Stable state detection method of power system under large disturbance

By constructing a joint neural network model, using neural differential equations and energy estimation error function, combined with time domain large disturbance data, the problem of stable state detection accuracy of power system under large disturbances is solved, and the rapid and accurate detection of complex power systems is achieved. .

CN119089310APending Publication Date: 2024-12-06ELECTRIC POWER RES INST CHINA SOUTHERN POWER GRID CO LTD +1
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411292847.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-14
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing state detection methods of power systems after large disturbances cannot accurately and quickly determine the stable state changes of complex and variable power systems.

Method used

By establishing the neural differential equation estimation error function and energy estimation error function, a joint loss function is constructed, a joint neural network model is established, and the model is trained using the time domain large disturbance data to obtain the stable state of the power system to be tested under large disturbances.

Benefits of technology

It improves the characterization ability of the complex nonlinear dynamics of the power system, enhances the accuracy of stable state detection, and can quickly and accurately detect the stable state of the power system under large disturbances, enhancing the interpretability of the detection and analysis results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089310A_ABST
    Figure CN119089310A_ABST
Patent Text Reader

Abstract

The invention discloses a stable state detection method of a power system under large disturbance. The method comprises the following steps: establishing a neural differential equation estimation error function and an energy estimation error function; establishing a joint loss function in combination with the neural differential equation estimation error function and the energy estimation error function; establishing a joint neural network model, and obtaining time domain large disturbance data; taking the time domain large disturbance data as a training sample, inputting the training sample into the joint neural network model, and training the joint neural network model by taking joint loss function minimization as a target to obtain a trained joint neural network model; and obtaining a state quantity of the to-be-tested power system under the large disturbance, and processing the state quantity by using the trained joint neural network model to obtain a stable state of the to-be-tested power system under the large disturbance. According to the scheme, a joint neural network model capable of outputting an accurate detection result is obtained, the model is used for processing the state quantity, and the stable state of the to-be-detected power system under large disturbance can be rapidly and accurately obtained.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of stable state detection, and in particular to a method for detecting the stable state of an electric power system under a large disturbance. Background Art

[0002] In today's society, the stable operation of the power system is of vital importance. With the rapid development of the economy and the continuous advancement of science and technology, the demand for electricity continues to grow, and the scale and complexity of the power system are also increasing. In order to ensure the stable operation of the power system, it is necessary to conduct in-depth detection and evaluation of the stability of the power system.

[0003] The existing methods for detecting the state of power systems after large disturbances have certain limitations. They can usually only detect the state of simple and single power systems. For power systems with complex and changeable states, they cannot accurately and quickly determine the changes in their stable states. Summary of the invention

[0004] In view of this, the present application provides a method for detecting the stable state of an electric power system under a large disturbance, which is used to solve the problem that the existing method for detecting the state of an electric power system after a large disturbance has certain limitations and can usually only detect the state of a simple and single electric power system. For an electric power system with a complex and changeable state, it is impossible to accurately and quickly determine the change in its stable state.

[0005] To achieve the above objectives, the proposed solution is as follows:

[0006] In a first aspect, a method for detecting a stable state of a power system under a large disturbance comprises:

[0007] Establish the neural differential equation estimation error function and energy estimation error function;

[0008] Combining the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function;

[0009] Establish a joint neural network model and obtain large disturbance data in the time domain;

[0010] Input the time-domain large disturbance data as training samples into the joint neural network model, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model;

[0011] The state quantity of the power system to be tested under a large disturbance is obtained, and the state quantity is processed by using the trained joint neural network model to obtain the stable state of the power system to be tested under the large disturbance.

[0012] Preferably, the joint neural network model includes: an input module, a feature extraction module, a neural differential equation module, an energy function module, a stability analysis module and an output module;

[0013] The output end of the input module is connected to the input end of the feature extraction module, the output end of the feature extraction module is respectively connected to the input end of the neural differential equation module and the input end of the energy function module, the output end of the neural differential equation module and the output end of the energy function module are respectively connected to the input end of the stability analysis module, and the output end of the stability analysis module is connected to the input end of the output module.

