Method and system for evaluating accuracy of aviation equipment full life cycle digital twin model
By using a method for evaluating the accuracy of digital twin models throughout the entire lifecycle of aviation equipment, uncertainties are identified and quantified, enabling a global accuracy evaluation of the digital twin model. This solves the problem of insufficient accuracy in digital twin models, improves the accuracy and reliability of the model, guides product design and process optimization, and promotes the efficient research and development and production of aviation equipment.
Patent Information
- Application Number
- CN202610740880.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-27
- Publication Date
- 2026-08-25
AI Technical Summary
In existing technologies, the accuracy of digital twin models is insufficient, resulting in distorted simulation results and excessive prediction bias. This makes it impossible to effectively manage and optimize the entire lifecycle of aviation equipment, thus limiting its application value in industrial scenarios.
This paper provides an accuracy evaluation method for a digital twin model of aviation equipment throughout its entire life cycle. By continuously collecting real data at each stage, it identifies and quantifies uncertainties, uses probability theory and Monte Carlo simulation methods to conduct a global accuracy evaluation, and outputs optimization guidance. This includes defining the digital twin model, uncertainty identification and quantification, accuracy evaluation, and optimization units, thus constructing a comprehensive and continuous data foundation to guide product design, testing methods, and process parameter optimization.
It significantly improves the fit between digital twin models and physical entities, enhances the accuracy and reliability of models, guides product design optimization, test improvement and process control, shortens the R&D cycle, and improves the quality and efficiency of aviation equipment R&D, production and operation.
Smart Images
Figure CN122634867A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of digital twin technology, specifically to a method and system for evaluating the accuracy of a digital twin model of an aviation equipment throughout its entire lifecycle. Background Technology
[0002] Driven by the Industry 4.0 wave, intelligent manufacturing and the Industrial Internet are accelerating their integration and evolution, and the manufacturing industry is undergoing a profound transformation towards digitalization, networking, and intelligence. In this process, digital twin technology, as a core hub technology bridging the physical and information worlds, has become a cutting-edge research hotspot focusing on both academia and industry due to its unique advantages of "virtual-real mapping, real-time linkage, and closed-loop optimization." Especially in the field of aerospace equipment, it is widely regarded as a key support for achieving intelligent manufacturing upgrades.
[0003] Specifically, the core logic of digital twin technology lies in: constructing a digital mirror model in virtual space that accurately matches all elements and processes of the physical entity; collecting real-time operational status data of the physical entity through technologies such as industrial sensors and the Internet of Things; conducting simulation, performance analysis, trend prediction, and process optimization based on the digital model; and feeding back the optimized decision instructions to the physical space, forming a closed-loop control link of "physical entity-virtual model-decision feedback," ultimately achieving intelligent management and efficient optimization of the entire lifecycle of a product, from design and development, production and manufacturing to operation and maintenance services.
[0004] It is important to emphasize that the full realization of the application value of digital twin technology hinges on the accuracy of the digital twin model—that is, whether the model can realistically and accurately replicate the operating state, behavioral characteristics, and core performance of the physical entity. Only when the digital twin model and the physical entity achieve a high-fidelity mapping can the reliability of subsequent simulation analysis, prediction optimization, and other stages be ensured, thereby providing effective support for intelligent decision-making throughout the entire lifecycle. Conversely, insufficient model accuracy will directly lead to distorted simulation results and excessive prediction deviations, making it impossible to effectively control and optimize the physical entity, ultimately limiting the practical application value of digital twin technology in various industrial scenarios. Summary of the Invention
[0005] To address the shortcomings of the existing technology, the present invention aims to provide an accuracy evaluation method and system for digital twin models of aviation equipment throughout their entire life cycle. This system can continuously collect real manufacturing and operational data at each stage, guide the reduction and correction of uncertainties in the digital twin model, and guide and optimize future product design, testing methods, and key process parameters.
