Power system harmonic detection method, system and medium based on SN interconvolution window

Abstract

The invention discloses a power system harmonic detection method, system and medium based on SN interconvolution window. The implementation steps of the power system harmonic detection method based on SN interconvolution window include: truncating and sampling the original power grid signal to obtain a digital sampling sequence of the power system voltage signal; using the SN interconvolution window to weight the sampling sequence to obtain a new sequence; performing discrete Fourier transform on the new sequence and searching its spectrum peak, combining the spectrum function of the SN interconvolution window to perform double spectrum interpolation on the signal spectrum; using the interpolation result to obtain the frequency, amplitude, phase and other parameter information of each harmonic of the power system; wherein the structure of one system includes a power module, a signal conditioning circuit, a low-pass filter, an analog-to-digital converter and a digital signal processor, and the output end of the signal conditioning circuit is connected through a low-pass filter, an analog-to-digital converter and a digital signal processor. The present invention can realize real-time detection and analysis of power system harmonic signals based on the double spectrum interpolation FFT method of the SN interconvolution window, and has the advantages of good spectrum leakage suppression effect, adaptive adjustment of the main lobe width and the side lobe attenuation ratio, strong adaptability and easy embedded implementation.

Description

Technical Field

[0001] The present invention relates to harmonic detection and analysis technology of power system. Specifically, a dual-spectrum interpolation FFT method based on SN interconvolution window is designed to realize accurate analysis of parameters such as frequency, amplitude, phase and the like of each harmonic of power system. Background Art

[0002] Harmonic detection and analysis is the process of separating different frequency components in distorted signals. With the continuous development of electronic technology and digital signal processing technology, a variety of harmonic analysis methods have emerged, including time domain, frequency domain, and time-frequency domain.

[0003] 1. Time domain analysis method

[0004] For power systems, harmonic time domain analysis mainly refers to the analysis based on various vector transformations ( α - β Transformation, d - q Transformation, symmetrical component transformation, etc.) and instantaneous reactive power theory (IRPT) are used to estimate power parameters. α - β Converts AC power from abc The three-phase coordinate transformation to α , β Two-phase coordinates, obtain frequency-invariant orthogonal AC components. d - q Convert the charge from static abc The three-phase coordinates are transformed into a rotating d , q Two-phase coordinates, two orthogonal DC components are obtained. The three-phase symmetrical positive sequence, negative sequence, and zero sequence components can be decomposed from the three-phase asymmetrical electrical quantities through symmetrical component transformation.

[0005] exist α - β Transformation, d - q Based on the transformation, Akagi H formally proposed IRPT under non-sinusoidal conditions. p and reactive power q Based on the definition of p - q Theory, and then added the definition of instantaneous active current and instantaneous reactive current and other physical quantities. Based on the three-phase instantaneous reactive power theory, to calculate p , q or i α , i βAs a starting point, two methods for harmonic detection of three-phase three-wire circuits can be obtained, namely p - q Fahe i α - i β The method is mainly used for harmonic detection and reactive power compensation in active power filters (APF). p - q Based on the theory, after coordinate transformation, a variety of harmonic detection methods have been developed, such as the harmonic detection method based on stationary reference coordinate transformation, the method based on synchronous reference coordinate transformation, etc.

[0006] 2. Frequency domain analysis method

[0007] (1) Harmonic analysis method based on modern spectrum estimation

[0008] Modern spectrum estimation methods can be roughly divided into two categories: parametric model method and non-parametric model method. Parametric model method includes rational parameter model and special parameter model. Rational parameter model can be expressed by rational system function, which includes autoregressive model, moving average model, autoregressive moving average model, and typical application is fuzzy autoregressive moving average model. Special parameter model assumes that the signal is a linear combination of some exponential signals, also known as exponential model, mainly including Prony method and extended Prony method. Non-parametric model method directly calculates power spectrum from observed data, typical examples include the minimum variance method proposed by Capon, the iterative filtering method proposed by Kay, and the multiple signal classification (MUSIC) method of mutual high-order cumulants. The Pisarenko harmonic decomposition method based on signal feature space decomposition estimates the sampled waveform to obtain the autocorrelation function, and divides the autocorrelation matrix into signal feature subspace and noise subspace by eigenvalue decomposition. By solving a characteristic polynomial and a set of linear algebraic equations, the frequency and amplitude of each harmonic component are obtained. Theoretically, this method can accurately determine the frequency and amplitude of any combination of alternating sinusoidal signals in the power system.