[0014] Preferably, the neural differential equation estimation error function is:

[0015] ;

[0016] in, represents the neural differential equation estimation error function, Indicates samples, is the sample size, Indicates The prediction error value of samples is For the The true error value of the sample.

[0017] Preferably, the energy estimation error function is:

[0018] ;

[0019] in, represents the energy estimation error function, Indicates samples, is the sample size, is the proportionality constant, Indicates The error value of the sample, Indicates that the power system to be tested has an error value of The state quantity corresponding to Indicates that the power system to be tested has an error value of The energy function value corresponding to .

[0020] Preferably, the combining the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function includes:

[0021] The neural differential equation estimation error function is expressed as , the energy estimation error function is expressed as , combined with and , establish the obtained joint loss function for:

[0022] ;

[0023] in, is the weight coefficient.

[0024] Preferably, the using the trained joint neural network model to process the state quantity to obtain the stable state of the power system to be tested under a large disturbance includes:

[0025] Extracting state features from the state quantity using a feature extraction module in the joint neural network model;

[0026] Using the neural differential equation module in the joint neural network model to perform neural constant differential error estimation on the state feature to obtain first state data;

[0027] Using the energy function module in the joint neural network model to perform energy estimation on the state feature to obtain second state data;

[0028] The stability analysis module in the joint neural network model is used to perform joint state estimation on the first state data and the second state data to obtain the stable state of the power system under large disturbance.

[0029] Preferably, the method of using the stability analysis module in the joint neural network model to perform joint state estimation on the first state data and the second state data to obtain the stable state of the power system under large disturbance includes:

[0030] Using the stability analysis module in the joint neural network model, jointly estimating the first state data and the second state data to obtain a joint estimation value;

[0031] Determine the maximum value of the energy function of the power system to be tested after being subjected to a large disturbance according to the joint estimated value;

[0032] Determine, according to the joint estimated value, an energy function value of the power system to be tested at any time point after being subjected to a large disturbance, as a first energy function value;

[0033] Determine, according to the joint estimated value, an energy function value of the power system to be tested at the initial equilibrium moment after being subjected to a large disturbance as a second energy function value;

[0034] The stable state of the power system to be tested under a large disturbance is obtained based on the maximum value of the energy function, the first energy function value and the second energy function value.

[0035] In a second aspect, a device for detecting a stable state of a power system under a large disturbance comprises:

[0036] Error function building module, used to build neural differential equation estimation error function and energy estimation error function;

[0037] A joint loss function establishment module, used to combine the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function;

[0038] Model building and data acquisition module, used to build a joint neural network model and obtain large disturbance data in the time domain;

[0039] A model training module, used to input the time-domain large disturbance data as training samples into the joint neural network model, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model;

[0040] The stable state determination module is used to obtain the state quantity of the power system to be tested under large disturbance, and use the trained joint neural network model to process the state quantity to obtain the stable state of the power system to be tested under large disturbance.

[0041] In a third aspect, a device for detecting a stable state of a power system under a large disturbance includes a memory and a processor;

[0042] The memory is used to store programs;

[0043] The processor is used to execute the program to implement the various steps of the method for detecting the stable state of the power system under large disturbance as described in any one of the first aspects.

[0044] In a fourth aspect, a storage medium stores a computer program thereon, wherein when the computer program is executed by a processor, the steps of the method for detecting a stable state of an electric power system under a large disturbance as described in any one of the first aspects are implemented.

