Bad data identification method for power system and related device
By adversarially training the GAN model and combining it with the state and residual constraints of the power system, the problem that existing algorithms are difficult to detect malicious data tampering is solved, and the accuracy and stability of power system state estimation are achieved.
Patent Information
- Application Number
- CN202510808107.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-17
- Publication Date
- 2025-09-19
AI Technical Summary
Existing bad data identification algorithms are difficult to detect bad data obtained through malicious tampering, resulting in insufficient stability and reliability of the power system.
By obtaining historical system parameters in the power system state estimation process, the initial GAN model is trained for bad data identification using preset attack vectors, and an optimized GAN model is constructed. Data identification analysis is performed by combining state constraints, residual constraints, and power system constraints.
It improves the ability to detect malicious data tampering, ensures the accuracy and reliability of power system state estimation, and enhances system stability.
Smart Images
Figure CN120675758A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of power systems, and in particular to a method and related apparatus for identifying bad data in power systems. Background Art
[0002] With the deep integration of power systems and information technology, the risk of cyberattacks has increased significantly. In particular, cyberattacks targeting the power system state estimation process can lead to erroneous control decisions, jeopardizing the safety and stability of the power system. Power system state estimation is one of the fundamental tools for ensuring the safe operation of power systems. The primary purpose of state estimation is to process data measured at multiple points within the system to determine the system's real-time operating status, such as node voltages and power flows, thereby providing accurate decision-making support for dispatchers and automatic control systems.
[0003] In practice, measurement data inevitably contains noise or even errors, known as bad data. To address this, bad data identification algorithms are used in power systems to eliminate anomalous data by analyzing measurement residuals, thereby ensuring the accuracy of state estimation results. Currently used bad data identification algorithms in power system state estimation, such as residual detection and the generalized maximum likelihood method within the least squares method, are based on statistical principles and identify and process anomalous data by detecting deviations between the data and the system model.
[0004] Existing bad data identification algorithms are effective in identifying abnormal data generated by natural faults or changes in the external environment because the distribution of abnormal data fluctuates randomly and greatly. However, if the abnormal data is maliciously generated by an attacker in order to undermine the stability of the power system, then the bad data is carefully designed and may meet multiple constraints of the power system. The data distribution fluctuates slightly and is not easily detected by traditional identification algorithms. This poses a greater security threat to the power system and makes it difficult to ensure its stable and reliable operation. Summary of the Invention
[0005] The present application provides a method and related devices for identifying bad data in power systems, which are used to solve the technical problem that existing bad data identification algorithms cannot adapt to the detection and analysis of bad data obtained by malicious tampering, resulting in poor actual identification effect and failure to ensure the stability and reliability of the power system.
[0006] In view of this, the first aspect of the present application provides a method for identifying bad data in a power system, comprising:
[0007] Acquiring historical system parameters during power system state estimation, wherein the historical system parameters include node phase angle, node voltage amplitude, line active power, and line reactive power;
[0008] Performing bad data identification adversarial training on the initial GAN model based on the historical system parameters and the preset attack vector to obtain an optimized GAN model, wherein the generator of the initial GAN model includes state constraints, residual constraints, and power system constraints;
[0009] The optimized GAN model is used to identify and analyze the real system parameters in the power system state estimation process to obtain identification results.
[0010] Preferably, the method further includes: performing bad data identification adversarial training on the initial GAN model based on the historical system parameters and the preset attack vector to obtain an optimized GAN model;
[0011] generating a forged measurement vector according to a preset attack vector and an actual measurement vector;
[0012] constructing a state estimation residual function according to the forged measurement vector, and generating a residual constraint based on a residual threshold;
[0013] Constructing a generator based on the residual constraint, state constraint, power system constraint and a preset generation loss function;
[0014] A discriminator is constructed based on a preset discriminant loss function, and an initial GAN model is constructed by combining the generator and the discriminator.
