Automatic identification methods, systems, and deployment methods for power system edge gateways

Abstract

This invention discloses an automatic identification method for power system edge gateways, comprising: acquiring and preprocessing existing edge gateway device data to construct a training dataset; constructing a primary automatic identification model for power system edge gateways and training it to obtain an automatic identification model; collecting data from the edge gateway devices to be identified in real time; and using the obtained automatic identification model to complete the automatic identification of the power system edge gateways. This invention also discloses a system for implementing the automatic identification method for power system edge gateways, and a deployment method incorporating the automatic identification method for power system edge gateways. This invention not only automates the automatic identification and deployment of power system edge gateways but also offers high reliability, accuracy, and efficiency.

Description

Technical Field

[0001] This invention belongs to the field of electrical automation, and specifically relates to an automatic identification method, system, and deployment method for power system edge gateways. Background Technology

[0002] With economic and technological development and the improvement of people's living standards, electricity has become an indispensable secondary energy source in people's production and daily life, bringing endless convenience. Therefore, ensuring a stable and reliable supply of electricity has become one of the most important tasks of the power system.

[0003] Currently, smart grid and Internet of Things (IoT) technologies are developing rapidly, and more and more smart devices are being integrated into the power system through edge gateway devices; the number and types of edge gateway devices are also increasing rapidly. These devices are manufactured by different companies and have their own hardware and software characteristics, thus requiring customized deployment on IoT management platforms.

[0004] Currently, the deployment scheme for power system edge gateways generally adopts a manual deployment method: power system personnel first manually identify the edge network devices, and then manually connect, set parameters, and install software for the edge network devices according to their characteristics and configuration requirements.

[0005] However, this method of manually identifying and deploying edge gateway devices is not only extremely labor-intensive and inefficient, but also prone to human error. Furthermore, the complex and ever-changing operating environment of power systems necessitates that edge gateway devices possess rapid response and flexible adaptability; however, existing manual configuration methods struggle to achieve real-time response and dynamic adjustment, failing to effectively address changes in system requirements. Summary of the Invention

[0006] One of the objectives of this invention is to provide an automatic identification method for power system edge gateways that is highly reliable, accurate, and efficient.

[0007] The second objective of this invention is to provide a system for implementing the automatic identification method of the power system edge gateway.

[0008] A third objective of this invention is to provide a deployment method that includes the automatic identification method for the power system edge gateway.

[0009] The automatic identification method for power system edge gateways provided by this invention includes the following steps:

[0010] S1. Obtain data information from existing edge gateway devices;

[0011] S2. Preprocess the data obtained in step S1 to construct a training dataset;

[0012] S3. Based on the self-attention mechanism, construct a primary model for automatic identification of power system edge gateways;

[0013] S4. Using the training dataset constructed in step S2, train the primary automatic identification model of the power system edge gateway constructed in step S3 to obtain the automatic identification model of the power system edge gateway.

[0014] S5. Real-time acquisition of data information from the edge gateway device to be identified;

[0015] S6. Based on the data information obtained in step S5, the automatic identification model of the power system edge gateway obtained in step S4 is used to complete the automatic identification of the power system edge gateway.

[0016] Step S1, which involves obtaining data information from existing edge gateway devices, specifically includes the following steps:

[0017] Obtain data information from existing edge gateway devices; the data information includes process data information, network data information, and system data information of the edge gateway devices.

[0018] The preprocessing described in step S2 specifically includes the following steps:

[0019] Statistical features are extracted from the data information obtained in step S1; the statistical features include average CPU utilization, average memory utilization, number of network connections, and system version.

[0020] The extracted statistical features and the data information obtained in step S1 are standardized.

[0021] The standardized statistical features and the data information obtained in step S1 are combined to construct a training dataset.

