Blockchain-based power information security storage supervision system

Abstract

The application discloses a kind of based on blockchain's electric power information security storage supervision system, the alarm information processing module of the present application introduces multi-head neural attention mechanism and fuzzy entropy technology.By pre-processing the alarm information of real-time collected electric power equipment, the system can extract time series features, environmental parameters and historical record information, and use fuzzy entropy calculation to quantify the complexity and uncertainty of each alarm information.Combined with multi-head neural attention mechanism, the system can intelligently identify alarm information that may cause major problems and prioritize allocation of processing resources.Compared with traditional alarm processing systems, the present application improves alarm information processing efficiency while effectively reducing false positives and false negatives, ensuring that critical alarm information is responded to in a timely manner, thereby significantly improving the safety and stability of the electric system.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of power information security storage supervision systems, in particular to a power information security storage supervision system based on a block chain. BACKGROUND

[0002] With the rapid development of the power system and the continuous deepening of intelligent application, real-time collection, transmission and storage of power data become increasingly important. Traditional power information systems usually rely on centralized storage and management methods, but this method has limitations when dealing with a large amount of real-time data. Although existing block chain technology is very mature, due to the large amount of power data, many systems have shifted to a distributed storage environment. However, in a distributed storage environment, data is distributed across multiple nodes, and how to verify the validity of the data while ensuring data security becomes a technical bottleneck for traditional methods.

[0003] Existing verification methods usually rely on public verification or exposure of partial data, which increases the risk of data leakage to some extent and is difficult to meet the demand for high security. In addition, the alarm information in the power system is not only large in quantity, but also involves complex environmental parameters, device states and historical records and many other factors. Traditional alarm processing systems lack intelligent analysis capabilities and are difficult to effectively identify potential major risks, often leading to false positives, false negatives, and even ignoring some alarm information that may trigger major accidents. In the face of multi-source alarm information, the system is difficult to achieve efficient alarm sorting and priority processing, resulting in a delay in responding to critical alarm information and endangering the stable operation of the power system.

[0004] Therefore, the present application is proposed. SUMMARY

[0005] The technical problem to be solved by the present application is to overcome the shortcomings of the prior art and provide a power information security storage supervision system based on a block chain, which solves the problems raised in the background.

[0006] To solve the above technical problems, the basic idea of the technical solution of the present application is:

[0007] A power information security storage supervision system based on a block chain, comprising: a data acquisition module, a delay tolerant network (DTN) module, a data slicing and encryption module, a block chain network module, a zero-knowledge proof (ZKP) module, an alarm information processing module, and a storage node management unit.

[0008] The data acquisition module is used to collect power data from power equipment in real time and transmit the data to the delay tolerant network module.

[0009] The delay-tolerant network module is configured to prioritize the power data according to the time sensitivity of the power data and transmit the prioritized data to the data slicing and encryption module;

[0010] The data slicing and encryption module is configured to slice the prioritized power data into a plurality of data blocks and encrypt each data block to generate encrypted data blocks;

[0011] The blockchain network module is configured to receive the encrypted data blocks and distribute the encrypted data blocks to storage nodes of a plurality of substations for distributed storage through blockchain technology, wherein the blockchain network is based on Hyperledger Fabric to realize distributed storage, management and record the storage path of the data blocks;

[0012] The zero-knowledge proof module is configured to verify the validity of the encrypted data blocks based on a zero-knowledge proof protocol in the blockchain network to ensure the confidentiality and integrity of the data in the distributed storage;

[0013] The alarm information processing module is configured to collect and store alarm information related to the power data, analyze historical alarm data, environmental parameters, device states and other information, predict possible future risks using a multi-head neural attention mechanism and fuzzy entropy technology, and finally dynamically adjust the weight of the alarm information to identify those alarms that may cause major problems in the future and prioritize the allocation of processing resources to these information;

[0014] The storage node management unit is configured to manage and maintain the storage nodes distributed in a plurality of substations to ensure the security and traceability of the distributed storage of the encrypted data blocks.

[0015] Optionally, when prioritizing the power data according to the time sensitivity of the power data, the time sensitivity of the collected power data is analyzed to obtain the sensitivity level of the power data, and the power data is classified according to the preset priority rules based on the time sensitivity analysis, and the classified data is transmitted to the data slicing and encryption module according to the priority of the data.

