Ring main unit digital twin construction system for smart power distribution network

Abstract

The application relates to the technical field of digital twin construction, and particularly discloses a ring network cabinet digital twin construction system for an intelligent power distribution network. The system sets up a semantic template library of ring network cabinet objects, interval objects, busbar objects, switch objects, feeder objects and operation measurement objects, parses account data, primary connection relationships, dispatching topologies, remote signaling data, remote measurement data and operation records into standard fields with source identifiers and time identifiers through a data processing program, generates a unified entity index and a semantic topology graph, and performs consistency checking according to busbar connectivity, switch conduction, upstream and downstream intervals, ring network breaking points and feeder attribution constraints; conflict relationships are marked and processed in combination with source credibility, update time, topology accessibility and historical version continuity to generate a ring network cabinet digital twin instance with version records. The system can reduce entity duplication, topology connection errors and measurement binding offsets.

Description

Technical Field

[0001] This invention relates to the field of digital twin construction technology, and specifically discloses a digital twin construction system for ring main units in smart distribution networks. Background Technology

[0002] In smart distribution networks, ring main units (RNBs) represent operational nodes related to feeder segmentation, interconnection and power transfer, and fault isolation. Their digital twins typically need to simultaneously reflect equipment objects, primary wiring relationships, dispatch topology, and operational measurement data. Existing construction methods are mostly based on asset ledgers, recording objects such as cabinets, bays, busbars, switches, and feeders into a database, and then attaching dispatch numbers, remote signaling point numbers, telemetry point numbers, and operation records as attributes to the corresponding objects. To form a visualization model, the system usually establishes inclusion and connection relationships between objects based on primary wiring diagrams or manually maintained topology tables, and periodically reads operational data using a data synchronization program to refresh the status fields in the twin. The above methods can complete the basic object display and operational status mapping, but the model generation process depends on the consistency of fields between multiple data sources and the quality of manual configuration. Data from different sources only maintains a loose correspondence through table fields or element numbers, lacking computable semantic constraints oriented towards electrical connection relationships.

[0003] The most common technical solution typically employs a template-based modeling approach to construct a digital twin of a ring main unit (RNB). Modelers first create an equipment template based on the RNB type, configuring the template with cabinet hierarchy, bay category, switch status fields, feeder name fields, and measurement point binding fields. Then, a data extraction program retrieves basic data from the ledger system, distribution network automation system, and dispatch system, generating object instances using name matching, number matching, or point number matching. For primary wiring relationships, connection edges are usually read from drawing analysis results, topology tables, or manually entered tables, and a tree structure or graphical topology is generated according to the parent-child relationships between object instances. When topology relationships are missing, maintenance personnel typically supplement or modify the mapping table. After RNB reconnection, feeder switching, or bay expansion, the system re-imports the tables or re-releases the model version to maintain the usability of the digital twin. The core processing in this approach remains object instantiation and field binding; topology relationships serve only as data carriers for model display or status queries.

[0004] Existing solutions typically employ field integrity checks, name duplication checks, point number existence checks, and general graph connectivity checks during data verification. For the same ring main unit object from different systems, if there are differences in name, number, or point number, the data retention is usually determined by fixed priority, manual confirmation, or update time overwrite. For inconsistencies between primary wiring relationships and scheduling topology, the conventional approach is to retain the connection relationships from the primary data source and mark the remaining sources as pending records. General graph verification can only determine whether nodes are isolated, edges are duplicated, or connections are broken; it struggles to express electrical topology semantics such as bus connectivity, switch conduction, upstream and downstream bays, ring network breakpoints, and feeder affiliation. While the resulting twins possess object hierarchy and status fields, the node, edge, and measurement binding relationships in the graph structure have not undergone electrical semantic consistency constraints, potentially leading to the continued use of incorrect connection relationships during subsequent operational data access.

[0005] As can be seen from the existing technologies described above, the main technical problem lies in the following: During the construction of the digital twin of the ring main unit, there are inconsistencies in naming rules, entity identifiers, connection source conflicts, and version updates among the ledger data, primary wiring relationships, dispatch topology, remote signaling data, telemetry data, and operation records. Existing construction methods lack a constraint mechanism to convert multi-source heterogeneous data into a unified entity index and semantic topology graph, and also lack a computational mechanism for consistency verification based on bus connectivity, switch conduction, upstream and downstream intervals, ring network breakpoints, and feeder affiliation. This leads to the generated digital twin being prone to duplicate equipment entities, incorrect topology edge connections, measurement data binding offsets, and broken topology version records. The essence of these problems is that the model construction process does not incorporate data source differences, entity merging, and electrical topology verification into the same calculation chain, making it difficult to ensure that the digital twin of the ring main unit maintains consistency with the actual electrical connection relationships of the smart distribution network. Summary of the Invention

[0006] The purpose of this invention is to provide a digital twin construction system for ring main units for smart distribution networks, which can solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A system for constructing digital twins of ring main units for smart distribution networks includes a semantic template library storing ring main unit objects, bay objects, bus objects, switch objects, feeder objects, and operational measurement objects, as well as a data processing program connected to the semantic template library. The data processing program is used to parse ledger data, primary wiring relationships, dispatch topology, remote signaling data, telemetry data, and operation records into standard fields with source identifiers and time identifiers, respectively. Based on the standard fields, a unified entity index and a semantic topology graph are generated. The semantic topology graph is checked for consistency according to bus connectivity, switch conduction, upstream and downstream bays, ring network breakpoints, and feeder affiliation constraints. Based on the checked nodes, edges, and data binding relationships, a digital twin instance of the ring main unit with a topology version record is generated.

[0008] Preferably, the unified entity index is generated as follows: the cabinet name, interval name, equipment number, and asset code in the ledger data are normalized; the terminal connection identifier in the primary wiring relationship and the scheduling number in the scheduling topology are associated with the same name but different codes; measurement type labels are established for the point numbers in the remote signaling data and the remote measurement data; and the normalized fields, association results, and measurement type labels are written into the same candidate entity set; in the candidate entity set, the main entity identifier is determined according to the equipment subordinate level and wiring adjacency relationship, and data items with the same main entity identifier are merged into the same unified entity index.

[0009] Preferably, the semantic topology graph includes nodes corresponding to the ring main unit object, the bay object, the busbar object, the switch object, the feeder object, and the operating measurement object, as well as edges representing cabinet inclusion, bay affiliation, busbar connection, switch connection, feeder power supply direction, and measurement binding; the data processing program generates node types according to the object hierarchy definition in the semantic template library, generates edge types according to the primary wiring relationship and the scheduling topology, and writes a source identifier, time identifier, and relationship confidence value for each edge.

[0010] Preferably, the consistency verification includes forming a bus connectivity subgraph, a switch status subgraph, a feeder affiliation subgraph, and a ring network breakpoint subgraph based on the semantic topology graph; the data processing program verifies the reachability of nodes of the same bus segment in the bus connectivity subgraph, verifies the consistency between the switch opening / closing state and the edge connectivity attribute in the switch status subgraph, verifies the directional continuity of upstream and downstream intervals in the feeder affiliation subgraph, verifies the correspondence between breakpoints and feeder boundaries in the ring network breakpoint subgraph, and writes the verification results of each subgraph into a topology verification table with node and edge identifiers.

[0011] Preferably, when a conflict arises during the consistency check, the data processing program configures a source credibility factor, an update time factor, a topology reachability factor, and a historical version continuity factor for the conflict relationship, and generates a conflict weight according to the source credibility factor, the update time factor, the topology reachability factor, and the historical version continuity factor; the conflict relationship is marked as a reserved relationship, a pending confirmation relationship, or an abnormal relationship according to the conflict weight, the reserved relationship participates in the generation of the ring network cabinet digital twin instance, and the pending confirmation relationship and the abnormal relationship are written into the check record.

[0012] Preferably, the unified entity index further forms multi-source matching features through name segmentation results, number fragments, point number prefixes, cabinet location descriptions, and feeder names; the data processing program constructs feature signatures for the multi-source matching features and writes the feature signatures together with the device hierarchy and wiring adjacency relationship into the entity merging matrix; in the entity merging matrix, data items with the same main entity identifier and wiring adjacency relationship satisfying the single reachable path condition from the cabinet node through the interval node and switch node to the bus node are merged, and data items that do not satisfy the single reachable path condition are assigned to the ambiguous entity set.

[0013] Preferably, the bus connectivity subgraph, the switch status subgraph, the feeder affiliation subgraph, and the ring network breakpoint subgraph are each configured with boundary nodes. The boundary nodes consist of feeder entry nodes, connecting switch nodes, segmented switch nodes, and end interval nodes. The data processing program generates a set of reachable paths using the boundary nodes as start and end constraints, and cross-compares the set of reachable paths with the edge types in the semantic topology graph. When the same node belongs to more than two sets of reachable paths, the path affiliation is determined based on the connectivity attribute in the switch status subgraph.

