Broadband oscillation monitoring method and apparatus

Abstract

The application discloses a broadband oscillation monitoring method and device, the method comprises the following steps: collecting initial signals of a power system line to be detected in real time, and pre-processing the initial signals; dividing the pre-processed signals into frequency intervals, and performing adaptive filtering; separating a plurality of key components from the filtered signals; performing parameter estimation on the plurality of key components, extracting target modal information, and reconstructing signals; performing multi-modal recognition on the reconstructed signals, and extracting a plurality of modal components; and determining whether a broadband oscillation phenomenon exists by using the plurality of modal components. The method realizes accurate separation and parameter estimation of broadband oscillation signals, and improves the recognition accuracy of oscillation frequency, amplitude and phase.

Description

Technical Field

[0001] This invention relates to the field of broadband oscillations in power systems, and more specifically, to a broadband oscillation monitoring method and apparatus based on TLS-ESPRIT with improved VMD filtering. Background Technology

[0002] In the context of modern power systems characterized by high power consumption and high voltage, a new type of oscillation problem has emerged. This is due to the complex interactions between power electronic devices and grid components, which induce oscillations with wide-frequency, time-varying, and significantly nonlinear characteristics. Unlike previous oscillations, this new form of oscillation is primarily rooted in the power electronic control mechanism, causing oscillations spanning over 10... -1 Up to 10 3 The electromagnetic oscillation in the Hertz frequency band is hence the name "wideband oscillation".

[0003] With the increasing frequency of grid connection of power electronic equipment such as wind and solar power and flexible DC transmission, power systems often encounter a series of thorny problems when managing and mitigating broadband oscillations. These include difficulties in monitoring and identification, complex and intertwined causes, widespread ripple effects, and limitations of existing control strategies. Specifically, the broad spectrum and variable dynamics of broadband oscillations exceed the detection capabilities of conventional monitoring systems; their causes involve multiple levels, including grid layout, new energy integration, and control strategies, requiring in-depth analysis and precise responses. Furthermore, such oscillations can easily spread rapidly throughout the grid, posing a severe challenge to the stability of the entire system. Summary of the Invention

[0004] To address the problems existing in the background technology, this invention proposes a broadband oscillation monitoring method and device, aiming to solve the problems of insufficient identification accuracy, weak noise resistance and limited ability to handle complex oscillation modes in the existing technology for broadband oscillation monitoring, and improve the monitoring and analysis capabilities of broadband oscillations in power systems.

[0005] This invention provides a broadband oscillation monitoring method, comprising:

[0006] The initial signal of the power system line under test is acquired in real time and preprocessed.

[0007] The preprocessed signal is divided into frequency ranges and then adaptively filtered.

[0008] Several key components were separated from the filtered signal;

[0009] The parameters of these key components are estimated, the target modal information is extracted, and the signal is reconstructed.

[0010] Multimodal identification is performed on the reconstructed signal to extract several modal components;

[0011] Using these modal components, we can determine whether broadband oscillations exist.

[0012] In one embodiment, the initial signal is preprocessed, including:

[0013] Wavelet analysis is used to denoise the initial signal, including:

[0014] By calculating each wavelet coefficient of the initial signal and applying soft thresholding to the wavelet coefficients, wavelet coefficients below a first threshold are removed, and wavelet coefficients greater than or equal to the first threshold are retained, i.e.:

[0015] T λ (x) = sign(x)max(|x|-λ,0)

[0016] In the formula: λ is the preset first threshold, T λ (x) is the wavelet coefficient obtained after applying the first threshold, and x is the original wavelet coefficient.

[0017] In one embodiment, preprocessing the initial signal further includes:

[0018] The initial signal is subjected to preliminary mode detection using Fast Fourier Transform to obtain initial mode information, which includes the frequency, amplitude, and initial phase of the resonant mode.

[0019] In one embodiment, the preprocessed signal is divided into frequency ranges and adaptive filtering is performed, including:

[0020] The preprocessed signal is initially divided into multiple frequency ranges, which include low frequency, medium frequency, high frequency, ultra-high frequency and interference frequency bands;

[0021] Based on these multiple frequency ranges, an adaptive bandpass filter is designed to dynamically adjust the center frequency and bandwidth, including:

[0022] Set the initial upper and lower frequency limits for bandpass filters in each frequency range: low frequency, mid frequency, high frequency, and ultra-high frequency.

[0023] Based on the initial upper and lower frequency limits of each bandpass filter, the initial center frequency and initial bandwidth of the bandpass filter are set. The initial center frequency is the average of the initial upper and lower frequency limits, and the initial bandwidth is the difference between the average of the initial upper and lower frequency limits.

[0024] Obtain real-time spectrum information, and dynamically adjust the center frequency and bandwidth of each bandpass filter based on this spectrum information, wherein:

[0025] The adjusted center frequency is the sum of the initial center frequency and the frequency adjustment amount, which is determined based on the difference between the initial center frequency and the main frequency components extracted from the spectrum information.

[0026] The adjusted bandwidth is the sum of the initial bandwidth and the bandwidth adjustment amount, which is determined based on the difference between the initial bandwidth and the main bandwidth component extracted from the spectrum information.

