A Deep Learning Scheme for Efficient Power System Waveform De-noising and Reconstruction

In the evolving landscape of modern power systems, the abundance of meters and digital communication technologies has led to an increase in data quality challenges, including noise and packet drops. This research paper introduces a deep learning model, specifically for the purposes of waveform de-noising and the reconstruction of missing packets in waveforms, with a focus on its application in power system power quality assessment. The proposed model’s performance was evaluated by comparing it with three established conventional approaches. Results demonstrate its superior capability in effectively de-noising the power system waveform and reconstructing missing data packets. Consequently, this model presents a valuable solution for enhanced power system monitoring and diagnosis.