| description abstract | Oil and gas pipeline transportation, the most common and economical mode of transport, poses significant risks of environmental pollution and human casualties due to corrosion failure. External corrosion, which accumulates over time, necessitates an accurate risk assessment method to evaluate these risks. While Bayesian networks (BNs) are a proven, flexible, and powerful analytical tool, traditional BN applications have been criticized for relying on fixed probabilities when assessing uncertainty. To address this limitation, this study proposes the use of a dynamic Bayesian network (DBN) to analyze external corrosion using dynamic probabilities. First, the influencing factors are categorized into natural and human factors, which are identified through a fault tree analysis. To account for the uncertainty between events, the fault tree is mapped to a DBN model. Secondly, due to limitations in data availability, the parameters of the Bayesian network are determined by combining expert investigation with fuzzy set theory. The uncertain causal relationships between nodes are modeled using the leaky noisy–OR gate within the Bayesian network. Finally, through dynamic probability analysis and the transmission of state transition probabilities in the DBN, the study reveals that the probability of pipeline failure gradually increases over time. Based on these findings, effective recommendations and preventive measures are proposed. The results demonstrate that the proposed model effectively represents external corrosion risks through dynamic probabilities and provides reliable predictions of pipeline failure probabilities. | |
