Broadband oscillation signal identification method, device, equipment, medium and product

Abstract

The invention discloses a broadband oscillation signal identification method, device and equipment, a medium and a product. The method comprises the following steps: acquiring a broadband oscillation signal to be identified; determining a target feature vector corresponding to the broadband oscillation signal; and inputting the target feature vector into a target network for classification and identification to obtain an identification result. Through the technical scheme of the invention, the accuracy of broadband oscillation type identification of the power system can be effectively improved, accurate identification of oscillation signals in different frequency ranges can be realized through an improved signal feature extraction technology, compared with the prior art, the embodiment of the invention can more effectively process non-stable and non-linear vibration signals, and the identification accuracy of the broadband oscillation type of the power system can be improved. And weak fault features are extracted in a complex noise environment, so that the precision and robustness of oscillation identification of the power system are improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of oscillation identification, and in particular to a method, device, equipment, medium and product for identifying a broadband oscillation signal. Background Art

[0002] Compared with traditional generators, new energy units are often connected to the grid through power electronic converters, forming a "double-high" power system with a high proportion of renewable energy and a high proportion of power electronic equipment. The interaction between power electronic equipment and the grid may cause a series of oscillation problems, with oscillation frequencies ranging from a few Hz to thousands of Hz, known as power system broadband oscillations, which seriously affect the stability of the power system and the safe operation of equipment. These oscillations not only affect the quality of power, but may also cause equipment damage and DC locking, seriously threatening the safe and stable operation of the power grid. Summary of the invention

[0003] Embodiments of the present invention provide a method, device, equipment, medium and product for identifying a broadband oscillation signal, so as to improve the accuracy of identifying broadband oscillations in a power system.

[0004] According to one aspect of the present invention, a method for identifying a broadband oscillation signal is provided, comprising:

[0005] Acquire a broadband oscillation signal to be identified;

[0006] Determining a target feature vector corresponding to the broadband oscillation signal;

[0007] The target feature vector is input into the target network for classification and identification to obtain an identification result.

[0008] In some embodiments of the present invention, determining a target feature vector corresponding to the broadband oscillation signal includes:

[0009] Decomposing the broadband oscillation signal to obtain an eigenmode component set;

[0010] Determining a target eigenmode component subset from the eigenmode component set;

[0011] Extracting the time domain characteristic parameters and the frequency domain characteristic parameters corresponding to each target intrinsic mode component in the target intrinsic mode component subset, and determining a high-dimensional characteristic vector according to the time domain characteristic parameters and the frequency domain characteristic parameters corresponding to each target intrinsic mode component;

[0012] Based on the kernel principal component analysis algorithm, the high-dimensional feature vector is reduced in dimension to obtain a target feature vector.

[0013] In some embodiments of the present invention, the broadband oscillation signal is decomposed to obtain an eigenmode component set, including:

[0014] Adding target white noise to the broadband oscillation signal to obtain an updated broadband oscillation signal;

[0015] Obtaining a residual corresponding to the updated broadband oscillation signal, and determining a current intrinsic mode component according to the residual and the updated broadband oscillation signal;

[0016] The residual is used as a new broadband oscillation signal, and the operation of adding target white noise to the broadband oscillation signal is returned to be executed until the number of iterations reaches the target number, thereby obtaining a set of intrinsic mode components.

[0017] In some embodiments of the present invention, determining a target eigenmode component subset from the eigenmode component set includes:

[0018] Determining the correlation between each eigenmode component in the set of eigenmode components and the broadband oscillation signal;

[0019] According to the correlation corresponding to each eigenmode component, a target number of target eigenmode components are screened out from the eigenmode component set to form a target eigenmode component subset.

[0020] In some embodiments of the present invention, determining the correlation between each eigenmode component in the set of eigenmode components and the broadband oscillation signal comprises:

[0021] Get the Pearson correlation coefficient;

[0022] The Pearson correlation coefficient is used to measure the correlation between each eigenmode component in the eigenmode component set and the broadband oscillation signal.

[0023] In some embodiments of the present invention, the target network consists of an input layer, a hidden layer, and an output layer, and the number of neurons in the input layer is the number of dimensions of the target feature vector;

[0024] Before inputting the target feature vector into the target network for classification and identification to obtain the identification result, the method further includes:

[0025] The target feature vector is normalized.

