Airborne system model modeling rule making method, device, equipment and medium

Abstract

The invention relates to the technical field of airborne system design, and discloses an airborne system model modeling rule making method and device, equipment and a medium, which can deeply fuse airborne system design multi-source heterogeneous knowledge, automatically generate executable modeling rules, improve airborne system design efficiency and model quality, and improve airborne system design efficiency and model quality. The problem that modeling rules are inconsistent due to the fact that traditional modeling rules depend on artificial experience is solved, and standardization, consistency and quality control efficiency of airborne system design are improved.

Description

Technical Field

[0001] This invention relates to the field of airborne system design technology, and discloses a method, apparatus, equipment and medium for formulating modeling rules for airborne system models. Background Technology

[0002] As aviation equipment develops towards greater integration, complexity, and intelligence, model-based systems engineering has become the mainstream paradigm for aviation equipment system design. The quality of the digital model of an airborne system, as the carrier of design ideas, directly determines the performance, safety, and reliability of the final product. However, in current engineering practice, the establishment of digital models for airborne systems heavily relies on the personal experience and subjective understanding of designers, leading to inconsistent modeling rules, styles, and quality across different projects and even different subsystems within the same project. This approach results in models with poor readability, low reusability, and difficulties in verification, becoming a core bottleneck in improving the efficiency and quality of airborne system design.

[0003] Currently, the formulation of airborne digital modeling rules mainly relies on domain experts manually reviewing documents and summarizing past experience. This method has significant limitations: First, the specification formulation process is time-consuming and labor-intensive, and it is difficult to fully cover the massive design requirements, professional standards, and tacit knowledge accumulated from historical projects, making it prone to omissions and deviations; second, the generated specifications are mostly static texts, which cannot be used for real-time, automated compliance checks with dynamically evolving design models; third, there is a lack of ability to automatically transform top-level specifications into executable verification methods (such as test cases), making the implementation of specifications heavily dependent on the designer's self-awareness and secondary interpretation, resulting in lagging quality control and high costs. Summary of the Invention

[0004] The purpose of this invention is to provide a method, apparatus, equipment, and medium for formulating modeling rules for airborne systems, which can effectively solve the problem of inconsistent modeling rules caused by traditional reliance on human experience, and improve the standardization, consistency, and quality control efficiency of airborne system design.

[0005] To achieve the above-mentioned technical effects, the technical solution adopted by the present invention is as follows: A method for formulating modeling rules for airborne systems includes: A knowledge graph for the airborne systems domain is constructed based on a multi-source knowledge base; the multi-source knowledge base includes design requirements, standards and specifications, and historical cases. The natural language processing model is trained using the knowledge graph of the airborne system domain. The trained natural language processing model is then used to parse the design requirement document of the target design object in the airborne system, identify and extract the entities corresponding to the key design elements of the target design object and the relationships between the entities, and form a complete set of key design elements. The modeling patterns of historical design digital models in the historical cases are learned using a machine learning model, and combined with the set of key design elements, a rule engine is used to generate digital modeling rules for the target design object; the digital modeling rules include modeling syntax rules, architecture style rules, and naming rules.

[0006] Furthermore, it also includes: integrating the digital model modeling rules into the modeling environment; using the model checking plugin of the modeling environment to perform static verification of the model structure of the digital model to be verified of the target design object; simultaneously, the test case generator of the modeling environment generates test cases based on the digital model modeling rules and performs dynamic behavior verification of the digital model to be verified; recording the verification results and feeding them back to the multi-source knowledge base, so as to optimize the natural language processing model and the rule engine through incremental learning algorithms and update the digital model modeling rules of the target design object.

[0007] Furthermore, methods for constructing knowledge graphs in the airborne systems domain include: An ontology modeling tool is used to construct an ontology model for the airborne system domain, and the core classes of the ontology model and the relationships between the core classes are defined. A natural language processing model is used to identify entities belonging to each of the core classes from text data in a multi-source knowledge base, and the relationships between the entities are extracted by combining the relationships between the core classes. Based on the airborne system domain ontology model and the identified entities and relationships between them, a knowledge graph for the airborne system domain is constructed using knowledge graph tools according to the semantic association of "requirements-functions-performance-components-constraints".

