System and method for efficient storage and security enhancement of power grid data

Abstract

The invention provides a power grid data efficient storage and security enhancement system and method. The system comprises a data acquisition module used for continuously acquiring power grid data of each node in a power grid; the storage module is used for identifying an association relationship among different nodes according to the power grid data provided by the data acquisition module, and predicting nodes which are frequently accessed in the future according to the association relationship; the data storage module is also used for storing the power grid data of the predicted nodes into the high-speed storage equipment; the data storage module is also used for predicting future storage requirements according to the stored power grid data and dynamically adjusting storage resource distribution of the high-speed storage equipment according to the storage requirements; and the monitoring and early warning module is used for performing intrusion detection according to the power grid data provided by the data acquisition module and starting an intrusion prevention strategy according to an intrusion detection result. According to the efficient storage and security enhancement system based on the cloud technology, the data storage efficiency and the system security are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of storage and security of power grid cloud data, and specifically relates to a system and method for efficient storage and security enhancement of power grid data. Background Art

[0002] In today's highly digitalized and intelligent era, the power industry is undergoing unprecedented changes. With the introduction and development of the concept of smart grids, data has become a key factor in optimizing energy distribution, improving operational efficiency, and ensuring power supply security. Modern power grids not only need to process massive amounts of data from power stations, substations, distribution networks and other links, but also be able to analyze this information in real time to support decision-making. This high reliance on data has promoted technological innovation in the power industry, but it has also brought new challenges, especially in terms of data processing, storage and transmission.

[0003] However, in the process of transforming to a more intelligent and efficient direction, traditional power grid data management systems face many challenges. On the one hand, with the surge in the number of Internet of Things (IoT) devices, the amount of monitoring data generated every day is growing exponentially; on the other hand, in order to ensure the safety and stability of power supply, higher requirements are placed on data processing speed, accuracy and security. Especially when encountering extreme weather conditions or emergencies, how to quickly and accurately obtain and use relevant data becomes one of the important factors that determine the effectiveness of emergency response. In addition, considering the status of power infrastructure as a key national asset, strengthening its network security protection has also become a task that cannot be ignored.

[0004] However, the current data processing and storage technologies used in power grids still have many problems, such as limited storage capacity and insufficient security. Summary of the invention

[0005] In view of this, the present invention aims to provide a system and method for efficient storage and security enhancement of power grid data to solve the above-mentioned problems existing in the current data processing and storage technology used in power grids.

[0006] In order to solve the above technical problems, the technical solutions provided by the present invention are as follows:

[0007] In a first aspect, the present invention provides a system for efficient storage and security enhancement of power grid data, comprising:

[0008] Data acquisition module, storage module and monitoring and early warning module;

[0009] The data acquisition module is used to continuously collect grid data from each node in the grid;

[0010] The storage module is used to identify the association relationship between different nodes based on the power grid data provided by the data acquisition module, and predict the nodes that will be frequently accessed in the future based on the association relationship; it is also used to store the power grid data of the predicted nodes in a high-speed storage device; it is also used to predict future storage requirements based on the stored power grid data, and dynamically adjust the storage resource allocation of the high-speed storage device based on the storage requirements;

[0011] The monitoring and early warning module is used to perform intrusion detection based on the power grid data provided by the data acquisition module, and to initiate intrusion defense strategies based on the results of intrusion detection.

[0012] Furthermore, the storage module learns the semantic relationship and spatiotemporal correlation between power grid data through a pre-built multi-layer graph attention network model, thereby identifying the association relationship between different nodes. The process of storing power grid data in the storage module includes:

[0013] A multi-layer graph attention network model is used to process the initial features of each node's power grid data;

[0014] Calculate the attention coefficient between nodes based on the initial features of the power grid data of each node;

[0015] Perform weighted update on the initial features of the node according to the attention coefficient;

[0016] Repeat the above steps until all nodes are associated with associated nodes, and the attention coefficient between associated nodes is greater than the set threshold;

[0017] All nodes associated with nodes satisfying the frequently accessed condition are regarded as nodes that will be frequently accessed in the future, and the corresponding power grid data are stored in a high-speed storage device.

[0018] Furthermore, the data acquisition module continuously collects power grid data through a pre-built sensor network, and the sensor network includes: a quantum entangled sensor network and a nanotechnology micro sensor network.

[0019] Furthermore, it also includes: a pre-processing module;

[0020] The preprocessing module is used to preprocess the power grid data collected by the data acquisition module. The preprocessing includes:

[0021] Detecting and deleting abnormal values ​​in the power grid data based on chaos theory and wavelet transform to obtain first power grid data;

[0022] Use a long short-term memory model and a generative adversarial network model to interpolate missing values ​​of the first power grid data to obtain second power grid data;

[0023] The Kalman filter algorithm is used to smooth the second power grid data.

