A broadband oscillation monitoring method based on waveform data

Abstract

The present invention discloses a broadband oscillation monitoring method based on waveform data, which relates to the field of power systems and includes: performing real-time sampling on the power grid system through a synchronized phasor measurement unit of a wide-area measurement system to obtain electrical signal waveform sequences for each time period; performing parameter identification on the electrical signal waveform sequences for each time period; comparing the amplitudes of harmonics and interharmonics in each time period, and if the amplitude of the harmonic or interharmonic continuously increases, it indicates that the power grid system has oscillated; if the amplitudes of both the harmonic and interharmonic do not continuously increase, it indicates that the power grid system has not oscillated. The present invention does not require the installation of additional equipment and can monitor oscillations in a broadband range only by using the data of the synchronized phasor measurement device, reducing the monitoring cost, solving the problem of online monitoring of oscillation signals in a broadband range, and can adaptively determine the window length of the required data according to the noise intensity to reduce the algorithm calculation amount, which has important significance in engineering.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and in particular to a wide-band oscillation monitoring method based on waveform data. Background Art

[0002] At present, vigorously developing renewable energy represented by wind power and photovoltaics will be an irreversible trend. The proportion of wind power / photovoltaic and other renewable energy generation, electric vehicles / energy storage and other high-power interactive multi-electrified loads connected to the power grid with power electronic equipment as the core technology is increasing, making the grid increasingly complex and the operating state changing. The power system shows a high degree of power electronics from source to network to load. However, the power electronics power system shows the characteristics of low inertia and weak damping, which makes the system prone to oscillation problems. In severe cases, it can cause chain failures. The disturbances that occur may spread to the main grid, endangering the safety and stability of the main grid. Therefore, it is necessary to monitor the oscillations in a wide frequency range.

[0003] Most existing solutions conduct disturbance source location research based on the resonance analysis method of the system model, including frequency scanning method, eigenvalue analysis method, impedance analysis method, etc. This type of method has a solid theoretical basis, but for a large power grid system, since the broadband oscillation of the power system is a complex system problem caused by control interactions of different types of equipment and different time scales, its precise parameters are not only difficult to obtain, but also the electromagnetic transient equivalent model is difficult to construct, and it has significant randomness and strong nonlinearity. Therefore, most of the existing methods are difficult to apply in practice. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a wide-band oscillation monitoring method based on waveform data, which performs wide-band monitoring of the power grid based on measurement data, and solves the problem of how to accurately extract signal parameters from waveform data and then monitor the oscillation signal.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] A broadband oscillation monitoring method based on waveform data comprises the following steps:

[0007] S1. Real-time sampling of the power grid system is performed through the synchronized phasor measurement unit of the wide-area measurement system to obtain the electrical signal waveform sequence of each time period;

[0008] S2. Performing parameter identification on the electrical signal waveform sequence in each time period to identify the harmonics and interharmonics of the electrical signal waveform, as well as the amplitudes of the harmonics and interharmonics in each time period;

[0009] S3. Compare the amplitudes of the harmonics and inter - harmonics in each time period. If the amplitude of the harmonic or inter - harmonic continues to increase, the power grid system has oscillated; if neither the amplitude of the harmonic nor the inter - harmonic continues to increase, the power grid system has not oscillated.

[0010] Further, in the step S2, the method for parameter identification of the electrical signal waveform sequence includes the following steps:

[0011] A1. Estimate the noise intensity of the electrical signal waveform sequence, and set the length of the data window according to the noise intensity;

[0012] A2. Intercept the electrical signal waveform sequence through the data window, construct the electrical signal Hankel matrix, and use the matrix pencil method to solve the rough estimate of the electrical signal frequency according to the electrical signal Hankel matrix;

[0013] A3. Use the Taylor Fourier transform to solve the refined estimate of the electrical signal parameters according to the rough estimate of the electrical signal frequency and the electrical signal waveform sequence intercepted by the data window;

[0014] A4. Identify the refined estimate of the electrical signal parameters to obtain the harmonics and inter - harmonics of the electrical signal waveform, and the amplitudes of the harmonics and inter - harmonics in the corresponding time periods of the electrical signal waveform sequence.

