Method, device and equipment for identifying high-frequency oscillation mode of power system and medium

Abstract

The invention discloses a high-frequency oscillation mode identification method and device of a power system, equipment and a medium. The method comprises the following steps: acquiring an original signal from the power system; performing parameter optimization on a preset variational mode decomposition model by adopting a black wing plinux optimization algorithm to obtain a target variational mode decomposition model; calling a target variational mode decomposition model to decompose the original signal to obtain a plurality of intrinsic mode components; calculating an autocorrelation coefficient corresponding to each intrinsic mode component, and selecting at least one oscillation characteristic component according to a comparison result of the autocorrelation coefficient and a preset noise threshold; and carrying out oscillation mode identification on each oscillation characteristic component by adopting a preset energy operator, and determining corresponding dominant high-frequency oscillation information. Therefore, high-frequency oscillation information identification of the broadband signal of the power grid is realized, and important oscillation information in the broadband signal is effectively prevented from being lost.

Description

Technical Field

[0001] The present invention relates to the technical field of high-frequency oscillation mode identification, and in particular to a method, device, equipment and medium for identifying high-frequency oscillation mode of an electric power system. Background Art

[0002] As the global energy mix shifts toward cleaner energy, new energy plants (such as photovoltaic and wind power) and flexible direct current (HVDC) transmission projects, specifically modular multilevel converter-high voltage direct current (MMC-HVDC), are gaining widespread adoption. Due to the interactions between power electronics devices, MMC-HVDC projects are prone to high-frequency oscillations, ranging from hundreds of hertz to several thousand hertz, during operation. These oscillatory signals often exhibit difficult-to-identify nonlinear and nonstationary characteristics.

[0003] Existing research mainly uses methods such as Fast Fourier Transform (FFT), Prony algorithm, wavelet transform and Hilbert-Huang Transform (HHT) for oscillation detection, but they are mainly aimed at low-frequency oscillations and subsynchronous oscillation signals in power systems.

[0004] However, these methods have limitations when processing high-frequency oscillation signals across a wide frequency domain. High-frequency oscillation signals can occur across a wide frequency range, and oscillation signals in MMC-HVDC systems have multiple oscillation frequencies and time-varying oscillations. Directly applying these methods can easily lead to the loss of important oscillation information in broadband signals. Summary of the Invention

[0005] The present invention provides a method, device, equipment and medium for identifying high-frequency oscillation modes in an electric power system, which solves the technical problem that high-frequency oscillation signals may appear in a wide frequency domain, and the oscillation signals of the MMC-HVDC system have the characteristics of multiple oscillation frequencies and time-varying oscillations. Directly adopting the above scheme easily leads to the loss of important oscillation information in the broadband signal.

[0006] A first aspect of the present invention provides a method for identifying high-frequency oscillation modes of an electric power system, comprising:

[0007] Collect raw signals from the power system;

[0008] The Black Kite optimization algorithm is used to optimize the parameters of the preset variational mode decomposition model to obtain the target variational mode decomposition model;

[0009] Decomposing the original signal by calling the target variational mode decomposition model to obtain a plurality of eigenmode components;

[0010] Calculating the autocorrelation coefficients corresponding to the eigenmode components, and selecting at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficients and a preset noise threshold;

[0011] A preset energy operator is used to perform oscillation mode identification on each of the oscillation characteristic components to determine the corresponding dominant high-frequency oscillation information.

[0012] Optionally, the method of optimizing parameters of a preset variational modal decomposition model using a Black Kite optimization algorithm to obtain a target variational modal decomposition model includes:

[0013] Initialize the algorithm parameters corresponding to the Black Kite optimization algorithm, and set the first upper and lower limits of the number of decomposition layers and the second upper and lower limits of the quadratic penalty factor in the preset variational mode decomposition model;

[0014] Initializing the number of decomposition layers and the quadratic penalty factor, and decomposing the original signal using the variational mode decomposition model to obtain a plurality of intrinsic mode training components;

[0015] Calculating the fitness value of each of the intrinsic mode training components at the current moment using an envelope entropy fitness function;

[0016] If the current number of iterations has not reached the maximum number of iterations and the fitness value has decreased, the number of decomposition layers is updated within the first upper and lower limits according to the preset gradient, and the quadratic penalty factor is updated within the second upper and lower limits until the maximum number of iterations is obtained, thereby obtaining the target variational mode decomposition model.