[0022] Step S3, which describes the construction of a primary model for automatic identification of power system edge gateways based on a self-attention mechanism, includes the following steps:

[0023] The constructed power system edge gateway automatic identification primary model includes an input layer, a self-attention mechanism layer, a first residual connection layer, a first Layer Norm layer, a progressive loop cross-attention layer, a second residual connection layer, a second Layer Norm layer, a feature fusion layer, a fully connected layer, a softmax layer, and an output layer, which are connected in sequence.

[0024] The input layer is used to receive the input feature data;

[0025] The self-attention mechanism layer is used to map the input feature data to a unified dimension, extract the corresponding query matrix, key matrix and value matrix, and process them with the self-attention mechanism to obtain self-attention features;

[0026] The first residual connection layer and the first Layer Norm layer are used to perform residual connections on the input of the self-attention mechanism layer and the obtained self-attention features, and then perform layer normalization to obtain intermediate features. The first residual connection layer and the first Layer Norm layer are used to alleviate the problem that gradients may vanish or explode during backpropagation, which may make the model difficult to train. At the same time, they can also accelerate model training and improve model stability and generalization ability.

[0027] The progressive loop cross-attention layer is used to perform cross-attention processing on the intermediate features output by the first Layer Norm layer in order to extract cross-combined features;

[0028] The second residual connection layer and the second Layer Norm layer are used to perform residual connections on the input of the progressive loop cross attention layer and the obtained cross-combination features, and then perform layer normalization to obtain cross attention features. The second residual connection layer and the second Layer Norm layer are used to alleviate the problem that gradients may vanish or explode during backpropagation, which may make the model difficult to train. At the same time, they can also accelerate model training and improve model stability and generalization ability.

[0029] The feature fusion layer is used to concatenate and linearize the cross-attention features output by the second Layer Norm layer to obtain fused features;

[0030] The fully connected layer is used to expand and linearly transform the obtained fused features to obtain an unnormalized score vector for each possible category;

[0031] The softmax layer is used to smooth the obtained score vector to obtain the probability of belonging to each category. The softmax layer can suppress small scores and amplify large scores, thus making the output probability distribution sharper, which is conducive to clearly identifying the most likely category.

[0032] The output layer is used to output the final power system edge gateway identification result.

[0033] Step S3 specifically includes the following steps:

[0034] Input layer: The process characteristics X of the input layer receiving input. p Network characteristics X n and system characteristics X s ;

[0035] Self-attention mechanism layer: The self-attention layer is used to apply a weight matrix W and a bias vector b to the input feature vector (X). p ,X n ,X s Mapping to a unified dimension yields the corresponding query matrix, key matrix, and value matrix, represented as (Q p ,K p V p ), (Q n ,K n V n ) and (Q s ,K s V s ), where Q p For the process feature query matrix, K p V is the process feature key matrix. p Let Q be the process eigenvalue matrix. n For the network feature query matrix, K n V is the network feature key matrix. n Let Q be the network eigenvalue matrix. s To query the feature query matrix, K s To query the feature key matrix, V s To query the feature matrix, a self-attention mechanism is then applied to the query matrix, key matrix, and value matrix to obtain the process feature self-attention feature Z. p Network features, self-attention features Z n and system features self-attention features Z s ;

[0036] First residual connection layer and first Layer Norm layer: for X p and Z p X n and Z n X s and Z s The process employs a first residual connection layer for residual connection, followed by a first Layer Normalization layer for layer normalization, to obtain the first intermediate feature X' of the process. p The first intermediate feature X' of the network n and the first intermediate feature X of the system s ';

[0037] Progressive ring cross-attention layer: X' p 、X' n and X s Perform progressive cyclic cross combinations to obtain the first combination (X') p ,X' n ), second combination (X' n ,X s ') and the third combination (X)s ',X' p ); targeting (X' p ,X' n ), (X' n ,X s ') and (X s ',X' p Each feature is processed using a self-attention mechanism to obtain the cross-combination feature Z. pn Z ns and Z sp ;

[0038] Second residual connection layer and second Layer Norm layer: used for processing intermediate feature X' p 、X' n and X s 'and the obtained cross-combination feature Z pn Z ns and Z sp A second residual connection layer is used for residual connection, followed by a second Layer Norm layer for layer normalization, to obtain the cross-attention feature X'. p ' n 、X' n ' s and X s " p ;

[0039] Feature fusion layer: The obtained cross-attention features X' p ' n 、X' n ' s and X s " p The features are then concatenated and linearized using a linear layer to obtain the fused feature C.