[0014] Preferably, the historical version continuity factor is generated from the node existence status, edge existence status, data binding status, and feeder affiliation status in the previous topology version. The data processing program couples the historical version continuity factor with the topology reachability factor for calculation, sets a version inheritance identifier for the retained relationship, sets a manual review identifier for the relationship to be confirmed, and sets a conflict chain identifier for the abnormal relationship. The conflict chain identifier includes a conflict node identifier, a conflict edge identifier, a conflict source identifier, and a conflict occurrence time.

[0015] Preferably, the reachable path set further includes candidate paths introduced by the ambiguous entity set. The data processing program establishes a path relationship table with the candidate paths and the reserved relationships, the unconfirmed relationships, and the abnormal relationships. In the path relationship table, when the edge corresponding to the same candidate path has both the reserved relationship and the unconfirmed relationship, the edge corresponding to the reserved relationship is retained and the data binding corresponding to the unconfirmed relationship is frozen. When the edge corresponding to the same candidate path has both the abnormal relationship, a local topology isolation segment associated with the conflict chain identifier is generated.

[0016] Preferably, the topology version record includes an entity index version, a semantic topology graph version, a consistency check version, and a data binding version; when the data processing program generates a new digital twin instance of the ring network cabinet, it writes the local topology isolation fragment, the path relationship table, and the conflict chain identifier into the incremental update queue, and limits the nodes to be updated according to the change range of the entity index version, limits the edges to be updated according to the change range of the semantic topology graph version, and limits the measurement relationships to be updated according to the change range of the data binding version.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. After adopting the aforementioned ring main unit digital twin construction scheme, ledger data, primary wiring relationships, scheduling topology, remote signaling data, telemetry data, and operation records are parsed into standard fields with source identifiers and time identifiers, and entered into the same semantic topology graph via a unified entity index. The inclusion, attribution, connection, conduction, power supply direction, and measurement binding relationships between cabinets, bays, buses, switches, feeders, and operational measurement objects can be expressed within the same data structure. The semantic topology graph is jointly constrained by bus connectivity, switch conduction, upstream and downstream bays, ring network breakpoints, and feeder attribution constraints during generation. Conflict relationships are weighted based on source credibility, update time, topology reachability, and historical version continuity before being included in relationship retention or anomaly marking. Therefore, the node, edge, and data binding relationships in the twin instance no longer rely solely on name matching or manual table configuration. The correspondence between topological edges and measurement points is written into the model after electrical semantic verification. This can reduce the situation of repeated modeling of the same equipment, incorrect connection of feeder boundaries, inconsistency between switch status and connectivity, and offset of remote signaling and telemetry point binding. This enables the digital twin to maintain a more stable and consistent relationship with the actual electrical topology of the ring main unit, and allows conflicting data to be structurally classified during the model generation stage.

[0018] 2. Regarding secondary technical effects, the unified entity index, semantic topology graph, consistency verification results, and data binding relationships are each recorded as version records. When ring main unit reconnects, feeder switches, intervals are extended, or data sources are updated, the model update scope can be limited based on the entity index version, semantic topology graph version, and data binding version. The settings for ambiguous entity sets, candidate paths, path relationship tables, and local topology isolation segments ensure that conflicting data items do not directly overwrite the original model or are simply discarded. Instead, they are recorded traceably along with conflicting nodes, conflicting edges, conflict sources, and conflict occurrence times. The incremental update queue organizes model updates according to the range of node changes, edge changes, and measurement relationship changes. This reduces the computational burden of fully rebuilding the semantic topology graph after each data change and ensures that abnormal relationships, pending confirmation relationships, and retained relationships are stored in the same version chain. For subsequent topology verification, model rollback, and rebinding of running data, the system can locate the change position based on the version chain, reducing data mismatches during repeated comparisons and manual searches, and providing a stable data update boundary for the continuous maintenance of the ring main unit digital twin. By including isolated local topology segments in the incremental update queue, conflicting connection edges and measurement relationships can be restricted to local processing, avoiding irrelevant nodes from being updated along with them. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the overall construction process of the multi-source heterogeneous data-driven digital twin of the ring main unit according to the present invention. Figure 2 A flowchart for generating the unified entity index and ambiguous entity set of this invention; Figure 3 This is a flowchart of the semantic topology graph subgraph splitting and consistency verification process of the present invention; Figure 4 This is a flowchart illustrating the conflict handling, version recording, and incremental update process of this invention. Detailed Implementation

[0020] refer to Figure 1In one embodiment, a ring main unit (RNB) digital twin construction system for smart distribution networks includes a semantic template library storing RNB objects, bay objects, busbar objects, switch objects, feeder objects, and operational measurement objects, and a data processing program connected to the semantic template library. The semantic template library stores the modeling rules for the RNB digital twin using object type, object attributes, object hierarchy, and object relationships as basic data structures. The data processing program receives ledger data, primary wiring relationships, dispatch topology, remote signaling data, telemetry data, and operation records. It then parses the equipment name, cabinet number, bay number, dispatch number, point number, operating time, status value, and measurement value from different sources into standard fields with source identifiers and time identifiers. The source identifier distinguishes between the ledger system, primary wiring diagram parsing results, dispatch topology data, remote signaling acquisition records, telemetry acquisition records, and operation logs. The time identifier records the data generation time, change time, or acquisition time. The standard fields are written into... After a unified data cache is established, the data processing program generates ring main unit object nodes, bay object nodes, bus object nodes, switch object nodes, feeder object nodes, and operational measurement object nodes according to the object type definitions in the semantic template library. It then generates edges in the semantic topology graph based on object dependency relationships, primary connection relationships, switch conduction relationships, feeder power supply direction relationships, and operational data binding relationships. After the semantic topology graph is generated, the data processing program calls the bus connectivity, switch conduction, upstream and downstream bays, ring network breakpoints, and feeder affiliation constraints for consistency verification. The verified nodes, edges, and data binding relationships are written into the ring main unit digital twin instance. The ring main unit digital twin instance simultaneously stores topology version records, data binding records, and verification records. This embodiment completes multi-source data parsing, entity index generation, semantic topology graph construction, electrical topology consistency verification, and twin instance generation through the same computational chain, avoiding equipment duplication, connection errors, and data binding offsets caused by relying solely on manual input or single-field matching.

[0021] Specifically, the semantic template library uses ring main unit (RMU) objects to represent the overall entity of the RMU as a distribution network node device; bay objects to represent incoming bays, outgoing bays, tie bays, metering bays, or other functional bays related to electrical connections within the RMU; busbar objects to represent busbar segments or busbar connection areas within the RMU that provide electrical connections; switch objects to represent circuit breakers, load switches, disconnectors, or switch entities related to conduction states; feeder objects to represent feeder entities that have a power supply direction relationship with the RMU; and operational measurement objects to represent remote signaling points, telemetry points, and status and analog quantities in operational records. The semantic template library does not limit the specific hardware structure of the equipment but uses the computable data relationships between objects as template content. When the data processing program parses the ledger data... The system extracts cabinet name, asset code, installation location description, bay name, and equipment number. When parsing primary wiring relationships, it extracts connection terminals, connection directions, and graphic element identifiers. When parsing scheduling topology, it extracts scheduling number, feeder name, upstream and downstream nodes, and breakpoint descriptions. When parsing remote signaling and telemetry data, it extracts point number, measurement type, acquisition time, and numerical fields. When parsing operation records, it extracts change events, pre-operation status, post-operation status, and recording time. After these fields enter the standard field structure, they do not directly form the final digital twin. Instead, a unified entity index first determines the same equipment entity to which each data item belongs. In this embodiment, the semantic template library serves as a constraint entry point before multi-source data enters the digital twin, ensuring that ledgers, topology, and operation data have a unified data interpretation foundation from the initial modeling stage.

