Rail transit operation and maintenance knowledge graph processing method and device and train

Abstract

The invention provides a rail transit operation and maintenance knowledge graph processing method, and relates to the technical field of knowledge graphs and artificial intelligence. The method comprises the following steps: processing rail transit operation and maintenance historical data to obtain a rail transit operation and maintenance ontology model with unified semantic representation; performing knowledge extraction operation and timestamp identification operation on the unstructured data associated with the rail transit operation and maintenance historical data to obtain an entity-relation-entity triad and a timestamp, and binding the entity-relation-entity triad and the timestamp to obtain a sequential entity-relation-entity triad; and based on the rail transit operation and maintenance ontology model and the operation and maintenance scene, performing knowledge fusion operation based on time sequence information guidance on the time sequence entity-relation-entity triple to obtain a rail transit operation and maintenance knowledge graph. The invention further provides a rail transit operation and maintenance knowledge graph processing device and a train.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of knowledge graph and artificial intelligence, and more particularly to a processing method and device of rail transit operation and maintenance knowledge graph and a train. BACKGROUND

[0002] With the rapid development of urban rail transit, the train operation control system becomes the core to ensure the safe and efficient operation of the train. However, due to the complex system structure and large amount of data, the traditional fault diagnosis method relies on manual experience, and it is difficult to achieve rapid and accurate fault positioning. At the same time, a large amount of unstructured fault log information generated during train operation increases the difficulty of knowledge extraction and analysis. In order to solve these problems, intelligent fault diagnosis and prediction methods have become an important research direction in the field of rail transit. Rail transit operation and maintenance is a strong time sequence dynamic process. The occurrence, development, maintenance and degradation of the equipment all have obvious time dimension characteristics.

[0003] In the process of implementing the present disclosure, the inventors found that at least the following problems exist in the related art: the existing rail transit operation and maintenance scheme based on knowledge graph lacks effective capture, representation and reasoning ability of time information in operation and maintenance knowledge, and cannot handle complex rail transit operation and maintenance problems based on time sequence. SUMMARY

[0004] Therefore, the present disclosure provides a processing method and device of rail transit operation and maintenance knowledge graph and a train.

[0005] One aspect of the present disclosure provides a processing method of rail transit operation and maintenance knowledge graph, comprising: processing rail transit operation and maintenance historical data to obtain a rail transit operation and maintenance ontology model with unified semantic representation; performing knowledge extraction and timestamp identification operations on unstructured data associated with the rail transit operation and maintenance historical data to obtain entity-relation-entity triples and timestamps, and binding the entity-relation-entity triples and the timestamps to obtain time sequence entity-relation-entity triples; based on the rail transit operation and maintenance ontology model and the operation and maintenance scene, performing knowledge fusion operations guided by time sequence information on the time sequence entity-relation-entity triples to obtain a rail transit operation and maintenance knowledge graph.

[0006] According to the embodiments of the present disclosure, processing the rail transit operation and maintenance historical data to obtain the rail transit operation and maintenance ontology model with unified semantic representation comprises: using a large language model to process multiple types of heterogeneous data in the rail transit operation and maintenance historical data in an interactive prompt word manner to obtain a resource description framework mode with unified semantic representation; using the resource description framework mode to define domain concepts and ontology structures for constructing the rail transit operation and maintenance knowledge graph to obtain the rail transit operation and maintenance ontology model.

[0007] According to an embodiment of the present disclosure, based on the rail transit operation ontology model and the operation scene, the knowledge fusion operation guided by the time sequence information is performed on the time sequence entity-relation-entity triplets to obtain the rail transit operation knowledge graph, which comprises: reprocessing the processing results of the rail transit operation historical data and the unstructured data by using a large language model to obtain secondary processing results, and determining the operation scene by entity type recognition based on the secondary processing results; in the case that the operation scene is a fault scene, based on the rail transit operation ontology model, knowledge retrieval and confidence calculation are performed on the fault diagnosis library by using the time sequence information guided fault propagation chain, and attribute filling and knowledge fusion are performed on the time sequence entity-relation-entity triplets by using the knowledge retrieval results and the confidence calculation results.

[0008] According to an embodiment of the present disclosure, based on the rail transit operation ontology model and the operation scene, the knowledge fusion operation guided by the time sequence information is performed on the time sequence entity-relation-entity triplets to obtain the rail transit operation knowledge graph, which further comprises: in the case that the operation scene is a maintenance scene, the time sequence information is verified by a fault-maintenance-recovery maintenance business chain, and based on the rail transit operation ontology model, attribute filling and knowledge fusion are performed on the time sequence entity-relation-entity triplets by using the verification results; in the case that the operation scene is a device state scene, the state information of the device entity is matched with the health degree mode in the device health degree mode library in real time sequence, and based on the rail transit operation ontology model, attribute filling and knowledge fusion are performed on the time sequence entity-relation-entity triplets by using the matching results to obtain the rail transit operation knowledge graph, wherein the health degree mode comprises normal state, early warning state, sub-health state and fault state.

[0009] According to an embodiment of the present disclosure, the processing method of the rail transit operation knowledge graph further comprises: performing an incremental knowledge extraction operation on the associated data of the new rail transit operation data to obtain an incremental knowledge extraction result; performing a conflict verification based on the time sequence relationship on the incremental knowledge extraction result to obtain a conflict verification result, and dynamically updating the rail transit operation knowledge graph by using the conflict verification result.

[0010] According to an embodiment of the present disclosure, the conflict verification based on the time sequence relationship is performed on the incremental knowledge extraction result to obtain the conflict verification result, which comprises: obtaining the fault propagation chain, the maintenance business chain and the device health degree mode library associated with the rail transit operation knowledge graph; performing a conflict verification based on the time sequence fault rule on the incremental knowledge extraction result and the fault propagation chain to obtain a first conflict verification result; performing a conflict verification based on the time sequence maintenance operation on the incremental knowledge extraction result and the maintenance business chain to obtain a second conflict verification result; and performing a conflict verification based on the time sequence health degree mode on the incremental knowledge extraction result and the device health degree mode library to obtain a third conflict verification result.

