| Typical aerospace vehicles (aircraft and spacecraft) experience a wide range of operating pressures during ascending and returning to earth. Compared to the sea-level atmospheric pressure (760 Torr), the pressure at about 60 km altitude is 2 Torr. The performance of the electric power system components of the aerospace vehicles must remain reliable even under such sub-atmospheric operating conditions. It is well known that the dielectric strength of gaseous insulators, while the electrode arrangement remains unchanged, is pressure dependent. Therefore, characterization of the performance and behavior of the electrical insulation in flight vehicles in low-pressure environments is extremely important. Partial discharge testing is one of the practical methods for evaluating the integrity of electrical insulation in aerospace vehicles.; This dissertation describes partial discharge (PD) measurements performed mainly with 60 Hz ac energization in air, argon and helium, for pressures between 2 and 760 Torr. Two main electrode arrangements were used. One was a needle-plane electrode arrangement with a Teflon insulating barrier. The other one was a twisted pair of insulated conductors taken from a standard aircraft wiring harness.; The measurement results are presented in terms of typical PD current pulse waveforms and waveform analysis for both main electrode arrangements. The evaluation criteria are the waveform polarity, magnitude, shape, rise time, and phase angle (temporal location) relative to the source voltage. Two-variable histograms and statistical averages of the PD parameters are presented. The PD physical mechanisms are analyzed.; For PD pattern recognition, both statistical methods (such as discharge parameter dot pattern representation, discharge parameter phase distribution, statistical operator calculations, and PD fingerprint development) and wavelet transform applications are investigated.; The main conclusions of the dissertation include: (1) The PD current pulse waveforms are dependent on the pressure. (2) The rise time of the waveform is another effective PD current pulse characteristic indicator. (3) PD fingerprint patterns that are already available for atmospheric pressure (760 Torr) conditions are inadequate for the evaluation of PD pulses at low pressures. (4) Various wavelet transform techniques can be used effectively for PD pulse signal denoising purposes, and for PD pulse waveform transient feature recognition. |
