A digital twin model evaluation method and system for intelligent manufacturing

Abstract

A kind of digital twin model evaluation method and system for intelligent manufacturing, comprising: obtaining n kinds of fidelity indexes of digital twin model, determining the weight of digital twin model fidelity primary index;N kinds of fidelity indexes correspond to n model evaluations, each model evaluation is divided into multiple secondary evaluation indexes, and the secondary index weight of each model evaluation is obtained according to the secondary evaluation index;According to all secondary evaluation indexes, the index evaluation scale is calculated, the secondary score is calculated by combining the secondary index weight, and the score of each model evaluation is obtained;According to the score of each model evaluation, the comprehensive score is obtained by comprehensive scoring.The present application starts from the digital twin model component, divides the evaluation of digital twin into five parts of geometry, electricity, kinematics, mechanics and virtual-real data, uses multi-angle full data to evaluate digital twin model, and improves the comprehensiveness and reliability of model evaluation.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of digital twin model evaluation, and particularly relates to a digital twin model evaluation method and system for intelligent manufacturing. BACKGROUND

[0002] Intelligent manufacturing: Intelligent manufacturing is a manufacturing method that comprehensively uses information technology, Internet of Things, big data analysis and other technologies, aiming to improve production efficiency and product quality. It covers the digitization and automation of production processes, as well as real-time monitoring and control.

[0003] Digital twin: Digital twin is an emerging concept that has attracted widespread attention in the fields of science and technology and industry. Its core idea is to replicate real-world entities, processes or systems into a virtual environment through digital means, so as to carry out simulation, analysis and optimization. The concept of digital twin is similar to the concept of biological twins, where one is real and the other is virtual.

[0004] Through digital twin, we can create virtual copies to understand the operation of actual objects or processes, conduct experiments and tests without directly intervening in the real world. This technology has a wide range of applications in various fields such as industrial production, urban planning, healthcare, etc. For example, in manufacturing, digital twin can be used to monitor the running state of equipment, predict maintenance needs, and improve production efficiency and reduce costs. In urban planning, digital twin can simulate traffic flow, environmental pollution and energy consumption in cities, helping decision-makers make more intelligent planning and management decisions.

[0005] Fidelity of digital twin model: Fidelity of digital twin model refers to the accuracy and precision of the model, that is, to what extent it can reflect the behavior, performance and characteristics of the actual physical entity or system. In the context of digital twin model, fidelity is a very important concept because it determines the usability and effectiveness of the model in different application fields.

[0006] Fidelity of digital twin model is usually related to the following factors:

[0007] Model accuracy: The accuracy of the digital twin model depends on the degree of accurate modeling of the actual system. This includes the physical properties, engineering parameters, operating conditions, etc. of the model. If the accuracy of the model is high, it can more accurately simulate the behavior of the actual system.

[0008] Data matching degree: The fidelity of the digital twin model also depends on the quality and matching degree of the data used to build the model. If the input data of the model is highly matched with the input data of the actual system, the fidelity of the model is usually higher.

[0009] Model updating and calibration: Maintaining the fidelity of the digital twin model requires regular updating and calibration to ensure that the model remains consistent with changes in the actual system. This may require using actual observational data to revise the model to reflect real-world changes.

[0010] Numerical methods and algorithms: The numerical methods and algorithms used have an important influence on the fidelity of the model. Some advanced numerical techniques can improve the accuracy and stability of the model.

[0011] Model evaluation: Model evaluation is the process of assessing the accuracy and credibility of the digital twin model in simulating the real physical system. It usually includes analyzing the performance, precision, and applicability of the model.

[0012] Despite the significant progress made in the above technical fields, there are still some obvious defects and deficiencies that may limit their effectiveness and reliability in practical applications:

[0013] Data source limitations: Existing digital twin model evaluation methods are often limited by data sources, limiting them to specific fields or single data types. This leads to insufficient comprehensive analysis of multi-source data, failing to fully consider the impact of various data.

[0014] Lack of specific credibility evaluation methods: There is a lack of a comprehensive credibility evaluation method to ensure that the digital twin system can accurately reflect the physical system. The existing technology proposes a digital twin device, model evaluation system and model running method including data model interaction module, model updating module and resource calling module, which can realize the establishment of digital twin model hierarchy, function and other physical properties, guide the verification of physical entity device research and development while the communication equipment manufacturer researches and develops physical entity device. However, the invention patent does not explicitly propose how to evaluate the model, and the evaluation details of the model are not thoroughly studied.

