CN116029219B - Aerodynamic heat prediction method, device, equipment and storage medium for aircraft - Google Patents
Aerodynamic heat prediction method, device, equipment and storage medium for aircraft Download PDFInfo
- Publication number
- CN116029219B CN116029219B CN202310166088.3A CN202310166088A CN116029219B CN 116029219 B CN116029219 B CN 116029219B CN 202310166088 A CN202310166088 A CN 202310166088A CN 116029219 B CN116029219 B CN 116029219B
- Authority
- CN
- China
- Prior art keywords
- aerodynamic
- heat
- prediction model
- aircraft
- matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Aerodynamic Tests, Hydrodynamic Tests, Wind Tunnels, And Water Tanks (AREA)
Abstract
The application discloses a method, a device, equipment and a storage medium for aerodynamic heat prediction of an aircraft, relates to the technical field of aerodynamic heat of the aircraft, and comprises the following steps: acquiring flight conditions of the aircraft and appearance characteristics of the aircraft; constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on a convolutional neural network; and inputting the flight condition and the appearance characteristic into a trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result. The pneumatic thermal prediction model is used for directly outputting the predicted pneumatic thermal result, the pneumatic thermal prediction model can be used for rapidly predicting the pneumatic thermal of aircrafts with different shapes, the thought of an image processing technology is used for reference, and the characteristic of weight sharing of the convolutional neural network is utilized, so that the training speed of the model is improved compared with that of a prediction model constructed based on a fully-connected neural network.
Description
Technical Field
The invention relates to the technical field of aerodynamic heat prediction of aircrafts, in particular to an aerodynamic heat prediction method, an aerodynamic heat prediction device, aerodynamic heat prediction equipment and a storage medium of an aircraft.
Background
The pneumatic thermal prediction method which is widely applied at present mainly comprises an engineering estimation method, a wind tunnel test method and CFD high-precision numerical simulation. The engineering estimation method has high speed and low cost, but has poor aerodynamic heat prediction precision for the aircraft with the complex appearance. Wind tunnel tests are also limited by test cycle time and test costs. CFD not only enables full-scale heat flow prediction for an aircraft, but also is relatively low cost compared to the first two methods. However, when a high-precision CFD simulation result is desired, a large amount of calculation time and calculation resources are consumed. On the other hand, the quality of the CFD simulation results is affected to some extent by the quality of the grid. In recent years, with the great improvement of the performance of computer hardware and the great development of machine learning technology, researchers in various fields begin to try to solve some of the conventional problems by using artificial intelligence technology. Researchers in the aerospace field have also begun to seek a more superior method to replace or optimize CFD techniques. Currently, the application of machine learning in the field of fluid mechanics focuses on combining physical knowledge, such as flow field control equations, to construct a neural network (Physics-Informed Neural Network, PINN) driven purely by physical information or driven jointly by physical information and data. However, the speed is slow in the machine learning process by the physical information neural network PINN.
In summary, how to implement the application of the convolutional neural network in machine learning in the pneumatic design field, accurately predict the pneumatic heat of various types of aircrafts under different flight conditions, and improve the efficiency of the aircraft design stage is a technical problem to be solved in the field.
Disclosure of Invention
In view of the above, the present invention aims to provide a method, a device and a storage medium for aerodynamic heat prediction of an aircraft, which can implement the application of a convolutional neural network in machine learning in the field of aerodynamic design, accurately predict aerodynamic heat of various types of aircraft under different flight conditions, and improve the efficiency of the aircraft design stage. The specific scheme is as follows:
in a first aspect, the present application discloses an aircraft aerodynamic thermal prediction method comprising:
acquiring flight conditions of the aircraft and appearance characteristics of the aircraft;
constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on a convolutional neural network;
and inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result.
Optionally, the inputting the flight condition and the appearance feature is before the trained aerodynamic thermal prediction model. Further comprises:
training a pneumatic thermal prediction model by utilizing a pneumatic thermal sampling data sample to determine the hidden layer number, the neuron node number, the activation function and the weight of each network of the pneumatic thermal prediction model, and acquiring the trained pneumatic thermal prediction model.
