Knowledge-based station digital twin model and system

Abstract

The knowledge-based field station digital twin model comprises a physical entity PS, a virtual model VS, a service set Ss, a data set DD, a data connection CN and knowledge cognition KC, the physical entity PS represents an actual physical object of the station; the virtual model VS comprises a geometric model Gv, a physical model Pv, a behavior model Bv, a rule model Rv and a cognitive model Cv; the service set Ss is a service set required by station business management and comprises functional services and business services; the data connection CN realizes interconnection and intercommunication among physical entities, virtual models, data and services; the knowledge cognition KC is a cognition function component based on information technology and knowledge data integration. The invention discloses a knowledge-based station digital twin system, which comprises a twin space layer, a data interaction layer and an application service layer. The method has the advantages that the requirement for advanced cognition in the production process can be met, operation can be optimized, and good practicability and efficiency are achieved.

Description

Technical Field

[0001] The present invention relates to the field of digital technology, and in particular to a knowledge-based digital twin model and system for a station. Background Art

[0002] Digital twins (DTs) are virtual models of physical entities created digitally. Using data to simulate the physical entity's behavior in the real world, they add or expand new capabilities to the physical entity through interactive feedback between the virtual and the real, data fusion analysis, and iterative decision optimization. Their development is driving the digitization, networking, and intelligence of human production activities. They provide the technical support needed to implement "industry mirroring digital solutions" across the entire process, business, and lifecycle in many industries, including manufacturing. In recent years, the knowledge-driven manufacturing paradigm has been widely researched. This manufacturing approach emphasizes the reuse of knowledge across design, process, management, and decision-making, driving the development of manufacturing by optimizing resource allocation through knowledge. A recent trend in digital twin research is combining digital twins with advanced knowledge modeling techniques to enhance their cognitive capabilities.

[0003] In the existing technology, WL et al. believe that digital twins are essentially carriers of orderly knowledge flow and interaction, and they play an important role in Industry 4.0 scenarios as high-quality knowledge extractors, high-density knowledge storage, and efficient knowledge interpreters. Gómez-Berbís et al. proposed a semantic digital twin framework based on IoT data management and knowledge graphs. Zhou et al. introduced a knowledge-driven digital twin manufacturing cell (KDTMC) framework for intelligent manufacturing, focusing on three key technologies (including digital twin models, dynamic knowledge bases, and knowledge-based intelligent skills) to achieve intelligent perception, simulation, understanding, prediction, optimization, and control strategies to support autonomous manufacturing. Akroyd et al. introduced a dynamic knowledge graph method that supports cross-domain interoperability and ensures data connectivity, portability, discoverability, and queryability through a unified interface. Wang et al. introduced a knowledge graph-based workshop multi-domain model integration method. By constructing a knowledge graph, they achieved multi-domain model integration for design, manufacturing, and simulation, supporting dynamic simulation and virtual-reality mapping. Su et al. introduced a knowledge-driven method for creating digital models. By integrating a three-layer model of knowledge, decision-making, and geometry, this method provides structured and semantic support. This method ensures semantic consistency between models and implements knowledge-driven and continuously updated closed-loop interaction of models, promoting self-cognition and self-update of the models. M. et al. constructed an Actionable Cognitive Twin (ACT) for manufacturing decision-making based on knowledge graph modeling. ACT aims to provide insights and decision support for complex systems by combining knowledge graphs and simulation models. While these studies have demonstrated the feasibility of using semantic technologies to enhance the cognitive capabilities of digital twins, they cannot meet the requirements for advanced process cognition in modern industrial systems, particularly in terms of practicality and efficiency in optimizing operational and maintenance decisions. Summary of the Invention

[0004] The present invention overcomes the above shortcomings and proposes a knowledge-based site digital twin model and system that can meet the needs of advanced cognition of the generation process, optimize operations, and has good practicality and efficiency.

[0005] The knowledge-based digital twin model of the station of the present invention, wherein: the digital twin model of the station M Gas It includes physical entity PS, virtual model VS, service set Ss, data set DD, data connection CN, and knowledge cognition KC, which can be expressed as:

[0006] M Gas =(PS,VS,Ss,DD,CN,KC)

[0007] The physical entity PS represents the actual physical objects of the station, including pipelines, control equipment, monitoring equipment, communication equipment, and electrical equipment. The pipelines include external transmission pipelines, venting pipelines, and sewage pipelines. Physical entities are managed hierarchically, and each physical entity can be decomposed into several subsystems or subcomponents.