[0016] Optionally, the sensitivity level includes high sensitivity, medium sensitivity and low sensitivity, wherein the high sensitivity data is assigned the highest priority, and the medium sensitivity and low sensitivity data correspond to medium priority and low priority, respectively.

[0017] Optionally, the data slicing needs to preset a slicing strategy before slicing, and after obtaining the power information data, the received power data is sliced into a plurality of independent data blocks according to the preset slicing strategy, and the details and execution results of the slicing strategy are recorded on the blockchain after each execution of the slicing strategy;

[0018] After obtaining the independent data blocks, each data block is encrypted using an asymmetric encryption algorithm, all data blocks are converted into encrypted data blocks, and each encrypted data block is attached with a unique encrypted identifier. Finally, the generated encrypted data blocks are transmitted to each storage node in the blockchain network for distributed storage.

[0019] Optionally, before uploading data to the blockchain network, the storage node of each transformer substation needs to be configured with the node of Hyperledger Fabric. When the blockchain network module receives the encrypted data block containing the meta information, the system uses a node selection algorithm to select a suitable node cluster from the available transformer substation storage nodes based on the current node's load, network bandwidth, storage capacity, and geographic location.

[0020] The blockchain network generates data block distribution instructions to distribute the encrypted data blocks to the selected storage nodes. After distribution is complete, the blockchain network records the storage path, storage node identifier, and related metadata of each encrypted data block in the distributed ledger of the blockchain.

[0021] Optionally, during distribution, the blockchain network dynamically selects transformer substation storage nodes suitable for storing these data blocks based on the metadata of the encrypted data blocks and the current state of the blockchain network. High-priority data blocks are allocated to nodes with better performance, and the system regularly verifies the storage path of the stored data blocks through the blockchain network to check their storage status in each node. If an anomaly is detected, the blockchain network will automatically reassign the storage node.

[0022] Optionally, after the encrypted data block is generated, the owning node of the data block calculates a hierarchical zero-knowledge proof using a generator. Based on the sensitivity and purpose of the data block, different levels of proof complexity are selected. For high-sensitivity data blocks, a multi-level proof mechanism is used, with each level providing additional security verification. A quantum-resistant encryption algorithm is used to generate a tag commitment σ i , a verification factor τ i , whose expression is: where ε lattice is an encryption function based on lattice theory, used to generate a quantum-secure tag commitment, M i is the i-th level data message or its digest, r i is a random number used to ensure the randomness and security of the encryption process, σ i is the i-th level generated tag commitment, which is the hash value of the encrypted message M i and the random number r i , V lattice is a verification function based on lattice theory, used to generate a quantum-secure verification factor, K iis the shared key or verification parameter for the i-th level, i is the verification factor generated for the i-th level, which is the hash value of the label commitment and the verification parameter, is a hash function used to generate a secure hash value;

[0023] After obtaining the hierarchical zero-knowledge proof, it is distributed to each storage node in the blockchain network together with the corresponding encrypted data block. After receiving the encrypted data block and its hierarchical zero-knowledge proof, the storage node starts the verification process. The verifier node independently verifies each level of proof using the shared verification parameter according to the hierarchical zero-knowledge proof protocol, and its expression is: where V layered is the verification result of the hierarchical zero-knowledge proof. If all levels are verified, the value is true (True), is the verification function for the i-th level, responsible for verifying whether the label commitment and the verification factor of this level are correct, and i , τ i is the label commitment and verification factor of the i-th level, and K i is the shared key or verification parameter for the i-th level;

[0024] If all levels are verified, the data block is formally stored in the distributed storage system, and the verification result and related information are recorded in the distributed ledger of the blockchain. If the verification fails, the storage node will refuse to store the data block and will not add it to the distributed storage of the system, and will immediately issue an alarm and send the alarm information to the alarm information processing module.

[0025] Optionally, the alarm information processing module performs the following steps when prioritizing:

[0026] The alarm information processing module collects the alarm information of the power equipment in real time, and pre-processes the data to extract time series features, environmental parameters and historical alarm record information. The fuzzy entropy of the time series data of the alarm information is calculated to obtain the complexity and uncertainty index of each alarm information, and the expression of the fuzzy entropy calculation is: where, is a time series segment with a length of m, r is the fuzziness of the sequence, n represents the number of data points in the sequence, r represents the allowed fuzziness threshold, which is used to control the similarity between two sequence segments, and N is the total number of data points, is a fuzziness function that calculates the similarity between sequence segments X i and .