[0011] According to an embodiment of the present disclosure, the dynamic updating of the rail transit operation and maintenance knowledge graph by using the conflict checking result comprises: in a case where the first conflict checking result represents that the incremental knowledge extraction result is consistent with the logic of the fault transmission chain and / or in a case where the second conflict checking result represents that the incremental knowledge extraction result is consistent with the standard operation and time logic of the maintenance business chain and / or in a case where the second conflict checking result represents that the incremental knowledge extraction result is consistent with the standard operation and time logic of the maintenance business chain, determining a new entity-relation-entity triple from the incremental knowledge extraction result, and dynamically updating the rail transit operation and maintenance knowledge graph by using the new entity-relation-entity triple.

[0012] According to an embodiment of the present disclosure, the processing method of the rail transit operation and maintenance knowledge graph further comprises: analyzing new rail transit operation and maintenance data by using a large language model to obtain entity information and scene information; based on the entity information and the scene information, performing semantic reasoning on the rail transit operation and maintenance knowledge graph to obtain attribute information of the entity information and an operation and maintenance scheme associated with the scene information, wherein the semantic reasoning comprises time sequence rule reasoning, time sequence graph neural network reasoning and time sequence causal chain reasoning.

[0013] Another aspect of the present disclosure provides a processing device of a rail transit operation and maintenance knowledge graph, characterized in that comprising: an ontology model acquisition module configured to process rail transit operation and maintenance historical data to obtain a rail transit operation and maintenance ontology model in a unified semantic representation; a time sequence triple acquisition module configured to perform knowledge extraction and timestamp identification operations on unstructured data associated with the rail transit operation and maintenance historical data to obtain entity-relation-entity triples and timestamps, and to perform binding processing on the entity-relation-entity triples and the timestamps to obtain time sequence entity-relation-entity triples; and a rail transit operation and maintenance knowledge graph construction module configured to perform knowledge fusion operations guided by time sequence information on the time sequence entity-relation-entity triples based on the rail transit operation and maintenance ontology model and an operation and maintenance scene to obtain the rail transit operation and maintenance knowledge graph.

[0014] Another aspect of the present disclosure provides a train comprising the above-mentioned processing device of a rail transit operation and maintenance knowledge graph.

[0015] Another aspect of the present disclosure provides an electronic device comprising: one or more processors; a memory configured to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described above.

[0016] Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions for implementing the method described above when executed.

[0017] Another aspect of the present disclosure provides a computer program product comprising computer-executable instructions for implementing the method described above when executed.

[0018] According to the embodiments of the present disclosure, because the technical means of binding timestamps with entity-relation-entity triples is adopted, the technical problem that the static rail transit operation and maintenance knowledge graph lacks effective capturing, representation and reasoning ability is at least partially overcome, and the technical effects of intelligent diagnosis and development trend prediction of rail transit faults are achieved. Meanwhile, according to the embodiments of the present disclosure, because the technical means of constructing a knowledge graph based on an operation and maintenance scenario is adopted, the rapid positioning of rail transit faults and the generation of targeted operation and maintenance solutions can be achieved. In addition, according to the embodiments of the present disclosure, because the unified semantic representation rail transit operation and maintenance ontology model is adopted, the problem of information fragmentation caused by the non-uniformity of data formats and entity concepts in existing knowledge graphs is at least partially overcome, and the standardization and specialization of rail transit operation and maintenance are achieved. BRIEF DESCRIPTION OF DRAWINGS

[0019] The above and other objects, features and advantages of the present disclosure will become more apparent from the following description of embodiments of the present disclosure taken in conjunction with the accompanying drawings, in which:

[0020] Figure 1 An exemplary system architecture to which the rail transit operation and maintenance knowledge graph processing method according to embodiments of the present disclosure can be applied is schematically shown.

[0021] Figure 2 A flowchart of the rail transit operation and maintenance knowledge graph processing method according to embodiments of the present disclosure is schematically shown.

[0022] Figure 3 A construction process diagram of the rail transit operation and maintenance knowledge graph based on a large language model according to embodiments of the present disclosure is shown.

[0023] Figure 4 A structural diagram of a rail transit operation and maintenance knowledge graph processing apparatus according to embodiments of the present disclosure is schematically shown.

[0024] Figure 5 A block diagram of an electronic device suitable for implementing the method described above according to embodiments of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0025] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood, however, that the description is merely exemplary and is not intended to limit the scope of the present disclosure. In the following detailed description of the embodiments of the present disclosure, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. However, it would be apparent to those skilled in the art that the embodiments of the present disclosure can be practiced without these specific details. In other instances, well-known structures and methods are not described in detail in order to avoid obscuring the concepts of the present disclosure.

[0026] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present disclosure. As used herein, the term "includes" and tautological expressions thereof, such as "including," means the inclusion of but not limited to, but also encompassing or more other non-specified items.

[0027] All terms used herein, including technical and scientific terms, have the meanings commonly understood by one of ordinary skill in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of the specification, and should not be interpreted in an idealized or overly formal way.

[0028] In the case of using expressions similar to "at least one of A, B, and C, etc.", it should generally be interpreted to include any of them, to include, for example, only A, only B, only C, or a combination of A and B, A and C, B and C, or A and B and C, etc.

[0029] In the embodiments of the present disclosure, the collection, updating, analysis, processing, use, transmission, provision, disclosure, storage, etc. of the data involved (for example, including but not limited to user personal information) comply with the relevant legal regulations, are used for legal purposes, and do not violate public order and good customs. In particular, necessary measures have been taken to prevent illegal access to user personal information data, to maintain the security of user personal information, network security and national security.

[0030] In the embodiments of the present disclosure, the authorization or consent of the user is obtained before the user personal information is acquired or collected, and when applicable, the user personal information is subjected to de-identification and / or anonymization and / or encryption technical processing.

