Power distribution network graph-real consistency checking system based on multi-source heterogeneous system data collaboration

Abstract

The invention relates to the technical field of data analysis, in particular to a multi-source heterogeneous system data collaborative power distribution network graph-real consistency checking system, which comprises a topological feature extraction module, a cross-system topological alignment module and an anomaly traceability tracking module, wherein the topological feature extraction module is used for extracting multi-dimensional attribute features and topological connection relations of equipment nodes from a multi-source heterogeneous system and generating dynamic fingerprint vectors; the cross-system topology alignment module receives the dynamic fingerprint vector, constructs an intermediate semantic layer and outputs a standard fingerprint vector; the method comprises the following steps: generating a unified topology model by analyzing a differential topology rule of a multi-source heterogeneous system, and calculating a comparison difference value of standard fingerprint vectors among different systems and an equivalent connection relationship of electrical connectivity characteristics; the anomaly traceability tracking module receives the comparison difference value, constructs an anomaly propagation tree and generates an anomaly propagation path; and an abnormal root node is positioned by analyzing and comparing the coupling weight of the equivalent connection relation and the comparison difference value.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a distribution network graph-to-physical consistency verification system for data collaboration of multiple-source heterogeneous systems. Background Art

[0002] In traditional distribution network consistency verification systems, data collaboration between multi-source heterogeneous systems faces significant technical bottlenecks. Due to differences in design goals and functional orientations, different business systems adopt heterogeneous topology modeling rules and data expression paradigms, resulting in semantic gaps and format conflicts in the attribute characteristics of device nodes (such as spatial coordinates and electrical connection relationships) when interacting across systems. Traditional methods usually rely on static rule bases or preset templates to perform one-way mapping of heterogeneous data, making it difficult to dynamically resolve implicit associations in multi-source topological expressions (such as hierarchical connection patterns and dynamic impedance changes), resulting in redundant topological models and distorted equivalence relationships.

[0003] Furthermore, when multi-source data is dynamically updated, real-time verification between systems lacks unified semantic support. Anomaly detection relies solely on a linear analysis mechanism based on local rule matching, failing to capture the cross-level coupling effects of anomaly propagation in complex power grids. Particularly in scenarios where device node states fluctuate frequently, traditional systems, lacking the ability to analyze dynamic fingerprint vectors and coupling weights, struggle to accurately model anomaly propagation paths. Root node identification relies on coarse-grained topology pruning strategies, resulting in low verification efficiency and a high rate of false positives. Summary of the Invention

[0004] The purpose of the present invention is to provide a distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration, so as to solve the problems of low verification efficiency and high misjudgment rate of traditional distribution network graph-to-physical consistency verification systems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a distribution network graph-to-real consistency verification system for multi-source heterogeneous system data collaboration, comprising a topology feature extraction module, a cross-system topology alignment module, and an anomaly tracing module, wherein:

[0006] The topology feature extraction module is used to extract multi-dimensional attribute features and topological connection relationships of device nodes from a multi-source heterogeneous system with differentiated topology expression rules, and generate a dynamic fingerprint vector containing spatial coordinate features and electrical connectivity features;

[0007] The cross-system topology alignment module receives the dynamic fingerprint vector and constructs an intermediate semantic layer, outputting a standard fingerprint vector; generates a unified topology model by analyzing the differentiated topology rules of the multi-source heterogeneous system, and calculates the comparison difference values of the standard fingerprint vectors between different systems in real time during the generation process; the unified topology model includes the equivalent connection relationship of the electrical connectivity feature;

[0008] The anomaly tracing module receives the comparison difference value and constructs an anomaly propagation tree, generates an anomaly propagation path based on the equivalent connection relationship of the electrical connectivity feature in the unified topology model and the coupling weight of the comparison difference value; and locates the root node of the anomaly in the anomaly propagation path in the anomaly propagation tree by analyzing the comparison coupling weight.

[0009] As a further improvement of the present technical solution, the multidimensional attribute characteristics of the device node include device type code, operating parameter characteristic values and topological level identifier; the multidimensional feature combination of device type code, operating parameters and topological level identifier enhances the comprehensiveness of node feature representation and supports the robustness of cross-system semantic alignment.

