CN117993255A - Multi-sample cross-stage collaborative operation method for digital-real fusion test - Google Patents
Multi-sample cross-stage collaborative operation method for digital-real fusion test Download PDFInfo
- Publication number
- CN117993255A CN117993255A CN202410198797.4A CN202410198797A CN117993255A CN 117993255 A CN117993255 A CN 117993255A CN 202410198797 A CN202410198797 A CN 202410198797A CN 117993255 A CN117993255 A CN 117993255A
- Authority
- CN
- China
- Prior art keywords
- numerical simulation
- data
- sample
- test
- virtual
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012360 testing method Methods 0.000 title claims abstract description 185
- 238000000034 method Methods 0.000 title claims abstract description 42
- 230000004927 fusion Effects 0.000 title claims abstract description 22
- 238000004088 simulation Methods 0.000 claims abstract description 174
- 238000013461 design Methods 0.000 claims abstract description 29
- 238000013507 mapping Methods 0.000 claims abstract description 25
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 13
- 238000004458 analytical method Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 7
- 238000012795 verification Methods 0.000 claims description 33
- 238000005457 optimization Methods 0.000 claims description 22
- 238000000605 extraction Methods 0.000 claims description 15
- 230000014509 gene expression Effects 0.000 claims description 12
- 239000000463 material Substances 0.000 claims description 12
- 239000013598 vector Substances 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 230000007246 mechanism Effects 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 9
- 238000000513 principal component analysis Methods 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 4
- 230000001351 cycling effect Effects 0.000 claims description 2
- 238000010801 machine learning Methods 0.000 claims description 2
- 238000004519 manufacturing process Methods 0.000 abstract description 12
- 238000005516 engineering process Methods 0.000 abstract description 4
- 230000009466 transformation Effects 0.000 abstract description 2
- 230000010354 integration Effects 0.000 abstract 1
- 238000006073 displacement reaction Methods 0.000 description 7
- 238000011156 evaluation Methods 0.000 description 4
- 238000013031 physical testing Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 238000012942 design verification Methods 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 230000000704 physical effect Effects 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000012938 design process Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 238000003466 welding Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method for performing a digital-real fusion test on multi-sample cross-stage collaborative operation, which belongs to the field of electric digital data processing and predicts the mechanical properties of samples by using numerical simulation technologies such as finite element analysis and the like; selecting a plurality of representative samples to manufacture sample objects for physical test, and acquiring actual test data such as stress, deformation and the like of each sample through a sensor; finally, establishing virtual quanta, fusing a data set of a numerical simulation result with an actual test data set, and establishing virtual representation of a sample through algorithms such as transformation, mapping and the like to realize consistency of the same sample in different test stages; the collaborative operation of a plurality of test stages is realized, the design and the test of the actual sample are guided by utilizing the numerical simulation result, and the numerical simulation model is verified and perfected by the actual test result, so that the multi-sample multi-stage collaborative closed loop is realized. The invention realizes effective integration of numerical simulation and actual test, reduces the number of samples to be manufactured, reduces the test cost and improves the accuracy of the simulation model.
Description
Technical Field
The invention belongs to the field of electric digital data processing, and particularly relates to a method for performing a digital-real fusion test on multi-sample cross-stage collaborative operation.
Background
Traditional product testing mainly relies on actual sample manufacturing and physical testing, and the method is long in period and high in cost. It is difficult to build a high-precision model using only numerical simulation. The existing method lacks an effective way to combine numerical simulation with actual test to realize complementary fusion and synergy of the numerical simulation and the actual test. The conventional product testing method has the following problems:
(1) Relying solely on actual sample testing
The traditional full-reliance practical sample testing method needs to design and manufacture a large number of sample objects for comprehensive verification, and the period from the design to the sample manufacture and the test is very long, usually several months or even longer, which is difficult to meet the requirement that the iteration speed of industries such as automobiles is increased. Meanwhile, the manpower and material costs of a large number of actual samples are extremely high, and the development of a brand new automobile can generally consume hundreds of millions or even hundreds of millions of yuan. In addition, there is also a risk of damage in the actual test, which may result in damage or rejection of the sample, thereby increasing the overall test cost.
