Dynamic sequential optimization method and device for aircraft performance prediction model

Abstract

The application provides a dynamic sequential optimization method and device for an aircraft performance prediction model, comprising: extracting an i-th test sample and placing it in a processing array, and performing the following update processing on the processing array: inputting each test sample in the processing array into a hybrid variable precision approximation model and a preset high precision approximation model respectively to determine an average aerodynamic error, performing optimization calculation on model coefficients corresponding to the hybrid scale variable precision approximation model by using the aerodynamic prediction average error, and obtaining an updated hybrid scale variable precision approximation model; verifying the updated hybrid scale variable precision approximation model based on an i+1-th test sample extracted from the test set; and if the verification is passed, determining the updated hybrid scale variable precision approximation model as the aircraft performance prediction model. The application iteratively optimizes the hybrid scale variable precision approximation model at each test point, reduces the calculation workload, and improves the calculation efficiency of model building.

Description

Technical Field

[0001] This application relates to the field of aircraft optimization design technology, and in particular to a dynamic sequential optimization method and apparatus for an aircraft performance prediction model. Background Technology

[0002] In recent years, the aerodynamic shape of aircraft has directly affected various flight performance characteristics, making efficient and accurate design paramount. Currently, there are various methods for aerodynamic calculations of aircraft. Aerodynamic engineering estimation methods based on the surface element method are fast and efficient, but have relatively low accuracy. Computational fluid dynamics (CFD) numerical simulations can obtain detailed flow field information during flight, providing high aerodynamic calculation accuracy, but require significant time and computational resources, resulting in low computational efficiency. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide at least one dynamic sequential optimization method and apparatus for an aircraft performance prediction model, which iteratively optimizes the hybrid scaling variable precision approximation model through each test point, thereby reducing the amount of computation and improving the computational efficiency of model building.

[0004] This application mainly includes the following aspects:

[0005] In a first aspect, embodiments of this application provide a dynamic sequential optimization method for an aircraft performance prediction model. The method includes: acquiring a test set, which includes multiple test samples, each test sample including the shape parameters and flight condition parameters corresponding to the reentry vehicle; extracting the i-th test sample and placing it into a processing array, and performing the following update processing on the processing array: inputting each test sample in the processing array into a hybrid scaling variable precision approximation model and a preset high precision approximation model respectively to determine the average aerodynamic error, where the aerodynamic error is the precision deviation between the first aerodynamic prediction result output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result output by the preset high precision approximation model; using the aerodynamic prediction average error to optimize the model coefficients corresponding to the hybrid scaling variable precision approximation model to obtain an updated hybrid scaling variable precision approximation model; verifying the updated hybrid scaling variable precision approximation model based on the (i+1)-th test sample extracted from the test set; if the verification passes, determining the updated hybrid scaling variable precision approximation model as the aircraft performance prediction model; if the verification fails, placing the (i+1)-th test sample into the processing array and re-performing the update processing on the processing array.

[0006] In one possible implementation, the average aerodynamic error corresponding to the processing array is determined by the following process: for each test sample in the processing array, the following processing is performed: obtaining the first aerodynamic prediction result corresponding to the test sample output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the test sample output by the preset high precision approximation model; calculating the first absolute difference between the first ratio between the first aerodynamic prediction result and the second aerodynamic prediction result and 1, wherein the first absolute difference is the absolute value of the difference between 1 and the first ratio; averaging the sums of the first absolute differences corresponding to each test sample in the processing array to obtain the aerodynamic average error corresponding to the processing array.

[0007] In one possible implementation, the step of optimizing the model coefficients of the hybrid scale variable precision approximation model using the aerodynamic prediction average error to obtain the updated hybrid scale variable precision approximation model includes: determining the optimization objective of the hybrid scale variable precision approximation model, wherein the optimization objective is the minimum average aerodynamic error; optimizing the model parameters of the hybrid scale variable precision approximation model using a genetic algorithm based on given constraints and optimization objective to obtain the optimal values ​​of the model parameters; and establishing the updated hybrid scale variable precision approximation model using the optimal values ​​of the model parameters.