[0008] Furthermore, the key design elements include functional units, performance parameters, interface units, and constraint indicators.

[0009] Furthermore, the method of using a machine learning model to learn the modeling patterns of historical design digital models in the historical cases, and combining them with the set of key design elements, to generate digital modeling rules for the target design object through a rule engine includes: The K-means clustering algorithm is used to analyze the modeling patterns of the historical design digital model and identify the architectural patterns and anti-patterns of the target design object. The Apriori association rule algorithm is used to analyze the dependencies between the key design elements to obtain the modeling constraints of the target design object. Based on the aforementioned architectural patterns and anti-patterns, and the aforementioned modeling constraints, the Drools rule engine generates digital modeling rules for the target design object and outputs them as machine-readable specification files based on XMLSchema.

[0010] Furthermore, the standard specifications include digital modeling rules, design standards, and airworthiness requirements for the target design object; the historical cases also include historical failure cases of the target design object.

[0011] An apparatus for formulating modeling rules for airborne system digital models, implemented based on any of the aforementioned methods for formulating modeling rules for airborne system models, includes: The multi-source knowledge base and knowledge graph construction module is used to construct a knowledge graph for the airborne systems domain based on a multi-source knowledge base in the airborne systems domain; the multi-source knowledge base includes design requirements, standards and specifications, and historical cases; The key design element extraction module is used to train the natural language processing model using the knowledge graph of the airborne system domain, and use the trained natural language processing model to parse the design requirement document of the target design object in the airborne system, identify and extract the entities corresponding to the key design elements of the target design object and the relationships between the entities, and form a complete set of key design elements. The modeling rule generation module is used to learn the modeling patterns of historical design digital models in the historical cases using a machine learning model, and combine them with the set of key design elements to generate digital modeling rules for the target design object through a rule engine; the digital modeling rules include modeling syntax rules, architecture style rules, and naming rules.

[0012] Furthermore, it also includes: a verification execution and closed-loop update module, used to integrate the digital model modeling rules into the modeling environment, use the model checking plugin of the modeling environment to perform static verification of the model structure of the digital model to be verified of the target design object, and at the same time, the test case generator of the modeling environment generates test cases based on the digital model modeling rules and performs dynamic behavior verification of the digital model to be verified; record the verification results and feed them back to the multi-source knowledge base, so as to optimize the natural language processing model and the rule engine through incremental learning algorithm and update the digital model modeling rules of the target design object.

[0013] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the airborne system modeling rule formulation method described in any of the preceding claims.

[0014] A computer-readable storage medium storing a computer program that performs the airborne system modeling rule formulation method described in any of the preceding claims.

[0015] Compared with the prior art, the beneficial effects of this invention are: This invention can deeply integrate multi-source heterogeneous knowledge of airborne system design, automatically generate executable modeling rules, improve the efficiency and quality of airborne system design, solve the problem of inconsistent modeling rules caused by traditional reliance on human experience, and improve the standardization, consistency and quality control efficiency of airborne system design. Attached Figure Description

[0016] Figure 1 The flowchart is shown below for the method of formulating modeling rules for airborne system models in Example 1. Figure 2 A structural block diagram of the device for formulating modeling rules for the airborne system digital model in Example 1; Figure 3 This is a structural block diagram of the computer device in Example 1; Wherein, 1-memory, 2-processor. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the embodiments and accompanying drawings. However, this should not be construed as limiting the scope of the above-described subject matter of the present invention to the following embodiments; all technologies implemented based on the content of the present invention fall within the scope of the present invention.

[0018] Example 1 See Figure 1 A method for formulating modeling rules for airborne systems includes: Step 1: Construct a knowledge graph for the airborne systems domain based on a multi-source knowledge base; the multi-source knowledge base includes design requirements, standards and specifications, and historical cases.

