A semantic enhancement and dynamic completion method and system for power market data graph

Abstract

The present application provides a method and system for semantic enhancement and dynamic completion of an electricity market data map. Among them, the present application collects target entity information and external data of the electricity market data map by responding to data query instructions; after converting external data into structured semantic labels, traverses the topological structure, establishes a mapping relationship between labels and entity attributes, combines the query instruction to influence the logic calculation topological completeness index and marks missing entities or associations to form a missing tag set; establishes attribute mapping based on the target entity and external data, generates a semantic vector that integrates cross-domain features, and projects it into the topological space according to the missing tag set to form cross-domain feature topological nodes and connections; embeds missing positions to reconstruct the topological structure and verify logical connectivity; parses the dynamic conduction chain and generates a feedback interface containing semantic relationships and completion structures to respond to queries. The present application realizes the closed-loop verification of the conduction logic of missing elements in the electricity market data map.

Description

Technical Field

[0001] The present application relates to the field of power market informatization technology, and in particular to a method and system for semantic enhancement and dynamic completion of power market data graphs. Background Art

[0002] Electricity market trading decisions require the establishment of a dynamic knowledge graph to integrate multi-source heterogeneous data. The graph must achieve semantic fusion of unstructured external data and structured operational data, dynamically repair missing associations between entities caused by market changes, and ensure that the completed topology conforms to the unique transmission rules of the electricity market.

[0003] At present, some adopt static graph completion technology based on ontology mapping, standardize entities through predefined power trading ontology library, use rule engine to match external data to ontology fields, supplement static association edges based on historical co-occurrence frequency when missing data is detected, and finally execute preset integrity verification rules to verify the completeness of basic data.

[0004] However, this technology has essential limitations: it relies on predefined ontological rules and static historical statistics, and is unable to resolve implicit relationships in cross-domain data. During the completion process, it ignores the constraints of the power market transmission mechanism, resulting in topological logic conflicts. Moreover, the response interface only displays isolated completed entities and cannot restore the dynamic transmission path, making it difficult to support decision traceability. Summary of the Invention

[0005] The present application provides a method and system for semantic enhancement and dynamic completion of power market data graphs, which is used to solve the problem in the existing technology that it is impossible to dynamically establish a closed loop of transmission paths between cross-domain data.

[0006] In a first aspect, the present application provides a method for semantic enhancement and dynamic completion of a power market data graph, including:

[0007] In response to the data query instruction sent by the user, the target entity information of structured operation and unstructured external data in the pre-established power market data map are collected;

[0008] Convert the external data into structured semantic tags, traverse the topological structure of the target entity information, establish a mapping relationship between the structured semantic tags and the target entity attributes, combine the impact logic in the data query instruction, calculate the completeness index of the topological structure based on the mapping relationship, and mark the missing entities or missing associations related to the data query instruction to generate a missing tag set;

[0009] Based on the target entity information and the external data, an attribute mapping is established between the external data attributes and the entity attributes, and a semantic vector integrating cross-domain features is generated by parsing the implicit relationship between the external data and the target entity information in the attribute mapping. The semantic vector is projected into a topological space according to the missing tag set to generate topological nodes and topological connections that inherit the cross-domain features.

[0010] Embedding the topological nodes and topological connections into the missing marked positions of the power market data map, reconstructing the topological structure, and verifying the logical connectivity of the completed topological structure based on the relationship conduction rules;

[0011] Based on the verified topological structure, the dynamic transmission chain formed by the completion path between the target entity information is parsed, and a feedback interface including semantic relationships and completion structures is generated to respond to data query instructions.

[0012] Optionally, it also includes:

[0013] For the target entity information, the ontology layer in the pre-built electricity market data map is called to identify the domain category to which the entity belongs and the domain relationship category between the entities;

[0014] In the data layer of the power market data map, neighbor node features and multi-hop path features associated with the identified entities and relationships are extracted as entity context features;

[0015] Fusing the domain categories to which entities belong, the domain relationship categories between entities, and the entity context features to generate a domain-enhanced semantic vector;

[0016] The domain-enhanced semantic vector is injected into the semantic vector of the cross-domain feature fusion, and the enhanced semantic vector is output.

[0017] Optionally, in the topological space of the power market data map, based on the structural similarity and attribute similarity between entities, the cross-domain entities having an association mapping with the target entity information are located, and the cross-domain entities and their association relationships are output;

[0018] Taking the cross-domain entity and the target entity as anchor points, dynamically constructing a reasoning sub-network including multi-hop neighbors in the topological space, performing logical rule reasoning along the multi-hop path in the dynamic reasoning sub-network, and outputting a reasoning result;

[0019] fusing the target entity information, the cross-domain entities and their associations, and the reasoning results to generate a joint semantic representation;

[0020] Decomposing the joint semantic representation into sub-vector sets according to entity levels, hierarchically projecting the sub-vector sets into a topological space based on voltage levels and service types, and generating completion nodes and associated edges corresponding to the missing tag sets;

[0021] The business logic compliance of the completed nodes and associated edges is verified through a rule engine, and cross-level backtracking correction is performed on the nodes that fail the verification until the logical connectivity requirements of the topology structure are met.

[0022] Optionally, in the process of establishing attribute mapping, a pre-built electric power field terminology library is called to identify professional terms in the external data and target entity information, perform semantic disambiguation, and output standardized entity attributes;

[0023] By using the standardized entity attributes, when traversing the target entity information and semantic labels, it is detected whether it hits the predefined atomic indicators, derived indicators or composite indicators. If it hits, the reconstructed node connection data completion path and calculation logic description are output according to the predefined indicator calculation rules;

[0024] Based on the node connection data completion path and calculation logic description, a sub-network containing relevant entities, attributes and calculation rules is dynamically constructed in the topological space, the indicator definition rules are executed in the sub-network, and the calculated predefined indicator value is output;

[0025] The predefined indicator value is injected into the semantic vector of the corresponding entity or relationship as a cross-domain feature, and based on the semantic vector containing the indicator value, the regularized diagnostic template associated with the indicator is called to perform abnormality analysis and interpretation.

[0026] Optionally, based on the target entity information and the external data, an attribute mapping between external data attributes and entity attributes is established; by parsing the implicit relationship between the external data and the target entity information in the attribute mapping, a semantic vector integrating cross-domain features is generated; the semantic vector is projected into a topological space according to the missing tag set, and topological nodes and topological connections that inherit the cross-domain features are generated, including:

[0027] Based on the target entity information and the external data, a bidirectional mapping dictionary between external data attributes and entity attributes is constructed, recording direct mapping rules, mathematical derivation formulas, and unit conversion coefficients between attributes;

[0028] By using a deep semantic association model, the implicit multi-order inference relationship between attributes in the bidirectional mapping dictionary is analyzed to identify cross-domain attribute combinations with common variation characteristics or causal constraints;

[0029] Generate a semantic vector for each cross-domain attribute combination, wherein the semantic vector integrates the encoding value of the associated attribute, the relationship weight coefficient and the time series mark;

[0030] According to the topological coordinates located by the missing tag set, a blank topological node is created in the power market data map, and the semantic vector is projected to the corresponding blank topological node to form a topological node carrying cross-domain features;

[0031] According to the relationship weight coefficient in the semantic vector, a weighted topological connection that inherits the cross-domain feature is established between the topological nodes carrying the cross-domain feature. The attributes and connection relationships of the topological nodes and topological connections must meet the topological constraint rules of the power market map.

[0032] Optionally, a deep semantic association model is used to parse implicit multi-order inference relationships between attributes in the bidirectional mapping dictionary and identify cross-domain attribute combinations with common variation characteristics or causal constraints, including:

[0033] Receiving a mapping entry set from the bidirectional mapping dictionary through a deep semantic association model, the mapping entry set records the correspondence rules and relationship identifiers between external data attributes and target entity attributes, and constructing an inference network with the mapping entry set;

[0034] Performing a traversal operation on the inference network, identifying consistency of time dimension fluctuation characteristics between nodes in the inference network, and recording a common change attribute group, detecting a conduction path that satisfies a causal constraint rule in the inference network and recording a response sequence;

[0035] Calculating the fluctuation consistency quantization value of the common change attribute group and measuring the trigger delay time of the response sequence; when the fluctuation consistency quantization value exceeds a set threshold and the trigger delay time is within a preset interval, outputting a cross-domain attribute combination with a relationship identifier.

[0036] Optionally, based on the verified topological structure, a dynamic transmission chain formed by the completion path between target entity information is parsed, and a feedback interface including semantic relationships and completion structures is generated to respond to data query instructions, including:

[0037] Based on the verified topological structure, a dynamic conduction chain formed by the completion path between target entity information is extracted, and the relationship weight value and time sequence marking trajectory of each conduction path in the dynamic conduction chain are analyzed;

[0038] Mapping the relationship weight value to a preset semantic strength level, converting the time sequence mark trajectory into a time axis, generating a node relationship chain with a conduction direction identifier, and marking the semantic strength level and the time axis on the node relationship chain;

[0039] The conduction paths that fail to pass the verification are converted into warning signs, the node relationship chain and the warning signs are integrated, and a visual feedback interface is generated to respond to data query instructions.

[0040] Optionally, the external data is converted into structured semantic tags, the topological structure of the target entity information is traversed, a mapping relationship between the structured semantic tags and the target entity attributes is established, and the completeness index of the topological structure is calculated based on the mapping relationship in combination with the impact logic in the data query instruction, and missing entities or missing associations related to the data query instruction are marked to generate a missing tag set, including:

[0041] Perform semantic annotation processing on the unstructured external data, extract attribute categories and numerical descriptions, generate structured semantic labels, traverse its associated topological structure starting from the target entity, and identify the conduction paths between all entities and the attribute associations on the paths;

[0042] Establishing a mapping relationship between the structured semantic tags and the target entity attributes, combining the impact logic in the data query instruction, calculating the completeness index of each conduction path in the mapping relationship, and simultaneously detecting the associated interruption points in the conduction path that violate the impact logic;

[0043] When the completeness index is lower than a preset threshold, or there is an association interruption point, the corresponding entity is marked as a missing entity or the association is marked as a missing association, and the topological coordinates of all the missing entities and missing associations are summarized to generate a missing mark set.