[0008] In one possible implementation, the updated hybrid scaling variable precision approximation model is verified as follows: the (i+1)th test sample extracted from the test set is input into the updated hybrid scaling variable precision approximation model and the preset high-precision approximation model, respectively, to obtain the first aerodynamic prediction result corresponding to the (i+1)th test sample output by the updated hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the (i+1)th test sample output by the preset high-precision approximation model; the second absolute difference between the second ratio and 1 between the first aerodynamic prediction result and the second aerodynamic prediction result corresponding to the (i+1)th test sample is calculated, where the second absolute difference is the absolute value of the difference between 1 and the second ratio; the updated hybrid scaling variable precision approximation model is verified based on the second absolute difference.

[0009] In one possible implementation, the step of verifying the updated hybrid scaling variable precision approximation model based on the second absolute difference includes: determining whether the second absolute difference corresponding to the (i+1)th test sample meets the preset precision; if the second absolute difference corresponding to the (i+1)th test sample meets the preset precision, then the verification is determined to be successful; if the second absolute difference of the (i+1)th test sample does not meet the preset precision, then the verification is determined to be successful.

[0010] Secondly, embodiments of this application also provide a dynamic sequential optimization device for an aircraft performance prediction model. The device includes: an acquisition module for acquiring a test set, the test set including multiple test samples, each test sample including the shape parameters and flight condition parameters corresponding to the reentry vehicle; an extraction module for extracting the i-th test sample and placing it into a processing array; and an error determination module for inputting each test sample in the processing array into a hybrid scaling variable precision approximation model and a preset high precision approximation model, respectively, to determine the average aerodynamic error, wherein the aerodynamic error is the difference between the first aerodynamic prediction result output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result output by the preset high precision approximation model. The module analyzes the accuracy deviation between aerodynamic prediction results; the optimization module optimizes the model coefficients of the hybrid scale variable accuracy approximation model using the average error of aerodynamic prediction to obtain the updated hybrid scale variable accuracy approximation model; the verification module verifies the updated hybrid scale variable accuracy approximation model based on the (i+1)th test sample extracted from the test set; the model determination module determines the updated hybrid scale variable accuracy approximation model as the aircraft performance prediction model if the verification passes; and the reprocessing module adds the (i+1)th test sample to the processing array and re-executes the update processing on the processing array if the verification fails.

[0011] In one possible implementation, the error determination module is further configured to: for each test sample in the processing array, perform the following processing: obtain the first aerodynamic prediction result corresponding to the test sample output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the test sample output by the preset high precision approximation model; calculate the first absolute difference between the first ratio between the first aerodynamic prediction result and the second aerodynamic prediction result and 1, wherein the first absolute difference is the absolute value of the difference between 1 and the first ratio; and average the sum of the first absolute differences corresponding to each test sample in the processing array to obtain the aerodynamic average error corresponding to the processing array.

[0012] In one possible implementation, the optimization module is further configured to: determine the optimization objective corresponding to the hybrid scaling variable precision approximation model, wherein the optimization objective is the minimum average aerodynamic error; based on the given constraints and optimization objective, use a genetic algorithm to optimize the model parameters corresponding to the hybrid scaling variable precision approximation model to obtain the optimal values ​​of the model parameters; and use the optimal values ​​of the model parameters to establish an updated hybrid scaling variable precision approximation model.

[0013] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to perform the steps of the dynamic sequential optimization method for the aircraft performance prediction model in the first aspect or any possible implementation of the first aspect.

[0014] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the dynamic sequential optimization method for the aircraft performance prediction model in the first aspect or any possible implementation of the first aspect.

[0015] This application provides a dynamic sequential optimization method and apparatus for an aircraft performance prediction model, comprising: extracting the i-th test sample and placing it into a processing array, and performing the following update processing on the processing array: inputting each test sample in the processing array into a hybrid scaling variable precision approximation model and a preset high precision approximation model respectively to determine the average aerodynamic error; using the aerodynamic prediction average error to optimize the model coefficients corresponding to the hybrid scaling variable precision approximation model to obtain an updated hybrid scaling variable precision approximation model; verifying the updated hybrid scaling variable precision approximation model based on the (i+1)-th test sample extracted from the test set; if the verification is successful, determining the updated hybrid scaling variable precision approximation model as the aircraft performance prediction model. This application iteratively optimizes the hybrid scaling variable precision approximation model at each test point, reducing the computational workload and improving the computational efficiency of model building.