[0009] (2) Harmonic analysis method based on Fourier transform

[0010] The energy center of gravity correction method is based on the amplitude correction of the discrete spectrum three-point convolution, and performs spectrum correction according to the characteristic that the energy center of gravity of the discrete spectrum of various window functions is infinitely close to the origin of the coordinates. The energy center of gravity correction method directly corrects the frequency and amplitude of the multi-segment average power spectrum. The algorithm is simple and the calculation speed is fast. The error caused by the leakage of negative frequency components and adjacent multi-frequency components has little effect on the accuracy. However, its correction accuracy is related to the window function and the truncation length, and it is not suitable for spectrum analysis with too dense spectrum peaks.

[0011] The virtual magnetic potential method transforms the three-phase problem into a single-phase problem, which simplifies the calculation amount and improves the calculation accuracy of the harmonic amount. However, it also has the following defects: 1) Since the zero-sequence components of the three-phase voltages are equal in magnitude and 120° apart in direction, the zero-sequence component of the synthetic magnetic potential is zero, resulting in that all zero-sequence harmonics cannot be directly calculated and must be solved through supplementary zero-sequence harmonic calculations; 2) The basic principle of the virtual magnetic potential method is based on three-phase balance, so large calculation errors will occur in the case of three-phase imbalance.

[0012] The discrete spectrum phase difference correction method is a method that uses the phase difference of discrete spectrum lines to correct the fundamental wave and each harmonic parameter. The phase difference correction method can directly use its phase difference to correct the frequency and initial phase angle without relying on the window function expression. It can be applied to various symmetrical window functions, with small calculation amount and high accuracy. However, it is affected by the performance of the window function and does not consider the time-varying characteristics of the signal and the influence of harmonic leakage. When the spectrum leakage is large, its accuracy decreases.

[0013] The windowed interpolation FFT method multiplies the time domain measured signal with a window function with low sidelobe characteristics, and then performs FFT and other data operations or processing. It is easy to cause spectrum leakage and fence effect due to asynchronous sampling, but choosing a window function with good performance can reduce the error caused by spectrum leakage.

[0014] 3. Time-frequency domain analysis method

[0015] As a localized time-frequency analysis method, the short-time Fourier transform (STFT) divides the measured signal into multiple small time intervals and analyzes each time interval through Fourier transform to determine the frequency existing in the time interval. Although STFT has overcome the defect of the standard FFT that it does not have local analysis capabilities to a certain extent, it also has its own insurmountable defects: 1) The STFT analysis results of the same signal using different window functions are very different, and a specific signal requires a specific window function to achieve better results, so selecting a suitable window function is the key to obtaining accurate results; 2) After the window function is determined, only the position of the window on the phase plane can be changed, but the shape of the window cannot be changed. At this time, the resolution is fixed, and the time resolution and frequency resolution cannot be adjusted according to needs unless the window function is reselected; 3) The discrete form of STFT does not have an orthogonal expansion, making it difficult to implement an efficient algorithm.

[0016] Wavelet Transform (WT) is a time-frequency localization analysis method with multi-resolution characteristics. By introducing variable scale factors and translation factors, WT has an adjustable time-frequency window in signal analysis, which cleverly solves the contradiction between time resolution and frequency domain resolution, and provides a dynamic analysis method under multi-resolution for signal processing. There are still many limitations in the application of WT in parameter quantitative detection: 1) The frequency domain resolution is coarse, and there may be serious frequency aliasing between frequency bands. Even if the scale factor is introduced, the result is still not a true time-frequency spectrum. In addition, wavelet functions of different scales interfere with each other in the frequency domain, and the influence of noise makes it difficult to separate harmonics and interharmonics with close frequencies; 2) It can only qualitatively analyze the signal amplitude or harmonic components, but cannot accurately detect; 3) It is not easy to effectively detect disturbance signals such as voltage surges and dips that are mainly characterized by changes in time domain characteristics; 4) The algorithm is complex and the amount of calculation is large, which is not conducive to real-time calculation. It cannot be implemented in embedded systems at present and is difficult to be applied in practice.