[0026] According to another aspect of the present invention, a broadband oscillation signal identification device is provided, the device comprising:

[0027] An acquisition module, used for acquiring a broadband oscillation signal to be identified;

[0028] A determination module, used to determine a target feature vector corresponding to the broadband oscillation signal;

[0029] The identification module is used to input the target feature vector into the target network for classification and identification to obtain an identification result.

[0030] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0031] at least one processor; and

[0032] a memory communicatively connected to the at least one processor; wherein,

[0033] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the wide-band oscillation signal identification method described in any embodiment of the present invention.

[0034] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the wide-band oscillation signal identification method described in any embodiment of the present invention when executed.

[0035] According to another aspect of the present invention, an embodiment of the present invention further provides a computer program product, the computer program product includes a computer program, and when the computer program is executed by a processor, the method for identifying a wide-band oscillation signal according to any embodiment of the present invention is implemented.

[0036] The embodiment of the present invention obtains the broadband oscillation signal to be identified, determines the target feature vector corresponding to the broadband oscillation signal, inputs the target feature vector into the target network for classification and identification, and obtains the identification result. The technical solution of the present invention can effectively improve the accuracy of the identification of broadband oscillation types in the power system, and realizes the accurate identification of oscillation signals in different frequency ranges through improved signal feature extraction technology. Compared with the prior art, the embodiment of the present invention can more effectively process non-stationary and nonlinear vibration signals, and extract weak fault features in complex noise environments, thereby improving the accuracy and robustness of power system oscillation identification.

[0037] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 is a flow chart of a method for identifying a broadband oscillation signal in an embodiment of the present invention;

[0040] Figure 2 is a schematic structural diagram of a broadband oscillation signal identification device in an embodiment of the present invention;

[0041] Figure 3 It is a structural schematic diagram of an electronic device for implementing the broadband oscillation signal identification method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme 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 described embodiments 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 creative work should fall within the scope of protection of the present invention.

[0043] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and their meanings are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0044] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0045] Embodiment 1

[0046] In view of the oscillation problem, the research on oscillation identification technology is very necessary. Power system oscillation identification can be divided into signal model estimation and digital signal decomposition. For complex AC / DC hybrid power grids, signal decomposition is easier than mechanism modeling, and can provide support for subsequent oscillation location and online suppression. The main existing technologies include EMD (Empirical Mode Decomposition), EEMD (Ensemble Empirical Mode Decomposition) and CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise). These technologies have their own advantages and disadvantages when extracting the IMF (Intrinsic Mode Function) of the signal. For example, the EMD method has serious modal aliasing, while EEMD has improved in reducing modal aliasing, but it is still difficult to handle non-stationary and nonlinear vibration signals, and it is difficult to extract weak fault features from complex noise. The CEEMDAN method performs well under complex noise interference, but there are problems such as too many feature dimensions and too long identification time.

[0047] Figure 1 is a flow chart of a method for identifying a broadband oscillation signal in an embodiment of the present invention. This embodiment is applicable to the case of identifying a broadband oscillation signal in a power system. The method can be executed by a broadband oscillation signal identification device in an embodiment of the present invention. The device can be implemented in software and / or hardware, such as Figure 1 As shown, the method specifically comprises the following steps:

[0048] S101, obtaining a broadband oscillation signal to be identified.

[0049] It can be known that a broadband oscillation signal refers to an oscillation signal with a wide frequency range. In this embodiment, the broadband oscillation signal can be a broadband oscillation signal collected in the power system. In the power system, the interaction between power electronic equipment and the power grid may cause a series of oscillation problems. The oscillation frequency ranges from a few Hz to thousands of Hz, which is called broadband oscillation of the power system. The signal collected therefrom is the broadband oscillation signal.

[0050] It is known that the identification of broadband oscillation signals is an important task in signal processing, which usually involves extracting useful information and features from complex signals.

[0051] S102: Determine a target feature vector corresponding to the broadband oscillation signal.

[0052] It should be noted that the target feature vector may be a feature vector corresponding to the broadband oscillation signal after dimension reduction.