(2) Depending on numerical simulation only
There are also a number of challenges to relying solely on computer-simulated numerical simulations. Firstly, the theoretical model is generally required to be simplified and assumed for the convenience of calculation, and all influencing factors in the actual situation are difficult to consider, so that the accuracy of the simulation model is difficult to guarantee, and a certain potential safety hazard exists when the simulation model is directly applied to the actual product design. Secondly, the actual materials and all physical effects in the process cannot be obtained only by theoretical calculation, and the reliability of the result is reduced.
(3) Both are disjointed
The existing technology faces the fracturing between numerical simulation and actual test, and lacks an effective combination mechanism. This fracturing results in an unsmooth join and iteration from the simulation design to the physical verification. Therefore, in the existing system, visual physical verification of the simulation result cannot be realized, and coherent iteration and optimization are difficult to realize in the design process. This limits the engineering field to comprehensive evaluation of system performance and reliability, and there is an urgent need for an innovative method capable of bridging the gap between simulation and actual testing.
In summary, the prior art has the problems of large number of actual samples and difficult guarantee of simulation precision, and the two are not effectively fused, so that the respective advantages are difficult to develop and the cooperation is difficult to realize.
Disclosure of Invention
In order to overcome the limitation of relative fracture between numerical simulation and actual test in the existing product test technology, the invention provides a multi-sample cross-stage collaborative operation method for a numerical fusion test, which predicts the mechanical properties of samples by using numerical simulation technologies such as finite element analysis and the like; selecting a plurality of representative samples to manufacture sample objects for physical test, and acquiring actual test data such as stress, deformation and the like of each sample through a sensor; finally, establishing virtual quanta, fusing a data set of a numerical simulation result with an actual test data set, and establishing virtual representation of a sample through algorithms such as transformation, mapping and the like to realize consistency of the same sample in different test stages; the collaborative operation of a plurality of test stages is realized, the design and the test of the actual sample are guided by utilizing the numerical simulation result, and the numerical simulation model is verified and perfected by the actual test result, so that the multi-sample multi-stage collaborative closed loop is realized. The invention realizes the penetration from design to verification, plays the advantages of the design and the verification, and improves the test efficiency.
In order to achieve the above purpose, the invention adopts the following technical scheme:
a multi-sample cross-stage collaborative operation method for a digital-real fusion test comprises the following steps:
Step 1, establishing a virtual test environment integrating data of a numerical simulation result and actual test data, wherein the virtual test environment is constructed based on a software platform and is used for storing and processing the data to realize fusion of the data of the numerical simulation result and the actual test data;
Step 2, predicting the performance of a plurality of different samples by adopting a finite element numerical simulation method, and performing physical test to obtain actual test data; then, extracting key features from the data of the numerical simulation result and the actual test data by using a feature extraction method, wherein the key features comprise the elastic modulus, the yield strength and the fatigue life of the material; the properties include mechanical properties, thermal properties; the actual test data comprises stress and deformation; the feature extraction method comprises a machine learning algorithm;
Step 3, establishing a corresponding relation between key features of data of a numerical simulation result and key features of actual test data in the same coordinate system by utilizing a principal component analysis mapping algorithm, and finally forming virtual features representing samples, namely virtual quanta;
step 4, importing virtual quanta of different samples in a virtual test environment, and comparing the data of the numerical simulation result of each sample with key features of actual test data to realize consistency verification of the same sample in different test stages;
And 5, establishing a closed loop iteration mechanism, evaluating a result after each round of actual test and updating a numerical simulation model, and then predicting based on the updated numerical simulation model, so that the loop is circulated until the precision requirement is met, mutual verification and optimization of simulation and test are realized, the whole process of the physical test from the numerical simulation in the design stage to the verification stage is cooperatively operated, and the comprehensive verification of multiple samples is carried out.