Optionally, the training the aerodynamic thermal prediction model by using the aerodynamic thermal sampling data sample includes:
carrying out data preprocessing on the aerodynamic heat sampling data sample to obtain an appearance matrix, a flight condition vector and a heat flow matrix with preset sizes, and training the aerodynamic heat prediction model by utilizing the appearance matrix, the flight condition vector and the heat flow matrix with preset sizes; the appearance matrix and the flight condition vector of the preset size are input sample data of the aerodynamic thermal prediction model, and the heat flow matrix is output sample data of the aerodynamic thermal prediction model.
Optionally, the training the aerodynamic thermal prediction model by using the profile matrix of the preset size, the flight condition vector and the heat flow matrix includes:
adding virtual units to the appearance matrix with the preset size and the heat flow matrix to obtain a target appearance matrix and a target heat flow matrix;
and training the aerodynamic thermal prediction model by using the target appearance matrix, the flight condition vector and the target heat flow matrix.
Optionally, before the data preprocessing is performed on the aerodynamic heat sampling data sample, the method further includes:
and obtaining aerodynamic heat sampling data samples of different flight conditions by using CFD numerical simulation calculation.
Optionally, the obtaining aerodynamic heat sampling data samples of different flight conditions by using CFD numerical simulation calculation includes:
and obtaining aerodynamic heat sampling data samples of different appearance characteristics and flight conditions including flight altitude, attack angle and Mach number by using CFD numerical simulation calculation.
Optionally, the aerodynamic heat prediction method of the aircraft further includes:
dividing the pneumatic thermal sampling data sample subjected to data pretreatment according to a preset proportion to obtain a test set;
and testing and evaluating the aerodynamic heat prediction model by using the test set.
In a second aspect, the present application discloses an aircraft aerodynamic thermal prediction device comprising:
the feature acquisition module is used for acquiring the flight condition of the aircraft and the appearance feature of the aircraft;
the model construction module is used for constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on the convolutional neural network;
and the aerodynamic heat prediction module is used for inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result.
In a third aspect, the present application discloses an electronic device comprising:
a memory for storing a computer program;
a processor for executing the computer program to implement the steps of the previously disclosed aircraft aerodynamic heat prediction method.
In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program; wherein the computer program when executed by a processor implements the steps of the previously disclosed method for aerodynamic thermal prediction of an aircraft.
It can be seen that the present application discloses a method for aerodynamic thermal prediction of an aircraft, comprising: acquiring flight conditions of the aircraft and appearance characteristics of the aircraft; constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on a convolutional neural network; and inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result. Therefore, the obtained appearance characteristics of the aircrafts and the flight conditions of the aircrafts are input into the obtained aerodynamic thermal prediction model after the convolutional neural network is trained, the aerodynamic thermal prediction model is used for directly outputting the predicted aerodynamic thermal results, the aerodynamic thermal of the aircrafts with different appearances can be rapidly predicted through the aerodynamic thermal prediction model, the thought of image processing technology is used for reference, and the training speed of the model is improved compared with that of a prediction model constructed based on the fully-connected neural network by utilizing the characteristic of weight sharing of the convolutional neural network.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method of aerodynamic thermal prediction of an aircraft disclosed herein;
FIG. 2 is a schematic diagram of a neural network according to the present disclosure;
FIG. 3 is a flow chart of a particular aircraft aerodynamic thermal prediction method disclosed herein;
FIG. 4 is a schematic view of a truncated cone shaped aircraft reconstruction grid as disclosed herein;
FIG. 5 is a schematic diagram of a rule for adding virtual units disclosed in the present application;
FIG. 6 is a schematic structural view of an aircraft aerodynamic thermal prediction device disclosed in the present application;
fig. 7 is a block diagram of an electronic device disclosed in the present application.
