Comprehensive uncertainty evaluation of dam break consequences considering multi-source information fusion

A comprehensive assessment of the consequences of dam-break is a critical strategic necessity for guaranteeing socio-economic development and lives for individuals. The consequences of dam-break are affected comprehensively by a multitude of uncertainties, resulting in multi-source and inconsistent relationships between indicators. It is extremely tough to integrate information from different sources adequately under multiple uncertainties, which often limit the assessment reliability. In this work, a comprehensive uncertainty evaluation methodology for the consequences of dam-break was developed through multi-source information fusion. Firstly, cloud model was employed to deal with randomness and fuzziness in the quantification of the grading of indicators and constructed the basic probability assignment function of the evidence corresponding to each data source. Then, in order to address the issue that conflicting evidence cannot be effectively fused utilizing traditional evidence theory. The basic probability assignment function was fused by the improved evidence theory. Furthermore, due to the differences in the importance of each data source in the assessment process. The corresponding weights were determined employing trapezoidal fuzzy analytic hierarchy process and entropy weight method. Finally, the effectiveness of the method was verified by taking five reservoirs in the Haihe River Basin. It shows that multiple uncertainties from different sources of information are combined and handled and the severity grades of consequences of dam-break can be quantitatively analyzed with our assessment method. Meanwhile, multi-source information with conflicts and uncertainties can be approached to produce more reliable risk assessment results in the situation of highly conflicting evidence.