A test result calculation and evaluation method based on virtual-real data fusion
By using a method for calculating and evaluating experimental results through virtual-physical data fusion, the challenges of fusing virtual and physical space data in existing technologies have been addressed. This method achieves high-precision, low-cost evaluation of experimental results and ensures the consistency of data-physical fusion experimental results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIHANG UNIV
- Filing Date
- 2026-04-02
- Publication Date
- 2026-06-16
AI Technical Summary
Existing testing methods struggle to balance comprehensive coverage, high accuracy, high reliability, and low cost when facing complex testing requirements, especially in terms of consistency in computation and evaluation of heterogeneous data from both virtual and physical spaces.
By acquiring physical and virtual test data, similarity analysis is performed and the data is divided into strongly correlated and weakly correlated data. A fusion data set is generated by adopting similar association and complementary association fusion strategies. The results of the fusion test are calculated by combining evaluation parameters and then compared with the results of the pure physical test.
It achieves full-scenario and full-condition test data coverage, improves the accuracy and reliability of test results, and reduces the reliance on purely physical tests, thus saving test costs.
Smart Images
Figure CN122221172A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the fields of electronic engineering and computer science, and specifically relates to a method for calculating and evaluating experimental results based on the fusion of virtual and real data. Background Technology
[0002] In the wave of industrial digitalization, the shift from purely physical testing to a data-real fusion testing model has become an inevitable trend. By fully utilizing digital and intelligent technologies such as sensor networks, modeling and simulation, digital twins, and artificial intelligence, data-real fusion testing can organically integrate virtual and physical spatial elements to achieve performance evaluation, better meeting the needs for comprehensiveness, accuracy, reliability, and economy, thereby comprehensively improving testing capabilities.
[0003] Existing testing methods often struggle to balance comprehensive coverage, high accuracy, high reliability, and low cost when facing increasingly complex testing requirements. Specifically, how to effectively integrate heterogeneous data from physical and digital (virtual) spaces, how to accurately calculate test results based on this integrated data, and how to scientifically evaluate the consistency between the results of fusion tests and purely physical tests are key technical issues that urgently need to be addressed in current fusion tests. Summary of the Invention
[0004] To address the aforementioned technical problems, this invention discloses a method for calculating and evaluating experimental results based on virtual-real data fusion. This method is applicable to experiments employing the data-real fusion approach. By fusing virtual and real experimental data, the method can calculate the data-real fusion experimental results and evaluate the consistency between the data-real fusion experimental results and the purely physical experimental results. This approach achieves accurate calculation of experimental results to a certain extent and saves experimental costs.
[0005] To achieve the above objectives, the present invention adopts the following technical solution:
[0006] A method for calculating and evaluating experimental results based on virtual-real data fusion includes the following steps:
[0007] Step 1: Obtain physical test data and virtual test data generated separately for the same experimental task through physical and virtual experiments;
[0008] Step 2: Perform similarity analysis on the physical test data and the virtual test data, divide the data into strongly correlated data and weakly correlated data according to a preset threshold, and perform similar association fusion and complementary association fusion respectively to generate a fused data set;
[0009] Step 3: Based on the fused data set, calculate the fused data test results according to the evaluation parameters and test result calculation method defined in the test task, and compare the results with the pure physical test results obtained based on the pure physical test data to complete the evaluation of the test results.
[0010] Beneficial effects:
[0011] 1. This invention obtains test data covering all scenarios and all working conditions by integrating test data from physical space and digital space, thereby improving the comprehensiveness of the data.
[0012] 2. This invention achieves accurate calculation of experimental results to a certain extent by employing fusion strategies of similar correlation and complementary correlation for strong and weak correlation data respectively, and combining them with the calculation of evaluation parameters.
[0013] 3. This invention can effectively utilize virtual data to supplement physical data, reducing reliance on purely physical experiments and thus saving experimental costs.
[0014] 4. This invention establishes a comparative evaluation mechanism between the results of data-real fusion and the results of pure physics, ensuring the credibility and consistency of the data-real fusion test results, and providing scientific methodological support for the promotion of data-real fusion tests. Attached Figure Description
[0015] Figure 1 This is a flowchart of a data-real fusion test scenario and model reconstruction method according to the present invention. Detailed Implementation
[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other. To achieve the above objectives, this invention adopts the following technical solution.
[0017] like Figure 1 As shown, the present invention provides a method for calculating and evaluating experimental results based on virtual-real data fusion, including virtual-real data fusion, calculating the fused experimental results, and evaluating the consistency of the experimental results. This method can achieve accurate calculation of experimental results to a certain extent and save experimental costs. Specifically, it includes the following steps:
[0018] S1: Virtual-real data fusion, including:
[0019] S1.1 For a specific experimental task, determine the calculation method for the experimental results. , The evaluation parameters required to calculate the experimental results are k, which is the number of evaluation parameters, and g is the experimental result function.
[0020] S1.2 Perform the data-real fusion experiment to obtain the experimental dataset. , Let N be the number of data points, and let N be the number of data points. Calculate the sum of any two data points. and data points similarity The calculation method is as follows This yields the similarity set S. This represents the L2 norm.
[0021] S1.3 Based on the data similarity calculation results, the experimental data are divided into two types: strongly correlated data and weakly correlated data; a threshold is set. ,like Then the data points and For strongly correlated data, if Then data points and data points This is a weakly correlated data.
