A quality management and control method, system, device and medium based on a multi-level partition power grid diagram

Abstract

The application discloses a kind of quality control methods, systems, equipment and medium based on multi-level partition power grid diagram, belong to intelligent power grid and digital twin field, including: obtaining multiple partition subgraphs according to power grid voltage level, the consistency of adjacent subgraph at boundary is verified to splice, form panoramic power grid diagram;Multi-dimensional quality assessment of fusion electrical, topology, carbon flow and management attribute is carried out to panoramic diagram, and integrated health status is obtained;If state is abnormal, then based on record model changes blood relationship's atlas traces root event, and triggers corresponding check and repair closed loop.Through the mechanism of checking first and splicing later, the topology breakpoint and attribute conflict after splicing of traditional method are actively avoided;Through multi-dimensional coupling evaluation, the health status of power grid data is quantified, and new business requirements such as electric-carbon collaboration are supported;Through atlas tracing and closed-loop management, automatic positioning and repair of abnormality are realized, and the paradigm shift from passive response to active prevention of data quality management is promoted.

Description

Technical Field

[0001] This invention relates to the field of smart grid and digital twin technology, specifically to a quality control method, system, device and medium based on a multi-level regional power grid map. Background Technology

[0002] In the process of power system digitization and modeling, the power grid model serves as the core foundation for dispatching, planning, simulation, and carbon flow analysis. Its accuracy and completeness directly determine the reliability of upper-level applications. Currently, mainstream power grid modeling generally uses CIM / XML or CIM / RFD formats to describe equipment ledgers, electrical parameters, and topological connections.

[0003] However, in practical engineering applications, power grid models are usually managed hierarchically according to dispatching jurisdiction, forming a multi-level independent modeling system such as national dispatch, provincial dispatch, and local dispatch. If these are to be merged and the various levels are spliced ​​together, inconsistencies in the interfaces between the models at voltage level boundaries (such as 220kV / 110kV substations) may occur, which can easily lead to problems such as topology breakpoints, duplicate nodes, terminal mismatches, or attribute conflicts after splicing.

[0004] Furthermore, traditional data quality assessments focus on a single dimension, such as the completeness of ledger fields or topological connectivity, without considering new coupling factors such as carbon emission factors and asset economic parameters. When the model is used for co-simulation of electricity, energy, and carbon, the simulation results will be biased due to the breakage of carbon flow paths or the distortion of economic attributes. At the same time, existing systems lack the ability to track the entire lifecycle of model data, and it is difficult to trace abnormal data back to specific business sources, operational events, or data versions, resulting in low efficiency in problem repair.

[0005] Although existing technologies have incorporated graph databases to store power grid models and support basic topology queries, they still cannot guarantee the self-consistency of model splicing, the quantitative evaluation of multi-dimensional quality indicators, or the automatic reasoning of anomaly propagation paths in cross-regional, multi-voltage-level, and multi-source heterogeneous scenarios. Therefore, there is an urgent need for a power grid model quality control technology architecture that can support full voltage-level coverage, multi-dimensional coupled verification, automatic source tracing, and closed-loop governance. Summary of the Invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] Therefore, the technical problem solved by this invention is: how to solve the problems of inconsistency in topology and attributes, single quality assessment dimension, difficulty in tracing abnormal data and lack of governance closed loop in the existing technology when the power grid model is spliced ​​across levels and regions, so as to achieve accurate, efficient and automated quality management of power grid model data throughout its entire life cycle.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a quality control method based on a multi-level regional power grid diagram, comprising, Multiple partition sub-graphs are obtained based on the hierarchical division of the power grid; By verifying the consistency of adjacent sub-maps at shared boundaries, the sub-maps are stitched together to form a panoramic power grid map; A multi-dimensional coupled quality assessment of the panoramic power grid diagram is performed to obtain the comprehensive health status; If the overall health status does not meet the preset requirements, the root cause change event that caused the quality abnormality will be traced back based on the map that records the lineage of power grid model changes. Based on the traced root cause change event, a targeted data verification and repair loop is triggered.

[0009] As a preferred embodiment of the quality control method based on a multi-level partitioned power grid map according to the present invention, wherein: the step of obtaining several partitioned sub-maps based on the hierarchical division of the power grid includes, The power grid is divided into multiple levels according to voltage level; For each level that is defined, a partitioned subgraph based on a public information model is constructed; Each subgraph in the partition contains the power equipment nodes and electrical connections within the current level of the power grid. Several sub-maps are used together to form a multi-level zoned power grid map covering all voltage levels.

[0010] As a preferred embodiment of the quality control method based on a multi-level partitioned power grid map according to the present invention, the step of stitching the partitioned sub-maps into a panoramic power grid map by verifying the consistency of adjacent partitioned sub-maps at shared boundaries includes: Identify and match device nodes at shared boundaries between adjacent partition subgraphs; Verify whether the topological connection relationship and multidimensional attributes of the matched device nodes in the two side subgraphs meet the preset consistency conditions; When the consistency condition is met, the adjacent partition subgraphs are stitched together at the matching device nodes to form a panoramic power grid diagram.

[0011] As a preferred embodiment of the quality control method based on a multi-level partitioned power grid map described in this invention, the step of performing a multi-dimensional coupled quality assessment on the panoramic power grid map to obtain a comprehensive health status includes: From the panoramic power grid map, extract several preset dimensions of quality indicators covering the electrical attributes of equipment, network topology attributes, carbon flow path attributes, and equipment management attributes. Based on preset rules, the quality indicators of several preset dimensions are quantitatively calculated; The quantified quality index values ​​of each dimension are weighted and fused to generate a comprehensive health status value that represents the overall data quality of the panoramic power grid map.

[0012] This invention achieves a unified quantitative perception of the comprehensive health status of the power grid model by constructing a multi-dimensional coupled quality assessment system that integrates equipment electrical attributes, network topology attributes, carbon flow path attributes, and equipment management attributes. It overcomes the limitations of traditional quality verification that only focuses on ledger completeness or topological connectivity, incorporating key requirements of new power systems such as carbon footprint traceability and asset data consistency into core quality indicators. This provides a reliable and complete data foundation for advanced applications such as electricity-carbon collaborative simulation, green trading, and lean asset management, avoiding decision-making biases caused by carbon flow path breaks or distorted management attributes.