[0015] Absolute difference is not clear: Existing evaluation methods do not necessarily provide a clear measurement method to measure the absolute difference between the virtual model of the digital twin model and the physical entity, i.e. fidelity. This makes it difficult to determine the accuracy of the model and the direction of improvement. SUMMARY

[0016] The purpose of the present invention is to provide a digital twin model evaluation method and system for intelligent manufacturing, to solve the problems of data source limitations, absolute difference not clear, and credibility evaluation method not specific in the prior art.

[0017] To achieve the above purpose, the technical scheme adopted by the present invention is as follows:

[0018] In a first aspect, the present application provides a digital twin model evaluation method for intelligent manufacturing, comprising:

[0019] n kinds of fidelity indexes of the digital twin model are obtained, and a primary index weight of the digital twin model fidelity is determined;

[0020] The n kinds of fidelity indexes correspond to n model evaluations, each model evaluation is divided into a plurality of secondary evaluation indexes, and a secondary index weight of each model evaluation is obtained according to the secondary evaluation indexes;

[0021] An index evaluation scale is calculated according to all the secondary evaluation indexes, a secondary score is calculated in combination with the secondary index weight, and a score of each model evaluation is obtained;

[0022] A comprehensive score is obtained by comprehensively scoring the score of each model evaluation.

[0023] Optionally, n kinds of fidelity indexes of the digital twin model are obtained:

[0024] The n kinds of fidelity indexes of the digital twin model include a geometric model fidelity, an electrical model fidelity, a mechanical model fidelity, a kinematics model fidelity and a twin virtual data fidelity.

[0025] Optionally, the primary index weight of the product digital twin model fidelity for intelligent manufacturing is:

[0026] F1=0.2G_mdq+0.2E_mfem+0.2K_me+0.2M_kme+0.2T_vdfem

[0027] Wherein, G_mdq is a geometric model fidelity evaluation index, E_mfem is an electrical model fidelity evaluation index, K_me is a kinematics model fidelity evaluation index, M_kme is a mechanical model fidelity evaluation index, and T_vdfem is a twin virtual data fidelity evaluation index.

[0028] Optionally, the n kinds of fidelity indexes correspond to n model evaluations, and each model evaluation is divided into a plurality of secondary evaluation indexes:

[0029] The n model evaluations include a geometric model, an electrical model, a mechanical model, a kinematics model and a twin virtual data model;

[0030] The geometric model is divided into a model basic feature Bmfe, a digital twin model information integrity Dsim and a user space cognitive feature Uscf;

[0031] The electrical model is divided into: demand layer model evaluation Dqem, logic layer model evaluation Lmem, software layer model evaluation Smev, communication layer model evaluation Cmev, hardware layer model evaluation Hmev, and geometric topology layer model evaluation Gmev.

[0032] The mechanical model is divided into: multi-body model verification evaluation rules NCmtmv, preprocessing process evaluation Ppev, and analysis result process evaluation Eapr

[0033] The kinematics model is divided into: dynamics model basic characteristic evaluation Dmce, STEP format part evaluation Egsc, XML description file evaluation Dexd, rigid-flexible coupled multi-body model construction and connection evaluation Cccm, model driven setting and result output evaluation Mdee, and multi-body model verification evaluation NCmtmv.

[0034] The twin virtual data model is divided into: data interaction reliability Dirl and data quality reliability Dqtr.

[0035] G_mdq = 0.25B mfe + 0.25D sim + 0.50U scf

[0036] E mfem = 0.210D qem + 0.177L mem + 0.162S mev + 0.162C mev + 0.178H mev + 0.111G mev

[0037] K me = 0.207D mce + 0.166E gsc + 0.189D exd + 0.170C ccm + 0.170M dee + 0.098N cmtmv

[0038] M kme = 0.4G dpv + 0.2P pev + 0.4E apr

[0039] T vdfem = 0.500D irl + 0.500D qtr .

[0040] Optionally, the index evaluation scale is calculated according to all the secondary evaluation indexes, the secondary score is calculated in combination with the secondary index weight, and the score of each model evaluation is obtained.

[0041] The index evaluation scale includes complete non-compliance, basic non-compliance, basic compliance, relatively compliance and complete compliance; each evaluation index in the model basic feature is scored, and an average value is obtained, and the score is the model basic feature score.

[0042] Optionally, a comprehensive score is obtained by comprehensive scoring according to the score of each model evaluation.

[0043] The secondary score is calculated into the overall score of the entire digital twin model evaluation by a weighting formula through the first weight.

[0044] In a second aspect, the present application provides a digital twin model evaluation system for intelligent manufacturing, comprising:

[0045] The data acquisition module is configured to acquire n kinds of fidelity indexes of the digital twin model, and determine the first index weight of the fidelity of the digital twin model.

[0046] The secondary index weight acquisition module is configured to correspond to n model evaluations for the n kinds of fidelity indexes, divide each model evaluation into a plurality of secondary evaluation indexes, and acquire the secondary index weight of each model evaluation according to the secondary evaluation indexes.

