Intelligent operation and maintenance method and system for power 5g network

Abstract

The application discloses a kind of electric power 5G network intelligent operation and maintenance method, including obtaining data information;Training start stage generates public-private key pair, retains private key and uploads public key;Select committee members and training members, download encrypted global model and train, committee members encrypt and save training results locally, training members upload block chain after encrypting training results;Cloud server receives result and normalizes verification, uploads result block chain;Committee members download verification results, score electric power terminal and upload block chain;Cloud server downloads result, aggregates and updates global model, and uploads zone block chain;Repeat the above steps to obtain intelligent operation and maintenance model;Intelligent operation and maintenance model is used to complete the intelligent operation and maintenance of electric power 5G network.The application also discloses a kind of system for realizing the electric power 5G network intelligent operation and maintenance method.The reliability of the application is high, safety is good, and efficiency is higher.

Description

Technical Field

[0001] This invention belongs to the field of electrical automation, specifically relating to a method and system for intelligent operation and maintenance of a power 5G network. 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 the stable and reliable operation of electricity has become one of the most important tasks of the power system.

[0003] Currently, the rapid development of 5G technology has greatly enhanced the transmission capabilities of communication networks; the power system has already begun to adopt 5G technology for communication. Due to the large number of power terminals connected to the power 5G network, the operational data, sensor monitoring data, and control command data from these terminals are rapidly increasing, gradually forming a large, diverse, and complex network big data volume. Therefore, intelligent operation and maintenance of the power 5G network has become one of the key research areas in the power system.

[0004] The realization of intelligent operation and maintenance of power 5G networks requires the support of a large amount of network data. For example, a large amount of historical data within the network can be collected and machine learning models trained for application to various types of operation and maintenance services. However, in this process, the servers used to train complex models require an extremely large amount of network data, which may lead to data security and terminal privacy issues in power 5G networks. Summary of the Invention

[0005] One of the objectives of this invention is to provide a highly reliable, secure, and efficient intelligent operation and maintenance method for power 5G networks.

[0006] The second objective of this invention is to provide a system for implementing the aforementioned intelligent operation and maintenance method for power 5G networks.

[0007] The intelligent operation and maintenance method for 5G power networks provided by this invention includes the following steps:

[0008] S1. Obtain data information from the target power 5G network;

[0009] S2. At the start of training, a public-private key pair is generated for the cloud server and each power terminal. The cloud server and each power terminal retain the private key and upload the public key to the cross-chain.

[0010] S3. According to the set election strategy, select committee members and training members in the power terminal, download the latest encrypted global model from the blockchain and train them respectively. The committee members encrypt the training results and save them locally, and the training members encrypt the training results and upload them to the blockchain.

[0011] S4. The cloud server receives the results uploaded by the training members, performs normalization verification, and uploads the verification results to the blockchain;

[0012] S5. Committee members download the verification results from the cross-chain database, score the power terminals based on the similarity calculation results, and upload the results to the blockchain;

[0013] S6. The cloud server downloads the results from the blockchain, performs aggregation and global model updates, and then uploads the encrypted results to the cross-chain zone.

[0014] S7. Repeat steps S3 to S6 until the set conditions are met, and finally obtain the intelligent operation and maintenance model;

[0015] S8. Using the intelligent operation and maintenance model obtained in step S7, complete the intelligent operation and maintenance of the power 5G network.

[0016] Step S1, which involves acquiring data information from the target power 5G network, specifically includes the following steps:

[0017] The target power 5G network includes several power terminals, cloud servers, edge servers, blockchain, and third-party public-private key pair generation institutions;

[0018] The set of power terminals K is represented as KK = {1, 2, ..., k, ..., K}, where k represents the k-th power terminal and K is the total number of power terminals; the local dataset of power terminal k is represented as DD. k This includes D training samples. k ;

[0019] During the t-th round of training, the committee members gathered. Represented as Let c be the c-th committee member, and C be the total number of committee members; training member set. Represented as Let B be the b-th training member, and B be the total number of training members; the set of training members who successfully participated in the aggregation. Represented as Let A be the a-th training member who successfully participated in the aggregation, and let A be the total number of training members who successfully participated in the aggregation. B > C; the global model used for training in round t is represented as W. t The loss of power terminal k in the t-th round of training is represented as F. k (W t ).

