Construction method and device of power terminal interconnection communication model, medium and program product

Abstract

The invention discloses a construction method and device of a power terminal interconnection communication model, a medium and a program product. The method comprises the following steps: acquiring terminal operation data of a plurality of power terminals in a power system, determining operation characteristic data in the terminal operation data, and classifying the terminal operation data according to the operation characteristic data to obtain a plurality of initial data sets; determining a plurality of target data sets according to the data similarity among the power terminal data in the plurality of initial data sets, and determining a target incidence matrix corresponding to the plurality of target data sets, determining the importance degree of the data category corresponding to the target data set in the communication network of the power system according to the target incidence matrix; and a data communication model is constructed according to the importance degree of the data category in the communication network of the power system, so that the operation efficiency, the stability and the response speed of the power system are improved, and the reliability and the intelligent level of the whole communication network are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method, device, medium and program product for constructing an electric power terminal interconnection communication model. Background Art

[0002] In the power system, with the development of smart grids, the number and complexity of power terminal equipment are increasing. How to efficiently manage and optimize the operating data of these terminals has become a key challenge.

[0003] In the related technology, data analysis methods lack a systematic method to identify and classify data with similar operating characteristics, and evaluate the importance of each data category in the network based on these classifications, which makes it difficult to effectively process large and complex terminal operation data, resulting in the inability to timely detect abnormal situations or optimize resource allocation. Summary of the invention

[0004] The present invention provides a method, device, medium and program product for constructing an electric power terminal interconnection communication model to solve the problem in related technologies that it is difficult to effectively process a large amount of complex terminal operation data.

[0005] According to one aspect of the present invention, a method for constructing a power terminal interconnection communication model is provided, comprising:

[0006] Acquire terminal operation data of a plurality of power terminals in a power system, determine operation characteristic data in the terminal operation data, classify the terminal operation data according to the operation characteristic data, and obtain a plurality of initial data sets; wherein the operation characteristic data includes voltage fluctuation data and power loss data;

[0007] Determine multiple target data sets according to the data similarity between the power terminal data in the multiple initial data sets, determine the target association matrix corresponding to the multiple target data sets, and determine the importance of the data category corresponding to the target data set in the communication network of the power system according to the target association matrix;

[0008] A data communication model is constructed according to the importance of the data categories in the communication network of the power system; wherein the data communication model is a digital model for simulating interconnected communication data between a plurality of the power terminals.

[0009] According to another aspect of the present invention, there is provided a device for constructing a power terminal interconnection communication model, comprising:

[0010] A terminal operation data classification module, used to obtain terminal operation data of multiple power terminals in the power system, and determine operation characteristic data in the terminal operation data, and classify the terminal operation data according to the operation characteristic data to obtain multiple initial data sets; wherein the operation characteristic data includes voltage fluctuation data and power loss data;

[0011] An importance determination module, used to determine multiple target data sets according to the data similarity between the power terminal data in the multiple initial data sets, and determine the target association matrix corresponding to the multiple target data sets, and determine the importance of the data category corresponding to the target data set in the communication network of the power system according to the target association matrix;

[0012] A data communication model building module is used to build a data communication model according to the importance of the data category in the communication network of the power system; wherein the data communication model is a digital model for simulating the interconnected communication data between multiple power terminals.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for constructing the power terminal interconnection communication model described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for constructing a power terminal interconnection communication model described in any embodiment of the present invention when executed.

[0018] According to another aspect of the present invention, an embodiment of the present disclosure further provides a computer program product, including a computer program, which, when executed by a processor, implements a method for constructing a power terminal interconnection communication model as described in any one of the embodiments of the present disclosure.

[0019] The technical solution of the embodiment of the present invention is as follows: first, by acquiring terminal operation data of multiple power terminals in the power system and determining operation characteristic data in the terminal operation data, the terminal operation data is classified according to the operation characteristic data to obtain multiple initial data sets. Since the operation characteristic data includes voltage fluctuation data and power loss data, accurate classification of terminal operation status can be achieved; then, multiple target data sets are determined according to the data similarity between the power terminal data in the multiple initial data sets, and the data association matrix corresponding to the multiple target data sets is determined. According to the data association matrix, the importance of the data category corresponding to the target data set in the communication network of the power system is determined, and the importance of the data category can be determined; finally, a data communication model is constructed according to the importance of the data category in the communication network of the power system. Since the data communication model is a digital model for simulating the interconnected communication data between the multiple power terminals, the network design and resource allocation can be optimized, and the problem that it is difficult to effectively process a large amount of complex terminal operation data in the related technology is solved, the operation efficiency, stability and response speed of the power system are improved, and the reliability and intelligence level of the overall communication network are enhanced.

[0020] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0022] Figure 1 is a flow chart of a method for constructing a power terminal interconnection communication model provided according to Embodiment 1 of the present invention;

[0023] Figure 2 This is a schematic diagram of the structure of a device for constructing an interconnected communication model of power terminals provided according to a second embodiment of the present invention;

[0024] Figure 3 It is a structural schematic diagram of an electronic device for implementing a method for constructing a power terminal interconnection communication model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme 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 described embodiments 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 should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0028] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0029] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0030] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0031] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0032] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0033] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0034] Embodiment 1

[0035] Figure 1 A flowchart of a method for constructing an interconnected communication model of electric power terminals is provided for the first embodiment of the present invention. This embodiment can be applied to the situation of optimizing the communication network of the electric power system to improve the efficiency of data transmission. The method can be executed by a device for constructing an interconnected communication model of electric power terminals. The device for constructing the interconnected communication model of electric power terminals can be implemented in the form of hardware and / or software, and optionally, can be implemented by an electronic device, which can be a mobile terminal, a PC or a server, etc.

[0036] like Figure 1 As shown, the method may specifically include:

[0037] S110, acquiring terminal operation data of a plurality of power terminals in a power system, and determining operation characteristic data in the terminal operation data, and classifying the terminal operation data according to the operation characteristic data to obtain a plurality of initial data sets; wherein the operation characteristic data includes voltage fluctuation data and power loss data;

[0038] Among them, the terminal operation data can be understood as data collected from various terminal devices (such as smart meters or distribution automation terminals, etc.) of the power system during operation, which can be used to detect and analyze the operating status of the power system. The terminal operation data may include but is not limited to at least one of electrical parameters such as voltage, current, power and frequency. The operation characteristic data can be understood as data extracted from the terminal operation data that can reflect the operating characteristics of the power terminal, which can be used to analyze different types of operating conditions or anomalies, and provide a data basis for the optimization and fault diagnosis of the power system. The operation characteristic data includes but is not limited to at least one of data such as voltage fluctuation data and power loss data. The voltage fluctuation data can be understood as data that records the change of voltage level in the power system over time, which is used to detect the stability of the power system and identify voltage instability that may cause damage to power equipment or performance degradation. The voltage fluctuation data may be caused by load changes, power supply problems or other factors. The power loss data can be understood as data on energy loss during power transmission and distribution, which is dissipated in the form of heat due to the thermal effect of resistors, and can be used to determine energy-inefficient nodes in the power network. Exemplarily, the power quality can be improved by analyzing the voltage fluctuation data, or the energy efficiency can be improved by evaluating the power loss data. The initial data set may be understood as a data set formed after preliminary classification of the original terminal operation data based on the operation characteristic data, which may be used to simplify the data processing flow so that subsequent analysis may be carried out according to the category.

[0039] On the basis of the above scheme, optionally, the terminal operation data includes at least operation voltage data; the operation characteristic data in the terminal operation data is determined, and the terminal operation data is classified according to the operation characteristic data to obtain multiple initial data sets, including: determining the voltage deviation value between the operation voltage data of multiple power terminals and the rated voltage, determining the power terminal whose voltage deviation value exceeds a preset deviation threshold as an abnormal terminal, and constructing a reference data set based on the terminal operation data corresponding to the abnormal terminal; obtaining the load rate data of the abnormal terminal, and fitting the load rate data of multiple abnormal terminals by the least squares method to obtain an electric energy loss distribution curve; by integrating the electric energy loss distribution curve, obtaining the accumulated electric energy loss value in the multiple abnormal terminals, and dividing the reference data set into multiple initial data sets according to the accumulated electric energy loss value.

[0040] Among them, the voltage deviation value can be understood as the difference between the operating voltage data of the power terminal and the rated voltage, which can be used to evaluate the voltage stability of the power terminal. The preset deviation threshold can be understood as a pre-set voltage deviation range, which is used to determine whether the voltage deviation value is within an acceptable range. In the case where the voltage deviation of the power terminal exceeds the preset deviation threshold, it is marked as an abnormal terminal. The abnormal terminal can be understood as a power terminal whose voltage deviation value exceeds the preset deviation threshold. The reference data set can be understood as a data set consisting of terminal operation data corresponding to the abnormal terminal. The load rate data can be understood as the ratio of the actual power consumption of the power terminal to its maximum possible power consumption (i.e., rated power), which can reflect the working intensity or load condition of the power terminal. The least squares fitting can be understood as a method for finding the function form of the best fit given data points, which can be achieved by minimizing the sum of squared errors. The power loss distribution curve can be understood as a graphical representation showing the change of power loss with load rate. The accumulated power loss value can be understood as the total power loss of multiple abnormal terminals.

