Intelligent Prediction Method for Multi-Field Coupling of Aerodynamics, Thermal and Structural Dynamics in High-Speed ​​Vehicles

Abstract

This invention relates to an intelligent prediction method for multi-field coupling of aerodynamics, heat, and structure in high-speed aircraft, belonging to the field of high-speed aircraft design technology. Based on the aircraft's flight trajectory, sample parameters for calculating aerodynamic and aerothermal states are extracted; the aircraft is modeled, and aerodynamic and aerothermal datasets are obtained through simulation calculations; the aircraft temperature field along the trajectory is calculated; the aerodynamic dataset is trained to obtain an aerodynamic prediction model; the structural stress-displacement field data of the aircraft model are calculated; from the aerodynamic dataset, the aircraft temperature field, and the structural stress-displacement field data of the aircraft model, corresponding data of the aircraft feature points of interest are extracted to obtain a feature point dataset; the feature point dataset is trained to obtain accurate prediction results; this invention uses an artificial intelligence algorithm for offline learning and modeling of simulated aerodynamic-thermal-structure coupling data, enabling the input flight trajectory to quickly predict and output data on the changes of aircraft feature points along the flight trajectory.

Description

Technical Field

[0001] This invention belongs to the field of high-speed aircraft design technology and relates to an intelligent prediction method for multi-field coupling of aerodynamics, heat and structure of high-speed aircraft. Background Technology

[0002] High-speed aircraft face complex multi-physics coupling problems, making it difficult to predict the structural state online. Each physical field has different characteristics and different analysis methods, making it difficult to obtain a relatively accurate structural response. Therefore, exploring the coupling relationship between different physical fields and accurately predicting the aerodynamic / thermal environment, structural temperature, stress and displacement state of the aircraft is of great significance for the development of high-speed aircraft.

[0003] Currently, high-speed aircraft face increasingly complex mission requirements and more intricate flight trajectories. Aerothermal heating generates high temperatures and temperature gradients on structures, leading to a decline in the mechanical properties of structural materials. Inhomogeneous thermal deformation and stress within the structure alter its stiffness characteristics and inherent vibrational properties. Under complex trajectories, conducting time-domain aeroelastic analysis of the entire trajectory is extremely costly, given the high dimensionality, strong nonlinearity, and large data volume of the fluid dynamics problems. While computational structural mechanics solutions are generally less time-consuming than computational fluid dynamics solutions for Navier-Stokes equations with tens of thousands to tens of millions of grids (often taking hours or even days), solving differential equations with tens to hundreds of thousands of grids also often takes minutes or even hours, a significant factor that cannot be ignored in multi-field coupled computations. Summary of the Invention

[0004] The technical problem solved by this invention is to overcome the shortcomings of the prior art and propose an intelligent prediction method for multi-field coupling of aerodynamics, heat and structure of high-speed aircraft. After performing offline learning and modeling of simulated aerodynamics-heat and structure coupling data using artificial intelligence algorithms, the method can quickly predict and output the changes of aircraft feature point data along the flight trajectory by inputting the flight trajectory.

[0005] The solution of the present invention is:

[0006] A multi-field coupled intelligent prediction method for aerodynamics, thermodynamics, and structure of high-speed aircraft includes:

[0007] Step 1: Based on the aircraft's flight trajectory, extract sample parameters for calculating aerodynamic and aerothermal states;

[0008] Step 2: Model the aircraft and simulate and calculate aerodynamic and aerothermal datasets based on the sample parameters extracted in Step 1.

[0009] Step 3: Based on the aerothermal dataset from Step 2, calculate the aircraft temperature field along the trajectory;

[0010] Step 4: Train the aerodynamic dataset from Step 2 to obtain the aerodynamic prediction model;

[0011] Step 5: Based on the aircraft temperature field along the trajectory in Step 3 and the aerodynamic prediction model in Step 4, calculate the structural stress-displacement field data of the aircraft model.

[0012] Step 6: Obtain the feature point dataset;

[0013] Step 7: Train the feature point dataset to obtain accurate prediction results.

[0014] In the above-mentioned intelligent prediction method for multi-field coupling of aerodynamics, heat and structure of high-speed aircraft, in step one, a uniform sampling method is used to extract sample parameters; the sample parameters include the aircraft's wall temperature, flight altitude, angle of attack and Mach number.