[0020] In the training start phase described in step S2, a public-private key pair is generated for the cloud server and each power terminal. The cloud server and each power terminal retain the private key and upload the public key to the cross-chain blockchain. Specifically, this includes the following steps:

[0021] Training begins; a third-party public-private key pair generator generates and distributes public-private key pairs for the cloud server. <pk s |sk s >,; The cloud server retains the private key sk s and PK the public key s Upload to cross-chain; a third-party public-private key pair generation organization generates and distributes public-private key pairs for each power terminal. <pk x |sk x > The power terminal retains the private key sk x and PK the public key x Upload area cross-chain.

[0022] Step S3, which involves selecting committee members and training members in the power terminal according to the set election strategy, downloading the latest encrypted global model from the blockchain and training them separately, with committee members encrypting and storing the training results locally, and training members encrypting and uploading the training results to the blockchain, specifically includes the following steps:

[0023] Select a number of power terminals as committee members, and the remaining power terminals as training members;

[0024] Both committee members and training members download the latest cryptographic global model from the blockchain. Using private key sk x After decryption, the model is then trained.

[0025] The committee members normalized the gradients obtained from training and then used public-key PK. s Encrypt and store locally;

[0026] The training members will use the gradients obtained during training. Using formulas After normalization, the normalization result will be... Using public key PK s Encryption And And local signature upload to the blockchain; among which Let L2 be the L2 norm of the gradient obtained during training.

[0027] The selection of several power terminals as committee members and the remaining power terminals as training members specifically includes the following steps:

[0028] In the first round of training, C power terminals are randomly selected from the power terminal set KK as committee members, and the remaining power terminals are used as training members.

[0029] In subsequent training rounds, the edge server retrieves the verification results stored in the blockchain and, based on the scores from the power terminals, assigns scores to the set. The elements in the array are sorted in descending order, and the C power terminals with a median value or higher are selected as committee members for the current round.

[0030] Step S4 involves the cloud server receiving the results uploaded by training members, performing normalization verification, and uploading the verification results to the blockchain. This process includes the following steps:

[0031] The cloud server performs normalization verification on the results uploaded by training members: it calculates the gradient inner product len ​​of the uploaded results. <> represents the vector inner product operation; the gradient inner product result is processed using the private key sk. s After decryption, the combination corresponding to len=1 Write to collection And upload to the cross-chain library; sign k for The local signature of the corresponding power terminal.

[0032] Step S5, which involves committee members downloading verification results from the cross-chain database, scoring the power terminal based on the similarity calculation results, and uploading the results to the blockchain, specifically includes the following steps:

[0033] Committee members download the collection from the blockchain.

[0034] Calculate the committee member k' and the set Cosine similarity of power terminal k in China [[Cs] (k,k') ]]for in The committee member k' will normalize the gradients obtained from training and then use the public key pk. s The result of encryption; For set The gradient obtained from training of the power terminal k Using formulas After normalization, the normalization result is then normalized. Using public key PK s The result of encryption;

[0035] Based on the obtained cosine similarity [[Cs (k,k') The set is calculated. Rating of China Power Terminal k for

[0036] Finally, the edge server update set The elements in are And upload it to the blockchain.

[0037] Step S6 involves the cloud server downloading results from the blockchain, aggregating and updating the global model, and then encrypting the results before uploading them to the cross-chain blockchain. Specifically, this includes the following steps:

[0038] Cloud server download collection And using the private key sk s Decrypt;

[0039] The summation result is: ReLU(Cs k ) represents the aggregate weight of power terminal k;

[0040] Aggregate the summation result (sum) to calculate the aggregate variable (g). t for

[0041] Finally, the global model is updated using the following formula:

[0042] W t+1 =W t -ηg t

[0043] In the formula W t+1 The updated global model for round t+1; W t The global model in round t before the update; η is the learning rate;

[0044] After the update is complete, the cloud server uses public key PK. x The updated global model is encrypted and uploaded to the cross-chain repository.

[0045] Step S7 involves repeating steps S3 to S6 until the set conditions are met, specifically including the following steps:

[0046] Repeat steps S3 to S6 until the set conditions are met;

[0047] The following formula is used as the set condition:

[0048]

[0049] In the formula F is the operation of finding the minimum value of the independent variable w; k (W t Let be the training loss of power terminal k in round t.

