A method for realizing secure sharing of multi-dimensional data

By adopting a blockchain architecture with multi-dimensional data security sharing in the smart energy Internet, the multi-dimensional data sharing and privacy protection problems are solved, data reliability and security are achieved, task execution efficiency and quality are improved, and data is avoided leakage.

CN117768086BActive Publication Date: 2024-11-05QINGDAO INST OF BIOENERGY & BIOPROCESS TECH CHINESE ACADEMY OF SCI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311775019.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-22
Publication Date
2024-11-05
Estimated Expiration
2043-12-22

AI Technical Summary

Technical Problem

The existing technology cannot effectively deal with the problems of multi-dimensional data sharing and privacy protection, especially in the smart energy Internet, data sharing methods cannot adapt to the heterogeneity of multiple types of data and complex business data chains, and fail to fully consider data privacy security.

Method used

Adopt a blockchain architecture with multi-dimensional data security sharing, including user layer, task layer, training layer and data layer. Through the authentication module, smart contract module, expert module, learning training module and private chain, data decomposition and division are realized. film and federated learning ensures secure sharing of data.

Benefits of technology

Through strict identity verification and federated learning, data reliability and security are ensured, task execution efficiency and quality are improved, data information is avoided leaked by central servers, and data authenticity and privacy protection are strengthened.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117768086B_ABST
    Figure CN117768086B_ABST
Patent Text Reader

Abstract

The present application relates to a method for realizing secure sharing of multi-dimensional data. The blockchain architecture includes a user layer, a task layer, a training layer, and a data layer. The user layer is based on the tasks issued by energy users, the task layer performs task decomposition, and the training layer learns the training tasks. The data layer is used to ensure the security and authenticity of shared data. The present application solves the problem that the existing technology cannot cope with multi-data sharing and does not pay attention to privacy protection during data sharing. The present application ensures the reliability of the data source, improves the efficiency of data sharing and the quality of task execution, and can effectively avoid data leakage.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application belongs to the field of blockchain technology, and specifically relates to a method for realizing secure sharing of multi-dimensional data. Background Art

[0002] In the energy field, pan-energy big data, with energy as the core, can be widely associated with open data systems in multiple dimensions such as economy, society, ecology, environment, climate, technology, policy, and security. It uses the most basic energy activities as data primitives and can connect multi-dimensional data. Smart Energy Internet refers to the emerging technology that uses advanced power electronics technology, information technology, and intelligent management technology to realize the interconnection of distributed energy collection devices, distributed energy storage devices, and distributed energy nodes. Smart Energy Internet can realize the collection and transmission of massive data. In the process of deeper and broader development of Smart Energy Internet, its requirements for data sharing methods are also higher.

[0003] In the existing technology, data sharing methods are mostly for a single type of data. Even if there is a heterogeneity problem, the duplication of data content and type cannot adapt to the characteristics of the smart energy Internet field, which has many energy subcategories, long business data chains, and wide data exposure. In addition, based on the important position of modern energy in social operation, the multi-dimensional data it contains should also focus on data privacy and security issues. Summary of the invention

[0004] The present invention provides a method for realizing secure sharing of multi-dimensional data, which solves the problem that the prior art cannot cope with multi-data sharing and does not pay attention to privacy protection during the sharing process.

[0005] The technical solution of the present invention is as follows:

[0006] A method for realizing secure sharing of multi-dimensional data is implemented based on a blockchain architecture for secure sharing of multi-dimensional data, and the blockchain architecture includes:

[0007] The user layer is where the energy users are located, and includes an identity authentication module, a task broadcast module, and a node screening module; the identity authentication module is used to verify the identity of the energy user to determine whether it can become a user node, the task broadcast module is used to publish tasks, and the node screening module is used to verify and select nodes that apply to join the task;

[0008] The task layer includes a smart contract module and an expert module. The smart contract module is used to initially decompose tasks, and the expert module can assist in decomposing tasks.

[0009] The training layer includes a learning and training module and a scoring module. The learning and training module is used for energy users to perform learning and training for specific tasks, and the scoring module is used to score user nodes participating in the training;

[0010] The data layer includes several private chains, each of which is connected to several smart sensors. The smart sensors collect data and upload it to the private chain;

[0011] Based on the above blockchain architecture, this method includes the following steps:

[0012] S1) The energy user presents his / her identity certificate and a description of his / her data type to the identity verification module;

[0013] S2) The identity verification module determines whether the energy user is qualified to be on-chain. If so, the energy user becomes a user node and is assigned to the corresponding shard according to the data type. If not, the energy user’s application is ignored.