Detailed Description
The following description of the technical solutions in the embodiments of the present application will be made clearly and completely with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The pneumatic thermal prediction method which is widely applied at present mainly comprises an engineering estimation method, a wind tunnel test method and CFD high-precision numerical simulation. The engineering estimation method has high speed and low cost, but has poor aerodynamic heat prediction precision for the aircraft with the complex appearance. Wind tunnel tests are also limited by test cycle time and test costs. CFD not only enables full-scale heat flow prediction for an aircraft, but also is relatively low cost compared to the first two methods. However, when a high-precision CFD simulation result is desired, a large amount of calculation time and calculation resources are consumed. On the other hand, the quality of the CFD simulation results is affected to some extent by the quality of the grid. In recent years, with the great improvement of the performance of computer hardware and the great development of machine learning technology, researchers in various fields begin to try to solve some of the conventional problems by using artificial intelligence technology. Researchers in the aerospace field have also begun to seek a more superior method to replace or optimize CFD techniques. At present, the application of the machine learning method in the field of fluid mechanics focuses on combining physical knowledge, such as a flow field control equation, to construct a neural network driven by physical information alone or driven by physical information and data together. However, the speed is slow in the machine learning process by the physical information neural network PINN.
Therefore, the application provides an aircraft aerodynamic heat prediction scheme, which can realize the application of a convolutional neural network in machine learning in the field of aerodynamic design, accurately predict aerodynamic heat of various types of aircrafts under different flight conditions, and improve the efficiency of an aircraft design stage.
Referring to fig. 1, the embodiment of the invention discloses a aerodynamic heat prediction method of an aircraft, which comprises the following steps:
step S11: the flight condition of the aircraft and the appearance characteristics of the aircraft are obtained.
In this embodiment, the flight condition and the appearance characteristic of the aircraft to be predicted are obtained, and it can be understood that, because the aerodynamic heat of the aircraft is affected by the flight condition of the aircraft and the appearance of the aircraft, when predicting the aerodynamic heat of the aircraft, the flight condition and the appearance characteristic of the aircraft need to be obtained first, and because the appearance characteristics of various aircraft are not nearly identical, the method specifically includes, but is not limited to: the air heat is the air heat generated by the hypersonic aircraft when the hypersonic aircraft flies in the atmosphere, the surrounding air is strongly compressed by the bow shock waves and generates a strong friction effect with the surface of the aircraft, most of the kinetic energy of the aircraft is converted into air heat energy, so that the air temperature rises sharply, the high-temperature air and the surface of the aircraft generate a huge temperature difference, and part of the heat energy is transmitted to the surface of the aircraft through a boundary layer, and the heat flow distribution of the surface of the aircraft is the air heat. It should be noted that the aerodynamic heat generated by the aircraft with different profiles may be different under the same flight condition, so, in order to predict more accurately, it is necessary to extract the profile features from the profile information of the aircraft, which are used as input data of the aerodynamic heat prediction process together with the flight condition, and the flight condition may specifically include, but is not limited to: flight conditions consist of altitude, angle of attack and mach number.
Step S12: and constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on the convolutional neural network.
In this embodiment, a aerodynamic thermal prediction model including a multi-layer network is constructed based on a Convolutional Neural Network (CNN), and the aerodynamic thermal prediction model includes three network layers, an appearance feature extraction network, an incoming flow information extraction network, and a heat flow prediction network, specifically referring to fig. 2, where the incoming flow information extraction network is used to extract a flight condition vector. The pneumatic thermal prediction model constructed at this time is a blank initial model, namely a pneumatic thermal prediction model which is not trained and debugged. The pneumatic thermal prediction model limits the form of a data sampling training set used for training, and the data sampling training set needs to be adjusted to generate the data sampling training set which accords with the input requirement form of the convolutional neural network so as to train the air-based initial model, and a trained pneumatic thermal prediction model is obtained, namely, a 'black box' model capable of directly predicting pneumatic thermal is generated.
Step S13: and inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result.
In this embodiment, the flight condition and the appearance characteristic of the aircraft are input into the trained aerodynamic thermal prediction model, so that the aerodynamic thermal prediction model predicts the aerodynamic heat of the aircraft to obtain a corresponding aerodynamic thermal prediction result, and the specific appearance characteristic and appearance condition of the aircraft are input into the aerodynamic thermal prediction model trained by using the appearance characteristics, the flight condition and the corresponding heat flow distribution of the aircraft to obtain the predicted aerodynamic heat of the aircraft.
