Railway work electricity supply knowledge graph construction method based on multi-source data fusion

Abstract

The application discloses a railway engineering, electric power and power supply knowledge graph construction method based on multi-source data fusion, relates to the railway comprehensive operation and maintenance technical field, obtains multi-source heterogeneous data of a railway engineering system, an electric power system and a power supply system, and carries out preprocessing on the obtained data to generate a standardized railway engineering, electric power and power supply basic data set; railway engineering, electric power and power supply knowledge entity sets are constructed based on the standardized railway engineering, electric power and power supply basic data set, and a whole-process event sequence is constructed according to the railway engineering, electric power and power supply knowledge entity sets; correlation degree analysis is carried out on the whole-process event sequence, cross-professional correlation event sequences are obtained, and an initial railway engineering, electric power and power supply knowledge graph is constructed based on the cross-professional correlation event sequences; the initial railway engineering, electric power and power supply knowledge graph is subjected to correlation relationship optimization and dynamic updating to form a railway engineering, electric power and power supply knowledge graph, and the application realizes unified correlation of railway engineering, electric power and power supply professional data and cross-professional collaborative analysis, and improves the intelligent operation and maintenance capability of railway equipment operation.

Description

Technical Field

[0001] This invention relates to the field of railway integrated operation and maintenance technology, specifically a method for constructing a railway power supply knowledge graph based on multi-source data fusion. Background Technology

[0002] With the rapid development of my country's railway industry, the maintenance and management model of railway infrastructure has undergone profound changes. Traditionally, the three major professional fields of railway engineering, signaling and power supply have been operated and maintained independently, resulting in problems such as information silos, low collaboration efficiency and waste of resources. Data from various disciplines, including engineering, electrical, and power supply, is stored in a scattered and heterogeneous format, making it difficult to achieve cross-disciplinary data sharing and collaborative analysis. A large amount of maintenance experience, fault cases, and maintenance procedures exist in unstructured text form, lacking an effective organizational structure. Furthermore, equipment fault diagnosis and maintenance decisions rely excessively on the personal experience of on-site personnel, lacking systematic knowledge support. Railway maintenance involves multi-source heterogeneous data such as equipment operation data, testing and monitoring data, meteorological and environmental data, and historical maintenance records, which traditional analysis methods struggle to handle. Therefore, how to break down the data barriers between engineering, electrical, and power supply specialties, integrate multi-source heterogeneous data, and build a shareable and reasonable knowledge system is a problem we need to solve. To this end, we now provide a method for constructing a railway engineering, electrical, and power supply knowledge graph based on multi-source data fusion. Summary of the Invention

[0003] The purpose of this invention is to provide a method for constructing a railway power supply knowledge graph based on multi-source data fusion.

[0004] The objective of this invention can be achieved through the following technical solution: a method for constructing a railway power supply knowledge graph based on multi-source data fusion, comprising the following steps: Step S1: Acquire heterogeneous multi-source railway power supply data and preprocess it to generate a standardized railway power supply basic dataset; Step S2: Construct a railway engineering and electrical supply knowledge entity set based on the standardized railway engineering and electrical supply basic dataset, and then construct a full-process event sequence based on the railway engineering and electrical supply knowledge entity set; Step S3: Perform correlation analysis on the entire process event sequence to obtain cross-professional related event sequences, and construct an initial railway engineering and electrical supply knowledge graph based on the cross-professional related event sequences; Step S4: Optimize the initial railway engineering and electrical supply knowledge graph to obtain a new railway engineering and electrical supply knowledge graph, and dynamically update the railway engineering and electrical supply knowledge graph.

[0005] Furthermore, the process of acquiring multi-source heterogeneous data for railway power supply includes: The railway engineering, electrical, and power supply systems establish data interfaces to obtain multi-source heterogeneous data. The multi-source heterogeneous data includes structured data and unstructured data. The structured data contains spatial location data and equipment identification data. In addition, structured data also includes job status data and scheduling-related data.

[0006] Furthermore, the preprocessing of the acquired heterogeneous data from multiple railway power supply sources includes: The railway line is divided into several sections according to the preset section length. Each section is identified by a unique section number, and the start and end range of the section and the equipment contained in the corresponding section are recorded. By reading the spatial location data from the heterogeneous multi-source data of railway power supply, the location of each device is spatially mapped, so that devices located in the same or adjacent sections can establish a positional correspondence under a unified line spatial coordinate system. Time synchronization processing is performed on heterogeneous data from multiple sources for railway power supply through a unified time benchmark. A unified entity coding system will be established so that when the same equipment in the railway engineering system, signaling system, and power supply system has different codes, a unified identification will be carried out and a unified entity code will be assigned. After completing the above preprocessing, a standardized railway power supply basic dataset is obtained.