[0016] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating an optimization method for an aircraft performance prediction model provided in an embodiment of this application is shown.

[0019] Figure 2This illustration shows a structural schematic diagram of an optimization device for an aircraft performance prediction model provided in an embodiment of this application;

[0020] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0022] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0023] The currently commonly used multi-precision proxy modeling method has the following drawbacks:

[0024] The multi-precision surrogate modeling method based on spatial mapping mainly transforms the design space of the low-precision simulation function so that the optimal solution of the low-precision simulation function can approximate the optimal solution of the high-precision simulation function. The core of this process is to find a suitable mapping relationship to transform the high-precision and low-precision design spaces. However, the training process of existing multi-precision surrogate modeling generally requires calculation of all test points. In this case, even if all test points have not been traversed, all points still need to be traversed, which reduces the construction efficiency of the multi-precision surrogate model.

[0025] Based on this, this application provides a dynamic sequential optimization method and apparatus for an aircraft performance prediction model. By iteratively optimizing a hybrid-scale variable-precision approximation model at each test point, the computational workload is reduced and the computational efficiency of model building is improved. Specifically, as follows:

[0026] Please see Figure 1 , Figure 1 A flowchart illustrating an optimization method for an aircraft performance prediction model provided in an embodiment of this application is shown. Figure 1 As shown in the embodiments of this application, the optimization method for the aircraft performance prediction model includes the following steps:

[0027] S100, Get the test set.

[0028] The test set includes multiple test samples. Each test sample includes the shape parameters and flight condition parameters corresponding to the reentry vehicle. This application needs to extract multiple samples from a preset sample pool using Latin hypercube sampling to obtain a sample set. The sample set includes multiple samples, each of which includes the shape parameters and flight condition parameters corresponding to the reentry vehicle. The sample set is divided into a training set and a test set according to a preset ratio. The training set includes multiple training samples.

[0029] S200, Extract the i-th test sample and put it into the processing array.

[0030] S300. Input each test sample in the processing array into the hybrid scaling variable precision approximation model and the preset high precision approximation model respectively to determine the average aerodynamic error.

[0031] Among them, the aerodynamic error is the accuracy deviation between the first aerodynamic prediction result output by the hybrid scaling variable accuracy approximation model and the second aerodynamic prediction result output by the preset high accuracy approximation model.

[0032] This application requires the prior construction of a hybrid scale variable precision approximation model using a training set. Specifically, the hybrid scale variable precision approximation model can be constructed in the following way:

[0033] The pre-built low-precision model is used to process each training sample in the training set to obtain the third aerodynamic prediction result YSL corresponding to each training sample. Multiple target training samples are randomly selected from the training set, and the high-precision model is used to process the multiple target training samples to obtain the second aerodynamic prediction result YSH1 corresponding to each target training sample. The high-precision model can be a CFD numerical simulation model.

[0034] For each target training sample, a scaling calculation is performed. Specifically, the difference between the second aerodynamic prediction result YSH1 and the third aerodynamic prediction result YSL is calculated to obtain the additive scaling ΔYSP, and the quotient of the second aerodynamic prediction result YSH1 and the third aerodynamic prediction result YSL is calculated and then subtracted by 1 to obtain the multiplicative scaling ΔYSM.

[0035] By combining a BP neural network, additive scaling and multiplicative scaling models are established using ΔYSP and ΔYSM corresponding to each target training sample. The weight coefficient a corresponding to ΔYSM is introduced to establish a hybrid scaling variable precision approximation model related to the weight coefficient a.

[0036] In a preferred embodiment, the average aerodynamic error corresponding to the processing array is determined by the following method:

[0037] For each test sample in the processing array, the following processing is performed: obtain the first aerodynamic prediction result corresponding to the test sample output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the test sample output by the preset high precision approximation model; calculate the first absolute difference between the first ratio between the first aerodynamic prediction result and the second aerodynamic prediction result and 1, where the first absolute difference is the absolute value of the difference between 1 and the first ratio; and average the sum of the first absolute differences corresponding to each test sample in the processing array to obtain the aerodynamic average error corresponding to the processing array.

