A real-time control method and system for transformer current test parameters

Abstract

The present invention provides a method and system for real-time control of transformer current-carrying test parameters, relating to the field of transformer current-carrying test control. The method comprises: collecting status data from the transformer current-carrying test, constructing the status data into a state model using an undirected graph data structure; detecting abnormal nodes in the state model using an anomaly detection model; activating anomaly retrieval upon detecting an abnormal node, searching for other potential abnormal nodes, sorting the other potential abnormal nodes in descending order of retrieval priority, and sequentially using the anomaly detection model to detect abnormalities in the other nodes; and dynamically adjusting the transformer current-carrying test parameters based on the anomaly detection results: when no other nodes are abnormal, single-node abnormal control is adopted; when other nodes are abnormal, multi-node associated abnormal control is adopted. The present invention utilizes a graph model to achieve multi-parameter associated quantification with a lower performance load, enabling earlier detection of overheating trends compared to traditional methods.

Description

Technical Field

[0001] The present invention relates to the field of transformer current flow test control, and in particular to a method and system for real-time control of transformer current flow test parameters. Background Art

[0002] Transformer current-through testing requires applying high current (up to 120% of the rated current) to verify the winding's load-carrying capacity. However, this presents two major risks: uneven current distribution can lead to sudden temperature rises in hot spots, causing local overheating and insulation degradation; and electromagnetic forces can cause excessive core vibration, cumulatively damaging the structure and causing mechanical damage. The core dilemma of transformer current-through testing lies in balancing high current requirements with equipment safety. Existing technologies fall into two main categories: traditional threshold methods rely on manual experience to set fixed thresholds, while intelligent detection methods such as LSTM prediction ignore the coupling relationship between multiple parameters.

[0003] Disadvantages of existing technologies: The single-parameter threshold method cannot provide early warning in the early stages of current harmonics, has hysteresis, and has a certain test interruption rate; multi-parameter methods such as the CNN-LSTM model have too high a computational load, ignore multi-parameter coupling, and have a false alarm rate; multi-parameter methods such as CNN extracting multi-sensor features also have the problem of large computational load and cannot be controlled in real time, and response delays lead to increased damage. Summary of the Invention

[0004] Purpose of the invention: To propose a method for real-time control of transformer current test parameters, and further provide a system capable of running and implementing the above method, so as to effectively solve the above problems mentioned in the background technology.

[0005] The technical solutions of the present invention are as follows:

[0006] In the first aspect, a method for real-time control of transformer current test parameters is proposed, the method comprising the following steps:

[0007] S1. Collecting status data of a transformer current-carrying test and constructing a status model using an undirected graph data structure;

[0008] S2. Detect abnormal nodes in the state model using an anomaly detection model;

[0009] S3. When an abnormal node is detected, an abnormal search is activated and other potential abnormal nodes are searched. Other potential abnormal nodes are sorted in descending order of search priority, and an abnormal detection model is used to detect whether other nodes are abnormal.

[0010] S4. Dynamically adjust the transformer current test parameters according to the abnormality detection result in step S3:

[0011] When no other nodes are abnormal, single node abnormality control is adopted;

[0012] When there are other node anomalies, multi-node associated anomaly control is adopted.

[0013] A further improvement of the present invention is that the state data of the transformer current-carrying test in step S1 includes: collecting winding current time series data through a current sensor; collecting hot spot temperature time series data through an infrared thermometer; collecting core vibration time series data through a vibration sensor; collecting discharge soundprint time series data through an acoustic sensor; and collecting environmental time series data through an environmental temperature and humidity sensor.

[0014] A further improvement of the present invention is that the state model in step S1 is an undirected graph , the undirected graph G includes a node set V, an edge set E, a node attribute matrix X and an adjacency matrix A, the node set ,in, Indicates current data, Represents temperature data, Indicates vibration data, Represents voiceprint data, Represents environmental data; the node attribute matrix ,in, Indicates the dimension The real number space of , where d represents the total number of time steps contained in the state data during a transformer current test cycle, and the i-th row vector of the node attribute matrix X ,in, Indicates the dimension The real space of , then the vector Represents the i-th node in the undirected graph G The attribute vector is the i-th state data collected during a transformer current test cycle, and the value of i is 1-5.