[0008] The virtual model VS includes a geometric model G v , physical model P v , Behavioral Model B v , rule model R v , cognitive model C v , that is, VE=(G v , P v , B v , R v , C v ); physical model P v The physical behavior and physical properties of the physical entities in the station, including natural gas flow, pipeline pressure changes, and temperature distribution; behavioral model B v Including equipment start and stop, action adjustment, and safety cutoff; rule model R v The rules and constraint logic of the entire station system, including process operation rules and constraint logic between skid-mounted partitions; cognitive model C v System cognition mechanisms and service agent behaviors based on other model inputs, including station status cognition and event service response;

[0009] The service set Ss is the service set required for station business management, including functional services and business services; the functional services include automated monitoring, safety and risk management, scheduling and production optimization; the business services include situation monitoring services, operation and maintenance decision-making services, and collaborative control services;

[0010] The data set DD includes physical data D p , virtual data D v , service data D s , knowledge data D k , fusion data D f , cognitive data D c , that is, DD=(D p ,D v ,D s ,D k ,D f ,D c ,), the physical data D p The static data and dynamic data of the physical entity include the physical entity's geometric structure, assembly relationship, physical characteristics, load constraint data, and the dynamic data of the physical entity include temperature, pressure, and flow; the virtual data D v Including geometric, physical, behavioral and rule model data of the virtual model, real-time simulation, prediction and evaluation data based on the virtual model; the service data D s The data for system operation and data for business needs; the knowledge data D k It is various information data such as station service procedures, rules and constraints, and expert knowledge; the fused data D f D P 、D v 、D s and D k The derived data generated by data fusion includes the output after preprocessing, classification and integration of multidimensional data; the cognitive data D c To associate data when implementing cognitive ability processes, including event, process and result data;

[0011] The data connection CN enables interconnection and intercommunication between physical entities, virtual models, data and services;

[0012] The knowledge cognition KC is a cognitive functional component based on the integration of information technology and knowledge data, and consists of three parts: integration of knowledge data information of all dimensions of the station, knowledge cognition based on the integrated station knowledge data information, and knowledge service application based on the cognitive results.

[0013] The above-mentioned knowledge-based site digital twin model, wherein: the knowledge cognition KC is a cognitive functional component based on the integration of information technology and knowledge data, and its construction process includes: constructing a multi-scale knowledge model and constructing a knowledge-based service; the construction of the multi-scale knowledge model includes demand boundary specification, development of ontology, and organization of services based on the ontology to form a multi-scale knowledge model.

[0014] In the above-mentioned knowledge-based digital twin model of a site, the requirement boundary specification includes clarifying the core functions of cognitive function components, service delivery objectives, and optimization requirements. The specific steps are as follows:

[0015] (a) Identify service procedures: In-depth analysis of the types and characteristics of service activities in the system;

[0016] (b) Determine functional components: Identify the components required to support service execution according to the service specifications;

[0017] (c) Determine boundary conditions: Define the boundary conditions and dependencies inside and outside the system to ensure subsequent service delivery goals and optimization requirements.

[0018] The above-mentioned knowledge-based site digital twin model, wherein: the services are organized based on ontology to form a multi-scale knowledge model, which includes a macro model, a meso model, and a micro model; the macro model includes functions, goals, and service processes; the meso model includes components, engines, entity engines, and entities; the micro model includes service data, data flows, and rules.

[0019] The above-mentioned knowledge-based digital twin model of the station, wherein: the specific implementation steps of constructing the multi-scale knowledge model are:

[0020] (a) Manually process the semantics of given information, extract knowledge data of all dimensions of the station through machine learning, and store the obtained knowledge data in the Neo4j database after processing;

[0021] (b) Use the large language model (LLM) to learn the source data and results of the knowledge integration process in the previous step to form an intelligent agent with real-time knowledge recognition capabilities.