Further, in the step 3, forming the virtual quantum of the sample includes:
Step 3.1, respectively extracting characteristics aiming at data of a numerical simulation result and actual test data, wherein the step comprises the following steps:
step 3.1.1, obtaining geometric dimensions and material properties in data of a numerical simulation result by adopting a parameter extraction method as characteristics;
step 3.1.2, a stress-strain curve and a deformation mode of actual test data are obtained by adopting a signal analysis method as characteristics;
and 3.2, selecting a principal component analysis mapping algorithm to realize the mapping relation between the characteristics of the data of the numerical simulation result and the characteristics of the actual test data, wherein the method comprises the following steps:
step 3.2.1, normalizing the characteristics of the data of the numerical simulation result and the characteristics of the actual test data, and eliminating the dimension influence;
Step 3.2.2, constructing a principal component analysis model, and finding out the principal component which can most represent the original data in the characteristics of the data of the numerical simulation result and the characteristics of the actual test data;
Step 3.2.3, determining a conversion relation between principal components, and establishing a linear or nonlinear mapping function;
Step 3.2.4, converting the characteristics of the data of the numerical simulation result and the characteristics of the actual test data into the same coordinate system by utilizing a linear or nonlinear mapping function;
step 3.3 generates virtual quanta representing virtual features of the sample, comprising:
Step 3.3.1, converting the characteristics of the data of the numerical simulation result extracted in the step 3.1 and the characteristics of the actual test data into the same expression system according to the mapping relation obtained in the step 3.2;
Step 3.3.2, describing virtual quanta of the sample by adopting a quantum state vector representation method; the virtual quantum of the sample integrates the characteristic information of the data of the numerical simulation result and the actual test data; the virtual quanta of a sample uniquely represent a virtual feature of the sample.
Further, the evaluating the result and updating the numerical simulation model after each round of actual test in the step 5 includes:
Step 5.1, collecting stress-strain curves and deformation modes of samples obtained by actual tests;
step 5.2, comparing the actual test data with the data of the current numerical simulation result to find a scene with larger error;
step 5.3, analyzing the difference between the actual test conditions and the simulation settings, and determining parameters of a numerical simulation model which causes errors;
Step 5.4, adjusting parameters of the numerical simulation model to conduct directional optimization, wherein the parameters comprise loading force, constraint and material properties;
And 5.5, rerun the simulation calculation based on the parameters of the optimized numerical simulation model to obtain the data of a new numerical simulation result for improving the precision.
Further, the predicting based on the updated numerical simulation model in the step 5 includes:
step 5.6, saving the parameter setting of the numerical simulation model and the numerical simulation model after the previous round of optimization;
step 5.7, running a new round of test of the numerical simulation model under the same setting as the actual test condition;
and 5.8, recording simulation prediction results of the optimized numerical simulation model under the same conditions.
Further, the cycling in step 5 until the accuracy requirement is met includes:
Step 5.9, evaluating the data of the numerical simulation result of the new round, and judging whether the error meets the precision requirement;
Step 5.10, if the data of the numerical simulation result meets the precision requirement, ending iteration; if the accuracy requirement is not met, the numerical simulation model is adjusted again;
Step 5.11 repeating the steps 5.2 to 5.10, and continuously iterating and optimizing the numerical simulation model until the error between the data of the numerical simulation result and the actual test data is smaller than a specified value;
And 5.12, obtaining a high-precision numerical simulation model for sample design optimization.
Compared with the prior art, the invention has the beneficial effects that:
(1) Different from the current experimental test and simulated fracture conditions, the invention realizes the effective fusion of numerical simulation and actual test through the virtual test environment, and truly realizes the full-flow through from design to verification.
(2) The invention establishes a closed loop iteration mechanism of simulation and test, can continuously optimize a simulation model instead of disposable use, and greatly improves the simulation accuracy.
(3) The virtual quantum concept provided by the invention realizes the correspondence between the numerical simulation result and the actual test data through feature extraction and mapping, and is a core innovation for realizing fusion.
(4) By means of virtual quanta, the invention realizes the consistency of the result expression of multiple samples in design simulation and actual test, and performs the collaborative verification of the multiple samples.
(5) The method realizes the integral closed loop optimization iteration of design, simulation and verification test, but not the current stage-by-stage disjoint state, and greatly improves the flow efficiency.
(6) Unlike the current test method requiring a large number of samples, the method reduces the number of actual samples, reduces the test cost and improves the economy.
Drawings
FIG. 1 is a flow chart of an implementation of a method for testing multi-sample cross-phase collaborative operation by digital-real fusion according to the present invention;
FIG. 2 is a flow chart of a method for testing multi-sample cross-phase collaborative operation by digital-to-real fusion according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
As shown in FIG. 1, the multi-sample cross-phase collaborative operation method for the digital-to-real fusion test comprises the following main parts:
Numerical simulation: the part is realized by two modules of a simulation model library and a solver, numerical simulation characteristics are obtained through characteristic extraction, and the numerical simulation characteristics are connected with the next flow.
Actual test part: the method is realized by two modules of a solid sample and test equipment, and the actual test characteristics are obtained through characteristic extraction and are connected with the next step of flow.