[0015] Further, the step A1 includes the following sub - steps:

[0016] A11. Fit and denoise the electrical signal waveform sequence through a neural network;

[0017] A12. Subtract the denoised electrical signal waveform sequence from the electrical signal waveform sequence before denoising to obtain the noise sequence;

[0018] A13. Calculate the signal - to - noise ratio according to the noise sequence and the denoised electrical signal waveform sequence through the following formula:

[0019]

[0020] where SNR is the signal - to - noise ratio, lg(·) is the logarithmic function with base 10, P N is the noise power solved according to the noise sequence, and P S is the signal power solved according to the denoised electrical signal waveform sequence;

[0021] A14. Set the length of the data window according to the signal - to - noise ratio through the following formula:

[0022]

[0023] where N is the length of the data window, and f s is the sampling frequency of the synchronized phasor measurement unit of the wide - area measurement system.

[0024] Furthermore, the step A2 comprises the following sub-steps:

[0025] A21, intercepting the electrical signal waveform sequence through the data window;

[0026] A22. Construct the electrical signal Hankel matrix based on the electrical signal waveform sequence intercepted by the data window:

[0027]

[0028] Wherein, Hk is the Hankel matrix, x(·) is the electrical signal waveform sampling value in the electrical signal waveform sequence, and L is the tap length of the matrix bundle method;

[0029] A23, copy the electrical signal Hankel matrix and delete its last row to obtain the first transfer matrix;

[0030] A24, copy the electrical signal Hankel matrix and delete its first row to obtain a second transfer matrix;

[0031] A25. According to the first transfer matrix and the second transfer matrix, M eigenvalues ​​are obtained by solving, where M is the total number of signal components of the electrical signal;

[0032] A26. Based on the M eigenvalues, a rough estimate of the frequency of the M signal components of the electrical signal is obtained.

[0033] Furthermore, in step A25, M eigenvalues ​​are obtained by solving the following formula according to the first transfer matrix and the second transfer matrix:

[0034]

[0035] Where λ=[λ 1 ,λ 2 , …λ m , …, λ M - 1 ,λ M ] is the eigenvalue sequence, λ m is the mth eigenvalue, m is an integer in the closed interval [1, M], H 1 is the first transfer matrix, H 2 is the second transfer matrix, is the Hermitian transpose operator, and eigenvalue(·) is the function for obtaining the eigenvalue.

[0036] Furthermore, in step A26, the rough frequency estimates of the M signal components of the electrical signal are obtained by solving the following formula according to the M eigenvalues:

[0037]

[0038] Among them, is the rough estimated frequency of the m-th signal component of the electrical signal, arctan(·) is the arctangent function, and T s is the sampling period of the synchronized phasor measurement unit of the wide area measurement system, Im(·) is the function to obtain the imaginary part, and Re(·) is the function to obtain the real part.

[0039] Furthermore, step A3 includes the following sub-steps:

[0040] A31. Set the order of the Taylor Fourier transform and the total number of iterations, and initialize the current iteration number to 0;

[0041] A32. Construct the Taylor Fourier basis phasor according to the rough estimated frequencies of the M signal components of the electrical signal:

[0042] B = [B 1 , B 2 , … B m , …, B M-1 , B M

[0043]

[0044]

[0045] Among them, B is the Taylor Fourier basis phasor, and B m is the m-th element of the Taylor Fourier basis phasor, n is the sampling result serial number of the synchronized phasor measurement unit of the wide area measurement system, the length of the sequence [-n, -(n - 1), …, 0, …, n - 1, n] is N, ω m is the angular frequency, p is the order of the Taylor Fourier transform, and T is the matrix transpose operator;

[0046] A33. Solve the dynamic phasor according to the Taylor Fourier basis phasor and the electrical signal waveform sequence after data window truncation;

[0047] A34. According to the solved dynamic phasor, solve the refined estimated values of the electrical signal parameters through the following formulas:

[0048]

[0049]

[0050]

[0051] Among them, is the refined estimated amplitude of the m-th signal component of the electrical signal, is the refined estimated phase of the m-th signal component of the electrical signal, is the refined frequency estimate of the m-th signal component of the electrical signal, is the zero-order derivative of the m-th element of the dynamic phasor, is the first-order derivative of the m-th element of the dynamic phasor, abs(·) is the function for obtaining the amplitude, angle(·) is the function for obtaining the phase angle, and j is the imaginary identifier;

[0052] A35. Increase the current iteration count by 1;

[0053] A36. Determine whether the current iteration count is greater than the total number of iterations. If so, end; if not, refresh the rough frequency estimates of all signal components of the electrical signal with the refined frequency estimates of all signal components of the electrical signal, and jump to step A32.

[0054] Further, in step A33, according to the Taylor Fourier basis phasor and the electrical signal waveform sequence after data window truncation, the least squares method is used to solve the dynamic phasor through the following formula:

[0055]

[0056] X = [x(-n), x(-(n - 1)), …, x(0), …, x(n - 1), x(n)] T

[0057] where, is the dynamic phasor, and X is the electrical signal waveform sequence after data window truncation.

[0058] The beneficial effects of the present invention are as follows:

[0059] 1) The present invention does not require the installation of additional equipment. It only needs to utilize the waveform recording function of the synchronized phasor measurement device to achieve the monitoring of oscillations in a wide frequency range, reducing the monitoring cost; at the same time, it does not require modeling of the equipment or the power grid system. It only needs to detect the current or voltage data generated by the system to monitor the wide-frequency oscillation signal; the present invention solves the problem of online monitoring of oscillation signals in a wide frequency range, and can adaptively determine the window length of the required data according to the noise intensity to reduce the algorithm calculation amount, which is of great significance in engineering.

[0060] 2) Compared with other methods, such as methods based on Fourier transform, wavelet analysis, Prony, etc., the present invention has excellent dynamic performance and higher monitoring accuracy.

[0061] 3) The amount of data required by the present invention can be adaptively adjusted according to the actual working conditions, and the calculation amount is small, which is suitable for practical engineering applications. Description of the Drawings

[0062] Figure 1 is a flowchart of a wide-frequency oscillation monitoring method based on waveform data provided by an embodiment of the present invention;

[0063] Figure 2 This is the synthetic signal waveform with 30 dB noise added in Experiment 1 of the embodiment of the present invention;

[0064] Figure 3 This is the overall vector error percentage graph of each component of the synthetic signal in the embodiment of the present invention under different noise intensities;

[0065] Figure 4 This is the frequency error graph of each component of the synthetic signal in the embodiment of the present invention under different noise intensities;

[0066] Figure 5 This is the waveform graph of the subsynchronous oscillation accident current in Guyuan, Hebei in Experiment 2 of the embodiment of the present invention;

[0067] Figure 6 This is the frequency dynamic change graph of the subsynchronous component and the fundamental wave component identified from the current waveform in Experiment 2 of the embodiment of the present invention;

[0068] Figure 7 This is the amplitude dynamic change graph of the subsynchronous component and the fundamental wave component identified from the current waveform in Experiment 2 of the embodiment of the present invention. Detailed implementation manners

[0069] The following describes the detailed implementation manners of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the detailed implementation manners. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are within the scope of protection.

[0070] As Figure 1 shown, in an embodiment of the present invention, a wide-frequency oscillation monitoring method based on waveform data includes the following steps:

[0071] S1. Real-time sampling of the power grid system is performed through the synchronized phasor measurement unit of the wide-area measurement system to obtain the electrical signal waveform sequences of each time period.

[0072] S2. Parameter identification is performed on the electrical signal waveform sequences of each time period to identify the harmonics and inter-harmonics of the electrical signal waveform, as well as the amplitudes of the harmonics and inter-harmonics in each time period.