[0007] Furthermore, the process of constructing a set of railway electrical and power supply knowledge entities based on a standardized railway electrical and power supply basic dataset includes: By using a unified entity coding system, the same equipment in the engineering system, electrical system, and power supply system can be mapped to ensure that the same equipment is identified as the same entity in different systems. Based on the standardized railway power supply basic dataset and combined with spatial location data, equipment entities are identified, and a name, section location, equipment type and line number are generated for each equipment entity. Based on the operation status data, identify the operation entities and map each operation process to an operation entity. Generate the operation start time, operation end time, associated equipment and execution unit attributes for each operation entity. Based on the scheduling association data, identify the personnel entities and generate job information, execution team, associated equipment and time range of participation for each personnel entity. Identify rule entities based on unstructured data. The rule entities include equipment maintenance procedures, work procedures, and safety specifications. Generate a rule number, applicable equipment type, execution conditions, and time constraint attributes for each rule entity. Based on unstructured data, fault event entities are obtained. For each fault event entity, an event number, event type, associated equipment, occurrence time, fault level, scope of impact, and recovery status attributes are generated. The identified equipment entities, operation entities, personnel entities, rule entities, and fault event entities are summarized to complete the construction of the railway power supply knowledge entity set.

[0008] Furthermore, the process of constructing a full-process event sequence based on the railway electrical and power supply knowledge entity set includes: The railway power supply knowledge entity set is divided into dynamic entities and static entities. The dynamic entities include operation entities and fault event entities, and the static entities include equipment entities, personnel entities and rule entities. Based on the operation entities and fault event entities, the corresponding dynamic entities are arranged in chronological order to form an initial event sequence. After obtaining the initial event sequence, the static entities are embedded as associated nodes into the initial event sequence to obtain the full process event sequence.

[0009] Furthermore, the process of conducting correlation analysis on the entire process event sequence to obtain cross-professional related event sequences includes: By using spatial location data, if the devices involved in different dynamic entities are located in the same or adjacent segments, a spatial association is established and quantified as the degree of spatial association. Based on equipment type and spatial location data, determine whether there is a physical connection between the equipment. If there is such a direct connection between the equipment involved in two dynamic entities, establish equipment connection association and quantify it as the degree of equipment connection association. A time relationship is established based on the time interval between the time attributes corresponding to each dynamic entity. When the time corresponding to two entities is within a preset time window, it is determined that there is a time relationship between the two entities, and the degree of time relationship is quantified. Business logic relationships are established based on rule entities. When the state changes between entities follow the constraints of equipment maintenance procedures, operating procedures and safety specifications, it is determined that there is a business logic relationship between them, and the degree of business logic relationship is quantified. The obtained spatial correlation degree, device connection correlation degree, temporal correlation degree, and business logic correlation degree are weighted and fused to obtain the event correlation value; Dynamic entities whose event association values ​​exceed a preset association threshold are connected in chronological order, and related static entities are embedded as association nodes to form a structured cross-disciplinary associated event sequence.

[0010] Furthermore, the process of constructing an initial railway electrical supply knowledge graph based on cross-professional related event sequences includes: Equipment entities, operation entities, personnel entities, rule entities, and fault event entities are set as equipment nodes, operation nodes, personnel nodes, rule nodes, and fault event nodes in the railway power supply knowledge graph. Each node retains the attribute information of the corresponding entity in the standardized railway power supply basic dataset. The established relationships serve as edges between nodes, including spatial relationships, device connection relationships, temporal relationships, and business logic relationships. The obtained event association values ​​are used as the weights of the corresponding edges to represent the degree of association between nodes. When the event association value exceeds the preset association threshold, an edge is established between the corresponding nodes; otherwise, no edge is established. The order in which the relationships between nodes are established is determined by the order of the time attributes of the entities, and the relationships are established sequentially according to the time order. After defining and establishing the nodes and edges as described above, an initial railway power supply knowledge graph is obtained.