Virtual test environment: the method is realized by two modules of virtual quantum construction and mapping conversion, and closed loop iteration and optimization evaluation are carried out on ice. And obtaining an optimized parameter model through the optimal parameters.
Data fusion: is positioned under the virtual test environment and serves as a bridge for connecting the numerical simulation part and the actual test part.
In this flow, both the numerical simulation and the actual test section independently perform feature extraction, and these features are then imported into the virtual test environment. And constructing and mapping conversion of the virtual quanta in the virtual environment, and finally realizing optimization and iteration of the model through an optimization feedback link.
As shown in fig. 2, the method for performing the multi-sample cross-phase collaborative operation in the digital-real fusion test comprises the following steps:
step one: establishing a virtual test environment integrating the numerical simulation result and the actual test data;
The virtual test environment constructed by the invention is built based on a computer software system, and the actual product test scene and process are simulated through an advanced program algorithm. The virtual test environment is similar to a same digital double world, and has complete test elements, including a sample geometric model library, a material physical property database, a numerical solver module, a sample optimization platform and other digital modules. The digital modules work cooperatively to perform theoretical calculation and optimization so as to achieve the aim of assisting the physical test.
In the virtual test environment, two types of core data can be imported for storage and fusion application. The first type is data of a numerical simulation result obtained through calculation methods such as finite element analysis or multi-body dynamics analysis. The data reflects the theoretical performance of the sample under a mathematical model, and provides digital information such as stress distribution curves, displacement deformation diagrams and the like. The other is the actual test data from the actual physical bench. Various responses of the sample under the action of loading force, such as strain, displacement, vibration modes and the like, are collected through the precision sensor.
The organic fusion of two types of data is one of the cores of the virtual test environment. The numerical simulation results provide theoretical predictions, while the physical test data gives true responses. The two types of information are fused in the system database, and a foundation is provided for subsequent mapping conversion and closed loop iteration optimization. In this way, the digital model can be continuously compared and corrected, and the digital model is continuously approximated to the actual situation, so as to achieve various purposes, such as reducing the number of sample tests and the like. I.e., the modeling process of the virtual environment can be represented by the formula E virtual＝integrate(Ssim,Dtest), E virtual represents the virtual test environment, S sim represents the storage structure of the simulation result, and D test represents the actual test data.
Step two: in a virtual test environment, constructing virtual quanta of a sample through a feature extraction and mapping algorithm, and realizing the correspondence of digital and real data;
In the established virtual test environment, aiming at the data of the imported numerical simulation result, a parameter extraction technology is adopted to acquire important input conditions such as geometric model configuration, grid setting information, loading force and the like of the sample, and the numerical simulation feature vector of the sample is formed in a collating mode. And meanwhile, extracting important characteristic information such as stress-strain relation curves, displacement deformation modes and the like from original physical test data obtained by an actual physical test bed by using technical means such as signal analysis and the like, and constructing test characteristic vectors of the physical test bed. I.e., by feature extraction V fea＝extract(Dsim,Dtest), where V fea represents feature vectors, D sim represents simulation data, and D test represents test data.
After the numerical simulation features and the test features of two groups of samples with different sources are obtained, the invention uses mathematical algorithms such as principal component analysis and the like, and takes the numerical simulation feature vector and the test feature vector as input data to perform normalization processing and principal component conversion. And finally establishing a one-to-one mapping relation model between the characteristics of the numerical simulation result and the characteristics of the actual test data through a series of mathematical operations. The mapping from two domains to one domain realizes the effective correspondence of the simulation numerical result and the actual test data at the characteristic level.
According to the established mapping relation model, the original two independent numerical simulation feature vectors and the test feature vector can be converted into the same coordinate space, and the new virtual feature expression of the sample is obtained. The virtual characteristics of the numerical simulation result and the actual test data are fused, so that the virtual quantum of the sample is formed. The virtual quantum provides an important basis for realizing the cooperative test and closed loop iterative optimization of multiple samples.