[0040] Fully connected layer: The obtained fused features C are expanded and linearly transformed to obtain the unnormalized score vector S for each possible category;

[0041] The softmax layer is used to smooth the obtained score vector S to obtain the probability P belonging to each category.

[0042] Output layer: The output layer is used to output the final power system edge gateway identification result.

[0043] This invention also provides a system for implementing the automatic identification method of the power system edge gateway, comprising a data acquisition module, a data processing module, a model building module, a model training module, a data acquisition module, and a gateway identification module; the data acquisition module, data processing module, model building module, model training module, data acquisition module, and gateway identification module are connected in series; the data acquisition module is used to acquire data information from existing edge gateway devices and upload the data information to the data processing module; the data processing module is used to preprocess the acquired data information based on the received data information, construct a training dataset, and upload the data information to the model building module; the model building module is used to, based on the received data information, perform self-attention... The mechanism involves constructing a primary model for automatic identification of power system edge gateways and uploading the data to the model training module. The model training module trains the primary model using the received data and a pre-constructed training dataset to obtain an automatic identification model for power system edge gateways, and then uploads the data to the data acquisition module. The data acquisition module collects data from the edge gateway devices to be identified in real time based on the received data and uploads the data to the gateway identification module. The gateway identification module uses the received data and the obtained automatic identification model to automatically identify the power system edge gateways.

[0044] The present invention also provides a deployment method including the automatic identification method of the power system edge gateway, which further includes the following steps:

[0045] S7. Based on the automatic identification result of the power system edge gateway obtained in step S6, configure the power system edge gateway, connect the power system edge gateway to the power system, and complete the deployment of the power system edge gateway.

[0046] Step S7, which involves configuring the power system edge gateway based on the automatic identification result obtained in step S6, connecting the power system edge gateway to the power system, and completing the deployment of the power system edge gateway, specifically includes the following steps:

[0047] The automatic identification results of the power system edge gateway obtained in step S6 are uploaded to the IoT management platform of the power system.

[0048] The IoT management platform of the power system issues configuration parameters based on the identification results;

[0049] Based on the received configuration parameters, confirm the corresponding configuration operation, generate a configuration script, and send it to the power system edge gateway;

[0050] The power system edge gateway executes the received configuration script, connects to the power system's IoT management platform, and completes the deployment of the power system edge gateway.

[0051] The automatic identification method, system, and deployment method for power system edge gateways provided by this invention collects data information from existing power system edge gateways, uses an automatic identification model for power system edge gateways built based on a self-attention mechanism to automatically identify power system edge gateways, and automatically deploys them based on the identification results. Therefore, this invention can not only automatically complete the automatic identification and deployment of power system edge gateways, but also has high reliability, good accuracy, and higher efficiency. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the automatic identification method of the present invention.

[0053] Figure 2 This is a schematic diagram of the functional modules of the system of the present invention.

[0054] Figure 3 This is a schematic diagram of the deployment method of the present invention. Detailed Implementation

[0055] like Figure 1 The diagram shown is a flowchart of the automatic identification method of the present invention: The automatic identification method for power system edge gateways disclosed in this invention includes the following steps:

[0056] S1. Obtain data information from existing edge gateway devices; specifically including the following steps:

[0057] Acquire data information from existing edge gateway devices; the data information includes process data information, network data information, and system data information of the edge gateway devices.

[0058] S2. Preprocess the data obtained in step S1 to construct a training dataset; specifically including the following steps:

[0059] Statistical features are extracted from the data information obtained in step S1; the statistical features include average CPU utilization, average memory utilization, number of network connections, and system version.

[0060] The extracted statistical features and the data information obtained in step S1 are standardized.