[0038] Specifically, the aerodynamic average error corresponding to the processing array can be determined using the following formula:

[0039]

[0040] In this formula, MET represents the aerodynamic mean error corresponding to the processing array, n represents the number of test samples in the processing array, and YT i YH represents the first aerodynamic prediction result for the i-th test sample output by the hybrid scaling variable accuracy approximation model. i This represents the second aerodynamic prediction result of the i-th test sample output by the preset high-precision approximation model.

[0041] S400: Optimize the model coefficients corresponding to the hybrid scale variable precision approximation model by using the aerodynamic prediction average error to obtain the updated hybrid scale variable precision approximation model.

[0042] The weight coefficient 'a' corresponding to ΔYSM is the model coefficient corresponding to the mixed-scale variable-precision approximation model.

[0043] In a preferred embodiment, step S400 includes:

[0044] The optimization objective for the hybrid scaling variable precision approximation model is determined, which is to minimize the average aerodynamic error. Based on the given constraints and optimization objective, the model parameters of the hybrid scaling variable precision approximation model are optimized using a genetic algorithm to obtain the optimal values ​​of the model parameters. The updated hybrid scaling variable precision approximation model is then established using the optimal values ​​of the model parameters.

[0045] In this application, the weight coefficient 'a' is determined as the optimization variable. Under given constraints and optimization objectives, a genetic algorithm is used to optimize the weight coefficient 'a' to obtain the optimal value of the weight coefficient 'a', thereby obtaining the updated hybrid scaling variable precision approximation model. At this point, the weight coefficient 'a' is no longer an unknown.

[0046] S500: Validate the updated hybrid scaling variable precision approximation model based on the (i+1)th test sample extracted from the test set.

[0047] In a preferred embodiment, the updated hybrid scaling variable precision approximation model is validated in the following manner:

[0048] The (i+1)th test sample extracted from the test set is input into the updated hybrid scaling variable precision approximation model and the preset high precision approximation model, respectively, to obtain the first aerodynamic prediction result corresponding to the (i+1)th test sample output by the updated hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the (i+1)th test sample output by the preset high precision approximation model. The second absolute difference between the second ratio corresponding to the (i+1)th test sample and 1 is calculated. The second absolute difference is the absolute value of the difference between 1 and the second ratio. The updated hybrid scaling variable precision approximation model is verified based on the second absolute difference.

[0049] Specifically, the second absolute difference can be determined using the following formula:

[0050]

[0051] In this formula, ET represents the second absolute difference, and YTT i+1 YH represents the first aerodynamic prediction result corresponding to the (i+1)th test sample output by the updated hybrid scaling variable precision approximation model. i+1 This represents the second aerodynamic prediction result corresponding to the (i+1)th test sample output by the preset high-precision approximation model.

[0052] In another preferred embodiment, the step of validating the updated hybrid scaling variable precision approximation model based on the second absolute difference includes:

[0053] Determine whether the second absolute difference corresponding to the (i+1)th test sample meets the preset precision. If the second absolute difference corresponding to the (i+1)th test sample meets the preset precision, the verification is confirmed to be successful. If the second absolute difference of the (i+1)th test sample does not meet the preset precision, the verification is confirmed to be successful.

[0054] S600: Determine whether the verification passed.

[0055] S700 If the verification is successful, the updated hybrid scale variable precision approximation model will be determined as the aircraft performance prediction model.

[0056] S710. If the verification fails, the (i+1)th test sample is added to the processing array, and the processing array is updated again.

[0057] In this application, the selection of weight coefficients is derived from an optimization problem. In each round of optimization, a sequential point-addition optimization method is adopted, that is, the model is first optimized through the i-th test point, and then the model accuracy is verified through the (i+1)-th test point. If the accuracy does not meet the requirements, the (i+1)-th test point is added to the processing array, and another round of parameter iteration optimization is performed. This method can greatly reduce the number of high-precision sample points used, reduce the workload of computation, and improve computational efficiency.

