A graphical data conversion model under a power distribution network graph multi-state architecture

Abstract

This invention discloses a graphical data conversion model under a multi-modal architecture for distribution networks, relating to the field of distribution network data analysis and visualization technology. It includes a data centralization module for collecting and storing distribution network data modalities to realize a multi-modal platform; a central display hall visualization platform connected to the data centralization module and providing a graphical user interface for users to view and analyze data; and an emergent graphical data conversion module connected to the central display hall visualization platform, which converts and represents the collected data as emergent graphical data. This invention integrates the emergent graphical data conversion model into the existing "multi-modal" system, achieving deep data fusion, improving data utilization efficiency and quality, and providing strong technical support for the intelligent system upgrade of distribution networks.

Description

Technical Field

[0001] This invention relates to the field of power distribution network data analysis and visualization technology, and in particular to an emergent graphical data transformation model and its application under the multi-morphic architecture of a power distribution network. Background Technology

[0002] In the current data-driven operation and maintenance system of the distribution network, due to historical and technological development reasons, relevant departments of the power grid have built multiple professional business systems. Most of these systems were independently built based on the specific responsibilities of each department, resulting in numerous, scattered systems with overlapping and isolated functions. This architecture struggles to meet the new requirements of a "smart service system characterized by comprehensive status awareness, efficient information processing, and convenient and flexible applications." Therefore, it is urgent to build upon the achievements of the power grid resource business platform construction, integrating business data from multiple systems to establish a comprehensive distribution network business management platform that connects the four stages of "planning, design, construction, and operation," supporting distribution network business applications. Simultaneously, by leveraging Geographic Information Systems (GIS), spatial and attribute information of electrical equipment can be integrated to provide integrated visual management of the entire process of distribution network planning, construction, and operation, making management more intuitive, operation more convenient, improving distribution network management efficiency, and ultimately achieving the management goals of "three-fold transformation and two-fold improvement" of the distribution network.

[0003] However, a major problem in existing power distribution network systems is the lack of deep data correlation and effective integration between different data modalities. For example, data from various business stages, such as power distribution network planning and construction, operation, maintenance and emergency repair, and potential power supply hazards, operate independently and are not interconnected. This data fragmentation and dispersion prevents data from working effectively together during fault analysis and management. Although existing technologies have achieved basic graphical representations of each part, these representations lack the ability to correlate and integrate different data modalities, thus failing to provide a comprehensive and integrated visualization.

[0004] Furthermore, with the rapid development of artificial intelligence technology, utilizing neural networks and machine learning platforms to learn and analyze power grid data has become a trend. However, in existing distribution network systems, due to data dispersion and quality issues, this data cannot be effectively used to train existing artificial intelligence platforms, thus limiting the development of intelligent applications.

[0005] Chinese patent application CN 117633317 A discloses a one-map-four-state visualization management system for power networks. This system includes a visualization module that integrates data reception, data analysis, and data mapping to establish four visualization states; an integration module that integrates the four visualization states to form an interactive network; a display device for visually displaying the interactive visualization states; and a login module that forms a user terminal for users to log in to the integration module, with different user terminals granting different access permissions. This invention enables different user terminals to perform management operations using visualization through the four visualization states, making some power data easier to understand. Furthermore, the four visualization states can interact through the established interactive network, improving the efficiency of visualization management. However, the one-map-four-state visualization management system developed by the above technology mainly focuses on simple data integration and consistency (and visualization in the central monitoring hall of the power grid), with less attention paid to the internal interactive relationships of the data, and the identifiable attributes of the data when connected to neural networks (or the trainability and quality of machine learning data). Therefore, it is necessary to further develop a data transformation model that integrates data from all stages of power distribution network planning, design, construction, operation, maintenance, emergency repair, and data security risks in a high-level and in-depth manner, in order to achieve deep data fusion and improve the efficiency and quality of data use.

[0006] In conclusion, it is particularly important to construct an emergent graphical data transformation model that can integrate multiple data modalities. This will not only enable deep data fusion and improve the efficiency and quality of data use, but also provide strong technical support for the intelligent system upgrade of power distribution networks. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide an emergent graphical data transformation model under the multi-state architecture of a distribution network, in order to achieve deep data integration, improve data utilization efficiency and quality, and provide strong technical support for the intelligent system upgrade of the distribution network.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: An emergent graphical data transformation model under the one-map-multi-mode architecture of a distribution network includes a data centralization module for collecting and storing distribution network data modes to realize a one-map-multi-mode platform, and a central display hall visualization platform connected to the data centralization module and providing a graphical user interface for users to view and analyze data. The model is characterized by further including an emergent graphical data transformation module connected to the central display hall visualization platform, which transforms and represents the collected data as emergent graphical data.

