Contribution of Bayesian networks as a robust tool in risk assessment under sustainability considerations, a case...

Refineries are among the industrial centers that supply the energy and raw materials to downstream industries. To achieve sustainable development goals, creating appropriate balance between economical and environmental goals has always been the focus of managers and policy makers in the societies. Bayesian Network model has become a robust tool in the field of risk assessment and uncertainty management in refineries. The focus of this research is to prioritizing different units from the point of view of social and ecological aspects for facilitating the decision making process in the context of waste material treatment in Bandarabbas refinery in line with the sustainable development goals.

The methodology of this research is based on risk assessment with the aid of Bayesian Networks. To this end, first material flow analysis of the processes procured risk identification, subsequently influence diagram and Bayesian Network structure were designed. After completing conditional probability tables, finally risk factors were prioritized. What is more, sensitivity analysis of the model performed by applying three approaches namely predictive, diagnostic, and considering only one risk.

According to the risk assessment results, Amine treatment and Fuel units were classified as the most significant risk factors, whereas Pipelines and Plant air & instrument air system were identified as the most environmental friendly units. In addition, sensitivity analysis of the model provided appropriate framework to shed some light on the circumstances of determining dominant risk factors whether only one or concurrently all of the endpoints are evaluated.