[0031] In recent years, although some researches use large language models to construct knowledge graphs to improve the efficiency of intelligent question answering, these methods construct knowledge graphs that are essentially static and lack a time dimension. The existing construction methods of rail transit operation and maintenance knowledge graphs mainly rely on the following methods: rule-based methods: through artificial rule and template, the text data is structured, such as knowledge extraction based on regular expressions or hand-written knowledge base; machine learning-based methods: use deep learning models for entity recognition and relationship extraction; graph database-based solutions: use graph database software or ontology editing and knowledge acquisition software to provide knowledge graph construction and storage capabilities.

[0032] However, the existing construction methods of rail transit operation and maintenance knowledge graphs lack effective capture, representation and reasoning capabilities for time information in operation and maintenance knowledge, and cannot answer complex questions based on time series, such as "how does the historical failure mode of a certain device evolve?" or "is this failure caused by a previous incomplete maintenance operation?" For example, some existing technical solutions lack the use of time series dimension, either completely ignoring the time information in the data or only storing and using the time information as ordinary data, lacking deep modeling and utilization of time information. The lack of time series dimension or superficial processing leads to the inability to capture the time series relationship and dynamic evolution process between "failure occurrence, development and maintenance" in operation and maintenance knowledge, and thus the graph cannot support complex reasoning queries based on time series. Some existing technical solutions construct static graphs, and their knowledge graphs are static snapshots. Since rail transit operation and maintenance is a strong time series dynamic process, static graphs cannot show the causal relationship in time series. This leads to a lack of ability to trace back to historical states and deduce future trends. Some existing technical solutions use large language models to construct knowledge graphs, but lack adaptability to industrial time series scenarios and do not consider the strong time series causal characteristics in rail transit operation and maintenance, which cannot meet the requirements of scenarios that require deep analysis of time series, such as fault prediction and maintenance decision optimization. In summary, although the existing technical solutions have improved the construction efficiency of knowledge graphs to some extent, they still lack effective capture, representation, updating and reasoning capabilities for time information in operation and maintenance knowledge. Therefore, developing a dynamic knowledge graph construction and reasoning method that can deeply integrate time series analysis capabilities is a technical problem that needs to be solved in this technical field.

[0033] Figure 1 An exemplary system architecture 100 to which the processing method of the rail transit operation and maintenance knowledge graph according to embodiments of the present disclosure can be applied is schematically shown. It should be noted that, Figure 1 The shown is only an example of a system architecture to which embodiments of the present disclosure can be applied, to help those skilled in the art understand the technical content of the present disclosure, but does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0034] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0035] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social media platform software, etc. (for example only).

[0036] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0037] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0038] It should be noted that the processing method of the rail transit operation knowledge graph provided in the embodiments of the present disclosure can generally be executed by the server 105. Correspondingly, the processing device of the rail transit operation knowledge graph provided in the embodiments of the present disclosure can generally be arranged in the server 105. The processing method of the rail transit operation knowledge graph provided in the embodiments of the present disclosure can also be executed by a server or a server cluster different from the server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Correspondingly, the processing device of the rail transit operation knowledge graph provided in the embodiments of the present disclosure can also be arranged in a server or a server cluster different from the server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Alternatively, the processing method of the rail transit operation knowledge graph provided in the embodiments of the present disclosure can also be executed by the first terminal device 101, the second terminal device 102 or the third terminal device 103, or by other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103. Correspondingly, the processing device of the rail transit operation knowledge graph provided in the embodiments of the present disclosure can also be arranged in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0039] For example, the rail transit operation related data or information to be processed can originally be stored in any one of the first terminal device 101, the second terminal device 102 or the third terminal device 103 (for example, the first terminal device 101, but not limited thereto), or stored on an external storage device and imported into the first terminal device 101. Then, the first terminal device 101 can execute the processing method of the rail transit operation knowledge graph provided in the embodiments of the present disclosure locally, or send the rail transit operation related data or information to be processed to other terminal devices, servers or server clusters, and execute the processing method of the rail transit operation knowledge graph provided in the embodiments of the present disclosure by other terminal devices, servers or server clusters receiving the rail transit operation related data or information to be processed.

[0040] It should be understood that Figure 1 The number of terminal devices, networks and servers in the above-mentioned system is only illustrative. According to the needs of implementation, there can be any number of terminal devices, networks and servers.

[0041] Figure 2 A flowchart of the processing method of the rail transit operation knowledge graph according to the embodiments of the present disclosure is schematically shown.

[0042] AsFigure 2 As shown, the processing method of the rail transit operation knowledge graph includes operations S210-S230.

[0043] In operation S210, rail transit operation history data is processed to obtain a rail transit operation ontology model in a unified semantic representation.

[0044] In the process of processing rail transit operation history data (such as equipment information, maintenance logs, etc.), a prompt word engineering module of a large language model can be used for processing, and related concepts in the rail transit operation field are unified in semantics and the structure of the operation knowledge graph to be constructed is determined, such as equipment type, fault type, maintenance scheme, etc.

[0045] In the process of obtaining the ontology model, the disclosed embodiments make full use of the zero-shot learning capability of the large language model, and can automatically extract relevant knowledge from structured or unstructured data, greatly improving the efficiency of constructing the knowledge graph.

[0046] In the process of obtaining the rail transit operation ontology model, prompt words are generated through the design of field concept layers to guide the large language model to construct the basic structure of the knowledge graph. Data extraction prompt words are designed, taking into account the timing information, to achieve efficient knowledge filling.

[0047] In operation S220, knowledge extraction and timestamp identification operations are performed on unstructured data associated with the rail transit operation history data to obtain entity-relation-entity triples and timestamps, and the entity-relation-entity triples and timestamps are bound to obtain time-series entity-relation-entity triples.