[0019] Specifically, firstly, a multi-source knowledge base is constructed based on heterogeneous data from multiple sources, including design requirements, standards and specifications, and historical cases of the target design object in the airborne system. The standards and specifications include digital modeling rules, design standards, and airworthiness requirements for the target design object. The historical cases include historical digital models of the target design object and historical failure cases.

[0020] Secondly, an ontology modeling tool is used to construct an ontology model of the airborne system domain, defining the core classes of the ontology model and the relationships between them. The ontology modeling tool can be software such as Protégé. The core classes of the target design object ontology model include requirement classes, functional classes, performance classes, component classes, and constraint classes. Specifically: requirement classes include functional and non-functional requirements, for example, requirement class instance: Req1 (functional requirement): "The aircraft must be able to calculate its real-time position"; functional classes include navigation and display functions, for example, functional class instance: Func1 (navigation function): "Calculates real-time position"; performance classes include positioning accuracy, for example, performance class instance: Perf1 (positioning accuracy): "Position calculation accuracy is within 10 meters"; component classes include the Global Positioning System (GPS) receiver and the Flight Management Computer (FMC), for example, component class instance: Comp1 (GPS receiver): "Receives satellite signals and calculates position"; constraint classes include weight constraints and power consumption constraints, for example, constraint class instance: Cons1 (weight constraint): "The GPS receiver weight does not exceed 1 kilogram".

[0021] The relationships between the core classes include inheritance, dependency, and constraint. Inheritance indicates a "one-to-one" relationship between core classes; for example, functional requirements and non-functional requirements both inherit from the requirement class. Similarly, functional classes can have sub-functionalities, and component classes can have sub-components. Dependency indicates that one core class uses another core class; for example, a functional class depends on a requirement class, a performance class depends on both functional and component classes, and a component class depends on both functional and performance classes. Constraint indicates that one core class imposes constraints on another core class; for example, a constraint class imposes constraints on a component class, such as a weight constraint on a GPS receiver; a constraint class imposes constraints on a functional class, such as a computational delay constraint on a navigation function.

[0022] Then, a natural language processing model is used to identify entities belonging to each of the core classes from the text data of the multi-source knowledge base, and the relationships between the entities are extracted based on the relationships between the core classes. The natural language processing model can be a semantic understanding model based on the BERT-Transformer architecture. For example, the semantic understanding model is used to parse the text data "The command calculation response time of the flight control computer should not be greater than 10 milliseconds.", and the entities belonging to each of the core classes are identified as follows: (1) the entity "flight control computer" belongs to the "component class", (2) the entity "command calculation response time" belongs to the "performance class", and (3) the entity "not greater than 10 milliseconds" belongs to the "constraint class". Then, based on the relationships between the "component class", "performance class", and "constraint class", the relationships between the entities "flight control computer", "command calculation response time", and "not greater than 10 milliseconds" are extracted, namely (flight control computer, has, command calculation response time), (command calculation response time, constrained by, not greater than 10 milliseconds).

[0023] Finally, based on the target design object ontology model and the identified entities and relationships between them, a knowledge graph for the airborne system domain is constructed using a knowledge graph tool such as Neo4j, according to the semantic association of "requirements-functions-performance-components-constraints".

[0024] Step Two: Train the Natural Language Processing (NLP) model using the airborne system domain knowledge graph. Use the trained NLP model to parse the design requirement document of the target design object in the airborne system, identify and extract entities corresponding to the key design elements of the target design object, and the relationships between these entities, forming a complete set of key design elements. It should be noted that the purpose of constructing the airborne system domain knowledge graph in Step One is to provide high-quality labeled data and domain prior knowledge for the NLP model in Step Two, thereby improving the accuracy of parsing and extracting entities and their relationships. The output of Step Two, in turn, enriches and expands the knowledge base of Step One, forming a mutually reinforcing closed loop.