[0015] A further improvement of the present invention is that any edge in the edge set E The value of is calculated by the following formula:

[0016]

[0017] in, For the i-th node With the jth node The edge between them, the value of j is 1-5, represents the L1 norm of the vector, is the normalization coefficient, is the physical association weight, is the physical association strength between nodes, where when i=1, j=2, Indicates the current-temperature correlation strength, with a value of 0.9. When i=1, j=3, Indicates the current-vibration correlation strength, with a value of 0.7. .

[0018] A further improvement of the present invention is that the adjacency matrix ,in, Indicates the dimension The real space of , the adjacency matrix Elements in .

[0019] A further improvement of the present invention is that S2 comprises the following specific steps:

[0020] S21, for each node Building an Autoencoder Anomaly Detection Model , the processing process is:

[0021]

[0022]

[0023] in, is the anomaly detection model for the i-th node The encoder, is the anomaly detection model for the i-th node decoder; for The output encoding result is express Output reconstruction results;

[0024] S22. Calculate the reconstruction error. The calculation formula is:

[0025]

[0026] Where k represents the kth time step;

[0027] S23. Determine whether the node is abnormal. The calculation formula is:

[0028]

[0029] in, is the anomaly detection threshold, For nodes The node risk coefficient is , , , , ;

[0030] S24, when When the i-th node is determined to be abnormal, an abnormal warning is triggered.

[0031] A further improvement of the present invention is that S3 includes the following specific steps:

[0032] S31, when When the i-th node is determined to be abnormal, abnormal retrieval is activated;

[0033] S32. Determine the set of potentially impacted nodes ;

[0034] S33, computing nodes The retrieval priority is calculated as follows:

[0035]

[0036] in, is the i-th node in the undirected graph G To the jth node The number of hops of the shortest path; is the maximum element value of the adjacency matrix A; Represents the jth node The search priority of

[0037] S34. Arrange the other nodes in descending order according to the retrieval priority, and use the anomaly detection model to detect whether the other nodes are abnormal.

[0038] A further improvement of the present invention is that step S4 includes the following specific steps:

[0039] S41. When there are no other node abnormalities, single node abnormality control is adopted. The control formula is:

[0040]

[0041] in, is the regulated current, is the current;

[0042] S42. When other nodes are abnormal, multi-node associated abnormality control is adopted. The control formula is:

[0043]

[0044] in, is the regulated current, is the current, Indicates the current abnormal node set;

[0045] S43. When all nodes meet When the recovery test current is 0.98 times the rated current of the transformer.

[0046] Secondly, a real-time control system for transformer current test parameters is proposed, which includes: a model building module, an abnormality retrieval module, a secondary retrieval module, and a parameter adjustment module;

[0047] The model building module is used to collect the state data of the transformer current test and construct the state data into a state model in the data structure of an undirected graph;

[0048] The anomaly retrieval module is used to detect abnormal nodes in the state model using the anomaly detection model;

[0049] The secondary retrieval module is used to activate anomaly retrieval when an abnormal node is detected, retrieve other potential abnormal nodes, sort other potential abnormal nodes in descending order according to the size of the retrieval priority, and use the anomaly detection model to detect whether other nodes are abnormal;

[0050] The parameter adjustment module is used to dynamically adjust the transformer current test parameters according to the abnormality detection results: when there are no other node abnormalities, single-node abnormality control is adopted; when there are other node abnormalities, multi-node associated abnormality control is adopted.

[0051] Beneficial effects:

[0052] (1) The present invention collects winding current time series data, hot spot temperature time series data, core vibration time series data, discharge soundprint time series data, and environmental time series data, and cross-validates multiple physical quantities to significantly solve the hysteresis of the traditional "single parameter threshold method" and reduce the false alarm rate.

