MBSE-based methods, apparatus, equipment, and storage media for aircraft metrological traceability.

By extracting spatiotemporal features and modeling requirements for aircraft metrology objects based on MBSE, the problem of neglecting the spatiotemporal characteristics of metrology traceability requirements was solved. This enabled multi-dimensional optimization of metrology design schemes and reduced the difficulty and cost of aircraft metrology traceability.

CN123068942APending Publication Date: 2026-11-10CHINA ELECTRONICS RELIABILITY AND ENVIRONMENTAL TESTING INSTITUTE ((THE FIFTH INSTITUTE OF ELECTRONICS MINISTRY OF INDUSTRY AND INFORMATION TECHNOLOGY) (CHINA SAIBAO LABORATORY)
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511983013.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-26
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

The current aircraft development process has not effectively considered the requirements for metrological traceability and the spatiotemporal characteristics of metrological attributes, resulting in blurred traceability boundaries of metrological objects and a lack of design support for testing/calibration schemes, which increases the difficulty and maintenance cost of metrological traceability.

Method used

The MBSE-based approach processes raw multimodal data, extracts the spatiotemporal characteristics of the measurement object, generates a measurement meta-model, constructs a measurement demand model, performs spatiotemporal design analysis, selects target measurement design schemes, and generates a summary table of aircraft traceability requirements.

Benefits of technology

It achieves accurate spatiotemporal feature extraction, efficiently filters out the core spatiotemporal attributes of the same measurement type, optimizes the measurement design scheme, reduces the difficulty and maintenance cost of measurement traceability, and ensures measurement compliance and traceability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123068942A_ABST
    Figure CN123068942A_ABST
Patent Text Reader

Abstract

This invention discloses a method, apparatus, device, and storage medium for aircraft metrological traceability based on MBSE (Metrological Detection and Analysis). The method includes: generating structural elements based on the spatiotemporal characteristics of similar metrological objects within the aircraft as a whole; generating a metrological meta-model based on the structural elements; dividing the aircraft as a whole into multiple sub-modules; forecasting demand based on stakeholders of each sub-module; constructing a metrological demand model; performing spatiotemporal design analysis on the aircraft based on the metrological meta-model and the metrological demand model; constructing a spatiotemporal metrological design model; selecting target metrological design schemes from the spatiotemporal metrological design model through trade-off analysis; and generating a summary table of aircraft traceability requirements based on the target metrological design schemes. By analyzing the spatiotemporal characteristics of metrological objects within the aircraft, this method effectively avoids the problems of static metrological attributes and metrological traceability gaps in metrological schemes, reduces the difficulty and maintenance cost of aircraft metrological traceability, and ensures metrological compliance and traceability.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of aircraft metrology, and in particular to an aircraft metrological traceability method, apparatus, equipment, and storage medium based on MBSE. Background Technology

[0002] Aircraft development is a complex systems engineering project, and modern aircraft design often employs Model-Based Systems Engineering (MBSE) methods. This approach outperforms traditional document-based and code-centric design methods. Driven by models, it uses graphical modeling languages ​​to provide unambiguous descriptions, analyses, and designs of the requirements, structure, and behavior of complex systems. This avoids inconsistencies in tracing back to the source information, reduces contradictions between system information elements, improves the collaborative efficiency of overall and component design, and eliminates system development problems caused by ambiguity. Furthermore, the model-based systems engineering approach covers the entire design lifecycle, enabling rapid iteration and testing of simulation designs, significantly accelerating system development.

[0003] However, in the current development of aircraft, the metrological traceability requirements throughout the entire lifecycle of the equipment (manufacturing, use, and maintenance) are often overlooked, and the spatiotemporal characteristics of metrological attributes changing dynamically with time and space are not considered. This disconnect between design and use, as well as the problem of only considering static metrological attributes, leads to blurred traceability boundaries for subsequent metrological objects and a lack of design support for testing / calibration schemes, significantly increasing the difficulty and maintenance cost of aircraft metrological traceability. Summary of the Invention

[0004] The main objective of this invention is to provide a method, apparatus, device, and storage medium for aircraft metrological traceability based on MBSE. This invention aims to address the technical problem that existing technologies have not effectively considered the metrological traceability requirements throughout the entire life cycle of aircraft and the spatiotemporal characteristics of metrological attributes, resulting in blurred traceability boundaries for subsequent metrological objects, a lack of design support for metrological schemes, and a significant increase in the difficulty and maintenance cost of aircraft metrological traceability.

[0005] To achieve the above objectives, this invention provides a method for the metrological traceability of aircraft based on MBSE, the method comprising the following steps: The original multimodal data is processed, and the spatiotemporal features of measurement objects of the same measurement type in the whole aircraft are extracted based on the processed original multimodal data. The spatiotemporal features include the temporal constraints and spatial constraints of the measurement objects, and the measurement objects are functional systems or functional devices in the whole aircraft. Based on the aforementioned spatiotemporal features, constructive elements are generated. Based on the constructive elements, metrological modeling is performed on each metrological object to generate a metrological meta-model. The metrological meta-model uses the detection equipment and / or calibration equipment of the metrological object as minimizing elements. The metrological meta-model includes a detection equipment meta-model and a calibration equipment meta-model. The aircraft is divided into multiple sub-modules. Demand forecasts are made based on the stakeholders corresponding to each sub-module, and a quantitative demand model is constructed based on the demand forecast results. Each sub-module corresponds to a quantitative demand. Based on the aforementioned metrological meta-model and metrological requirement model, a spatiotemporal design analysis of the aircraft is conducted, and multiple spatiotemporal metrological design models are constructed. The spatiotemporal metrological design models include detection schemes and / or calibration schemes for the time and space dimensions. A trade-off analysis is performed on the multiple spatiotemporal metrological design models to select the target metrological design scheme, and a summary table of aircraft traceability requirements is generated based on the target metrological design scheme.

[0006] Optionally, the step of processing the original multimodal data and extracting the spatiotemporal features of measurement objects of the same measurement type in the entire aircraft based on the processed original multimodal data includes: The original multimodal data is processed into different modalities to obtain initial multimodal features; The initial multimodal features are input into a multimodal pre-trained model for multimodal fusion, and a multimodal feature vector of uniform dimension is output. The multimodal pre-trained model is configured to learn the association weights of different modalities based on the modal attention mechanism, and to fuse the initial multimodal features based on the association weights. Semantic analysis is performed on the multimodal feature vectors, and multimodal entity recognition and multimodal relation extraction are performed based on the semantic analysis results to obtain multimodal entities and multimodal relations; Multiple valid quadruplets are generated based on the multimodal entities and the multimodal relationships. The entities in the valid quadruplets include the aircraft as a whole, the functional systems in the aircraft as a whole, the measurement objects in the functional systems, and the measurement attribute information of the measurement objects. The measurement attribute information includes measurement attributes and the time attributes of the measurement attributes. The measurement objects are the functional systems or the functional devices in the functional systems. A dynamic hypergraph is constructed based on the effective quadruples, and spatiotemporal features of measurement objects of the same measurement type in the whole spacecraft are extracted based on the dynamic hypergraph. The nodes in the dynamic hypergraph are the entities in the effective quadruples. The hyperedges of the dynamic hypergraph include structural association hyperedges, measurement association hyperedges and temporal association hyperedges. The structural association hyperedges are used to connect measurement objects with structural associations in the same system or across systems in the whole spacecraft. The measurement association hyperedges are used to connect measurement objects with attribute associations in the same system or across systems in the whole spacecraft. The temporal association hyperedges are used to connect measurement objects with temporal dimension associations.

[0007] Optionally, generating multiple valid quadruplets based on the multimodal entities and the multimodal relationships includes: A set of quadruples is generated based on the multimodal entities and the multimodal relationships, and the set of quadruples includes multiple original quadruples; Traverse the set of quadruples to determine the valid and invalid quadruples in the set of quadruples, wherein the invalid quadruples are missing at least one entity. Based on the overall system topology of the aircraft with the invalid quadruple, a candidate entity set is generated. The candidate entity set includes multiple candidate entities, and each candidate entity includes the overall functional system of the aircraft, the measurement object, and the measurement attribute information of the measurement object. Semantic analysis is performed on the invalid quadruplets, and multiple entities to be scored are selected from the candidate entity set based on the semantic analysis results; Analyze the structural and / or quantitative correlations between the entities currently existing in the invalid quadruples and each entity to be scored to obtain the correlation analysis results; Based on the correlation analysis results, each entity to be scored is scored, and the target completion entity is selected from the entities to be scored based on the scoring results; The target completion entity is embedded into the invalid quad, and the invalid quad after entity embedding is marked as a valid quad.

[0008] Optionally, the process of dividing the entire aircraft into multiple sub-modules, forecasting demand based on the stakeholders corresponding to each sub-module, and constructing a quantitative demand model based on the demand forecast results includes: Based on the overall system architecture, structural type, usage scenarios and / or users of the aircraft, the aircraft is divided into multiple sub-modules; Demand time series data is obtained based on the stakeholders corresponding to each sub-module, and the demand time series data is preprocessed to obtain candidate time series data. The demand time series data includes the demands of stakeholders and the timestamps of the demands. The preprocessing includes normalization processing. The candidate time-series data is input into a pre-built demand time-series prediction model to obtain demand prediction results. The demand time-series prediction model includes a multi-feature extraction branch structure, a target feature fusion structure, and a multi-task time-series decoding layer. The multi-feature extraction branch structure is used to extract global demand features and local demand features from the candidate time-series data. The target feature fusion structure is used to fuse the global demand features and the local demand features to output target fused features. The multi-task time-series decoding layer is used to decode the target fused features to output demand prediction results. Architectural analysis, behavioral analysis, and indicator analysis are performed on each submodule to obtain multidimensional analysis results for each submodule; Based on the multidimensional analysis results, demand forecasting results are captured, and a quantitative demand model is constructed based on the captured demand results.

[0009] Optionally, the feature extraction branch structure includes a local feature extraction structure and a global feature extraction structure. The local feature extraction structure includes a channel splitting structure, a first branch structure, a second branch structure, a third branch structure, and a candidate feature fusion module. The global feature extraction structure includes a fourth branch structure. The channel splitting structure is configured to perform semantic awareness analysis on candidate time series data to obtain semantic awareness features, and divide the candidate time series data into multiple channel data based on the semantic awareness features, and input the channel data into the first branch structure, the second branch structure and the third branch structure respectively based on the semantic awareness features corresponding to each channel data. The first branch structure is built on a Bi-LSTM network and is configured to extract bidirectional time-series features based on channel data, and convert the bidirectional time-series features into a query vector matrix; The second branch structure is built based on a graph attention network and is configured to extract the demand association features between each sub-module based on channel data, and convert the demand association features into a key vector matrix; The third branch structure is constructed based on a spatiotemporal graph convolutional network and is configured to extract spatiotemporal constraint features based on channel data and convert the spatiotemporal constraint features into a value vector matrix. The candidate feature fusion module is configured to split the query vector matrix, key vector matrix, and value vector matrix into multiple sub-matrices along the feature dimension, generate multiple attention heads based on the sub-matrices, calculate cross-branch attention scores through the multiple attention heads, concatenate the output results of each attention head, map the concatenation result to the original dimension, and output local demand features. The fourth branch structure is built based on the Transformer model and is configured to extract global requirement features of the entire life cycle of the aircraft based on candidate time series data.

[0010] Optionally, the step of performing a trade-off analysis on the multiple spatiotemporal econometric design models, selecting a target econometric design scheme, and generating a summary table of aircraft traceability requirements based on the target econometric design scheme includes: Feature encoding is performed on each spatiotemporal metrological design model from multiple trade-off feature dimensions to obtain multi-dimensional metrological trade-off features. The multi-dimensional metrological trade-off features include the metrological cost features, metrological time features, metrological space features, and metrological method features of each metrological object in the spatiotemporal metrological design model. The metrological method features include off-site metrological features and in-situ metrological features. Based on the multi-dimensional measurement trade-off features, the spatiotemporal measurement design model is converted into a measurement graph structure. The nodes in the measurement graph structure represent the measurement objects in the spatiotemporal measurement design model, and the node attributes of each node are the multi-dimensional measurement trade-off features. The edges of the measurement graph structure include structural dependencies, measurement dependencies, and time dependencies between nodes. The metrology graph structure is input into a graph attention network to obtain the node embedding vectors of each metrology object. According to the usage scenarios of each stage of the overall life cycle of the aircraft, configure the measurement weights corresponding to the node embedding vectors of each measurement object, and aggregate the node embedding vectors of each measurement object based on the measurement weights to obtain the global measurement features of each spatiotemporal measurement design model. Based on the global metrological characteristics, a target metrological design scheme is selected from the metrological design schemes corresponding to the multiple spatiotemporal metrological design models, and a summary table of aircraft traceability requirements is generated based on the target metrological design scheme.