[0025] The key design elements include functional units, performance parameters, interface units, and constraint indicators. Specifically: A functional unit refers to a subsystem or module designed to complete a specific task, possessing a clearly defined logical boundary. It defines what the target design object in the airborne system "does." For example, the functional unit "Flight Management Module" is responsible for core flight management functions. The Flight Management Module includes a navigation management subunit, a performance management subunit, and a guidance law calculation subunit. The navigation management subunit calculates and tracks the flight trajectory, the performance management subunit calculates optimal speed, thrust, and other parameters, and the guidance law calculation subunit generates specific lateral and longitudinal guidance commands. Performance parameters are a series of measurable indicators used to quantitatively characterize the dynamic and static characteristics and capability boundaries of a functional unit or system. For example, the performance parameter of the functional unit "Flight Management Module" can be defined as "reliability not less than 0.999 under a standard 5-hour mission profile." Interface units define the channels, protocols, and specifications for information, energy, or material exchange between functional units or with the external environment. For example, the ARINC 429 transmit / receive interface between the functional unit "Navigation Management Subunit" and the functional unit "Inertial Reference System." The aforementioned safety indicators are quantitative or qualitative requirements used to assess and ensure that the target design object in an airborne system will not lead to catastrophic consequences in the event of a failure. For example, for a flight control system, the safety indicator may require that "the computer processor of the flight control system has a detection coverage of no less than 99.9% for transient single-event upset events."

[0026] Step 3: Use a machine learning model to learn the modeling pattern of the historical design digital model, and combine it with the set of key design elements to generate digital modeling rules for the target design object through a rule engine, and output a machine-readable specification document based on XMLSchema.

[0027] Specifically, firstly, the K-means clustering algorithm is used to analyze the modeling patterns of the historical design digital model, identifying the general architectural patterns and common anti-patterns of the target design object. Then, the Apriori association rule algorithm is used to analyze the dependencies between the key design elements, obtaining the modeling constraints of the target design object. Finally, based on the architectural patterns and anti-patterns, and the modeling constraints, the Drools rule engine generates the modeling syntax rules, architectural style rules, and naming rules for the target design object, outputting a machine-readable specification file based on XMLSchema.

[0028] Step 4: Integrate the digital modeling rules into the modeling environment. Utilize the model checking plugin within the modeling environment to perform static structural verification of the digital model of the target design object to be verified, thereby verifying whether the structure of the digital model conforms to the digital modeling rules. Simultaneously, the test case generator in the modeling environment generates test cases based on the digital modeling rules and performs dynamic behavior verification on the digital model to be verified, thereby verifying whether the dynamic response characteristics of the digital model conform to the digital modeling rules.

[0029] After verification, the verification results are recorded and fed back to the multi-source knowledge base to optimize the natural language processing model and the rule engine through incremental learning algorithms, and to update the digital modeling rules of the target design object.

[0030] See Figure 2 This embodiment also provides an apparatus for formulating modeling rules for airborne system digital models, which implements the aforementioned method for formulating modeling rules for airborne system models, including: The multi-source knowledge base and knowledge graph construction module is used to construct a knowledge graph for the airborne systems domain based on a multi-source knowledge base in the airborne systems domain; the multi-source knowledge base includes design requirements, standards and specifications, and historical cases; The key design element extraction module is used to train the natural language processing model using the knowledge graph of the airborne system domain, and to use the trained natural language processing model to parse the design requirement document of the target design object in the airborne system, identify and extract the entities and relationships between the key design elements of the target design object, and form a complete set of key design elements; the key design elements include functional units, performance parameters, interface units, and constraint indicators; The modeling rule generation module is used to learn the modeling patterns of historical design digital models in the historical cases using a machine learning model, and combine them with the set of key design elements to generate digital modeling rules for the target design object through a rule engine; the digital modeling rules include modeling syntax rules, architecture style rules, and naming rules.

[0031] The verification execution and closed-loop update module is used to integrate the digital model modeling rules into the modeling environment. It utilizes the model checking plugin of the modeling environment to perform static verification of the model structure of the digital model to be verified for the target design object. Simultaneously, the test case generator of the modeling environment generates test cases based on the digital model modeling rules and performs dynamic behavior verification on the digital model to be verified. The verification results are recorded and fed back to the multi-source knowledge base to optimize the natural language processing model and the rule engine through incremental learning algorithms, and to update the digital model modeling rules for the target design object.

