Abstract
The main purpose of this study was to investigate if the application of very low frequencies for PD measurements in type II generator bar insulation is relevant when compared to applying power line frequency at different test temperatures. Laboratory experiments were performed on two stator bar types rated 6.4 kV and 7.4 kV, labeled as Group A and Group B. The stator bars were taken from service and spare part storage (pristine). The samples were preconditioned at the maximum test voltage of 1.5 U 0 before sweeping the AC high voltage supply from 0.1 Hz and 50 Hz. The test temperature was varied between 20 °C and 155 °C. Investigation of cross sections of the generator bars was done to qualitatively describe the voids. The main results showed that the partial discharge inception voltage (PDIV) for Group A was temperature independent at 50 Hz and decreased by 40 % at 0.1 Hz with increased temperatures, whereas the PDIV for Group B was frequency independent and increased by 70 % with temperature. Similarly, the PD magnitude and total apparent charge decreased by 90 % for Group B with increased temperature, whereas it more than doubled for Group A with temperature. The tested bars from Group A showed significantly larger voids, located close to the copper conductors, compared to Group B, where voids were mainly located within the mainwall insulation. Results indicated that tests performed at 0.1 Hz cannot directly represent results obtained at 50 Hz. Using a wide range of frequencies and test temperatures broadens the understanding of the behavior of the voids inside the mainwall insulation.