[0047] The model evaluation score acquisition module is configured to calculate the index evaluation scale according to all the secondary evaluation indexes, calculate the secondary score in combination with the secondary index weight, and obtain the score of each model evaluation.

[0048] The evaluation output module is configured to obtain a comprehensive score by comprehensive scoring according to the score of each model evaluation.

[0049] In a third aspect, the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to realize the steps of the digital twin model evaluation method for intelligent manufacturing.

[0050] In a fourth aspect, the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize the steps of the digital twin model evaluation method for intelligent manufacturing.

[0051] Compared with the prior art, the present application has the following technical effects:

[0052] 1. Multi-modal multi-directional data collection for digital twin model. The present application starts from the components of the digital twin model, divides the evaluation of the digital twin into five parts of geometry, electricity, kinematics, mechanics and virtual-real data, uses multi-angle full data for digital twin model evaluation, and improves the comprehensiveness and reliability of model evaluation.

[0053] 2. Concretize the reliability evaluation method. The evaluation process of the digital twin model evaluation method proposed in the present application is analyzed and evaluated layer by layer according to the analytic hierarchy process, and the whole flow of evaluation is complete and effective. The reliability degree of the digital twin model can be more accurately reflected.

[0054] 3. Full display of the absolute difference between virtual model and physical entity. The existing evaluation method does not necessarily provide a clear measurement method to measure the absolute difference between the virtual model of the digital twin model and the physical entity, which makes it difficult to determine the accuracy of the model and the direction of improvement. The present application divides the whole digital twin model into three layers for evaluation, proposes a scoring evaluation method containing multiple evaluation details, and proposes an evaluation grade suitable for each evaluation detail, which has a more comprehensive display in evaluating physical models and twin virtual models. Compared with the existing patent, the present application has more comprehensive advantages. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 The flowchart of the present application.

[0056] Figure 2 The logic block diagram of the present application.

[0057] Figure 3 The two-level index division diagram of the present application. DETAILED DESCRIPTION

[0058] The present application is further described below in conjunction with the drawings:

[0059] Please refer to Figures 1 to 3 A digital twin model evaluation method for intelligent manufacturing, comprising:

[0060] Obtain n kinds of fidelity indexes of the digital twin model, and determine the primary index weight of the digital twin model fidelity;

[0061] The n kinds of fidelity indexes correspond to n model evaluations, each model evaluation is divided into multiple secondary evaluation indexes, and the secondary index weight of each model evaluation is obtained according to the secondary evaluation indexes;

[0062] The index evaluation scale is calculated according to all the secondary evaluation indexes, the secondary score is calculated combined with the secondary index weight, and the score of each model evaluation is obtained;

[0063] The comprehensive score is obtained by synthesizing the scores evaluated according to each model.

[0064] Specifically,

[0065] The first index determination method

[0066] For the fidelity of the digital twin model, the importance of the three indexes is basically the same. The geometric model fidelity, the electrical model fidelity, the mechanical kinematics model fidelity, and the twin virtual data fidelity.

[0067] The geometric, electrical mechanism, mechanical model, and data model support each other and are indispensable. Therefore, the weight of the first index of the product digital twin model fidelity facing intelligent manufacturing is determined as:

[0068] F1=0.2G_mdq+0.2E_mfem+0.2K_me+0.2M_kme+0.2T_vdfem

[0069] Wherein: G_mdq is the geometric model fidelity evaluation index, E_mfem is the electrical model fidelity evaluation index, K_me is the kinematics model fidelity evaluation index, M_kme is the mechanical model fidelity evaluation index, and T_vdfem is the twin virtual data fidelity evaluation index.

[0070] Second evaluation index details

[0071] In the evaluation of geometric-electrical-kinematics-mechanical-virtual-real data model, each model evaluation is divided into multiple second evaluation indexes and third evaluation details. The second index division is as follows Figure 3 .

[0072] Second index weight determination method

[0073] The geometric model is divided into: model basic feature Bmfe, digital twin model information integrity Dsim, and user space cognitive feature Uscf;

[0074] The electrical model is divided into: demand layer model evaluation Dqem, logic layer model evaluation Lmem, software layer model evaluation Smev, communication layer model evaluation Cmev, hardware layer model evaluation Hmev, and geometric topology layer model evaluation Gmev;

[0075] The mechanical model is divided into: multi-body model verification evaluation details NCmtmv, preprocessing process evaluation Ppev, and analysis result process evaluation Eapr

[0076] Kinematics model is divided into: dynamic model basic characteristics evaluation Dmce, STEP format parts evaluation Egsc, XML description file evaluation Dexd, rigid-flexible coupling multi-body model construction and connection evaluation Cccm, model driven setting and result output evaluation Mdee, multi-body model verification evaluation NCmtmv.