[0041] An optional implementation method is to collect voltage fluctuation data from the operation data with an interval of 300 seconds from the power terminal, construct a fluctuation data set based on the voltage deviation value of each data point in the voltage fluctuation data and the rated voltage of 220 volts, and use a convolutional neural network to mark the data points whose voltage deviation value exceeds the limit of 4% and enter them into the voltage fluctuation characteristic data set. According to the voltage fluctuation characteristic data set, current and power records are extracted from the terminal operation data, and the ratio of the current power product and the rated power at the corresponding moment of each marked data point is calculated to obtain the load rate data, and the load rate data is fitted by the least squares method to obtain the power loss distribution curve. For the power loss distribution curve, the cumulative value of power loss in the 1800 second interval before and after each marked data point is calculated by integral operation, and the data points in the voltage fluctuation characteristic data set are sorted and grouped according to the cumulative loss value, and multiple initial data sets are obtained according to the cumulative loss value. For multiple initial data sets, the standard deviation of the voltage fluctuation data points in each subset with respect to the adjacent four data points is calculated, the peak point is extracted from the standard deviation sequence, and the decision tree is used to cluster the occurrence pattern of the peak point to obtain the terminal operation characteristic index reflecting the correlation between the voltage fluctuation characteristics of the power terminal and the power loss. The collection of voltage fluctuation records involves the continuous monitoring of the voltage values ​​in the actual operation of the power terminal. When the voltage deviates from the rated value of 220 volts, a fluctuation record point is generated. The voltage value is recorded at a fixed sampling interval of 300 seconds. If the voltage is measured at a certain moment as 229 volts, the fluctuation deviation is 4.1%, which exceeds the preset threshold of 4%. This data point is marked and written into the voltage fluctuation characteristic data set. The current and power records reflect the actual load conditions of the power terminal. For the marked fluctuation points, the ratio of the actual power to the rated power at that moment is calculated to obtain the load rate. For example, when the rated power of a terminal is 2000 watts and the actual power is 1800 watts, the load rate is 0.9. The loss distribution curve is drawn by fitting these load rate data points by the least squares method. The trend of the curve reflects the law of power loss changing with load. Based on the loss distribution curve, the cumulative loss value in the 1800-second time window before and after each fluctuation point is calculated. The size of the cumulative value represents the degree of power loss in this period. For example, the cumulative loss value of a fluctuation point in the corresponding period is 0.85 kWh. According to the size of the loss value, all fluctuation points are divided into three subsets of low loss, medium loss, and high loss to form data classifications in different operating states. The standard deviation of the fluctuation points in the three subsets is calculated with respect to the adjacent four points. The standard deviation reflects the severity of voltage fluctuations. A larger standard deviation value corresponds to a rapid change in voltage. For example, the standard deviation of a fluctuation point and the adjacent points is 8.5 volts, indicating that the voltage changes violently at this location. The decision tree is used to identify the regularity of the peak points in the standard deviation sequence, and the corresponding relationship between the fluctuation characteristics and the degree of loss is established.In actual operation, power terminals may experience different operating conditions such as no-load, light load, and rated load. Load changes cause voltage fluctuations and loss changes. Larger voltage fluctuations often correspond to higher losses. For example, when no-load suddenly changes to rated load, the voltage drop can reach 12 volts, and the power loss rises rapidly. By analyzing the correlation between this fluctuation and loss, abnormal operating conditions can be identified and potential fault hazards can be discovered. In the operation of the distribution network, instantaneous high-power fluctuations caused by the start and stop of terminal equipment can cause voltage drops. For example, when a motor starts, a shock of 5 times the rated current is generated, resulting in a 3% to 5% reduction in the voltage of the adjacent terminal and additional losses. In this case, the fluctuation characteristics are significantly correlated with the loss index. Identifying this type of fluctuation loss law is helpful in evaluating the terminal operation status.

[0042] By adopting this technical solution, the operating voltage data of the power terminal is analyzed, the voltage deviation value is determined and the abnormal terminal is identified, a reference data set is constructed, the load rate data of the abnormal terminal is fitted using the least squares method, an energy loss distribution curve is generated, and the accumulated energy loss value is calculated by integral operation. The reference data set is divided into multiple initial data sets, which can accurately locate terminals with abnormal voltage and high energy loss, help to timely discover and solve potential problems, optimize the operating efficiency of the power terminal, improve the overall power supply quality and reliability, and ensure the stable and efficient operation of the power terminal.

[0043] S120. Determine multiple target data sets based on the data similarity between the power terminal data in the multiple initial data sets, and determine the target association matrix corresponding to the multiple target data sets, and determine the importance of the data categories corresponding to the target data sets in the communication network of the power system based on the target association matrix.

[0044] Among them, the data similarity can be understood as an indicator to measure the similarity between the operating data of two or more power terminals, and can be used to identify data sets with similar characteristics. The data similarity can be determined by Euclidean distance, cosine similarity or Manhattan distance. The target data set can be understood as a data set formed by clustering the similarities between the power terminal data in the initial data set. The target association matrix can be understood as a matrix structure representing the relationship between target data sets. The target association matrix is ​​used to describe the association information between data categories corresponding to multiple target data sets. The importance can be understood as the relative importance of the target data set determined according to the data association matrix in the power system communication network.

[0045] On the basis of the above scheme, optionally, the method of determining multiple target data sets according to the data similarity between the power terminal data in the multiple initial data sets, and determining the target association matrices corresponding to the multiple target data sets, includes: determining the Pearson correlation coefficient between the power terminal data in the initial data set, obtaining the cluster center coordinates through single nearest neighbor clustering according to the Pearson correlation coefficient, calculating the Manhattan distance for the power terminal data using the cluster center coordinates, and determining the intermediate data set according to the Manhattan distance; determining the cosine similarity matrix according to the multiple power terminal data in the intermediate data set, and obtaining the target data set by calculating the Mahalanobis distance for the power terminal data whose cosine similarity is higher than or equal to a preset preselected threshold; using a support vector machine to respectively determine the feature space distribution position corresponding to each of the target data sets, constructing a Euclidean distance matrix through the multiple feature space positions, and determining the target association matrices corresponding to the multiple target data sets according to the Euclidean distance matrix.

[0046] Among them, the Pearson correlation coefficient can be understood as a statistical indicator that measures the strength and direction of the linear relationship between two power terminal data, and the value range is from -1 to +1. The Pearson correlation coefficient can be used to quantify the degree of linear association between power terminal data in the initial data set, and a positive value indicates a positive correlation, a negative value indicates a negative correlation, and a value close to 0 means that there is no obvious linear relationship. The single nearest neighbor clustering can be understood as a hierarchical clustering algorithm that forms a larger cluster by connecting the closest data points or clusters. The Manhattan distance can be used to calculate the distance between the power terminal data and the cluster center coordinates to determine which data should belong to which cluster, thereby forming an intermediate data set. The cosine similarity matrix can be understood as a measure of the angular difference between multiple power terminal data, which is used to evaluate the directional similarity between power terminal data in the intermediate data set, and the value range is [-1,1], where 1 indicates exactly the same direction, -1 indicates exactly the opposite direction, and 0 indicates orthogonal (no correlation). The preset pre-selected threshold can be understood as a pre-set cosine similarity threshold, which is used to filter out qualified data points. In this embodiment, the preset preselected threshold value can be used to filter out the power terminal data whose cosine similarity is higher than or equal to the threshold value, ensuring that the data in the target data set finally obtained has a high degree of similarity. The Mahalanobis distance can be understood as a distance measurement method for processing the covariance structure of data distribution, which is used to further refine the target data set to ensure that the selected power terminal data are not only similar in direction (determined by cosine similarity), but also have similar distribution characteristics in multidimensional space. The support vector machine can be understood as a supervised learning model applied to classification and regression analysis. In this embodiment, the support vector machine is used to determine the feature space distribution position corresponding to each target data set, that is, to identify the optimal separation boundary of each data set in multidimensional space. The feature space distribution position can be understood as the position of the target data set in the feature space, and the position of each data point reflects the characteristics of the target data set. The Euclidean distance matrix can be understood as a matrix reflecting the Euclidean distance (i.e., straight-line distance) between multiple points.