[0048] Through the use of another prompt word engineering module of the large language model, entities and relationships are extracted from unstructured data (such as text work orders, fault logs, and equipment manuals) associated with the rail transit operation history data, and all time information, such as fault occurrence time and maintenance duration, is identified. The timestamps are bound to the extracted triples (entity-relation-entity). Natural Language Processing (NLP) methods are used for entity recognition, relationship extraction, and attribute filling to ensure data integrity. A knowledge graph conforming to the RDF (Resource Description Framework) syntax is generated, and optimized storage is performed through relevant query interfaces.

[0049] In operation S230, based on the rail transit operation ontology model and the operation scenario, time-series entity-relation-entity triples are subjected to knowledge fusion operations guided by timing information to obtain a rail transit operation knowledge graph.

[0050] According to the embodiment of the present disclosure, because the technical means of binding timestamps with entity-relation-entity triples is adopted, the technical problem that the static rail transit operation and maintenance knowledge graph lacks effective capture, representation and reasoning ability is at least partially overcome, and the technical effect of intelligent diagnosis and development trend prediction of rail transit faults is achieved. At the same time, according to the embodiment of the present disclosure, because the construction technology of the knowledge graph based on the operation and maintenance scene is adopted, the rapid positioning of rail transit faults and the generation of targeted operation and maintenance schemes can be realized. In addition, according to the embodiment of the present disclosure, therefore, the unified semantic representation of the rail transit operation and maintenance ontology model is adopted, so that the problem of information fragmentation caused by the non-uniformity of data formats and entity concepts existing in the existing knowledge graph is at least partially overcome, and the standardization and specialization of rail transit operation and maintenance are realized.

[0051] According to the embodiment of the present disclosure, the rail transit operation and maintenance historical data is processed to obtain a unified semantic representation of a rail transit operation and maintenance ontology model, which includes: using a large language model to process multiple types of heterogeneous data in the rail transit operation and maintenance historical data in an interactive prompt word manner to obtain a unified semantic representation of a resource description framework mode; using the resource description framework mode to define the domain concepts and ontology structure for constructing the rail transit operation and maintenance knowledge graph to obtain the rail transit operation and maintenance ontology model.

[0052] Through the natural language processing capability of the large language model, unstructured data such as text, image, video and sensor log can be automatically parsed and converted into RDF triples with unified semantics, solving the problem of large data format difference and weak correlation in traditional operation and maintenance. At the same time, the resource description framework mode with unified semantic representation can clearly define the core concepts and their relationships in the rail transit field such as equipment, fault and maintenance, form an extensible knowledge graph framework, and support cross-industry collaborative analysis. In addition, the interactive prompt word technology allows operation and maintenance personnel to query historical data through natural language, and the large language model can generate semantic description in real time and feed back to the ontology model, realizing the continuous iteration of the knowledge base.

[0053] According to an embodiment of the present disclosure, based on the rail transit operation ontology model and the operation scene, a knowledge fusion operation guided by time sequence information is performed on the time sequence entity-relation-entity triplets to obtain a rail transit operation knowledge graph, including: reprocessing the processing results of the rail transit operation historical data and the processing results of the unstructured data by using a large language model to obtain secondary processing results, and determining the operation scene by entity type recognition based on the secondary processing results; in the case that the operation scene is a fault scene, based on the rail transit operation ontology model, knowledge retrieval and confidence calculation are performed in the fault diagnosis library by using the time sequence information guided fault transmission chain, and attribute filling and knowledge fusion are performed on the time sequence entity-relation-entity triplets by using the knowledge retrieval results and the confidence calculation results.

[0054] The construction of the knowledge graph of the embodiment of the present disclosure is a construction method of a knowledge graph based on time sequence, that is, time information is taken as an important basis for the construction of the knowledge graph, and the construction of a dynamic knowledge graph is realized: by reprocessing the processing results of the rail transit operation historical data and the processing results of the unstructured data, entity types are identified, and the operation business scene of the entity types is intelligently judged. And according to the operation business scene, the related processing flow is called.

[0055] Through the secondary processing of the historical data and the unstructured data (such as sensor logs, maintenance records) by the large language model, the multi-modal information such as text and image can be converted into time sequence triplets with unified semantics; at the same time, based on the fault transmission chain constructed by the ontology model, knowledge retrieval and confidence calculation can be performed in the diagnosis library combined with the time sequence information; in addition, through the attribute filling and knowledge fusion operation, the knowledge graph can be dynamically expanded.

[0056] In the construction process of the rail transit operation knowledge graph, data fusion can be performed in combination with the existing operation knowledge base (for example, a fault diagnosis knowledge base), so as to improve the coverage rate of the operation knowledge base.

[0057] According to an embodiment of the present disclosure, based on the rail transit operation ontology model and the operation scene, the knowledge fusion operation guided by the time sequence information is performed on the time sequence entity-relation-entity triplets to obtain the rail transit operation knowledge graph, which further includes: in the case that the operation scene is a maintenance scene, the time sequence information is verified through a maintenance business chain of fault-maintenance-recovery, and based on the rail transit operation ontology model, the verification result is used to perform attribute filling and knowledge fusion on the time sequence entity-relation-entity triplets; in the case that the operation scene is a device state scene, the state information of the device entity is matched with the health degree mode in the device health degree mode library in real time sequence, and based on the rail transit operation ontology model, the matching result is used to perform attribute filling and knowledge fusion on the time sequence entity-relation-entity triplets to obtain the rail transit operation knowledge graph, wherein the health degree mode includes a normal state, a warning state, a sub-health state and a fault state.

[0058] The above embodiments of the present disclosure will be further described in detail through specific embodiments.

[0059] In the case that the operation scene is a fault scene, a fault transmission knowledge base is queried, and a leadsTo relationship is created under the condition of meeting a time window and a rule. A fault transmission time window knowledge base is built in, and according to the type and time of a new fault entity, other faults occurring within the time window are automatically queried in the graph database. If there is a transmission rule in the knowledge base, a leadsTo relationship is automatically created, and the confidence is calculated.