[0015] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, heat, and structure of high-speed aircraft, step two involves simulating and calculating aerodynamic forces and aerothermal effects as follows:

[0016] The aerodynamic mesh is divided into the aircraft model; the sample parameters are used as boundary conditions; numerical simulation is performed to obtain aerodynamic and aerothermal datasets.

[0017] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, heat, and structure of high-speed aircraft, the calculation method for the aircraft temperature field along the trajectory in step three is as follows:

[0018] S31, obtained through interpolation calculation of aerodynamic thermal dataset. tn At any given moment, the aerodynamic heat;

[0019] S32, Settings t0 Given that the initial structural temperature field and radiation boundary conditions of the spacecraft are known at any given moment; and that the thermal load is known based on the spacecraft's flight trajectory; the results are obtained iteratively. tn Temperature field of the initial structure of the spacecraft at any given moment;

[0020] S33, Applying S31 to the aircraft model tn Nodal aerodynamic heat and radiation boundary conditions at time points tn The initial structural temperature field of the spacecraft was obtained by transient structural heat transfer analysis using NASTRAN software. tn+1 The structural temperature field at any given time;

[0021] S34, from tn+1 The surface temperature of the spacecraft in the structural temperature field at time t is used as tn+1 The nodal aerodynamic heat at time will tn+1 The structural temperature field at time t is used as tn+1Get the initial temperature field of the spacecraft structure at any time; repeat steps S31-S33 to obtain the temperature field of the spacecraft along the trajectory.

[0022] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft, in step S33, the results are obtained iteratively. tn The method for determining the initial structural temperature field of the spacecraft is as follows:

[0023] exist t0 Based on the initial structural temperature field of the spacecraft at any given time, the thermal load is input as the load variable of the spacecraft model, and the results are obtained iteratively. tn Temperature field of the initial structure of the spacecraft at any given moment.

[0024] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, heat, and structure of high-speed aircraft, the method for training the aerodynamic dataset from step two in step four is as follows:

[0025] S41. Use the POD method to reduce the order of aerodynamic data;

[0026] S42. Based on the mapping relationship between sample parameters and reduced-order aerodynamic data, a BP neural network is used for training to obtain an aerodynamic prediction model.

[0027] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft, step five involves calculating the structural stress and displacement of the aircraft model as follows:

[0028] S51. Based on the aerodynamic prediction model, calculate... tn Aerodynamic forces at specific moments;

[0029] S52, will tn At each time point, the aerodynamic forces and the temperature field of the aircraft along the trajectory in step three are applied to the aircraft model. The structural stress and displacement field are obtained by structural strength analysis using NASTRAN software.

[0030] S53, Enter tn+1 Cycle through steps S51-S52 to obtain stress and displacement field data of the aircraft structure along the trajectory.

[0031] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft, in step S51, the following is obtained: tn The method for determining the aerodynamic forces at specific time points is as follows:

[0032] Will tn The altitude, angle of attack, and Mach number of the aircraft at any given time are used as inputs to the aerodynamic prediction model, and the output is t. n Aerodynamic forces at specific moments.

[0033] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft, the method for obtaining the feature point dataset in step six is ​​as follows:

[0034] In the aerodynamic dataset of step two, the temperature field of the aircraft in step three, and the structural stress-displacement field data of the aircraft model in step five, the corresponding data of the aircraft feature points of interest are extracted to obtain the feature point dataset.

[0035] In the aforementioned intelligent prediction method for multi-field coupling of aerodynamics, heat, and structure of high-speed aircraft, the method for training the feature point dataset in step seven is as follows:

[0036] S71. Use the POD method to reduce the order of the feature point dataset;

[0037] S72. A BP neural network is used to train the mapping relationship between the flight trajectory and the reduced feature point dataset.

[0038] The advantages of this invention compared to the prior art are:

[0039] (1) Based on the aerodynamic-thermal-structural multi-field dataset obtained by simulation, this invention uses the Proper Orthogonal Decomposition (POD) and Backpropagation Neural Network (BPNN) methods to train and learn the multi-field dataset offline, and establishes a fast and high-precision intelligent prediction method for high-speed aerodynamic-thermal-structural multi-field coupling effects.

[0040] (2) The present invention uses a large dataset obtained by simulation, and is based on machine learning methods which are good at processing and analyzing large datasets, and can discover patterns and correlations that are difficult for humans to perceive. The BP network is introduced to train the dataset.