[0045] It can be seen from the above technical scheme that the present application establishes a neural differential equation estimation error function and an energy estimation error function; establishes a joint loss function by combining the neural differential equation estimation error function and the energy estimation error function; establishes a joint neural network model, and obtains large disturbance data in the time domain; uses the large disturbance data in the time domain as training samples, inputs them into the joint neural network model, and trains the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model; obtains the state quantity of the power system to be tested under large disturbance, and uses the trained joint neural network model to process the state quantity to obtain the stable state of the power system to be tested under large disturbance. This scheme builds and solves the model by establishing a neural differential equation estimation error function and an energy estimation error function, and combines them to construct a joint loss function. This can improve the model's ability to characterize the complex nonlinear dynamics of the power system to be tested. The simultaneous optimization of the joint loss function can improve the accuracy of stable state detection. The model is trained with large disturbance data in the time domain, which can fully capture the dynamic behavior and energy characteristics of different power systems under different types of large disturbances, thereby obtaining a trained joint neural network model that can output accurate detection results. By using this model to process the state quantity of the power system to be tested under large disturbances, the stable state of the power system to be tested under large disturbances can be quickly and accurately obtained, thereby enhancing the interpretability of the detection and analysis results. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0047] Figure 1 An optional flow chart of a method for detecting a stable state of a power system under a large disturbance provided in an embodiment of the present application;

[0048] Figure 2 A schematic diagram of the structure of a joint neural network model provided in an embodiment of the present application;

[0049] Figure 3 A schematic diagram of the structure of a device for detecting the stable state of a power system under a large disturbance provided in an embodiment of the present application;

[0050] Figure 4 A schematic diagram of the structure of a stable state detection device for an electric power system under a large disturbance provided in an embodiment of the present application. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0052] The present invention can be used in many general or special computing device environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multi-processor devices, distributed computing environments including any of the above devices or devices, etc.

[0053] The embodiment of the present invention provides a method for detecting the stable state of a power system under a large disturbance. The method can be applied to various computer terminals or smart terminals. The execution subject can be a processor or a server of the computer terminal or the smart terminal. The method flow chart of the method is as follows: Figure 1 As shown, specifically including:

[0054] S1: Establish the neural differential equation estimation error function and the energy estimation error function.

[0055] Considering the interaction and dynamic behavior of various electrical components in the power system to be tested, the state of the power system to be tested after a large disturbance can be described by a nonlinear differential equation, which can also be regarded as a derivative function:

[0056] ;

[0057] in, The state quantity of the power system to be measured. The selection of specific state quantities depends on the specific conditions of the network equipment, structure, modeling method, etc. Generally, it may include the phase angle, speed, transient potential, etc. of the generator; is the input vector, which represents the control input of the system or the external large disturbance, is the network parameter.

[0058] When using neural differential equations for modeling and prediction, the error function plays a key role. The error function is used to measure the difference between the predicted value of the neural differential equation and the actual observed value. By minimizing the error function, the parameters of the neural network can be adjusted to make the prediction results of the neural differential equation more accurate. The neural differential equation estimation error function can also be regarded as the system dynamic prediction error function.

[0059] The energy estimation error function refers to the energy function estimation error. The energy function can be used to evaluate the stability of the power system under test under different operating conditions and predict the dynamic response of the power system under test after being disturbed.

[0060] S2: Combine the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function.

[0061] In order to more effectively simulate the dynamic behavior of the power system under test, this application introduces the neural differential equation estimation error function, and the energy function estimation error function can be used to evaluate the energy characteristics of the power system under test in different states, which helps to accurately detect the stable state of the system. Therefore, the two are combined to establish a joint loss function for building a model, which can achieve the simultaneous optimization of the dynamic and energy functions of the power system under test.

[0062] S3: Establish a joint neural network model and obtain large disturbance data in the time domain.

[0063] After the joint neural network model is established, it is necessary to obtain large disturbance data as training samples to train the model. The method of obtaining large disturbance data in the time domain may include: using monitoring equipment to collect first time domain data samples of multiple selected power systems under large disturbances, extracting second time domain data samples containing large disturbance events from the database, aggregating the first time domain data samples and the second time domain data samples to obtain third time domain data samples, and then preprocessing the third time domain data samples to obtain fourth time domain data samples, and finally performing sequence expansion processing on the fourth data samples to obtain large disturbance data in the time domain.