[0015] Preferably, the method further includes: performing bad data identification adversarial training on the initial GAN model based on the historical system parameters and the preset attack vector to obtain an optimized GAN model;
[0016] The power system constraints are generated by combining the power balance equation and the transmission upper limit of the line transmission power. The power balance equation is expressed as:
[0017]
[0018] in, 、 、 They represent generator power, load power and network loss power respectively.
[0019] Preferably, the optimized GAN model is used to identify and analyze the real system parameters in the power system state estimation process to obtain an identification result, and then the method further includes:
[0020] The real-time state estimation process of the power system is updated and adjusted according to the identification result to obtain an optimized system state analysis result.
[0021] A second aspect of the present application provides a bad data identification device for a power system, comprising:
[0022] A parameter acquisition unit is used to acquire historical system parameters in the power system state estimation process, wherein the historical system parameters include node phase angle, node voltage amplitude, line active power and line reactive power;
[0023] a model training unit, configured to perform bad data identification adversarial training on an initial GAN model based on the historical system parameters and a preset attack vector to obtain an optimized GAN model, wherein the generator of the initial GAN model includes state constraints, residual constraints, and power system constraints;
[0024] An attack identification unit is used to use the optimized GAN model to identify and analyze the real system parameters in the power system state estimation process to obtain an identification result.
[0025] Preferably, it also includes:
[0026] a vector construction unit, configured to generate a forged measurement vector based on a preset attack vector and an actual measurement vector;
[0027] a residual constraint unit, configured to construct a state estimation residual function according to the forged measurement vector and generate a residual constraint based on a residual threshold;
[0028] A generator construction unit, configured to construct a generator according to the residual constraint, the state constraint, the power system constraint and a preset generation loss function;
[0029] A model generation unit is used to construct a discriminator based on a preset discriminant loss function, and to construct an initial GAN model by combining the generator and the discriminator.
[0030] Preferably, it also includes:
[0031] The power constraint generation unit is used to generate power system constraints by combining the power balance equation and the transmission upper limit value of the line transmission power. The power balance equation is expressed as:
[0032]
[0033] in, 、 、 They represent generator power, load power and network loss power respectively.
[0034] Preferably, it also includes:
[0035] The state updating unit is used to update and adjust the real-time state estimation process of the power system according to the identification result to obtain an optimized system state analysis result.
[0036] A third aspect of the present application provides a bad data identification device for a power system, the device comprising a processor and a memory;
[0037] The memory is used to store program code and transmit the program code to the processor;
[0038] The processor is configured to execute the bad data identification method for the power system described in the first aspect according to the instructions in the program code.
[0039] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store program code, and the program code is used to execute the bad data identification method for the power system described in the first aspect.
[0040] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:
[0041] In this application, a method for identifying bad data for a power system is provided, including: obtaining historical system parameters in the power system state estimation process, the historical system parameters including node phase angle, node voltage amplitude, line active power and line reactive power; performing bad data identification adversarial training on an initial GAN model based on the historical system parameters and a preset attack vector to obtain an optimized GAN model, the generator of the initial GAN model including state constraints, residual constraints and power system constraints; using the optimized GAN model to identify and analyze the real system parameters in the power system state estimation process to obtain an identification result.
[0042] The bad data identification method for the power system provided by the present application simulates attack data that cannot be detected by the existing bad data identification algorithm through a preset attack vector, and then generates adversarial training for the GAN model in combination with the actual historical system parameters and the preset attack vector; in this process, the GAN model can continuously learn the differences between real data and forged data, thereby identifying the characteristics of such attack data, and then obtaining data identification results; moreover, the GAN model can ensure that the data generation process of the model is more consistent with the data characteristics of the actual power system based on state constraints, residual constraints and power system constraints, thereby ensuring the accuracy of the GAN model and the identification results obtained based on the model. Therefore, the present application can solve the technical problem that the existing bad data identification algorithm cannot adapt to the detection and analysis of bad data obtained by malicious tampering, resulting in poor actual identification effect and failure to ensure the stability and reliability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 A flowchart of a method for identifying bad data in a power system according to an embodiment of the present application;
[0044] Figure 2 A schematic diagram of the structure of a bad data identification device for a power system provided in an embodiment of the present application;
[0045] Figure 3 IEEE 14-node system topology diagram provided for this application example;
[0046] Figure 4 This is a graph showing the residual error during the GAN model training process, as provided in this application example. DETAILED DESCRIPTION
[0047] In order to help those skilled in the art better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 those skilled in the art without creative work are within the scope of protection of this application.