[0013] As a preferred embodiment of the quality control method based on a multi-level partitioned power grid diagram according to the present invention, wherein: if the overall health status does not meet the preset requirements, the root cause change event leading to the quality anomaly is traced based on the map recording the power grid model change lineage, including: In response to the overall health status not meeting the preset requirements, at least one specific data object that fails to meet the quality assessment is located from the panoramic power grid map as the starting point of the anomaly to be traced. In the graph, a dependency structure for reverse tracing is constructed based on the generation dependency relationship between the abnormal starting point and historical change events; Perform traceability analysis on the dependency structure to locate the root cause of the change events by quantifying the impact of historical change events on quality anomalies.

[0014] This invention establishes a spatiotemporal-business lineage map of power grid model changes and, based on reverse dependency analysis and tracing algorithms, achieves automatic root cause localization of quality anomalies. This method overcomes the shortcomings of existing systems, such as broken data lineages and difficulty in tracing anomalies. It can automatically trace data quality issues like splicing errors and attribute conflicts back to specific historical change operations (such as a model import, update, or manual modification), accurately pinpointing "when, which system, who, and why" the anomaly was introduced. This changes the inefficient model of relying on manual experience for layer-by-layer investigation, providing a core basis for proactive and precise data governance.

[0015] As a preferred embodiment of the quality control method based on a multi-level partitioned power grid diagram described in this invention, the step of triggering a targeted data verification and repair closed loop based on the traced root cause change event includes: Push relevant information about the root cause change event to the corresponding business system to trigger a re-verification of the data entities associated with the root cause change event; Receive feedback information from the business system that is associated with the re-verification result; Based on the feedback information, the quality status of the panoramic power grid map is updated, completing the closed loop from anomaly location to repair verification.

[0016] As a preferred embodiment of the quality control method based on a multi-level partitioned power grid diagram described in this invention, the quality indicators of the several preset dimensions include indicators for evaluating basic equipment parameters, indicators for evaluating the integrity of network connectivity, indicators for evaluating the continuity of carbon emission transmission paths, and indicators for evaluating static management attributes of equipment.

[0017] This invention provides a quality control system based on a multi-level partitioned power grid diagram.

[0018] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a quality control system based on a multi-level partitioned power grid map, comprising: a data collection module, a splicing module, an evaluation module, a judgment module, and a source tracing module; The data collection module obtains multiple partition sub-graphs based on the hierarchical division of the power grid; The stitching module stitches adjacent sub-maps into a panoramic power grid map by verifying the consistency of adjacent sub-maps at the shared boundary. The evaluation module performs a multi-dimensional coupled quality assessment of the panoramic power grid map to obtain a comprehensive health status. The judgment module is used to trace the root cause change event that caused the quality abnormality based on the map that records the lineage of power grid model changes if the overall health status does not meet the preset requirements. The source tracing module triggers a targeted data verification and repair loop based on the root cause change event traced back.

[0019] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the quality control method based on a multi-level partitioned power grid diagram.

[0020] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the quality control method based on a multi-level partitioned power grid diagram.

[0021] The beneficial effects of this invention are as follows: This invention, through a hierarchical partitioning and boundary consistency verification mechanism, actively identifies and resolves conflicts in the topology connections and multidimensional attributes of shared devices before splicing, fundamentally avoiding the problems of topology breakpoints, node duplication, terminal mismatch and attribute contradictions that occur after splicing using traditional methods, and ensuring that the constructed panoramic power grid map has both topological integrity and physical interpretability.

[0022] This invention achieves a unified quantification of the comprehensive health status of power grid data by constructing a coupled evaluation system that integrates equipment electrical attributes, network topology attributes, carbon flow path attributes, and equipment management attributes. It overcomes the limitations of traditional single-dimensional (such as ledger or topology) inspections, incorporating key requirements of new power systems, such as carbon footprint traceability and asset data consistency, into the core quality monitoring scope. This provides high-quality data assurance for electricity-carbon collaborative analysis, green trading accounting, and lean asset management, effectively avoiding simulation and decision-making biases caused by missing or distorted data dimensions.

[0023] By constructing a spatiotemporal-business lineage graph that records changes throughout the entire lifecycle of a model, and based on reverse dependency analysis and graph tracing algorithms, it can automatically and accurately pinpoint data problems to historical operational events that triggered the anomalies (such as changes in specific times, systems, or operators) after quality anomalies are detected. This changes the inefficient traditional troubleshooting model that relies on manual experience and layer-by-layer backtracking, automating the process from "phenomenon discovery" to "source location," greatly improving the accuracy and efficiency of problem diagnosis.

[0024] By automatically pushing identified root cause events to relevant business systems to trigger re-verification and updating the overall status based on feedback, a technical closed loop for automatic governance and continuous optimization of data quality issues has been established. This has shortened the response cycle from problem discovery to remediation, and promoted a paradigm shift in power grid data quality management from "passive response" to proactive prevention and closed-loop control. Attached Figure Description

[0025] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a flowchart illustrating a quality control method based on a multi-level partitioned power grid diagram, as provided in one embodiment of the present invention. Detailed Implementation

[0027] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0028] Example 1, referring to Figure 1This is one embodiment of the present invention, which provides a quality control method based on a multi-level partitioned power grid diagram, including: To address the issues that existing technologies for power grid models often encounter when splicing multiple levels, such as topology and attribute conflicts, limited quality assessment dimensions, difficulty in tracing the source of anomalies, and lack of a closed-loop governance mechanism, Traditional methods, when integrating multi-level independent modeling systems, often result in topological breakpoints and semantic inconsistencies in the stitched panoramic image due to the lack of boundary consistency verification mechanisms. Their quality assessment is mostly limited to a single dimension such as electrical or topological, which cannot meet the requirements of new businesses such as carbon flow analysis and asset management for the quality of multi-attribute coupled data.

[0029] Meanwhile, due to the lack of lineage records and automatic traceability of the entire data change process, problem localization is inefficient, and there is no closed-loop governance process, resulting in data quality issues being unable to be systematically discovered and repaired.

[0030] This invention provides a quality control method based on a multi-level partitioned power grid diagram.