[0011] Optionally, generating the aircraft traceability requirements summary table based on the target metrological design scheme includes: Based on the target metrological design scheme, determine the testing equipment and / or calibration equipment for each metrological object; Identify multiple candidate suppliers associated with the testing equipment and / or calibration equipment; Collect historical supply data, supplier profile data, and industry dynamic data of the candidate suppliers, and generate an evaluation time series dataset based on the historical supply data, supplier profile data, and industry dynamic data; The evaluation time series dataset is preprocessed to obtain target time series data, and supply dynamic risk prediction is performed on each candidate supplier based on the target time series data to obtain supply dynamic risk prediction results. Send requests to multiple device quality sharing nodes so that each device quality sharing node sends the device quality data of the candidate supplier; The equipment quality data is preprocessed, and the equipment quality risk assessment of each candidate supplier is performed based on the preprocessed equipment quality data to obtain the quality risk assessment results. Target suppliers are selected from the candidate suppliers based on the supply dynamic risk prediction results and the quality risk assessment results; A summary table of aircraft traceability requirements is generated based on the target suppliers associated with the testing and / or calibration equipment for each metrological object and the target metrological design scheme.

[0012] Furthermore, to achieve the above objectives, the present invention also proposes an MBSE-based aircraft metrological traceability device, wherein the MBSE-based aircraft metrological traceability device comprises: The spatiotemporal feature extraction module is used to process the raw multimodal data and extract the spatiotemporal features of measurement objects of the same measurement type in the whole aircraft based on the processed raw multimodal data. The spatiotemporal features include the temporal constraints and spatial constraints of the measurement objects. The measurement objects are functional systems or functional devices in the whole aircraft. The meta-model construction module is used to generate constructive elements based on the spatiotemporal features, perform metrological modeling on each metrological object based on the constructive elements, and generate a metrological meta-model. The metrological meta-model uses the detection equipment and / or calibration equipment of the metrological object as the minimization element. The metrological meta-model includes a detection equipment meta-model and a calibration equipment meta-model. The demand forecasting module is used to divide the entire aircraft into multiple sub-modules, forecast the demand based on the stakeholders corresponding to each sub-module, and build a quantitative demand model based on the demand forecasting results. Each sub-module corresponds to a quantitative demand. The design modeling module is used to perform spatiotemporal design analysis on the aircraft based on the metrological meta-model and the metrological requirement model, and to construct multiple spatiotemporal metrological design models, which include detection schemes and / or calibration schemes for time and space dimensions. The traceability and summary module is used to perform trade-off analysis on the multiple spatiotemporal metrological design models, select target metrological design schemes, and generate a summary table of aircraft traceability requirements based on the target metrological design schemes.

[0013] Furthermore, to achieve the above objectives, this application also proposes an MBSE-based aircraft metrological traceability device, the device comprising: a memory, a processor, and an MBSE-based aircraft metrological traceability program stored in the memory, the processor being used to run the MBSE-based aircraft metrological traceability program, the computer program being configured to implement the steps of the MBSE-based aircraft metrological traceability method as described above.

[0014] In addition, to achieve the above objectives, this application also proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the MBSE-based aircraft metrological tracing method described above.

[0015] This invention addresses the issues of inconsistent multimodal data and varying data quality by extracting spatiotemporal features of measurement objects of the same measurement type from processed multimodal data. It achieves accurate spatiotemporal feature extraction, efficiently filters out the core spatiotemporal attributes of measurement objects of the same measurement type, and enables efficient and accurate spatiotemporal modeling of the metrological meta-model. This effectively avoids the problem of static metrological attributes in metrological schemes. By forecasting the needs of stakeholders in each sub-module of the aircraft, the metrological requirements and constraints of each sub-module are clearly defined. Based on the metrological meta-model and metrological requirement model, spatiotemporal design modeling of the aircraft is performed, achieving multi-dimensional optimization of the metrological design scheme. This fully balances the multi-objective design of time, space, and needs, avoiding the defects caused by single-dimensional design. Through trade-off analysis, the target metrological design scheme is selected, improving the objectivity of the metrological scheme, reducing the influence of subjective factors on scheme selection, lowering the difficulty and maintenance cost of aircraft metrological traceability, and ensuring metrological compliance and traceability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the structure of an MBSE-based aircraft metrological traceability device in the hardware operating environment involved in the embodiments of the present invention; Figure 2 This is a flowchart illustrating the first embodiment of the MBSE-based aircraft metrological traceability method of the present invention. Figure 3 This is a flowchart illustrating the second embodiment of the MBSE-based aircraft metrological traceability method of the present invention. Figure 4 This is a flowchart illustrating the third embodiment of the MBSE-based aircraft metrological traceability method of the present invention. Figure 5 This is a schematic diagram of the demand time series prediction model in one embodiment of the MBSE-based aircraft metrological tracing method of the present invention; Figure 6 This is a schematic diagram of the structure of multiple feature extraction branches in one embodiment of the MBSE-based aircraft metrological tracing method of the present invention; Figure 7 This is a flowchart illustrating the fourth embodiment of the MBSE-based aircraft metrological traceability method of the present invention. Figure 8 This is a structural block diagram of the first embodiment of the aircraft metrological traceability device based on MBSE of the present invention.

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0020] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of an MBSE-based aircraft metrological traceability device in the hardware operating environment of an embodiment of the present invention.

[0021] like Figure 1 As shown, the MBSE-based aircraft metrology and traceability device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0022] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on MBSE-based aircraft metrological traceability equipment and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0023] like Figure 1As shown, the memory 1005, which is a computer-readable storage medium, may include an operating system, a network communication module, a user interface module, and an MBSE-based aircraft metrological tracing program.

[0024] exist Figure 1 In the MBSE-based aircraft metrological traceability device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and memory 1005 in the MBSE-based aircraft metrological traceability device of the present invention can be set in the MBSE-based aircraft metrological traceability device. The MBSE-based aircraft metrological traceability device calls the MBSE-based aircraft metrological traceability program stored in the memory 1005 through the processor 1001 and executes the MBSE-based aircraft metrological traceability method provided in the embodiment of the present invention.

[0025] This invention provides a method for aircraft metrological traceability based on MBSE, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the MBSE-based aircraft metrological traceability method of the present invention.

[0026] In this embodiment, the MBSE-based aircraft metrological traceability method includes the following steps: Step S10: Process the original multimodal data and extract the spatiotemporal features of measurement objects of the same measurement type in the whole aircraft based on the processed original multimodal data.

[0027] It should be understood that the executing entity of this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or a terminal electronic device capable of performing the above functions. The following description uses an MBSE-based aircraft metrological traceability device (hereinafter referred to as the traceability device) as an example to illustrate this embodiment and the following embodiments.

[0028] It should be noted that the original multimodal data can be a collection of heterogeneous data related to measurement throughout the entire life cycle of the aircraft. It can include data from different modalities such as vision, sensing, and text, such as image data of various components of the aircraft, operational parameter data collected by sensors, text data of CAD design drawings, and operation and maintenance record data.

[0029] It should be noted that the measurement type can be a measurement category divided according to the measurement purpose and object attributes, and may include accuracy measurement, performance measurement, safety measurement, life measurement, etc. For example, the power measurement of an engine belongs to performance measurement, and the signal accuracy measurement of an avionics system belongs to accuracy measurement.

[0030] The measurement objects can be functional systems or equipment within the aircraft that require measurement operations. Functional systems include power systems, avionics systems, and navigation systems, while functional equipment includes engines, pressure sensors, temperature and humidity sensors, gyroscopes, and communication terminals.

[0031] In some embodiments, the traceability device can classify objects of the same measurement category under different systems into the same type of measurement object. For example, the temperature and humidity sensor in the cockpit and the temperature and humidity sensor in the cargo hold can be the same type of measurement object.

[0032] It should be noted that the spatiotemporal characteristics are used to reflect the core attributes of the measurement object related to the measurement operation in the time and space dimensions, and may include the time constraints and space constraints of the measurement object. The measurement object is a functional system or functional device in the overall aircraft.

[0033] The aforementioned time constraints refer to the time-related limitations that must be followed when the measurement object is tested or calibrated, including the measurement cycle, the time consumed per measurement, and the time window for measurement operations (such as the requirement to complete the measurement within a specific time period when the aircraft is shut down for maintenance).

[0034] The aforementioned spatial constraints refer to the spatial limitations that must be followed when a metrological object is subjected to testing or calibration operations. These limitations include the physical space required for the metrological operation, the spatial constraints imposed by the installation location of the metrological object on the operation, and the space requirements for the placement of testing / calibration equipment.

[0035] In practical implementation, the tracing device can preprocess the raw multimodal data collected from the aircraft, then extract multimodal features, and fuse the extracted features across modalities. For example, image data is denoised using algorithms (such as Gaussian filtering) to remove noise and image registration is used to unify the coordinate system; sensor data is filtered to remove invalid data using outlier detection algorithms and interpolation is used to supplement missing data; text data is processed using natural language processing techniques (such as word segmentation and entity recognition) to extract key information. Subsequently, data standardization (such as min-max standardization) is used to unify the dimensions of different data types, and then attention mechanisms are used to fuse multi-source data to obtain structured processed data, thereby improving the quality of the original data. Through preprocessing and fusion operations, invalid and redundant information is removed, solving the problems of inconsistent multimodal data formats and noise, providing a reliable data foundation for subsequent feature extraction.

[0036] In some embodiments, the traceability device may use knowledge graph technology to construct a graph structure for the processed standardized data, thereby enabling accurate analysis of the measurement objects, accurately structuring the dependencies and spatiotemporal measurement attributes among the massive number of measurement objects in the aircraft, and improving the efficiency of subsequent measurement traceability.

[0037] In some embodiments, the traceability device, based on processed standardized data, first classifies measurement objects according to measurement type using K-Means clustering, and then filters out a set of measurement objects of the same measurement type. For this set, a hybrid model of CNN convolution and LSTM can be used to extract spatiotemporal features. CNN extracts spatially relevant features of the measurement objects (such as the spatial coordinates of the installation location and the spatial parameters required for operation), while LSTM extracts time-related features (such as historical measurement cycles and fluctuations in measurement time under different operating conditions), achieving accurate spatiotemporal feature extraction. By combining classification clustering and deep learning models, the core spatiotemporal attributes of measurement objects of the same measurement type are efficiently filtered out, enabling measurement analysis from a spatiotemporal dimension and avoiding the problem of static measurement attributes.

[0038] Step S20: Generate structural elements based on the spatiotemporal features, perform metrological modeling on each metrological object based on the structural elements, and generate a metrological meta-model.

[0039] It should be noted that the structural elements can be basic building blocks extracted based on the spatiotemporal characteristics of the measurement object, used to support metrological modeling. The metrological meta-model can be an abstract and standardized descriptive template model of the metrological process of the measurement object, with the detection equipment and / or calibration equipment of the measurement object as the minimized elements. The metrological meta-model includes a detection equipment meta-model and a calibration equipment meta-model.

[0040] It should be noted that the testing equipment metamodel describes the relevant attributes of the testing equipment required by the metrological object, including testing equipment model adaptation rules, equipment operating parameters, equipment operating constraints, and the compatibility relationship between the equipment and the metrological object. The calibration equipment metamodel can describe the relevant attributes of the calibration equipment required by the metrological object, including equipment accuracy level requirements, calibration process specifications, equipment maintenance standards, and the calibration logic between the equipment and the metrological object.

[0041] In some embodiments, the econometric metamodel may include custom elements, which are auxiliary elements for building the econometric metamodel and are used to assist in constructing the stereotyped element model. Custom elements can customize the use cases of the constructed stereotyped elements, and the customized content includes, but is not limited to, the target, the owner, etc.

[0042] Taking the measurability design of an airborne communication system as an example, the communication system of an aircraft is used for communication between the aircraft and the ground or other aircraft. The transmission and reception of communication signals are achieved through antennas. Therefore, the antenna is a crucial component in the measurability design of the communication system. Using the antenna as the object of measurement, and following the antenna's traceability path, the device completes traceability through testing, and the corresponding meta-model is the testing device meta-model. Aircraft antennas are generally distributed in the fuselage, wings, tail, and top, and their technical specifications vary depending on the type and function of the antenna. By defining the meta-classes of the antenna testing device meta-model, the structural type is customized according to different antenna types or the needs of different stakeholders, and different targets or owners are defined according to the application scenarios in different locations.

[0043] Step S30: Divide the aircraft into multiple sub-modules, forecast the demand based on the stakeholders corresponding to each sub-module, and build a quantitative demand model based on the demand forecast results.

[0044] It should be noted that a submodule can be several independent but interconnected functional units formed by disassembling the entire aircraft according to its functional structure, assembly logic, or management boundaries. Examples include power submodules, avionics modules, navigation submodules, and braking submodules. Each submodule corresponds to a measurement requirement.

[0045] It should be noted that stakeholders can be various entities related to the metering operation of aircraft sub-modules, including aircraft designers, manufacturers, operators, regulators, metering service providers, etc., and the metering needs of different entities may differ.

[0046] It should be noted that the metrological requirements model can be a standardized and structured description model of the metrological requirements of each sub-module, which clarifies the core metrological requirements of each sub-module, including metrological accuracy requirements, metrological cost constraints, metrological efficiency requirements, compliance standards, and other content.

[0047] In some embodiments, the traceability device can divide the overall aircraft system into several sub-modules, and form a measurable requirement model based on the needs of stakeholders in each sub-module from a measurable perspective. The entire process includes conducting requirement analysis on the measurable objects corresponding to the sub-modules, then capturing the requirements, use cases, and performing use case analysis. Based on the requirement capture function, a complete description of the input requirements of the measurable objects is provided. Based on the use case capture function, a complete coverage of the functional requirements or application scenarios of the measurable objects is provided. Based on the use case analysis function, the participants in the use case scenarios and the functions that the participants should perform in completing the use case scenarios are displayed. Requirement capture is divided into functional requirements and non-functional requirements.