[0017] Optionally, calling the target variational mode decomposition model to decompose the original signal to obtain multiple eigenmode components includes:

[0018] Converting the target variational mode decomposition model into an augmented Lagrangian function model according to the quadratic penalty factor and the Lagrangian multiplier;

[0019] Inputting the original signal into the augmented Lagrangian function model;

[0020] Taking the minimum point as the target, the augmented Lagrangian function model is iteratively solved by using the alternating direction multiplier method until the preset convergence condition is met, thereby obtaining a plurality of eigenmode components.

[0021] Optionally, calculating the autocorrelation coefficients corresponding to the eigenmode components, and selecting at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficients and a preset noise threshold, includes:

[0022] Calculating the autocorrelation coefficients corresponding to the eigenmode components respectively;

[0023] comparing the autocorrelation coefficient with a preset noise threshold;

[0024] At least one eigenmode component whose autocorrelation coefficient is greater than the preset noise threshold is selected as an oscillation characteristic component.

[0025] Optionally, the using a preset energy operator to perform oscillation mode identification on each of the oscillation characteristic components to determine the corresponding dominant high-frequency oscillation information includes:

[0026] Calculating the local energy value of each of the oscillation characteristic components at the corresponding sampling moment using a preset energy operator;

[0027] According to the local energy value and the discrete expression of the oscillation characteristic component, the dominant high-frequency oscillation information corresponding to the oscillation characteristic component is determined.

[0028] Optionally, the dominant high-frequency oscillation information includes amplitude, normalized frequency and damping ratio:

[0029]

[0030] in, is the damping ratio of the kth oscillation characteristic component, is the amplitude of the kth oscillation characteristic component, is the normalized frequency of the kth oscillation characteristic component, is the normalized attenuation coefficient of the kth oscillation characteristic component, is the local energy value of the mth sampling point, is the local energy difference between the m+1th sampling point and the m-1th sampling point.

[0031] Optionally, the method further includes:

[0032] Reconstructing the signal according to all the dominant high-frequency oscillation information to obtain a reconstructed signal;

[0033] Calculating the root mean square error, mean absolute error, and signal correlation between the original signal and the reconstructed signal;

[0034] According to the root mean square error, the mean absolute error and the signal correlation, it is determined whether the dominant high-frequency oscillation information meets the extraction requirements.

[0035] A second aspect of the present invention provides a device for identifying high-frequency oscillation modes of an electric power system, comprising:

[0036] A signal acquisition module, used to collect original signals from the power system;

[0037] The model optimization module is used to optimize the parameters of the preset variational mode decomposition model using the Black Kite optimization algorithm to obtain the target variational mode decomposition model;

[0038] A signal decomposition module, configured to call the target variational mode decomposition model to decompose the original signal to obtain a plurality of eigenmode components;

[0039] a modal component screening module, configured to calculate the autocorrelation coefficient corresponding to each of the eigenmodal components, and select at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficient and a preset noise threshold;

[0040] The oscillation mode identification module is used to use a preset energy operator to perform oscillation mode identification on each of the oscillation characteristic components to determine the corresponding dominant high-frequency oscillation information.

[0041] A third aspect of the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the high-frequency oscillation mode identification method of the power system as described in any one of the first aspects of the present invention.

[0042] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed, the method for identifying high-frequency oscillation modes of an electric power system as described in any one of the first aspects of the present invention is implemented.

[0043] It can be seen from the above technical solutions that the present invention has the following advantages:

[0044] The present invention collects raw signals from the power system; uses the Black Kite optimization algorithm to optimize the parameters of a preset variational modal decomposition model to obtain a target variational modal decomposition model; then uses the target variational modal decomposition model to decompose the raw signals to obtain multiple intrinsic modal components; calculates the autocorrelation coefficient corresponding to each intrinsic modal component, and selects at least one oscillation characteristic component based on the comparison between the autocorrelation coefficient and a preset noise threshold; and uses a preset energy operator to perform oscillation mode identification on each oscillation characteristic component to determine the corresponding dominant high-frequency oscillation information. This method enables the identification of high-frequency oscillation information in broadband power grid signals, effectively preventing the loss of important oscillation information in broadband signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 A flowchart of a method for identifying high-frequency oscillation modes in an electric power system according to an embodiment of the present invention;

[0047] Figure 2 A flow chart of a parameter optimization process for a variational mode decomposition model provided by an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of a process for identifying dominant high-frequency oscillation information provided by an embodiment of the present invention;

[0049] Figure 4 This is a structural block diagram of a high-frequency oscillation mode identification device for an electric power system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0050] Embodiments of the present invention provide a method, apparatus, device, and medium for identifying high-frequency oscillation modes in a power system, which are used to address the technical problem that high-frequency oscillation signals may appear in a wide frequency domain, and the oscillation signals of the MMC-HVDC system have the characteristics of multiple oscillation frequencies and time-varying oscillations. Directly adopting the above-mentioned scheme is likely to cause the loss of important oscillation information in the wide-band signal.