[0047] An optional implementation method is to centralize the data and calculate the Pearson correlation coefficient for the initially classified power terminal data, set the data similarity threshold of 0.8 according to the correlation coefficient, obtain the cluster center coordinates through single nearest neighbor clustering, calculate the Manhattan distance between the attribute characteristic values ​​of the power terminal data and the cluster center point, and obtain the intermediate data set from the Manhattan distance distribution of the cluster center point. According to the Manhattan distance of each data point in the intermediate data set, calculate the cosine similarity matrix of the data point, set the similarity judgment threshold of 0.85, mark the data points whose cosine similarity is higher than the preset pre-selected threshold, calculate the Mahalanobis distance based on the marked data to obtain the spatial distribution range of the data point, and divide the target data set from the spatial distribution range. For the data points in the target data set, use a support vector machine to calculate the feature space distribution position of the data point in the feature space, construct the Euclidean distance matrix of the data point from the feature space distribution position, calculate the correlation coefficient between the data points based on the Euclidean distance matrix, extract the data category identifier from the correlation coefficient matrix and obtain the inter-class association matrix. According to the inter-class association matrix, each association value in the matrix is ​​subtracted from the row mean and then divided by the row standard deviation, the main diagonal elements are extracted from the standardized matrix, the eigenvalue decomposition is calculated based on the main diagonal elements, and the normalized association matrix is ​​obtained by eigenvalue decomposition. The power terminal data contains multiple attribute features such as voltage, current, and power. After data centering, the Pearson correlation coefficient is calculated to measure the strength of the association between attributes. For example, the correlation coefficient between voltage and current is 0.92, indicating that the two have a strong positive correlation. Based on this, the similarity threshold of 0.8 is set to distinguish between strong and weak associated data points. Single nearest neighbor clustering is classified by calculating the Manhattan distance between data points. In actual operation, the operating data of a distribution transformer shows that the voltage deviation is 3% and the power factor is 0.92 when the load rate is 80%. The Manhattan distance between this set of data and the cluster center is 0.15, which is less than the judgment threshold of 0.2 and is classified into the same category. Cosine similarity reflects the angle relationship between data points in high-dimensional space. The cosine similarity of adjacent moments under normal operation is calculated to be 0.95, which is higher than the judgment threshold of 0.85, indicating that the operation state is stable. When a fault occurs, the cosine similarity drops to 0.75, which is lower than the threshold and is marked as an abnormal data point. The Mahalanobis distance takes into account the covariance structure of the data. In the distribution network operation data, the voltage and current data of multiple measuring points on a feeder are calculated to have a Mahalanobis distance of 2.8, which is greater than the typical value of 1.5 under normal operation, reflecting that the load distribution of the feeder is uneven and needs to be classified and marked. The support vector machine constructs a classification hyperplane in the feature space. When processing the distribution network operation data, the three features of voltage, current and power are used as coordinate axes to map the data to a three-dimensional space, obtain the distribution area of ​​different operation states, calculate the distance from the data point to the classification boundary, and determine the category to which the data belongs.Eigenvalue decomposition is used to extract the main features of the association matrix. In practical applications, the association matrix of the 10 kV bus voltage and feeder load data of a substation has a maximum eigenvalue of 3.2 after normalization, and the corresponding eigenvector reflects the main association pattern between voltage and load. In the distribution automation system, when classifying the operating data of multiple measuring points, the association matrix is ​​constructed by calculating the Euclidean distance between the measuring points. For example, the voltage data of the three measuring points A, B, and C of the substation are calculated two by two to obtain the distance value to form a 3x3 matrix. After normalization, the normalized association strength is obtained to evaluate the degree of association between the measuring points. The correlation analysis of the operating data of equipment at different voltage levels shows that the 110 kV bus voltage fluctuation is significantly correlated with the 10 kV feeder load change. The calculated correlation coefficient is 0.88. The voltage transfer characteristics are characterized by the association matrix, providing a basis for the analysis of the power grid operation status.

[0048] This technical solution is adopted to calculate the Pearson correlation coefficient between the power terminal data and use single nearest neighbor clustering to determine the cluster center, combine the Manhattan distance to screen the intermediate data set, further use the cosine similarity matrix and Mahalanobis distance to accurately locate the target data set, and use the support vector machine to determine the feature space distribution position to construct the Euclidean distance matrix to generate the target association matrix. It can efficiently identify and classify power terminals with similar operating characteristics, improve data analysis accuracy, help discover potential abnormal patterns, optimize resource allocation of power terminals, and enhance system stability and response speed.

[0049] On the basis of the above scheme, optionally, determining the importance of the data category corresponding to the target data set in the communication network of the power system according to the target association matrix includes: screening out target data groups whose mutual information is greater than a preset information threshold from the target association matrix, and determining association rules between the target data groups, constructing a category association rule network according to the association rules, determining adjacent data categories according to the category association rule network, and determining the mutual information entropy between adjacent data categories, and determining category connection weights according to the mutual information entropy; using a breadth-first search algorithm to identify the shortest transmission path between the data categories, and determining the category connection weights according to the mutual information entropy; Determine the degree centrality value and the betweenness centrality value of the data category based on the shortest transmission path between the data category and the data category, and determine the centrality value based on the degree centrality value and the betweenness centrality value; for each of the data categories, determine the fluctuation standard deviation of the category connection weight of the data category within a continuous preset number of cycles, obtain the total amount of data transmission of the data category with the fluctuation standard deviation less than a preset fluctuation threshold, determine the clustering coefficient of the data category based on the total amount of data transmission of the data category, multiply the clustering coefficient by the centrality value, and obtain the importance of the data category in the communication network of the power system.

[0050] Among them, the mutual information can be understood as a statistical indicator to measure the amount of shared information between two target data groups, which is used to evaluate the correlation between target data groups. The target data groups whose mutual information is greater than the preset information threshold indicate that there is a significant correlation or dependency between them. The preset information threshold can be understood as a pre-set standard or limit for screening the size of the mutual information. The association rule can be understood as a rule describing the interaction between target data groups. The category association rule network can be understood as a network structure constructed based on association rules, where nodes represent different data categories and edges represent association rules between categories, which are used to analyze the relationship network between different data categories. The mutual information entropy can be understood as a measure of shared information between two adjacent data categories, which is used to quantify the strength of association between adjacent data categories. The category connection weight can be understood as a numerical value for determining the strength of association between adjacent data categories, which is used to define the weight of the edge in the category association rule network. The breadth-first search algorithm can be understood as an algorithm for traversing or searching a graph structure, starting from the root node and expanding outward layer by layer until all reachable nodes are found. In this embodiment, the breadth-first search algorithm can be used to identify the shortest transmission path between data categories to ensure that information can be transmitted most efficiently in the communication network of the power terminal. The degree centrality value can be understood as the number of direct connections between a data category and other data categories, reflecting the degree of direct connection of the data category in the network. A higher degree centrality means that the data category has more direct connections in the network. The betweenness centrality value can be understood as the frequency of a node appearing in the shortest transmission path. The fluctuation standard deviation can be understood as a measure of the difference between the category connection weight and its average value, which is used to reflect the stability of the category connection weight over time. The total amount of data transmission can be understood as the sum of the amount of data transmitted through a certain data category within a preset time period. The clustering coefficient can be understood as a parameter that describes the degree of interconnection between adjacent nodes of a node, reflecting the tightness of the network around the node.

[0051] An optional implementation method is to use an Apriori association rule miner to extract association rules from data with a mutual information amount greater than the threshold value for data categories in a standardized association matrix, and obtain a category node association rule network; based on the category node association rule network, the node connection weights are obtained by calculating the mutual information entropy between adjacent nodes, and a breadth-first search algorithm is used to identify the shortest transmission path between the nodes; for the shortest transmission path, the importance of nodes is ranked by calculating the node degree centrality value and the betweenness centrality value to obtain a node centrality data set; based on the node centrality data set, if the standard deviation of the node connection weight fluctuation is less than the threshold value, the total amount of data transmission is obtained from a subset of stable connection nodes, and the node importance ranking is obtained by calculating the product of the clustering coefficient and the centrality value.

[0052] Specifically, for the association value of each pair of data categories in the data association matrix, the mutual information is obtained by calculating the ratio of the probability of the two categories appearing at the same time to the probability of their appearing alone. The mutual information threshold of 0.6 is set to judge the significance of the association. The Apriori association rule miner is used to extract the transmission rules with support greater than 0.3 and confidence greater than 0.7 from the significant association data to construct a category node association rule network. According to the category node association rule network, the mutual information entropy is calculated for the data transmission volume between each pair of adjacent nodes, the category connection weight is extracted from the entropy value distribution, the breadth-first search algorithm is used to identify the shortest transmission path between the category connection weights, the path connectivity is calculated by multiplying the category connection weights on the transmission path, and the node connection distribution function is constructed from the connectivity value. For the node connection distribution function, the number of direct connections of each node is calculated to obtain the degree centrality value, the betweenness centrality value is calculated for the frequency of the node appearing in the shortest path, the node importance is sorted based on the centrality value, a topological relationship directed graph is constructed from the sorting result, and the node connection strength in the directed graph is normalized to obtain the node centrality data set. According to the node centrality data set, the standard deviation of the connection weight fluctuation within the continuous monitoring cycle is calculated for each node, and the standard deviation threshold of 0.15 is set to determine the connection stability. The total amount of data transmission is extracted from the node subset with stable connection, and the node clustering coefficient is calculated by the total amount of data transmission. The product of the clustering coefficient and the centrality value is used to obtain the importance ranking of the node in the communication network. In the classification of power terminal data, the mutual information reflects the degree of correlation between different categories of data. For example, in the transformer operation data, the mutual information between oil temperature and load rate is 0.72, which is higher than the threshold of 0.6, indicating that the two are significantly correlated. The rules with support of 0.35 and confidence of 0.82 are obtained through Apriori mining, indicating that there is an obvious transmission relationship between the increase in load rate and the increase in oil temperature. The data transmission between data categories is reflected in the mutual influence of the operating parameters of power equipment. Taking the distribution transformer as an example, the mutual information entropy of the primary side voltage and the secondary side voltage is calculated to be 1.25, indicating that the information transmission is relatively sufficient. The shortest path is identified through breadth-first search, including two intermediate nodes: transformer body impedance and tap switch position, and the path connectivity is 0.78. In practical applications, degree centrality and betweenness centrality are respectively reflected in the number of direct connections and information transmission capabilities of the equipment. For example, a 110 kV busbar node connects 4 outgoing lines, with a degree centrality value of 4. It appears 3 times in the 5 shortest monitoring paths, with a betweenness centrality value of 0.6. The node centrality after normalization is 0.85, ranking in the top 10%. Connection stability is reflected in the degree of fluctuation of equipment operating parameters. During the monitoring period, the standard deviation of the connection weight between a 10 kV feeder and a transformer in the substation area is 0.12, which is lower than the threshold of 0.15 and is determined to be a stable connection. The total monthly data transmission of this feeder reaches 86,400, and the node clustering coefficient is calculated to be 0.72.In the distribution automation terminal, the bus voltage node often has a high centrality and maintains a stable connection with multiple feeder load nodes. For example, the bus voltage of a substation and 8 feeder loads form a star topology. The clustering coefficient of the bus node is 0.95, the centrality value is 0.92, and the comprehensive importance ranks first. The associated network of the power equipment group often presents multi-level characteristics. Taking the distribution line as an example, the cable segmented ring network cabinet is associated with multiple branch boxes, and the branch boxes are connected to multiple user transformers to form a tree topology. The mutual information between nodes gradually decreases with the increase of the level. The mutual information of the first-level branch nodes is generally above 0.7, and the second-level branch drops to about 0.5. The substation bus is the core node, which is electrically connected to each outgoing line through circuit breakers, transformers and other equipment to form a radial topology. The circuit breaker switch state, current transformer measurement value, etc. participate in the shortest path calculation as intermediate nodes. In typical cases, the path connectivity is between 0.75 and 0.85, reflecting the good information transmission characteristics between substation equipment. From the perspective of power supply reliability, the clustering coefficient of feeder terminal nodes is low but the connection stability requirements are high. For example, the average clustering coefficient of transformer nodes in the substation is 0.45, but the standard deviation of connection weights is generally lower than 0.1, which meets the requirements of stable power supply and belongs to the terminal stable node in the importance ranking.