[0022] A knowledge-based digital twin system for a site, comprising: a twin space layer, a data interaction layer, and an application service layer;

[0023] The twin space layer provides the interactive evolution process of virtual and real stations. According to the hierarchical division of the physical entities of the digital twin model of the station, a high-fidelity digital station is established using 3D modeling software and the UE engine. The physical properties, behavioral properties and logical relationships of each part of the model are edited and set through the UE engine.

[0024] The data interaction layer uses the twin space layer data and pre-integrated knowledge data as the driving force for interaction between layers. Based on the station digital twin model, sensors are used to collect station data in real time and transmit it to the system. After pre-processing, the data is connected to each data interface.

[0025] The application service layer encapsulates the service set in the digital twin model of the station, and provides functions including station visualization virtual model, station equipment status monitoring, equipment information query, remote control, and security policy decision-making; at the same time, it embeds a human-computer interaction collaboration mechanism in the system interaction logic, integrates artificial intelligence and the composite decision-making of human experts, realizes reliable driving of the service engine, and provides the required functional services for station operation management tasks.

[0026] Compared to existing technologies, the present invention offers significant advantages. As can be seen from the above scheme, the present invention proposes a knowledge-enhanced six-dimensional digital twin model to comprehensively describe all aspects of the digital twin. Essentially, it leverages technologies such as knowledge graphs and artificial intelligence to enable more comprehensive, in-depth, and intelligent digital modeling and simulation of entities, processes, or systems in the physical world. This approach focuses not only on the precise mapping of physical entities in digital space but also on the integration and utilization of knowledge about their operational patterns, relationship networks, and historical experience to achieve higher-level decision support, predictive optimization, and intelligent control. The physical entity (PE) in the model is the actual physical object, while the virtual entity (VE) is the digital representation of the physical entity (PE). The physical and virtual entities (VE) interact bidirectionally through data collection and feedback mechanisms. The data layer (DD) is the core driver of the digital twin model, acting as a bridge between the physical and virtual entities (PE). By facilitating data collection, transmission, storage, and processing, the data layer (DD) enables real-time interaction and synchronization between the physical and virtual entities (PE) and VE, and provides data support for the knowledge computing layer (KC). The service layer (Ss), as the output component of the digital twin model, encapsulates the model's functionality into operational services to meet the needs of different users. The service layer (Ss) interacts with the physical entities (PE), virtual entities (VE), data layer (DD), and knowledge computing layer (KC) through the connection layer (CN), enabling data input and output. As the connection hub of the digital twin model, the connection layer (CN) enables bidirectional interaction between the physical entities (PE) and data layer (DD), between the physical entities (PE) and virtual entities (VE), between the physical entities (PE) and service layer (Ss), between virtual entities (VE) and data layer (DD), between virtual entities (VE) and service layer (Ss), between the service layer (Ss) and data layer (DD), and between the service layer (Ss) and knowledge computing layer (KC) through various protocols and interfaces, including data acquisition, data transmission, protocol standards, and interface services. In practical applications, the knowledge computing layer (KC) processes the data layer (DD), provides advanced decision support for the service layer (Ss), and enables real-time interaction with physical entities (PE) and virtual entities (VE) through the connection layer (CN). The knowledge computing layer (KC) can be considered a cognitive functional component, a cognitive technology system based on new information technology and knowledge data integration.

[0027] The framework of the digital twin system was designed based on the model hierarchy, and the human-computer interaction mechanism within the system based on the domain model was clearly defined. The framework consists of a twin space layer, a data interaction layer, and an application service layer. The twin space layer is a critical component of the digital twin system, serving as the underlying digital environment for deep interaction and real-time mapping between physical and virtual scenes. As the middle layer of the system framework, the data interaction layer integrates the full-dimensional data backplane provided by the twin space layer. Through efficient data transmission protocols and processing algorithms, it connects data across all dimensions of the system, supporting data analysis, simulation, and cognitive computing within the system. The application service layer provides a key business operation and management engine for system operation, including specific functional services or service core components. A human-computer interaction collaborative mechanism is embedded in the system interaction logic. Essentially, it is a hybrid decision-making approach that integrates artificial intelligence and human experts. Its core purpose is to leverage the unique advantages of both humans and machines in data processing and problem-solving, achieving reliable driving of the service engine. To improve the operability of the service, a service-based multi-scale knowledge model is further proposed. In summary, the present invention meets the requirements of advanced cognition in the generation process, optimizes operations, and possesses excellent practicality and efficiency.

