Charging pile DC signal harmonic detection method and system based on Prony algorithm

By introducing singular value decomposition-overall least squares denoising technology in Prony algorithm, the problems of low accuracy and noise sensitivity in DC signal detection are solved, and high-precision and noise-impaired DC signal harmonic detection is realized, and DC power is improved Accuracy of measurement.

CN119064670APending Publication Date: 2024-12-03STATE GRID ZHEJIANG ELECTRIC POWER CO MARKETING SERVICE CENT
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202411560773.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-04
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing DC signal harmonic detection method based on Prony algorithm has problems such as low accuracy and noise sensitivity, which affects the accuracy of DC power metering.

Method used

The Prony algorithm combined with the singular value decomposition-popular least squares (SVD-TLS) denoising technology is used to resample and singular value decomposition of the collected DC signals, extract harmonic components, and obtain the harmonic amplitude and Frequency information.

Benefits of technology

It improves detection accuracy, enhances noise resistance, effectively realizes the separation of DC signals and harmonics, and improves the accuracy of DC power metering.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119064670A_ABST
    Figure CN119064670A_ABST
Patent Text Reader

Abstract

The invention discloses a charging pile direct current signal harmonic detection method and system based on a Prony algorithm, belongs to the field of direct current metering, and aims to solve the problem of inaccurate metering of an electric energy meter caused by rapid change of charging voltage and current due to a large number of nonlinear loads in an existing electric vehicle charging pile. The charging pile direct current signal harmonic detection method comprises the following steps: collecting a charging pile direct current signal containing noise and harmonic components; selecting a proper sampling frequency, and carrying out resampling processing on the collected direct current signal; carrying out singular value decomposition-total least square denoising on the resampled direct current signal, and extracting a direct current signal containing a harmonic component; and carrying out Prony analysis on the de-noised DC signal containing the harmonic components to obtain amplitude and frequency information of each harmonic component. According to the invention, direct-current signals and harmonic waves can be effectively separated.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of direct current electric energy metering, and relates to a charging pile direct current signal harmonic detection method and system based on a Prony algorithm. Background Art

[0002] Although the research on AC power measurement is relatively mature, the DC power measurement has been slow to make a breakthrough. The DC charging piles supporting electric vehicles contain a large number of nonlinear loads, which cause harmonics in the output voltage and current of the charging piles, causing the charging voltage and current to change rapidly in a short time, seriously affecting the accuracy of DC power measurement. Therefore, it is urgent to propose a method that can accurately detect the frequency and amplitude of harmonics, which has a far-reaching impact on the optimization of DC power measurement algorithms and the further development of DC distribution networks.

[0003] Since the Fourier transform (FFT) was proposed, after decades of development, many methods for harmonic detection have been proposed, such as wavelet transform, Prony algorithm, Hilbert-Huang transform (HHT), empirical mode decomposition (EMD), variational mode decomposition (VMD), empirical wavelet transform (EWT), etc. FFT and its improved algorithms cannot adapt to non-stationary signals, and the data window will limit the accuracy of the algorithm; it is difficult to select the wavelet basis for wavelet transform; HHT, EMD, and EWT have endpoint effects and mode aliasing; the penalty factor and layering parameter selection of VMD will affect the decomposition effect of VMD, and the parameter setting method cannot obtain the optimal parameter combination; the Prony algorithm is sensitive to noise. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the defects of the existing harmonic detection method based on the Prony algorithm, and to provide a charging pile DC signal harmonic detection method and system with high precision and good noise resistance. The Prony algorithm is used to add a denoising link while detecting harmonics to effectively separate the DC signal from the harmonics.

[0005] To this end, the present invention adopts the following technical solution.

[0006] In a first aspect, the present invention provides a charging pile DC signal harmonic detection method based on the Prony algorithm, which comprises: Step 1: Collect the DC signal of the charging pile containing noise and harmonic components; Step 2: Select a suitable sampling frequency and resample the collected DC signal; Step 3, performing singular value decomposition-total least squares method (SVD-TLS) denoising on the resampled DC signal to extract the DC signal containing harmonic components; Step 4: Perform Prony analysis on the DC signal containing harmonic components after denoising to obtain the amplitude and frequency information of each harmonic component.