[0048] For easier understanding, see Figure 1 , an embodiment of a method for identifying bad data in a power system provided by this application includes:
[0049] Step 101: Acquire historical system parameters in the process of power system state estimation, where the historical system parameters include node phase angle, node voltage amplitude, line active power, and line reactive power.
[0050] It is understood that the historical system parameters here refer to the actual power system data related to the power system state estimation, which can be collected in the system according to actual needs. Therefore, the historical system parameters include but are not limited to node phase angles, node voltage amplitudes, line active power, and line reactive power. The state estimation operation process of the power system is achieved by solving the nonlinear measurement equation, which is:
[0051]
[0052] in, is the state vector, To measure noise, is the measurement vector, which represents the actual measurement value of the power system, including parameters such as node voltage amplitude, Represents a nonlinear measurement function, used to describe the state vector With the measurement vector The state estimation operation of the power system is to solve the state vector in this equation , and used for subsequent system control scheduling.
[0053] Step 102: Perform bad data identification adversarial training on the initial GAN model based on historical system parameters and preset attack vectors to obtain an optimized GAN model. The generator of the initial GAN model includes state constraints, residual constraints, and power system constraints.
[0054] Furthermore, before step 102, the following steps are also included:
[0055] generating a forged measurement vector according to a preset attack vector and an actual measurement vector;
[0056] Constructing a state estimation residual function based on the forged measurement vector and generating a residual constraint based on a residual threshold;
[0057] Construct a generator based on residual constraints, state constraints, power system constraints and a preset generation loss function;
[0058] The discriminator is constructed based on the preset discriminant loss function, and the generator and discriminator are combined to construct the initial GAN model.
[0059] Furthermore, before step 102, the following steps are also included:
[0060] The power balance equation and the upper limit of the line transmission power are combined to generate the power system constraints. The power balance equation is expressed as:
[0061]
[0062] in, 、 、 They represent generator power, load power and network loss power respectively.
[0063] It should be noted that in order to maximize the research on the detection and analysis of malicious attack data that cannot be detected by traditional bad data identification algorithms, this embodiment proposes a scheme to construct a total vector; the purpose of designing the attack vector is to change the output of the power system state estimation, and the bad data generated based on this cannot be detected by traditional bad data identification algorithms. The preset attack vector is denoted as , then combined with the actual measurement vector You can generate a fake measurement vector :
[0064]
[0065] Fake measurement vector It can be regarded as a measurement vector that has been maliciously tampered with. Moreover, the attacker hopes to forge the measurement vector In the process of state estimation, deviations are generated while keeping the residual of the system from triggering the anomaly detection mechanism. Based on this, the state estimation residual function can be constructed:
[0066]
[0067] in, represents the system residual, Indicates that the forged measurement vector The state vector under influence The corresponding nonlinear estimate can also be regarded as a forged measurement vector Then, based on the state estimation residual function, the design standard of the preset attack vector can be determined, that is, the residual constraint is satisfied:
[0068]
[0069] in, is the residual threshold.
[0070] In addition, the preset attack vectors in the generator After injecting the measurement value, in addition to satisfying the residual constraint, the state constraint must also be satisfied:
[0071]
[0072] This constraint refers to the preset attack vector used This should cause a bias in the state estimation results.
