Study On Long-term Optimal Scheduling And Application Of Data Mining

Energy is an important material basis for the survival and development of human society,which is related to the people's livelihood and national strategic competitiveness.It is a major historical mission of energy development and reform to establish and implement the concept of innovation,coordination,green,open and shared development,and to promote the transformation of energy production and utilization,and to build clean,low-carbon,safe and efficient modern energy system.Hydropower energy has the advantages of renewable,clean and efficient,and the development of hydropower energy is an important strategy of national energy system modernization goal.With the rapid development of social economy,large-scale hydropower station reservoir(group)gradually formed,the status and role of reservoir dispatching more and more prominent,this paper focuses on the long-term optimal dispatching of reservoir,and combines the current hotspot technology to carry out in-depth study on reservoir planning guidance.The main results of this paper are as follows:(1)Study on the Optimization of Reservoir Scheduling.According to the idea of implicit random optimization,the optimization model of conventional operation chart is constructed with the goal of maximizing the power generation with the guarantee rate.The model is solved by using the progressive optimal algorithm of corridor search strategy and adaptive step size exponential contraction strategy.The study of the Three Gorges Hydropower Station and Xiluodu Hydropower Station shows that the proposed method can effectively improve the guiding effect of the operation chart on reservoir operation and provide theoretical guidance and scientific basis for the efficient operation of the reservoir.(2)Aiming at the shortcomings of the conventional operation chart without using the runoff information,this paper presents a three-dimensional operation chart of the reservoir in the form of dispatching time-water level and runoff frequency.The three-dimensional operation chart is initialized with the optimized operation chart,and the three-dimensional operation chart optimization model is constructed.Aiming at the problem that the computational scale is multiplied and the time consuming is increased,the parallel strategy is proposed to reduce the computation time.The research example of Xiluodu Hydropower Station shows that the new form of three-dimensional operation chart can effectively use the runoff information to optimize the scheduling decision,reduce the amount of discarded water and improve the power generation efficiency of the reservoir,which can provide a new form for the reservoir operation chart.(3)Combined with the reservoir scheduling rules and data mining classification technology,carried out reservoir power generation scheduling research based on the decision tree model.The data of the optimal scheduling results are processed by the output control decision and the final water level control decision classification method,and the decision tree is generated by the C4.5 algorithm and the decision tree scheduling rules are extracted.Finally,the reservoir scheduling model is constructed to carry out the scheduling simulation,and compared with the conventional scheduling method,the validity of the decision tree scheduling rules is verified.The research example of the Three Gorges Power Station shows that the decision tree method of data mining can extract intuitive,concise and efficient scheduling rules.The simulation results show that the guidance effect of the reservoir is better than that of the conventional scheduling method,which can provide a new idea for long-term power generation scheduling.