A method and system for identifying low-frequency oscillation of a power system

By constructing multiple oscillation identification models and utilizing singular value decomposition and subspace clustering methods, combined with the transformer model to optimize sparse attention, the problem of rapid identification of low-frequency oscillations in power systems was solved, improving the accuracy and stability of the analysis.

CN114722962BActive Publication Date: 2024-11-26HEILONGJIANG ELECTRIC POWER SCIENCE RESEARCH INSTITUTE +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210444517.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-04-26
Publication Date
2024-11-26
Estimated Expiration
2042-04-26

AI Technical Summary

Technical Problem

Existing technologies lack methods for rapidly analyzing the contribution of generating units to stability in power systems at ultra-low frequencies. Inaccurate load modeling leads to unreliable results in free oscillation analysis, and it is difficult to locate the source of forced oscillation disturbances.

Method used

Multiple oscillation identification models based on different damping frequencies are constructed. Feature vectors are extracted using singular value decomposition and subspace clustering methods. Load data are trained using a transformer model, and the sparse attention mechanism is optimized to improve identification accuracy.

Benefits of technology

This approach fully explores multi-perspective low-frequency oscillation models, improves the dynamics and accuracy of low-frequency oscillation identification, and enhances the robustness of the model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114722962B_ABST
    Figure CN114722962B_ABST
Patent Text Reader

Abstract

The present invention relates to a method and system for identifying low-frequency oscillations of an electric power system, the method comprising: constructing a plurality of oscillation identification models based on different damping frequencies, determining the characteristic matrix of each oscillation identification model; extracting the characteristic vector of each oscillation identification model by using a singular value decomposition method, mapping the plurality of characteristic vectors into multidimensional vectors by a subspace clustering method, and training a transformer model based thereon; inputting the load data of the target electric power system into the transformer model to obtain identification parameters. The present invention comprehensively extracts the characteristic values ​​of a plurality of low-frequency oscillation models of an electric power system by a subspace clustering method, and then improves the dynamics and accuracy of the identification model by a transformer model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of electric power measurement, and in particular relates to a low-frequency oscillation method and system. Background Art

[0002] Low-frequency oscillation seriously threatens the safe and stable operation of power systems, and has therefore always been one of the important topics in the field of power system research. In a broad sense, low-frequency oscillation can be divided into low-frequency oscillations between 0.1 and 2.5 Hz and ultra-low-frequency oscillations less than 0.1 Hz according to the oscillation frequency. Ultra-low-frequency oscillations are mainly caused by the water hammer effect of turbines and improper governor parameters. However, there are many generators in the system, and there is currently a lack of methods for quickly analyzing the contribution of units to stability in the ultra-low frequency band based on measured data.

[0003] Low-frequency oscillations with a frequency of 0.1 to 2.5 Hz are mainly divided into two categories according to the mechanism: free oscillation and forced oscillation. A key problem of free oscillation is that the accuracy of the load model seriously affects the analysis results of free oscillation (small disturbance analysis results). However, the existing load modeling method cannot solve the problem of inaccurate load modeling due to the time-varying nature of the load, which will lead to unreliable free oscillation analysis results. The cause of forced oscillation is the existence of periodic disturbance sources in the system. However, the scale of the power system is large, so a key issue of forced oscillation is how to locate the location of the disturbance source. Summary of the invention

[0004] In order to solve the problem of improving the anti-noise performance and accuracy of the load model in the oscillation identification process, a first aspect of the present invention provides a method for identifying low-frequency oscillations of an electric power system, including: constructing a plurality of oscillation identification models based on different damping frequencies, and determining the characteristic matrix of each oscillation identification model; using a singular value decomposition method to extract the characteristic vector of each oscillation identification model, mapping the plurality of characteristic vectors into a multidimensional vector through a subspace clustering method, and training a transformer model based on the vector; inputting the load data of the target power system into the transformer model to obtain identification parameters.

[0005] In some embodiments of the present invention, the singular value decomposition method is used to extract the characteristic vector of each oscillation identification model, and multiple characteristic vectors are mapped into multidimensional vectors through a subspace clustering method, including: using an extended complex Kalman filter to filter the dynamic noise of the oscillation identification model; using the Henkel singular value decomposition method to extract the characteristic vector of each oscillation identification model.