[0058] The optimized aircraft performance prediction model in this application can obtain the flight performance prediction results of the reentry vehicle based on the input shape parameters and flight condition parameters of the reentry vehicle. Furthermore, the aircraft performance prediction model obtained in this application is constructed based on hybrid scaling, which further improves the accuracy and precision of reentry vehicle performance prediction.

[0059] Based on the same application concept, this application also provides an apparatus for determining an aircraft performance prediction model corresponding to the method for determining an aircraft performance prediction model provided in the above embodiments. Since the principle of the apparatus in this application for solving the problem is similar to the method for determining an aircraft performance prediction model in the above embodiments of this application, the implementation of the apparatus can refer to the implementation of the method, and the repeated parts will not be described again.

[0060] Please see Figure 2 , Figure 2 A schematic diagram of the structure of an optimization device for an aircraft performance prediction model provided in an embodiment of this application is shown. Figure 2 As shown, the device includes:

[0061] The acquisition module 800 is used to acquire a test set, which includes multiple test samples. Each test sample includes the shape parameters and flight condition parameters corresponding to the reentry vehicle.

[0062] The extraction module 810 is used to extract the i-th test sample and put it into the processing array.

[0063] The error determination module 820 inputs each test sample in the processing array into the hybrid scaling variable precision approximation model and the preset high precision approximation model respectively to determine the average aerodynamic error. The aerodynamic error is the precision deviation between the first aerodynamic prediction result output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result output by the preset high precision approximation model.

[0064] The optimization module 830 is used to optimize the model coefficients corresponding to the hybrid scale variable precision approximation model by using the aerodynamic prediction average error, so as to obtain the updated hybrid scale variable precision approximation model.

[0065] The verification module 840 is used to verify the updated hybrid scaling variable precision approximation model based on the (i+1)th test sample extracted from the test set.

[0066] The model determination module 850 is used to determine the updated hybrid scaling variable precision approximation model as the aircraft performance prediction model if the verification is successful.

[0067] The reprocessing module 860 is used to add the (i+1)th test sample to the processing array and re-execute the update processing on the processing array if the verification fails.

[0068] The error determination module 820 is further configured to: for each test sample in the processing array, perform the following processing: obtain the first aerodynamic prediction result corresponding to the test sample output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the test sample output by the preset high precision approximation model; calculate the first absolute difference between the first ratio between the first aerodynamic prediction result and the second aerodynamic prediction result and 1, wherein the first absolute difference is the absolute value of the difference between 1 and the first ratio; and average the sum of the first absolute differences corresponding to each test sample in the processing array to obtain the aerodynamic average error corresponding to the processing array.

[0069] The optimization module 630 is also used to: determine the optimization objective corresponding to the mixed-scale variable-precision approximation model, the optimization objective being the minimum average aerodynamic error; based on the given constraints and optimization objective, use a genetic algorithm to optimize the model parameters corresponding to the mixed-scale variable-precision approximation model to obtain the optimal values ​​of the model parameters; and use the optimal values ​​of the model parameters to establish an updated mixed-scale variable-precision approximation model.

[0070] Preferably, the verification module 640 is further configured to: input the (i+1)th test sample extracted from the test set into the updated hybrid scaling variable precision approximation model and the preset high precision approximation model respectively, to obtain the first aerodynamic prediction result corresponding to the (i+1)th test sample output by the updated hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the (i+1)th test sample output by the preset high precision approximation model; calculate the second absolute difference between the second ratio and 1 between the first aerodynamic prediction result and the second aerodynamic prediction result corresponding to the (i+1)th test sample, wherein the second absolute difference is the absolute value of the difference between 1 and the second ratio; and verify the updated hybrid scaling variable precision approximation model based on the second absolute difference.

[0071] Preferably, the verification module 640 is further configured to: determine whether the second absolute difference corresponding to the (i+1)th test sample meets the preset precision; if the second absolute difference corresponding to the (i+1)th test sample meets the preset precision, then the verification is confirmed to be passed; if the second absolute difference of the (i+1)th test sample does not meet the preset precision, then the verification is confirmed to be passed.