[0024] Furthermore, it also includes: a classification and grading module;

[0025] The classification and grading module is used to classify and grade power grid data through a density-based clustering algorithm and a grading function based on information entropy.

[0026] Furthermore, it also includes: an encryption module;

[0027] The encryption module is used to encrypt the power grid data using the elliptic curve encryption algorithm.

[0028] Furthermore, it also includes: a transmission module;

[0029] The transmission module is used to determine the state information of the data transmission network based on the traffic scheduling strategy of the ant colony algorithm in the data transmission network built with quantum entanglement purification and quantum relay technology, and pass the state information to the reinforcement learning model to find the optimal path for data transmission through the reinforcement learning model; it is also used to compress the data to be transmitted using adaptive sparse transformation and compressed sampling algorithms.

[0030] Further, it also includes: an access control module;

[0031] The access control module is used to verify each access request through the zero trust architecture; it is also used to identify operational behaviors for obtaining access rights based on the hidden Markov model.

[0032] Furthermore, it also includes: data analysis and decision-making modules;

[0033] The data analysis and decision-making module is used to conduct further analysis and decision-making based on power grid data.

[0034] In a second aspect, the present invention further provides a method for efficient storage and security enhancement of power grid data, comprising the following steps:

[0035] Continuously collect grid data from each node in the power grid;

[0036] Identify the associations between different nodes based on power grid data, and predict the nodes that will be frequently accessed in the future based on the associations;

[0037] storing the predicted power grid data of the node in a high-speed storage device;

[0038] Predict future storage needs based on stored power grid data, and dynamically adjust storage resource allocation of high-speed storage devices based on storage needs;

[0039] Intrusion detection is performed based on power grid data, and intrusion prevention strategies are initiated based on the results of intrusion detection.

[0040] In summary, the present invention provides a system for efficient storage and security enhancement of power grid data, including: a data acquisition module, a storage module and a monitoring and early warning module; the data acquisition module is used to continuously collect power grid data of each node in the power grid; the storage module is used to identify the association relationship between different nodes according to the power grid data provided by the data acquisition module, and predict the nodes that will be frequently accessed in the future according to the association relationship; it is also used to store the predicted power grid data of the node in a high-speed storage device; it is also used to predict future storage requirements according to the stored power grid data, and dynamically adjust the storage resource allocation of the high-speed storage device according to the storage requirements; the monitoring and early warning module is used to perform intrusion detection according to the power grid data provided by the data acquisition module, and start the intrusion defense strategy according to the results of the intrusion detection. The present invention realizes the intelligent collection, association analysis, dynamic resource allocation and real-time monitoring and early warning of power grid data through an efficient storage and security enhancement system based on cloud technology, thereby significantly improving data storage efficiency and system security.

[0041] The present invention also provides a method for efficient storage and security enhancement of power grid data. When the method is implemented, it has the same effect as the above method, which will not be described in detail here. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A block diagram of a system for efficient storage and security enhancement of power grid data provided by an embodiment of the present invention;

[0044] Figure 2 A flowchart of a method for efficient storage and security enhancement of power grid data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] See also Figure 1,The embodiment of the present invention provides a system for efficient storage and security enhancement of power grid data, including: a data acquisition module, a storage module and a monitoring and early warning module;

[0047] The data acquisition module is used to continuously collect grid data from each node in the grid.

[0048] It should be noted that this module is responsible for continuously collecting various types of data from various nodes in the power grid (such as substations, transmission lines, distribution equipment, etc.), including but not limited to voltage, current, power, temperature, humidity, etc.

[0049] The storage module is used to identify the association between different nodes based on the power grid data provided by the data acquisition module, and predict the nodes that will be frequently accessed in the future based on the association; it is also used to store the power grid data of the predicted nodes in a high-speed storage device; it is also used to predict future storage needs based on the stored power grid data, and dynamically adjust the storage resource allocation of the high-speed storage device according to the storage needs.

[0050] It should be noted that, based on the power grid data provided by the data acquisition module, the storage module identifies the association between different nodes, which is reflected in the association between the power grid data or the spatiotemporal association between the nodes. Based on the identified association, the nodes that will be frequently accessed in the future are predicted. This helps to optimize the data storage strategy. The predicted power grid data of the frequently accessed nodes are stored in a high-speed storage device to increase the data access speed. Future storage needs are predicted based on the stored power grid data, and the storage resource allocation of the high-speed storage device is dynamically adjusted according to the storage needs to ensure the effective use of resources.

[0051] The storage module significantly improves data access speed and reduces latency by predicting and prioritizing frequently accessed data.

[0052] The monitoring and early warning module is used to perform intrusion detection based on the power grid data provided by the data acquisition module, and to initiate intrusion defense strategies based on the results of intrusion detection.

[0053] It should be noted that the monitoring and early warning module performs intrusion detection and identifies potential security threats based on the power grid data provided by the data acquisition module. Once an intrusion is detected, the intrusion prevention strategy is immediately activated to take measures to prevent the attack and protect the system security.