[0053] (2) The present invention associates and quantifies the status data of the transformer current-carrying test in the form of an undirected graph. Compared with the traditional CNN-LSTM model, this method achieves multi-parameter coupling with a lower performance load.

[0054] (3) The present invention establishes an autoencoder anomaly detection model for each node in the undirected graph, uses the autoencoder anomaly detection model to detect abnormal nodes and activate anomaly retrieval, and dynamically adjusts the transformer current test parameters to capture overheating trends in advance.

[0055] (4) The present invention adopts a hierarchical dynamic control strategy. When there are no other node abnormalities, single-node abnormality control is adopted; when there are other node abnormalities, multi-node associated abnormality control is adopted to avoid the traditional "one-size-fits-all" power outage. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of a method for real-time control of transformer current test parameters according to Example 1 of the present invention.

[0057] Figure 2 This is a structural diagram of a real-time control system for transformer current test parameters according to Example 2 of the present invention. DETAILED DESCRIPTION

[0058] In the following description, numerous specific details are provided to provide a more thorough understanding of the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced without one or more of these details. In other instances, certain technical features well known in the art have not been described to avoid confusion with the present invention.

[0059] Example 1:

[0060] This embodiment constructs a real-time control method for transformer current-carrying test parameters. This method collects transformer current-carrying test status data to construct a state model. It then uses an anomaly detection model to detect abnormal nodes in the state model. Upon detecting an abnormal node, it activates anomaly retrieval to retrieve other potentially abnormal nodes. The transformer current-carrying test parameters are then dynamically adjusted based on the anomaly detection results. A graphical model quantifies the associations between multiple parameters, enabling the detection of overheating trends earlier than existing methods. A priority retrieval mechanism reduces computational effort. Furthermore, closed-loop control employs gradual current adjustment to avoid test interruptions.

[0061] Figure 1 The following are the specific steps of the method for real-time control of transformer current test parameters in this embodiment:

[0062] S1. Collecting status data of a transformer current-carrying test and constructing a status model using an undirected graph data structure;

[0063] S2. Detect abnormal nodes in the state model using an anomaly detection model;

[0064] S3. When an abnormal node is detected, an abnormal search is activated and other potential abnormal nodes are searched. Other potential abnormal nodes are sorted in descending order of search priority, and an abnormal detection model is used to detect whether other nodes are abnormal.

[0065] S4. Dynamically adjust the transformer current test parameters according to the abnormality detection result in step S3:

[0066] When no other nodes are abnormal, single node abnormality control is adopted;

[0067] When there are other node anomalies, multi-node associated anomaly control is adopted.

[0068] In this embodiment, the status data collected in S1 of the transformer current-carrying test includes: collecting winding current time series data through a current sensor; collecting hot spot temperature time series data through an infrared thermometer; collecting core vibration time series data through a vibration sensor; collecting discharge soundprint time series data through an acoustic sensor; and collecting environmental time series data through an environmental temperature and humidity sensor.

[0069] In this embodiment, the state model in S1 is an undirected graph , the undirected graph G contains the node set V, the edge set E, the node attribute matrix X and the adjacency matrix A, the node set ,in, Indicates current data, Represents temperature data, Indicates vibration data, Represents voiceprint data, Represents environmental data; node attribute matrix ,in, Indicates the dimension The real number space of , where d represents the total number of time steps contained in the state data during a transformer current test cycle, and the i-th row vector of the node attribute matrix X ,in, Indicates the dimension The real space of , then the vector Represents the i-th node in the undirected graph G The attribute vector is the i-th state data collected during a transformer current test cycle, and the value of i is 1-5.

[0070] In this embodiment, any edge in the edge set E The value of is calculated by formula (1):

[0071]

[0072] in, For the i-th node With the jth node The edge between them, the value of j is 1-5, represents the L1 norm of the vector, is the normalization coefficient, is the physical association weight, is the physical association strength between nodes, where when i=1, j=2, Indicates the current-temperature correlation strength, with a value of 0.9. When i=1, j=3, Indicates the current-vibration correlation strength, with a value of 0.7. .

