| Summary: | 碩士 === 國立勤益技術學院 === 資訊與電能科技研究所 === 93 === The goal of this thesis is to study the applications of extension theory to load forecasting and partial discharge (PD) pattern recognition in power systems. In the load forecasting, a novel extension clustering method based on the extension theory combined with multi-regression analysis method (MRAM) are introduced to build the load forecasting models for long-term and short-term loads forecasting. Based on the extension theory, the matter-element model can build the every load type model, and then the changed range of long-term and short-term loads at forecasting time can be forecasted according to correlation degree between the built models and the forecasting models. Second, according to the load data of every load type,using the multi-regression analysis method (MRAM) to build the load forecasting models of every load type, then the forecasting models can be used to forecast the values of long-term and short-term loads at forecasting time. To verify the proposed forecasting methods, the statistics data of the real operation in Taiwan have been tested and the methods have given rather encouraging results.
In the PD recognition, this thesis proposes a novel extension based clustering method to recognize the three dimensional (3D) PD patterns of the high voltage cast-resin current transformers (CRCT). First, three data preprocessing schemes that extract relevant features from the raw 3D-PD patterns are presented for the proposed PD recognition method. Second, the matter-element models of the PD defect types are built according to PD patterns derived from practical experimental results, then, the PD defect in a tested CRCT can be directly identified by degrees of correlation between the tested pattern and the matter-element models which have been built up.
To verify the proposed PD recognition method, 150 sets of field-test PD patterns are tested with rather successful results.
The study results verify that the proposed method is computing fast,high recognize and predicting accuracy in electricity load forecasting or the PD recognition.
|
|---|