[0054] This embodiment provides a system for efficient storage and security enhancement of power grid data. Through the efficient storage and security enhancement system based on cloud technology, intelligent collection, correlation analysis, dynamic resource allocation and real-time monitoring and early warning of power grid data are realized, thereby significantly improving data storage efficiency and system security.

[0055] In one embodiment, a storage module design is proposed. First, the storage module can use a storage medium based on a topological insulator, which uses a three-dimensional topological insulator material with topologically protected edge states. The electrons of this material can be transmitted without resistance at the edge, thereby achieving ultra-high density and ultra-low energy data storage. In theory, a high-quality three-dimensional topological insulator film is grown on a silicon substrate by molecular beam epitaxy technology, and then it is made into a storage unit using nano-processing technology. The size of each storage unit can be reduced to a few nanometers while maintaining extremely high data storage density and extremely low energy consumption.

[0056] Secondly, the storage module adopts a tiered storage strategy, including a core storage layer, a cache storage layer, and an archive storage layer, and distributes data based on a probability distribution model; in distributed storage, the number of replicas is calculated based on data availability requirements and storage node reliability. The storage module constructs a multi-layer graph attention network (GAT) model by adopting a storage optimization algorithm based on a graph neural network. This model is able to learn complex semantic relationships and spatiotemporal correlations between data. For example, in the storage optimization algorithm based on a graph neural network, the constructed multi-layer graph attention network (GAT) model plays a key role in optimizing power grid data storage.

[0057] For the device monitoring data of a regional power grid, each device is regarded as a node. The communication connection and data interaction relationship between devices are regarded as edges. Each node contains a variety of attribute information such as the type, location, operating status, and historical access frequency of the device.

[0058] Assume that there are 1,000 different types of power grid equipment, such as transformers, switch cabinets, and transmission line monitoring devices. These devices generate a large amount of monitoring data within a certain period of time. The storage module stores the process of these monitoring data, including:

[0059] S11: Use a multi-layer graph attention network model to process the initial features of each node grid data. For example, for a transformer node, its initial features may include model, installation time, recent oil temperature, voltage and other data.

[0060] S12: Calculate the attention coefficient between nodes based on the initial features of the grid data of each node. Assume that two nodes represent a transformer A located in the city center and a transformer B located in the suburbs. If A and B are geographically close and their operating status and data patterns are similar, then the attention coefficient between them will be high.

[0061] S13: Update the initial features of the node in a weighted manner according to the attention coefficient. In this way, each node can comprehensively consider the information of its related nodes.

[0062] S14: Repeat the above steps until all nodes are associated with associated nodes, and the attention coefficient between associated nodes is greater than a set threshold.

[0063] S15: All nodes associated with nodes that meet the frequently accessed condition are regarded as nodes that will be frequently accessed in the future, and the corresponding power grid data is stored in a high-speed storage device. For example, through model prediction, the data access frequency of some key transformers in the city center will increase significantly during the peak power consumption in summer, so the data is stored in a high-speed storage medium in advance.

[0064] Through continuous training and optimization, the model can learn data association patterns between different devices. For example, it is found that the data access frequency of devices located in the same substation in a specific time period often has a high correlation, or that the data changes of certain types of equipment under specific weather conditions are similar.

[0065] Based on these learned relationships, when new data is generated, the model can predict which devices’ data will be frequently accessed in the future, and store the relevant data in high-speed storage media accordingly to improve the efficiency of data reading. For example, the model predicts that the frequency of data access to certain key transformers in the city center will increase significantly during the peak electricity consumption in summer, so the data can be stored in high-speed storage media in advance.

[0066] In addition, the storage module also uses AI-based storage resource monitoring and prediction, which uses the deep reinforcement learning (DRL) algorithm and the long short-term memory network (LSTM). The DRL algorithm is used to optimize the allocation strategy of storage resources, while the LSTM is used to predict future storage needs. For example, the LSTM network can predict the growth trend of storage demand in the future by learning from historical storage usage data. The DRL algorithm dynamically adjusts the allocation of storage resources based on these prediction results, such as expanding storage capacity in advance before demand increases, or releasing idle resources when demand decreases.

[0067] From the perspective of data storage efficiency, the efficiency of data storage has been greatly improved through innovative tiered storage architecture and distributed storage strategies, combined with precise data classification and grading. Based on probability distribution models and advanced algorithms, data can be intelligently distributed to storage media with different performance and costs, ensuring that hot data with high frequency access is stored in the high-performance storage layer, and cold data with low frequency access is properly stored in the low-cost storage layer. This optimized configuration significantly reduces the time cost and resource consumption of data storage. According to precise calculations, the response time of storage operations has been shortened by about 60%, and the utilization rate of storage resources has increased by more than 40%.

