| dc.description.abstract | The determination of whether high voltage (HV) devices suffer from high levels of partial dis charge (PD) has received significant attention to ascertain the safety of neighboring utilities and achieve economic satisfaction. Such HV devices include gas-insulated switchgear (GIS), power transformers, rotating electric machines, and power transmission lines. Albeit such capital assets are rarely damaged due to their high robustness, such devices can still experience significant degradation, primarily due to PD events. Therefore, monitoring high voltage systems against PD has become of paramount significance. This has contributed to developing a plethora of PD-based detection techniques, including acoustics, optical, electromagnetic (using Ultra-high frequency sensors and high-frequency current transformers), and chemical techniques.
The recent advances in ultra-high frequency-based (UHF) techniques have led to the utilization of such techniques in PD detection for many high voltage devices like GIS, transformers, and power cables. This is attributed to the superior immunity and high sensitivity of UHF techniques in detecting, localizing, and classifying different PD defects. UHF detection was predominantly concerned with implementing antennas or sensors to detect PD activities and localizing PD defects based on the time-of-flight (ToF) or the time-difference-of-arrival (TDoA) in GIS and power transformers. The work done hitherto covered either simplified models of disconnecting parts inside GIS enclosures or addressed a single disconnecting part at a time. This work, on the other hand, analyzes EM waves due to multiple disconnecting parts simultaneously.
First, this work utilizes CIGRE sensitivity verification recommendations to overcome the treeing issue associated with partial discharge and properly analyze EM waves inside two different GIS structures. A 145kV L-structured GIS model was implemented in COMSOL multi-physics, and EM wave propagation has been analyzed. Moreover, a 550kV Siemens π model GIS has also been implemented and analyzed. The obtained results have been compared with experimental results to verify the model’s accuracy. It has been shown that GIS systems are very complex structures for propagating electromagnetic waves due to the multiple barriers existing within such devices. Hence, many reflections and signal attenuation are experienced by electromagnetic waves.
Then, a disk-cone (DISCONE) planar class of antennas has been optimized, implemented, and tested against partial discharge. A size-reduction technique, which utilizes the structural symmetry of the antenna, has also been proposed and discussed to reduce the cost of implementation and improve the antenna’s directivity. The experimental results largely agree with the simulation results verifying the modeling accuracy. A maximum gain of 6.25dBi has been achieved using the proposed antenna, and 47% size-reduction has been accomplished without significant performance degradation. After testing the antenna performance, the device has also been tested against PD activities within ceramic insulators and induction machines. The obtained results show that the antenna can be used to obtain different PD signatures, and hence, PD defect classification can be easily performed. | |