[0048] In some embodiments, the traceability device can collect historical demand time series data from stakeholders corresponding to each sub-module, preprocess the historical demand time series data to construct a demand time series, perform time series prediction based on the demand time series, and obtain demand prediction results.

[0049] Step S40: Based on the metrological meta-model and the metrological demand model, perform spatiotemporal design analysis on the aircraft and construct multiple spatiotemporal metrological design models.

[0050] It should be noted that the spatiotemporal metrological design model can be a model that is built based on the spatiotemporal design analysis results and includes specific metrological schemes. It covers detection and / or calibration schemes in both time and space dimensions and can clearly describe the core contents of metrological operation, such as execution process, equipment configuration, time arrangement, spatial layout, and equipment suppliers.

[0051] It should be noted that the time-dimensional testing / calibration scheme is used to clarify the time planning of the metrology scheme, including the metrology sequence of each metrology object, the metrology cycle arrangement, the time node of a single metrology, and the time window adaptation strategy. The spatial-dimensional testing / calibration scheme is used to clarify the spatial planning of metrology operations, including the placement of testing / calibration equipment, the spatial layout of metrology operations, the spatial operation priority of different metrology objects, and the spatial resource allocation strategy.

[0052] In some embodiments, the traceability equipment can be designed and analyzed using a multi-objective optimization algorithm combined with simulation analysis. First, the equipment information and operational logic in the metrological meta-model are correlated and matched with the demand indicators and constraints in the metrological demand model to clarify the core objectives of the design analysis (such as meeting metrological accuracy requirements, improving time efficiency, and reducing space occupation) and spatiotemporal constraints (such as cost limitations and compliance standards). Then, a spatiotemporal optimization model is constructed, with time efficiency and spatial adaptability as the core optimization objectives. Intelligent optimization algorithms (such as particle swarm optimization) are used to solve for the feasible region of the optimal design scheme. Simultaneously, a digital twin technology is used to build an aircraft metrological simulation environment, verifying the optimized design scheme in the simulation environment. The execution effect of the scheme under different operating conditions is analyzed, and problems such as time conflicts and spatial interference are identified. This achieves multi-dimensional optimization of the metrological scheme. Through spatiotemporal design analysis, multiple objectives such as time, space, and demand satisfaction are fully balanced, avoiding the defects caused by single-dimensional design.

[0053] Step S50: Perform a trade-off analysis on the multiple spatiotemporal metrological design models, select the target metrological design scheme, and generate a summary table of aircraft traceability requirements based on the target metrological design scheme.

[0054] It should be noted that the aircraft traceability requirements summary table can be a standardized document generated based on the target metrological design scheme to record aircraft metrological-related traceability information. It covers core contents such as metrological objects, testing / calibration equipment, target suppliers, metrological schemes, requirement fulfillment status, and compliance information. It serves as a carrier to ensure that metrological operations are traceable and verifiable.

[0055] In some embodiments, trade-off analysis can be a process of comprehensively comparing, analyzing and evaluating multiple spatiotemporal econometric design models from multiple evaluation dimensions, and selecting the target econometric design scheme that best meets the core requirements.

[0056] For example, the delivery time, delivery distance, and in-situ or off-site measurement of the aircraft's metrology objects can be used as the basis for a trade-off strategy. The following rules are set: The testing time is divided into three levels: completion within 5 days, completion within 5-10 days, and completion within more than 10 days, and assigned values ​​of 10, 5, and 1 respectively; the testing distance is divided into three levels: round trip within 2 days, round trip within 2-5 days, and round trip within more than 5 days, and assigned values ​​of 10, 5, and 1 respectively; in-situ measurement is assigned a value of 8, and off-site measurement is assigned a value of 0.

[0057] This embodiment solves the problems of inconsistent multimodal data and uneven data quality by extracting the spatiotemporal features of measurement objects of the same measurement type from the processed multimodal data. It achieves accurate spatiotemporal feature extraction, efficiently filters out the core spatiotemporal attributes of measurement objects of the same measurement type, and realizes efficient and accurate spatiotemporal modeling of the metrological meta-model. It effectively avoids the problem of static metrological attributes in metrological schemes. By forecasting the needs of stakeholders of each sub-module of the aircraft, the metrological requirements and constraints of each sub-module are clearly defined. Based on the metrological meta-model and metrological demand model, the aircraft is spatiotemporally designed and modeled, realizing multi-dimensional optimization of metrological design schemes. It fully balances the multi-objective design of time, space, and demand, avoids the scheme defects caused by single-dimensional design, and selects the target metrological design scheme through trade-off analysis. It improves the objectivity of metrological schemes, reduces the influence of subjective factors on scheme selection, reduces the difficulty and maintenance cost of aircraft metrological traceability, and ensures metrological compliance and traceability.

[0058] refer to Figure 3 , Figure 3 This is a flowchart illustrating the second embodiment of the MBSE-based aircraft metrological traceability method of the present invention.

[0059] Based on the first embodiment described above, in this embodiment, step S10 further includes: Step S101: Perform modal sub-processing on the original multimodal data to obtain initial multimodal features.

[0060] It is understood that this embodiment adopts preprocessing and initial feature extraction adapted to the data characteristics of different modalities to achieve preliminary structuring and feature representation of each modal data. After modal processing, modality-specific initial multimodal features are extracted from each single modal data, such as texture and contour features of image modality, temporal change features of sensor data, and semantic unit features of text modality, to avoid information distortion or feature redundancy caused by direct processing of cross-modal data.

[0061] In some embodiments, the initial multimodal features include image modal initial features, sensor modal initial features, text modal initial features, and audio modal initial features.

[0062] Image modal data (such as assembly drawings of aircraft parts and images of equipment appearance) can be extracted using a feature extraction framework based on convolutional neural networks (CNN). First, noise reduction and enhancement preprocessing are completed through techniques such as Gaussian filtering and adaptive histogram equalization. Then, pre-trained CNN models (such as ResNet and VGG) are used to extract shallow texture features and deep semantic features of the image to form the initial features of the image modality.

[0063] Sensing modal data (such as engine speed, pressure sensor monitoring values, and vibration signals) can be preprocessed using time-series data to remove outliers and fill in missing values ​​with linear interpolation. Then, long short-term memory networks (LSTM) or gated recurrent units (GRU) can be used to extract time-series trend features and fluctuation features to generate initial features of the sensing modal data.

[0064] Text modal data (such as design specification documents and metrological standard texts) can be preprocessed using text processing techniques adapted to the field of aircraft metrology. This can be done through word segmentation, stop word removal, and dictionary matching in the field of aircraft technology. Then, word embedding models (such as Word2Vec and BERT) can be used to extract semantic features of the text to form the initial features of the text modality.

[0065] Audio modal data (such as equipment operating noise) can be converted from time-domain signals to frequency-domain signals using short-time Fourier transform (STFT), and spectral features and Mel-frequency cepstral coefficients (MFCC) can be extracted as initial features of the audio modality.

[0066] Step S102: Input the initial multimodal features into the multimodal pre-trained model for multimodal fusion and output a multimodal feature vector of uniform dimension.

[0067] It should be noted that the multimodal pre-trained model, by embedding a modal attention mechanism, adaptively learns the association weights of different modalities, thereby achieving effective fusion of heterogeneous modal features. The multimodal pre-trained model is configured to learn the association weights of different modalities based on the modal attention mechanism, and then fuse the initial multimodal features based on these association weights.

[0068] It should be noted that multimodal feature vectors can be structured feature expressions with a unified dimension and containing cross-modal correlation information, output after fusion by a multimodal pre-trained model. This enables homogeneous representation of data in different forms and provides a foundation for subsequent semantic analysis.

[0069] It is understood that this embodiment constructs a multimodal pre-trained model. The multimodal pre-trained model calculates the contribution of different modal features to the target task (such as feature extraction of aircraft measurement objects) through a modal attention mechanism, assigns differentiated association weights, strengthens the representation of high-value modal information, and suppresses the interference of redundant or noisy modalities.

[0070] In some embodiments, the tracing device may select a pre-trained model (such as the CLIP model) adapted to the cross-modal fusion task as the basic framework and introduce a modal attention mechanism to optimize the fusion effect. Specifically, this includes: constructing an original model; building a modal attention layer on the original model; this layer learns the association weights of different modalities by calculating the mutual information entropy between the initial features of each modality and the measurement target task (such as measurement object recognition); performing dimensional normalization on the initial features of each modality to eliminate dimensional differences; assigning association weights through the modal attention layer; and achieving cross-modal feature alignment by combining weighted summation and feature concatenation; inputting the aligned features into the fusion layer of the model; and mapping them to a unified-dimensional multimodal feature vector through a fully connected network to ensure that the output features contain both the unique information of each modality and the cross-modal association information.

[0071] Step S103: Perform semantic analysis on the multimodal feature vector, and perform multimodal entity recognition and multimodal relation extraction based on the semantic analysis results to obtain multimodal entities and multimodal relations.

[0072] It should be noted that multimodal entity recognition can combine knowledge graphs in the aircraft domain to identify key entities from the fused multimodal features. Multimodal entities can include the aircraft as a whole, functional systems, measurement objects, measurement attributes, etc., to achieve accurate location and extraction of entity information.

[0073] Multimodal relation extraction can be based on multimodal entity recognition, mining and extracting the relationships between different entities. Multimodal relationships can include the subordinate relationship between the aircraft as a whole and its functional systems, the attribution relationship between measurement objects and measurement attributes, etc., constructing a semantic association network between entities.

[0074] In some embodiments, the tracing device may employ a Transformer-based semantic parsing model to perform deep semantic mining on multimodal feature vectors of a unified dimension, identify the domain semantic information contained in the features (such as the measurement attributes of "engine pressure sensor" and the functional affiliation of "avionics system"), establish a mapping relationship between features and domain semantics, and obtain semantic analysis results.

[0075] In some embodiments, the tracing device adjusts a pre-built entity extraction model based on an entity annotation dataset in the aircraft domain to accurately identify core entities from semantic analysis results. These entities include the aircraft as a whole (e.g., a certain type of passenger aircraft or UAV), functional systems (e.g., power systems, avionics systems, navigation systems), measurement objects (e.g., sensors, engines, gyroscopes), and measurement attribute information (e.g., accuracy ±0.1%, measurement cycle 3 months), outputting an entity set. Based on the extracted entity set, a relation extraction model is used, combined with preset relation types in the domain knowledge graph (e.g., subordinate relations, measurement attribute attribution relations, structural association relations, time constraint relations), to mine and extract the associations between different entities, such as "aircraft as a whole - subordinate - power system", "pressure sensor - measurement attribute - accuracy ±0.1%", "engine - structural association - fuel pump", etc., forming a multimodal entity-relationship set.

[0076] Step S104: Generate multiple valid quadruplets based on the multimodal entities and the multimodal relationships.

[0077] It should be noted that an effective quadruple can be a structured data unit with clear semantic logic constructed based on the results of multimodal entity and relation extraction. It contains key entities in the aircraft domain, can accurately represent the association and attribute characteristics between entities, and eliminate invalid or conflicting semantic combinations.

[0078] The entities in the effective quadruple include the aircraft as a whole, the functional system in the aircraft as a whole, the measurement object in the functional system, and the measurement attribute information of the measurement object. The measurement attribute information includes the measurement attribute and the time attribute of the measurement attribute. The measurement object is the functional system or the functional device in the functional system.

[0079] The measurement object can be the functional system itself, which has no subdivided functional devices and is an integral, inseparable module (e.g., an integrated radio transponder system or an integrated emergency positioning system for aircraft); or it can be one or more functional devices contained in the functional system (e.g., fuel flow sensor, engine speed sensor, oil pressure sensor, vibration sensor, etc. in an aircraft power system condition monitoring system, and GPS receiver, Beidou positioning module, attitude sensor, etc. in an aircraft multi-source navigation system).

[0080] It should be noted that the measurement attribute information is the core information used to describe the measurement characteristics of the measurement object. It includes measurement attributes (such as accuracy, range, and error range) and the time attributes of these attributes (such as measurement cycle, effective duration, and data collection timestamp), and is the core basis for spatiotemporal feature extraction.

[0081] In practical implementation, the traceability device can define a unified four-tuple structure based on the semantic logic of the aircraft domain as [head entity, system entity, measurement object entity, measurement attribute information], where the measurement attribute information consists of the measurement attribute and time attribute corresponding to the measurement object entity.

[0082] For example, when the measurement object is equipment, the quaternion is [a certain type of passenger aircraft, power system, temperature sensor, accuracy ±0.1%, measurement cycle 3 months]; when the measurement object is a functional system, the quaternion is [a certain type of general aviation helicopter, integrated radio transponder system, integrated radio transponder system, signal modulation depth error ≤ ±3%, response delay ≤ 20ms, anti-interference capability meets XX standard, measurement cycle 6 months], etc.

[0083] Furthermore, to improve the completeness of quadruples and avoid incomplete hypergraph construction due to missing quadruple entities, and to achieve accurate completion of missing quadruple entities, step S104 above may include: Step S1041: Generate a set of quadruples based on the multimodal entities and the multimodal relationships.