Step three: the consistency of expressions of multiple samples in a numerical simulation domain and an actual test domain is realized by utilizing virtual quanta;
The present invention further extends to the case of covering a plurality of different samples after obtaining a virtual quantum representation of a single sample. In the product design stage, engineers may propose multiple solutions, corresponding to different sample options. These alternative samples are then subjected to in turn computer simulation numerical analysis, predicting their structural properties. The characteristics of the digital simulation result of each sample are obtained through technical means such as parameter extraction and the like, and the expression of the virtual quanta of the sample is further constructed. At the same time, several key samples representative thereof are selected for actual fabrication and physical testing. And acquiring various original data information such as strain, displacement and the like of the actual samples on an actual physical test bed. Through feature extraction and conversion, corresponding features of the actual test data and mapped virtual quanta are also formed. This process may be expressed as a consistency metric formula c=consistency (V sim,Vtest), where C represents a consistency score, and V sim and V test represent virtual quanta for simulation and testing, respectively.
To this end, in a previously established virtual test environment, data of virtual quanta of a plurality of different samples are imported. These data contain two types of information: firstly, the numerical calculation characteristics of the samples in the design simulation stage are expressed as virtual quanta; and the characteristics obtained by the physical test of the actual manufactured sample on the actual physical test bed are also expressed as virtual quanta. Each sample has both types of expression, and contains comprehensive information from design to reality.
Consistency verification can be performed by means of data of virtual quanta of multiple sets of samples. The virtual quanta obtained in the design simulation stage of the same specific sample are compared with the virtual quanta obtained in the actual test. If the difference of the characteristic values of the two is larger, the characteristic values indicate that the test and the simulation result have deviation. If the expressions are close to unity, the correctness of the test and simulation is verified. The consistency check realizes the effective correspondence of multiple samples in a numerical simulation domain and an actual test domain.
Step four: establishing a closed loop iteration mechanism of simulation and test, and realizing continuous optimization of the simulation and the test;
the invention establishes a closed loop iteration mechanism for mutual verification and optimization of simulation and test. The flow of the mechanism is: after each time the physical test is completed on the actual sample, the characteristics of the obtained actual test data are imported into a virtual test environment and are compared with the characteristics of the numerical simulation result predicted by the current simulation model. If the two are obviously inconsistent, the simulation link causing the error is positioned, and an error source such as loading force error setting is determined. And aiming at the error source, adjusting relevant parameters of the simulation model to optimize. And then running simulation trial calculation again based on the adjusted simulation model to generate a new round of prediction result. And performing actual sample test again to check the accuracy of the new round of numerical simulation result, and completing one iteration closed loop. The formulas M updated＝update(Mcurrent,Dtest, P) can be updated by the model, where M updated is the updated model, M current is the current model, D test is the test data, and P is the adjusted parameter.
Through continuous closed loop iterative optimization, the accuracy of the simulation model is gradually improved, the result is continuously close to the actual situation, and finally, the high consistency is achieved. This enables an efficient fusion of numerical simulations with actual testing. The iteration mechanism enables simulation not to be disposable any more, and can form a closed loop of verification and feedback with physical testing, so that design verification testing becomes an optimized and improved continuous evolution process.
Step five: and finishing the synergy and iteration of the multiple samples in the design-simulation-verification full link.
The invention can realize the whole-flow collaborative optimization from sample design to simulation to real object verification test. In the design stage, samples of multiple alternatives can be presented in parallel; and predicting the performance index of each sample by using the simulation model, and carrying out preliminary screening and comparison. And selecting a priority sample to manufacture a real object for a first round of actual test to obtain actual test data. And constructing results of each stage as virtual quanta of the sample, and effectively fusing data of each link of design, simulation and test to form a closed loop. The result of the actual test data can guide the gradual optimization iteration of the simulation model, and the optimized simulation model is used for the updated prediction evaluation of each sample scheme. And repeating the steps to finally determine the design sample with optimal comprehensive performance.
According to the invention, effective connection and cooperative work are realized in each flow stage, and the design can be optimized rapidly by means of simulation and test results; the simulation is no longer disposable but forms a closed loop iteration with the physical test. The multi-sample cooperation and iteration in the full link greatly improves the efficiency and quality of design verification.
According to the invention, by establishing a virtual test environment integrating the digital real data, constructing a virtual quantum of a sample to realize the correspondence between a numerical simulation result and actual test data, utilizing the virtual quantum to realize the consistency expression and verification of multiple samples, establishing a closed-loop iteration mechanism to continuously optimize a simulation model, and realizing five key steps of the coordination and iteration of multiple samples in a design-simulation-verification full-link, the numerical simulation and the actual test are effectively combined, the whole closed-loop flow from the predictive simulation of a design stage to the experimental verification of an intermediate test stage and the comprehensive evaluation of a final verification stage is opened, and compared with the limitations of the current numerical simulation and experimental test fracture, the invention realizes a more efficient, economic and accurate product design and verification scheme, reduces the test cost and improves the design quality.