[0053] In this embodiment, after the broadband oscillation signal is collected, a series of processing can be performed on the broadband oscillation signal to determine the target feature vector corresponding to the broadband oscillation signal. Exemplarily, ICEEMDAN (Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise, an improved ensemble empirical mode decomposition method, is a signal decomposition technology for processing nonlinear and non-stationary time series data) can be used to decompose the broadband oscillation signal into several intrinsic mode components IMF; then, the five IMFs with the strongest correlation with the broadband oscillation signal are screened out by the Pearson correlation coefficient method, and their characteristic parameters in the time domain and frequency domain are extracted to form a high-dimensional feature vector; then, the KPCA (Kernel Principal Component Analysis) method is used to reduce the dimension of the feature vector to obtain the target feature vector corresponding to the broadband oscillation signal.

[0054] S103, inputting the target feature vector into the target network for classification and identification to obtain an identification result.

[0055] In this embodiment, the target network may be a DNN (Deep Neural Network). DNN is a type of deep learning model that extracts hierarchical features of data by setting multiple hidden layers between the input layer and the output layer. This structure makes DNN more effective in dealing with large-scale nonlinear problems. The training of DNN involves forward propagation, in which data is processed layer by layer to produce output; and back propagation, in which weights are adjusted layer by layer through error feedback to achieve the purpose of optimizing model performance.

[0056] In this embodiment, the identification result may be, for example, a fitting / classification value corresponding to the broadband oscillation signal.

[0057] Specifically, the target feature vector is input into the DNN for classification and identification, and the fitting / classification value corresponding to the broadband oscillation signal is output, thereby obtaining the identification result corresponding to the broadband oscillation signal.

[0058] The embodiment of the present invention obtains the broadband oscillation signal to be identified, determines the target feature vector corresponding to the broadband oscillation signal, inputs the target feature vector into the target network for classification and identification, and obtains the identification result. The technical solution of the present invention can effectively improve the accuracy of the identification of broadband oscillation types in the power system, and realizes the accurate identification of oscillation signals in different frequency ranges through improved signal feature extraction technology. Compared with the prior art, the embodiment of the present invention can more effectively process non-stationary and nonlinear vibration signals, and extract weak fault features in complex noise environments, thereby improving the accuracy and robustness of power system oscillation identification.

[0059] Optionally, determining a target feature vector corresponding to the broadband oscillation signal includes:

[0060] The broadband oscillation signal is decomposed to obtain a set of intrinsic mode components.

[0061] In this embodiment, the broadband oscillation signal may be decomposed by using the ICEEMDAN signal decomposition algorithm, wherein the intrinsic mode component set may be a collection of several intrinsic mode components IMF obtained by decomposing the broadband oscillation signal by using the ICEEMDAN signal decomposition algorithm.

[0062] Specifically, the broadband oscillation signal is decomposed by using the ICEEMDAN signal decomposition algorithm to obtain several intrinsic mode components IMFs, and these intrinsic mode components IMFs constitute an intrinsic mode component set.

[0063] A target eigenmode component subset is determined from the eigenmode component set.

[0064] It should be noted that the target intrinsic mode component may be an eigenmode component with a high correlation with the original broadband oscillation signal selected from the eigenmode component set, and the target eigenmode component subset may be a subset consisting of all target eigenmode components.

[0065] In actual operation, among the decomposed IMF components (i.e., intrinsic mode components), the signal features contained in some IMF components have a weak correlation with the original broadband oscillation signal. Therefore, it is necessary to screen out several IMF components with strong correlation from the eigenmode component set to retain sufficient signal features while avoiding excessive refinement and loss of signal features due to too many IMF components.

[0066] The time domain characteristic parameters and frequency domain characteristic parameters corresponding to each target intrinsic mode component in the target intrinsic mode component subset are extracted, and a high-dimensional characteristic vector is determined according to the time domain characteristic parameters and frequency domain characteristic parameters corresponding to each target intrinsic mode component.

[0067] In actual operation, since the collected broadband oscillation signal is a time domain signal, the most direct method is to perform time domain analysis on the target intrinsic mode component. Exemplarily, the time domain characteristic parameters corresponding to the target intrinsic mode component may include, for example, mean, variance, maximum value, etc. The specific time domain characteristic parameters are shown in Table 1.

[0068] Table 1

[0069]

[0070] In Table 1, X i is a broadband oscillation time domain signal sequence, i=1,2,…,N, N is the number of sample points.

[0071] In order to fully extract the signal features, further frequency domain analysis is required. Frequency domain analysis is to obtain a spectrum diagram by fast Fourier transform of the target intrinsic mode component to achieve the transformation of the time domain to the frequency domain signal. The specific frequency domain feature parameters can be seen in Table 2.