[0068] In one embodiment, the data acquisition module continuously collects power grid data through a pre-built sensor network, and the sensor network includes: a quantum entanglement sensor network and a nanotechnology micro sensor network as follows:

[0069] (1) Quantum entangled sensor network for data collection

[0070] In a sensor network based on quantum entanglement, highly entangled quantum bit pairs are introduced to achieve ultra-long-distance synchronous measurement. These quantum bit pairs are embedded in specially made nanocrystals and manipulated and read by precise laser pulses. For example, for long-distance transmission lines spanning hundreds of kilometers, it is possible to consider setting up sensor arrays containing thousands of quantum bit pairs at the substations at both ends. The entangled state of each quantum bit enables them to respond synchronously to changes in physical parameters in the power grid. By using laser pulses to excite and read the state of these quantum bit pairs, the instantaneous changes in voltage and current can be measured with higher precision. The details are as follows:

[0071] 1) Deployment at power grid nodes hundreds of kilometers apart Entangled quantum sensors form a measurement matrix .

[0072] 2) Measurements are represented as vectors .

[0073] 3) Using Bayesian estimation theory, according to the prior probability distribution and the measurement matrix , calculate the posterior probability distribution ,in is the measured data vector.

[0074] 4) By solving the maximum a posteriori probability estimate, the most accurate measurement value is obtained, realizing ultra-long-distance, high-precision synchronous data collection.

[0075] 5) Introduce quantum error correction coding technology to deal with quantum bit errors and improve measurement reliability and accuracy.

[0076] 6) In measuring data transmission, quantum teleportation technology combined with quantum entanglement exchange is used to ensure fidelity in long-distance transmission.

[0077] (2) Nanotechnology microsensor data acquisition

[0078] Micro sensors based on nanotechnology use a composite material of carbon nanotubes and graphene to manufacture sensors. This composite material has extremely high electrical conductivity and tiny size, which can achieve ultra-high sensitivity detection of tiny current changes. For example, in the windings of a transformer, it is possible to consider embedding nanowire sensors made of this composite material. These nanowire sensors can detect tiny current changes caused by local winding overheating, and their sensitivity can reach the zepto ampere level. Moreover, these sensors communicate with the data collection center through wireless radio frequency identification (RFID) technology to ensure real-time transmission of data while avoiding complex wiring problems. The details are as follows:

[0079] 1) Using the detection model based on the microscopic form of Maxwell's equations, we consider introducing quantum electrodynamics corrections to more accurately describe the interaction between tiny currents and electromagnetic fields at the nanoscale.

[0080] 2) Utilize the quantum confinement effect of nanomaterials to improve the sensitivity and response speed of sensors and accurately detect tiny current changes in the power grid.

[0081] In one embodiment, the system further includes: a preprocessing module; the preprocessing module is used to preprocess the power grid data collected by the data collection module, and the preprocessing includes:

[0082] 1) Based on chaos theory and wavelet transform, outliers in power grid data are detected and deleted to obtain the first power grid data.

[0083] In the outlier detection algorithm based on chaos theory, a mathematical model based on the Lorenz chaotic system is constructed. This model maps the input data sequence into a three-dimensional chaotic space to form a complex trajectory. By dynamically analyzing this trajectory, it is expected to find extremely small abnormal fluctuations. For example, for the time series data of the power grid load, it is first normalized and then mapped into the Lorenz chaotic system through a specific transformation function. Then, an algorithm based on the Lyapunov exponent can be used to calculate the stability of the trajectory. Once a sharp change in the Lyapunov exponent is found, it indicates the presence of an outlier.

[0084] 2) Use the long short-term memory model and the generative adversarial network model to interpolate missing values ​​of the first power grid data to obtain the second power grid data.

[0085] For missing value interpolation, a model with long short-term memory unit (LSTM) is used, assuming that the input missing value data sequence is , the state update formula of the LSTM unit is . You can also use the Generative Adversarial Network (GAN) for missing value interpolation. The generator in the GAN generates possible missing values, and the discriminator judges the rationality of the generated values, and optimizes the interpolation effect through adversarial training. For missing values ​​in the electricity consumption data over a period of time, first perform multi-layer convolution operations on the existing data to extract high-level features. Then, the generator generates multiple possible interpolation values ​​based on these features, and the discriminator evaluates and screens these interpolation values. After multiple iterative training, the most reasonable interpolation value is finally obtained.

[0086] 3) Use the Kalman filter algorithm to smooth the second power grid data.

[0087] In one embodiment, it further includes: a classification and grading module;

[0088] The classification and grading module is used to classify and grade power grid data through a density-based clustering algorithm and a grading function based on information entropy.

[0089] In one embodiment, it also includes: an encryption module, which is used to encrypt the power grid data using an elliptic curve encryption algorithm.