[0017] Stockwell et al. proposed and developed a new time-frequency analysis method, namely S transform, based on the localization ideas of STFT and WT. This transform is equivalent to STFT with normalized Gaussian window or WT with phase correction, and can be derived from these two transforms. The result of S transform reflects the time-frequency characteristics of the signal and can be intuitively represented by graphics, but the resolution of the signal S transform is related to the frequency, and its result is directly related to its Fourier transform spectrum, and the basic wavelet does not have to meet the admissibility conditions. Therefore, compared with time-frequency analysis methods such as continuous WT and STFT, S transform is more suitable for power quality disturbance signal extraction.

[0018] Hilbert spectrum can describe the signal more intuitively than Fourier spectrum, and has the characteristics of high resolution in time and frequency domain, simple calculation and easy implementation. Since Fourier and wavelet transforms rely on the decomposition method of prior function basis, and the empirical mode decomposition law decomposes according to the time scale characteristics of the data itself, this method is more suitable for processing non-stationary and nonlinear data.

[0019] 4. Other power harmonic analysis methods

[0020] The application of neural networks in harmonic detection and analysis is to use the approximation and learning ability of neural networks for any continuous function, and to establish the corresponding harmonic measurement circuit by constructing a special multi-layer feedforward neural network, which mainly involves network construction, sample determination and algorithm selection. The application of neural networks in power system harmonic detection is still in its early stages, with a short research and application time, and the implementation technology needs to be improved. It has not yet been given priority in engineering applications.

[0021] The advantage of the harmonic detection method based on SVM is its small sample processing capability, that is, it can perform statistical learning well even with a small number of samples. The disadvantage of SVM applied to harmonic and interharmonic analysis is that the selection of harmonic order requires prior knowledge of the model, which is not suitable for embedded system implementation and difficult to meet the real-time requirements of harmonic detection in power systems.

[0022] Intelligent optimization algorithms for harmonic detection and analysis, such as Kalman filtering, genetic algorithms, and simulated annealing algorithms, are combined with DFT, STFT, WT, and error theory. Based on the transformation of harmonic currents by DFT, STFT, and WT, they make full use of their respective optimization algorithms to optimize and compare unknown harmonic frequency components with known harmonic components, so as to minimize the mean square error and achieve the purpose of detecting and identifying unknown frequency components. The above methods are applied to power harmonic prediction, harmonic power flow fluctuation analysis, etc., and have certain noise resistance, but poor real-time performance and low accuracy. Summary of the invention

[0023] The technical problem to be solved by the present invention is as follows: In view of the above-mentioned problems in the prior art, a method, system and medium for detecting harmonics in an electric power system based on an SN inter-convolution window are provided. The SN inter-convolution window function constructed by the present invention has excellent sidelobe characteristics and the ability to suppress mutual leakage between harmonic spectra, and can realize fast and accurate analysis of harmonics of any order, and has the advantages of accurate detection, small amount of calculation and easy embedded implementation.

[0024] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0025] A method for detecting harmonics in a power system based on a SN interconvolution window, characterized in that the implementation steps include:

[0026] 1) Obtain the filtered digital voltage signal of the power system;

[0027] 2) The measured voltage digital signal is sent to the digital signal processor (DSP) to complete the weighted processing of the SN cross-convolution window on the digital signal sampling sequence to obtain a new sampling sequence;

[0028] 3) Perform discrete Fourier transform on the new sequence and search for its spectrum peak, combine the spectrum function of the SN interconvolution window, and perform dual spectrum line interpolation on the spectrum of the signal;

[0029] 4) Perform dual-spectrum interpolation fast Fourier transform (FFT) based on SN interconvolution window to obtain the frequency, amplitude, phase and other parameter information of each harmonic in the power system.

[0030] Optionally, before step 1), a step of preprocessing the measured voltage analog signal is also included, and the signal preprocessing includes low-pass filtering and analog-to-digital conversion.

[0031] Optionally, the detailed steps of step 2) include:

[0032] 2.1) The discrete form of the Slepian window function is shown as follows:

[0033]

[0034] In the above formula, , N is the length of the discrete ellipsoidal wave function; h ( m ) is a Toeplitz form matrix.

[0035] The abbreviated form of the above formula is as follows:

[0036]

[0037] In the above formula, For the matrix The characteristic value of For the eigenvalue The associated eigenvector. Arrange in descending order, then the maximum eigenvalue The corresponding eigenvector It is a time-constrained discrete sequence with maximum energy concentration within the frequency band, that is, the Slepian window function. The Slepian window function represented by can be rewritten as of the form, in which k / N is the main lobe cutoff frequency, k is the maximum spectral line value in the main lobe of the Slepian window function.

