Research On Harmonic Analysis Method Based On Ratio Method And Improved Incomplete S-Transform

The rapid development of high-speed railways has brought great convenience to our lives and it has also become a major factor affecting the safe operation of the power grid.The mixed operation of AC-DC electric locomotives and AC-DC locomotives causes wide-frequency domain harmonics to become the main power quality problem of the power grid.The International Electrotechnical Commission standard IEC61000-4-7 recommends the use of FFT to detect harmonics.However,due to spectral leakage,the FFT cannot obtain accurate values during non-simultaneous sampling,and there is no time-frequency localization capability.In this paper,we use the ratio correction method and the improved incomplete S-transform method to study harmonic parameter estimation methods under non-simultaneous sampling conditions.The ratio correction method is implemented on the basis of FFT windowing.The center of the main lobe of the window function is where the true frequency is located.The corrected frequency is obtained by correcting the spectrum of the center of the main lobe by the ratio method,and then the correction phase is obtained.Then the window function is used.The main lobe function gives the amplitude estimation formula.Because of the intuitive time-frequency characteristics and strong anti-jamming,the S-transform becomes a hotspot to analyze the power quality.In this paper,by improving the kernel function of incomplete S-transformation,the computational efficiency is improved and the time-frequency resolution is tunable.The real-time accurate estimation of non-stationary harmonic parameters can be achieved.It mainly solves three key problems of incomplete S-transform kernel function:1)Determination of harmonic frequency point:Incomplete S-transform adopts dynamic measurement of power spectrum envelope to detect harmonic frequency points,although it can locate peaks,but Computation is more complicated.When the harmonic frequencies are close and non-simultaneously sampled,it is easy to miss the adjacent harmonic frequency points.This paper extracts the diagonal elements of the TT transform based on the S transform,and finds out that the FFT achieves the exact detection of the frequency points;2)The determination of the optimal window width coefficient of each harmonic frequency point:improper selection of the window width coefficient will result in no The full S-transform eigenvector amplitude is either not flat or the transition time is too long to track signal fluctuations in real time.In this paper,based on the principle of direct amplitude measurement,taking into account the straightness and dynamic effects of the amplitude vector,the optimal window width coefficient of each harmonic frequency point is determined;3)The elimination of the end effect of the incomplete S-transform amplitude vector:Under non-simultaneous sampling conditions,distortion occurs at both ends of the incomplete S-transform modulo vector,which is called the end effect.In this paper,end-effect overlap elimination is used to achieve the end effect.Finally,according to improving the linear phase properties of incomplete S-transforms,frequency and phase estimation expressions are obtained to achieve complete estimation of harmonic parameters.Through simulation analysis,the ratio correction method can accurately estimate the stationary harmonic parameters with an estimated error within 0.1%.The improved incomplete S-transform method has excellent detection performance for both stationary harmonics and non-stationary harmonics.The parameter estimation error is within 0.1%.