[0053] By adopting this technical solution, the mutual information in the target association matrix is ​​analyzed, the relevant target data groups are screened out, and a category association rule network is constructed to determine the mutual information entropy and category connection weights of adjacent data categories. Then, the breadth-first search algorithm is used to identify the shortest transmission path, and the degree centrality and betweenness centrality of the data category are calculated to obtain the centrality value, evaluate the standard deviation of the fluctuation of the category connection weight in continuous periods, determine the clustering coefficient in combination with the total amount of data transmission, and multiply the clustering coefficient by the centrality value. Finally, the importance of the data category is quantified, which can accurately identify the key data categories in the power system communication network, optimize resource allocation, and improve network stability and transmission efficiency.

[0054] S130. Construct a data communication model according to the importance of the data category in the communication network of the power system; wherein the data communication model is a digital model for simulating interconnected communication data between multiple power terminals.

[0055] Among them, the data communication model can be understood as a digital model that simulates the actual interconnection communication process between power terminals, aiming to simulate and optimize the performance of the communication network, to predict communication problems that may occur, and to test different network configuration schemes, thereby improving the operation and maintenance level of the power terminal communication network.

[0056] On the basis of the above scheme, optionally, constructing a data communication model according to the importance of the data categories in the communication network of the power system includes: constructing a data category importance matrix according to the importance of the data categories in the communication network of the power system, and using Euclidean distance metric to calculate the connection strength values ​​between adjacent data categories for the data category importance matrix, and obtaining a normalized connectivity matrix by dividing the connection strength values ​​by the maximum communication bandwidth; clustering the data categories according to the normalized connectivity matrix using a spectral clustering method of a symmetric regularized Laplace matrix to obtain a target data cluster, and determining the data within the target data cluster. The Manhattan distance between the data categories is determined based on the Manhattan distance. The total amount of data transmission of the data categories is determined based on the spatial vector difference and the communication frequency between the data categories, and the initial importance score of each data category is determined by the total amount of data transmission, the spatial vector difference and the Gaussian process regressor. The data transmission delay between the data categories is determined, and the communication bandwidth value is determined based on the data transmission delay. The communication timing score of the data category is determined based on the initial importance score and the communication bandwidth value, and a data communication model is constructed based on the communication timing score of the data category.

[0057] Among them, the data category importance matrix can be understood as a matrix representing the importance of different data categories in the power system communication network, which is used to quantify and compare the importance of each data category and provide a basis for subsequent analysis. The Euclidean distance metric can be understood as a method for calculating the straight-line distance between two points, which is used to calculate the connection strength value between adjacent data categories, reflecting the degree of proximity or similarity between them. The connection strength value can be understood as a numerical value that measures the degree of association between two data categories, which is used to evaluate the relationship strength between data categories. The normalized connectivity matrix can be understood as a matrix obtained by normalizing the connection strength value, which can ensure that the connection strength value is within a preset range, which is convenient for further analysis. The symmetric regularized Laplace matrix can be understood as a mathematical tool for processing graph structure data in graph theory and machine learning, and is suitable for spectral clustering methods. The spectral clustering method can be understood as a clustering algorithm based on graph theory, which is used to extract information from the normalized connectivity matrix, cluster data categories, and form target data clusters. The Manhattan distance can be understood as calculating the distance between data categories in the target data cluster, which helps to determine the distribution of data categories in the feature space. The spatial vector difference can be understood as representing the difference between two data categories in a multidimensional data space, and can be used to quantify the relative position changes between data categories in combination with the Manhattan distance. The Gaussian process regressor can be understood as a non-parametric Bayesian method for regression and classification tasks that can provide a probability distribution of predicted values. In this embodiment, the Gaussian process regressor is used to determine the initial importance score of each data category based on the spatial vector difference and the communication frequency. The initial importance score can be understood as the importance score of the data category determined based on a combination of multiple factors. The data transmission delay can be understood as the time required for the data in each pair of data categories to go from the transmitter to the receiver. The communication bandwidth value can be understood as referring to the maximum data transmission rate of the communication channel. The communication timing score can be understood as a score after comprehensively considering the data transmission delay and the communication bandwidth, reflecting the time scheduling priority of the data category in the communication network.

[0058] An optional implementation method is to use the Euclidean distance metric to calculate the connection strength between adjacent data categories for the standardized data category importance matrix, divide the connection strength value by the maximum communication bandwidth of the category data to obtain the normalized connectivity, use the spectral clustering of the symmetric regularized Laplace matrix to divide the data categories with normalized connectivity greater than 0.75 into the same data cluster, and determine the size of the data cluster by calculating the monthly average communication frequency between the data categories in the cluster. For the data categories in each data cluster, the Manhattan distance is used to calculate the spatial vector difference between the data category and the cluster median coordinates, and the total amount of data transmission is calculated for each data category by weighted summation according to the communication frequency. The initial importance score of each data category is extracted from the total amount of data transmission and the spatial vector difference based on the Gaussian process regressor. According to the initial importance score, the communication transmission delay is recorded between each pair of data categories in the data cluster, and the communication bandwidth is obtained by dividing the transmission data packet size by the average transmission delay. The real-time response index is obtained from the product of the communication bandwidth and the initial importance score, and the data category priority queue is generated according to the real-time response index in descending order. For the data category priority queue, the data packet transmission time interval between data categories is extracted from the communication record, the periodic transmission probability is calculated based on the time interval sequence, and the communication timing score is obtained by weighting the priority through the periodic transmission probability. The data communication model of the power terminal data category is constructed from the communication timing score. The classification of power terminal data categories involves multiple levels of features. Taking the distribution automation terminal as an example, the connection strength between voltage data, current data, and power data is usually above 0.8. After normalization, these three types of data are divided into the same data cluster, and the monthly communication frequency reaches 720 times, reflecting the close correlation of basic measurement data. At the substation level, the spatial vector difference between bus voltage data and feeder power data is small, the Manhattan distance between the two is 0.25, and the monthly data transmission volume exceeds 43,200. The importance scores calculated by Gaussian process regression are 0.95 and 0.92, respectively, reflecting the core position of these two types of data in substation operation. The data transmission characteristics have an important impact on the communication order. Taking the switchgear status data as an example, the data packet size is 256 bytes, the average transmission delay is 0.1 seconds, and the communication bandwidth is calculated to be 2.56 megabits per second. Multiplying it with the importance score of 0.88, the real-time response index is 2.25, ranking at the top of the priority queue. Periodic transmission is reflected in the regularity of data collection. The bus voltage data is collected every 5 minutes, the transmission time interval is stable, and the probability of periodic transmission reaches 0.98. The switch status data is uploaded on demand, and the probability of periodic transmission is only 0.15. This difference leads to a significant distinction in the communication timing score.In the feeder terminal unit, short-circuit fault data has the highest communication priority. The data packet delay is usually controlled within 50 milliseconds, the communication bandwidth is required to reach 10 megabits per second, and the importance score is 0.96, which is at the top of the data category priority queue to ensure that the fault information is transmitted in the first time. The telemetry data of the distribution terminal equipment shows obvious cluster characteristics. The connection strength of the four types of data, voltage, current, active power, and reactive power, exceeds 0.85. The communication frequency between them is maintained at once every 5 minutes, and the total amount of data transmission is similar, all around 86,400 per day. The data categories of equipment of different voltage levels also present a hierarchical cluster structure. The importance of the 110 kV equipment data cluster is generally higher than that of the 10 kV equipment, which is reflected in the communication bandwidth configuration. The communication bandwidth of the 110 kV circuit breaker status data is 5 megabits per second, while the 10 kV switch status data is 2 megabits per second. Although the transmission frequency of operating status data and alarm data is low, their importance scores are maintained above 0.85 because they are related to the safe operation of the equipment. After being weighted by the periodic transmission probability, they are given a higher priority in the data communication model to ensure a quick response when the status is abnormal.

[0059] By adopting this technical solution, by constructing a data category importance matrix and calculating the connection strength values ​​of adjacent data categories, the target data cluster is obtained by normalization processing and spectral clustering method. Then, the Manhattan distance between data categories in the cluster is analyzed to determine the spatial vector difference, and the total data transmission amount is calculated in combination with the communication frequency. The Gaussian process regressor is used to evaluate the initial importance score of each data category, and the communication timing score is determined by comprehensively considering the data transmission delay and communication bandwidth value. Finally, a data communication model is constructed, which can accurately quantify the importance and time scheduling priority of each data category in the network, optimize resource allocation, improve the efficiency and reliability of the power system communication network, and ensure the efficient transmission of terminal operation data.