[0032] This embodiment also provides a computer device, such as... Figure 3 As shown, the system includes a memory 1, a processor 2, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned method for formulating modeling rules for an airborne system model. Specifically, the computer device can be a computer terminal, a server, or a similar computing device.

[0033] This embodiment also provides a computer-readable storage medium storing a computer program that executes the above-described method for formulating modeling rules for airborne system models.

[0034] Specifically, computer-readable storage media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer-readable storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable storage media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0035] Example 2 The fuel shut-off valve is a core actuating component of an aircraft fuel system, and its model must accurately reflect the valve's dynamic characteristics and meet stringent safety and reliability requirements. Traditional modeling methods, due to inconsistent interpretations of specifications and strong reliance on experience, easily lead to significant deviations between the model and actual physical characteristics, affecting the accuracy of system-level simulations. Therefore, this invention takes the modeling of a fuel shut-off valve in an aircraft's airborne system as the target design object, and uses the formulation of modeling rules for the fuel shut-off valve as an example to further illustrate the solution of this invention. (See [link to relevant documentation]). Figure 2 .

[0036] Step 1: Construction of Multi-Source Knowledge Base and Knowledge Graph First, we acquire multi-source heterogeneous data in the field of airborne systems, such as design requirements, standards and specifications, and historical cases for avionics systems, electromechanical systems, and flight control systems, to form the original database. The electromechanical systems database includes data related to fuel systems (including fuel shut-off valves). For example, the design requirement for a certain aircraft's airborne system fuel shut-off valve is "opening or closing time not exceeding 2 seconds," the design standard is HB 8505 "Design Requirements for Civil Aircraft Fuel Systems," and the airworthiness clause is the special requirements for fuel system reliability in CCAR 25.1189. The database also includes digital models of historical fuel shut-off valves and historical case studies of failures from previous projects.

[0037] Secondly, the Protégé ontology modeling tool was used to construct an airborne domain ontology model, defining core classes such as requirement classes, functional classes, performance classes, component classes, and constraint classes, and determining the inheritance, dependency, and constraint relationships between these core classes. It should be noted that requirement classes include performance requirements and safety requirements, functional classes include cockpit display and control, fuel control, and flight control, performance classes include response time and display accuracy, component classes include displays, computers, and actuators, and constraint classes include electrical interfaces and fail-safety mechanisms.

[0038] Then, a natural language processing model is used to identify entities belonging to each of the core classes from the text data of the multi-source knowledge base, and the relationships between the entities are extracted based on the relationships between the core classes. For example, based on the design document of the "fuel cut-off valve" component of the fuel subsystem in the electromechanical system, which states "opening or closing time not greater than 2s", the following entities can be extracted: "fuel cut-off valve", "response time", and "not greater than 2s". The relationships between these entities are: (fuel cut-off valve, has, response time) and (response time, constrained by, not greater than 2s).

[0039] Finally, based on the airborne system domain ontology model and the identified entities and relationships between them, an airborne system domain knowledge graph is constructed using the Neo4j graph database. The nodes of the knowledge graph include fuel cut-off valve, response time, and no more than 2 seconds, etc., and the edges of the knowledge graph include "have", "constrained by", etc.

[0040] Step 2: Extraction of Key Design Elements Deploy the BERT-Transformer model and use the airborne system domain knowledge graph from step 1 for domain-adaptive training and fine-tuning. This enables the BERT-Transformer model to accurately understand the contextual meaning of airborne domain terminology and learn the relationships between entities. Then, use the BERT-Transformer model to perform deep semantic understanding on the design requirements text for an aircraft fuel cut-off valve, identifying and extracting entities and relationships corresponding to the key design elements of the fuel cut-off valve, thus forming a set of key design elements for the fuel cut-off valve. For example, identify "open or close time" as a performance parameter and "1.5s" as a numerical constraint. Use the BiLSTM-CRF sequence labeling model to identify entity boundaries and types. For example, label "open or close time" as a performance parameter entity and "must not exceed 1.5s" as a constraint entity. Use a graph neural network relationship extraction module to parse the relationships between entities. For example, establish the relationship between (fuel cut-off valve, performance parameter, open or close time) and (open or close time, constraint, not greater than 1.5s).