[0072] Table 2

[0073]

[0074] In Table 2, s k is the broadband oscillation signal X i spectrum, k = 1, 2, ..., K, K is the number of spectral lines; f k is the frequency value of the kth spectrum line. Frequency feature 1 reflects the intensity of the frequency domain vibration energy; frequency features 2 to 4, 6, and 10 reflect the dispersion / concentration of the spectrum; and frequency features 5, 7 to 9 reflect the change in the position of the main frequency band.

[0075] The high-dimensional feature vector may refer to a feature vector formed by time domain feature parameters and frequency domain feature parameters corresponding to each target intrinsic mode component.

[0076] In this embodiment, 14 time domain feature parameters and 10 frequency domain feature parameters corresponding to each target intrinsic mode component in the target intrinsic mode component subset are extracted to form a 144-dimensional high-dimensional feature vector.

[0077] Based on the kernel principal component analysis algorithm, the high-dimensional feature vector is reduced to obtain the target feature vector.

[0078] It is known that the kernel principal component analysis algorithm, or KPCA, is an extension of the principal component analysis PCA and is used for nonlinear dimensionality reduction. It maps the data into a high-dimensional space so that the data is linearly separable in the space, and then performs PCA.

[0079] In actual operation, if the high-dimensional feature vector is directly input into the DNN model, it may not only cause feature redundancy and reduce the importance of effective information, but also make the model run too long. Therefore, it is necessary to reduce the dimensionality of the high-dimensional feature vector.

[0080] For example, let the sample points be {x1,x2,…,x N}, N is the total number of samples, then the specific steps of the KPCA method are as follows:

[0081] Step 1: Select the kernel function k. The kernel function is selected as Gaussian kernel, which is defined as:

[0082]

[0083] Among them, ||x i -x j || is a sample x i and x j The Euclidean distance between them, σ is the parameter of the kernel function, which controls the width of the Gaussian function.

[0084] Step 2: Calculate the kernel matrix K. K is an N×N matrix, where each element K ij Represents the inner product of the i-th sample and the j-th sample in the feature space. The calculation formula is:

[0085] K ij = k(x i ,x j );

[0086] Step 3: Center the kernel matrix. KPCA requires the data to be centered in the feature space, so the kernel matrix needs to be centered. The calculation formula is:

[0087]

[0088] in,

[0089] Step 4: Calculate eigenvalues ​​and eigenvectors. Eigenvectors represent the main directions of the data in the original high-dimensional space, while eigenvalues ​​represent the variance of the data in these directions. Perform eigenvalue decomposition and calculate eigenvalues ​​and eigenvectors:

[0090]

[0091] Among them, λ is the eigenvalue of the kernel matrix and α is the corresponding eigenvector.

[0092] Step 5: Select the principal components and normalize them. Select the first d principal components according to the size of the eigenvalues. The corresponding eigenvectors are α1, α2, …, αd. Normalize them and the calculation formula is:

[0093]

[0094] in, is the mth normalized eigenvector.

[0095] Step 6: Project the original sample data into the principal component space. The projection calculation formula in the principal component space is:

[0096]

[0097] Among them, y im is the sample x i The projection on the mth principal component, is the jth element of the mth eigenvector, k(x i ,x j ) is the value of the Gaussian kernel function.

[0098] Step 7: Final representation. All sample points {x1,x2,...,x N After projecting to the principal component space, we get its low-dimensional representation: {y1,y2,...,y N}. Among them, each y i is the sample point x i Representation in principal component space.

[0099] Through the above steps, the high-dimensional features of the broadband oscillation signal can be reduced to a lower dimension, thereby removing redundant features and improving the computational efficiency of the data and the generalization ability of the model.

[0100] Optionally, the broadband oscillation signal is decomposed to obtain a set of intrinsic mode components, including:

[0101] The target white noise is added to the broadband oscillation signal to obtain an updated broadband oscillation signal.

[0102] Different from the traditional EMD method, ICEEMDAN combines adaptive noise and complete integration strategy to improve the stability and accuracy of the decomposition process. Compared with CEEMDAN, which directly adds Gaussian white noise during the decomposition process, the improved ICEEMDAN method uses EMD to decompose white noise, and adds the kth IMF component obtained as auxiliary noise to the previous round of residuals. The local mean of the new residual is subtracted from the previous round of residuals to finally obtain the kth IMF component.