In this embodiment, before inputting the flight condition and the profile feature into the trained aerodynamic thermal prediction model, the method further includes: carrying out data preprocessing on the aerodynamic heat sampling data sample to obtain an appearance matrix, a flight condition vector and a heat flow matrix with preset sizes, and training the aerodynamic heat prediction model by utilizing the appearance matrix, the flight condition vector and the heat flow matrix with preset sizes; the appearance matrix and the flight condition vector of the preset size are input sample data of the aerodynamic thermal prediction model, and the heat flow matrix is output sample data of the aerodynamic thermal prediction model. It can be understood that the data preprocessing is performed on the aerodynamic heat sampling data sample, that is, the data preprocessing is performed on the data sampling training set, and the specific preprocessing process is as follows: the form of the pneumatic heat sampling data samples is adjusted to be in accordance with the input requirement of the convolutional neural network, specifically, the appearance characteristic, the flight condition and the corresponding heat flow distribution of each pneumatic heat sampling data sample are obtained, the appearance of the aircraft is processed into an appearance matrix with a preset size, the flight condition is processed into a flight condition vector, the corresponding heat flow distribution is processed into a heat flow matrix, and the heat flow distribution of each pneumatic heat sampling data sample is noted as the pneumatic heat.
In this embodiment, before the data preprocessing is performed on the pneumatic thermal sampling data sample, the method further includes: and obtaining aerodynamic heat sampling data samples of different flight conditions by using CFD numerical simulation calculation. Specifically, aerodynamic thermal sampling data samples of different appearance characteristics and flight conditions including flight altitude, attack angle and Mach number are obtained by CFD numerical simulation calculation. It can be appreciated that aerodynamic thermal numerical simulation by HyFlow in self-lapping NNW series software, the flight conditions include: fly height, angle of attack and Mach number, wherein the flight conditions are set in the range of 20km to 60km, the angle of attack is set in the range of 0 ° to 20 °, and the Mach number is set in the range of 5 to 20. And respectively simulating different appearance characteristics to obtain heat flow results under different flight conditions, wherein the appearance characteristics, the flight conditions and the corresponding heat flow results obtained by simulation are pneumatic heat sampling data samples.
In this embodiment, the flight condition and the profile feature are input before the trained aerodynamic thermal prediction model. Further comprises: training a pneumatic thermal prediction model by utilizing a pneumatic thermal sampling data sample to determine the hidden layer number, the neuron node number, the activation function, the weight and the like of each network of the pneumatic thermal prediction model, and acquiring the trained pneumatic thermal prediction model. In this embodiment, according to the complexity of the training set of aerodynamic heat sampling data and by using the aerodynamic heat sampling data samples in the training set of aerodynamic heat sampling data, the blank initial model is trained, and the number of hidden layers, the number of neuronal nodes, the activation function, the weight and the like of the three networks of the appearance feature extraction network, the incoming flow information extraction network and the heat flow prediction network are determined, so that the training speed of the aerodynamic heat prediction model is improved through the characteristic of weight sharing of the convolutional neural network.
It can be seen that the present application discloses a method for aerodynamic thermal prediction of an aircraft, comprising: acquiring flight conditions of the aircraft and appearance characteristics of the aircraft; constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on a convolutional neural network; and inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result. Therefore, the obtained appearance characteristics of the aircrafts and the flight conditions of the aircrafts are input into the obtained aerodynamic thermal prediction model after the convolutional neural network is trained, the aerodynamic thermal prediction model is used for directly outputting the predicted aerodynamic thermal results, the aerodynamic thermal of the aircrafts with different appearances can be rapidly predicted through the aerodynamic thermal prediction model, the thought of image processing technology is used for reference, and the training speed of the model is improved compared with that of a prediction model constructed based on the fully-connected neural network by utilizing the characteristic of weight sharing of the convolutional neural network.