[0060] In the case that the operation scene is a maintenance scene, a state machine and a business rule verifier are started, the time sequence logic compliance is verified, and the result is marked as an entity attribute. The state change event of the work order is listened to, whether the time stamp is strictly increasing is verified, and the time consumption of each link is calculated. If the rule is violated or the time is exceeded, the verification failure is automatically marked.

[0061] In the case that the operation scene is a device state, real-time sequence matching is performed with the device health degree mode library. If the matching is successful, predictive knowledge and early warning work orders are generated.

[0062] The knowledge generated in the above operation scenes is enhanced and verified, and the above knowledge is stored in the graph database.

[0063] The embodiments of the present disclosure realize deep understanding and active guidance of operation time sequence information, which goes beyond simple time stamp recording, and realizes intelligent reasoning and prediction based on domain knowledge.

[0064] According to an embodiment of the present disclosure, the processing method of the rail transit operation and maintenance knowledge graph further includes: performing an incremental knowledge extraction operation on associated data of the new rail transit operation and maintenance data to obtain an incremental knowledge extraction result; performing conflict checking based on a time sequence relationship on the incremental knowledge extraction result to obtain a conflict checking result, and dynamically updating the rail transit operation and maintenance knowledge graph by using the conflict checking result.

[0065] Through input of real-time operation and maintenance data and incremental knowledge extraction (or incremental learning), the timeliness and accuracy of the knowledge graph can be ensured, and time sequence, scene and knowledge fusion and updating of the knowledge graph can be realized.

[0066] According to an embodiment of the present disclosure, the conflict checking based on the time sequence relationship on the incremental knowledge extraction result to obtain the conflict checking result includes: obtaining a fault transmission chain, a maintenance business chain and a device health degree mode library associated with the rail transit operation and maintenance knowledge graph; performing conflict checking based on a time sequence fault rule on the incremental knowledge extraction result and the fault transmission chain to obtain a first conflict checking result; performing conflict checking based on a time sequence maintenance operation on the incremental knowledge extraction result and the maintenance business chain to obtain a second conflict checking result; and performing conflict checking based on a time sequence health degree mode on the incremental knowledge extraction result and the device health degree mode library to obtain a third conflict checking result.

[0067] In the field of rail transit operation and maintenance, as time goes by, a series of subsequent faults occur in other associated devices when a device fails, forming a "fault transmission chain". Through the time sequence information guiding module, the graph can automatically construct a fault root cause analysis chain to realize active root cause analysis of the fault. The core of operation and maintenance is a closed loop of "fault-maintenance-recovery". Each time point in the life cycle of an effective maintenance activity must comply with the business logic. Through the time sequence information guiding module, the business rules are converted into calculable and monitorable time sequence indexes to realize digital management of the operation and maintenance process. Operation and maintenance pay attention to the health state change of the whole life cycle of the device, which is a unique time sequence mode. The device state is not simply "normal / fault", but has a degradation process, for example: normal-> warning-> sub-health-> fault. Based on the time sequence information, historical data and expert evaluation, the device is comprehensively evaluated, and the fault handling is changed from "after-the-fact remediation" to "pre-emptive warning" predictive maintenance.

[0068] Therefore, it is necessary to perform conflict checking on the newly added knowledge for the dynamic updating of the rail transit operation and maintenance knowledge graph based on the time sequence.

[0069] According to an embodiment of the present disclosure, the dynamic updating of the rail transit operation and maintenance knowledge graph by using the conflict checking result comprises: in a case where the first conflict checking result represents that the incremental knowledge extraction result is consistent with the logic of the fault transmission chain and / or in a case where the second conflict checking result represents that the incremental knowledge extraction result is consistent with the standard operation and time logic of the maintenance business chain and / or in a case where the second conflict checking result represents that the incremental knowledge extraction result is consistent with the standard operation and time logic of the maintenance business chain, a new entity-relation-entity triple is determined from the incremental knowledge extraction result, and the rail transit operation and maintenance knowledge graph is dynamically updated by using the new entity-relation-entity triple.

[0070] Embodiments of the present disclosure support automatic adaptation and updating of new data through an incremental learning mechanism, ensuring continuous optimization of the knowledge graph. When the newly extracted time-based facts conflict with existing knowledge, a timestamp-based latest priority strategy is adopted for resolution, ensuring that the graph state is consistent with the latest reality. The technical solution of the present disclosure continuously monitors new operation and maintenance data streams. Once new data is input, the incremental extraction process is triggered, and the pre-constructed rail transit equipment topology knowledge base and fault propagation rule base are queried to check whether the time sequence relationship between the new data and the existing facts is reasonable. For example, whether the cause is earlier than the result and whether the fault propagation time is within a reasonable interval. For device state conflicts, the health degree prediction engine is called. The engine checks whether the conflict data conforms to the known health degree degradation mode of the device. Readings that are more consistent with historical patterns are given higher confidence and retained, while abnormal readings that deviate significantly from the pattern may be considered as noise data and rejected or down-weighted. For fault transmission conflicts, the fault transmission engine and its knowledge base are called. The engine analyzes whether the conflicting fault event is in a reasonable transmission chain, and preferentially adopts evidence consistent with the existing fault transmission chain logic. For maintenance operation conflicts, the maintenance verification engine is called. The engine makes a decision according to the maintenance state machine rules. Data that conforms to the Standard Operating Procedure (SOP) and time logic is preferentially adopted.

[0071] The new triple that passes the verification is finally updated to the graph, realizing the self-evolution of the knowledge graph. Embodiments of the present disclosure ensure that the knowledge graph always maintains consistency in time sequence logic and high domain accuracy during dynamic updating, laying a solid foundation for its reliability in advanced applications such as fault deduction.