[0006] In some embodiments of the present invention, mapping the plurality of eigenvectors into a multidimensional vector by a subspace clustering method comprises: cascading the plurality of eigenvectors into a matrix, and then decomposing the matrix to obtain a multidimensional vector.

[0007] In some embodiments of the present invention, the transformer model is trained by the following method: the load data of the power system is cleaned and normalized; the normalized load data is used as a sample and its corresponding multidimensional vector is used as a label to construct a training data set; and the transformer model is trained using the training data set until its error tends to be stable and is lower than a threshold.

[0008] Furthermore, the transformer model is optimized through a sparse attention mechanism.

[0009] In the above-mentioned embodiment, the oscillation identification model includes: a first-order mechanical transient induction motor parallel constant impedance model and a third-order electromechanical transient induction motor parallel constant impedance model.

[0010] The second aspect of the present invention provides a low-frequency oscillation identification system for an electric power system, comprising: a determination module for constructing a plurality of oscillation identification models based on different damping frequencies, and determining a characteristic matrix of each oscillation identification model; a mapping module for extracting a characteristic vector of each oscillation identification model using a singular value decomposition method, mapping the plurality of characteristic vectors into a multidimensional vector using a subspace clustering method, and training a transformer model based thereon; and an identification module for inputting the load data of the target electric power system into the transformer model to obtain identification parameters.

[0011] Furthermore, the mapping module includes: a filtering unit for filtering the dynamic noise of the oscillation identification model using an extended complex Kalman filter; and an extraction unit for extracting the characteristic vector of each oscillation identification model using a Henkel singular value decomposition method.

[0012] According to a third aspect of the present invention, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method for identifying low-frequency oscillations of the power system provided in the first aspect of the present invention.

[0013] A fourth aspect of the present invention provides a computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for identifying low-frequency oscillations of an electric power system provided in the first aspect of the present invention.

[0014] The beneficial effects of the present invention are:

[0015] 1. The present invention comprehensively extracts the eigenvalues ​​of various low-frequency oscillation models of the power system through the subspace clustering method, obtains the low-dimensional subspace representation of each perspective through singular value decomposition, and then cascades these representations to fully mine the composite model data of multiple perspectives and multiple sources;

[0016] 2. The transformer model improves the dynamics and accuracy of the recognition model through sparse optimization;

[0017] 3. It is superior to low-frequency oscillation in that the sample size is small and the robustness of the model is further improved through sparse representation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A basic flow chart of a method for identifying low-frequency oscillation of an electric power system in some embodiments of the present invention;

[0019] Figure 2 is a schematic diagram of the structure of a transformer model in some embodiments of the present invention;

[0020] Figure 3 Schematic diagram of the encoder structure of the transformer model in some embodiments of the present invention;

[0021] Figure 4 A schematic diagram of a decoder structure of a transformer model in some embodiments of the present invention;

[0022] Figure 5 It is a schematic diagram of the principle of the transformer model after sparseness in some embodiments of the present invention;

[0023] Figure 6 is a parallel constant impedance equivalent circuit diagram of an induction motor in some embodiments of the present invention;

[0024] Figure 7 It is a structural schematic diagram of a method system for identifying low-frequency oscillation of an electric power system in some embodiments of the present invention;

[0025] Figure 8 It is a schematic diagram of the structure of an electronic device in some embodiments of the present invention. DETAILED DESCRIPTION

[0026] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0027] refer to Figure 1In the first aspect of the present invention, a method for identifying low-frequency oscillations of an electric power system is provided, comprising: S100. constructing a plurality of oscillation identification models based on different damping frequencies, and determining a characteristic matrix of each oscillation identification model; S200. extracting a characteristic vector of each oscillation identification model by using a singular value decomposition method, mapping the plurality of characteristic vectors into a multidimensional vector by using a subspace clustering method, and training a transformer model based thereon; S300. inputting the load data of the target electric power system into the transformer model to obtain identification parameters.