Referring to fig. 3, an embodiment of the present invention discloses a specific aerodynamic heat prediction method for an aircraft, and compared with the previous embodiment, the present embodiment further describes and optimizes a technical solution. Specific:
step S21: the flight condition of the aircraft and the appearance characteristics of the aircraft are obtained.
Step S22: and constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on the convolutional neural network.
Step S23: and carrying out data preprocessing on the aerodynamic heat sampling data sample to obtain an outline matrix, a flight condition vector and a heat flow matrix with preset sizes.
In this embodiment, the aerodynamic thermal sampling data sample is subjected to appearance feature extraction, where the process of extracting appearance features of an aircraft is as follows, when appearance features of the aircraft to be extracted are truncated cones, an mxn structured grid is reconstructed on the surface of each truncated cone, a distance between each grid point in the reconstructed grid and an x axis is calculated, and normalization processing is performed, so that the following appearance feature matrix can be obtained:
wherein M is the number of circumferential grids, N is the number of grids in the x-axis direction,
representing the matrix value corresponding to the grid point (M, N),
represents the distance of grid point (M, N) to x-axis, +.>
And->
Representing the minimum x-value and the maximum x-value of the outline in the coordinate system, respectively. Since the uniform distribution grid cannot accurately express the regions such as the head, the grid encryption processing is considered to be performed on the key head position, and the grid sparseness processing is performed on the body, as shown in fig. 4. Distributing all the aerodynamic heat sampling data samples according to the same gridThe matrix capable of expressing the appearance characteristics can be obtained by processing, the three-dimensional appearance can be restored according to the matrix, and the size of the appearance characteristic matrix is determined according to the complexity degree of the appearance of the target aircraft in practical application, in this embodiment, the size of the appearance characteristic matrix corresponding to the selected blunt cone appearance is 128×128.
In this embodiment, after the grid reconstruction is completed and the appearance feature matrix is obtained by all the pneumatic thermal sampling data samples, the heat flow matrix corresponding to the appearance feature matrix is interpolated to obtain the heat flow value corresponding to the grid point in the reconstructed grid. The heat flow values of all the points are arranged according to the grid sequence to obtain the following heat flow matrix:
wherein M is the number of circumferential grids, N is the number of grids in the x-axis direction,
the CFD heat flow value corresponding to the grid point (M, N) is represented.
In this embodiment, the flight condition of each aerodynamic thermal sampling data sample is combined into a vector according to [ flight height, mach number, attack angle ], so as to obtain a flight condition vector, which is used as the input of the aerodynamic thermal prediction model with respect to the flight condition.
Step S24: adding virtual units to the appearance matrix with the preset size and the heat flow matrix to obtain a target appearance matrix and a target heat flow matrix; and training the aerodynamic thermal prediction model by using the target appearance matrix, the flight condition vector and the target heat flow matrix to obtain the trained aerodynamic thermal prediction model.
In this embodiment, a virtual unit is added to the profile feature matrix and the heat flow matrix corresponding to each aerodynamic thermal sampling data sample, so as to reduce the edge prediction error. The values of the virtual cells are copied from the neighboring area inside the edge, the specific rule is shown in fig. 5. The sizes of the target outline characteristic matrix and the target heat flow matrix corresponding to the blunt cone outline after adding 6 layers of virtual units are changed from 128×128 to 140×140. And then designing a aerodynamic thermal prediction model according to the preset size of the appearance matrix of the aircraft. And training the model by taking the average absolute percentage error (MAPE) as a loss function, and stopping training after reaching the target precision. And testing the trained aerodynamic thermal prediction model by using a test set, and evaluating the performance of the model according to the prediction error. Since virtual cells are added, the output heat flow prediction matrix size is 140×140, and the really needed effective area is the 128×128 area in the middle of the matrix, the loss function can ignore the edge area and be calculated by the following formula:
wherein S represents the number of the aerodynamic heat sampling data samples in the data sampling training set,
and->
Respectively represent +.>
CFD hot-current values and predicted hot-current values corresponding to grid points (i, j) in the samples.