[0006] Specifically, in a first aspect, the present invention provides a method for evaluating the accuracy of a digital twin model of an aviation equipment throughout its entire life cycle, comprising the following steps: S1. Define the digital twin model: The digital twin model includes physical space, virtual space, twin data, and connectivity and services, and covers the entire life cycle of aviation equipment from design, testing, mass production and service. S2. Identify and quantify uncertainties. Identify uncertainties in digital twins, including equipment uncertainties, data processing uncertainties, virtual space modeling uncertainties, and numerical simulation uncertainties. Quantify uncertainties using probability theory-based methods along a pre-defined uncertainty propagation path. S3. Evaluate the global accuracy of the digital twin model, specifically including: S31. Random sampling is performed based on the probability distribution of key state parameters of aviation equipment, and virtual simulation is conducted using the established digital twin model. After repeated multiple times, the virtual model prediction probability density function of static key state parameters is obtained. And the probability density surface of the virtual model prediction of dynamic key state parameters ; S32. Instruct the randomly sampled physical entities of the aviation equipment from step S31 to perform actual operations, and obtain the physical entity measurement probability density function of the static key state parameters. And the probability density surface of physical entity measurement of dynamic key state parameters ; S33. Utilizing the Globally Weighted Accuracy Index Evaluating the global accuracy of digital twins: ; in, To determine the consistency between the virtual model prediction probability density function and the physical entity measurement probability density function for the i-th static critical state parameter, To ensure consistency between the virtual model prediction probability density surface for the j-th dynamic critical state parameter and the physical entity measurement probability density surface. for The weight, for The weight.
[0007] Preferably, the virtual model prediction probability density function of the static key state parameters in step S33 With the probability density function of physical entity measurement The formula for calculating consistency is: ; in, These are static key state parameters.
[0008] Preferably, the formula for calculating the consistency between the virtual model prediction probability density curve and the physical entity measurement probability density curve of the dynamic key state parameters in step S33 is as follows: ; in, These are dynamic key state parameters. For time.
[0009] Preferably, in step S33, .
[0010] Preferably, in step S33, The value range is [0,1]. A value of 1 indicates that the global accuracy of the digital twin is 100%, and a value of 0 indicates that the global accuracy of the digital twin is 0.
[0011] Preferably, the method further includes step S4, optimizing the virtual space and the physical space, wherein optimizing the virtual space is to guide the design optimization of the virtual space model, and optimizing the physical space is to guide the optimization of physical operations.
[0012] Preferably, the virtual space in step S1 includes a prediction model and a performance mapping model. The prediction model is used to predict the geometric parameters of the operation process, and the performance mapping model is used to correlate the predicted geometric parameters with the core machine performance parameters.
[0013] Preferably, the simulation method in step S3 is Monte Carlo simulation.
[0014] Preferably, the key parameters in step S1 are operation parameters whose impact on operation accuracy is greater than the accuracy threshold.
[0015] Secondly, this invention provides an accuracy evaluation system for a digital twin model of aviation equipment throughout its entire life cycle. The system includes a digital twin unit, an uncertainty identification and quantification unit, an accuracy evaluation unit, and an output and optimization unit. The digital twin unit performs digital twin operations on the evaluation object; the uncertainty identification and quantification unit identifies and quantifies uncertainties; the accuracy evaluation unit evaluates the overall accuracy of the digital twin; and the optimization unit outputs the accuracy of the digital twin and provides optimization guidance for the virtual space model and physical operation process based on predicted key parameters.
[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The accuracy evaluation method of the full life cycle digital twin model of aviation equipment of the present invention proposes a brand-new concept and specific method for evaluating the accuracy of digital twin models, filling the current technical gap on whether digital twins are accurate, enabling digital twin technology to be better applied in the full life cycle prediction of aviation equipment, and through accuracy feedback, it helps to further improve the consistency between digital twins and physical space.
[0017] (2) The accuracy evaluation method of the full life cycle digital twin model of aviation equipment of the present invention can continuously collect real data of each stage of manufacturing, operation and other aspects of the full life cycle of aviation equipment, and build a comprehensive, continuous and objective data foundation for the accuracy evaluation of digital twin model, avoid evaluation deviation caused by data fragmentation and lag in the evaluation process, and ensure the reliability of evaluation results.
[0018] (3) The accuracy evaluation method of the full life cycle digital twin model of aviation equipment of the present invention can provide targeted guidance to reduce and correct the uncertainty of the digital twin model, effectively reduce the deviation between the model and the real state of aviation equipment, significantly improve the fit between the digital twin model and the physical entity, enhance the accuracy and credibility of the model, and provide reliable guarantee for various subsequent model-based applications.