[0027] In one embodiment, an objective function is constructed based on the adjusted center frequency and bandwidth, and the objective function is expressed as:

[0028]

[0029] Where: φ k (t) represents the k-th mode function, a k (t) is the time-varying amplitude function of the k-th mode, f H f is the upper limit of frequency. L δ is the lower limit of frequency; β and γ are regularization parameters used to control the smoothness of the mode and the degree to which the bandwidth approaches the preset target δ.

[0030] In one embodiment, all bandpass filters are cascaded, and the center frequency and bandwidth of the bandpass filter in the next stage are finely adjusted according to the output of the bandpass filter in the previous stage.

[0031] In one embodiment, a matching pursuit algorithm is used to separate the key components of the filtered signal layer by layer, including:

[0032] Set the initial residual to the filtered signal;

[0033] In each iteration, an atom is selected from a predefined atom library such that the atom matches the current residual to the highest degree under a preset metric condition;

[0034] Calculate a coefficient between the atom and the current residual;

[0035] The next residual is obtained based on the difference between the current residual and the coefficient multiplied by the atom.

[0036] In one embodiment, the key component is parameter estimated using the TLS-ESPRIT algorithm, including:

[0037] Using these key components as input, construct a Hankel matrix;

[0038] The eigenvalue decomposition of the constructed Hankel matrix is ​​performed using the total least squares method;

[0039] Frequency components are extracted from the obtained feature values.

[0040] In one embodiment, multimodal identification of the reconstructed signal and extraction of modal components includes:

[0041] Calculate the autocorrelation function of the reconstructed signal and set a threshold for the autocorrelation function to distinguish the modal components and noise of the reconstructed signal;

[0042] Different modal features are extracted and classified from the identified modal components. These modal features include frequency, amplitude, and phase.

[0043] In one embodiment, determining whether a broadband oscillation phenomenon exists includes:

[0044] Determine whether there is a frequency band exceeding a preset width frequency range in the frequency of each modal component. If there is a frequency band exceeding the preset width frequency range, it is identified as having broadband oscillation.

[0045] and / or

[0046] Calculate the power spectral density of each modal component to determine the energy distribution; if there is an abnormal increase in energy in the mid-frequency band, it is identified as a broadband oscillation.

[0047] Observe the frequency changes of each modal component over time. If there are rapid fluctuations or irregular changes in frequency, it is identified as a broadband oscillation.

[0048] In one embodiment, before determining whether a broadband oscillation phenomenon exists, the method further includes:

[0049] Applying the Hilbert transform to each modal component determines the instantaneous frequency and instantaneous amplitude of the signal, where: the instantaneous frequency is determined based on the ratio of the derivative of the instantaneous phase to 2π, i.e.:

[0050]

[0051] Where f(t) represents the instantaneous frequency, and the instantaneous phase Φ(t) is determined by the phase of the instantaneous complex signal z(t) = x(t) + jy(t) composed of the imaginary part y(t) obtained by the Hilbert transform of the real part x(t), and Φ(t) = arg(z(t)).

[0052] The instantaneous amplitude is determined by the magnitude of the instantaneous complex signal amplitude, that is:

[0053]

[0054] Where a(t) represents the instantaneous amplitude.

[0055] In one embodiment, before applying the Hilbert transform to each modal component to determine the instantaneous frequency and instantaneous amplitude of the signal, the method further includes:

[0056] Adaptive filtering is performed on modal components with complex frequency structures; and

[0057] In this adaptive filtering process, an iterative optimization algorithm is used to update the filter weights under the minimum mean square error criterion, where: the filter weight vector w n The update in the nth iteration is represented as:

[0058] w[n+1]=w[n]+μe[n]x[n]

[0059] In the formula: e[n] is the current error, d[n] is the expected output, y[n] is the actual output of the filter, x[n] is the input vector, and μ is the learning rate.

[0060] In another aspect, the present invention provides a broadband oscillation monitoring device, which employs the above-described broadband oscillation monitoring method, the broadband oscillation monitoring device comprising:

[0061] The preprocessing module is used to acquire the initial signal of the power system line under test in real time and preprocess the initial signal.

[0062] The adaptive filtering module is used to divide the preprocessed signal into frequency ranges and perform adaptive filtering.

[0063] The key component extraction module is used to separate several key components from the filtered signal;

[0064] The reconstruction module is used to estimate the parameters of the key components, extract the target modal information, and reconstruct the signal.

[0065] The multimodal recognition module is used to perform multimodal recognition on the reconstructed signal and extract several modal components;

[0066] The wideband oscillation determination module is used to determine whether wideband oscillations exist by utilizing these modal components.

[0067] Furthermore, another aspect of the present invention provides a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the aforementioned broadband oscillation monitoring method and achieve the same technical effect.