[0022] refer to Figure 2In this embodiment, the unified entity index is generated according to the candidate entity set. The data processing program performs field normalization on the cabinet name, interval name, equipment number, and asset code from the ledger data. Field normalization includes removing irrelevant spaces, unifying Chinese full-width and half-width characters, unifying the expression of connectors in the numbering, extracting numerical fragments and Chinese position fragments from the feeder name, and retaining words used to distinguish object levels. The data processing program performs homonymous and heteronymous code association on the terminal connection identifiers in the primary wiring relationship and the scheduling numbers in the scheduling topology. The homonymous and heteronymous code association refers to the merging of multiple identifiers generated by different numbering systems in different data sources for the same ring network cabinet or the same interval into the candidate entity set. The data processing program establishes point number construction in the remote signaling data and telemetry data. The measurement type label includes switch open / closed status, protection action status, current, voltage, power, and other types related to the operating measurement object. Each candidate data item in the candidate entity set includes a normalized field, source identifier, time identifier, measurement type label, and subordinate level information. The data processing program determines the main entity identifier based on the device subordinate level and wiring adjacency relationship. Data items with the same main entity identifier are merged into the same unified entity index. During merging, the original fields and standard fields of each data item are retained to avoid losing source information when tracing subsequent versions. In this embodiment, the unified entity index incorporates name matching, number matching, point number matching, and wiring relationship into the entity merging process, so that object merging is not limited to a single field being the same.

[0023] In the generation of a unified entity index, the following entity matching score formula can be used to calculate the attribution relationship between candidate data items, and the formula is written as:

[0024] in, Represents candidate data items With candidate data items Entity matching score between them Indicates normalized name similarity. This indicates the consistency of the device's hierarchical hierarchy. Indicates the consistency of adjacent wiring connections. This indicates the consistency between the point number and the scheduling number. , , and Indicates a preset weight and satisfies In the example, take , , , When the normalized name similarity of two candidate data items is The consistency of equipment hierarchical level is The consistency of wiring adjacency is The consistency of the point number or scheduling number is At that time, the calculation yielded The calculation results are used to determine whether candidate data items enter the same main entity identifier candidate range. The physical meaning is to convert text similarity, hierarchical consistency, connection adjacency and number association into a calculable basis for entity attribution.

[0025] In one embodiment, the semantic topology graph expresses the data structure of the ring main unit's digital twin using a set of nodes, a set of edges, and a set of bindings. The node set corresponds to ring main unit objects, bay objects, busbar objects, switch objects, feeder objects, and operational measurement objects. The edge set corresponds to cabinet inclusion, bay affiliation, busbar connection, switch connection, feeder power supply direction, and measurement binding. The binding set corresponds to the mapping relationship between remote signaling points, telemetry points, and operation records and specific equipment entities. The data processing program generates node types according to the object hierarchy definition in the semantic template library, generates edge types according to the primary wiring relationship and scheduling topology, and writes a source identifier, time identifier, and relationship confidence value for each edge. The relationship confidence value is not solely determined by the data source level. The decision is not made by the edge source, but by the entity matching score of the node connected by the edge, the occurrence of the edge in a single wiring relationship, the occurrence of the edge in the scheduling topology, and the response relationship between the edge and the running measurement object. When the semantic topology graph is saved, each edge is set with edge start point, edge end point, edge type, edge direction, edge status, source identifier, time identifier, relationship confidence value, and version identifier. The edge direction is used to express the power supply direction of the feeder or the upstream and downstream interval relationship. The edge status is used to express the switch conduction relationship or whether the edge participates in the current topology instance. The version identifier is used to distinguish the connection relationship of the same object in different time periods. The semantic topology graph in this embodiment transforms the ring network cabinet digital twin from a static object set into a verifiable graph data structure.

[0026] Table 1 shows an object relationship configuration method in a semantic template library, which is used to define the basis for generating nodes and edges in a semantic topology graph.

[0027] Table 1. Object Relationship Configuration Methods in the Semantic Template Library

[0028] Based on the configuration shown in Table 1, when generating the semantic topology diagram, the data processing program uses the ring main unit object node as the upper-level node of the object level, the bay object node as the functional partition node under the cabinet node, the bus object node and the switch object node as the wiring relationship calculation node, the feeder object node as the power supply direction and belonging range calculation node, and the operating measurement object node as the status value and measurement value access node. The cabinet containing edge is only generated between the ring main unit object node and the bay object node. The bay belonging edge is generated between the bay object node and the switch object node or bus object node inside it. The bus connection edge is generated according to the primary wiring relationship. The switch connection edge is generated according to the opening and closing status of the switch object. The feeder power supply direction edge is generated according to the upstream and downstream description in the scheduling topology. The measurement binding edge is generated according to the correspondence between the remote signaling point number, the telemetry point number and the main entity identifier. The association verification constraints in Table 1 are called in the subsequent consistency verification and participate in the determination of whether the edge is retained. This embodiment limits the source and usage boundary of different types of edges through the object relationship configuration method to avoid mixing connection relationships with different semantics.

[0029] refer to Figure 3 In one embodiment, consistency verification includes forming a bus connectivity subgraph, a switch state subgraph, a feeder affiliation subgraph, and a ring network breakpoint subgraph based on the semantic topology graph. The data processing program verifies the reachability of nodes within the same bus segment in the bus connectivity subgraph, verifies the consistency between the switch open / closed state and edge connectivity attributes in the switch state subgraph, verifies the directional continuity of upstream and downstream intervals in the feeder affiliation subgraph, and verifies the correspondence between breakpoints and feeder boundaries in the ring network breakpoint subgraph. The verification results of each subgraph are then written into a topology verification table with node and edge identifiers. The bus connectivity subgraph extracts bus object nodes, switch object nodes connected to the bus, and other nodes connected to the bus from the semantic topology graph. The data processing program uses different constraint rules to verify each subgraph and writes the verification results into a unified topology verification table. This embodiment reduces the interference between different semantic relationships during full-map verification by splitting the subgraphs, allowing different electrical topology semantics to be calculated within their respective boundaries.

[0030] The switch conduction relationship can be expressed using the following formula for calculating edge connectivity, which is written as: in, Represents a node With nodes The connectivity value between them at the current verification time. Representing nodes in a semantic topology graph With nodes If there is a basic connecting edge between them, then take the... When there are no basic connecting edges, take , This represents the switch object that controls the connection edge. The corresponding state value is taken when the switch is in the ON state. When the switch is in the non-conducting state, take In the example, if there is a basic connection edge between a certain interval node and a bus node, and the corresponding switch object is in the on state, then... If the basic connection edge exists but the switch object is in a non-conducting state, then The calculation results are used for cross-validation between the switch state subgraph and the bus connection subgraph. The physical meaning is to convert the static connection in the primary wiring and the dynamic conduction in the remote signaling state into a passable edge in the current semantic topology graph.

[0031] Preferably, when the data processing program performs reachability traversal of nodes within the same bus segment in the bus connectivity subgraph, it uses the bus object node as the reference node and searches for adjacent nodes along the bus connection edge and the open / closed connection edge with a connectivity value of (1). If there is no reachable path between two bus object nodes under the same bus segment identifier, the corresponding bus connection edge or open / closed connection edge is written into the disconnected record in the topology check table. If there is a reachable path between two bus object nodes under different bus segment identifiers that does not pass through the connecting switch or sectionalizing switch, the corresponding connection edge is written into the cross-segment connectivity record. In the switch state subgraph, the data processing program performs the connection / disconnection status of the remote signaling measurement node and the edge status of the open / closed connection edge. If the remote signaling status is non-conductive but the corresponding edge in the semantic topology graph is still marked as connected, the edge is written into the state conflict record. In the feeder affiliation subgraph, the data processing program forms a directed path according to the power supply direction edge from the feeder object node to the interval object node, and checks whether the direction between adjacent intervals in the path is continuous. In the ring network breakpoint subgraph, the data processing program checks whether the breakpoint node is located at the boundary of the affiliation range of two feeders, and checks whether the paths on both sides of the breakpoint belong to different feeder objects respectively. In this embodiment, the subgraph verification results are uniformly saved through the topology verification table, so that the subsequent conflict weight calculation can directly read the conflict node, conflict edge, conflict source and conflict generation time.

[0032] refer to Figure 4In one embodiment, when a conflict arises during consistency verification, the data processing program configures a source credibility factor, an update time factor, a topology reachability factor, and a historical version continuity factor for the conflict relationship. It then generates a conflict weight based on these factors. The conflict relationship is marked as a retained relationship, a pending confirmation relationship, or an abnormal relationship according to the conflict weight. Retained relationships participate in the generation of the ring main unit's digital twin instance. Pending confirmation relationships and abnormal relationships are written into the verification record. The source credibility factor is determined by the data source category, source record integrity, and cross-references between sources. The verification results confirm that the update time factor is determined by the difference between the time identifier of the data item corresponding to the conflict relationship and the time of the current construction version. The topology reachability factor is determined by whether the edge where the conflict relationship is located satisfies the reachability path of the corresponding subgraph. The historical version continuity factor is determined by the node existence status, edge existence status, data binding status and feeder affiliation status in the previous topology version. The data processing program does not cover all conflict relationships with a single source, but incorporates multiple factors into the same conflict weight calculation model to obtain the processing category of the conflict relationship in the current twin construction. In this embodiment, the conflict weight transforms multi-source relationship conflicts into a computable relationship filtering process.