[0009] Furthermore, the self-emerging graphical data conversion module is linked to the existing central display hall visualization platform of the power grid, and further graphically topologically constructs different data modalities such as distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points, thereby improving the visualization and analysis capabilities of the data.

[0010] Further optimized, at the basic data transformation level, the basic data transformation framework of the emergent graphical data transformation module is set to include: Initial modal plot: A blank plot used for data visualization; Data Modal Regions: The initial modal map is divided into multiple regions, each region corresponding to a specific data modality. The data modality includes, but is not limited to, distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points. Data mapping module: Based on the location of the data in the original database, it maps the data in each data mode to the corresponding region of the initial mode map.

[0011] Further optimized, at the core data association processing level, the emergent graphical data transformation module specifically includes: a. Extract data from the data centralization module, and identify and extract intra-modal and cross-modal correlation data based on the objective correlation and holistic characteristics at the physical level; b. Construct an associated data array module, which assigns the identified associated data to different partitions in a global graph according to their respective data modes. Each partition corresponds to an independent data mode and stores the associated data within the mode. c. Generate multiple correlation matrix modules, and construct an correlation matrix R within each independent modal partition. i , where i represents the specific number of the corresponding modalities, and a global association matrix R is constructed for cross-modal data association; d. A graphical modality building module that displays data in the correlation matrix R. i The position in the matrix is ​​fixed using the correlation matrix R. i The data points are sorted and connected according to their correlation strength, that is, according to their values ​​from 0 to 1, to form a dot-line graph with cross structures. e. The emergent graphical data conversion module uses the point and line graphs generated by the graphical modality construction module as interactive, machine-recognizable emergent graphical maps to represent the operating status of the distribution network under the following conditions: distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and its data focus points, i.e., other subsequently added data modalities. It is also particularly compatible with subsequent machine learning access and potential problem identification.

[0012] Further optimization involves integrating the graphical maps generated by the emergent graphical data conversion module into an existing artificial intelligence platform for data training. Based on the data in the graphical maps generated by the emergent graphical data conversion module and the annotation information of power grid operation events, an intelligent model for predicting and monitoring distribution network events is trained. Further optimization, at the executable data process level, the construction process of this data transformation model can optionally include: initial modality map construction, associated data model construction, and graphical modality construction.

[0013] Further optimized, the initial modality map construction includes: first, initializing a blank map by setting a blank map for data visualization as the initial modality map; dividing the initial modality map into n regions, where n is determined according to the number of data modalities that need to be processed; each region corresponds to a single data modality; and then performing data mapping, whereby for each selected data modality, its discrete data points are sequentially mapped to the corresponding regions.

[0014] Further optimization involves constructing the associated data model as follows: for all data modes or their mapped graphical multi-region maps, a data array is constructed based on the correlation of data within a mode or across modes; the data array is represented as a matrix, which can be divided into n square matrices through matrix partitioning operations, each square matrix corresponding to the data correlation within a mode; the data array as a whole corresponds to global data correlation or cross-modal correlation; the data array is mutated according to the number of modes in the current analysis and the data characteristics of each mode; based on the data array, an association matrix R is further constructed within each independent modal partition. i To characterize the data correlation within mode i, for cross-modal correlated data, a global correlation matrix R is constructed to characterize the correlation between all data points; the correlation matrix is ​​constructed as a parameter of the linear correlation strength between any two data points when changes occur, and its value is between 0 and 1.

[0015] Further optimization of the graphical modality construction includes: ordering and cross-referencing point-line graphing based on the correlation matrix. To this end, the position of the data in the local or global correlation matrix is ​​first fixed. Then, for the ordering and cross-referencing of the graph, the implementation path is: sorting and connecting the data points according to the correlation strength, that is, according to the value from 0 to 1, to form a point-line graph containing cross-structure.