Examples:
and confirming the product sample to be designed and the verification requirement, and performing verification.
(1) According to the size parameters of the sample, three-dimensional modeling software such as CAD and the like is selected to construct an accurate three-dimensional geometric model of the sample, so that a foundation is laid for subsequent simulation calculation.
(2) And comprehensively setting key parameters such as material properties, boundary constraint modes, loading force distribution modes and the like of the simulation model by comprehensively referencing a material property parameter table, actual working conditions and the like.
(3) And determining a numerical solution algorithm such as a finite element method, setting parameters such as convergence error limit, grid density and the like of control solution, and ensuring the accuracy of calculation.
(4) And (3) solving the model by using a determined simulation algorithm by means of solver software, and finally outputting numerical results of each point of the sample, such as stress distribution, node displacement and the like.
(5) And extracting key characteristics such as a maximum stress value, a main deformation mode and the like of the sample from a large amount of calculation result data, and establishing a foundation for subsequent mapping and analysis.
Actual test part:
(1) Under the guidance of simulation results, a certain number of actual samples are selected and manufactured, and the parameters of the samples represent typical working conditions and key load conditions.
(2) Specific manufacturing schemes such as a cutting method and a welding process of a sample are formulated in detail, and meanwhile, the types and parameters of required manufacturing equipment and processing tools are determined.
(3) And (3) processing and manufacturing by using the determined tool and equipment strictly according to a manufacturing scheme, and carefully manufacturing the actual physical sample with controllable quality and accurate precision.
(4) And (3) constructing an actual physical test bed, so that the physical test bed can be loaded in different forms, wherein the platform comprises a mechanical loading device and a sample fixing mechanism.
(5) And installing and fixing the manufactured actual sample on a designated position of the test platform.
(6) The test is carried out by loading different levels of force or pressure multiple times, and the response of the sample under each typical loading is obtained.
(7) And acquiring displacement, strain and vibration mode results of the sample under loading by using a strain gauge, a displacement sensor and the like.
Virtual test environment section:
(1) And extracting geometric parameters, material mechanical properties and the like of the sample from simulation result data to construct a feature vector representing simulation features.
(2) And analyzing the original actual test data acquired by the actual test, extracting key stress-strain relation, deformation mode and the like, and forming a feature vector representing the features of the actual test data.
(3) And selecting mathematical models such as principal component analysis and the like, inputting two sets of characteristic data, and establishing a mapping conversion model between the characteristics.
(4) And eliminating dimension and magnitude differences of the two sets of characteristic data through pretreatment means such as standardization and the like.
(5) And inputting the preprocessed feature vectors into a mapping model, running an algorithm code, and executing conversion between the two groups of features.
(6) And finally outputting the conversion characteristic data of the virtual quanta of the expression sample, and realizing the correspondence between the simulation result and the test data.
(7) A virtual test environment simulating an actual test process is built in a software system, and an integrated platform is provided.
(8) And introducing the virtual quanta obtained by converting the parameters of different samples into the virtual test environment.
(9) And comparing the virtual quantum characteristic values of different samples, and verifying the consistency of the same sample in a digital-real domain.
(10) A plurality of samples having different parameter ranges were selected as the verification samples.
(11) And loading virtual quanta corresponding to the selected verification sample in the virtual test environment.
(12) And comparing and verifying the expression of the virtual quanta of the sample, and determining the accuracy of the conversion result.
(13) The first round selects the actual sample for physical testing and imports the test data.
(14) And comparing the result of the actual test data with the current numerical simulation result to find the cause of error generation.
(15) And aiming at the reason of error generation determined in the last step, adjusting a simulation model and carrying out simulation prediction again.
(16) And (5) circularly iterating the steps (13) - (15) until the numerical simulation result and the result of the actual test data reach the required consistent precision.
(17) And through multiple closed loop iteration comparison, the numerical simulation result is highly matched with the result of actual test data.
(18) And finally, the validity and reliability verification of the method is completed, and a foundation is laid for practical application.