[0073] S3. Compare the amplitudes of the harmonics and inter-harmonics in each time period. If the amplitude of the harmonic or inter-harmonic continuously increases, the power grid system has an oscillation; if the amplitudes of both the harmonics and inter-harmonics do not continuously increase, the power grid system does not have an oscillation.

[0074] In step S2, the method for parameter identification of the electrical signal waveform sequence includes the following steps:

[0075] A1. Estimate the noise intensity of the electrical signal waveform sequence, and set the length of the data window according to the noise intensity.

[0076] Step A1 includes the following sub-steps:

[0077] A11. Fit and denoise the electrical signal waveform sequence through a neural network;

[0078] A12. Subtract the denoised electrical signal waveform sequence from the electrical signal waveform sequence before denoising to obtain a noise sequence;

[0079] A13. Calculate the signal-to-noise ratio according to the noise sequence and the denoised electrical signal waveform sequence through the following formula:

[0080]

[0081] where SNR is the signal-to-noise ratio, lg(·) is the logarithmic function with base 10, P N is the noise power obtained by solving according to the noise sequence, and P S is the signal power obtained by solving according to the denoised electrical signal waveform sequence;

[0082] A14. Set the length of the data window according to the signal-to-noise ratio through the following formula:

[0083]

[0084] where N is the length of the data window, and f s is the sampling frequency of the synchronized phasor measurement unit of the wide area measurement system.

[0085] A2. Intercept the electrical signal waveform sequence through the data window, construct an electrical signal Hankel matrix, and use the matrix pencil method to solve for the rough estimate of the electrical signal frequency according to the electrical signal Hankel matrix.

[0086] Step A2 includes the following sub-steps:

[0087] A21. Intercept the electrical signal waveform sequence through the data window;

[0088] A22. Construct an electrical signal Hankel matrix according to the electrical signal waveform sequence intercepted by the data window:

[0089]

[0090] where Hk is the Hankel matrix, x(·) is the sampling value of the electrical signal waveform in the electrical signal waveform sequence, and L is the tap length of the matrix pencil method;

[0091] A23. Copy the Hankel matrix of the electrical signal and delete its last row to obtain the first transfer matrix;

[0092] A24. Copy the Hankel matrix of the electrical signal and delete its first row to obtain the second transfer matrix;

[0093] A25. According to the following formula, solve for M eigenvalues based on the first transfer matrix and the second transfer matrix, where M is the total number of signal components of the electrical signal:

[0094]

[0095] where λ = [λ 1 , λ 2 , … λ m , …, λ M-1 , λ M is the eigenvalue sequence, λ m is the m-th eigenvalue, m is an integer in the closed interval [1, M], H 1 is the first transfer matrix, H 2 is the second transfer matrix, is the Hermitian transpose operator, and eigenvalue(·) is the function for obtaining eigenvalues.

[0096] A26. According to the following formula, solve for the rough frequency estimates of the M signal components of the electrical signal based on the M eigenvalues:

[0097]

[0098] where is the rough frequency estimate of the m-th signal component of the electrical signal, arctan(·) is the arctangent function, T s is the sampling period of the synchronized phasor measurement unit of the wide area measurement system, Im(·) is the function for obtaining the imaginary part, and Re(·) is the function for obtaining the real part.

[0099] A3. Use the Taylor Fourier transform to solve for the refined estimates of the electrical signal parameters based on the rough frequency estimates of the electrical signal and the waveform sequence of the electrical signal after data window truncation.

[0100] Step A3 includes the following sub-steps:

[0101] A31. Set the order and total number of iterations of the Taylor Fourier transform, and initialize the current iteration number to 0;

[0102] A32. Construct the Taylor Fourier basis phasors according to the rough frequency estimates of the M signal components of the electrical signal:

[0103] B = [B 1 , B 2 , … Bm , …, B M-1 , B M

[0104]

[0105]

[0106] where B is the Taylor - Fourier - basis phasor, and B m is the m - th element of the Taylor - Fourier - basis phasor, n is the sampling result serial number of the synchronized phasor measurement unit of the wide - area measurement system, the length of the sequence [-n, -(n - 1), …, 0, …, n - 1, n] is N, ω m is the angular frequency, p is the order of the Taylor - Fourier transform, and T is the matrix transpose operator;

[0107] A33. According to the Taylor - Fourier - basis phasor and the electrical signal waveform sequence after data window truncation, using the least - squares method, solve the dynamic phasor through the following formula:

[0108]

[0109] X = [x(-n), x(-(n - 1)), …, x(0), …, x(n - 1), x(n)] T

[0110] where, is the dynamic phasor, and X is the electrical signal waveform sequence after data window truncation.

