Investigation on the Geometrical Characteristics of Secondary Arc by Image Edge Detection

The image of secondary arc, which is a common phenomenon in ultrahigh voltage power system, contains a wealthy of information. To determine its geometrical characteristics, the secondary arc image is systematically processed involving gray scale, noise filtering, and image segmentation. To overcome the discontinuous and false edge caused by conventional operators, an improved algorithm is employed; the unrelated edges are fully removed, and the interrupted endpoints are dilated. Based on the results, the component and evolution of secondary arc are discussed; the correlation of the arc luminance and arc current is revealed. Occasionally, the arc roots and arc column have irregular motion behavior. Using the edge-detection technique, the detailed jumping location and velocity of anode arc root are tracked; the life expectancy and stability of coexistence of multicathode arc roots are determined; the morphology of short circuit of arc column is ascertained. Furthermore, the secondary arc dimension is estimated in this paper, including the arc radius and arc length. A number of formulas have been presented to predict the dimension. The obtained results are more accuracy than previous works.