[0051] In order to make the purpose, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0052] See also Figure 1 , Figure 1 A flowchart of the steps of a method for identifying high-frequency oscillation modes of an electric power system provided by an embodiment of the present invention.

[0053] The present invention provides a method for identifying high-frequency oscillation modes of an electric power system, comprising:

[0054] Step 101, collecting original signals from the power system;

[0055] The original signal refers to the current or voltage sequence signal collected by the data acquisition device in the power grid system according to a certain sampling frequency and within a certain sampling time, which is expressed as , the sampling time is T, the sampling frequency is The power system may be a flexible DC transmission system or other power systems that require high-frequency oscillation identification.

[0056] In this embodiment, by setting up data acquisition equipment in the power system to obtain grid voltage or current signals, the sampling frequency can meet the Nyquist sampling theorem (usually not less than twice the highest frequency of the signal).

[0057] Step 102: Optimize the parameters of the preset variational mode decomposition model using the Black Kite optimization algorithm to obtain a target variational mode decomposition model;

[0058] In this embodiment, parameters are initialized, including the Black-winged Kite Algorithm (BKA) parameters, the population size (pop), and the maximum number of iterations (T). After setting the number of decomposition levels (K) and the upper and lower limits of the quadratic penalty factor (α) for the variational mode decomposition (VMD) model, K and α are optimized until the maximum number of iterations is met, resulting in the target variational mode decomposition model. During the optimization process, the values of K and α are integers.

[0059] In one example of the present invention, step 102 may include the following sub-steps:

[0060] Initialize the algorithm parameters corresponding to the Black Kite optimization algorithm, and set the first upper and lower limits of the number of decomposition layers in the preset variational mode decomposition model and the second upper and lower limits of the quadratic penalty factor;

[0061] Initialize the number of decomposition layers and the quadratic penalty factor, and decompose the original signal using the variational mode decomposition model to obtain multiple intrinsic mode training components;

[0062] The envelope entropy fitness function is used to calculate the fitness value of each eigenmode training component at the current moment;

[0063] If the current number of iterations has not reached the maximum number of iterations and the fitness value has decreased, the number of decomposition layers is updated within the first upper and lower limits according to the preset gradient, and the quadratic penalty factor is updated within the second upper and lower limits until the maximum number of iterations is obtained, and the target variational mode decomposition model is obtained.

[0064] The Black Kite Optimization Algorithm is a new population intelligence optimization algorithm. Its idea is derived from the attack and migration behavior of the Black Kite population. Based on the highly adaptable and intelligent behavior of the Black Kite, the algorithm can handle complex problems and has the advantages of strong evolutionary ability and fast search speed.

[0065] After initializing the population, the algorithm selects the individual with the lowest fitness (using the minimum fitness problem as an example) as the leader of the population, which corresponds to the optimal solution in the initial state. The algorithm then solves the optimal value of the objective function based on the behavior of the black kite population, which mainly includes two stages: attack behavior and migration behavior:

[0066] (1) Aggressive behavior:

[0067] The BKA algorithm includes different attack behaviors, corresponding to the black kite's global exploration and search for prey. Its mathematical model is shown in the following formula:

[0068]

[0069]

[0070] Where, represents the value of the jth dimension of the i-th black kite during the t-th iteration, r is a random number in the interval [0,1], p is a constant of 0.9, and T is the total number of iterations set.

[0071] (2) Migration behavior

[0072] Migration is led by the aforementioned leader, and the migration principle is as follows: if the fitness of the current population is better than that of the random population, the population will reach the destination and remain the leader; if the fitness of the current population is worse than that of the random population, the current leader will give up the leadership position and join the migrating population. This process improves the global search capability of the BKA algorithm, and its corresponding mathematical model is shown in the following formula:

[0073]

[0074]

[0075] Where, Represents the data value of the j-th dimension of the population leader during the t-th iteration; represents the fitness of the i-th black kite in the t-th iteration; represents the fitness of an arbitrarily selected black kite; C(0,1) is the Cauchy mutation.

