| Summary: | 博士 === 國立臺灣科技大學 === 電機工程系 === 101 === Partial discharge (PD) is the main cause of degradation of the insulation in gas-insulated switchgear (GIS). PD phenomena include: surface discharge, cavity discharge, corona discharge, and treeing channel discharge. Previous research has shown that different types of defects in GIS generate different symptoms of PD, which are associated with various degrees of damage to the GIS. Hence, PD detection is essential to the reliable evaluation of insulation systems and the identification of defects in GIS. In this research, the experimental objects were GIS defect models, which were filled with SF6 gas. Three models were designed based on the results of investigations of numerous power equipment failures. Statistical features were extracted from the PD pattern data and were inputs of adaptive neuro-fuzzy inference system (ANFIS). The results reveal that ANFIS classification has a high success rate, reaching an acceptable classification accuracy 90%. In addition to accumulating a huge mass of PD data of GIS, the procedure that proposed in this study can be also used to develop a data base of defect recognition for other power equipment.
|
|---|