[0028] The beneficial effects of the present invention are further illustrated below through specific implementation methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a flow chart of the present invention;

[0030] Figure 2 A schematic diagram of the evaluation results of the device failure scenario experiment in a specific embodiment;

[0031] Figure 3 Schematic diagram of the evaluation results of the security incident scenario experiment in a specific implementation method. DETAILED DESCRIPTION

[0032] The following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation methods, features and functions of a knowledge-based station digital twin model and system proposed in accordance with the present invention.

[0033] like Figure 1 As shown in the figure, a knowledge-based digital twin model of a station is constructed, where: the digital twin model M Gas It includes physical entity PS, virtual model VS, service set Ss, data set DD, data connection CN, and knowledge cognition KC, which can be expressed as:

[0034] M Gas =(PS,VS,Ss,DD,CN,KC)

[0035] The physical entity PS represents the actual physical objects of the station, including pipelines, control equipment, monitoring equipment, communication equipment, and electrical equipment; the pipelines include external transmission pipelines, venting pipelines, and sewage pipelines; physical entities are managed hierarchically, and each physical entity can be decomposed into several subsystems or subcomponents; the hierarchical management of physical entities is divided into unit level, system level, and complex system level. These levels can correspond to equipment and systems of different granularities, and characterize the multi-granularity and multi-level physical entity characteristics of the station.

[0036] The virtual model VS includes a geometric model G v , physical model P v , Behavioral Model B v , rule model R v and cognitive model C v , that is, VE=(G v , P v , B v , R v , C v ); physical model P v The physical behavior and physical properties of the physical entities in the station, including natural gas flow, pipeline pressure changes, and temperature distribution; behavioral model B v It is the relationship, movement, control and response mechanism between entity components in the station, including equipment start and stop, action adjustment, and safety cutoff; rule model R v The rules and constraint logic of the entire station system, including process operation rules and constraint logic between skid-mounted partitions; cognitive model C v This includes system recognition mechanisms and service agent behaviors based on input from other models, including station status recognition and event service response. For example, it can predict and identify potential equipment failures or efficiency degradation, and trigger maintenance or adjustment measures within the rules.

[0037] The service set Ss is the service set required for station business management, including functional services (supporting the internal functional operation of the digital twin system) and business services (meeting the operation and decision-making needs of external users); the functional services include automated monitoring, safety and risk management, scheduling and production optimization, and the business services include situation monitoring services, operation and maintenance decision-making services, and collaborative control services;

[0038] The data set DD includes physical data D p , virtual data D v , service data D s , knowledge data D k , fusion data D f , cognitive data D c , that is, DD=(D p ,D v ,Ds ,D k ,D f ,D c ,), the physical data D p The static data and dynamic data of the physical entity include the physical entity's geometric structure, assembly relationship, physical characteristics, load constraint data, and the dynamic data of the physical entity include temperature, pressure, and flow; the virtual data D v Including geometric, physical, behavioral and rule model data of the virtual model, real-time simulation, prediction and evaluation data based on the virtual model; the service data D s The data for system operation and business needs (such as station management, equipment maintenance, production planning, resource scheduling); the knowledge data D k It is various information data such as station service procedures, rules and constraints, and expert knowledge; the fused data D f D P 、D v 、D s and D k The derived data generated by data fusion includes the output after preprocessing, classification and integration of multidimensional data; the cognitive data D c To associate data when implementing cognitive capabilities, including event, process and result data.

[0039] The data connection (CN) enables interconnection between physical entities, virtual models, data, and services. Various sensors collect real-time data from the field, which is then transmitted to the data collection (DD) via a network that adheres to the same protocol standards. Processed data or instructions are then transmitted using the same protocol specifications and fed back to each module to drive system operation.