[0073] In this embodiment, the adjacency matrix ,in, Indicates the dimension The real space of , the adjacency matrix Elements in .

[0074] In this embodiment, S2 includes the following specific steps:

[0075] S21, for each node Building an Autoencoder Anomaly Detection Model , the processing process is shown in formula (2) and formula (3):

[0076]

[0077]

[0078] in, is the anomaly detection model for the i-th node The encoder, is the anomaly detection model for the i-th node decoder; for The output encoding result is express Output reconstruction results;

[0079] S22. Calculate the reconstruction error according to the following formula (4):

[0080]

[0081] Where k represents the kth time step;

[0082] S23. Determine whether the node is abnormal. The calculation formula is shown in the following formula (5):

[0083]

[0084] in, is the anomaly detection threshold, For nodes The node risk coefficient is , , , , ;

[0085] S24, when When the i-th node is determined to be abnormal, an abnormal warning is triggered.

[0086] In this embodiment, S3 includes the following specific steps:

[0087] S31, when When the i-th node is determined to be abnormal, abnormal retrieval is activated;

[0088] S32. Determine the set of potentially impacted nodes ;

[0089] S33, computing nodes The retrieval priority is calculated by the following formula (6):

[0090]

[0091] in, is the i-th node in the undirected graph G To the jth node The number of hops of the shortest path; is the maximum element value of the adjacency matrix A; Represents the jth node The search priority of

[0092] S34. Arrange the other nodes in descending order according to the retrieval priority, and use the anomaly detection model to detect whether the other nodes are abnormal.

[0093] In this embodiment, S4 includes the following specific steps:

[0094] S41. When there are no other node anomalies, single node anomaly control is adopted. The control formula is shown in formula (7):

[0095]

[0096] in, is the regulated current, is the current;

[0097] S42. When there are other node anomalies, multi-node associated anomaly control is adopted. The control formula is shown in formula (8):

[0098]

[0099] in, is the regulated current, is the current, Indicates the current abnormal node set;

[0100] S43. When all nodes meet When the recovery test current is 0.98 times the rated current of the transformer.

[0101] Example 2:

[0102] This embodiment proposes a real-time control system for transformer current test parameters, such as Figure 2As shown in the figure, the system consists of four parts: model building module, anomaly retrieval module, secondary retrieval module, and parameter adjustment module.

[0103] The model building module is used to collect the status data of the transformer current test and build the status model;

[0104] The anomaly retrieval module is used to detect abnormal nodes in the state model using the anomaly detection model;

[0105] The secondary retrieval module is used to activate anomaly retrieval when an abnormal node is detected and retrieve other potential abnormal nodes;

[0106] The parameter adjustment module is used to dynamically adjust the transformer current test parameters according to the abnormality detection results.

[0107] In this embodiment, collecting status data of the transformer current-carrying test includes: collecting winding current time series data through a current sensor; collecting hot spot temperature time series data through an infrared thermometer; collecting core vibration time series data through a vibration sensor; collecting discharge soundprint time series data through an acoustic sensor; and collecting environmental time series data through an environmental temperature and humidity sensor.

[0108] The above-mentioned parameters and steps for each unit module to achieve corresponding functions in a real-time control system for transformer current test parameters of the present invention can refer to the parameters and steps in the embodiment of a real-time control method for transformer current test parameters in Example 1 above.

[0109] for example:

[0110] The status data of the transformer current-through test includes winding current time series data collected by current sensors, hot spot temperature time series data collected by infrared thermometers, core vibration time series data collected by vibration sensors, discharge soundprint time series data collected by acoustic sensors, and environmental time series data collected by ambient temperature and humidity sensors.

[0111] For example:

[0112] The state model constructed by the model construction module is an undirected graph , the undirected graph G includes a node set V, an edge set E, a node attribute matrix X and an adjacency matrix A.