[0068] This invention has outstanding substantive features and significant progress compared to the prior art, specifically:

[0069] This invention proposes a broadband oscillation monitoring method. This method preprocesses the initial signal from the real-time acquired power system line under test, divides the preprocessed signal into frequency ranges, and performs adaptive filtering using an improved VMD filter. Simultaneously, it separates several key components from the filtered signal and uses the TLS-ESPRIT algorithm to estimate the parameters of these key components for signal reconstruction. The reconstructed signal then undergoes multi-mode identification to extract several mode components. Finally, these mode components are used to determine whether broadband oscillations exist. This method achieves accurate separation and parameter estimation of broadband oscillation signals, improving the accuracy of identifying oscillation frequency, amplitude, and phase. This method is applicable to various oscillation phenomena in modern power systems, especially broadband, time-varying, and nonlinear oscillations caused by complex interactions between power electronic devices and grid components. Attached Figure Description

[0070] Figure 1 This is a schematic diagram of the overall flow of a broadband oscillation monitoring method according to an embodiment of the present invention;

[0071] Figure 2 This is a schematic block diagram of the overall structure of a broadband oscillation monitoring device according to an embodiment of the present invention;

[0072] The attached figures are labeled as follows:

[0073] 300-Wideband Oscillation Monitoring Device;

[0074] 301 - Preprocessing Module;

[0075] 302-Adaptive Filtering Module;

[0076] 303 - Key Component Extraction Module;

[0077] 304 - Refactoring module;

[0078] 305 - Multimodal Recognition Module;

[0079] 306 - Wideband Oscillation Detection Module. Detailed Implementation

[0080] To make the above features and effects of the present invention clearer and easier to understand, specific embodiments are described below, and detailed descriptions are provided in conjunction with the accompanying drawings.

[0081] Please see Figure 1 As shown, Figure 1 A schematic diagram of the overall flow of a broadband oscillation monitoring method disclosed in an embodiment of the present invention is shown.

[0082] A broadband oscillation monitoring method includes at least the following steps:

[0083] Step S1: Acquire the initial signal of the power system line under test in real time and preprocess the initial signal.

[0084] In one embodiment, a PMU (Phasor Measurement Unit) is used to monitor the initial signals of the power system line under test in real time, including voltage and current, and the raw signals are sampled to ensure that the signals are digitized for further analysis.

[0085] Furthermore, signal quality is improved through preprocessing. Specifically, in one embodiment, wavelet analysis, particularly the Daubechies wavelet basis, is used to denoise the initial signal. This is achieved by calculating each wavelet coefficient of the initial signal and applying a soft threshold to these coefficients, i.e.:

[0086] T λ (x) = sign(x)max(|x|-λ,0)

[0087] In the formula: λ is the preset first threshold, T λ (x) is the wavelet coefficient obtained after applying the first threshold, and x is the original wavelet coefficient.

[0088] In this formula, max(|x|-λ, 0) ensures that if |x|-λ is negative (i.e., x is less than the first threshold λ), it is set to 0, so that small values ​​are often regarded as noise. sign(x) multiplied by max(|x|-λ, 0) ensures that the processed value retains the same sign as the original value, only its absolute value is thresholded.

[0089] Soft thresholding can "smooth" away wavelet coefficients below a first threshold while retaining wavelet coefficients above or equal to the first threshold, although there will be some attenuation. Therefore, it is very useful for noise reduction and data compression, especially for large power systems containing a large number of microsecond-level power electronic devices, ensuring that the quality of signal processing is not affected by noise.

[0090] Furthermore, in one embodiment, the preprocessing process further utilizes Fast Fourier Transform to perform preliminary mode detection on the initial signal, obtaining initial mode information, which includes the frequency, amplitude, and initial phase of the resonant modes. This process provides basic data references to guide subsequent bandpass filter parameter settings and the setting of signal processing thresholds.

[0091] Step S2: Divide the preprocessed signal into frequency ranges and perform adaptive filtering.

[0092] In one embodiment, the K-means clustering algorithm is used to perform cluster analysis on the frequencies of the resonance modes in the initial modal information obtained by the Fast Fourier Transform (FFT), initially dividing the preprocessed signal into multiple frequency intervals according to frequency range. Specifically, these multiple frequency intervals include low-frequency f... l , intermediate frequency f m High frequency f h Extremely high frequency f he and interference frequency band f o There are five categories in total. Among them, the interference frequency band f o Poor clustering results are attributed to spectral feature points that are difficult to clearly categorize into low-frequency, mid-frequency, high-frequency, or ultra-high-frequency categories. These feature points may frequently change their cluster affiliation or be far from all cluster centers. o As interference, it was excluded, and further noise reduction was performed on the basis of wavelet noise reduction processing.

[0093] Furthermore, based on the initial modal information provided by FFT modal detection, an adaptive bandpass filter is designed for adaptive filtering. This adaptive bandpass filter can dynamically adjust parameters such as center frequency and bandwidth to extract or separate voltage / current signals in each frequency band. This not only improves processing efficiency but also ensures the accuracy and real-time performance of signal decomposition.

[0094] In one embodiment, an improved VMD filter is designed. Based on the traditional VMD (Variational Mode Decomposition) filter, this improved VMD filter dynamically adjusts the lower frequency limit f. L and frequency upper limit f H and with the lower limit f of that frequency L With the upper limit of this frequency f H Using the average value as the center frequency, the bandwidth and other parameters are automatically adjusted to achieve adaptive signal separation, especially within their respective frequency bands, making the filtering process more closely match the true characteristics of the signal.

