Grid impedance estimation using low power signal injection in noisy measurement condition based on wavelet denoising

Abstract

Grid impedance estimation is used in many power system applications such as power quality analysis of smart grids and grid connected renewable energy systems. In this paper a low power and high frequency range signal injected to grid, and the ratio of voltage to current signal is calculated as impedance. Accuracy of impedance estimation in noiseless systems is independent to energy of injected signal. However, in noisy system, energy of injected signal must be sufficient for an accurate approximation. In this paper, two groups of noise (source harmonic voltage noise and measurement noise) are considered. The additive white Gaussian noise is assumed as the measurement noise. In this paper, we proposed to denoising the current and voltage signals using stationary wavelet denoising without increasing the energy of the injected signal. The injection signal is a short duration rectangular pulse. Accurate impedance estimation using signal injection can be done in frequencies that the energy of signal is bigger enough than the energy of noise. The simulation results show that the proposed method can properly estimate grid impedance in a wide frequency range 2 kHz to 150 kHz.