A method and apparatus for detecting inter-harmonics in a power system

Abstract

This invention discloses a method and apparatus for detecting interharmonics in a power system. The method includes: preprocessing discrete sampled signals with full-phase data to obtain full-phase data; performing a fast Fourier transform on the full-phase data to obtain a full-phase amplitude spectrum and a phase spectrum; combining the full-phase amplitude spectrum and phase spectrum to identify the existence of dense spectrum and obtain peak spectral line values; using a full-phase spectral line correction method to obtain interharmonic parameters; obtaining constraint conditions based on the obtained interharmonic parameters; and using a least-squares fitting method with constraint conditions to obtain accurate values ​​of the interharmonic parameters. This invention can effectively determine the existence of dense spectrum, even for dense spectrum with very small frequency intervals; it can effectively suppress spectral leakage, significantly reducing the interference caused by spectral leakage of the fundamental frequency and harmonics to interharmonic detection, and also reducing interference between interharmonics in dense spectrum; this invention also has strong noise resistance.

Description

Technical Field

[0001] This invention belongs to the field of power system technology, specifically relating to a method and apparatus for detecting inter-harmonics in power systems. Background Technology

[0002] With the increasing application of nonlinear devices in power grids, people have higher and higher requirements for the stability and reliability of transmission lines and power systems. However, with the vigorous development of new energy sources such as wind power and photovoltaics in recent years, and the connection of more and more nonlinear loads to the power system, the harmonic and interharmonic problems of the power system have become increasingly serious; a large number of harmonics and interharmonics greatly endanger the safety and stability of the power system.

[0003] Currently, the main methods used for detecting harmonics and interharmonics in power grids include Fast Fourier Transform (FFT), wavelet transform, Independent Component Analysis (ICA), and neural network methods. FFT is one of the most widely used and reliable methods, but it can cause spectral leakage and the picket fence effect in asynchronous sampling. Wavelet transform offers the advantage of high time and frequency domain resolution, but it may cause wavelet aliasing, reducing stability. ICA often fails to completely separate the components and has a high computational cost. Neural network-based methods require a large number of training samples and extremely high computational complexity. Summary of the Invention

[0004] The technical problem to be solved by the embodiments of the present invention is to provide a method and apparatus for detecting interharmonics in a power system, so as to accurately reduce the detection error of interharmonic parameters under conditions of low signal-to-noise ratio and adjacent fundamental / harmonic frequencies.

[0005] To solve the above-mentioned technical problems, the present invention provides a method for detecting inter-harmonics in a power system, comprising:

[0006] Step A: Perform full-phase data preprocessing on the discrete sampled signal to obtain full-phase data;

[0007] Step B: Perform FFT transformation on the full-phase data to obtain the full-phase amplitude spectrum and phase spectrum;

[0008] Step C: Combining the full-phase amplitude spectrum and phase spectrum, identify the existence of dense spectrum and obtain peak spectral line values, and use the full-phase spectral line correction method to obtain interharmonic parameters;

[0009] Step D: Based on the obtained interharmonic parameters, obtain the constraints, and use the least squares fitting method with constraints to obtain the accurate values ​​of the interharmonic parameters.

[0010] Further, step A specifically includes:

[0011] The discrete sequence x(n) (n = 0, ..., N-1) obtained by discrete sampling of the original signal is used to construct an N×N matrix by retaining the first N-1 data of x(0) for weighting:

[0012] x0 = [x(0), x(1), ... x(N-1)]

[0013] x1 = [x(-1), x(0), ... x(N-2)] ......

[0015] x N-1 = [x(-N+1),x(-N+2),...x(0)]

[0016] Circularly shift each vector to the left so that x(0) is in the first position, resulting in a new N×N matrix:

[0017] x0 = [x(0), x(1), ... x(N-1)]

[0018] x1′=[x(0),x(1)…x(-1)] ......