[0019] (4) The accuracy evaluation method of the full life cycle digital twin model of aviation equipment of the present invention can effectively guide the product design optimization, test method improvement and key process parameter control of future aviation equipment through the results accumulated by accurate evaluation and model correction, help shorten the product development cycle, improve the test effectiveness and optimize the process stability, and thus comprehensively improve the overall quality and efficiency of aviation equipment research and development, production and operation and maintenance, and has significant engineering application value. Attached Figure Description
[0020] Figure 1 This is a schematic diagram of the overall method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to the present invention. Figure 2 This is a flowchart illustrating the steps of a specific embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the full lifecycle digital twin uncertainty quantification of a specific embodiment of the present invention; Figure 4 This is a schematic diagram of the overall system structure of the accuracy evaluation system for the full life cycle digital twin model of aviation equipment according to the present invention. Detailed Implementation
[0021] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
[0022] Specifically, in a first aspect, the present invention provides a method for evaluating the accuracy of a digital twin model of an aviation equipment throughout its entire life cycle, such as... Figure 1 As shown, it includes the following steps: S1. Define the digital twin model: The digital twin consists of five dimensions: physical space, virtual space, twin data, connectivity, and services, covering the entire lifecycle of a product, from design and testing to mass production and service. The virtual space includes a predictive model and a performance mapping model. The predictive model is used to predict the geometric parameters of the operational process, while the performance mapping model is used to correlate the predicted geometric parameters with the core machine's performance parameters. The entire lifecycle of aerospace equipment refers to the design phase → testing phase → mass production phase → service phase. The operations here can include design, manufacturing, assembly, or lifespan prediction during service.
[0023] In practical applications, the evaluation of digital twins in this invention should consist of five dimensions: physical space, virtual space, twin data, connectivity, and services. Physical space includes physical objects such as devices, products, systems, and environments that objectively exist in the real world. Virtual space includes a complete mapping and digital mirror of entities in the physical space, including but not limited to three-dimensional geometric models, physical models, behavioral models, and rule models. Twin data is a data set generated by the interaction and fusion between physical entities, virtual entities, and services, integrating data from physical entities such as sensor data and operational history, virtual entities such as simulation data and prediction results, and domain knowledge such as design parameters and process standards. Connectivity refers to the carriers and mechanisms for communication, data transmission, and interaction among these dimensions: physical entities, virtual entities, twin data, and services. Services refer to various applications and functions provided based on the digital twin model, designed to meet the specific needs of different fields and users, and are typically provided to users in the form of software such as apps or application programming interfaces (APIs).
[0024] Throughout the entire lifecycle of aerospace equipment, the design phase primarily determines basic design information and assesses the product, materials, or components in a controlled laboratory environment to ensure they meet design standards and expected performance requirements. The testing phase includes the early stages of product development, involving the initial construction of conceptual prototypes based on the design, verifying their feasibility and technical solutions, and subsequent systematic verification processes designed to ensure the product meets design specifications and operates as expected, laying the foundation for mass production. The mass production phase, following the completion of initial certification testing, encompasses the entire lifecycle of mass production from raw material processing to final assembly. The service phase refers to the actual deployment and performance tracking of the product under real-world usage conditions, including operational status monitoring, performance tracking and reliability testing, fault analysis, and user feedback collection. Key parameters are those identified as having a significant impact from an expert database.
[0025] S2. Uncertainty Identification and Quantification: This involves systematically identifying uncertainties in the virtual operation process, including equipment uncertainties, data processing uncertainties, virtual space modeling uncertainties, and numerical simulation uncertainties. Simultaneously, uncertainties are quantified along a pre-defined uncertainty propagation path. Uncertainty quantification methods include at least one of Monte Carlo simulation, Bayesian inference, and sensitivity analysis.