[0027] The multi-head neural attention model is established and trained using historical data, the complexity and uncertainty indexes calculated by the fuzzy entropy, and other alarm features are input into the multi-head neural attention model, different attention heads of the multi-head neural attention model pay attention to different features, and the importance and correlation of these features are learned independently according to the output results of the fuzzy entropy;

[0028] According to the results calculated by the fuzzy entropy and the multi-head neural attention mechanism, the weight of the alarm information is dynamically adjusted, and the alarm sorting is optimized, and the expression is: W alert W is the dynamic weight of the alarm information, and a i is the weight coefficient of the i-th attention head, which is calculated by the multi-head attention mechanism, and b is the weight coefficient of the fuzzy entropy, which is used to balance the influence of the fuzzy entropy in the maximum weight;

[0029] The alarm information processing module processes the alarm information according to the optimized sorting, preferentially responds to the alarm that may cause a major problem, and continuously monitors the processing effect of the alarm information, re-evaluates the stored alarm information by the fuzzy entropy technology, and if new abnormalities or risks are monitored, the alarm information processing module dynamically adjusts the alarm processing strategy and triggers the necessary emergency response.

[0030] After the above technical scheme is adopted, the present application has the following beneficial effects compared with the prior art. Of course, any product implementing the present application does not necessarily need to achieve all the advantages described below:

[0031] 1. The alarm information processing module of the present application introduces the multi-head neural attention mechanism and the fuzzy entropy technology. By preprocessing the real-time collected power equipment alarm information, the system can extract time series features, environmental parameters and historical record information, and use fuzzy entropy calculation to quantify the complexity and uncertainty of each alarm information. Combined with the multi-head neural attention mechanism, the system can intelligently identify the alarm information that may cause a major problem and preferentially allocate processing resources. Compared with the traditional alarm processing system, the present application improves the efficiency of alarm information processing while effectively reducing the false positives and false negatives, ensuring that critical alarm information can be responded in time, thereby greatly improving the safety and stability of the power system.

[0032] 2. By combining the zero-knowledge proof technology and the blockchain technology, the system can realize efficient data management and secure storage in a distributed storage environment. The storage path and verification result of each encrypted data block are recorded in the distributed ledger of the blockchain, ensuring the data's non-tamperability and traceability. In the event of an abnormality or failure of a storage node, the system can quickly redistribute the storage nodes to ensure the continuous availability and security of the data

[0033] The specific embodiments of the present application will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0034] The drawings in the following description are only some embodiments, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings. In the drawings

[0035] In the drawings:

[0036] Figure 1 The structure of the safe storage supervision system is shown.

[0037] It should be noted that these drawings and written descriptions are not intended to limit the scope of the concept of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0038] The present application will now be further described in detail with reference to the accompanying drawings.

[0039] Please refer to Figure 1 It is shown that in this embodiment, a power information security storage supervision system based on blockchain is provided, which includes a data acquisition module, a delay tolerant network (DTN) module, a data slicing and encryption module, a blockchain network module, a zero-knowledge proof (ZKP) module, an alarm information processing module, a storage node management unit;

[0040] The data acquisition module is used to acquire power data from power equipment in real time and transmit the data to the delay tolerant network module;

[0041] The delay tolerant network module is used to classify the power data according to the time sensitivity of the data, and transmit the classified data to the data slicing and encryption module;

[0042] The data slicing and encryption module is used to divide the power data classified by priority into multiple data blocks, and encrypt each data block to generate encrypted data blocks;

[0043] The blockchain network module is used to receive the encrypted data blocks and distribute them to the storage nodes of multiple substations for distributed storage through blockchain technology. The blockchain network is based on Hyperledger Fabric to realize distributed storage, management and record the storage path of the data blocks. The blockchain network module realizes distributed storage and management of data through Hyperledger Fabric technology, ensuring the data's non-tamperability and traceability. The storage path of the data is also recorded in the blockchain, which is convenient for later auditing and verification.

[0044] A zero-knowledge proof module is configured to verify the validity of the encrypted data block based on a zero-knowledge proof protocol in the blockchain network, to ensure the confidentiality and integrity of the data in distributed storage; the zero-knowledge proof module ensures the confidentiality and integrity of the data in distributed storage, and the specific content of the data does not need to be exposed in the verification process, which protects the privacy of the data while improving the security of the system. In the power trading system, zero-knowledge proof can be used to verify the legality of the transaction without disclosing the specific details of the transaction, thereby protecting the business secrets of the participants.

