Research And Desigh Of Flood Disaster Monitoring And Pre-warning System

Currently,global climate change leads to frequent flood disaster events.It is of great significance to build a more accurate flood disaster monitoring and warning model based on the geographical and other congenital conditions in disaster-prone areas.The flood disasters caused by rainstorm and typhoon severely affect the security of people’s lives and property.When the flood disasters occur,especially regional flood,which are not only triggered by uncontrollable natural factors,but also are interfered by human beings,thus producing a highly non-linear relationship in runoff prediction.As a result,the hydrophysical model which only relies on rainfall and topographic data for runoff analysis is not satisfactory.How to predict runoff more accurately has become a tough nut to crack in flood disaster research.This paper mainly takes "upstream river basin of Tan Jia Fang hydrological station as study area,which is located in Shandong province.this region is prone to heavy rains in summer under the influences of climate,special physiognomy as well as sea-land thermal effect.The convergence of these runoffs in this area causes the leap of water storage in the upstream reservoirs,such as the Ren river and the Black Tiger Hill reservoirs.It will lead to the increasing in runoff and then floods that hit coastal villages when the reservoir reserves are close to or exceed the waterline,Based on the construction of flood hydrological model as well as the research and development of flood disaster monitoring and warning system,this paper conducts research on flood disaster prevention and control in this region.The specific work and innovations are as follows:1.Combined the research processes and experiences of flood hydrological model and monitoring as well as warning system at home and abroad,this paper collected the data required for constructing flood hydrological model and developing flood disaster monitoring and warning system in this research area.In addition,in order to meet the format requirements of the model data input,the script interface is designed and written,the data are processed in batches and the corresponding data are uploaded to the database in large number.2.This paper proposes a combination between hydrophysical model and neural network model,and constructs the CSSA＿LSTM hydrological forecasting model based on the weaknesses and strengths of hydrophysical model and neural network model in runoff forecasting.The model not only has the ability of CREST hydrological model to simulate the real runoff,but also has the advantages of neural network model to predict nonlinear runoff data.The CSSA＿LSTM hydrological prediction model drives CREST module to operate with rainfall and potential transpiration data,and outputs various runoff related hydrological basic data through the measured runoff rate-setting CREST module of Tanjiafang Hydrological station.The hydrological basic data and measured runoff data are coupled into the neural network module for multi-value coupled runoff model prediction.3.In order to improve the prediction accuracy of CSSA＿LSTM hydrological model further,this paper improves the design of neural network module and adds sparse self-attention mechanism into the network structure.By comparing the prediction effect with simple LSTM hydrological model,it is proved that CSSA＿LSTM hydrological model has better prediction effect.4.A flood disaster monitoring and warning system centered on the research area is designed.The system takes CSSA＿LSTM hydrological model as the core,Web GIS,Openlayers and Geo Server as the supports,and Postgre SQL database and Geo Server as the tools of data storage and processing.On the basis of the geographical base map of the system,the functions of rain monitoring and warning,water monitoring and warning,line chart used to analyze the runoff station,satellite cloud map and so on are developed and designed.Monitoring and warning of flood disasters will be realized.