| In recent years, the overall layout of multi-reservoir joint optimization operation has been forming in large river basins of China. How to make multi-reservoir joint optimization operation scientifically and reasonably is the key point of water resources planning and management and.real-time operation. With large-scale multi-reservoir systems becoming more and more complicated, multi-reservoir operation meets new challenges. For instance, how to allocate common water supply tasks of each member reservoir; how to resolve optimization operation problem of the complicated multi-reservoir systems; and how to make informed decisions to achieve a striking balance among different objectives of multi-reservoir systems. This thesis carries out research on the construction and the solving of multi-reservoir operation model, and multi-objective decision accordingly. It takes the East-to-West Middle-line Inter-basin Water Transfer Project (IBWTS) in Liaoning Province of China as a study case. And the main research results are as follows.(1) Firstly, the water resources allocation pattern of Middle-line IBWTS is analyzed from many aspects such as the development situation of society and economy, the character of water resources, and the actual situation of water supply. Then, the structure of Middle-line IBWTS is generalized. And then, the long-term observed runoff data and water demand data are analyzed. Finally, based on all of the above materials and analyses, this thesis confirms the problems and the goals of the project.(2) The multi-reservoir operation model for water supply is established. Firstly, the basic form of reservoir water supply rules are illustrated. Secondly, the objective functions is determined according to the goal of Middle-line IBWTS. Thirdly, different types of constraints and the common decision variables are given. Finally, the problem of existing techniques for solving reservoir optimization operation problem is illustrated and the corresponding solution strategy is put forward.(3) A sensitivity-informed dimensionality reduction method is presented to resolve the "dimension disaster" problem in the optimizaton operations of multi-reservoir systems. Firstly, Sobo''s global sensitivity analysis is used to screen insensitive decision variables and guide the formulation of the optimization problems with a significantly reduced number of decision variables. Then, the reduced complexity multi-objective reservoir operation problems is solved to give the initial search results. Finally the search results are used as pre-condition of the full search of the original optimization problem. The Middle-line IBWTS sensitivity analysis results show Sobol''s global sensitivity analysis can identify the sensitivity of the decision variables, reduced number of decision variables and improves the efficiency and effectiveness of multi-objective evolutionary optimization.(4) Two methods for allocating common water supply tasks are presented to solve the problem of allocating common water supply tasks in complex multi-reservoir system. One is a compensation adjustment method considering the limit curve of supply. This method can avoid the problem of oversupply or less supply which exists in existing method. The other is the water ratio method combined with compensative regulation method. This method can allocate common water supply tasks according to the properties and water supply tasks of each reservoir. This two methods are applied to DaHuoFang-GuanYinGe-ShenWo multi-reservoir system, which is a part of Middle-line IBWTS. The results reveal both two methods can exert the effects of the underlying hydrological compensation and operation compensation and lead to better allocation of common tasks for water supply of complex multi-reservoir system.(5) Pareto solutions can be obtained by multi-objective evolutionary algorithm which brings new a challenge to decision making, according to this problem, this Paper put forward a weight combination design method based on improved crowding distance, using the Euclidean distance to calculate the crowding distance, the farther the distance, the less similar, which will ensure the diversity of weightings; Then 13 solutions were selected from among 968 solution of Pareto solution set by the generated weight combination, which is provided for decision makers as a decision making table. Eventually, a final decision is recommended based on the principle of fairness.(6) In order to reveal how controlled inflow (CI) mitigates the effect of natural inflow (NI) variability to exert the compensation on reservoir operation, Dahuofang-Biliuhe reservoir system is taken as a case study. The results show that:CI complements NI in terms of water supply reliability through a nonlinear relationship, as shown by an inverse S curve; the two turning points in the curve specify a range of CI with respect to the stochastic nature of NI and the reservoir operation objectives; for the case study reservoir, if the amount of annual CI increases to about 54% of the NI, the primary reservoir storage capacity can be reduced by 40% while providing the same level of water supply reliability. |
PDF Full Text Request