[0064] S4: Input the large time-domain disturbance data into the joint neural network model as training samples, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model.

[0065] The first time domain data samples collected by the monitoring equipment cover data of many different types of power systems, and the second time domain data samples extracted from the database contain various time domain data. Sequence expansion processing of the data samples can increase the length or dimension of the data samples to obtain more information or meet specific detection and analysis needs. Then, by obtaining large-scale time domain disturbance data with rich data volume in the above manner, and training the joint neural network model with the goal of minimizing the joint loss function, the trained joint neural network model can quickly and efficiently process input data from different power systems.

[0066] S5: Acquire the state quantity of the power system to be tested under the large disturbance, and use the trained joint neural network model to process the state quantity to obtain the stable state of the power system to be tested under the large disturbance.

[0067] By using the trained joint neural network to process the input data of the power system to be tested, the stable state of the power system to be tested under large disturbances can be obtained, such as maintaining stability, tending to instability, instability, etc.

[0068] It can be seen from the above technical scheme that the present application establishes a neural differential equation estimation error function and an energy estimation error function; establishes a joint loss function by combining the neural differential equation estimation error function and the energy estimation error function; establishes a joint neural network model, and obtains large disturbance data in the time domain; uses the large disturbance data in the time domain as training samples, inputs them into the joint neural network model, and trains the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model; obtains the state quantity of the power system to be tested under large disturbance, and uses the trained joint neural network model to process the state quantity to obtain the stable state of the power system to be tested under large disturbance. This scheme builds and solves the model by establishing a neural differential equation estimation error function and an energy estimation error function, and combines them to construct a joint loss function. This can improve the model's ability to characterize the complex nonlinear dynamics of the power system. The simultaneous optimization of the joint loss function can improve the accuracy of stable state detection. The model is trained with large disturbance data in the time domain, which can fully capture the dynamic behavior and energy characteristics of different power systems under different types of large disturbances, thereby obtaining a trained joint neural network model that can output accurate detection results. By using this model to process the state quantity of the power system to be tested under large disturbances, the stable state of the power system to be tested under large disturbances can be quickly and accurately obtained, thereby enhancing the interpretability of the detection and analysis results.

[0069] In one embodiment provided in the present application, the structure diagram of the joint neural network model is as follows: Figure 2 As shown, including:

[0070] Input module, feature extraction module, neural differential equation module, energy function module, stability analysis module and output module.

[0071] The output end of the input module is connected to the input end of the feature extraction module, the output end of the feature extraction module is respectively connected to the input end of the neural differential equation module and the input end of the energy function module, the output end of the neural differential equation module and the output end of the energy function module are respectively connected to the input end of the stability analysis module, and the output end of the stability analysis module is connected to the input end of the output module.

[0072] Specifically, the input module is used to receive the state quantity data input by the power system to be tested, and then use the feature extraction module to extract the feature data in the state quantity data, and transmit them to the neural differential equation module and the energy function module for processing respectively to obtain different state data. The neural differential equation can extract the nonlinear dynamic characteristics under large disturbances in the feature data, and the energy function module can extract the energy characteristics under different states, and input them together into the stability analysis module for joint state estimation to obtain the stable state of the power system to be tested.

[0073] The structural equation of the joint neural network model can be expressed as:

[0074] ;

[0075] in, are the network parameters of the neural differential equation module, is the network parameter of the energy function module, are the network parameters of the feature extraction module.

[0076] In the method provided by the embodiment of the present invention, the neural differential equation estimation error function can be expressed as:

[0077] ;

[0078] in, represents the neural differential equation estimation error function, Indicates samples, is the sample size, Indicates The prediction error value of samples is For the The true error value of the sample.

[0079] The energy estimation error function can be expressed as:

[0080] ;

[0081] in, represents the energy estimation error function, Indicates samples, is the sample size, is a proportional constant used to adjust the weight ratio of the system dynamics and energy function estimation errors. Indicates The error value of the sample, Indicates that the power system to be tested has an error value of The state quantity corresponding to Indicates that the power system to be tested has an error value of The energy function value corresponding to .