[0010] As a further improvement of this technical solution, the process of obtaining the topological connection relationship of the device nodes is as follows:

[0011] The topological expression rules of each system in the multi-source heterogeneous system are expressed using the adjacency matrix A k Represented as follows, where k is the index of each system in the multi-source heterogeneous system, and a fusion function ψ is used to process the differentiation rules of each system to generate the fused adjacency matrix A unified , the fusion formula is: A unified =ψ(A1,A2,...,A k ); Using the multi-source adjacency matrix fusion method, the adjacency matrices with differentiated topological expression rules in multi-source heterogeneous systems are fused to resolve the connection relationship conflicts under heterogeneous topological rules and build a globally consistent topological connection network to obtain the topological connection relationship of device nodes.

[0012] As a further improvement of this technical solution, the generation process of the dynamic fingerprint vector is as follows:

[0013] Construct the spatial coordinate feature s according to the topological level identifier in the multidimensional attribute feature of the device node i , where i represents the index of the device node;

[0014] According to the fused adjacency matrix A in the topological connection relationship of the device nodes unified , obtain the electrical connection weight; and obtain the power transmission mode derived from the topological connection path to obtain the power flow distribution characteristics; and obtain the dynamic impedance according to the equivalent impedance value of the topological connection state change; at the same time, combine the electrical connection weight, power flow distribution and dynamic impedance calculation to generate the electrical connectivity index as the electrical connectivity characteristic e i ;

[0015] Weighted fusion of spatial coordinate features s through dynamic attention mechanism i Electrical connection characteristics i , generate dynamic fingerprint vector v i,This process improves the semantic discrimination and anti-interference ability of ,dynamic fingerprint vectors.

[0016] As a further improvement of the present technical solution, the cross-system topology alignment module includes a standard fingerprint vector output unit, which is used for the standard fingerprint vector and specifically includes:

[0017] Dynamic fingerprint vector v for each device node i i Contains spatial coordinate features s i and electrical connectivity characteristics i , that is, v i =(s i ,e i );

[0018] Construct an intermediate semantic layer that transforms the dynamic fingerprint vector v into i Convert to standard fingerprint vector u i , the specific formula is u i =φ(v i );

[0019] The mapping function of the intermediate semantic layer realizes the standardized conversion of heterogeneous fingerprint vectors, eliminates the semantic deviation between multi-source systems, and supports unified modeling.

[0020] As a further improvement of the present technical solution, the cross-system topology alignment module includes a unified topology model generation unit, which is used to generate a unified topology model, specifically including:

[0021] Based on the standard fingerprint vector u after conversion from the intermediate semantic layer i And the fused adjacency matrix A unified , generate a unified topology model G, which includes the equivalent connection relationship of electrical connectivity features: G = (V, E), where V = {u1, u2, ..., u i}, the standard fingerprint vector u converted by the intermediate semantic layer i Composition, E represents the electrical connectivity feature e in the standard fingerprint vector i The equivalent connection relationship;

[0022] A unified topology model is constructed based on standard fingerprint vectors and fused adjacency matrix to achieve equivalent relationship mapping of electrical connectivity and enhance the scalability of the model.

[0023] As a further improvement of the present technical solution, the cross-system topology alignment module includes a comparison difference value calculation unit, which is used to calculate the comparison difference values of the standard fingerprint vectors between different systems in real time during the process of generating a unified topology model, specifically including:

[0024] Use the Euclidean distance metric to measure the difference between the two systems and compare the difference value d(u i,1 ,u i,2 ) is calculated as follows:

[0025] d(u i,1 ,u i,2 )=||u i,1 -u i,2 ||, where u i,1 and u i,2 The same standard fingerprint vector u after conversion from the intermediate semantic layer of two different systems i ;

[0026] Euclidean distance quantifies the topological differences between multi-source systems in real time, providing a dynamic quantitative benchmark for anomaly detection and improving the sensitivity of anomaly identification.

[0027] As a further improvement of the present technical solution, the anomaly tracing module includes an anomaly path generation unit, which is used to generate an anomaly propagation path, specifically including:

[0028] For each device node i, calculate the comparison difference value of its standard fingerprint vectors between different systems according to the calculation formula of the comparison difference value. Select the device node corresponding to the standard fingerprint vector with the largest comparison difference value and use it as the root node of the anomaly propagation tree. Set this root node to n0;

[0029] Create an empty propagation tree T and add n0 to T; according to the fused adjacency matrix A unified , find all directly connected neighbor nodes of n0, and calculate the coupling weight between each neighbor node and n0. The coupling weight is calculated based on the equivalent connection relationship and the comparison difference value of the electrical connectivity characteristics in the unified topology model. Using the coupling weight function, the two are weighted and summed to obtain the corresponding coupling weight; select the neighbor node with the highest coupling weight as the next propagation node and add it to the propagation tree T; repeat the above process until there is no new neighbor node with the highest coupling weight;

[0030] The anomaly propagation path generation mechanism based on coupling weights accurately captures the diffusion law of anomalies in the topological network and optimizes the traceability efficiency.