[0026] In practical applications, the main evaluation factors addressed by this method include uncertainties introduced by testing equipment, data transmission and communication, data processing, virtual modeling, and numerical simulation. The mapping relationship between each evaluation factor and the constituent elements of the digital twin is as follows: Figure 3 As shown. Uncertainty introduced by testing equipment mainly analyzes and quantifies the measurement errors and uncertainties introduced by physical entity hardware such as sensors and measuring devices during the data acquisition phase due to factors such as their own accuracy, drift, and environmental influences. Uncertainty introduced by data transmission and communication mainly assesses the data distortion or incomplete information caused by problems such as delays, packet loss, and signal interference that may occur during the transmission of twin data from physical space to virtual space via the network. Quantifying the uncertainty introduced by data processing involves errors and information loss introduced by algorithm selection and parameter settings during the preprocessing of the acquired raw data, such as cleaning, noise reduction, and feature extraction. The uncertainty assessment caused by virtual model modeling mainly addresses the structural deviations and uncertainties of the model resulting from simplification, abstraction, or assumptions about the geometry, physical properties, and operating mechanisms of physical entities when constructing the digital twin virtual model. The analysis of uncertainty introduced by numerical simulation mainly evaluates the uncertainty of the calculation results introduced by factors such as discretization errors, convergence criteria, and computational accuracy limitations of the solver algorithm during numerical simulation calculations.
[0027] In specific applications, such as Figure 3 As shown, this method considers the overall accuracy of the digital twin model as a comprehensive result of the cumulative transmission of uncertainties from multiple stages. To achieve systematic quantitative evaluation, this method defines two key uncertainty propagation paths: a vertical single-node propagation path and a horizontal multi-mapping node propagation path. The vertical single-node propagation path follows the sequence of test equipment → data transmission → data processing → simulation analysis model → numerical calculation model. The evaluation method will identify, quantify, and transmit uncertainties unidirectionally and node by node along this path, and evaluate the uncertainty of the digital twin model through the final output simulation results. The horizontal multi-mapping node propagation path quantifies the positive transmission and accumulation of uncertainties between the design stage → testing stage → mass production stage → service stage, while coupling a closed-loop feedback loop formed when measured data from subsequent stages is fed back to preceding stages for model correction and iterative optimization.
[0028] S3. Evaluate the global accuracy of digital twins based on uncertainty. This evaluation includes the following sub-steps: S31. Random sampling is performed based on the probability distribution of key state parameters of aviation equipment. Virtual simulation is conducted using digital twins combined with Monte Carlo methods. After repeating the simulation multiple times, typically tens of thousands of times, the virtual model prediction probability density function containing accumulated uncertainty of the static key state parameters is obtained. And the probability density surface of the virtual model prediction of dynamic key state parameters .
[0029] S32. The randomly sampled physical entities of the aviation equipment from step S31 are put into actual operation. After the operation is completed, key parameters are measured, and the probability density function of the physical entity measurement of static key state parameters is obtained. And the probability density surface of physical entity measurement of dynamic key state parameters .
[0030] S33. Utilizing the Globally Weighted Accuracy Index Evaluating the global accuracy of digital twins: ; in, To determine the consistency between the virtual model prediction probability density function and the physical entity measurement probability density function for the i-th static critical state parameter, To ensure consistency between the virtual model prediction probability density surface for the j-th dynamic critical state parameter and the physical entity measurement probability density surface. for The weight, for The weight, , The value range is [0,1], and the closer the value is to 1, the higher the accuracy. A value of 1 indicates that the global accuracy of the digital twin is 100%, and a value of 0 indicates that the global accuracy of the digital twin is 0.
[0031] Among them, the virtual model prediction probability density function of static key state parameters With the probability density function of physical entity measurement The formula for calculating the consistency of ) is: ; in, Let D(x) represent the static critical state parameters, and let P(x) represent the probability density of predictions from multiple virtual models for the static critical state parameters. To differentiate the static key state parameters. The value range is [0,1]. The closer the value is to 1, the higher the consistency, and the higher the accuracy of the digital twin in static response.
[0032] The formula for calculating the consistency between the virtual model prediction probability density curve and the physical entity measurement probability density curve of the dynamic key state parameters is as follows: ; Where y is a dynamic key state parameter. For time. The value range is [0,1]. The closer the value is to 1, the higher the accuracy of the digital twin in dynamic response.