[0040] The knowledge cognition KC is a cognitive function component based on information technology and knowledge data integration, which consists of three parts: the integration of knowledge data information of all dimensions of the station. i , carry out knowledge recognition based on integrated station knowledge data information c , knowledge service application based on cognitive results K a， Right now K c , K a KC uses advanced artificial intelligence technology to allow the system to simulate human expert decision-making, with the goal of building a functional dimension to act as the system brain and then developing digital twin intelligent service capabilities. i， K c， K a The three interdependent parts form a continuous intelligent knowledge enhancement process.

[0041] The knowledge cognition KC is a cognitive function component based on the integration of information technology and knowledge data. Its construction process includes: building a multi-scale knowledge model and building a knowledge-based service;

[0042] The construction of the multi-scale knowledge model includes requirement boundary specification and ontology development. The requirement boundary specification includes clarifying the core functions of cognitive function components, service delivery goals, and optimization requirements to define boundaries. The specific steps are as follows:

[0043] a. Identify service procedures: In-depth analysis of the types and characteristics of service activities in the system.

[0044] b. Identify functional components: Identify the components required to support service execution based on the service specifications.

[0045] c. Determine boundary conditions: Define the boundary conditions and dependencies inside and outside the system to ensure subsequent service delivery goals and optimization requirements.

[0046] The ontology development method is as follows: constructing a service-oriented multi-dimensional knowledge model top-level ontology as shown in the following formula. Table 1 is the definition of the ontology.

[0047]

[0048] Table 1 Definition of ontology

[0049]

[0050] By reorganizing services based on the top-level ontology, it evolves into a service-oriented multi-scale knowledge model.

[0051] Knowledge model = {macro model, meso model, micro model}

[0052] The model consists of three sub-models at the macro, meso and micro scales, each of which describes a different emphasis on each dimension from its own scale.

[0053] Macro model = {function, goal, service process}

[0054] Meso-model = {component, engine, entity engine, entity}

[0055] Micro model = {service data, data flow, rules}

[0056] Based on the requirement boundary specification and ontology, the specific implementation steps of building the multi-scale knowledge model are as follows:

[0057] (a) Manually process the semantics of given information, extract knowledge data of all dimensions of the station through machine learning, and store the obtained knowledge data in the Neo4j database after processing.

[0058] (b) Use the large language model (LLM) to learn the source data and results of the knowledge integration process in the previous step to form an intelligent agent with real-time knowledge recognition capabilities.

[0059] The construction of knowledge-based services: Based on the results of the previous step, the site is supported to dynamically implement knowledge push and user question-and-answer services. The specific steps are as follows:

[0060] (a) Migrate the diagnostic algorithm model from the supervisory control and data acquisition (SCADA) system to the station digital twin system.

[0061] (b) The historical monthly daily report data of the station equipment is used as the data source, and a data distribution script is constructed to transmit the data in JSON format to the diagnostic algorithm model in real time.

[0062] (c) Execute “knowledge recognition” services based on the diagnostic results of the model.

[0063] Knowledge push refers to the system's ability to autonomously make decisions based on real-time data and push them to users. This includes decisions on equipment failure diagnosis and equipment operation optimization. User question-and-answer refers to the system's ability to perform tasks based on user input, including knowledge retrieval and decision-making based on questions.

[0064] A knowledge-based digital twin system for a site, comprising: a twin space layer, a data interaction layer, and an application service layer;

[0065] The twin space layer provides the interactive evolution process of virtual and real stations. According to the hierarchical division of the physical entities of the digital twin model of the station, a high-fidelity digital station is established using three-dimensional modeling software and the UE engine. The physical properties, behavioral properties and logical relationships of each part of the model are edited and set through the UE engine.

[0066] The data interaction layer uses the full-dimensional data of the twin space and the pre-integrated knowledge data as the driving force for interaction between layers. The station data collected by the sensors in real time is transmitted to the system through communication methods such as optical fiber, and is processed with the help of its mature data capabilities before being connected to various data interfaces.

[0067] The application service layer encapsulates the functional services required for station operation and management tasks, providing system functions such as displaying and visualizing precisely mapped digital physical station scenes, monitoring station equipment status, querying equipment information, remote control, and security policy decision-making. Furthermore, a human-computer interaction and collaborative mechanism is embedded in the system's interaction logic, integrating the combined decision-making of artificial intelligence and human experts to reliably drive the service engine.