[0077] Twin virtual data model is divided into: data interaction credibility Dirl, data quality credibility Dqtr.

[0078] 1. Model basic characteristics, model information integrity, user space cognition

[0079] Compared with model basic characteristics and model information integrity, user space cognition is more result-oriented, which is the comparison of the results after the system is basically completed, and can best reflect the fidelity of digital twin geometric model. Based on the above theory, the comparison matrix is shown in the table.

[0080] Table 1 Geometric model weight

[0081]

[0082] The result passes the consistency test, and the weight of the secondary index of geometric model fidelity Fl is:

[0083] G_mdq=0.25B mfe +0.25D sim +0.50U scf

[0084] 2. Demand layer, logic layer, software layer, communication layer, hardware layer, geometric topology layer

[0085] Table 2 Electrical model weight

[0086]

[0087] E mfem =0.210D qem +0.177L mem +0.162S mev +0.162C mev +0.178H mev +0.111G mev

[0088] 3. Dynamic model basic characteristics evaluation, STEP format parts evaluation, XML description file evaluation, rigid-flexible coupling multi-body model construction and connection evaluation, model driven setting and result output evaluation, multi-body model verification evaluation of numerical control machine tool.

[0089] Table 3 Kinematics model weight

[0090]

[0091] K me = 0.207D mce + 0.166E gsc + 0.189D exd + 0.170C ccm + 0.170M dee + 0.098N cmtmv

[0092] 4. Meshing process, pre-processing process, analysis result process

[0093] Table 4. Weight of mechanical model

[0094]

[0095] M kme = 0.4G dpv + 0.2P pev + 0.4E apr

[0096] 5. Data interaction credibility Dir1, data quality credibility Dqtr

[0097]

[0098] T vdfem = 0.500D irl + 0.500D qtr

[0099] Determination of evaluation scale of three-level indicators

[0100] For 57 evaluation details, each detail has five levels of evaluation. Through the evaluation of each detail, the final overall evaluation score is obtained.

[0101] Evaluation level Score Not at all 0.0 Not very much 0.3 Quite a lot 0.6 Very much 0.8 Figure 1 1.0

[0102] All fidelity evaluation details have the same evaluation scale based on the model basic features, as shown in the table. The average of the scores of the 57 items is the model information integrity score.

[0103] By setting the first-level weight, the second-level weight and the third-level evaluation details, the second-level score is converted from the third-level indicator score based on the second-level weight, and then the second-level score is calculated into the overall score of the entire digital twin model evaluation by using the weighted formula based on the first-level weight.

[0104] Model basic features Bmfe

[0105] In the construction process of digital twin model, mechanical geometry model plays a key role because it directly maps the actual physical structure of the CNC machine. This model not only provides the basis for subsequent performance analysis, but more importantly, it lays the necessary foundation for the functional implementation of the entire digital twin system. The visualization modeling process of mechanical geometry model is divided into several main driving parts, including sketch, reference surface, feature, and association. Such division makes the model construction process more systematic and controllable. The accuracy and scientificity of the basic features of the model are crucial, as they not only provide the interface and storage location for various information in the model evolution, but also demonstrate the authenticity of the model through eye-catching visual effects. To ensure the consistency of virtual entities with actual physical entities, the construction method of model basic features must be consistent with the organization structure of the actual object, and the association of key information also needs to be consistent.

[0106] To ensure the establishment of digital twin geometry model with high quality, we have developed detailed evaluation rules covering sketch, reference surface / reference axis, feature, association, assembly relationship, and user experience, totaling 27 items. The development of these rules helps to ensure that the basic features of the model can be accurately represented in the model, and also helps to ensure the portability of the model in different application scenarios.

[0107] Digital twin model information integrity Dsim

[0108] In the construction process of digital twin model, it is particularly important to ensure the integrity of information. This measure can ensure that we can accurately simulate, analyze and predict the key features and behaviors in the actual system. Information integrity is particularly important in this context, covering multiple key aspects, including process information, quality information, resource information, file information, personnel information, and environmental information. The comprehensive integration of these aspects builds the foundation of the digital twin model, ensuring that the model can reliably simulate and predict in various different situations.

[0109] Electrical model fidelity evaluation index E_mfem

[0110] The evaluation of the digital twin model will be conducted from the perspectives of demand management, logical function architecture, system software architecture and hardware network topology, communication network design, hardware electrical principle, wire harness details, and installation location and wire harness geometry topology.

