| dc.description.abstract | Transmitting power using overhead lines is heavily used by power utilities due to their lower cost compared with underground cables. However, overhead lines have shown great influence on power system due to frequent power interruption caused insulator failure. Power interruption is a critical issue facing utility companies in terms of huge financial loss by both utility, unpaid interruption hours, and customers, especially industrial customers. Insulator failure is the core reason of power interruption; therefore, failure of outdoor insulator should receive great attention and analysis. Overhead line insulators are mainly of two types: ceramic and composite insulators. Composite insulators are now replacing ceramic insulators especially in extra high voltage applications because of their low weight, vandalism resistance, and excellent performance against contamination. However, recent studies have shown that composite insulators suffer from aging phenomena that deteriorates their properties. It was found that aging of composite insulators causes a decrease in surface resistance and consequently, insulator surface would become more prone to contamination accumulation and water filming. The main causes of composite insulator aging are: dry band arcing, partial discharge, UV radiation, chemical reactions, etc. Among these, partial discharge was found to be the most severe factor affecting the aging process. Partial discharge is considered to be the first alarm for composite insulator degradation. Therefore, it is of important to perform online detection of such degradation before the insulator is completely damaged. Partial discharge may be detected using many different techniques such as: acoustical sensors, Radio Frequency (RF) antennas and Radio Frequency Current Transformer (RFCT). Each sensor has advantages and disadvantages depending on the type of application it will be used for. In overhead lines, partial discharges are detected more efficiently using RF sensor due to its high signal to noise ratio and its high sensitivity compared with other sensors. Although partial discharge detection is successfully implemented, it is not the only burden for utilities to stop aging of composite insulators. The most important thing after detecting partial discharge signals is to identify sources of partial discharge, so that proper corrective actions could be implemented to completely stop the aging process. The purpose of this study is not only detecting partial discharges in outdoor insulators, but also to identify sources of partial discharge using artificial intelligence. In outdoor insulators, partial discharges have four main sources: surface discharge, corona discharge from energized end, corona discharge from dead end and combination of both corona and surface discharge happening at the same time. Using feed forward back propagation neural network, a clear distinction between the four classes of partial discharges is achieved with 96.3% recognition rate. | en_US |