[0020] x N-1 = [x(0), x(-N+1)…x(-1)]

[0021] Add all the N row vectors of the new N×N matrix and average them to obtain the full-phase data vector:

[0022]

[0023] Furthermore, step B specifically includes:

[0024] For the sampled N-point discrete single-frequency complex exponential signal x(n), expressing the angular frequency 2πf0 in terms of β times the frequency interval 2π / N, we get:

[0025]

[0026] in, This is the initial phase of the signal;

[0027] The unwindowed Fourier transform spectrum of the sequence {x(n)} is:

[0028]

[0029] After transformation, we get:

[0030]

[0031] Where k = 0, 1, ..., N-1;

[0032] Based on the shift property of the Discrete Fourier Transform, in step A, x i DFT transform X i (k) and x i ' ... i The relationship of '(k) is:

[0033]

[0034] The full-phase Fourier transform spectrum is obtained from the above:

[0035]

[0036] |X ap (k)| represents the full-phase amplitude spectrum, arg(X) ap (k) is the full-phase phase spectrum.

[0037] Further, step C specifically includes:

[0038] Full-phase FFT transformations were performed on single-frequency signals x1(n), x2(n) and composite signal x(n) respectively to obtain the corresponding amplitude spectrum and phase spectrum;

[0039] The amplitude, phase, and frequency parameters of the interharmonic wave were obtained using the full-phase spectral line correction method.

[0040] Furthermore, step D specifically includes:

[0041] For complex exponential signals containing multiple frequencies

[0042] Determine the number of useful frequencies, denoted as M; let the fitting function be...

[0043]

[0044] The sum of squared errors is

[0045] J = J(a1 a2…a M ,b1b2…b M ,c1 c2…c M ) is the variable a i b i c i The function, according to the principle of least squares, minimizes the sum of squared errors J by taking parameters (a1a2…a ... M ,b1b2…b M ,c1c2…c M That is, the value we are looking for;

[0046] The parameter b in the fitting function δ(t) i Use the calculated frequency value directly, i.e., b i =f0(i), then J = J(a1a2…a M ,c1c2…c M It's just a i and c i The function;

[0047] Set inequality constraints

[0048]

[0049] The fitting function δ(t) is expressed in linear form.

[0050]

[0051] Among them, a hi =a i cos(c i ),b hi =-a i sin(c i Its matrix form is: X = FI

[0052] Among them, X=[δ(t1)δ(t2)……δ(t N )] T It is an N×1 dimensional matrix; I = [a h1 b h1 a h2 b h2 ...a hM b hM ] T It is a 2M×1 dimensional matrix;

[0053]

[0054] It is an N×2M dimensional matrix;

[0055] The linear rule of least squares is used to predict I.

[0056] I = (F T F) -1 F T X

[0057] get:

[0058]

[0059] When the aforementioned constraint condition is met, a i and c i These are the final amplitude and phase values.

[0060] The present invention also provides a power system inter-harmonic detection device, comprising:

[0061] The preprocessing module is used to perform full-phase data preprocessing on the discrete sampled signal to obtain full-phase data;

[0062] The Fourier transform module is used to perform FFT transform on the full-phase data to obtain the full-phase amplitude spectrum and phase spectrum;

[0063] The identification module is used to combine the full-phase amplitude spectrum and the phase spectrum to identify the existence of the dense spectrum and obtain the peak spectral line value, and to obtain the interharmonic parameters using the full-phase spectral line correction method.

[0064] The acquisition module is used to obtain constraints based on the obtained interharmonic parameters, and to obtain the accurate values ​​of the interharmonic parameters using the least squares fitting method with constraints.

[0065] The present invention has the following beneficial effects: the present invention can effectively determine the existence of dense spectrum, even for dense spectrum with very small frequency intervals; the present invention can effectively suppress spectrum leakage, thus greatly reducing the main lobe interference caused by harmonics to interharmonic detection, and also reducing the interference between interharmonics in dense spectrum; the present invention also has strong noise resistance. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0067] Figure 1 This is a flowchart illustrating a method for detecting inter-harmonics in a power system according to an embodiment of the present invention.