[0076] In this embodiment, the decomposition effect of the variational modal decomposition model is affected by the number of decomposition layers K and the quadratic penalty factor α. A smaller K value cannot guarantee the complete decomposition of the dominant oscillation signal in the original signal, while a larger K value may lead to the generation of false components and the appearance of modal aliasing. The selection of α mainly affects the bandwidth of the intrinsic mode function (IMF). The above two parameters have irregular effects on the decomposition results. Therefore, it is necessary to further find the optimal parameter combination of VMD to ensure the decomposition and extraction of the original signal oscillation information. In this embodiment, the black kite optimization algorithm is used to optimize the parameters of the variational modal decomposition model to obtain the target variational modal decomposition model.

[0077] like Figure 2 As shown, Figure 2 A flow chart of a parameter optimization process of a variational mode decomposition model provided by an embodiment of the present invention is shown.

[0078] The original signal, i.e., the sampled data f(t), the sampling frequency fs, and the number of sampling points N, are input into the variational mode decomposition model. The parameters of the BKS algorithm are initialized, i.e., the population size pop, the maximum number of iterations T, and the current number of iterations t = 0. The first upper and lower bounds of the number of decomposition levels K of the variational mode decomposition model and the second upper and lower bounds of the quadratic penalty factor α are set for initialization.

[0079] Using envelope entropy as the fitness function, the BKA algorithm optimizes the number of decomposition levels K and the quadratic penalty factor α. In each iteration, the BKA algorithm continuously updates the individual positions of black kites based on its own attack and migration behavior simulation mechanism (corresponding to the position update formula in the algorithm), thereby tentatively updating the values of K and α. During this process, the decision to retain the new parameter values is determined by determining whether the fitness function (IMF envelope entropy) decreases. If the fitness function decreases, the new K and α parameter combination is retained and the iteration continues; if not, the original parameter values are retained and the iteration continues. This process continues until the maximum number of iterations T is reached. After the maximum number of iterations is reached, the optimal K and α parameter combination optimized by the BKA algorithm is output, resulting in the target variational mode decomposition model for subsequent decomposition of the original signal.

[0080] Among them, the minimum envelope entropy of the IMF component after VMD decomposition is used as the fitness function for optimization. The envelope entropy represents the sparse characteristics of the signal. The smaller the envelope entropy, the smaller the noise in the IMF component, the more characteristic information, and the better the VMD decomposition effect. The envelope entropy value E of the IMF component is p The equation to be solved is as follows:

[0081]

[0082] Where, or is the envelope signal obtained by Hilbert transform. It is the normalization processing of the envelope signal.

[0083] Step 103: calling the target variational mode decomposition model to decompose the original signal to obtain multiple eigenmode components;

[0084] In this embodiment, after completing the parameter optimization, the target variational modal decomposition model is called to construct an augmented Lagrangian function model. After the original signal is input into the augmented Lagrangian function model, the augmented Lagrangian function model is iteratively solved using the alternating direction multiplier method to achieve the decomposition of the original signal and generate multiple eigenmode components.

[0085] In one example of the present invention, step 103 may include the following sub-steps:

[0086] According to the quadratic penalty factor and Lagrangian multiplier, the target variational mode decomposition model is converted into an augmented Lagrangian function model;

[0087] Input the original signal into the augmented Lagrangian function model;

[0088] Taking the minimum point as the target, the augmented Lagrangian function model is iteratively solved by the alternating direction multiplier method until the preset convergence conditions are met and multiple eigenmode components are obtained.

[0089] In this embodiment, after obtaining the quadratic penalty factor and the number of decomposition layers, the target variational mode decomposition model established is as follows:

[0090]

[0091] Where t is time; f(t) is the original input signal; u k (t) is the k intrinsic mode function IMF components; ω k is the center frequency of each IMF component; ∂t is the unilateral spectrum of each IMF component obtained by Hilbert transform HT, where δ(t) is the pulse unit function.

[0092] To solve the above constrained variational model, a quadratic penalty factor α and a Lagrangian multiplier λ are introduced to transform the original constrained problem into an unconstrained problem, that is, to convert the target variational mode decomposition model into an augmented Lagrangian function model, as shown below:

[0093]

[0094] in, Express request and The inner product of .