[0060] On the basis of the above scheme, optionally, after constructing the data communication model according to the importance of the data category in the communication network of the power system, it also includes: determining the communication bandwidth occupancy rate corresponding to the data category through the transmission data packet size and sampling time interval corresponding to the data category, and determining the link quality score of the communication link corresponding to the data category according to the communication bandwidth occupancy rate; for the communication link whose link quality score is greater than a preset score threshold, performing a bandwidth test within a fixed detection period, and determining the maximum transmission capacity of the channel according to the round-trip delay and the number of packet losses of the test data packet in the bandwidth test; determining the cross-cluster distance value between the target data clusters according to the maximum transmission capacity of the channel, determining the cluster cohesion value from the average connectivity of the data category in the target data cluster, determining the channel bandwidth allocation ratio through the cross-cluster distance value and the cluster cohesion value, and establishing a direct communication channel between the target data clusters in the data communication model based on the channel bandwidth allocation ratio.

[0061] Among them, the communication bandwidth occupancy rate can be understood as the proportion of communication bandwidth used by the data category within the preset time length, which is used to quantify the degree of demand for communication resources by the data category. The link quality score can be understood as an indicator for measuring the performance of the communication link, which is used to reflect the actual transmission quality and reliability of the communication link. A higher link quality score indicates a more stable communication environment. The preset score threshold can be understood as a pre-set link quality score standard, which is used to screen out communication links with good performance. The fixed detection period can be understood as a predetermined time period, during which detection activities such as bandwidth testing are regularly performed to detect the trend or abnormal situation of bandwidth changes over time. The bandwidth test can be understood as a method for measuring the actual performance of the communication link, including but not limited to sending a test data packet and recording its round-trip delay and the number of packet losses. The maximum transmission capacity of the channel can be understood as the maximum data transmission rate that the communication link can support under optimal conditions, which can be used as a basis for evaluating the transmission capacity of the link. The cross-cluster distance value can be understood as the average distance or difference measurement between different target data clusters, which is used to quantify the relative position relationship between target data clusters and affect the efficiency of cross-cluster communication. The cluster cohesion value can be understood as an indicator describing the degree of connection between data categories within the target data cluster, which is used to reflect the strength of data association within the data cluster and helps to evaluate the stability and internal communication efficiency of the data cluster. The channel bandwidth allocation ratio can be understood as a bandwidth allocation scheme between different communication links determined according to the cross-cluster distance value and the cluster cohesion value to ensure the effective use of bandwidth resources. The direct communication channel can be understood as a dedicated communication path established between two or more target data clusters to ensure that data can be accurately transmitted in the shortest time, which is used to improve communication efficiency and reduce intermediate links.

[0062] An optional implementation method is to use a long short-term memory network to calculate the number of data packet transmissions and the upper limit of the channel capacity between two adjacent data clusters for the cross-cluster communication traffic data recorded by the data cluster communication data collector at a sampling time interval of 60 seconds, and obtain the communication bandwidth occupancy rate by dividing the transmission data packet size by the sampling time interval. The bandwidth occupancy rate is divided into communication time slots at intervals of 10 milliseconds, and a cross-layer communication time slot table is generated from the time slot division result. According to the cross-layer communication time slot table, a data transmission rate curve is obtained from the inter-cluster communication bandwidth monitor, and independent cache space is allocated to data packets whose transmission rate exceeds 75% of the channel capacity. A sliding time window is used to accumulate statistical data transmission conflicts and link interruptions, and a link quality score is calculated by the number of conflicts and interruptions, and a cross-layer channel state table is constructed from the link quality score. For channels with a link quality score greater than 0.8 in the cross-layer channel state table, uplink and downlink bandwidth tests are performed respectively within a fixed monitoring period, and the round-trip delay and packet loss number of the test data packet are recorded. Based on the adaptive boosting tree, the maximum transmission capacity of the channel and the link retention time are predicted from the delay and packet loss data, and the prediction results are graded and marked to generate a cross-layer link capacity table. According to the cross-layer link capacity table, the Euclidean distance between the data cluster center coordinates is calculated to obtain the cross-cluster distance value, the cluster cohesion value is calculated for the average connectivity between the data categories within the cluster, the channel bandwidth allocation ratio is determined by the weighted sum of the cross-cluster distance value and the cluster cohesion value, the data cache space is divided according to the allocation ratio, and the cross-cluster data transmission control rule (direct communication channel between data clusters) is constructed. The cross-cluster communication flow data reflects the information interaction characteristics between data clusters of equipment at different voltage levels. Taking the 110 kV substation and the 10 kV distribution line as an example, the data collector records the communication data volume every 60 seconds. The peak communication volume is predicted to be 2.5 megabits per second through the long short-term memory network, and the channel capacity upper limit is 3 megabits per second. Communication resource allocation is achieved according to the 10 millisecond time slot division. The dynamic change of data transmission rate directly affects the communication quality. During the transmission of substation telemetry data to the distribution automation master station, the bandwidth monitor recorded that the transmission rate curve had multiple peaks exceeding 2.25 megabits per second, accounting for more than 75% of the channel capacity. At this time, the cache mechanism was triggered. Within the 1-second time window, there were 5 transmission conflicts and 2 link interruptions. The link quality score was calculated to be 0.85. The channel status assessment involves bidirectional bandwidth testing. For the substation communication channel with a link quality score of 0.9, uplink and downlink tests were performed alternately within a 5-minute monitoring cycle. The test data packet size was 64 bytes, and the average round-trip delay was recorded to be 15 milliseconds, with a packet loss rate of 0.1%. The adaptive boosting tree predicted that the maximum transmission capacity of the channel was 2.8 megabits per second, and the link could be maintained for up to 4 hours.The spatial distribution characteristics of data clusters affect the bandwidth allocation strategy. The distance between the substation data cluster and the distribution line data cluster is calculated to be 0.65, the average connectivity of data categories within the substation cluster is 0.82, and the average connectivity within the distribution line cluster is 0.75. Based on these parameters, it is determined that the substation data occupies 75% of the channel bandwidth. In the scenario where the feeder terminal data is uploaded to the substation, the communication traffic shows obvious periodicity and burstiness. Conventional telemetry data is uploaded every 5 minutes, with a data volume of about 200 kilobytes. Fault alarm data occurs randomly, and the single data volume can reach 1 megabyte. 20% of the bandwidth resources are reserved for burst data transmission through the time slot table. The cross-layer communication requirements between the substation and the distribution line are reflected in many aspects. The protection action information requires real-time upload, the data packet delay is limited to within 10 milliseconds, the measurement data allows short-term caching, and the maximum delay is limited to 1 second. The delay requirement for the operating status data is relatively loose, which can reach 10 seconds. The dynamic allocation of channel bandwidth resources is based on the data transmission priority. When protection action data and measurement data are applied for transmission at the same time, the cross-cluster transmission control rule determines that the protection data takes priority in occupying channel resources, and the measurement data enters the buffer pool to wait, so as to achieve the priority transmission guarantee of key business data. The clustering characteristics of data clusters affect the allocation of communication resources. The cohesion of the local control data cluster of the substation is as high as 0.95, and the cross-cluster distance with the adjacent interval switch cabinet is only 0.25, so it is allocated a larger communication bandwidth share. The cluster cohesion of the remote measurement data cluster is 0.68, and the cross-cluster distance reaches 0.85, so it obtains relatively less bandwidth resources.

[0063] By adopting this technical solution, the communication bandwidth occupancy rate is determined by analyzing the transmission data packet size and sampling time interval corresponding to the data category, and the link quality score is evaluated accordingly. For communication links with link quality scores higher than the preset threshold, bandwidth tests are performed within a fixed detection period, and the round-trip delay and packet loss number of the test data packet are used to determine the maximum transmission capacity of the channel. Based on this, the cross-cluster distance value and cluster cohesion value between the target data clusters are calculated, the channel bandwidth allocation ratio is determined, and a direct communication channel is established. It can dynamically optimize bandwidth resource allocation, ensure the efficiency and stability of key communication links, improve the overall performance and service quality of the power terminal communication network, and enhance the reliability and response speed of the power terminal.

[0064] On the basis of the above scheme, optionally, after constructing the data communication model according to the importance of the data category in the communication network of the power system, it also includes: determining the inter-cluster association relationship between the target data clusters according to the data communication model, establishing a data cluster interconnection graph according to the inter-cluster association relationship, and determining the target path and connectivity index between the target data clusters in the data cluster interconnection graph, if the path connectivity index is lower than the preset index threshold, constructing a link supplement list between the target data clusters according to the target path; calculating the number of direct links of each target data cluster according to the link supplement list to obtain a node degree value, constructing an adjacency matrix according to the node degree value and the path redundancy, and constructing a topology structure graph based on the adjacency matrix; for each of the supplementable communication links in the topology structure graph, determining the data throughput and link bandwidth capacity of the supplementable communication link, dividing the data throughput by the link bandwidth capacity to obtain link utilization, scoring the supplementable communication link based on the link utilization and the node degree value, generating a link enhancement rule set according to the scoring result, so as to update the data communication model based on the link enhancement rule set.