[0073] Moreover, the preset attack vector added in this embodiment It is also necessary to meet the physical constraints in the power system, namely the power system constraints, which can be divided into power balance constraints and line capacity constraints. Power balance constraints are the attack vectors to ensure the design Comply with the power balance equation to avoid causing system operation abnormalities. The line capacity constraint is to ensure the line transmission power Do not exceed the physical limit, that is, the upper limit of the transmission , specifically expressed as:
[0074]
[0075] Line capacity constraints can prevent line protection measures from being triggered due to power overloads. Power system constraints can constrain designed attack vectors to conform to the power system's operational logic, ensuring that the proposed method is more consistent with the actual power system. The constructed GAN model and the resulting identification results are also more accurate, reliable, and convincing.
[0076] Furthermore, the measurement vector and They are all input into the GAN generator, and then the generator generates sample data close to it. ; The loss function of the generator designed in this embodiment is the preset generation loss function:
[0077]
[0078] in, Represents the discriminator evaluation sample data is the true probability, Represents the measurement vector The probability distribution of the data, represents the expected value of the data distribution, 、 Represents the weight coefficient, which is used to balance different loss terms.
[0079] The task of the discriminator is to distinguish the authenticity of the sample data generated by the generator, which is reflected in the form of true probability; this embodiment mainly distinguishes the actual true measurement vector and the forged measurement vector ; The loss function of the discriminator is the preset discriminant loss function:
[0080]
[0081] in, Represents the discriminator evaluation state vector is the true probability, is the state vector By optimizing the parameters of the discriminator, the model can improve its ability to detect forged data and help the generator learn attack vectors that are more difficult to detect.
[0082] The initial GAN model constructed by combining the generator and the discriminator can perform data recognition adversarial training based on randomly generated preset attack vectors in the early stage of training, continuously optimize the two loss functions, and continuously update the preset attack vectors to satisfy multiple constraints. After multiple alternating optimization iterative training, the optimized generator and discriminator can be obtained, that is, the optimized GAN model.
[0083] Step 103: Use the optimized GAN model to identify and analyze the real system parameters in the power system state estimation process to obtain identification results.
[0084] Furthermore, step 103 further includes:
[0085] The real-time state estimation process of the power system is updated and adjusted according to the identification results to obtain the optimized system state analysis results.
[0086] It is understandable that the real system parameters may include normal actual system parameters and may also include maliciously tampered attack data. They are parameters extracted from the real power system during the state estimation process. In order to ensure the accuracy and reliability of the state estimation, the real system parameters are subjected to data identification and analysis by optimizing the GAN model. Based on the probabilistic identification results, it can be clearly determined whether they are normal actual system parameters or malicious attack data.
[0087] Based on the identification results, the state estimation process of the power system can be updated and adjusted to optimize the system state analysis and provide theoretical support for subsequent system energy management. The update operation in the state estimation process can be expressed as:
[0088]
[0089] in, express The inverse function of can be used to solve the true state vector x by calculating the true measurement value z based on the attack vector a.
[0090] For ease of understanding, this application provides an application example of a bad data identification method for a power system. Figure 3 Taking the IEEE 14-bus power system as an example, the state estimation process of the system can be analyzed by combining the generation process of GAN with the state estimation mechanism of the IEEE 14-bus system. The state estimation model is still expressed as:
[0091]
[0092] in, is the state vector, To measure noise, is the measurement vector, which represents the actual measurement value of the power system, including parameters such as node voltage amplitude, Represents a nonlinear measurement function, used to describe the state vector With the measurement vector The relationship between them.
[0093] Assume that the attacker wants to change the state estimation result of a specific node, including:
[0094] (1) Voltage at node 4 is reduced to unreasonable values;
[0095] (2) Power injection value of node 5 be enlarged;
[0096] (3) Phase angle of node 6 Deviation occurs.
[0097] Attack vectors can be designed to change the measurement values of these nodes while keeping the entire system from triggering bad data identification algorithms. By performing identification adversarial training on the model, the generator can obtain attack vectors that meet the attack requirements:
[0098]
[0099] The fake data are injected into the measurement system to replace the real measurement values. The state estimation results after the attack are shown in Table 1.