[0033] Conflict weights can be calculated using the following formula, which is written as:

[0034] in, Indicates conflict relationship Conflict weights This represents the source credibility factor. Indicates the update time factor. Represents the topological reachability factor. Indicates the continuity factor of historical versions. , , and Represents the corresponding weight and satisfies In the example, take , , , The source credibility factor of a certain conflict relationship is The update time factor is Topological reachability factor is Historical version continuity factor is ,but The physical meaning is to convert source quality, temporal proximity, topological legality, and version continuity into the calculation basis for relation preservation.

[0035] Table 2 shows a conflict relationship processing table, which is used to illustrate the data writing method between conflict weights and relationship tags.

[0036] Table 2 Conflict Resolution Table

[0037] In the processing method shown in Table 2, the data processing program calculates the conflict weights for multiple candidate relationships within the same conflict group and writes the connection relationships that meet the retention conditions into the current version of the semantic topology graph. If there are more than two candidate relationships in the same conflict group that both meet the retention conditions, the topology reachability factor and historical version continuity factor are read for sorting, and the relationship consistent with the current feeder's belonging path is retained. Relationships to be confirmed do not participate in the generation of node edges of the current twin instance, but are saved in the verification record and version record so that they can participate in the calculation again after the data source is updated. Abnormal relationships are written into the conflict chain record and enter the local topology isolation segment. Measurement relationships with data binding offsets do not cover the original measurement binding edges before the verification is completed. Retained relationships, pending confirmation relationships, and abnormal relationships are saved using different tags in the version chain. This embodiment enables the structured storage and diversion of conflict relationships through the processing table shown in Table 2, avoiding the need to manually fix the priority of all conflicts.

[0038] In a preferred embodiment, the unified entity index further forms multi-source matching features based on name segmentation results, number fragments, point prefixes, cabinet location descriptions, and feeder names. The data processing program constructs feature signatures for the multi-source matching features and writes the feature signatures, along with the equipment hierarchy and wiring adjacency relationships, into the entity merging matrix. In the entity merging matrix, data items with the same main entity identifier and whose wiring adjacency relationships satisfy the single reachable path condition from the cabinet node through the interval node and switch node to the bus node are merged. Data items that do not satisfy the single reachable path condition are assigned to the ambiguous entity set. The name segmentation results are obtained by segmenting the cabinet name, interval name, and feeder name into words. The number fragments are obtained by segmenting the asset code, equipment number, and scheduling number. The hierarchical meaning of numbers or character fragments is extracted. The point number prefix is ​​obtained by extracting the prefix of the corresponding station, cabinet or interval from the remote signaling point number and the telemetry point number. The cabinet location description is obtained by parsing the location field in the ledger. The feeder name is obtained by the feeder field in the scheduling topology. The feature signature is generated by concatenating the above multi-source matching features in a fixed field order and then performing digest encoding. The rows and columns in the entity merging matrix correspond to candidate data items respectively. The matrix elements record whether two candidate data items satisfy the same main entity identifier, whether they have the same or adjacent subordinate level, whether there is a single reachable path, and whether there is a measurement type conflict. In this embodiment, the field similarity relationship and the topological adjacency relationship are used for entity merging simultaneously through feature signature and entity merging matrix.

[0039] The elements in the entity merging matrix can be expressed using the following formula, written as:

[0040] in, Represents candidate data items With candidate data items The merge marker in the entity merge matrix takes the value of The time indicates that merging is allowed, and the value is [value]. This indicates that merging is not allowed. and Representing candidate data items With candidate data items The main entity identifier, This indicates whether the signatures of the two entities satisfy the consistency condition; if they do, then the signature is taken. If not satisfied, take , This indicates whether there exists a single reachable path from the cabinet node through the bay node and switch node to the bus node; if so, it is taken. If it does not exist, take In the example, the main entity identifiers of the two candidate data items are both the same interval object, the feature signatures satisfy the consistency condition and there exists a single reachable path, then... If two names are similar but their wiring adjacency cannot form a single reachable path, then And obtained The physical meaning is that entity merging only occurs when the entity identifier, feature signature, and topological path are all valid.

[0041] In this embodiment, the ambiguous entity set is used to store data items with the same main entity identifier but whose single reachability path condition is not met. Each data item in the ambiguous entity set retains the candidate main entity identifier, original fields, standard fields, feature signature, adjacent node identifier, and the reason for not meeting the single reachability path condition. The reasons include the existence of more than two candidate paths from the cabinet node to the bus node, the lack of interval nodes, the lack of switch nodes, bus node affiliation conflict, and the measurement type being inconsistent with the equipment object type. When the data processing program generates the reachability path set later, it uses the candidate data items in the ambiguous entity set as input for the candidate paths and does not directly merge the ambiguous entities into the twin instance. If the subsequent topology verification result makes a certain candidate path satisfy the bus connectivity constraint and feeder affiliation constraint, the corresponding data item can enter the candidate merging state from the ambiguous entity set. If the subsequent conflict weight still does not meet the retention condition, the data item continues to be retained in the ambiguous entity set and written into the verification record. This embodiment isolates high-risk entity merging through the ambiguous entity set to avoid data items with similar names or numbers but inconsistent electrical topologies from entering the same digital twin object.

[0042] Furthermore, boundary nodes are set in the bus connectivity subgraph, switch status subgraph, feeder affiliation subgraph, and ring network breakpoint subgraph. These boundary nodes consist of feeder entry nodes, tie switch nodes, segment switch nodes, and end-interval nodes. The data processing program generates a set of reachable paths using these boundary nodes as start and end constraints. It then cross-compares the reachable path set with the edge types in the semantic topology graph. When the same node belongs to more than two reachable path sets, the path affiliation is determined based on the connectivity attribute in the switch status subgraph. The feeder entry node is determined by the feeder access relationship in the scheduling topology. The tie switch node is determined by the node with tie attributes or cross-feeder connection relationships among the switch object nodes. The segment switch node is determined by the node of the switch object node. The nodes with feeder segmentation attributes or busbar segmentation connection relationships are determined. The terminal interval nodes are determined by the interval objects without downstream interval nodes in the feeder belonging subgraph. The data processing program starts from each type of boundary node and expands the path along the edge that meets the edge type constraint in the semantic topology graph. During the path expansion process, only the edge type that matches the semantics of the current subgraph is allowed to pass through. For example, in the bus connection subgraph, the bus connection edge and the open connection edge with the connection state being conductive are allowed to pass through. In the feeder belonging subgraph, the feeder power supply direction edge and the upstream and downstream interval edge are allowed to pass through. In the ring network breakpoint subgraph, the tie edge and the segment edge are allowed to pass through. In this embodiment, the path search range is limited by the boundary nodes so that the subgraph verification and the specific electrical semantics are maintained.

[0043] The set of reachable paths can be generated by calculating the path validity using the following formula, which is written as:

[0044] in, Indicates the first The validity value of a reachable path. Indicates the first The set of edges contained in a reachable path. Representing an edge The connectivity value is taken when connectivity is achieved. When not connected, take , Representing an edge The direction is a valid value, and it is taken when the edge direction is consistent with the current subgraph requirement. If there is a discrepancy, take the... In the example, a path contains 3 edges, and the connectivity value of all 3 edges is 0. The legal values ​​for the directions are respectively , , ,but If the direction of the second side is inconsistent with the power supply direction of the feeder, then And obtained In physical terms, a path is only written into the reachable path set if all edges in the path are connected and their directions conform to the electrical semantics of the current subgraph.

[0045] Preferably, when the data processing program cross-compares the reachable path set with the edge types in the semantic topology graph, it records the starting boundary node, ending boundary node, intermediate node sequence, edge type sequence, switch state sequence, and feeder assignment sequence for each path. If there is an edge type in the path that does not match the current subgraph, the path is marked as an edge type conflict. If there is a switch-connected edge with inconsistent conduction states in the path, the path is marked as a switch state conflict. If there is a feeder power supply direction edge with reversed upstream and downstream directions in the path, the path is marked as a direction conflict. If the same node belongs to two or more reachable path sets and the feeder objects of the two path sets are different, the data processing program reads the connectivity attribute in the switch state subgraph, assigns the node on the conduction path to the corresponding path set, and marks the node on the non-conducting path as a candidate assignment. If both path sets meet the conduction conditions, the node is written into the conflict chain record. In this embodiment, by comparing the edge type sequence and state sequence at the path level, the node assignment is not determined by the static graphic position, but by the edge type, edge direction, and switch connectivity attribute in the current semantic topology graph.