[0007] Furthermore, the step 2 specifically includes: First, the DC signal is resampled with a high sampling frequency to extract the high-frequency signal component, and the frequency and amplitude sequence of the high-frequency signal is obtained after setting the threshold and filtering out the redundant items; Secondly, a low-pass filter is set to filter out high-frequency signals, and the filtered signal after filtering out the high-frequency signals is sampled at a low sampling frequency. The threshold is set again, and the frequency and amplitude sequence of the low-frequency signal is obtained after filtering out redundant signals. Finally, the signal frequency bands obtained by the two samplings are combined to obtain the frequency and amplitude sequence of the complete signal.

[0008] Furthermore, the step 3 specifically includes: The DC signal of the charging pile collected in step 1 is used to generate a The augmented matrix of : , , In the formula, is the order of the matrix, is the number of sampling points, Pick , is the input signal sequence; It means taking conjugate; Step 3.2, augment the matrix Perform singular value decomposition , and store the singular values ​​and matrices , and the order is determined using the normalized singular value method: , In the formula, and The augmented matrices are The left singular array and the right singular array, is a A diagonal matrix of order; Step 3.3, determine the effective rank of the augmented matrix and construct a dimensional diagonal matrix : , , in, is a matrix No. Row, No. Elements of a column; The effective rank Row, No. Elements of a column; p Represents the order of the matrix; H is the matrix conjugate transpose; Step 3.4, find The inverse matrix , calculate the overall least squares estimates of the unknown parameters: .

[0009] Furthermore, in step 3.2, when there is noise interference during the measurement process, the diagonal matrix The elements in have certain errors. The larger the value of the diagonal element, the smaller the relative error. Set the relevant threshold to filter out the error term: , In the formula, is the threshold for determining the order. When the effective rank is less than When , the noise can be effectively suppressed and the modeling error can be reduced; when i is less than 0.001, then the order is determined to be i -1st level.

[0010] Furthermore, the step 4 specifically includes: Step 4.1, find The root is ,use Generate Vandermonde matrix : , Step 4.2, calculate the amplitude of the harmonic components , Phase ,frequency and attenuation factor : First, the DC signal is represented by a series of attenuated sinusoidal frequency components: , In the n The estimated value of the DC signal at a sampling time point is expressed as: , In the formula, represents a time interval; Assume that the DC signal contains a DC component and i Attenuated harmonic components, use Euler's formula to expand a certain harmonic component: , The discrete form function of the DC signal is expressed as: , Using the value obtained in step 4.1 , the discrete form function of the DC signal is expressed in matrix form:

[0011] In the formula, , , The least squares solution of is: ; Finally, the amplitude, phase, attenuation factor, and frequency of the harmonic components are calculated using the following formula: , In the formula, represents the real part; Represents the imaginary part.

[0012] In a second aspect, the present invention provides a charging pile DC signal harmonic detection system based on the Prony algorithm, which includes: Signal acquisition unit: collects the DC signal of the charging pile containing noise and harmonic components; Resampling processing unit: selects the appropriate sampling frequency and performs resampling processing on the collected DC signal; Component extraction unit: performs singular value decomposition-total least squares denoising on the resampled DC signal to extract the DC signal containing harmonic components; Prony analysis unit: Prony analysis is performed on the DC signal containing harmonic components after denoising to obtain the amplitude and frequency information of each harmonic component.

[0013] The beneficial effects of the present invention are as follows: when detecting harmonics, the present invention adopts the Prony algorithm and simultaneously adds a denoising link, thereby effectively realizing the separation of DC signals and harmonics; the present invention has high detection accuracy and good noise resistance. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0015] Figure 1 This is a flow chart of the method for detecting harmonics in a DC signal of a charging pile according to the present invention; Figure 2This is a structural diagram of the charging pile DC signal harmonic detection system of the present invention; Figure 3 This is a diagram showing the structure of a component extraction unit of a charging pile DC signal harmonic detection system of the present invention; Figure 4 This is a diagram showing the structure of the Prony analysis unit of the charging pile DC signal harmonic detection system of the present invention; Figure 5 A comparison diagram of the relative errors of the amplitudes of DC signals measured by using the traditional Prony algorithm and the improved Prony algorithm (i.e., the Prony algorithm with resampling preprocessing added) in a specific embodiment of the present invention; Figure 6 A comparison diagram of relative frequency errors of DC signals measured by using a traditional Prony algorithm and an improved Prony algorithm (i.e., a Prony algorithm with resampling preprocessing added) in a specific embodiment of the present invention; Figure 7 A comparison diagram of the relative errors of the amplitudes of DC signals measured by using no noise reduction processing and SVD-TLS noise reduction processing in a specific embodiment of the present invention; Figure 8 This is a comparison diagram of relative frequency errors of DC signals measured without noise reduction processing and with SVD-TLS noise reduction processing in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the embodiments described are only part of the embodiments of the present invention, rather than all of the embodiments.

