A Data Security Assessment Method and System Based on Clustering Processing

Abstract

This invention discloses a data security assessment method and system based on clustering processing. The method includes data acquisition, data coherence detection, clustering processing, and data security assessment. This invention belongs to the field of data processing, specifically referring to a data security assessment method and system based on clustering processing. This scheme calculates the correlation between each pair of buses and introduces a graph model to represent the power network, which helps to identify critical paths and potential fault points. Adaptive genetic operations further optimize the partitioning of the initial anomaly region, dynamically updating the weight matrix based on real-time data and the influence of external factors, improving the ability to cope with complex and changing power system environments. Anomalies in the detected region are accurately assessed based on adaptive projection metrics, thereby improving the data security assessment effect.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a data security assessment method and system based on clustering processing. Background Technology

[0002] Data security assessment methods analyze and evaluate the normality of data areas to identify potential security risks. Once an anomaly is detected, the system issues an early warning to help managers take appropriate measures to ensure the safe operation of the system. However, general data security assessment methods suffer from several drawbacks: they cannot adapt to changes in the operating conditions and environmental conditions of the power system in a timely manner, leading to poor accuracy in security assessments; they also have low precision in detecting fault areas, resulting in misjudgments; and they cannot reflect the real-time operating status of the power system, leading to poor security assessment results. Summary of the Invention

[0003] To address the aforementioned issues and overcome the shortcomings of existing technologies, this invention provides a data security assessment method and system based on clustering processing. Addressing the problem that general data security assessment methods cannot adapt to changes in the operating conditions and environmental conditions of the power system in a timely manner, leading to poor accuracy in security assessments, this solution calculates the correlation between each pair of buses and introduces a graph model to represent the power network, which helps identify critical paths and potential fault points. By introducing dynamic edge weight coefficients and factors related to environmental changes, the accuracy of data security detection is improved. Addressing the problems of low accuracy in detecting fault areas, leading to misjudgments, and the inability to reflect the real-time operating conditions of the power system, resulting in poor security assessment performance, this solution further optimizes the partitioning of the initial anomaly region through adaptive genetic operations. Based on real-time data and the influence of external factors, the weight matrix is ​​dynamically updated to improve the ability to cope with complex and changing power system environments. Finally, based on adaptive projection metrics, the anomalies in the area to be detected are accurately assessed, thereby improving the effectiveness of data security assessment.

[0004] The technical solution adopted by this invention is as follows: The data security assessment method based on clustering processing provided by this invention includes the following steps:

[0005] Step S1: Data Acquisition;

[0006] Step S2: Data coherence detection;

[0007] Step S3: Clustering process;

[0008] Step S4: Data security assessment.

[0009] Further, in step S1, the data acquisition involves collecting power system operation data and performing data conversion, noise reduction, and normalization to obtain a dataset; the power system operation data includes bus operation data, system topology data, meteorological data, equipment aging data, load data, and communication network data.

[0010] Further, in step S2, the data coherence detection involves constructing a correlation coefficient sensitivity matrix; extracting bus operation data from the power system, and calculating the correlation between each pair of buses, expressed as: A graph model is constructed based on the correlation matrix, where the main axis represents the node, and weighted edge betweenness is introduced to construct the edge weights. The edge weights are expressed as: Clustering is performed using the Newman algorithm, calculating edge betweenness numbers, and removing edges with the highest betweenness numbers until the clusters split into k² subclusters. The clustering effect is evaluated, and the normal and abnormal data regions of the power system are output. The evaluation of the clustering effect is expressed as follows: Clustering and correlation thresholds are pre-defined. Clusters with correlation values ​​higher than the correlation threshold and clustering results higher than the clustering threshold are marked as normal regions, while those with correlation values ​​lower than the correlation threshold are marked as preliminary abnormal regions. It is to remove and The dataset; data points correspond to bus data; It is the data of the i-th bus. and the j-th bus data The correlation coefficient; and The standard deviation of the bus data is denoted as A; Cov(·|·) is the partial covariance; Q is the clustering effect; m is the total number of edges in the graph; A ij It is bus correlation; k i and k j It is the sum of the edge weights connecting the busbar to other busbars; It is a set of busbars; It is an indicator function; if the busbars belong to the same cluster, it is 1, otherwise it is 0. and It is the cluster to which the busbar belongs; and These are the edge weight coefficients; s, t, and v are nodes; It is the number of paths that pass through node v among all the shortest paths from node s to node t; It is the number of all shortest paths from node s to node t.