[0011] Furthermore, the process of optimizing the initial railway electrical and power supply knowledge graph to obtain a new railway electrical and power supply knowledge graph includes: Based on spatial location data and equipment type, the equipment nodes located in the same or adjacent sections are traversed and judged. When two equipment nodes are located in the same or adjacent sections, and the physical connection relationship between the equipment is judged based on the equipment type and spatial location data, but the corresponding edge has not yet been established in the current initial railway power supply knowledge graph, the association relationship is established and the weight of the corresponding edge is assigned to the preset basic connection value. The initial railway power supply knowledge graph, which has been supplemented with the equipment node association relationship, is traversed to check for the existence of isolated nodes. If an isolated node exists, then based on the spatial location data and time attributes of the node, find other nodes that are located in the same segment or adjacent segment and whose time interval is less than the preset time window. If a matching node is found, an association is established and the corresponding edge is obtained; if no other matching node is found, the node remains isolated. Regenerate the event association value corresponding to the obtained edge. If the corresponding event association value exceeds the preset association threshold, then use the event association value as the final weight of the corresponding edge; otherwise, delete the corresponding edge. For any two nodes with multiple duplicate edges, only the edge with the largest weight is retained, and the remaining duplicate edges are deleted. After completing the above process of supplementing the association relationships and processing isolated nodes in the initial railway engineering, electrical and power supply knowledge graph, the railway engineering, electrical and power supply knowledge graph is obtained.

[0012] Furthermore, the process of dynamically updating the railway electrical supply knowledge graph includes: Set a dynamic update cycle to continuously acquire new railway power supply multi-source heterogeneous data, and perform time synchronization processing, resampling and unified entity coding processing according to the aforementioned railway power supply multi-source heterogeneous data preprocessing process to form a new standardized railway power supply basic dataset. Read the newly added standardized railway power supply basic dataset, and identify new equipment entities, operation entities, personnel entities, rule entities, and fault event entities based on equipment identification data, spatial location data, operation status data, scheduling association data, and unstructured data; For each new entity, determine whether it already exists in the current railway engineering and electrical supply knowledge graph; When the unified entity code corresponding to the newly added entity does not exist in the railway engineering, electrical and power supply knowledge graph, a corresponding node is created and the attribute information corresponding to the newly added entity is written into the node. When the unified entity code corresponding to the new entity already exists in the railway engineering and electrical supply knowledge graph, the existing node attributes are read and compared with the corresponding attributes of the new entity. If there is a change, the attribute information in the existing node is updated. After the node update is completed, the event association values ​​between the new entities are regenerated according to the association relationship between the new entities and the business logic association. When the event association value between newly added entities exceeds the preset association threshold, a new edge is established between the corresponding nodes, and the event association value is used as the weight of the corresponding edge. If the updated event association value is lower than the preset association threshold, the corresponding edge is deleted to eliminate the invalid association relationship. After completing node and edge updates, the railway power supply knowledge graph is dynamically updated.

[0013] Compared with the prior art, the beneficial effects of the present invention are: to acquire multi-source heterogeneous data from railway engineering system, electrical system and power supply system, and to preprocess the acquired data to generate a standardized railway engineering, electrical and power supply basic dataset. By uniformly preprocessing the multi-source heterogeneous data of railway engineering, electrical and power supply, the data differences between different professional systems can be eliminated, and the consistency and reliability of multi-source data fusion can be improved. Based on the standardized railway engineering, electrical and power supply basic dataset, a set of railway engineering, electrical and power supply knowledge entities is constructed, and a full-process event sequence is constructed based on the set of railway engineering, electrical and power supply knowledge entities. By uniformly representing equipment, operations, personnel, rules and fault events and forming a full-process event sequence, a unified description of the operation process of railway engineering, electrical and power supply professionals can be achieved, and the cross-professional information association capability can be improved. By performing correlation analysis on the entire process event sequence, cross-professional correlation event sequences are obtained, and an initial railway engineering, electrical and power supply knowledge graph is constructed based on the cross-professional correlation event sequences. This correlation analysis method can accurately identify the implicit correlation between the three major professions of engineering, electrical and power supply, reveal the equipment fault propagation path and operation collaboration logic, and improve the accuracy and comprehensiveness of fault diagnosis. The initial railway engineering, electrical, and power supply knowledge graph is optimized and dynamically updated to form a railway engineering, electrical, and power supply knowledge graph. By supplementing the knowledge graph with relationships, correcting isolated nodes, and dynamically updating it, the integrity and timeliness of the knowledge graph structure can be improved, making it adaptable to the continuously changing data environment in railway operation scenarios.