[0111] Demand layer model evaluation Dqem

[0112] Subdivision of demand layer modeling for two demand description perspectives

[0113] The first one is Customer Feature, which is the classification and abstraction of requirements from the perspective of user usage, such as engine automatic start-stop, constant speed cruise, etc. The focus of customer feature design is to reflect the difference of configuration requirements for intelligent manufacturing variants of different levels, and to realize the management of high and low configuration variants of target vehicle intelligent manufacturing.

[0114] The second one is Requirements, which is the functional description given to developers from the perspective of engineering implementation, such as the specific working temperature range of a part. In cross-layer mapping, the functional requirement analysis biased towards technical description is usually mapped to the specific software or hardware modules at the lower level. The two together form the requirement layer model. In addition, the functional requirements can be further improved by linking the two according to the relevant specifications.

[0115] Logical layer model evaluation Lmem

[0116] The logical layer refers to the transition layer that uses functional logic models for verification in the design stage of intelligent manufacturing. It connects the requirement model and the software and hardware model, and realizes the description of system architecture and subsystem functional logic through abstract modeling, and provides a basis for subsequent software and hardware development. In the logical layer, the system is divided into a series of functional abstractions, which mainly come from the division of system functions in the requirement layer. This part will use PREEvision software as an example to use some of its professional terms.

[0117] The logical architecture diagram is the main form of the logical function architecture layer. In the actual modeling work of the logical layer model, model components are generated by instantiating meta-types and defined and constructed according to the specific development process. Components and logical relationships at the same level in the diagram layer show the same logical hierarchy, and more detailed functional module modeling is expanded through containment relationships.

[0118] The logical layer generally takes user operation behavior (Actor), sensor module (Sense) and actuator module (Actuator) as the logical input or output endpoints in the system function diagram.

[0119] Evaluation: division and definition of meta-types.

[0120] Software layer model evaluation Smev

[0121] The main purpose of the software layer model is to realize the graphical visualization of system software architecture design using software engineering methods. In addition to basic UML diagram support, this layer uses software system architecture diagrams to define and model the behavior of software components and their interfaces. In this project, component type diagrams are mainly used to model the communication behavior between software architecture and functional components.

[0122] Communication layer model evaluation Cmev

[0123] The communication layer model defines how software layer components interact with each other across the boundaries of hardware domains through gateways. It maps the upper layer to specific functional logic modules and the lower layer to specific carrier harnesses and ECUs. It spans across the logical, software, and hardware layers and associates them through mapping relationships and signal interfaces. The main modeling objects in this layer include signals, message attribute information, and corresponding signal routing settings in the communication matrix. It supports bus network protocols such as CAN, CANFD, LIN, FlexRay, and Ethernet communication.

[0124] Hardware layer model evaluation Hmev

[0125] The digital twin hardware layer includes three sub-layers: network architecture topology, electrical schematic layer, and harness schematic layer. This layer models and abstracts physical entities such as ECUs, sensors, actuators, batteries, junction boxes, and ground points. The network topology layer generally does not include power and ground-related elements, while the electrical schematic layer and harness layer reflect power supply and ground point design. The electrical modules used in network topology diagrams, electrical schematics, and harness schematics are reusable, but the connection relationships between modules within each layer are determined by the nature of the diagrams.

[0126] Network topology diagrams are used to model the communication network physical architecture between ECUs involved in communication buses, including connectors and bus systems. Electrical schematics reflect power supply configurations, ground point designs, and electrical properties within and outside the subsystems. The harness schematic in the harness layer refines the abstract connection relationships and ports in the electrical schematic to specific details such as the number and type of actual harnesses. Gray represents hard-wired connections, green represents bus connections, orange represents power connections, and brown represents ground connections. The harness layer defines physical details such as pin, connector, and cable properties.

[0127] Geometric topology layer model evaluation Gmev

[0128] The topology layer, also known as the geometric topology layer, mainly models the installation space and relative installation positions of system components and harnesses. Corresponding to the upper layer, the geometric topology diagram in this layer shows the installation position of ECU units in the actual vehicle physical space, and the harness topology diagram is an abstract mapping of the harness design diagram in the actual physical space. The geometric model of the vehicle contains real component size information and can exchange model information with other CAD tools through compatible data formats.

[0129] Mechanical model evaluation (finite element analysis method) M_kme

[0130] Fidelity evaluation of digital twin kinematic models plays a crucial role in intelligent manufacturing, robotics, and CNC machine tools. Its significance lies in the following key aspects:

[0131] Accuracy verification: Fidelity evaluation of digital twin kinematic models is an important means to verify the accuracy between virtual models and actual systems. By comparing the prediction results of virtual models with the motion behavior of actual systems, the reliability and accuracy of the model can be objectively evaluated to ensure its effectiveness in real applications.