[0061] The standardized statistical features and the data information obtained in step S1 after standardization are combined to construct a training dataset.

[0062] By processing the multi-source data obtained in step S1 and constructing the dataset, the statistical characteristics of different types of data (process, network, system information) can be comprehensively analyzed, thereby more comprehensively reflecting the operating status and performance of the equipment and enhancing the identification ability of subsequent models.

[0063] S3. Based on the self-attention mechanism, construct a primary model for automatic identification of power system edge gateways; including the following steps:

[0064] The constructed power system edge gateway automatic identification primary model includes an input layer, a self-attention mechanism layer, a first residual connection layer, a first Layer Norm layer, a progressive loop cross-attention layer, a second residual connection layer, a second Layer Norm layer, a feature fusion layer, a fully connected layer, a softmax layer, and an output layer, which are connected in sequence.

[0065] The input layer is used to receive the input feature data;

[0066] The self-attention mechanism layer is used to map the input feature data to a unified dimension, extract the corresponding query matrix, key matrix and value matrix, and process them with the self-attention mechanism to obtain self-attention features;

[0067] The first residual connection layer and the first Layer Norm layer are used to perform residual connections on the input of the self-attention mechanism layer and the obtained self-attention features, and then perform layer normalization to obtain intermediate features. The first residual connection layer and the first Layer Norm layer are used to alleviate the problem that gradients may vanish or explode during backpropagation, which may make the model difficult to train. At the same time, they can also accelerate model training and improve model stability and generalization ability.

[0068] The progressive ring cross attention layer is used to perform cross attention processing on the intermediate features output by the first Layer Norm layer in order to extract cross-combined features. The progressive ring cross attention layer introduces progressive and ring cross combination into the attention mechanism to deeply explore the relationship between features, enhance the interaction between features, and improve the expressive power and recognition accuracy of the model.

[0069] The second residual connection layer and the second Layer Norm layer are used to perform residual connections on the input of the progressive loop cross attention layer and the obtained cross-combination features, and then perform layer normalization to obtain cross attention features. The second residual connection layer and the second Layer Norm layer are used to alleviate the problem that gradients may vanish or explode during backpropagation, which may make the model difficult to train. At the same time, they can also accelerate model training and improve model stability and generalization ability.

[0070] The feature fusion layer is used to concatenate and linearize the cross-attention features output by the second Layer Norm layer to obtain fused features;

[0071] The fully connected layer is used to expand and linearly transform the obtained fused features to obtain an unnormalized score vector for each possible category;

[0072] The softmax layer is used to smooth the obtained score vector to obtain the probability of belonging to each category. The softmax layer can suppress small scores and amplify large scores, thus making the output probability distribution sharper, which is conducive to clearly identifying the most likely category.

[0073] The output layer is used to output the final power system edge gateway identification result;

[0074] In practice, the following steps are taken:

[0075] Input layer: The process characteristics X of the input layer receiving input. p Network characteristics X n and system characteristics X s ;

[0076] Self-attention mechanism layer: The self-attention layer is used to apply a weight matrix W and a bias vector b to the input feature vector (X). p ,X n ,X s Mapping to a unified dimension yields the corresponding query matrix, key matrix, and value matrix, represented as (Q p ,K p V p ), (Q n ,K n V n ) and (Q s ,K s V s ), where Q p For the process feature query matrix, K p V is the process feature key matrix. p Let Q be the process eigenvalue matrix. n For the network feature query matrix, K n V is the network feature key matrix. n Let Q be the network eigenvalue matrix. s To query the feature query matrix, K s To query the feature key matrix, V s To query the feature matrix, a self-attention mechanism is then applied to the query matrix, key matrix, and value matrix to obtain the process feature self-attention feature Z. p Network features, self-attention features Z n and system features self-attention features Zs ;

[0077] First residual connection layer and first Layer Norm layer: for X p and Z p X n and Z n X s and Z s The process employs a first residual connection layer for residual connection, followed by a first Layer Normalization layer for layer normalization, to obtain the first intermediate feature X' of the process. p The first intermediate feature X' of the network n and the first intermediate feature X of the system s ';