[0016] The beneficial effects of adopting the above technical solution are as follows: Earlier developed multi-modal distribution data systems mainly focused on simple data integration and consistency (and visualization in the central monitoring hall of the power grid), with less attention paid to the internal interaction and correlation of data, and the identifiable attributes of data when connected to neural networks (or the trainability and quality of data in machine learning). In contrast, this invention achieves deep data fusion by integrating data from various stages of distribution network planning, design, construction, operation, maintenance, emergency repair, and data security risks, thereby improving data utilization efficiency and quality. Furthermore, this invention enhances the comprehensive utilization capability of data by identifying and integrating the correlation of cross-modal data, and cross-modal data fusion helps to build a more robust and efficient distribution network. A comprehensive and accurate power grid operation model improves the effectiveness of fault prediction and response. Furthermore, based on the interaction and correlation of multi-modal distribution network data, the newly developed data model significantly enhances the identifiable attributes when facing neural networks. Its data trainability and quality for machine learning support direct access to existing data for training neural networks, and seamless integration with existing artificial intelligence data training platforms. Therefore, in backend applications, it can support in-depth and correlated analysis, training, and optimization of multi-modal power grid data using machine learning technology. Based on the specific data annotation format in the early stages of data training, it is expected to identify key events in power grid operation, provide real-time analysis and early warnings, and so on. Detailed Implementation

[0017] The following embodiments illustrate the present invention in detail. In the description of the following embodiments, specific details such as particular system structures and techniques are set forth for illustrative purposes and not for limitation, so as to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary detail. It should be understood that, as used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. It should also be understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0018] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance. References to "one embodiment" or "some embodiments" in this application mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0019] First, the technical concept of this invention is summarized as follows: Existing technology development has already achieved data centralization and visualization of the central display hall in a multi-format map platform. Furthermore, based on the already constructed data centralization model for multi-format maps and its central visualization platform, we are developing an innovative data system: an emergent graphical data transformation model under the multi-format map architecture of the power distribution network.

[0020] Specifically, for the various data modalities related to distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network, and other data concerns, the aforementioned multi-modal platform has already achieved data centralization and visualization in a central exhibition hall. Based on this, in our newly developed data model, these different data modalities can be graphically topologically constructed separately.

[0021] The most basic data model is to first use a sufficiently large blank map as the initial modal map, and then divide this initial modal map into multiple regions, each region corresponding to a single data modality, such as distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points. Then, all the data in each data modality is mapped to different regions on the initial modal map according to their data positions in the original database. This is an initial modal map that lists all the data.

[0022] The core data model is based on the objective interconnectedness and holistic nature of the distribution network at the physical level. For distribution network planning and construction, operation, maintenance and emergency repair, potential power supply hazards, and related data, the data within each of these modes, as well as the data that is interconnected across different modes, often provide the most realistic and substantial representation of the actual operating status of the distribution network and its various maintenance events. This data with intra-modal or cross-modal correlations has genuine data content and the greatest data mining value. Therefore, our data model development is guided by the principle of constructing graphical and atlas-like modes and conducting data mining based on this type of interconnected data. Under this technical approach, we can first consider constructing an intra-modal or cross-modal data array, or extracting data with intra-modal or cross-modal correlations from the initial modal graph to construct a data array. This data array, in its overall structure, is somewhat similar to the initial modal graph; that is, it constructs several partitions on a global graph, each partition corresponding to an independent mode, and each partition stores intra-modal correlated data. The next step, relatively naturally, is to introduce the data tool of the association matrix, which allows us to construct an association matrix R within each independent modality partition. i Here, the value of i corresponds to the specific number of modes; simultaneously, a global correlation matrix R is constructed for cross-modal correlation data. Next, mathematically, we know that the correlation matrix is ​​a square matrix. For example, for m data points, it represents the correlation between any one data point and the remaining m-1 data points. Therefore, this matrix is ​​an m-order square matrix, and generally, the values ​​of its internal data are between 0 and 1, used to represent the linear correlation between two data points. At this point, we construct a graphical mode. First, the position of the data in the correlation cluster is fixed. Then, lines are connected sequentially according to the values ​​from 0 to 1. After all the lines are connected, a dot-line graph (often with an intersecting structure) is formed. This dot-line graph constitutes an emergent graphical map capable of machine recognition and representing the power grid's operating status.

[0023] Based on such a data graph model, it is easy to connect to existing artificial intelligence platforms for data training. Through the initial data annotation, the typical power grid operation events are nonlinearly fused with the corresponding point-line intersection structure graph in the data graph model. As the amount of data accumulates and is optimized, a graphical intelligent data model with increasingly accurate and strong generalization capabilities is trained, thus having practical application value and enabling power grid multimodal recognition and early warning. Example 1

[0024] Based on existing technologies, a data centralization module for collecting and storing distribution network data modes is established to realize a multi-mode platform. This module connects with the data centralization module and provides a central display hall visualization platform with a graphical user interface for users to view and analyze data. An emergent graphical data transformation model under the multi-mode architecture of distribution network is constructed, which includes an emergent graphical data transformation module.