[0072] Based on the same application concept, please refer to Figure 3 , Figure 3 The diagram illustrates the structure of an electronic device 900 according to an embodiment of this application. The electronic device 900 includes a processor 910, a memory 920, and a bus 930. The memory 920 stores machine-readable instructions executable by the processor 910. When the electronic device 900 is running, the processor 910 and the memory 920 communicate via the bus 930. The machine-readable instructions are executed by the processor 910 to perform the steps of the dynamic sequential optimization method for the aircraft performance prediction model provided in any of the above embodiments.

[0073] Based on the same concept, this application also provides a computer-readable storage medium storing a computer program, which, when run by a processor, executes the steps of the dynamic sequential optimization method for the aircraft performance prediction model provided in the above embodiments.

[0074] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0075] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0076] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0077] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0078] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A dynamic sequential optimization method for an aircraft performance prediction model, characterized in that, The method includes: Obtain a test set, which includes multiple test samples, each of which includes the shape parameters and flight condition parameters corresponding to the reentry vehicle; Extract the i-th test sample and put it into the processing array, and perform the following update processing on the processing array: input each test sample in the processing array into the hybrid scale variable precision approximation model and the preset high precision approximation model respectively to determine the average aerodynamic error. The aerodynamic error is the precision deviation between the first aerodynamic prediction result output by the hybrid scale variable precision approximation model and the second aerodynamic prediction result output by the preset high precision approximation model. Optimize the model coefficients corresponding to the hybrid scale variable precision approximation model using the average aerodynamic prediction error to obtain the updated hybrid scale variable precision approximation model. The updated hybrid scaling variable precision approximation model is validated based on the (i+1)th test sample extracted from the test set. If the verification is successful, the updated hybrid scaling variable precision approximation model will be determined as the aircraft performance prediction model. If the verification fails, the (i+1)th test sample is added to the processing array, and the update process is re-executed on the processing array. The hybrid scaling variable precision approximation model is constructed in the following way: The pre-built low-precision model is used to process each training sample in the training set to obtain the third aerodynamic prediction result corresponding to each training sample. Multiple target training samples are randomly selected from the training set, and the high-precision model is used to process the multiple target training samples to obtain the second aerodynamic prediction result corresponding to each target training sample. For each target training sample, a scaling calculation is performed. Specifically, the difference between the second and third aerodynamic prediction results is calculated to obtain an additive scaling, and the quotient between the second and third aerodynamic prediction results is subtracted to obtain a multiplicative scaling. By combining a backpropagation neural network, an additive scaling model and a multiplicative scaling model are established using the additive and multiplicative scaling corresponding to each target training sample. The weight coefficients corresponding to the multiplicative scaling are introduced to establish a hybrid scaling variable precision approximation model related to the weight coefficients corresponding to the multiplicative scaling.

2. The method according to claim 1, characterized in that, The average aerodynamic error corresponding to the processing array is determined in the following manner: For each test sample in the processing array, perform the following processing: Obtain the first aerodynamic prediction result corresponding to the test sample output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the test sample output by the preset high precision approximation model; Calculate the first absolute difference between the first ratio between the first aerodynamic prediction result and the second aerodynamic prediction result and 1, where the first absolute difference is the absolute value of the difference between 1 and the first ratio; The aerodynamic average error corresponding to the processing array is obtained by averaging the sums of the first absolute differences corresponding to each test sample in the processing array.

3. The method according to claim 1, characterized in that, The steps for optimizing the model coefficients of the hybrid scaling variable-precision approximation model using the aerodynamic prediction average error to obtain the updated hybrid scaling variable-precision approximation model include: Determine the optimization objective corresponding to the hybrid scaling variable accuracy approximation model, wherein the optimization objective is the minimum average aerodynamic error; Based on the given constraints and the optimization objective, the genetic algorithm is used to optimize the model parameters corresponding to the hybrid scaling variable precision approximation model to obtain the optimal values ​​of the model parameters; Using the optimal values ​​of the model parameters, an updated hybrid scaling variable precision approximation model is established.

4. The method according to claim 1, characterized in that, The updated hybrid scaling variable precision approximation model is validated using the following methods: The (i+1)th test sample extracted from the test set is input into the updated hybrid scaling variable precision approximation model and the preset high precision approximation model, respectively, to obtain the first aerodynamic prediction result corresponding to the (i+1)th test sample output by the updated hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the (i+1)th test sample output by the preset high precision approximation model. Calculate the second absolute difference between the second ratio of the first aerodynamic prediction result corresponding to the (i+1)th test sample and 1, where the second absolute difference is the absolute value of the difference between 1 and the second ratio. The updated hybrid scaling variable precision approximation model is validated based on the second absolute difference.

