Complex environment adaptability assessment method based on digital-real collaborative test

Abstract

The invention discloses a complex environment adaptability assessment method based on a digital-real collaborative test, and belongs to the technical field of complex environment adaptability assessment of equipment test identification. Comprising the following steps: collecting complex environment data and physical data of an evaluated object to construct and generate a digital twinborn model of the evaluated object in a complex environment, and performing digital model verification on the digital twinborn model to obtain a target model; establishing interactive connection between the target model and an actual physical system, and monitoring interactive data in real time to realize data-real collaboration; if the difference between the interactive data of the two is greater than a preset difference threshold value, correcting the target model through the generated data of the actual physical system until a preset condition is completed, and generating a corrected model; and simulating simulation working conditions of the evaluated object in different complex environments through the correction model, evaluating the complex environment adaptability of the evaluated object based on the simulation working conditions, and generating an evaluation report.

Description

Technical Field

[0001] This invention belongs to the field of complex environment adaptability assessment technology for equipment testing and evaluation, and specifically relates to a complex environment adaptability assessment method based on numerical and real collaborative testing. Background Technology

[0002] Complex equipment characterized by electronic information technology is developing rapidly, leading to unprecedented changes in the traditional battlefield environment. In actual combat, the combat effectiveness of equipment is often reduced due to the complex battlefield environment. The complex electromagnetic environment, as a crucial component of the battlefield environment, is caused by various natural environments and civilian facilities, as well as malicious interference from non-cooperative parties. Therefore, to ensure the combat effectiveness of equipment, simulation verification of complex electromagnetic environments is particularly important for evaluating system performance indicators, technical effects, and combat effectiveness.

[0003] Currently, the testing and evaluation methods for equipment in battlefield electromagnetic environments mainly rely on field tests. This involves placing equipment in a realistic battlefield environment and using collected real data to assess the impact of the electromagnetic environment on the equipment. While this method is closer to actual combat and yields highly reliable results, it is also costly, time-consuming, non-repeatable, and its complexity and parameters are difficult to change. The test results are highly unpredictable, and the number of data samples obtained is very limited. This method is also susceptible to weather and geographical factors, sometimes requiring waiting for specific weather conditions to meet the testing requirements. Furthermore, this method struggles to guarantee diverse battlefield environments, cannot simulate the enemy's operational spectrum and characteristics, has low flexibility, is difficult to coordinate, and is unsuitable for testing and evaluating equipment in boundary, extreme, and complex environments.

[0004] Virtual electromagnetic environments are an important supplement and alternative to physical electromagnetic environments. They can replace real equipment systems or environments through virtual simulation models, effectively overcoming difficulties arising from testing conditions and environments, and are characterized by high efficiency and low risk. However, due to the large spatiotemporal range and dynamic changes in battlefield environment modeling during equipment testing and evaluation, high requirements are placed on resolution accuracy, operational performance, and dynamism. If only a simple simulation system formed by traditional digital modeling is used, it is necessary to establish state models, effect models, and military action models of environmental elements such as terrain, radio frequency, and infrared under typical or extreme conditions. This makes it difficult to adapt to the evaluation of complex environmental factors.

[0005] In summary, existing equipment testing and evaluation methods, when facing complex environmental assessments, suffer from oversimplification in handling environmental impact issues or completely disregard environmental factors and their effects. They lack quantitative analysis and mechanistic research on the impact of the environment on equipment and combat, and cannot objectively and effectively reflect the environmental change process and its effects. They can only meet the basic application issues of simple performance compliance testing, and cannot meet the assessment of equipment's combat effectiveness and operational suitability. Summary of the Invention

[0006] In view of this, the purpose of this invention is to overcome the problem that existing technologies, when assessing the environmental adaptability of equipment-level frequency products in complex environments through simple simulation systems formed by traditional digital modeling, suffer from the uncontrollability of simulation methods when facing actual environmental interference factors. This results in a large deviation between the simulation results and the actual situation, and the accuracy of the assessment is difficult to meet the stringent requirements of high precision and high reliability for complex equipment. Therefore, this invention proposes a complex environment adaptability assessment method based on digital-real collaborative testing.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A method for assessing adaptability to complex environments based on combined numerical and real-world experiments includes: Collect data on complex environments and physical data of the objects being evaluated to obtain target data; After preprocessing the target data, a digital twin model of the object being evaluated in a complex environment is constructed and validated to obtain the target model; Establish a data interaction connection between the target model and the actual physical system, and monitor the interaction data in real time to achieve data-real collaboration; If the difference between the interaction data of the two is greater than the preset difference threshold, the target model is corrected by the generated data of the actual physical system until the preset conditions are met and the corrected model is generated. The model is modified to simulate the working conditions of the evaluated object under different complex environments, and the adaptability of the evaluated object to complex environments is evaluated based on the simulated working conditions, generating an evaluation report.

[0008] Furthermore, complex environmental data and physical data of the object being evaluated are collected to obtain target data, including: By using the standards for assessing the adaptability of equipment to complex environments, the test content and methods for the evaluated objects are determined. Based on the experimental content and methods, determine the data collection methods for complex environments; Data on complex environments is collected according to the methods used for collecting such data; this data includes both human-caused interference data and natural environment data. Human interference data includes malicious electromagnetic interference data released by non-cooperative parties, and natural environment data includes electromagnetic noise caused by different natural environments and civilian facilities; Meanwhile, based on the test content and methods, the physical data acquisition method for the evaluated object is determined, which is used to collect the physical data of the evaluated object accordingly; among them, the evaluated object is a device-level frequency-using product.