[0084] It should be noted that the set of quadruples includes multiple original quadruples.

[0085] In practical implementation, the traceability device can set up structured mapping rules for four-tuples [head entity, system entity, measurement object entity, measurement attribute information] based on the semantic logic of the aircraft domain, clarify the type range of each entity (including the aircraft as a whole, functional system, measurement object, and measurement attribute information in the multimodal entity set), and classify and integrate all the generated original four-tuples according to entity type (aircraft as a whole, functional system, measurement object, and measurement attribute information) to form a set of four-tuples.

[0086] Step S1042: Traverse the set of quadruples to determine the valid and invalid quadruples in the set of quadruples.

[0087] It should be noted that a valid quadruple can be a quadruple in the quadruple set that conforms to the semantic logic of the aircraft domain after being checked for completeness and semantic consistency. Its core characteristics are entity completeness (including head entity, tail entity, and necessary measurement attribute information), relationship type matching with entity, and no semantic conflict. It can be directly used for subsequent dynamic hypergraph construction.

[0088] An invalid quadruple can be a quadruple in the set of quadruples that has data defects. The core criteria for judgment are that at least one key entity is missing (such as a missing head entity, an empty tail entity, or a missing core field in the measurement attribute information), or there is a semantic conflict where the entity and the relation do not match.

[0089] In some embodiments, the traceability device may pre-build a verification rule base, and based on the verification rule base, traverse the set of quadruplets to determine the valid quadruplets and invalid quadruplets in the set of quadruplets. The verification rule base may include entity integrity verification rules (head entity, system entity, and metering object entity are all indispensable; metering object entity must be associated with complete metering attribute information) and semantic consistency verification rules (whether the metering object entity and system entity types in the quadruplet match, such as "cabin system" and "wind speed sensor" entities do not match semantically because the relationship between the two does not match in the actual architecture).

[0090] Step S1043: Generate a set of candidate entities based on the overall system topology of the aircraft based on the invalid quadruple.

[0091] It should be noted that the candidate entity set is a set of potential completion entities generated based on the aircraft system topology for missing entities in invalid quadruplets. It includes functional systems, measurement objects, and measurement attribute information that have structural or metrological associations with existing entities in the invalid quadruplets. The candidate entity set includes multiple candidate entities, each comprising the overall functional systems of the aircraft, measurement objects, and the metrological attribute information of those measurement objects.

[0092] For example, candidate entities may include various functional systems of the aircraft (such as the fuel system and braking system), measurement objects within the functional systems (such as temperature sensors and speed sensors), and measurement attribute information (such as accuracy level and measurement cycle) corresponding to the measurement objects.

[0093] It should be noted that the overall system topology of an aircraft can be a structured model used to describe the hierarchical affiliation and relationship between various functional systems (such as power systems, avionics systems, and navigation systems) and measurement objects (such as sensors, engines, and gyroscopes) within a functional system. It covers the hierarchical architecture of system-subsystem-equipment-measurement object, as well as the cross-system correlation logic (such as signal transmission and energy supply correlation).

[0094] In practical implementation, the traceability device can combine the system topology and domain knowledge graph to generate a targeted set of candidate entities (adapted to the four-tuple structure [head entity, system entity, measurement object entity, measurement attribute information]).

[0095] First, load the overall system topology model of the aircraft and parse the hierarchical position and associated entities of existing entities (i.e., existing head entities, system entities, or measurement object entities) in the invalid quadruple. If the existing entity contains a head entity (such as "a certain type of passenger aircraft"), then parse its subordinate system entities (such as power system, avionics system); if the existing entity contains a system entity (such as "power system"), then parse the measurement object entities it contains (such as pressure sensor, engine).

[0096] Then, based on the topology analysis results and the missing entity types of invalid quadruples (such as missing system entities, missing measurement object entities, or missing measurement attribute information), related entities are screened from the multimodal entity library. For missing system entities, the system entities corresponding to the head entity are screened; for missing measurement object entities, the measurement object entities corresponding to the system entities are screened; and for missing measurement attribute information, the measurement attribute information (such as accuracy level and measurement period) of the corresponding measurement object entities is supplemented by combining the measurement domain dictionary, thus forming an initial candidate entity pool.

[0097] Finally, entities in the initial candidate entity pool that have no hierarchical or measurement association with the existing entities of the invalid quadruple are removed (e.g., if the existing entity is "avionics system", remove measurement objects such as "fuel pump" that do not belong to the avionics system). This results in a final candidate entity set, ensuring that all entities in the set are adaptation and completion options for the missing fields of the invalid quadruple, and that they are feasible for completion.

[0098] Step S1044: Perform semantic analysis on the invalid quadruples, and select multiple entities to be scored from the candidate entity set based on the semantic analysis results.

[0099] It should be noted that the entities to be scored can be a subset of entities that are highly semantically compatible with the invalid quadruple after semantic analysis of the invalid quadruple. This subset is used to eliminate candidate entities that have semantic conflicts with or have very low relevance to the existing entities in the invalid quadruple, thus narrowing the selection range of the supplementary entities.

[0100] In some embodiments, the traceability device can configure a semantic analysis model based on a thesaurus of aircraft metrology to achieve accurate screening of candidate entities. For example, it can construct a domain semantic analysis model based on Transformer, and train and optimize the model based on aircraft metrology corpus. It can perform deep semantic analysis on existing entities and relation types of invalid quadruples to extract core semantic features (e.g., the semantic features of the existing entity "pressure sensor" are "power system, metrology object, precision sensitive", and the semantic feature of the relation type "metrology attribute affiliation" is "associated metrology attribute information"). Then, it can calculate the matching degree between each candidate entity in the candidate entity set and the core semantic features of the invalid quadruples, and use cosine similarity to quantify the matching degree. By setting a semantic matching degree threshold (e.g., ≥0.6), it can screen out candidate entities with matching degrees higher than the threshold to form a set of entities to be scored, and eliminate candidate entities with semantic conflicts or low matching degrees (e.g., the core semantics of the invalid quadruples is "metrology attribute association", eliminating candidate entities of the "structural component class") to narrow down the subsequent scoring range.

[0101] Step S1045: Analyze the structural and / or quantitative correlations between the entities currently existing in the invalid quadruple and each entity to be scored, and obtain the correlation analysis results.

[0102] It should be noted that structural correlation is used to describe the degree of hierarchical or assembly correlation between existing entities and candidate entities in an invalid quadruple, based on the aircraft system topology. For example, if the existing entity is "engine" (power system), then "fuel pump" (assembly correlation within the power system) and "avionics controller" (cross-system signal correlation) both have structural correlation. The correlation strength is determined by the hierarchical distance of the topology and the correlation type.

[0103] Metering correlation can be used to describe the degree of correlation between existing entities and candidate entities in an invalid quadruple based on metering attributes. For example, if the existing entity is "pressure sensor" (metering attribute: accuracy ±0.1%), then "temperature sensor of the same accuracy level" and "flow sensor that needs to be calibrated synchronously" both have metering correlation. The correlation strength is determined by the matching degree of metering attributes and the correlation of metering process.

[0104] It should be noted that the correlation analysis results can be standardized analysis data obtained by quantifying the structural correlation strength and / or quantitative correlation strength between existing entities in invalid quadruples and each entity to be scored. The results can be presented in the form of correlation strength score or correlation level (high / medium / low) to provide a basis for entity scoring.

[0105] In practical implementation, the results of correlation analysis can include structural correlation analysis results and quantitative correlation analysis results. Structural correlation analysis can be based on the aircraft system topology, calculating the hierarchical distance between existing entities and each entity to be scored (e.g., the hierarchical distance for direct subordinate relationships is 1, and the hierarchical distance for cross-system indirect relationships is 2), and combining the correlation type weights (direct relationships have a higher weight than indirect relationships) to quantify the structural correlation strength score.

[0106] Metrological correlation analysis can extract metrological attribute information (such as accuracy level, metrological cycle, calibration standard) between existing entities and each entity to be scored, calculate the matching degree of metrological attributes (such as a matching degree of 1 for the same accuracy level and a matching degree of 0.6 for similar accuracy), combine the metrological process correlation (such as a correlation weight of 0.8 for synchronous calibration and a correlation weight of 0.3 for asynchronous calibration), and quantify the strength score of metrological correlation.

[0107] If the invalid quadruple involves structural relationships, the structural correlation analysis result is output; if it involves quantitative relationships, the quantitative correlation analysis result is output; if it involves mixed relationships, the two-dimensional scores are integrated to form a standardized correlation analysis result (e.g., structural correlation score 0.8, quantitative correlation score 0.7, and overall score 1.5).

[0108] Step S1046: Based on the correlation analysis results, score each entity to be scored, and based on the scoring results, select the target completion entity from the entities to be scored.

[0109] It should be noted that the target completion entity can be the entity with the highest score and the best fit with the existing entities of the invalid quadruple after quantitative scoring of each entity to be scored based on the correlation analysis results. This entity is used to fill in the missing part of the entity in the invalid quadruple, so that the invalid quadruple is transformed into a valid quadruple.

[0110] In some embodiments, the traceability device can set a scoring index system based on the correlation analysis results. The structural correlation score and the quantitative correlation score are weighted according to the relationship type of invalid quadruples (e.g., structural correlation weight 0.8 and quantitative correlation weight 0.2 for structural relationships; the opposite is true for quantitative relationships). The weighted summation formula is used to calculate the comprehensive score of each entity to be scored. The entities to be scored are sorted in descending order of comprehensive score. A scoring threshold (e.g., ≥0.7) is set, and the entity with the highest score that meets the threshold requirement is selected as the target completion entity. If there are multiple entities with the same score, further screening is carried out by combining domain expert rules (e.g., prioritizing entities with complete quantitative attributes).

[0111] Step S1047: Embed the target completion entity into the invalid quad, and mark the invalid quad after entity embedding as a valid quad.

[0112] In practical implementation, the traceability device can accurately embed the target completion entity into the corresponding position according to the missing entity type of the invalid quadruple (e.g., if the head entity is missing, it is embedded into the head entity field; if the measurement attribute information is missing, it is embedded into the measurement attribute information field). For example, embedding "pressure sensor" into the invalid quadruple "(aircraft A, power system, empty, accuracy ±0.1%)" which is missing the tail entity, results in "(aircraft A, power system, pressure sensor, accuracy ±0.1%)".

[0113] Step S105: Construct a dynamic hypergraph based on the effective quadruple, and extract the spatiotemporal features of measurement objects of the same measurement type in the whole spacecraft based on the dynamic hypergraph.

[0114] It should be noted that dynamic hypergraphs differ from traditional graph structures. They can form complex network models by connecting multiple nodes through hyperedges. Combined with the dynamic characteristics of the time dimension, they can accurately represent the complex relationships between multiple entities in the field of aircraft. Their nodes are entities in effective quadruplets, and hyperedges connect multiple entities through different types of associations. Moreover, their attributes can be dynamically updated over time.

[0115] It should be noted that the nodes in the dynamic hypergraph are entities in the effective quadruples. The hyperedges of the dynamic hypergraph include structural association hyperedges, measurement association hyperedges, and temporal association hyperedges. The structural association hyperedges are used to connect measurement objects with structural associations in the same system or across systems in the whole aircraft. The measurement association hyperedges are used to connect measurement objects with attribute associations in the same system or across systems in the whole aircraft. The temporal association hyperedges are used to connect measurement objects with temporal dimension associations.

[0116] It should be noted that structural correlation hyperedges are used to connect measurement objects with the same functional systems or cross-systems that have structural correlations within the overall aircraft, such as the assembly structural correlation between the engine and fuel pump in the power system, and the signal transmission structural correlation between the avionics system and the navigation system.

[0117] Metrological association hyperedges are used to connect metrological objects with the same functional systems or across systems within an aircraft that have metrological attribute associations, such as sensors of the same accuracy level or equipment groups that need to be calibrated synchronously.

[0118] Time-related hyperedges are used to connect measurement objects that have time-related relationships, such as devices with the same measurement cycle or objects with measurement time constraints (testing before calibration), reflecting the time-related characteristics of measurement operations.

[0119] In some embodiments, the traceability device uses entities in the effective quadruple as hypergraph nodes and constructs three types of hyperedges based on the relationships between entities: structural association hyperedges connect measurement objects with structural relationships (such as associated devices of engines and fuel pumps, avionics systems and navigation systems), measurement association hyperedges connect measurement objects with measurement attribute relationships (such as sensors with the same accuracy level and calibration standard), and time association hyperedges connect measurement objects with time dimension relationships (such as devices with the same measurement period and time constraints). At the same time, a timestamp mechanism is introduced so that the hypergraph can be dynamically updated with the time attributes of the measurement attributes (such as time association hyperedges being updated synchronously after the measurement period is adjusted), forming a dynamic hypergraph model.

[0120] The traceability equipment can classify metrological objects (such as precision metrology, performance metrology, safety metrology, lifespan metrology, etc.) according to metrological purpose and object attributes based on the metrological attribute information in the effective quadruple set, and filter out a set of metrological objects of the same metrological type. It uses a hypergraph convolutional neural network (HGCN) to learn features of the dynamic hypergraph, and captures the structural features and metrological association features of nodes (metrological objects) under the three types of hyperedge association through hypergraph convolution operations. At the same time, it extracts time constraint features (such as metrological period, time window, and time series dependency) by combining the dynamic characteristics of the hypergraph (time-related hyperedge). Finally, it outputs a set of spatiotemporal features of metrological objects of the same metrological type, including spatial constraints (such as structural layout and installation location association) and time constraints (such as metrological period and time series dependency).