[0078] Progressive ring cross-attention layer: X' p 、X' n and X s Perform progressive cyclic cross combinations to obtain the first combination (X') p ,X' n ), second combination (X' n ,X s ') and the third combination (X) s ',X' p ); targeting (X' p ,X' n ), (X' n ,X s ') and (X s ',X' p Each feature is processed using a self-attention mechanism to obtain the cross-combination feature Z. pn Z ns and Z sp ;

[0079] Second residual connection layer and second Layer Norm layer: used for processing intermediate feature X' p 、X' n and X s 'and the obtained cross-combination feature Z pn Z ns and Z sp A second residual connection layer is used for residual connection, followed by a second Layer Norm layer for layer normalization, to obtain the cross-attention feature X'. p ' n 、X' n ' s and X s " p ;

[0080] Feature fusion layer: The obtained cross-attention features X' p' n 、X' n ' s and X s " p The features are then concatenated and linearized using a linear layer to obtain the fused feature C.

[0081] Fully connected layer: The obtained fused features C are expanded and linearly transformed to obtain the unnormalized score vector S for each possible category;

[0082] The softmax layer is used to smooth the obtained score vector S to obtain the probability P belonging to each category.

[0083] Output layer: The output layer is used to output the final power system edge gateway identification result;

[0084] S4. Using the training dataset constructed in step S2, train the primary automatic identification model of the power system edge gateway constructed in step S3 to obtain the automatic identification model of the power system edge gateway.

[0085] S5. Real-time acquisition of data information from the edge gateway device to be identified;

[0086] S6. Based on the data information obtained in step S5, the automatic identification model of the power system edge gateway obtained in step S4 is used to complete the automatic identification of the power system edge gateway.

[0087] like Figure 2The diagram shows the functional modules of the system of the present invention: The system for implementing the automatic identification method of the power system edge gateway disclosed in this invention includes a data acquisition module, a data processing module, a model building module, a model training module, a data acquisition module, and a gateway identification module; the data acquisition module, data processing module, model building module, model training module, data acquisition module, and gateway identification module are connected in series; the data acquisition module is used to acquire data information from existing edge gateway devices and upload the data information to the data processing module; the data processing module is used to preprocess the acquired data information according to the received data information, construct a training dataset, and upload the data information to the model building module; the model building module is used to preprocess the acquired data information according to the received data information. Based on a self-attention mechanism, a primary model for automatic identification of power system edge gateways is constructed, and the data information is uploaded to the model training module. The model training module trains the primary model for automatic identification of power system edge gateways using the received data and the constructed training dataset, obtaining an automatic identification model for power system edge gateways, and uploads the data information to the data acquisition module. The data acquisition module collects data information of the edge gateway devices to be identified in real time based on the received data information and uploads the data information to the gateway identification module. The gateway identification module uses the obtained automatic identification model for power system edge gateways based on the received data information to complete the automatic identification of power system edge gateways.

[0088] like Figure 3 The diagram shown is a flowchart of the deployment method of the present invention: The deployment method disclosed in this invention, which includes the automatic identification method of the power system edge gateway, specifically includes the following steps:

[0089] S1. Obtain data information from existing edge gateway devices;

[0090] S2. Preprocess the data obtained in step S1 to construct a training dataset;

[0091] S3. Based on the self-attention mechanism, construct a primary model for automatic identification of power system edge gateways;

[0092] S4. Using the training dataset constructed in step S2, train the primary automatic identification model of the power system edge gateway constructed in step S3 to obtain the automatic identification model of the power system edge gateway.

[0093] S5. Real-time acquisition of data information from the edge gateway device to be identified;

[0094] S6. Based on the data information obtained in step S5, the automatic identification model of the power system edge gateway obtained in step S4 is used to complete the automatic identification of the power system edge gateway;

[0095] S7. Based on the automatic identification result of the power system edge gateway obtained in step S6, configure the power system edge gateway, connect the power system edge gateway to the power system, and complete the deployment of the power system edge gateway; specifically including the following steps:

[0096] The automatic identification results of the power system edge gateway obtained in step S6 are uploaded to the IoT management platform of the power system.