[0031] S1. Obtain multiple partition sub-graphs based on the hierarchical division of the power grid; S2. By verifying the consistency of adjacent partition subgraphs at the shared boundary, the partition subgraphs are stitched together into a panoramic power grid diagram; S3. Perform a multi-dimensional coupled quality assessment on the panoramic power grid diagram to obtain the comprehensive health status; S4. If the overall health status does not meet the preset requirements, then based on the map that records the lineage of power grid model changes, trace the root cause change event that led to the quality abnormality. S5. Based on the traced root cause change event, trigger a targeted data verification and repair loop.

[0032] Example 2, an embodiment of the present invention, provides a quality control method based on a multi-level partitioned power grid diagram, based on the previous embodiment, including: S1. Obtaining multiple partition subgraphs based on the hierarchical division of the power grid includes the following steps: S11. Divide the power grid into multiple levels according to voltage level.

[0033] The power grid is divided into multiple levels based on voltage levels, resulting in multiple sub-maps.

[0034] This invention includes three levels, namely L0, L1 and L2.

[0035] L0 corresponds to the 500kV and above national power grid backbone network, while L1 corresponds to the 220kV provincial backbone network and the 110kV / 66kV urban / county main distribution network.

[0036] S12. For each level that is divided, construct a partition subgraph based on the public information model.

[0037] For each level, a subgraph based on the Common Information Model (CIM) is constructed. The CIM is an abstract model that describes all major objects of a power enterprise, particularly those related to power operation. It provides a standard method for representing power system resources using object classes and attributes and the relationships between them.

[0038] Each subgraph in the partition contains the power equipment nodes and electrical connections within the current level of the power grid. S13. Several sub-maps are used to form a multi-level regional power grid map covering the entire voltage level.

[0039] S2. By verifying the consistency of adjacent sub-maps at shared boundaries, the sub-maps are stitched together to form a panoramic power grid map, including the following steps: S21. Identify and match device nodes at shared boundaries between adjacent partition subgraphs.

[0040] Identify and extract all power equipment nodes that bridge the voltage levels of this partition and adjacent partitions in the first partition subgraph to form the boundary equipment node set B1.

[0041] Identify and extract all power equipment nodes that bridge the voltage levels of this partition and adjacent partitions in the second partition subgraph to form the boundary equipment node set B2.

[0042] Based on the globally unique identifier of the device, the power equipment nodes in B1 and B2 are matched, and a bidirectional node mapping relationship is established between the first partition subgraph and the second partition subgraph to complete the alignment of shared device nodes in the two subgraphs.

[0043] S22. Verify whether the topological connection relationship and multidimensional attributes of the matched device nodes in the two side subgraphs meet the preset consistency conditions.

[0044] S23. When the consistency condition is met, the adjacent partition sub-graphs are spliced ​​together at the matching device nodes to form a panoramic power grid diagram.

[0045] When the topological equivalence condition is met and all the verified multidimensional attributes satisfy the corresponding preset tolerance threshold, it is determined that the topological connection relationship and multidimensional attributes of the shared device node in the two side subgraphs are consistent.

[0046] In this embodiment of the invention, the k-order neighborhood subgraph is a local subgraph (k≥1) formed by expanding k layers of connections outward from the aligned shared device nodes. The critical conductive path refers to the main energy transmission path through which electrical energy flows from the device to the system backbone. Terminal connection methods include the number and type of device terminals and their connected objects. Phase sequence configuration refers to the identification order of the three phases A / B / C on the terminals and connected devices.

[0047] The topological equivalence condition is equal to the following three conditions being met simultaneously: 1. The key conductive paths are consistent.

[0048] 2. The terminal connection methods are consistent.

[0049] 3. The phase sequence configuration is consistent.

[0050] The key conductive paths are consistent, including: terminals of each voltage level of the shared equipment are connected to the corresponding bus nodes in both sub-diagrams (i.e., the first and second sub-diagrams). The connection sequence of the main energy transmission path (e.g., main transformer -> bus -> outgoing line) is consistent with the electrical flow direction.

[0051] The terminal connection methods are consistent: the number and type of terminals corresponding to the voltage levels of the shared equipment nodes are consistent on both sides of the sub-diagram. All terminals have established effective electrical connections, which means that the terminals are connected to busbars, lines, or other conductive equipment nodes that match other voltage levels.

[0052] Phase sequence configuration consistency: The phase sequence identifiers of the shared equipment terminals have the same semantics and the same order in both sub-diagrams, and the phases corresponding to the targets or lines they are connected to match (ensuring three-phase topology matching).

[0053] This invention extracts a k-order neighborhood subgraph (k≥1) with the shared device as the root node, which not only verifies the device itself, but also covers its local electrical context (such as bus-switch-transformer link), thereby improving the reliability of power grid subgraph splicing.

[0054] Compared to methods that only verify topological connectivity, this invention achieves a leap from structural consistency to multidimensional semantic consistency by simultaneously verifying the consistency of electrical parameters, carbon emission factors, and economic attributes. This ensures that the spliced ​​power grid model has data reliability in various application scenarios such as power flow calculation, carbon flow analysis, and cost accounting.

[0055] In this embodiment of the invention, the first partition subgraph and the second partition subgraph are spliced ​​together at the shared device node to generate a spliced ​​power grid subgraph.

[0056] Specifically, it is based on a pre-established bidirectional node mapping relationship.

[0057] Determine the shared device nodes in the first partition subgraph Shared device nodes in the second partition subgraph .

[0058] Shared device nodes and shared device nodes Characterizes the same physical device.

[0059] Shared device nodes and shared device nodes Merged into a single shared device node .

[0060] Among them, the shared device node Inherited shared device nodes and shared device nodes The common attributes, and, provided that the multidimensional attributes are consistent, adopt the attribute value of either party; connect the original to All adjacent edges are redirected to the shared device node. and connect the original to All adjacent edges are redirected to the shared device node. Remove and The original node instances are only retained in the graph storage structure. This results in a topologically connected and non-redundant grid subgraph.

[0061] It should be noted that multidimensional attribute consistency refers to those... and All of these properties have the same values ​​or differences that are within a preset acceptable range.

[0062] In the embodiment of the present invention, the consistency condition of S22 is to take the aligned shared device node as the root node, and extract its k-order neighborhood subgraphs in the first partition subgraph and the second partition subgraph respectively, where k≥1, and the k-order neighborhood subgraph includes the shared device node and its directly or indirectly connected device nodes and connection relationships.

