In Finland, overhead covered conductors (CC) are commonly used in the medium voltage (MV) networks because the distribution loads in rural areas are not heavy and are widely distributed in the forested terrain. Such parts of the network are exposed to leaning trees and therefore partial discharging into the overhead CC frequently occurs. In this paper, a multi-end correlation-based technique for locating partial discharge (PD) sources on CC lines is presented. The theory of maximum correlation factor is used in order to find the time differences between signal arrivals at three synchronized-measuring points. This helps to solve the problem of varying parameters of the CC line that can change the propagation velocity and the problem of identifying the arrival times of the corresponding PD pulses at different measuring points. The algorithm is developed and tested using a simulated study and experimentally verified using an alternating current (ac) energized system. The Electromagnetic Transient Program-Alternative Transient Program (EMTP-ATP) is used to implement and analyze a PD monitoring system. The Discrete Wavelet Transform (DWT) is used to de-noise PD signals measured from a Rogowski coil utilized as a high frequency transducer. The noise model is added to the simulated-power network signals and then a DWT-based de-noising process is carried out. The proposed algorithm performance is evaluated by extracting the high frequency features using a windowed-standard deviation. The simulated and experimental evaluations prove that accurate PD source location can be achieved using the proposed algorithm technique.
|IEEE Transactions on Dielectrics and Electrical Insulation
|Published - 2012
|MoE publication type
|A1 Journal article-refereed