Application of Sliding DFT Filtration to Improve Frequency Measurement Accuracy of Zero Crossing Technique

Abstract

Among the digital methods for frequency measurement in electrical power systems the method based on zero crossings of the input signal (zero crossing technique) is the most popular. This method is easy to implement and characterizes a high measurement accuracy. The main reason for the error is the influence of input noise. One way to improve measurement accuracy is application of a digital filter. This paper is devoted to the application of a sliding discrete Fourier transform (DFT) as an input bandpass filter. This approach allows to effectively suppress noise and non-fundamental spectral components of the signal, which increase the sensitivity of the measurement method to input noise. An approach to the implementation of the sliding DFT (based on the direct implementation of the real part of the DFT) is proposed, which significantly reduces its complexity in the problem being solved. A method is proposed to reduce the spectrum leakage effect, which makes it possible to increase the frequency measurement accuracy. Dependencies are obtained empirically, which make it possible to estimate the frequency measurement error for a given input noise level and parameters of applied filter (sliding DFT). By applying Simulink software package, a simulation model of the considered measurement method and represented modification has been developed. The proposed model makes it possible to estimate the static and dynamic errors of frequency measurement for the input signals of arbitrary form.