[0063] Perform bidirectional structural alignment between the k-order neighbor subgraph in the first partition subgraph and the k-order neighbor subgraph in the second partition subgraph, and determine whether the two satisfy the topological equivalence conditions in terms of critical conductive paths, terminal connection methods and phase sequence configurations.

[0064] Obtain the multidimensional attributes associated with the shared device node in the two side subgraphs, including electrical parameters, carbon emission factors, and economic attributes.

[0065] For each type of multidimensional attribute, calculate the attribute value deviation in the two side subgraphs, and determine whether the deviation is less than or equal to the corresponding preset tolerance threshold.

[0066] In one optional implementation, the consistency condition of S22 is to determine the device importance level based on the voltage level of the matched device node and its topological centrality in the corresponding partition subgraph. Devices whose voltage level reaches or exceeds the lower limit of the standard voltage of the transmission network (e.g., according to the State Grid standard, often 220kV is the boundary), or devices identified by the topology analysis algorithm as being located at the intersection of multiple critical electrical paths, are marked as high importance.

[0067] Subsequently, differential verification is performed: for devices with high importance, the k+1 order neighborhood subgraphs (i.e., the connection range of the extended layer) in the subgraphs on both sides are extracted for bidirectional structural alignment, and the first tolerance threshold after compression based on the device manufacturing standard tolerance is adopted (for example, 50% of the national standard or manufacturer's nominal tolerance range is taken as the consistency verification threshold).

[0068] For other devices, a standard k-order neighborhood subgraph is extracted, and a second baseline tolerance threshold set based on general engineering experience is used for verification.

[0069] Finally, for all devices, the deviations of their management attributes, such as carbon emission factors and original asset value, in the subgraphs on both sides must be less than the general tolerance thresholds for each attribute derived from business rules or historical data (e.g., an allowable deviation of ±5% for original asset value, and identical commissioning dates). Only when the topology alignment meets the equivalence condition and all attribute deviations meet their corresponding, defined threshold requirements is the device node deemed to meet the consistency condition. This solution optimizes overall verification efficiency while ensuring the reliability of the splicing process by imposing stricter verification on critical devices.

[0070] In another optional implementation, the consistency condition of S22 is to compare the multidimensional attribute values ​​of the device in both side subgraphs with a predefined device attribute consistency rule base. This rule base contains explicit logic, such as: "The 'rated voltage' attribute value of the same device must be exactly the same," and "The 'commissioning date' must not be later than the release date of the model version of the opposite side subgraph," etc. If a device attribute violates any rule, an attribute inconsistency flag is generated.

[0071] Extract k-order neighborhood subgraphs centered on the device node from both subgraphs, and abstract them into attribute graphs labeled with device type and terminal connection relationships. Calculate the minimum editing operation cost between the two attribute graphs using a graph edit distance algorithm, and convert this cost into a topological similarity score between 0 and 1 using a preset normalization function. If this score is higher than a preset topological similarity threshold (e.g., 0.85, which can be determined by the similarity statistical distribution of historical correct splicing cases), a topological consistency flag is generated.

[0072] A device node is considered to meet the consistency condition if and only if it receives both the attribute consistency flag and the topology consistency flag.

[0073] Furthermore, this invention achieves accurate matching and alignment of shared boundary devices through globally unique device identifiers, solving the misalignment problem caused by traditional methods that rely on name or local encoding matching. Then, taking the aligned device as the root node, it verifies topological equivalence by extracting and bidirectionally comparing its k-order neighborhood subgraphs, and simultaneously calculates the deviations of multi-dimensional attributes such as electrical, carbon emission, and management, ensuring the consistency between local topological semantics and global attributes, thus breaking through the limitation of traditional methods that only verify connectivity.

[0074] By making consistency verification a mandatory prerequisite for splicing operations, a proactive prevention mechanism of "verification first, then fusion" is constructed. This eliminates legacy issues such as topology breakpoints, terminal mismatches, and attribute conflicts after splicing, and generates a panoramic power grid diagram that maintains electrical continuity while satisfying semantic consistency across multiple services such as carbon flow and asset management.

[0075] S3. Perform a multi-dimensional coupled quality assessment on the panoramic power grid diagram to obtain the comprehensive health status, including the following steps: S31. Extract quality indicators for several preset dimensions of the covered equipment electrical attributes, network topology attributes, carbon flow path attributes, and equipment management attributes from the panoramic power grid map.

[0076] From the stitched panoramic power grid data, four core data sets for quality quantification were analyzed and separated: The first category is the set of electrical attributes of equipment, which refers to the key electrical parameter fields of conductive equipment in the power grid (such as transformers, lines, busbars, switches, etc.).

[0077] The second category is the network topology attribute set, which refers to the topology connection data that describes the electrical connection relationships between device terminals.

[0078] The third category is the carbon flow path attribute set, which refers to the carbon emission factors configured for nodes such as power sources and loads, as well as the topological relationships characterizing the carbon flow tracking path.

[0079] The fourth category is the equipment management attribute set, which refers to the attribute data of equipment in terms of assets and costs, such as the original value of the asset and the date of commissioning. This step clarifies the scope of input data for subsequent quantitative calculations.

[0080] S32. Based on preset rules, perform quantitative calculations on quality indicators of several preset dimensions.

[0081] For the electrical attributes of the equipment, the electrical parameter integrity (E) is calculated by statistically analyzing the ratio of the number of missing or invalid key electrical fields in all conductive equipment nodes within the area to be evaluated to the total number that should exist.

[0082] For network topology attributes, network topology connectivity (N) is calculated by statistically analyzing the ratio of the number of terminals with connection breaks to the total number of terminals that should have valid electrical connections.

[0083] For carbon flow path attributes, the carbon flow path closure degree (C) is calculated by statistically analyzing the number of load nodes whose carbon flow paths can be completely traced back to the effective carbon emission factor power source, and the ratio of this number to the total number of load nodes that need to be traced for carbon flow.

[0084] For equipment management attributes, the matching degree is calculated by comparing the numerical differences of the corresponding economic attributes of aligned shared equipment in the multi-source model. Each calculation outputs a standardized quality score.

