A Comparative Study of Empirical and Variational Mode Decomposition on High Voltage Discharges

Abstract

Signal quality is the key issue for maintaining effective power transmission in electrical networks. In most cases, a high voltage (HV) is transmitted in power systems to decrease power loss. Power quality disturbances are monitored by observing the noise degradation of HV signals. Increased oscillations and high-frequency components of power signals exhibit nonstationary signal characteristics. In this study, a comparative analysis of empirical mode decomposition (EMD) and variational mode decomposition (VMD) was conducted on noisy discharge signals. These techniques were used for adaptive signal decomposition in the time domain, facilitating the evaluation of deeper characteristics of the investigated signal. The HV discharges were obtained using 0.4/40 kV and 8 kVA transformers in a laboratory, and all the current and voltage signal waveforms were recorded using high-frequency current and high-voltage probes. The results demonstrate distinct calculations of EMD and VMD techniques in terms of signal decomposition and extracting intrinsic mode functions (IMFs), which define low- and high-frequency components.