A novel integrated multi-source heterogeneous uncertainty fusion framework for reducing the complexity of uncertainty characterization in domino effect risk assessment

The risk assessment result of domino accidents in chemical industry can be influenced by uncertainty, and the multi-source heterogeneous data fusion makes the uncertainty analysis, especially the characterization, complex and difficult. This paper proposed an integrated framework of uncertainty analysis considering multi-source heterogeneous data, and provided different characterization methods for common multi-source heterogeneous data types in order to reduce the complexity of uncertainty analysis. In order to deal with the influence of expert cognitive bias on fuzzy information, the interval type-2 fuzzy set suitable for domino accidents in chemical industry was established based on questionnaire data. The proposed framework was applied to a real tank farm to obtain the intervals and the corresponding probability envelopes of time to failure and escalation probability, and the most likely accident scenario. Compared with the competitive method, the superiority of the proposed method was verified. The sensitivity analysis of inherent safety design parameter was carried out to obtain the priority of the material, size and storage of LNG storage tank considering the fire-induced domino risk. The integrated framework can reduce the complexity of uncertainty analysis of chemical domino accident risk assessment and provide reference information for decision makers to control domino effect.