[0014] This invention realizes the unified association and cross-professional collaborative analysis of railway engineering, signaling and power supply data, thereby improving the railway equipment fault prediction capability and operation and maintenance resource coordination capability, and improving railway operation safety and operation and maintenance efficiency. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0017] like Figure 1 As shown, the method for constructing a railway power supply knowledge graph based on multi-source data fusion includes the following steps: Step S1: Acquire heterogeneous multi-source railway power supply data and preprocess it to generate a standardized railway power supply basic dataset; Step S2: Construct a railway engineering and electrical supply knowledge entity set based on the standardized railway engineering and electrical supply basic dataset, and then construct a full-process event sequence based on the railway engineering and electrical supply knowledge entity set; Step S3: Perform correlation analysis on the entire process event sequence to obtain cross-professional related event sequences, and construct an initial railway engineering and electrical supply knowledge graph based on the cross-professional related event sequences; Step S4: Optimize the initial railway engineering and electrical supply knowledge graph to obtain a new railway engineering and electrical supply knowledge graph, and dynamically update the railway engineering and electrical supply knowledge graph.

[0018] It should be further explained that, in the specific implementation process, the process of acquiring multi-source heterogeneous data for railway power supply includes: The railway engineering, electrical, and power supply systems establish data interfaces to acquire multi-source heterogeneous data, which includes structured and unstructured data. The structured data includes engineering data, electrical data, and power supply data. The engineering data includes track geometry inspection data, turnout status data, bridge and tunnel structural status data, and rail flaw detection data. The electrical engineering data includes signal equipment operation data, interlocking status data, track circuit electrical parameters, and signal centralized monitoring system logs; The power supply data includes overhead contact line detection data, feeder load current, and substation operating parameters. The structured data includes spatial location data and device identification data; The spatial location data is used to characterize the positional relationship of the equipment in the line, including the line number to which the equipment belongs, the section location, the mileage marker, and the spatial coordinates of the equipment. The equipment identification data is used to characterize the unique identity information of the equipment, including the equipment number, the equipment type, and the professional category to which it belongs. In addition, structured data also includes job status data and scheduling-related data; The operational status data is used to characterize the status changes of the equipment during inspection, repair, maintenance, and fault handling. The scheduling association data is used to characterize the scheduling commands, execution teams, and job information during equipment operation. The unstructured data includes maintenance records, inspection logs, fault reports, work orders, equipment history documents, construction records, and video surveillance records.

[0019] It should be further explained that, in the specific implementation process, the preprocessing of the acquired multi-source heterogeneous railway power supply data to generate a standardized railway power supply basic dataset includes: The railway line is divided into several sections according to the preset section length. Each section is identified by a unique section number, and the start and end range of the section and the equipment contained in the corresponding section are recorded. By reading the spatial location data from the heterogeneous multi-source data of railway power supply, the location of each device is spatially mapped, so that devices located in the same or adjacent sections can establish a positional correspondence under a unified line spatial coordinate system. By using a unified time benchmark, time synchronization processing is performed on the heterogeneous data from multiple sources in railway engineering, electrical engineering, and power supply systems. For the different data collection cycles of the railway engineering system, electrical engineering system, and power supply system, a preset time window is used for resampling. For data with a sampling period shorter than the window length, the mean or median of all sampling points within the window is calculated as the representative value of the window. For data with a sampling period longer than the window length, forward padding is used to assign the previous valid value to the current window. A unified entity coding system will be established so that when the same equipment in the railway engineering system, signaling system, and power supply system has different codes, a unified identification will be carried out and a unified entity code will be assigned. The unified entity coding system includes line number, equipment type number and equipment location number. The line number determines the line to which the equipment belongs, the equipment type number distinguishes the equipment category, and the equipment location number determines the specific location of the equipment in the line, so as to achieve a unique identifier for each equipment. After completing the above preprocessing, a standardized railway power supply basic dataset is obtained.