[0014] S3) Update the data type list at the task layer;

[0015] S4) One of the nodes in the user layer publishes a complex task. The smart contract module in the task layer performs a preliminary decomposition of the complex task. The expert module further decomposes the preliminarily decomposed task into several simple tasks and adds tags containing the required data types to the simple tasks.

[0016] S5) The simple task is returned to the task publishing node, which broadcasts the simple task at the user layer. The communication node broadcasts the task requirements to the adjacent shards, and other shards broadcast the task requirements in turn.

[0017] S6) The nodes that meet the task requirements send task application information to the node screening module;

[0018] S7) The node screening module screens the user nodes that meet the task requirements and selects the best user nodes to perform the task;

[0019] S8) The node screening module creates a dedicated channel for executing tasks, forms a training layer, and sends a dedicated certificate to the user node that executes the task;

[0020] S9) The user node that executes the task completes the task training based on federated learning. In each round of iterative training, the publishing node scores the executing node according to its performance. Finally, the task publishing node integrates the training results, completes the task, and releases the dedicated channel.

[0021] S10) The transaction records that do not contain private information during the training process reach a consensus and are recorded on the chain, mainly including the number of training iterations and the score of the user node. The score is given by the task issuing node based on the performance of the user node during this training process. The consensus is recorded on the chain;

[0022] A single consensus mechanism is used in the user layer. There is a shard specifically responsible for consensus. All shards share the same consensus mechanism and maintain the same ledger.

[0023] Furthermore, after the user is verified by the identity authentication module, he / she is added to the corresponding shard according to the data type of the energy user. A shard includes at least a communication node and at least one user node. The task broadcast module publishes tasks based on the decomposed simple tasks; energy users with multiple data types can join multiple shards.

[0024] Furthermore, the task layer has a data type list included in the maintenance user layer. When the user layer adds or deletes a slice, the corresponding data type is updated in the data type list.

[0025] Furthermore, the smart contract module performs preliminary task decomposition based on the NLP smart contract, and the basis for the preliminary decomposition of the NLP smart contract is the data type; the expert module includes experts in the expert think tank in various fields. The experts verify and analyze the results of the preliminary task decomposition, and then mark the data type required for the task to finally decompose the task into simple tasks. The final decomposition result of the task is returned to the node that publishes the task in the user layer.

[0026] Furthermore, the task broadcast module includes task requirements, training content requirements, data set labels, computing power requirements, application deadline, and the number of required training nodes when the task is released; after the user node releases the task, the communication node forwards the task to the adjacent shard, and the communication node of the shard that meets the task requirements broadcasts the task requirements in the shard, and the user node that meets the task requirements sends an application to join the task.

[0027] Furthermore, the node screening module screens the energy users before the deadline based on the matching degree of the data set attributes, computing power, historical task scores, and data set authenticity, and selects the best set of energy users to perform the task.

[0028] Furthermore, the private chain in the data layer is independently maintained by energy users in the user layer. When the energy users perform model training, they verify the authenticity of their data sets through the data stored in the private chain.

[0029] Furthermore, after determining the user node for learning and training, the node screening module forms a training layer by creating a channel, and sends a certificate that can join the channel to the user node that performs the task. The user nodes that perform the same simple task have the same certificate.

[0030] The training layer is formed only when the user node performs learning training, and the training layer implements task requirements based on federated learning.

[0031] Furthermore, the energy user who issues the task obtains the result of the simple task by reverse fusion according to the logic of task decomposition based on the learning and training results, uploads the result of the initial task to the blockchain network, and then releases the channel.

[0032] A single consensus mechanism is used in the user layer. There is a shard specifically responsible for consensus. All shards share the same consensus mechanism and maintain the same ledger.

[0033] Due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0034] 1. This application conducts strict audits on energy users based on the identity authentication module, ensures the reliability of data from the source of the data, and introduces federated learning to realize user data sharing, ensuring the security of user data. Finally, the smart sensor data obtained through the private chain enhances the authenticity of the data.

