Signal processing of on-line PDs captured by Rogowski coil for enhanced fault detection in smart distribution networks

On-line partial discharge (PD) measurements can be used as a condition monitoring tool to detect falling/leaning trees on the covered-conductor (CC) overhead distribution networks. PD measurements conducted in the high voltage (HV) laboratory are less affected by electromagnetic disturbances (EMD). However, on the other hand, on-line/on-site PD measurements conducted in smart distribution networks are often affected by several disturbances. Extracting low level PD signal from noisy backgrounds is a major challenge for on-line condition monitoring. In this paper, wavelet transform (WT) technique is proposed as a powerful tool to de-noise on-line PD signals in CC overhead distribution lines, which are completely buried by electromagnetic interference (EMI). The on-line PD signals are captured in the laboratory environment and on-site PD measurements are simulated. The principle of denoising based on multi-resolution signal decomposition (MSD) is implemented. The validation of WT technique is carried out by comparing its effectiveness with other available filtering techniques. Rogowski coil is used as a PD sensor to capture PD signals in this specific application. In the future, the proposed method will be implemented in a real-life environment to get more stable and reliable on-line/on-site PD measurement results for enhanced fault detection in smart distribution networks.