[0111] A34. According to the solved dynamic phasor, solve the refined estimated values of the electrical signal parameters through the following formulas:

[0112]

[0113]

[0114]

[0115] where, is the refined estimated value of the amplitude of the m - th signal component of the electrical signal, is the refined estimated value of the phase of the m - th signal component of the electrical signal, is the refined estimated value of the frequency of the m - th signal component of the electrical signal, is the zero - order derivative of the m - th element of the dynamic phasor, is the first - order derivative of the m - th element of the dynamic phasor, abs(·) is the function for obtaining the amplitude value, angle(·) is the function for obtaining the phase angle, and j is the imaginary identifier;

[0116] A35. Increase the current iteration count by 1; ​

[0117] A36. Determine whether the current iteration count is greater than the total iteration count. If so, end; if not, refresh the rough frequency estimates of all signal components of the electrical signal with the refined frequency estimates of all signal components of the electrical signal, and jump to step A32.

[0118] A4. Identify the refined estimates of the electrical signal parameters to obtain the harmonics and inter - harmonics of the electrical signal waveform, as well as the amplitudes of the harmonics and inter - harmonics in the corresponding time periods of the electrical signal waveform sequence.

[0119] Currents or voltages that are sinusoidal and whose frequencies are integer multiples of the fundamental wave are called harmonics, and those that are non - integer multiples are collectively called inter - harmonics. Based on this concept, the attribution relationships of each signal component of the electrical signal with harmonics and inter - harmonics can be identified, and the amplitudes of the harmonics and inter - harmonics in the corresponding time periods of the electrical signal waveform sequence can be confirmed.

[0120] To verify the effects of the embodiments of the present invention, two rounds of experiments were designed for verification.

[0121] Experiment 1:

[0122] Perform simulation verification with a synthetic signal:

[0123] The synthetic signal is as follows:

[0124] y = A 1 cos(2πf 1 t + φ 1 (t)) + A 2 coS(2πf 2 t + φ 2 (t)) + A 3 cos(2πf 3 t + φ 3 (t)) + A 4 coS(2πf 4 t + φ 4 (t)) + ω(t)

[0125] Where ω(t) is Gaussian white noise, and the other parameters are shown in Table 1.

[0126] Figure 2 is a waveform diagram with 30dB Gaussian white noise added, and the sampling frequency is 12800Hz.

[0127] Collect data for two power - frequency cycles (512 points) and use a neural network for noise estimation. The noise estimation result is approximately 30dB. Therefore, the data length is about 4 power - frequency cycles (1024 points). If the sampling rate is low, the window length can be appropriately extended according to the actual situation.

[0128] Table 1 Signal Parameters

[0129]

[0130] In this experiment, the order of the Taylor Fourier transform is set to 2, and the total number of iterations is 3. Figure 4 And Figure 5 are the percentage of the Total Vector Error (TVE) and the Frequency Error (FE) of the synthesized signal from 15 dB to 45 dB. The total vector error and the frequency error are defined as follows:

[0131]

[0132] FE = |f true - f measured |

[0133] In the formula, and V r (n) are the real parts of the vector estimated value and the vector true value respectively, and V i (n) are the imaginary parts of the vector estimated value and the vector true value respectively, f measured and f true are the estimated value and the true value of the frequency respectively.

[0134] From Figure 3 and Figure 4 it can be seen that except for the case of extremely low signal-to-noise ratio, the percentage of the total vector error will exceed 5%, and the estimation effect in other cases is relatively accurate.