[0041] (3) This invention introduces the POD method to reduce the order of high-dimensional data while ensuring data characteristics. The BP network trains the reduced-order data, which greatly reduces the training time and memory usage of the model.

[0042] (4) Based on the trained intelligent prediction model, the present invention inputs the flight trajectory of the aircraft and realizes the prediction output of the aerodynamic force, aerothermal force, temperature, stress and displacement of the aircraft's characteristic points within seconds. Attached Figure Description

[0043] Figure 1 This is a flowchart of the multi-field coupled intelligent prediction process of the present invention.

[0044] Figure 2 This is a flowchart of the aircraft temperature field calculation process according to the present invention;

[0045] Figure 3 This is a flowchart of the structural stress-displacement calculation for the present invention.

[0046] Figure 4 This is a schematic diagram of the feature point prediction data and simulation data of the present invention;

[0047] Figure 5 This is a schematic diagram illustrating the relative error between the feature point prediction data and the simulation data of this invention;

[0048] Figure 6 This is a schematic diagram illustrating the predicted delay time of the present invention. Detailed Implementation

[0049] The present invention will be further described below with reference to the embodiments.

[0050] This invention addresses the problem of high-speed aerodynamic-thermal-structural multi-field coupling and proposes an intelligent prediction method for the high-speed aerodynamic-thermal-structural multi-field coupling effect, aiming to reduce the online prediction time of the aircraft while ensuring computational accuracy. This method uses an artificial intelligence algorithm to learn and model the simulated aerodynamic-thermal-structural coupling data offline, and then quickly predicts and outputs the changes in the aircraft's feature point data along the flight trajectory from the input flight trajectory.

[0051] Intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft, such as Figure 1 As shown, the specific steps include the following:

[0052] Step 1: Based on the aircraft's flight trajectory, extract sample parameters for calculating aerodynamic and aerothermal states.

[0053] The sample parameters were extracted using a uniform sampling method. The sample parameters included the aircraft's wall temperature, flight altitude, angle of attack, and Mach number.

[0054] Step 2: Model the aircraft and, based on the sample parameters extracted in Step 1, simulate and calculate aerodynamic and aerothermal datasets. The method for simulating and calculating aerodynamic and aerothermal forces is as follows:

[0055] The aerodynamic mesh is divided into the aircraft model; the sample parameters are used as boundary conditions; numerical simulation is performed to obtain aerodynamic and aerothermal datasets.

[0056] Step 3: Calculate the aircraft temperature field along the trajectory based on the aerothermal dataset from Step 2.

[0057] like Figure 2 The method for calculating the temperature field of the aircraft along the trajectory is as follows:

[0058] S31, obtained through interpolation calculation of aerodynamic thermal dataset. tn aerodynamic heat at a specific moment.

[0059] S32, Settings t0 Given that the initial structural temperature field and radiation boundary conditions of the spacecraft are known at any given moment; and that the thermal load is known based on the spacecraft's flight trajectory; the results are obtained iteratively. tn Temperature field of the initial structure of the spacecraft at any given moment.

[0060] S33, Applying S31 to the aircraft model tn Nodal aerodynamic heat and radiation boundary conditions at time points tn The initial structural temperature field of the spacecraft was obtained by transient structural heat transfer analysis using NASTRAN software. tn+1 The structural temperature field at time t.

[0061] Iterative acquisition tn The method for determining the initial structural temperature field of the spacecraft is as follows:

[0062] exist t0 Based on the initial structural temperature field of the spacecraft at any given time, the thermal load is input as the load variable of the spacecraft model, and the results are obtained iteratively. tn Temperature field of the initial structure of the spacecraft at any given moment.

[0063] S34, from tn+1 The surface temperature of the spacecraft in the structural temperature field at time t is used as tn+1 The nodal aerodynamic heat at time will tn+1 The structural temperature field at time t is used as tn+1 Get the initial temperature field of the spacecraft structure at any time; repeat steps S31-S33 to obtain the temperature field of the spacecraft along the trajectory.

[0064] Step 4: Train the aerodynamic dataset from Step 2 to obtain the aerodynamic prediction model.

[0065] The method for training the aerodynamic dataset from step two is as follows:

[0066] S41. The POD method is used to reduce the order of aerodynamic data.