After the aerodynamic thermal prediction model is trained, predicting the calculation example in the test set by using the aerodynamic thermal prediction model which is well trained and debugged; and (3) examining the prediction error of the test set, evaluating the prediction precision of the aerodynamic thermal prediction model, and counting the prediction error of the statistical model on the test set, wherein the relative error is 2.460621%. The errors of the four profiles are shown in table 1.
TABLE 1
| Appearance of the shape | Global heat flow prediction relative error (%) | Maximum heat flow relative error (%) |
| Blunt cone | 1.7163 | 1.360172 |
| Blunt bipyramid | 2.2697 | 2.031807 |
| Double ellipsoids | 2.5601 | 2.094667 |
| Lifting body | 3.3255 | 2.266938 |
The test set is obtained by dividing the aerodynamic heat sampling data sample subjected to data pretreatment according to a preset proportion; and testing and evaluating the aerodynamic heat prediction model by using the test set. Specifically, all the aerodynamic heat sampling data samples subjected to data preprocessing are randomly divided into a data sampling training set and a test set according to a ratio of 4:1.
Step S25: and inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result.
In the steps S21, S22, and S25, the detailed processing procedure is referred to the above disclosed embodiments, and will not be described herein.
It can be seen that after training is completed, rapid prediction of aerodynamic heat of aircrafts of different shapes can be achieved, and prediction errors on the test set can be within 3%. The virtual units are added to the appearance matrix and the heat flow matrix with preset sizes, so that the prediction error of the edge area is reduced, and the prediction precision is effectively improved.
Referring to fig. 6, the embodiment of the invention also correspondingly discloses a aerodynamic heat prediction device of an aircraft, which comprises:
a feature acquisition module 11 for acquiring a flight condition of the aircraft and an appearance feature of the aircraft;
the model construction module 12 is used for constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on the convolutional neural network;
and the aerodynamic heat prediction module 13 is configured to input the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predict aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model to obtain a corresponding prediction result.
It can be seen that the present application discloses acquiring the flight conditions of an aircraft and the appearance characteristics of the aircraft; constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on a convolutional neural network; and inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result. Therefore, the obtained appearance characteristics of the aircrafts and the flight conditions of the aircrafts are input into the obtained aerodynamic thermal prediction model after the convolutional neural network is trained, the aerodynamic thermal prediction model is used for directly outputting the predicted aerodynamic thermal results, the aerodynamic thermal of the aircrafts with different appearances can be rapidly predicted through the aerodynamic thermal prediction model, the thought of image processing technology is used for reference, and the training speed of the model is improved compared with that of a prediction model constructed based on the fully-connected neural network by utilizing the characteristic of weight sharing of the convolutional neural network.
Further, the embodiment of the present application further discloses an electronic device, and fig. 7 is a block diagram of the electronic device 20 according to an exemplary embodiment, where the content of the figure is not to be considered as any limitation on the scope of use of the present application.
Fig. 7 is a schematic structural diagram of an electronic device 20 according to an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input output interface 25, and a communication bus 26. Wherein the memory 22 is adapted to store a computer program to be loaded and executed by the processor 21 for implementing the relevant steps in the aircraft aero-thermal prediction method disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in the present embodiment may be specifically an electronic computer.
In this embodiment, the power supply 23 is configured to provide an operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and an external device, and the communication protocol to be followed is any communication protocol applicable to the technical solution of the present application, which is not specifically limited herein; the input/output interface 25 is used for acquiring external input data or outputting external output data, and the specific interface type thereof may be selected according to the specific application requirement, which is not limited herein.
Processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 may be implemented in at least one hardware form of DSP (Digital Signal Processing ), FPGA (Field-Programmable Gate Array, field programmable gate array), PLA (Programmable Logic Array ). The processor 21 may also comprise a main processor, which is a processor for processing data in an awake state, also called CPU (Central Processing Unit ); a coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 21 may integrate a GPU (Graphics Processing Unit, image processor) for rendering and drawing of content required to be displayed by the display screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence ) processor for processing computing operations related to machine learning.