[0035] 2. This application can decompose complex tasks according to data types, and use NLP smart contracts to perform preliminary decomposition of complex tasks during the decomposition process. Experts then decompose the preliminary decomposed tasks into simple tasks from a professional and authoritative perspective, thereby achieving systematic training of complex tasks. Since complex tasks are decomposed into simple tasks, this application can enable energy users to quickly identify training tasks that are suitable for them, which helps to improve the execution efficiency and quality of tasks, and can also make full use of expert resources.

[0036] 3. This application combines blockchain technology with federated learning training methods. The user node that publishes the task can also be the user node that executes the task. The user node that publishes the task communicates directly with the user node that executes the task, which can effectively avoid the central server forwarding and leaking data information. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0038] Figure 1 A schematic diagram of a blockchain architecture that can realize multi-dimensional data secure sharing provided by this application;

[0039] Figure 2 A schematic diagram of the on-chain process for energy users;

[0040] Figure 3 Flow chart of the task breakdown of this application;

[0041] Figure 4 Publish a flow chart for this application task;

[0042] Figure 5 A schematic diagram of the task release process for this application;

[0043] Figure 6 This is a flowchart of the blockchain architecture data sharing in this application. DETAILED DESCRIPTION

[0044] There is a close relationship between energy and various dimensions of society. By explaining the connection between energy and many dimensions, we can reveal the inherent laws of the operation of modern society.

[0045] As attached Figure 1 As shown, a method for realizing secure sharing of multi-dimensional data is implemented based on a blockchain architecture for secure sharing of multi-dimensional data, and the blockchain architecture includes:

[0046] The user layer is where the energy users are located, and includes an identity authentication module, a task broadcast module, and a node screening module. The identity authentication module is used to verify the identity of the energy user to determine whether he can become a user node, the task broadcast module is used to publish tasks, and the node screening module is used to verify and select nodes that apply to join the task. The task layer includes a smart contract module and an expert module. The smart contract module is used to preliminarily decompose tasks, and the expert module can assist in decomposing tasks. The training layer includes a learning and training module and a scoring module. The learning and training module is used for energy users to perform learning and training for specific tasks, and the scoring module is used to score user nodes participating in the training. The data layer includes several private chains, each of which is connected to several smart sensors, and the smart sensors collect data and upload it to the private chain.

[0047] As attached Figure 3 After the user is verified by the identity authentication module, the energy user is segmented according to the data type, and a segment includes at least a communication node and at least one other node. In addition, the task layer has a list of data types contained in the maintenance user layer. When the user layer adds or deletes a segment, the corresponding data type is updated in the data type list.

[0048] Smart Energy Internet can design sharding rules based on business needs and invite business-related experts. In specific implementation, considering the circulation attributes of various dimensions of society, sharding can be performed according to the regional attributes and industry attributes of the data contained in energy users.

[0049] In specific implementation, the identity of energy users can be verified based on the user company's business license, registration information, etc., and the credibility of energy users can be verified based on corporate credit reports, reports from credit rating agencies, etc.

[0050] In the specific implementation, the sharding principles are designed according to the characteristics of the data, such as the industry or province to which it belongs, and the data is added to the corresponding shard according to the user characteristics and the sharding principles. For example, the principle of sharding according to the province is: 1. Pre-sharding is formed according to the characteristics of the node data. The pre-sharding is the set of all possible shards. 2. Node joining is to add it to the specified shard based on its data characteristics. A node containing data from multiple provinces can be added to multiple shards. 3. If the node does not have a suitable position in the existing shards, a new shard is formed according to the pre-sharding. 4. When the last user node of a shard exits, the shard is also deleted.

[0051] As attached Figure 4 and attached Figure 5 As shown, the task broadcast module publishes tasks based on the data type of energy users at the user node; energy users with multiple data types can join multiple shards. Energy users in the user layer publish complex tasks to the task layer, which energy users cannot solve by themselves. The smart contract module in the task layer performs preliminary decomposition of complex tasks based on NLP smart contracts. The basis for the preliminary decomposition of tasks by NLP smart contracts is the data type. Experts verify, discuss, and analyze the preliminary decomposition results of the tasks, and then mark the data type required for the tasks to finally decompose the tasks into simple tasks. Simple tasks are returned to the user layer, and energy users conduct learning and training according to the task requirements.