[0068] Effect analysis

[0069] To verify the effectiveness of this invention, we collected historical data on real equipment failures and safety incidents at an unmanned natural gas station with the assistance of experts. This data served as the foundational data source for the experiment and was integrated into various systems using simulation scripts. Based on this data, the experimental scenarios shown in Table 2 were constructed under the guidance of experts.

[0070] Table 2 Experimental scenarios

[0071]

[0072] Four key indicators based on the historical performance of the Supervisory Control and Data Acquisition (SCADA) system in equipment management and emergency response to safety incidents were designed: (1) Prediction accuracy: measures the accuracy of the system in predicting equipment failures and safety incidents; (2) Diagnosis accuracy: evaluates the accuracy of the system in analyzing the root cause of the problem; (3) Response time: records the time interval from the discovery of the problem to the start of problem processing, which is used to evaluate the efficiency of task execution; (4) Solution scoring: based on expert evaluation, evaluates the rationality and effectiveness of the solutions provided in the experimental scenario. Five station staff with practical experience were invited to conduct the experiment, and the average performance of the experimental staff in each indicator under different scenarios was calculated. The results are as follows: Figure 2 and Figure 3 As shown, the performance of different systems (digital twin (DT) and actionable knowledge augmented digital twin (AKEDT) of the present invention) in terms of the above indicators is further evaluated.

[0073] The experimental results show that the AKEDT of the present invention has the following outstanding properties:

[0074] In a single scenario, the system demonstrated the following: (1) Higher prediction and diagnostic accuracy: AKEDT was able to identify equipment failures and safety risks earlier and more accurately. (2) Shorter response time: AKEDT leveraged knowledge reasoning to quickly generate decisions, significantly shortening response time and demonstrating faster service delivery capabilities. However, it should be emphasized that there is still room for improvement in AKEDT's decision-making process. (3) Better solutions: AKEDT's higher solution scores indicate a stronger service delivery capability when handling complex incidents.

[0075] In complex scenarios, the system demonstrates: (1) Better prediction and diagnosis accuracy: When dealing with complex scenarios, such as simultaneous failures of pressure sensors (PT4101) and temperature sensors (TT6102), and simultaneous equipment overload and pipeline leakage, AKEDT has higher prediction and diagnosis accuracy than traditional methods. This shows that AKEDT is more capable of identifying and diagnosing multiple faults. Although the accuracy may be slightly lower than that of a single scenario, AKEDT still maintains high performance under complex conditions. (2) Shorter response time: Although the response time of AKEDT in complex scenarios is slightly longer than that of a single scenario, it is still significantly shorter than that of traditional methods. This shows that AKEDT can quickly use knowledge reasoning results to generate decisions and maintain fast service delivery capabilities even in the face of multiple faults and risks. The slight increase in response time may be because the system needs to process more data and complex logic. (3) Higher quality solutions: In complex scenarios, the solution effectiveness score provided by AKEDT is still higher than that of traditional methods, although it may be slightly lower than that of a single scenario. This shows that AKEDT can still provide relatively effective solutions when dealing with multiple complex events, demonstrating its deep service delivery capabilities.