[0009] Furthermore, the target data is preprocessed, including: using edge computing to perform preliminary cleaning of complex environmental data and physical data of the object being evaluated, and using median filtering and wavelet denoising to perform noise reduction on the pre-cleaned data, and using max-min standardization to normalize the denoised data to obtain the second target data, which serves as the preprocessed data.

[0010] Furthermore, a digital twin model of the object being evaluated in a complex environment is constructed and validated to obtain the target model, including: Acquire preprocessed complex environmental data and physical data of the object being evaluated, which are denoted as the first data and the second data, respectively. Using the second data, a geometric model of the object being evaluated is constructed using 3D modeling software. Material properties are assigned and meshes are generated using finite element analysis software. Simultaneously, the interaction between the environment and the object being evaluated is simulated using the first data and CFD methods to construct a digital twin model of the object being evaluated in a complex environment. The equivalence principle must be met when simulating the interaction between the complex environment and the object being evaluated using the first data and CFD methods. The key parameter combinations were selected from a pre-built database of relevant parameters using the Latin hypercube sampling method to perform simulation calculations on the digital twin model, and the calculation results were compared with historical experimental data. Based on the difference between the comparison results and the preset comparison conditions, the genetic algorithm is used to optimize the model parameters of the digital twin model, complete the model calibration and verification, and obtain the target model.

[0011] Furthermore, a data interaction connection is established between the target model and the actual physical system, and the interaction data during their collaborative operation is monitored in real time, including: Establish an actual physical system, deploy the verified digital twin model to a cloud computing platform, and use communication technology to establish a two-way data transmission channel between the digital twin model and the actual physical system for data interaction; During data interaction, real-time data generated by the actual physical system is input into the digital twin model in Protobuf format to drive model state updates. The simulation results generated by the digital twin model are fed back to the actual physical system via the OPC UA protocol for difference assessment.

[0012] Furthermore, if the difference between the interaction data of the two is greater than a preset difference threshold, the target model is corrected using the generated data from the actual physical system until the preset conditions are met, and a corrected model is generated, including: A difference analysis mechanism was constructed based on root mean square error; Periodically monitor the interaction data between the digital twin model and the actual physical system during collaborative operation; the interaction data includes real-time data generated by the actual physical system and simulation results generated by the digital twin model; Based on the difference analysis mechanism, the difference between the interaction data of the digital twin model and the actual physical system is analyzed in real time to obtain the difference results, and it is determined whether the difference results are greater than the preset difference threshold to obtain the judgment result. If the judgment result is that the difference is greater than the preset difference threshold, the source of the difference is determined by tracing the data of the actual physical system, and the relevant correction data is obtained based on the source of the difference to improve and correct the digital twin model until the preset conditions are met, the corrected model is generated, and the digital-physical collaborative verification is realized. The correction data includes model structure data, model parameter settings and model data input, and the preset condition is that the difference does not exceed the preset difference threshold.

[0013] Furthermore, by modifying the model to simulate the working conditions of the evaluated object under different complex environments, and based on the simulated working conditions, the adaptability of the evaluated object to complex environments is assessed, generating an assessment report, including: Multiple combinations of environmental parameters are generated based on the Monte Carlo method; each combination of environmental parameters corresponds to a complex environment. Based on a combination of various environmental parameters, a modified model is used to simulate the working conditions of the evaluated object in a complex environment, and the evaluation index parameters of the evaluated object under the simulated state are monitored in real time. Among them, the evaluation index parameters include key performance parameters, structural integrity parameters and functional reliability parameters. For the evaluation index parameters of the evaluated object, the corresponding evaluation index weights are determined by the analytic hierarchy process, and the fuzzy comprehensive evaluation method is combined to quantitatively evaluate the object's adaptability to complex environments. An evaluation report is generated based on the evaluation results.

[0014] Furthermore, a method for assessing the adaptability of a complex environment based on a combined numerical and real-world experiment also includes: acquiring the latest physical data of the object being assessed and new complex environmental data, or only the latest physical data, for dynamically updating the modified model.

[0015] Furthermore, a method for assessing adaptability to complex environments based on combined numerical and real-world experiments also includes: Obtain historical physical data for several objects being evaluated; Based on the attribute type of the physical data, several sets of physical data are split to obtain several types; In each type set, the data are of the same attribute type, and each data carries a unique data ID. The data ID generation logic is related to the collection time of the corresponding data and the number of updates of the corresponding digital twin model. Several data trees are constructed from several types of collections to form a data forest. Each data tree corresponds to a root node, each root node corresponds to an attribute type, and each data tree includes a trunk and branches, or only a trunk. Several nodes on the trunk or branches represent data in the type collection corresponding to the root node, and the nodes on the trunk or branches satisfy the principles of time ordering and call priority. Obtain the latest physical data of the object being evaluated, and perform a missing value search on the latest physical data to obtain the missing data and prompt words, or only obtain the missing data; The target data tree in the data forest is determined based on the attribute type of the missing data, and the data ID is determined by the prompt words as clues to trace back to the corresponding trunk or branch of the data tree to obtain the target data. Among them, the identification and selection of prompt words adopts a fuzzy matching algorithm that combines data ID generation logic and semantic understanding to transform the prompt word input into matching conditions of data ID features to determine the data ID. If only missing data is obtained, determine the target data tree in the data forest based on the attribute type of the missing data; The length limit of the short-term memory list is determined, and the data on the target data tree are added to the short-term memory list in sequence according to the temporal order to form short-term memory. The short-term memory is added to the pre-trained prompt word generation model to generate prompt words, thus completing the acquisition of target data. The latest physical data of the evaluated object is supplemented based on all target data to dynamically update the corrected digital twin model.