[0038] 2.2) The Nuttall window function is shown as follows:

[0039]

[0040] In the above formula, M is the number of terms in the window function, n =1, 2, 3, ..., N -1, b m The following constraints should be satisfied:

[0041]

[0042] 2.3) The SN interconvolution window is the result of the interconvolution operation of the Slepian window and the Nuttall window. Its function expression is shown in the following formula:

[0043]

[0044] In the above formula, w SN ( t ) is the SN interconvolution window, w S ( t ) is the Slepian window function, w N ( t ) is the Nuttall window function.

[0045] 2.4) Discretize the continuous SN interconvolution window function, and obtain its discrete function expression as shown in the following formula:

[0046]

[0047] 2.5) Two lengths are Q Discrete Slepian window and Nuttall window are used for discrete convolution operation to obtain a discrete convolution window with a length of 2. Q -1 sequence, and padding with zeros at the end of the sequence to obtain a sequence with a length of 2 Q Discrete SN interconvolution window.

[0048] 2.6) Perform weighted processing on the sampling sequence using the constructed SN inter-convolution window.

[0049] Optionally, the detailed steps of step 3) include:

[0050] 3.1) Weighted New Sequence x ( n ) is shown in the following formula:

[0051]

[0052] In the above formula, k 0= f 0 N / f s Representative frequency f 0 is the position of the discrete spectrum, and N is the data truncation length.

[0053] 3.2) Perform peak search on the discrete Fourier spectrum of the weighted sequence to obtain the peak point k h The spectral line values ​​with the largest and second largest amplitudes nearby are and and .

[0054] 3.3) Assume that the amplitudes of the two largest and second largest spectral lines are y 1=| X ( k h1 )|and y 2=| X ( k h2 )|. Considering 0≤ k h1 - k h2 ≤1, definition, definition , then . Introducing parameters The definition is as follows:

[0055]

[0056] 3.4) The least squares spectrum interpolation polynomial constructed based on the SN mutual convolution window is shown as follows:

[0057]

[0058] Optionally, the detailed steps in step 4) are:

[0059] 4.1) Calculate the parameters by least squares spectral interpolation polynomial After that, h The function expression of subharmonic frequency is as follows:

[0060]

[0061] In the above formula, is the frequency resolution.

[0062] 4.2) Section h The function expression of subharmonic amplitude is as follows:

[0063]

[0064] 4.3) Section h The functional expression of subharmonic phase is shown as follows:

[0065]

[0066] In addition, the present invention also provides a power system harmonic detection system based on SN interconvolution window, comprising:

[0067] A signal input program unit is used to obtain a filtered digital signal of the voltage being measured;

[0068] The SN interconvolution window construction program unit is used to perform weighted operations on the measured voltage digital signal;

[0069] The dual-spectrum interpolation FFT is used to perform the dual-spectrum interpolation FFT method based on the SN interconvolution window to obtain the frequency, amplitude, phase and other parameter information of each harmonic of the power system.

[0070] In addition, the present invention also provides an electric power system harmonic detection system based on an SN inter-convolution window, comprising a digital signal processing device, characterized in that the digital signal processing device is programmed or configured to execute the steps of the electric power system harmonic detection method based on an SN inter-convolution window.

[0071] In addition, the present invention also provides an electric power system harmonic detection system based on an SN inter-convolution window, comprising a digital signal processing device, characterized in that a memory of the digital signal processing device stores an embedded program programmed or configured to execute the electric power system harmonic detection method based on an SN inter-convolution window.

[0072] In addition, the present invention also provides a digital signal readable storage medium, characterized in that the digital signal readable storage medium stores an embedded program programmed or configured to execute the power system harmonic detection method based on the SN interconvolution window.

[0073] In addition, the present invention also provides an electric power system harmonic detection system based on an SN interconvolution window, comprising a power module, a signal conditioning circuit, an analog-to-digital converter and a digital signal processor, wherein the output end of the power module is electrically connected to the signal conditioning circuit, the analog-to-digital converter and the digital signal processor, respectively, and the output end of the signal conditioning circuit is connected through a low-pass filter, the analog-to-digital converter and the digital signal processor, and the digital signal processor is programmed or configured to execute the steps of the electric power system harmonic detection method based on the SN interconvolution window.

[0074] Optionally, the digital signal processor is also connected to a synchronous dynamic random access memory, a flash memory, an active crystal oscillator, a reset module and a simulation debugging interface respectively.