[0041] Step 3: Generating Modeling Rules Using the aforementioned set of key design elements and the historical design digital model of the fuel cut-off valve as input, K-means clustering analysis was used to analyze the historical design digital model of the fuel cut-off valve, identifying common architectural patterns (such as using state machine modeling) and anti-patterns (such as avoiding the use of continuous-time integrators). The Apriori algorithm was used to analyze the association rules between key design elements; for example, the performance parameter "open or close time" often appears simultaneously with the constraint "not greater than 2s" and is related to the step response test in the test cases. Modeling rules were generated using the Drools rule engine, including: syntax rule: the fuel cut-off valve model must contain a state machine submodule; architectural rule: the fuel cut-off valve model must not contain a continuous-time integrator; naming rule: the state machine submodule of the fuel cut-off valve must be named "ValveStateMachine"; constraint rule: the step response simulation result of the fuel cut-off valve must reach stability within 2 seconds. The output is an XMLSchema-based specification file containing the above rules, forming a machine-readable document.

[0042] Step 4: Verify Execution and Closed-Loop Update The aforementioned XML specification file is integrated into Simulink, and SimulinkCheck is used to check whether the fuel cut-off valve model to be verified conforms to the rules. For example, it checks whether the model contains a state machine named "ValveStateMachine" and whether a prohibited continuous-time integrator is used. SimulinkTest is used to generate test cases according to the rules, such as automatically generating a step input signal and verifying whether the model switches from the open state to the closed state within 2 seconds. The verification results are fed back to a multi-source knowledge base. For example, if a model structure is found to perform well even though it does not conform to the naming rules, the rule engine is adjusted through incremental learning to update the modeling rules. Through the above steps, modeling rules can be automatically generated from the design requirements of the fuel cut-off valve, and then verified and optimized.

[0043] This invention is based on an intelligent specification generation mechanism that integrates multi-source knowledge, breaking through the traditional model that relies on experts to manually compile specifications. By constructing a multi-source knowledge base that integrates design requirements, standard specifications, and historical cases, and using knowledge graph and natural language processing technologies, it realizes semantic association and intelligent mining of knowledge in the field of airborne systems, automatically generating machine-readable specifications that cover the entire modeling process, and solving the industry pain point of incomplete and inaccurate specification formulation.

[0044] This invention learns modeling paradigms from historical digital design models using machine learning algorithms, transforming abstract text specifications into specific modeling syntax rules, architectural constraints, and naming conventions, and simultaneously and automatically generating corresponding test cases. This achieves integrated output from specification formulation to verification and execution, significantly improving the efficiency of specification implementation.

[0045] This invention establishes a real-time verification channel between specifications and models. By integrating an inspection plugin into the modeling environment, it performs static and dynamic verification of the design model and automatically feeds the verification results back to the knowledge base, driving the continuous iterative optimization of AI models and specification content, forming an intelligent control closed loop with self-learning and self-evolution capabilities.

[0046] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for formulating modeling rules for airborne systems, characterized in that, include: A knowledge graph for the airborne systems domain is constructed based on a multi-source knowledge base; the multi-source knowledge base includes design requirements, standards and specifications, and historical cases. The natural language processing model is trained using the knowledge graph of the airborne system domain. The trained natural language processing model is then used to parse the design requirement document of the target design object in the airborne system, identify and extract the entities corresponding to the key design elements of the target design object and the relationships between the entities, and form a set of key design elements. The modeling patterns of historical design digital models in the historical cases are learned using a machine learning model, and combined with the set of key design elements, a rule engine is used to generate digital modeling rules for the target design object. The digital modeling rules include modeling syntax rules, architecture style rules, and naming rules.

