Study On Optimal Flood Control Operation Of Cascade Reservoirs In Henan Section Of The Middle Reaches Of The Yellow River Under Changing Environment

The Sanmenxia and Xiaolangdi Reservoirs are two large reservoirs located in the middle reaches of the main stream of the Yellow River.They are an important part of the flood control project in the "San-Hua" interval reach.How to control and use existing water conservancy projects to store and detain floods,and reduce flood peaks,scientifically and rationally,has important practical significance for flood control and disaster reduction in the middle and lower reaches of the Yellow River.In recent years,due to continuous changes in climate and underlying surface environment,floods in the middle and lower reaches of the Yellow River Basin are also evolving.At the same time,when the two reservoirs’ existing flood control and dispatching plans were formulated,there were problems such as less calculation of design floods or rough consideration of peak volume correlation and flood area composition.In order to improve the rationality of the cascade flood control operation plan,based on the diagnosis and restoration of the inconsistency of the flood sequence,this paper introduces the Copula function to analyze the combined distribution of flood peaks and the regional composition of the design flood,and establishes the flood control optimization of the SanmenxiaXiaolangdi cascade reservoir.The dispatching model is solved by using an improved genetic algorithm to explore the optimal dispatching plan for flood control of cascade reservoirs in a changing environment.The main research contents and results are as follows:(1)Variation diagnosis and restoration of inconsistent flood sequence under changing environment.Taking floods at Huayuankou Station and Heishiguan Station in the middle reaches of the Yellow River as the research objects,the mutation diagnosis system is used to analyze the trend and abrupt changes of the flood sequence.On this basis,a variety of reduction calculation methods are introduced to calculate the design flood,and the design flood is compared with the existing design flood.The comparative analysis of the results indirectly proves that the method of restoring inconsistent floods adopted in this paper is more reasonable.(2)Design flood calculation based on multivariate joint distribution.The traditional design flood estimation method assumes that the flood peak and the flood volume are at the same frequency,without considering the correlation between the two flood elements.Therefore,in response to this shortcoming,this paper introduces the Copula function to derive the design flood during the return period of the joint variable,that is,by selecting the marginal function that conforms to the peak and flood distribution,and then selecting the most suitable Copula function for the peak-volume joint distribution,taking into account the flood peak Under the combined action of the flood volume,the design flood peak and the flood volume under different frequencies are calculated,and then the design flood process of different frequencies is obtained.(3)Analysis of design flood area composition of main and tributary streams in different return periods.There are multiple sets of inter-regional flood inflows in this study area.If we only rely on the previous assumptions of the same frequency among the flood components in a specific return period,it is possible to ignore the interaction between floods in the main and tributary regions of the basin.Therefore,based on the existing flood data and the spatial distribution of rivers,this paper quantitatively analyzes the correlation degree of floods in the design sections of each area;introduces the Copula function to establish the joint distribution of floods among the various components of the flood;uses an improved genetic algorithm for optimization,Obtain the most likely design flood area composition in each specific return period.(4)Construction and solution of optimized dispatching model for cascade reservoirs.According to the actual situation of the study area,the SanmenxiaXiaolangdi cascade reservoir joint flood control optimization dispatch model was established.A more efficient and applicable genetic algorithm is used to solve the problem,and improvements are made in the generation of genetic initial populations,parameter adaptation,and quasi-annealing acceleration calculations.The results show that after the cascade joint optimization operation,when the same frequency of the river basin encounters a once-in-a-hundred-year flood,the peak shaving rates at the Sanmenxia dam site,Xiaolangdi dam site,and Huayuankou section reached 64.25%,78.58%,and 34.20%,respectively.The results of conventional dispatch have increased by 6.88%,9.14%and 2.56%respectively;when Huayuankou encounters floods in a century and once in two hundred years,after the optimized dispatch of the SanmenxiaXiaolangdi cascade reservoirs,the highest peak flood discharge of Huayuankou section was 15136.47 m3/s and 17339.58 m3/s,which are 4.54%and 3.19%lower than conventional scheduling.This method gives full play to the joint flood control function of cascade reservoirs,and reduces the risk of flood disasters in Huayuankou and downstream areas.