A dynamic assessment method for risk evolution in chemical processes based on MFM-HAZOP-FDBN

At present, chemical accidents have occurred frequently with disastrous consequences. In order to effectively prevent chemical accidents and their secondary disasters, it is urgent to develop detailed risk assessment for chemical processes. However, traditional methods tend to ignore the dynamic correlation coupling of chemical processes and devices; this results in the assessment results having poor accuracy. Therefore, a dynamic assessment method for risk evolution in chemical processes based on MFM-HAZOP-FDBN is proposed in this study. First, multilevel flow model (MFM) is established according to the transfer paths of material flow and energy flow in a chemical process, and further combined with hazard and operability analysis (HAZOP) to form the functional HAZOP technology. Second, risk identification results from the MFM-HAZOP are applied as the input of a fuzzy dynamic Bayesian network (FDBN); the device states and process errors are used as dynamic and static nodes of FDBN to present the risk propagation and evolution mechanism in the chemical process, respectively. Finally, the dynamic risk evolutionary trend throughout the chemical process is assessed quantitatively by determining the occurrence likelihood and consequence severity based on the risk matrix. The BP America (Texas City) refinery explosion is taken as a case study. Results show that the proposed method can provide a detailed characterization and dynamic assessment of risk evolution in the chemical process and the assessment results are more reasonable and effective.