[0011] Furthermore, in step S3, the clustering process specifically includes the following steps:

[0012] Step S31: Embedded clustering. A multi-layer graph model is constructed for the initial anomaly region data. The main line and data are represented as a single vertex set, and different attributes are treated as edges of independent layers. Each layer's graph is represented by an adjacency matrix, and a normalized Laplacian matrix is ​​constructed. ; indicates as: A multi-objective optimization function is constructed by minimizing the impact of data anomalies, including minimizing active power flow interruptions, expressed as: Minimizing reactive power flow interruptions is represented as: Solving for the eigenvectors of the normalized Laplacian matrix, we get: Where A is the adjacency matrix, representing the connection relationship between busbars; D is the degree matrix, a diagonal matrix of node degree values; It is the set of busbars in the initial anomaly region; and These are the active power flow and reactive power flow between the buses, respectively; E is the eigenvector matrix.

[0013] Step S32: Perform k-means clustering on the row vectors of the eigenvector matrix; specifically including:

[0014] Step S321: Initialize the population; use the row vectors of the feature vector matrix E as initial data; randomly generate K1 initial cluster centers C. k ;

[0015] Step S322: Define the objective function, which represents: Define individual fitness values. , is represented as: ; It represents an individual; It is the value of the clustering objective function; It is the eigenvalue of the j1-th dimension of the i1-th data point; d is the center value of the j1-th dimension of the k-th cluster; d is the total number of dimensions; n is the total number of data points; It is an indicator variable for whether a data point belongs to a cluster;

[0016] Step S323: Adaptive genetic operation; using roulette wheel selection to choose the next generation of individuals based on fitness values, represented as: Dynamically adjust crossover probability , is represented as: Dynamically adjust mutation probability , is represented as: ;in, It is the probability that an individual is selected. It is the individual fitness value; and These are the maximum and minimum crossover probabilities, respectively. and These are the minimum and maximum fitness values ​​of the original population, respectively. It is the average fitness value of the population; This is the crossover probability before adjustment;

[0017] Step S324: Clustering optimization; retain the top 30% of individuals with the highest fitness values ​​in the current population to the next generation, without participating in crossover mutation; if the fitness value of the mutated individual is higher than that of the individual before mutation, retain the mutation; otherwise, perform dynamic replacement; dynamic replacement is expressed as: ;in, It is the position after dynamic replacement; It is the current maximum fitness value of the population; M is the amplitude coefficient;

[0018] Step S33: Adjust the weight matrix according to changes in the data stream. Based on real-time data, introduce the influence of external factors and dynamically update the weight matrix, as shown below: ;in, This is the updated inter-bus weight; and These are the weighting coefficients for meteorological factors and equipment aging factors, respectively. and It is the average value of the characteristics of meteorological factors and equipment aging factors corresponding to the busbar;

[0019] Step S34: Data anomaly identification; specifically: calculating the adaptive projection metric of the region to be detected. ; indicates as: ; ; A distance threshold is preset, and areas to be detected with a projection distance higher than the distance threshold are identified as abnormal areas, while those with a projection distance lower than the distance threshold are identified as normal areas; where N is the number of busbars in the area to be detected. tr(·) is the feature vector matrix of the region to be detected; tr(·) is the trace of the matrix; It is a neighborhood set; These are the adaptive coefficients; Var(·) is the variance; It is the weight between buses; x k It is the feature of the k-th data point.

[0020] Furthermore, in step S4, the data security assessment is based on the final clustering results, assessing the data corresponding to normal areas as safe data and the data corresponding to abnormal areas as abnormal data; and issuing early warnings to management personnel.

[0021] The data security assessment system based on clustering processing provided by this invention includes a data acquisition module, a data coherence detection module, a clustering processing module, and a data security assessment module.

[0022] The data acquisition module collects power system operation data and sends the data to the data coherence detection module;

[0023] The data coherence detection module calculates the correlation between power system buses, constructs a graph model, and detects potential data anomaly areas through cluster analysis; then sends the data to the cluster processing module.