[0121] This embodiment constructs quadruples and dynamic hypergraphs to accurately represent complex relationships and time dimensions, effectively captures the spatiotemporal constraint information of measurement objects of the same measurement type, and realizes in-depth mining of the spatiotemporal dynamics of measurement characteristics, laying a reliable foundation for subsequent measurement modeling, demand analysis and other stages.

[0122] refer to Figure 4 , Figure 4 This is a flowchart illustrating the third embodiment of the MBSE-based aircraft metrological traceability method of the present invention.

[0123] Based on the above embodiments, in this embodiment, step S30 further includes: Step S301: Divide the aircraft into multiple sub-modules based on the overall system architecture, structure type, usage scenario and / or users.

[0124] In practical implementation, the traceability equipment can divide the entire aircraft into n sub-modules according to classification methods such as aircraft system architecture, structural type, usage scenario, and users.

[0125] For example, aircraft systems can be classified into external fuselage and wing structure systems, hydraulic systems, flight control systems, landing gear systems, cockpit environmental control systems, fuel systems, fire protection systems, etc.

[0126] Based on the usage scenarios of aircraft, they can be classified into the stationary phase, taxiing phase, takeoff phase, climb phase, cruise phase, descent phase, landing phase, taxiing phase, maintenance phase, and repair phase.

[0127] Based on the personnel used by the aircraft, they can be divided into pilots, flight attendants, cargo handling teams, ground control teams, aircraft support teams, aircraft maintenance teams, and aircraft research and development teams.

[0128] Based on the structure of the aircraft, it can be classified into main structure, power system, landing gear, airborne system, interior and protective structure, etc.

[0129] Understandably, this embodiment significantly improves system manageability through structured decomposition. The independence of sub-modules facilitates subsequent targeted analysis and reduces overall complexity; multi-dimensional classification ensures coverage of requirements throughout the entire lifecycle, avoiding omissions of key scenarios; at the same time, clear module boundaries lay the foundation for accurate collection of requirement time-series data and reduce cross-module interference.

[0130] Step S302: Obtain demand time series data based on the stakeholders corresponding to each sub-module, and preprocess the demand time series data to obtain candidate time series data.

[0131] It should be noted that in this embodiment, the aircraft is divided into multiple sub-modules, each corresponding to specific requirements. Stakeholders for each requirement can originate from users, relevant standards or policies, technical specifications, etc. These stakeholder requirements must include the overall technical specifications for the aircraft's development.

[0132] Taking an aircraft's GPS antenna as an example, in a communication system, the stakeholders of the antenna can be the design department (S1), the research and development department (S2), the user department (S3), the metrology department (S4), and the maintenance and support department (S5). S1 to S5 are corresponding department codes for easy reference in subsequent designs. The design department (S1) proposes requirements for antenna weight control, the user department (S3) proposes requirements for gain, VSWR, and bandwidth, and the metrology department (S4) proposes requirements for metrological traceability rate, metrological equipment standardization rate, and the classification of testing equipment by importance level.

[0133] It should be noted that demand time-series data can be serialized data containing demand content and its occurrence timestamps, reflecting the dynamic changes of demand over time. The demand time-series data includes stakeholder demands and the timestamps of those demands, and the preprocessing includes normalization.

[0134] It is understood that this embodiment can use linear transformation algorithms (such as min-max normalization or Z-score normalization) to transform demand indicators with different dimensions and different numerical ranges (such as measurement accuracy error, calibration cycle, cost budget) to a unified numerical range (such as [0,1] or [-1,1]), thereby eliminating the interference of dimension differences on model training.

[0135] In some embodiments, to adapt to the input standards of the demand time series forecasting model, the traceability device can perform preprocessing on the original demand time series data, including data cleaning, format unification, missing value filling, outlier removal and normalization, with the aim of eliminating data noise, unifying data format and improving data quality.

[0136] Step S303: Input the candidate time series data into the pre-built demand time series prediction model to obtain the demand prediction result.

[0137] It should be noted that the demand time series prediction model can be a pre-built and trained deep learning prediction model designed for the time series characteristics of aircraft sub-module metering demand. Through multi-branch feature extraction, adaptive feature fusion and multi-task decoding, it can achieve accurate prediction of sub-module metering demand in the future, and is corely adapted to the long-term trend and short-term fluctuation characteristics of aircraft metering demand.

[0138] It should be noted that, referring to Figure 5 , Figure 5 This is a schematic diagram of the demand time series prediction model in one embodiment. The demand time series prediction model includes a channel splitting structure, a multi-feature extraction branch structure, a target feature fusion structure, and a multi-task time series decoding layer. The channel splitting result is used to divide the input candidate time series data, and the divided data streams are respectively input to each feature extraction branch structure. The multi-feature extraction branch structure is used to extract global demand features and local demand features from the candidate time series data. The target feature fusion structure is used to fuse the global demand features and the local demand features to output target fused features. The multi-task time series decoding layer is used to decode the target fused features to output the demand prediction result.

[0139] Understandably, the multi-feature extraction branch structure is the core feature extraction component of the demand time series prediction model, containing two parallel independent branches: one is the global demand feature extraction branch (which can adopt the Transformer architecture), which captures long-term global features in candidate time series data (such as demand changes driven by annual metrology standard updates and the evolution trend of demand throughout the entire life cycle) through a self-attention mechanism; the other is the local demand feature extraction branch (for example, a multi-branch architecture that can adopt a hybrid of LSTM and CNN convolutional networks), which captures short-term local features (such as calibration cycle adjustments caused by quarterly maintenance pressure and temporary metrology needs caused by sudden failures) through gating units.

[0140] It should be noted that global demand characteristics can reflect the overall demand change trend and long-term demand patterns in candidate time series data, covering macro-level characteristics such as total demand and demand type distribution trends in different time periods.

[0141] Local demand features can reflect the details of short-term demand fluctuations, sudden changes in demand at specific time points, and changes in single demand types in candidate time series data, covering micro-features such as demand peaks within a certain time period and the correlation characteristics of the causes of demand changes.

[0142] Furthermore, in order to improve the accuracy of demand forecasting and accurately mine demand characteristics in the spatiotemporal dimensions, refer to Figure 6 , Figure 6 This is a schematic diagram of a multi-feature extraction branch structure in one embodiment. In one embodiment, the feature extraction branch structure includes a local feature extraction structure and a global feature extraction structure. The local feature extraction structure includes a channel splitting structure, a first branch structure, a second branch structure, a third branch structure, and a candidate feature fusion module. The global feature extraction structure includes a fourth branch structure. The channel splitting structure is configured to perform semantic awareness analysis on candidate time series data to obtain semantic awareness features, and divide the candidate time series data into multiple channel data based on the semantic awareness features, and input the channel data into the first branch structure, the second branch structure and the third branch structure respectively based on the semantic awareness features corresponding to each channel data. The first branch structure is built on a Bi-LSTM network and is configured to extract bidirectional time-series features based on channel data, and convert the bidirectional time-series features into a query vector matrix; The second branch structure is built based on a graph attention network and is configured to extract the demand association features between each sub-module based on channel data, and convert the demand association features into a key vector matrix; The third branch structure is constructed based on a spatiotemporal graph convolutional network and is configured to extract spatiotemporal constraint features based on channel data and convert the spatiotemporal constraint features into a value vector matrix. The candidate feature fusion module is configured to split the query vector matrix, key vector matrix, and value vector matrix into multiple sub-matrices along the feature dimension, generate multiple attention heads based on the sub-matrices, calculate cross-branch attention scores through the multiple attention heads, concatenate the output results of each attention head, map the concatenation result to the original dimension, and output local demand features. The fourth branch structure is built based on the Transformer model and is configured to extract global requirement features of the entire life cycle of the aircraft based on candidate time series data.

[0143] It should be noted that, considering the coupling characteristics between aircraft sub-modules, this embodiment uses a differentiated network to match data from different semantic channels, and employs a three-branch local feature extraction structure for local feature extraction: The first branch uses a bidirectional long short-term memory (Bi-LSTM) network to capture the bidirectional dependencies of time-series data. It can effectively model the sequential correlation of requirements within a single submodule (such as the impact of early maintenance requirements of a certain type of engine on later procurement requirements). By directly connecting the output query vector matrix to the attention mechanism, it can model the time-dimensional requirements of the submodule.

[0144] The second branch employs Graph Attention Networks (GAT), which can quantify the strength of the demand correlation between sub-modules through attention weights (such as the coupling degree between the upgrade requirements of the navigation sub-module and the adaptation requirements of the flight control sub-module), and generate a key vector matrix to provide a basis for cross-branch feature interaction, thereby realizing the modeling of the spatial dimension demand features of sub-modules.

[0145] The third branch employs a Spatial-Temporal Graph Convolutional Network (STGCN) to simultaneously model the temporal evolution of the time dimension and the physical / logical relationships of modules in the spatial dimension. This allows it to capture the spatiotemporal constraints of aircraft requirements (such as the synchronous impact of flight scenario restrictions in a certain airspace on the requirements of multiple sub-modules). The generated value vector matrix provides spatiotemporal constraint information for feature fusion, enabling simultaneous modeling of the requirement features of sub-modules in both the time and spatial dimensions.

[0146] It should be noted that the time span of the demand for the entire life cycle of an aircraft (research and development, flight testing, operation and maintenance, and decommissioning) is large (up to several years to several decades). In this embodiment, the Transformer is used to construct the first branch structure for global feature extraction. The self-attention mechanism of the Transformer can overcome the gradient vanishing bottleneck of LSTM-like networks and achieve accurate modeling of long-sequence global trends (such as the long-term impact of the demand distribution in the overall aircraft development stage on the demand type in the operation and maintenance stage), thereby adapting to the demand for global demand feature extraction.

[0147] It is understandable that this embodiment performs semantic classification (such as "functional upgrade requirements", "fault maintenance requirements", "compliance adaptation requirements" etc.) and channel data division on candidate time series data through semantic awareness analysis, and then diverts channel data with different semantics to the appropriate feature extraction branch. Essentially, it is a combination of prior business knowledge and data-driven feature extraction, which can greatly improve the targeting of feature extraction. For example, semantic awareness and channel division can be achieved through a lightweight semantic classifier (such as a pre-trained small TransformerEncoder or CNN).

[0148] It should be understood that this embodiment uses cross-branch multi-head attention fusion to split the Q, K, V matrices (i.e., query vector matrix, key vector matrix, and value vector matrix) into sub-matrices to generate multiple attention heads, thereby achieving fine-grained cross-branch feature interaction (such as weighted alignment of bidirectional temporal features and module-related features). Then, the feature dimensions are restored by concatenation and dimension mapping, which can avoid the information bottleneck of single-head attention and achieve deep coupling of multi-branch features.

[0149] In its implementation, the demand time series prediction model employs a parallel architecture of local and global feature extraction for demand feature extraction. The dual-branch parallel structure works together to achieve comprehensive mining of candidate time series data features. In the local feature extraction branch structure, the candidate time series data is first preprocessed by the channel splitting structure. Semantic perception analysis is performed by a semantic analysis model (such as the BERT lightweight model) to extract semantic perception features from the data and obtain the required attributes and correlation dimensions of each data segment. Based on the similarity and attribute classification of the semantic perception features, the candidate time series data is divided into multiple channel data. Then, according to the semantic attributes of each channel data, it is distributed to the first branch structure, the second branch structure and the third branch structure respectively.

[0150] The first branch structure is built on a Bi-LSTM network. It performs bidirectional temporal traversal on the input channel data, captures the evolutionary relationship between the past and future of the data, extracts bidirectional temporal features, and then converts the features into a fixed-dimensional query vector matrix through a fully connected layer.

[0151] The second branch structure is built based on a graph attention network. It constructs a graph structure with each sub-module as a node and the demand association as an edge. After inputting channel data, it calculates the demand dependency relationship between nodes through attention weights and extracts the demand association features. It is also converted into a key vector matrix through a fully connected layer.

[0152] The third branch structure is built on a spatiotemporal graph convolutional network. It combines the evolution of demand in the time dimension with the distribution of sub-modules in the spatial dimension to construct a spatiotemporal graph. After inputting channel data, it extracts spatiotemporal constraint features and converts them into a value vector matrix through a fully connected layer.

[0153] After the first, second, and third branch structures are output, the candidate feature fusion module integrates the features: the query vector matrix, key vector matrix, and value vector matrix are evenly split into multiple sub-matrices along the feature dimensions, and each sub-matrix corresponds to an attention head; cross-branch attention scores are calculated for each attention head to filter out key correlation information between features of different branches; the outputs of all attention heads are concatenated, and then the concatenated features are mapped back to the original dimensions through a dimension mapping layer to finally output the local required features.

[0154] Global feature extraction is handled by the fourth branch of the global feature extraction structure. This branch is built on the Transformer model and introduces a multi-head attention mechanism and positional encoding technology. Candidate time-series data are sequentially encoded according to time order and then input into the model. The multi-head attention mechanism captures the demand correlations across different time spans and uncovers the demand evolution patterns throughout the entire lifecycle of aircraft design, production, operation, maintenance, and upgrades. After the features are enhanced and filtered by the feedforward neural network, global demand features reflecting the overall demand trend are extracted.