5. The method according to claim 4, characterized in that, The steps for validating the updated hybrid scaling variable precision approximation model based on the second absolute difference include: Determine whether the second absolute difference corresponding to the (i+1)th test sample meets the preset precision. If the second absolute difference corresponding to the (i+1)th test sample meets the preset precision, then the verification is deemed successful. If the second absolute difference of the (i+1)th test sample does not meet the preset precision, then the verification is determined to be unsuccessful.

6. A dynamic sequential optimization device for predicting aircraft performance models, characterized in that, The device includes: The acquisition module is used to acquire a test set, which includes multiple test samples. Each test sample includes the shape parameters and flight condition parameters corresponding to the reentry vehicle. The extraction module is used to extract the i-th test sample and put it into the processing array; The error determination module inputs each test sample in the processing array into the hybrid scaling variable precision approximation model and the preset high precision approximation model respectively to determine the average aerodynamic error. The aerodynamic error is the precision deviation between the first aerodynamic prediction result output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result output by the preset high precision approximation model. The optimization module is used to optimize the model coefficients of the hybrid scale variable precision approximation model by using the aerodynamic prediction average error, so as to obtain the updated hybrid scale variable precision approximation model. The verification module is used to verify the updated hybrid scaling variable precision approximation model based on the (i+1)th test sample extracted from the test set. The model determination module is used to determine the updated hybrid scaling variable accuracy approximation model as the aircraft performance prediction model if the verification is successful. The reprocessing module is used to add the (i+1)th test sample to the processing array and re-execute the update processing on the processing array if the verification fails. The hybrid scaling variable precision approximation model is created as follows: The pre-built low-precision model is used to process each training sample in the training set to obtain the third aerodynamic prediction result corresponding to each training sample. Multiple target training samples are randomly selected from the training set, and the high-precision model is used to process the multiple target training samples to obtain the second aerodynamic prediction result corresponding to each target training sample. For each target training sample, a scaling calculation is performed. Specifically, the difference between the second and third aerodynamic prediction results is calculated to obtain an additive scaling, and the quotient between the second and third aerodynamic prediction results is subtracted to obtain a multiplicative scaling. By combining a backpropagation neural network, an additive scaling model and a multiplicative scaling model are established using the additive and multiplicative scaling corresponding to each target training sample. The weight coefficients corresponding to the multiplicative scaling are introduced to establish a hybrid scaling variable precision approximation model related to the weight coefficients corresponding to the multiplicative scaling.

7. The apparatus according to claim 6, characterized in that, The error determination module is also used for: For each test sample in the processing array, perform the following processing: Obtain the first aerodynamic prediction result corresponding to the test sample output by the hybrid scaling variable precision approximation model and the second aerodynamic prediction result corresponding to the test sample output by the preset high precision approximation model; Calculate the first absolute difference between the first ratio between the first aerodynamic prediction result and the second aerodynamic prediction result and 1, where the first absolute difference is the absolute value of the difference between 1 and the first ratio; The aerodynamic average error corresponding to the processing array is obtained by averaging the sums of the first absolute differences corresponding to each test sample in the processing array.

8. The apparatus according to claim 6, characterized in that, The optimization module is also used for: Determine the optimization objective corresponding to the hybrid scaling variable accuracy approximation model, wherein the optimization objective is the minimum average aerodynamic error; Based on the given constraints and the optimization objective, the genetic algorithm is used to optimize the model parameters corresponding to the hybrid scaling variable precision approximation model to obtain the optimal values ​​of the model parameters; Using the optimal values ​​of the model parameters, an updated hybrid scaling variable precision approximation model is established.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. The machine-readable instructions are executed by the processor to perform the steps of the dynamic sequential optimization method for the aircraft performance prediction model as described in any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the dynamic sequential optimization method for the aircraft performance prediction model as described in any one of claims 1 to 5.

Images