[0067] S42. Based on the mapping relationship between the sample parameters and the reduced aerodynamic data, a BP neural network is used for training.

[0068] S43. Check whether the trained aerodynamic prediction model meets the requirements using test set data. If it does, output the final result; if it does not, modify the reduction parameters and network parameters, retrain until it meets the design requirements, and save the qualified aerodynamic prediction model.

[0069] Step 5: Based on the aircraft temperature field along the trajectory in Step 3 and the aerodynamic prediction model in Step 4, calculate the structural stress-displacement field data of the aircraft model.

[0070] like Figure 3 As shown, the method for calculating the structural stress displacement of the aircraft model is as follows:

[0071] S51. Based on the aerodynamic prediction model, calculate... tn Aerodynamic forces at specific time points. In step S51, t is obtained. n The method for determining the aerodynamic forces at specific time points is as follows:

[0072] Will tn The altitude, angle of attack, and Mach number of the aircraft at any given time are used as inputs to the aerodynamic prediction model, and the output is t. n Aerodynamic forces at specific moments.

[0073] S52, will tn At each time point, the aerodynamic forces and the temperature field of the aircraft along the trajectory in step three are applied to the aircraft model. The structural stress and displacement field are obtained by structural strength analysis using NASTRAN software.

[0074] S53, Enter tn+1 Cycle through steps S51-S52 to obtain stress and displacement field data of the aircraft structure along the trajectory.

[0075] Step 6: Obtain the feature point dataset.

[0076] The method for obtaining the feature point dataset is as follows:

[0077] In the aerodynamic dataset of step two, the temperature field of the aircraft in step three, and the structural stress-displacement field data of the aircraft model in step five, the corresponding data of the aircraft feature points of interest are extracted to obtain the feature point dataset.

[0078] Step 7: Train the feature point dataset to obtain accurate prediction results.

[0079] The method for training on the feature point dataset is as follows:

[0080] S71. Use the POD method to reduce the order of the feature point dataset.

[0081] S72. A BP neural network is used to train the mapping relationship between the flight trajectory and the reduced feature point dataset.

[0082] S73. Check whether the trained intelligent prediction model meets the requirements using test set data. If it does, output the final result; if it does not, modify the reduction parameters and network parameters, retrain until it meets the design requirements, and save the intelligent prediction model that meets the requirements.

[0083] By inputting the test trajectory through an intelligent prediction model, the intelligent prediction program obtains the time-varying curves of aerodynamic heat, aerodynamic force, temperature, displacement, and stress data of the feature nodes, as shown in the figure. Figure 4 , Figure 5 As shown in the figure, the prediction results of the intelligent prediction model deviate little from the simulation data, with the maximum relative error not exceeding 10%. Statistics on 100 prediction times are as follows. Figure 6 As shown, the longest prediction time is 0.6423s, the shortest time is 0.3376s, and the average prediction time is 0.3975s. The prediction delay is relatively short.

[0084] This invention, based on aerodynamic-thermal-structural multi-field datasets obtained through simulation, employs Proper Orthogonal Decomposition (POD) and Backpropagation Neural Network (BPNN) methods for offline training of the multi-field datasets, establishing a fast and high-precision intelligent prediction method for high-speed aerodynamic-thermal-structural multi-field coupling effects. Unlike traditional high-speed aircraft multi-field coupling effect prediction, this invention leverages the large datasets obtained from simulations. Machine learning methods excel at processing and analyzing large-scale datasets, enabling the discovery of patterns and correlations that are difficult for humans to perceive. A BP network is introduced to train the dataset. Furthermore, considering that the data obtained from aircraft simulation calculations are high-dimensional, direct training would be memory-intensive and time-consuming. By introducing the POD method, the high-dimensional data is reduced in order while preserving its characteristics. The BP network is then used to train the reduced-dimensional data, significantly reducing training time and memory consumption. Based on the trained intelligent prediction model, this invention, by inputting the aircraft's flight trajectory, can predict and output the aerodynamic forces, aerothermal effects, temperature, stress, and displacement of the aircraft's feature points within seconds.

[0085] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solutions of the present invention by utilizing the methods and techniques disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention shall fall within the protection scope of the technical solutions of the present invention.