[0095] After obtaining the above augmented Lagrangian function model, the alternating direction multiplier method is used to iterate it multiple times to search for the minimum point of the function. The variable u k (ω), λ(ω), and ω k The iterative update expression is:

[0096]

[0097]

[0098]

[0099] In the iterative formula, the original time domain variables f(t), u(t) and λ(t) are converted into frequency domain forms f(ω), u(ω) and λ(ω), n is the number of iterative updates, and τ is the fidelity parameter.

[0100] When the IMF components in the iteration meet the following convergence conditions, the iteration ends and VMD decomposes the output original signal into K intrinsic mode components:

[0101]

[0102] in, is the state of the kth eigenmode component at the nth iteration, is the state of the k+1th eigenmode component at the nth iteration, is a pre-set convergence threshold.

[0103] Step 104, calculating the autocorrelation coefficient corresponding to each eigenmode component, and selecting at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficient and a preset noise threshold;

[0104] In this embodiment, since the primary objective is to identify high-frequency oscillation signals within the broadband range of the power grid, no noise reduction processing is performed on the original signal before analysis to avoid loss of the oscillation signal. Therefore, the acquired IMF component combination contains both the signal's characteristic components and noise components. Using the autocorrelation coefficient calculation formula, the autocorrelation coefficient corresponding to each intrinsic mode component is calculated. The autocorrelation coefficient is compared with a preset noise threshold, and based on the comparison result, the eigenmode component with an autocorrelation coefficient greater than the preset noise threshold is selected as the oscillation characteristic component.

[0105] In one example of the present invention, step 104 may include the following sub-steps:

[0106] Calculate the autocorrelation coefficient corresponding to each eigenmode component;

[0107] Comparing the autocorrelation coefficient with a preset noise threshold;

[0108] At least one eigenmode component whose autocorrelation coefficient is greater than a preset noise threshold is selected as an oscillation characteristic component.

[0109] In this embodiment, the autocorrelation coefficient of the signal is used to filter out the noise components in the IMF. The noise in the power grid appears randomly at any time, and its autocorrelation coefficient is small; while the IMF component containing characteristic information will oscillate periodically according to the center frequency, and its autocorrelation coefficient is large. A threshold Corr_min is set before screening, and the IMF components with autocorrelation coefficients less than Corr_min are regarded as noise components and filtered out. The calculation formula is as follows:

[0110]

[0111] Where, is the sampling duration of the IMF signal, is the arithmetic mean of the eigenmode components, is the signal value of the intrinsic mode function at time t, that is, the eigenmode component at time t; is the autocorrelation coefficient between the IMF signal and the signal after time shift τ.

[0112] Step 105 : Using a preset energy operator to perform oscillation mode identification on each oscillation characteristic component, and determining the corresponding dominant high-frequency oscillation information.

[0113] In this embodiment, the Teager-Kaiser energy operator algorithm is used to identify the oscillation modes of the nonlinear and non-stationary oscillation characteristic components to determine the dominant high-frequency oscillation information corresponding to each oscillation characteristic component. The dominant high-frequency oscillation information includes but is not limited to signal frequency, amplitude and damping ratio.

[0114] In one example of the present invention, step 105 may include the following sub-steps:

[0115] The local energy value of each oscillation characteristic component at the corresponding sampling moment is calculated using a preset energy operator;

[0116] According to the local energy value and the discrete expression of the oscillation characteristic component, the dominant high-frequency oscillation information corresponding to the oscillation characteristic component is determined.

[0117] In this embodiment, since the actual original signal is usually discrete data, the TKEO energy operator in discrete form is:

[0118]

[0119] Where,d (c(m)) is the TKEO energy operator value, m is the sampling point sequence, Δt is the sampling interval time, and c(m) is the signal data of the mth sampling point.

[0120] Taking the signal attenuation coefficient into consideration, the discrete expression of each oscillation characteristic component is as follows:

[0121]

[0122] in, is the amplitude of the kth oscillation characteristic component, is the amplitude of the kth oscillation characteristic component, is the normalized frequency of the kth oscillation characteristic component, is the normalized attenuation coefficient of the kth oscillation characteristic component.

[0123] Combining the above two equations, the parameter values of the oscillation characteristic components can be calculated as follows, that is, the dominant high-frequency oscillation information including amplitude, normalized frequency and damping ratio is:

[0124]

[0125] in, is the damping ratio of the kth oscillation characteristic component, is the amplitude of the kth oscillation characteristic component, is the normalized frequency of the kth oscillation characteristic component, is the normalized attenuation coefficient of the kth oscillation characteristic component, is the local energy value of the mth sampling point, is the local energy difference between the m+1th sampling point and the m-1th sampling point.