[0016] The beneficial effects of this invention are as follows: This invention provides a method for assessing the adaptability of equipment to complex environments based on digital-physical collaborative testing. Compared with existing technologies that only use digital models for simulation, this invention adopts a digital-physical collaborative verification approach, allowing dynamic comparison and verification between the physical system and the digital system formed by the digital twin model. This enables high-precision assessment of the adaptability of complex equipment to complex environments, solving the problems of existing technologies that tend to simplify complex physical phenomena, idealize parameter settings, and are uncontrollable when facing actual environmental interference factors during model construction, resulting in significant deviations between simulation results and reality. This method is beneficial to improving the accuracy of assessments of the adaptability of complex equipment to complex environments in existing type approval testing scenarios.

[0017] Other advantages, objectives, and features of the invention will be set forth in the following description and will be apparent to those skilled in the art in some respects, or may be learned by practice of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating a method for assessing the adaptability of a complex environment based on a combined numerical and real-object test, as described in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the complex data acquisition process in a complex environment adaptability assessment method based on a combined numerical and real-reality experiment, as described in an embodiment of the present invention. Figure 3 This is a system architecture diagram of the test standard for the adaptability of equipment to complex electromagnetic environments, which is involved in a complex environment adaptability assessment method based on numerical and real collaborative testing in an embodiment of the present invention. Figure 4 This is a classification diagram of human interference factors involved in a complex environment adaptability assessment method based on numerical and real collaborative testing in an embodiment of the present invention. Figure 5 This is a flowchart illustrating the automatic data tracing process for model updates in a complex environment adaptability assessment method based on data-real collaborative experiments, as described in an embodiment of the present invention. Detailed Implementation

[0020] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0021] like Figure 1 As shown, this invention proposes a method for assessing adaptability to complex environments based on numerical-real collaborative experiments, comprising: S101. Collect complex environmental data and physical data of the object being evaluated to obtain target data; S102. After preprocessing the target data, construct and generate a digital twin model of the object being evaluated in a complex environment, and verify it to obtain the target model. S103. Establish a data interaction connection between the target model and the actual physical system, and monitor the interaction data in real time to achieve data-physical collaboration. If the difference between the interaction data of the two is greater than a preset difference threshold, S104. Correct the target model using the generated data from the actual physical system until the preset conditions are met, and generate the corrected model. S105. Simulate the simulated working conditions of the evaluated object under different complex environments by modifying the model, and evaluate the adaptability of the evaluated object to complex environments based on the simulated working conditions, and generate an evaluation report. The working principle of the above technical solution is as follows: This application is mainly used in complex environment test scenarios during the finalization stage of complex equipment. It proposes a complex environment adaptation assessment method based on digital-physical collaborative testing to complete the adaptability assessment of complex equipment to complex battlefield environments and determine its equipment performance in complex battlefield environments (especially complex electromagnetic environments). Here, complex equipment refers to equipment-level frequency-using products, including but not limited to radar seekers. Compared with the existing technology that only uses digital models for evaluation, this application innovatively proposes a new evaluation method of digital-physical collaborative verification in this technical field to improve evaluation accuracy. Digital-physical collaboration relies on technologies such as digital twins, the Internet of Things, and big data to build a linkage mode between digital systems and physical systems. Through real-time data interaction, it achieves dynamic state mapping, enabling the digital system to accurately replicate the operating state, structural characteristics, and behavioral patterns of the physical system. At the same time, the physical system receives feedback information from the digital system, forming a dynamic relationship of two-way response and collaborative linkage between the virtual and physical systems, achieving precise docking and synchronous linkage between the two during operation. Digital twin technology can create virtual models of physical entities in a digital way, and use historical data, real-time data and algorithm models to simulate, verify, predict and control physical entities, that is, the entire life cycle process.

[0022] Specifically, this application includes the following process: First, it is necessary to collect data on the complex environment to be evaluated. This data mainly includes complex electromagnetic environment data from the battlefield and natural environment data corresponding to different battlefield environments. Simultaneously, physical data of the complex equipment to be evaluated also needs to be collected. The type of physical data collection is targeted according to the performance type of the equipment to be evaluated, allowing for targeted evaluation of different equipment performance characteristics. The collected data is then integrated to obtain the target data. After obtaining the target data, to reduce the impact of noise and other interference on the final results, the target data needs to be preprocessed to obtain the final data, which serves as model building data. This data is used to construct a digital twin model of the evaluated object in the complex environment. Compared to a simple simulation model, the digital twin model facilitates subsequent interactive verification with the real physical system. After the digital-real co-verification is completed, a higher-precision simulation model can be generated for adapting the complex equipment to the complex environment. Adaptability assessment; After obtaining the above digital twin model, it is necessary to calibrate and verify the model to obtain the target model; then, establish a data interaction connection between the target model and the pre-built actual physical system, and monitor the interaction data generated by the two in real time after successful connection. During the monitoring process, if the difference between the interaction data of the two is found to be greater than the preset difference threshold, the target model is corrected by the data generated by the actual physical system until the preset condition is met, that is, when the difference between the interaction data generated by the two is no longer greater than the preset difference threshold within a set period of time, a corrected model is generated; finally, the corrected model is used to simulate the working conditions of the evaluated object under different complex environments, and the adaptability of the evaluated object to complex environments is evaluated based on the simulated working conditions, and an evaluation report is generated. The evaluation report is preferably displayed in a visual format, and the individual performance of each complex environment and each complex equipment is displayed separately. The beneficial effects of the above technical solution are as follows: Compared with the existing technology that only uses digital models for simulation, this application adopts a digital-real collaborative test method, which allows the physical system and the digital system formed by the digital twin model to be dynamically compared and tested. This enables a high-precision evaluation of the adaptability of complex equipment to complex environments, and solves the problems that existing technologies tend to simplify complex physical phenomena, idealize parameter settings, and are uncontrollable when facing actual environmental interference factors during model construction, resulting in a large deviation between simulation results and reality. This is beneficial to improving the evaluation accuracy of complex equipment in complex environment adaptability assessment in existing equipment type approval test scenarios.