[0020] It should be further explained that, in the specific implementation process, the process of constructing a set of railway electrical and electronic supply knowledge entities based on the standardized railway electrical and electronic supply basic dataset includes: By using a pre-built unified entity coding system, the same equipment in the engineering system, electrical system, and power supply system is mapped to ensure that the same equipment is identified as the same entity in different systems. Read the standardized railway electrical and power supply basic dataset, and identify the equipment entity based on the equipment number, equipment type and professional category in the equipment identification data, combined with the line number, section location and mileage mark in the spatial location data; The equipment entities include track sections, turnouts, bridges, tunnels, signals, interlocking equipment, track circuits, overhead contact line supports, transformers, and substations. For each equipment entity, a name, section location, equipment type, and line number are generated. Identify work entities based on work status data and map each work process to a work entity. The work entities include inspection work, maintenance work, repair work, and fault handling work. For each work entity, the work start time, work end time, associated equipment, and execution unit attributes are generated. Identify personnel entities based on the relationships between execution teams, job information, and dispatch commands in the scheduling association data; The personnel entities include engineering maintenance personnel, electrical maintenance personnel, power supply maintenance personnel, and dispatch personnel. For each personnel entity, job information, the work team to which they belong, associated equipment, and the time range for participating in the operation are generated. Identify rule entities based on equipment history documents and work orders in unstructured data; The rule entities include equipment maintenance procedures, work procedures, and safety specifications. For each rule entity, a rule number, applicable equipment type, execution conditions, and time constraint attributes are generated. Based on the maintenance records and fault reports in the unstructured data, identify fault event entities and map fault reports to fault event entities; The fault event entities include track deviation, turnout jamming, signal interlocking abnormality, track circuit fault, catenary disconnection and power supply abnormality events. For each fault event entity, an event number, event type, associated equipment, occurrence time, fault level, scope of impact and recovery status attributes are generated. The identified equipment entities, operation entities, personnel entities, rule entities, and fault event entities are summarized to complete the construction of the railway power supply knowledge entity set.

[0021] It should be further explained that, in the specific implementation process, the process of constructing the entire event sequence based on the railway engineering and electrical supply knowledge entity set includes: The knowledge entity set of railway engineering, electrical, and supply systems is divided into dynamic entities and static entities; The dynamic entities include operation entities and fault event entities; The static entities include equipment entities, personnel entities, and rule entities; Based on the start and end times of the work entities and the occurrence time of the fault event entities, the corresponding dynamic entities are arranged in chronological order to form an initial event sequence. The initial event sequence is used to describe the basic temporal relationship of the evolution of each dynamic entity in the railway engineering system, signaling system and power supply system over time. After obtaining the initial event sequence, static entities are embedded as associated nodes into the initial event sequence to obtain the full-process event sequence; Specifically: Associate the equipment entity with the corresponding fault event entity or operation entity based on the associated equipment included in the fault event entity and operation entity; Associate personnel entities with their corresponding work entities based on the execution units included in the work entity; The work entity is associated with equipment maintenance procedures, work procedures, safety regulations, and rule entities.

[0022] It should be further explained that, in the specific implementation process, the process of conducting correlation analysis on the entire process event sequence to obtain cross-professional related event sequences includes: By using spatial location data, if the devices involved in different dynamic entities are located in the same or adjacent segments, a spatial association is established and quantified as the degree of spatial association, denoted as... ; Based on equipment type and spatial location data, determine whether a physical connection exists between the equipment. If such a direct connection exists between the equipment involved in two dynamic entities, establish a device connection association and quantify it as the degree of device connection association, denoted as . ; A time correlation is established based on the time intervals between the time attributes corresponding to each dynamic entity. When the corresponding times of two entities fall within a preset time window, a time correlation is determined to exist between the two entities, and this correlation is quantified as the degree of correlation, denoted as . ; Business logic relationships are established based on rule entities. When state changes between entities follow equipment maintenance procedures, operating procedures, and safety regulations, a business logic relationship is determined to exist between them, and this is quantified as the degree of business logic relationship, denoted as [insert degree here]. ; The obtained spatial correlation, device connectivity correlation, temporal correlation, and business logic correlation are weighted and fused to calculate the event correlation value, which is denoted as... ,Right now: ; in, , , , The preset weighting coefficients satisfy... Relationship; Dynamic entities whose event association values ​​exceed a preset association threshold are connected in chronological order, and related static entities are embedded as association nodes to form a structured cross-disciplinary associated event sequence.

[0023] It should be further explained that, in the specific implementation process, the process of constructing the initial railway electrical and power supply knowledge graph based on cross-professional related event sequences includes: Read the equipment entities, operation entities, personnel entities, rule entities, and fault event entities in the cross-professional related event sequence, as well as the established spatial associations, equipment connection associations, temporal associations, and business logic associations, and use the equipment entities, operation entities, personnel entities, rule entities, and fault event entities as nodes in the railway engineering and electrical supply knowledge graph. Specifically, equipment entities are set as equipment nodes, operation entities are set as operation nodes, personnel entities are set as personnel nodes, rule entities are set as rule nodes, and fault event entities are set as fault event nodes. Each node retains the attribute information of the corresponding entity in the standardized railway power supply basic dataset. The attribute information includes unified entity code, name, location coordinates, line to which it belongs, operating status, and time attribute. The established relationships serve as edges between nodes, including spatial relationships, device connection relationships, temporal relationships, and business logic relationships. The obtained event association values ​​are used as the weights of the corresponding edges to represent the degree of association between nodes. When the event association value exceeds the preset association threshold, an edge is established between the corresponding nodes; otherwise, no edge is established. The order in which the relationships between nodes are established is determined by the order of the time attributes of the entities, and the relationships are established sequentially according to the time order. After defining and establishing the nodes and edges as described above, an initial railway power supply knowledge graph is obtained.

