Font Size: a A A

Study On Load Classification In Power System

Posted on:2008-12-17Degree:MasterType:Thesis
Country:ChinaCandidate:Z H ZhangFull Text:PDF
GTID:2132360245991986Subject:Power system and its automation
Abstract/Summary:PDF Full Text Request
Load classification is an essential task in economic analysis, operation and planning of distribution systems, especially for the rapid development of power deregulation and demand side management (DSM). Nowadays, the load classification has become the base of price setting, load forecasting, system planning, load modeling, etc. Traditional load classification used in the power utility is usually accomplished based on the economic activities of customers. However, characteristics of the load curves with similar economic activities may be absolutely different. So it is very important to find more effective and precise load classification method not only for the theoretical study but for the power system application.In this thesis, data mining and fuzzy clustering analysis are used to find more effective load classification method. Load forecasting based on load classification is further discussed. Work of this thesis is as following:1. Load classification method based on the clustering analysis is deeply studied. Firstly, comparison is performed between several clustering validity functions to find the optimal one. Then, fuzzy clustering analysis and fuzzy weighted clustering methods are used to classify the measured load profiles. The loads in the same group are guaranteed to be homogenous.2. Two pattern recognition methods, i.e. maximum degree of membership and mean square deviation of series, are deeply studied to recognize the pattern of load customers.3. Finally, load forecasting method based on load classification is discussed. Two forecasting methods, typical load curve superposition method and support vector regression method, are introduced. Their effectiveness is then validated by the numerical results.Studies in this thesis are helpful to find more scientific and effective load classification and forecasting method.
Keywords/Search Tags:Load Classification, Clustering Analysis, Pattern Recognition, Load Forecasting, Power System
PDF Full Text Request
Related items