In summary, the method for performing the multi-sample cross-stage collaborative operation in the digital-real fusion test designed by the invention can organically combine numerical simulation with actual test, realize data correspondence and fusion expression of the numerical simulation and the actual test by constructing a virtual test environment and providing a virtual quantum concept, and establish a closed-loop iteration mechanism to continuously optimize a simulation model, thereby realizing the collaborative and efficient iteration of multiple samples in the whole flow of design, simulation and verification. The method is sufficient to overcome the problems of disjoint numerical simulation and experimental test, single sample and the like in the prior art, can remarkably improve the economy and accuracy of the test, and provides important support for product design and verification.
What is not described in detail in the present specification belongs to the prior art known to those skilled in the art.
The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.
Claims (5)
1. A multi-sample cross-stage collaborative operation method for a digital-real fusion test is characterized by comprising the following steps:
Step 1, establishing a virtual test environment integrating data of a numerical simulation result and actual test data, wherein the virtual test environment is constructed based on a software platform and is used for storing and processing the data to realize fusion of the data of the numerical simulation result and the actual test data;
Step 2, predicting the performance of a plurality of different samples by adopting a finite element numerical simulation method, and performing physical test to obtain actual test data; then, extracting key features from the data of the numerical simulation result and the actual test data by using a feature extraction method, wherein the key features comprise the elastic modulus, the yield strength and the fatigue life of the material; the properties include mechanical properties, thermal properties; the actual test data comprises stress and deformation; the feature extraction method comprises a machine learning algorithm;
Step 3, establishing a corresponding relation between key features of data of a numerical simulation result and key features of actual test data in the same coordinate system by utilizing a principal component analysis mapping algorithm, and finally forming virtual features representing samples, namely virtual quanta;
step 4, importing virtual quanta of different samples in a virtual test environment, and comparing the data of the numerical simulation result of each sample with key features of actual test data to realize consistency verification of the same sample in different test stages;
And 5, establishing a closed loop iteration mechanism, evaluating a result after each round of actual test and updating a numerical simulation model, and then predicting based on the updated numerical simulation model, so that the loop is circulated until the precision requirement is met, mutual verification and optimization of simulation and test are realized, the whole process of the physical test from the numerical simulation in the design stage to the verification stage is cooperatively operated, and the comprehensive verification of multiple samples is carried out.
2. The method for performing a multi-sample cross-phase collaborative operation according to claim 1, wherein in step 3, forming virtual quanta of samples comprises:
Step 3.1, respectively extracting characteristics aiming at data of a numerical simulation result and actual test data, wherein the step comprises the following steps:
step 3.1.1, obtaining geometric dimensions and material properties in data of a numerical simulation result by adopting a parameter extraction method as characteristics;
step 3.1.2, a stress-strain curve and a deformation mode of actual test data are obtained by adopting a signal analysis method as characteristics;
and 3.2, selecting a principal component analysis mapping algorithm to realize the mapping relation between the characteristics of the data of the numerical simulation result and the characteristics of the actual test data, wherein the method comprises the following steps:
step 3.2.1, normalizing the characteristics of the data of the numerical simulation result and the characteristics of the actual test data, and eliminating the dimension influence;
Step 3.2.2, constructing a principal component analysis model, and finding out the principal component which can most represent the original data in the characteristics of the data of the numerical simulation result and the characteristics of the actual test data;
Step 3.2.3, determining a conversion relation between principal components, and establishing a linear or nonlinear mapping function;
Step 3.2.4, converting the characteristics of the data of the numerical simulation result and the characteristics of the actual test data into the same coordinate system by utilizing a linear or nonlinear mapping function;
step 3.3 generates virtual quanta representing virtual features of the sample, comprising:
Step 3.3.1, converting the characteristics of the data of the numerical simulation result extracted in the step 3.1 and the characteristics of the actual test data into the same expression system according to the mapping relation obtained in the step 3.2;
Step 3.3.2, describing virtual quanta of the sample by adopting a quantum state vector representation method; the virtual quantum of the sample integrates the characteristic information of the data of the numerical simulation result and the actual test data; the virtual quanta of a sample uniquely represent a virtual feature of the sample.
3. The method for multi-sample cross-phase collaborative operation according to claim 2, wherein the evaluating results and updating the numerical simulation model after each round of actual testing in step 5 comprises:
Step 5.1, collecting stress-strain curves and deformation modes of samples obtained by actual tests;
step 5.2, comparing the actual test data with the data of the current numerical simulation result to find a scene with larger error;
step 5.3, analyzing the difference between the actual test conditions and the simulation settings, and determining parameters of a numerical simulation model which causes errors;
Step 5.4, adjusting parameters of the numerical simulation model to conduct directional optimization, wherein the parameters comprise loading force, constraint and material properties;
And 5.5, rerun the simulation calculation based on the parameters of the optimized numerical simulation model to obtain the data of a new numerical simulation result for improving the precision.