[0024] The clustering processing module performs multi-level graph model optimization on the data of the initial abnormal regions, minimizes power flow interruption using the normalized Laplace matrix, and optimizes the clustering results through K-means clustering and genetic algorithms; and sends the data to the data security assessment module.

[0025] The data security assessment module assesses normal and abnormal areas in the power system based on clustering results and provides early warning processing to management personnel.

[0026] The beneficial effects achieved by the present invention using the above solution are as follows:

[0027] (1) In view of the problem that general data security assessment methods cannot adapt to changes in the operating status and environmental conditions of the power system in a timely manner, resulting in poor accuracy of security assessment, this scheme calculates the correlation between each pair of buses and introduces a graph model to represent the power network, which helps to identify critical paths and potential fault points; by introducing dynamic edge weight coefficients and environmental change factors, the accuracy of data security detection is improved.

[0028] (2) In view of the problems that general data security assessment methods have low accuracy in detecting fault areas, which leads to misjudgment; and cannot reflect the operation of the power system in real time, which leads to poor security assessment effect; this scheme further optimizes the partitioning of the initial abnormal area through adaptive genetic operation, dynamically updates the weight matrix based on real-time data and the influence of external factors, and improves the ability to cope with complex and changing power system environment; and accurately assesses the abnormal situation of the area to be detected based on adaptive projection metric, thereby improving the data security assessment effect. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating the data security assessment method based on clustering processing provided by the present invention.

[0030] Figure 2 This is a schematic diagram of the data security assessment system based on clustering processing provided by the present invention;

[0031] Figure 3 This is a flowchart illustrating step S3.

[0032] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0033] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0034] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the system or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0035] Example 1, see Figure 1 The present invention provides a data security assessment method based on clustering processing, which includes the following steps:

[0036] Step S1: Data Acquisition; Acquire power system operation data;

[0037] Step S2: Data coherence detection; By calculating the correlation between power system buses, a graph model is constructed and potential data anomaly areas are detected through cluster analysis;

[0038] Step S3: Clustering; perform multi-level graph model optimization on the data of the initial abnormal regions, minimize power flow interruption using the normalized Laplace matrix, and optimize the clustering results through K-means clustering and genetic algorithm;

[0039] Step S4: Data security assessment; Based on the clustering results, assess the normal and abnormal areas in the power system and provide early warning processing to management personnel.

[0040] Example 2, see Figure 1This embodiment is based on the above embodiment. In step S1, the collected data is processed by data conversion, noise reduction, and normalization to obtain a dataset. The power system operation data includes bus operation data, system topology data, meteorological data, equipment aging data, load data, and communication network data. The bus operation data includes voltage, current, power flow, frequency, and phase. The system topology data includes bus connection relationships, transformer and transmission line connections. The meteorological data includes temperature, humidity, wind speed, and precipitation. The equipment aging data includes equipment service life, equipment operating hours, and failure rate. The load data includes peak load and base load. The communication network data includes latency and packet loss rate.

[0041] Example 3, see Figure 1 This embodiment is based on the above embodiment. In step S2, data coherence detection identifies potential data anomaly regions by detecting low-coherence groups in the dataset; specifically, it involves: constructing a correlation coefficient sensitivity matrix; extracting bus operation data from the power system, calculating the correlation between each pair of buses, and expressing it as: A graph model is constructed based on the correlation matrix, where the main axis represents the node, and weighted edge betweenness is introduced to construct the edge weights. This is used to reflect the critical path of data flow and relationships between buses in a power network, dynamically adapting to environmental changes. The edge weights are represented as follows: Clustering is performed using the Newman algorithm, calculating edge betweenness numbers, and removing edges with the highest betweenness numbers until the clusters split into k² subclusters. The clustering effect is evaluated, and the normal and abnormal data regions of the power system are output. The evaluation of the clustering effect is expressed as follows: Clustering and correlation thresholds are pre-defined. Clusters with correlation values ​​higher than the correlation threshold and clustering results higher than the clustering threshold are marked as normal regions, while those with correlation values ​​lower than the correlation threshold are marked as preliminary abnormal regions. It is to remove and The dataset; data points correspond to bus data; It is the data of the i-th bus. and the j-th bus data The correlation coefficient; and The standard deviation of the bus data is denoted as A; Cov(·|·) is the partial covariance; Q is the clustering effect; m is the total number of edges in the graph; A ij It is bus correlation; k i and k j It is the sum of the edge weights connecting the busbar to other busbars; It is a set of busbars; It is an indicator function; if the busbars belong to the same cluster, it is 1, otherwise it is 0. and It is the cluster to which the busbar belongs; and These are the edge weight coefficients; s, t, and v are nodes; It is the number of paths that pass through node v among all the shortest paths from node s to node t; It is the number of all shortest paths from node s to node t.