[0044] Optionally, embedding the topological nodes and topological connections into missing marked positions of the power market data map, reconstructing the topological structure, and verifying the logical connectivity of the completed topological structure based on relationship conduction rules, including:

[0045] Locating entity missing positions and associated missing positions in the power market data map according to the missing mark set, embedding the topological nodes into the entity missing positions and writing attribute coding values, embedding the topological connections into the associated missing positions and configuring weight coefficients, and reconstructing the topological structure;

[0046] Traverse the reconstructed topological structure based on the relational conduction rules and verify the conduction direction, weight attenuation gradient and time sequence continuity of the newly added conduction path;

[0047] If the conduction direction meets the preset constraints, the weight attenuation gradient is within the allowable range, and the timing mark is continuous, the conduction path is determined to be logically connected, and the conduction paths and failure rules that fail the logical connectivity verification are recorded.

[0048] In a second aspect, the present application provides a semantic enhancement and dynamic completion system for power market data graphs, including:

[0049] The acquisition module, in response to the data query instruction sent by the user, collects the target entity information of structured operation and unstructured external data in the pre-established power market data map;

[0050] a mapping module that converts the external data into structured semantic tags, traverses the topological structure of the target entity information, establishes a mapping relationship between the structured semantic tags and the target entity attributes, calculates a completeness index of the topological structure based on the mapping relationship in combination with the impact logic in the data query instruction, and marks missing entities or missing associations related to the data query instruction to generate a missing tag set;

[0051] A generation module, based on the target entity information and the external data, establishes an attribute mapping between external data attributes and entity attributes, generates a semantic vector that integrates cross-domain features by parsing the implicit relationship between the external data and the target entity information in the attribute mapping, projects the semantic vector into a topological space according to the missing tag set, and generates topological nodes and topological connections that inherit the cross-domain features;

[0052] a reconstruction module, which embeds the topological nodes and topological connections into the missing marked positions of the power market data map, reconstructs the topological structure, and verifies the logical connectivity of the completed topological structure based on the relationship transmission rules;

[0053] The verification module analyzes the dynamic transmission chain formed by the completion path between the target entity information based on the verified topological structure, and generates a feedback interface including semantic relationships and completion structures to respond to data query instructions.

[0054] This application collects target entities and external data in a targeted manner in response to data query instructions, establishes a mapping relationship between structured semantic labels and entity attributes to achieve cross-domain data semantic integration; combines the influencing logic to calculate the topological integrity and mark the missing set, breaking through the limitations of static detection to accurately locate the conduction fault; constructs attribute mapping to parse implicit relationships to generate semantic vectors that integrate cross-domain features, and projects to form topological elements that inherit external characteristics; after embedding the missing positions to reconstruct the topological structure, the logical connectivity is verified based on the relationship conduction rules to ensure that the completion path complies with the constraints of the power market; finally, the dynamic conduction chain is analyzed to generate visual feedback, forming a technical closed loop from cross-domain data fusion, rule constraint completion to interpretable output of the conduction path.

[0055] Furthermore, by calling the pre-built ontology layer to identify the domain categories and relationship types of power market entities, the associated neighbor nodes and multi-hop path features are extracted based on the data layer to form entity context features, and the domain category information and context features are integrated to generate domain-enhanced semantic vectors, which are injected into the cross-domain semantic vectors to achieve dual feature enhancement; this mechanism eliminates potential conflicts between external data and power market rules through domain knowledge constraints, such as correcting the deviation between meteorological load forecasts and the actual regulation capabilities of units, and optimizing the transmission capacity constraints of cross-regional topological connections, ensuring that the generated topological nodes and connections inherit cross-domain data features and comply with power operation rules, and finally completing domain adaptability enhancement in the semantic vector layer, providing dual verification guarantees of domain rules and external features for topological projection.

[0056] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0058] Figure 1 A flowchart of a method for semantic enhancement and dynamic completion of a power market data graph provided by the present application is shown;

[0059] Figure 2 A scenario diagram showing a method for semantic enhancement and dynamic completion of a power market data graph provided by this application is shown;

[0060] Figure 3 A structural diagram of a semantic enhancement and dynamic completion system for an electricity market data graph provided by the present application is shown. DETAILED DESCRIPTION

[0061] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0062] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0063] Researchers have discovered that existing methods for analyzing electricity market data have serious shortcomings: they rely on predefined rules and static data, making it difficult to integrate heterogeneous information from multiple sources; when completing data, they ignore the market's unique transmission mechanisms, resulting in frequent logical conflicts in the reconstructed topology; and the completion results are isolated, failing to clearly demonstrate the dynamic transmission path of information, making it difficult to provide traceability support for decision-making. Therefore, a dynamic completion method that can deeply integrate data, ensure logical consistency, and visualize the transmission chain is urgently needed.

[0064] In response to the above problems, the present invention proposes a method for semantic enhancement and dynamic completion of the power market data map. The core of the method is to use attribute mapping to generate a semantic vector that integrates cross-domain features, and to achieve logically self-consistent dynamic completion and conduction chain analysis under the constraints of conduction rules. Specifically, the method responds to query instructions, collects target entity information and external data; structures external data and analyzes topological completeness, marking missing points; establishes attribute mapping and analyzes implicit relationships, generates fused semantic vectors, and intelligently projects them to topological missing locations; applies relationship conduction rules to strictly verify the logical connectivity of the new structure; and finally analyzes and displays the dynamic conduction chain formed by the completed entities and their semantic relationships. This method effectively overcomes the limitations of predefined rules, achieves deep cross-domain data fusion, avoids topological logic conflicts caused by ignoring the conduction mechanism, and accurately restores the dynamic conduction path, significantly improving the traceability support capabilities for market decision-making.

[0065] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0066] Figure 1 The present invention provides a flowchart of a method for semantic enhancement and dynamic completion of a power market data map, such as Figure 1 As shown, the method includes:

[0067] 101. In response to a data query instruction sent by a user, collect structured target entity information and unstructured external data in a pre-established power market data map;

[0068] In the above steps, a data query instruction refers to a command sent by a user to request electricity market data. A power market data graph refers to a pre-established graph structure describing power market entities and their relationships. Structured operational entity information refers to data entities stored and manipulated in a structured format. Target entity information refers to the specific data entity content specified by the user instruction. Unstructured external data refers to data from external sources that does not adhere to a fixed format. The specific external data collected can be obtained based on the data the user is querying.

[0069] In an embodiment of the present application, first, in response to a data query instruction sent by a user, the instruction is received and parsed to determine the target entity information and unstructured external data requirements that need to be collected; secondly, based on the parsing result, the structured target entity information is collected from a pre-established electricity market data map to extract the required data; then, according to the specific content of the user's query instruction, the corresponding unstructured external data is further obtained, for example, when the user asks about the correlation between power load and meteorology, external data is obtained from the meteorological bureau's data source, or when the user asks about the scope of influence of electricity price policy, relevant external policy documents are obtained; finally, the entire collection process is completed to ensure that the target entity information and external data are successfully integrated and obtained.

[0070] In actual applications, on a certain power market operation platform, the user enters a data query instruction about the recent regional power supply and demand through the interface. The system immediately responds to the instruction and automatically extracts the structured operation information of seven target entities from the pre-established power market data map, including the maximum output value and node electricity price data of the power plant. At the same time, it uses crawler tools to collect three unstructured market analysis reports and policy documents from external websites. The entire collection process is completed within 15 milliseconds, successfully integrating key data, greatly enhancing the information analysis and decision support capabilities, and helping users to efficiently identify potential market risks.

[0071] After receiving a user's data query instruction, this application system can obtain structured target entity information from a pre-established power market data map, while simultaneously collecting unstructured external data, effectively integrating the two data types to achieve a comprehensive query response, thereby improving the efficiency of data mining and the depth of the results. This process ensures that the query results integrate internally structured operational objectives and rich unstructured supplementary information, providing users with accurate and detailed power market data analysis support.

[0072] 102. Convert the external data into structured semantic tags, traverse the topological structure of the target entity information, establish a mapping relationship between the structured semantic tags and the target entity attributes, combine the impact logic in the data query instruction, calculate the completeness index of the topological structure based on the mapping relationship, and mark the missing entities or missing associations related to the data query instruction to generate a missing tag set;

[0073] Optionally, step 102 may specifically include the following steps:

[0074] 1021. Perform semantic annotation processing on the unstructured external data, extract attribute categories and numerical descriptions, generate structured semantic labels, traverse the associated topological structure starting from the target entity, and identify the conductive paths between all entities and the attribute associations on the paths;

[0075] 1022. Establish a mapping relationship between the structured semantic tag and the target entity attribute, calculate the completeness index of each conduction path in the mapping relationship in combination with the influence logic in the data query instruction, and simultaneously detect the associated interruption points in the conduction path that violate the influence logic;

[0076] 1023. When the completeness index is lower than a preset threshold, or there is an association interruption point, the corresponding entity is marked as a missing entity or the association is marked as a missing association, and the topological coordinates of all the missing entities and missing associations are summarized to generate a missing mark set.

