Spectral analysis techniques with Kalman filtering for estimating power quality indices

Abstract

This paper presents the theory, design and implementation of two real time virtual instruments to measure the power quality indices such as current harmonic distortion, voltage harmonic distortion and power factor as given by the IEEE standard. Different windowing techniques are explored in the estimation of the spectra. Cross spectral analysis is used for the measurement of phase angles in voltage and current signal. The virtual instrument integrates digital signal processing (DSP) methods such as windowing techniques and Cross-spectral density estimation with Welch spectral estimation to calculate voltage and current phasors. Moreover a tunable Kalman filter with a forgetting factor is applied to further enhance the results due to noise in the measurement. Simulations followed by a real time experiment using a rectified AC waveform applied across an RL circuit is followed to support the theoretical claims and further to show the robustness of the virtual instruments. Finally the two instruments are compared with each other in terms of performance and speed.