[0028] In step S100 of some embodiments of the present invention, specifically, the free oscillation oscillation identification model includes: using a load model of small disturbance analysis, using a load model of generator mechanical power disturbance source positioning based on power spectrum density in forced oscillation, and using a load model of generator mechanical power disturbance source positioning using a decoupling observer. For example, the discrete state space model of the load established using small disturbance analysis:

[0029] X(k+1)=A d x(k)+B d u(k),

[0030] Y(k)=C d x(k)+D d u(k);

[0031] In the above formula, the state matrix A d The dimension is n×n, and the input matrix B d The dimension is n×m, and the output matrix C d The dimension is l×n, and the feedforward matrix D d The dimension is l×m. The input of the load model u(k)=[△V(k)△θ(k)] T ∈2x1, the output of the load model. Among them, △V(k) represents the deviation of the load voltage amplitude, △θ(k) represents the deviation of the load voltage phase angle, △P(k) represents the deviation of the load active power, and △θ(k) represents the deviation of the load reactive power. The above variables can be obtained by removing the mean of the actually measured load signals V(k), θ(k), P(k), θ(k).

[0032] The subspace method is an identification method that directly determines the state space model of a linear system. The state space matrix determined by the subspace method is denoted as A d ,B d ,C d ,D d The actual load model discrete state space matrix is ​​denoted as A c ,B c ,C c ,Dc The matrix A determined by the subspace method d ,B d ,C d ,D d It is not equal to the state space matrix A of the actual load model c , B c , C c , D c , there is an unknown linear transformation T between the two. It can be understood that A d ,B d ,C d ,D d It can show the load time, voltage, active power and idle time parameters waiting to be identified or already identified.

[0033] In step S200 of some embodiments of the present invention, the singular value decomposition method is used to extract the characteristic vector of each oscillation identification model, and the plurality of the characteristic vectors are mapped into a multidimensional vector by the subspace clustering method, including: S201. Using an extended complex Kalman filter (ECKF) to filter the dynamic noise of the oscillation identification model; S202. Using the Henkel singular value decomposition method to extract the characteristic vector of each oscillation identification model. Specifically, taking the above subspace method as an example, assuming A c and C c Known, define the extended observable matrix M c ∈k1l×n:

[0034]

[0035] In the above formula, k1 is the extended observable matrix M c The coefficients of the matrix M c The first l rows of the instant matrix C c , the eigenvector of each oscillation identification model is extracted using the Hankel singular value decomposition method; it can be understood that the above only uses the oscillation identification model of the subspace to illustrate the extraction process of the Hankel singular value decomposition method (Hankel Singular Value Decomposition, HSVD), which does not affect its application in other oscillation identification models.

[0036] In some embodiments of the present invention, mapping the plurality of eigenvectors into a multidimensional vector by a subspace clustering method comprises: cascading the plurality of eigenvectors into a matrix, and then decomposing the matrix to obtain a multidimensional vector.

[0037] In step S200 of some embodiments of the present invention, the transformer model is trained by the following method: S203. The load data of the power system is cleaned and normalized; S204. The normalized load data is used as a sample and its corresponding multidimensional vector is used as a label to construct a training data set; S205. The transformer model is trained using the training data set until its error tends to be stable and is lower than a threshold.

[0038] refer to Figure 2 To Figure 5 Specifically, the transformer model includes at least one encoder and decoder, and the decoder and encoder are connected through an attention mechanism (transferring information or residuals); each encoder includes at least a position embedding layer, one or more self-attention layers, a regularization layer, and one or more self-attention layer fully connected layers; the position embedding layer can pass the training information to the last regularization layer and then output it. Each decoder includes at least a position embedding layer, a mask self-attention layer, a regularization layer, an encoding and decoding attention layer, and a fully connected layer.