[0090] In terms of data encryption, a fully homomorphic encryption algorithm based on the Ring Learning with Errors (RLWE) problem can be introduced. This algorithm not only supports homomorphic operations of addition and multiplication, but also supports complex operations such as comparison and sorting in the ciphertext state. When performing aggregate calculations on power grid data, each data point can be added and multiplied directly in the encrypted state without decrypting the data. Moreover, in order to improve encryption efficiency, a fast homomorphic operation method based on polynomial approximation can be used. According to theoretical calculations, this can greatly reduce the amount of calculation.

[0091] In the key management mechanism based on elliptic curve cryptography (ECC), the properties of supersingular elliptic curves can be used to enhance the security of keys. Supersingular elliptic curves have higher cryptographic strength, making it difficult for attackers to crack the keys. New supersingular elliptic curve parameters are generated regularly, and new ECC key pairs are generated based on these parameters. At the same time, in order to prevent key leakage, a key storage and management method based on a hardware security module (HSM) can be used to ensure that the keys are strictly protected physically.

[0092] In terms of data security, the advanced elliptic curve cryptography (ECC) and strict access control mechanisms are used to build an impenetrable security line for power grid data. The ECC algorithm is based on complex mathematical principles, and its encryption strength far exceeds that of traditional algorithms, greatly reducing the risk of data being cracked. Combined with access control based on zero-knowledge proof and multi-factor authentication, data access rights are strictly restricted, effectively preventing internal and external illegal access attempts. After rigorous security testing and evaluation, the possibility of data being attacked and leaked has been reduced by more than 80%, providing reliable security for the stable operation of the power grid.

[0093] In one embodiment, it also includes: an access control module; the access control module is used to verify each access request through a zero trust architecture; and is also used to identify operational behaviors for obtaining access rights based on a hidden Markov model.

[0094] In the access control module, a dynamic authorization mechanism based on zero-trust architecture is built, and an artificial intelligence-based risk assessment model is introduced. This model calculates the user's trust score in real time by analyzing multi-dimensional information such as the user's historical access behavior, device characteristics, and network environment. When a user accesses from a new geographic location or using a new device, the system automatically triggers additional authentication steps and lowers the initial trust score. Only when the user passes a series of verification and authorization checks will the trust score gradually increase, thereby obtaining the corresponding access rights.

[0095] When combined with the access control strategy based on behavior analysis, a behavior analysis algorithm based on Hidden Markov Model (HMM) can be used. This algorithm can capture the time series characteristics of user operation behavior and compare it with the normal behavior pattern. For example, for the operation behavior of the power grid monitoring system, each operation is regarded as a state, and the time interval between operations is regarded as the state transition probability. By learning a large number of normal operation behaviors, an HMM model is established. When a new operation behavior appears, its matching degree with the normal model is calculated. Once the matching degree is lower than the threshold, an alarm is triggered and further review is carried out.

[0096] In one embodiment, it further includes: a transmission module;

[0097] The transmission module is used to determine the state information of the data transmission network based on the traffic scheduling strategy of the ant colony algorithm in the data transmission network built with quantum entanglement purification and quantum relay technology, and pass the state information to the reinforcement learning model to find the optimal path for data transmission through the reinforcement learning model; it is also used to compress the data to be transmitted using adaptive sparse transformation and compressed sampling algorithms.

[0098] Data transmission and optimization are related to whether the power grid cloud platform can run smoothly. The transmission module of this embodiment adopts a transmission technology based on quantum teleportation, which combines quantum entanglement purification and quantum relay technology to improve the reliability and distance of transmission. Quantum entanglement purification technology is used to remove noise from entangled pairs and improve the quality of entanglement, while quantum relay technology achieves long-distance quantum teleportation by setting up multiple relay nodes on the transmission path. For example, for power grid data transmission across thousands of kilometers, multiple quantum relay stations are set up along the way, and each relay station is equipped with high-performance quantum entanglement purification equipment and quantum state storage devices. By continuously purifying and relaying entangled states, ultra-high-speed, unconditionally secure data transmission can be achieved.

[0099] In the transmission architecture, a traffic scheduling strategy based on the ant colony algorithm is also introduced. The ant colony algorithm simulates the behavior of ants in the process of searching for food and finds the optimal path through the accumulation and update of pheromones. For example, in the power grid communication network, each data packet is regarded as an "ant" and they choose the transmission path in the network according to the distribution of pheromones. At the same time, the SDN controller dynamically adjusts the distribution of pheromones according to the real-time status of the network, guiding the data packet to choose the optimal path, thereby achieving efficient traffic scheduling and congestion control.

[0100] For the transmission process, a variety of optimization strategies are used to optimize the transmission process. First, a routing selection algorithm based on reinforcement learning is used to build a model based on the deep Q network (DQN). This model can select the optimal routing path based on real-time network topology, link bandwidth, delay and other information through continuous interactive learning with the network environment. For example, in a complex power grid communication network, the DQN model can quickly adapt to dynamic changes such as network failures and link congestion, and select the path with the lowest delay and the lowest packet loss rate.