[0103] The target white noise may be white noise added to the broadband oscillation signal. For example, the white noise may be Gaussian white noise.

[0104] Let M(·) be the operator for computing the local mean, E k (·) is the operator that generates the kth IMF component through the EMD algorithm. The specific decomposition process is as follows:

[0105] Step 1: When k = 1, E1[w (i) ] is added to the original signal x. In this embodiment, the original signal x is a broadband oscillation signal.

[0106] x (i) =x+p0E1(w (i) )(i=1,2,...,N);

[0107] Among them, x (i) is the ith signal with white noise added, w (i) is the ith white noise, p0 is the signal-to-noise ratio, and N is the number of white noises added.

[0108] A residual corresponding to the updated broadband oscillation signal is obtained, and a current intrinsic mode component is determined according to the residual and the updated broadband oscillation signal.

[0109] Step 2: Calculate x by EMD (i) The local mean of N signals is calculated, and then the overall mean is calculated for the N signals to obtain the first residual r1, thereby calculating the first IMF component c1.

[0110]

[0111] c1=x-r1.

[0112] The residual is used as a new broadband oscillation signal, and the operation of adding the target white noise to the broadband oscillation signal is returned until the number of iterations reaches the target number, thereby obtaining a set of intrinsic mode components.

[0113] In this embodiment, the target number may be k.

[0114] Step 3: When k ≥ 2, repeat Step 1 and 2, and convert the residual r k-1 As the original signal, add E k (w (i) ), calculate the new residual r k and the kth IMF component c k .

[0115]

[0116] c k =rk-1 -r k .

[0117] Through the above steps, the original signal x is equivalently decomposed into the sum of n IMF components and the residual.

[0118] Optionally, determining a target eigenmode component subset from the eigenmode component set includes:

[0119] A correlation between each eigenmode component in the set of eigenmode components and the broadband oscillating signal is determined.

[0120] In this embodiment, the correlation between each eigenmode component in the eigenmode component set and the broadband oscillation signal may be calculated by using the Pearson correlation coefficient method.

[0121] According to the correlation corresponding to each eigenmode component, a target number of target eigenmode components are screened out from the eigenmode component set to form a target eigenmode component subset.

[0122] In actual operation, the target number may be determined based on actual needs or empirical values, and may be the number of intrinsic mode components with the strongest correlation. Preferably, the target number may be 5.

[0123] Specifically, among the decomposed IMF components, the signal features contained in some components have a weak correlation with the original broadband oscillation signal. Therefore, the five IMF components with strong correlation can be screened by the correlation coefficient method to retain sufficient signal features while avoiding excessive refinement and loss of signal features due to too many IMF components.

[0124] Optionally, determining the correlation between each eigenmode component in the eigenmode component set and the broadband oscillation signal includes:

[0125] Get the Pearson correlation coefficient.

[0126] It can be known that the Pearson correlation coefficient value is between [-1,1]. The closer it is to 1, the stronger the correlation. The formula of the Pearson correlation coefficient is as follows:

[0127]

[0128] Among them, X i is the i-th IMF signal, Y is the original signal, ρ Xi,Y For X i , Pearson correlation coefficient of Y, M is the number of signal samples, and E(·) is the expectation operator.

[0129] The Pearson correlation coefficient is used to measure the correlation between each eigenmode component in the eigenmode component set and the broadband oscillation signal.

[0130] Specifically, based on the Pearson correlation coefficient, the correlation between each intrinsic mode component and the broadband oscillation signal is measured, and then the five IMF components with the strongest correlation are screened out.

[0131] Optionally, the target network consists of an input layer, a hidden layer, and an output layer, and the number of neurons in the input layer is the number of dimensions of the target feature vector.

[0132] In this embodiment, the target network can be a DNN network with two hidden layers. The input layer has m neurons, where m is the dimension of the target feature vector after dimensionality reduction, and the input feature X. The output layer has n neurons, where n can be 1, to output the broadband oscillation category and the fitting / classification value y; since the output is a continuous value, the Softmax activation function can be used. The number of hidden layers is 2, and the number of neurons is p and q respectively; the hidden layer can use the ReLU function as the activation function.

[0133] In the forward propagation stage of DNN, the input of layer i is the output of layer i-1, and the relationship between them is:

[0134] a i =σ(ω i a i-1 +b i );

[0135] Among them, a i is the output vector of the i-th layer; σ(·) is the activation function; ω i and b i The weight matrix and bias vector of the i-th layer are respectively expressed as:

[0136]

[0137]

[0138] Among them, S i is the number of neurons in the i-th layer.