4. The method for multi-sample cross-phase collaborative operation according to claim 3, wherein the predicting based on the updated numerical simulation model in step 5 comprises:
step 5.6, saving the parameter setting of the numerical simulation model and the numerical simulation model after the previous round of optimization;
step 5.7, running a new round of test of the numerical simulation model under the same setting as the actual test condition;
and 5.8, recording simulation prediction results of the optimized numerical simulation model under the same conditions.
5. The method for multi-sample cross-phase collaborative operation according to claim 4, wherein the cycling in step 5 until the accuracy requirement is met comprises:
Step 5.9, evaluating the data of the numerical simulation result of the new round, and judging whether the error meets the precision requirement;
Step 5.10, if the data of the numerical simulation result meets the precision requirement, ending iteration; if the accuracy requirement is not met, the numerical simulation model is adjusted again;
Step 5.11 repeating the steps 5.2 to 5.10, and continuously iterating and optimizing the numerical simulation model until the error between the data of the numerical simulation result and the actual test data is smaller than a specified value;
And 5.12, obtaining a high-precision numerical simulation model for sample design optimization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410198797.4A CN117993255A (en) | 2024-02-22 | 2024-02-22 | Multi-sample cross-stage collaborative operation method for digital-real fusion test |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410198797.4A CN117993255A (en) | 2024-02-22 | 2024-02-22 | Multi-sample cross-stage collaborative operation method for digital-real fusion test |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117993255A true CN117993255A (en) | 2024-05-07 |
Family
ID=90889293
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410198797.4A Pending CN117993255A (en) | 2024-02-22 | 2024-02-22 | Multi-sample cross-stage collaborative operation method for digital-real fusion test |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117993255A (en) |
-
2024
- 2024-02-22 CN CN202410198797.4A patent/CN117993255A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106682350B (en) | Three-dimensional model-based multi-attribute decision quality detection method | |
| CN107563067A (en) | Analysis of structural reliability method based on Adaptive proxy model | |
| CN111723440B (en) | Thin-wall part machining precision prediction hybrid modeling method | |
| CN111125946B (en) | Method for optimizing structure of boarding body based on MDO technology | |
| CN116305522A (en) | Digital twin-based aircraft structure reliability simulation test model calibration method, device and medium | |
| Chen | FE model validation for structural dynamics | |
| CN112231786B (en) | General evaluation method and system for evaluating effectiveness of simulation model | |
| CN110716520B (en) | Flight control servo actuator reliability evaluation modeling method based on multi-source information fusion | |
| CN112287484B (en) | Complex engineering system reliability design method based on vector agent model | |
| CN117171896A (en) | Method and system for optimizing machining of input shaft of automobile steering gear | |
| CN117217020A (en) | Industrial model construction method and system based on digital twin | |
| CN117993255A (en) | Multi-sample cross-stage collaborative operation method for digital-real fusion test | |
| CN116090063A (en) | Finite element modeling method of power transmission tower considering node slipping effect | |
| CN115577856A (en) | Method and system for predicting construction cost and controlling balance of power transformation project | |
| CN114329805A (en) | Connecting piece multidisciplinary collaborative design optimization method based on self-adaptive agent model | |
| CN115455754A (en) | Mine hydraulic support design method based on digital twinning | |
| CN113239495A (en) | Complex structure reliability design method based on vector hybrid agent model | |
| Xie et al. | Lightweight design of hinge beam based on Kriging agent model | |
| CN112906235A (en) | Method for three-dimensional factory construction and engineering application | |
| Iriondo Pascual et al. | Optimizing ergonomics and productivity by connecting digital human modeling and production flow simulation software | |
| CN111159797A (en) | BIM-based general modeling method for mechanical model | |
| CN111506998B (en) | Method for constructing parameter drift fault feature sample library in manufacturing process of electromechanical product | |
| CN117855613A (en) | Lithium battery production and assembly method | |
| Hu et al. | The use of function-driven shape definition method and GRA–EMM-based Taguchi method to multi-criterion blade preform optimization | |
| Babu et al. | Minimum weight optimization of a gear train by using GA |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |