Partial discharge (PD) may have a significant effect on the insulation performance of power apparatus. Insulation performance and properties can be influenced by PD activity from different types of PD sources. Therefore, PD source identification and diagnosis is of interest to both power equipment manufacturers and utilities.
The main aim of this research is to investigate approaches that may be used to facilitate on-line condition monitoring of high voltage assets. The application of machine based learning techniques to partial discharge (PD) discrimination will be researched. The use of support vector machines (SVM) has been assessed as a potential tool for PD source identification. A comprehensive automatic PD identification system has been developed and assessed. The approach has also been applied to PD monitoring of power transformers using an electro-optic modulator based PD data transmission system.
Current work is concentrating on methods of identifying PD signal buried in measurement noise and methods of discriminating between PD signals from multiple discharge sites within an item of power plant.