[0085] In this embodiment of the invention, electrical parameter integrity measures whether the basic electrical attributes of power grid equipment are complete and accurate. The formula for calculating electrical parameter integrity is: ; Where D is the set of all conductive equipment nodes to be evaluated within the area to be evaluated (such as transformers, lines, busbars, switches, etc.); d is a single device. The total number of key electrical fields corresponding to the type of device d (determined by preset rules); This represents the number of key fields that are not provided or are invalid in the area to be evaluated for device d.

[0086] The formula for calculating network topology connectivity is: ; in, It is the total number of terminals (or the total number of terminal pairs associated with shared equipment) that should have valid electrical connections established within the area to be evaluated. It is the number of terminals (or the number of breakpoint connection pairs) with topological breakpoints.

[0087] The formula for calculating the carbon flow path closure is: ; Among them, when When C=1, we define C=1.

[0088] Where C represents the carbon flow path closure degree, This indicates the number of load nodes that form a complete closed carbon flow path. This indicates the total number of load nodes for carbon flow tracing.

[0089] When calculating the carbon flow path closure, the total number of load nodes in the model that require carbon flow tracing should be determined first. Then, through topological traversal and carbon intensity propagation, the number of load nodes with completely closed carbon flow paths is identified. A closed path requires that the load be traceable back to a power source node with an effective carbon emission factor, and that there are no topological breaks or missing carbon attributes along the path. It should be understood that the carbon flow path closure is a key indicator for measuring whether the carbon emission transmission path from power source to load in a power grid model is complete, traceable, and without breaks; this indicator forms the basis for supporting "electricity-carbon synergy" analysis. Since the carbon flow path itself depends on topological connectivity and the completeness of carbon attributes, its closure needs to be quantified on the set of effective carbon flow paths.

[0090] This invention uses carbon flow path integrity as a core indicator of data health, providing a reliable data foundation for green electricity trading, carbon footprint accounting, and other activities.

[0091] In one embodiment, the formula for calculating the economic attribute matching degree is: ; in, S represents the matching degree of economic attributes; S is the set of aligned shared equipment; K is the quantity of economic attributes involved in the comparison (such as original asset value, commissioning date, unit capacity investment, net value, etc.). The value of the k-th economic attribute of the shared device in the first partition subgraph; The value of the k-th economic attribute of the shared device in the second partition subgraph; It is a very small constant to avoid a denominator of 0; An indicator function for whether the k-th attribute applies to the shared device s (if an attribute does not apply, such as a capacitor with no unit capacity investment, then...). =0).

[0092] S33. The quality index values ​​of each dimension after quantitative calculation are weighted and integrated to generate a comprehensive health status value that represents the overall data quality of the panoramic power grid map.

[0093] The independent quality scores for each dimension, including electrical parameter integrity (E), network topology connectivity (N), carbon flow path closure (C), and device management attribute matching degree, calculated by S32, are weighted and summed according to weight coefficients reflecting the business importance of each dimension. These weight coefficients can be obtained by gradually converging consensus weights through multiple rounds of anonymous consultation with domain experts, or by calculating using the entropy weight method. This application does not limit the weight calculation method in its embodiments.

[0094] Generate a single, quantifiable, and comparable comprehensive data health index. This index is a quantitative value characterizing the current overall data quality level and health status of the panoramic power grid map, and can be directly used for quality rating, trend analysis, and anomaly warning.

[0095] In this embodiment of the invention, the multidimensional coupling quality assessment of S3 is achieved by calculating the quality indicators of four preset dimensions: electrical parameter integrity, network topology connectivity, carbon flow path closure, and equipment management attribute matching degree, and then synthesizing a comprehensive data health index based on preset weights.

[0096] Electrical parameter integrity is calculated by statistically analyzing the proportion of missing key electrical fields; network topology connectivity is calculated by analyzing the proportion of terminal connection interruptions; carbon flow path closure is calculated by statistically analyzing the proportion of load nodes where carbon flow can be completely traced back to the power source; and equipment management attribute matching degree is calculated by comparing the differences in economic attribute values ​​of the same shared equipment in the multi-source model.

[0097] The four quantitative results are combined according to preset weights to generate a comprehensive data health index as a comprehensive health status value.

[0098] In one alternative implementation, the multidimensional coupling quality assessment of S3 is a dynamic assessment and trend analysis method.

[0099] It not only calculates the instantaneous quality indicators of the above four dimensions of the panoramic power grid map at the current moment, but also extracts the quality indicator data of the same power grid map at multiple time points in the past from the historical version library to form the time series of indicators of each dimension.

[0100] By analyzing time series trends (e.g., using moving averages or fitted curves), we can identify deterioration trends in indicator values ​​(such as a continuous decline in carbon flow path closure) and predict their future changes. The final health status value will be composed of the current instantaneous index and a health prognosis score based on trend analysis, achieving an enhancement from static assessment to dynamic early warning.

[0101] In another alternative implementation, S3’s multidimensional coupling quality assessment is a simplified and optimized assessment method for specific business scenarios.

[0102] Based on the core needs of the target business application, one or more key dimensions are dynamically selected from the four basic dimensions for focused evaluation.

[0103] For example, when conducting carbon quota accounting simulations, the evaluation process will focus on carbon flow path attributes, deeply calculate path closure and the accuracy of carbon emission factor data, and significantly simplify the computational complexity of other dimensions. The weighting scheme will also be adjusted to be strongly correlated with the scenario (e.g., the carbon flow dimension weight is set to 0.8, and other dimensions account for 0.2). By adapting the evaluation dimensions and weights to the scenario, the relevance and efficiency of quality assessment are improved while ensuring the reliability of core business data.

[0104] This invention constructs a multi-dimensional coupled quality assessment system that integrates equipment electrical attributes, network topology attributes, carbon flow path attributes, and equipment management attributes, and quantifies them into a unified comprehensive health status value, thus completely changing the limitation of traditional methods that only conduct isolated assessments of a single dimension (such as ledgers or topology).

[0105] It not only achieves a comprehensive perception of power grid data quality from local to overall and from qualitative to quantitative, but also directly solves the business pain point of distorted carbon footprint tracing and asset value analysis caused by the lack of data dimensions by introducing new indicators such as carbon flow path closure degree and equipment management attribute matching degree. It provides a unique and reliable data quality evaluation benchmark and quantitative early warning basis for new core power system businesses such as electricity-carbon collaborative simulation, green trading and lean asset management.