[0068] Figure 2a This is a schematic diagram of the amplitude spectrum and phase spectrum of signal x1(n) in an embodiment of the present invention. Figure 2b This is a schematic diagram of the amplitude spectrum and phase spectrum of signal x2(n) in an embodiment of the present invention. Figure 2c This is a schematic diagram of the amplitude spectrum and phase spectrum of signal x(n) in an embodiment of the present invention. Detailed Implementation

[0069] The following description of the embodiments is taken with reference to the accompanying drawings, which illustrate specific embodiments in which the invention can be implemented.

[0070] Traditional FFT struggles to identify dense interharmonic frequencies and exhibits significant errors in detecting interharmonics adjacent to the fundamental / harmonic frequencies. Furthermore, noise interference greatly hinders the detection of low-amplitude interharmonics. Therefore, please refer to... Figure 1 As shown, Embodiment 1 of the present invention provides a method for detecting inter-harmonics in a power system, comprising:

[0071] Step A: Perform full-phase data preprocessing on the discrete sampled signal x(n) to obtain full-phase data x. ap (n);

[0072] Step B, for the full-phase data x ap (n) Perform a fast Fourier transform to obtain the full-phase amplitude spectrum and phase spectrum;

[0073] Step C: Combining the full-phase amplitude spectrum and phase spectrum, identify the existence of dense spectrum and obtain peak spectral line values, and use the full-phase spectral line correction method to obtain interharmonic parameters;

[0074] Step D: Based on the obtained interharmonic parameters, obtain the constraints, and use the least squares fitting method with constraints to obtain the accurate values ​​of the interharmonic parameters.

[0075] Specifically, step A first preprocesses the data to obtain full-phase data, and then performs an FFT transform. Let the discrete sequence obtained by discrete sampling of the original signal be x(n) (n = 0, ..., N-1). Retaining the first N-1 data points of x(0) for weighting, we can construct an N×N matrix as follows:

[0076] x0 = [x(0), x(1), ... x(N-1)]

[0077] x1=[x(-1),x(0),...x(N-2)] (1) ......

[0079] x N-1 = [x(-N+1),x(-N+2),...x(0)]

[0080] Then, by cyclically shifting each vector to the left so that x(0) is in the first position, we can obtain a new N×N matrix:

[0081] x0 = [x(0), x(1), ... x(N-1)]

[0082] x1′=[x(0),x(1)…x(-1)] (2) ......

[0084] x N-1 = [x(0), x(-N+1)…x(-1)]

[0085] The full-phase data vector can be obtained by summing all N row vectors of matrix (2) and taking the average.

[0086]

[0087] A full-phase FFT can be obtained by performing an FFT transformation on a full-phase data vector.

[0088] In step B, for the sampled N-point discrete single-frequency complex exponential signal:

[0089]

[0090] Where f0 is the signal frequency; Assuming the initial phase of the signal, the angular frequency 2πf0 can be expressed as β times the frequency interval 2π / N, resulting in...

[0091]

[0092] Where β can be a decimal; then the unwindowed Fourier transform spectrum of the sequence {x(n)} is:

[0093]

[0094] After transformation, we get:

[0095]

[0096] Where k = 0, 1, ..., N-1.

[0097] Based on the shift property of the Discrete Fourier Transform (DFT), x in equation (1) i DFT transform X i (k) and x in formula (2) i ' ... i The relationship of '(k) is:

[0098]

[0099] Combining the properties of formulas (3), (7), and (8), the full-phase Fourier transform spectrum can be obtained:

[0100]

[0101] |X ap (k)| represents the full-phase amplitude spectrum, arg(X) ap (k) is the full-phase phase spectrum.