[0045] An alarm information processing module is configured to collect and store alarm information related to power data, analyze historical alarm data, environmental parameters, device status and other information, predict possible future risks using a multi-head neural attention mechanism and fuzzy entropy technology, and finally dynamically adjust the weight of the alarm information to identify those alarms that may cause major problems in the future and prioritize the allocation of processing resources for these information; the alarm information processing module uses a multi-head neural attention mechanism and fuzzy entropy technology to not only analyze current alarm information but also predict possible future risks and prioritize high-risk alarms. This greatly improves the system's predictive ability and emergency handling capacity. By analyzing past device failure records and current operating status, the system can predict that a device may soon fail and schedule maintenance in advance to avoid power outages.

[0046] A storage node management unit is configured to manage and maintain storage nodes distributed in multiple substations to ensure the security and traceability of distributed storage of encrypted data blocks.

[0047] In this embodiment, when classifying the power data according to the time sensitivity of the power data, the time sensitivity of the collected power data is analyzed to obtain the sensitivity level of the power data, and the power data is classified according to the preset priority rules based on the time sensitivity analysis. The classified data is transmitted to the data slicing and encryption module according to the priority of the data. By analyzing the time sensitivity of the power data and classifying it, the system can allocate resources reasonably according to the importance and timeliness of the data to ensure that critical data is processed first. This approach improves the system's response speed to high-priority data and optimizes resource utilization efficiency. The sensitivity level includes high sensitivity, medium sensitivity and low sensitivity, where high sensitivity data is assigned the highest priority, medium sensitivity and low sensitivity data correspond to medium and low priority, respectively. The introduction of sensitivity level and priority rules enables the system to flexibly adjust the processing strategy according to the characteristics of different data. For example, in the case of limited resources, the system can prioritize high-sensitivity data and delay processing of low-sensitivity data. This flexibility helps the system better adapt to changing power environments and real-time requirements.

[0048] For example, in a smart grid, assume that a substation in a certain area detects an abnormality in the equipment, and the acquisition module obtains the relevant power data in real time. The time sensitivity of these data is very high, because any delay can lead to greater power failure or even power outage. The system analyzes the time sensitivity of these data, determines that they are high-sensitivity data, and gives them the highest priority. Subsequently, these data are transmitted to the data slicing and encryption module in priority, processed quickly, and stored in the most secure storage node for subsequent analysis and fault handling. This priority processing mechanism can ensure that measures are taken in the shortest possible time to avoid widespread power failure.

[0049] In this embodiment, the data slicing needs to preset the slicing strategy before slicing. After obtaining the power information data, the received power data is sliced into multiple independent data blocks according to the preset slicing strategy, and the details of the slicing strategy and the execution results are recorded on the blockchain after each execution of the data slicing strategy. By presetting the slicing strategy before data slicing and recording the execution details on the blockchain, the transparency and traceability of the entire data processing process are enhanced.

[0050] After obtaining the independent data blocks, each sliced data block is encrypted using an asymmetric encryption algorithm, all data blocks are converted into encrypted data blocks, and each encrypted data block is attached with a unique encryption identifier. Finally, the generated encrypted data blocks are transmitted to each storage node in the blockchain network for distributed storage. After data slicing, each data block is encrypted using an asymmetric encryption algorithm, and each encrypted data block is assigned a unique encryption identifier. This mechanism ensures the security of data during transmission and storage. Even if a data block is intercepted or tampered with, the attacker cannot decrypt or tamper with the data, thereby protecting the integrity and confidentiality of the data.

[0051] Before uploading the data to the blockchain network, the storage node of each substation needs to be configured with the node of Hyperledger Fabric. When the blockchain network module receives the encrypted data block containing the meta-information, the system uses a node selection algorithm to select a suitable node cluster from the available substation storage nodes based on the current load, network bandwidth, storage capacity, and geographic location of the node. The expression is: where S * is the final selected node cluster, S i represents the i-th candidate node, S is the current set of all available substation storage nodes, L i is the current load of node S i , which is represented as CPU usage or I / O load, C i is the network bandwidth of node S ithe available storage capacity of the node S, i is a weight factor to adjust the impact of the load-to-storage capacity ratio in the overall formula, μ, v are non-linear adjustment parameters to adjust the relative importance of the load and the storage capacity i the network bandwidth of the node S i , δ is a weight factor to adjust the impact of the network bandwidth in the overall formula, η is a non-linear adjustment parameter to adjust the impact of the network bandwidthG i is a geographical weighting factor representing the geographical importance of the node S i , possibly based on certain strategic requirements, ζ is a non-linear adjustment parameter to adjust the impact of the geographical weighting factor, d i，j is the physical distance between the node S i and the data source j, κ is a decay constant of the distance impact to control the degree of the impact of the distance on the selection, λ is a weight factor to adjust the proportion of the geographical location and the distance impact in the overall formula;