[0139] Before inputting the target feature vector into the target network for classification and identification and obtaining the identification result, it also includes:

[0140] Normalize the target feature vector.

[0141] In order to avoid gradient explosion / vanishing and improve the training speed, stability and generalization ability of the model, this embodiment uses the [0,1] normalization method to normalize the original samples.

[0142] The technical solution of the embodiment of the present invention effectively provides a method for accurately identifying the type of broadband oscillation signals in power systems, and through improved signal decomposition and feature extraction techniques, accurate identification of broadband oscillation signals in different frequency ranges is achieved. Compared with the prior art, the embodiment of the present invention can more effectively process non-stationary and nonlinear vibration signals, and extract weak fault features in complex noise environments, thereby improving the accuracy and robustness of the identification of broadband oscillation signals in power systems. In addition, the embodiment of the present invention has also made improvements in feature dimension optimization and identification time, significantly reducing the occupation of computing resources and time costs, facilitating online real-time monitoring and rapid response, and providing reliable protection for the stability of the power system and the safe operation of equipment.

[0143] In order to verify the effectiveness of the method proposed in the embodiment of the present invention, a Matlab / Simulink simulation model of a four-machine two-zone system including a wind farm is built in the experimental part to verify the reasonable accuracy of the proposed algorithm:

[0144] The number of sampling points is set to 80 / cycle (i.e., the sampling frequency is 4800Hz), and the sampling time is set to 10 seconds. Under different load conditions, step disturbances or sinusoidal disturbance signals of different amplitudes and frequencies are applied to the grid-side controller of the wind farm to form a broadband oscillation data set with a frequency range of 10-40Hz, 70-100Hz, and 130-160Hz. According to the actual signal conditions, 10dB to 20dB of noise is randomly added to the simulation samples.

[0145] The four types of broadband oscillation signals are numbered and labeled 0 to 4 respectively, as shown in Table 3:

[0146] Table 3

[0147]

[0148] The broadband oscillation signal is input into the ICEEMDAN algorithm for decomposition, and the original signal is decomposed into 6 IMF components with different time scales. The Pearson correlation coefficients of these IMF component signals and the original signal are calculated, as shown in Table 4.

[0149] Table 4

[0150]

[0151] The first five IMF components with higher correlation coefficients are selected: IMF4, IMF5, IMF3, IMF2, and IMF1, and their 14 time domain feature parameters and 10 frequency domain feature parameters are calculated to form a 144-dimensional high-dimensional feature vector.

[0152] The experiment uses the cumulative explained variance ratio to select the number of principal components and draws a curve showing that the cumulative explained variance ratio changes with the increase in the number of principal components.

[0153] When the number of principal components is 13, the cumulative explained variance ratio reaches 90%. These 13 features are selected as the final feature vector, reducing the original 144-dimensional features to 5 dimensions, thus achieving data redundancy removal. The first three dimensions are selected to visualize the data.

[0154] The features after dimensionality reduction are normalized and input into the DNN classifier for classification, with the ratio of training set to test set being 4:1. After classification by the DNN classifier, the type of broadband oscillation signal is output.

[0155] In order to verify the superiority of the proposed ICEEMDAN-KPCA-DNN model, five different models are further constructed for effect comparison, including:

[0156] 1) EMD, KPCA and DNN are used to decompose, reduce dimension and classify broadband oscillation signals respectively, which is referred to as EMD-KPCA-DNN model.

[0157] 2) EEMD, KPCA and DNN are used to decompose, reduce dimension and classify broadband oscillation signals respectively, which is referred to as EEMD-KPCA-DNN model.

[0158] 3) CEEMDAN, KPCA and DNN are used to decompose, reduce dimension and classify broadband oscillation signals respectively, which is referred to as CEEMDAN-KPCA-DNN model.

[0159] 4) Use ICEEMDAN, PCA, and DNN to decompose, reduce dimension, and classify broadband oscillation signals, referred to as the ICEEMDAN-PCA-DNN model.

[0160] 5) ICEEMDAN, KPCA and DNN are used to decompose, reduce dimension and classify broadband oscillation signals respectively, which is referred to as ICEEMDAN-KPCA-DNN model.