[0075] Compared with the prior art, the present invention has the following advantages: the present invention obtains the measured voltage digital signal after filtering, performs SN interconvolution window weighting on the measured voltage digital signal to obtain a new sampling sequence, performs discrete Fourier transform on the new sequence and searches for its spectrum peak, performs dual-line interpolation on the spectrum of the signal in combination with the spectrum function of the SN interconvolution window, and performs dual-line interpolation FFT based on the SN interconvolution window to obtain parameter information such as frequency, amplitude, phase, etc. of each harmonic of the power system, and can realize real-time detection of parameters such as frequency, amplitude, phase, etc. of each harmonic of the power system based on the dual-line interpolation FFT method of the SN interconvolution window, and has the advantages of accurate detection, small calculation amount, and easy embedded implementation. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 Schematic diagram of the basic flow of the method of the embodiment of the present invention.

[0077] Figure 2 SN is the main lobe characteristic of the cross-convolution window in the embodiment of the present invention.

[0078] Figure 3 SN cross-convolution window sidelobe characteristics in the embodiment of the present invention.

[0079] Figure 4 1 is the simulation result of each harmonic amplitude in the embodiment of the present invention.

[0080] Figure 5 1 is the simulation result of each harmonic phase in the embodiment of the present invention.

[0081] Figure 6 Schematic diagram of the basic structure of the system of the embodiment of the present invention.

[0082] Figure 7 Schematic diagram of the framework structure of the system according to the embodiment of the present invention. DETAILED DESCRIPTION

[0083] like Figure 1 The implementation steps of the power system harmonic detection method based on SN interconvolution window include:

[0084] 1) Obtain the filtered digital voltage signal of the power system;

[0085] 2) The measured voltage digital signal is sent to the digital signal processor (DSP) to complete the weighted processing of the digital signal sampling sequence by the SN cross-convolution window to obtain a new sampling sequence;

[0086] 3) Perform discrete Fourier transform on the new sequence and search for its spectrum peak, combine the spectrum function of the SN interconvolution window, and perform dual spectrum line interpolation on the spectrum of the signal;

[0087] 4) Perform dual-spectrum interpolation fast Fourier transform (FFT) based on SN interconvolution window to obtain the frequency, amplitude, phase and other parameter information of each harmonic in the power system.

[0088] In this embodiment, before step 1), there is also a step of performing signal preprocessing on the measured voltage analog signal, and the signal preprocessing includes low-pass filtering and analog-to-digital conversion.

[0089] In this embodiment, the detailed steps of step 2) include:

[0090] 2.1) The discrete form of the Slepian window function is shown as follows:

[0091]

[0092] In the above formula, , N is the length of the discrete ellipsoidal wave function; h ( m ) is a Toeplitz form matrix.

[0093] The abbreviated form of the above formula is as follows:

[0094]

[0095] In the above formula, For the matrix The characteristic value of For the eigenvalue The associated eigenvector. Arrange in descending order, then the maximum eigenvalue The corresponding eigenvector It is a time-constrained discrete sequence with maximum energy concentration within the frequency band, that is, the Slepian window function. The Slepian window function represented by can be rewritten as of the form, in which k / N is the main lobe cutoff frequency, k is the maximum spectral line value in the main lobe of the Slepian window function.

[0096] 2.2) The Nuttall window function is shown as follows:

[0097]

[0098] In the above formula, M is the number of terms in the window function,n =1, 2, 3, ..., N -1, b m The following constraints should be satisfied:

[0099]

[0100] 2.3) The SN interconvolution window is the result of the interconvolution operation of the Slepian window and the Nuttall window. Its function expression is shown in the following formula:

[0101]

[0102] In the above formula, w SN ( t ) is the SN interconvolution window, w S ( t ) is the Slepian window function, w N ( t ) is the Nuttall window function.

[0103] 2.4) Discretize the continuous SN interconvolution window function, and obtain its discrete function expression as shown in the following formula:

[0104]

[0105] 2.5) Two lengths are Q Discrete Slepian window and Nuttall window are used for discrete convolution operation to obtain a discrete convolution window with a length of 2. Q -1 sequence, and padding with zeros at the end of the sequence to obtain a sequence with a length of 2 Q Discrete SN interconvolution window.

[0106] 2.6) Perform weighted processing on the sampling sequence using the constructed SN inter-convolution window.