[0077] In the above steps, structured semantic labeling refers to the conversion of unstructured external data into structured identifiers containing attribute categories and numerical descriptions through semantic annotation. The topological structure of target entity information refers to the network structure that describes its associated entities and transmission relationships, starting from the target entity. Transmission paths refer to the attribute transmission links between entities in the topological structure. Attribute associations refer to the dependencies between entity attributes along the transmission paths. Mapping relationships refer to the corresponding connections between structured semantic labels and target entity attributes. Influence logic refers to the attribute change transmission rules defined in data query instructions. The completeness index is a quantitative assessment of the attribute coverage of a single transmission path in the mapping relationship. Association breakpoints refer to transmission path breakpoints that violate the influence logic. Missing entities refer to entities that do not meet the completeness requirements or are located at breakpoints. Missing associations refer to association edges in the transmission path that do not satisfy the influence logic. Topological coordinates refer to the locations of missing entities or associations in the topological structure. The missing tag set is the set of all topological coordinates.

[0078] In an embodiment of the present application, first, step 1021 performs semantic annotation processing on unstructured external data, uses an entity recognition algorithm to extract attribute categories and numerical descriptions, and generates structured semantic labels; secondly, starting from the target entity, a breadth-first traversal algorithm is used to scan the topological structure, identify the conduction paths between all entities, and record the attribute associations on the paths; finally, the generated structured semantic labels and the topological data set containing the conduction path attribute associations are output as input for subsequent steps.

[0079] Then, in step 1022, a mapping relationship table is established between structured semantic labels and target entity attributes using the topological dataset. Next, each transmission path is traversed and, in conjunction with the impact logic rules in the data query instruction, an attribute matching algorithm is used to calculate the attribute coverage of the path in the mapping relationship as a completeness indicator. To calculate the attribute coverage, the algorithm first traverses each node attribute in the transmission path, checks the mapping relationship table for any matching attributes, and then counts the number of matching attributes and divides it by the total number of attributes in the mapping table to obtain a ratio. The formula is: attribute coverage = (number of matching attributes / total number of attributes) × 100%). Next, the transmission path is simultaneously detected for any associated breakpoints that violate the impact logic, such as by checking for missing or conflicting variables using the logic rule engine. Finally, a transmission path analysis set with completeness indicators and associated breakpoints is output.

[0080] Finally, when the integrity index of a path in the conduction path analysis set in step 1023 is lower than a preset threshold or there is an associated interruption point, a marking algorithm is used to mark the corresponding uncovered entity as a missing entity and the broken associated edge as a missing association; secondly, the topological coordinates of all missing entities and missing associations in the topological structure are aggregated, for example, the positions are recorded through a coordinate identification algorithm; finally, all topological coordinates are summarized to generate a missing marking set.

[0081] In a practical application, a power market analysis system, after receiving a user-submitted query for regional power balance data, first performs semantic parsing on three collected unstructured policy reports, extracting three structured semantic tags: "Coal-fired power plant environmental protection production limit ratio," "Interprovincial transmission channel capacity," and "New energy subsidy policy adjustment." Starting with the four target power plants in the query, the system traverses the connected transmission network topology, identifying 12 transmission paths containing attributes such as power plant output, line transmission limits, and node loads. The system automatically establishes a mapping relationship between semantic tags and target entity attributes. For example, it maps "Environmental protection production limit ratio" to the coal-fired power plant's available output limit attribute. Based on the "policy change impacts power supply reliability" logic in the user query, the system calculates the completeness index of each transmission path using the following formula: Transmission path completeness index = Σ(mapped attribute existence flag × weight coefficient) / total number of required attributes for the path. When it was discovered that a path containing inter-provincial transmission lines had a mapping completeness of only 75% due to missing grid structure adjustment data, which was lower than the preset 85% threshold, and that another path had a disconnect between environmental protection production restrictions and unit maintenance plans, the system immediately marked two coal-fired power plants as missing entities and three inter-regional interconnection lines as missing associations. It ultimately generated a missing marker set containing five topological coordinates, accurately locating data gaps that affect power supply reliability assessments.

[0082] In the overall solution of step 102 above, by converting unstructured external data into structured semantic labels, establishing a mapping relationship based on the target entity's topological structure, and combining it with influence logic to calculate the conduction path completeness index, data missing issues can be systematically detected and marked. After extracting attribute features through semantic annotation to generate semantic labels during the data conversion phase, this solution traverses the topological structure starting from the target entity to identify all conduction paths and their attribute associations. In the mapping analysis phase, the association continuity of each path is verified based on influence logic and the completeness index is calculated. Finally, topological coordinates that fall below the completeness threshold or have association interruptions are marked as missing. The entire process achieves a deep fusion analysis of unstructured data and structured graphs. Through quantitative assessment of conduction path integrity and location of interruption points, it accurately outputs missing entities and associated tag sets with topological coordinates.

[0083] 103. Based on the target entity information and the external data, establish an attribute mapping between external data attributes and entity attributes, generate a semantic vector that integrates cross-domain features by parsing the implicit relationship between the external data and the target entity information in the attribute mapping, project the semantic vector into a topological space according to the missing tag set, and generate topological nodes and topological connections that inherit the cross-domain features;

[0084] Optionally, step 103 may specifically include the following steps:

[0085] 1031. Based on the target entity information and the external data, construct a bidirectional mapping dictionary between external data attributes and entity attributes, and record direct mapping rules, mathematical derivation formulas, and unit conversion coefficients between attributes;

[0086] 1032. Analyze the implicit multi-order inference relationship between attributes in the bidirectional mapping dictionary through a deep semantic association model, and identify cross-domain attribute combinations with common variation characteristics or causal constraints;

[0087] Among them, step 1032 can also specifically include the following processes: through the deep semantic association model, receiving a mapping entry set from the bidirectional mapping dictionary, the mapping entry set records the corresponding rules and relationship identifiers between external data attributes and target entity attributes, and constructing an inference network with the mapping entry set; performing a traversal operation on the inference network, identifying the consistency of time dimension fluctuation characteristics between nodes in the inference network, and recording the common change attribute group, detecting the conduction path that meets the causal constraint rules in the inference network and recording the response sequence; calculating the fluctuation consistency quantization value of the common change attribute group, and measuring the trigger delay time of the response sequence; when the fluctuation consistency quantization value exceeds the set threshold and the trigger delay time is in a preset interval, outputting a cross-domain attribute combination with a relationship identifier.

[0088] 1033. Generate a semantic vector for each cross-domain attribute combination, wherein the semantic vector integrates the encoding value of the associated attribute, the relationship weight coefficient, and the time series label;

[0089] 1034. Create a blank topological node in the power market data map according to the topological coordinates located by the missing tag set, and project the semantic vector to the corresponding blank topological node to form a topological node carrying a cross-domain feature;

[0090] 1035. Based on the relationship weight coefficient in the semantic vector, a weighted topological connection that inherits the cross-domain feature is established between the topological nodes carrying the cross-domain feature. The attributes and connection relationships of the topological nodes and topological connections must meet the topological constraint rules of the power market map.

[0091] In the above steps, attribute mapping refers to the set of association rules between external data attributes and target entity attributes. A bidirectional mapping dictionary is a structured dictionary that records the mathematical derivation formulas and unit conversion coefficients for direct mapping rules between attributes. Implicit multi-order derivation relationships refer to multi-level mathematical or logical conductive associations between attributes that are not explicitly defined. A cross-domain attribute combination refers to a group of attributes from different data domains that share common variation characteristics or causal constraints. A derivation network refers to a network of derivation relationships between attributes constructed from mapping entries. A co-varying attribute group refers to a set of attributes with consistent fluctuation characteristics over time. A response sequence refers to a sequence of attributes in a transmission path that satisfies causal constraints. A fluctuation consistency quantification value refers to a numerical measure of the temporal fluctuation similarity of a co-varying attribute group. A trigger delay refers to the response time interval of a causal constraint transmission path. A semantic vector refers to a feature vector that integrates attribute encoding values, relationship weight coefficients, and time series labels. A blank topological node refers to an unassigned node created in the power market data graph based on a missing label set. A topological connection that inherits cross-domain characteristics refers to a node connection link established based on the semantic vector relationship weight coefficients. Topological constraints refer to the logical restrictions on node and connection relationships predefined in the power market data graph.

[0092] In this embodiment, step 1031 first extracts attribute pairs based on target entity information and external data. A key-value pair matching algorithm is then used to construct a bidirectional mapping dictionary, recording the mathematical derivation formulas and unit conversion coefficients for direct mapping rules between attributes. Structured storage technology is then used to form a set of mapping entries. For example, a mapping relationship is established between the temperature attribute of meteorological data and the load attribute of the distribution network, and a temperature-to-load conversion formula is recorded. This set of mapping entries is then output as input for subsequent steps. Finally, the bidirectional mapping dictionary construction process is completed.

[0093] Secondly, the operation process is executed based on the mapping entry set through step 1032: first, the inference network is constructed through the graph structure conversion algorithm, and the source attribute and target attribute of each mapping entry are converted into network nodes with metadata. The metadata includes value type, range, and unit. For example, when assigning node identifier N15 to meteorological temperature, the binding metadata value type is defined as float32, and the range is defined as , the unit is defined as ℃, the binding metadata value type is defined as float32 when the power load is assigned to node N22, and the range is defined as , the unit is defined as MW; establish directed edges with mathematical relationship expressions between nodes, where the relationship expression is such as N15 pointing to N22 to store the conversion formula: In the formula, the letter load represents the load value, and the letter temp represents the temperature value. Then, the dynamic time warping algorithm is used to traverse the network to identify the consistency of the time dimension fluctuation characteristics between nodes. The specific processing process includes extracting the corresponding time series data of the node, such as the temperature series T is equal to the sequence value and the load sequence L is equal to the sequence value , 290.0; perform linear interpolation on the missing values, and use the formula at the missing position k Complete the formula letters Indicates the value before position k, letter Indicates the value after position k, letter Indicates the value after interpolation; the standardization process uses the formula , where the letter x represents the original data value, and the letter Represents the mean of the sequence, letters Represents the standard deviation of the series, letters Represents the standardized value; construct a dynamic time warping matrix to calculate the path distance, and the matrix elements are recursively deduced using the formula: , the letters in the formula The first temperature series data points, letters Indicates the jth data point of the load sequence, letter Represents the sequence from the start point to the position The minimum cumulative distance, the letter min represents the minimum value operation; output fluctuation - consistency quantization value In the formula, the letter d represents the cumulative distance calculated by dynamic time warping. When d is equal to 0.087, Sim is calculated to be 0.92. Finally, the attribute groups with fluctuation consistency are recorded, such as the combination of N15 and N22.