[0135] Experiment 2:

[0136] Now, the subsynchronous oscillation accident that actually occurred in Guyuan, Hebei is used for verification. The current data in the accident is selected for parameter identification. The waveform is as Figure 5 shown, and the dynamic change processes of the identified frequency and amplitude are as Figure 6 and Figure 7 shown. According to the dynamic change process, the subsynchronous component can be judged as an oscillating signal.

[0137] To sum up, the present invention does not require the installation of additional equipment. It only needs to utilize the waveform recording function of the synchronized phasor measurement device to realize the monitoring of oscillations in a wide frequency range, reducing the monitoring cost; at the same time, it also does not require modeling of the equipment or the power grid system. It only needs to detect the current or voltage data generated by the system to monitor the wide-frequency oscillation signal; the present invention solves the problem of online monitoring of oscillation signals in a wide frequency range, and can adaptively determine the window length of the required data according to the noise intensity to reduce the algorithm calculation amount, which has important significance in engineering.

[0138] Compared with other methods, such as those based on Fourier transform, wavelet analysis, Prony, etc., the present invention has excellent dynamic performance and higher monitoring accuracy.

[0139] The amount of data required by the present invention can be adaptively adjusted according to the actual working conditions, and the calculation amount is small, which is suitable for practical engineering applications.

[0140] Specific embodiments are applied in the present invention to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0141] Those of ordinary skill in the art will realize that the embodiments described herein are for helping the reader understand the principle of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not depart from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. A broadband oscillation monitoring method based on waveform data, characterized in that, it includes the following steps: S1. Real-time sampling of the power grid system is carried out through the synchronized phasor measurement unit of the wide area measurement system to obtain the electrical signal waveform sequences in each period; S2. Parameter identification is carried out on the electrical signal waveform sequences in each period to identify the harmonics and inter-harmonics of the electrical signal waveform, and the amplitudes of the harmonics and inter-harmonics in each period; S3. Compare the amplitudes of the harmonics and inter-harmonics in each period. If the amplitude of a harmonic or inter-harmonic continuously increases, the power grid system has an oscillation; if the amplitudes of both the harmonics and inter-harmonics do not continuously increase, the power grid system does not have an oscillation; The method for parameter identification of the electrical signal waveform sequence includes the following steps: A1. Estimate the noise intensity of the electrical signal waveform sequence, and set the length of the data window according to the noise intensity; A2. Intercept the electrical signal waveform sequence through the data window, construct the electrical signal Hankel matrix, and use the matrix pencil method to solve the rough estimate of the electrical signal frequency according to the electrical signal Hankel matrix; A3. Use the Taylor Fourier transform to solve the refined estimate of the electrical signal parameters according to the rough estimate of the electrical signal frequency and the electrical signal waveform sequence after being intercepted by the data window; A4. Identify the refined estimate of the electrical signal parameters to obtain the harmonics and inter-harmonics of the electrical signal waveform, and the amplitudes of the harmonics and inter-harmonics in the corresponding periods of the electrical signal waveform sequence.

2. The broadband oscillation monitoring method based on waveform data according to claim 1, characterized in that, the step A1 includes the following sub-steps: A11. Fit and denoise the electrical signal waveform sequence through a neural network; A12. Subtract the denoised electrical signal waveform sequence from the electrical signal waveform sequence before denoising to obtain a noise sequence; A13. Calculate the signal-to-noise ratio according to the noise sequence and the denoised electrical signal waveform sequence through the following formula: where SNR is the signal-to-noise ratio, lg(·) is the logarithmic function with base 10, P N is the noise power solved according to the noise sequence, and P S is the signal power solved according to the waveform sequence of the denoised electrical signal; A14. Set the length of the data window according to the signal-to-noise ratio through the following formula: where N is the length of the data window, and f s is the sampling frequency of the synchronized phasor measurement unit of the wide area measurement system.