[0082] Then, combining the neural differential equation estimation error function and the energy estimation error function, the neural differential equation estimation error function is expressed as , the energy estimation error function is expressed as , combined with and , establish the obtained joint loss function for:

[0083] ;

[0084] in, is the weight coefficient.

[0085] In this way, the neural differential equation estimation error function and the energy estimation error function can be combined to train a joint neural network model.

[0086] The following embodiment describes in detail the process of using the trained joint neural network model in the present application to process the state quantity and obtain the stable state of the power system to be tested under a large disturbance.

[0087] Extracting state features from the state quantity using a feature extraction module in the joint neural network model;

[0088] Using the neural differential equation module in the joint neural network model to perform neural constant differential error estimation on the state feature to obtain first state data;

[0089] Using the energy function module in the joint neural network model to perform energy estimation on the state feature to obtain second state data;

[0090] The stability analysis module in the joint neural network model is used to perform joint state estimation on the first state data and the second state data to obtain the stable state of the power system under large disturbance.

[0091] The process of using the stability analysis module in the joint neural network model to perform joint state estimation on the first state data and the second state data to obtain the stable state of the power system under large disturbance includes:

[0092] Using the stability analysis module in the joint neural network model, jointly estimating the first state data and the second state data to obtain a joint estimation value;

[0093] Determine the maximum value of the energy function of the power system to be tested after being subjected to a large disturbance according to the joint estimated value;

[0094] Determine, according to the joint estimated value, an energy function value of the power system to be tested at any time point after being subjected to a large disturbance, as a first energy function value;

[0095] Determine, according to the joint estimated value, an energy function value of the power system to be tested at the initial equilibrium moment after being subjected to a large disturbance as a second energy function value;

[0096] The stable state of the power system to be tested under a large disturbance is obtained based on the maximum value of the energy function, the first energy function value and the second energy function value.

[0097] Specifically, the maximum value of the energy function of the power system under test after a large disturbance can refer to the fault trajectory of the power system under test. The energy value in the fault trajectory will increase first and then decrease. The highest point (maximum point) of the fault trajectory is the critical energy of the power system under test, which can also be regarded as the energy function value of the critical point of stability. Among them, the stable state can be determined according to the energy function values ​​at the three points, such as calculating the stability margin.

[0098] and Figure 1 Corresponding to the method described above, the embodiment of the present invention also provides a stable state detection device for a power system under a large disturbance, which is used to detect the stable state of the power system under a large disturbance. Figure 1 In the specific implementation of the method, the stable state detection device of the power system under large disturbance provided by the embodiment of the present invention can be used in a computer terminal or various mobile devices, combined with Figure 3 , the stable state detection device of the power system under large disturbance is introduced, such as Figure 3 As shown, the device may include:

[0099] An error function establishment module 10, used to establish a neural differential equation estimation error function and an energy estimation error function;

[0100] A joint loss function establishment module 20, used to establish a joint loss function by combining the neural differential equation estimation error function and the energy estimation error function;

[0101] The model building and data acquisition module 30 is used to build a joint neural network model and acquire large disturbance data in the time domain;

[0102] A model training module 40 is used to input the time domain large disturbance data as training samples into the joint neural network model, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model;

[0103] The stable state determination module 50 is used to obtain the state quantity of the power system under test under large disturbance, and use the trained joint neural network model to process the state quantity to obtain the stable state of the power system under test under large disturbance.