[0094] Next, based on cross-domain attribute combinations, such as records containing combined weather temperature and power load attributes, step 1033 is performed. First, word embedding technology is used to extract the associated attributes of each combination and generate attribute encoding values. The word embedding technology uses a pre-trained Word2Vec model to process attribute names and relationship description texts to generate fixed-dimensional numerical vectors. For example, a code value of 0.72 is generated for the weather temperature attribute, and a code value of 1.15 is generated for the power load attribute, forming an attribute encoding vector [0.72, 1.15]. Secondly, a weight calculation algorithm is used to generate a relationship weight coefficient based on the relationship identification field. The weight calculation algorithm is based on the relationship indicator data in the combination, such as the fluctuation-consistency Sim value and the delay time, and the cosine similarity score is calculated. The formula is weight coefficient. , where letter A represents the Sim value vector and letter B represents the delay time vector. Then add a time series tag to record the time characteristics of attribute changes, extract the minimum time interval from the timestamp field of the combined data and calculate the time difference, for example, calculate the time characteristics based on the difference in the time points of the series changes of meteorological temperature and power load Equal to 8 minutes. Finally, the attribute encoding value relationship weight coefficient and time series tag are fused into a semantic vector through feature fusion technology, where the feature fusion adopts the weighted average formula , where the letter V represents the output semantic vector, the letter w represents the weight coefficient, the letter E represents the attribute encoding vector, and the letter Represents the time feature value. For example, the encoding vector [0.72, 1.15] formed by integrating the meteorological temperature encoding value 0.72 and the power load encoding value 1.15 is combined with the weight 0.92 and the time series feature 8 minutes. After weighted averaging, a 32-dimensional semantic vector is output (for example, the specific value can be expressed as , ... to represent cross-domain semantic features), forming a semantic representation that can be directly used in subsequent knowledge graph construction. Conversely, if the attribute encoding value is missing or the weight coefficient is below the 0.8 threshold, the fusion is skipped and the data is marked as incomplete.

[0095] Then, in step 1034, based on the missing tag set and the semantic vector output from step 1033, a graph database indexing algorithm is first used to locate the topological coordinates in the missing tag set. Next, a blank topological node is created at the corresponding location in the power market data graph using the node creation interface. A vector projection algorithm is then used to map the semantic vector to the blank topological node. Finally, a topological node carrying cross-domain features is formed. For example, a blank node with load meteorological features is created at the location of a missing distribution network node, and the vector projection process is completed.

[0096] Finally, in step 1035, the semantic vector field of the topological node is accessed. The weight coefficient field at indexes 128 to 135 is located according to the predefined data structure, and the value in this field is directly read using the array indexing method. For example, when processing the meteorological temperature node, the value 0.92 stored at index 130 in its associated semantic vector is extracted as the weight coefficient. The connecting line generation algorithm is invoked to perform node unique identifier matching, identifying the corresponding relationship between source node ID = N15 and target node ID = N22. The spatial position of the node is calculated based on the topological coordinate system, for example, the source node coordinates are (120, 80) and the target node coordinates are (310, 150). The Bresenham line algorithm is used to generate a pixel set between the two points, creating a 1-pixel black solid connecting line segment. The algorithm actually generates a directional connecting line entity containing 210 pixels. The weight coefficient value 0.92 obtained from the analysis is written into the weight attribute field of the connecting line entity. At the same time, the visualization parameters are automatically configured according to preset rules: the weight value 0.92 falls between 0.9 and 1.0, and the connecting line is formatted as a 5-pixel wide red solid line. The topology constraint engine loads the "Environmental Domain → Power Domain Unidirectional Constraint" rule (ID=R007) from the rule base. The source node metadata is verified. The domain type field value for the meteorological temperature node is "Environmental Domain"; the domain type field value for the power grid load node is "Power Domain." Connection direction verification confirms that the arrow points from the environmental domain to the power grid domain. All three verification results meet the rule requirements. A unique connection ID=CONN_001 is assigned, and the correspondence between the source node {ID:N15, attribute: "Meteorological Temperature", domain type: "Environmental Domain"} and the target node {ID:N22, attribute: "Power Grid Load", domain type: "Power Domain"} is recorded. A binding weight of 0.92 is assigned, the verification status is marked as PASS, and the rule ID=R007 is associated. Finally, a 5-pixel-wide red connecting line is rendered in the topological space, generating a topological connection entity with complete metadata.

[0097] In a practical application, a smart grid modeling system, based on structured operational information from four target power plants and three collected external policy reports, first constructs a bidirectional mapping dictionary containing five entries. This dictionary details the direct correspondence rules, mathematical conversion formulas, and unit conversion coefficients between external data attributes and entity attributes. A pretrained deep semantic association model then parses the dictionary data. This model automatically constructs a 28-node inference network, traversing the network to identify temporal fluctuation consistency characteristics between attributes. It also detects transmission paths that satisfy causal constraints, calculates fluctuation consistency quantification values, and measures response delays. When the quantification value exceeds a set threshold and the delay time falls within a preset interval, it automatically outputs four cross-domain attribute combinations with causal identification. For each attribute combination, a semantic vector is generated that incorporates multi-domain features. The encoding value includes the attribute numerical code and the relationship strength coefficient, along with additional temporal tagging information. Based on the five topological coordinates located by the previously generated missing tag set, the system creates five blank topological nodes in the power market data graph and accurately projects the semantic vectors to the corresponding node locations, forming entity node units that carry policy analysis features. Finally, based on the relationship strength coefficients embedded in the semantic vectors, weighted topological connections are automatically constructed between newly created nodes. This process strictly adheres to the topological rules of the power market graph, such as node degree constraints and thermal stability thresholds, to ensure the compliance of all connections. This entire process achieves a deep integration of unstructured policy data and structured grid entity characteristics, providing cross-domain feature support for market risk prediction.

[0098] In the overall solution of step 103 above, a topological traversal is performed by constructing an inference network to identify common change attribute groups with time-dimensional fluctuation consistency between nodes, and detect the conduction path response sequence that satisfies the causal constraint rules; at the same time, the fluctuation consistency quantization value is calculated and the trigger delay time is measured. When the quantization value exceeds the set threshold and the delay is within the preset interval, a cross-domain attribute combination with a relationship identifier is generated. Based on each cross-domain attribute combination, a semantic vector is dynamically generated that integrates the associated attribute encoding value, the relationship weight coefficient, and the time series tag. For the topological coordinates of the missing tag set, a blank node is created in the power market data map, and the semantic vector is projected to the corresponding node to form a topological node entity carrying cross-domain characteristics. Finally, based on the relationship weight coefficient in the semantic vector, a weighted topological connection is established between the nodes to ensure that all node attributes and connection relationships strictly follow the topological constraint rules of the power market map, thereby completing the topological reconstruction of the cross-domain characteristics in the power market data map.

[0099] 104. Embed the topological nodes and topological connections into the missing marked positions of the power market data map, reconstruct the topological structure, and verify the logical connectivity of the completed topological structure based on the relationship transmission rule;

[0100] Optionally, step 104 may specifically include the following steps:

[0101] 1041. Locate entity missing locations and associated missing locations in the power market data map according to the missing tag set, embed the topological nodes in the entity missing locations and write attribute code values, embed the topological connections in the associated missing locations and configure weight coefficients, and reconstruct the topological structure.

[0102] 1042. Traverse the reconstructed topological structure based on the relational conduction rules and verify the conduction direction, weight attenuation gradient, and time sequence continuity of the newly added conduction path;

[0103] 1043. If the conduction direction meets the preset constraints, the weight attenuation gradient is within the allowable range, and the timing mark is continuous, the conduction path is determined to be logically connected, and the conduction paths and failure rules that fail the logical connectivity verification are recorded.

[0104] In the above steps, relational transmission rules refer to the predefined directional constraints, weight decay gradients, and temporal continuity conditions in the power market. Logical connectivity refers to the effective transmission state of a newly added transmission path after it satisfies the relational transmission rules. Attribute encoding values ​​refer to the feature encoding values ​​embedded in topological nodes. Weight coefficients refer to the relationship weight values ​​of topological connection configurations. Reconstructed topology refers to the updated network structure after embedding nodes and connections. New transmission paths refer to newly generated inter-entity transmission links in the reconstructed topology. Transmission direction refers to the direction of attribute changes in a transmission path. Weight decay gradient refers to the rate of change of weight coefficients on a transmission path. Temporal continuity refers to the smooth continuity of time stamps in a transmission path. Preset constraints refer to the directional matching requirements set in relational transmission rules. Allowable range refers to the tolerance range for changes in the weight decay gradient. Temporal stamp continuity refers to a transmission state in which time intervals change smoothly. Transmission paths that fail logical connectivity verification refer to transmission links that fail verification. Failure rules refer to the specific constraint rule identifiers that cause verification failures.

[0105] In an embodiment of the present application, first, through step 1041, the entity missing position and the associated missing position are located according to the missing mark set, and the Cypher query language of the graph database is used to retrieve the topological coordinates to be completed in the power market data map; secondly, the graph node creation API is called to embed the topological node into the entity missing position and write the attribute code value, for example, a meteorological feature node is created at the missing coordinate of the distribution network node and the code value [0.72, 1.15] is written; then, the topological connection is embedded into the associated missing position and the weight coefficient is configured through the edge creation interface, for example, a connection edge with a weight of 0.92 is established between the meteorological node and the load node; finally, the topological structure is reconstructed to generate a network structure containing the newly added nodes and connections, and the updated power market data map is output.