[0017] Example 1 This embodiment is a charging pile DC signal harmonic detection method based on the Prony algorithm. Figure 1 As shown, the steps are as follows: Step 1: Collect the DC signal of the charging pile containing noise and harmonic components; Step 2: Select a suitable sampling frequency and resample the collected DC signal; Step 3, performing singular value decomposition-total least squares denoising on the resampled DC signal to extract the DC signal containing harmonic components; Step 4: Perform Prony analysis on the DC signal containing harmonic components after denoising to obtain the amplitude and frequency information of each harmonic component.

[0018] Specifically, the step 2 specifically includes: The collected DC signal is sampled and processed. First, the DC signal is sampled at a high sampling frequency to extract the high-frequency signal component. Since redundant items may be generated in the process, it is necessary to set a threshold to filter out the redundant items. After filtering out the redundant items, the frequency and amplitude sequence of the high-frequency signal is obtained; Secondly, a low-pass filter is set to filter out high-frequency signals, and the filtered signal after filtering out the high-frequency signals is sampled at a low sampling frequency. The threshold is set again, and the frequency and amplitude sequence of the low-frequency signal is obtained after filtering out redundant signals. Finally, the signal frequency bands obtained by the two samplings are combined to obtain the frequency and amplitude sequence of the complete signal.

[0019] Specifically, the step 3 specifically includes: Step 3.1, generate a DC signal of the charging pile collected in step 1 The augmented matrix of : , , In the formula, is the order of the matrix, is the number of sampling points, Pick , is the input signal sequence; It means taking conjugate; Step 3.2, augment the matrix Perform singular value decomposition , and store the singular values ​​and matrices , and the order is determined using the normalized singular value method: , In the formula, and The augmented matrices are The left singular array and the right singular array, is a A diagonal matrix of order; Step 3.3, determine the effective rank of the augmented matrix and construct a dimensional diagonal matrix : , , in, is a matrix No. Row, No. Elements of a column; The effective rank Row, No. Elements of a column; p Represents the order of the matrix; H is the matrix conjugate transpose; Step 3.4, find The inverse matrix , calculate the overall least squares estimates of the unknown parameters: .

[0020] More specifically, in step 3.2, when there is noise interference during the measurement process, the diagonal matrix The elements in have certain errors. The larger the value of the diagonal element, the smaller the relative error. Set the relevant threshold to filter out the error term: , In the formula, is the threshold for determining the order. When the effective rank is less than When , the noise can be effectively suppressed and the modeling error can be reduced; when i is less than 0.001, then the order is determined to be i -1st level.

[0021] Specifically, the step 4 specifically includes: Step 4.1, find The root is ,use Generate Vandermonde matrix : , Step 4.2, calculate the amplitude of the harmonic components , Phase ,frequency and attenuation factor : First, the DC signal is represented by a series of attenuated sinusoidal frequency components: , In the n The estimated value of the DC signal at a sampling time point is expressed as: , In the formula, represents a time interval; Assume that the DC signal contains a DC component and i Attenuated harmonic components, use Euler's formula to expand a certain harmonic component: , The discrete form function of the DC signal is expressed as: , Using the value obtained in step 4.1 , the discrete form function of the DC signal is expressed in matrix form:

[0022] In the formula, , , The least squares solution of is: ; Finally, the amplitude, phase, attenuation factor, and frequency of the harmonic components are calculated using the following formula: , In the formula, represents the real part; Represents the imaginary part.

[0023] Example 2 This embodiment provides a charging pile DC signal harmonic detection system based on the Prony algorithm. Figure 2 As shown, it includes: Signal acquisition unit: collects the DC signal of the charging pile containing noise and harmonic components; Resampling processing unit: selects the appropriate sampling frequency and performs resampling processing on the collected DC signal; Component extraction unit: performs singular value decomposition-total least squares denoising on the resampled DC signal to extract the DC signal containing harmonic components; Prony analysis unit: Prony analysis is performed on the DC signal containing harmonic components after denoising to obtain the amplitude and frequency information of each harmonic component.