[0104] It can be seen from the above technical scheme that the present application establishes a neural differential equation estimation error function and an energy estimation error function; establishes a joint loss function by combining the neural differential equation estimation error function and the energy estimation error function; establishes a joint neural network model, and obtains large disturbance data in the time domain; uses the large disturbance data in the time domain as training samples, inputs them into the joint neural network model, and trains the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model; obtains the state quantity of the power system to be tested under large disturbance, and uses the trained joint neural network model to process the state quantity to obtain the stable state of the power system to be tested under large disturbance. This scheme builds and solves the model by establishing a neural differential equation estimation error function and an energy estimation error function, and combines them to construct a joint loss function. This can improve the model's ability to characterize the complex nonlinear dynamics of the power system. The simultaneous optimization of the joint loss function can improve the accuracy of stable state detection. The model is trained with large disturbance data in the time domain, which can fully capture the dynamic behavior and energy characteristics of different power systems under different types of large disturbances, thereby obtaining a trained joint neural network model that can output accurate detection results. By using this model to process the state quantity of the power system to be tested under large disturbances, the stable state of the power system to be tested under large disturbances can be quickly and accurately obtained, thereby enhancing the interpretability of the detection and analysis results.

[0105] Furthermore, the embodiment of the present application provides a device for detecting the stable state of a power system under a large disturbance. Figure 4 The hardware structure diagram of the stable state detection device of the power system under large disturbance is shown. Figure 4 The hardware structure of the stable state detection device of the power system under large disturbance may include: at least one processor 01, at least one communication interface 02, at least one memory 03 and at least one communication bus 04.

[0106] In the embodiment of the present application, the number of the processor 01 , the communication interface 02 , the memory 03 , and the communication bus 04 is at least one, and the processor 01 , the communication interface 02 , and the memory 03 communicate with each other through the communication bus 04 .

[0107] The processor 01 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0108] The memory 03 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), etc., such as at least one disk memory.

[0109] The memory stores a program, and the processor can call the program stored in the memory, and the program is used to execute the following method for detecting the stable state of the power system under a large disturbance, including:

[0110] Establish the neural differential equation estimation error function and energy estimation error function;

[0111] Combining the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function;

[0112] Establish a joint neural network model and obtain large disturbance data in the time domain;

[0113] Input the time-domain large disturbance data as training samples into the joint neural network model, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model;

[0114] The state quantity of the power system to be tested under a large disturbance is obtained, and the state quantity is processed by using the trained joint neural network model to obtain the stable state of the power system to be tested under the large disturbance.

[0115] Optionally, the detailed functions and extended functions of the program may refer to the description of the method for detecting the stable state of the power system under a large disturbance in the method embodiment.

[0116] The embodiment of the present application further provides a storage medium, which may store a program suitable for execution by a processor, and when the program is executed, controls the device where the storage medium is located to execute the following method for detecting the stable state of a power system under a large disturbance, including:

[0117] Establish the neural differential equation estimation error function and energy estimation error function;

[0118] Combining the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function;

[0119] Establish a joint neural network model and obtain large disturbance data in the time domain;

[0120] Input the time-domain large disturbance data as training samples into the joint neural network model, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model;

[0121] The state quantity of the power system to be tested under a large disturbance is obtained, and the state quantity is processed by using the trained joint neural network model to obtain the stable state of the power system to be tested under the large disturbance.

[0122] Specifically, the storage medium may be a computer-readable storage medium, and the computer-readable storage medium may be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read-Only Memory), an EPROM, a hard disk or a ROM.

[0123] Optionally, the detailed functions and extended functions of the program may refer to the description of the method for detecting the stable state of the power system under a large disturbance in the method embodiment.

[0124] In addition, the functional modules in the various embodiments of the present disclosure can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a computer device (which can be a personal computer, a live broadcast device, or a network device, etc.) to perform all or part of the steps of the methods of the various embodiments of the present disclosure.

[0125] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0126] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0127] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting the stable state of a power system under a large disturbance, characterized in that: include: Establish the neural differential equation estimation error function and energy estimation error function; Combining the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function; Establish a joint neural network model and obtain large disturbance data in the time domain; Input the time-domain large disturbance data as training samples into the joint neural network model, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model; The state quantity of the power system to be tested under a large disturbance is obtained, and the state quantity is processed by using the trained joint neural network model to obtain the stable state of the power system to be tested under the large disturbance.