[0106] Next, in step 1042, the reconstructed topology is traversed based on the relational conduction rules. First, a breadth-first traversal algorithm is used to scan all newly added conduction paths. Secondly, the dot product of the directional vectors is used to verify whether the conduction direction meets the preset constraints, for example, verifying the "meteorological domain → power domain" unidirectional rule. Next, the weight decay gradient difference value δ = |(w1-w2) / d| (w is the weight, d is the path length) is calculated to verify whether it is within the preset allowable range of 0.05-0.2. Finally, a time series dynamic regularization algorithm is used to check time series continuity to ensure that the time stamp interval meets the Δt ≤ 1 hour continuity threshold, for example, checking the continuity of the 15-minute sampling interval of meteorological data and load data.

[0107] Finally, in step 1043, if the conduction direction meets the preset constraints, the weight attenuation gradient is within the allowable range, and the timing mark is continuous, the conduction path is marked as logically connected; if the direction vector dot product value is less than 0.8, the direction constraint failure is recorded; if δ>0.2, the gradient over-limit failure is recorded; if the timing fault is greater than 1 hour, the timing interruption failure is recorded; and finally, a verification report containing the failed path ID and the corresponding failure rule is output, for example, the weight attenuation failure of path P35 is recorded because δ=0.28 exceeds the threshold.

[0108] In practical applications, within a regional power grid digital twin platform, the system performs the following operations: Based on a missing marker set containing the missing coordinates of two coal-fired power plants and three transmission lines, it accurately locates five missing locations in the power market data map. Topological node units with cross-domain characteristics are embedded into the missing physical locations of the two coal-fired power plants, and the output code value of 0.75, mapped to the environmental protection production restriction policy, is written. Simultaneously, topological connection links with weighted coefficients are embedded into the associated missing locations of the three transmission lines, with weights of 0.62, 0.79, and 0.85, respectively, completing the topological reconstruction. A full topology traversal is then initiated based on the power supply and demand transmission rules: The newly added interprovincial transmission paths are verified to ensure that their transmission direction complies with the "power plant → load center" constraint, that the weight decay gradient along the transmission distance is within the permissible threshold of ±0.1 per 100 kilometers, and that the timestamp continuity between the renewable energy output data and the actual load curve is verified. If a reconstructed tie line fails directional verification due to unassociated regional load data, and another line's output fluctuation mark has a four-hour time gap, the system automatically records the two failed paths and the corresponding "flow unidirectionality rule" and "timing synchronization rule" violations, ensuring the reconstructed map has strict interpretability and business logic consistency.

[0109] In the overall solution of step 104 above, the integrity of the topological structure is reconstructed by embedding topological nodes carrying cross-domain features into the missing entity locations and writing attribute code values, while also embedding weighted topological connections into the missing associated locations and configuring weight coefficients. The reconstructed topological structure is then traversed based on the relationship conduction rules, and a three-dimensional logical check is performed on the newly added conduction paths: verifying whether the conduction direction complies with the preset constraint rules, checking whether the weight decay gradient is within the allowable range, and verifying whether there are continuity breaks in the timing marks. Only when the conduction direction is compliant, the weight decay gradient is controllable, and the timing marks are continuous is the conduction path considered logically connected. For conduction paths that fail the check, the system will accurately record their path coordinates and the specific failure rules that are violated, and generate a logical connectivity verification report. Ultimately, the closed-loop completion and verification of the data map is completed, ensuring that the newly added topological elements meet the constraints of the power market map ontology.

[0110] 105. Based on the verified topological structure, the dynamic transmission chain formed by the completion path between the target entity information is parsed, and a feedback interface including semantic relationships and completion structures is generated to respond to data query instructions.

[0111] Optionally, step 105 may specifically include the following steps:

[0112] 1051. Based on the verified topological structure, extract the dynamic conduction chain formed by the completed paths between the target entity information, and analyze the relationship weight value and time sequence marking trajectory of each conduction path in the dynamic conduction chain;

[0113] 1052. Map the relationship weight value to a preset semantic strength level, convert the time sequence mark trajectory into a time axis, generate a node relationship chain with a conduction direction identifier, and mark the semantic strength level and the time axis on the node relationship chain;

[0114] 1053. Convert the conduction path that has not passed the verification into a warning mark, integrate the node relationship chain and the warning mark, and generate a visual feedback interface to respond to the data query instruction.

[0115] In the above steps, the verified topology structure refers to the topology structure output from step 104 after logical connectivity verification. The target entity information refers to the entity data specified by the user data query instruction. The completed path refers to the newly added conductive link formed through topology reconstruction and verification in steps 103 and 104. The dynamic conductive chain refers to the conductive sequence of attribute changes formed on the completed path. The semantic relationship refers to the semantic association between nodes in the conductive chain. The completed structure refers to the updated graph structure containing the newly added elements. The feedback interface refers to the visual display interface that responds to user queries. The relationship weight value refers to the relationship weight value configured in the topological connection. The time series mark trajectory refers to the sequence of time marks in the semantic vector. The preset semantic strength level refers to the predefined high, medium, and low level mapping table. The timeline refers to the visual timeline of the time series mark conversion. The conductive direction indicator refers to the directional arrow mark in the node relationship chain. The node relationship chain refers to the visual representation element of the dynamic conductive chain. The warning indicator refers to the warning icon for the conductive path that failed verification. The visual feedback interface refers to the visual display interface that integrates all elements. The conductive path that failed verification refers to the conductive path that failed logical connectivity verification in step 104.

[0116] In the embodiment of the present application, first, the following operation process is performed based on the verified topological structure through step 1051: First, a depth-first traversal algorithm is used to extract the dynamic conduction chain formed by the completed path between the target entity information starting from the selected node. For example, the generator node identifier G1 is selected, and the traversal and access to the adjacent nodes are started, including the transformer node T1 and the load node L1 to form a complete path. The generator G1 points to the transformer T1, and the transformer T1 points to the load L1; when the intermediate nodes in the detection path are all completed links, a conduction chain identifier is generated, for example, the distribution chain ID is DC001, corresponding to the complete path list Secondly, the parsing algorithm is used to scan each conduction path in the dynamic conduction chain one by one, and the relationship weight attribute in the path text field is parsed through the regular expression matching algorithm to extract the numerical value, such as The text description field matches the weight value 0.85, the conduction path The matching weight value is 0.93. Then, through the time series, the extraction technology analyzes the time mark trajectory in the timestamp vector corresponding to the conduction chain. Based on the time point sequence value, for example, the specific timestamp sequence value is 10:00, 10:03, and 10:08, the formula is used to calculate the interval difference between adjacent time points. , where the letters Indicates that the time point number starts counting from 1, and the letters Indicates the Time point value, letter Indicates the value of the previous time point, letter Indicates the Time interval differences, thereby generating a time stamp sequence Unit is hour. Finally, the structured dynamic conduction chain dataset is output.

[0117] Secondly, through step 1052, the relationship weight value is input into the mapping algorithm, and the preset semantic strength level table is matched to convert it into high, medium and low semantic strength levels; at the same time, the timing mark trajectory is converted into a time axis visualization component using timing processing technology; then, a node relationship chain with a conduction direction identifier is constructed through a visualization generation tool, and the direction is indicated by an arrow; finally, the semantic strength level and time axis are marked on the node relationship chain to generate a complete conduction path display element.

[0118] Finally, in step 1053, the conduction paths that have not passed the verification are input into the conversion logic module and converted into warning signs, such as red warning icons; the node relationship chain and the warning signs are integrated through the interface integration module; then the visualization engine is called to generate a feedback interface containing semantic relationships and completion structures; and finally, the feedback interface is output in response to data query instructions, for example, when the user queries the correlation between power load and weather, the complete conduction chain is displayed and the warning mark of the interrupted path is highlighted.

[0119] In practical application, within a regional energy decision-making platform, the system, based on a validated reconstruction graph, identified three newly added interprovincial dynamic transmission chains from four target power plant entities. First, core path features were extracted: a coal-fired power transmission channel with a weight of 0.82 and a continuous 72-hour output curve was identified. Simultaneously, a new energy consumption chain with a weight of 0.67 and a fluctuation trajectory containing five time markers was captured. Semantic transformation was then performed: the weights were mapped to three levels of intensity (weak / medium / strong), and the time series trajectory was converted into a three-segment timeline labeled "early peak," "midday valley," and "late peak." When generating transmission chains with arrowed flow directions in the visualization interface, coal-fired power channels were labeled with a deep red "strong correlation" indicator and overlaid with a timeline. Consumption chains with four-hour data gaps were transformed into flashing lightning warning icons. Finally, the three dynamic transmission chains and two failure warning channels were integrated to generate an interactive feedback interface with a colored flow topology map, a timeline scale, and a floating warning box, dynamically responding to user-submitted power balance assessment commands.

[0120] In the overall solution of step 105 above, a closed query-response loop is achieved by parsing the dynamic transmission chain formed by the completed paths between target entities. Specifically, the dynamic transmission chain generated in the verified topology is first extracted, and the relationship weight values ​​and time-series marking trajectory of each transmission path are accurately parsed. The relationship weight values ​​are then mapped to preset semantic strength levels, and the time-series marking trajectory is converted into a visual timeline. A node relationship chain with transmission direction identifiers is constructed, and the semantic strength level and timeline coordinates are annotated on the chain. Transmission paths that fail logical verification are converted into eye-catching warning signs (such as flashing signs or highlighted borders). Finally, the complete node relationship chain and warning signs are integrated into a unified visualization panel, generating an interactive feedback interface containing three levels of information.