[0022] S1.4 performs similarity association fusion on the strongly correlated data identified in S1.3, and the calculation method is as follows: , The fused data yields strongly correlated fused data. ;
[0023] S1.5 performs complementary correlation and fusion on the weakly correlated data identified in S1.4, and the calculation method is as follows: , For the merged data, and To determine the fusion weights, weakly correlated fused data is obtained. .
[0024] S1.6 combines the strongly correlated fused data from S1.4 and the weakly correlated fused data from S1.5 to obtain the fused data set. .
[0025] S2: Calculation of test results, including:
[0026] S2.1 refers to the fused data set obtained from S1 Calculate the evaluation parameters of the test results determined in S1. , Functions for calculating experimental parameters;
[0027] S2.2 Based on the evaluation parameters obtained in S2.1, the experimental results of the data-real fusion experiment are calculated according to the experimental result calculation method specified in S1. ;
[0028] S3: Evaluation of test results, including:
[0029] S3.1 For the experimental task in S1, the experiment is performed using a purely physical experimental method to obtain purely physical experimental data. ;
[0030] S3.2 Calculate the purely physical test data according to the test result calculation method specified in S1. The results of the purely physical experiments were obtained. ;
[0031] The results of the data-real fusion experiment obtained by comparing S3.3 and S2 and results of pure physical experiments To assess the consistency of the data-real fusion test results If the results are consistent, then the results of the data-real fusion experiment are valid.
[0032] Example:
[0033] Example: Application of virtual and real data fusion in fatigue life testing of aero-engine blades
[0034] The high-pressure compressor blades of a certain type of aero-engine require fatigue life assessment. Traditional physical tests require repeated loading under high temperature and high speed conditions, which is costly, time-consuming, and difficult to cover all operating conditions. Therefore, a fusion of physical and digital testing methods is adopted: key data are obtained through a limited number of physical tests, and virtual test data is generated by combining them with a high-fidelity digital twin model to achieve accurate assessment of the blade fatigue life.
[0035] S1: Virtual and Real Data Fusion
[0036] S1.1 Determine the calculation method for test results:
[0037] The test results are the blade fatigue life R, and the evaluation parameters include: maximum stress amplitude, average stress, and number of cycles (k=3 parameters). The test result function is a life prediction model based on Miner's linear cumulative damage theory.
[0038] S1.2 Obtain the test dataset:
[0039] Physical experiment: Bench tests were conducted under 5 typical working conditions to obtain 5 sets of measured strain and temperature data.
[0040] Virtual experiment: Based on the calibrated digital twin model, simulations were performed under 20 identical and extended operating conditions to obtain 20 sets of simulation data, which were then merged to obtain the total dataset D, N=25.
[0041] S1.3 Calculate similarity and classify:
[0042] For any two data points, a weighted Euclidean distance is used to calculate their similarity. Features can be stress, temperature, and rotational speed, with weights of [0.5, 0.3, 0.2]. A threshold is set. =0.15.
[0043] S1.4 Fusion of strongly correlated data, such as the first set of measured strain and temperature data and the first set of working conditions;
[0044] S1.5 For weakly correlated data (such as high temperature and low speed vs. normal temperature and high speed), a complementary fusion strategy is adopted to retain their respective extreme value characteristics and extract common information through principal component analysis (PCA).
[0045] S1.6 Obtain the fused dataset.
[0046] S2: Calculation of test results.
[0047] S3: Conduct a consistency assessment.
[0048] This embodiment only conducted 5 physical tests and achieved high-precision life prediction by integrating 20 sets of virtual data, verifying the practicality and economy of the invention in the reliability assessment of aviation equipment.
[0049] Contents not described in detail in this specification are prior art known to those skilled in the art. The above descriptions are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A method for calculating and evaluating experimental results based on virtual-real data fusion, characterized in that, This includes the following steps: Step 1: Obtain physical test data and virtual test data generated separately for the same experimental task through physical and virtual experiments; Step 2: Perform similarity analysis on the physical test data and the virtual test data, divide the data into strongly correlated data and weakly correlated data according to a preset threshold, and perform similar association fusion and complementary association fusion respectively to generate a fused data set; Step 3: Based on the fused data set, calculate the fused data test results according to the evaluation parameters and test result calculation method defined in the test task, and compare the results with the pure physical test results obtained based on the pure physical test data to complete the evaluation of the test results.
2. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The physical test data and virtual test data are output data obtained by the physical test system and the virtual simulation system respectively for the same test task under the same input conditions.
3. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The similarity analysis includes calculating the similarity value between any two data points and comparing the similarity value with a preset threshold to determine the correlation category.
4. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The strongly correlated data are data pairs with a similarity value greater than or equal to a preset threshold, and the weakly correlated data are data pairs with a similarity value less than a preset threshold.
5. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The similarity association fusion includes aggregating strongly correlated data to generate fused data that retains high consistency features.
6. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The complementary correlation fusion involves retaining the distinct features of weakly correlated data while extracting common information to generate fused data.
7. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The fused data set is composed of strongly correlated fused data obtained through similarity association fusion and weakly correlated fused data obtained through complementary association fusion.
8. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The results of the pure physical experiments are obtained by processing the physical experimental data according to the same evaluation parameters and experimental result calculation methods as the data-real fusion experiments.
9. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, The consistency comparison includes determining whether the results of the data-real fusion test and the results of the pure physical test are consistent within the allowable error range.
10. The method for calculating and evaluating experimental results based on virtual-real data fusion according to claim 1, characterized in that, When the consistency comparison result meets the consistency judgment condition, the data-real fusion test result is confirmed to be valid.