Claims

1. A smart prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft, characterized by: include: Step 1: Based on the aircraft's flight trajectory, extract sample parameters for calculating aerodynamic and aerothermal states; Step 2: Model the aircraft and simulate and calculate aerodynamic and aerothermal datasets based on the sample parameters extracted in Step 1. Step 3: Based on the aerothermal dataset from Step 2, calculate the aircraft temperature field along the trajectory; Step 4: Train the aerodynamic dataset from Step 2 to obtain the aerodynamic prediction model; In step four, the method for training the aerodynamic dataset from step two is as follows: S41. Use the POD method to reduce the order of aerodynamic data; S42. Based on the mapping relationship between sample parameters and reduced aerodynamic data, a BP neural network is used for training to obtain an aerodynamic prediction model. Step 5: Based on the aircraft temperature field along the trajectory in Step 3 and the aerodynamic prediction model in Step 4, calculate the structural stress-displacement field data of the aircraft model. Step 6: Obtain the feature point dataset; In step six, the method for obtaining the feature point dataset is as follows: In the aerodynamic dataset of step two, the temperature field of the aircraft in step three, and the structural stress-displacement field data of the aircraft model in step five, the corresponding data of the aircraft feature points of interest are extracted to obtain the feature point dataset. Step 7: Train the feature point dataset to obtain accurate prediction results.

2. The intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft according to claim 1, characterized in that: In step one, a uniform sampling method is used to extract sample parameters; the sample parameters include the aircraft's wall temperature, flight altitude, angle of attack, and Mach number.

3. The intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft according to claim 1, characterized in that: In step two, the method for simulating and calculating aerodynamic forces and aerothermal activity is as follows: The aerodynamic mesh is divided into the aircraft model; the sample parameters are used as boundary conditions; numerical simulation is performed to obtain aerodynamic and aerothermal datasets.

4. The intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft according to claim 1, characterized in that: In step three, the method for calculating the aircraft temperature field along the trajectory is as follows: S31. t is obtained by interpolation calculation using aerodynamic thermal dataset. n At any given moment, the aerodynamic heat; S32. Set the initial structural temperature field and radiation boundary conditions of the spacecraft at time t0 to be known; based on the spacecraft's flight trajectory, the thermal load is known; iterate to obtain t n Temperature field of the initial structure of the spacecraft at any given moment; S33, Apply S31's t to the aircraft model n Nodal aerodynamic heat, radiation boundary conditions at time t n The initial structural temperature field of the spacecraft at time t was obtained by transient structural heat transfer analysis using NASTRAN software. n+1 The structural temperature field at any given time; S34, from t n+1 The surface temperature of the spacecraft in the structural temperature field at time t is taken as t n+1 The nodal aerodynamic heat at time t n+1 The structural temperature field at time t n+1 Get the initial temperature field of the spacecraft structure at any time; repeat steps S31-S33 to obtain the temperature field of the spacecraft along the trajectory.

5. The intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft according to claim 4, characterized in that: In S33, t is obtained iteratively. n The method for determining the initial structural temperature field of the spacecraft is as follows: Based on the initial structural temperature field of the spacecraft at time t0, the thermal load is input as the load variable of the spacecraft model, and the temperature at time t is obtained iteratively. n Temperature field of the initial structure of the spacecraft at any given moment.

6. The intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft according to claim 1, characterized in that: In step five, the method for calculating the structural stress displacement of the aircraft model is as follows: S51. Calculate t based on the aerodynamic prediction model. n Aerodynamic forces at specific moments; S52, t n At each time point, the aerodynamic forces and the temperature field of the aircraft along the trajectory in step three are applied to the aircraft model. The structural stress and displacement field are obtained by structural strength analysis using NASTRAN software. S53, Enter t n+1 Cycle through steps S51-S52 to obtain stress and displacement field data of the aircraft structure along the trajectory.

7. The intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft according to claim 6, characterized in that: In S51, t is obtained. n The method for determining the aerodynamic forces at specific time points is as follows: t n The altitude, angle of attack, and Mach number of the aircraft at any given time are used as inputs to the aerodynamic prediction model, and the output is t. n Aerodynamic forces at specific moments.

8. The intelligent prediction method for multi-field coupling of aerodynamics, thermodynamics, and structure of high-speed aircraft according to claim 1, characterized in that: In step seven, the method for training the feature point dataset is as follows: S71. Use the POD method to reduce the order of the feature point dataset; S72. A BP neural network is used to train the mapping relationship between the flight trajectory and the reduced feature point dataset.

Images