[0046] In one embodiment, the historical version continuity factor is generated from the node existence status, edge existence status, data binding status, and feeder affiliation status of the previous topology version. The data processing program couples the historical version continuity factor with the topology reachability factor for calculation, sets a version inheritance identifier for retained relationships, sets a manual review identifier for relationships to be confirmed, and sets a conflict chain identifier for abnormal relationships. The conflict chain identifier includes a conflict node identifier, a conflict edge identifier, a conflict source identifier, and a conflict occurrence time. The node existence status in the previous topology version indicates whether the nodes involved in the current candidate relationship existed in the previous version, the edge existence status indicates whether the edges involved in the current candidate relationship existed in the previous version, and the data binding status... The measurement bindings involved in the current candidate relationship exist in the previous version. The feeder attribution status is used to indicate whether the nodes or edges involved in the current candidate relationship belong to the same feeder range in the previous version. The data processing program reads the difference records between the previous version and the current version and converts the difference records into historical version continuity factors. If the candidate relationship has the same nodes, the same edges, the same data bindings, and the same feeder attribution in the previous version, the historical version continuity factor takes a higher value. If the candidate relationship only has the same nodes but the edges and feeder attributions have changed, the historical version continuity factor takes a lower value. This embodiment avoids unnecessary jumps in topology relationships due to short-term data source conflicts by using historical version continuity factors.

[0047] The historical version continuity factor can be calculated using the following formula, which is written as:

[0048] in, Indicates conflict relationship The historical version continuity factor, This indicates that the node has a consistent state value; the value is taken when the node exists in both the previous topology version and the current candidate relationship. Otherwise take , This indicates that the edge has a consistent state value; the edge is considered to exist in both the previous topological version and the current candidate relation. Otherwise take , This indicates a consistent data binding state value; the value is taken when the measurement binding exists in both the previous topology version and the current candidate relationship. Otherwise take , This represents the consistency value of the feeder's attribution status. The value is taken when the feeder's attribution is consistent in the previous topology version and the current candidate relationship. Otherwise take , , , and Represents the corresponding weight and satisfies In the example, take , , , If a conflicting relationship node has a consistent state, an edge has a consistent state, a data binding state is inconsistent, and a feeder affiliation state is consistent, then The physical meaning is to convert the existence of objects, connections, measurement bindings, and feeder affiliations in the previous version into the calculation basis for whether the current relationship continues.

[0049] Furthermore, the version inheritance identifier is used to indicate that the node identifier, edge identifier, or data binding identifier in the previous topology version should be retained. The manual review identifier is used to indicate that the original fields, standard fields, candidate entity indexes, and conflict reasons should be retained for the relationship to be confirmed. The conflict chain identifier is used to organize the conflict nodes, conflict edges, conflict sources, and conflict occurrence times involved in the abnormal relationship into a directed chain structure. When the data processing program generates a new ring network cabinet digital twin instance, it reads the version inheritance identifier and directly puts the nodes, edges, and measurement bindings corresponding to the inheritance identifier into the instance data layer. The data items corresponding to the manual review identifier enter the verification record layer but do not cover the instance data layer. The data items corresponding to the conflict chain identifier enter the local topology isolation segment. The conflict chain is sorted by the conflict occurrence time, associated by the conflict nodes and conflict edges, and grouped by the conflict source. If the same conflict node generates different connecting edges in multiple sources, the conflict chain records the edge start point, edge end point, edge type, and time identifier given by each source. In this embodiment, the version inheritance identifier, manual review identifier, and conflict chain identifier are used to store the three types of relationships in a layered manner, so that the new twin instance has a traceable version structure.

[0050] In a preferred embodiment, the reachable path set further includes candidate paths introduced by the ambiguous entity set. The data processing program establishes a path relationship table by associating candidate paths with reserved relationships, pending confirmation relationships, and abnormal relationships. In the path relationship table, when an edge corresponding to the same candidate path simultaneously has both reserved and pending confirmation relationships, the edge corresponding to the reserved relationship is retained and the data binding corresponding to the pending confirmation relationship is frozen. When an edge corresponding to the same candidate path simultaneously has abnormal relationships, a local topological isolation fragment associated with a conflict chain identifier is generated. The candidate path is generated by the temporary connection relationship between candidate data items in the ambiguous entity set and boundary nodes. The temporary connection relationship is not directly written into the current version of the semantic topology graph, but is instead written into... The path relationship table records candidate path identifiers, candidate entity identifiers, path start points, path end points, edges in the path, relationship tags for each edge, candidate measurement bindings, and associated conflict chain identifiers. When processing multiple types of relationships in the same candidate path, the data processing program uses retained relationships as the basis for generating the current instance, unconfirmed relationships as frozen objects, and abnormal relationships as isolated objects. Frozen data bindings maintain the original measurement binding version and are not overwritten. Local topology isolation segments store nodes and edges related to abnormal relationships. In this embodiment, the path relationship table associates ambiguous entities, candidate paths, and conflict relationships within the same data structure, avoiding ambiguous data directly affecting the current twin instance.

[0051] The relationship selection of candidate paths can be expressed by the following formula, which is written as:

[0052] in, Indicates the first The relationship selection value of the candidate paths, Indicates the first The set of relations that are compared on each candidate path. Representing relations Conflict weights Indicates the first The validity value of each reachable path, and the validity value of a candidate path in the example. The conflict weights of the three candidate relationships on the path are respectively , and ,but If the validity value of the same path is Then the product of all candidate relations is In physical terms, only candidate paths that satisfy the path validity requirement are allowed to participate in the generation of the current instance according to the conflict weight.

[0053] Table 3 shows a configuration method for the data fields of a path relationship table, which is used to illustrate the correspondence between candidate paths, relationship tags, and isolated fragments.

[0054] Table 3: Data Field Configuration Method for the Path Relationship Table

[0055] In the path relationship table shown in Table 3, the data processing program generates a candidate path identifier for each candidate path and writes the candidate entity identifier, the feature signature corresponding to the candidate entity, the path start point, and the path end point in the same row or the same record group. The relationship marker is determined by the conflict weight calculation result. The frozen binding identifier is generated only when there is a pending relationship in the path and the pending relationship involves the running measurement object. The isolated segment identifier is generated only when there is an abnormal relationship in the path and the abnormal relationship destroys the bus connection, switch conduction, feeder affiliation, or ring network break point constraint. The version association identifier is used to associate the candidate path, frozen binding, and isolated segment with the entity index version, semantic topology graph version, and data binding version. The path relationship table is stored in the verification record layer and read in the next construction. If the new data source causes the pending relationship to be converted into a retained relationship, the frozen binding identifier is released and enters the data binding update scope. If the abnormal relationship still does not meet the topology constraint in the next construction, the isolated segment identifier continues to be retained. In this embodiment, the field organization method of the path relationship table enables candidate paths to continuously participate in verification across versions.

[0056] In one embodiment, the topology version record includes an entity index version, a semantic topology graph version, a consistency check version, and a data binding version. When generating a new ring network cabinet digital twin instance, the data processing program writes local topology isolation fragments, path relationship tables, and conflict chain identifiers into the incremental update queue. It also limits the nodes to be updated according to the change range of the entity index version, the edges to be updated according to the change range of the semantic topology graph version, and the measurement relationships to be updated according to the change range of the data binding version. The entity index version records newly created, merged, split, and frozen entities in the unified entity index. The semantic topology graph version records changes in the node set, edge set, and edge attributes. The consistency verification version records the subgraphs, constraint rules, conflict weights, and relationship tags used in each verification. The data binding version records the binding changes between remote signaling data, telemetry data, and operation records and the twin object. The incremental update queue organizes data according to node update items, edge update items, measurement relationship update items, isolated fragment update items, and conflict chain update items. After comparing the current version with the previous version, the data processing program only puts the entities, edges, and binding relationships that have changed into the incremental update queue. The unchanged nodes and edges enter the new twin instance through the version inheritance identifier. This embodiment limits the calculation scope of each model update through version records and incremental update queues.

[0057] The incremental update range can be determined using the following formula, which is written as:

[0058] in, This represents the set of nodes to be updated. This represents the set of nodes in the current entity index version. This represents the set of nodes in the previous entity index version. This represents the set of nodes involved, identified by the conflict chain. This represents the set of edges to be updated. This represents the set of edges in the current version of the semantic topology graph. This represents the set of edges in the previous semantic topology graph version. This represents the set of edges involved by a locally topologically isolated segment. This represents the set of measurement relationships to be updated. This represents the measurement binding set in the current data binding version. This represents the measurement binding set in the previous data binding version. This represents the set of measurement bindings involved by the frozen binding identifier. In the example, the current version adds 2 nodes and the conflict chain involves 1 existing node. It contains 3 nodes. The current version adds 1 edge and the isolation segment involves 2 edges. It contains 3 edges. The current version adds 1 measurement binding and the frozen binding flag involves 1 existing binding. It includes two measurement relationships, and its physical meaning is to use version differences and conflict isolation ranges together as the boundary of twin incremental updates.

