Structure Identification And Its Application Of Cascade Reservoirs Optimal Operation System In Jinsha River Middle Reaches

With the development of the social economy, the problems of the energy sources and environment are increasingly prominent. The rich hydropower resources in our country provide the energy security for the economic social development. Developing the hydropower resources vigorously, which is an important measure for preserving the security of the energy、protecting the ecological environment and realizing the sustainable development. The operation of the hydropower stations is an important link in the reservoir operation and management. The scheduling level will affect the play of the comprehensive utilization efficiency. The high efficient hydropower stations operation can obtain considerable social and economic benefit without adding the investment of the hardware, which is an efficient measure to optimize the energy structure and realize the energy conservation and emission reduction.On the basic of the research status of the hydropower station reservoir scheduling abroad, this paper takes the recognition and application of the optimal operation of hydropower stationsystem structure as the main line and proposes the reservoir optimal operation hydropower station system structure on the basics of cluster analysis and the strategy-making method of the hydropower station stochastic optimization scheduling on the basics of the optimal operation system structure. This paper identifies the optimization of reservoir operation for the cascade hydropower stations of the middle Jinsha River system structure by usingthe proposed method and formulate the implicit stochastic scheduling rule of this cascade hydropower station reservoirs and prediction model of Longpan reservoir’s year-end level under the guidance of the identified system structure. The research results of this paper mainly includes:(1) Proposing the identification method of optimization of reservoir operation of the hydropower stations based on the implicit stochastic thought.Firstly, solving the optimal scheduling model and obtaining the sample data which represent the characteristics of the system. Secondly, primary electing the clustering index from the sample data and pretreating clustering index according to the construction method of the crossed indexes based on the proposed distance Gini coefficient the descending dimension method of indexes time series based on Gini coefficient and the structure descending dimension method of the target system based on the correlation coefficient. And clustering the target system according to improved K-means algorithm.Then extracting the system characteristic of different clustering through the deep analysis of the scheduling process during typical years.At last, summarizing the system characteristic of each clustering and identifying the system structure.(2) The settled output trial model of cascade hydropower stations on the consideration of the constraint of (system) critical quotative discharge.Defining the (system) critical quotative discharge as the maximum quotative discharge of the hydropower stations (system) when there is no surplus water and analyzing the relationship between the maximum discharge and critical quotative discharge. On the foundation of the settled output model of traditional cascade hydropower stations, this paper adds the (system) critical quotative discharge constraints module in order to compatible with the scheduling time when it should take the (system) critical quotative discharge as upper limit.(3) Structure identification and optimality analysis of optimization of reservoir operation of cascade hydropower stations of the middle Jinsha River.Using the system structure identification thought, this paper identifies the optimization of reservoir operation of cascade hydropower stations of the middle Jinsha River structure. The result shows that the system is a piecewise linearity structure system. The scheduling process in any water conservancy year of this system consists of "firm capacity "、"increasing flow" and "critical quotative discharge" in some certain rules. And this paper analyzes the optimality of the three basic elements which constitute the system scheduling process.(4) Proposing the strategy-making method of the implicit stochastic optimization rules hydropower stations based on the optimal operation system structure.Aiming at the shortcomings of lack of mechanism support of implicit stochastic optimization rules, this paper adds a optimal operation system structure identification module based on the thought of selecting different kinds of model structure to match sample data and making implicit stochastic optimization rules which is on the foundation of the sample data provided by traditional optimal operation results. It means that establishing the implicit stochastic rules which have mechanism support under the guidance of identified optimal operation system structure. Take the cascade hydropower stations of the middle Jinsha River for example, this paper sets up the linear dispatching function which takes the annual total available energy as independent variable and the output of each month as decision variable and draw the total energy scheduling figure according to it. At the same time, in order to give full play to the advantage of traditional scheduling graph and the proposed total energy scheduling graph, this paper puts forward the combined scheduling discipline and analyzes the benefit of the combined scheduling discipline. (5) Establishing the prediction model of carry-over storage year-end water level based on the optimization scheduling system. This paper establishes the thought of building the prediction model of carry-over storage year-end water level which have mechanism support under the guidance of identified optimization scheduling system structure. Take the cascade hydropower stations of the middle Jinsha River for example, under the guidance of the identified linear system structure of this cascade hydropower stations, this paper sets up the water conservancy year-end reservoir storage level prediction model, which take the annual total available energy as independent variable, of Longpan reservoir. Through the coupling transforming model of the water conservancy year-end storage level to the calendar year-end level, it can obtain the calendar year-end level prediction model of Longpan reservoir. Compared it with the prediction effect of existing model, this paper evaluates the prediction effect of this model.