Risk assessment of tunnel flood based on the weighting of index reliability measurement and set pair cloud

Tunnel flood assessment is a nonlinear and complex system with uncertainty. Scientific and reasonable flood grading evaluation and corresponding prevention and treatment measures have become the primary problem to be solved in the construction and operation of karst tunnels. Firstly, aimed at the uncertainty, fuzziness and randomness of evaluation index parameters, the correlation between indexes by the multi-evidence correlation coefficient of Jousselme distance are analyzed in this study. The dynamic weighting theory of index reliability measurement based on Jousselme distance is also put forward. With the weighting method, the accurate weight of each index can be obtained according to the reliability classification of indexes. Thus, the dynamic weighting of the whole system by the measured values of different cases and indexes is realized, and the risk of deviation of evaluation results caused by the error or error of measured values of indexes in practice is reduced; thus, the robustness of the evaluation model will be enhanced.Secondly, based on the theory and idea of uncertain artificial intelligence, and with the combination of the eigenvalues in cloud theory and the set pair theory, the cloud theory to optimize the set pair connection degree is used to obtain risk assessment indexes of karst tunnel flood. The index cloud connection degree and index reliability measurement are weighted by the dynamic weighting method to obtain the degree of system comprehensive cloud connection, and the risk value is obtained by the expected weighted average of the grading evaluation interval. At the same time, the corresponding grade cloud map is generated to determine the flood grade of karst tunnel. Accordingly, the flood state of karst tunnel is determined, and the visualization of flood grade determination is realized.Thirdly, the karst tunnel flood occurs in a complex system composed of both the tunnel and the external environment; therefore, its risk assessment is affected by many factors, and the evaluation index system should be established in a comprehensive and concise way. From the perspective that atmospheric precipitation is the main source of flood in karst tunnel, five correlation indexes-annual precipitation, infiltration coefficient, catchment area, permeability coefficient and unit water inflow-are constructed as evaluation indexes of set pair cloud of karst tunnel flood in this study. On this basis, the grading standard of each evaluation index and its corresponding cloud map of flood grade are established. In order to verify the accuracy and effectiveness of the set pair cloud model and its evaluation index system constructed based on the weighting of the index reliability measurement, the data of six typical karst tunnel samples for the model test is collected. Consistency between evaluation results and those of other methods proves the reliability and effectiveness of the model proposed in this study. Finally, this model has been applied to the flood accident of Pishuangao karst tunnel of Beijing-Zhuhai expressway, the results of which are also consistent with the actual situation of the project. At the same time, the treatment measures corresponding to Grade I have achieved effective flood control. The result shows that the set pair cloud model constructed in this study based on the weighting of the index reliability measurement takes into account the correlation between the evaluation indexes, the uncertainty of the evaluation system and the ambiguity of the evaluation index grade. The model can improve the accuracy of flood risk assessment in karst tunnels, and the evaluation process of this model is highly maneuverable. In fact, this model can provide a new method for rapid and effective analysis of flood in karst tunnels, and can also provide reference and guidance for the prediction and prevention of flood in karst tunnels in China.