A numerical-real fusion test method based on element fusion

Abstract

The application discloses a kind of based on element fusion's digital-real fusion test method, method includes: executing digital test and real installation test;Process fusion module receives data from digital test and real installation test, data is output to result fusion module in the module by process fusion mechanism processing,;Result fusion module receives the data output from process fusion module after, by result fusion module iteration processing, transmission is given in result fusion module next iteration;Comprehensive test result module receives real installation test and digital test data, the result of result fusion module, after calculation processing, generates comprehensive test result.According to the technical scheme of the application, for the situation that digital test model is low in reliability, real installation test model is high in cost and has less available information, a kind of digital-real fusion test comprehensive result calculation method is designed, and then high-reliable, wide-coverage test result for equipment or system is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer simulation, in particular to a numerical-real fusion test method based on element fusion. BACKGROUND

[0002] Traditional equipment or system testing often faces problems such as incomplete test information, non-repetitive measurement of part of the information, and high cost of testing part of the information. The digital test model formed with the development of information technology also faces problems such as low credibility and unreasonable digital test methods. At present, there is still a lack of effective methods, processes and ideas for integrating real equipment testing and digital testing. Therefore, the following problems exist:

[0003] High cost of real equipment testing: Real equipment testing usually requires a large amount of manpower and material resources, and the debugging, maintenance and testing process of equipment involves high costs and huge time costs, making the testing inefficient and insufficient.

[0004] It is difficult to obtain real equipment testing information: In a complex testing environment, it is difficult to obtain comprehensive and accurate real equipment testing data due to equipment limitations and external conditions.

[0005] Unreasonable digital test methods: Existing digital test methods often lack pertinence and rationality, and cannot fully simulate the complexity in real environments, resulting in unreliable test results.

[0006] Low credibility of digital testing: The construction and testing process of digital models may have uncertainties, lack effective verification and feedback mechanisms, and thus affect their credibility.

[0007] There is a lack of effective ideas and methods for integrating real equipment testing and digital testing: There is still a lack of systematic methods and processes to effectively combine real equipment testing and digital testing to achieve an innovative testing mode of information complementation and data integration.

[0008] In summary, there are still significant deficiencies in real equipment testing and digital testing, and the gap between the two is still significant, lacking effective integration methods, processes and ideas. This not only limits the progress of testing technology, but also to some extent restricts the improvement and optimization of equipment or system performance. SUMMARY

[0009] The technical problem solved by the present application is: in view of the problems of missing real installation test environment and small amount of data in current equipment or system test, a process model is designed, which integrates real installation test, digital test, process integration and result integration; in view of the problem that real installation test data, digital test data and process integration test data cannot be effectively integrated, a data credibility compensation mechanism is established to solve the problems of small amount of real installation test data and low credibility of digital test; in view of the problems of low credibility of digital test model, high cost of real installation test model and less available information, a numerical-real integration test method based on element integration is designed.

[0010] The technical problem solved by the present application is solved by the following technical scheme: a numerical-real integration test method based on element integration, comprising:

[0011] Step one, the digital model of equipment and system is tested digitally, and after the digital test is completed, the process step data and key parameter data are sent to the next iteration and process integration module of digital test at the same time and in parallel;

[0012] Step two, the actual physical equipment of equipment and system is tested by real installation, and after the real installation test is completed, the process step data and key parameter data are sent to the next iteration and process integration module of real installation test at the same time and in parallel;

[0013] Step three, after the process step data and key parameter data from digital test and real installation test are received by the process integration module, the process integration mechanism is processed in the module, and the output data is transmitted to the result integration module;

[0014] Step four, after the data output from the process integration module is received by the result integration module, the result integration module is iterated and processed, and the result integration module is transmitted to the next iteration;

[0015] Step five, the process step data and key parameter data of real installation test after multiple iterations, the process step data and key parameter data of digital test after multiple iterations, and the results of result integration module after multiple iterations are received by the comprehensive test result module, and the comprehensive test result is generated after calculation and processing.

[0016] The beneficial effects of the present application compared with the prior art are:

[0017] (1) When the test elements are missing, the data is small, and other conditions, the missing of some data that must be obtained but cannot be directly obtained will lead to the inaccuracy of the overall test and result calculation. The traditional method cannot directly obtain the data through the sensor, and even if it can, the installation of the sensor will affect the mechanism and performance of the test object itself, so the traditional method cannot realize reliable testing under the condition of missing test object elements, small data, etc. The method of the present application solves the above problems by comprehensively utilizing real installation testing, digital testing, process fusion, and result fusion, and designing a multi-element fusion digital-real fusion test model.