[0023] like Figure 2 As shown, in one embodiment, complex environmental data and physical data of the object being evaluated are collected to obtain target data, including: S201. Determine the test content and methods for the evaluated object by using the equipment adaptability assessment method standard for complex environments; S202. Based on the experimental content and methods, determine the data collection method for complex environments; S203. Collect complex environment data according to the methods for collecting complex environment data; whereby complex environment data includes human interference data and natural environment data; Human interference data includes malicious electromagnetic interference data released by non-cooperative parties, and natural environment data includes electromagnetic noise caused by different natural environments and civilian facilities; at the same time, S204. Based on the test content and methods, determine the physical data acquisition method for the evaluated object, which is used to collect the physical data of the evaluated object accordingly; wherein, the evaluated object is a device-level frequency-using product; The working principle of the above technical solution is as follows: Typically, when inspecting equipment, the system structure of the test standard architecture for testing the equipment's adaptability to complex electromagnetic environments mainly includes general requirements for equipment adaptability to complex electromagnetic environments, standards for testing the basic electromagnetic characteristics of equipment, standards for constructing complex electromagnetic environments, standards for testing methods for adaptability to complex electromagnetic environments, and standards for evaluating adaptability to complex electromagnetic environments, etc., specifically divided as follows: Figure 3 As shown; this technical solution is mainly used to describe the collection of environmental and equipment data. Therefore, this technical solution mainly involves the standard for the assessment method of complex electromagnetic environment adaptability in the above-mentioned general requirements. Based on this standard, the test content and methods of the assessed object (i.e., complex equipment, preferably referring to equipment-level frequency-using products) are determined, i.e., the assessment means. According to the assessment means, the types of complex environmental data that need to be collected are determined, which are used to collect the corresponding environmental data to construct the complex environment. The complex environmental data usually includes man-made interference data and natural environment data. Among them, man-made interference data includes malicious electromagnetic interference data released by non-cooperative parties, and natural environment data includes data from different natural environments and civilian sources. Electromagnetic noise caused by facilities; furthermore, in addition to assessing the adaptability of complex environments, which are of primary importance in complex battlefield environments, this application can also extend to assessing the adaptability of equipment-level products to complex environments. In this case, human-caused interference data includes electromagnetic interference and optoelectronic interference. Specifically, electromagnetic interference and optoelectronic interference are further divided into intentional interference and unintentional interference. Unintentional electromagnetic interference includes radar interference from both sides and interference from radio equipment, while intentional interference includes active suppression interference, disruptive interference, and passive decoy interference. Unintentional optoelectronic interference includes smoke screen interference and optoelectronic camouflage, while intentional interference includes infrared interference and laser interference, etc. Detailed classifications are as follows... Figure 4As shown, natural data should also include one or more of the following: surface environmental data, meteorological environmental data, and hydrological environmental data. The specific environmental data to be collected for the adaptability assessment of the complex equipment will be determined based on the aforementioned test content and methods, in order to complete the adaptability assessment of a certain type of function on the complex equipment. Of course, in addition to collecting environmental data, it is also necessary to determine the data types that need to be monitored for the corresponding functions of the assessed object, based on the test content and methods, and to collect them specifically to complete the overall data collection process for subsequent assessment. The beneficial effects of the above technical solution are as follows: through the above technical solution, effective and targeted data collection is beneficial to improving the accuracy of subsequent adaptability assessment and providing reliable data support for subsequent complex environmental adaptability assessment of the assessed object.

[0024] In one embodiment, preprocessing the target data includes: using edge computing to perform preliminary cleaning of complex environmental data and physical data of the object being evaluated, and using median filtering and wavelet denoising to perform noise reduction on the pre-cleaned data, and using max-min standardization to normalize the denoised data to obtain second target data as preprocessed data. The beneficial effects of the above technical solution are as follows: by completing the data processing of the target data through the above technical solution, it is beneficial to provide reliable data support for the subsequent adaptive evaluation process.