[0059] Specifically, when executing the incremental update queue, the data processing program reads the entity index version differences. If a node is generated by a newly added entity index, a new twin object node is generated and an edge containing the parent object is established. If a node is generated by entity merging, the main entity identifier is retained and the source identifier, time identifier, and original fields of the merged entity are written into the historical mapping record. If a node is generated by entity splitting, the node affiliation is redistributed according to the single reachable path condition in the entity merging matrix. If a node is in a frozen state, it does not enter the instance data layer to overwrite the original node. The data processing program reads the semantic topology graph version differences. If an edge is generated by a retained relationship, it is written into the current semantic topology graph version. If an edge is generated by a pending confirmation relationship, it is written into the path relationship table and kept frozen. If an edge is generated by an abnormal relationship, it is written into the local topology isolation fragment. The data processing program reads the data binding version differences. If a measurement binding corresponds to a retained relationship, the binding relationship is updated. If a measurement binding corresponds to a pending confirmation relationship, a frozen binding identifier is generated. This embodiment realizes the hierarchical update of nodes, edges, and measurement relationships through the incremental update queue, and keeps conflicting data from directly overwriting the current twin instance.

[0060] In a preferred embodiment, the ring main unit digital twin instance consists of an object layer, a topology layer, a state layer, a verification layer, and a version layer. The object layer stores node information of ring main unit objects, bay objects, busbar objects, switch objects, feeder objects, and operational measurement objects. The topology layer stores the cabinet containment edges, bay ownership edges, busbar connection edges, switch connection edges, feeder power supply direction edges, and measurement binding edges. The state layer stores the current state values ​​and historical records corresponding to telemetry data, teleindication data, and operational records. The verification layer stores the verification results of the busbar connectivity sub-graph, switch state sub-graph, feeder ownership sub-graph, and ring main unit breakpoint sub-graph. The version layer stores the entity index version. The data processing program, when generating instances, writes edges marked as reserved relationships in the semantic topology graph to the topology layer, writes unconfirmed relationships and abnormal relationships to the verification layer, writes version inheritance identifiers to the version layer, and writes measurement bindings between running measurement object nodes and device object nodes to the state layer. The object layer, topology layer, state layer, verification layer, and version layer are associated using the same main entity identifier. This embodiment uses a hierarchical instance structure to enable the object expression, topology expression, running status, verification record, and version record of the ring network cabinet digital twin to be read under the same identifier system.

[0061] In this embodiment, when the data processing program reads the digital twin instance of the ring main unit, it does not directly use the visualized primitives as the calculation object. Instead, it reads the main entity identifier from the object layer, the edges connected to the main entity identifier from the topology layer, the corresponding measurement bindings and state values ​​from the state layer, the conflict records related to the main entity identifier from the verification layer, and the change records of the object in each version from the version layer. If the current topology relationship needs to be output, only the edges that are in the reserved relationship and not isolated in the current semantic topology graph version are read. If the topology fragment with conflict needs to be output, the local topology isolation fragment corresponding to the conflict chain identifier is read. If model rollback is required, the nodes, edges and data bindings of the previous version are restored according to the version inheritance identifier and historical mapping records. If the frozen binding needs to be reviewed, the original point number, standard field, candidate entity index and candidate path corresponding to the frozen binding identifier are read. In this embodiment, the same main entity identifier runs through the object layer, topology layer, state layer, verification layer and version layer, so that subsequent reading and maintenance do not need to be repeatedly matched between multiple data sources.

[0062] In a preferred embodiment, the data processing program can also participate in the generation of topology version records for change events in the operation record. The change events include switch status changes, feeder affiliation changes, interval association changes, data binding changes, and manual review result changes. After the change event enters the system, it is parsed into an event field with event type, associated main entity identifier, occurrence time, and source identifier. The data processing program compares the event field with the current semantic topology map version. If the main entity identifier involved in the event field exists in the current instance, the event field is written into the corresponding version record. If the main entity identifier involved in the event field is in the ambiguous entity set, the event field is written into the path relationship table and participates in candidate path calculation. If the connection edge involved in the event field is in a local topology isolation segment, the event field is written into the time series corresponding to the conflict chain identifier. The processing does not change the hardware structure of the ring network cabinet, nor does it introduce additional acquisition equipment. Instead, it uses the existing operation record as the version update input. This embodiment enables the topology version change to maintain data correspondence with the state change during operation through the association between the operation record and the version record.

[0063] In one embodiment, when the data processing program parses ledger data, primary wiring relationships, scheduling topology, remote signaling data, telemetry data, and operation records, it uses a unified field dictionary. The unified field dictionary stores field names, field sources, field types, field levels, field belonging objects, and applicable verification constraints. For example, the cabinet name field belongs to the ring network cabinet object and participates in object hierarchy constraints; the interval number field belongs to the interval object and participates in upstream and downstream interval constraints; the terminal connection identifier belongs to the bus or switch object and participates in bus connectivity constraints and switch conduction constraints; the feeder name belongs to the feeder object and participates in feeder belonging constraints; the remote signaling point number belongs to the operation measurement object and participates in switch conduction constraints; the telemetry point number belongs to the operation measurement object and participates in data binding consistency constraints; and the change time in the operation record belongs to the version layer and participates in the calculation of historical version continuity factors. During the parsing phase, the data processing program writes the field belonging object and applicable constraints for each standard field. Subsequently, when generating the unified entity index and semantic topology diagram, the field is filtered based on the field dictionary. This embodiment limits the scope of use of various data fields through the unified field dictionary, preventing irrelevant fields from entering the topology verification process.

[0064] Specifically, the unified field dictionary can be implemented using a key-value structure, a relational table structure, or a graph attribute structure. In a key-value structure, the field name serves as the key, and the field source, field type, field level, and belonging object serve as the value. In a relational table structure, each row corresponds to a standard field, and each column corresponds to a field attribute. In a graph attribute structure, fields serve as attribute nodes and are connected to object nodes via belonging edges. When the data processing program parses the raw data, it first reads the field dictionary. If the raw field can find a corresponding standard field in the field dictionary, the raw field value is written to the standard field. If the raw field cannot find a corresponding standard field, it is written to the unidentified field cache. Fields in the unidentified field cache do not participate in the current twin instance generation, but retain the source identifier and time identifier. If the field can be identified after the field dictionary is expanded, the field re-enters the standard field cache. The participation constraints in the field dictionary are used to limit the field's entry into the corresponding subgraph verification. For example, the telemetry current value does not directly participate in the single reachability path judgment of the entity merging matrix, but it can participate in the measurement binding consistency check. In this embodiment, the constraints of the field dictionary enable data parsing to have stable field boundaries.

[0065] In a preferred embodiment, after the data processing program generates a unified entity index, it also performs index integrity checks on objects at different levels. The index integrity checks include whether a ring main unit object is associated with at least one bay object, whether a bay object has a cabinet affiliation, whether a switch object has a bay affiliation and connection edge, whether a busbar object has a busbar connection edge or a connection edge adjacent to a switch object, whether a feeder object has a power supply direction edge, and whether a running measurement object has a measurement binding candidate. If an object lacks the necessary association relationship for its level, it is not directly deleted, but written into the ambiguous entity set or verification record. The data processing program distinguishes between missing entities and missing relationships through index integrity checks. If an object itself has a stable main entity identifier but missing connection edges, the node is retained and the missing edge is marked. If the main entity identifier of the object itself is unstable and the connection edge is missing, the object is written into the ambiguous entity set. If the main entity identifier of the object is stable but the data binding is missing, the node is retained and the relevant measurement binding update is frozen. This embodiment supplements the object quality control after the unified entity index is generated through index integrity checks.

