overcome data uncertainties and the dynamic consideration is utilized to introduce time function into the model to accommodate the time-dependent variables. Several parameters are considered in the model development, such as pipeline content, size and material grade, environmental conditions, operational conditions, internal and external corrosion rates mitigation methods and in-line inspection data on corrosion rates. The application of the proposed model to an industrial case study is presented in this paper, along with the basic event prioritization analysis using the sensitivity approach. The proposed dynamic Bayesian model provides an efficient option for reliability assessment, to predict the future condition of the corroded pipeline based on the current and historical data, leading to rational risk assessment.