[0132] Optimization and improvement: Fidelity evaluation can find the differences and deviations between virtual models and actual systems. These differences may be due to the limitations of the model or may reflect errors in the actual system parameters. By accurately understanding the differences, the virtual model can be optimized and improved to make it closer to the actual system motion behavior.

[0133] Control and planning reliability: The fidelity of digital twin kinematic models directly affects the reliability of control algorithms and path planning. By establishing an accurate virtual model, it can provide reliable input data to improve the motion accuracy and stability of the system, providing reliable protection for the control and planning process.

[0134] Cost reduction: Fidelity evaluation is tested and optimized in a virtual environment, avoiding trial and error and experiments on actual systems, thereby saving valuable time and resources. Such cost-effectiveness will help improve the efficiency and economy of the manufacturing process.

[0135] Predictive ability: Accurate digital twin kinematic models can be used to predict the motion behavior of the system, such as the attitude and motion trajectory of the robot under different working conditions. This provides a strong basis for system design and decision-making, helping to solve potential problems and optimize the manufacturing process in advance.

[0136] Fault diagnosis: Fidelity evaluation helps to find the differences between virtual models and actual systems, making it easier to detect and diagnose faults and abnormal conditions in actual systems. This will help improve the reliability and stability of the equipment

[0137] The development of the arch dam digital twin health assessment platform relies on the arch-beam load distribution method, finite element simulation method, and statistical data analysis method. The arch-beam load distribution method is widely used in the field of arch dam design in China and has been proven by engineering, highly recognized by industry experts and engineers, and has unique advantages, but also has some shortcomings, i.e., it cannot consider the local mechanical state of the arch dam. With the development of arch dam calculation methods and computer technology, using the finite element method to calculate arch dam stress can overcome the shortcomings of the arch-beam load distribution method and better reflect the true stress state of the local arch dam. In the arch dam design specification, it is clearly stated that for high arch dams and relatively complex arch dams, the finite element equivalent stress method should be used for calculation.

[0138] In finite element analysis, first, the structure or object to be studied or designed is divided into multiple small units, and relevant parameters such as the geometric shape, material properties, and boundary conditions of each small unit are established to obtain a mathematical model representing the small unit. Next, by solving the equations of each small unit, the solution of the entire system is obtained. Finally, based on the solution results, the required stress, strain, displacement, and other information are obtained through post-processing.

[0139] The structural finite element open-source mechanical solver Code_aster is used for structural state calculation and analysis of complex products. Code_aster is widely used in EDF's nuclear energy engineering field and has good engineering applicability. For example, in seismic analysis of concrete dams, Code_aster's substructure method can be used for fluid-structure coupling to evaluate the seismic performance of complex product structures. Code_aster is good at solving mechanical problems such as structural thermodynamics analysis, linear and nonlinear static and dynamic analysis, and acoustic analysis. Code_aster also supports user secondary development, providing good development functions for users with different engineering needs.

[0140] In finite element analysis, first, the structure or object to be studied or designed is divided into multiple small units, and relevant parameters such as the geometric shape, material properties, and boundary conditions of each small unit are established to obtain a mathematical model representing the small unit. Next, by solving the equations of each small unit, the solution of the entire system is obtained. Finally, based on the solution results, the required stress, strain, displacement, and other information are obtained through post-processing.

[0141] Data quality credibility Dqtr

[0142] As an innovative data analysis method, digital twin models have shown strong application potential in various fields. However, to fully utilize the advantages of digital twin models and ensure their accurate simulation and analysis of actual objects, attention should be paid to several key data management principles: standardization, completeness, accuracy, consistency, timeliness, and accessibility.

[0143] Standardization: The standardization of data refers to ensuring that data is collected, stored, and processed according to uniform standards and definitions. This includes consistency in data naming, units, classification, format, and more. Standardization helps eliminate data ambiguity, ensuring that data can be correctly understood and interpreted, resulting in accurate results in digital twin models.

[0144] Completeness: The completeness of data ensures that every data element in the model meets the required rules and is not missing. This involves not only the presence of data but also its accuracy, comprehensiveness, and timeliness. Maintaining data completeness is the foundation of building reliable and practical digital twin models, as missing data can affect the accuracy and value of the model.

[0145] Accuracy: The accuracy of data is a core attribute of digital twin models. Ensuring the accuracy of data input, as well as the careful consideration of model design, algorithms, and parameters, is essential to improve the model's ability to accurately describe and predict the actual object. Any data errors or inaccuracies can lead to inaccurate results in the model output.

[0146] Consistency: The consistency of data requires that data remains uniform across multiple dimensions, whether in terms of specified standards or logical relationships between data. In digital twin models, consistency is particularly important, especially when facing the challenges of data evolution and updates, to ensure that the model's output is always reliable and consistent.