[0106] S4. If the overall health status does not meet the preset requirements, then based on the map recording the lineage of power grid model changes, the root cause change event leading to the quality anomaly is traced through the following steps: S41. In response to the overall health status not meeting the preset requirements, locate at least one specific data object that does not meet the quality assessment from the panoramic power grid map as the starting point of the anomaly to be traced.

[0107] When the comprehensive data health index obtained from step S3 is lower than the preset threshold, the anomaly tracing process is triggered.

[0108] When the comprehensive data health index obtained from step S3 falls below a preset first threshold or any dimension quality indicator falls below its corresponding second threshold, the anomaly tracing process is triggered. In response to this anomaly detection result, at least one anomalous data entity node is identified. This node may be a device with missing electrical parameters, a connection with a topological break, or a load node where the carbon flow path cannot be closed. The initial anomaly score of this anomalous data entity node is set to 1, and the initial anomaly scores of the remaining data entity nodes are set to 0, thus completing the tracing initialization.

[0109] S42. In the graph, based on the generation dependency relationship between the abnormal starting point and historical change events, construct a dependency structure for reverse tracing.

[0110] Based on the anomalous data entity nodes identified by S41, the system searches within the constructed spatiotemporal-business dual-dimensional graph. This graph records all historical model change events and their impact on data entities in the form of nodes. To perform reverse tracing, the system extracts all change event nodes directly or indirectly related to the anomalous data entity node and reverses the original edge direction of "change event generating data entity," constructing a temporary reverse dependency subgraph with edges pointing from data entity nodes to change event nodes. This subgraph clearly depicts the causal dependency chain of "which historical operations might have caused the current anomaly."

[0111] It should be further explained that the steps for constructing a spatiotemporal-business dual-dimensional map include: In response to performing a model change operation on the multi-level partitioned power grid diagram, the metadata of the model change event is extracted. The metadata includes at least a timestamp, source system identifier, business event type, operator ID, and a list of affected data entities.

[0112] A unique change event node is created based on the metadata. The change event node is configured with spatiotemporal attributes and business attributes. The spatiotemporal attributes include the timestamp and the source system identifier, and the business attributes include the business event type and the operator ID.

[0113] For each data entity in the list of affected data entities, find the corresponding data entity node in the graph; if it exists, establish a first association edge from the change event node to the data entity node and a second association edge from the data entity node to the change event node; if it does not exist, create a new data entity node and establish the first and second association edges.

[0114] In this embodiment of the invention, the first associated edge represents the impact of the change event on the data entity, and the second associated edge represents the generation of the data entity by the change event. This invention automatically extracts metadata and creates change event nodes each time a model change operation is performed, while simultaneously establishing bidirectional associated edges between these nodes and the affected data entities. This fully records the entire process of the power grid model from initial modeling to historical evolution, achieving end-to-end auditing capabilities of "who did it, when, from which system, what, and which devices were modified." Traditional ledger-based management only records the current state of equipment and the last change information, without saving the historical evolution process. This invention explicitly models the power grid model change process as a graph structure with spatiotemporal and business-related semantics. This not only solves the pain point of traditional ledger-based management's inability to trace the evolution process but also provides a foundation for subsequent anomaly root cause localization based on graph neural networks.

[0115] Based on the spatiotemporal-business dual-dimensional graph, a reverse dependency subgraph is constructed. The reverse dependency subgraph contains data entity nodes and change event nodes, and the edge direction points from the data entity node to its corresponding change event node to represent the generation dependency relationship.

[0116] S43. Perform traceability analysis on the dependency structure, and locate the root cause of the change event by quantifying the impact of historical change events on quality anomalies.

[0117] Perform multi-hop graph neural network message passing on the inversely dependent subgraph, where the first... The anomaly score of any change event node is updated based on the anomaly scores of its downstream data entity nodes, the time decay factor, and the business event weight. Specifically, for any change event node v, the update expression for its anomaly score is: For any change event node v, its new anomaly score is: ; in, Weighting of business events The time decay coefficient, It is the difference between the current time and the timestamp of the change event node. As the attenuation factor, Represents a node The anomaly score is calculated. It should be noted that the business event weights are pre-configured based on the event type of the change event node. The weights reflect the historical risk level of data anomalies introduced by various operations. Specifically, these weights are set based on power system operation and maintenance experience and historical accident statistics. The time decay coefficient γ is a configurable parameter used to control the intensity of the impact of historical change events on the current anomaly score. Preferably, this parameter can be adjusted according to the specific power grid system's operation and maintenance strategy and data update frequency.

[0118] After a preset number of hops, the final anomaly score of all change event nodes is obtained, and the change event node with the highest anomaly score is determined as the original node that generated the anomaly data.

[0119] Furthermore, it should be noted that constructing the reverse dependency subgraph includes: starting with the abnormal data entity node, traversing upstream along the generation dependency relationship in the spatiotemporal-business dual-dimensional graph, extracting all data entity nodes and change event nodes that have direct or indirect generation dependency relationships with the abnormal data entity node, and reversing the generation edges in the spatiotemporal-business dual-dimensional graph that originally pointed from the change event node to the data entity node, to construct a local subgraph composed of directed edges pointing from the data entity node to the change event node; wherein, the reverse dependency subgraph is used for causal backtracking propagation from the abnormal data entity node to the historical change event node.

[0120] In this embodiment of the invention, the tracing of the root cause change event of the quality anomaly in S4 is achieved by locating the starting point of the abnormal data, constructing a reverse dependency subgraph, and performing message passing analysis based on a graph neural network.

[0121] First, in response to the failure of the overall health status to meet the requirements, specific abnormal data entity nodes are identified from the panoramic power grid map as the starting point for tracing, and their abnormal scores are initialized. Second, in the spatiotemporal-business dual-dimensional graph that records the full history of changes, the system traverses backward from this starting point to construct a reverse dependency subgraph that depicts the generation dependency relationship of "data entity → change event". Finally, multi-hop graph neural network message passing is performed on this subgraph. Taking into account the scores of downstream nodes, time decay, and event type weights, the abnormal contribution score of each historical change event is dynamically calculated. The event with the highest score is identified as the root cause change event, and its complete meta-information is output, thereby completing the automated and accurate attribution from phenomenon to source.