[0102] By comparing formulas (7) and (9), the phase values ​​corresponding to each spectral line of the Fourier transform spectrum can be obtained. It is related to the frequency deviation (β-k); while the phase value corresponding to each spectral line of the full-phase Fourier transform spectrum is This is the true initial phase, independent of frequency deviation; this is "phase invariance," which makes the phase detection of all-phase FFT (apFFT) highly accurate because it is not affected by spectral leakage. Furthermore, the spectral values ​​of apFFT are quadratic, resulting in rapid sidelobe decay and a more prominent main lobe, significantly reducing spectral leakage.

[0103] Step C specifically includes the following steps:

[0104] Step C1, dense spectrum identification;

[0105] Let the single-frequency signals be respectively The composite signal is x(n) = x1(n) + x2(n) (n = 0, 1, ..., N-1); where f1 = 50.5 Hz, f2 = 53 Hz, the number of sampling points N = 16, and the sampling frequency f s =80Hz;

[0106] The frequency resolution can then be calculated as Δf = f s / N=5HZ. Since |f1-f2|=2.5HZ<Δf, the composite signal x(n) belongs to the dense spectral distribution.

[0107] The apFFT transform was performed on x1(n), x2(n), and x(n) respectively to obtain the amplitude spectrum and phase spectrum, as shown in Figure 2. Figures 2(a), 2(b), and 2(c) correspond to signals x1(n), x2(n), and x(n), respectively. It can be seen from the figures that the apFFT phase spectrum of the single-frequency signal exhibits obvious "flatness". Figure 2(c) shows the mixed signal x(n) with a peak spectral line value k = 10. The basis for determining whether a dense spectrum exists near the spectral lines 9 and 11 is whether the phase spectrum corresponding to the peak spectral line value has "flatness". As shown in the figure, the phase spectrum at k = 9, 10, and 11 fluctuates greatly, thus indicating the presence of a dense spectrum. However, it is difficult to determine the existence of a dense spectrum by directly observing the amplitude spectrum.

[0108] Step C2: Use the full-phase spectral line correction method to obtain the amplitude, phase, and frequency parameters of the interharmonic wave.

[0109] Let the peak spectral line be k p From formulas (7) and (9), we can obtain

[0110]

[0111] Therefore, there is

[0112]

[0113] Therefore, the spectral line k p The corresponding frequency f0 satisfies

[0114]

[0115] Given the frequency correction value β, the amplitude A0 can be directly calculated using equation (8), and the peak spectral line k p The corresponding spectral line value

[0116]

[0117] therefore

[0118]

[0119] Based on the phase invariance of apFFT, the peak spectral line k can be used directly. p The corresponding phase is used as the initial phase of the signal.

[0120] In step D, for complex exponential signals containing multiple frequencies...

[0121]

[0122] The above steps allow us to determine the number of useful frequencies, which we denot as M. Therefore, let the fitting function be...

[0123]

[0124] Then the sum of squared errors

[0125]

[0126] Therefore, J = J(a1 a2…a M ,b1 b2…b M ,c1 c2…c M ) is the variable a i b i c i The function, according to the principle of least squares, minimizes the sum of squared errors J by taking parameters (a1a2…a ... M ,b1b2…b M ,c1c2…c M The desired value is J. According to multivariate calculus, J can reach its minimum value when the partial derivatives of J with respect to each parameter are 0.

[0127] From formulas (16) and (17), we can obtain formula (18):

[0128]

[0129] Furthermore, this holds true for all k = (1, 2, ..., M), and solving this system of linear equations yields estimates for 3M parameters. However, directly solving a system of 3M equations is extremely complex, thus simplification is necessary. Since the apFFT method already possesses high accuracy for frequency detection, the parameter b in the fitting function δ(t) is... i The calculated frequency value, b, can be used directly. i =f0(i), so J = J(a1a2…a M ,c1c2…c M It's just a i and c i The function. Considering that the relative error of the apFFT method in detecting amplitude and phase under noise will not exceed 5%, inequality constraints are set.

[0130]

[0131] The fitting function δ(t) can be expressed in linear form.