[0095] Specifically, firstly, based on the aforementioned multiple frequency ranges, the initial upper and lower frequency limits of the bandpass filters for each frequency range—low frequency, mid frequency, high frequency, and ultra-high frequency—are set, i.e.:

[0096] Low-frequency bandpass filter: Initial frequency range Min(f) l )-Max(f l )

[0097] Intermediate frequency bandpass filter: Initial frequency range Min(f) m )-Max(f m )

[0098] High-frequency bandpass filter: Initial frequency range Min(f) h )-Max(f h )

[0099] Extremely high frequency bandpass filter: Initial frequency range Min(f) he )-Max(f he )

[0100] Furthermore, based on the initial upper and lower frequency limits of each bandpass filter, the initial center frequency and initial bandwidth of the bandpass filter are determined. The initial center frequency is the average of the initial upper and lower frequency limits, and the initial bandwidth is the difference between the initial upper and lower frequency limits.

[0101] Furthermore, in practical applications, the frequency characteristics of a signal may change with time or environment. Therefore, this adaptive bandpass filter further refines and dynamically adjusts the frequency range based on the initially defined frequency intervals to more accurately separate and extract frequency components from the signal. In one embodiment, the signal's spectral information, particularly changes in frequency components, is monitored in real time. Specifically, a sliding window FFT can be used to obtain real-time spectral information. The length of the sliding window is set to L, and the sliding step size is S. An FFT is performed on each segment of the original signal within each sliding window length to obtain the spectrum within that time window:

[0102]

[0103] X can be obtained n (k) The signal values ​​within the window are then calculated, and the window slides with a step size S to obtain a series of spectrum diagrams X. n (f).

[0104] Then, based on the real-time monitored spectrum information, the center frequency and bandwidth of each bandpass filter are dynamically adjusted, wherein:

[0105] For each bandpass filter, the adjusted center frequency is the sum of the initial center frequency and a frequency adjustment amount, which is determined based on the difference between the initial center frequency and the main frequency components extracted from the spectral information. That is, the adjusted center frequency is:

[0106] f c (t+1)=f c (t)+Δf c In the formula: fc is the initial center frequency of the bandpass filter, Δf c Δf is the frequency adjustment amount. c =α(f main -f c), where α is the adjustment coefficient used to control the speed and magnitude of the adjustment.

[0107] For each bandpass filter, the adjusted bandwidth is the sum of the initial bandwidth and the bandwidth adjustment amount, which is determined based on the difference between the initial bandwidth and the main bandwidth component extracted from the spectral information. That is, the adjusted bandwidth is: Bandwidth(t+1) = Bandwidth(t) + ΔBW, where: Bandwidth(t) is the initial bandwidth of the bandpass filter, and ΔBW is the bandwidth adjustment amount, ΔBW = β(BW) / ΔBW. main -BW), where β is the adjustment coefficient used to control the speed and magnitude of the adjustment.

[0108] After adjusting the center frequency and bandwidth, an initial frequency range is provided for VMD optimization. Simultaneously, the results of the sliding window FFT are used to update the center frequency and bandwidth of the modal function in real time. This dynamic adjustment ensures that the modal function remains synchronized with signal changes.

[0109] Based on this, further optimization of VMD processing is carried out.

[0110] The objective function of Variational Mode Decomposition (VMD) is used to decompose a complex signal into a set of modal components with specific frequency characteristics. In one embodiment, the objective function is constructed based on the adjusted center frequency and bandwidth, and is expressed as follows:

[0111]

[0112] Where: φ k (t) represents the k-th mode function, a k (t) is the time-varying amplitude function of the k-th mode, f H f is the upper limit of frequency. L δ is the lower limit of frequency; β and γ are regularization parameters that control the smoothness and bandwidth constraints of the mode, used to control the degree to which the smoothness and bandwidth of the mode approach the preset target δ.

[0113] Wherein, the mode function φ k (t) and magnitude function a k The initial state of (t) is determined based on the preceding steps. β is used to control the smoothness of the mode function, and an empirical initial value can be determined through theoretical analysis. For example, for signals with high noise, a larger β value is chosen to suppress high-frequency noise. Similarly, an initial value of γ is set. γ is used to control the frequency domain bandwidth of the mode component to approach the preset target bandwidth δk. For signals with a wide spectrum, a larger γ value is chosen for bandwidth control. For the k-th mode, its target bandwidth δk is... k This is determined through the preceding fast FFT and dynamic adjustments.

[0114] Furthermore, in one embodiment, the Alternating Directional Multiplier Method (ADMM) is employed to iteratively optimize the objective function, performing iterative optimization of the amplitude function, mode function, and frequency upper and lower limits respectively, ultimately converging to obtain a local optimum, thus completing the VMD decomposition process. The ultimate goal is to ensure that the characteristics of each mode (such as frequency, bandwidth, sparsity, etc.) accurately reflect the properties of the original signal when separating and extracting different modes of the signal.