[0025] In one embodiment, a digital twin model of the object being evaluated in a complex environment is constructed and validated to obtain the target model, including: Acquire preprocessed complex environmental data and physical data of the object being evaluated, which are denoted as the first data and the second data, respectively. Using the second data, a geometric model of the object being evaluated is constructed using 3D modeling software. Material properties are assigned and meshes are generated using finite element analysis software. Simultaneously, the interaction between the environment and the object being evaluated is simulated using the first data and CFD methods to construct a digital twin model of the object being evaluated in a complex environment. The equivalence principle must be met when simulating the interaction between the complex environment and the object being evaluated using the first data and CFD methods. The key parameter combinations were selected from a pre-built database of relevant parameters using the Latin hypercube sampling method to perform simulation calculations on the digital twin model, and the calculation results were compared with historical experimental data. Based on the difference between the comparison results and the preset comparison conditions, the genetic algorithm is used to optimize the model parameters of the digital twin model, complete the model verification, and obtain the target model. The working principle of the above technical solution is as follows: After obtaining the preprocessed data, the data is first distinguished, and the preprocessed complex environment data and the physical data of the object being evaluated are respectively recorded as the first data and the second data. This facilitates subsequent operations based on the corresponding data and reduces the data identification time during subsequent processing. Then, 3D modeling software is used to construct and generate the geometric model of the functional components of the object being evaluated using the second data. Next, finite element analysis software (FEA) is used to assign material properties and perform mesh generation. Finally, the first data is used in conjunction with computational fluid dynamics (CFD) to simulate the interaction between the environment and the object being evaluated, thereby constructing a digital twin model of the object being evaluated in a complex environment. Regarding the core technical content disclosed above, the remaining specific construction methods can be constructed using conventional methods of existing technology, and will not be elaborated here. It is worth noting that when using the first data combined with CFD methods to simulate the interaction between a complex environment and the evaluated object, the equivalence principle must be met. This equivalence principle means that the impact of the constructed environment on the tested equipment should be basically equivalent to the impact of the predicted complex battlefield environment on the tested equipment. Then, in the calibration stage, the Latin hypercube sampling method is used to select several combinations of key parameters from the test content and methods of the evaluated object required by the current equipment complex environment adaptability assessment method standard, based on statistical principles, to simulate and calculate the digital twin model, aiming to cover more than 90% of extreme working conditions. Finally, based on the difference between the comparison results and the preset comparison conditions, the simulation calculation results are compared with the historical test data of the corresponding functional components of the complex equipment, and the model parameters of the digital twin model are optimized using a genetic algorithm to complete the model verification and obtain the target model. The beneficial effects of the above technical solution are as follows: By constructing a digital twin model in a complex environment, and by using equivalence simulation, key parameter sampling and genetic algorithm optimization, the model can be accurately verified, efficiently cover extreme working conditions, improve the reliability of evaluation, and reduce data identification time and test costs.

[0026] In one embodiment, a data interaction connection is established between the target model and the actual physical system, and the interaction data during their collaborative operation is monitored in real time, including: Establish an actual physical system, deploy the verified digital twin model to a cloud computing platform, and use communication technology to establish a two-way data transmission channel between the digital twin model and the actual physical system for data interaction; During data interaction, real-time data generated by the actual physical system is input into the digital twin model in Protobuf format to drive model state updates; compression in Protobuf format can reduce the amount of data transmission. The simulation results generated by the digital twin model are fed back to the actual physical system via the OPC UA protocol for difference assessment; OPC UA has cross-platform, security and reliability characteristics, and can realize data interaction between different systems; It is worth noting that the data format and transmission protocol mentioned above are not limited to the Protobuf format and OPC UA protocol proposed in this application. JSON format or Modbus protocol may also be used as alternatives. The beneficial effects of the above technical solution are as follows: By establishing a two-way data channel between the digital twin model and the physical system, a driving model can be updated in real time; at the same time, the simulation results can be used to evaluate differences through standard protocols, realizing collaborative monitoring between digital and physical systems, ensuring efficient and compatible data transmission, and improving the real-time performance and accuracy of system status assessment.

[0027] In one embodiment, if the difference between the interaction data of the two is greater than a preset difference threshold, the target model is corrected using the generated data from the actual physical system until the preset conditions are met, and a corrected model is generated, including: A difference analysis mechanism was constructed based on the root mean square error (RMSE). Periodically monitor the interaction data between the digital twin model and the actual physical system during collaborative operation; the interaction data includes real-time data generated by the actual physical system and simulation results generated by the digital twin model; Based on the difference analysis mechanism, the difference between the interaction data of the digital twin model and the actual physical system is analyzed in real time to obtain the difference results, and it is determined whether the difference results are greater than the preset difference threshold to obtain the judgment result. If the difference is greater than the preset difference threshold, the source of the difference is determined by tracing the data of the actual physical system. Based on the source of the difference, relevant correction data is obtained. Using a Bayesian optimization algorithm, combined with prior knowledge and new data, the parameter is refitted to improve and correct the digital twin model until the preset conditions are met, generating a corrected model and realizing digital-real collaborative verification. The correction data includes model structure data, model parameter settings, and model data input. The preset condition is that the difference does not exceed the preset difference threshold. The beneficial effects of the above technical solution are as follows: By constructing an error analysis mechanism to monitor differences in interactive data, and tracing and locating the source, the model structure, parameters and inputs are corrected and continuously optimized until the preset conditions are met, thereby realizing digital-real collaborative verification and ensuring that the model continuously and accurately reflects the actual operating status of the equipment-level frequency-using products. This is beneficial to ensuring the reliability of the virtual-real mapping when correcting the digital twin model in this application.

[0028] In one embodiment, a modified model is used to simulate the working conditions of the object under evaluation in different complex environments, and the adaptability of the object to complex environments is evaluated based on the simulated working conditions to generate an evaluation report, including: Multiple combinations of environmental parameters are generated based on the Monte Carlo method; each combination of environmental parameters corresponds to a complex environment, preferably an extreme electromagnetic environment. Based on a combination of various environmental parameters, a modified model is used to simulate the working conditions of the evaluated object in a complex environment, and the evaluation index parameters of the evaluated object under the simulated state are monitored in real time. Among them, the evaluation index parameters include key performance parameters, structural integrity parameters and functional reliability parameters. For the evaluation index parameters of the evaluated object, the corresponding evaluation index weights are determined by the analytic hierarchy process, and the fuzzy comprehensive evaluation method is combined to quantitatively evaluate the complex environment adaptability of the evaluated object. An evaluation report is generated based on the evaluation results. The evaluation results can be displayed using 3D visualization technology; The beneficial effects of the above technical solution are as follows: by using the above technical solution, Monte Carlo is used to generate multi-environment parameter combination simulation working conditions for real-time monitoring of multi-dimensional indicators, and combined with hierarchical analysis and fuzzy evaluation for quantitative assessment, a scientific assessment report is generated. This enables the complex environment adaptability method provided in this application to fully cover complex environment scenarios, achieve accurate quantification of adaptability, and provide reliable decision support for researchers.