[0025] The self-emerging graphical data conversion module is linked to the existing central exhibition hall visualization platform of the power grid, and further graphically topologically constructs different data modalities such as distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points, thereby improving the visualization and analysis capabilities of the data.

[0026] At the basic data transformation level, the basic data transformation framework of the emergent graphical data transformation module includes: Initial modal plot: A blank plot used for data visualization; Data Modal Regions: The initial modal map is divided into multiple regions, each region corresponding to a specific data modality. Data modalities include, but are not limited to, distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points. Data mapping module: Based on the location of the data in the original database, it maps the data in each data mode to the corresponding region of the initial mode map.

[0027] At the core data association processing level, the self-emergent graphical data transformation module specifically includes: a. Extract data from the data centralization module, and identify and extract intra-modal and cross-modal correlation data based on the objective correlation and holistic characteristics at the physical level; b. Construct an associated data array module, which assigns the identified associated data to different partitions in a global graph according to their respective data modes. Each partition corresponds to an independent data mode and stores the associated data within the mode. c. Generate multiple correlation matrix modules, and construct an correlation matrix R within each independent modal partition. i , where i represents the specific number of the corresponding modalities, and a global association matrix R is constructed for cross-modal data association; d. A graphical modality building module that displays data in the correlation matrix R. i The position in the matrix is ​​fixed using the correlation matrix R. i The data points are sorted and connected according to their correlation strength, that is, according to their values ​​from 0 to 1, to form a dot-line graph with cross structures. e. The emergent graphical data conversion module transforms the point and line graph generated by the graphical modality construction module into an interactive, machine-recognizable emergent graphical map, which represents the operating status of the distribution network under the following conditions: distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and its data focus points, i.e., other subsequently added data modalities. It is also particularly compatible with subsequent machine learning access and potential problem identification.

[0028] The graphical maps generated by the emergent graphical data conversion module are connected to the existing artificial intelligence platform for data training. Based on the data in the graphical maps generated by the emergent graphical data conversion module and the annotation information of power grid operation events, an intelligent model for predicting and monitoring distribution network events is trained.

[0029] As can be seen, the innovative technical solution described above can achieve deep data integration, improve data utilization efficiency and quality, and provide strong technical support for the intelligent system upgrade of power distribution networks. Example 2

[0030] Based on the emergent graphical data transformation model under the one-map multi-state architecture of the distribution network described in Example 1, we further constructed its instantiated executable data process. The main framework of this data process includes: initial modal graph construction, associated data model construction, and graphical modal construction.

[0031] Initial Modal Map Construction: First, a blank map is initialized, setting a blank map for data visualization as the initial modal map. The initial modal map is divided into n regions, where n is determined according to the number of data modes that need to be processed. For example, for the following common modes: distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points, then n=4. Each region corresponds to a single data mode. Then, data mapping is performed. For each selected data mode, its discrete data points are sequentially mapped to the corresponding regions.

[0032] Construction of the correlation data model: For all data modes or their mapped graphical multi-region maps, a data array is constructed based on the correlation of data within a mode or across modes. Preferably, the data array can be represented as a matrix, which can be divided into n square matrices through matrix partitioning, each square matrix corresponding to the data correlation within a mode; the data array as a whole corresponds to global data correlation (cross-modal correlation); the data array can be varied according to the number of modes in the current analysis and the data characteristics of each mode; based on the data array, a correlation matrix R is further constructed within each independent modal partition. iTo characterize the data correlation within mode i, for cross-modal correlated data, a global correlation matrix R is constructed to characterize the correlation between all data points; the correlation matrix is ​​constructed as a parameter of the linear correlation strength between any two data points when changes occur, and its value is between 0 and 1.