[0024] The specific processing process of the resampling processing unit is as follows: First, the DC signal is sampled at a high sampling frequency to extract the high-frequency signal component, and the frequency and amplitude sequence of the high-frequency signal is obtained after setting the threshold and filtering out the redundant items. Secondly, a low-pass filter is set to filter out high-frequency signals, and the filtered signal after filtering out the high-frequency signals is sampled at a low sampling frequency. The threshold is set again, and the frequency and amplitude sequence of the low-frequency signal is obtained after filtering out redundant signals. Finally, the signal frequency bands obtained by the two samplings are combined to obtain the frequency and amplitude sequence of the complete signal.

[0025] The component extraction unit, such as Figure 3 As shown, specifically including: Augmented matrix generation subunit: The DC signal of the charging pile collected by the signal acquisition unit is generated into a The augmented matrix of : , , In the formula, is the order of the matrix, is the number of sampling points, Pick , is the input signal sequence; It means taking conjugate; Singular value decomposition subunit: augmented matrix Perform singular value decomposition , and store the singular values ​​and matrices , and the order is determined using the normalized singular value method: , In the formula, and The augmented matrices are The left singular array and the right singular array, is a A diagonal matrix of order; Matrix construction subunit: Determine the effective rank of the augmented matrix and construct a dimensional diagonal matrix : , , in, is a matrix No. Row, No. Elements of a column; The effective rank Row, No. Elements of a column; p Represents the order of the matrix; H is the matrix conjugate transpose; Least squares estimate calculation unit: Find The inverse matrix , calculate the overall least squares estimates of the unknown parameters: .

[0026] Specifically, in the matrix construction subunit, when there is noise interference during the measurement process, the diagonal matrix The elements in have certain errors. The larger the value of the diagonal element, the smaller the relative error. Set the relevant threshold to filter out the error term: , In the formula, is the threshold for determining the order. When the effective rank is less than When , it can effectively suppress noise and reduce modeling errors. ; If the above formula i is less than 0.001, then the order is determined to be i -1st level.

[0027] The Prony analysis unit, such as Figure 4 As shown, specifically including: Vandermonde matrix generation subunit: Get The root is ,use Generate Vandermonde matrix : ; Parameter calculation subunit: used to calculate the amplitude of the harmonic component , Phase ,frequency and attenuation factor , the specific steps are as follows: First, the DC signal is represented by a series of attenuated sinusoidal frequency components: , In the n The estimated value of the DC signal at a sampling time point is expressed as: , In the formula, represents a time interval; Assume that the DC signal contains a DC component and i Attenuated harmonic components, use Euler's formula to expand a certain harmonic component: , The discrete form function of the DC signal is expressed as: , Using the value obtained in step 4.1 , the discrete form function of the DC signal is expressed in matrix form:

[0028] In the formula, , , The least squares solution of is: ; Finally, the amplitude, phase, attenuation factor, and frequency of the harmonic components are calculated using the following formula: , In the formula, represents the real part; Represents the imaginary part.

[0029] Application Examples In practical applications, the DC signal in the charging pile often exhibits time-varying properties. In order to further verify and compare the performance of the method described in the present invention, the following simulation experiment is performed. The simulation parameter settings are shown in Table 1 below: Table 1 Signal parameter setting table

[0030] The components of the ideal signal are shown in the table above. It can be seen that the components of the signal are divided into 0.1V 4Hz, 0.15V 8Hz, 1V50Hz, 0.2V 100Hz, 0.1V 200Hz, 0.1V 300Hz, 0.1V 600Hz and 220V DC signals.

[0031] Example 1: The traditional Prony algorithm and the Prony algorithm with resampling preprocessing are used to compare and analyze the ideal DC signal without noise after adding harmonics. The ideal signal is constructed as follows:

[0032] The sampling frequency of the Prony algorithm is set to 1500 Hz and the number of sampling points is 1500.

[0033] The actual amplitude and frequency measurements of the 4Hz and 8Hz signals after resampling the Prony algorithm are shown in the following table.

[0034]

[0035] In the identification of 4Hz and 8Hz signals, Figure 5 and Figure 6 It can be concluded that the improved Prony algorithm has greatly improved the fitting accuracy in both amplitude and frequency compared with the traditional Prony algorithm, and the relative error accuracy of the amplitude has been improved by 10 2 The relative error accuracy of the frequency is also improved by 10 3 above.