[0121] Regarding steps 101 to 105 above, the following is a specific embodiment:

[0122] For example, when a user queries the monthly transaction volume data of the electricity trading market in 2024 (such as Figure 2 As shown in the bar chart), the original graph is missing data for June. This application dynamically executes the following process:

[0123] First, in response to a user query for "trading volume by month in 2024," the system collects structured operational data from the electricity market data graph (e.g., 8 million kWh in May and 9 million kWh in July). It also acquires external unstructured information (e.g., news reports stating that "persistent high temperatures in June 2024 led to a 20% year-on-year increase in electricity load"). Semantic analysis transforms this external data into structured labels {time: 2024-06, event: high temperature, impact: 20% increase in electricity load}. Traversing the monthly transmission chain (January → February → ... → December), the system discovers that the June node is missing and adjacent paths are interrupted. Combined with the logical association in electricity market rules that "increased electricity load typically leads to increased trading volume," the system calculates a topological completeness of 91.6% (below the 95% threshold). Finally, missing entities (June's trading volume) and missing relationships (transmission relationships between May and June, and between June and July) are marked.

[0124] Secondly, attribute association rules were established based on the cross-domain fusion mechanism: historical data shows that for every 10% increase in electricity load, transaction volume increases by an average of 8% (i.e., an impact coefficient of 0.8). Deep analysis identified a strong statistical correlation between high temperature events and transaction volume (a historical correlation coefficient of 0.85), generating a semantic vector of fused features: {high temperature intensity: 0.9, load increase: 0.2, impact coefficient: 0.8, time: 2024-06}. A June node was created at the missing location in the topology. Based on the May data (8 million kWh) and the load increase, the transaction volume complement was calculated: 800 × (1 + 0.2 × 0.8) = 9.28 million kWh. Transmission relationships were also established from May to June (weight 0.8) and from June to July (weight 0.7).

[0125] Next, the newly generated June node and its transmission relationship were embedded into the original graph, forming a complete time chain: January → … → May → June → July → … → December. The rationality of the newly added path was verified: the transmission direction conformed to the chronological order (May → June → July), the weight change was within the allowable range (0.8 → 0.7, attenuation rate 12.5% ​​< 15% threshold), and the month labels were continuous without gaps, confirming the logical completeness of the topology.

[0126] Finally, generate Figure 2 The visual feedback interface shown:

[0127] Bar chart: Displays data for the full 12 months, with the June bar (9.28 million kWh) marked with the completed value in a special color.

[0128] Conduction chain identification: Mark the path from May to June to July with “dynamic completion based on high temperature events” and an impact coefficient of 0.8 (not shown in the figure).

[0129] Data interpretation: A floating prompt states, "June data is calculated based on a 20% increase in electricity load and historical association rules, and the transmission intensity with adjacent months is >0.7" (not shown in the figure).

[0130] This application converts unstructured external data (high temperature events) into structured transaction volume data through cross-domain attribute fusion and dynamic topology completion, inherits cross-domain semantic relationships (such as the highlighted path in the attached figure) in the reconstructed transmission chain, and finally outputs a visualization result that is logically complete and includes an explanation of the completion process.

[0131] Optionally, the method also includes: for the target entity information, calling the ontology layer in the pre-constructed electricity market data map to identify the domain category to which the entity belongs and the domain relationship category between entities; in the data layer of the electricity market data map, extracting neighbor node features and multi-hop path features associated with the identified entities and relationships as entity context features; fusing the domain category to which the entity belongs, the domain relationship category between entities and the entity context features to generate a domain-enhanced semantic vector; injecting the domain-enhanced semantic vector into the semantic vector of the fused cross-domain features, and outputting the enhanced semantic vector.

[0132] In the above steps, the ontology layer refers to the structured description system of domain entity categories and relationship categories predefined in the power market data map. Domain category refers to the classification of the professional field of the power market to which the entity belongs. Domain relationship category refers to the identification of the relationship type between entities that complies with the rules of the power market. Domain entity type refers to the entity domain classification result identified by the ontology layer. Neighbor node features refer to the set of node attributes directly connected to the target entity. Multi-hop path features refer to the sequence of node attributes that the target entity can reach through multiple associated paths. Entity context features refer to the composite attribute set that integrates neighbor node features and multi-hop path features. Domain enhanced semantic vector refers to the enhanced feature vector that is infused with domain knowledge. The enhanced semantic vector refers to the optimized feature expression that integrates the domain enhanced semantic vector and the original semantic vector.

[0133] In an embodiment of the present application, the ontology layer of the power market data map is first called to identify the domain category to which the target entity belongs and the domain relationship category between entities, and the domain entity type and relationship type are output; secondly, based on the identified domain entity type and relationship type, the neighbor node features are extracted through the graph neural network in the data layer, and the multi-hop path traversal algorithm is simultaneously used to extract the path features of the associated nodes within three hops, and the two are fused to generate entity context features; then the domain entity type relationship type and entity context features are input into the feature fusion model to generate a domain enhanced semantic vector; finally, the domain enhanced semantic vector is injected into the semantic vector of the original fused cross-domain feature, and the enhanced semantic vector is output through the vector splicing layer, for example, the generator entity type and the transmission relationship type are injected into the meteorological-load semantic vector to form an optimized vector containing the domain weight coefficient.

[0134] In a practical application, a power grid intelligent analysis system invokes the ontology layer of a pre-built electricity market data graph to identify three of four target power plants as belonging to the "coal-fired power sector" and one as belonging to the "new energy sector." It then analyzes six domain relationships, including "fuel supply relationships" and "power transmission relationships." Feature extraction then occurs at the data layer: for a coal-fired power plant entity, load data for four neighboring nodes, including its coal supplier and transmission and transformation facilities, is extracted. The three-hop path features of "coal-fired power-transmission-industrial user" are extracted along the ontology path, capturing price fluctuation characteristics at seven nodes along the path. Domain entity type identifiers and entity context features are then integrated to generate a domain-enhanced semantic vector containing a domain code of 0.88 and a supply stability score. This vector is then injected into a cross-domain semantic vector generated by external policy features, increasing the output code value from 0.75 to 0.82. The reconstructed topological nodes display the coupled effects of fuel cost fluctuations and environmental policies in the feedback interface, making market risk assessment results more relevant to real-world business scenarios.

[0135] This application accurately identifies the domain category to which the target entity belongs and the domain relationship category between entities by calling the electricity market data graph ontology layer, and clearly outputs professional entity types such as power generators and electricity sellers, as well as relationship types such as transmission contracts and dispatch instructions. Based on this domain knowledge, the multi-hop neighbor node features and conduction path features of the target entity are deeply extracted at the data layer, and entity context features containing topological constraint relationships are dynamically output. Subsequently, the domain entity type identifier, relationship type tag and entity context features are integrated to generate a domain-enhanced semantic vector that carries the market rule logic. Finally, the domain-enhanced semantic vector is injected into the original cross-domain feature semantic vector, and an enhanced semantic vector that carries both external data relevance and professional rules of the electricity market is output. This process ensures that the semantic feature expression in the subsequent topological projection step has both the breadth of cross-domain data relevance and meets the professional depth requirements of the electricity market field, significantly improving the adaptation accuracy of the feature representation to the electricity trading scenario.

[0136] Optionally, the method also includes: in the topological space of the electricity market data map, based on the structural similarity and attribute similarity between entities, locating the cross-domain entities that have an association mapping with the target entity information, and outputting the cross-domain entities and their association relationships; using the cross-domain entities and the target entity as anchor points, dynamically constructing a reasoning sub-network containing multi-hop neighbors in the topological space, and performing logical rule reasoning along the multi-hop path in the dynamic reasoning sub-network, and outputting the reasoning results; fusing the target entity information, the cross-domain entities and their association relationships, and the reasoning results to generate a joint semantic representation, and completing the reasoning results to the target entity information, and outputting the completed target entity information.

[0137] In the above steps, structural similarity refers to the degree of matching between the connection structures of entities in the topological space of the power market data graph. Attribute similarity refers to the cosine similarity calculation result of the numerical or categorical features between entities. Cross-domain entities refer to external entities identified through association mapping that have cross-domain associations with the target entity information. Association relationships refer to the attribute transmission or logical dependency connection between cross-domain entities and target entities. Anchor points refer to the target entity and cross-domain entities that serve as the starting point for reasoning. The reasoning subnetwork refers to a local network consisting of multi-hop neighboring nodes extending from the anchor point. A multi-hop path refers to a transmission path in the reasoning subnetwork that spans two or more association edges. Logical rule reasoning refers to the calculation process of executing preset power market rules (such as the supply and demand balance formula) along a multi-hop path. The reasoning result refers to the attribute derivation value or state judgment output by the logical rule reasoning. The joint semantic representation refers to the feature vector that integrates the target entity information, cross-domain entity association relationships, and the reasoning result. The completed target entity information refers to the enhanced entity data that is injected with the reasoning result.

[0138] In an embodiment of the present application, first, in the topological space of the power market data map, the domain category of the target entity information is parsed through the ontology layer, and the structural similarity and attribute similarity between entities are calculated using the cosine similarity algorithm, the cross-domain entities with associated mappings are located and their association relationships are output; secondly, with the target entity and the cross-domain entity as anchor points, a random walk algorithm with an attenuation factor is used to dynamically construct an inference sub-network containing three-hop neighbors; then, in the inference sub-network, the rule engine is called along the multi-hop path to execute the preset logical rule reasoning, for example, the supply and demand balance formula is applied to the "power plant, transmission network, user" path to derive the load gap value, and the inference result is output; finally, the target entity information, the cross-domain entity association relationship and the inference result are input into the feature fusion model to generate a joint semantic representation and complete it to the target entity information, and the target entity information containing the inference result is output, such as the load forecast data with the meteorological influence coefficient completed.