[0029] In one embodiment, a complex environment adaptability assessment method based on data-real collaborative experiments further includes: acquiring the latest physical data of the object being assessed and new complex environment data, or only the latest physical data, for dynamically updating the correction model; The working principle and beneficial effects of the above technical solution are as follows: In order to improve the dynamism and timeliness of the correction model, this technical solution also provides a dynamic update strategy to support the data assimilation of real-time battlefield data fusion. By continuously collecting new complex environmental data and physical data of the evaluated object, the digital twin model and evaluation results are dynamically updated and optimized based on the new data to achieve real-time response to environmental changes and object state evolution.

[0030] like Figure 5 As shown, in one embodiment, a method for assessing adaptability to complex environments based on numerical-real collaborative experiments further includes: S301. Obtain historical physical data of several objects being evaluated; S302. Based on the attribute type of the physical data, split several physical data into several type sets; In each type set, the data are of the same attribute type, and each data carries a unique data ID. The data ID generation logic is related to the collection time of the corresponding data and the number of updates of the corresponding digital twin model. S303. Several data trees are constructed from several types of collections to form a data forest. Each data tree corresponds to a root node, each root node corresponds to an attribute type, and each data tree includes a trunk and branches, or only a trunk. Several nodes on the trunk or branches represent data in the type collection corresponding to the root node, and the nodes on the trunk or branches satisfy the principles of time ordering and call priority. S304. Obtain the latest physical data of the object being evaluated, and perform a missing value search on the latest physical data to obtain the missing data and prompt words, or only obtain the missing data; S305. Determine the target data tree in the data forest based on the attribute type of the missing data, and trace the data tree to the corresponding trunk or branch by using prompt words as clues to obtain the target data; wherein, the identification and selection of prompt words adopts a fuzzy matching algorithm that combines data ID generation logic and semantic understanding, which is used to transform the prompt word input into matching conditions of data ID features to determine the data ID; S306. If only missing data is obtained, determine the target data tree in the data forest based on the attribute type of the missing data. S307. Determine the length limit of the short-term memory list, and add the data on the target data tree to the short-term memory list in sequence according to the temporal order to form a short-term memory. Add the short-term memory to the pre-trained prompt word generation model to generate prompt words and complete the acquisition of target data. S308. Based on all target data, complete the latest physical data of the evaluated object to dynamically update the corrected digital twin model; The working principle of the above technical solution is as follows: At present, there are two situations when evaluating the adaptability of complex equipment (such as equipment-level frequency products) to complex environments. One is that the physical data of the complex equipment has been fixed, and only the final adaptability verification needs to be performed, that is, the equipment finalization stage verification, to determine whether it meets the preset conditions. Usually, this verification is based on the static data after the "design freeze". Another type is equipment development stage verification, which verifies countless physical data to determine which set of data meets the preset conditions, thereby determining that set of data. That is, in the conceptual design or parameter optimization stage of complex equipment, it is necessary to select the optimal solution from the combination of massive physical data. Under normal circumstances, the development stage of a complex piece of equipment requires tens of thousands of combined simulations to complete the final optimization iteration. This application primarily targets the verification phase of complex equipment development. In practice, to ensure the constructed complex equipment meets pre-defined adaptability to complex environments, researchers often need to iteratively fine-tune the previous set of data as the current set of data in scenarios involving dynamic filtering of multiple sets of physical data for complex equipment. Finally, after numerous fine-tuning iterations, the physical data obtained is used to construct the final complex equipment. This process requires researchers to input data several times. However, in reality, during data fine-tuning, thousands of approximate data sets may be generated within a single optimization cycle. To reduce the workload for researchers, traditional methods typically employ a self-save and fixed index model, automatically saving each set of data and generating a corresponding index. When researchers need to input repetitive data, they only need to input the corresponding index to replace the data input. For example, using a timestamp plus data number as an index... When constructing a model, once a certain data is input, the data is automatically saved and an index 20220504-0002 is generated. However, when the amount of historical data input is too large, researchers still need to remember a large amount of index information to meet the actual requirements. This can easily lead to deviations in simulation results due to index confusion. At the same time, a large amount of data can increase the computational load of index filtering under complex working conditions. To solve this problem, this application provides a method for automatically tracing target data based on missing value self-checking and prompt words. Researchers only need to input variable data and relevant prompt words (unlike fixed indexes, relevant prompt words can be described from multiple perspectives) to determine the ID of the target data. Then, the data can be automatically traced and added to the data group for model correction. At the same time, the spontaneous dynamic arrangement of nodes in the data forest is utilized to avoid the length limitations of the long short-term memory network used in the existing technology. When implementing this solution, it is necessary to obtain some historical physical data for the object being evaluated. If the system corresponding to this method is introduced from the very beginning of the evaluation, the physical data input each time will be automatically obtained from the very beginning of the evaluation. Then, based on the attribute type of the physical data, the aforementioned physical data is split into several type sets. The data in each set are of the same attribute type, such as a material thickness set, a thermal conductivity set, etc. At the same time, a unique ID is generated for each data in the set and it is marked. It is worth noting that the generation logic of this unique ID can be optimized to be related to the data acquisition time and the update frequency of the digital twin model corresponding to the data. In order to improve the correlation between the ID and subsequent prompts and the accuracy of data filtering, the memory preferences of the corresponding researchers can also be incorporated into the ID generation. Several data trees are constructed from several types of collections to form a data forest. Each data tree corresponds to a root node, each root node corresponds to an attribute type, and each data tree includes a trunk and branches, or only a trunk. Several nodes on the trunk or branches represent data in the type collection corresponding to the root node, and the nodes on the trunk or branches satisfy the principles of time ordering and call priority. For example, if there are 5 attribute types, then 5 data trees will be generated in the data forest. The root node of each data tree corresponds to one attribute type. For the convenience of those skilled in the art, only the trunk of each data tree is generated, and no branches are set. In this case, the nodes on each trunk are in a series form, that is, a linear structure. In layman's terms, when there are 8 nodes on a certain data tree, except for the root node and the last node, each node has a parent node and a child node. And because it satisfies the temporal order, in terms of the distribution of nodes, the closer the data collection time of the data corresponding to the node is, the closer it is to the root node. That is, the temporal order of the nodes is positively correlated with the hierarchical depth of the tree. Meanwhile, to avoid the length limitations of existing technologies using Long Short-Term Memory (LSTM) networks, a self-dynamic node arrangement function for the data tree is constructed by utilizing the principle of priority based on the number of calls and combining it with temporal order. Specifically, for each node in the data tree, a comprehensive coefficient concept is introduced. The calculation of this comprehensive coefficient is related to the number of calls and the data acquisition time. Whenever new data is added or a node is called, a full node reordering is triggered. That is, the positions of child nodes on each branch are dynamically arranged according to the comprehensive coefficient, thus constructing a dynamic data tree. The higher the comprehensive coefficient, the closer it is to the root node. The specific formula for calculating the comprehensive coefficient can be customized according to the actual situation. This technical solution is mainly used to provide direction and does not impose specific limitations. To enhance the understanding of this application by those skilled in the art, the concept of branches is further explained here. The emergence of branches is to prevent excessive data on the trunk, which would require a large amount of computation each time data is dynamically arranged and data is traced back on the data tree. Therefore, this application preferably triggers a hierarchical splitting mechanism automatically when the number of nodes on the trunk exceeds 50. The splitting method is selected according to the generation logic of the unique ID. For example, it can split into multi-level child root nodes according to the parameter range. The construction method of the multi-level child root nodes is the same as that of the trunk, still adopting a linear structure, and the nodes satisfy the principle of temporal ordering and priority of call count in the sorting to complete the tree expansion. The comprehensive coefficient of each child node is the sum of the comprehensive coefficients of all the nodes it splits. After the data forest is built, the latest physical data of the evaluated object currently input by the researcher is obtained. The missing values ​​of the latest physical data are automatically searched based on the baseline design scheme of the evaluated object, and the prompt words input by the researcher are identified. The missing data and prompt words are obtained. If the researcher does not input prompt words, only the missing data is obtained. When it is determined that the obtained data includes missing data and prompt words, the target data tree in the data forest is determined according to the attribute type of the missing data, and the data ID is determined by using the prompt words as clues to trace back to the corresponding trunk or branch of the data tree to obtain the target data. It is worth noting that the prompt word identification and selection adopts a fuzzy matching algorithm that combines data ID generation logic and semantic understanding. This algorithm is used to transform the prompt word input into matching conditions for data ID features to determine the data ID. The specific algorithm construction logic for the core technical solution proposed above can be obtained through existing technologies and will not be elaborated here. If only missing data is obtained, the target data tree in the data forest is determined according to the attribute type of the missing data. The length limit of the short-term memory list is determined by the Long Short-Term Memory (LSTM) network used. According to the temporal order, the data on the target data tree is added to the short-term memory list in sequence to form short-term memory. The short-term memory is added to the pre-trained prompt word generation model to generate prompt words, thus completing the acquisition of target data. Finally, all target data are integrated to complete the latest physical data of the evaluated object for dynamic updating of the corrected digital twin model; The beneficial effects of the above technical solution are as follows: By constructing a data forest to store historical physical data, and through attribute splitting and dynamic node arrangement (time + call count priority), combined with semantic fuzzy matching and prompt word generation model, missing data can be automatically traced; it helps to solve the problems of memory burden and large amount of computation in traditional indexes. At the same time, through dynamic filtering in equipment development, it can also improve data retrieval efficiency and traceability accuracy, while reducing human input error and input workload of R&D personnel.

