An original partial discharge generated in oil insulation identification methodology based on simultaneously conducted measurements using electrical method, ultra high frequency method and acoustic emission method is presented in the paper. Three different partial discharge model sources as well as measuring instruments commonly applied for partial discharge detection in electrical power transformers are yielded within a laboratory research. Total of 45 scenarios, including proposed spark gap configurations, selected supply voltage levels and UHF frequencies are analyzed during measurements series. Furthermore, form among total of 93 descriptors assigned for every applied partial discharge model source configuration there are 24 proposed as potentially useful for partial discharge identification applications with their 95% confidence bounds. Attempt of discriminative descriptors selection for partial discharge source analysis in on-site transformer applications as well as a proposal of unique descriptors according to every selected spark gap configuration that could be potentially useful for partial discharge identification purposes are the main purpose of the presented paper. The proposed methodology verification on a real life transformer with particular consideration of the selected descriptors potential utility in the fields of partial discharge detection and identification in electrical power industry applications confirmed a proposed methodology usefulness.
Kunicki, M., Cichoń, A., & Nagi, Ł. (2018). Statistics based method for partial discharge identification in oil paper insulation systems. Electric Power Systems Research, 163, 559–571. https://doi.org/10.1016/j.epsr.2018.01.007