[0031] As a further improvement of this technical solution, the anomaly tracing module includes an anomaly root source analysis unit, which is used to locate the root node of the anomaly, specifically including:

[0032] Starting from the root node n0, perform a depth-first traversal, record the access path and cumulative coupling weight of each node; evaluate the cumulative coupling weight on each path, and find the path with the highest cumulative coupling weight; and use the starting node of the path with the highest cumulative coupling weight as the root node of the anomaly.

[0033] Depth-first traversal combined with cumulative coupling weight evaluation can achieve multi-dimensional and precise positioning of abnormal root nodes, reducing the risk of misjudgment.

[0034] As a further improvement of this technical solution, the spatial coordinate feature s i The construction process is as follows:

[0035] According to the topological level identifier in the multidimensional attribute characteristics of the device node, the number of layers of each device node is determined, and a radius is assigned to each layer. The device nodes in the same layer are evenly distributed with angles. If there are x device nodes in a layer, the angle interval of each device node is 2π / x. The spatial coordinates of each device node are calculated based on the radius and angle to obtain the relative coding of the geographic coordinates. The spatial coordinate feature s is constructed through the relative coding of the geographic coordinates and the spatial distribution density analysis. i , where i represents the index of the device node;

[0036] The dynamic encoding method of spatial coordinates based on the topological hierarchy ensures the uniqueness and resolvability of the spatial distribution characteristics of device nodes and supports high-precision verification.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] By building dynamic fingerprint vectors and an intermediate semantic layer, we achieve in-depth analysis and semantic unification of the topological features of multi-source heterogeneous systems, breaking through the limitations of traditional static rule mapping. Dynamic fingerprint vectors, generated based on multidimensional attribute features and topological connectivity, effectively integrate the dynamic correlation between spatial coordinates and electrical connectivity, eliminating the semantic gap and addressing the low verification efficiency and high error rate of traditional distribution network consistency verification systems.

[0039] By calculating the difference between standard fingerprint vectors in real time and constructing an anomaly propagation tree, combined with correlation analysis of equivalent connectivity relationships and coupling weights, the root node of the anomaly can be accurately located, significantly improving the efficiency of fault tracing. Furthermore, the unified topology model supports dynamic synchronization and adaptive updating of multi-source data, strengthening the modeling capabilities of cross-level anomaly propagation in complex power grids and ensuring high accuracy and robustness of the verification process. This method provides a full-dimensional, adaptive technical solution for the management of graph-to-physical consistency in multi-source heterogeneous distribution networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of the overall module of the present invention;

[0041] Figure 2 A schematic diagram of a cross-system topology alignment module unit of the present invention;

[0042] Figure 3 This is a schematic diagram of the abnormality tracing module unit of the present invention;

[0043] In the figure: 100, topological feature extraction module; 200, cross-system topological alignment module; 201, standard fingerprint vector output unit; 202, unified topological model generation unit; 203, comparison difference value calculation unit; 300, anomaly tracing module; 301, anomaly path generation unit; 302, anomaly root cause analysis unit. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making any creative efforts shall fall within the scope of protection of the present invention.

[0045] Next, see Figure 1 The present invention provides a technical solution: a distribution network graph consistency verification system for multi-source heterogeneous system data collaboration, including a topology feature extraction module 100, a cross-system topology alignment module 200 and an anomaly tracing module 300.

[0046] The topology feature extraction module 100 is used to extract the multi-dimensional attribute features and topological connection relationships of device nodes from a multi-source heterogeneous system with differentiated topology expression rules, and generate a dynamic fingerprint vector containing spatial coordinate features and electrical connectivity features, specifically including:

[0047] Multi-source heterogeneous systems include but are not limited to geographic information systems, data acquisition and monitoring systems, and asset management systems; by extracting device type codes, operating parameter characteristic values, and topological level identifiers in multi-source heterogeneous systems, the multi-dimensional attribute characteristics of device nodes are obtained.