[0107] In this embodiment, the detailed steps of step 3) include:

[0108] 3.1) Sequence x ( n )The weighted discrete Fourier spectrum is shown as follows:

[0109]

[0110] In the above formula, k 0= Nf 0 / f s Representative frequency f The position of 0 in the discrete spectrum,N Truncated length for data.

[0111] 3.2) Perform peak search on the discrete Fourier spectrum of the weighted sequence to obtain the peak point k h The spectral line values ​​with the largest and second largest amplitudes nearby are and and .

[0112] 3.3) Assume that the amplitudes of the two largest and second largest spectral lines are y 1=| X ( k h1 )|and y 2=| X ( k h2 )|. Considering 0≤ k h1 - k h2 ≤1, definition, definition , then . Introducing parameters The definition is as follows:

[0113]

[0114] 3.4) The least squares spectrum interpolation polynomial constructed based on the SN mutual convolution window is shown as follows:

[0115]

[0116] In this embodiment, the detailed steps in step 4) are:

[0117] 4.1) Calculate the parameters by least squares spectral interpolation polynomial After that, h The function expression of subharmonic frequency is as follows:

[0118]

[0119] In the above formula, is the frequency resolution.

[0120] 4.2) Section h The function expression of subharmonic amplitude is as follows:

[0121]

[0122] 4.3) Section h The functional expression of subharmonic phase is shown as follows:

[0123]

[0124] The following will further simulate the power system harmonic detection method based on the SN interconvolution window of this embodiment. The simulation experiment is performed on a set of signals containing high-order harmonic components and fundamental frequency changes. The simulation experiment is performed on a set of signals containing high-order harmonic components and fundamental frequency changes. During the simulation process, the fundamental frequency of the signal changes between 49.5 and 50.5 Hz. Sampling frequency f s =6.4kHz, sampling length N =1024, the amplitude of each harmonic , Phase As shown in Table 1. The simulation results are shown in Table 1. Figure 4-5 As shown, Figure 4 is the relative error between the fundamental wave and each harmonic amplitude; Figure 5 The relative error of the fundamental wave and each harmonic amplitude. Figure 4-5 It can be seen that when the signal frequency fluctuates, the SN interconvolution window function is used to perform weighted processing on the signal, and the relative errors of the fundamental wave and each harmonic amplitude and the initial phase change smoothly, which are no more than 7.9×10 -9 % and 4.8×10 -5 %. Therefore, the simulation experiment results show that the SN inter-convolution window of the power system harmonic detection method based on the SN inter-convolution window in this embodiment has excellent sidelobe performance, strong spectrum leakage suppression capability, effectively suppresses mutual interference between harmonics, realizes accurate estimation of signal harmonic parameters and even weak amplitude frequency components, and can effectively overcome the influence of fundamental frequency fluctuation on harmonic parameter estimation.

[0125] Table 1: Amplitude and phase information of each harmonic

[0126]

[0127] In addition, this embodiment also provides a power system harmonic detection system based on SN interconvolution window, including:

[0128] A signal input program unit is used to obtain a filtered digital signal of the voltage being measured;

[0129] The SN interconvolution window construction program unit is used to perform weighted operations on the measured voltage digital signal;

[0130] The dual-spectrum interpolation FFT is used to obtain the frequency, amplitude, phase and other parameter information of each harmonic of the power system based on the SN interconvolution window dual-spectrum interpolation FFT method.

[0131] In addition, this embodiment also provides an electric power system harmonic detection system based on SN inter-convolution window, including a digital signal processing device, characterized in that the digital signal processing device is programmed or configured to execute the steps of the electric power system harmonic detection method based on SN inter-convolution window.

[0132] In addition, this embodiment also provides an electric power system harmonic detection system based on an SN inter-convolution window, comprising a digital signal processing device, characterized in that a memory of the digital signal processing device stores an embedded program programmed or configured to execute the electric power system harmonic detection method based on an SN inter-convolution window.

[0133] In addition, the present invention also provides a digital signal readable storage medium, characterized in that the digital signal readable storage medium stores an embedded program programmed or configured to execute the power system harmonic detection method based on the SN interconvolution window.

[0134] like Figure 6 and Figure 7 As shown, the power system harmonic detection system (power system harmonic analyzer) based on the SN interconvolution window of this embodiment includes a power module 1, a signal conditioning circuit 2, an analog-to-digital converter 3 and a digital signal processor 4, the output end of the power module 1 is electrically connected to the signal conditioning circuit 2, the analog-to-digital converter 3 and the digital signal processor 4 respectively, the output end of the signal conditioning circuit 2 is connected to the digital signal processor 4 through the analog-to-digital converter 3, and the digital signal processor 4 is programmed or configured to execute the steps of the power system harmonic detection method based on the SN interconvolution window of this embodiment. The power system harmonic detection system (power system harmonic analyzer) based on the SN interconvolution window of this embodiment has the advantages of simple structure and reasonable layout.