[0126] After the above values are calculated, the modal identification of the broadband oscillation signal of the power grid is realized.

[0127] In one example of the present invention, the method further includes the following steps S11-S13:

[0128] S11, reconstructing the signal according to all the dominant high-frequency oscillation information to obtain a reconstructed signal;

[0129] In this embodiment, the dominant high-frequency oscillation information obtained by step 105, i.e., the amplitude ,frequency , damping ratio The original signal is linearly superimposed and reconstructed to obtain the reconstructed signal.

[0130] Furthermore, if the original signal is in discrete form, the reconstructed signal is converted from a continuous-time signal to discrete sampling point data to obtain a reconstructed signal in discrete form with the same dimension as the original signal.

[0131] S12, calculating the root mean square error, the mean absolute error, and the signal correlation between the original signal and the reconstructed signal;

[0132] In this embodiment, by performing error analysis and correlation analysis on the original signal and the reconstructed signal, the specific index calculation process is as follows:

[0133] The root mean square error RMSE is:

[0134] The mean absolute error MAE is:

[0135] Signal correlation R 2 for:

[0136] Where n is the length of the original signal, is the i-th sampling point data of the original signal, is the i-th data of the reconstructed signal, is the average value of the original signal.

[0137] S13. Determine whether the dominant high-frequency oscillation information meets the extraction requirements according to the root mean square error, the mean absolute error, and the signal correlation.

[0138] In this embodiment, the root mean square error (RMSE) measures the average square of the difference between two signals and is obtained by taking the square root. It can intuitively reflect the size of the error and has a high sensitivity to large errors. The mean absolute error (MAE) measures the average absolute error of the two signals. The smaller the above two indicators, the closer the signals are. Signal correlation (R 2 ) represents the goodness of fit of the signal. The closer it is to 1, the higher the correlation between the two signals.

[0139] Specifically, if the root mean square error is less than or equal to the first preset threshold, the absolute error mean is less than or equal to the second preset threshold, and the absolute difference between the signal correlation and 1 is less than or equal to the third preset threshold, then it indicates that the dominant high-frequency oscillation information at this time meets the extraction requirements, and the dominant high-frequency oscillation information at the current moment is output.

[0140] If the extraction requirements are not met, the process may return to step 102 to further optimize the VMD parameters, or perform a gradient adjustment on the screening threshold of the autocorrelation coefficient and then re-execute the process.

[0141] See also Figure 3 , Figure 3A schematic diagram of a process for identifying dominant high-frequency oscillation information provided by an embodiment of the present invention.

[0142] In this embodiment, the target variational mode decomposition model is used to obtain the optimal K and α parameter combination, and the sampled signal f(t) is subjected to VMD decomposition using this parameter combination to obtain K IMFs (intrinsic mode components). min , calculate the autocorrelation coefficient Corr of each of the K IMF components, let k = 1. Determine whether the Corr value of the kth IMF is greater than Corr min If greater than, the kth IMF is determined to be an oscillatory component. If not, the kth IMF is determined to be a noise component. Determine whether k > K: If so, proceed to the next step. If not, set k = k + 1 and return to determine the next IMF. Use the TKEO algorithm to calculate the amplitude, frequency, and damping ratio of the oscillatory component. Calculate the root mean square error, mean absolute error, and signal correlation to evaluate signal identification performance.

[0143] In an embodiment of the present invention, raw signals are collected from the power system; a Black Kite optimization algorithm is used to optimize the parameters of a preset variational modal decomposition model to obtain a target variational modal decomposition model; the target variational modal decomposition model is used to decompose the raw signals to obtain multiple intrinsic modal components; the autocorrelation coefficient corresponding to each intrinsic modal component is calculated, and at least one oscillation characteristic component is selected based on the comparison result of the autocorrelation coefficient with a preset noise threshold; and a preset energy operator is used to perform oscillation mode identification on each oscillation characteristic component to determine the corresponding dominant high-frequency oscillation information. This enables the identification of high-frequency oscillation information in the broadband signal of the power grid, effectively preventing the loss of important oscillation information in the broadband signal.

[0144] See also Figure 4 , Figure 4 The figure shows a structural block diagram of a high-frequency oscillation mode identification device for a power system according to an embodiment of the present invention.