[0161] After the broadband oscillation signal is decomposed, dimensionally reduced, and classified by the above five different models, the classification accuracy corresponding to the above five different models is determined according to the final classification results. The comparison results are shown in Table 5:

[0162] Table 5

[0163]

[0164] According to Table 5, the model proposed in the embodiment of the present invention has the highest accuracy of 99.55%. The accuracy of EMD, EEMD, and CEEMDAN decomposition increases successively, because the modal aliasing phenomenon of the algorithm decreases successively, so that features can be better extracted and classification accuracy can be improved; KPCA dimensionality reduction is better than PCA, because KPCA can capture the nonlinear structure in the data and solve the limitation of PCA in processing nonlinear data.

[0165] In summary, the broadband oscillation identification method based on ICEEMDAN-KPCA-DNN proposed in the embodiment of the present invention is aimed at the problem of broadband oscillation identification in power system and can quickly and accurately identify broadband oscillation signals.

[0166] Embodiment 2

[0167] Figure 2 Schematic diagram of the structure of a broadband oscillation signal identification device in an embodiment of the present invention. This embodiment is applicable to the case of broadband oscillation signal identification in a power system. The device can be implemented in software and / or hardware. The device can be integrated in any device that provides the function of broadband oscillation signal identification, such as Figure 2 As shown, the broadband oscillation signal identification device specifically includes: an acquisition module 201 , a determination module 202 and an identification module 203 .

[0168] Wherein, the acquisition module 201 is used to acquire the broadband oscillation signal to be identified;

[0169] A determination module 202, configured to determine a target feature vector corresponding to the broadband oscillation signal;

[0170] The identification module 203 is used to input the target feature vector into the target network for classification and identification to obtain an identification result.

[0171] Optionally, the determining module 202 includes:

[0172] A decomposition unit, used for decomposing the broadband oscillation signal to obtain an eigenmode component set;

[0173] A determination unit, configured to determine a target eigenmode component subset from the eigenmode component set;

[0174] An extraction unit, used to extract the time domain characteristic parameters and frequency domain characteristic parameters corresponding to each target intrinsic mode component in the target intrinsic mode component subset, and determine a high-dimensional feature vector according to the time domain characteristic parameters and frequency domain characteristic parameters corresponding to each target intrinsic mode component;

[0175] The dimension reduction unit is used to reduce the dimension of the high-dimensional feature vector based on the kernel principal component analysis algorithm to obtain a target feature vector.

[0176] Optionally, the decomposition unit is specifically used for:

[0177] Adding target white noise to the broadband oscillation signal to obtain an updated broadband oscillation signal;

[0178] Obtaining a residual corresponding to the updated broadband oscillation signal, and determining a current intrinsic mode component according to the residual and the updated broadband oscillation signal;

[0179] The residual is used as a new broadband oscillation signal, and the operation of adding target white noise to the broadband oscillation signal is returned to be executed until the number of iterations reaches the target number, thereby obtaining a set of intrinsic mode components.

[0180] Optionally, the determining unit includes:

[0181] A determination subunit, configured to determine a correlation between each intrinsic mode component in the intrinsic mode component set and the broadband oscillation signal;

[0182] The screening subunit is used to screen out a target number of target eigenmode components from the set of eigenmode components according to the correlation corresponding to each eigenmode component, so as to form a target eigenmode component subset.

[0183] Optionally, the determining subunit is specifically used for:

[0184] Get the Pearson correlation coefficient;

[0185] The Pearson correlation coefficient is used to measure the correlation between each eigenmode component in the eigenmode component set and the broadband oscillation signal.

[0186] Optionally, the target network consists of an input layer, a hidden layer and an output layer, and the number of neurons in the input layer is the number of dimensions of the target feature vector;

[0187] The device also includes:

[0188] A normalization module is used to normalize the target feature vector.

[0189] The above-mentioned product can execute the broadband oscillation signal identification method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0190] Embodiment 3

[0191] Figure 3A schematic diagram of the structure of an electronic device 30 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0192] like Figure 3 As shown, the electronic device 30 includes at least one processor 31, and a memory connected to the at least one processor 31, such as a read-only memory (ROM) 32, a random access memory (RAM) 33, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 31 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 32 or the computer program loaded from the storage unit 38 to the random access memory (RAM) 33. In the RAM 33, various programs and data required for the operation of the electronic device 30 can also be stored. The processor 31, the ROM 32, and the RAM 33 are connected to each other through a bus 34. An input / output (I / O) interface 35 is also connected to the bus 34.