[0135] In this embodiment, the signal conditioning circuit 2 performs low-pass filtering on the measured voltage signal, and determines the resistance and capacitance values ​​of the analog low-pass filter and the sampling rate and bit number of the analog-to-digital converter according to the time-frequency resolution requirements of the measured signal.

[0136] The analog-to-digital converter 3 is used to perform high-speed analog-to-digital conversion on the measured voltage signal, and convert the measured signal into a digital signal. In this embodiment, the analog-to-digital converter 3 adopts ADS8556 produced by TI.

[0137] The digital signal processor 4 is programmed or configured to execute the steps of the power system harmonic detection method based on the SN interconvolution window described above in this embodiment. In this embodiment, the CPU of the digital signal processor 4 adopts TMS320C6745 produced by TI, and the main parameters are selected as follows: (1) sampling rate: 10kHz; (2) sampling length of Fourier transform data N =1024. Sampling length NIt is the data length for a power system harmonic parameter detection based on SN interconvolution window, and the sampling length N It can be determined based on comprehensive considerations of signal detection accuracy and the operating speed of the computer or embedded system device.

[0138] like Figure 7 As shown, the digital signal processor 4 is also respectively connected to a synchronous dynamic random access memory (SDRAM) 5, a flash memory 6, an active crystal oscillator 7, a reset module 8 and an emulation debugging interface 9. In this embodiment, the emulation debugging interface 9 specifically adopts a JTAG emulation debugging interface, and other types of emulation debugging interfaces may also be adopted as needed.

[0139] In this embodiment, an example of the measurement results of each harmonic obtained by using the power system harmonic detection system (power system harmonic analyzer) based on the SN interconvolution window of this embodiment is shown in Table 2. Among them, Indicates the relative error of each harmonic amplitude; Indicates the relative phase error of each harmonic.

[0140] Table 2: Measurement results of each harmonic

[0141] Harmonic order 1 2 3 4 5 7 9 11 #timg# #timg# #timg# #timg# #timg# #timg# #timg# #timg# #timg# #timg# 0 #timg# #timg# #timg# #timg# #timg# #timg# #timg#