[0066] In a preferred embodiment, the measurement binding relationship is represented by the measurement binding edges between the operating measurement object node and the ring main unit object node, bay object node, switch object node, bus object node, or feeder object node. When the data processing program generates the measurement binding edges, it reads the point number prefix, measurement type label, candidate main entity identifier, and topological adjacency relationship. If the measurement type label of the telemetry point number is switch on / off state, then it prioritizes generating candidate measurement binding edges with the switch object node and checks whether the switch object node belongs to the corresponding bay object. If the measurement type label of the telemetry point number is current or power, then it prioritizes... Candidate measurement binding edges are generated for feeder object nodes or interval object nodes, and it is checked whether the node is located on the corresponding feeder belonging path. If the measurement type label of the telemetry point number is voltage, then candidate measurement binding edges are generated with the bus object node or interval object node first, and it is checked whether the node is located in the bus connected subgraph. After the candidate measurement binding edges are written into the semantic topology graph, they participate in the data binding consistency constraint. If the point number prefix is ​​consistent with the main entity identifier but the topological adjacency relationship is inconsistent, the measurement binding edge is marked as a relationship to be confirmed. In this embodiment, the measurement binding object is defined by the measurement type label and the topological adjacency relationship.

[0067] In one embodiment, when the data processing program verifies the consistency of measurement binding, it aligns the acquisition time of the running measurement object node with the current version time in the switch status subgraph or feeder attribution subgraph. If the time identifier of the measurement data is earlier than the current topology version time, the measurement data will not participate in the current measurement binding coverage and will only be saved as a historical record. If the time identifier of the measurement data is within the current topology version time range, the data processing program checks whether the measurement type label matches the type of the bound object, checks whether the point number prefix is ​​associated with the main entity identifier, and checks whether the path where the bound object is located meets the legality value in the reachable path set. If all the above conditions are met, the measurement binding edge is marked as a reserved relationship. If the measurement type matches but the path legality value is not met, the measurement binding edge is marked as a pending confirmation relationship and a frozen binding identifier is generated. If the measurement type does not match the object type, the measurement binding edge is marked as an abnormal relationship. This embodiment restricts the measurement binding relationship through time identifier, measurement type, and path legality, reducing the situation where the running measurement object is bound to the wrong device entity.

[0068] In a preferred embodiment, the verification of the ring network breakpoint subgraph uses tie switch nodes and segment switch nodes as the main boundary nodes. The data processing program reads the breakpoint description, feeder power supply direction edge, and upstream and downstream interval edge in the scheduling topology to generate a candidate breakpoint set. Each candidate breakpoint in the candidate breakpoint set includes a candidate node identifier, adjacent feeder identifier, adjacent interval identifier, corresponding switch status, and source identifier. The data processing program extends the path from the candidate breakpoint to both sides and checks whether the paths on both sides belong to different feeder objects or the same feeder. For different segment ranges, if the path attribution relationship on both sides is consistent with the feeder boundary in the scheduling topology, the breakpoint relationship is written into the reserved relationship. If the path attribution relationship on both sides is inconsistent but there is an entity to be confirmed or frozen binding, the breakpoint relationship is written into the unconfirmed relationship. If the paths on both sides cannot form a boundary segmentation or the switch status conflicts with the breakpoint description, the breakpoint relationship is written into the abnormal relationship. In this embodiment, the topology boundary related to the connection transfer is calculated through the ring network breakpoint subgraph, so that the ring network cabinet digital twin has a verifiable data structure in the feeder boundary expression.

[0069] In a preferred embodiment, when the data processing program verifies the feeder attribution subgraph, it takes the feeder inlet node as the starting point and the end interval node, tie switch node, or break point node as the ending point, and generates a directed path along the feeder power supply direction edge and the upstream and downstream interval edges. If there is an edge with reversed direction in the path, the edge is written into the direction conflict record. If two different feeder objects in the path give an attribution relationship to the same interval object, the data processing program reads the connectivity attribute in the switch status subgraph and determines the current conducting path. If two feeder objects are separated by a non-conducting tie switch, the interval object belongs to the conducting side feeder object. If two feeder objects are connected by a conducting tie switch and no break point boundary is formed, the interval object enters the conflict chain record. If there is a missing interval object in the same feeder path but the paths of adjacent switch objects and bus objects are continuous, the missing interval is written into the pending confirmation relationship. In this embodiment, the feeder attribution is determined by the feeder inlet, power supply direction, break point, and switch connectivity attribute.

[0070] In one embodiment, when the data processing program verifies the bus connectivity subgraph, it reads the bus object nodes, the switch object nodes connected to the bus, the interval object nodes connected to the bus, and the bus connection edges. The basic connection edges in the bus connectivity subgraph are generated by the primary wiring relationship, and the dynamic connection edges are generated by the connectivity values ​​in the switch state subgraph. If all bus object nodes in the same bus segment can form a connected component through the basic connection edges and the dynamic connection edges, the bus segment connectivity relationship is written into the reserved relationship. If more than two connected components appear in the same bus segment, the data processing program checks whether the missing edge exists in the pending confirmation relationship or the ambiguous entity set. If it exists, the corresponding bus connection relationship is written into the pending confirmation relationship. If it does not exist, the corresponding bus segment is written into the abnormal relationship. If different bus segments form a connected component without passing through the connecting switch node or the segment switch node, the data processing program will write the cross-segment connected edge into the conflict chain identifier. In this embodiment, the bus connection edges are verified through the bus connectivity components, so that the static primary wiring relationship and the switch state jointly participate in the bus topology expression.

[0071] In a preferred embodiment, when the data processing program sets a manual review identifier for the relationship to be confirmed, the manual review identifier is only stored as a data status marker in the verification record layer. The manual review identifier includes candidate entity identifier, candidate path identifier, frozen binding identifier, conflict relationship type, generation source, and generation time. The manual review identifier does not change the nodes, edges, and measurement bindings in the current twin instance. If the review result is written back to the system in the form of structured data, the review result enters the runtime record parsing process and generates new standard fields. The data processing program recalculates the unified entity index, semantic topology graph, and consistency verification result. If the review result is consistent with the retained relationship, the corresponding relationship to be confirmed is converted into a version inheritance relationship. If the review result is consistent with the abnormal relationship, the corresponding relationship to be confirmed enters the local topology isolation segment. If the review result introduces a new connection edge, the new connection edge enters the conflict weight calculation according to the source identifier and time identifier. In this embodiment, the manual review identifier is limited to traceable data status, and manual review is not used as a direct modeling means to replace topology verification.

[0072] In one embodiment, when the data processing program generates the topology verification table, it sets a verification object, verification subgraph, verification constraint, participating node, participating edge, source identifier, time identifier, relationship identifier, and version identifier for each verification record. The verification object corresponds to one of the following: ring network cabinet object, bay object, bus object, switch object, feeder object, or operating measurement object. The verification subgraph corresponds to the bus connectivity subgraph, switch status subgraph, feeder affiliation subgraph, or ring network breakpoint subgraph. The verification constraint corresponds to bus connectivity, switch conduction, upstream and downstream bays, ring network breakpoint, feeder affiliation, or data binding consistency. Participating nodes and participating edges are used to locate the location of the conflict. The source identifier and time identifier are used to trace the source of the conflicting data. The relationship identifier is used to record retained relationships, pending confirmation relationships, or abnormal relationships. The version identifier is used to associate the entity index version, semantic topology graph version, consistency verification version, and data binding version. In this embodiment, the intermediate data in the verification process is saved as a readable structure through the topology verification table, avoiding the verification results from existing only in log text form.

[0073] In a preferred embodiment, after the ring main unit digital twin instance is generated, the data processing program performs consistency encapsulation on the instance. The consistency encapsulation includes writing the reserved relationships in the current version into the instance topology layer, writing the version inheritance identifier into the instance version layer, writing the reserved bindings between the running measurement object and the device object into the instance state layer, and writing the pending confirmation relationships, abnormal relationships, conflict chain identifiers, frozen binding identifiers, and local topology isolation fragments into the instance verification layer. During the instance encapsulation process, pending confirmation relationships or abnormal relationships are not written into the current runnable topology. If a node only exists in a local topology isolation fragment and does not belong to a reserved relationship, then the node does not participate in the current topology path calculation. If a measurement binding is in a frozen state, then the current instance continues to reference the binding relationship corresponding to the measurement object in the previous data binding version. If there is no referenceable binding in the previous data binding version, then the measurement object is stored in the unbound set of the state layer. This embodiment controls the writing position of different data states in the twin instance through consistency encapsulation, so that the current instance can distinguish between available topology, pending verification topology, and isolated topology.

[0074] In a preferred embodiment, the data processing program stores instances using a graph database, a relational database, or a combination of both. When using a graph database, the node set, edge set, and edge attributes are directly saved as graph data, while version identifiers and relationship tags are saved as node or edge attributes. When using a relational database, the node table, edge table, measurement binding table, verification table, and version table each store corresponding data, with the main entity identifier serving as the association key between tables. When using a combination of both, the current semantic topology graph is stored in the graph database, while version records, verification records, and running records are stored in the relational database. The choice of storage method does not change the system's data processing logic. When the data processing program reads data, it associates the data using the main entity identifier, edge identifier, version identifier, and relationship tag. If the current instance needs to output a local topology isolation fragment, it reads the associated nodes and edges according to the conflict chain identifier. If the current instance needs to output the measurement binding status, it reads the binding relationship according to the data binding version. This embodiment unifies the data reading path under different storage methods through the main entity identifier and version identifier.