The memory 22 may be a carrier for storing resources, such as a read-only memory, a random access memory, a magnetic disk, or an optical disk, and the resources stored thereon may include an operating system 221, a computer program 222, and the like, and the storage may be temporary storage or permanent storage.
The operating system 221 is used for managing and controlling various hardware devices on the electronic device 20 and the computer program 222, so as to implement the operation and processing of the processor 21 on the mass data 223 in the memory 22, which may be Windows Server, netware, unix, linux, etc. The computer program 222 may further comprise a computer program capable of performing other specific tasks in addition to the computer program capable of performing the aircraft aero-thermal prediction method performed by the electronic device 20 as disclosed in any of the preceding embodiments. The data 223 may include, in addition to data received by the electronic device and transmitted by the external device, data collected by the input/output interface 25 itself, and so on.
Further, the application also discloses a computer readable storage medium for storing a computer program; wherein the computer program, when executed by the processor, implements the previously disclosed method of aerodynamic thermal prediction of an aircraft. For specific steps of the method, reference may be made to the corresponding contents disclosed in the foregoing embodiments, and no further description is given here.
In this specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, so that the same or similar parts between the embodiments are referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of functionality in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application. The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software modules may be disposed in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
Finally, it is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above description of the method, the device, the equipment and the storage medium for aerodynamic thermal prediction of an aircraft provided by the invention applies specific examples to illustrate the principles and the implementation of the invention, and the description of the above examples is only used for helping to understand the method and the core idea of the invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.
Claims (7)
1. A method of aerodynamic thermal prediction of an aircraft, comprising:
acquiring flight conditions of the aircraft and appearance characteristics of the aircraft;
constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on a convolutional neural network;
carrying out data preprocessing on the aerodynamic heat sampling data sample to obtain an appearance matrix, a flight condition vector and a heat flow matrix with preset sizes, and training the aerodynamic heat prediction model by utilizing the appearance matrix, the flight condition vector and the heat flow matrix with preset sizes; the appearance matrix and the flight condition vector of the preset size are input sample data of the aerodynamic thermal prediction model, the heat flow matrix is output sample data of the aerodynamic thermal prediction model, so that the number of hidden layers, the number of neuron nodes, an activation function and a weight of each network of the aerodynamic thermal prediction model are determined, and the trained aerodynamic thermal prediction model is obtained;
inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model to obtain a corresponding prediction result;
the training the aerodynamic thermal prediction model by using the appearance matrix with the preset size, the flight condition vector and the heat flow matrix comprises the following steps:
adding virtual units to the appearance matrix with the preset size and the heat flow matrix to obtain a target appearance matrix and a target heat flow matrix; and training the aerodynamic thermal prediction model by using the target appearance matrix, the flight condition vector and the target heat flow matrix.
2. The method of claim 1, wherein prior to data preprocessing the aerodynamically and thermally sampled data samples, further comprising:
and obtaining aerodynamic heat sampling data samples of different flight conditions by using CFD numerical simulation calculation.
3. The method of claim 2, wherein obtaining aerodynamic heat sample data samples for different flight conditions using CFD numerical simulation calculations comprises:
and obtaining aerodynamic heat sampling data samples of different appearance characteristics and flight conditions including flight altitude, attack angle and Mach number by using CFD numerical simulation calculation.
4. A method of aerodynamic thermal prediction of an aircraft according to any one of claims 1 to 3, further comprising:
dividing the pneumatic thermal sampling data sample subjected to data pretreatment according to a preset proportion to obtain a test set;
and testing and evaluating the aerodynamic heat prediction model by using the test set.