2. The method according to claim 1, characterized in that The joint neural network model includes: an input module, a feature extraction module, a neural differential equation module, an energy function module, a stability analysis module and an output module; The output end of the input module is connected to the input end of the feature extraction module, the output end of the feature extraction module is respectively connected to the input end of the neural differential equation module and the input end of the energy function module, the output end of the neural differential equation module and the output end of the energy function module are respectively connected to the input end of the stability analysis module, and the output end of the stability analysis module is connected to the input end of the output module.

3. The method according to claim 1, characterized in that The neural differential equation estimation error function is: ; in, represents the neural differential equation estimation error function, Indicates samples, is the sample size, Indicates The prediction error value of samples is For the The true error value of the sample.

4. The method according to claim 1, characterized in that: The energy estimation error function is: ; in, represents the energy estimation error function, Indicates samples, is the sample size, is the proportionality constant, Indicates The error value of the sample, It means that the power system is in the error value The state quantity corresponding to It means that the power system is in the error value The energy function value corresponding to .

5. The method according to claim 1, characterized in that The combining the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function includes: The neural differential equation estimation error function is expressed as , the energy estimation error function is expressed as , combined with and , establish the obtained joint loss function for: ; in, is the weight coefficient.

6. The method according to claim 1, characterized in that The step of processing the state quantity using the trained joint neural network model to obtain a stable state of the power system under large disturbance includes: Extracting state features from the state quantity using a feature extraction module in the joint neural network model; Using the neural differential equation module in the joint neural network model to perform neural constant differential error estimation on the state feature to obtain first state data; Using the energy function module in the joint neural network model to perform energy estimation on the state feature to obtain second state data; The stability analysis module in the joint neural network model is used to perform joint state estimation on the first state data and the second state data to obtain the stable state of the power system under large disturbance.

7. The method according to claim 6, characterized in that The method of using the stability analysis module in the joint neural network model to perform joint state estimation on the first state data and the second state data to obtain the stable state of the power system under large disturbance includes: Using the stability analysis module in the joint neural network model, jointly estimating the first state data and the second state data to obtain a joint estimation value; Determine the maximum value of the energy function of the power system to be tested after being subjected to a large disturbance according to the joint estimated value; Determine, according to the joint estimated value, an energy function value of the power system to be tested at any time point after being subjected to a large disturbance, as a first energy function value; Determine, according to the joint estimated value, an energy function value of the power system to be tested at the initial equilibrium moment after being subjected to a large disturbance as a second energy function value; The stable state of the power system to be tested under a large disturbance is obtained based on the maximum value of the energy function, the first energy function value and the second energy function value.

8. A device for detecting the stable state of a power system under large disturbances, characterized in that: include: Error function building module, used to build neural differential equation estimation error function and energy estimation error function; A joint loss function establishment module, used to combine the neural differential equation estimation error function and the energy estimation error function to establish a joint loss function; Model building and data acquisition module, used to build a joint neural network model and obtain large disturbance data in the time domain; A model training module, used to input the time-domain large disturbance data as training samples into the joint neural network model, and train the joint neural network model with the goal of minimizing the joint loss function to obtain a trained joint neural network model; The stable state determination module is used to obtain the state quantity of the power system to be tested under large disturbance, and use the trained joint neural network model to process the state quantity to obtain the stable state of the power system to be tested under large disturbance.

9. A device for detecting the stable state of a power system under large disturbances, characterized in that: including memory and processor; The memory is used to store programs; The processor is used to execute the program to implement the various steps of the method for detecting the stable state of the power system under large disturbances as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the method for detecting a stable state of an electric power system under a large disturbance as described in any one of claims 1 to 7 is implemented.

Citation Information

Patent Citations

  • Optimization control method for frequency stability of island microgrid containing high-proportion renewable energy

    CN112186767A

  • Method, device and equipment for judging operation stability of power grid after fault, and medium

    CN115940253A

  • Transient voltage stability evaluation method based on residual stochastic differential equation network

    CN118095082A

  • Analysis and design of dynamical system controllers using neural differential equations

    US20210374531A1