[0139] In a practical application, within a power dispatch decision-making system, the system uses a topological analysis module within a power market data graph to perform the following operations: Based on the structural and attribute similarities of entities within the topological space, it accurately locates three cross-domain coal transportation companies with fuel supply ties to a target coal-fired power plant and outputs the "coal price transmission" relationship. Using these three coal companies and the four target power plants as anchor points, it automatically constructs a six-hop inference subnetwork encompassing nine entity types, including generators, coal transportation railway nodes, and port storage and transportation facilities. Within this network, a pre-set inventory warning rule is implemented: if a coal company's railway capacity saturation exceeds 90% and its port inventory is less than 200,000 tons, a "fuel supply risk" inference conclusion is triggered. Finally, the target power plant's output value of 0.82, coal company supply-related parameters, and three inference conclusions, including two warning conclusions and one safety conclusion, are integrated to generate a joint semantic representation containing the fuel risk level identifier. The risk level is dynamically added to the target power plant's operating indicators. For example, a new "Fuel Risk Level III" field is added to the attribute table of a power plant, which improves the accuracy of power generation plan forecasts by 0.12 and significantly optimizes the robustness of day-ahead market clearing decisions.

[0140] This application accurately locates cross-domain entities with potential association mappings based on multi-dimensional similarity matching calculations, and outputs a mapping set covering cross-domain entity identifiers and their association relationships. Taking the target entity and cross-domain entity as topological anchors, a reasoning sub-network containing three-hop neighbor nodes is dynamically extracted and constructed. This sub-network completely retains the original conduction path and attribute constraint relationship between entities. In the constructed dynamic reasoning sub-network, the power market transaction rule library and operation constraint library are loaded along the multi-hop conduction path to execute the logical reasoning engine, and the reasoning results that comply with the power field specifications are generated layer by layer. This process strictly ensures the logical compliance and time consistency of the derived relationship. Finally, the target entity native data, cross-domain entity association relationship and reasoning results are integrated to generate a joint semantic representation and automatically complete it to the target entity information, forming an enhanced entity information set that contains both original features and cross-domain derived features, thereby significantly improving the entity information completeness and decision support value in complex association scenarios in the power market.

[0141] Optionally, the method also includes: in the process of establishing attribute mapping, calling a pre-built power field terminology library, identifying professional terms in the external data and target entity information, performing semantic disambiguation, and outputting standardized entity attributes; using the standardized entity attributes, when traversing the target entity information and semantic labels, detecting whether it hits the predefined atomic indicators, derived indicators or composite indicators, and if so, outputting the reconstructed node connection data completion path and calculation logic description according to the predefined indicator calculation rules; based on the node connection data completion path and calculation logic description, dynamically constructing a subnetwork containing relevant entities, attributes and calculation rules in the topological space, executing the indicator definition rules in the subnetwork, and outputting the calculated predefined indicator value; injecting the predefined indicator value as a cross-domain feature into the semantic vector of the corresponding entity or relationship, and based on the semantic vector containing the indicator value, calling the regularized diagnostic template associated with the indicator to perform abnormality analysis and interpretation.

[0142] In the above steps, the attribute mapping process refers to the operational flow when external data is mapped to the attributes of the target entity. The pre-built power domain terminology library refers to a pre-established database containing power domain vocabulary. Terminology refers to domain-specific vocabulary in the external data and target entity information. Semantic disambiguation refers to the process of eliminating ambiguous terms to determine a unique meaning. Standardized entity attributes refer to attribute data in a unified format after disambiguation. Atomic indicators refer to the most basic independent calculated indicators. Derived indicators refer to secondary indicators derived from atomic indicators. Composite indicators refer to complex indicators composed of multiple indicators. Predefined indicator calculation rules refer to the preset formulas or logic for calculating indicators. The reconstructed node connection data completion path refers to the description of the supplementary path generated based on indicator calculation. The calculation logic description refers to the steps for indicator calculation. A subnetwork refers to a local network dynamically constructed in the topological space, containing relevant entity attributes and calculation rules. Indicator definition rules refer to the set of instructions for executing indicator calculations. Predefined indicator values ​​refer to the calculated numerical results of indicators. Cross-domain feature injection refers to the process of incorporating indicator values ​​into semantic vectors. The regularized diagnostic template refers to the preset diagnostic rule framework. Anomaly analysis and interpretation refers to the operations of performing diagnostics and generating analytical results.

[0143] In an embodiment of the present application, first, in the process of establishing attribute mapping, the terminology recognition interface of the pre-built power field terminology library is called, and the context-based natural language processing algorithm is used to identify professional terms in external data and target entity information, and semantic disambiguation is performed to output standardized entity attributes; secondly, the standardized entity attributes are used to adopt a depth-first traversal algorithm when traversing the target entity information and semantic labels to detect whether the predefined atomic indicator derivative indicator or composite indicator is hit. If it is hit, the rule engine is called to output the reconstructed node connection data completion path and calculation logic description according to the predefined indicator calculation rules; then, based on the reconstructed calculation logic description, a sub-network containing relevant entity attributes and calculation rules is dynamically constructed in the topological space, and the indicator definition rules are used in the sub-network to execute the rule engine and calculate and output the predefined indicator value; then the predefined indicator value is injected into the semantic vector of the corresponding entity or relationship as a cross-domain feature; finally, based on the semantic vector containing the indicator value, the template matching function of the regularized diagnostic template associated with the indicator is called to perform abnormal analysis and interpret the output diagnosis results.

[0144] In practical applications, within a smart grid analysis platform, the system utilizes a pre-built power sector terminology library to standardize terminology for external data related to coal supply policies and target power plant entity information. For example, "maximum unit output" is disambiguated as "rated capacity," and "coal calorific value fluctuation" is standardized as "fuel calorific value deviation." The output contains six standardized entity attributes. While traversing the target power plant's topology, it detects that the "environmental protection production restriction compliance rate" matches a predefined atomic indicator library, and that the "inter-regional transmission margin" matches a derived indicator rule. This automatically triggers the calculation logic: Based on the atomic indicator calculation rules, the node connection path of a coal-fired power plant is reconstructed, completing the calculation logic of "production restriction compliance rate = output / approved capacity × 100%." ​​Based on the derived indicator formula, a subnetwork containing transmission lines and load nodes is constructed in the topological space, dynamically calculating "margin = (line thermal stability limit - actual power flow) / thermal stability limit." The system automatically executes the rule calculation, outputting a margin value of 62.3% for a specific inter-provincial channel and the production restriction compliance rates of three power plants. After injecting these indicator values ​​into the semantic vector as cross-domain features, the preset diagnostic template is called: when it is detected that the execution rate of a power plant is 77%, which is lower than the threshold of 80% and the associated channel margin is less than 70%, the "section blocking risk" alarm logic is activated, and a highlighted blockage cause explanation is generated on the feedback interface: "Unit #1 has failed to implement environmental protection production restrictions, resulting in a decrease in regional power transmission capacity," realizing automatic attribution of business anomalies.

[0145] This application calls a pre-built terminology library in the power field to identify professional terms in external data and target entity information, performs standardized disambiguation processing, and outputs normalized entity attributes. Standardized entity attributes are used to traverse the target entity topology and semantic labels to detect whether the definition threshold of atomic indicators, derived indicators or composite indicators is hit. If hit, the node connection path is automatically reconstructed to complete the data breakpoints based on the predefined indicator calculation rules, and the calculation rule chain with mathematical logic description is simultaneously output. Then, based on the calculation rule chain, a dedicated sub-network containing relevant entity attributes and calculation factors is dynamically constructed in the topological space, and the indicator definition rules are enforced in the sub-network to output the calculation results. Finally, the calculated indicator value is injected into the semantic vector of the corresponding entity or relationship as a cross-domain feature, triggering the association rule diagnosis template to perform anomaly analysis. This process generates an interpretation report with confidence by mapping the results to the standard warning event library, realizing an automated indicator diagnosis closed loop in the power market data fusion scenario.

[0146] Figure 3 The present invention provides a structural diagram of a semantic enhancement and dynamic completion system for a power market data graph, such as Figure 3 As shown, the system includes:

[0147] The acquisition module 31 collects the target entity information of structured operation and unstructured external data in the pre-established power market data map in response to the data query instruction sent by the user;

[0148] A mapping module 32 converts the external data into structured semantic tags, traverses the topological structure of the target entity information, establishes a mapping relationship between the structured semantic tags and the target entity attributes, calculates a completeness index of the topological structure based on the mapping relationship in combination with the impact logic in the data query instruction, and marks missing entities or missing associations related to the data query instruction to generate a missing tag set;

[0149] A generation module 33 establishes an attribute mapping between external data attributes and entity attributes based on the target entity information and the external data, generates a semantic vector that integrates cross-domain features by parsing the implicit relationship between the external data and the target entity information in the attribute mapping, projects the semantic vector into a topological space based on the missing tag set, and generates topological nodes and topological connections that inherit the cross-domain features;

[0150] A reconstruction module 34 is configured to embed the topological nodes and topological connections into the missing marked positions of the power market data map, reconstruct the topological structure, and verify the logical connectivity of the completed topological structure based on the relationship conduction rules;

[0151] The verification module 35 parses the dynamic transmission chain formed by the completion path between the target entity information based on the verified topological structure, and generates a feedback interface including semantic relationships and completion structures to respond to the data query instruction.

