| Compared with open substation, Gas Insulated Substation (GIS) has the advantages of covering the smaller area, running more safely and reliably, less constant-maintenance, overhauling in a longer period and so on. Because of these outstanding advantages, GIS has been regarded as a revolutionary apparatus in electric power domain since it was invented. GIS develops very quickly and gets more and more applications when it came out. In our country, the application of GIS becomes more and more popular from 80's of 20th century. Especially in recent years, GIS has been widely used in the construction and reconstruction of urban electric networks. Operation experiences show that intrinsic defects in GIS still cause accidents though its high reliability. Insulation breakdown is primary reason for GIS failures. Partial discharge(PD) phenomena is most common characteristic before insulation breakdown in GIS. By PD on-line monitoring systems, GIS internal defects can be discovered in time to avoid the occurrence of fatal failure. So the research of PD detection and its pattern recognition for GIS is very important to guarantee its safe operation, find out its insulation condition and defects type in time and guide its maintenance.In this paper, based on analyzing known researches about PD detection in GIS and its pattern recognition at home and abroad, inner-sensing method of partial discharge in GIS is explored, estimation method for image fractal dimension and fractal feature of GIS PD gray intensity image are deeply studied, new pattern recognition method for GIS PD is presented. Main researches and conclusions in the paper are as following:1. Detection of PD in GIS by inner-sensing method is studied based on electrical measurement. Two kinds of inner sensors including disk sensor and cirque sensor, are developed. Their performance is studied deeply. The results of PD detection show that these inner sensors perform well and can meet the demands of PD detection in GIS and cirque sensor is more sensitive than disk sensor.2. Difference box-counting(DBC) method which is used to estimate image fractal dimension is studied deeply. It is discovered that "empty" box may exist and be counted when DBC method is used to estimate the box-number which covers fractal surface and that the box-number by DBC method may be larger than normal. It is pointed out that DBC method is not universal to fractal dimension estimation for all kinds of images, especially for images whose surfaces vary violently.3. Minimum box-counting(MBC) method which is used to estimate image fractal dimension is presented. MBC method overcomes the shortcoming of DBC method. In MBC method, only those boxes which cover image surface indeed are counted and "empty box" is not counted. It is proved that MBC method is an effective method by fractal dimension estimation for GIS PD gray intensity images and Brodatz texture images.4. A piecewise hunting method is invented to search fractal scale independent range. It searches fractal scale independent range by piecewise hunting mode based onthat the mean-square-sum error of fitting line gets least. It is verified that the piecewise hunting method is effective through searching fractal scale independent ranges of Brodatz texture images by this method.5. Five kinds of typical artificial defect models are designed to simulate insulation defects in GIS. These artificial defect models include needle on the conductor model, free particle model, particle on the surface of insulator model, mixed defect model which combines needle on the conductor defect with free particle defect, mixed defect model which combines particle on the surface of insulator with free particle defect. PD sample data of these artificial defect models are got by PD detection system in laboratory. Utilizing these sample data, GIS PD gray intensity images are constructed. 6. Fractal features of GIS PD gray intensity images are studied deeply. Box and information dimension of PD image caused by five kinds of typical artificia...
|
PDF Full Text Request