[0036] In Example 2, the resampling Prony algorithm and the Prony algorithm with TLS-SVD noise reduction (ie, the present invention) are used to test and analyze the signal of Example 1 after adding 50 dB noise.

[0037] The sampling frequency of the Prony algorithm is set to 1500 Hz and the number of sampling points is 1500.

[0038] The actual amplitude and frequency measurements of the 4Hz and 8Hz signals after adding the TLS-SVD noise reduction algorithm are shown in the following table.

[0039]

[0040] After adding 50dB noise, the recognition of 4Hz and 8Hz signals is done by Figure 7 and Figure 8 It can be concluded that the addition of TLS-SVD denoising algorithm processing (i.e., the present invention) has greatly improved the recognition of amplitude and frequency compared with the non-addition of denoising method processing. The relative error accuracy of amplitude has been improved by more than 10 times, and the relative error accuracy of frequency has also been improved by 20-10. 3 above.

[0041] The embodiments of the present invention are given for the purpose of illustration and description, and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

Claims

1. A charging pile DC signal harmonic detection method based on Prony algorithm, characterized in that: include: Step 1: Collect the DC signal of the charging pile containing noise and harmonic components; Step 2: Select a suitable sampling frequency and resample the collected DC signal; Step 3, performing singular value decomposition-total least squares denoising on the resampled DC signal to extract the DC signal containing harmonic components; Step 4: Perform Prony analysis on the DC signal containing harmonic components after denoising to obtain the amplitude and frequency information of each harmonic component.

2. According to the method for detecting harmonics of a charging pile DC signal based on the Prony algorithm in claim 1, it is characterized in that: The step 2 specifically includes: Firstly, the DC signal is resampled with a high sampling frequency to extract the high-frequency signal component, and the frequency and amplitude sequence of the high-frequency signal is obtained after setting the threshold and filtering out the redundant items. Secondly, a low-pass filter is set to filter out high-frequency signals, and the filtered signal after filtering out the high-frequency signals is sampled at a low sampling frequency. The threshold is set again, and the frequency and amplitude sequence of the low-frequency signal is obtained after filtering out redundant signals. Finally, the signal frequency bands obtained by the two samplings are combined to obtain the frequency and amplitude sequence of the complete signal.

3. According to the method for detecting harmonics of a charging pile DC signal based on the Prony algorithm in claim 1, it is characterized in that: The step 3 specifically includes: Step 3.1, generate a DC signal of the charging pile collected in step 1 The augmented matrix of : , , In the formula, is the order of the matrix, is the number of sampling points, Pick , is the input signal sequence; It means taking conjugate; Step 3.2, augment the matrix Perform singular value decomposition , and store the singular values ​​and matrices , and the order is determined using the normalized singular value method: , In the formula, and The augmented matrices are The left singular array and the right singular array, is a A diagonal matrix of order; Step 3.3, determine the effective rank of the augmented matrix and construct a dimensional diagonal matrix : , , in, is a matrix No. Row, No. Elements of a column; represents the effective rank matrix Line Elements of a column; p Represents the order of the matrix; H is the matrix conjugate transpose; Step 3.4, find The inverse matrix , calculate the overall least squares estimates of the unknown parameters: 。 4. A charging pile DC signal harmonic detection method based on Prony algorithm according to claim 3, characterized in that: In step 3.2, when there is noise interference during the measurement process, the diagonal matrix The elements in have certain errors. The larger the value of the diagonal element, the smaller the relative error. Set the relevant threshold to filter out the error term: , In the formula, is the threshold for determining the order. When the effective rank is less than When , the noise can be effectively suppressed and the modeling error can be reduced; when i is less than 0.001, then the order is determined to be i -1st level.

5. The method for detecting harmonics of a charging pile DC signal based on the Prony algorithm according to claim 3 is characterized in that: The step 4 specifically includes: Step 4.1, find The root is ,use Generate Vandermonde matrix : , Step 4.2, calculate the amplitude of the harmonic components , Phase ,frequency and attenuation factor : First, the DC signal is represented by a series of attenuated sinusoidal frequency components: , In the n The estimated value of the DC signal at a sampling time point is expressed as: , In the formula, represents a time interval; Assume that the DC signal contains a DC component and i Attenuated harmonic components, use Euler's formula to expand a certain harmonic component: , The discrete form function of the DC signal is expressed as: , Using the value obtained in step 4.1 , the discrete form function of the DC signal is expressed in matrix form: , In the formula, , , The least squares solution of is: ; Finally, the amplitude, phase, attenuation factor, and frequency of the harmonic components are calculated using the following formula: , In the formula, represents the real part; Represents the imaginary part.