[0155] The target feature fusion structure can use an attention mechanism or a combination of feature concatenation and fully connected layers to fuse global and local demand features, select key features that contribute significantly to the prediction results, and output the target fusion features. The multi-task temporal decoding layer (e.g., it can be built based on a multilayer perceptron (MLP)) performs tasks such as demand type classification, demand intensity regression, and demand time prediction in parallel for the target fusion features to obtain the final demand prediction results.

[0156] Step S304: Perform architecture analysis, behavior analysis, and indicator analysis on each submodule to obtain multidimensional analysis results for each submodule.

[0157] It should be noted that architecture analysis can be conducted on the internal structure, component composition, interface design, and interaction with other sub-modules of each sub-module, focusing on the architectural rationality and compatibility of the sub-module. For example, a GPS antenna consists of three parts: a passive antenna, a low-noise amplifier module, and a connector.

[0158] In some embodiments, architectural analysis begins with functional grouping and allocation, as well as the definition of logical components. Then, the logical architecture is designed, and interface characteristics are developed. Finally, a highly secure system logical architecture scheme is formed through continuous iteration.

[0159] It should be noted that behavioral analysis can be an analysis of the behavioral characteristics of each submodule under different usage scenarios, such as its running status, function execution process, response speed, and fault behavior, focusing on the functional implementation capability and stability of the submodule.

[0160] In some embodiments, behavioral analysis can begin by creating a behavioral analysis activity diagram, identifying the control flow between sub-functional blocks, sorting them according to their coupling relationships, and identifying and analyzing other possible functional paths. Then, the interface interaction relationships between functional blocks are clarified, forming a functional interface block diagram. Finally, based on the functional control flow in the functional flow diagram, the interaction content between functional blocks is clarified, forming a functional interface block diagram and functional interface files, thus completing the functional behavioral analysis.

[0161] For example, after capturing GPS antenna use cases, a series of behavioral use cases will be captured, the most important of which is to transmit or receive communication signals. Taking the implementation of communication signal reception as an example, the main components closely related to this process are passive antennas, low-noise amplifier modules, and connectors.

[0162] In some embodiments, the traceability device builds a submodule simulation test environment to simulate the operating status of the submodule under different usage scenarios (such as extreme weather, high load operation), collects real-time operating data of the submodule (such as response time, execution success rate) through sensors, retrieves historical operating logs and fault reports, statistically analyzes the fault types, fault occurrence frequency and fault handling effects of the submodule, analyzes the behavioral characteristics and stability performance of the submodule under different scenarios, and obtains behavioral analysis results.

[0163] It should be noted that the index analysis can be an analysis of the core performance indicators (such as the thrust of the power submodule and the control accuracy of the flight control submodule), reliability indicators (such as mean time between failures), and safety indicators (such as fault tolerance) of each submodule, with a focus on the submodule's index compliance and optimization potential.

[0164] In some embodiments, index analysis can first involve inheriting, decomposing, and assigning overall indices to various systems. These indices are then described using system and device / subsystem mean-value attributes, and parametric diagrams are used to depict the relationships between overall system indices and individual system indices. By creating new module instances, simulation analysis is performed on the system indices to obtain the parameters and indices decomposed and assigned to devices or subsystems.

[0165] For example, a GPS antenna is divided into a passive antenna, a low-noise amplifier module, and a connector. The value attributes of the GPS antenna are weight, gain, VSWR, and bandwidth. The value attributes of the GPS antenna will be decomposed or assigned to the passive antenna, the low-noise amplifier module, and the connector. The value attributes of the passive antenna include weight, the value attributes of the low-noise amplifier module include weight, gain, VSWR, and bandwidth, and the value attributes of the connector include weight and bandwidth.

[0166] Step S305: Capture the demand forecast results based on the multidimensional analysis results, and construct a quantitative demand model based on the demand capture results.

[0167] It should be understood that this embodiment can screen and confirm effective requirements that meet the actual capabilities and future development needs of the sub-modules based on multi-dimensional analysis results and demand forecast results, eliminate invalid and unreasonable requirements, and clarify the priority of requirements and implementation constraints, thereby obtaining the requirement capture results. The requirement capture results may include key information such as requirement content, requirement priority, implementation constraints (such as technical constraints, cost constraints, and time constraints), corresponding sub-modules, and stakeholders.

[0168] It should be noted that a quantitative demand model can be a quantifiable and assessable demand model formed by defining quantitative indicators and constructing mathematical relationships based on the demand capture results. It can clearly present the quantitative correlation between demands and the correspondence between demands and sub-module indicators.

[0169] In some embodiments, requirement capture is divided into functional requirements and non-functional requirements; that is, stakeholder requirements are divided into functional and non-functional, and requirements are defined from a measurable perspective. Functional requirements are analyzed using behavioral use cases through use case capture, while non-functional requirements are analyzed using metrics. Based on the requirements capture results, modeling tools are used to model the requirements, utilizing system context to capture the requirements and establish traceability relationships, including the source of the requirements analysis and the impact of requirement changes, thereby constructing a measurable requirements model.

[0170] Understandably, this embodiment divides the overall aircraft system into several sub-modules, and based on the needs of stakeholders in each sub-module, forms a measurable demand model from a measurable perspective. The entire process includes conducting demand analysis for the measurable objects corresponding to the sub-modules, collecting and preprocessing the time-series demand data of stakeholders, inputting the processed candidate time-series data into a pre-built demand time-series prediction model for demand prediction, obtaining demand prediction results, and realizing demand prediction from both time and spatial dimensions. This accurately captures global and local demand characteristics, and fully covers the application scenarios throughout the entire lifecycle of measurable demand.

[0171] Because this embodiment uses a multi-layered architecture for metering demand analysis, it decomposes sub-modules from the business layer, performs data preprocessing from the data layer, and extracts global and local features from multiple branches in the model layer. Then, it fuses and decodes the features from multiple branches to output the results, realizing a full-link design of the business layer, data layer, and model layer. This effectively adapts to the demand analysis business scenario of aircraft metering, effectively uncovers potential metering needs, and improves the accuracy of demand prediction and capture.

[0172] refer to Figure 7 , Figure 7 This is a flowchart illustrating the fourth embodiment of the MBSE-based aircraft metrological traceability method of the present invention.

[0173] Based on the above embodiments, in this embodiment, step S50 further includes: Step S501: Encode the features of each spatiotemporal econometric design model from multiple trade-off feature dimensions to obtain multi-dimensional econometric trade-off features.

[0174] It should be noted that the measurement trade-off features are multi-dimensional feature data obtained by encoding the spatiotemporal measurement design model based on the trade-off feature dimensions. These features are used to quantitatively describe the model's performance in each trade-off dimension, and primarily include four categories of features: measurement cost, time, space, and method. The multi-dimensional measurement trade-off features include the measurement cost characteristics, measurement time characteristics, measurement space characteristics, and measurement method characteristics of each measurement object in the spatiotemporal measurement design model. The measurement method characteristics include off-site measurement features and in-situ measurement features.

[0175] It should be noted that the measurement cost characteristics are used to reflect the characteristics of the resources (such as manpower, equipment, and funds) required for each measurement object to carry out measurement operations in the spatiotemporal measurement design model, and cover quantitative information such as single measurement cost and long-term maintenance cost.

[0176] Metering time characteristics are used to reflect the time required for each metering object to carry out metering operations, including time-related information such as metering preparation time, single metering duration, and metering cycle.

[0177] Metering space characteristics are used to reflect the physical or virtual space required for each metering object to carry out metering operations, including space-related information such as the space occupied by metering equipment, data storage space, and metering operation site requirements.

[0178] The metrology method characteristics are used to reflect the characteristics of the metrology implementation mode adopted by each metrology object. The core is divided into two categories: off-site metrology (metrology is carried out after the metrology object is removed from the aircraft) and in-situ metrology (metrology is carried out in the original installation state of the aircraft without removing the metrology object).

[0179] In some embodiments, for each spatiotemporal metrological design model, the original data of each metrological object under multiple trade-off feature dimensions are extracted one by one. Differentiated coding methods are adopted for different types of metrological trade-off features: numerical features such as metrological cost, time, and space are converted to a unified numerical range using the min-max normalization method to eliminate dimensional differences; categorical features such as metrological methods are converted into structured data using label coding (e.g., 0 for out-of-situ metrology and 1 for in-situ metrology) or one-hot coding; finally, the encoded features of each dimension are integrated to form a multi-dimensional metrological trade-off feature matrix to ensure that the feature format is uniform and adaptable to subsequent processing.

[0180] Step S502: Based on the multi-dimensional metrological trade-off characteristics, the spatiotemporal metrological design model is converted into a metrological graph structure.

[0181] It should be noted that the econometric graph structure can be a structured data form built based on graph theory, used to structurally model the relationships between various econometric objects in a spatiotemporal econometric design model, including three core elements: nodes, node attributes, and edges. The nodes in the econometric graph structure represent the econometric objects in the spatiotemporal econometric design model, the node attributes of each node are the multi-dimensional econometric trade-off features, and the edges of the econometric graph structure include structural dependencies, econometric dependencies, and temporal dependencies between nodes.

[0182] It should be noted that structural dependency refers to the dependency relationship between different metering objects in terms of physical structure or system architecture (e.g., a metering object needs to rely on the installation location of other objects for metering). Metering dependency refers to the sequential or collaborative relationship between different metering objects in the metering process (e.g., the metering of one object needs to be based on the metering results of another object), and metering dependency includes direct dependency and indirect dependency. Time dependency refers to the constraint relationship between different metering objects in terms of metering time (e.g., one object needs to be metered synchronously with another object or sequentially at fixed time intervals).

[0183] In some embodiments, each measurement object in the spatiotemporal econometric design model is taken as an independent node, and the multi-dimensional measurement trade-off features corresponding to each node are taken as node attributes. The edges of the graph structure are constructed based on the relationship between measurement objects: structural dependency edges are constructed based on physical / architectural relationships, measurement dependency edges are constructed based on measurement relationships, and time dependency edges are constructed based on measurement time constraints. A correlation strength weight is added to each edge (e.g., a strong dependency weight is set to 1, and a weak dependency weight is set to 0.5) to clarify the degree of correlation between nodes. The above nodes, node attributes, edges and edge weights are integrated by a graph structure construction tool to generate a standardized econometric graph structure.

[0184] Step S503: Input the metrology graph structure into the graph attention network to obtain the node embedding vector of each metrology object.

[0185] It should be noted that the node embedding vector is a vector representation generated for each node (metric object) after the graph attention network processes the econometric graph structure. It can contain the node's own attributes and the relationship information with other nodes.

[0186] In some embodiments, the tracing device can pre-build a graph attention network model, using the generated metrology graph structure as model input; configure core model parameters (such as the number of attention heads, hidden layer dimension, learning rate, etc.), using node attributes as basic features, and calculate the attention weights of each node and its neighboring nodes through the model's attention mechanism to strengthen the feature contributions of key related nodes; train the model using the stochastic gradient descent algorithm, using the node feature reconstruction error as the loss function, and iteratively optimize the model parameters until the model converges; after the model training is completed, extract the output vector corresponding to each metrology object, which is the node embedding vector containing the node's own features and related features.

[0187] Step S504: Configure the measurement weights corresponding to the node embedding vectors of each measurement object according to the usage scenarios of each stage of the overall life cycle of the aircraft, and aggregate the node embedding vectors of each measurement object based on the measurement weights to obtain the global measurement features of each spatiotemporal measurement design model.

[0188] It should be noted that the measurement weights can be importance coefficients configured for the node embedding vectors of each measurement object based on the usage scenario requirements of different life stages of the aircraft, used to reflect the priority differences of measurement objects under different scenarios. The aforementioned global measurement features can be comprehensive features obtained by weighted aggregation of the node embedding vectors of each measurement object, which can comprehensively reflect the overall performance and adaptability of the spatiotemporal measurement design model.

[0189] In some embodiments, the traceability device can acquire the core usage scenarios of the entire life cycle of the aircraft (such as R&D design, production and manufacturing, operation and maintenance support, and scrapping and recycling), and combine the priority of the measurement requirements of each scenario. It can then use the analytic hierarchy process or expert scoring method to configure measurement weights for each measurement object (e.g., in the operation and maintenance support scenario, the measurement object of the power submodule has a higher weight than that in the R&D stage). The device can then perform weighted calculations on the node embedding vectors of each measurement object and their corresponding measurement weights to obtain the weighted node features. Finally, it can use feature splicing or mean aggregation to integrate all the weighted node features into global measurement features.

[0190] Step S505: Based on the global metrological characteristics, select the target metrological design scheme from the metrological design schemes corresponding to the multiple spatiotemporal metrological design models, and generate a summary table of aircraft traceability requirements based on the target metrological design scheme.

[0191] It should be noted that the target metrological design scheme can be the optimal scheme that best meets the full life-cycle usage requirements of the aircraft, selected through a trade-off analysis from multiple spatiotemporal metrological design models. The aforementioned aircraft traceability requirements summary table can be structured data generated based on the target metrological design scheme, used to summarize the traceability-related requirements of various metrological objects of the aircraft, covering core information such as metrological standards, traceability paths, responsible entities, and time requirements.