[0039] Since the signal or sample of low-frequency oscillation is small, in order to improve accuracy and effectively transmit sparse information, the sparse attention mechanism is used to optimize the transformer model; specifically, the attention mechanism of the self-attention layer or the masked self-attention layer is expressed by:

[0040] Attend(X, S) = (a(x i , S i )) i∈{1,...,n}

[0041]

[0042]

[0043] Where S = {j:j<=i}, W q ,、W k ,、W v is a matrix for q, k, v; and each attendance has p heads of attention; for each head, define A i (m) It is a subset of S={j:j<=i}.

[0044] Furthermore, the first is to define two such attention heads:

[0045] A i (1) ={t,t+1,…,i}for t=max(0,il),A i(2) = {j: (ij) modl = 0}; where l = sqrt(n).

[0046] The second method is similar, defining two attention heads:

[0047] A i (1) ={j:floor(j / l)=floor(i / l)},A i (2) ={j:j:modl∈{t,t+1,…,l}}; where t=lc, c is a hyperparameter. This is called fixed attention. In general, l takes the value of {128,256} and c takes the value of {8,16,32}. Floor represents the floor function.

[0048] refer to Figure 5 , multidimensional vector (x1,x2,…,x n ) is superimposed with the emded multi-dimensional vector through regularization, attention mechanism and dropout (residual module), and after multiple such iterations, it is output through softmax; the kth layer is obtained by adding the result of the k-1 layer to the current layer;

[0049] H0=embed(X,W e )

[0050] H k =H k-1 +resblock(H k-1 )

[0051] y=softmax(norm(H N )W out ),

[0052] Where H0 represents the multidimensional vector after emded, X is the matrix representation of the multidimensional vector, and W e Its weight; H k represents the output after the k-th layer residual module, y represents the output, norm represents regularization, and W out It should be noted that the linear regularization layer (linear) and feed-forward layer (feed-forward) in the figure belong to the general layer and will not be described here.

[0053] refer to Figure 6 In the above-mentioned embodiment, the first-order mechanical transient induction motor parallel constant impedance model and the third-order electromechanical transient induction motor parallel constant impedance model using the comprehensive identification method are used. Specifically, the comprehensive compliance model of the third-order electromechanical transient induction motor parallel constant impedance is used for parameter identification;

[0054] Exemplarily, the state equation of the third-order electromechanical transient induction motor shunt constant impedance model is as follows:

[0055]

[0056]

[0057]

[0058] Among them, e′=ex′+je y ′ represents the transient potential of the induction motor (the subscripts x and y are its real and imaginary parts respectively); u = us + ju y Represents the voltage excitation of the load; ω B represents the synchronous angular velocity, ω represents the operating angular frequency of the power system; s represents the slip rate of the induction motor; x′ represents the stator and rotor transient reactance, x s Expressed as the stator and rotor synchronous reactance, r s Represents stator resistance; T j represents the rotor inertia time constant; T0′ represents the decay time constant of transient potential; P m It represents the mechanical load power of the induction motor, which is determined by the mechanical load characteristics of the induction motor. The first-order mechanical transient induction motor parallel constant impedance model can also be obtained based on the above-mentioned related parameters and circuit deformation.

[0059] Example 2

[0060] refer to Figure 7 According to a second aspect of the present invention, a low-frequency oscillation identification system 1 for an electric power system is provided, comprising: a determination module 11 for constructing a plurality of oscillation identification models based on different damping frequencies, and determining a characteristic matrix of each oscillation identification model; a mapping module 12 for extracting a characteristic vector of each oscillation identification model by using a singular value decomposition method, mapping the plurality of characteristic vectors into a multidimensional vector by using a subspace clustering method, and training a transformer model based thereon; and an identification module 13 for inputting the load data of the target electric power system into the transformer model to obtain identification parameters.

[0061] Furthermore, the mapping module 12 includes: a filtering unit for filtering the dynamic noise of the oscillation identification model using an extended complex Kalman filter; and an extraction unit for extracting the feature vector of each oscillation identification model using a Henkel singular value decomposition method.

[0062] Example 3

[0063] refer to Figure 8According to a third aspect of the present invention, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method of the first aspect of the present invention.