6. A charging pile DC signal harmonic detection system based on Prony algorithm, characterized in that: include: Signal acquisition unit: collects the DC signal of the charging pile containing noise and harmonic components; Resampling processing unit: selects the appropriate sampling frequency and performs resampling processing on the collected DC signal; Component extraction unit: performs singular value decomposition-total least squares denoising on the resampled DC signal to extract the DC signal containing harmonic components; Prony analysis unit: Prony analysis is performed on the DC signal containing harmonic components after denoising to obtain the amplitude and frequency information of each harmonic component.

7. A charging pile DC signal harmonic detection system based on Prony algorithm according to claim 6, characterized in that: The specific processing process of the resampling processing unit is as follows: First, the DC signal is sampled at a high sampling frequency to extract the high-frequency signal component, and the frequency and amplitude sequence of the high-frequency signal is obtained after setting the threshold and filtering out the redundant items. Secondly, a low-pass filter is set to filter out high-frequency signals, and the filtered signal after filtering out the high-frequency signals is sampled at a low sampling frequency. The threshold is set again, and the frequency and amplitude sequence of the low-frequency signal is obtained after filtering out redundant signals. Finally, the signal frequency bands obtained by the two samplings are combined to obtain the frequency and amplitude sequence of the complete signal.

8. A charging pile DC signal harmonic detection system based on Prony algorithm according to claim 6, characterized in that: The component extraction unit specifically includes: Augmented matrix generation subunit: The DC signal of the charging pile collected by the signal acquisition unit is generated into a The augmented matrix of : , , In the formula, is the order of the matrix, is the number of sampling points, Pick , is the input signal sequence; It means taking conjugate; Singular value decomposition subunit: augmented matrix Perform singular value decomposition , and store the singular values ​​and matrices , and the order is determined using the normalized singular value method: , In the formula, and The augmented matrices are The left singular array and the right singular array, is a A diagonal matrix of order; Matrix construction subunit: Determine the effective rank of the augmented matrix and construct a dimensional diagonal matrix : , , in, is a matrix No. Row, No. Elements of a column; represents the effective rank matrix Line Elements of a column; p Represents the order of the matrix; H is the matrix conjugate transpose; Least squares estimate calculation unit: beg The inverse matrix , calculate the overall least squares estimates of the unknown parameters: 。 9. A charging pile DC signal harmonic detection system based on Prony algorithm according to claim 8, characterized in that: In the matrix construction subunit, when there is noise interference during the measurement process, the diagonal matrix The elements in have certain errors. The larger the value of the diagonal element, the smaller the relative error. Set the relevant threshold to filter out the error term: , In the formula, is the threshold for determining the order. When the effective rank is less than When , the noise can be effectively suppressed and the modeling error can be reduced; when i is less than 0.001, then the order is determined to be i -1st level.

10. A charging pile DC signal harmonic detection system based on Prony algorithm according to claim 8, characterized in that: The Prony analysis unit specifically comprises: Vandermonde matrix generation subunit: Get The root is ,use Generate Vandermonde matrix : ; Parameter calculation subunit: used to calculate the amplitude of the harmonic component , Phase ,frequency and attenuation factor , the specific steps are as follows: First, the DC signal is represented by a series of attenuated sinusoidal frequency components: , In the n The estimated value of the DC signal at a sampling time point is expressed as: , In the formula, represents a time interval; Assume that the DC signal contains a DC component and i Attenuated harmonic components, use Euler's formula to expand a certain harmonic component: , The discrete form function of the DC signal is expressed as: , Using the Vandermonde matrix to generate the subunit, we can obtain , the discrete form function of the DC signal is expressed in matrix form: , In the formula, , , The least squares solution of is: ; Finally, the amplitude, phase, attenuation factor, and frequency of the harmonic components are calculated using the following formula: , In the formula, represents the real part; Represents the imaginary part.

Citation Information

Patent Citations

  • Power system low-frequency oscillation monitoring method adopting improved Prony algorithm

    CN114374210A