[0145] An embodiment of the present invention provides a device for identifying high-frequency oscillation modes of an electric power system, comprising:

[0146] The signal acquisition module 401 is used to collect original signals from the power system;

[0147] A model optimization module 402 is configured to optimize parameters of a preset variational mode decomposition model using a Black Kite optimization algorithm to obtain a target variational mode decomposition model;

[0148] The signal decomposition module 403 is used to call the target variational mode decomposition model to decompose the original signal to obtain multiple eigenmode components;

[0149] The modal component screening module 404 is configured to calculate the autocorrelation coefficient corresponding to each eigenmodal component and select at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficient and a preset noise threshold;

[0150] The oscillation mode identification module 405 is used to perform oscillation mode identification on each oscillation characteristic component using a preset energy operator to determine the corresponding dominant high-frequency oscillation information.

[0151] Optionally, the model optimization module 402 is specifically configured to:

[0152] Initialize the algorithm parameters corresponding to the Black Kite optimization algorithm, and set the first upper and lower limits of the number of decomposition layers in the preset variational mode decomposition model and the second upper and lower limits of the quadratic penalty factor;

[0153] Initialize the number of decomposition layers and the quadratic penalty factor, and decompose the original signal using the variational mode decomposition model to obtain multiple intrinsic mode training components;

[0154] The envelope entropy fitness function is used to calculate the fitness value of each eigenmode training component at the current moment;

[0155] If the current number of iterations has not reached the maximum number of iterations and the fitness value has decreased, the number of decomposition layers is updated within the first upper and lower limits according to the preset gradient, and the quadratic penalty factor is updated within the second upper and lower limits until the maximum number of iterations is obtained, and the target variational mode decomposition model is obtained.

[0156] Optionally, the signal decomposition module 403 is specifically configured to:

[0157] According to the quadratic penalty factor and Lagrangian multiplier, the target variational mode decomposition model is converted into an augmented Lagrangian function model;

[0158] Input the original signal into the augmented Lagrangian function model;

[0159] Taking the minimum point as the target, the augmented Lagrangian function model is iteratively solved by the alternating direction multiplier method until the preset convergence conditions are met and multiple eigenmode components are obtained.

[0160] Optionally, the modal component screening module 404 is specifically configured to:

[0161] Calculate the autocorrelation coefficient corresponding to each eigenmode component;

[0162] Comparing the autocorrelation coefficient with a preset noise threshold;

[0163] At least one eigenmode component whose autocorrelation coefficient is greater than a preset noise threshold is selected as an oscillation characteristic component.

[0164] Optionally, the oscillation mode identification module 405 is specifically configured to:

[0165] The local energy value of each oscillation characteristic component at the corresponding sampling moment is calculated using a preset energy operator;

[0166] According to the local energy value and the discrete expression of the oscillation characteristic component, the dominant high-frequency oscillation information corresponding to the oscillation characteristic component is determined.

[0167] Optionally, the dominant high-frequency oscillation information includes the amplitude, normalized frequency and damping ratio as follows:

[0168]

[0169] in, is the damping ratio of the kth oscillation characteristic component, is the amplitude of the kth oscillation characteristic component, is the normalized frequency of the kth oscillation characteristic component, is the normalized attenuation coefficient of the kth oscillation characteristic component, is the local energy value of the mth sampling point, is the local energy difference between the m+1th sampling point and the m-1th sampling point.

[0170] Optionally, the device further includes an identification and evaluation module, specifically configured to:

[0171] Reconstruct the signal according to all the dominant high-frequency oscillation information to obtain a reconstructed signal;

[0172] Calculate the root mean square error, mean absolute error and signal correlation between the original signal and the reconstructed signal;

[0173] According to the root mean square error, absolute error mean and signal correlation, it is judged whether the dominant high-frequency oscillation information meets the extraction requirements.

[0174] An embodiment of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the high-frequency oscillation mode identification method of the power system as described in any embodiment of the present invention.

[0175] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed, the method for identifying high-frequency oscillation modes of a power system as described in any embodiment of the present invention is implemented.

[0176] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0177] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is merely a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0178] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of the present embodiment according to actual needs.

[0179] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing module, or each module may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or software functional modules.

[0180] If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing an electronic device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0181] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying high-frequency oscillation modes of an electric power system, characterized in that: include: Collect raw signals from the power system; The Black Kite optimization algorithm is used to optimize the parameters of the preset variational mode decomposition model to obtain the target variational mode decomposition model; Decomposing the original signal by calling the target variational mode decomposition model to obtain a plurality of eigenmode components; Calculating the autocorrelation coefficients corresponding to the eigenmode components, and selecting at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficients and a preset noise threshold; A preset energy operator is used to perform oscillation mode identification on each of the oscillation characteristic components to determine the corresponding dominant high-frequency oscillation information.