[0018] (2) For the problem that real installation test data, digital test data and process fusion test data are not easy to effectively fuse, the traditional method usually analyzes and evaluates separately, and it is difficult to achieve reliable and effective fusion. The method of the present application solves the problems of small amount of real installation test data and low reliability of digital test by establishing a data credibility compensation mechanism based on real and digital data.

[0019] (3) For the condition that the digital test model has low reliability, the real installation test model has high cost and less available information, the traditional method usually uses complex and time-consuming theoretical derivation, and it is difficult to efficiently form reliable results. The present application designs a digital-real fusion test comprehensive result calculation method, which realizes high reliable and wide coverage test results of equipment or system through multi-element iterative fusion. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The overall process diagram of the digital-real fusion test method based on element fusion is described.

[0021] Figure 2 The data flow diagram of real installation test is described.

[0022] Figure 3 The data flow diagram of digital test is described.

[0023] Figure 4 The data flow diagram of the process fusion module is described.

[0024] Figure 5 The data flow diagram of the result fusion module is described. DETAILED DESCRIPTION

[0025] The present application will be described in further detail below with reference to the accompanying drawings.

[0026] The application designs a numerical-real fusion test model based on element fusion, including a process model of fusing multiple elements of real equipment test, digital test, process fusion and result fusion to solve the problems of missing real equipment test environment and insufficient data in current equipment or system test; a data credibility compensation mechanism of real-to-numerical and numerical-to-real data to solve the problem of ineffective fusion of real test data, digital test data and process fusion test data; and a numerical-real fusion test comprehensive result calculation method to solve the problems of low credibility of digital test model, high cost of real equipment test model and less available information.

[0027] Figure 1 The overall process diagram of the numerical-real fusion test method based on element fusion shows a process model of fusing multiple elements of real equipment test, digital test, process fusion and result fusion. Figure 1 As shown in the figure, the process model includes four element threads: digital test (sd), real equipment test (sp), process fusion (sf) and result fusion (sr). Finally, a comprehensive test result (PDEO) of the equipment or system object is formed.

[0028] The numerical-real fusion test method based on element fusion includes the following steps:

[0029] Step one, digital test is performed on the digital model of the equipment and system. After the digital test is completed, process step data and key parameter data are sent to the next iteration of the digital test and the process fusion module simultaneously and in parallel; the process step data includes test target, condition, instrument configuration, instrument state and other information data, and the key parameter data includes performance data such as power, response time, load capacity, vibration, noise, life and mean time between failures of the equipment and system.

[0030] Step two, real equipment test is performed on the actual physical equipment of the equipment and system. After the real equipment test is completed, process step data and key parameter data are sent to the next iteration of the real equipment test and the process fusion module simultaneously and in parallel;

[0031] Step three, the process fusion module receives the process step data and key parameter data from the digital test and the real equipment test, processes them through the process fusion mechanism in the module, and transmits the output data to the result fusion module;

[0032] Step four, the result fusion module receives the data output from the process fusion module, iteratively processes them in the result fusion module, and transmits them to the next iteration of the result fusion module;

[0033] In step five, the comprehensive test result module receives the process step data and key parameter data from the actual installation test after multiple iterations, the process step data and key parameter data from the digital test after multiple iterations, and the results of the result fusion module after multiple iterations, and generates a comprehensive test result after calculation and processing.

[0034] The following describes in detail the specific implementation of steps 1 to 5. This specific implementation enables the establishment of a data credibility compensation mechanism that uses reality to generate numbers and numbers to test reality, addressing the problem of ineffective integration of actual test data, digital test data, and process fusion test data.

[0035] Step 1 includes: designing a multi-port parallel transmission mechanism for the test data of each step or beat in the digital test process, and outputting them to the next digital test step or beat and process fusion module respectively, such as Figure 3 shown.

[0036] Step 2 includes: designing a multi-port parallel transmission mechanism for the test data of each step or beat in the actual installation test process, and outputting them to the next actual installation test step or beat and process fusion module, such as Figure 2 As shown;

[0037] The process fusion model is built based on the specific implementation characteristics. Its architecture is as follows Figure 4 The third step of its execution includes: the process fusion module receives Figure 2 and Figure 3 The process step data, key parameter data of the actual installation test and the process step data, key parameter data of the digital test are substituted into the data for difference comparison, feature extraction, feature dimension reduction, feature correlation analysis and other data processing operations. This process also includes: for the missing parts of the actual installation data, the process fusion module is used to generate supplementary data, which is recorded as There are two sources of compensation data. One is to rely on other actual test data input and process fusion module calculation and generation, which is recorded as The other is to rely on other digital test data input and process fusion module calculation, which is recorded as Other actual installation test data and other digital test data refer to the actual installation test data and digital test data collected during other iterations. The supplementary data generated from the two sources are analyzed for differences using the KS algorithm. If the P value of the KS algorithm is less than 0.05, they are weighted and fused separately. If it is not less than 0.05, it means that there is no significant difference between the supplementary data from the two sources, and both are equally weighted and fused, that is:

[0038] ;

[0039] The designed respectively empowered fusion mechanism is as follows: if the P value of KS algorithm difference test analysis is less than 0.05, there are:

[0040]

[0041] Among them, is a statistical quantity for measuring whether the sample data conforms to the consistency hypothesis, a larger value means that there is no significant difference between the two samples, a smaller value means that there is a difference between the two samples.

[0042] In the test process, multi-threading technology is used to realize parallel processing of each test module, so as to significantly improve the overall efficiency of the test. The scheduling control system is responsible for reasonably allocating computing resources, ensuring the coordination and synchronization between threads, avoiding data conflicts and competition conditions. For the task arrangement of each thread, a priority strategy is designed for adjustment. In the case of conflict, the priority is: the real installation test process is higher than the process fusion process, and the process fusion process is higher than the digital test process.

[0043] Step four includes: the result fusion module fuses the results of the process fusion module and the results of the previous result fusion module, and the fusion process is as shown in Figure 5 .

[0044] In the result fusion module, the results of the previous step result fusion module and the to-be-fused object are subjected to respectively empowered operation, and the weight of the to-be-fused object is greater than the result of the previous step result fusion . In order to achieve better update and emphasis effect, the result fusion formula is designed as:

[0045]

[0046] Step five includes: for the case that the digital test model has low reliability, a comprehensive result calculation method based on digital-real fusion test is designed. In the digital-real fusion test process, the comprehensive test result will be updated with the increasing number of iterations. In the updating process, the new comprehensive test result is integrated into the last iteration result. The specific implementation is as follows:

[0047] a. The expression of the comprehensive test result is:

[0048]

[0049] Among them, is the result of the result fusion module of the nth iteration, is the real installation test result of the nth iteration, ​​​The digital test result for the nth iteration, f, represents the PDEO is determined by 、 、 Three independent variables determine the proxy of the fusion function.

[0050] The iteration process formula for the comprehensive test result is:

[0051] ;

[0052] ;

[0053] Wherein, 、 、 The weights of the result fusion module, the implementation test result, and the digital test result reflect the importance in different iterations, Indicates the iteration number.

[0054] The data set weight changes with the iteration number, and the weight formula is set as follows:

[0055] ;

[0056] ;

[0057] ;

[0058] Wherein, k is an adjustment coefficient, and by adjusting the value of k, the change trend of each weight can be controlled, so as to adjust the adaptation of each weight when facing different iteration orders of magnitude. This weight design ensures that the weight proportion of different data sets can be reasonably reflected in the contribution to the final result in the multiple iteration process. In the early stage of iteration, the result fusion result weight is small, and the growth speed is slow, the implementation test and digital test result weight is large, and the growth speed is fast; in the later stage of iteration, the result fusion result weight is large, and the growth speed is fast, the implementation test and digital test result weight is small, and the growth speed is slow.

[0059] In the weight expression formula shown, the size of the k parameter can be adjusted according to the specific iteration predetermined condition. For example, when the iteration number is predetermined to be about 200 times, set k = 100, and the weight values are as follows:

[0060] ;

[0061] ;

[0062] ;

[0063] In summary, a numerical-real fusion test method based on element fusion is provided, which comprises the following steps:

[0064] In the process of testing the equipment or system, first, the digital model of the equipment or system object and the process fusion model of the process fusion module are established in advance, so that digital testing can be carried out by using digital means. During the testing process, the equipment or system object is respectively and iteratively tested (sd), the actual equipment is tested (sp), the process fusion is tested (sf), and the result fusion is tested (sr), and finally the comprehensive test result (PDEO) is formed.

[0065] The digital testing is testing the digital model of the equipment and system, and after the testing is completed, the data is respectively and simultaneously sent to the next iteration of the digital testing and the process fusion module; the actual equipment testing is testing the actual physical equipment of the equipment and system, and after the testing is completed, the data is respectively and simultaneously sent to the next iteration of the actual equipment testing and the process fusion module; the process fusion module comprises a process fusion model and a process fusion mechanism, and after receiving the process step data and key parameter data information from the digital testing and the actual equipment testing, the module processes and transmits the output data to the result fusion module; the result fusion module receives the information data output from the process fusion module, iteratively processes the information data in the module, and transmits the information data to the next iteration of the result fusion module. The comprehensive test result module receives the actual equipment testing, digital testing and result fusion information data after multiple iterations, and generates a comprehensive test result after calculation and processing.