[0075] In one embodiment, the overall execution flow of the data processing program is carried out in units of build versions. A build version is determined by the time range of the current input data, the data source set, and the semantic template library version. Within the same build version, the data processing program completes standard field parsing, unified entity index generation, semantic topology graph generation, and sub-builds. Figure 1The process includes consistency verification, conflict weight calculation, relationship marking, path relationship table generation, version record update, and twin instance encapsulation. If a semantic template library is found to be missing a certain object type or edge type during the construction process, the corresponding data item will be entered into the unidentified field cache or ambiguous entity set and will not participate in the current instance generation. If a data source is found to be missing a certain running measurement object during the construction process, the corresponding device object can still generate nodes and edges through ledger data, primary wiring relationships, and scheduling topology. Measurement binding relationships are written into the unbound set or pending confirmation relationships. If a conflict is found between the scheduling topology and primary wiring relationships during the construction process, the conflict weight determines the retained relationship, pending confirmation relationship, and abnormal relationship. This embodiment organizes the data processing flow by constructing versions, so that each twin construction has a defined data input range and version output results.

[0076] In this embodiment, the ring main unit digital twin construction system formed by each implementation method operates around the same core data chain. That is, multi-source heterogeneous data is parsed into standard fields, the standard fields are included in a unified entity index, the unified entity index is converted into a semantic topology graph, the semantic topology graph is split into subgraphs with electrical semantics for consistency verification, conflict relationships are diverted through conflict weights and path relationship tables, retained relationships are written into the digital twin instance, and unconfirmed relationships and abnormal relationships are written into verification records, conflict chain identifiers, and local topology isolation fragments. The entity index version, semantic topology graph version, consistency verification version, and data binding version are used as the basis for subsequent incremental updates. The data chain enables the differences between ledgers, primary wiring, scheduling topology, remote signaling, telemetry, and operation records to be uniformly processed during the twin construction stage. The advantage of this embodiment is that the entity merging, topology generation, measurement binding, conflict isolation, and version update of the ring main unit digital twin have a unified calculation basis, and it can reduce model construction errors caused by field inconsistencies, connection conflicts, and version breaks.

Claims

1. A digital twin construction system for ring main units in smart distribution networks, characterized in that, It includes a semantic template library that stores ring main unit objects, bay objects, busbar objects, switch objects, feeder objects and operation measurement objects, as well as a data processing program connected to the semantic template library; The data processing program is used to parse ledger data, primary wiring relationships, scheduling topology, remote signaling data, telemetry data, and operation records into standard fields with source identifiers and time identifiers, respectively. Based on the standard fields, a unified entity index and semantic topology graph are generated. The semantic topology graph is checked for consistency according to bus connectivity, switch conduction, upstream and downstream intervals, ring network breakpoints, and feeder affiliation constraints. Based on the checked nodes, edges, and data binding relationships, a digital twin instance of the ring network cabinet with a topology version record is generated.

2. The ring main unit digital twin construction system for smart distribution networks according to claim 1, characterized in that, The unified entity index is generated in the following manner: The cabinet name, interval name, equipment number and asset code in the ledger data are normalized. The terminal connection identifier in the primary wiring relationship and the scheduling number in the scheduling topology are associated with the same name but different codes. Measurement type labels are established for the point numbers in the remote signaling data and remote measurement data. The normalized fields, association results and measurement type labels are written into the same candidate entity set. In the candidate entity set, the main entity identifier is determined based on the device hierarchy and wiring adjacency. Data items with the same main entity identifier are merged into the same unified entity index.

3. The ring main unit digital twin construction system for smart distribution networks according to claim 1, characterized in that, The semantic topology graph includes nodes corresponding to the ring main unit object, the bay object, the busbar object, the switch object, the feeder object, and the operating measurement object, as well as edges representing cabinet inclusion, bay affiliation, busbar connection, switch connection, feeder power supply direction, and measurement binding; The data processing program generates node types according to the object hierarchy definition in the semantic template library, generates edge types according to the primary wiring relationship and the scheduling topology, and writes a source identifier, time identifier and relationship confidence value for each edge.

4. The ring main unit digital twin construction system for smart distribution networks according to claim 3, characterized in that, The consistency check includes forming a bus connectivity subgraph, a switch status subgraph, a feeder attribution subgraph, and a ring network breakpoint subgraph based on the semantic topology graph. The data processing program verifies the reachability of nodes in the same bus segment in the bus connectivity subgraph, verifies the consistency between the switch opening / closing state and the edge connectivity attribute in the switch state subgraph, verifies the directional continuity of upstream and downstream intervals in the feeder affiliation subgraph, verifies the correspondence between the break point and the feeder boundary in the ring network break point subgraph, and writes the verification results of each subgraph into a topology verification table with node and edge identifiers.

5. The ring main unit digital twin construction system for smart distribution networks according to claim 4, characterized in that, When a conflict arises during the consistency check, the data processing program configures a source credibility factor, an update time factor, a topology reachability factor, and a historical version continuity factor for the conflict relationship, and generates a conflict weight according to the source credibility factor, the update time factor, the topology reachability factor, and the historical version continuity factor. The conflicting relationships are marked as reserved relationships, pending confirmation relationships, or abnormal relationships according to the conflict weight. The reserved relationships participate in the generation of the digital twin instance of the ring main unit, and the pending confirmation relationships and the abnormal relationships are written into the verification record.

6. The ring main unit digital twin construction system for smart distribution networks according to claim 5, characterized in that, The unified entity index also forms multi-source matching features through name segmentation results, number fragments, point prefixes, cabinet location descriptions, and feeder names; The data processing program constructs a feature signature for the multi-source matching features and writes the feature signature, along with the device hierarchy and wiring adjacency relationship, into the entity merging matrix. In the entity merging matrix, data items with the same main entity identifier and whose wiring adjacency satisfies the single reachable path condition from the cabinet node through the interval node and the switch node to the bus node are merged, while data items that do not satisfy the single reachable path condition are assigned to the ambiguous entity set.

7. The ring main unit digital twin construction system for smart distribution networks according to claim 6, characterized in that, The bus connection sub-diagram, the switch status sub-diagram, the feeder ownership sub-diagram, and the ring network break point sub-diagram are each equipped with boundary nodes. The boundary nodes consist of feeder entry nodes, tie switch nodes, segment switch nodes, and end interval nodes. The data processing program generates a set of reachable paths using the boundary nodes as start and end constraints, and cross-compares the set of reachable paths with the edge types in the semantic topology graph. When the same node belongs to more than two reachable path sets, the path affiliation is determined based on the connectivity attribute in the switch state subgraph.

8. The ring main unit digital twin construction system for smart distribution networks according to claim 7, characterized in that, The historical version continuity factor is generated from the node existence status, edge existence status, data binding status, and feeder affiliation status in the previous topology version. The data processing program couples the historical version continuity factor with the topology reachability factor for calculation, sets a version inheritance identifier for the retained relationship, sets a manual review identifier for the relationship to be confirmed, and sets a conflict chain identifier for the abnormal relationship. The conflict chain identifier includes the conflict node identifier, the conflict edge identifier, the conflict source identifier, and the conflict occurrence time.

9. The ring main unit digital twin construction system for smart distribution networks according to claim 8, characterized in that, The reachable path set also includes candidate paths introduced by the ambiguous entity set, and the data processing program establishes a path relationship table with the candidate paths, the reserved relationships, the unconfirmed relationships, and the abnormal relationships; In the path relationship table, when the edge corresponding to the same candidate path has both the reserved relationship and the pending confirmation relationship, the edge corresponding to the reserved relationship is retained and the data binding corresponding to the pending confirmation relationship is frozen; When the edges corresponding to the same candidate path simultaneously exhibit the aforementioned abnormal relationships, a local topological isolation segment associated with the conflict chain identifier is generated.

10. The ring main unit digital twin construction system for smart distribution networks according to claim 9, characterized in that, The topology version record includes entity index version, semantic topology graph version, consistency check version, and data binding version; When generating a new digital twin instance of the ring network cabinet, the data processing program writes the local topology isolation segment, the path relationship table, and the conflict chain identifier into the incremental update queue, and limits the nodes to be updated according to the change range of the entity index version, limits the edges to be updated according to the change range of the semantic topology graph version, and limits the measurement relationships to be updated according to the change range of the data binding version.

Images