[0100] Table 1 State estimation results after the system is attacked
[0101]
[0102] The state estimation results show that the voltage at node 4 is reduced, which may lead to incorrect system scheduling decisions, resulting in the assumption that the node's load has increased. The increased power injection at node 5 misleads the system into believing that the node's load has decreased. The phase angle at node 6 is artificially adjusted, potentially interfering with system stability analysis.
[0103] The attack vector generated by GAN successfully changed the state estimation results of the IEEE 14-node power system. Based on this attack vector, the discriminator model parameters were trained to increase its ability to identify attack features based on the residual, and ultimately achieve a successful identification of the vector with an accuracy rate of over 95%. The specific training process is as follows: Figure 4 shown.
[0104] The bad data identification method for the power system provided by the embodiment of the present application simulates attack data that cannot be detected by the existing bad data identification algorithm through a preset attack vector, and then performs generative adversarial training on the GAN model in combination with the actual historical system parameters and the preset attack vector; in this process, the GAN model can continuously learn the differences between real data and forged data, thereby identifying the characteristics of such attack data, and then obtaining data identification results; moreover, the GAN model can ensure that the data generation process of the model is more consistent with the data characteristics of the actual power system based on state constraints, residual constraints and power system constraints, thereby ensuring the accuracy of the GAN model and the identification results obtained based on the model. Therefore, the embodiment of the present application can solve the technical problem that the existing bad data identification algorithm cannot adapt to the detection and analysis of bad data obtained by malicious tampering, resulting in poor actual identification effect and failure to ensure the stability and reliability of the power system.
[0105] For easier understanding, see Figure 2 The present application provides an embodiment of a bad data identification device for a power system, including:
[0106] The parameter acquisition unit 201 is used to acquire historical system parameters in the power system state estimation process, where the historical system parameters include node phase angle, node voltage amplitude, line active power, and line reactive power;
[0107] A model training unit 202 is configured to perform bad data identification adversarial training on the initial GAN model based on historical system parameters and preset attack vectors to obtain an optimized GAN model, wherein the generator of the initial GAN model includes state constraints, residual constraints, and power system constraints;
[0108] The attack identification unit 203 is used to use the optimized GAN model to identify and analyze the real system parameters in the power system state estimation process to obtain identification results.
[0109] Furthermore, it also includes:
[0110] A vector construction unit 204 is configured to generate a forged measurement vector based on a preset attack vector and an actual measurement vector;
[0111] A residual constraint unit 205 is configured to construct a state estimation residual function based on the forged measurement vector and generate a residual constraint based on a residual threshold;
[0112] A generator construction unit 206 is used to construct a generator according to the residual constraint, the state constraint, the power system constraint and the preset generation loss function;
[0113] The model generation unit 207 is used to construct a discriminator based on a preset discriminant loss function, and to construct an initial GAN model by combining the generator and the discriminator.
[0114] Furthermore, it also includes:
[0115] The power constraint generation unit 208 is configured to generate power system constraints by combining the power balance equation and the transmission upper limit of the line transmission power. The power balance equation is expressed as:
[0116]
[0117] in, 、 、 They represent generator power, load power and network loss power respectively.
[0118] Furthermore, it also includes:
[0119] The state updating unit 209 is used to update and adjust the real-time state estimation process of the power system according to the identification result to obtain an optimized system state analysis result.
[0120] The present application also provides a bad data identification device for a power system, the device including a processor and a memory;
[0121] The memory is used to store program codes and transmit the program codes to the processor;
[0122] The processor is configured to execute the bad data identification method for the power system in the above method embodiment according to instructions in the program code.
[0123] The present application also provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute the bad data identification method for the power system in the above method embodiment.
[0124] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0125] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.
[0126] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.
[0127] If the integrated unit is implemented in the form of a software functional unit 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 application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for executing all or part of the steps of the method described in each embodiment of the present application through a computer device (which can be a personal computer, server, or network device, etc.). The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (full name: Read-Only Memory, English abbreviation: ROM), random access memory (full name: Random Access Memory, English abbreviation: RAM), disk or optical disk, and other media that can store program code.