[0033] The construction of the graphical modality includes: ordering and cross-referencing the point-line graph based on the association matrix. To this end, the position of the data in the local or global association matrix is ​​first fixed. Then, the ordering and cross-referencing of the graph is achieved by: sorting the data points according to the association strength, that is, according to the value from 0 to 1, and connecting them to form a point-line graph with cross-structure. The point-line graph is represented as a graph G(V,E), where V is the set of vertices and E is the set of edges. Example 3

[0034] In addition to the above embodiments, as a numerical mode-oriented approach outside the main model, an auxiliary data process based on additional parameter indicators is constructed outside the above-mentioned graphical atlas model. This auxiliary data integrates the output parameter indicators, either alone or in combination with the graphical atlas data in the main model, and serves as a graphical data mode for parameter numerical enhancement under the multi-mode power grid (usually requiring combination, because data merging in the parameter indicator data process reduces the connotation of power grid operation data under many modes). Specifically, this auxiliary data process involves: constructing a topological mode based on the correlation matrix, constructing the determinant of the correlation matrix and using it for multi-indicator or single-indicator data analysis, and attaching it to the graphical atlas data in the main model to merge it into a graphical data mode for parameter numerical enhancement. Example 4

[0035] Furthermore, after constructing the aforementioned emergent graphical data transformation model, the constructed graphical atlas model can be integrated into an existing artificial intelligence platform for data training. Through prior data annotation, significant power grid operation events are nonlinearly fused with their corresponding point-line intersection structure graphs in the constructed data atlas model using multi-parameter methods. Example 5

[0036] Based on the above data processing model, a computer data processing process is constructed (based on Python and NumPy, Pandas, SciPy, etc.); its data structure framework is as follows: I. Initial Modal Map Data Structure Blank map initialization: # Define a sufficiently large blank spectrum G as the initial modality spectrum. # Assume there are 4 data modal regions n = 4 # Initialize blank graph G = [None] * n Data mapping: # Collect data and store it in a CSV file # Data mapping function # Map each data mode to its corresponding region II. Data Structure of the Related Data Model Data array construction: # Identify and extract data with intramodal or cross-modal correlations, and construct data array A. Introduction of the correlation matrix: # Construct an association matrix Ri within each independent modal partition. # For cross-modal association data, construct a global association matrix R. III. Graphical Modal Data Structure Graphical representation of the correlation matrix: # Fix the position of the data in the correlation matrix and connect them sequentially according to the value from 0 to 1. # Draw a graph of the global correlation matrix; IV. Data Processing Structure Topology mode construction: # Constructing topological modes based on the correlation matrix # Calculate the determinant / volume of a matrix Further Expansion Examples include data training, introducing machine learning models, or expanding data modules according to current job requirements. Example 6

[0037] Integrating the above model into the existing "one map, multiple states" system enables real-time data synchronization and automated processing. Through the visualization platform in the central exhibition hall, the operating status and potential hazards of the power grid can be displayed and monitored in real time. Furthermore, the "one map, multiple states" operation and maintenance model can be gradually promoted on a wider scale to improve the efficiency of distribution network construction and operation. Based on feedback from practical applications, the data model and algorithms will be continuously optimized and upgraded. Example 7

[0038] Based on feedback from various levels of power grid units during the application and promotion, subsequent expansion and construction were carried out. Currently, the following independent data modules have been constructed, and these expanded data modules can be integrated into the data model constructed above.

[0039] Expanding data module ①, three-level index mode: In the initial modal map construction, the regional division is further refined to construct a three-level index mode for substations, operation and maintenance unit jurisdictions and distribution lines.

[0040] Expand data module ②, real-time data display: Enables real-time display of data such as the health status of distribution network equipment, equipment operation history, special investigation tasks, and work order progress status.

[0041] Expanding data module ③, physical ID, drone inspection and temperature measurement cable technology: Integrating data from technologies such as physical ID, drone inspection and temperature measurement cables into the system to form multi-source data fusion.

[0042] Extended data module ④: Online tracking and management of three types of proactive work orders: Enables online tracking and management of three types of proactive work orders (planning, operation, and maintenance) to improve power supply reliability. Let W represent the work order status matrix, where each column represents the status of planning, operation, and maintenance work orders, respectively.

[0043] This invention achieves deep data fusion by integrating data from various stages, including planning, design, construction, and operation, thereby improving data utilization efficiency and quality. Furthermore, by identifying and integrating the correlations between cross-modal data, this invention enhances the comprehensive utilization capability of data. Cross-modal data fusion helps construct a more comprehensive and accurate power grid operation model, improving the effectiveness of fault prediction and response. This invention supports seamless connection and integration with artificial intelligence data training platforms, utilizing machine learning technology to conduct in-depth analysis of power grid data. Through training and optimization, the model can automatically identify key events in power grid operation and provide real-time analysis and early warning.