[0033] S4. Output the accuracy of the digital twin and provide optimization guidance for the virtual space model and physical operation process based on the predicted key parameters. Specifically, optimizing the virtual space model guides the design optimization of the virtual space model, and optimizing the physical operation process guides the optimization of the physical operation.
[0034] Secondly, this invention provides an accuracy evaluation system for digital twin models of aviation equipment throughout their entire lifecycle, such as... Figure 4 As shown, it includes a digital twin unit 1, an uncertainty identification and quantification unit 2, an accuracy evaluation unit 3, and an output and optimization unit 4. The digital twin unit 1 is used to create a digital twin of the evaluation object; the uncertainty identification and quantification unit 2 is used to identify and quantify uncertainties; the accuracy evaluation unit 3 is used to evaluate the global accuracy of the digital twin; and the optimization unit 4 is used to output the accuracy of the digital twin and provide optimization guidance for the virtual space model and physical operation process based on the predicted key parameters. Specific Implementation This embodiment uses the evaluation of the accuracy of digital twins for predicting the geometric accuracy and performance impact of aero-engine core engines as an example to provide a detailed and reproducible description of the complete implementation process of the present invention. This embodiment applies the method of the present invention to evaluate the accuracy of a digital twin system used to predict the final state of key geometric parameters of a certain type of core engine after assembly, such as high-pressure turbine blade tip clearance and rotor coaxiality, and to estimate the impact of these geometric deviations on initial performance, such as vibration. Figure 2 As shown, this embodiment includes the following steps: S1. Perform digital twin assembly on the evaluation object: Select the aviation equipment to be assembled as the evaluation object. Before the evaluation object is physically assembled, virtual operations are performed in virtual space by integrating the actual measurement data of individual parts, and the assembly accuracy is predicted probabilistically to obtain the predicted probability of key parameters.
[0036] The core task of this digital twin is to perform virtual assembly in virtual space by integrating the actual measurement data of individual parts before physical assembly, so as to predict the final assembly accuracy probabilistically, thereby guiding the selection of parts and optimizing the assembly process.
[0037] In this embodiment, the evaluation object is a batch of core machine components to be assembled, such as blades, discs, casings, sealing structures, etc. of high-pressure compressors / turbines.
[0038] The equipment used for physical entity measurement includes high-precision coordinate measuring machines (CMMs) and laser scanners, which are used to precisely measure the critical dimensions and geometric tolerances of each part. It also includes tooling and fixtures used on the assembly site.
[0039] The core model of the virtual space is a three-dimensional tolerance analysis and variation model. This model is based on the CAD model of the part, but includes all the key geometric dimensions and tolerance information. It can simulate the positioning, contact, and "dimensional chain" accumulation of the part during the assembly process.
[0040] The performance mapping model is a simplified or surrogate model that correlates predicted geometric deviations, such as a 0.1 mm increase in tip clearance, with core engine performance parameters, such as a 0.05% decrease in high-pressure turbine efficiency.
[0041] Twin data: Physical entity data: The CMM full-size inspection report for each part to be assembled constitutes the core input of the digital twin.
[0042] Virtual entity data: The predicted probability distribution of key parameters output after virtual assembly, such as the predicted distribution of high-pressure turbine blade tip clearance and the predicted distribution of rotor radial runout.
[0043] Data integration: nominal dimensions and tolerance ranges on the design blueprints, as well as material properties; material properties are used to analyze deformation caused by assembly stress.
[0044] Data connection: The CMM measurement equipment has a data interface with the Manufacturing Execution System (MES), which automatically imports the measured data of the parts into the digital twin system.
[0045] Virtual Space Output: Virtual Assembly Verification App: Given a set of actual measurement data for a specific part, run the virtual assembly with one click and give a prediction report of "pass / fail / warning".
[0046] Intelligent Selection API: Automatically calculates and recommends the optimal combination of parts from a database of hundreds of similar parts, such as turbine blades, in the warehouse to minimize or maximize the tip clearance of the final assembly.
[0047] S2. Uncertainty Identification and Quantification: This involves identifying uncertainties in the virtual assembly process, including equipment uncertainties, data processing uncertainties, virtual space modeling uncertainties, and numerical simulation uncertainties.