[0065] Among them, the inter-cluster association relationship can be understood as a way to describe the interaction and dependence between different target data clusters, which is used for the logical connection between the clusters and provides a data basis for the subsequent data cluster interconnection diagram construction and path planning. The data cluster interconnection diagram can be understood as a graphical table showing the connection relationship between the target data clusters, which can provide an intuitive network structure view. The target path can be understood as a communication path from one target data cluster to another target data cluster in the data cluster interconnection diagram, which is used to determine the specific route of data transmission. The connectivity index can be understood as an index to measure the reliability and efficiency of the connection between nodes or clusters in the network. The link supplement list can be understood as an additional communication link used to enhance the path connectivity, providing alternative links to improve the redundancy and reliability of the network. Among them, the link supplement list can include multiple complementary communication links between the target data clusters. The number of direct links can be understood as the number of other clusters directly connected to each target data cluster, which can be used to reflect the degree of connection of the target data cluster in the network. The node degree value can be understood as the number of direct connections between each target data cluster and other clusters. The higher the node degree value, the more direct connections the node has in the network. The path redundancy can be understood as the degree to which there are multiple independent target paths to achieve the same communication, which can increase the fault tolerance and stability of the network and prevent communication interruption caused by the failure of a single target path. The adjacency matrix can be understood as a matrix structure used to represent the connection relationship between target data clusters in the data cluster interconnection diagram. The topology diagram can be understood as a graphical representation of all target data clusters in the network and their connection methods. The data throughput can be understood as the amount of data transmitted through a communication link per unit time, which is used to measure the actual load of the link. The link bandwidth capacity can be understood as the maximum data transmission rate that a communication link can support. The link utilization can be understood as the ratio of data throughput to link bandwidth capacity, which is used to evaluate whether the link is fully utilized. The link enhancement rule set can be understood as a series of rules that specify how to update and optimize the data communication model based on the scoring results.

[0066] An optional implementation method is to calculate the average number of communication bytes per second and the number of communications between data clusters based on the graph convolution network according to the communication data of each data cluster within a 300-second statistical period recorded by the power terminal data communication monitor, divide the number of communication bytes by the maximum bandwidth to obtain the link occupancy rate, use the convolution kernel to perform a two-dimensional convolution operation on the link occupancy rate and the number of communications to obtain the association weight, and construct a data cluster interconnection graph from the normalized association weight. For the data cluster interconnection graph, the bandwidth-weighted shortest path length is calculated for each pair of data cluster nodes by the shortest path calculator, the path connectivity score is calculated from the path length and the number of node hops, and the connectivity score threshold of 0.8 is set as the communication quality standard. Based on the monthly communication data statistics of the number of interruptions and the number of delays exceeding the standard, a logical link tag is added to the node pairs whose connectivity scores are lower than the preset indicator threshold, and a link supplement list is constructed. According to the link supplement list, the number of direct links of the node is calculated to obtain the node degree value, the node degree value is divided by the total number of nodes to obtain the node connectivity, the path redundancy is calculated by subtracting the number of shortest path links from the number of links of each path, and the adjacency matrix is ​​constructed based on the node connectivity and path redundancy. The topological structure diagram is obtained by multiplying the element value of the adjacency matrix by the bandwidth capacity of the corresponding link. For the topological structure diagram, the link utilization is calculated by dividing the link data throughput by the link bandwidth capacity, the node distribution degree is calculated based on the distance from the node to the center of the graph, the topological hierarchy is divided according to the number of shortest path hops between nodes, the transmission path scoring function is constructed from the link utilization, node distribution degree, and topological hierarchy, and the transmission path scoring results are normalized to generate a link enhancement rule set. The communication mode between the power terminal data clusters shows obvious hierarchical and periodic characteristics. Within the 300-second statistical period, the average communication volume between the substation control data cluster and the switchgear data cluster reached 1.2 megabytes per second, the number of communications reached 240 times, and the association weight after normalization was 0.85, showing a strong inter-cluster association. The path connectivity between data cluster nodes directly affects the communication quality. Taking the 110 kV substation and the 10 kV distribution line terminal as an example, the bandwidth-weighted shortest path length between the two is 3 hops. Monthly statistics show that the link was interrupted 12 times and the delay exceeded 35 times. The calculated connectivity score is 0.75, which is lower than the threshold standard of 0.8. It is necessary to increase the direct connection channel. The node degree reflects the connection complexity of the data cluster. The substation main controller node is connected to 8 measurement and control terminal nodes, with a node degree of 8 and a total number of nodes of 15. The calculated node connectivity is 0.53, while the actual shortest path only needs 3 hops to cover all nodes. The path redundancy is 1.67, indicating that the network structure has sufficient communication backup capabilities. The link utilization reveals the efficiency of communication resource utilization. The peak throughput of the core control data link reaches 8 megabits per second, the link bandwidth capacity is 10 megabits per second, and the link utilization is 0.8, which is in a high load state. The shortest distance from the node to the center of the graph is 2 hops, and the node distribution degree is 0.4, which belongs to the edge node of the network.In the substation automation system, there were originally two independent communication links between the protection device data cluster and the measurement and control device data cluster. Monthly statistics show that the number of link interruptions exceeded 20 times. After adding a third backup link, the path redundancy was increased to 1.5, and the number of link interruptions was reduced to less than 5 times. The communication topology between distribution automation terminals presents a radial structure. The master station node is the center of the graph, connecting multiple feeder terminal nodes. The node distribution degree decreases from the center to the outside, and the distribution degree of the peripheral nodes is as low as 0.2. The connectivity of the edge nodes is improved by adding ring network connections. The hierarchical attributes of the terminal equipment affect the selection of the communication path. The on-site digital measurement and control unit of the primary equipment belongs to the bottom node. The measurement and control data needs to be forwarded through three layers before it can reach the master station. After adding a direct connection channel, the communication path is shortened to a single hop, realizing the rapid upload of data. The application of link enhancement rules is reflected in many aspects. For links with a connectivity score lower than 0.8, direct connection channels are added first. For paths with a link utilization rate of more than 0.75, the communication load is shared by adding backup links. For edge nodes with a node distribution degree lower than 0.3, a ring network structure is introduced to improve connectivity.

[0067] An optional implementation method, according to the power terminal interconnection communication data model, collects data flow records at 10 millisecond intervals from each communication link, constructs three test scenarios: a sudden increase of communication load by 2 times, a change in network topology, and a node downtime recovery, uses a circular convolutional network to extract features of the number of error bits, the number of packet losses, and the number of throughput bytes in the communication data packet, and calculates the communication quality score of each communication link from the feature data. For the communication link quality score, the link bandwidth usage value and the data packet processing delay value are recorded every 60 seconds, a load trend curve is constructed through the bandwidth usage record, a probability distribution diagram is drawn for the processing delay value, a polynomial regressor is used to fit the load and delay curve, and the mean square error value of the fitting curve is calculated as a link stability index, and a communication robustness score table is generated based on the stability index. According to the communication robustness score table, a score threshold of 0.85 is set as the communication quality qualification standard, communication data is extracted from the period where the score is lower than the threshold, and data with a packet arrival time interval of less than 1 millisecond is marked as a congestion point. A communication load distribution heat map is drawn based on the congestion point record, and links with a load exceeding 80% of the preset bandwidth are identified from the heat map as objects to be optimized. For the link to be optimized, the peak bandwidth, average bandwidth, and minimum bandwidth are calculated through the historical communication records of the link, and the link resource allocation weights are corrected based on the three bandwidth values. A resource allocation curve is constructed from the corrected weight value, and the bandwidth of the communication link is dynamically adjusted according to the resource allocation curve to generate an interconnection enhanced communication data model. The collection of power terminal communication data requires high time accuracy. Taking substation monitoring as an example, data flow is recorded every 10 milliseconds. During normal operation, the data packet size is maintained at about 64 bytes. When the protection device is activated, the data volume suddenly increases to 128 bytes, and the duration is about 300 milliseconds, which constitutes a typical load surge scenario. The network topology change scenario is reflected in the fact that when the state of the switch device changes, the original two communication links are suddenly reduced to one, and the communication load jumps from 50% to 95%. The number of error bits extracted from the data packet by the cyclic convolutional network increases to 3 bits per packet, the packet loss rate rises to 0.5%, and the throughput drops to 85% of the original communication volume. The link bandwidth usage shows obvious periodicity. In daily operation, the bandwidth utilization rate fluctuates between 45% and 55% within a 60-second statistical period. The packet processing delay is concentrated in the range of 5 to 8 milliseconds. The mean square error value obtained by polynomial regression fitting is 0.05, and the corresponding link stability index is 0.92. The load distribution heat map reveals the location and degree of communication bottlenecks. During the upload of substation telemetry data, the core switching node experienced multiple data congestion between 8 and 10 in the morning, and the arrival interval of adjacent data packets dropped to 0.8 milliseconds. The bandwidth utilization rate reached 85%, exceeding the warning threshold of 80%.The dynamic allocation of bandwidth resources is based on historical data statistics. The peak bandwidth of a communication link reaches 9.5 megabits per second, the average bandwidth is 4.2 megabits per second, and the minimum bandwidth is 1.8 megabits per second. The resource allocation weight calculated based on this is 0.65, which is lower than the original configuration of 0.8. The communication characteristics during the start and stop of the equipment change significantly. Taking the closing of the circuit breaker as an example, the measurement and control data flow in the first second of the operation increases from 2 kilobytes per second to 12 kilobytes per second, and the communication quality score drops from 0.95 to 0.82, triggering the dynamic bandwidth adjustment mechanism. The robustness of the communication model is reflected in its adaptability to abnormal working conditions. When a short circuit occurs in the feeder, the amount of fault recording data transmission increases to 8 times the normal value within 10 milliseconds. The communication quality score is kept above 0.88 through the pre-allocated spare bandwidth resources. The communication load mutation caused by the bus differential protection action is the most drastic. Within 50 milliseconds from the start of the protection to the completion of the action, control instructions, status information, and fault data are transmitted simultaneously. The instantaneous bandwidth occupancy rate reaches 95%. Low-priority data is transferred to the backup channel through a dynamic allocation algorithm to maintain the quality of core business communications.