2. The method for formulating modeling rules for airborne system models according to claim 1, characterized in that, Also includes: The digital modeling rules are integrated into the modeling environment. The model checking plugin of the modeling environment is used to perform static verification of the model structure of the digital model to be verified of the target design object. At the same time, the test case generator of the modeling environment generates test cases based on the digital modeling rules and performs dynamic behavior verification of the digital model to be verified. The verification results are recorded and fed back to the multi-source knowledge base to optimize the natural language processing model and the rule engine through incremental learning algorithms, and to update the digital modeling rules of the target design object.

3. The method for formulating modeling rules for airborne system models according to claim 2, characterized in that, Methods for constructing knowledge graphs in the airborne systems domain include: An ontology modeling tool is used to construct an ontology model for the airborne system domain, and the core classes of the ontology model and the relationships between the core classes are defined. A natural language processing model is used to identify entities belonging to each of the core classes from text data in a multi-source knowledge base, and the relationships between the entities are extracted by combining the relationships between the core classes. Based on the airborne system domain ontology model and the identified entities and relationships between them, a knowledge graph for the airborne system domain is constructed using knowledge graph tools according to the semantic association of "requirements-functions-performance-components-constraints".

4. The method for formulating modeling rules for airborne system models according to claim 3, characterized in that, The key design elements include functional units, performance parameters, interface units, and constraint indicators.

5. The method for formulating modeling rules for airborne system models according to claim 4, characterized in that, The method of using a machine learning model to learn the modeling patterns of historical design digital models in the historical cases, and combining them with the set of key design elements, to generate digital modeling rules for the target design object through a rule engine includes: The K-means clustering algorithm is used to analyze the modeling patterns of the historical design digital model and identify the architectural patterns and anti-patterns of the target design object. The Apriori association rule algorithm is used to analyze the dependencies between the key design elements to obtain the modeling constraints of the target design object. Based on the aforementioned architectural patterns and anti-patterns, and the aforementioned modeling constraints, the Drools rule engine generates digital modeling rules for the target design object and outputs them as machine-readable specification files based on XMLSchema.

6. The method for formulating modeling rules for airborne system models according to claim 5, characterized in that, The standard specifications include digital modeling rules, design standards, and airworthiness requirements for the target design object; the historical cases also include historical failure cases of the target design object.

7. A device for formulating modeling rules for airborne system digital models, implemented based on the airborne system modeling rule formulating method according to any one of claims 1-6, characterized in that, include: The multi-source knowledge base and knowledge graph construction module is used to construct a knowledge graph for the airborne systems domain based on a multi-source knowledge base in the airborne systems domain; the multi-source knowledge base includes design requirements, standards and specifications, and historical cases; The key design element extraction module is used to train the natural language processing model using the knowledge graph of the airborne system domain, and to use the trained natural language processing model to parse the design requirement document of the target design object in the airborne system, identify and extract the entities corresponding to the key design elements of the target design object and the relationships between the entities, forming a complete set of key design elements; the key design elements include functional units, performance parameters, interface units, and constraint indicators. The modeling rule generation module is used to learn the modeling patterns of historical design digital models in the historical cases using a machine learning model, and combine them with the set of key design elements to generate digital modeling rules for the target design object through a rule engine. The digital modeling rules include modeling syntax rules, architecture style rules, and naming rules.

8. The airborne system digital modeling rule formulation device according to claim 7, characterized in that, Also includes: The verification execution and closed-loop update module is used to integrate the digital model modeling rules into the modeling environment, use the model checking plugin of the modeling environment to perform static verification of the model structure of the digital model to be verified of the target design object, and at the same time, the test case generator of the modeling environment generates test cases based on the digital model modeling rules and performs dynamic behavior verification of the digital model to be verified. The verification results are recorded and fed back to the multi-source knowledge base to optimize the natural language processing model and the rule engine through incremental learning algorithms, and to update the digital modeling rules of the target design object.

9. A computer device comprising a memory (1), a processor (2), and a computer program stored in the memory (1) and executable on the processor (2), characterized in that, When the processor (2) executes the computer program, it implements the airborne system modeling rule formulation method as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that executes the airborne system modeling rule formulation method according to any one of claims 1 to 6.

Images