[0072] According to an embodiment of this disclosure, the processing method of the rail transit operation and maintenance knowledge graph further includes: parsing the newly generated rail transit operation and maintenance data using a large language model to obtain entity information and scene information; and performing semantic reasoning based on the entity information and scene information using the rail transit operation and maintenance knowledge graph to obtain attribute information associated with the entity information and operation and maintenance solutions associated with the scene information. The semantic reasoning includes temporal rule reasoning, temporal graph neural network reasoning, and temporal causal chain reasoning.

[0073] The embodiments disclosed above utilize a semantic reasoning engine to achieve intelligent analysis of operation and maintenance data, such as predicting equipment failure trends and recommending optimal maintenance solutions. Combined with a rail transit operation and maintenance management system, it provides functions such as intelligent question answering, equipment health value early warning, compliance monitoring and optimization of the maintenance process, and proactive root cause analysis of failures.

[0074] The following describes specific implementation methods in conjunction with appendices. Figure 3 The construction of the rail transit operation and maintenance knowledge graph provided by this invention will be described in further detail.

[0075] Figure 3 This is a diagram illustrating the construction process of a knowledge graph for rail transit operation and maintenance based on a large language model, according to an embodiment of the present invention.

[0076] like Figure 3 As shown, rail transit operation and maintenance text is obtained using rail transit operation and maintenance data tables, and prompt data is extracted using operation and maintenance cases and rail transit operation and maintenance text; the rail transit operation and maintenance data tables and operation and maintenance cases are processed using a large language model to generate a prompt domain concept layer; a knowledge graph is generated through prompt data extraction, RDF data layer definition, and time-series information guidance module. Figure 3 The embodiments of this disclosure shown are based on a large language model for knowledge extraction and knowledge graph construction. However, those skilled in the art can choose other knowledge graph construction methods based on the embodiments of this disclosure. For example, they can use a Transformer-based bidirectional coding model or a deep learning model based on forward LSTM (Long Short-Term Memory) and backward LSTM sequence models for entity recognition and relation extraction; or they can use an expert system for knowledge extraction and relation recognition; or they can use multimodal fusion technology to combine multiple data sources such as images and sensor data to construct an operation and maintenance knowledge graph, which can improve the completeness of the graph; or they can use reinforcement learning to assist knowledge extraction, using reinforcement learning to optimize the knowledge graph construction strategy and improve extraction efficiency.

[0077] The processing method of the rail transit operation and maintenance knowledge graph provided by the present disclosure integrates a time sequence guiding processing mechanism and a dynamic updating mechanism based on operation and maintenance scenarios (such as fault conduction, maintenance verification, and device health prediction), realizes the time sequencing and dynamic updating of the rail transit operation and maintenance knowledge graph, and uses a large language model to solve the problems of inaccurate entity recognition, high redundancy of relationship extraction, long construction and reasoning process time, and the like in the traditional knowledge graph construction process in the construction process of the rail transit operation and maintenance knowledge graph.

[0078] In order to illustrate the advantages and effectiveness of the method provided by the embodiments of the present disclosure, the present disclosure compares the processing method of the knowledge graph of the present disclosure with other knowledge graph construction methods through specific experiments.

[0079] Among them, Table 1 is a test design of a fault operation and maintenance knowledge graph construction prompt word engineering module. In order to show the effect of generating prompt words in the field concept layer in the large language model, tests are conducted on three kinds of large language models respectively, and the model output results are shown.

[0080] Table 1: Field concept layer generation prompt word test table

[0081]

[0082] The A, B and C models can successfully identify and generate core concepts related to rail transit operation and maintenance, and the generated concepts basically cover the classes and attributes required for constructing a knowledge graph.

[0083] In order to further verify the effectiveness of the method proposed by the present disclosure, the present disclosure compares it with some common Chinese natural language processing tools, including entity relationship extraction models, Chinese natural language processing models, pre-training language models, and the like.

[0084]

[0085] Among them, the method provided by the present disclosure performs well in the triple extraction task, with a precision rate and recall rate of 0.94, an F1 value of 0.93, and a running time significantly lower than other deep learning models. Compared with other deep learning-based tools, the large language model scheme based on prompt words (i.e. the method provided by the present disclosure) has lower computational cost and higher processing efficiency, especially in low-code platforms, it has higher operability and responsiveness.

[0086] Figure 4 The structure of the processing device of the rail transit operation and maintenance knowledge graph according to the embodiments of the present disclosure is schematically shown.

[0087] As Figure 4As shown, the processing device 400 of the rail transit operation knowledge graph according to the embodiments of the present disclosure includes an ontology model obtaining module 410, a time-series triple obtaining module 420, and a rail transit operation knowledge graph constructing module 430.

[0088] The ontology model obtaining module 410 is configured to process the rail transit operation historical data to obtain the rail transit operation ontology model in a unified semantic representation.

[0089] The time-series triple obtaining module 420 is configured to perform a knowledge extraction operation and a timestamp identification operation on the unstructured data associated with the rail transit operation historical data to obtain entity-relation-entity triples and timestamps, and perform a binding process on the entity-relation-entity triples and the timestamps to obtain time-series entity-relation-entity triples.

[0090] The rail transit operation knowledge graph constructing module 430 is configured to perform a knowledge fusion operation based on time-series information guidance on the time-series entity-relation-entity triples based on the rail transit operation ontology model and the operation and maintenance scene to obtain the rail transit operation knowledge graph.

[0091] Another aspect of the present disclosure provides a train comprising the processing device of the rail transit operation knowledge graph.

[0092] Any one or more of the modules, sub-modules, units, sub-units according to the embodiments of the present disclosure, or at least part of the functions of any one or more of the modules, sub-modules, units, sub-units can be implemented in one module. Any one or more of the modules, sub-modules, units, sub-units according to the embodiments of the present disclosure can be split into multiple modules for implementation. Any one or more of the modules, sub-modules, units, sub-units according to the embodiments of the present disclosure can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or any other reasonable way of hardware or firmware through integration or packaging of circuits, or in any one of software, hardware, and firmware, or in any appropriate combination of any one or more of them. Alternatively, one or more of the modules, sub-modules, units, sub-units according to the embodiments of the present disclosure can be at least partially implemented as computer program modules that can perform corresponding functions when the computer program modules are run.