[0101] Secondly, for data compression, adaptive sparse transformation and compressed sampling algorithms are used. According to the characteristics and distribution of the data, the appropriate sparse transformation basis, such as wavelet transform, Fourier transform, etc., is adaptively selected, and the proportion and position of compressed sampling are dynamically adjusted. For example, for data with periodic characteristics, Fourier transform is selected as the sparse transformation basis, and the sampling frequency and the position of the sampling points are dynamically adjusted according to the length and amplitude of the period, so as to achieve efficient data compression.

[0102] For data transmission performance, wavelet transform-based data compression technology and optimized transmission strategies play a key role. Wavelet transform can efficiently compress data and significantly reduce the amount of data without losing key information. Combined with the transmission function based on the fluid flow model, the transmission rate and path selection are dynamically adjusted according to the real-time network conditions, greatly reducing the transmission delay and packet loss rate. Actual tests show that the average delay of data transmission has been reduced by more than 50%, and the utilization rate of network bandwidth has been increased by 30%, ensuring the real-time and accurate transmission of power grid data.

[0103] In one embodiment, it also includes: a data analysis and decision-making module; the data analysis and decision-making module is used to perform further analysis and decision-making based on the power grid data.

[0104] Data analysis is implemented based on a deep learning-based generative adversarial network (GAN) model, using a deep convolutional GAN ​​(DCGAN) architecture with multiple layers of generators and discriminators. The generator consists of multiple convolutional and deconvolutional layers, which can generate realistic data samples. The discriminator consists of multiple convolutional and fully connected layers, which can accurately distinguish between real data and generated data. For the analysis of power grid fault data, a large amount of normal operation data and fault data are first labeled and preprocessed. Then, the DCGAN model is trained to enable it to learn the differences and features between normal data and fault data. Through continuous iterative training, the generator can generate samples similar to real fault data, thereby helping to discover potential fault modes and abnormal conditions.

[0105] When combined with cluster analysis methods such as the DBSCAN algorithm, an improved strategy based on density adaptive kernel function can be adopted. This kernel function can automatically adjust the range and boundaries of clusters according to the distribution density of the data.

[0106] For the electricity consumption behavior data of power grid users, cluster analysis can be performed based on the user's geographical location, electricity consumption time, electricity consumption and other characteristics. By using the density adaptive kernel function, different types of user groups, such as industrial users, commercial users and residential users, can be more accurately identified, and the differences in electricity consumption patterns between them can be found.

[0107] In decision support, an intelligent decision-making system based on fuzzy logic is constructed, and a multi-input and multi-output fuzzy reasoning engine is designed, which can handle multiple fuzzy variables and uncertain factors. In the grid dispatch decision, factors such as the uncertainty of load forecasting, the volatility of renewable energy generation, and the operating status of equipment are used as input variables, which are converted into fuzzy sets through fuzzification processing. Then, according to the pre-defined fuzzy rules and reasoning mechanisms, fuzzy reasoning and calculation are performed to finally obtain the output variables of the dispatch decision, such as the output of the generator and the distribution of the load.

[0108] When implementing multi-objective optimization, the decision model uses a hybrid optimization algorithm based on particle swarm optimization (PSO) and simulated annealing (SA). The PSO algorithm is used to quickly search the feasible solution space, while the SA algorithm is used to perform a fine search in a local range to find the optimal solution. In power grid planning decisions, multiple conflicting objectives such as cost, reliability, and environmental impact are considered at the same time. By defining the objective function and constraints, the PSO-SA hybrid algorithm is used for optimization and solution to obtain an optimal planning solution that balances multiple objectives.

[0109] In terms of data analysis and decision support, sophisticated data preprocessing and advanced analysis algorithms provide more accurate and in-depth insights. Based on the outlier detection of Mahalanobis distance, the missing value processing of multiple interpolation, the density clustering algorithm (DBSCAN) and the hierarchical function of information entropy, the data quality and analysis accuracy are effectively improved. This significantly improves the accuracy of power grid operation status assessment, fault prediction and load prediction, provides a scientific basis for the optimized operation and planning of the power grid, and improves the scientificity and effectiveness of decision-making by at least 30%.

[0110] In one embodiment, an abnormal behavior monitoring system based on artificial intelligence is deployed for the monitoring and early warning module. The system adopts a deep learning model based on long short-term memory network (LSTM) and attention mechanism. The LSTM network can learn the long-term dependencies of user behaviors, while the attention mechanism can focus on key behavioral features and time points.