[0050] This invention also provides a system for implementing the aforementioned intelligent operation and maintenance method for a power 5G network, comprising a data acquisition module, a key generation module, a training module, a verification module, a scoring module, an update module, a loop module, and an intelligent operation and maintenance module; the data acquisition module, key generation module, training module, verification module, scoring module, and update module are connected in series; the input of the loop module is connected to the update module, the output of the loop module is connected to the training module, and the output of the loop module is also connected to the input of the intelligent operation and maintenance module; the output of the intelligent operation and maintenance module serves as the system's output; the data acquisition module acquires data information from the target power 5G network and uploads the data information to the key generation module; the key generation module, based on the received data information, generates public-private key pairs for the cloud server and each power terminal during the initial training phase, with the cloud server and each power terminal retaining the private key and uploading the public key to the blockchain, and also uploads the data information to the training module; the training module, based on the received data information and a set election strategy, selects committee members and training members among the power terminals, downloads the latest encrypted global model from the blockchain, and trains them individually. The committee members encrypt and store the training results locally. The training members encrypt the training results and upload them to the blockchain, and also upload the data information to the verification module. The verification module receives the results uploaded by the training members, performs normalization verification, uploads the verification results to the blockchain, and uploads the data information to the scoring module. The scoring module downloads the verification results from the blockchain based on the received data, scores the power terminals based on the similarity calculation results, uploads the results to the blockchain, and uploads the data information to the update module. The update module downloads the results from the blockchain based on the received data, performs aggregation and global model updates, encrypts the results, uploads them to the blockchain, and uploads the data information to the loop module. The loop module controls the training, verification, scoring, and update modules to work in a loop until the set conditions are met, ultimately obtaining the intelligent operation and maintenance model, and uploads the data information to the intelligent operation and maintenance module. The intelligent operation and maintenance module uses the intelligent operation and maintenance model to complete the intelligent operation and maintenance of the power 5G network based on the received data.

[0051] The intelligent operation and maintenance method and system for 5G power networks disclosed in this invention uses a committee and a cloud server to cross-validate uploaded data, resulting in higher accuracy in identifying malicious data. Furthermore, it uses blockchain to record intermediate data and verification results, ensuring transparency and traceability of the training process and facilitating timely intervention. Therefore, this invention has high reliability, good security, and high efficiency. Attached Figure Description

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

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

[0054] like Figure 1 The diagram shown is a flowchart of the method of the present invention: The intelligent operation and maintenance method for 5G power networks disclosed in this invention includes the following steps:

[0055] S1. Obtain data information from the target power 5G network; specifically including the following steps:

[0056] The target power 5G network includes several power terminals, cloud servers, edge servers, blockchain, and third-party public-private key pair generation institutions;

[0057] The set of power terminals K is represented as KK = {1, 2, ..., k, ..., K}, where k represents the k-th power terminal and K is the total number of power terminals; the local dataset of power terminal k is represented as DD. k This includes D training samples. k ;

[0058] During the t-th round of training, the committee members gathered. Represented as Let c be the c-th committee member, and C be the total number of committee members; training member set. Represented as Let B be the b-th training member, and B be the total number of training members; the set of training members who successfully participated in the aggregation. Represented as Let A be the a-th training member who successfully participated in the aggregation, and let A be the total number of training members who successfully participated in the aggregation. B > C; the global model used for training in round t is represented as W. t The loss of power terminal k in the t-th round of training is represented as F. k (W t );

[0059] S2. At the start of training, a public-private key pair is generated for the cloud server and each power terminal. The cloud server and each power terminal retain the private key and upload the public key to the blockchain. This includes the following steps:

[0060] Training begins; a third-party public-private key pair generator generates and distributes public-private key pairs for the cloud server. <pk s |sk s >,; The cloud server retains the private key sk s and PK the public key sUpload to cross-chain; a third-party public-private key pair generation organization generates and distributes public-private key pairs for each power terminal. <pk x |sk x > The power terminal retains the private key sk x and PK the public key x Upload area cross-chain;

[0061] S3. According to the set election strategy, committee members and training members are selected in the power terminal. The latest encrypted global model is downloaded from the blockchain and trained by each member. Committee members encrypt and save the training results locally, while training members encrypt and upload their training results to the blockchain. Specifically, the steps include:

[0062] Select a number of power terminals as committee members, and the remaining power terminals as training members;

[0063] Both committee members and training members download the latest cryptographic global model from the blockchain. Using private key sk x After decryption, the model is then trained.