[0122] In one alternative implementation, S4 traces the root cause change event leading to the quality anomaly using a rapid tracing method based on direct association and rule matching. This is suitable for scenarios where the anomaly can be clearly associated with a specific data entity (such as the loss of a device parameter) and the change history is clear.

[0123] The system first directly associates and matches anomalous data entities with change event nodes in the graph. If an entity is generated by a single recent change event, that event is directly identified as a suspected root cause. Simultaneously, a set of predefined root cause determination rules is used (e.g., "If the anomalous attribute is 'carbon emission factor' and its value was modified by the most recent model import event, then that import event is prioritized as the root cause") to cross-validate and rank the directly associated results. This eliminates complex graph propagation calculations and, under certain conditions, enables root cause localization within seconds, making it suitable for simple anomaly scenarios with high real-time requirements.

[0124] In another alternative implementation, tracing the root cause change event leading to the quality anomaly in S4 is an interactive, progressive tracing method that integrates automated reasoning with human feedback. In this approach, the system first performs automated tracing analysis based on a graph neural network to obtain a "list of suspected root cause events" sorted by anomaly score.

[0125] When the score of the top-ranked event does not reach the absolute confidence threshold, or when there are multiple events with similar scores in the list, the system will not force a single result. Instead, it will push the list along with relevant contextual evidence (such as comparison of changed content and impact analysis) to domain experts. Experts can review, select, or supplement information based on their experience. Their feedback (such as confirming an event as the root cause) will be recorded by the system and used to optimize subsequent model weights and rules. This method combines the efficiency of automated analysis with the judgment of human experts, making it particularly suitable for complex and highly coupled anomaly scenarios, improving the acceptability of source tracing results and the reliability of decision-making.

[0126] This invention constructs a spatiotemporal-business dual-dimensional graph recording the entire lifecycle changes of a data model, automatically builds a reverse dependency subgraph based on abnormal data entities, and finally uses a graph neural network message passing algorithm to quantify and score the abnormal contribution of historical change events, achieving automatic, accurate, and efficient location of the root causes of quality anomalies. This completely changes the inefficient backtracking mode that relies on manual experience and layer-by-layer querying of logs and change records, solving the inherent pain point of "anomalies are visible, but the root cause is difficult to find" caused by broken data lineage.

[0127] S5. Based on the traced root cause change event, trigger a targeted data verification and repair closed loop, including the following steps: S51. Push relevant information about the root cause change event to the corresponding business system to trigger a re-verification of the data entity associated with the root cause change event.

[0128] The metadata of the root cause change event (i.e., the original generating node) located in step S43, including the timestamp, source system identifier, business event type, and operator ID, is pushed to the source business system to which the event belongs via an inter-system interface (e.g., in the form of a repair work order). This push operation will automatically trigger the target business system to initiate a targeted re-verification and data repair process for the specific data entities associated with the event (such as incorrectly modified device parameters or connection relationships).

[0129] S52, Receive feedback information from the business system that is associated with the re-verification result.

[0130] After the business system completes its internal data repair and verification, it receives feedback information directly related to the results of this re-verification. This feedback information is used to clearly inform the quality control system of the result of the repair operation, such as "repair successful and data has been corrected", "repair failed and manual intervention is required", or "data has been verified to be correct".

[0131] S53. Based on the feedback information, update the quality status of the panoramic power grid map to complete the closed loop from anomaly location to repair verification.

[0132] Based on the feedback received in S52, the system executes the final verification action of the closed loop: if the feedback indicates successful repair, it automatically triggers an update to the panoramic power grid diagram (e.g., re-retrieves the repaired data) and re-executes the multi-dimensional coupled quality assessment in step S3 to verify whether the overall health status has returned to normal; if the feedback indicates other statuses, it updates the issue tracking status and may trigger an escalation process. Thus, the system completes a full technical closed loop from quality anomaly detection (S3), automatic root cause tracing (S4), to repair triggering and verification (S5), achieving autonomous governance of data quality issues.

[0133] It should be further explained that in this embodiment of the invention, the economic attribute matching degree is a quantitative indicator used to measure whether the asset and cost-related economic attributes recorded for the same physical device in the multi-source power grid model are consistent or reasonable. If the economic attributes of the same device differ too much between two sub-graphs, it indicates that the data is conflicting or unreliable, affecting business operations such as asset lifecycle management, green finance, and cost allocation.

[0134] For the closed-loop governance interface, this invention pushes the original generating node to the corresponding business system through a repair work order. After the power grid staff repairs the original generating node, a repair completion message is sent. Upon detecting the repair completion message, a re-verification is automatically triggered, forming a closed-loop technology of detection-diagnosis-repair-verification, which significantly shortens the response and repair cycle of data quality issues.

[0135] This invention divides the power grid into multi-level sub-graphs according to voltage levels. Before splicing the sub-graphs, it verifies the consistency of topological connections and multi-dimensional attributes. Only when the topological connections and multi-dimensional attributes of the shared equipment nodes in both sub-graphs are consistent are the first and second sub-graphs spliced ​​at the shared equipment nodes. This solves the problem of inconsistent interfaces at voltage level boundaries between different levels of models, which leads to topological breakpoints, node duplication, terminal mismatch, or attribute conflicts after splicing. This improves the topological integrity and physical interpretability of the spliced ​​panoramic power grid diagram.

[0136] Traditional quality verification methods are limited to ledger fields or topological connectivity, failing to support the modeling needs of new power systems for coupled factors such as carbon flow and economics. This invention utilizes a four-dimensional coupled data quality evaluator to calculate electrical parameter integrity, network topological connectivity, carbon flow path closure, and economic attribute matching, synthesizing a comprehensive data health index. This provides a unified, quantifiable, comparable, and predictable indicator system for model quality. Furthermore, by considering novel coupled factors such as carbon emission factors and asset economic parameters, it avoids simulation result deviations caused by carbon flow path breaks or distorted economic attributes.

[0137] To address the lack of data lineage tracing capabilities in existing systems, this invention constructs a spatiotemporal-business dual-dimensional graph. When the comprehensive data health index falls below a preset first threshold or any dimension quality indicator falls below its corresponding second threshold, the root cause reasoning engine uses a graph neural network to trace back along the spatiotemporal-business dual-dimensional graph, locating the original generation node. This enables automatic attribution from abnormal phenomena to root cause errors. Combined with a closed-loop governance interface, the original generation node can be pushed to the corresponding business system and automatically triggered for re-verification, forming a closed-loop technology of detection-diagnosis-repair-verification, significantly shortening the response and repair cycle for data quality issues.

[0138] Example 3 is an embodiment of the present invention, which provides a quality control system based on a multi-level partitioned power grid map, including a data collection module, a splicing module, an evaluation module, a judgment module, and a source tracing module; The data collection module obtains multiple partition sub-graphs based on the hierarchical division of the power grid; The stitching module stitches adjacent sub-maps into a panoramic power grid map by verifying the consistency of adjacent sub-maps at the shared boundary. The evaluation module performs a multi-dimensional coupled quality assessment of the panoramic power grid map to obtain a comprehensive health status. The judgment module is used to trace the root cause change event that caused the quality abnormality based on the map that records the lineage of power grid model changes if the overall health status does not meet the preset requirements. The source tracing module triggers a targeted data verification and repair loop based on the root cause change event traced back.

[0139] This embodiment also provides an electronic device applicable to a quality control method based on a multi-level partitioned power grid diagram, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the quality control method based on a multi-level partitioned power grid diagram as proposed in the above embodiment.

[0140] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a quality control method based on a multi-level partitioned power grid diagram as proposed in the above embodiments.

[0141] The storage medium proposed in this embodiment and the quality control method based on a multi-level partitioned power grid diagram proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0142] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0143] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A quality control method based on a multi-level regional power grid diagram, characterized in that: include, Multiple partition sub-graphs are obtained based on the hierarchical division of the power grid; By verifying the consistency of adjacent sub-maps at shared boundaries, the sub-maps are stitched together to form a panoramic power grid map; A multi-dimensional coupled quality assessment of the panoramic power grid diagram is performed to obtain the comprehensive health status; If the overall health status does not meet the preset requirements, the root cause change event that caused the quality abnormality will be traced back based on the map that records the lineage of power grid model changes. Based on the traced root cause change event, a targeted data verification and repair loop is triggered.

2. The quality control method based on a multi-level regional power grid diagram as described in claim 1, characterized in that: The hierarchical division based on the power grid to obtain several partition sub-graphs includes, The power grid is divided into multiple levels according to voltage level; For each level that is defined, a partitioned subgraph based on a public information model is constructed; Each subgraph in the partition contains the power equipment nodes and electrical connections within the current level of the power grid. Several sub-maps are used together to form a multi-level zoned power grid map covering all voltage levels.

3. The quality control method based on a multi-level regional power grid diagram as described in claim 2, characterized in that: The step of stitching together the partitioned subgraphs into a panoramic power grid map by verifying the consistency of adjacent partitioned subgraphs at shared boundaries includes: Identify and match device nodes at shared boundaries between adjacent partition subgraphs; Verify whether the topological connection relationship and multidimensional attributes of the matched device nodes in the two side subgraphs meet the preset consistency conditions; When the consistency condition is met, the adjacent partition subgraphs are stitched together at the matching device nodes to form a panoramic power grid diagram.

4. The quality control method based on a multi-level regional power grid diagram as described in claim 3, characterized in that: The multi-dimensional coupled quality assessment of the panoramic power grid map yields a comprehensive health status, including: From the panoramic power grid map, extract several preset dimensions of quality indicators covering the electrical attributes of equipment, network topology attributes, carbon flow path attributes, and equipment management attributes. Based on preset rules, the quality indicators of several preset dimensions are quantitatively calculated; The quantified quality index values ​​of each dimension are weighted and fused to generate a comprehensive health status value that represents the overall data quality of the panoramic power grid map.

5. The quality control method based on a multi-level partitioned power grid diagram as described in claim 4, characterized in that: If the overall health status does not meet the preset requirements, then based on the map recording the lineage of power grid model changes, the root cause change events leading to the quality anomalies are traced, including: In response to the overall health status not meeting the preset requirements, at least one specific data object that fails to meet the quality assessment is located from the panoramic power grid map as the starting point of the anomaly to be traced. In the graph, a dependency structure for reverse tracing is constructed based on the generation dependency relationship between the abnormal starting point and historical change events; Perform traceability analysis on the dependency structure to locate the root cause of the change events by quantifying the impact of historical change events on quality anomalies.

6. The quality control method based on a multi-level partitioned power grid diagram as described in claim 5, characterized in that: The targeted data verification and repair closed loop triggered based on the traced root cause change event includes: Push relevant information about the root cause change event to the corresponding business system to trigger a re-verification of the data entities associated with the root cause change event; Receive feedback information from the business system that is associated with the re-verification result; Based on the feedback information, the quality status of the panoramic power grid map is updated, completing the closed loop from anomaly location to repair verification.

7. The quality control method based on a multi-level regional power grid diagram as described in claim 6, characterized in that: The aforementioned quality indicators across several preset dimensions include indicators for evaluating basic equipment parameters, indicators for evaluating the integrity of network connectivity, indicators for evaluating the continuity of carbon emission transmission paths, and indicators for evaluating the static management attributes of equipment.

8. A quality control system based on a multi-level partitioned power grid diagram, employing the quality control method based on a multi-level partitioned power grid diagram as described in any one of claims 1 to 7, characterized in that, include: The system includes a data collection module, a data assembly module, an evaluation module, a judgment module, and a source tracing module. The data collection module obtains multiple partition sub-graphs based on the hierarchical division of the power grid; The stitching module stitches adjacent sub-maps into a panoramic power grid map by verifying the consistency of adjacent sub-maps at the shared boundary. The evaluation module performs a multi-dimensional coupled quality assessment of the panoramic power grid map to obtain a comprehensive health status. The judgment module is used to trace the root cause change event that caused the quality abnormality based on the map that records the lineage of power grid model changes if the overall health status does not meet the preset requirements. The source tracing module triggers a targeted data verification and repair loop based on the root cause change event traced back.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the quality control method based on a multi-level partitioned power grid diagram as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the quality control method based on a multi-level partitioned power grid diagram as described in any one of claims 1 to 7.

Images