Research On Multi-objective Optimization Methods For Complex Operation Requirements Of Cascade Hydropower Stations

The "Electricity Transmission from West to East" project implementation for more than ten years,there have been built the Lancang River,Jinsha River,Wujiang River,Hongshui River and other extra-large river basin cascade hydropower stations with over 10GW installed capacity in the southwest of China.Different from small and medium-sized hydropower systems,the extra-large river basin cascade hydropower stations is typically characterized by multiple hydropower plants,large installed capacity,wide transmission range and close hydraulic and electrical connections.It presents many new operation features and requirements,including more attention to the overall benefits and sophisticated control requirements,demands of considering the power generation,storage energy,shipping,peak shaving and other comprehensive interests and requirements of more efficient and practical new multi-objective optimal operation methods.Therefore,taking a number of engineering projects in different extra-large river basins as the background and its new actual scheduling demands being oriented,focusing on the multi-objective problem of responsing cascade hydropower stations complex operation requirements,the multi-objective optimization methods are studied on four aspects,respectively the sophisticated control of overall energy storage,the response to peak shaving and navigation needs;the efficient calculation for multi-objective optimization and the practical calculation of scheduling plans.The main content of present paper is summarized as follows:(1)For the sophisticated scheduling demand of cascade hydropower stations under the overall control conditions of hydropower in the power grids,an optimal operation model of cascaded hydropower stations under the storage energy control is presented and a refinement method based on iso-storage energy line is provided to optimiza the model.In the presented model,the objectives are to maximize the cascade total generating capacity and to achieve the cascade energy storage control.In the proposed method,firstly,the feasible region precompression strategy is introduced to reduce the search space to some extent.Then the indicators of energy storage and the iso-storage energy line are used to identify the feasible state combinations at each period.In this way,the problem is translated into a constrained optimization problem.Finally,the traditional dynamic programming is used to solve the problem.The above model and mehod are implemented on the Lancang River hydropower system.The results show that the proposed method can balance cascade power generation efficiency and storage energy reserves effectively,so that to get equitable distribution between the present and future hydropower resources of grids.It can real service for the engineering practice of sophisticated operation for cascade hydropower stations.(2)For the widespread huge peak shaving pressure in most power grids and the flourishing shipping needs in many large rivers.a multi-objective optimal operation model for cascaded hydropower stations with peak load regulation and navigation demands is presented and a hybrid search method of coupling intelligent algorithm and heuristic correction strategy is proposed to optimal the model.In the model,the peak shaving and shipping requirements are taken into consideration.The former is to minimize the maximum residual loads of the power grid and the latter is to minimize the variance of water level in navigable river.In order to improve the performance and efficiency of the presented method,based on non-dominated sorting genetic algorithm(NSGA-?).improved genetic operator combined with the characteristics of the model is proposed.Moreover,some heuristic solution strategies to effectively deal with complex temporal coupling constraints and the specified level control targets are provided.The proposed method has been applied to the joint optimal operation of Jinghong and Ganlanba cascade hydroplants on Lancang River.The results show that the proposed model and method can give full play to the role of the re-regulation hydropower station and meet the demands for peak load regulation and navigation simultaneously.(3)In order to adapt the rapid growth of the system size of cascade hydropower stations and increasingly diverse scheduling requirements,effective calculating methods of multi-objective optimization operations for cascaded hydropower stations are proposed.It mainly includes two aspects of the work.On the one hand,multi-objective quantum-behaved particle swarm optimization(MOQPSO)is proposed.It is based on quantum-behaved particle swarm optimization(QPSO)and achieves the development of QPSO in the field of multi-objective optimization scheduling by external file set and other search strategies.On the other hand,parallel multi-objective genetic algorithm(PMOGA)is presented.It is based on Fork/Join parallel computation framework and realizes the parallel computing of the multi-objective genetic algorithm(MOGA).The above two methods are from swarm intelligence algorithm and implement the new method development and advanced technology integration respectively.The provided methods are applied to the cascade hydropower stations in the Wujiang River and the Lancang River and have achieved good calculation results.They enrich the method theory system of multi-objective optimization operation for cascade hydropower stations.(4)To meet practical pressing needs of the optimization scheduling solution for extra-large river basin cascade hydropower stations,which based on the background of increasing large computational scale and increasing complex running environment,an instruction operation optimization method based on data mining for cascade hydropower stations is proposed.It combines data mining technology and is divided into two layers to solve.Firstly,some key indexes are selected from the massive operation data of hydropower system,and the fuzzy clustering is used to build the decision-making database.Then,the large system decomposition coordination model and progressive optimization algorithm are employed to optimal output curves and amplitude change and search for the optimal decision.The proposed method was applied to the cascaded hydropower stations located on the Lancang River.The results show that it can availably build the decision-making database of typical output curves for cascade hydropower station group and quickly obtain the feasible cascade joint operation scheme under the dispatch instructions,which is a practical approach for the large-scale hydropower system optimal operation.