[0147] Timeliness: Timeliness means that data should be collected, updated, and applied in a timely manner. For digital twin models, real-time or near-real-time data is crucial for simulating the actual object's state and behavior. Outdated data may cause the model to lose its accurate reflection of the actual situation, affecting the model's practicality.

[0148] Accessibility: The accessibility of data ensures that those who need it can easily access and use it. This requires appropriate data storage and sharing mechanisms, as well as controlling access rights to prevent unauthorized access. Data accessibility is crucial for the cooperation and application of digital twin models.

[0149] As an abstract entity, the data model is difficult to evaluate directly. This paper mainly evaluates the effect of the application business of the digital twin model to indirectly reflect the key properties of the digital twin data model. So far, the most authoritative data model quality evaluation index standard is GB / T36344-2018 ICS 35.24.01. Combined with the characteristics and needs of the digital twin model, the following indicators are extracted: (1) Integrity: The degree to which elements are assigned values according to data rules. The integrity of the digital twin data model is the basis and premise for realizing the function. The missing data set will accumulate in the system, gradually reducing the available value of the system. (2) Consistency: Requires data to follow uniform standards and maintain uniform formats, mainly reflected in standardization and logicality, and data should have standard coding rules. Consistency is particularly important for digital twin models with complex evolution processes. (3) Accuracy: Indicates the degree of true value of the real entity (actual object) described. (4) Timeliness: The time interval (data delay time) from data generation to viewable. If the data analysis period is too long, the overall complex product digital twin model loses its timeliness, resulting in a loss of reference significance for analysis conclusions. The evaluation of the data model needs to be assisted by professional data analysis software.

[0150] In another embodiment of the present application, a digital twin model evaluation system for intelligent manufacturing is provided, which can be used to realize the above-mentioned digital twin model evaluation method for intelligent manufacturing. Specifically, the system comprises:

[0151] A data acquisition module is configured to acquire n kinds of fidelity indicators of the digital twin model and determine the primary indicator weight of the digital twin model fidelity;

[0152] A secondary indicator weight acquisition module is configured to correspond n kinds of fidelity indicators to n model evaluations, divide each model evaluation into a plurality of secondary evaluation indicators, and acquire the secondary indicator weight of each model evaluation according to the secondary evaluation indicators;

[0153] A model evaluation score acquisition module is configured to calculate the indicator evaluation scale according to all secondary evaluation indicators, combine the secondary indicator weight to calculate the secondary score, and obtain the score of each model evaluation;

[0154] An evaluation output module is configured to obtain a comprehensive score by comprehensively scoring each model evaluation score.

[0155] The division of the modules in the embodiments of the present application is illustrative, and is merely logical function division. In actual implementation, another division manner can be used. In addition, each function module in each embodiment of the present application can be integrated in one processor, or can be physically separated, or two or more modules can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software function module.

[0156] In still another embodiment of the present application, a computer device is provided, which includes a processor and a memory. The memory is configured to store a computer program, and the computer program includes program instructions. The processor is configured to execute the program instructions stored in the computer storage medium. The processor can be a central processing unit (CPU), and can also be another general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or another programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and the like. The processor is a computing core and a control core of the terminal, and is suitable for implementing one or more instructions, and is specifically suitable for loading and executing one or more instructions in the computer storage medium to implement a corresponding method flow or a corresponding function. The processor in the embodiments of the present application can be used for the operation of the intelligent manufacturing-oriented digital twin model evaluation method.

[0157] In still another embodiment of the present application, the present application also provides a storage medium, specifically a computer readable storage medium (Memory), which is a memory device in a computer device, used for storing programs and data. It can be understood that the computer readable storage medium herein can include a built-in storage medium in the computer device, and of course can also include an extended storage medium supported by the computer device. The computer readable storage medium provides a storage space, which stores an operating system of the terminal. Moreover, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and the instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM memory, or a non-volatile memory such as at least one disk memory. The one or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to realize the corresponding steps of the method for evaluating the digital twin model for intelligent manufacturing in the above embodiment.

[0158] Those skilled in the art should understand that embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0159] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks Figure 1 The means for performing the functions specified in one or more flows and / or blocks.

[0160] These computer program instructions can also be stored in a computer readable memory capable of directing the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including instruction means, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocksFigure 1 the function specified in the one or more blocks.

[0161] These computer program instructions can also be loaded into computer or other programmable data processing devices, so that a series of operation steps are performed on the computer or other programmable data processing devices to generate computer-implemented processes, thus the instructions executed on the computer or other programmable data processing devices provide processes for implementing the flows Figure 1 the flows or the plurality of flows and / or blocks ​ the steps of the function specified in the one or more blocks.