[0066] The above description covers the specific implementation details of a multi-element fusion process model, a data credibility compensation mechanism and a numerical-real fusion test comprehensive result calculation method, fully considers various technical details and their mutual relationships, and ensures efficient and accurate data processing and analysis in actual application.

[0067] The contents not described in detail in the specification of the present application belong to the prior art known to those skilled in the art.

[0068] The above description is only the preferred embodiment of the present application, and it should be pointed out that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. A digital-real fusion testing method based on element fusion, characterized in that: include: Step 1: Perform digital testing on the digital models of the equipment and system. After the digital testing is completed, the process step data and key parameter data are sent to the next iteration of the digital testing and process fusion module in parallel. Step 2: Perform installation testing on the actual physical equipment of the equipment and system. After the installation test is completed, the process step data and key parameter data are sent to the next iteration of the installation test and the process fusion module in parallel. Step 3: After receiving the process step data and key parameter data from the digital test and the actual installation test, the process fusion module processes them through the process fusion mechanism in the module and transmits the output data to the result fusion module; Step 3 includes: the process fusion module receives the process step data, key parameter data of the actual installation test and the process step data, key parameter data of the digital test, and substitutes the process step data and key parameter data of the two into the process fusion module to perform difference comparison, feature extraction, feature dimension reduction, and feature correlation analysis. The process fusion module generates supplementary data for the missing parts in the actual installation data, which is recorded as There are two sources of compensation data. One is to rely on other actual test data input and process fusion module calculation and generation, which is recorded as The other is to rely on other digital test data input and process fusion module calculation, which is recorded as , the KS algorithm is used to perform difference analysis on the supplementary data generated by the two sources. If the P value of the KS algorithm is less than 0.05, the weighted fusion is performed separately: ； in, It is a statistic that measures whether the sample data conforms to the hypothesized distribution. The smaller the p-value, the greater the difference between the two samples. If it is not less than 0.05, the two samples are equally weighted and merged, that is: ； Step 4: After the result fusion module receives the data output from the process fusion module, it is iteratively processed by the result fusion module and transmitted to the next iteration of the result fusion module; In step five, the comprehensive test result module receives the process step data and key parameter data from the actual installation test after multiple iterations, the process step data and key parameter data from the digital test after multiple iterations, and the results of the result fusion module after multiple iterations, and generates a comprehensive test result after calculation and processing.

2. The digital-real fusion testing method based on element fusion according to claim 1 is characterized in that: Step 1 includes: designing a multi-port parallel transmission mechanism for the test data of each step or beat in the digital test process, and outputting them to the next digital test step or beat and process fusion module respectively; Step 2 includes: designing a multi-port parallel transmission mechanism for the test data of each step or beat in the actual installation test process, and outputting them to the next actual installation test step or beat and process fusion module respectively; Step 4 includes: the result fusion module fuses the result of the process fusion module with the result of the previous step result fusion module: and the object to be fused Implement separate empowerment operations, waiting for fusion objects The weight is greater than the result of the previous step fusion , the design result fusion formula is: ； Step 5 includes: updating the comprehensive test results as the number of iterations increases, integrating the new comprehensive test results into the previous iteration results during the updating process, including: a. The expression of the test comprehensive result is: ； in, is the result of the result fusion module of the nth iteration, is the actual test result of the nth iteration, is the numerical test result of the nth iteration; The iterative process calculation formula for the comprehensive test results is: ； ； in, 、 、 The weights of the result fusion module results, the actual installation test results, and the digital test results; The weight changes with the number of iterations, and the weight setting formula is as follows: ； ； ； Among them, k is the adjustment coefficient, and the change trend of each weight is controlled by adjusting the value of k.

3. The digital-real fusion testing method based on element fusion according to claim 2 is characterized in that: The adjustment coefficient k is set as follows: in the early stage of iteration, the weight of the result fusion module is small and increases slowly with the number of iterations, while the weight of the actual test results and the digital test results is large and increases quickly with the number of iterations; In the later stage of iteration, the weight of the result fusion module is larger and increases faster with the number of iterations, while the weight of the actual test results and the digital test results is smaller and increases slower with the number of iterations.

4. The digital-real fusion testing method based on element fusion according to claim 3 is characterized in that: The adjustment coefficient k is set according to a predetermined number of iterations.

Images