Research On Estimation For The Level Of Hydrophobicity Of Silicon Rubber Insulator's

When the silicon insulator's hydrophobicity in power network operation decreases or even completely lost, flashover voltage of silicon insulator would be lowered down and chances of pollution flashover accidents would be increased, which greatly threaten the safety and stable operation of power system. Therefore, it is urgently needed to carry out effective researches on the detection technique of silicon insulator's hydrophobicity. This detection technique would help us to find out the aged silicon rubber insulators then improve the stability of transmission line's operation. At present, the existed detection methods at home and abroad are still incompleted. What's more, the only method to identify the level of hydrophobicity quantitatively, indicator function, can not identify different levels accurately and easy to mix the similar levels and so on. According to this situation, this paper makes a research on estimation for the level of hydrophobicity of silicon rubber insulators, to make quantitative judgments on the level of hydrophobicity and guarantee the classification accuracy.In this paper, the correlative research on the level of hydrophobicity is based on hydrophobic image. At the beginning, the noise in the high frequency coefficient can be filtered by Haar wavelet function, then, hydrophobicity image is segmented to binary image through adopting a method based on maximum between-cluster variance.Considering that both the pollution degree and hydrophobicity are the basic parameters the pollution performance of silicon rubber insulators, the equivalent resistance model is built on the hydrophobic image(binary image), After analyzing the law that the equivalent resistance of different levels of hydrophobic image changes along with pollution class, there comes out the results show that the resistance mode can show a relatively clear hydrophobic image when the surface of silicon rubber insulator is mild polluted. However, the resistance mode cannot effectively distinguish the hydrophobic image, because the equivalent resistance of images HC2 and HC3, HC4 and HC5 are very close to each other when the surface of silicon rubber insulator is seriously contaminated.According to the distribution of water droplets or traces of water on the surface of hydrophobic image, it is founded that the distribution of water droplets or traces of water on the surface of the hydrophobic image could be indicated by the surfacefractal dimension situation. On the Criterion of Combining the box dimension of the hydrophobic image(D),the box dimension of maximum water droplets(d) with the ratio of the box dimension of maximum water droplets( df)is proposed in paper, and the hydrophobic image is effectively classified into good(HC1-HC3), fair(HC4-HC5)and bad(HC6-HC7)states.Considering that box fractal dimension could only measure the fractal characteristics of hydrophobic image from a single aspect instead of overall aspects,A judgment mechanism for hydrophobic image classification based on multifractal is proposed. It can measure the hydrophobic image from different aspects and construct the standard Dq-q curve for hydrophobic image samples, and then identify the hydrophobic images of unknown levels.Finally, the results state that the method proposed in this paper can effectively identify and judge the level of hydrophobic. Meanwhile, factors that affect the judgment is also analyzed.