[0052] The blockchain network generates data block distribution instructions, instructing the distribution of encrypted data blocks to selected storage nodes. After the distribution is completed, the blockchain network records the storage path, storage node identifier, and related metadata of each encrypted data block in the distributed ledger of the blockchain.

[0053] For example: In a smart grid system, a large amount of power data is collected from multiple substations and sensors every day. These data need to be cut, encrypted, and distributed in multiple nodes of the substation to ensure the security and reliability of the data. Through the preset cutting strategy, the system cuts the power data into multiple small blocks and uses asymmetric encryption algorithm for encryption processing. The encrypted data blocks are assigned unique identifiers and selected by the node selection algorithm in the blockchain network. The details of each cutting and encryption operation, including the storage path and node information of the data block, are recorded in the distributed ledger of the blockchain. In this way, even if an error occurs during data transmission or a node fails, system administrators can track the flow path of the data block through the blockchain ledger and take appropriate measures to recover or retransmit the data, ensuring the security of the power data and the reliability of the system.

[0054] In the distribution process, the blockchain network dynamically selects the substation storage node suitable for storing the encrypted data block based on the metadata of the data block and the current state of the blockchain network, allocates data blocks with high priority to nodes with better performance, and the system regularly verifies the path of the stored data block through the blockchain network to check its storage status in each node. If an anomaly is detected, the blockchain network will automatically redistribute the storage node. By dynamically selecting the storage node based on the metadata of the encrypted data block and the state of the blockchain network, the system can intelligently allocate high-priority data to nodes with better performance. This intelligent allocation mechanism ensures that critical data is given priority and improves the efficiency of data processing and the response speed of the system.

[0055] For example, in a smart grid system, some data (such as load scheduling instructions, fault alarms, etc.) are crucial to the stable operation of the power grid. These data usually have high priority and require fast storage and retrieval. Based on this system design, when these high-priority data blocks are generated, the blockchain network will allocate them to storage nodes with higher performance based on the priority of the data and the current network load. These nodes may have faster processing speed, larger storage capacity, and lower network delay, ensuring timely processing and storage of critical data. After data storage, the system regularly verifies the path through the blockchain network to check the storage status of these data blocks in each node. For example, if a node's storage state is unstable, it may cause data to be unavailable or lost. At this time, the system detects the anomaly and automatically redistributes these critical data blocks to other healthy nodes, ensuring that the data is always in a safe and available state. This dynamic adjustment mechanism greatly improves the security and reliability of power grid data management.

[0056] In this embodiment, after the encrypted data block is generated, the node that owns the data block calculates a hierarchical zero-knowledge proof through the generator, selects different levels of proof complexity based on the sensitivity and purpose of the data block, and uses a quantum-resistant encryption algorithm to generate a tag commitment σ i , a verification factor τ i , whose expression is: where ε lattice is an encryption function based on lattice theory, used to generate a quantum-secure tag commitment, M i is the i-th level data message or its digest, r i is a random number used to ensure the randomness and security of the encryption process, σ i is the tag commitment generated at the i-th level, which is the hash value of the encrypted message M i and the random number r i , and Vlattice K is a quantum secure verification factor generated by the verification function based on lattice theory i K is a shared key or verification parameter of the i-th level i K is a verification factor generated by the i-th level, which is a hash value of the label commitment and the verification parameter H is a hash function used to generate secure hash values; by introducing a hierarchical zero-knowledge proof mechanism and a quantum-resistant encryption algorithm, the system provides multiple security verifications at different levels. This multi-level verification mechanism ensures that even if the security of a certain level is threatened, other levels can still provide protection, thereby improving the overall security of the data.