[0113] The node set ,in, Indicates current data, Represents temperature data, Indicates vibration data, Represents voiceprint data, Represents environmental data; the node attribute matrix ,in, Indicates the dimension The real number space of , where d represents the total number of time steps contained in the state data during a transformer current test cycle, and the i-th row vector of the node attribute matrix X ,in, Indicates the dimension The real space of , then the vector Represents the i-th node in the undirected graph G The attribute vector is the i-th state data collected during a transformer current test cycle, and the value of i is 1-5.

[0114] Any edge in the edge set E The value of is calculated by formula (1) in Example 1 and will not be repeated here.

[0115] Adjacency Matrix ,in, Indicates the dimension The real space of , the adjacency matrix Elements in .

[0116] For example:

[0117] The anomaly retrieval module first performs Building an Autoencoder Anomaly Detection Model The processing process is shown in formula (2) and formula (3) in Example 1 and will not be repeated here.

[0118] Then, the reconstruction error is calculated according to formula (4) in embodiment 1, and then whether the node is abnormal is determined according to formula (5) in embodiment 1. If a node is determined to be abnormal, an abnormal warning is triggered.

[0119] For example:

[0120] The implementation of the secondary retrieval module includes the following specific steps: when the i-th node is determined to be abnormal, activate the abnormal retrieval; then determine the set of potential affected nodes ; Further, according to formula (6) in Example 1, the node is calculated Finally, the other nodes are sorted in descending order according to the size of the retrieval priority, and the anomaly detection model is used to detect whether the other nodes are abnormal.

[0121] For example:

[0122] The parameter adjustment module is used to dynamically adjust the transformer current test parameters according to the abnormality detection results: when there are no other node abnormalities, single-node abnormality control is adopted, and the control formula is shown in Formula (7) in Example 1. When there are other node abnormalities, multi-node associated abnormality control is adopted, and the control formula is shown in Formula (8) in Example 1.

[0123] Example 3

[0124] This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor; the processor executes the above-mentioned method for real-time control of transformer current test parameters by calling the computer program stored in the memory.

[0125] This electronic device may vary significantly depending on its configuration or performance. It can include one or more processors (Central Processing Units, CPUs) and one or more memories, wherein the memories store at least one computer program, which is loaded and executed by the processor to implement the method for real-time control of transformer current test parameters provided in the above-mentioned method embodiment. The electronic device can also include other components for implementing the device's functions. For example, the electronic device can also include components such as wired or wireless network interfaces and input / output interfaces for data input and output. This embodiment is not described in detail here.

[0126] Those skilled in the art will appreciate that the present invention may be implemented as a system, method, or computer program product. Therefore, the present disclosure may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, the present invention may be implemented as a computer program product embodied in one or more computer-readable media containing computer-readable program code.

[0127] Any combination of one or more computer-readable media can be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, 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 thereof. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or device.

[0128] The present invention is described with reference to flowcharts and block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process or block in the flowcharts and block diagrams, as well as combinations of processes and blocks in the flowcharts or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts. Figure 1 A process or multiple processes and boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 A process or multiple processes and boxes Figure 1 A step that specifies a function in one or more boxes.

[0130] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which are all protected by the present invention.

Claims

1. A method for real-time control of transformer current test parameters, characterized in that: The steps include: S1. Collecting status data of a transformer current-carrying test and constructing a status model using an undirected graph data structure; S2. Detect abnormal nodes in the state model using an anomaly detection model; S21, for each node Building an Autoencoder Anomaly Detection Model , the processing process is: in, is the anomaly detection model for the i-th node The encoder, is the anomaly detection model for the i-th node decoder; for The output encoding result is express Output reconstruction results; S22. Calculate reconstruction error : Where k represents the kth time step; S23. Determine whether the node is abnormal. The calculation formula is: in, is the anomaly detection threshold, For nodes The node risk coefficient is , , , , ; when When the i-th node is judged to be abnormal, an abnormal warning is triggered; S3. When an abnormal node is detected, an abnormal search is activated and other potential abnormal nodes are searched. Other potential abnormal nodes are sorted in descending order of search priority, and an abnormal detection model is used to detect whether other nodes are abnormal. S4. Dynamically adjust the transformer current test parameters according to the abnormality detection result in step S3: When there are no other node abnormalities, single node abnormality control is adopted, and the control formula is: in, is the regulated current, is the current; When other nodes are abnormal, multi-node associated abnormality control is adopted. The control formula is: in, is the regulated current, is the current, Indicates the current abnormal node set; When all nodes meet When the recovery test current is 0.98 times the rated current of the transformer.