[0042] By performing the above operations, this solution addresses the problem that general data security assessment methods cannot adapt to changes in the operating conditions and environmental conditions of the power system in a timely manner, resulting in poor accuracy of security assessments. This solution calculates the correlation between each pair of buses and introduces a graph model to represent the power network, which helps to identify critical paths and potential fault points. Furthermore, by introducing dynamic edge weight coefficients and factors of environmental change, the accuracy of data security detection is improved.

[0043] Example 4, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S3, the clustering process specifically includes the following steps:

[0044] Step S31: Embedded clustering further optimizes data partitioning for the initial anomaly regions to detect potential fault areas in the power system; specifically: a multi-layer graph model is constructed for the initial anomaly region data, representing the bus and data as a single vertex set, with different attributes as edges of independent layers, and each layer's graph is represented by an adjacency matrix, constructing a normalized Laplace matrix. ; indicates as: A multi-objective optimization function is constructed by minimizing the impact of data anomalies, including minimizing active power flow interruptions, expressed as: Minimizing reactive power flow interruptions is represented as: Solving for the eigenvectors of the normalized Laplacian matrix, we get: Where A is the adjacency matrix, representing the connection relationship between busbars; D is the degree matrix, a diagonal matrix of node degree values; It is the set of busbars in the initial anomaly region; and These are the active power flow and reactive power flow between the buses, respectively; E is the eigenvector matrix.

[0045] Step S32: Perform k-means clustering on the row vectors of the eigenvector matrix; specifically including:

[0046] Step S321: Initialize the population; use the row vectors of the feature vector matrix E as initial data; randomly generate K1 initial cluster centers C. k ;

[0047] Step S322: Define the objective function, which represents: Define individual fitness values. , is represented as: ; It represents an individual; It is the value of the clustering objective function; It is the eigenvalue of the j1-th dimension of the i1-th data point; d is the center value of the j1-th dimension of the k-th cluster; d is the total number of dimensions; n is the total number of data points; It is an indicator variable for whether a data point belongs to a cluster;

[0048] Step S323: Adaptive genetic operation; using roulette wheel selection to choose the next generation of individuals based on fitness values, represented as: Dynamically adjust crossover probability , is represented as: Dynamically adjust mutation probability , is represented as: ;in, It is the probability that an individual is selected. It is the individual fitness value; and These are the maximum and minimum crossover probabilities, respectively. and These are the minimum and maximum fitness values ​​of the original population, respectively. It is the average fitness value of the population; This is the crossover probability before adjustment;

[0049] Step S324: Clustering optimization; retain the top 30% of individuals with the highest fitness values ​​in the current population to the next generation, without participating in crossover mutation; if the fitness value of the mutated individual is higher than that of the individual before mutation, retain the mutation; otherwise, perform dynamic replacement; dynamic replacement is expressed as: ;in, It is the position after dynamic replacement; It is the current maximum fitness value of the population; M is the amplitude coefficient;

[0050] Step S33: Adjust the weight matrix according to changes in the data stream. Based on real-time data, introduce the influence of external factors and dynamically update the weight matrix, as shown below: ;in, This is the updated inter-bus weight; and These are the weighting coefficients for meteorological factors and equipment aging factors, respectively. and It is the average value of the characteristics of meteorological factors and equipment aging factors corresponding to the busbar;

[0051] Step S34: Data anomaly identification; utilizing detection and clustering results, combined with real-time data, to achieve security assessment; specifically: calculating the adaptive projection metric of the region to be detected. ; indicates as: ; ; A distance threshold is preset, and areas to be detected with a projection distance higher than the distance threshold are identified as abnormal areas, while those with a projection distance lower than the distance threshold are identified as normal areas; where N is the number of busbars in the area to be detected. tr(·) is the feature vector matrix of the region to be detected; tr(·) is the trace of the matrix; It is a neighborhood set; These are the adaptive coefficients; Var(·) is the variance; It is the weight between buses; x k It is the feature of the k-th data point.