[0057] After obtaining the hierarchical zero-knowledge proof, it is distributed together with the corresponding encrypted data block to each storage node in the blockchain network. After receiving the encrypted data block and its hierarchical zero-knowledge proof, the storage node starts the verification process. The verifier node independently verifies each level of proof using shared verification parameters according to the hierarchical zero-knowledge proof protocol, and its expression is: V layered V is the verification result of the hierarchical zero-knowledge proof. If all levels pass the verification, the value is true (True), K is a verification function of the i-th level, responsible for verifying whether the label commitment and verification factor of this level are correct i K is a shared key or verification parameter of the i-th level i K is a label commitment and verification factor of the i-th level i K is a shared key or verification parameter of the i-th level; through the hierarchical verification mechanism, only data blocks that pass all verifications will be officially stored in the distributed storage system, greatly improving the reliability and integrity of the stored data. The system rejects data blocks that do not pass the verification, preventing incomplete or tampered data from entering the system.

[0058] If all levels pass the verification, the data block is officially stored in the distributed storage system, and the verification result and related information are recorded in the distributed ledger of the blockchain. If the verification fails, the storage node will reject the storage of the data block and will not add it to the distributed storage of the system, and will immediately issue an alarm and send the alarm information to the alarm information processing module.

[0059] For example, suppose the system collects a set of highly sensitive power load data that needs to be distributed and stored in multiple nodes of substations. The system first generates encrypted data blocks for these data blocks and generates multi-level zero-knowledge proofs according to the sensitivity of the data blocks. Each proof level uses different encryption algorithms and verification mechanisms, especially introducing quantum-resistant encryption algorithms to defend against future quantum computing attacks. After generating the zero-knowledge proof, the system distributes the data blocks and corresponding proofs to the storage nodes in the blockchain network. After receiving these data, the storage nodes start the hierarchical verification process. By verifying the label commitment and verification factor layer by layer, the system ensures the integrity and authenticity of each data block. If all levels of verification pass, the data block will be officially stored in the distributed storage system and recorded in the blockchain ledger. If the verification fails, for example, the label commitment of a certain level does not match, the system will refuse to store the data block and immediately issue an alarm to notify the administrator to check the problem. The system also records the alarm information and the data block that fails the verification in the blockchain ledger, ensuring that the administrator can trace back and analyze the cause of the failure to prevent similar problems from happening again.

[0060] In this embodiment, the alarm information processing module performs the following steps when prioritizing processing:

[0061] The alarm information processing module collects real-time alarm information of power equipment and pre-processes the data, extracts time series features, environmental parameters and historical alarm record information, and calculates the fuzzy entropy of the time series data of the alarm information to obtain the complexity and uncertainty indicators of each alarm information, fuzzy entropy The expression for calculating is: wherein, is a time series segment with a length of m, r is the fuzziness of the sequence, n represents the number of data points in the sequence, r represents the allowed fuzziness threshold, which is used to control the similarity between two sequence segments, and N is the total number of data points, is the fuzziness function, which calculates the similarity between sequence segments X i and By calculating the fuzzy entropy of the alarm information, the system can quantify the complexity and uncertainty indicators of each alarm information. This quantitative evaluation can help the system more accurately determine which alarm information needs to be prioritized, avoiding potential risks caused by neglecting complex alarms.

[0062] The multi-head neural attention model is established and trained using historical data. The complexity and uncertainty indicators calculated by fuzzy entropy and other alarm features are input into the multi-head neural attention model. Different attention heads of the multi-head neural attention model focus on different features, and independently learn the importance and relevance of these features according to the output results of fuzzy entropy. The multi-head neural attention model intelligently sorts the alarm information according to the weights of different features. The system not only focuses on the current state of the alarm information, but also dynamically adjusts the weight of the alarm information through historical data and the calculation results of fuzzy entropy, ensuring that the alarm that may cause major problems is prioritized.

[0063] According to the results calculated by fuzzy entropy and multi-head neural attention mechanism, the weight of the alarm information is dynamically adjusted, and the alarm sorting is optimized, and the expression is: W alert is the dynamic weight of the alarm information, and i is the weight coefficient of the i-th attention head, which is calculated by the multi-head attention mechanism, and is the weight coefficient of fuzzy entropy, which is used to balance the influence of fuzzy entropy in the maximum weight; The alarm information processing module can not only process the current alarm in real time, but also dynamically adjust the alarm processing strategy through continuous learning of fuzzy entropy technology and multi-head neural attention model. The system regularly re-evaluates the stored alarm information to ensure that any new abnormality or risk can be discovered and responded to in a timely manner.

[0064] The alarm information processing module processes the alarm information according to the optimized sorting, prioritizes the alarm that may cause major problems, and continuously monitors the processing effect of the alarm information. Through fuzzy entropy technology, the stored alarm information is regularly re-evaluated. If new abnormalities or risks are detected, the alarm information processing module will dynamically adjust the alarm processing strategy and trigger the necessary emergency response.