[0052] In the specific implementation, the smart contract module of the task decomposition part mainly refers to the pre-trained BRET model to decompose the task; the expert module includes experts in various fields, who are from expert think tanks or have certification certificates from national authoritative institutions.

[0053] The specific steps of task decomposition are as follows:

[0054] First, the BRET model is trained using a large-scale text corpus, and then the trained BRET model is used to perform a preliminary decomposition of the task;

[0055] Experts verified the preliminary decomposition results and reached a consensus;

[0056] If the decomposition results are inappropriate, the experts will decompose and broadcast them to reach a consensus;

[0057] If the decomposition results are appropriate, experts label the simple tasks and reach a consensus;

[0058] Task decomposition is complete.

[0059] In the specific implementation, the task broadcast module includes task requirements when the task is released. The task requirements include training content requirements, data set labels, computing power requirements, application deadline, and the number of required training nodes. After the user node releases the task, the communication nodes of each shard forward the task to the adjacent shard. The communication nodes of the shard that meet the task requirements broadcast the task requirements in the shard to be received by each user node. The user node that meets the task requirements sends an application to join the task. The task application information includes its own data set attributes, computing power, and the scores of historical tasks and log analysis results.

[0060] In specific implementation, the information of joining the task application sent by the energy user node does not mean that the user node can be selected as the user node to perform the task. The node screening module will screen the user node before the deadline based on the matching degree of the data set type, data set range, computing power, historical task scores and data set authenticity, and finally select the best set of user nodes to perform the task.

[0061] The private chain in the data layer is independently maintained by the energy users in the user layer. Each private chain is equipped with a log analysis tool to monitor the normal operation of IoT devices and upload data on time to ensure the authenticity and reliability of the collected data. When conducting model training, energy users can verify the authenticity of their data sets by checking the log analysis report in the private chain.

[0062] As attached Figure 6 As shown, in the specific implementation, the node screening module forms a training layer by creating a channel after determining the energy users for learning and training. The task broadcast module sends certificates that can join the channel to the energy users used to perform tasks. Energy users who perform the same simple task have the same certificate. The training layer is only formed when the energy users perform learning and training, and the training layer implements the task requirements based on federated learning.

[0063] The energy user who issues the task obtains the result of the initial task by reverse fusion according to the logic of task decomposition based on the learning and training results, and releases the channel after uploading the result of the initial task to the blockchain network. In the specific implementation, the task broadcast module performs the reverse fusion process, which can be represented by social institutions or specific energy users.

[0064] Taking the State Grid Shandong Company joining the blockchain provided by this application as an example, the blockchain architecture and blockchain operation method are explained in detail.

[0065] The specific steps for State Grid Shandong Company to join the blockchain (the result of the example is that it has joined) are as follows:

[0066] SA1) State Grid Shandong Company applies to join the blockchain, and State Grid Shandong Company provides proof of identity to the identity authentication module, as well as a description of the type of data it has. The proof of identity can be an authoritative certificate such as a business license.

[0067] SA2) The identity verification module determines whether the applicant is qualified to be on the chain based on the provided identity documents, corporate credit report, and credit rating agency report. If qualified, the applicant is allowed to become a user node. If not, the application is ignored.

[0068] SA3) Add the data of State Grid Shandong Company to Shandong Electric Power Data Slice according to its data type. The data types in this step are: province and industry.

[0069] SA4) If a new shard is generated due to the addition of State Grid Shandong Company, the data type list is updated in the task layer, that is, the Shandong Electric Power data type is added.

[0070] SA5) After State Grid Shandong Company joins the user layer and becomes a user node, the chain operation is completed.

[0071] As attached Figure 1 and attached Figure 2 As shown in the figure, after the State Grid Shandong Company becomes a user node, the 0 node of one of the shards on the chain publishes Task A, the content of which is: "How much more is the energy consumption per unit GDP of power plants in Shandong Province than that of power plants in Shanxi Province?" The specific steps for the blockchain architecture of this application to process this task are as follows:

[0072] SB1) Node 0 of the above shard sends Task A to the task layer. Task A is a complex task. In the task layer, the NLP smart contract first performs a preliminary decomposition of the task based on the two data types of "Shandong Province Electricity Data Type" and "Shanxi Province Electricity Data Type". Task A is decomposed into two tasks, a1 and a2. a1 is "calculate the unit GDP energy consumption of power plants in Shandong Province" and a2 is "calculate the unit GDP energy consumption of power plants in Shanxi Province".

