Study On The Flood Forecast Of The Basin Above The Nenjiang-Kumotun Statio

Flood disasters are one of the most serious disasters in nature.Every year,they affect people's production and life to varying degrees and bring losses to my country's economic and social development.As an important non-engineering measure,flood forecasting plays an important role in flood prevention and disaster reduction.Although there have been studies on flood forecasting in the Nenjiang River Basin,the research on the upper reaches of the Nenjiang River is still limited.In addition,some new observation stations have been added in the upper reaches of the Nenjiang River in recent years.Therefore,this article takes the basin above Kumotun as the research area to conduct flood calculations on the river course.,And from the perspective of extending the foreseeable period,the BP neural network and the Xin'anjiang model are used for flood forecasting respectively.The main research contents and results are as follows:(1)Analyze the natural geographic and hydrological conditions of the basin above Kumotun,count the historical flood characteristics of the basin,and combine the distribution of water systems and measuring stations in the basin.It is planned to adopt the Muskingum method,BP neural network,and Xin'anjiang model to analyze the Kumo basin.Research on flood forecasting in the basin above the Tun.(2)The lime kiln-Guli-Kumotun section is taken out separately,and the flow calculation of the lime kiln and the Guli station is carried out by using the Maskingen method of first performance and later integration.Set up two modes of continuous calculation and calculation by flood,and then divide floods into small,medium and large categories,and calibrate the parameters and flow calculations respectively;from the relative error of flood volume,relative error of flood peak,time difference of peak present,and determination Evaluate the calculation results in four aspects,analyze the causes of unqualified floods;finally,compare the classification and non-classification conditions,and conclude that the calculation accuracy can be effectively improved after flood classification,especially for large floods,but Muskingum The method cannot solve the problem of inaccurate calculation of peak current time.(3)Construct a three-layer BP neural network model,use the flow of Limekiln and Guli in the previous period to forecast the flow of Kumotun in the next period,set up two situations of flood classification and non-classification,and in the case of flood classification The following subdivided the two modes of non-interval rainfall and inter-interval rainfall,and finally analyzed the simulation results of each situation,and concluded that: without classification,the flood peak simulation in most years was low,and the overall pass rate was 53.8%;After the flood classification and rainfall are added,the eligible years have increased significantly,and the flood peak,peak-occurrence time difference,and certainty coefficient have all improved to varying degrees.(4)Use the Xin'an River model to simulate the watershed above Kumotun,and use genetic algorithms to calibrate runoff generation and confluence parameters respectively,and finally analyze the parameters and unqualified floods.The qualified rate of runoff generation is 94.5%.The qualified rate of confluence is 77.8%.It is concluded that it is feasible to use the Xin'anjiang model for flood forecasting in the basin above Kumotun,which has certain reference significance.