Font Size: a A A

Study On Short-Term Load Forecasting For Power System Based On Complementation Of Fuzzy-Rough Set Theory And Artificial Neural Network

Posted on:2009-07-30Degree:MasterType:Thesis
Country:ChinaCandidate:P ChenFull Text:PDF
GTID:2178360242487257Subject:Agricultural Electrification and Automation
Abstract/Summary:PDF Full Text Request
When the Power system is controlled on line, we should use the short-term load forecasting to realize reasonable dispatch of sending and supplying power. The short-term load forecasting is the important basis of security dispatching and economical operation in the power system, so the precision of the load forecasting directly affects not only the reliability and the economical efficiency of the power system operation, but also the quality of supplying power. With the development of power cause, the power department paid more and more attention to the forecasting precision for the maximum profit.The Artificial Neural Network is based on the physiological research results of the human brain. The purpose is simulating the mechanism of the brain,and realizing a certain function. The ANN is a nonlinear system which is consitituted by a great deal of simple computing units. They imitates the human brain's functions in information processing,storage services and searching, so they have the intelligent functions of learning, remembering and calculating. Lots of results show that nerve network model using for load forecasting is effective. In this paper, through analysis and comparison, the author chooses the improved BP neutral network that has only one hidden layer to do the research on STLF.In order to raise the forecast accuracy of ANN, for one thing, we can increase the number of training samples. For another thing, we can increase the influencing factors introduction. But both the neural network's structure and decision rule are greatly influenced by the dimension of input data. In the worst, it also may decide the net's astringency. So, simplify the input attributes is an effective way to improve the quality of forecasting. According to the characteristics of electric short-term load forecasting, a complementation method based on Fuzzy -Rough Set theory and BP NN is proposed to deal with this problem in the paper.The Fuzzy -Rough Set theory introduces the Fuzzy Set into the research of Rough Set theory by the fuzzy membership function. It will discriminate the elements in any class by the fuzzy membership function, and take the membership degree instead of the attribute value. Firstly, this method can consider more influencing factors, cast out the inessential overfull input information by reduction, and put the attribute importance to be the net's input initial weights. In this way, it will greatly simplify the net's structure and enhance the forecasting precision. Secondly, use the Fuzzy -Rough Set theory can deal with the continuous variable directly and get the more reasonable input, which can avoid the information loss in attribute discretization.The testing results on a real power system show that the proposed model isfeasible.
Keywords/Search Tags:Short-term load forecasting, Artificial Neural Network, Fuzzy -Rough Set theory, Fuzzy membership function, Reduction
PDF Full Text Request
Related items