2. The method according to claim 1, characterized in that The method of using the Black Kite optimization algorithm to optimize the parameters of a preset variational modal decomposition model to obtain a target variational modal decomposition model includes: Initialize the algorithm parameters corresponding to the Black Kite optimization algorithm, and set the first upper and lower limits of the number of decomposition layers and the second upper and lower limits of the quadratic penalty factor in the preset variational mode decomposition model; Initializing the number of decomposition layers and the quadratic penalty factor, and decomposing the original signal using the variational mode decomposition model to obtain a plurality of intrinsic mode training components; Calculating the fitness value of each of the intrinsic mode training components at the current moment using an envelope entropy fitness function; If the current number of iterations has not reached the maximum number of iterations and the fitness value has decreased, the number of decomposition layers is updated within the first upper and lower limits according to the preset gradient, and the quadratic penalty factor is updated within the second upper and lower limits until the maximum number of iterations is obtained, thereby obtaining the target variational mode decomposition model.

3. The method according to claim 2, characterized in that The calling of the target variational mode decomposition model to decompose the original signal to obtain multiple eigenmode components includes: Converting the target variational mode decomposition model into an augmented Lagrangian function model according to the quadratic penalty factor and the Lagrangian multiplier; Inputting the original signal into the augmented Lagrangian function model; Taking the minimum point as the target, the augmented Lagrangian function model is iteratively solved by using the alternating direction multiplier method until the preset convergence condition is met, thereby obtaining a plurality of eigenmode components.

4. The method according to claim 1, wherein The calculating the autocorrelation coefficients corresponding to the eigenmode components, and selecting at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficients and a preset noise threshold, includes: Calculating the autocorrelation coefficients corresponding to the eigenmode components respectively; comparing the autocorrelation coefficient with a preset noise threshold; At least one eigenmode component whose autocorrelation coefficient is greater than the preset noise threshold is selected as an oscillation characteristic component.

5. The method according to claim 1, wherein The method of using a preset energy operator to perform oscillation mode identification on each of the oscillation characteristic components to determine the corresponding dominant high-frequency oscillation information includes: Calculating the local energy value of each of the oscillation characteristic components at the corresponding sampling moment using a preset energy operator; According to the local energy value and the discrete expression of the oscillation characteristic component, the dominant high-frequency oscillation information corresponding to the oscillation characteristic component is determined.

6. The method according to claim 5, characterized in that The dominant high-frequency oscillation information includes amplitude, normalized frequency and damping ratio: in, is the damping ratio of the kth oscillation characteristic component, is the amplitude of the kth oscillation characteristic component, is the normalized frequency of the kth oscillation characteristic component, is the normalized attenuation coefficient of the kth oscillation characteristic component, is the local energy value of the mth sampling point, is the local energy difference between the m+1th sampling point and the m-1th sampling point.

7. The method according to claim 1, characterized in that The method further comprises: Reconstructing the signal according to all the dominant high-frequency oscillation information to obtain a reconstructed signal; Calculating the root mean square error, mean absolute error, and signal correlation between the original signal and the reconstructed signal; According to the root mean square error, the mean absolute error and the signal correlation, it is determined whether the dominant high-frequency oscillation information meets the extraction requirements.

8. A high-frequency oscillation mode identification device for an electric power system, characterized in that: include: A signal acquisition module, used to collect original signals from the power system; The model optimization module is used to optimize the parameters of the preset variational mode decomposition model using the Black Kite optimization algorithm to obtain the target variational mode decomposition model; A signal decomposition module, configured to call the target variational mode decomposition model to decompose the original signal to obtain a plurality of eigenmode components; a modal component screening module, configured to calculate the autocorrelation coefficient corresponding to each of the eigenmodal components, and select at least one oscillation characteristic component according to a comparison result between the autocorrelation coefficient and a preset noise threshold; The oscillation mode identification module is used to use a preset energy operator to perform oscillation mode identification on each of the oscillation characteristic components to determine the corresponding dominant high-frequency oscillation information.

9. An electronic device, characterized in that: The invention comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the high-frequency oscillation mode identification method of the power system according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the method for identifying high-frequency oscillation modes of an electric power system according to any one of claims 1 to 7 is implemented.

Images