[0048] Equipment uncertainty or measurement uncertainty: Source: The measurement accuracy of CMM is not infinite; its measurement results have inherent uncertainties, such as ±2 micrometers.
[0049] Quantification: Based on the CMM calibration certificate, each measurement dimension is modeled as a probability distribution with the mean of the measurement value and the standard deviation or range of the equipment accuracy.
[0050] Data processing uncertainty or data fitting uncertainty: Source: CMM fits a geometric feature, such as a plane or cylinder, by measuring a series of discrete points. The algorithm for fitting a point cloud to a geometric feature inherently contains uncertainties.
[0051] Quantification: The uncertainty of the fitting results is evaluated by the covariance matrix of the algorithm and then applied to the geometric model of the part.
[0052] Uncertainty in virtual modeling: Source 1: Simplification of part contact relationships in 3D tolerance models. For example, the model may assume that two planes fit perfectly, but in reality, there are minute gaps or deformations due to surface roughness and assembly forces.
[0053] Quantization 1: Introduce contact random elements to model the behavior of the contact surface as a spring system with a probability distribution.
[0054] Source 2: Residual stress inside the part introduced during the manufacturing process, which is released during assembly and causes deformation, may not have been considered in the initial model.
[0055] Quantization 2: The deformation caused by residual stress is modeled as a random field and attached to the geometric model of the part.
[0056] Numerical simulation uncertainty or statistical convergence uncertainty: Source: This twin typically employs the Monte Carlo method for tens of thousands of virtual assembly simulations. The limited number of simulations inherently introduces statistical convergence errors.
[0057] Quantification: The magnitude of statistical uncertainty is assessed by analyzing the convergence of key output quantities, such as the mean blade tip clearance, as the number of simulations increases.
[0058] S3. Global accuracy evaluation of digital twins, specifically including: S31. Randomly sample from the probability distribution of each dimension of each part to be assembled, perform virtual assembly using simulation, repeat multiple times, and obtain the virtual model prediction probability density function of the static key state parameters. And the probability density surface of the virtual model prediction of dynamic key state parameters .
[0059] S32. Perform actual physical assembly on the randomly sampled parts from step S31. After assembly, measure the key parameters and obtain the physical entity measurement probability density function of the static key state parameters. And the probability density surface of physical entity measurement of dynamic key state parameters .
[0060] S33. Utilizing the Globally Weighted Accuracy Index Evaluating the global accuracy of digital twins: ; in, To determine the consistency between the virtual model prediction probability density function and the physical entity measurement probability density function for the i-th static critical state parameter, To ensure consistency between the virtual model prediction probability density surface for the j-th dynamic critical state parameter and the physical entity measurement probability density surface. for The weight, for The weight.
[0061] Uncertainty propagation: Run Monte Carlo simulations. In each simulation, random sampling is performed from the probability distribution of each dimension of each part, with the center of the distribution being the CMM measurement, the width being the measurement uncertainty, and combined with the uncertainty of the model itself, to perform a complete virtual assembly.
[0062] After repeating the process tens of thousands of times, the key parameter f_D, a virtual model prediction probability density function for the tip clearance of a high-pressure turbine blade, is obtained.
[0063] Physical test comparison (actual assembly measurement): Perform actual physical assembly using the same selected batch of parts.
[0064] After assembly, use a special endoscope or feeler gauge and other precision tools to measure the actual blade tip clearance at multiple locations.
[0065] Due to limitations in the accuracy of measuring tools and differences in operation, the actual measurement result should also be considered as a distribution with uncertainty, denoted as the probability density function of the physical entity's measured value. .
[0066] Accuracy evaluation: Apply the Area Validation Metric (AVM) method to calculate and The overlapping area S_AVM.
[0067] In this embodiment, the calculation result is S_AVM = 0.92. The conclusion is that this digital twin achieves a prediction accuracy of up to 92% for the tip clearance of the high-pressure turbine blades in the core engine, demonstrating very high reliability. Since there are no dynamic parameters in this embodiment, no further calculations are performed.
[0068] S4. Output the accuracy of the digital twin and optimize the virtual space model and physical assembly process based on the predicted key parameters.
[0069] Sensitivity analysis: The results show that the factors contributing most to the uncertainty of the final blade tip clearance prediction are the roundness tolerance of the turbine casing and the end face runout tolerance of the rotor disk.