[0064] The committee members normalized the gradients obtained from training and then used public-key PK. s Encrypt and store locally;

[0065] The training members will use the gradients obtained during training. Using formulas After normalization, the normalization result will be... Using public key PK s Encryption And And local signature upload to the blockchain; among which The L2 norm of the gradient obtained during training;

[0066] In practice, the selection of several power terminals as committee members and the remaining power terminals as training members includes the following steps:

[0067] In the first round of training, C power terminals are randomly selected from the power terminal set KK as committee members, and the remaining power terminals are used as training members.

[0068] In subsequent training rounds, the edge server retrieves the verification results stored in the blockchain and, based on the scores from the power terminals, assigns scores to the set. The elements in the array are sorted in descending order, and the C power terminals with a median value or higher are selected as the committee members for the current round.

[0069] S4. The cloud server receives the results uploaded by training members, performs normalization verification, and uploads the verification results to the blockchain; specifically, it includes the following steps:

[0070] The cloud server performs normalization verification on the results uploaded by training members: it calculates the gradient inner product len ​​of the uploaded results. <> represents the vector inner product operation; the gradient inner product result is processed using the private key sk. s After decryption, the combination corresponding to len=1 Write to collection And upload to the cross-chain library; sign k for The local signature of the corresponding power terminal;

[0071] S5. Committee members download the verification results from the cross-chain database, score the power terminals based on the similarity calculation results, and upload the results to the blockchain; specifically, this includes the following steps:

[0072] Committee members download the collection from the blockchain.

[0073] Calculate the committee member k' and the set Cosine similarity of power terminal k in China [[Cs] (k,k') ]]for in The committee member k' will normalize the gradients obtained from training and then use the public key pk. s The result of encryption; For set The gradient obtained from training of the power terminal k Using formulas After normalization, the normalization result is then normalized. Using public key PK s The result of encryption;

[0074] Based on the obtained cosine similarity [[Cs (k,k') The set is calculated. Rating of China Power Terminal k for

[0075] Finally, the edge server update set The elements in are And upload it to the blockchain;

[0076] S6. The cloud server downloads the results from the blockchain, performs aggregation and global model updates, and then uploads the encrypted results to the cross-chain blockchain; specifically, it includes the following steps:

[0077] Cloud server download collection And using the private key sks Decrypt;

[0078] The summation result is: ReLU(Cs k ) represents the aggregate weight of power terminal k;

[0079] Aggregate the summation result (sum) to calculate the aggregate variable (g). t for

[0080] Finally, the global model is updated using the following formula:

[0081] W t+1 =W t -ηg t

[0082] In the formula W t+1 The updated global model for round t+1; W t The global model in round t before the update; η is the learning rate;

[0083] After the update is complete, the cloud server uses public key PK. x The updated global model is encrypted and uploaded to the cross-chain repository;

[0084] S7. Repeat steps S3 to S6 until the set conditions are met, and finally obtain the intelligent operation and maintenance model; specifically including the following steps:

[0085] Repeat steps S3 to S6 until the set conditions are met;

[0086] The following formula is used as the set condition:

[0087]

[0088] In the formula F is the operation of finding the minimum value of the independent variable w; k (W t Let ) be the training loss of power terminal k in round t;

[0089] S8. Using the intelligent operation and maintenance model obtained in step S7, complete the intelligent operation and maintenance of the power 5G network.

[0090] This invention uses blockchain to record the public keys of all power terminals and cloud servers. Model uploads and downloads rely on the blockchain, and training members are required to upload local signatures to ensure transparency and traceability of the training process, preventing malicious power terminals and servers from tampering with intermediate results. This invention uses an edge server as a full node to store the entire blockchain content, while the power terminals store the block headers. The edge server is also responsible for the election and scheduling of committee members. To protect the original data of the power terminals and prevent back-pull attacks, this invention uses homomorphic encryption algorithms for both local and global models.