[0162] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application, but not to limit it, although the above embodiments of the present application have been described in detail, those skilled in the art should understand: the specific embodiments of the present application can be modified or replaced by the same, without departing from the spirit and scope of the present application, any modification or equivalent replacement, which should be covered within the scope of protection of the claims of the present application.

Claims

1. A method for evaluating digital twin models for intelligent manufacturing, characterized in that, include: Obtain n fidelity indices for the digital twin model and determine the weights of the primary fidelity indices for the digital twin model. n fidelity indicators correspond to n model evaluations. Each model evaluation is divided into multiple secondary evaluation indicators. The weight of each secondary evaluation indicator is obtained based on the secondary evaluation indicators. The evaluation scale is calculated based on all secondary evaluation indicators, and the secondary score is calculated by combining the weights of the secondary indicators to obtain the score of each model evaluation. A comprehensive score is obtained by combining the scores of each model evaluation. Obtain n fidelity metrics for digital twin models: The n fidelity metrics of digital twin models include geometric model fidelity, electrical model fidelity, mechanical model fidelity, kinematic model fidelity, and twin virtual data fidelity; The primary indicator weight for the fidelity of digital twin models of products for intelligent manufacturing is as follows: in: As an evaluation index for the fidelity of geometric models, As an evaluation index for the fidelity of electrical models, As an evaluation index for the fidelity of kinematic models, As an evaluation index for the fidelity of mechanical models, As an evaluation index for the fidelity of twin virtual data; There are n fidelity metrics corresponding to n model evaluations, and each model evaluation is divided into multiple secondary evaluation metrics: The n model evaluations include geometric models, electrical models, mechanical models, kinematic models, and twin virtual data models; The geometric model is divided into: basic model features (Bmfe), digital twin model information integrity (Dsim), and user spatial cognitive features (Uscf). The electrical model is divided into: Demand Layer Model Evaluation (Dqem), Logic Layer Model Evaluation (Lmem), Software Layer Model Evaluation (Smev), Communication Layer Model Evaluation (Cmev), Hardware Layer Model Evaluation (Hmev), and Geometric Topology Layer Model Evaluation (Gmev). The mechanical model is divided into: Multibody Model Validation and Evaluation Details (NCmtmv), Preprocessing Process Evaluation (Ppev), and Analysis Result Process Evaluation (Eapr). The kinematic model is divided into: evaluation of basic characteristics of dynamic model (Dmce), evaluation of STEP format components (Egsc), evaluation of XML description file (Dexd), evaluation of rigid-flexible coupling multibody model construction and connection (Cccm), evaluation of model-driven setting and result output (Mdee), and evaluation of multibody model verification (NCmtmv). The twin virtual data model is divided into: data interaction credibility Dirl and data quality credibility Dqtr.

2. The evaluation method for digital twin models in intelligent manufacturing according to claim 1, characterized in that, The weights of the secondary evaluation indicators for each model are obtained based on the secondary evaluation indicators: 。 3. The evaluation method for digital twin models in intelligent manufacturing according to claim 1, characterized in that, The evaluation scale is calculated based on all secondary evaluation indicators, and the secondary scores are calculated by combining the weights of the secondary indicators, thus obtaining the score for each model evaluation: The evaluation scales for indicators include completely non-compliant, basically non-compliant, basically compliant, fairly compliant, and completely compliant; each evaluation indicator in the basic features of the model is scored, and the average score is calculated. This score is the basic feature score of the model.

4. The evaluation method for digital twin models in intelligent manufacturing according to claim 1, characterized in that, A comprehensive score is obtained by combining the scores from each model evaluation. The secondary scores are calculated using a weighted formula based on the primary weights to obtain the overall score for the entire digital twin model evaluation.

5. A digital twin model evaluation system for intelligent manufacturing that implements the digital twin model evaluation method for intelligent manufacturing as described in claim 1, characterized in that, include: The data acquisition module is used to acquire n fidelity indicators of the digital twin model and determine the weight of the primary fidelity indicator of the digital twin model. The secondary indicator weight acquisition module is used to evaluate n models corresponding to n fidelity indicators. It divides each model evaluation into multiple secondary evaluation indicators and obtains the weight of each secondary indicator for each model evaluation based on the secondary evaluation indicators. The model evaluation score acquisition module is used to calculate the indicator evaluation scale based on all secondary evaluation indicators, and calculate the secondary score by combining the weights of the secondary indicators, thus obtaining the score for each model evaluation. The evaluation output module is used to generate a comprehensive score based on the scores of each model evaluation.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the digital twin model evaluation method for intelligent manufacturing as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the digital twin model evaluation method for intelligent manufacturing as described in any one of claims 1 to 4.

Images