[0070] The design optimization guidance is fed back to the design phase: The design department can assess whether it is necessary to tighten these two tolerance requirements in the next generation of designs, or to add new positioning benchmarks to reduce their impact.
[0071] Guiding process optimization and providing feedback to mass production: The manufacturing department should invest more resources to ensure the roundness of the casing and the end face accuracy of the rotor disk. Simultaneously, assembly engineers can be guided to proactively compensate for these critical deviations during assembly by rotating the casing or adjusting the phase of the disk—this is precisely precision assembly guided by digital twins.
[0072] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.
Claims
1. A method for evaluating the accuracy of a digital twin model of an aircraft equipment throughout its entire life cycle, characterized in that: It includes the following steps: S1. Define the digital twin model: The digital twin model includes physical space, virtual space, twin data, and connectivity and services, and covers the entire life cycle of aviation equipment from design, testing, mass production and service. S2. Identify and quantify uncertainties, including uncertainties in digital twins, such as equipment uncertainties, data processing uncertainties, virtual space model modeling uncertainties, and numerical simulation uncertainties. Uncertainty is quantified using a probability theory-based method along a pre-defined uncertainty propagation path. S3. Evaluate the global accuracy of the digital twin model, specifically including: S31. Random sampling is performed based on the probability distribution of key state parameters of aviation equipment, and virtual simulation is conducted using the established digital twin model. After repeated multiple times, the virtual model prediction probability density function of static key state parameters is obtained. And the probability density surface of the virtual model prediction of dynamic key state parameters ; S32. Instruct the randomly sampled physical entities of the aviation equipment from step S31 to perform actual operations, and obtain the physical entity measurement probability density function of the static key state parameters. And the probability density surface of physical entity measurement of dynamic key state parameters ; S33. Utilizing the Globally Weighted Accuracy Index Evaluating the global accuracy of digital twins: ； in, To determine the consistency between the virtual model prediction probability density function and the physical entity measurement probability density function for the i-th static critical state parameter, To ensure consistency between the virtual model prediction probability density surface for the j-th dynamic critical state parameter and the physical entity measurement probability density surface. for The weight, for The weight.
2. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: The virtual model prediction probability density function of static key state parameters in step S33 With the probability density function of physical entity measurement The formula for calculating consistency is: ； in, These are static key state parameters.
3. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: The formula for calculating the consistency between the virtual model prediction probability density curve and the physical entity measurement probability density curve of the dynamic key state parameters in step S33 is as follows: ； in, These are dynamic key state parameters. For time.
4. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: In step S33, 。 5. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: In step S33, The value range is [0,1]. A value of 1 indicates that the global accuracy of the digital twin is 100%, and a value of 0 indicates that the global accuracy of the digital twin is 0.
6. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: It also includes step S4, optimizing the virtual space and the physical space, wherein optimizing the virtual space is to guide the design optimization of the virtual space model, and optimizing the physical space is to guide the optimization of physical operations.
7. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: The virtual space in step S1 includes a prediction model and a performance mapping model. The prediction model is used to predict the geometric parameters of the operation process, and the performance mapping model is used to correlate the predicted geometric parameters with the core machine performance parameters.
8. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: The simulation method in step S3 is Monte Carlo simulation.
9. The method for evaluating the accuracy of the digital twin model of the entire life cycle of aviation equipment according to claim 1, characterized in that: The key parameters in step S1 are those that have a greater impact on operational accuracy than the accuracy threshold.
10. An accuracy evaluation system for a digital twin model of aerospace equipment's entire life cycle, used in the accuracy evaluation method for the digital twin model of aerospace equipment's entire life cycle as described in claim 9, characterized in that: It includes a digital twin unit, an uncertainty identification and quantification unit, an accuracy evaluation unit, and an output and optimization unit; Digital twin units are used to perform digital twin operations on the evaluation object; The uncertainty identification and quantification unit is used to identify and quantify uncertainties. The accuracy evaluation unit is used to evaluate the overall accuracy of the digital twin; The optimization unit is used to output the accuracy of the digital twin and provide optimization guidance for the virtual space model and physical operation process based on the predicted key parameters.