[0132]

[0133] Among them, a hi =a i cos(c i ),b hi =-a i sin(c i It can also be written in matrix form, as in formula (21):

[0134] X = FI (21)

[0135] Among them, X=[δ(t1)δ(t2)……δ(t N )] T It is an N×1 dimensional matrix; I = [a h1 b h1 a h2 b h2 ...a hM b hM ] T It is a 2M×1 dimensional matrix;

[0136]

[0137] It is an N×2M dimensional matrix.

[0138] The linear rule of least squares is used to predict I.

[0139] I = (F T F) -1 FT X (23)

[0140] From formulas (20) and (21), we can obtain

[0141]

[0142] When constraint (19) is satisfied, a i and c i These are the final amplitude and phase values.

[0143] For multi-frequency signals, the mathematical model is very simple, and the model obtained after dense spectrum determination is also consistent with reality. Therefore, under the condition of sufficient sample size and no excessive noise interference, the fitting method can be further improved in terms of constraints.

[0144] First, only apply to a i Set constraints for c i No constraints are set; that is...

[0145]

[0146] Find (a1a2…a) M This is the corrected amplitude, and then c is adjusted separately. i Set constraints and perform fitting.

[0147]

[0148] Find (c1c2…c M This is the corrected phase value.

[0149] The following simulation experiments verify that the present invention can accurately reduce the detection error of interharmonic parameters under conditions of low signal-to-noise ratio and adjacent fundamental / harmonic frequencies.

[0150] The signal model constructed in MATLAB is shown in equation (27):

[0151]

[0152] The specific parameters of each signal component are shown in Table 1, where n = 0, ..., N-1, and the sampling frequency f s =5120Hz. According to the IEC testing standard, the frequency resolution should be 5Hz, therefore N=1024. This invention is used to detect interharmonics. The calculation results and calculation errors are compared with other methods, as shown in Table 2-7.

[0153] Table 1 Parameters of each signal component

[0154]

[0155] First, only constrain a. i ,Right now Find the fitting result a′ i Then, without substituting the result into the equation, we obtain a′. i Only for c i Perform constraint solving, i.e. Find the fitting result c′ i .

[0156] This embodiment relaxes the constraints on phase and amplitude during the fitting process, resulting in higher accuracy. The following tables show the interharmonic parameters estimation and relative errors of the signal using four algorithms without added noise and with a signal-to-noise ratio (SNR) of 20 dB. In Table 6, e-12 is an abbreviation for relative errors of the same order of magnitude. Method 1 is the ordinary full-phase method, Method 2 is the three-spectral-line interpolation method, Method 3 is the spectral-line equation method, and Method 4 is the method described in this embodiment.

[0157] Table 2 Comparison of relative errors in interharmonic frequency detection (without added noise)

[0158]

[0159] Table 3 Comparison of relative errors in interharmonic frequency detection (Snr=20)

[0160]

[0161] Table 4 Comparison of relative errors in interharmonic amplitude detection (without added noise)

[0162]

[0163] Table 5 Comparison of relative errors in interharmonic amplitude detection (Snr=20)

[0164]

[0165] Table 6 Comparison of relative errors in interharmonic phase detection (without added noise)

[0166]

[0167] Table 7 Comparison of relative errors in interharmonic phase detection (Snr=20)

[0168]

[0169] A comparison in Tables 2-7 reveals that, in the absence of noise, the traditional apFFT method (Method 1) demonstrates significantly higher detection accuracy for frequency, amplitude, and phase compared to the three-line interpolation method (Method 2) and the spectral equation system method (Method 3). However, the detection of amplitude and phase is greatly affected by noise. Method 4, by correcting the traditional apFFT method, overcomes the impact of noise on detection accuracy. When the signal-to-noise ratio is 20 dB, compared to Method 1, this embodiment improves the relative error of signal amplitude and phase detection by approximately 1 to 2 orders of magnitude. This indicates that, compared to the traditional FFT spectral method, apFFT improves the detection accuracy of signal parameters by suppressing spectral leakage. Furthermore, it demonstrates that the amplitude and phase corrections performed on the basis of apFFT are effective.