[0128] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A method for identifying bad data in a power system, characterized in that: include: Acquiring historical system parameters during power system state estimation, wherein the historical system parameters include node phase angle, node voltage amplitude, line active power, and line reactive power; Performing bad data identification adversarial training on the initial GAN model based on the historical system parameters and the preset attack vector to obtain an optimized GAN model, wherein the generator of the initial GAN model includes state constraints, residual constraints, and power system constraints; The optimized GAN model is used to identify and analyze the real system parameters in the power system state estimation process to obtain identification results.
2. The method for identifying bad data in a power system according to claim 1, wherein: The method further includes: performing bad data identification adversarial training on the initial GAN model based on the historical system parameters and the preset attack vector to obtain an optimized GAN model; generating a forged measurement vector according to a preset attack vector and an actual measurement vector; constructing a state estimation residual function according to the forged measurement vector, and generating a residual constraint based on a residual threshold; Constructing a generator based on the residual constraint, state constraint, power system constraint and a preset generation loss function; A discriminator is constructed based on a preset discriminant loss function, and an initial GAN model is constructed by combining the generator and the discriminator.
3. The method for identifying bad data in a power system according to claim 1, wherein: The method further includes: performing bad data identification adversarial training on the initial GAN model based on the historical system parameters and the preset attack vector to obtain an optimized GAN model; The power system constraints are generated by combining the power balance equation and the transmission upper limit of the line transmission power. The power balance equation is expressed as: in, 、 、 They represent generator power, load power and network loss power respectively.
4. The method for identifying bad data in a power system according to claim 1, wherein: The method further includes: using the optimized GAN model to identify and analyze the real system parameters in the power system state estimation process to obtain an identification result, and then further including: The real-time state estimation process of the power system is updated and adjusted according to the identification result to obtain an optimized system state analysis result.
5. A bad data identification device for a power system, characterized in that: include: A parameter acquisition unit is used to acquire historical system parameters in the power system state estimation process, wherein the historical system parameters include node phase angle, node voltage amplitude, line active power and line reactive power; a model training unit, configured to perform bad data identification adversarial training on an initial GAN model based on the historical system parameters and a preset attack vector to obtain an optimized GAN model, wherein the generator of the initial GAN model includes state constraints, residual constraints, and power system constraints; An attack identification unit is used to use the optimized GAN model to identify and analyze the real system parameters in the power system state estimation process to obtain an identification result.
6. The bad data identification device for the power system according to claim 5, characterized in that: Also includes: a vector construction unit, configured to generate a forged measurement vector based on a preset attack vector and an actual measurement vector; a residual constraint unit, configured to construct a state estimation residual function according to the forged measurement vector and generate a residual constraint based on a residual threshold; A generator construction unit, configured to construct a generator according to the residual constraint, the state constraint, the power system constraint and a preset generation loss function; A model generation unit is used to construct a discriminator based on a preset discriminant loss function, and to construct an initial GAN model by combining the generator and the discriminator.
7. The bad data identification device for the power system according to claim 5, characterized in that: Also includes: The power constraint generation unit is used to generate power system constraints by combining the power balance equation and the transmission upper limit value of the line transmission power. The power balance equation is expressed as: in, 、 、 They represent generator power, load power and network loss power respectively.
8. The bad data identification device for the power system according to claim 5, characterized in that: Also includes: The state updating unit is used to update and adjust the real-time state estimation process of the power system according to the identification result to obtain an optimized system state analysis result.
9. Bad data identification equipment for power systems, characterized in that: The device includes a processor and a memory; The memory is used to store program code and transmit the program code to the processor; The processor is configured to execute the bad data identification method for a power system according to any one of claims 1 to 4 according to instructions in the program code.
10. A computer-readable storage medium, characterized in that The computer-readable storage medium is used to store program code, and the program code is used to execute the bad data identification method for the power system according to any one of claims 1 to 4.