Empirical Mode Decomposition-Based DC Measurement for Photovoltaic Power Generation System

Abstract

Due to the advantages such as being clean and safe, photovoltaic (PV) power generation has become an emerging industry of widespread concern in the world. However, the connection of PV inverters and the unknowability of environmental factors cause effective operation status information submerged in interference. To this end, this paper proposes a PV-DC signal de- noising method based on empirical mode decomposition (EMD) combined with wavelet adaptive threshold. The frequency-domain waveform characteristics of different operating conditions are investigated by analyzing theoretical model and fault recording data. The fast Fourier transform (FFT) based on Hanning window is used to extract the frequency-domain features. Extensive test results using simulation data demonstrate the feasibility of the proposed method.