[0073] SB2) The task broadcast module broadcasts a1 and a2. In the expert module, experts in the energy field discuss a1 and a2 and form simple tasks, a1, a2, and a3. The task of a3 is "a1 calculated value minus a2 calculated value". After discussion, the experts add a label "Shandong Province Electric Power Data Type" to the data type required by a1, and add a label "Shanxi Province Electric Power Data Type" to the data type required by a2. Finally, the three simple tasks are given to node 0.

[0074] SB3) After node 0 receives the three simple tasks, the communication node broadcasts the task requirements at the user layer, including training content requirements, data set type requirements, data set range requirements, computing power requirements, training node data requirements, and application deadlines. Then the "Shandong Province Electric Power Data Type" shard and the "Shanxi Province Electric Power Data Type" shard forward a1 and a2 to their own shards respectively, and other shards, such as the "Shandong Hydropower Data Type" and "Shanxi Hydropower Data Type" shards in the user layer, update the ledger. The task broadcast is completed.

[0075] SB4) After the task broadcast is completed, nodes 1, 2, 3, 4, 5, 6, and 7 in the "Shandong Electric Power Data Type" shard that meet the requirements of task a1 send application information to join task a1. The application information includes its own data type, data set range, historical task score, and computing power. The node screening module in the user layer queries the results of the log tool in the data layer to determine the authenticity of the data set, the degree of matching of the data set type, the computing power, and the historical task score of the application node based on the application information received within the task application deadline. It selects the best from the nodes that meet the requirements and determines the final execution node set 1, 2, 3, 4, and 5 of task a1. Nodes 1, 7, 8, 9, 10, 11, 12, 13, and 14 in the "Shanxi Electric Power Data Type" shard that meet the requirements of task a2 finally determine the execution node set 1, 8, 9, 10, 11, 12, and 13 of task a2.

[0076] SB5) The node screening module creates the C1 channel and the C2 channel. The node screening module sends the C1 and C2 channel certificates to the nodes that execute the a1 task and the nodes that execute the a2 task respectively. The C1 and C2 channel certificates are used for training in C1 and C2 respectively.

[0077] SB6) The node screening module sends the initial global model and local model and uploads the deadline. The node executing the task uses the received global model and its own data set to perform model training locally, and uploads the model to node 0 within the deadline. Node 0 uses the federated average algorithm to filter and aggregate the received local model to obtain a new global model. Then check whether the global model meets the final training requirements. If not, send the updated global model to the node executing the task for the next round of iterative training; if so, the federated learning task training is completed, and the C1 and C2 channels are released.

[0078] SB7) Integrate training results: After receiving the training results of tasks a1 and a2, node 0 performs the calculation of "a1-a2" and finally completes task A.

[0079] SB8) Transaction records are uploaded to the chain: After the transaction is completed, the transaction records that do not contain private information during the training process reach a consensus and are recorded on the chain. The recorded information includes: number of iterations, score of user nodes, and the corresponding channels C1 and C2 are released.

[0080] In the above steps, the node screening module can be replaced by node 0 that publishes the task.

[0081] The above description is only the basic principle and preferred embodiments of the present invention. Improvements and substitutions made by those skilled in the art based on the present invention belong to the protection scope of the present invention.