[0142] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for detecting harmonics in a power system based on SN cross-convolution windows, characterized in that Implementation steps include: 1) Obtaining the filtered power system voltage digital signal; 2) The measured voltage digital signal is sent to the digital signal processor DSP to complete the weighted processing of the digital signal sampling sequence by the SN cross-convolution window to obtain a new sampling sequence. The detailed steps include: 2.1) The discrete form of the Slepian window function is shown as follows: In the above formula, n = -N / 2, -N / 2+1, ..., N / 2, N is the length of the discrete ellipsoidal wave function; h(m) is the Toeplitz form matrix; The abbreviated form of the above formula is as follows: Hψ i =λ i (c)ψ i In the above formula, λ i (c) is the eigenvalue of the matrix H; ψ i is the eigenvalue λ i (c) The associated eigenvector; if the eigenvalue λ i (c) are arranged in descending order, then the eigenvector ψ0(n,c) corresponding to the maximum eigenvalue λ0(c) is a time-limited discrete sequence with the maximum energy concentration in the frequency band, that is, the Slepian window function; the Slepian window function represented by ψ0(n,c) can be rewritten as w S (n, k), where k / N is the main lobe cutoff frequency and k is the maximum spectral line value of the Slepian window function in the main lobe; 2.2) The Nuttall window function is shown below: In the above formula, M is the number of terms of the window function, n = 1, 2, 3, ..., N-1, b m The following constraints should be satisfied: 2.3) The SN interconvolution window is the result of the interconvolution operation of the Slepian window and the Nuttall window, and its function expression is shown in the following formula: In the above formula, w SN (t) is the SN interconvolution window, w S (t) is the Slepian window function, w N (t) is the Nuttall window function; 2.4) Discretize the continuous SN interconvolution window function, and obtain its discrete function expression as shown in the following formula: 2.5) Perform discrete convolution operation on two discrete Slepian windows and Nuttall windows of length Q to obtain a sequence of length 2Q-1, and perform zero padding operation at the end of the sequence to obtain a discrete SN inter-convolution window of length 2Q; 2.6) Performing weighted processing on the sampling sequence using the constructed SN interconvolution window; 3) Perform discrete Fourier transform on the new sampling sequence and search for its spectrum peak. Combined with the spectrum function of the SN interconvolution window, perform double spectrum line interpolation on the spectrum of the signal. The detailed steps include: 3.1) The weighted discrete Fourier spectrum of the sequence x(n) is as follows: In the above formula, k0=Nf0 / f s Represents the position of frequency f0 in the discrete spectrum, and N is the data truncation length; 3.2) Perform peak search on the discrete Fourier spectrum of the weighted sequence to obtain the peak point k h The spectral line values ​​with the largest and second largest amplitudes nearby are k h1 and k h2 And k h1 ≤k h ≤k h2 =k h1 +1; 3.3) Let the amplitudes of the two largest and second largest spectral lines be y1 = |X(k h1 )| and y2=|X(k h2 )|, considering 0≤k h1 -k h2 ≤1, definition, definition α=k h -k h1 -0.5, then α∈[-0.5,0.5], and the parameter β is introduced and defined as follows: 3.4) The least squares spectrum interpolation polynomial constructed based on the SN mutual convolution window is shown as follows: α＝2.9549β+0.1768β 3 +0.0898β 5 +0.0568β 7 4) Perform a dual-spectrum interpolation fast Fourier transform FFT based on the SN interconvolution window to obtain the frequency, amplitude, and phase parameter information of each harmonic in the power system. The detailed steps include: 4.1) After calculating the parameter α through the least squares spectrum interpolation polynomial, the function expression of the hth harmonic frequency is as follows: f h =k h Δf=(α+k h1 +0.5)Δf In the above formula, Δf=f s / N is the frequency resolution; 4.2) The function expression of the hth harmonic amplitude is as follows: 4.3) The function expression of the hth harmonic phase is as follows:

2. The power system harmonic detection method based on SN interconvolution window according to claim 1 is characterized in that: Before step 1), the method further includes a step of preprocessing the measured voltage analog signal, wherein the signal preprocessing includes low-pass filtering and analog-to-digital conversion.

3. A power system harmonic detection system based on SN interconvolution window, the system is used to perform the harmonic detection method as claimed in claim 1, characterized in that include: A signal input program unit is used to obtain a filtered digital signal of the voltage being measured; The SN interconvolution window construction program unit is used to perform weighted operations on the measured voltage digital signal; The dual-spectrum interpolation FFT is used to obtain the frequency, amplitude and phase parameter information of each harmonic of the power system by using the dual-spectrum interpolation FFT method based on the SN interconvolution window.

4. A power system harmonic detection system based on SN interconvolution window, comprising a digital signal processing device, characterized in that: The digital signal processing device is programmed or configured to execute the steps of the power system harmonic detection method based on SN interconvolution window as described in any one of claims 1 to 2.

5. A power system harmonic detection system based on SN interconvolution window, comprising a digital signal processing device, characterized in that: The memory of the digital signal processing device stores an embedded program programmed or configured to execute the power system harmonic detection method based on SN interconvolution window as described in any one of claims 1 to 2.

6. A digital signal readable storage medium, characterized in that: The digital signal readable storage medium stores an embedded program programmed or configured to execute the power system harmonic detection method based on SN interconvolution window as described in any one of claims 1 to 2.

7. A power system harmonic detection system based on SN interconvolution window, characterized in that: The invention comprises a power module (1), a signal conditioning circuit (2), an analog-to-digital converter (3) and a digital signal processor (4), wherein the output end of the power module (1) is electrically connected to the signal conditioning circuit (2), the analog-to-digital converter (3) and the digital signal processor (4) respectively, the output end of the signal conditioning circuit (2) is connected to the digital signal processor (4) via a low-pass filter, the analog-to-digital converter (3) and the digital signal processor (4), and the digital signal processor (4) is programmed or configured to execute the steps of the power system harmonic detection method based on the SN interconvolution window as claimed in any one of claims 1 to 2.

8. The power system harmonic detection system based on SN interconvolution window according to claim 7 is characterized in that: The digital signal processor (4) is also respectively connected to a synchronous dynamic random access memory (5), a flash memory (6), an active crystal oscillator (7), a reset module (8) and a simulation debugging interface (9).

Images