| In recent years,with the global climate change,intra-city rainstorms have been frequent.The backward means of urban waterlogging prevention and control do not match with the high-speed urbanization process,bringing serious waterlogging disasters to major cities in China.The current mainstream urban flooding early warning system is an integrated system integrating various information technologies with Urban Storm Water Models as the theoretical basis.However,the Urban Storm Water Models are greatly affected by insufficient basic data information,complicated modeling process that is not easy to implement,and poor flexibility in analyzing the temporal characteristics of actual urban flooding;while another existing method based on data-driven technology cannot adapt itself to different urban waterlogging-prone areas,and establishing different prediction models for different waterlogging-prone areas will increase a lot of resources and time costs.Therefore,for impervious areas within the city that are prone to waterlogging during short-duration rainstorms,this paper combines the respective advantages of data-driven technology and Urban Storm Water Models to design and implement a waterlogging early warning system based on an adaptive urban flooding model.The main work and results of this study are as follows:A hybrid neural network-based feature extraction and prediction method for waterlogging data is proposed for the problem that the Urban Storm Water Models have poor flexibility in analyzing the time-series features of actual urban waterlogging.The depth of water accumulation is collected in real time using millimeter wave radar sensors,and meteorological data and water depth are input to the hybrid neural network as input time series features.The prediction model utilizes the respective advantages of Gated Recurrent Units and 1D-Convolutional Neural Networks to achieve accurate prediction of water accumulation and rapid warning of waterlogging disasters.The experimental results show that the proposed method has higher speed and accuracy in real collected multi-dimensional data sets of waterlogging compared with other mainstream methods.An adaptive waterlogging prediction method incorporating urban rainfall models is proposed to address the problems of poor generalization ability of waterlogging prediction models for different waterlogging-prone areas in cities and the diversity of spatial characteristics of waterlogging causes.The method uses irregular structureless grids and buffer zone analysis methods to generalize the spatial feature information of the city and divide it into grids of different sparsity levels,thus reducing the complex diversity of urban spatial features.The traditional Urban Storm Water Models are used as the theoretical basis,and the mathematical model of urban flooding is used to analyze meteorological characteristics,urban topographic features and drainage network characteristics,and transformed into runoff time series data distributed with rainfall time.It is input into the hybrid neural network together with the water depth and meteorological data to achieve accurate prediction of future water depth changes in waterlogging-prone urban areas.The experimental results show that the method well reduces the complex diversity of urban spatial characteristics,effectively combines the spatial generalization characteristics and meteorological characteristics of the city with the characteristics of waterlogging time series,and further improves the generalization ability and prediction efficiency of the prediction model on the basis of the original improvement.Finally,based on the above research results,a waterlogging early warning system based on the adaptive urban flooding model is designed and implemented.The visual monitoring of waterlogging data and the back-end management system adopt the architecture technology of front and back-end separation.An efficient and high-precision waterlogging early warning system based on the adaptive urban flooding model is realized,which provides data support for urban flooding prevention and control decisions and ensures the smooth development of urban construction and the travel safety of people. |
