Flood Intelligent Calculation Model And Disaster Risk Assessment Of The Huaihe River Mainstream Flood Control Area

The Huaihe River was affected by the Yellow River for a long time,and complicated by natural,economic and social conditions.The floods were characterized by large floods,high flood peaks and long durations.The flood disasters in the basin were serious.The intelligent calculation of the flood of the Huaihe River and the rapid analysis of disaster risk are of great significance in the decision-making of emergency disasters.The specific research contents and results are as follows:(1)Comprehensive statistical methods are used to analyze the annual peak characteristics and interannual variation characteristics of the main stream of the Huaihe River in the Huaihe River Basin Wangjiaba,Lutaizi,Wujiadu and Xiaoliuxiang four hydrological stations.Focus on the series of trends,mutations,and periodic variation characteristics,which laid the foundation for the study of the Huaihe River mainstream flood calculation model.(2)According to the multi-year observation data of the typical hydrological station of the Huaihe River,the convolutional neural network and the attention mechanism are used to optimize the gated loop unit neural network(GRU),and the spatial and temporal characteristics of the flood are considered to study the intelligent composite algorithm to construct the Huaihe Huaibin-The intelligent calculation model of Xiaoliuxiang flooding uses the firefly algorithm to optimize the model parameters and calculate the water level and flow process of the main hydrological stations such as Wangjiaba.Taking the Wangjiaba site as an example,the 10-fold cross-validation result after optimizing the GRU model and the flood period 3h and6 h foreseeing period The RMSE of the water level process prediction results decreased by 27.5% and 14.63%,respectively.The RMSE of the flow process prediction results decreased by 11.19% and 6.37%,respectively.The RMSE of the water level process prediction during the 3h and 6h flood seasons in the flood season decreased by 33.02% and 37.50%,respectively.The RMSE was reduced by 35.75%and 39.48%,respectively.(3)Construct a rapid prediction model of flood evolution based on graph convolutional neural network and attention mechanism,and consider the space-time process of flood evolution to achieve rapid and accurate prediction of flood evolution.Using the CPU/GPU synergy method,a one and two-dimensional coupled hydrodynamic rapid calculation model for the Cinanfeizuo flood control area and the Huaihe River was established,and a flood submerged sample database was generated.When the flood is predicted,the root mean square error of the prediction result is as low as 0.08 m,and the accuracy is above 97.53%.The predicted submerged water depth distribution and the submerged area are close to the hydrodynamic model to simulate the submergence result.(4)An improved method of flood risk assessment based on combination weighting and fuzzy clustering is proposed.The subjective and objective weighting of evaluation index adopts intuitionistic fuzzy analytic hierarchy process considering hesitation and abstention,and VC based on objective data information and variability.The CRITIC method optimizes the game theory combination weighting method to optimize the combination,determines the optimal combination weight,and uses the Gaussian mixture model fuzzy clustering algorithm to classify the regional flood disaster risk level,and applies the method to the Cinanfeizuo In the flood disaster risk assessment of flood protection areas,the evaluation results are reasonable and reliable,and the research results can provide technical support for flood disaster risk assessment and disaster reduction decision-making.