[0093] For example, any multiple of the ontology model obtaining module 410, the time-series triple obtaining module 420, and the rail transit operation and maintenance knowledge graph constructing module 430 can be combined in one module / unit / sub-unit for implementation, or any one of the modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of the modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units, and implemented in one module / unit / sub-unit. According to an embodiment of the present disclosure, at least one of the ontology model obtaining module 410, the time-series triple obtaining module 420, and the rail transit operation and maintenance knowledge graph constructing module 430 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on chip, a system on board, a system in package, an application specific integrated circuit (ASIC), or any other reasonable manner of integrating or packaging a circuit, etc. hardware or firmware, or implemented in any one of software, hardware, and firmware or in a proper combination of any of them. Alternatively, at least one of the ontology model obtaining module 410, the time-series triple obtaining module 420, and the rail transit operation and maintenance knowledge graph constructing module 430 can be at least partially implemented as a computer program module that can perform corresponding functions when the computer program module is run.

[0094] It should be noted that the rail transit operation and maintenance knowledge graph processing device part in the embodiments of the present disclosure corresponds to the rail transit operation and maintenance knowledge graph processing method part in the embodiments of the present disclosure, and the description of the rail transit operation and maintenance knowledge graph processing device part is specifically referred to the rail transit operation and maintenance knowledge graph processing method part, which will not be repeated here.

[0095] Figure 5 A block diagram of an electronic device suitable for implementing the above-described method according to an embodiment of the present disclosure is schematically shown. Figure 5 The electronic device shown is only an example and should not bring any limitation to the functions and use range of the embodiments of the present disclosure.

[0096] As Figure 5As shown, the electronic device 500 according to embodiments of the present disclosure includes a processor 501 that can perform various appropriate actions and processes in accordance with programs stored in a read only memory (ROM) 502 or loaded from a storage section 508 into a random access memory (RAM) 503. The processor 501 can include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor, and / or a related chipset, and / or a special purpose microprocessor (e.g., an application specific integrated circuit (ASIC)), and so on. The processor 501 can also include an on-board memory for cache use. The processor 501 can include a single processing unit or multiple processing units for executing different actions of the method processes according to embodiments of the present disclosure.

[0097] In the RAM 503, various programs and data required for the operation of the electronic device 500 are stored. The processor 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. The processor 501 performs various operations of the method processes according to embodiments of the present disclosure by executing the programs in the ROM 502 and / or the RAM 503. Note that the programs can also be stored in one or more memories other than the ROM 502 and the RAM 503. The processor 501 can also perform various operations of the method processes according to embodiments of the present disclosure by executing the programs stored in the one or more memories.

[0098] According to embodiments of the present disclosure, the electronic device 500 can further include an input / output (I / O) interface 505 that is also connected to the bus 504. The electronic device 500 can further include one or more of the following components connected to the input / output (I / O) interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a display such as a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as necessary. A removable recording medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc. is attached to the drive 510 as necessary, so that a computer program read out therefrom is installed in the storage section 508 as necessary.

[0099] According to an embodiment of the present disclosure, the method flow according to the embodiments of the present disclosure can be implemented as a computer software program. For example, the embodiments of the present disclosure include a computer program product including a computer program carried on a computer-readable storage medium, the computer program containing program codes for executing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network through the communication part 509, and / or installed from the detachable medium 511. When the computer program is executed by the processor 501, the above-mentioned functions defined in the system / apparatus according to the embodiments of the present disclosure are executed. According to an embodiment of the present disclosure, the system, device, apparatus, module, unit, etc. described above can be implemented by computer program modules.

[0100] The present disclosure also provides a computer-readable storage medium, which can be included in the device / apparatus / system described in the above embodiments, or can exist separately without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, which, when executed, implement the methods according to the embodiments of the present disclosure.

[0101] According to an embodiment of the present disclosure, the computer-readable storage medium can be a non-volatile computer-readable storage medium. For example, it can include but is not limited to portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any appropriate combination of the above. In the present disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus or device.

[0102] For example, according to an embodiment of the present disclosure, the computer-readable storage medium can include the ROM 502 and / or the RAM 503 described above and / or one or more memories other than the ROM 502 and the RAM 503.

[0103] The embodiments of the present disclosure also include a computer program product including a computer program containing program codes for executing the methods provided by the embodiments of the present disclosure, which are used to make the electronic device implement the processing method of the rail transit operation and maintenance knowledge graph provided by the embodiments of the present disclosure when the computer program product is running on the electronic device.

[0104] When the computer program is executed by the processor 501, the above-mentioned functions defined in the system / apparatus according to the embodiments of the present disclosure are executed. According to an embodiment of the present disclosure, the system, device, apparatus, module, unit, etc. described above can be implemented by computer program modules.

[0105] In one embodiment, the computer program can be embodied on a tangible memory device, such as a magnetic storage device, an optical storage device, etc. In another embodiment, the computer program can be transmitted in a signal over a network, distributed across networks, downloaded and installed, and / or installed from a removable memory media 511. The computer program comprising the program code can be transmitted using any suitable network medium, including, but not limited to wireless, wired, etc., or any suitable combination of the foregoing.

[0106] According to an embodiment of the disclosure, program code for execution by a computer program can be written in any combination of one or more programming languages, and can specifically be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages include, but are not limited to, Java, C++, python, “C”, or similar programming languages. Program code can execute entirely on the user’s computing device, partly on the user’s device, partly on a remote computing device, or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user’s computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computing device, such as through the Internet using an Internet Service Provider.