[0064] The electronic device 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0065] Typically, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 8 The electronic device 500 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead. Figure 8 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0066] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, the above functions defined in the method of the embodiment of the present disclosure are executed. It should be noted that the computer-readable medium described in the embodiment of the present disclosure can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more conductors, 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 above. In an embodiment of the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, an apparatus, or a device. In an embodiment of the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in combination with an instruction execution system, an apparatus, or a device. The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to: wire, optical cable, RF (radio frequency), etc., or any suitable combination of the foregoing.

[0067] The computer-readable medium may be included in the electronic device, or may exist independently without being installed in the electronic device. The computer-readable medium carries one or more computer programs. When the one or more programs are executed by the electronic device, the electronic device:

[0068] Computer program code for performing the operations of embodiments of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, Python, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on a user's computer, partially on a user's computer, as a separate software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0069] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0070] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. 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 low-frequency oscillations in an electric power system, characterized in that: include: Construct multiple oscillation identification models based on different damping frequencies and determine the characteristic matrix of each oscillation identification model; The singular value decomposition method is used to extract the characteristic vectors of each oscillation identification model, and the plurality of the characteristic vectors are mapped into a multidimensional vector by a subspace clustering method, and the transformer model is trained according to the plurality of the characteristic vectors. The mapping of the plurality of the characteristic vectors into a multidimensional vector by the subspace clustering method comprises: cascading the plurality of the characteristic vectors into a matrix, and then decomposing the matrix to obtain a multidimensional vector; The load data of the target power system is input into the transformer model to obtain identification parameters; the transformer model is trained by the following method: the load data of the power system is cleaned and normalized; the normalized load data is used as a sample and its corresponding multidimensional vector is used as a label to construct a training data set; the transformer model is trained using the training data set until its error tends to be stable and is lower than a threshold.

2. The method for identifying low-frequency oscillation of an electric power system according to claim 1, characterized in that: The method of extracting the characteristic vector of each oscillation identification model by using the singular value decomposition method and mapping the plurality of characteristic vectors into a multidimensional vector by using the subspace clustering method comprises: Using extended complex Kalman filter to filter dynamic noise of oscillation identification model; The Henkel singular value decomposition method is used to extract the eigenvectors of each oscillation identification model.

3. The method for identifying low-frequency oscillation of an electric power system according to claim 1, characterized in that: The transformer model is optimized via a sparse attention mechanism.

4. The method for identifying low-frequency oscillation of an electric power system according to any one of claims 1 to 3, characterized in that: The oscillation identification model includes: a first-order mechanical transient induction motor parallel constant impedance model and a third-order electromechanical transient induction motor parallel constant impedance model.

5. A low-frequency oscillation identification system for an electric power system, characterized in that: include: A determination module is used to construct a plurality of oscillation identification models based on different damping frequencies and determine a characteristic matrix of each oscillation identification model; A mapping module, used for extracting the characteristic vector of each oscillation identification model by using a singular value decomposition method, mapping a plurality of the characteristic vectors into a multidimensional vector by using a subspace clustering method, and training a transformer model according to the multidimensional vector; the mapping of the plurality of the characteristic vectors into a multidimensional vector by using a subspace clustering method comprises: cascading a plurality of the characteristic vectors into a matrix, and then decomposing the matrix to obtain a multidimensional vector; The identification module is used to input the load data of the target power system into the transformer model to obtain the identification parameters; the transformer model is trained by the following method: the load data of the power system is cleaned and normalized; the normalized load data is used as a sample and its corresponding multidimensional vector is used as a label to construct a training data set; the transformer model is trained using the training data set until its error tends to be stable and is lower than a threshold.

6. The low frequency oscillation identification system of the power system according to claim 5, characterized in that: The mapping module comprises: A filtering unit, for filtering dynamic noise of the oscillation identification model using an extended complex Kalman filter; The extraction unit is used to extract the feature vector of each oscillation identification model by using the Henkel singular value decomposition method.

7. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method for identifying low-frequency oscillations of the power system as described in any one of claims 1 to 4.

8. A computer-readable medium, characterized in that A computer program is stored thereon, wherein when the computer program is executed by a processor, the method for identifying low-frequency oscillation of an electric power system as claimed in any one of claims 1 to 4 is implemented.