[0097] The IoT management platform of the power system issues configuration parameters based on the identification results;

[0098] Based on the received configuration parameters, confirm the corresponding configuration operation, generate a configuration script, and send it to the power system edge gateway;

[0099] The power system edge gateway executes the received configuration script, connects to the power system's IoT management platform, and completes the deployment of the power system edge gateway.

[0100] The deployment method of this invention automatically configures the edge gateway device based on the identification results and synchronously updates the IoT management platform. By remotely executing configuration commands and API requests, it realizes the linkage configuration between the edge gateway device and the IoT management platform, enhancing the automation and intelligence level of the system. At the same time, the deployment method of this invention can respond to changes in device status in a timely manner, optimize device operation, and reduce manual intervention through intelligent management, thereby improving the overall efficiency and reliability of the system.

Claims

1. An automatic identification method for power system edge gateways, comprising the following steps: S1. Obtain data information from existing edge gateway devices; S2. Preprocess the data obtained in step S1 to construct a training dataset; S3. Based on the self-attention mechanism, construct a primary model for automatic identification of power system edge gateways; including the following steps: The constructed power system edge gateway automatic identification primary model includes an input layer, a self-attention mechanism layer, a first residual connection layer, a first Layer Norm layer, a progressive loop cross-attention layer, a second residual connection layer, a second Layer Norm layer, a feature fusion layer, a fully connected layer, a softmax layer, and an output layer, which are connected in sequence. The input layer is used to receive the input feature data; The self-attention mechanism layer is used to map the input feature data to a unified dimension, extract the corresponding query matrix, key matrix and value matrix, and process them with the self-attention mechanism to obtain self-attention features; The first residual connection layer and the first Layer Norm layer are used to perform residual connections on the input of the self-attention mechanism layer and the obtained self-attention features, and then perform layer normalization to obtain intermediate features; The progressive loop cross-attention layer is used to perform cross-attention processing on the intermediate features output by the first Layer Norm layer in order to extract cross-combined features; The second residual connection layer and the second Layer Norm layer are used to perform residual connection on the input of the progressive loop cross attention layer and the obtained cross combination features, and then perform layer normalization to obtain cross attention features; The feature fusion layer is used to concatenate and linearize the cross-attention features output by the second Layer Norm layer to obtain fused features; The fully connected layer is used to expand and linearly transform the obtained fused features to obtain an unnormalized score vector for each possible category; The softmax layer is used to smooth the obtained score vector to obtain the probability of belonging to each category; The output layer is used to output the final power system edge gateway identification result; S4. Using the training dataset constructed in step S2, train the primary automatic identification model of the power system edge gateway constructed in step S3 to obtain the automatic identification model of the power system edge gateway. S5. Real-time acquisition of data information from the edge gateway device to be identified; S6. Based on the data obtained in step S5, the automatic identification model of the power system edge gateway obtained in step S4 is used to complete the automatic identification of the power system edge gateway.

2. The automatic identification method for power system edge gateways according to claim 1, characterized in that... Step S1, which involves obtaining data information from existing edge gateway devices, specifically includes the following steps: Obtain data information from existing edge gateway devices; the data information includes process data information, network data information, and system data information of the edge gateway devices.

3. The automatic identification method for power system edge gateways according to claim 2, characterized in that... The preprocessing described in step S2 specifically includes the following steps: Statistical features are extracted from the data information obtained in step S1; the statistical features include average CPU utilization, average memory utilization, number of network connections, and system version. The extracted statistical features and the data information obtained in step S1 are standardized. The standardized statistical features and the data information obtained in step S1 are combined to construct a training dataset.

