Insulator condition analysis for overhead distribution lines using combined wavelet support vector machine (SVM)

Condition analysis of overhead power distribution system insulators using combined support vector machine (SVM) and wavelet multi-resolution analysis (MRA) seems to be promising for distribution system monitoring (DSM) automation to cope with the increasing system complexity. Though system well-being analysis for engineering applications has been used mostly for electric power system reliability studies, the same principle has been extended for assessing the condition of insulators in a distribution system based on the extent of their damage. Video surveillance with fixed cameras provide the required images of power lines along with insulators at regular intervals and same is sent to a control room using remote terminal units (RTUs) for analysis. Not only the health of the insulators, but also the sagging of the lines, breakage of both insulators and lines can be captured with such cameras. This paper mainly focuses on application of wavelet-transform based feature extraction for digital image processing and SVM for subsequent condition analysis of insulators. The most significant contribution of the paper is to compute the condition indices for overhead power distribution line insulators to overcome difficulties related to vehicular applications in video surveillance. The results contained in this paper validate the efficacy of the proposed methodology for wide-scale applications in overhead power distribution system monitoring (DSM) automation.