[0076] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A knowledge-based digital twin model of a station, characterized by: Station digital twin model M Gas It includes physical entity PS, virtual model VS, service set Ss, data set DD, data connection CN, and knowledge cognition KC, which can be expressed as: M Gas =(PS,VS,Ss,DD,CN,KC) The physical entity PS represents the actual physical objects of the station, including pipelines, control equipment, monitoring equipment, communication equipment, and electrical equipment. The pipelines include external transmission pipelines, venting pipelines, and sewage pipelines. Physical entities are managed hierarchically, and each physical entity can be decomposed into several sub-physical entities. The virtual model VS includes a geometric model G v , physical model P v , Behavioral Model B v , rule model R v , cognitive model C v , that is, VE=(G v , P v , B v , R v , C v ); physical model P v The physical behavior and physical properties of the physical entities in the station, including natural gas flow, pipeline pressure changes, and temperature distribution; behavioral model B v Including equipment start and stop, action adjustment, and safety cutoff; rule model R v The rules and constraint logic of the entire station system, including process operation rules and constraint logic between skid-mounted partitions; cognitive model C v System cognition mechanisms and service agent behaviors based on other model inputs, including station status cognition and event service response; The service set Ss is the service set required for station business management, including functional services and business services; the functional services include automated monitoring, safety and risk management, scheduling and production optimization; the business services include situation monitoring services, operation and maintenance decision-making services, and collaborative control services; The data set DD includes physical data D p , virtual data D v , service data D s , knowledge data D k , fusion data D f , cognitive data D c , that is, DD=(D p ,D v ,D s ,D k ,D f ,D c ,), the physical data D p The static data and dynamic data of the physical entity include the physical entity's geometric structure, assembly relationship, physical characteristics, load constraint data, and the dynamic data of the physical entity include temperature, pressure, and flow; the virtual data D v Including geometric, physical, behavioral and rule model data of the virtual model, real-time simulation, prediction and evaluation data based on the virtual model; the service data D s The data for system operation and data for business needs; the knowledge data D k It is various information data such as station service procedures, rules and constraints, and expert knowledge; the fused data D f D P 、D v 、D s and D k Derivative data generated by data fusion, including the output after preprocessing, classification, and integration of multidimensional data; The cognitive data D c To link data when implementing cognitive ability processes; The data connection CN enables interconnection and intercommunication between physical entities, virtual models, data and services; The knowledge cognition KC is a cognitive functional component based on the integration of information technology and knowledge data, and consists of three parts: integration of knowledge data information of all dimensions of the station, knowledge cognition based on the integrated station knowledge data information, and knowledge service application based on the cognitive results.

2. The knowledge-based station digital twin model according to claim 1, characterized in that: The knowledge cognition KC is a cognitive functional component based on the integration of information technology and knowledge data. Its construction process includes: building a multi-scale knowledge model and building knowledge-based service identification; the construction of the multi-scale knowledge model includes requirement boundary specification, developing ontology, organizing services based on the ontology, and forming a multi-scale knowledge model.

3. The knowledge-based station digital twin model according to claim 2, characterized in that: The requirement boundary specification includes clarifying the core functions of cognitive function components, service delivery goals, and optimization requirements to define boundaries. The specific steps are as follows: (a) Identify service procedures: In-depth analysis of the types and characteristics of service activities in the system; (b) Determine functional components: Identify the components required to support service execution according to the service specifications; (c) Determine boundary conditions: Define the boundary conditions and dependencies inside and outside the system to ensure subsequent service delivery goals and optimization requirements.

4. The knowledge-based station digital twin model according to claim 2, characterized in that: The services are organized based on ontology to form a multi-scale knowledge model, which includes a macro model, a meso model, and a micro model; the macro model includes functions, goals, and service processes; the meso model includes components, engines, entity engines, and entities; the micro model includes service data, data flows, and rules.

5. The knowledge-based station digital twin model according to claim 2, characterized in that: The specific implementation steps of constructing the multi-scale knowledge model are: (a) Manually process the semantics of given information, extract knowledge data of all dimensions of the station through machine learning, and store the obtained knowledge data in the Neo4j database after processing; (b) Use the large language model (LLM) to learn the source data and results of the knowledge integration process in the previous step to form an intelligent agent with real-time knowledge recognition capabilities.

6. The knowledge-based station digital twin model system according to any one of claims 1 to 5, characterized in that: Includes twin space layer, data interaction layer, and application service layer; The twin space layer provides the interactive evolution process of virtual and real stations. According to the hierarchical division of the physical entities of the digital twin model of the station, a high-fidelity digital station is established using 3D modeling software and the UE engine. The physical properties, behavioral properties and logical relationships of each part of the model are edited and set through the UE engine. The data interaction layer uses the twin space layer data and pre-integrated knowledge data as the driving force for interaction between layers. Based on the digital twin model of the station, sensors are used to collect station data in real time and transmit it to the system. After pre-processing, the data is connected to each data interface. The application service layer encapsulates the service set in the digital twin model of the station, and provides functions including station visualization virtual model, station equipment status monitoring, equipment information query, remote control, and security policy decision-making; at the same time, it embeds a human-computer interaction collaboration mechanism in the system interaction logic, integrates artificial intelligence and the composite decision-making of human experts, realizes reliable driving of the service engine, and provides the required functional services for station operation management tasks.

Images