[0065] For example: Suppose a transformer in the past alarm records, when the environmental temperature and load exceed a certain critical value at the same time, it often leads to device overload and failure. The multi-head neural attention model will automatically identify this feature and increase the weight of the current alarm information. The system then dynamically adjusts the alarm sorting according to the optimized weight, ensuring that the key alarm is prioritized. In addition, the system will regularly re-evaluate the stored alarm information. If new environmental changes or device abnormalities are detected, the system will dynamically adjust the alarm processing strategy and may trigger new emergency response measures. In this way, the system can maintain efficient alarm management under different operating conditions and ensure the stable operation of the smart grid.

[0066] Specifically, through fuzzy entropy calculation, multi-head neural attention model and dynamic alarm processing, the system realizes intelligent management and response optimization of power equipment alarm information. This design not only improves the safety and reliability of the system, but also effectively prevents key alarms from being ignored or delayed, ensuring the stable operation of the power system and the distributed energy system.

[0067] The present application is not limited to the above-mentioned embodiments, and any person should know that the structural changes made under the inspiration of the present application fall within the protection scope of the present application. The technical, shape and structure parts not described in detail in the present application are well-known technologies.

Claims

1. A blockchain-based power information security storage and supervision system, characterized by: include: Data acquisition module, delay-tolerant network (DTN) module, data slicing and encryption module, blockchain network module, zero-knowledge proof (ZKP) module, alarm information processing module, storage node management unit; A data acquisition module, configured to collect power data from the power equipment in real time and transmit the data to the delay-tolerant network module; a delay-tolerant network module, configured to prioritize the power data according to its time sensitivity and transmit the classified data to the data slicing and encryption module; The data slicing and encryption module is used to slice the priority-classified power data into multiple data blocks and encrypt each data block to generate encrypted data blocks; A blockchain network module, configured to receive encrypted data blocks and distribute them to storage nodes in multiple substations for distributed storage using blockchain technology. The blockchain network implements distributed storage and management of data based on Hyperledger Fabric, and records the storage path of the data blocks. During the distribution process, the blockchain network will dynamically select substation storage nodes suitable for storing these data blocks based on the metadata of the encrypted data blocks and the current status of the blockchain network, and allocate high-priority data blocks to nodes with better performance. The system will also regularly verify the paths of stored data blocks through the blockchain network and check their storage status at each node. If an anomaly is detected, the blockchain network will automatically reallocate the storage node. A zero-knowledge proof module is used to verify the validity of the encrypted data block based on the zero-knowledge proof protocol in the blockchain network, ensuring the confidentiality and integrity of the data in distributed storage; The alarm information processing module is used to collect and store alarm information related to power data. By analyzing historical alarm data, environmental parameters, equipment status, and other information, it uses a multi-head neural attention mechanism and fuzzy entropy technology to predict possible future risks. Finally, it dynamically adjusts the weight of alarm information, identifies alarms that may cause major problems in the future, and prioritizes the allocation of processing resources for these alarms. The storage node management unit is used to manage and maintain storage nodes distributed in multiple substations, ensuring the security and traceability of distributed storage of encrypted data blocks.

2. The blockchain-based power information security storage and supervision system according to claim 1 is characterized in that: When the data is prioritized according to the time sensitivity of the power data, a time sensitivity analysis is performed on the collected power data to obtain the sensitivity level of the power data, and the power data is classified according to preset priority rules based on the time sensitivity analysis, and the classified data is transmitted to the data slicing and encryption modules respectively according to the priority of the data.

3. The blockchain-based power information security storage and supervision system according to claim 2 is characterized in that: Sensitivity levels include high sensitivity, medium sensitivity and low sensitivity, among which high sensitivity data is given the highest priority, medium sensitivity and low sensitivity data correspond to medium priority and low priority respectively.

4. The blockchain-based power information security storage and supervision system according to claim 2 is characterized in that: Data slicing requires a preset slicing strategy before slicing. After obtaining the power information data, the received power data will be divided into multiple independent data blocks according to the preset slicing strategy. After each execution of the data slicing strategy, the system will record the details of the slicing strategy and the execution results on the blockchain; After obtaining independent data blocks, an asymmetric encryption algorithm is used to encrypt each data block after segmentation, and all data blocks are converted into encrypted data blocks. Each encrypted data block is accompanied by a unique encryption identifier. Finally, the generated encrypted data blocks will be transmitted to each storage node in the blockchain network for distributed storage.