[0048] Since multi-source heterogeneous systems (geographic information systems, data acquisition and monitoring systems, and asset management systems) have differentiated topological expression rules, the multi-source adjacency matrix fusion method is used to analyze these differentiated rules, and the topological expression rules of each system in the multi-source heterogeneous system are expressed using the adjacency matrix A. k Represented as follows, where k is the index of each system in the multi-source heterogeneous system, and a fusion function ψ is used to process the differentiation rules of each system to generate the fused adjacency matrix A unified , the fusion formula is: A unified=ψ(A1,A2,...,A k ); Using the multi-source adjacency matrix fusion method, the adjacency matrices with differentiated topology expression rules in the multi-source heterogeneous system are fused to obtain the topological connection relationship of the device nodes.

[0049] Based on the multi-dimensional attribute characteristics and topological connection relationship of the device nodes, a dynamic fingerprint vector containing spatial coordinate characteristics and electrical connectivity characteristics is generated. The generation process is as follows:

[0050] According to the topological level identifier in the multidimensional attribute characteristics of the device node, the number of layers of each device node is determined, and a radius is assigned to each layer. The device nodes in the same layer are evenly distributed with angles. If there are x device nodes in a layer, the angle interval of each device node is 2π / x. The spatial coordinates of each device node are calculated based on the radius and angle to obtain the relative coding of the geographic coordinates. The spatial coordinate feature s is constructed through the relative coding of the geographic coordinates and the spatial distribution density analysis. i , where i represents the index of the device node.

[0051] According to the fused adjacency matrix A in the topological connection relationship of the device nodes unified , obtain the electrical connection weight; and obtain the power transmission mode derived from the topological connection path to obtain the power flow distribution characteristics; and obtain the dynamic impedance according to the equivalent impedance value of the topological connection state change; at the same time, combine the electrical connection weight, power flow distribution and dynamic impedance calculation to generate the electrical connectivity index as the electrical connectivity characteristic e i .

[0052] Weighted fusion of spatial coordinate features s through dynamic attention mechanism i Electrical connection characteristics i , and introduces a time window function to capture topological state changes, and finally generates a dynamic fingerprint vector v that can characterize the spatial position, electrical connection and real-time topological evolution of the device i , realizing the collaborative representation and dynamic update of multi-dimensional heterogeneous features.

[0053] See also Figure 2 The standard fingerprint vector output unit 201 in the cross-system topology alignment module 200 receives the dynamic fingerprint vector and constructs an intermediate semantic layer, and outputs the standard fingerprint vector, specifically including:

[0054] Dynamic fingerprint vector v for each device node i i Contains spatial coordinate features s i and electrical connectivity characteristics i , that is, v i =(s i ,e i ), where the electrical connectivity characteristic e iIncluding electrical connection weights, power flow distribution characteristics and equivalent impedance values.

[0055] In order to process the differentiated data of multi-source heterogeneous systems, an intermediate semantic layer is constructed, which transforms the dynamic fingerprint vector v into i Convert to standard fingerprint vector u i , the specific formula is u i =φ(v i ); The purpose of the intermediate semantic layer is to standardize the data in multi-source heterogeneous systems to facilitate subsequent unified processing.

[0056] The unified topology model generation unit 202 in the cross-system topology alignment module 200 generates a unified topology model by parsing the differentiated topology rules of the multi-source heterogeneous systems, specifically including:

[0057] The fused adjacency matrix A unified Used to reflect the differentiated topology rules of multi-source heterogeneous systems; based on the standard fingerprint vector u converted from the intermediate semantic layer i And the fused adjacency matrix A unified , generate a unified topology model G, which includes the equivalent connection relationship of electrical connectivity features: G = (V, E), where V = {u1, u2, ..., u i}, the standard fingerprint vector u converted by the intermediate semantic layer i Composition, E represents the electrical connectivity feature e in the standard fingerprint vector i The equivalent connection relationship.

[0058] The comparison difference value calculation unit 203 in the cross-system topology alignment module 200 calculates the comparison difference value of the standard fingerprint vectors between different systems in real time during the process of generating the unified topology model, and uses the Euclidean distance metric to measure the difference between the two systems. The comparison difference value d(u i,1 ,u i,2 ) is calculated as follows:

[0059] d(u i,1 ,u i,2 )=||u i,1 -u i,2 ||, where u i,1 and u i,2 The same standard fingerprint vector u after conversion from the intermediate semantic layer of two different systems i .

[0060] See also Figure 3The abnormal path generation unit 301 in the abnormality tracing module 300 receives the comparison difference value and constructs an abnormality propagation tree. The abnormality propagation path is generated based on the equivalent connection relationship of the electrical connectivity feature in the unified topology model and the coupling weight of the comparison difference value, specifically including:

[0061] For each device node i, the comparison difference value of its standard fingerprint vector between different systems is calculated according to the calculation formula of the comparison difference value. The device node corresponding to the standard fingerprint vector with the largest comparison difference value is selected and used as the root node of the anomaly propagation tree, and the root node is set to n0.

