Natural Gas Purification Plants Based on Interpretive Structural Models and Bayesian Networks

Natural gas (NG) purification plants play a crucial role in the supply system. Some high-level risk equipment and processes are involved, and the determination of the corresponding risk factors is essential for the purification plant. A comprehensive risk analysis method was developed by integrating the decision-making trial and evaluation laboratory (DEMATEL), interpretive structure modeling (ISM), and a Bayesian Network (BN). Based on the hierarchy determined by the DEMATEL and ISM model, a BN of a natural gas purification plant was established to determine the key factors. The results showed that a lack of visual signboards, inadequate safety education training and emergency drills, insufficient safety management, and unauthorized work were the fundamental causative factors in the DEMATEL and ISM analyses. According to the BN analysis, the probability of the natural gas purification plant being in an unsafe state was approximately 66%, which was higher than that of the underground natural gas pipe gallery. Additionally, it was found that illegal operations had the highest probability, and faulty pipes had the highest sensitivity, indicating that the two factors were key causative factors. Based on the above results, corresponding countermeasures for the safety management of natural gas purification plants have been proposed.