[0192] In some embodiments, the traceability device can construct a scheme evaluation system, using global metrological characteristics as the core indicator, combined with auxiliary indicators such as demand matching degree, technical feasibility, and economic efficiency, and employing a comprehensive scoring method to quantitatively evaluate the schemes corresponding to multiple spatiotemporal metrological design models; setting an evaluation threshold, and selecting the scheme with the highest score that meets all constraints as the target metrological design scheme; based on the target scheme, sorting out the traceability requirements of each metrological object: clarifying the metrological standard basis, traceability implementation path, responsible entity, metrological time node, anomaly handling mechanism, and other core information; performing structured processing and analysis on the above information, generating an aircraft traceability requirement summary table containing fields such as metrological object name, traceability requirement content, constraints, related sub-modules, and life cycle stage.

[0193] Furthermore, to improve the stability and reliability of metering equipment, perfect the closed-loop management of the metering traceability system, and enhance the safety of aircraft metering, step S505 above may include: Step S5051: Determine the detection equipment and / or calibration equipment for each metrological object based on the target metrological design scheme.

[0194] It should be noted that the testing equipment is used to test the performance parameters and operational status of the aircraft being measured. The calibration equipment can be used to calibrate the testing equipment or the object being measured, correcting errors to ensure that the measurement accuracy of the testing equipment meets metrological standards and to guarantee the accuracy of the measurement results.

[0195] In some embodiments, the traceability device can compile a list of testing and / or calibration equipment required for each metrological object based on the metrological standards, accuracy requirements, and spatiotemporal constraints of each metrological object in the target metrological design scheme; combine the technical parameters of the metrological object (such as measurement range and accuracy threshold) and the industry metrological equipment selection standards to clarify the core technical indicators of each device (such as detection accuracy, measurement speed, and applicable environment); verify the compatibility of the selected equipment with the metrological object and the existing metrological system through equipment compatibility testing, and finally form a standardized equipment requirement list, based on which the testing and / or calibration equipment for each metrological object is determined.

[0196] Step S5052: Identify multiple candidate suppliers associated with the testing equipment and / or calibration equipment.

[0197] It should be noted that candidate suppliers can be entities that provide testing and / or calibration equipment, including equipment manufacturers, authorized distributors, third-party service providers, etc.

[0198] In some embodiments, the traceability device can conduct a preliminary qualification review of the collected suppliers (such as checking the company's business license and relevant certification certificates), eliminate suppliers with incomplete qualifications or those that do not meet industry standards, and finally form a list of candidate suppliers, clarifying the types of equipment that each supplier can provide and their core advantages.

[0199] Step S5053: Collect the historical supply data, supplier profile data, and industry dynamic data of the candidate suppliers, and generate an evaluation time series dataset based on the historical supply data, the supplier profile data, and the industry dynamic data.

[0200] It should be noted that historical supply data refers to relevant records of equipment supplied by candidate suppliers in the past, covering data such as supply cycle, delivery qualification rate, after-sales service quality, and price fluctuations.

[0201] Supplier profile data can be characteristic data reflecting the comprehensive capabilities of candidate suppliers, including information such as enterprise qualifications (e.g., production licenses, metrological certification certificates), production scale, technological R&D capabilities, financial status, and industry reputation. Supplier profile data can be locally built based on historical supply records, or it can be profile data shared from other shared nodes.

[0202] Industry dynamic data can be industry-level information that affects equipment supply, including raw material price fluctuations, changes in industry capacity, policy and regulatory adjustments (such as updates to industry standards for metering equipment), and market supply and demand data.

[0203] The evaluation time series dataset can be a time-series structured dataset that integrates historical supply data of candidate suppliers, supplier profile data, and industry dynamic data, organized according to the time dimension, and used for supply dynamic risk prediction and analysis.

[0204] In some embodiments, the traceability device conducts multi-dimensional data collection for each candidate supplier: extracting historical supply data from the enterprise's procurement system and supplier cooperation ledger, covering historical (e.g., the most recent 3-5 years) supply records, delivery cycles, quality feedback, etc.; obtaining supplier profile data through supplier questionnaires, enterprise website information, third-party credit assessment platforms, etc., clarifying enterprise qualifications, production capacity, technical level, financial status, etc.; collecting industry dynamic data from industry information platforms, regulatory platform release channels, and raw material market reports, including raw material price trends, industry capacity changes, relevant policy adjustments, etc.; organizing the three types of data collected by timestamp, and after data cleaning, standardization, and other processing, integrating the data into a unified data format to form an evaluation time-series dataset, ensuring the time-series integrity and traceability of the data.

[0205] Step S5054: Preprocess the evaluation time series dataset to obtain target time series data, and perform supply dynamic risk prediction on each candidate supplier based on the target time series data to obtain supply dynamic risk prediction results.

[0206] It should be noted that the supply dynamic risk forecast result can be a quantitative result of the supply dynamic risk forecast output, including the risk level (such as high, medium, low), probability of risk occurrence, and main risk characteristics of each candidate supplier.

[0207] In some embodiments, the traceability device preprocesses the evaluation time series dataset, which may include: removing duplicate data and invalid data (such as records with missing key fields), using interpolation to fill in the missing time series data; performing min-max normalization on numerical data (such as price and delivery cycle), and labeling and encoding categorical data (such as qualification level).

[0208] In some embodiments, the traceability device divides the preprocessed target time series data into a training set and a validation set, and inputs them into a pre-built time series prediction model (such as a Bi-LSTM model); configures the model parameters (such as time step, hidden layer dimension, and learning rate), takes supply risk events (such as delayed delivery or supply disruption) as the prediction target, trains the model, and optimizes the parameters through the validation set; after the model converges, it inputs the complete target time series data to obtain the supply dynamic risk prediction results of each candidate supplier.

[0209] Step S5055: Send a request to multiple device quality sharing nodes so that each device quality sharing node sends the device quality data of the candidate supplier.

[0210] It should be noted that the equipment quality sharing nodes can be distributed nodes that store supplier equipment quality-related data (such as regulatory platforms, third-party quality testing agencies, enterprise shared databases, etc.). For example, equipment quality data may include equipment testing reports, quality complaint records, fault repair data, quality certification documents, and other information. Equipment quality sharing nodes can establish communication connections through a consortium blockchain.

[0211] In some embodiments, the traceability device can obtain a list of equipment quality sharing nodes (such as an industry quality supervision platform, an inter-enterprise quality data sharing alliance, or a third-party quality testing agency database); send data acquisition requests to each equipment quality sharing node, the requests including quality testing reports, fault records, quality certification documents, etc. of the candidate supplier's relevant equipment; and connect to the data interfaces of each sharing node to receive feedback equipment quality data.

[0212] Step S5056: Preprocess the equipment quality data, and conduct equipment quality risk assessment on each candidate supplier based on the preprocessed equipment quality data to obtain the quality risk assessment results.

[0213] It should be noted that the quality risk assessment results can be obtained by quantitatively evaluating the quality reliability of the equipment provided by the candidate supplier based on equipment quality data, including information such as equipment quality level, probability of quality failure risk, and quality assurance capability score.

[0214] In some embodiments, the traceability device can preprocess the acquired equipment quality data: remove false and invalid data, standardize and convert quality reports of different formats; extract core quality indicators (such as failure rate, mean time between failures, and quality certification level), and quantify each indicator using a weighted scoring method; construct a quality risk assessment system, and set the weight of each indicator in conjunction with industry quality standards (such as the highest weight for factory inspection accuracy); calculate the equipment quality score of each candidate supplier through comprehensive scoring, classify the quality risk level, form a quality risk assessment result, and clarify the quality advantages and potential risk points of each supplier's equipment.

[0215] Step S5057: Select target suppliers from the candidate suppliers based on the supply dynamic risk prediction results and the quality risk assessment results.

[0216] It should be noted that the target supplier can be the optimal supplier selected from the candidate suppliers, which has low supply risk, reliable equipment quality, and strong comprehensive adaptability. The target supplier provides supply services for the testing / calibration equipment of various metrology objects.

[0217] In practice, the traceability equipment combines the results of supply dynamic risk prediction and quality risk assessment for screening: setting supply risk thresholds and quality risk thresholds, and eliminating suppliers whose results are below either threshold; for the remaining suppliers, the analytic hierarchy process (AHP) is used to determine the weights of supply risk and quality risk (e.g., the weight of quality risk is higher than that of supply risk), and calculating the comprehensive score of each supplier; combining the cost constraints of the target quantitative design scheme, the cost suitability assessment is performed on suppliers with qualified comprehensive scores; finally, the supplier with the highest comprehensive score and strong cost suitability is selected as the target supplier.

[0218] Step S5058: Generate a summary table of aircraft traceability requirements based on the target suppliers associated with the testing equipment and / or calibration equipment of each metrology object and the target metrology design scheme.

[0219] In practical implementation, the traceability equipment, based on the target metrological design scheme, the list of testing / calibration equipment for each metrological object, and the corresponding target supplier information, sorts out the complete traceability requirements: clearly defining the name of each metrological object, the corresponding testing / calibration equipment model, the target supplier name, the equipment quality standard, and the metrological testing process; supplementing key traceability information, including equipment purchase batch, factory serial number, quality inspection report number, supplier after-sales service contact information, metrological data recording requirements, traceability path (tracing from metrological results to equipment, supplier, and quality inspection link), responsible entities (purchasing department, metrology department, supplier), and data storage cycle; and using a standardized table format, organizing the fields according to "metrological object - testing / calibration equipment - target supplier - quality standard - traceability path - responsible entity - constraints" to generate a summary table of aircraft traceability requirements.

[0220] Understandably, this embodiment avoids risks such as delayed delivery and supply disruptions by predicting supply dynamic risks in advance, and selects high-quality equipment through equipment quality risk assessment, thus providing dual protection for the stability and reliability of equipment supply, reducing potential faults in subsequent measurement work, strengthening the standardization of traceability management, optimizing supply chain configuration, and achieving optimal allocation of equipment supply resources through the scientific selection of target suppliers. Combined with cost suitability assessment, it reduces equipment procurement and subsequent maintenance costs while ensuring quality and supply stability.

[0221] Because this embodiment uses multi-dimensional weighted feature encoding to convert different types of measurement features into a unified format, it eliminates data heterogeneity, effectively integrates multi-dimensional measurement features, avoids the loss of key information, focuses on core related nodes through graph attention networks, can accurately capture the core features and key associations of measurement objects, improve feature expression capabilities, and configure measurement weights in combination with lifecycle scenarios to make global measurement features more in line with actual usage needs, enhance the scenario adaptability of solution selection, and ensure the optimality of solutions and the standardization of traceability.

[0222] Furthermore, this embodiment of the invention also proposes a computer-readable storage medium storing an MBSE-based aircraft metrological traceability program. When executed by a processor, the MBSE-based aircraft metrological traceability program implements the steps of the MBSE-based aircraft metrological traceability method described above. In this embodiment, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0223] The aforementioned computer-readable storage medium may be included in an MBSE-based aircraft metrological traceability device; or it may exist independently and not be assembled into an MBSE-based aircraft metrological traceability device.

[0224] The specific implementation of the computer program product of the present invention is basically the same as the embodiments of the above-mentioned MBSE-based aircraft metrological traceability method, and will not be repeated here.

[0225] Reference Figure 8 , Figure 8 This is a structural block diagram of the first embodiment of the aircraft metrological traceability device based on MBSE of the present invention.

[0226] like Figure 8 As shown, the MBSE-based aircraft metrological traceability device proposed in this embodiment of the invention includes: The spatiotemporal feature extraction module 10 is used to process the original multimodal data and extract the spatiotemporal features of the same measurement type of measurement objects in the whole aircraft based on the processed original multimodal data. The spatiotemporal features include the time constraints and spatial constraints of the measurement objects. The measurement objects are functional systems or functional devices in the whole aircraft. The meta-model construction module 20 is used to generate constructive elements based on the spatiotemporal features, perform metrological modeling on each metrological object based on the constructive elements, and generate a metrological meta-model. The metrological meta-model uses the detection equipment and / or calibration equipment of the metrological object as the minimization element. The metrological meta-model includes a detection equipment meta-model and a calibration equipment meta-model. The demand forecasting module 30 is used to divide the entire aircraft into multiple sub-modules, forecast the demand based on the stakeholders corresponding to each sub-module, and build a quantitative demand model based on the demand forecasting results. Each sub-module corresponds to a quantitative demand. Design modeling module 40 is used to perform spatiotemporal design analysis on the aircraft based on the metrological meta-model and the metrological requirement model, and to construct multiple spatiotemporal metrological design models, wherein the spatiotemporal metrological design models include detection schemes and / or calibration schemes for time and space dimensions. The traceability summary module 50 is used to perform a trade-off analysis on the multiple spatiotemporal metrological design models, select the target metrological design scheme, and generate a summary table of aircraft traceability requirements based on the target metrological design scheme.

