Research On Quantitative Analysis Methods For Risk And Uncertainty Of Oil And Gas Transmission Pipelines

As the "lifeline" of energy safety for China' s national economy and social development,oil and gas pipelines and their safe and stable operation possess self-evident significance Scientific and reasonable quantitative risk analysis can effectively identify risk factors that will harm the pipelines,ascertain credible scenarios of secondary consequences,predict and quantify the level of risk,and provide decision support for safety management of oil and gas pipelines.Focusing on the three difficulties in traditional quantitative risk analysis for oil and gas pipelines,namely the lack of dynamics,the neglecting of uncertainty and the joint treatment of both,a series of researches,including the realization of dynamic risk analysis for specific oil and gas pipelines,the uncertainty analysis in quantitative risk analysis for oil and gas pipelines,and the fusion of dynamic risk analysis and uncertainty analysis for oil and gas pipelines,have been carried out in this dissertation.The main works are as follows(1)Aiming at the problem of incomplete identification of causes and consequences of oil and gas pipeline accidents in the past,a causal analysis is conducted by using qualitative Bow-tie analysis,based on the analysis of accident cases and data.Meanwhile,considering the distinctiveness of a specific object,namely the possible occurrence of atypical or redundant causes and consequences,the dynamic updating mechanism is introduced to build adaptive qualitative Bow-tie analysis,which will provide adaptive dynamic update of accident causal model for quantitative risk analysis of specific oil and gas pipelines(2)On the need of dynamic risk analysis for water-submerged natural gas pipelines,the adaptive Bow-tie model is updated.Then,to overcome the restriction of static attribute of Bow-tie model and to cope with the absence of occurrence probabilities of basic events,a dynamic risk analysis method combining Bayesian network modeling and fuzzy probability elicitation is proposed.The case study manifests that the method can accurately assess the risk of the target pipeline and trace the key basic events(3)In view of the problem that uncertainty handling is neglected in traditional quantitative risk analysis for oil and gas pipelines,the source and nature of uncertainty and the process of uncertainty analysis are introduced.Then,the parameter uncertainty characterization and propagation methods based probability analysis,interval analysis and fuzzy set theory,and the model uncertainty handling method based on Frank model,are studied.The study on uncertainty analysis methods provides a methodological support for quantitative risk analysis of oil and gas pipelines that based on uncertainty analysis(4)In order to cope with hybrid uncertainty in quantitative risk analysis for oil and gas pipelines,the principle to select methods for parameter uncertainty characterization and the method to unify heterogeneous characterization are proposed.The way to jointly propagate parameter and model uncertainties is also derived.Besides,to reasonably characterize,quantify and moderately reduce model uncertainty,a method to quantitatively evaluate the dependency between events is proposed based on evidence theory.At the end,the uncertainty analysis-based quantitative risk analysis method for oil and gas pipelines is formed.The case study manifests that the method can accurately assess the risk and uncertainty of the target pipeline under the condition of uncertain parameters and dependencies(5)So as to retrofit the shortage that discrete variable nodes of traditional Bayesian network cannot characterize uncertainty,a dynamic risk analysis method for oil and gas pipelines is proposed based on continuous Bayesian network.The equation nodes of continuous Bayesian network provide a tool to characterize uncertainty,and the rule to map Frank model to continuous Bayesian network offers a way to deal with model uncertainty.The case study indicates that the method can effectively achieve the fusion of dynamic risk analysis and uncertainty analysis for oil and gas pipelines.