[0052] By performing the above operations, this solution addresses the problems of low accuracy in detecting fault areas, leading to misjudgments, and the inability to reflect the real-time operation of the power system, resulting in poor security assessment results, which are inherent in general data security assessment methods. This solution further optimizes the partitioning of the initial abnormal area through adaptive genetic operations, dynamically updates the weight matrix based on real-time data and the influence of external factors, and improves the ability to cope with complex and changing power system environments. Furthermore, it accurately assesses the anomalies in the area to be detected based on adaptive projection metrics, thereby improving the effectiveness of data security assessment.

[0053] Example 5, see Figure 1 This embodiment is based on the above embodiment. In step S4, the data security assessment is based on the final clustering result. The data corresponding to the normal area is assessed as safe data, and the data corresponding to the abnormal area is assessed as abnormal data. And the management personnel are given an early warning.

[0054] Example 6, see Figure 2 This embodiment is based on the above embodiments. The data security assessment system based on clustering processing provided by the present invention includes a data acquisition module, a data coherence detection module, a clustering processing module and a data security assessment module.

[0055] The data acquisition module collects power system operation data and sends the data to the data coherence detection module;

[0056] The data coherence detection module calculates the correlation between power system buses, constructs a graph model, and detects potential data anomaly areas through cluster analysis; then sends the data to the cluster processing module.

[0057] The clustering processing module performs multi-level graph model optimization on the data of the initial abnormal regions, minimizes power flow interruption using the normalized Laplace matrix, and optimizes the clustering results through K-means clustering and genetic algorithms; and sends the data to the data security assessment module.

[0058] The data security assessment module assesses normal and abnormal areas in the power system based on clustering results and provides early warning processing to management personnel.

[0059] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0060] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0061] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A data security assessment method based on clustering processing, characterized in that: The method includes the following steps: Step S1: Data Acquisition; Acquire power system operation data; Step S2: Data coherence detection; Step S3: Clustering; Step S4: Data security assessment; Step S3 includes step S31: Embedded clustering, constructing a multi-layer graph model for the initial abnormal region data, representing the bus and data as a single vertex set, using different attributes as edges of independent layers, representing the graph of each layer through an adjacency matrix, and constructing a normalized Laplacian matrix. ; indicates as: A multi-objective optimization function is constructed by minimizing the impact of data anomalies, including minimizing active power flow interruptions, expressed as: Minimizing reactive power flow interruptions is represented as: Solving for the eigenvectors of the normalized Laplacian matrix, we get: Where A is the adjacency matrix, representing the connection relationship between busbars; D is the degree matrix, a diagonal matrix of node degree values; It is the set of busbars in the initial anomaly region; and These represent the active power flow and reactive power flow between buses, respectively; E is the eigenvector matrix; i and j are the bus indices; In step S2, the data coherence detection involves constructing a correlation coefficient sensitivity matrix; extracting bus operation data from the power system, and calculating the correlation between each pair of buses, expressed as: A graph model is constructed based on the correlation matrix, where the main axis represents the node, and weighted edge betweenness is introduced to construct the edge weights. The edge weights are expressed as: Clustering is performed using the Newman algorithm, calculating edge betweenness numbers, and removing edges with the highest betweenness numbers until the clusters split into k² subclusters. The clustering effect is evaluated, and the normal and abnormal data regions of the power system are output. The evaluation of the clustering effect is expressed as: Clustering and correlation thresholds are pre-defined. Clusters with correlation values ​​higher than the correlation threshold and clustering results higher than the clustering threshold are marked as normal regions, while those with correlation values ​​lower than the correlation threshold are marked as preliminary abnormal regions. It is to remove and The dataset; data points correspond to bus data; It is the data of the i-th bus. and the j-th bus data The correlation coefficient; and The standard deviation of the bus data is denoted as A; Cov(·|·) is the partial covariance; Q is the clustering effect; m is the total number of edges in the graph; A ij It is bus correlation; k i and k j It is the sum of the edge weights connecting the busbar to other busbars; It is a set of busbars; It is an indicator function; if the busbars belong to the same cluster, it is 1, otherwise it is 0. and It is the cluster to which the busbar belongs; and These are the edge weight coefficients; s, t, and v are nodes; It is the number of paths that pass through node v among all the shortest paths from node s to node t; It is the number of all shortest paths from node s to node t.