[0024] It should be further explained that, in the specific implementation process, the process of optimizing the initial railway engineering and electrical supply knowledge graph to obtain the railway engineering and electrical supply knowledge graph includes: Read the equipment nodes, operation nodes, personnel nodes, rule nodes, and fault event nodes in the initial railway power supply knowledge graph, as well as the established edges and their corresponding weights; Based on spatial location data and equipment type, traverse and judge equipment nodes located in the same or adjacent segments; When two device nodes are located in the same or adjacent segments, and a physical connection is determined to exist between the devices based on the device type and spatial location data, but no corresponding edge has been established in the current initial railway power supply knowledge graph, the association relationship is established and the weight of the corresponding edge is assigned to the preset basic connection value. The initial railway power supply knowledge graph, which has been supplemented with the association relationships of equipment nodes, is traversed to check for the existence of isolated nodes, which are nodes that have not yet established association relationships with any other nodes. If an isolated node exists, then based on the spatial location data and time attributes of the node, find other nodes that are located in the same segment or adjacent segment and whose time interval is less than the preset time window. If a node that meets the criteria is found, an association is established to obtain the corresponding edge, and it is marked as "weak association pending verification" in the processing log. If no other node that meets the criteria can be found, the node remains isolated and is marked as "node to be associated". Recalculate the event association value corresponding to the obtained edge. If the corresponding event association value exceeds the preset association threshold, the event association value is used as the final weight of the corresponding edge; otherwise, the corresponding edge is deleted. For any two nodes that have multiple duplicate edges, keep only the edge with the largest weight and delete the rest of the duplicate edges. After completing the above steps to supplement the relationships and process isolated nodes in the initial railway engineering, electrical, and power supply knowledge graph, the railway engineering, electrical, and power supply knowledge graph is obtained.

[0025] It should be further explained that, in the specific implementation process, the dynamic updating of the railway electrical supply knowledge graph includes: Set a dynamic update cycle to continuously acquire new railway power supply multi-source heterogeneous data, and perform time synchronization processing, resampling and unified entity coding processing according to the aforementioned railway power supply multi-source heterogeneous data preprocessing process to form a new standardized railway power supply basic dataset. Read the newly added standardized railway power supply basic dataset, and identify new equipment entities, operation entities, personnel entities, rule entities, and fault event entities based on equipment identification data, spatial location data, operation status data, scheduling association data, and unstructured data; For each new entity, determine whether it already exists in the current railway engineering and electrical supply knowledge graph; The newly added entities refer to entities identified from the newly added standardized railway engineering and electrical supply basic dataset within the current dynamic update cycle that have not yet been recorded in the existing railway engineering and electrical supply knowledge graph or whose attributes need to be updated. When the unified entity code corresponding to the newly added entity does not exist in the railway engineering, electrical and power supply knowledge graph, a corresponding node is created and the attribute information corresponding to the newly added entity is written into the node. When the unified entity code corresponding to the newly added entity already exists in the railway engineering, electrical and power supply knowledge graph, the existing node attributes are read and compared with the corresponding attributes of the newly added entity. If the attributes corresponding to the newly added entity change, then update the attribute information in the existing nodes; After the node update is completed, the event association values ​​between the newly added entities are recalculated based on the spatial association, device connection association, temporal association, and business logic association between the newly added entities. When the event association value between newly added entities exceeds the preset association threshold, a new edge is established between the corresponding nodes, and the event association value is used as the weight of the corresponding edge. If the updated event association value is lower than the preset association threshold, the corresponding edge is deleted to eliminate the invalid association relationship; After completing the above node and edge updates, the dynamic updating of the railway power supply knowledge graph is realized.