[0068] This technical solution is adopted to determine the target paths and connectivity indicators between target data clusters by constructing a data cluster interconnection diagram, and to identify and supplement links with insufficient connectivity. The node degree value of each target data cluster is calculated, and an adjacency matrix is ​​constructed in combination with the path redundancy to generate a topological structure diagram, evaluate the data throughput and link bandwidth capacity of each supplementary communication link, calculate the link utilization and score the links, and generate a link enhancement rule set to update the data communication model. This can optimize the network structure, improve path connectivity and redundancy, ensure the efficiency and stability of key communication links, enhance the reliability and response speed of the power system communication network, and improve resource utilization efficiency.

[0069] The technical solution of the embodiment of the present invention is as follows: first, by acquiring terminal operation data of multiple power terminals in the power system and determining operation characteristic data in the terminal operation data, the terminal operation data is classified according to the operation characteristic data to obtain multiple initial data sets. Since the operation characteristic data includes voltage fluctuation data and power loss data, accurate classification of terminal operation status can be achieved; then, multiple target data sets are determined according to the data similarity between the power terminal data in the multiple initial data sets, and the data association matrix corresponding to the multiple target data sets is determined. According to the data association matrix, the importance of the data category corresponding to the target data set in the communication network of the power system is determined, and the importance of the data category can be determined; finally, a data communication model is constructed according to the importance of the data category in the communication network of the power system. Since the data communication model is a digital model for simulating the interconnected communication data between the multiple power terminals, the network design and resource allocation can be optimized, and the problem that it is difficult to effectively process a large amount of complex terminal operation data in the related technology is solved, the operation efficiency, stability and response speed of the power system are improved, and the reliability and intelligence level of the overall communication network are enhanced.

[0070] Embodiment 2

[0071] Figure 2 This is a schematic diagram of a structure of a device for constructing a power terminal interconnection communication model provided in the second embodiment of the present invention. Figure 2 As shown, the device includes: a terminal operation data classification module 210, an importance determination module 220 and a data communication model construction module 230.

[0072] The terminal operation data classification module 210 is used to obtain the terminal operation data of multiple power terminals in the power system, and determine the operation characteristic data in the terminal operation data, and classify the terminal operation data according to the operation characteristic data to obtain multiple initial data sets; wherein the operation characteristic data includes voltage fluctuation data and power loss data; the importance determination module 220 is used to determine multiple target data sets according to the data similarity between the power terminal data in the multiple initial data sets, and determine the target association matrix corresponding to the multiple target data sets, and determine the importance of the data category corresponding to the target data set in the communication network of the power system according to the target association matrix; the data communication model construction module 230 is used to construct a data communication model according to the importance of the data category in the communication network of the power system; wherein the data communication model is a digital model for simulating the interconnected communication data between multiple power terminals.

[0073] The technical solution of the embodiment of the present invention is as follows: first, the terminal operation data of multiple power terminals in the power system are obtained through the terminal operation data classification module 210, and the operation characteristic data in the terminal operation data are determined, and the terminal operation data are classified according to the operation characteristic data to obtain multiple initial data sets. Since the operation characteristic data includes voltage fluctuation data and power loss data, accurate classification of terminal operation status can be achieved; then, the importance determination module 220 determines multiple target data sets according to the data similarity between the power terminal data in the multiple initial data sets, and determines the data association matrix corresponding to the multiple target data sets. The data association matrix determines the importance of the data category corresponding to the target data set in the communication network of the power system, and the importance of the data category can be determined; finally, the data communication model is constructed according to the importance of the data category in the communication network of the power system through the data communication model construction module 230. Since the data communication model is a digital model for simulating the interconnected communication data between multiple power terminals, it can optimize network design and resource allocation, solve the problem of difficulty in effectively processing a large amount of complex terminal operation data in related technologies, improve the operation efficiency, stability and response speed of the power system, and enhance the reliability and intelligence level of the overall communication network. On the basis of the above scheme, optionally, the terminal operation data at least includes operating voltage data; the terminal operation data classification module includes: a construction submodule, a fitting submodule and a division submodule. Among them, the construction submodule is used to determine the voltage deviation value between the operating voltage data of multiple power terminals and the rated voltage, determine the power terminal whose voltage deviation value exceeds the preset deviation threshold as an abnormal terminal, and construct a reference data set based on the terminal operation data corresponding to the abnormal terminal; the fitting submodule is used to obtain the load rate data of the abnormal terminal, and fit the load rate data of multiple abnormal terminals by the least squares method to obtain the power loss distribution curve; the division submodule is used to obtain the cumulative power loss value in the multiple abnormal terminals by integrating the power loss distribution curve, and divide the reference data set into multiple initial data sets according to the cumulative power loss value.

[0074] On the basis of the above scheme, optionally, the importance determination module includes: an intermediate data set determination submodule, a target data set determination submodule and a target association matrix determination submodule. Wherein, the intermediate data set determination submodule is used to determine the Pearson correlation coefficient between the power terminal data in the initial data set, obtain the cluster center coordinates through single nearest neighbor clustering according to the Pearson correlation coefficient, calculate the Manhattan distance of the power terminal data using the cluster center coordinates, and determine the intermediate data set according to the Manhattan distance; the target data set determination submodule is used to determine the cosine similarity matrix according to the multiple power terminal data in the intermediate data set, and for the power terminal data whose cosine similarity is higher than or equal to the preset pre-selected threshold, the target data set is obtained by Mahalanobis distance calculation; the target association matrix determination submodule is used to use a support vector machine to respectively determine the feature space distribution position corresponding to each target data set, construct a Euclidean distance matrix through multiple feature space positions, and determine the target association matrix corresponding to multiple target data sets according to the Euclidean distance matrix; wherein, the target association matrix is ​​used to describe the association information between the data categories corresponding to the multiple target data sets.

[0075] On the basis of the above scheme, optionally, the importance determination module includes: a category connection weight determination submodule, a centrality value determination submodule and an importance determination submodule. Among them, the category connection weight determination submodule is used to screen out target data groups whose mutual information is greater than a preset information threshold from the target association matrix, and determine the association rules between the target data groups, construct a category association rule network according to the association rules, determine adjacent data categories according to the category association rule network, and determine the mutual information entropy between adjacent data categories, and determine the category connection weight according to the mutual information entropy; the centrality value determination submodule is used to use a breadth-first search algorithm to identify the shortest transmission path between the data categories, determine the degree centrality value and the betweenness centrality value of the data category according to the data category and the shortest transmission path between the data categories, and determine the centrality value according to the degree centrality value and the betweenness centrality value; the importance determination submodule is used to determine, for each of the data categories, the fluctuation standard deviation of the category connection weight of the data category within a preset number of consecutive cycles, obtain the total amount of data transmission of the data category with the fluctuation standard deviation less than the preset fluctuation threshold, determine the clustering coefficient of the data category according to the total amount of data transmission of the data category, and multiply the clustering coefficient by the centrality value to obtain the importance of the data category in the communication network of the power system.

[0076] On the basis of the above scheme, optionally, the data communication model construction module includes: a normalized connectivity matrix determination submodule, a spatial vector difference determination submodule, an initial importance score determination submodule and a data communication model construction submodule. Among them, the normalized connectivity matrix determination submodule is used to construct a data category importance matrix according to the importance of the data category in the communication network of the power system, and for the data category importance matrix, the Euclidean distance metric is used to calculate the connection strength value between adjacent data categories, and the normalized connectivity matrix is ​​obtained by dividing the connection strength value by the maximum communication bandwidth; the spatial vector difference determination submodule is used to cluster the data categories according to the normalized connectivity matrix using the spectral clustering method of the symmetric regularized Laplace matrix to obtain a target data cluster, determine the Manhattan distance between the data categories in the target data cluster, and determine the Manhattan distance according to the Manhattan distance. The spatial vector difference between the data categories; the initial importance score determination submodule, used to determine the total amount of data transmission of the data category based on the spatial vector difference and the communication frequency between the data categories, and determine the initial importance score of each data category by using the total amount of data transmission, the spatial vector difference and the Gaussian process regressor; the data communication model construction submodule, used to determine the data transmission delay between the data categories, determine the communication bandwidth value based on the data transmission delay, determine the communication timing score of the data category based on the initial importance score and the communication bandwidth value, and construct a data communication model based on the communication timing score of the data category.

[0077] On the basis of the above scheme, optionally, the construction device of the interconnected communication model of the power terminal further includes: a link quality score determination module, a channel maximum transmission capacity determination module and a direct communication channel establishment module. Among them, the link quality score determination module is used to determine the communication bandwidth occupancy rate corresponding to the data category through the transmission data packet size and sampling time interval corresponding to the data category, and determine the link quality score of the communication link corresponding to the data category according to the communication bandwidth occupancy rate; the channel maximum transmission capacity determination module is used to perform a bandwidth test within a fixed detection period for the communication link whose link quality score is greater than a preset score threshold, and determine the channel maximum transmission capacity according to the round-trip delay and packet loss number of the test data packet in the bandwidth test; the direct communication channel establishment module is used to determine the cross-cluster distance value between the target data clusters according to the channel maximum transmission capacity, determine the cluster cohesion value from the average connectivity of the data category in the target data cluster, determine the channel bandwidth allocation ratio through the cross-cluster distance value and the cluster cohesion value, and establish the direct communication channel between the target data clusters in the data communication model based on the channel bandwidth allocation ratio.