[0193] A number of components in the electronic device 30 are connected to the I / O interface 35, including: an input unit 36, such as a keyboard, a mouse, etc.; an output unit 37, such as various types of displays, speakers, etc.; a storage unit 38, such as a disk, an optical disk, etc.; and a communication unit 39, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 39 allows the electronic device 30 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0194] The processor 31 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 31 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 31 performs the various methods and processes described above, such as the broadband oscillation signal recognition method:

[0195] Acquire a broadband oscillation signal to be identified;

[0196] Determining a target feature vector corresponding to the broadband oscillation signal;

[0197] The target feature vector is input into the target network for classification and identification to obtain an identification result.

[0198] In some embodiments, the broadband oscillation signal identification method can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 38. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 30 via the ROM 32 and / or the communication unit 39. When the computer program is loaded into the RAM 33 and executed by the processor 31, one or more steps of the broadband oscillation signal identification method described above can be performed. Alternatively, in other embodiments, the processor 31 can be configured to perform the broadband oscillation signal identification method in any other appropriate manner (for example, by means of firmware).

[0199] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0200] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0201] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0202] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0203] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0204] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0205] In one embodiment, the present invention further includes a computer program product, the computer program product includes a computer program, and when the computer program is executed by a processor, the method for identifying a broadband oscillation signal according to any embodiment of the present invention is implemented.

[0206] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0207] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0208] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying a broadband oscillation signal, characterized in that: include: Acquire a broadband oscillation signal to be identified; Determining a target feature vector corresponding to the broadband oscillation signal; The target feature vector is input into the target network for classification and identification to obtain an identification result.

2. The method according to claim 1, characterized in that Determining a target feature vector corresponding to the broadband oscillation signal includes: Decomposing the broadband oscillation signal to obtain an eigenmode component set; Determining a target eigenmode component subset from the eigenmode component set; Extracting the time domain characteristic parameters and the frequency domain characteristic parameters corresponding to each target intrinsic mode component in the target intrinsic mode component subset, and determining a high-dimensional characteristic vector according to the time domain characteristic parameters and the frequency domain characteristic parameters corresponding to each target intrinsic mode component; Based on the kernel principal component analysis algorithm, the high-dimensional feature vector is reduced in dimension to obtain a target feature vector.

3. The method according to claim 2, characterized in that The broadband oscillation signal is decomposed to obtain an eigenmode component set, including: Adding target white noise to the broadband oscillation signal to obtain an updated broadband oscillation signal; Obtaining a residual corresponding to the updated broadband oscillation signal, and determining a current intrinsic mode component according to the residual and the updated broadband oscillation signal; The residual is used as a new broadband oscillation signal, and the operation of adding target white noise to the broadband oscillation signal is returned to be executed until the number of iterations reaches the target number, thereby obtaining a set of intrinsic mode components.

4. The method according to claim 2, characterized in that: Determining a target eigenmode component subset from the eigenmode component set includes: Determining the correlation between each eigenmode component in the set of eigenmode components and the broadband oscillation signal; According to the correlation corresponding to each eigenmode component, a target number of target eigenmode components are screened out from the eigenmode component set to form a target eigenmode component subset.

5. The method according to claim 4, characterized in that Determining the correlation between each eigenmode component in the eigenmode component set and the broadband oscillation signal comprises: Get the Pearson correlation coefficient; The Pearson correlation coefficient is used to measure the correlation between each eigenmode component in the eigenmode component set and the broadband oscillation signal.

6. The method according to claim 1, characterized in that The target network consists of an input layer, a hidden layer and an output layer, and the number of neurons in the input layer is the number of dimensions of the target feature vector; Before inputting the target feature vector into the target network for classification and identification to obtain the identification result, the method further includes: The target feature vector is normalized.

7. A broadband oscillation signal identification device, characterized in that: include: An acquisition module, used for acquiring a broadband oscillation signal to be identified; A determination module, used to determine a target feature vector corresponding to the broadband oscillation signal; The identification module is used to input the target feature vector into the target network for classification and identification to obtain an identification result.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the wide-frequency oscillation signal identification method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the broadband oscillation signal identification method according to any one of claims 1 to 6 when executed.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for identifying a broadband oscillation signal according to any one of claims 1 to 6 is implemented.

Images