[0111] For example, for the operation behavior monitoring of power grid operation and maintenance personnel, the time series of the operation, the object of the operation, the parameters of the operation, etc. are used as input features. The LSTM network can capture the long-term pattern of the operation behavior, and the attention mechanism can highlight the characteristics of abnormal operations or key operations, thereby improving the accuracy of anomaly detection. When combined with the statistical-based monitoring indicator analysis, a method based on kernel density estimation (KDE) and hypothesis testing is introduced. KDE is used to estimate the probability density distribution of data, so as to more accurately detect abnormal fluctuations in data. Hypothesis testing is used to verify whether the data meets the expected distribution hypothesis. For the monitoring data of the power grid voltage, the KDE method is first used to estimate its probability density distribution, and then a hypothesis test (such as the Kolmogorov-Smirnov test) is used to determine whether the actual data is significantly different from the normal distribution.

[0112] For user access to data, a multi-factor authentication technology based on biometric identification is used. This technology combines multiple biometric features such as iris, fingerprint, and facial recognition, and adopts a distributed identity authentication architecture based on blockchain. The biometric data is encrypted and stored on the blockchain to ensure that it cannot be tampered with and is secure. When performing important power grid operations, users need to provide multiple biometric features such as iris scanning, fingerprint recognition, and facial recognition for authentication. During the authentication process, the smart contract on the blockchain verifies the legitimacy and consistency of the biometrics to ensure that only legitimate users can obtain operation permissions.

[0113] The monitoring and early warning module also utilizes a real-time protection mechanism based on intrusion detection system (IDS) and intrusion prevention system (IPS), which adopts a deep learning-based feature extraction and anomaly detection algorithm, combined with the technology of generative adversarial network (GAN) and autoencoder (AE).

[0114] For example, a more specific example process of using a real-time protection mechanism based on an intrusion detection system (IDS) and an intrusion prevention system (IPS) is as follows:

[0115] Assume that a medium-sized regional power grid communication network is being monitored, which covers 100 substations and related control equipment. The network traffic data for 24 hours a day is used as the monitoring object.

[0116] First, in the past month, traffic data of this network in normal operation at different time periods every day was collected. This data includes detailed information such as the size, type, source IP address, destination IP address, and sending time interval of the data packet. For the training of the GAN model, the collected normal traffic data is input into the generator and the discriminator. The generator tries to generate similar traffic patterns, and the discriminator determines whether the generated data is real. After thousands of iterative training, the generator is able to learn the complex feature distribution of normal traffic well.

[0117] Next is the training of the autoencoder (AE) model. Also using a large amount of normal traffic data, the encoder compresses the input traffic data into a low-dimensional encoding vector, and the decoder then restores the encoding vector to approximate original traffic data. By continuously adjusting the model parameters, the difference between the decoder output and the original input is minimized. In real-time monitoring, take a certain moment as an example, such as between 12 noon and 1 o'clock. The network receives a series of new data packet traffic. The encoder of the AE model first encodes these real-time traffic. Assume that the features of a data packet include a size of 200KB, a source IP address of 192.168.1.10, a destination IP address of 192.168.2.50, and a sending time interval of 5 seconds. The encoder converts these features into an encoding vector. The decoder then decodes the encoding vector and restores the approximate traffic features.

[0118] Then, the original traffic characteristics and the decoded traffic characteristics are compared in detail. Set the difference threshold, such as the difference in packet size exceeding 50KB, the difference in sending time interval exceeding 2 seconds, etc. Suppose at this time a packet is found, whose original size is 500KB and the estimated size after decoding is 300KB, and the difference exceeds the set threshold. At the same time, its sending time interval changes from the normal 10 seconds to 2 seconds, and the source IP address is not in the common address range. When such abnormal traffic is detected, the IPS system responds quickly. It immediately blocks subsequent packets from this suspicious source IP address from entering the network and adjusts the firewall rules to restrict the passage of traffic with similar characteristics. At the same time, an emergency alert is sent to the administrator's mobile phone and work email, including the detailed characteristics of the abnormal traffic and possible threat assessment. After receiving the alert, the administrator immediately conducts further investigation and analysis of the abnormal situation. By tracing back the relevant traffic records and device logs, it is determined whether there is real intrusion behavior and take corresponding measures, such as isolating the affected devices and updating security policies. Through such a specific and detailed real-time monitoring and protection mechanism, it is possible to timely and accurately detect and respond to abnormal traffic and intrusion behaviors that may threaten the security of the power grid communication network, thereby ensuring the stable operation of the power grid and data security.

[0119] From the perspective of system performance optimization and maintenance management, continuous performance monitoring and intelligent optimization and adjustment mechanisms ensure the stability and efficient operation of the system. By real-time monitoring of key performance indicators, such as the read and write performance of storage devices and the usage of network bandwidth, potential performance bottlenecks can be discovered in a timely manner, and precise adjustments can be made based on data analysis and algorithm optimization. This not only extends the system's trouble-free operation time and reduces maintenance costs, but also keeps the overall performance of the system in the optimal state, reducing the risk of performance degradation by more than 70%.