5. The blockchain-based power information security storage and supervision system according to claim 1 is characterized in that: Before uploading data to the blockchain network, each substation's storage node must first be configured as a Hyperledger Fabric node. When the blockchain network module receives an encrypted data block containing metadata, the system uses a node selection algorithm to select a suitable node cluster from the available substation storage nodes based on the current node's load, network bandwidth, storage capacity, and geographic location. The blockchain network generates data block distribution instructions, instructing the distribution of encrypted data blocks to the selected storage nodes. After the distribution is completed, the blockchain network records the storage path, storage node identifier and related metadata of each encrypted data block in the distributed ledger of the blockchain.

6. The blockchain-based power information security storage and supervision system according to claim 1 is characterized in that: After the encrypted data block is generated, the node that owns the data block calculates a hierarchical zero-knowledge proof through the generator. Different levels of proof complexity are selected according to the sensitivity and purpose of the data block. For highly sensitive data blocks, a multi-level proof mechanism is adopted, each layer provides additional security verification, and uses a quantum-resistant encryption algorithm to generate a label commitment. , verification factor , whose expression is: ,in, is a lattice-theoretic encryption function used to generate quantum-secure label commitments. For the The data message of the layer or its summary, is a random number used to ensure the randomness and security of the encryption process. For the The tag commitment generated by the layer is the data message or its digest and random numbers The hash value of is a verification function based on lattice theory, used to generate quantum-safe verification factors. For the shared secrets or authentication parameters at the level, For the The verification factor generated by the hierarchy is the hash value of the tag commitment and the verification parameter. A hash function used to generate a secure hash value; After obtaining the layered zero-knowledge proof, it is distributed together with the corresponding encrypted data block to each storage node in the blockchain network. After receiving the encrypted data block and its layered zero-knowledge proof, the storage node starts the verification process. The validator node uses the shared verification parameters according to the layered zero-knowledge proof protocol to independently verify the proof at each level. The expression is: ,in, The verification result of the layered zero-knowledge proof. If all layers are verified, the value is true. For the The layer verification function is responsible for verifying whether the label commitment and verification factor of the layer are correct. For the Hierarchical label commitment and verification factors, For the shared secrets or authentication parameters at each level; If all levels of verification pass, the data block will be officially stored in the distributed storage system, and the verification results and related information will be recorded in the distributed ledger of the blockchain. If the verification fails, the storage node will refuse to store the data block and will not add it to the system's distributed storage. It will immediately issue an alarm and notify and send the alarm information to the alarm information processing module.

7. The blockchain-based power information security storage and supervision system according to claim 1 is characterized in that: The alarm information processing module performs the following steps when performing priority processing: The alarm information processing module collects the alarm information of power equipment in real time, pre-processes the data, extracts time series characteristics, environmental parameters and historical alarm record information, and performs fuzzy entropy on the time series data of the alarm information. Calculate and obtain the complexity and uncertainty index of each alarm information, fuzzy entropy The expression to be calculated is: ,in, To indicate the length A time series segment of is the fuzziness of the sequence, represents the number of data points in the series, Indicates the allowed fuzziness threshold, which is used to control the similarity between two sequence segments. The total number of data points, is the fuzziness function, calculating the sequence fragment and similarities between; Use historical data to build and train a multi-head neural attention model. Input the complexity and uncertainty indicators calculated by fuzzy entropy, as well as other alarm features, into the multi-head neural attention model. Different attention heads of the multi-head neural attention model focus on different features and independently learn the importance and relevance of these features based on the output results of fuzzy entropy. According to the results calculated by fuzzy entropy and multi-head neural attention mechanism, the weight of the alarm information is dynamically adjusted to optimize the alarm sorting. The expression is: , is the dynamic weight of the alarm information, For the The weight coefficient of each attention head is calculated by the multi-head attention mechanism. is the weight coefficient of fuzzy entropy, which is used to balance the influence of fuzzy entropy in the final weight; The alarm information processing module processes alarm information according to the optimized sorting, giving priority to responding to alarms that may cause major problems and continuously monitoring the processing effect of alarm information. It regularly re-evaluates the stored alarm information through fuzzy entropy technology. If new anomalies or risks are detected, the alarm information processing module will dynamically adjust the alarm processing strategy and trigger the necessary emergency response.

Images