Research And Application Of Data Mining In Power Dispatching Automation System

Data Mining is a newly developing technology based on machine study in artificial intelligence and database. After analyzing and processing a large amount of data, useful information and knowledge in those data can be obtained to meet people's increasing demand of knowledge. Applications of using data mining technology in power dispatching automation system will be more widely applied and make more practical value.As we all known, short-term load forecasting (STLF) is a part of advanced functions of power dispatching automation system. Accurate load forecasting is helpful to the security and stability of power system as well as generation costs. With the development of electric industry, STLF will play more and more role in power system.In this paper, data mining technology and the theory of STLF are discussed at first, and then the thesis puts emphasis on studying of artificial neural network (ANN) method. Following that a hybrid model of STLF is established based on back-propagation neural network. In order to improve the learning speed of ANN, a revised BP algorithm is adopted by using smooth coefficient and forgetting coefficient. To improve accuracy of forecasting, all the facts, such as influence of basic load, temperature, weather related sensitive factors and festive national holidays are considered systematically and simultaneously.In final part, the thesis presents a design and realization of power dispatching automation system. The hybrid neural network forecasting model has been applied in the "Power Dispatching Automation System of Tian Men Power Supply Bureau-Subsystem with load forecasting" and has obtained preferable forecasting results.This article has six chapters:Chapter 1 mainly describes the content and significance of the article, the theory and technology of data mining as well as the applied actuality of data mining in power dispatching automation system.Chapter 2 introduces the theory and common forecasting method of STLF. Comparing with many other kinds of forecasting algorithm, a revised BPalgorithm based on ANN is put forward.Chapter 3 mainly presents the establishment of a hybrid model of STLF. And emphasizes on the introduction of the choosing strategy of neural network structure,the method of smoothing and generalizing load data and the studying and forecasting process of the STLF model which adopts the revised BP algorithm.Chapter 4 introduces the design and realization of application platform of the hybrid STLF model, i.e. the power dispatching automation system, which mainly includes the architecture of the system, hardware enviroment, database design as well as the software component and so on.Chapter 5 introduces the application and forecasting effect of the hybrid STLF model in actuality. Comparing with the common BP model, it is proved that the hybrid model has better studying speed and forecasting accuracy, and makes well practical value.Chapter 6 sums up of the paper and puts forward the working plan in the future.