[0115] Furthermore, in some embodiments, to process different frequency components of the signal more precisely, a method of cascading multiple improved VMD filters is employed. This involves cascading bandpass filters for the low-frequency, mid-frequency, high-frequency, and ultra-high-frequency ranges, with each bandpass filter optimized for a different frequency band. By setting multiple VMD filters with different parameters, each bandpass filter is responsible for separating signal components within a specific frequency range, such as low-frequency, mid-frequency, high-frequency, and ultra-high-frequency. The advantage of cascading processing is that it can improve the accuracy and real-time performance of signal separation through multi-stage refinement.

[0116] The subsequent filter adjusts its center frequency, bandwidth, and other parameters based on the output of the previous filter. The overall process can be viewed as an iterative optimization chain as follows:

[0117] x (i) (t)→VMD i (f L，i f H,i params i )→y (i) (t)

[0118] x (i+1) (t)=y (i) (t)-Processed Component

[0119] In the formula: x (i) (t) and y (i) (t) represents the input and output signals of the i-th stage of processing, respectively; VMDi represents the signal component of a specific frequency band that the i-th filter processes and removes in each iteration until the required processing accuracy or a predetermined number of cascades is achieved; Processed Component represents the signal from y (i) The specific frequency components separated in (t) actually involve the analysis of y. (i) Appropriate processing of (t) is required to prepare it for the next stage of analysis or further filtering.

[0120] Specifically, firstly (1) for the divided frequency band, perform the first-level VMD decomposition to obtain the first-level modal components; (2) calculate the signal-to-noise ratio to evaluate the effect of the first-level decomposition; (3) based on the effect of the first-level VMD decomposition, fine-tune the parameters such as the center frequency and bandwidth; (4) repeat the steps of the first level, perform further VMD processing on the output signal of each level, and continuously adjust the VMD parameters according to the decomposition effect until the set decomposition accuracy or the maximum number of cascades is reached.

[0121] In this embodiment, wavelet denoising technology and an improved VMD filter are employed to effectively remove various noises from the signal, including thermal noise, shot noise, and flicker noise, thereby enhancing the quality of signal processing. The improved VMD filter can dynamically adjust its bandwidth and center frequency, adapting to the frequency characteristics of the signal and making it suitable for signal processing in different frequency bands. Furthermore, by cascading multiple bandpass filters, multi-stage refinement can be achieved, improving the accuracy and real-time performance of signal separation.

[0122] Step S3: Separate several key components from the filtered signal.

[0123] In this embodiment, after filtering is completed through step S2, the filtered signal is further subjected to key component separation. In one embodiment, a matching pursuit algorithm is used to separate several key components of the signal layer by layer, especially handling transient and non-stationary parts, thereby improving the clarity of the signal structure.

[0124] Matching Pursuit (MP) is an iterative signal decomposition method, particularly suitable for processing non-stationary signals and signals with transient characteristics, such as broadband electromagnetic oscillations in power systems. The core idea of ​​this algorithm is to extract the key components that best represent the signal structure from the signal through layer-by-layer decomposition, thereby improving the signal's resolution and clarity. Applying MP after cascading Virtual Machine Decomposition (VMD) can further optimize the signal decomposition results and improve reconstruction accuracy, especially when dealing with noisy and complex signals. Its main steps are as follows:

[0125] First, at the beginning, the initial residual r0 of the signal is defined as the filtered signal x[n], that is, r0 = x[n].

[0126] Then, in each iteration, an atom g[n] is selected from a predefined library of atoms (usually a redundant dictionary, such as wavelet basis, Fourier basis, or Gabor basis, etc.) such that this atom is related to the current residual r. n The highest degree of matching is achieved under a certain metric (such as inner product). That is, finding an atom such that...

[0127]

[0128] In the formula: It is an atomic library, and <., .> represents the inner product.

[0129] Calculate the selected atom g[n] and the current residual r n The coefficient a between n The length of the projection of the signal along the direction of that atom determines the contribution of that atom to the signal representation.

[0130]

[0131] Based on the difference between the current residual and the coefficient multiplied by the atom, the residual signal is updated to obtain the next residual r. n+1 This is to reflect the remaining information beyond the extracted components, namely:

[0132] r n+1 =r n -a n g n

[0133] Repeat the above iterative process until a certain preset stopping criterion is met. In some embodiments, the preset stopping condition is set as the residual energy being lower than a set threshold, reaching the maximum number of iterations, or the residual change being small enough.

[0134] In this embodiment, the matching pursuit algorithm is used to separate several key components of the signal layer by layer, thereby improving the clarity and resolution of the signal decomposition.

[0135] Step S4: Estimate the parameters of the key components, extract the target mode information, and reconstruct the signal. The target mode information includes frequency, amplitude, and phase information.

[0136] In this embodiment, after processing with cascaded VMD filtering and matched pursuit algorithms, several key components of the signal are extracted, and the signal is decomposed into multiple mode functions. These mode functions contain signal characteristics of different frequency components. Furthermore, the TLS-ESPRIT algorithm is used to estimate the parameters of these key components. By constructing a Hankel matrix and performing total least squares estimation, the target mode information is accurately extracted for signal reconstruction. Compared to the initial mode information, this target mode information provides more accurate frequency, amplitude, and phase information, thus completing the signal's spectral reconstruction.

