A new approach to condition monitoring of high voltage overhead lines using neural networks

In view of increasing demand for reliable and high quality power supply, electrical utilities are interested in avoiding transmission failures by identifying problems at an early stage. However, degradation as a result of ageing and operational service conditions is inevitable. The monitoring and location of incipient failures on overhead lines, known as condition monitoring (CM) is thus imperative. This paper describes the design of a novel CM technique for overhead transmission lines using artificial neural networks (ANNs). The technique is based on utilising the high frequency phenomenon associated with the partial discharge (PD) activity emanating from various line conductor defects