[0152] Figure 3 The semantic enhancement and dynamic completion system of the power market data map can perform Figure 1 The implementation principles and technical effects of the semantic enhancement and dynamic completion method for the power market data graph described in the illustrated embodiment will not be elaborated on here. The specific manner in which each module and unit performs operations in the semantic enhancement and dynamic completion system for the power market data graph in the above embodiment has been described in detail in the relevant embodiments of the method and will not be elaborated on here.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for semantic enhancement and dynamic completion of power market data graph, characterized in that: include: In response to the data query instruction sent by the user, the target entity information of structured operation and unstructured external data in the pre-established power market data map are collected; Convert the external data into structured semantic tags, traverse the topological structure of the target entity information, establish a mapping relationship between the structured semantic tags and the target entity attributes, combine the impact logic in the data query instruction, calculate the completeness index of the topological structure based on the mapping relationship, and mark the missing entities or missing associations related to the data query instruction to generate a missing tag set; Based on the target entity information and the external data, an attribute mapping is established between the external data attributes and the entity attributes, and a semantic vector integrating cross-domain features is generated by parsing the implicit relationship between the external data and the target entity information in the attribute mapping. The semantic vector is projected into a topological space according to the missing tag set to generate topological nodes and topological connections that inherit the cross-domain features. Embedding the topological nodes and topological connections into the missing marked positions of the power market data map, reconstructing the topological structure, and verifying the logical connectivity of the completed topological structure based on the relationship conduction rules; Based on the verified topological structure, the dynamic transmission chain formed by the completion path between the target entity information is parsed, and a feedback interface containing semantic relationships and completion structures is generated to respond to the data query instruction; Also includes: For the target entity information, the ontology layer in the pre-built electricity market data map is called to identify the domain category to which the entity belongs and the domain relationship category between the entities; In the data layer of the power market data map, neighbor node features and multi-hop path features associated with the identified entities and relationships are extracted as entity context features; Fusing the domain categories to which entities belong, the domain relationship categories between entities, and the entity context features to generate a domain-enhanced semantic vector; The domain-enhanced semantic vector is injected into the semantic vector fused with cross-domain features, and the enhanced semantic vector is output.

2. The method according to claim 1, characterized in that Also includes: In the topological space of the power market data map, based on the structural similarity and attribute similarity between entities, the cross-domain entities having an association mapping with the target entity information are located, and the cross-domain entities and their association relationships are output; Taking the cross-domain entity and the target entity as anchor points, dynamically constructing a reasoning sub-network including multi-hop neighbors in the topological space, performing logical rule reasoning along the multi-hop path in the dynamic reasoning sub-network, and outputting a reasoning result; fusing the target entity information, the cross-domain entities and their associations, and the reasoning results to generate a joint semantic representation; Decomposing the joint semantic representation into sub-vector sets according to entity levels, hierarchically projecting the sub-vector sets into a topological space based on voltage levels and service types, and generating completion nodes and associated edges corresponding to the missing tag sets; The business logic compliance of the completed nodes and associated edges is verified through a rule engine, and cross-level backtracking correction is performed on the nodes that fail the verification until the logical connectivity requirements of the topology structure are met.

3. The method according to claim 1, characterized in that Also includes: In the process of establishing attribute mapping, a pre-built electric power domain terminology library is called to identify professional terms in the external data and target entity information, perform semantic disambiguation, and output standardized entity attributes; By using the standardized entity attributes, when traversing the target entity information and semantic labels, it is detected whether it hits the predefined atomic indicators, derived indicators or composite indicators. If so, the reconstructed node connection data completion path and calculation logic description are output according to the predefined indicator calculation rules; Based on the node connection data completion path and calculation logic description, a sub-network containing relevant entities, attributes and calculation rules is dynamically constructed in the topological space, the indicator definition rules are executed in the sub-network, and the calculated predefined indicator value is output; The predefined indicator value is injected into the semantic vector of the corresponding entity or relationship as a cross-domain feature, and based on the semantic vector containing the indicator value, the regularized diagnostic template associated with the indicator is called to perform abnormality analysis and interpretation.

4. The method according to claim 1, wherein Based on the target entity information and the external data, an attribute mapping between external data attributes and entity attributes is established; by parsing the implicit relationship between the external data and the target entity information in the attribute mapping, a semantic vector integrating cross-domain features is generated; the semantic vector is projected into a topological space according to the missing tag set, and topological nodes and topological connections that inherit the cross-domain features are generated, including: Based on the target entity information and the external data, a bidirectional mapping dictionary between external data attributes and entity attributes is constructed, and direct mapping rules, mathematical derivation formulas, and unit conversion coefficients between attributes are recorded; By using a deep semantic association model, the implicit multi-order inference relationship between attributes in the bidirectional mapping dictionary is analyzed to identify cross-domain attribute combinations with common variation characteristics or causal constraints; Generate a semantic vector for each cross-domain attribute combination, wherein the semantic vector integrates the encoding value of the associated attribute, the relationship weight coefficient and the time series mark; According to the topological coordinates located by the missing tag set, a blank topological node is created in the power market data map, and the semantic vector is projected to the corresponding blank topological node to form a topological node carrying cross-domain features; According to the relationship weight coefficient in the semantic vector, a weighted topological connection that inherits the cross-domain feature is established between the topological nodes carrying the cross-domain feature. The attributes and connection relationships of the topological nodes and topological connections must meet the topological constraint rules of the power market map.

5. The method according to claim 4, characterized in that Through the deep semantic association model, the implicit multi-order inference relationship between attributes in the bidirectional mapping dictionary is analyzed to identify cross-domain attribute combinations with common variation characteristics or causal constraints, including: Receiving a mapping entry set from the bidirectional mapping dictionary through a deep semantic association model, the mapping entry set records the correspondence rules and relationship identifiers between external data attributes and target entity attributes, and constructing an inference network with the mapping entry set; Performing a traversal operation on the inference network, identifying consistency of time dimension fluctuation characteristics between nodes in the inference network, and recording a common change attribute group, detecting a conduction path that satisfies a causal constraint rule in the inference network and recording a response sequence; Calculating a fluctuation consistency quantization value of the common variation attribute group and measuring a trigger delay time of the response sequence; When the fluctuation consistency quantization value exceeds a set threshold and the trigger delay time is within a preset interval, a cross-domain attribute combination with a relationship identifier is output.

6. The method according to claim 1, characterized in that Based on the verified topological structure, the dynamic transmission chain formed by the completion path between the target entity information is parsed, and a feedback interface containing semantic relationships and completion structures is generated to respond to data query instructions, including: Based on the verified topological structure, a dynamic conduction chain formed by the completion path between target entity information is extracted, and the relationship weight value and time sequence marking trajectory of each conduction path in the dynamic conduction chain are analyzed; Mapping the relationship weight value to a preset semantic strength level, converting the time sequence mark trajectory into a time axis, generating a node relationship chain with a conduction direction identifier, and marking the semantic strength level and the time axis on the node relationship chain; The conduction paths that fail to pass the verification are converted into warning signs, the node relationship chain and the warning signs are integrated, and a visual feedback interface is generated to respond to data query instructions.

7. The method according to claim 1, characterized in that The external data is converted into structured semantic labels, the topological structure of the target entity information is traversed, a mapping relationship between the structured semantic labels and the target entity attributes is established, and the completeness index of the topological structure is calculated based on the mapping relationship in combination with the impact logic in the data query instruction, and missing entities or missing associations related to the data query instruction are marked to generate a missing mark set, including: Perform semantic annotation processing on the unstructured external data, extract attribute categories and numerical descriptions, generate structured semantic labels, traverse its associated topological structure starting from the target entity, and identify the conduction paths between all entities and the attribute associations on the paths; Establishing a mapping relationship between the structured semantic tags and the target entity attributes, calculating the completeness index of each conduction path in the mapping relationship in combination with the influence logic in the data query instruction, and simultaneously detecting the associated interruption points in the conduction path that violate the influence logic; When the completeness index is lower than a preset threshold, or there is an association interruption point, the corresponding entity is marked as a missing entity or the association is marked as a missing association, and the topological coordinates of all the missing entities and missing associations are summarized to generate a missing mark set.

8. The method according to claim 1, characterized in that Embedding the topological nodes and topological connections into the missing marked positions of the power market data map, reconstructing the topological structure, and verifying the logical connectivity of the completed topological structure based on the relationship conduction rule, including: Locating entity missing positions and associated missing positions in the power market data map according to the missing mark set, embedding the topological nodes in the entity missing positions and writing attribute coding values, embedding the topological connections in the associated missing positions and configuring weight coefficients, and reconstructing the topological structure; Traverse the reconstructed topological structure based on the relational conduction rules and verify the conduction direction, weight attenuation gradient and time sequence continuity of the newly added conduction path; If the conduction direction meets the preset constraints, the weight attenuation gradient is within the allowable range, and the timing mark is continuous, the conduction path is determined to be logically connected, and the conduction paths and failure rules that fail the logical connectivity verification are recorded.

9. A semantic enhancement and dynamic completion system for power market data graphs, used to execute the semantic enhancement and dynamic completion method for power market data graphs according to any one of claims 1 to 8, characterized in that: include: The acquisition module is used to respond to the data query instruction sent by the user and collect the structured target entity information and unstructured external data in the pre-established power market data map; a mapping module, configured to convert the external data into structured semantic tags, traverse the topological structure of the target entity information, establish a mapping relationship between the structured semantic tags and the target entity attributes, calculate a completeness index of the topological structure based on the mapping relationship in combination with the impact logic in the data query instruction, and mark missing entities or missing associations related to the data query instruction to generate a missing tag set; A generation module is configured to establish an attribute mapping between external data attributes and entity attributes based on the target entity information and the external data, generate a semantic vector integrating cross-domain features by parsing the implicit relationship between the external data and the target entity information in the attribute mapping, project the semantic vector into a topological space according to the missing tag set, and generate topological nodes and topological connections that inherit the cross-domain features; a reconstruction module, configured to embed the topological nodes and topological connections into the missing marked positions of the power market data map, reconstruct the topological structure, and verify the logical connectivity of the completed topological structure based on the relationship conduction rules; The verification module is used to parse the dynamic transmission chain formed by the completion path between the target entity information based on the verified topological structure, and generate a feedback interface including semantic relationships and completion structures to respond to data query instructions.

Images