Application Of Flood Forecasting Model In Mountainous Rivers

Flood disaster has always been the main factor threatening people's lives and property,especially with the global warming in recent years,the frequent occurrence of extreme weather,coupled with China's mountainous terrain,and the small and medium watershed in mountainous areas affected by topography and other factors,resulting in the construction of preventive engineering measures is difficult and difficult to implement,and the non engineering measures of flood prevention are becoming more and more important.Mountainous rivers are characterized by short duration,high intensity,steep rise and fall of rainstorm flood,and are greatly affected by topography,vegetation types and underlying surface conditions.Due to the lack of hydrometeorological data in these rivers,there is almost no forecast scheme at present,which makes the early warning and forecast difficult.In this study,Liusha River,a mountainous river in Hanyuan County,Ya'an,is taken as the main research object to carry out flood forecasting.The main research work and achievements are as follows:1.Based on the measured hydrological data of Liusha River Basin from 2009 to 2020,the hydrological analysis and calculation of Liusha River are carried out by using Xin'anjiang model and BP model,and the prediction accuracy is evaluated according to the relevant provisions of hydrological code.The prediction accuracy of Xin'anjiang model is above grade B,and the prediction accuracy of BP model is 22.3% higher than that of Xin'anjiang model.2.BP and GA-BP models are different from the traditional non-linear forecasting methods of hydrological forecasting.They are a supplement to the traditional flood forecasting models and effectively improve the accuracy of flood forecasting.3.Aiming at the limitation of BP model itself,which has the problems of non unique structure,slow convergence speed and local optimization,the BP model is optimized by using genetic algorithm,and the GA-BP flood forecasting model of Liusha river is established by using MATLAB(2020b)as the main research means.4.Comparing the prediction results of GA-BP model and BP model,the results show that GA-BP model not only uses global search method to train the model,but also overcomes the limitations of single BP model.Moreover,the prediction accuracy of the improved model is improved,which has good applicability and can be applied to flood prediction.The research results provide important data support for the flood forecasting scheme of small and medium-sized basins in mountainous rivers,solve the problem of flood forecasting in small and medium-sized basins,and provide important guarantee for the prediction and early warning of mountain flood disasters and the implementation of prevention and control projects in mountainous rivers.