[0026] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications or equivalent substitutions made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for constructing a railway electrical and power supply knowledge graph based on multi-source data fusion, characterized in that, Includes the following steps: Step S1: Acquire heterogeneous multi-source railway power supply data and preprocess it to generate a standardized railway power supply basic dataset; Step S2: Construct a railway engineering and electrical supply knowledge entity set based on the standardized railway engineering and electrical supply basic dataset, and then construct a full-process event sequence based on the railway engineering and electrical supply knowledge entity set; Step S3: Perform correlation analysis on the entire process event sequence to obtain cross-professional related event sequences, and construct an initial railway engineering and electrical supply knowledge graph based on the cross-professional related event sequences; Step S4: Optimize the initial railway engineering and electrical supply knowledge graph to obtain a new railway engineering and electrical supply knowledge graph, and dynamically update the railway engineering and electrical supply knowledge graph. The preprocessing process for the acquired heterogeneous multi-source railway power supply data includes: The railway line is divided into several sections according to the preset section length. Each section is identified by a unique section number, and the start and end range of the section and the equipment contained in the corresponding section are recorded. By reading the spatial location data from the heterogeneous multi-source railway power supply data, the location of each device is spatially mapped, so that devices located in the same or adjacent sections can establish a positional correspondence under a unified line spatial coordinate system. Time synchronization processing is performed on heterogeneous data from multiple sources for railway power supply through a unified time benchmark. A unified entity coding system will be established so that when the same equipment in the railway engineering system, signaling system, and power supply system has different codes, a unified identification will be carried out and a unified entity code will be assigned. After completing the above preprocessing, a standardized railway power supply basic dataset is obtained; The process of constructing a full-process event sequence based on the railway power supply knowledge entity set includes: The railway power supply knowledge entity set is divided into dynamic entities and static entities. The dynamic entities include operation entities and fault event entities, and the static entities include equipment entities, personnel entities and rule entities. According to the operation entities and fault event entities, the corresponding dynamic entities are arranged in chronological order to form an initial event sequence. After obtaining the initial event sequence, the static entities are embedded as associated nodes into the initial event sequence to obtain the full process event sequence. The process of performing correlation analysis on the entire process event sequence to obtain cross-professional related event sequences includes: By using spatial location data, if the devices involved in different dynamic entities are located in the same or adjacent segments, a spatial association is established and quantified as the degree of spatial association. Based on equipment type and spatial location data, determine whether there is a physical connection between the equipment. If there is such a direct connection between the equipment involved in two dynamic entities, establish equipment connection association and quantify it as the degree of equipment connection association. A time relationship is established based on the time interval between the time attributes corresponding to each dynamic entity. When the time corresponding to two entities is within a preset time window, it is determined that there is a time relationship between the two entities, and the degree of time relationship is quantified. Business logic relationships are established based on rule entities. When the state changes between entities follow the constraints of equipment maintenance procedures, operating procedures and safety specifications, it is determined that there is a business logic relationship between them, and the degree of business logic relationship is quantified. The obtained spatial correlation degree, device connection correlation degree, temporal correlation degree, and business logic correlation degree are weighted and fused to obtain the event correlation value; Dynamic entities whose event association values ​​exceed a preset association threshold are connected in chronological order, and related static entities are embedded as association nodes to form a structured cross-disciplinary associated event sequence.

2. The method for constructing a railway power supply knowledge graph based on multi-source data fusion according to claim 1, characterized in that, The process of acquiring heterogeneous data from multiple sources for railway power supply includes: The railway engineering, electrical, and power supply systems establish data interfaces to obtain multi-source heterogeneous data. The multi-source heterogeneous data includes structured data and unstructured data. The structured data contains spatial location data and equipment identification data. In addition, structured data also includes job status data and scheduling-related data.

3. The method for constructing a railway power supply knowledge graph based on multi-source data fusion according to claim 1, characterized in that, The process of constructing a set of railway electrical and power supply knowledge entities based on a standardized railway electrical and power supply basic dataset includes: By using a unified entity coding system, the same equipment in the engineering system, electrical system, and power supply system can be mapped to ensure that the same equipment is identified as the same entity in different systems. Based on the standardized railway power supply basic dataset and combined with spatial location data, equipment entities are identified, and a name, section location, equipment type and line number are generated for each equipment entity. Based on the operation status data, identify the operation entities and map each operation process to an operation entity. Generate the operation start time, operation end time, associated equipment and execution unit attributes for each operation entity. Based on the scheduling association data, identify the personnel entities and generate job information, execution team, associated equipment and time range of participation for each personnel entity. Identify rule entities based on unstructured data. The rule entities include equipment maintenance procedures, work procedures, and safety specifications. Generate a rule number, applicable equipment type, execution conditions, and time constraint attributes for each rule entity. Based on unstructured data, fault event entities are obtained. For each fault event entity, an event number, event type, associated equipment, occurrence time, fault level, scope of impact, and recovery status attributes are generated. The identified equipment entities, operation entities, personnel entities, rule entities, and fault event entities are summarized to complete the construction of the railway power supply knowledge entity set.