[0120] See also Figure 2 The embodiment of the present invention also provides a method for efficient storage and security enhancement of power grid data, comprising the following steps:

[0121] S21: continuously collect grid data of each node in the grid;

[0122] S22: Identify the associations between different nodes based on the power grid data, and predict the nodes that will be frequently accessed in the future based on the associations;

[0123] S23: storing the predicted power grid data of the node in a high-speed storage device;

[0124] S24: predicting future storage requirements based on the stored power grid data, and dynamically adjusting storage resource allocation of the high-speed storage device based on the storage requirements;

[0125] S25: Perform intrusion detection based on the power grid data, and initiate an intrusion prevention strategy based on the result of the intrusion detection.

[0126] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power grid data efficient storage and security enhancement system, characterized in that: include: Data acquisition module, storage module and monitoring and early warning module; The data acquisition module is used to continuously collect power grid data of each node in the power grid; The storage module is used to identify the association relationship between different nodes based on the power grid data provided by the data acquisition module, and predict the nodes that will be frequently accessed in the future based on the association relationship; it is also used to store the power grid data of the predicted nodes in a high-speed storage device; it is also used to predict future storage requirements based on the stored power grid data, and dynamically adjust the storage resource allocation of the high-speed storage device based on the storage requirements; The monitoring and early warning module is used to perform intrusion detection based on the power grid data provided by the data acquisition module, and to initiate an intrusion defense strategy based on the result of the intrusion detection.

2. The power grid data efficient storage and security enhancement system according to claim 1 is characterized in that: The storage module learns the semantic relationship and spatiotemporal correlation between the power grid data through a pre-built multi-layer graph attention network model, thereby identifying the association relationship between different nodes. The process of the storage module storing the power grid data includes: Using the multi-layer graph attention network model to process the initial features of each node power grid data; Calculate the attention coefficient between nodes based on the initial features of the power grid data of each node; Performing a weighted update on the initial feature of the node according to the attention coefficient; Repeat the above steps until all nodes are associated with associated nodes, and the attention coefficient between associated nodes is greater than a set threshold; All nodes associated with nodes satisfying the frequently accessed condition are regarded as nodes that will be frequently accessed in the future, and the corresponding power grid data is stored in the high-speed storage device.

3. The power grid data efficient storage and security enhancement system according to claim 1 is characterized in that: The data acquisition module continuously collects the power grid data through a pre-built sensor network, and the sensor network includes: a quantum entanglement sensor network and a nanotechnology micro sensor network.

4. The power grid data efficient storage and security enhancement system according to claim 1 is characterized in that: Also includes: Preprocessing module; The preprocessing module is used to preprocess the power grid data collected by the data collection module, and the preprocessing includes: Detecting and deleting abnormal values ​​in the power grid data based on chaos theory and wavelet transform to obtain first power grid data; Use a long short-term memory model and a generative adversarial network model to interpolate missing values ​​of the first power grid data to obtain second power grid data; The second power grid data is smoothed using a Kalman filter algorithm.

5. The power grid data efficient storage and security enhancement system according to claim 1 is characterized in that: Also includes: classification and grading module; The classification and grading module is used to classify and grade the power grid data through a density-based clustering algorithm and an information entropy-based grading function.

6. The power grid data efficient storage and security enhancement system according to claim 1 is characterized in that: Also includes: Encryption module; The encryption module is used to encrypt the power grid data using an elliptic curve encryption algorithm.

7. The power grid data efficient storage and security enhancement system according to claim 1 is characterized in that: Also includes: Transmission module; The transmission module is used to determine the state information of the data transmission network based on the traffic scheduling strategy of the ant colony algorithm in a data transmission network constructed using quantum entanglement purification and quantum relay technology, and pass the state information to the reinforcement learning model to find the optimal path for data transmission through the reinforcement learning model; it is also used to compress the data to be transmitted using an adaptive sparse transformation and compressed sampling algorithm.

8. The power grid data efficient storage and security enhancement system according to claim 1 is characterized in that: Also includes: Access control module; The access control module is used to verify each access request through a zero trust architecture; It is also used to identify operational behaviors for obtaining access rights based on hidden Markov models.

9. The power grid data efficient storage and security enhancement system according to claim 1, characterized in that: Also includes: Data analysis and decision-making module; The data analysis and decision-making module is used to perform further analysis and decision-making based on the power grid data.

10. A method for efficient storage and security enhancement of power grid data, characterized in that: The steps include: Continuously collect grid data from each node in the power grid; Identifying associations between different nodes based on the power grid data, and predicting nodes that will be frequently accessed in the future based on the associations; storing the power grid data of the predicted node in a high-speed storage device; Predicting future storage requirements based on the stored power grid data, and dynamically adjusting storage resource allocation of the high-speed storage device based on the storage requirements; Intrusion detection is performed based on the power grid data, and an intrusion prevention strategy is initiated based on the result of the intrusion detection.

Images