[0091] like Figure 2The diagram shows the functional modules of the system of the present invention: The system for implementing the intelligent operation and maintenance method of the power 5G network disclosed in this invention includes a data acquisition module, a key generation module, a training module, a verification module, a scoring module, an update module, a loop module, and an intelligent operation and maintenance module; the data acquisition module, key generation module, training module, verification module, scoring module, and update module are connected in series; the input end of the loop module is connected to the update module, the output end of the loop module is connected to the training module, and the output end of the loop module is also connected to the input end of the intelligent operation and maintenance module; the output end of the intelligent operation and maintenance module serves as the output end of the system; the data acquisition module is used to acquire data information of the target power 5G network and upload the data information to the key generation module; the key generation module is used to generate public and private key pairs for the cloud server and each power terminal at the beginning of the training phase based on the received data information, the cloud server and each power terminal retain the private key, and upload the public key to the blockchain, and upload the data information to the training module; the training module is used to select committee members and training members in the power terminals according to the received data information and the set election strategy, and download the latest encrypted global key from the blockchain. Each model is trained independently. Committee members encrypt and store the training results locally, while training members encrypt their results and upload them to the blockchain, along with the data information to the verification module. The verification module receives the results uploaded by the training members, performs normalization verification, uploads the verification results to the blockchain, and uploads the data information to the scoring module. The scoring module downloads the verification results from the blockchain, scores the power terminals based on similarity calculations, uploads the results to the blockchain, and uploads the data information to the update module. The update module downloads the results from the blockchain, aggregates and updates the global model, encrypts the results, uploads them to the blockchain, and uploads the data information to the loop module. The loop module controls the training, verification, scoring, and update modules to work in a loop until the set conditions are met, ultimately obtaining the intelligent operation and maintenance model, and uploads the data information to the intelligent operation and maintenance module. The intelligent operation and maintenance module uses the received data and the intelligent operation and maintenance model to complete the intelligent operation and maintenance of the power 5G network.

Claims

1. A method for intelligent operation and maintenance of a power 5G network, comprising the following steps: S1. Obtain data information from the target power 5G network; specifically including the following steps: The target power 5G network includes several power terminals, cloud servers, edge servers, blockchain, and third-party public-private key pair generation institutions; Power terminal collection Represented as , Let k represent the k-th power terminal, where K is the total number of power terminals; the local dataset for power terminal k is represented as follows. The number of training samples included is ; During the t-th round of training, the committee members gathered. Represented as , Let c be the c-th committee member, and C be the total number of committee members; training member set. Represented as , Let b be the b-th training member, and B be the total number of training members; The set of training members who successfully participated in the aggregation Represented as , Let A be the a-th training member who successfully participated in the aggregation, and let A be the total number of training members who successfully participated in the aggregation. , , , The global model used for training in round t is represented as: The loss of power terminal k in the t-th round of training is expressed as ; S2. At the start of training, a public-private key pair is generated for the cloud server and each power terminal. The cloud server and each power terminal retain the private key and upload the public key to the blockchain. S3. According to the set election strategy, select committee members and training members in the power terminal, download the latest encrypted global model from the blockchain and train them respectively. The committee members encrypt the training results and save them locally, and the training members encrypt the training results and upload them to the blockchain. Specifically, the steps include the following: Select a number of power terminals as committee members, and the remaining power terminals as training members; Both committee members and training members download the latest cryptographic global model from the blockchain. Using private key After decryption, the model is then trained. The committee members normalized the gradients obtained from the training and then used public keys. Encrypt and store locally; The training members will use the gradients obtained during training. Using formulas After normalization, the normalization result will be... Using public key Encryption and will Upload local signatures to the blockchain; in The L2 norm of the gradient obtained during training; S4. The cloud server receives the results uploaded by training members, performs normalization verification, and uploads the verification results to the blockchain; specifically, it includes the following steps: The cloud server performs normalization validation on the results uploaded by training members: it calculates the gradient inner product of the uploaded results. for , This is a vector inner product operation; the gradient inner product result is processed using a private key. After decryption, The corresponding combination Write to collection And upload it to the blockchain; for The local signature of the corresponding power terminal; S5. Committee members download the verification results from the blockchain, score the power terminals based on the similarity calculation results, and upload the results to the blockchain; specifically, this includes the following steps: Committee members download the collection from the blockchain. ; Members of the Computing Committee With sets Cosine similarity of power terminal k for ,in As a member of the committee The gradients obtained during training are normalized before using the public key. The result of encryption; For set The gradient obtained from training of the power terminal k Using formulas After normalization, the normalization result is then normalized. Using public key The result of encryption; Based on the obtained cosine similarity The set is calculated Rating of China Power Terminal k for ; Finally, the edge server update set The elements in are And upload it to the blockchain; S6. Cloud servers download collections from the blockchain. It performs aggregation and global model updates, and uploads the encrypted results to the blockchain; S7. Repeat steps S3 to S6 until the set conditions are met, and finally obtain the intelligent operation and maintenance model; S8. Using the intelligent operation and maintenance model obtained in step S7, complete the intelligent operation and maintenance of the power 5G network.