2. The data security assessment method based on clustering processing according to claim 1, characterized in that: In step S3, the clustering process specifically includes the following steps: Step S31: Embedded clustering; Step S32: Perform k-means clustering on the row vectors of the feature vector matrix; Step S33: Adjust the weight matrix according to changes in the data stream. Based on real-time data, introduce the influence of external factors and dynamically update the weight matrix, as shown below: ;in, This is the updated inter-bus weight; and These are the weighting coefficients for meteorological factors and equipment aging factors, respectively. and It is the average value of the characteristics of meteorological factors and equipment aging factors corresponding to the busbar; Step S34: Data anomaly identification; specifically: calculating the adaptive projection metric of the region to be detected. ; indicates as: ; ; A distance threshold is preset, and areas to be detected with a projection distance higher than the distance threshold are identified as abnormal areas, while those with a projection distance lower than the distance threshold are identified as normal areas; where N is the number of busbars in the area to be detected. tr(·) is the feature vector matrix of the region to be detected; tr(·) is the trace of the matrix; It is a neighborhood set; These are the adaptive coefficients; Var(·) is the variance; It is the weight between buses; x k It is the feature of the k-th data point.

3. The data security assessment method based on clustering processing according to claim 2, characterized in that: In step S32, the k-means clustering of the row vectors of the feature vector matrix specifically includes: Step S321: Initialize the population; use the row vectors of the feature vector matrix E as initial data; randomly generate K1 initial cluster centers C. k ; Step S322: Define the objective function, which represents: Define individual fitness values , is represented as: ; It represents an individual; It is the value of the clustering objective function; It is the eigenvalue of the j1-th dimension of the i1-th data point; d is the center value of the j1-th dimension of the k-th cluster; d is the total number of dimensions; n is the total number of data points; It is an indicator variable for whether a data point belongs to a cluster; Step S323: Adaptive genetic operation; using roulette wheel selection to choose the next generation of individuals based on fitness values, represented as: Dynamically adjust crossover probability , is represented as: Dynamically adjust mutation probability , is represented as: ;in, It is the probability that an individual is selected. It is the individual fitness value; and These are the maximum and minimum crossover probabilities, respectively. and These are the minimum and maximum fitness values ​​of the original population, respectively. It is the average fitness value of the population; This is the crossover probability before adjustment; Step S324: Clustering optimization; retain the top 30% of individuals with the highest fitness values ​​in the current population to the next generation, without participating in crossover mutation; if the fitness value of the mutated individual is higher than that of the individual before mutation, retain the mutation; otherwise, perform dynamic replacement; dynamic replacement is expressed as: ;in, It is the position after dynamic replacement; It is the current maximum fitness value of the population; M is the amplitude coefficient.

4. The data security assessment method based on clustering processing according to claim 3, characterized in that: In step S1, the collected data is processed by data conversion, noise reduction and normalization to obtain a dataset; the power system operation data includes bus operation data, system topology data, meteorological data, equipment aging data, load data and communication network data.

5. The data security assessment method based on clustering processing according to claim 4, characterized in that: In step S4, the data security assessment is based on the final clustering results, in which data corresponding to normal regions are assessed as safe data and data corresponding to abnormal regions are assessed as abnormal data. And issue early warnings to management personnel.

6. A data security assessment system based on clustering processing, used to implement the data security assessment method based on clustering processing as described in any one of claims 1-5, characterized in that: It includes a data acquisition module, a data coherence detection module, a clustering processing module, and a data security assessment module; The data acquisition module collects power system operation data and sends the data to the data coherence detection module; The data coherence detection module calculates the correlation between power system buses, constructs a graph model, and detects potential data anomaly areas through cluster analysis; then sends the data to the cluster processing module. The clustering processing module performs multi-level graph model optimization on the data of the initial abnormal region, minimizes the power flow interruption by using the normalized Laplace matrix, and optimizes the clustering results by K-means clustering and genetic algorithm; And send the data to the data security assessment module; The data security assessment module assesses normal and abnormal areas in the power system based on clustering results and provides early warning processing to management personnel.

Images