[0137] First, using these key components as input, a Hankel matrix H is constructed. If the sampling points of these key components are x... n If n = 0, 1, 2, 3, ..., N-1, then the Hankel matrix can usually be represented as...

[0138]

[0139] In the formula: M is the number of rows in the Hankel matrix (which also determines the lag length), N is the total number of sampling points of the signal, and usually N>M.

[0140] Then, the eigenvalue decomposition of the constructed Hankel matrix is ​​performed using the Total Least Squares (TLS) method. Unlike traditional least squares methods that only minimize the sum of squared errors, the TLS method considers the errors of both the observation vector and the coefficient matrix, typically providing a more robust estimate. After performing TLS-ESPRIT, the eigenvalue decomposition of the matrix is ​​obtained:

[0141] H=UΛU H

[0142] In the formula: U is a matrix containing orthogonal eigenvectors, Λ is a diagonal matrix, and its diagonal elements are the eigenvalues ​​of the Hankel matrix, which contain the frequency information of the signal.

[0143] Finally, the frequency components of the signal can be further extracted from the obtained eigenvalues. For the case of discrete spectral peaks, the frequency f... k The following relationship can be used to calculate based on the k-th eigenvalue:

[0144]

[0145] In the formula: T represents the sampling period, and k is an index based on the eigenvalue sorting, reflecting the position of different frequency components in the signal.

[0146] In this embodiment, the target mode information of the signal is accurately estimated using the TLS-ESPRIT algorithm, through the total least squares method and the principle of rotation invariance.

[0147] Step S5: Perform multimodal recognition on the reconstructed signal and extract several modal components.

[0148] In some embodiments, after reconstructing the signal through step S4 above, the autocorrelation function (ACF) of the reconstructed signal is further calculated, and a threshold for the ACF is set to distinguish the modal components and noise of the reconstructed signal, further verifying the accuracy of the identification results. For broadband oscillation monitoring, the calculation of the ACF can be expressed as follows:

[0149]

[0150] In the formula: x[n] represents the sample value of the reconstructed signal, and τ is the lag time. To distinguish between modal components and noise in the reconstructed signal, a threshold T for the autocorrelation function is set, and the screening condition can be expressed as: R xx If (τ) > T, then the modal component exists.

[0151] Then, the identified modal components are further extracted and classified for different modal features, providing a reliable data foundation for multimodal identification. Once the signal modes are initially identified, the next step is to extract and classify the features of these modes. The main features include frequency, amplitude, and phase, which can be obtained directly or indirectly from the frequency domain or time-frequency analysis of the signal. The parameters in this case can be obtained from the TLS-ESPRIT algorithm, where:

[0152] The extracted frequency features are: In the formula: k is the frequency index, N is the number of sampling points, and T is the sampling period.

[0153] The extracted amplitude features are: A k =|X[k]|, where: X[k] is the kth coefficient of the discrete Fourier transform.

[0154] The extracted phase feature is: φ k =∠X[k].

[0155] Then, after extracting the frequency, amplitude, and phase features, a clustering algorithm is used for modal classification.

[0156] Cluster = ClusteringMethod({A k ,φ k ,...})

[0157] In the formula: set {A} k φ k , ...} represents the set of feature vectors of a modality, while ClusteringMethod is the representation of a clustering algorithm.

[0158] Step S6: Use these modal components to determine whether broadband oscillations exist.

[0159] In step S6, Hilbert transform is first applied to these modal components to determine the instantaneous frequency and instantaneous amplitude of the signal.

[0160] In in-depth analysis of broadband oscillations, the Hilbert transform is a powerful tool for understanding the instantaneous characteristics of signals. Calculating the instantaneous frequency f(t) allows direct observation of frequency changes at different points in time, which is crucial for capturing nonlinear or non-steady-state behavior in broadband oscillations. It helps identify and track the evolution of oscillation modes over time.

[0161] The instantaneous frequency ft is determined by the ratio of the derivative of the instantaneous phase Φ(t) to 2π, that is:

[0162]

[0163] The instantaneous phase Φ(t) is determined by the phase of the complex signal z(t) = x(t) + jy(t) composed of the imaginary part y(t) obtained by the Hilbert transform of the real part x(t), Φ(t) = argz(t)).

[0164] The instantaneous amplitude a(t) is the modulus of the instantaneous complex signal amplitude, that is:

[0165]

[0166] Furthermore, in some embodiments, before applying Hilbert transform to these modal components to determine the instantaneous frequency and amplitude of the signal, adaptive filtering is performed on the modal component portion with complex frequency structures to further reduce noise and ensure signal purity. This allows for dynamic adjustment of filter parameters to better adapt to signal characteristics when processing modal components with complex frequency structures. In one embodiment, an iterative optimization algorithm is used to update the filter weights under the minimum mean square error (LMS) criterion, where: the filter weight vector w n The update in the nth iteration is represented as:

[0167] w[n+1]=w[n]+μe[n]x[n]

[0168] In the formula: e[n] is the current error, d[n] is the expected output, y[n] is the actual output of the filter, x[n] is the input vector, and μ is the learning rate.

[0169] In this embodiment, adaptive filtering is applied to signals with complex frequency structures to further reduce noise, ensuring signal purity and processing effectiveness, thereby enhancing the stability and reliability of the system.