4. The automatic identification method for power system edge gateways according to claim 3, characterized in that... Step S3 specifically includes the following steps: Input layer: The process characteristics of the input layer receiving input. Network characteristics and system characteristics ; Self-attention mechanism layer: The self-attention layer is used to employ the weight matrix. and bias vector , input feature vector Mapping to a unified dimension yields the corresponding query matrix, key matrix, and value matrix, represented as follows: , and ,in For process feature query matrix, The process feature key matrix, For the process eigenvalue matrix, For network feature query matrix, The network feature key matrix, The network eigenvalue matrix, To query the feature query matrix, To query the feature key matrix, To query the feature matrix, a self-attention mechanism is then applied to the query matrix, key matrix, and value matrix to obtain the process feature self-attention features. Network features and self-attention features System features and self-attention features ; First residual connection layer and first Layer Norm layer: targeting and , and , and The process employs a first residual connection layer for residual connection, followed by a first Layer Normalization layer for layer normalization, to obtain the first intermediate feature of the process. First intermediate features of the network and the first intermediate feature of the system ; Progressive ring cross attention layer: , and By performing progressive cyclic cross combinations, the first combination is obtained. The second combination and the third combination ;against , and Each feature is processed using a self-attention mechanism to obtain cross-combination features. , and ; Second residual connection layer and second Layer Norm layer: used for intermediate features , and and the obtained cross-combination features , and A second residual connection layer is used for residual connections, followed by a second Layer Norm layer for layer normalization, to obtain the cross-attention features. , and ; Feature fusion layer: This layer combines the obtained cross-attention features. , and The data is then spliced ​​together and linearized using a linear layer to obtain the fused features. ; Fully connected layer: This layer will fuse the obtained features. The unnormalized score vector S for each possible category is obtained by performing expansion and linear transformation. The softmax layer is used to smooth the obtained score vector S to obtain the probability P belonging to each category. Output layer: The output layer is used to output the final power system edge gateway identification result.

5. A system for implementing the automatic identification method for power system edge gateways as described in any one of claims 1 to 4, characterized in that... It includes a data acquisition module, a data processing module, a model building module, a model training module, a data acquisition module, and a gateway identification module; the data acquisition module, data processing module, model building module, model training module, data acquisition module, and gateway identification module are connected in series; the data acquisition module is used to acquire data information from existing edge gateway devices and upload the data information to the data processing module; The data processing module is used to preprocess the acquired data information based on the received data information, construct the training dataset, and upload the data information to the model building module; The model building module is used to build a primary model for automatic identification of the power system edge gateway based on the received data information and the self-attention mechanism, and upload the data information to the model training module. The model training module is used to train the constructed primary model for automatic identification of power system edge gateways based on the received data information and the constructed training dataset, thereby obtaining the automatic identification model for power system edge gateways, and uploading the data information to the data acquisition module. The data acquisition module is used to collect data information of the edge gateway device to be identified in real time based on the received data information, and upload the data information to the gateway identification module; The gateway identification module is used to automatically identify power system edge gateways based on the received data and the obtained power system edge gateway automatic identification model.

6. A deployment method comprising the automatic identification method for a power system edge gateway as described in any one of claims 1 to 4, characterized in that... It also includes the following steps: S7. Based on the automatic identification result of the power system edge gateway obtained in step S6, configure the power system edge gateway, connect the power system edge gateway to the power system, and complete the deployment of the power system edge gateway.

7. The deployment method according to claim 6, characterized in that... Step S7, which involves configuring the power system edge gateway based on the automatic identification result obtained in step S6, connecting the power system edge gateway to the power system, and completing the deployment of the power system edge gateway, specifically includes the following steps: The automatic identification results of the power system edge gateway obtained in step S6 are uploaded to the IoT management platform of the power system. The IoT management platform of the power system issues configuration parameters based on the identification results; Based on the received configuration parameters, confirm the corresponding configuration operation, generate a configuration script, and send it to the power system edge gateway; The power system edge gateway executes the received configuration script, connects to the power system's IoT management platform, and completes the deployment of the power system edge gateway.

Images