[0227] This embodiment solves the problems of inconsistent multimodal data and uneven data quality by extracting the spatiotemporal features of measurement objects of the same measurement type from the processed multimodal data. It achieves accurate spatiotemporal feature extraction, efficiently filters out the core spatiotemporal attributes of measurement objects of the same measurement type, and realizes efficient and accurate spatiotemporal modeling of the metrological meta-model. It effectively avoids the problem of static metrological attributes in metrological schemes. By forecasting the needs of stakeholders of each sub-module of the aircraft, the metrological requirements and constraints of each sub-module are clearly defined. Based on the metrological meta-model and metrological demand model, the aircraft is spatiotemporally designed and modeled, realizing multi-dimensional optimization of metrological design schemes. It fully balances the multi-objective design of time, space, and demand, avoids the scheme defects caused by single-dimensional design, and selects the target metrological design scheme through trade-off analysis. It improves the objectivity of metrological schemes, reduces the influence of subjective factors on scheme selection, reduces the difficulty and maintenance cost of aircraft metrological traceability, and ensures metrological compliance and traceability.

[0228] The MBSE-based aircraft metrological traceability device provided in this application employs the MBSE-based aircraft metrological traceability method described in the above embodiments, and can solve the technical problems of MBSE-based aircraft metrological traceability. Compared with the prior art, the beneficial effects of the MBSE-based aircraft metrological traceability device provided in this application are the same as those of the MBSE-based aircraft metrological traceability method provided in the above embodiments, and other technical features in the MBSE-based aircraft metrological traceability device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0229] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.

[0230] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0231] In addition, for technical details not described in detail in this embodiment, please refer to the MBSE-based aircraft metrological traceability method provided in any embodiment of the present invention, which will not be repeated here.

[0232] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0233] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0234] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for the metrological traceability of aircraft based on MBSE, characterized in that, The MBSE-based aircraft metrological tracing method includes: The original multimodal data is processed, and the spatiotemporal features of measurement objects of the same measurement type in the whole aircraft are extracted based on the processed original multimodal data. The spatiotemporal features include the temporal constraints and spatial constraints of the measurement objects, and the measurement objects are functional systems or functional devices in the whole aircraft. Based on the aforementioned spatiotemporal features, constructive elements are generated. Based on the constructive elements, metrological modeling is performed on each metrological object to generate a metrological meta-model. The metrological meta-model uses the detection equipment and / or calibration equipment of the metrological object as minimizing elements. The metrological meta-model includes a detection equipment meta-model and a calibration equipment meta-model. The aircraft is divided into multiple sub-modules. Demand forecasts are made based on the stakeholders corresponding to each sub-module, and a quantitative demand model is constructed based on the demand forecast results. Each sub-module corresponds to a quantitative demand. Based on the aforementioned metrological meta-model and metrological requirement model, a spatiotemporal design analysis of the aircraft is conducted, and multiple spatiotemporal metrological design models are constructed. The spatiotemporal metrological design models include detection schemes and / or calibration schemes for the time and space dimensions. A trade-off analysis is performed on the multiple spatiotemporal metrological design models to select the target metrological design scheme, and a summary table of aircraft traceability requirements is generated based on the target metrological design scheme.

2. The MBSE-based aircraft metrological traceability method as described in claim 1, characterized in that, The process of processing the raw multimodal data and extracting the spatiotemporal features of measurement objects of the same measurement type in the entire aircraft based on the processed raw multimodal data includes: The original multimodal data is processed into different modalities to obtain initial multimodal features; The initial multimodal features are input into a multimodal pre-trained model for multimodal fusion, and a multimodal feature vector of uniform dimension is output. The multimodal pre-trained model is configured to learn the association weights of different modalities based on the modal attention mechanism, and to fuse the initial multimodal features based on the association weights. Semantic analysis is performed on the multimodal feature vectors, and multimodal entity recognition and multimodal relation extraction are performed based on the semantic analysis results to obtain multimodal entities and multimodal relations; Multiple valid quadruplets are generated based on the multimodal entities and the multimodal relationships. The entities in the valid quadruplets include the aircraft as a whole, the functional systems in the aircraft as a whole, the measurement objects in the functional systems, and the measurement attribute information of the measurement objects. The measurement attribute information includes measurement attributes and the time attributes of the measurement attributes. The measurement objects are the functional systems or the functional devices in the functional systems. A dynamic hypergraph is constructed based on the effective quadruples, and spatiotemporal features of measurement objects of the same measurement type in the whole spacecraft are extracted based on the dynamic hypergraph. The nodes in the dynamic hypergraph are the entities in the effective quadruples. The hyperedges of the dynamic hypergraph include structural association hyperedges, measurement association hyperedges and temporal association hyperedges. The structural association hyperedges are used to connect measurement objects with structural associations in the same system or across systems in the whole spacecraft. The measurement association hyperedges are used to connect measurement objects with attribute associations in the same system or across systems in the whole spacecraft. The temporal association hyperedges are used to connect measurement objects with temporal dimension associations.

3. The MBSE-based aircraft metrological traceability method as described in claim 2, characterized in that, The generation of multiple valid quadruplets based on the multimodal entities and the multimodal relationships includes: A set of quadruples is generated based on the multimodal entities and the multimodal relationships, and the set of quadruples includes multiple original quadruples; Traverse the set of quadruples to determine the valid and invalid quadruples in the set of quadruples, wherein the invalid quadruples are missing at least one entity. Based on the overall system topology of the aircraft with the invalid quadruple, a candidate entity set is generated. The candidate entity set includes multiple candidate entities, and each candidate entity includes the overall functional system of the aircraft, the measurement object, and the measurement attribute information of the measurement object. Semantic analysis is performed on the invalid quadruplets, and multiple entities to be scored are selected from the candidate entity set based on the semantic analysis results; Analyze the structural and / or quantitative correlations between the entities currently existing in the invalid quadruples and each entity to be scored to obtain the correlation analysis results; Based on the correlation analysis results, each entity to be scored is scored, and the target completion entity is selected from the entities to be scored based on the scoring results; The target completion entity is embedded into the invalid quad, and the invalid quad after entity embedding is marked as a valid quad.

4. The MBSE-based aircraft metrological traceability method as described in any one of claims 1 to 3, characterized in that, The process of dividing the entire aircraft into multiple sub-modules, forecasting demand based on the stakeholders corresponding to each sub-module, and constructing a quantitative demand model based on the demand forecast results includes: Based on the overall system architecture, structural type, usage scenarios and / or users of the aircraft, the aircraft is divided into multiple sub-modules; Demand time series data is obtained based on the stakeholders corresponding to each sub-module, and the demand time series data is preprocessed to obtain candidate time series data. The demand time series data includes the demands of stakeholders and the timestamps of the demands. The preprocessing includes normalization processing. The candidate time-series data is input into a pre-built demand time-series prediction model to obtain demand prediction results. The demand time-series prediction model includes a multi-feature extraction branch structure, a target feature fusion structure, and a multi-task time-series decoding layer. The multi-feature extraction branch structure is used to extract global demand features and local demand features from the candidate time-series data. The target feature fusion structure is used to fuse the global demand features and the local demand features to output target fused features. The multi-task time-series decoding layer is used to decode the target fused features to output demand prediction results. Architectural analysis, behavioral analysis, and indicator analysis are performed on each submodule to obtain multidimensional analysis results for each submodule; Based on the multidimensional analysis results, demand forecasting results are captured, and a quantitative demand model is constructed based on the captured demand results.

5. The MBSE-based aircraft metrological traceability method as described in claim 4, characterized in that, The feature extraction branch structure includes a local feature extraction structure and a global feature extraction structure. The local feature extraction structure includes a channel splitting structure, a first branch structure, a second branch structure, a third branch structure, and a candidate feature fusion module. The global feature extraction structure includes a fourth branch structure. The channel splitting structure is configured to perform semantic awareness analysis on candidate time series data to obtain semantic awareness features, and divide the candidate time series data into multiple channel data based on the semantic awareness features. Based on the semantic awareness features corresponding to each channel data, the channel data is respectively input to the first branch structure, the second branch structure and the third branch structure. The first branch structure is built on a Bi-LSTM network and is configured to extract bidirectional time-series features based on channel data, and convert the bidirectional time-series features into a query vector matrix; The second branch structure is built based on a graph attention network and is configured to extract the demand association features between each sub-module based on channel data, and convert the demand association features into a key vector matrix; The third branch structure is constructed based on a spatiotemporal graph convolutional network and is configured to extract spatiotemporal constraint features based on channel data and convert the spatiotemporal constraint features into a value vector matrix. The candidate feature fusion module is configured to split the query vector matrix, key vector matrix, and value vector matrix into multiple sub-matrices along the feature dimension, generate multiple attention heads based on the sub-matrices, calculate cross-branch attention scores through the multiple attention heads, concatenate the output results of each attention head, map the concatenation result to the original dimension, and output local demand features. The fourth branch structure is built based on the Transformer model and is configured to extract global requirement features of the entire life cycle of the aircraft based on candidate time series data.

6. The MBSE-based aircraft metrological traceability method as described in any one of claims 1 to 3, characterized in that, The process involves a trade-off analysis of the multiple spatiotemporal econometric design models to select a target econometric design scheme, and generating a summary table of aircraft traceability requirements based on the target econometric design scheme, including: Feature encoding is performed on each spatiotemporal metrological design model from multiple trade-off feature dimensions to obtain multi-dimensional metrological trade-off features. The multi-dimensional metrological trade-off features include the metrological cost features, metrological time features, metrological space features, and metrological method features of each metrological object in the spatiotemporal metrological design model. The metrological method features include off-site metrological features and in-situ metrological features. Based on the multi-dimensional measurement trade-off features, the spatiotemporal measurement design model is converted into a measurement graph structure. The nodes in the measurement graph structure represent the measurement objects in the spatiotemporal measurement design model, and the node attributes of each node are the multi-dimensional measurement trade-off features. The edges of the measurement graph structure include structural dependencies, measurement dependencies, and time dependencies between nodes. The metrology graph structure is input into a graph attention network to obtain the node embedding vectors of each metrology object. According to the usage scenarios of each stage of the overall life cycle of the aircraft, configure the measurement weights corresponding to the node embedding vectors of each measurement object, and aggregate the node embedding vectors of each measurement object based on the measurement weights to obtain the global measurement features of each spatiotemporal measurement design model. Based on the global metrological characteristics, a target metrological design scheme is selected from the metrological design schemes corresponding to the multiple spatiotemporal metrological design models, and a summary table of aircraft traceability requirements is generated based on the target metrological design scheme.

7. The MBSE-based aircraft metrological traceability method as described in claim 6, characterized in that, The process of generating a summary table of aircraft traceability requirements based on the target metrological design scheme includes: Based on the target metrological design scheme, determine the testing equipment and / or calibration equipment for each metrological object; Identify multiple candidate suppliers associated with the testing equipment and / or calibration equipment; Collect historical supply data, supplier profile data, and industry dynamic data of the candidate suppliers, and generate an evaluation time series dataset based on the historical supply data, supplier profile data, and industry dynamic data; The evaluation time series dataset is preprocessed to obtain target time series data, and supply dynamic risk prediction is performed on each candidate supplier based on the target time series data to obtain supply dynamic risk prediction results. Send requests to multiple device quality sharing nodes so that each device quality sharing node sends the device quality data of the candidate supplier; The equipment quality data is preprocessed, and the equipment quality risk assessment of each candidate supplier is performed based on the preprocessed equipment quality data to obtain the quality risk assessment results. Target suppliers are selected from the candidate suppliers based on the supply dynamic risk prediction results and the quality risk assessment results; A summary table of aircraft traceability requirements is generated based on the target suppliers associated with the testing and / or calibration equipment for each metrological object and the target metrological design scheme.

8. A measurement and traceability device for aircraft based on MBSE, characterized in that, The MBSE-based aircraft metrological traceability device includes: The spatiotemporal feature extraction module is used to process the raw multimodal data and extract the spatiotemporal features of measurement objects of the same measurement type in the whole aircraft based on the processed raw multimodal data. The spatiotemporal features include the temporal constraints and spatial constraints of the measurement objects. The measurement objects are functional systems or functional devices in the whole aircraft. The meta-model construction module is used to generate constructive elements based on the spatiotemporal features, perform metrological modeling on each metrological object based on the constructive elements, and generate a metrological meta-model. The metrological meta-model uses the detection equipment and / or calibration equipment of the metrological object as the minimization element. The metrological meta-model includes a detection equipment meta-model and a calibration equipment meta-model. The demand forecasting module is used to divide the entire aircraft into multiple sub-modules, forecast the demand based on the stakeholders corresponding to each sub-module, and build a quantitative demand model based on the demand forecasting results. Each sub-module corresponds to a quantitative demand. The design modeling module is used to perform spatiotemporal design analysis on the aircraft based on the metrological meta-model and the metrological requirement model, and to construct multiple spatiotemporal metrological design models, which include detection schemes and / or calibration schemes for time and space dimensions. The traceability and summary module is used to perform trade-off analysis on the multiple spatiotemporal metrological design models, select target metrological design schemes, and generate a summary table of aircraft traceability requirements based on the target metrological design schemes.

9. An aircraft metrological traceability device based on MBSE, characterized in that, The MBSE-based aircraft metrological traceability device includes: a memory, a processor, and an MBSE-based aircraft metrological traceability program stored in the memory. The processor is used to run the MBSE-based aircraft metrological traceability program, which is configured to implement the MBSE-based aircraft metrological traceability method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an MBSE-based aircraft metrological traceability program, which, when executed by a processor, implements the MBSE-based aircraft metrological traceability method as described in any one of claims 1 to 7.