[0078] On the basis of the above scheme, optionally, the device for constructing the power terminal interconnection communication model further includes: a link supplement list construction module, a link supplement list construction module and a data communication model update module. The link supplement list construction module is used to determine the inter-cluster association relationship between the target data clusters according to the data communication model, establish a data cluster interconnection graph according to the inter-cluster association relationship, and determine the target path and connectivity index between the target data clusters in the data cluster interconnection graph. If the path connectivity index is lower than the preset index threshold, a link supplement list between the target data clusters is constructed according to the target path; wherein the link supplement list includes multiple supplementable communication links between the target data clusters; the link supplement list construction module is used to calculate the number of directly connected links of each target data cluster according to the link supplement list to obtain a node degree value, construct an adjacency matrix according to the node degree value and the path redundancy, and construct a topology graph based on the adjacency matrix; the data communication model update module is used to determine the data throughput and link bandwidth capacity of the supplementable communication link for each supplementable communication link in the topology graph, divide the data throughput by the link bandwidth capacity to obtain the link utilization, score the supplementable communication link based on the link utilization and the node degree value, and generate a link enhancement rule set according to the scoring result to update the data communication model based on the link enhancement rule set.

[0079] The device for constructing the electric power terminal interconnection communication model provided in the embodiment of the present invention can execute the method for constructing the electric power terminal interconnection communication model provided in any embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method.

[0080] Embodiment 3

[0081] Figure 3 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0082] like Figure 3As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0083] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0084] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a method for constructing a power terminal interconnection communication model.

[0085] In some embodiments, a method for constructing a power terminal interconnection communication model may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the method for constructing a power terminal interconnection communication model described above may be executed. Alternatively, in other embodiments, the processor 11 may be configured to execute a method for constructing a power terminal interconnection communication model in any other appropriate manner (e.g., by means of firmware).

[0086] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0087] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0088] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0089] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0090] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0091] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0092] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0093] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing an interconnection communication model of power terminals, characterized in that: include: Acquire terminal operation data of a plurality of power terminals in a power system, determine operation characteristic data in the terminal operation data, classify the terminal operation data according to the operation characteristic data, and obtain a plurality of initial data sets; wherein the operation characteristic data includes voltage fluctuation data and power loss data; Determine multiple target data sets according to the data similarity between the power terminal data in the multiple initial data sets, determine the target association matrix corresponding to the multiple target data sets, and determine the importance of the data category corresponding to the target data set in the communication network of the power system according to the target association matrix; A data communication model is constructed according to the importance of the data categories in the communication network of the power system; wherein the data communication model is a digital model for simulating interconnected communication data between a plurality of the power terminals.

2. The method for constructing a power terminal interconnection communication model according to claim 1, characterized in that: The terminal operation data at least includes operation voltage data; the operation characteristic data in the terminal operation data is determined, and the terminal operation data is classified according to the operation characteristic data to obtain multiple initial data sets, including: Determine voltage deviation values ​​between operating voltage data of a plurality of the power terminals and a rated voltage, determine the power terminals whose voltage deviation values ​​exceed a preset deviation threshold as abnormal terminals, and construct a reference data set based on the terminal operating data corresponding to the abnormal terminals; Acquire the load rate data of the abnormal terminal, and fit the load rate data of multiple abnormal terminals by least square method to obtain an electric energy loss distribution curve; The power loss cumulative values ​​in the plurality of abnormal terminals are obtained by performing an integral operation on the power loss distribution curve, and the reference data set is divided into a plurality of initial data sets according to the power loss cumulative values.

3. The method for constructing a power terminal interconnection communication model according to claim 1, characterized in that: The step of determining a plurality of target data sets according to the data similarities between the power terminal data in the plurality of initial data sets, and determining a target association matrix corresponding to the plurality of target data sets, comprises: Determine the Pearson correlation coefficient between the power terminal data in the initial data set, obtain cluster center coordinates through single nearest neighbor clustering according to the Pearson correlation coefficient, calculate the Manhattan distance of the power terminal data using the cluster center coordinates, and determine the intermediate data set according to the Manhattan distance; Determine a cosine similarity matrix according to the plurality of power terminal data in the intermediate data set, and obtain a target data set by calculating the Mahalanobis distance for the power terminal data whose cosine similarity is higher than or equal to a preset preselected threshold; A support vector machine is used to determine the feature space distribution position corresponding to each target data set, a Euclidean distance matrix is ​​constructed through multiple feature space positions, and a target association matrix corresponding to multiple target data sets is determined according to the Euclidean distance matrix; wherein the target association matrix is ​​used to describe the association information between data categories corresponding to multiple target data sets.

4. The method for constructing a power terminal interconnection communication model according to claim 1, characterized in that: The determining, according to the target association matrix, the importance of the data category corresponding to the target data set in the communication network of the power system comprises: Filtering target data groups whose mutual information is greater than a preset information threshold from the target association matrix, determining association rules between the target data groups, constructing a category association rule network according to the association rules, determining adjacent data categories according to the category association rule network, determining mutual information entropy between adjacent data categories, and determining category connection weights according to the mutual information entropy; A breadth-first search algorithm is used to identify the shortest transmission path between the data categories, a degree centrality value and a betweenness centrality value of the data category are determined according to the data category and the shortest transmission path between the data categories, and a centrality value is determined according to the degree centrality value and the betweenness centrality value; For each of the data categories, determine the fluctuation standard deviation of the category connection weight of the data category within a continuous preset number of periods, obtain the total data transmission amount of the data category whose fluctuation standard deviation is less than the preset fluctuation threshold, determine the clustering coefficient of the data category based on the total data transmission amount of the data category, multiply the clustering coefficient by the centrality value, and obtain the importance of the data category in the communication network of the power system.

5. The method for constructing a power terminal interconnection communication model according to claim 1, characterized in that: The step of constructing a data communication model according to the importance of the data category in the communication network of the power system includes: Constructing a data category importance matrix according to the importance of the data category in the communication network of the power system, calculating the connection strength values ​​between adjacent data categories using the Euclidean distance metric for the data category importance matrix, and obtaining a normalized connectivity matrix by dividing the connection strength values ​​by the maximum communication bandwidth; Clustering the data categories using a spectral clustering method of a symmetric regularized Laplace matrix according to the normalized connectivity matrix to obtain a target data cluster, determining the Manhattan distance between the data categories in the target data cluster, and determining the spatial vector difference between the data categories according to the Manhattan distance; Determine the total amount of data transmission of the data category according to the spatial vector difference and the communication frequency between the data categories, and determine the initial importance score of each of the data categories by using the total amount of data transmission, the spatial vector difference and a Gaussian process regressor; Determine the data transmission delay between the data categories, determine the communication bandwidth value according to the data transmission delay, determine the communication timing score of the data category according to the initial importance score and the communication bandwidth value, and construct a data communication model according to the communication timing score of the data category.

6. The method for constructing a power terminal interconnection communication model according to claim 1, characterized in that: After constructing the data communication model according to the importance of the data categories in the communication network of the power system, the method further includes: Determine the communication bandwidth occupancy rate corresponding to the data category by using the transmission data packet size and the sampling time interval corresponding to the data category, and determine the link quality score of the communication link corresponding to the data category according to the communication bandwidth occupancy rate; For the communication link whose link quality score is greater than a preset score threshold, a bandwidth test is performed within a fixed detection period, and a maximum transmission capacity of the channel is determined according to the round-trip delay and the number of packet losses of the test data packets in the bandwidth test; The cross-cluster distance value between the target data clusters is determined according to the maximum transmission capacity of the channel, the cluster cohesion value is determined from the average connectivity of the data categories in the target data cluster, the channel bandwidth allocation ratio is determined by the cross-cluster distance value and the cluster cohesion value, and a direct communication channel between the target data clusters is established in the data communication model based on the channel bandwidth allocation ratio.

7. The method for constructing a power terminal interconnection communication model according to claim 1, characterized in that: After constructing the data communication model according to the importance of the data categories in the communication network of the power system, the method further includes: Determine the inter-cluster association relationship between the target data clusters according to the data communication model, establish a data cluster interconnection graph according to the inter-cluster association relationship, and determine the target path and connectivity index between the target data clusters in the data cluster interconnection graph, and if the path connectivity index is lower than a preset index threshold, construct a link supplement list between the target data clusters according to the target path; wherein the link supplement list includes a plurality of supplementable communication links between the target data clusters; Calculate the number of directly connected links of each target data cluster according to the link supplement list to obtain a node degree value, construct an adjacency matrix according to the node degree value and path redundancy, and construct a topology structure graph based on the adjacency matrix; For each of the supplementable communication links in the topology diagram, the data throughput and link bandwidth capacity of the supplementable communication link are determined, the data throughput is divided by the link bandwidth capacity to obtain link utilization, the supplementable communication links are scored based on the link utilization and the node degree values, and a link enhancement rule set is generated according to the scoring results, so as to update the data communication model based on the link enhancement rule set.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for constructing the power terminal interconnection communication model described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for constructing a power terminal interconnection communication model according to any one of claims 1 to 7 when executed.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the computer program implements the method for constructing a power terminal interconnection communication model as described in any one of claims 1 to 7.

Images