[0170] In step S6, after applying Hilbert transform to each modal component to determine the instantaneous frequency and instantaneous amplitude of the signal, the existence of broadband oscillation is further determined based on the instantaneous frequency and instantaneous amplitude of each modal component.

[0171] In some embodiments, specifically, the existence of broadband oscillation is determined from three perspectives: frequency distribution analysis, modal energy analysis, and time-varying characteristic analysis.

[0172] For frequency distribution analysis, the presence of frequency bands exceeding a preset width is examined among the identified modal components. This involves focusing on frequencies outside the traditional narrowband oscillation range (such as the 0.1Hz to 2Hz range commonly found in power systems). Wideband oscillations typically involve multiple frequencies occurring simultaneously or sequentially, spanning a wide frequency range. If a frequency band exceeding the preset width exists, it is identified as a wideband oscillation.

[0173] For modal energy analysis, the energy distribution is assessed by calculating the power spectral density (PSD) of each modal component. Broadband oscillations are often accompanied by abnormally high energy levels in certain frequency bands. This can be identified by comparing the data with historical data from normal operation. If an abnormally high energy level is found in a specific frequency band, broadband oscillations are identified. The power spectral density is expressed as:

[0174] PSD(f k )=|X k | 2 In the formula: X k For the signal at frequency f k The spectral coefficients at that location.

[0175] For time-varying characteristic analysis, the instantaneous frequency information obtained by Hilbert transform is used to observe the frequency changes of each modal component over time. Broadband oscillations may manifest as rapid frequency fluctuations or irregular changes, which are related to nonlinear interactions within the system or external disturbances. Therefore, if rapid frequency fluctuations or irregular changes exist, they are identified as broadband oscillations.

[0176] In summary, the broadband oscillation monitoring method disclosed in this invention, through the combination of improved VMD filtering and the TLS-ESPRIT algorithm, achieves accurate separation and parameter estimation of broadband oscillation signals, improving the identification accuracy of oscillation frequency, amplitude, and phase. This method is applicable to various oscillation phenomena in modern power systems, especially broadband, time-varying, and nonlinear oscillations caused by complex interactions between power electronic devices and grid components.

[0177] Furthermore, corresponding to the aforementioned broadband oscillation monitoring method, another embodiment of the present invention also discloses a broadband oscillation monitoring device, referring to... Figure 2 , Figure 2 The diagram shows the architecture of the broadband oscillation monitoring device 300, which can achieve monitoring via, for example... Figure 1 The process of the broadband oscillation monitoring method is shown.

[0178] A broadband oscillation monitoring device 300, the device comprising at least:

[0179] The preprocessing module 301 is used to acquire the initial signal of the power system line under test in real time and preprocess the initial signal.

[0180] The adaptive filtering module 302 is used to divide the preprocessed signal into frequency ranges and perform adaptive filtering.

[0181] The key component extraction module 303 is used to separate several key components from the filtered signal.

[0182] The reconstruction module 304 is used to estimate the parameters of the key components, extract the target modal information, and reconstruct the signal.

[0183] The multimodal recognition module 305 is used to perform multimodal recognition on the reconstructed signal and extract several modal components.

[0184] The wideband oscillation determination module 306 is used to determine whether a wideband oscillation phenomenon exists by utilizing the several modal components.

[0185] Furthermore, corresponding to the aforementioned broadband oscillation monitoring method, another embodiment of the present invention discloses an electronic device, including a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the steps of the aforementioned broadband oscillation monitoring method and achieve the same technical effect.

[0186] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0187] Another embodiment of the present invention also provides a readable storage medium storing a program or instructions that, when executed by a processor, implement the above-described broadband oscillation monitoring steps and achieve the same technical effect.

[0188] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be applied, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0189] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0190] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A broadband oscillation monitoring method, characterized in that, Include: The initial signal of the power system line under test is acquired in real time and preprocessed. The preprocessed signal is divided into frequency ranges and then adaptively filtered. Several key components were separated from the filtered signal; The parameters of these key components are estimated, the target modal information is extracted, and the signal is reconstructed. Multimodal identification is performed on the reconstructed signal to extract several modal components; Using these modal components, it can be determined whether broadband oscillations exist. This includes dividing the preprocessed signal into frequency ranges and performing adaptive filtering, which includes: The preprocessed signal is initially divided into multiple frequency ranges, which include low frequency, medium frequency, high frequency, ultra-high frequency and interference frequency bands; Based on these multiple frequency ranges, an adaptive bandpass filter is designed to dynamically adjust the center frequency and bandwidth, including: Set the initial upper and lower frequency limits for bandpass filters in each frequency range: low frequency, mid frequency, high frequency, and ultra-high frequency. Based on the initial upper and lower frequency limits of each bandpass filter, the initial center frequency and initial bandwidth of the bandpass filter are set. The initial center frequency is the average of the initial upper and lower frequency limits, and the initial bandwidth is the difference between the average of the initial upper and lower frequency limits. Obtain real-time spectrum information, and dynamically adjust the center frequency and bandwidth of each bandpass filter based on this spectrum information, wherein: The adjusted center frequency is the sum of the initial center frequency and the frequency adjustment amount, which is determined based on the difference between the initial center frequency and the main frequency components extracted from the spectrum information. The adjusted bandwidth is the sum of the initial bandwidth and the bandwidth adjustment amount, which is determined based on the difference between the initial bandwidth and the main bandwidth components extracted from the spectral information. Also includes: Based on the adjusted center frequency and bandwidth, an objective function is constructed, which is expressed as: ； In the formula: Represents the k-th mode function, f is the time-varying amplitude function of the k-th mode. H f is the upper limit of frequency. L This is the lower limit of the frequency. and The regularization parameter is used to control the smoothness and bandwidth of the modality to a degree that approaches the preset target δ.

