Design And Implementation Of Flood Forecasting System Based On Xin'anjiang Model

The change of water in nature affects the development of society,which provides essential guarantee for the production of all walks of life,but it threatens people's lives and property at all times.The flood forecast scheme before the formation of floods played an extremely significant role in flood control,and it is an effective means to alleviate the flood.The research is based on the project of the “Lijiang Water Conservation and Guilin Flood Control Reservoir Group Ecological Study” of the China Institute of Water Resources and Hydropower Research.The flood forecasting of the Qingshitan reservoir section can reduce the losses caused by the flood.The thesis drew lessons from the relevant research experiences of existing flood forecasting systems,and analyzed the feasibility of the system based on the software engineering standards,clarified its functional and non-functional requirements,and gradually designed the architecture?problem domain?persistence and interface parts.This thesis used UML(Unified Modeling Language)to model and analyze the business functions of the system.Based on this,it implemented the management functions of basic information such as rainfall value,flood flow,and scheme parameters.At the same time,in order to deal with issues such as flood control and flood alleviation,this thesis used the Xin'anjiang hydrological model as a forecasting model,and analyzed current hydrological model calibration method,this thesis took Nash's efficiency coefficient as the objective function,and used the improved dual-population PSO-DE(Particle Swarm Optimization& Differential Evolution)algorithm to calibrate the model parameters.The thesis is based on the B/S architecture,uses SSM(Spring,SpringMVC,Mybatis)integration framework,and adopts Bootstrap,Echarts,Jquery and other technologies to implement business functions such as data management and flood forecasting.Simulation experiment showed that compared with the traditional parameter calibration method,the hybrid particle swarm algorithm optimization and differential evolution dual-population algorithm used in this thesis can obtain a faster rate;In the system of this thesis,a prediction experiment was conducted on the section of the Qingshitan reservoir in Guilin,by collecting the rainfall information monitored by the rainfall station in the upper river basin and comparing the calculated simulated flowwith the measured flow,it can be known that the forecasting results of the flood forecasting algorithm based on the Xin'anjiang model are consistent with the actual results.