3. The broadband oscillation monitoring method based on waveform data according to claim 2, characterized in that, the step A2 includes the following sub-steps: A21. Intercept the electrical signal waveform sequence through the data window; A22. Construct the electrical signal Hankel matrix according to the electrical signal waveform sequence after being intercepted by the data window: where Hk is the Hankel matrix, x(·) is the electrical signal waveform sampling value in the electrical signal waveform sequence, and L is the tap length of the matrix pencil method; A23. Copy the electrical signal Hankel matrix and delete its last row to obtain the first transfer matrix; A24. Copy the electrical signal Hankel matrix and delete its first row to obtain the second transfer matrix; A25. Solve for M eigenvalues according to the first transfer matrix and the second transfer matrix, where M is the total number of signal components of the electrical signal; A26. Solve for the rough estimate of the frequencies of the M signal components of the electrical signal according to the M eigenvalues.

4. The broadband oscillation monitoring method based on waveform data according to claim 3, characterized in that, the step A25 solves for M eigenvalues according to the first transfer matrix and the second transfer matrix through the following formula: where λ = [λ 1 , λ 2 , … λ m , …, λ M-1 , λ M is the eigenvalue sequence, λ m is the m-th eigenvalue, m is an integer within the closed interval [1, M], H 1 is the first transition matrix, H 2 is the second transition matrix, is the Hermitian transpose operator, eigenvalue(·) is the function for obtaining eigenvalues.

5. The broadband oscillation monitoring method based on waveform data according to claim 4, characterized in that, in step A26, according to the following formula, the rough frequency estimates of the M signal components of the electrical signal are obtained by solving based on M eigenvalues: Among them, is the rough estimated value of the frequency of the m-th signal component of the electrical signal, arctan(·) is the arctangent function, T s is the sampling period of the synchronized phasor measurement unit of the wide area measurement system, Im(·) is the function for obtaining the imaginary part, and Re(·) is the function for obtaining the real part.

6. The broadband oscillation monitoring method based on waveform data according to claim 5, characterized in that, step A3 includes the following sub-steps: A31. Set the order and total number of iterations of the Taylor Fourier transform, and initialize the current iteration number to 0; A32. Construct a Taylor Fourier basis phasor according to the rough frequency estimates of the M signal components of the electrical signal; B = [B 1 , B 2 , … B m , …, B M-1 , B M ​ where B is the Taylor-Fourier basis phasor, and B m is the m-th element of the Taylor-Fourier basis phasor, n is the sampling result serial number of the synchronized phasor measurement unit of the wide area measurement system, the length of the sequence [-n, -(n - 1), …, 0, …, n - 1, n] is N, ω m is the angular frequency, p is the order of the Taylor-Fourier transform, and T is the matrix transpose operator; A33. Solve the dynamic phasor according to the Taylor Fourier basis phasor and the electrical signal waveform sequence after data window truncation; A34. According to the solved dynamic phasor, solve the refined estimates of the electrical signal parameters through the following formulas: Among them, is the amplitude refined estimate of the m-th signal component of the electrical signal, is the phase refined estimate of the m-th signal component of the electrical signal, is the frequency refined estimate of the m-th signal component of the electrical signal, is the zero-order derivative of the m-th element of the dynamic phasor, is the first-order derivative of the m-th element of the dynamic phasor, abs(·) is the function for obtaining the amplitude, angle(·) is the function for obtaining the phase angle, and j is the imaginary identifier; A35. Increase the current iteration number by 1; A36. Determine whether the current iteration number is greater than the total number of iterations. If so, end; if not, refresh the rough frequency estimates of all the signal components of the electrical signal with the refined frequency estimates of all the signal components of the electrical signal, and jump to step A32.

7. The broadband oscillation monitoring method based on waveform data according to claim 6, characterized in that, in step A33, according to the Taylor Fourier basis phasor and the electrical signal waveform sequence after data window truncation, the least squares method is used to solve the dynamic phasor through the following formula: X = [x(-n), x(-(n - 1)), …, x(0), …, x(n - 1), x(n)] T Among them, is a dynamic phasor, and X is the electrical signal waveform sequence after data window truncation.

Images