[0044] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A graphical data conversion system for a distribution network with a multi-mode architecture, comprising a data centralization module for collecting and storing data from different modes of the distribution network to realize a multi-mode platform, and a central display hall visualization platform connected to the data centralization module and providing a graphical user interface for users to view and analyze the data, characterized in that, It also includes an emergent graphical data conversion module connected to the visualization platform of the central exhibition hall; the emergent graphical data conversion module specifically includes: a. Extract data from the data centralization module, and identify and extract intra-modal and cross-modal correlation data based on the objective correlation and holistic characteristics at the physical level; b. Construct an associated data array module, which assigns the identified associated data to different partitions in a global graph according to their respective data modes. Each partition corresponds to an independent data mode and stores the associated data within the mode. c. generating a plurality of correlation matrix modules, one correlation matrix R is constructed within each independent modality partition i wherein i represents the specific number of corresponding modality, and one global correlation matrix R is constructed for cross-modality correlation data; d. A graphical modal construction module, which fixes the position of data in the correlation matrix R i , uses the data in the correlation matrix R i , sorts and connects the data points according to the correlation strength, i.e. according to the size of the numerical value from 0 to 1, to form a dot-line graph containing cross structures; e. The emergent graphical data conversion module uses the point and line graphs generated by the graphical modality construction module as interactive, machine-recognizable emergent graphical maps to characterize the operating status of the distribution network under the following conditions: distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points, and other subsequently added data modalities. It is also compatible with subsequent machine learning access and potential problem identification.

2. The graphical data conversion system under the multi-mode architecture of a distribution network according to claim 1, characterized in that, The self-emerging graphical data conversion module is linked to the existing central exhibition hall visualization platform of the power grid, and further constructs graphical topologies for different modalities of data, such as distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points, thereby improving the visualization and analysis capabilities of the data.

3. The graphical data conversion system under the multi-mode architecture of a distribution network according to claim 2, characterized in that, At the basic data transformation level, the basic data transformation framework of the self-emergent graphical data transformation module includes: Initial modal plot: A blank plot used for data visualization; Data Modal Regions: The initial modal map is divided into multiple regions, each corresponding to a data modality. The data modality includes, but is not limited to, distribution network planning and construction, distribution network operation, distribution network maintenance and emergency repair, potential power supply hazards in the distribution network and their data focus points. Data mapping module: Based on the location of the data in the original database, it maps the data in each data mode to the corresponding region of the initial mode map.

4. The graphical data conversion system under the multi-mode architecture of a distribution network according to claim 3, characterized in that, The graphical maps generated by the emergent graphical data conversion module are connected to the existing artificial intelligence platform for data training. Based on the data in the graphical maps generated by the emergent graphical data conversion module and the annotation information of power grid operation events, an intelligent model for predicting and monitoring distribution network events is trained.

5. The graphical data conversion system under the multi-mode architecture of a distribution network according to any one of claims 1-4, characterized in that, At the executable data process level, the construction process of this data transformation system can optionally include: initial modality map construction, associated data model construction, and graphical modality construction.

6. The graphical data conversion system under the multi-mode architecture of a distribution network according to claim 5, characterized in that, The initial modality map construction includes: first, initializing a blank map by setting a blank map for data visualization as the initial modality map; dividing the initial modality map into n regions, where n is determined according to the number of data modalities that need to be processed; each region corresponds to a single data modality; and then performing data mapping by sequentially mapping the discrete data points within each selected data modality to the corresponding region.

7. The graphical data conversion system under the multi-mode architecture of a distribution network according to claim 5, characterized in that, The construction of the associated data model includes: for all data modes and their mapped graphical multi-region maps, constructing a data array based on the correlation of data within a mode or across modes; the data array is represented as a matrix, which can be divided into n square matrices through matrix partitioning, each square matrix corresponding to the data correlation within a mode; the data array as a whole corresponds to global data correlation or cross-modal correlation; the data array is mutated according to the number of modes in the current analysis and the data characteristics of each mode; based on the data array, a correlation matrix R is further constructed within each independent modal partition. i To characterize the data correlation within mode i, for cross-modal correlated data, a global correlation matrix R is constructed to characterize the correlation between all data points; the correlation matrix is ​​constructed as a parameter of the linear correlation strength between any two data points when changes occur, and its value is between 0 and 1.

8. The graphical data conversion system under the multi-mode architecture of a distribution network according to claim 5, characterized in that, The graphical modality construction includes: ordering and cross-referencing point-line graphing based on the correlation matrix. To this end, the position of the data in the local or global correlation matrix is ​​first fixed. Then, the ordering and cross-referencing of the correlation matrix is ​​implemented by: sorting and connecting the data points according to the correlation strength, that is, according to the value from 0 to 1, to form a point-line graph containing cross-structures.