2. The method for real-time control of transformer current test parameters according to claim 1, characterized in that: The status data of the transformer current-carrying test in step S1 includes: collecting winding current time series data through a current sensor; collecting hot spot temperature time series data through an infrared thermometer; collecting core vibration time series data through a vibration sensor; collecting discharge soundprint time series data through an acoustic sensor; and collecting environmental time series data through an environmental temperature and humidity sensor.

3. The real-time control method for transformer current test parameters according to claim 2 is characterized in that: The state data is converted into an undirected graph The data structure is composed of a state model, where V is the node set, E is the edge set, X is the node attribute matrix, and A is the adjacency matrix.

4. The method for real-time control of transformer current test parameters according to claim 3, characterized in that: The node set ,in, Derived from the winding current timing data, Derived from the hotspot temperature time series data, Derived from the core vibration time series data, Derived from the discharge voiceprint time series data, Derived from the environmental time series data.

5. The method for real-time control of transformer current test parameters according to claim 3, characterized in that: The node attribute matrix ,in, Indicates the dimension The real number space of , d represents the total number of time steps contained in the state data during a transformer current test cycle; The i-th row vector of the node attribute matrix X ,in, Indicates the dimension The real space of vector Represents the i-th node in the undirected graph G The attribute vector is the i-th state data collected during a transformer current test cycle.

6. The method for real-time control of transformer current test parameters according to claim 3, characterized in that: Any edge in the edge set E The value of is calculated by the following formula: Where, For the i-th node With the jth node The edge between them, the value of j is 1-5; Represents the L1 norm of the vector; is the normalization coefficient; is the physical association weight; is the physical association strength between nodes, where when i=1, j=2, Indicates the current-temperature correlation strength, with a value of 0.

9. When i=1, j=3, Indicates the current-vibration correlation strength, with a value of 0.

7. ; The adjacency matrix ,in, Indicates the dimension The real space of , the adjacency matrix Elements in .

7. The method for real-time control of transformer current test parameters according to claim 6, characterized in that: Step S3 specifically includes: S31, when When the i-th node is determined to be abnormal, abnormal retrieval is activated; S32. Determine the set of potentially impacted nodes ; S33, computing nodes Search priority : ； in, is the i-th node in the undirected graph G To the jth node The number of hops in the shortest path; is the maximum element value of the adjacency matrix A; Represents the jth node The search priority of S34. Arrange the other nodes in descending order according to the retrieval priority, and use the anomaly detection model to detect whether the other nodes are abnormal.

8. A real-time control system for transformer current test parameters, used to execute the real-time control method for transformer current test parameters according to any one of claims 1 to 7, characterized in that: The real-time control system for transformer current test parameters includes: A model building module is used to collect state data of a transformer current-carrying test and construct a state model from the state data in an undirected graph data structure; Anomaly retrieval module, used to detect abnormal nodes in the state model using an anomaly detection model; The secondary retrieval module is used to activate anomaly retrieval when an abnormal node is detected, retrieve other potential abnormal nodes, sort other potential abnormal nodes in descending order according to the size of the retrieval priority, and use the anomaly detection model to detect whether other nodes are abnormal; The parameter adjustment module is used to dynamically adjust the transformer current test parameters according to the abnormality detection results: when there are no other node abnormalities, single-node abnormality control is adopted; when there are other node abnormalities, multi-node associated abnormality control is adopted.

Images