[0062] Create an empty propagation tree T and add n0 to T; according to the fused adjacency matrix A unified , find all directly connected neighbor nodes of n0, and calculate the coupling weight between each neighbor node and n0. The coupling weight is calculated based on the equivalent connection relationship and the comparison difference value of the electrical connectivity characteristics in the unified topology model. Using the coupling weight function, the two are weighted and summed to obtain the corresponding coupling weight; select the neighbor node with the highest coupling weight as the next propagation node and add it to the propagation tree T; repeat the above process until there is no new neighbor node with the highest coupling weight.

[0063] The abnormality root source analysis unit 302 in the abnormality tracing module 300 locates the root node of the abnormality in the abnormality propagation path in the abnormality propagation tree by analyzing and comparing the coupling weights. The specific steps are as follows:

[0064] Starting from the root node n0, a depth-first traversal is performed to record the access path and cumulative coupling weight of each node; the cumulative coupling weight on each path is evaluated to find the path with the highest cumulative coupling weight; the starting node of the path with the highest cumulative coupling weight is used as the root node of the anomaly.

[0065] The above is an embodiment of the distribution network graph consistency verification system for multi-source heterogeneous system data collaboration provided by the present invention. As can be seen from the description of the above embodiment, it realizes the deep analysis and semantic unification of the topological characteristics of multi-source heterogeneous systems through the construction of dynamic fingerprint vectors and intermediate semantic layers, breaking through the limitations of traditional static rule mapping. The dynamic fingerprint vectors generated based on multi-dimensional attribute features and topological connection relationships effectively integrate the dynamic correlation characteristics of spatial coordinates and electrical connectivity, eliminating the semantic gap and solving the problems of low verification efficiency and high error rate of traditional distribution network graph consistency verification systems.

[0066] Furthermore, by calculating the difference values of standard fingerprint vectors in real time and constructing an anomaly propagation tree, combined with correlation analysis of equivalent connection relationships and coupling weights, the root node of the anomaly can be accurately located, significantly improving the efficiency of fault tracing. At the same time, the unified topology model supports dynamic synchronization and adaptive updating of multi-source data, strengthening the modeling capabilities of cross-level anomaly propagation in complex power grids and ensuring high accuracy and robustness of the verification process. This method provides a full-dimensional, adaptive technical solution for the management of graph-to-physical consistency in multi-source heterogeneous distribution networks.

[0067] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration, characterized by: include: A topological feature extraction module (100) is used to extract multi-dimensional attribute features and topological connection relationships of device nodes from a multi-source heterogeneous system with differentiated topological expression rules, and generate a dynamic fingerprint vector including spatial coordinate features and electrical connectivity features; A cross-system topology alignment module (200) is configured to receive the dynamic fingerprint vector and construct an intermediate semantic layer, and output a standard fingerprint vector; generate a unified topology model by analyzing the differentiated topology rules of the multi-source heterogeneous system, and calculate the comparison difference values of the standard fingerprint vectors between different systems in real time during the generation process; the unified topology model includes an equivalent connection relationship of electrical connectivity features; An abnormality tracing module (300) is used to receive the comparison difference value and construct an abnormality propagation tree, and generate an abnormality propagation path based on the equivalent connection relationship of the electrical connectivity feature in the unified topology model and the coupling weight of the comparison difference value; By analyzing and comparing the coupling weights, the root node of the anomaly is located in the anomaly propagation path in the anomaly propagation tree.

2. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The multi-dimensional attribute characteristics of the device node include a device type code, an operating parameter characteristic value, and a topology level identifier.

3. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The process of obtaining the topological connection relationship of the device nodes is as follows: The topological expression rules of each system in the multi-source heterogeneous system are expressed using the adjacency matrix A k Represented as follows, where k is the index of each system in the multi-source heterogeneous system, and a fusion function ψ is used to process the differentiation rules of each system to generate the fused adjacency matrix A unified , the fusion formula is: A unified =ψ(A1,A2,...,A k ); Using the multi-source adjacency matrix fusion method, the adjacency matrices with differentiated topology expression rules in the multi-source heterogeneous system are fused to obtain the topological connection relationship of the device nodes.

4. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The generation process of the dynamic fingerprint vector is as follows: Construct the spatial coordinate feature s according to the topological level identifier in the multidimensional attribute feature of the device node i , where i represents the index of the device node; According to the fused adjacency matrix A in the topological connection relationship of the device nodes unified , obtain the electrical connection weight; and obtain the power flow distribution characteristics through the power transmission mode derived from the topological connection path; And according to the equivalent impedance value of the topological connection state change, the dynamic impedance is obtained; at the same time, the electrical connection weight, power flow distribution and dynamic impedance calculation are combined to generate the electrical connectivity index as the electrical connectivity feature e i ; Weighted fusion of spatial coordinate features s through dynamic attention mechanism i Electrical connection characteristics i , generate dynamic fingerprint vector v i .

5. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The cross-system topology alignment module (200) comprises a standard fingerprint vector output unit (201), and the standard fingerprint vector output unit (201) is used for standard fingerprint vectors, specifically comprising: Dynamic fingerprint vector v for each device node i i Contains spatial coordinate features s i and electrical connectivity characteristics i , that is, v i =(s i ,e i ); Construct an intermediate semantic layer that transforms the dynamic fingerprint vector v into i Convert to standard fingerprint vector u i , the specific formula is u i =φ(v i ).

6. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The cross-system topology alignment module (200) comprises a unified topology model generation unit (202), wherein the unified topology model generation unit (202) is used to generate a unified topology model, specifically comprising: Based on the standard fingerprint vector u after conversion from the intermediate semantic layer i And the fused adjacency matrix A unified , generate a unified topology model G, which includes the equivalent connection relationship of electrical connectivity features: G = (V, E), where V = {u1, u2, ..., u i }, the standard fingerprint vector u converted by the intermediate semantic layer i Composition, E represents the electrical connectivity feature e in the standard fingerprint vector i The equivalent connection relationship.

7. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The cross-system topology alignment module (200) comprises a comparison difference value calculation unit (203), which is used to calculate the comparison difference values of standard fingerprint vectors between different systems in real time during the process of generating a unified topology model, specifically comprising: Use the Euclidean distance metric to measure the difference between the two systems and compare the difference value d(u i,1 ,u i,2 ) is calculated as follows: d(u i,1 ,u i,2 )=||u i,1 -u i,2 ||, where u i,1 and u i,2 The same standard fingerprint vector u after conversion from the intermediate semantic layer of two different systems i .

8. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The abnormality tracing module (300) includes an abnormality path generating unit (301), and the abnormality path generating unit (301) is used to generate an abnormality propagation path, specifically including: For each device node i, calculate the comparison difference value of its standard fingerprint vectors between different systems according to the calculation formula of the comparison difference value. Select the device node corresponding to the standard fingerprint vector with the largest comparison difference value and use it as the root node of the anomaly propagation tree. Set this root node to n0; Create an empty propagation tree T and add n0 to T; according to the fused adjacency matrix A unified , find all directly connected neighbor nodes of n0, and calculate the coupling weight between each neighbor node and n0. The coupling weight is calculated based on the equivalent connection relationship and the comparison difference value of the electrical connectivity characteristics in the unified topology model. Using the coupling weight function, the two are weighted and summed to obtain the corresponding coupling weight; select the neighbor node with the highest coupling weight as the next propagation node and add it to the propagation tree T; repeat the above process until there is no new neighbor node with the highest coupling weight.

9. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 1 is characterized in that: The abnormality tracing module (300) includes an abnormality root source analysis unit (302), and the abnormality root source analysis unit (302) is used to locate the root node of the abnormality, specifically including: Starting from the root node n0, a depth-first traversal is performed to record the access path and cumulative coupling weight of each node; the cumulative coupling weight on each path is evaluated to find the path with the highest cumulative coupling weight; the starting node of the path with the highest cumulative coupling weight is used as the root node of the anomaly.

10. The distribution network graph-to-physical consistency verification system for multi-source heterogeneous system data collaboration according to claim 4 is characterized in that: The spatial coordinate feature s i The construction process is as follows: According to the topological level identifier in the multi-dimensional attribute characteristics of the device node, the layer number of each device node is determined, a radius is assigned to each layer, and the device nodes in the same layer are evenly distributed in angle; If there are x device nodes in a layer, the angular interval of each device node is 2π / x. The spatial coordinates of each device node are calculated based on the radius and angle to obtain the relative coding of the geographic coordinates. The spatial coordinate feature s is constructed by analyzing the relative coding of the geographic coordinates and the spatial distribution density. i , where i represents the index of the device node.

Images