Research Of High Precision Detection Algorithm For Interharmonics Analysis In Power System

As a large number of power electronics, such as non-linear devices connected to the power system, harmonic and interharmonics content increasing in electrical power systems,however interharmonics will not only increase energy consumption, but also produce zero-offset voltage, voltage flicker and a series of new issues, a serious threat to the safe operation of the power system, but also affects the user’s normal use.Accurate the phase, frequency and amplitude of hamonics/interharmonics analysis and measurement and further understanding the mechanism of harmonics in electrical power systems is the premise of the governance of harmonic, to improve the power quality has very important significance.A lot of methods for the harmonic analysis have been developed in recent decades, such as DFT, advanced spectrum estimation and the wavelet transform. Although these mestods have some advantages.Although the traditional Fourier transform has achieved great success in the harmonic detection, but due to the uncertainty and randomness of interharmonics, at the same time by using Fourier transform in the case of asynchronous sampling for harmonic analysis produce spectral leakage and fence effect,modern spectral estimation methods are as Pisarenko, Music, Prony, Esprit etc., spectral estimation method for limited time domain sampling signal has infinite frequency resolution, it can detect two adjacent frequency components, and to overcome the traditional DFT methods shortcomings of low frequency resolution.The basic idea of the signal subspace method, the signal is decomposed into mutually orthogonal signal subspace and noise subspace, and finally using the least squares method of curve fitting.First,this paper introduces the traditional Fourier transform application in harmonic analysis, due to its limitations has a frequency spectrum leakage and fence effect.Respectively introduces the causes of the two errors, then by comparing the window function to choose a more suitable Hanning window.To improve the method of harmonic analysis,given the windowed FFT algorithm and the windowed interpolation FFT algorithm,get a higher harmonic and harmonic detection accuracy, the simulation verify the effectiveness of the proposed method.Second,taking Pisarenko as an example,introduces the modem spectral estimation between harmonic and interharmonics detection algorithm,demonstrates the random initial phase obey uniform distribution of0-27π range,under the additive white noise conditions the generalized stationary of the models of power harmonic and interharmonic.Focuses on the analysis of the spectrum for signal estimation in order the various solutions.Prove that the singular value decomposition and eigenvalue decompositionin the Music and Esprit algorithm in the "order" is essentially equivalent to.Last,as to further improve the detection accuracy of harmonics and interharmonics in power system, this paper deeply studied the Fourier transform method and envelope spectrum estimation method.An envelope Prony-based method for interharmonics analysis in power system is proposed.Interharmonic estimation information enveloping the method focuses on that signals of power system containing.First of all, through the narrow Hilbert filter to extract the envelope.Then, the spectrum for signal envelope, and then according to a specific formula of frequency, amplitude of each interharmonic,at this time can be accurate to solve the "meet the minimum frequency resolution" between harmonic and interharmonic components, and adopts frequency analysis and position between adjacent harmonic and interharmonic in the frequency range;Finally, the time domain filtering residual after component using Prony algorithm calculation, determine the real harmonic and inter harmonic components, eliminating the pseudo spectral. The results showed that the effect of harmonic spectrum, caused by a leak (or pseudo harmonic) can use the proposed method to avoid.MATLAB simulation verifies that the method is more accurate for interharmonics analysis in the real network.