2. The method according to claim 1, characterized in that, The initial signal is preprocessed, including: Wavelet analysis is used to denoise the initial signal, including: By calculating each wavelet coefficient of the initial signal and applying soft thresholding to the wavelet coefficients, wavelet coefficients below a first threshold are removed, and wavelet coefficients greater than or equal to the first threshold are retained, i.e.: ； In the formula: For the preset first threshold, is the wavelet coefficient obtained after applying the first threshold, and x is the original wavelet coefficient.

3. The method according to claim 1 or 2, characterized in that, Preprocessing of the initial signal also includes: The initial signal is subjected to preliminary mode detection using Fast Fourier Transform to obtain initial mode information, which includes the frequency, amplitude, and initial phase of the resonant mode.

4. The method according to claim 1, characterized in that, Also includes: All bandpass filters are cascaded, and the center frequency and bandwidth of each subsequent bandpass filter are finely adjusted based on the output of the previous bandpass filter.

5. The method according to claim 1, characterized in that, The matching pursuit algorithm is used to separate the key components of the filtered signal layer by layer, including: Set the initial residual to the filtered signal; In each iteration, an atom is selected from a predefined atom library such that the atom matches the current residual to the highest degree under a preset metric condition; Calculate a coefficient between the atom and the current residual; The next residual is obtained based on the difference between the current residual and the coefficient multiplied by the atom.

6. The method according to claim 1, characterized in that, The key component is parameter estimated using the TLS-ESPRIT algorithm, including: Using these key components as input, construct a Hankel matrix; The eigenvalue decomposition of the constructed Hankel matrix is ​​performed using the total least squares method; Frequency components are extracted from the obtained feature values.

7. The method according to claim 1, characterized in that, Multimodal identification is performed on the reconstructed signal to extract modal components, including: Calculate the autocorrelation function of the reconstructed signal and set a threshold for the autocorrelation function to distinguish the modal components and noise of the reconstructed signal; Different modal features are extracted and classified from the identified modal components. These modal features include frequency, amplitude, and phase.

8. The method according to claim 1, characterized in that, Determining whether broadband oscillation exists includes: Determine whether there is a frequency band exceeding a preset width range among the frequencies of each modal component. If a frequency band exceeding the preset width range exists, it is identified as having broadband oscillation; and / or Calculate the power spectral density of each modal component to determine the energy distribution; if there is an abnormal increase in energy in the mid-frequency band, it is identified as a broadband oscillation. Observe the frequency changes of each modal component over time. If there are rapid fluctuations or irregular changes in frequency, it is identified as a broadband oscillation.

9. The method according to claim 8, characterized in that, Before determining whether broadband oscillation exists, the following steps are also included: Applying the Hilbert transform to each modal component determines the instantaneous frequency and instantaneous amplitude of the signal, where: the instantaneous frequency is determined based on the ratio of the derivative of the instantaneous phase to 2π, i.e.: ； Where f(t) represents the instantaneous frequency and the instantaneous phase. From the real part The imaginary part obtained by Hilbert transform Composition of instantaneous complex signals The phase is determined. ; The instantaneous amplitude is determined by the magnitude of the instantaneous complex signal amplitude, that is: ； in, Indicates the instantaneous amplitude.

10. The method according to claim 9, characterized in that, Before applying the Hilbert transform to each modal component to determine the instantaneous frequency and instantaneous amplitude of the signal, the following steps are also included: Adaptive filtering is performed on modal components with complex frequency structures; and In this adaptive filtering process, an iterative optimization algorithm is used to update the filter weights under the minimum mean square error criterion, where: the filter weight vector The update in the nth iteration is represented as: ； In the formula: It is the current error. It is the expected output. It is the actual output of the filter. It is the input vector. It is the learning rate.

11. A broadband oscillation monitoring device, characterized in that, The broadband oscillation monitoring method according to any one of claims 1-10, wherein the broadband oscillation monitoring device comprises: The preprocessing module is used to acquire the initial signal of the power system line under test in real time and preprocess the initial signal. The adaptive filtering module is used to divide the preprocessed signal into frequency ranges and perform adaptive filtering. The key component extraction module is used to separate several key components from the filtered signal; The reconstruction module is used to estimate the parameters of the key components, extract the target modal information, and reconstruct the signal. The multimodal recognition module is used to perform multimodal recognition on the reconstructed signal and extract several modal components; The wideband oscillation determination module is used to determine whether wideband oscillations exist by utilizing these modal components.

Images