[0170] Corresponding to the inter-power system harmonic detection method of Embodiment 1 of the present invention, Embodiment 2 of the present invention also provides an inter-power system harmonic detection device, comprising:

[0171] The preprocessing module is used to perform full-phase data preprocessing on the discrete sampled signal to obtain full-phase data;

[0172] The Fourier transform module is used to perform FFT transform on the full-phase data to obtain the full-phase amplitude spectrum and phase spectrum;

[0173] The identification module is used to combine the full-phase amplitude spectrum and the phase spectrum to identify the existence of the dense spectrum and obtain the peak spectral line value, and to obtain the interharmonic parameters using the full-phase spectral line correction method.

[0174] The acquisition module is used to obtain constraints based on the obtained interharmonic parameters, and to obtain the accurate values ​​of the interharmonic parameters using the least squares fitting method with constraints.

[0175] Furthermore, step B specifically includes:

[0176] For the sampled N-point discrete single-frequency complex exponential signal x(n), expressing the angular frequency 2πf0 in terms of β times the frequency interval 2π / N, we get:

[0177]

[0178] in, This is the initial phase of the signal;

[0179] The unwindowed Fourier transform spectrum of the sequence {x(n)} is:

[0180]

[0181] After transformation, we get:

[0182]

[0183] Where k = 0, 1, ..., N-1;

[0184] Based on the shift property of the Discrete Fourier Transform, in step A, x i DFT transform X i (k) and x i The DFT transformation X′ i The relationship of (k) is:

[0185]

[0186] The full-phase Fourier transform spectrum is obtained from the above:

[0187]

[0188] |X ap (k)| represents the full-phase amplitude spectrum, arg(X) ap (k) is the full-phase phase spectrum.

[0189] Furthermore, step D specifically includes:

[0190] For complex exponential signals containing multiple frequencies

[0191] Determine the number of useful frequencies, denoted as M; let the fitting function be...

[0192]

[0193] The sum of squared errors is

[0194] J = J(a1 a2…a M ,b1b2…b M ,c1 c2…c M ) is the variable a i b i c i The function, according to the principle of least squares, minimizes the sum of squared errors J by taking parameters (a1a2…a ... M ,b1b2…b M ,c1c2…c M That is, the value we are looking for;

[0195] The parameter b in the fitting function δ(t) i Use the calculated frequency value directly, i.e., b i =f0(i), then J = J(a1a2…a M ,c1c2…c M It's just a i and c i The function;

[0196] Set inequality constraints

[0197]

[0198] The fitting function δ(t) is expressed in linear form.

[0199]

[0200] Among them, a hi =a i cos(c i ),b hi =-a i sin(c i Its matrix form is: X = FI

[0201] Among them, X=[δ(t1)δ(t2)……δ(t N )] T It is an N×1 dimensional matrix; I = [a h1 b h1 a h2 b h2 ...a hM b hM ] T It is a 2M×1 dimensional matrix;

[0202]

[0203] It is an N×2M dimensional matrix;

[0204] The linear rule of least squares is used to predict I.

[0205] I = (F T F) -1 F T X

[0206] get:

[0207]

[0208] When the aforementioned constraint condition is met, a i and c i These are the final amplitude and phase values.

[0209] Furthermore, regarding a i Set constraints for c i No constraints are set; that is...

[0210]

[0211] Find (a1a2…a) M This is the corrected amplitude, and then c is adjusted separately.i Set constraints and perform fitting:

[0212]

[0213] Find (c1c2…c M This is the corrected phase value.

[0214] For the working principle and process of this embodiment, please refer to the description of Embodiment 1 of the present invention, which will not be repeated here.