4. The method for constructing a railway power supply knowledge graph based on multi-source data fusion according to claim 1, characterized in that, The process of constructing an initial railway electrical supply knowledge graph based on cross-disciplinary related event sequences includes: Equipment entities, operation entities, personnel entities, rule entities, and fault event entities are set as equipment nodes, operation nodes, personnel nodes, rule nodes, and fault event nodes in the railway power supply knowledge graph. Each node retains the attribute information of the corresponding entity in the standardized railway power supply basic dataset. The established relationships serve as edges between nodes, including spatial relationships, device connection relationships, temporal relationships, and business logic relationships. The obtained event association values ​​are used as the weights of the corresponding edges to represent the degree of association between nodes. When the event association value exceeds the preset association threshold, an edge is established between the corresponding nodes; otherwise, no edge is established. The order in which the relationships between nodes are established is determined by the order of the time attributes of the entities, and the relationships are established sequentially according to the time order. After defining and establishing the nodes and edges as described above, an initial railway power supply knowledge graph is obtained.

5. The method for constructing a railway power supply knowledge graph based on multi-source data fusion according to claim 4, characterized in that, The process of optimizing the initial railway electrical and power supply knowledge graph to obtain the railway electrical and power supply knowledge graph includes: Based on spatial location data and equipment type, the equipment nodes located in the same or adjacent sections are traversed and judged. When two equipment nodes are located in the same or adjacent sections, and the physical connection relationship between the equipment is judged based on the equipment type and spatial location data, but the corresponding edge has not yet been established in the current initial railway power supply knowledge graph, the association relationship is established and the weight of the corresponding edge is assigned to the preset basic connection value. The initial railway power supply knowledge graph, which has been supplemented with the equipment node association relationship, is traversed to check for the existence of isolated nodes. If an isolated node exists, then based on the spatial location data and time attributes of the node, find other nodes that are located in the same segment or adjacent segment and whose time interval is less than the preset time window. If a matching node is found, an association is established and the corresponding edge is obtained; if no other matching node is found, the node remains isolated. Regenerate the event association value corresponding to the obtained edge. If the corresponding event association value exceeds the preset association threshold, then use the event association value as the final weight of the corresponding edge; otherwise, delete the corresponding edge. For any two nodes with multiple duplicate edges, only the edge with the largest weight is retained, and the remaining duplicate edges are deleted. After completing the above process of supplementing the association relationships and processing isolated nodes in the initial railway engineering, electrical and power supply knowledge graph, the railway engineering, electrical and power supply knowledge graph is obtained.

6. The method for constructing a railway power supply knowledge graph based on multi-source data fusion according to claim 5, characterized in that, The process of dynamically updating the railway electrical supply knowledge graph includes: Set a dynamic update cycle to continuously acquire new railway power supply multi-source heterogeneous data, and perform time synchronization processing, resampling and unified entity coding processing according to the aforementioned railway power supply multi-source heterogeneous data preprocessing process to form a new standardized railway power supply basic dataset. Read the newly added standardized railway power supply basic dataset, and identify new equipment entities, operation entities, personnel entities, rule entities, and fault event entities based on equipment identification data, spatial location data, operation status data, scheduling association data, and unstructured data; For each new entity, determine whether it already exists in the current railway engineering and electrical supply knowledge graph; When the unified entity code corresponding to the newly added entity does not exist in the railway engineering, electrical and power supply knowledge graph, a corresponding node is created and the attribute information corresponding to the newly added entity is written into the node. When the unified entity code corresponding to the new entity already exists in the railway engineering and electrical supply knowledge graph, the existing node attributes are read and compared with the corresponding attributes of the new entity. If there is a change, the attribute information in the existing node is updated. After the node update is completed, the event association values ​​between the new entities are regenerated according to the association relationship between the new entities and the business logic association. When the event association value between newly added entities exceeds the preset association threshold, a new edge is established between the corresponding nodes, and the event association value is used as the weight of the corresponding edge. If the updated event association value is lower than the preset association threshold, the corresponding edge is deleted to eliminate the invalid association relationship. After completing node and edge updates, the railway power supply knowledge graph is dynamically updated.

Images