Short-term Energy-saving Power Generation Operation Of Hydro-thermal Power System

Energy-saving power generation operation is about optimization of unit commitment and load distribution by giving priority to the use of renewable and clean power generation resources without suffering safety and reliability of system, according to type of generating sets, energy consumption level and environmental protection index. Short-term power generation operation of hydro-thermal system covers many aspects from load forecasting, maintenance of generating sets, unit operating method, load distribution, power flow calculation to transmission losses analysis etc, and can be used to guide daily operation. Introduction of energy saving and consumption reducing to the making of short-term power generation scheduling is of great significance for carrying out energy-saving power generation operation.In this paper, the short-term energy-saving power generation operation problem is studied from the three aspects: short-term load forecasting, load distribution optimization and hydroelectric plant economic operation. The main contents and research process are as follows:(1) By analyzing of the features of short-term power load, Support Vector Machines (SVM) model, which parameters are estimated by SCE-UA algorithm, is presented as promising method for short-term load forecasting. The Radial basis function is selected in the establishment of the model, and SCE-UA algorithm is used to identify the parameters of SVM, which uses load data of similar day as training sample and daily load forecasting accuracy maximization as object. The optimized SVM model is applied to short-term load forecasting, and compared with the forecasting results of artificial neural network model. The results showed its effectiveness and accuracy.(2) A new method, which combines the improved cutting load method and the equal increment rate method, is presented to solve load distribution of hydro-thermal power system for saving energy and reducing energy consumption. The improved cutting load method is used to get hydropower stations' working position in system daily load map and their power generation scheduling according to their sorting principle. The equal increment rate method is used to optimize the system remaining load among the affirmatory running units according to their sorting principle. The character of proposed method is that hydropower resource is adequately used and coal consumption falls to minimum. To the examples of Guizhou Power Grid, the study indicated that the proposed method is practical and feasible, and presents a feasible project for energy-saving power generation operation of power grid.(3) A model has been introduced to solve the economic operation of hydroelectric plant, and the improved mixed Particle Swarm Optimization (PSO) algorithm embedded with Tabu Search (TS) is used in this model for results. The TS-PSO algorithm uses the mixed encoding technology involving discrete binary variables and the continuous variables. The discrete binary PSO is used to solve the unit commitment sub-problem, and the standard PSO is used for loading economic dispatch sub-problem. Parallel optimization is made possible by combining the two sub-problems for the reason they are interrelated. To avoid entrapment in local optima, the flexible memory system of TS is embedded into PSO to keep particles' diversities and in turn enlarge the search space and enhance convergence property. The experimental results by using the TS-PSO method in hydroelectric plant reveal that, compared with PSO, it effectively overcomes the premature phenomenon and improves the global convergence and global optimization searching capability.(4) A web-based short-term power generation operation system for power grid is successfully designed and developed on J2EE platform by means of JBuilder as the IDE and Oracle9i database as the data saving and managing warehouse. Decision support system embedded with intelligent models is introduced to short-term power generation operation and management. The key techniques, such as system architecture and function, design of work flow, development method of EJB component, and design of database et are covered in this article. The system can process a variety functions from load forecasting, equipment maintenance, generating set opening, load arranging, electric power, energy balance calculating, to making power generation plan and report forms. The system stands out for its easiness to use, strong functions, good stability, transplant-ability and extensibility.Finally, a summary is given and some problems to be further studied are discussed.