[0215] As can be seen from the above description, compared with the prior art, the beneficial effects of the present invention are as follows: the present invention can effectively determine the existence of dense spectrum, and has accurate judgment ability even for dense spectrum with very small frequency intervals; the present invention can effectively suppress spectrum leakage, thus greatly reducing the main lobe interference caused by harmonics to interharmonic detection, and also reducing the interference between interharmonics in dense spectrum; the present invention also has strong noise resistance.

[0216] The above description is merely a preferred embodiment of the present invention and should not be construed as limiting the scope of the invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method of inter-harmonic detection for a power system, characterized by, include: Step A: Perform full-phase data preprocessing on the discrete sampled signal to obtain full-phase data; Step B: Perform a fast Fourier transform on the full-phase data to obtain the full-phase amplitude spectrum and phase spectrum; Step C: Combining the full-phase amplitude spectrum and phase spectrum, identify the existence of dense spectrum and obtain peak spectral line values, and use the full-phase spectral line correction method to obtain interharmonic parameters; Step D: Based on the obtained interharmonic parameters, obtain the constraints, and use the least squares fitting method with constraints to obtain the accurate values ​​of the interharmonic parameters. Step A specifically includes: obtaining a discrete sequence by performing discrete sampling on the original signal. ,reserve The front These data points are used for weighting to construct a... Matrix; circularly shift each vector to the left, so that Being in first place, gaining a new Matrix; to put the new Matrix The average value of all row vectors is obtained by adding them together to get the full-phase data vector. In step A: initial The matrix is ​​as follows: New The matrix is ​​as follows: The obtained full-phase data vector is as follows: ； Step C specifically includes: processing the single-frequency signals respectively. , and composite signals Perform a full-phase FFT transform to obtain the corresponding amplitude and phase spectra; use the full-phase spectral line correction method to obtain the amplitude, phase, and frequency parameters of the interharmonics; Step D specifically includes: For complex exponential signals containing multiple frequencies Determine the number of useful frequencies. Let the parameters in the fitting function be... Use the frequency value obtained in step C directly, that is Set amplitude and phase The inequality constraints are determined; the fitted function is expressed in linear form and solved using matrix operations and the least squares method to find the condition that satisfies the constraints. These are the final amplitude and phase values; In step D: The fitting function is: The sum of squared errors is For variables , , The function, according to the principle of least squares, makes the sum of squared errors... Minimum Parameter That is, the value we are looking for; exist hour, Just and The function; The inequality constraints are set as follows: Fitting function It is represented in linear form, as shown below: in, Its matrix form is: ; in, yes 3D matrix; yes 3D matrix; for 3D matrix; Linearity rule of least squares To make a forecast get: When the aforementioned constraints are met. and These are the final amplitude and phase values.

2. The method for detecting inter-harmonics in a power system according to claim 1, characterized in that, Step B specifically includes: For the sampled Discrete single-frequency complex exponential signal angular frequency Frequency interval of The multiple represents the result: in, This is the initial phase of the signal; Then the sequence The unwindowed Fourier transform spectrum is: After transformation, we get: in, ; Based on the shift property of the Discrete Fourier Transform, in step A... DFT transformation and DFT transformation The relationship is: The full-phase Fourier transform spectrum is obtained from the above: It is the full-phase amplitude spectrum. It is a full-phase phase spectrum.

3. A power system inter-harmonic detection device, characterized in that, include: The preprocessing module is used to perform full-phase data preprocessing on the discrete sampled signal to obtain full-phase data; The Fourier transform module is used to perform a fast Fourier transform on the full-phase data to obtain the full-phase amplitude spectrum and phase spectrum; The identification module is used to combine the full-phase amplitude spectrum and the phase spectrum to identify the existence of the dense spectrum and obtain the peak spectral line value, and to obtain the interharmonic parameters using the full-phase spectral line correction method. The acquisition module is used to obtain constraints based on the obtained interharmonic parameters and to obtain the accurate values ​​of the interharmonic parameters using the least squares fitting method with constraints. The power system inter-harmonic detection device is used to perform the power system inter-harmonic detection method as described in any one of claims 1-2.

Images