Claims

1. A method for realizing secure sharing of multi-dimensional data, characterized in that: This method is implemented based on a blockchain architecture for secure sharing of multi-dimensional data, which includes: The user layer is where the energy users are located, and includes an identity authentication module, a task broadcast module, and a node screening module; the identity authentication module is used to verify the identity of the energy user to determine whether it can become a user node, the task broadcast module is used to publish tasks, and the node screening module is used to verify and select nodes that apply to join the task; The task layer includes a smart contract module and an expert module. The smart contract module is used to initially decompose tasks, and the expert module can assist in decomposing tasks. The training layer includes a learning and training module and a scoring module. The learning and training module is used for energy users to perform learning and training for specific tasks, and the scoring module is used to score user nodes participating in the training; The data layer includes several private chains, each of which is connected to several smart sensors. The smart sensors collect data and upload it to the private chain; Based on the above blockchain architecture, this method includes the following steps: S1) The energy user presents his / her identity certificate and a description of his / her data type to the identity verification module; S2) The identity verification module determines whether the energy user is qualified to be on-chain. If so, the energy user becomes a user node and is assigned to the corresponding shard according to the data type. If not, the energy user’s application is ignored. S3) Update the data type list at the task layer; S4) One of the nodes in the user layer publishes a complex task. The smart contract module in the task layer performs a preliminary decomposition of the complex task. The expert module further decomposes the preliminarily decomposed task into several simple tasks and adds tags containing the required data types to the simple tasks. S5) The simple task is returned to the task publishing node, which broadcasts the simple task at the user layer. The communication node broadcasts the task requirements to the adjacent shards, and other shards broadcast the task requirements in turn. S6) The nodes that meet the task requirements send task application information to the node screening module; S7) The node screening module screens the user nodes that meet the task requirements and selects the best user nodes to perform the task; S8) The node screening module creates a dedicated channel for executing tasks, forms a training layer, and sends a dedicated certificate to the user node that executes the task; S9) The user node that executes the task completes the task training based on federated learning. In each round of iterative training, the publishing node scores the executing node according to its performance. Finally, the task publishing node integrates the training results, completes the task, and releases the dedicated channel. S10) The transaction records that do not contain private information during the training process reach a consensus and are recorded on the chain, mainly including the number of training iterations and the score of the user node. The score is given by the task issuing node based on the performance of the user node during this training process. The consensus is recorded on the chain; A single consensus mechanism is used in the user layer. There is a shard specifically responsible for consensus. All shards share the same consensus mechanism and maintain the same ledger.

2. A method for realizing secure sharing of multi-dimensional data according to claim 1, characterized in that: After the user is verified by the identity authentication module, he / she is added to the corresponding shard according to the data type of the energy user. A shard includes at least a communication node and at least one user node. The task broadcast module publishes tasks based on the decomposed simple tasks; energy users with multiple data types can join multiple shards.

3. A method for realizing secure sharing of multi-dimensional data according to claim 2, characterized in that: The task layer has a data type list included in the maintenance user layer. When the user layer adds or deletes a slice, the corresponding data type is updated in the data type list.

4. A method for realizing secure sharing of multi-dimensional data according to claim 3, characterized in that: The smart contract module performs preliminary task decomposition based on the NLP smart contract, and the basis for the preliminary decomposition of the NLP smart contract is the data type; the expert module includes experts from the expert think tank in various fields. The experts verify and analyze the preliminary task decomposition results, and then mark the data type required for the task to finally decompose the task into simple tasks. The final task decomposition results are returned to the node that publishes the task in the user layer.

5. A method for realizing secure sharing of multi-dimensional data according to claim 4, characterized in that: The task broadcast module includes task requirements, training content requirements, data set labels, computing power requirements, application deadline, and the number of required training nodes when publishing a task. After the user node publishes the task, the communication node forwards the task to the adjacent shard. The communication node of the shard that meets the task requirements broadcasts the task requirements in the shard, and the user node that meets the task requirements sends an application to join the task.

6. A method for realizing secure sharing of multi-dimensional data according to claim 5, characterized in that: The node screening module screens energy users before the deadline based on their matching degree of data set attributes, computing power, historical task scores, and data set authenticity, and selects the best set of energy users to perform the task.

7. A method for realizing secure sharing of multi-dimensional data according to claim 6, characterized in that: The private chain in the data layer is independently maintained by energy users in the user layer. When performing model training, energy users verify the authenticity of their data sets through the data stored in the private chain.

8. A method for realizing secure sharing of multi-dimensional data according to claim 7, characterized in that: After determining the user node for learning and training, the node screening module creates a channel to form a training layer, and sends a certificate that can join the channel to the user node that performs the task. The user nodes that perform the same simple task have the same certificate. The training layer is formed only when the user node performs learning training, and the training layer implements task requirements based on federated learning.

9. A method for realizing secure sharing of multi-dimensional data according to claim 8, characterized in that: The energy user who issues the task obtains the result of the simple task by reverse fusion according to the logic of task decomposition based on the learning and training results, uploads the result of the initial task to the blockchain network, and then releases the channel.