Application of Window Functions to Improve the Accuracy of the Phase-Increment-Based Frequency Measurement Method

Abstract

Accurate frequency measurement is one of the most important tasks in power engineering. Signals in real power grids contain harmonics, noises, flicker and finite frequency deviation. All these factors lead to a signal frequency measurement error. For this reason, frequency measurement methods must be resistant to the effects of aforementioned factors on the accuracy of measurement. This article describes a popular approach to frequency measurement based on the signal phase increment analysis. The main reason of frequency measurement error in this method is the spectral leakage effect. This article proposes a modification of this method based on the use of window functions, which reduce the spectral leakage effect and improve the accuracy of frequency measurement. The application of various window functions to perform this task is considered and the final error is determined. Window functions application is considered in the context of various discrete Fourier transform (DFT) algorithm implementations such as direct implementation of DFT, DFT as discrete filtering, Goertzel's algorithm and the fast Fourier transform. The case of the time-varying number of samples and special aspects of the window functions method implementation in this case are considered. The evaluation of the efficiency of implementation of various window functions in terms of achieving maximum accuracy and complexity of implementation, including the case of time-varying number of samples is performed.