[0107] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions. Those skilled in the art will understand that the features described in the various embodiments of the present disclosure can be combined and / or combined in various ways, even if such combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments of this disclosure can be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0108] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A method for processing a knowledge graph of rail transit operation and maintenance, characterized in that, include: By processing historical data on rail transit operation and maintenance, a rail transit operation and maintenance ontology model with unified semantic representation is obtained; Knowledge extraction and timestamp identification operations are performed on the unstructured data associated with the historical data of rail transit operation and maintenance to obtain entity-relationship-entity triples and timestamps. The entity-relationship-entity triples and timestamps are then bound together to obtain time-series entity-relationship-entity triples. Based on the rail transit operation and maintenance ontology model and operation and maintenance scenario, a knowledge fusion operation guided by time-series information is performed on the temporal entity-relationship-entity triple to obtain a rail transit operation and maintenance knowledge graph.

2. The method according to claim 1, characterized in that, By processing historical data on rail transit operation and maintenance, a rail transit operation and maintenance ontology model with unified semantic representation is obtained, including: By using a large language model to process the heterogeneous data of various types in the historical data of rail transit operation and maintenance in the form of interactive prompt words, a resource description framework pattern with unified semantic representation is obtained. The domain concepts and ontology structure used to construct the rail transit operation and maintenance knowledge graph are defined using the resource description framework pattern, resulting in the rail transit operation and maintenance ontology model.

3. The method according to claim 1, characterized in that, Based on the aforementioned rail transit operation and maintenance ontology model and operation and maintenance scenario, a knowledge fusion operation guided by time-series information is performed on the temporal entity-relationship-entity triples to obtain a rail transit operation and maintenance knowledge graph, including: The processing results of the historical rail transit operation and maintenance data and the processing results of the unstructured data are processed again using a large language model to obtain a secondary processing result. Based on the secondary processing result, the operation and maintenance scenario is determined by entity type recognition. When the operation and maintenance scenario is a fault scenario, based on the rail transit operation and maintenance ontology model, knowledge retrieval and confidence calculation are performed in the fault diagnosis database using the fault transmission chain guided by time-series information. The knowledge retrieval results and confidence calculation results are then used to fill attributes and fuse knowledge in the time-series entity-relationship-entity triple.

4. The method according to claim 3, characterized in that, Also includes: When the operation and maintenance scenario is a maintenance scenario, the time-series information is verified through the maintenance business chain of fault-maintenance-recovery, and based on the rail transit operation and maintenance ontology model, the verification results are used to perform attribute filling and knowledge fusion on the time-series entity-relationship-entity triple. In the case that the operation and maintenance scenario is an equipment status scenario, the status information of the equipment entity is matched with the health status patterns in the equipment health status pattern library in real time. Based on the rail transit operation and maintenance ontology model, the matching results are used to fill attributes and fuse knowledge in the temporal entity-relationship-entity triples to obtain the rail transit operation and maintenance knowledge graph. The health status patterns include normal status, early warning status, sub-health status and fault status.

5. The method according to claim 1, characterized in that, Also includes: Incremental knowledge extraction is performed on the associated data of the new rail transit operation and maintenance data to obtain the incremental knowledge extraction results; The incremental knowledge extraction results are subjected to conflict verification based on time sequence relationships to obtain conflict verification results, and the conflict verification results are used to dynamically update the rail transit operation and maintenance knowledge graph.

6. The method according to claim 5, characterized in that, The incremental knowledge extraction results are subjected to conflict verification based on temporal relationships to obtain conflict verification results including: Obtain the fault propagation chain, maintenance business chain, and equipment health pattern library associated with the rail transit operation and maintenance knowledge graph; The incremental knowledge extraction result is compared with the fault propagation chain to perform a conflict check based on the time-series fault rules, and a first conflict check result is obtained. The incremental knowledge extraction result is compared with the maintenance business chain to perform a conflict check based on the time sequence of maintenance operations, resulting in a second conflict check result. The incremental knowledge extraction result is compared with the device health mode library to perform a conflict check based on the time-series health mode, resulting in a third conflict check result.

7. The method according to claim 6, characterized in that, Dynamically updating the rail transit operation and maintenance knowledge graph using the conflict verification results includes: If the first conflict verification result indicates that the incremental knowledge extraction result is consistent with the logic of the fault propagation chain, and / or if the second conflict verification result indicates that the incremental knowledge extraction result is consistent with the standard operation and time logic of the maintenance business chain, and / or if the second conflict verification result indicates that the incremental knowledge extraction result is consistent with the standard operation and time logic of the maintenance business chain, a new entity-relationship-entity triple is determined from the incremental knowledge extraction result, and the new entity-relationship-entity triple is used to dynamically update the rail transit operation and maintenance knowledge graph.

8. The method according to claim 1, characterized in that, Also includes: The newly generated rail transit operation and maintenance data is analyzed using a large language model to obtain entity information and scene information; Based on the entity information and the scene information, semantic reasoning is performed using a rail transit operation and maintenance knowledge graph to obtain attribute information associated with the entity information and operation and maintenance solutions associated with the scene information. The semantic reasoning includes temporal rule reasoning, temporal graph neural network reasoning, and temporal causal chain reasoning.

9. A processing device for a knowledge graph of rail transit operation and maintenance, characterized in that, include: The ontology model acquisition module is used to process historical data of rail transit operation and maintenance to obtain a rail transit operation and maintenance ontology model with unified semantic representation. The temporal triple acquisition module is used to perform knowledge extraction and timestamp recognition operations on unstructured data associated with the historical data of rail transit operation and maintenance, to obtain entity-relationship-entity triples and timestamps, and to bind the entity-relationship-entity triples and the timestamps to obtain temporal entity-relationship-entity triples; The rail transit operation and maintenance knowledge graph construction module is used to perform knowledge fusion operations guided by time-series information on the time-series entity-relationship-entity triples based on the rail transit operation and maintenance ontology model and operation and maintenance scenarios, so as to obtain the rail transit operation and maintenance knowledge graph.

10. A train, comprising a processing device for a rail transit operation and maintenance knowledge graph as described in claim 9.

Images