Research On Multi-objective Interaction Relationship And Decision-making Methods Of Cascade Reservoirs

Reservoir is an important engineering measure for the spatial-temporal allocation of water resources.It can reduce or avoid the loss of flood disaster,alleviate the crisis of water resources,improve the ecological environment of upstream and downstream,and increase the benefit of hydropower generation by storage and discharge regulation of inflow.In recent years,with the gradual improvement of water conservancy project construction and the rapid development of social economy,the relationship among flood control,power generation,water supply,ecological protection and other departments is becoming more and more complex,and the contradictions are becoming more serious.In addition to the comprehensive impact of global extreme climate change on river basins,it is of great practical significance to carry out the research on multi-objective intraction relationship and decision-making method of cascade reservoirs,and to seek for multi-objective optimal operation scheme under the complex engineering and environment of river basins for disaster prevention and efficient utilization of water resources.Centered on the four objectives of flood control,power generation,water supply and ecological protection,this paper makes a further study on four aspects—flood control risk assessment,multi-objective optimal operation of power generation-water supply-ecological protection,interaction relationship analysis and multi-attribute decision-making of flood control-benefit making—using multidisciplinary theories and methods such as system engineering,big data,probability and statistics,management and operational research.The main research contents and achievements are as follows.(1)Flood control risk analysis considering flood peak type and its frequency.In order to solve the problem that the existing flood simulation methods do not consider the influence of a flood's characteristic variables on the shape of its hydrograph,a stochastic flood simulation method considering peak type and its frequency is proposed.The Copula function is used to simulate flood characteristic variables,which takes the correlation between the variables into account.Cluster analysis is employed and dimensionless flood hydrographs are generated,which considers the variability of flood peak types.The proximity degree with peak type coefficient and peak time as the main variables is calculated,which considers the influence of flood peak,flood amount and flood duration on the frequency of different peak types' flood hydrographs.Therefore,compared with traditional methods,the flood simulated by this means is closer to the natural flood process,which can effectively improve the calculation accuracy of cascade reservoirs' flood control risk.(2)INSGA-? algoritHm and its application in multi-objective optimal operation of cascade reservoirs.Aiming at the shortcomings of NSGA-II algorithm in the initial population uniformity and high-dimensional characteristics processing of multi-objective optimization,an INSGA-? algorithm is put forward,which adopts both the population initialization strategy based on orthogonal design to ensure the quality of the initial population and the problem transformation strategy to ensure the search strength of the algorithm in a high-dimensional space.According to the comprehensive utilization requirements of cascade reservoirs aimed at promoting benefits,a multi-objective joint optimal operation model of cascade reservoirs is cons·tructed,whose objective is to optimize the weighted value of each index that needs to be considered for power generation,water supply and ecological operation,and Using INSGA-II algorithm to solve the model.The advantage of the method in multi-objective optimal operation over other algorithms is proved by comparative analysis.(3)Research on the interaction relationship among benefit objectives of cascade reservoirs based on structural equation modeling.In view of the high dimension,complex relationship of variables and the difficulty to quantify in classical statistical methods in the study of the multi-objective interaction relationship of cascade reservoirs,structural equation modeling is introduced to develop a cascade reservoirs' multi-objective interaction relationship model.The inflow data of different typical years are input to the multi-objective optimal operation model of cascade reservoirs,which is solved by the INSGA-II algorithm to obtain a non-inferior solution set.Through qualitative analysis of the non-inferior solution set,a preliminary assumption of the interaction relationship among power generation,water supply and ecological protection is put forward.The index value in the non-inferior solution set is used as data source to carry out high-dimensional confirmatory factor analysis.The quantitative calculation of the interaction relationship among cascade reservoirs' benefit objectives in different typical years is realized.And with the calculation results,the mutual response mechanism of power generation-water supply-ecological protection is explored in depth.(4)INGTDM-MAS and its application in multi-attribute decision-making of flood control-benefit making.In order to solve the problem that the traditional interval number decision-making model only takes the upper and lower bounds of indexes as decision criteria and deems the interval's interior to be uniformly distributed,a multi-dimensional association sampling algorithm is proposed and incorporated into the grey target decision-making model,to establish an Interval Number Grey Target Decision-making Model Based on Multi-dimensional Association Sampling(INGTDM-MAS).In order to reduce the fuzzy error in the process of expert subjective weighting,the set-valued statistical method is introduced to weight the target layer,the path coefficient in the calculation results of the structural equation model is used to weight the index layer.INGTDM-MAS is applied to the multi-attribute decision-making of cascade reservoirs'flood control risk-benefit making,compared with the results of the traditional model,the improved model is more refined in dealing with interval number,and can reduce information loss in decision-making process.