Partial discharge monitoring of power transformers

This thesis is concerned with an approach to the monitoring and measurement of partial discharges within power transformers, including the classification and recognition of partial discharge signals in the both frequency and time domains from bushing tap points via RFCT (Radio Frequency Current Transducer), the propagation characteristics of high frequency signals and real partial discharge signals within a transformer model with a bandwidth of up to 200 MHz, and also the techniques of de-noising, diagnosis and localization of internal partial discharge. The principle proposed based on the fact that an internal partial discharge signal of an oil-filled power transformer may propagate along a winding and its leads to reach a bushing core bar, the bushing acting like a capacitor can transfer the high frequency components of the partial discharge signal across to its tap point, using a sensor the transferred partial discharge signal can be detected. The developed technique allows the partial discharge monitoring and measurement of power transformers to be carried out by using a RFCT as a sensor at taps of transformer bushings. The obtained data was stored, displayed and analysed using digital equipment. This approach was applied to the auto-transformer SGT3A at Northfleet Substation, West London, and was also verified on the transformer model at the High Voltage Laboratory, UMIST. Centred on this principle, the sensitivities of a 60kV transformer bushing, RFCTs, transformer oil, transformer windings and a transformer model with an interleaved disc winding and a plain disc winding were investigated by studying the propagation characteristics of high frequency signals within them. The classification and characteristics of partial discharges within power transformers and out of the 60kV bushing tap were also studied in both time and frequency domains. A major aspect of this work was the experimentally confirmed hypothesis that high frequency components of internal partial discharge signals pass through the winding and bushing with little attenuation and can be detected at the bushing tap point. The feasibility of software techniques required to realize de -noising from measured signals in the field was investigated. A transfer function has been developed for the transformer model based on its frequency response from 0-200 MHz. This linear model has a significantly wider bandwidth than previous published models and greatly improves the prospect of diagnosis and localization of partial discharges within power transformers.