Quantitative assessment of leakage orifices within gas pipelines using a Bayesian network

Leakage is a significant threat to gas pipeline reliability, causing serious safety incidents such as fire and explosion, which is determined by the leakage rate. Leakage rate estimation are critical to predict leakage level and take effective control measures for a response strategy. Many experimental and numerical studies have been conducted focusing on leakage estimation, with theoretical and empirical equations being determined. However, the dependency of leakage rate on the leakage orifice diameter, which is unknown in most situations, makes it difficult to apply established models in the field. To solve the problem, this paper presents a probabilistic simulation of leakage orifice diameters of gas pipelines based on a Bayesian network model, with experimental verification of the results. According to the established Bayesian network model the leakage orifice diameter can be assessed precisely, which is verified using a range of gas pipeline leakage experiments. Four different diameters of leakage orifice were studied under five different operating pressure levels. The results indicated that the proposed Bayesian network can assess leakage orifice diameters, which can be further applied to the leakage rate estimation in the actual situations.