2. The intelligent operation and maintenance method for power 5G networks according to claim 1, characterized in that... In the training start phase described in step S2, a public-private key pair is generated for the cloud server and each power terminal. The cloud server and each power terminal retain the private key and upload the public key to the blockchain. Specifically, this includes the following steps: Training begins; a third-party public-private key pair generator generates and distributes public-private key pairs for the cloud server. The cloud server retains the private key. and the public key Upload to the blockchain; a third-party public-private key pair generator generates and distributes public-private key pairs for each power terminal. The power terminal retains the private key. and the public key Upload to the blockchain.

3. The intelligent operation and maintenance method for power 5G networks according to claim 2, characterized in that... The selection of several power terminals as committee members and the remaining power terminals as training members specifically includes the following steps: In the first round of training, from the power terminal set In this process, C power terminals are randomly selected as committee members, and the remaining power terminals are used as training members. In subsequent training rounds, the edge server retrieves the verification results stored in the blockchain and, based on the scores from the power terminals, assigns scores to the set. The elements in the array are sorted in descending order, and the C power terminals with a median value or higher are selected as committee members for the current round.

4. The intelligent operation and maintenance method for power 5G networks according to claim 3, characterized in that... Step S6 involves the cloud server downloading results from the blockchain, aggregating and updating the global model, and then encrypting the results before uploading them to the blockchain. Specifically, this includes the following steps: Cloud server download collection and using a private key Decrypt; Calculate the summation result for , Let k be the aggregate weight of the power terminal. Based on the summation result Perform aggregation to calculate the aggregate variable. for ; Finally, the global model is updated using the following formula: In the formula This is the updated global model for round t+1; This is the global model for round t before the update; The learning rate; After the update is complete, the cloud server will use a public key. The updated global model is encrypted and uploaded to the blockchain.

5. The intelligent operation and maintenance method for power 5G networks according to claim 4, characterized in that... Step S7 involves repeating steps S3 to S6 until the set conditions are met, specifically including the following steps: Repeat steps S3 to S6 until the set conditions are met; The following formula is used as the set condition: In the formula This is an operation to find the minimum value of the independent variable w; Let be the training loss of power terminal k in round t.

6. A system for implementing the intelligent operation and maintenance method for a power 5G network as described in any one of claims 1 to 5, characterized in that... The system includes a data acquisition module, a key generation module, a training module, a verification module, a scoring module, an update module, a loop module, and an intelligent operation and maintenance module. These modules are connected in series. The input of the loop module is connected to the update module, and its output is connected to the training module. The output of the loop module is also connected to the input of the intelligent operation and maintenance module. The output of the intelligent operation and maintenance module serves as the system's output. The data acquisition module acquires data information from the target power 5G network and uploads this data to the key generation module. The key generation module is used to generate public and private key pairs for the cloud server and each power terminal at the beginning of the training phase based on the received data information. The cloud server and each power terminal retain the private key and upload the public key to the blockchain, and upload the data information to the training module. The training module is used to select committee members and training members in the power terminal according to the received data information and the set election strategy. The latest encrypted global model is downloaded from the blockchain and trained by each member. The committee members encrypt the training results and save them locally, while the training members encrypt the training results and upload them to the blockchain, and upload the data information to the verification module. The verification module is used to receive the results uploaded by the training members from the cloud server based on the received data information, perform normalization verification, upload the verification results to the blockchain, and upload the data information to the scoring module. The scoring module is used to download verification results from the blockchain based on the received data information, score the power terminal based on the similarity calculation results, upload the results to the blockchain, and upload the data information to the update module; The update module is used to download the results from the blockchain based on the received data information, perform aggregation and global model updates, encrypt the results and upload them to the blockchain, and upload the data information to the loop module; The loop module controls the training module, verification module, scoring module, and update module to work in a loop based on the received data information until the set conditions are met, and finally obtains the intelligent operation and maintenance model, and uploads the data information to the intelligent operation and maintenance module; the intelligent operation and maintenance module is used to complete the intelligent operation and maintenance of the power 5G network based on the received data information and the intelligent operation and maintenance model.

Images