Abstract:  In order to evaluate the risk of gas leakage in heating furnace of rolling steel production, a Bayesian network model of gas leakage in heating furnace is established based on the bow-tie model.The prior probability of Bayesian network is determined by using the membership function of fuzzy theory, lambda section set theory and the left and right fuzzy sorting method, and the dynamic factors affecting gas leakage in heating furnace are analyzed. The dynamic characteristics of gas leakage risk in heating furnace with time change are studied by introducing time factor.The risk of gas leakage in a heating furnace is analyzed by an example, and the sensitivity analysis (ROV) value curve is obtained by using the proportional change method, so as to determine the important root nodes and obtain the dynamic change curve of gas leakage risk in the heating furnace. The results can provide reference for dynamic risk analysis and safety management of gas leakage in heating furnace. 

Key words: bow- tie model, Bayesian network, fuzzy theory, sensitivity analysis, dynamic risk analysis