[0031] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. A method for assessing adaptability to complex environments based on numerical-real collaborative experiments, characterized in that, include: Collect data on complex environments and physical data of the objects being evaluated to obtain target data; After preprocessing the target data, a digital twin model of the object being evaluated in a complex environment is constructed and validated to obtain the target model; Establish a data interaction connection between the target model and the actual physical system, and monitor the interaction data in real time to achieve data-real collaboration; If the difference between the interaction data of the two is greater than the preset difference threshold, the target model is corrected by the generated data of the actual physical system until the preset conditions are met and the corrected model is generated. The model is modified to simulate the working conditions of the evaluated object under different complex environments, and the adaptability of the evaluated object to complex environments is evaluated based on the simulated working conditions, generating an evaluation report.

2. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 1, characterized in that, Collect complex environmental data and physical data of the object being evaluated to obtain target data, including: By using the standards for assessing the adaptability of equipment to complex environments, the test content and methods for the evaluated objects are determined. Based on the experimental content and methods, determine the data collection methods for complex environments; Data on complex environments is collected according to the methods used for collecting such data; this data includes both human-caused interference data and natural environment data. Human interference data includes malicious electromagnetic interference data released by non-cooperative parties, and natural environment data includes electromagnetic noise caused by different natural environments and civilian facilities; Meanwhile, based on the test content and methods, the physical data acquisition method for the evaluated object is determined, which is used to collect the physical data of the evaluated object accordingly; among them, the evaluated object is a device-level frequency-using product.

3. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 1, characterized in that, The target data is preprocessed, including: using edge computing to perform preliminary cleaning of complex environmental data and physical data of the object being evaluated; using median filtering and wavelet denoising to reduce noise in the pre-cleaned data; and using max-min standardization to normalize the denoised data to obtain the second target data, which serves as the preprocessed data.

4. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 1, characterized in that, Construct a digital twin model of the object being evaluated in a complex environment, and validate the model to obtain the target model, including: Acquire preprocessed complex environmental data and physical data of the object being evaluated, which are denoted as the first data and the second data, respectively. Using the second data, a geometric model of the object being evaluated is constructed using 3D modeling software. Material properties are assigned and meshes are generated using finite element analysis software. Simultaneously, the interaction between the environment and the object being evaluated is simulated using the first data and CFD methods to construct a digital twin model of the object being evaluated in a complex environment. The equivalence principle must be met when simulating the interaction between the complex environment and the object being evaluated using the first data and CFD methods. The key parameter combinations were selected from a pre-built database of relevant parameters using the Latin hypercube sampling method to perform simulation calculations on the digital twin model, and the calculation results were compared with historical experimental data. Based on the difference between the comparison results and the preset comparison conditions, the genetic algorithm is used to optimize the model parameters of the digital twin model, complete the model calibration and verification, and obtain the target model.

5. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 1, characterized in that, Establish data interaction connections between the target model and the actual physical system, and monitor the interaction data during their collaborative operation in real time, including: Establish an actual physical system, deploy the verified digital twin model to a cloud computing platform, and use communication technology to establish a two-way data transmission channel between the digital twin model and the actual physical system for data interaction; During data interaction, real-time data generated by the actual physical system is input into the digital twin model in Protobuf format to drive model state updates. The simulation results generated by the digital twin model are fed back to the actual physical system via the OPC UA protocol for difference assessment.

6. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 1, characterized in that, If the difference between the interaction data of the two is greater than a preset difference threshold, the target model is corrected using the generated data from the actual physical system until the preset conditions are met, and a corrected model is generated, including: A difference analysis mechanism was constructed based on root mean square error; Periodically monitor the interaction data between the digital twin model and the actual physical system during collaborative operation; the interaction data includes real-time data generated by the actual physical system and simulation results generated by the digital twin model; Based on the difference analysis mechanism, the difference between the interaction data of the digital twin model and the actual physical system is analyzed in real time to obtain the difference results, and it is determined whether the difference results are greater than the preset difference threshold to obtain the judgment result. If the judgment result is that the difference is greater than the preset difference threshold, the source of the difference is determined by tracing the data of the actual physical system, and the relevant correction data is obtained based on the source of the difference to improve and correct the digital twin model until the preset conditions are met, the corrected model is generated, and the digital-physical collaborative verification is realized. The correction data includes model structure data, model parameter settings and model data input, and the preset condition is that the difference does not exceed the preset difference threshold.

7. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 1, characterized in that, The evaluation report is generated by simulating the working conditions of the evaluated object under different complex environments using a modified model, and assessing the object's adaptability to complex environments based on the simulated working conditions. The report includes: Multiple combinations of environmental parameters are generated based on the Monte Carlo method; each combination of environmental parameters corresponds to a complex environment. Based on a combination of various environmental parameters, a modified model is used to simulate the working conditions of the evaluated object in a complex environment, and the evaluation index parameters of the evaluated object under the simulated state are monitored in real time. Among them, the evaluation index parameters include key performance parameters, structural integrity parameters and functional reliability parameters. For the evaluation index parameters of the evaluated object, the corresponding evaluation index weights are determined by the analytic hierarchy process, and the fuzzy comprehensive evaluation method is combined to quantitatively evaluate the object's adaptability to complex environments. An evaluation report is generated based on the evaluation results.

8. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 1, characterized in that, Also includes: Obtain the latest physical data and new complex environmental data of the object being evaluated, or only the latest physical data, to dynamically update the modified model.

9. The method for assessing adaptability to complex environments based on numerical-real collaborative experiments according to claim 8, characterized in that, Also includes: Obtain historical physical data for several objects being evaluated; Based on the attribute type of the physical data, several sets of physical data are split to obtain several types; In each type set, the data are of the same attribute type, and each data carries a unique data ID. The data ID generation logic is related to the collection time of the corresponding data and the number of updates of the corresponding digital twin model. Several data trees are constructed from several types of collections to form a data forest. Each data tree corresponds to a root node, each root node corresponds to an attribute type, and each data tree includes a trunk and branches, or only a trunk. Several nodes on the trunk or branches represent data in the type collection corresponding to the root node, and the nodes on the trunk or branches satisfy the principles of time ordering and call priority. Obtain the latest physical data of the object being evaluated, and perform a missing value search on the latest physical data to obtain the missing data and prompt words, or only obtain the missing data; The target data tree in the data forest is determined based on the attribute type of the missing data, and the data ID is determined by the prompt words as clues to trace back to the corresponding trunk or branch of the data tree to obtain the target data. Among them, the identification and selection of prompt words adopts a fuzzy matching algorithm that combines data ID generation logic and semantic understanding to transform the prompt word input into matching conditions of data ID features to determine the data ID. If only missing data is obtained, determine the target data tree in the data forest based on the attribute type of the missing data; The length limit of the short-term memory list is determined, and the data on the target data tree are added to the short-term memory list in sequence according to the temporal order to form short-term memory. The short-term memory is added to the pre-trained prompt word generation model to generate prompt words, thus completing the acquisition of target data. The latest physical data of the evaluated object is supplemented based on all target data to dynamically update the corrected digital twin model.

Images