Sinusoidal signal frequency estimation method based on interpolation DFT

Abstract

The invention discloses a sinusoidal signal frequency estimation method based on interpolation DFT, being applicable to electric energy quality assessment of a power system. A real sinusoidal signal has a mirrored negative frequency component on spectrum, so that the directly use of DFT generates a relatively large estimation error due to the spectrum leakage of the negative frequency component under an incoherent condition. Aiming at the problem, a two-point IpDFT algorithm based on signal adjacent spectral lines is provided. The method comprises the steps of firstly performing discrete Fourier transformation on a time domain signal x(n), and acquiring an index value corresponding a peak spectral line through searching DFT amplitude spectrum, namely a rough estimated value k0 of frequency; then positioning a secondary spectral line of the amplitude through comparing the left and right spectral line amplitudes of the peak spectral line; and at last, constructing an interpolation proportion by the two maximum DFT spectral lines, computing a deviation [delta]0 of the rough estimated value of the frequency and thus acquiring the complete signal frequency. The method can effectively suppress the spectrum leakage and interference, has the relatively high estimation accuracy and anti-noise performance, and has more significant superiority in the scene with severe spectrum interference.

Description

technical field [0001] The invention relates to the field of signal processing and power system, in particular to a sinusoidal signal frequency estimation method based on interpolation DFT. Background technique [0002] The frequency estimation of sinusoidal signals in the background of noise is not only an important research content in modern digital signal processing, but also has been widely used in many engineering fields such as mobile communication and power quality monitoring with the development of information technology. In the power system, the system frequency is a key parameter to control the unbalanced power generation load in the power grid, and can be used to determine the harmonic components injected by nonlinear loads and other electronic components, etc. For grid signals, only by maintaining the nominal frequency value can the stability of the system and the normal operation of electrical equipment be maintained. Therefore, accurate frequency estimation is...

AI Technical Summary

Problems solved by technology

[0004] DFT estimator has become a widely concerned and expanded method in the field of sinusoidal signal frequency estimation due to its clear physical meaning, simple implementation, and high computational efficiency. However, it is accompanied by spectrum leakage and fence effects in an incoherent sampling environment. Inherent defects will affect the estimation accuracy of its direct use, especially for real sinusoidal signals containing image negative frequency components on the spectrum, additional steps are required to further improve the performance of the DFT frequency estimator

In the DFT-based improved frequency estimation algorithm, although improvements such as windowing and interpolation can improve the accuracy of the DFT frequency estimator to a certain extent, they do not fundamentally consider the source of spectrum leakage, but try to ignore or default to add The window can suppress it to a negligible level, which will lead to a certain error in the estimation results of the algorithm, especially when the positive and negative frequency components are closely spaced, the spectral leakage interference between components will make this type of improvement scheme produce significant estimates deviation

In addition, in order to further improve the estimation accuracy, the iterative DFT algorithm came into being. However, the introduction of iteration will generate additional calculations, resulting in a huge calculation burden, which limits the real-time processing capability of the algorithm.

Images