5. An aircraft aerodynamic thermal prediction device, comprising:
the feature acquisition module is used for acquiring the flight condition of the aircraft and the appearance feature of the aircraft;
the model construction module is used for constructing a aerodynamic thermal prediction model comprising an appearance feature extraction network, an incoming flow information extraction network and a thermal flow prediction network based on the convolutional neural network;
the model training module is used for carrying out data preprocessing on the aerodynamic heat sampling data sample to obtain an appearance matrix, a flight condition vector and a heat flow matrix with preset sizes, and training the aerodynamic heat prediction model by utilizing the appearance matrix, the flight condition vector and the heat flow matrix with preset sizes; the appearance matrix and the flight condition vector of the preset size are input sample data of the aerodynamic thermal prediction model, the heat flow matrix is output sample data of the aerodynamic thermal prediction model, so that the number of hidden layers, the number of neuron nodes, an activation function and a weight of each network of the aerodynamic thermal prediction model are determined, and the trained aerodynamic thermal prediction model is obtained;
the aerodynamic heat prediction module is used for inputting the flight condition and the appearance characteristic into the trained aerodynamic heat prediction model, and predicting the aerodynamic heat of the aircraft by using the trained aerodynamic heat prediction model so as to obtain a corresponding prediction result;
the model training module is specifically configured to add virtual units to the outline matrix and the heat flow matrix with preset dimensions, so as to obtain a target outline matrix and a target heat flow matrix; and training the aerodynamic thermal prediction model by using the target appearance matrix, the flight condition vector and the target heat flow matrix.
6. An electronic device, comprising:
a memory for storing a computer program;
a processor for executing the computer program to carry out the steps of the aircraft aerodynamic heat prediction method according to any one of claims 1 to 4.
7. A computer-readable storage medium storing a computer program; wherein the computer program when executed by a processor implements the steps of the aircraft aerodynamic heat prediction method according to any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310166088.3A CN116029219B (en) | 2023-02-27 | 2023-02-27 | Aerodynamic heat prediction method, device, equipment and storage medium for aircraft |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310166088.3A CN116029219B (en) | 2023-02-27 | 2023-02-27 | Aerodynamic heat prediction method, device, equipment and storage medium for aircraft |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116029219A CN116029219A (en) | 2023-04-28 |
| CN116029219B true CN116029219B (en) | 2023-07-07 |
Family
ID=86079632
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310166088.3A Active CN116029219B (en) | 2023-02-27 | 2023-02-27 | Aerodynamic heat prediction method, device, equipment and storage medium for aircraft |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116029219B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116227389A (en) * | 2023-05-08 | 2023-06-06 | 中国空气动力研究与发展中心计算空气动力研究所 | Method and device for predicting aerodynamic heat data |
| CN116296239A (en) * | 2023-05-26 | 2023-06-23 | 中国空气动力研究与发展中心计算空气动力研究所 | Wind tunnel test method, device, equipment and medium based on convolutional neural network |
| CN116534246B (en) * | 2023-07-05 | 2023-09-12 | 中国空气动力研究与发展中心计算空气动力研究所 | Flow direction vortex modulation device |
| CN117309079B (en) * | 2023-11-28 | 2024-03-12 | 中国空气动力研究与发展中心计算空气动力研究所 | Ultrasonic flying time measuring method, device, equipment and medium based on time difference method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021103135A1 (en) * | 2019-11-25 | 2021-06-03 | 中国科学院深圳先进技术研究院 | Deep neural network-based traffic classification method and system, and electronic device |
| CN113609596A (en) * | 2021-09-29 | 2021-11-05 | 中国空气动力研究与发展中心计算空气动力研究所 | Aircraft aerodynamic characteristic prediction method based on neural network |
| CN113901594A (en) * | 2021-12-09 | 2022-01-07 | 中国空气动力研究与发展中心计算空气动力研究所 | Intelligent prediction method for aerodynamic thermal environment on surface of aircraft |
| CN114996658A (en) * | 2022-07-20 | 2022-09-02 | 中国空气动力研究与发展中心计算空气动力研究所 | Projection-based hypersonic aircraft aerodynamic heat prediction method |
-
2023
- 2023-02-27 CN CN202310166088.3A patent/CN116029219B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021103135A1 (en) * | 2019-11-25 | 2021-06-03 | 中国科学院深圳先进技术研究院 | Deep neural network-based traffic classification method and system, and electronic device |
| CN113609596A (en) * | 2021-09-29 | 2021-11-05 | 中国空气动力研究与发展中心计算空气动力研究所 | Aircraft aerodynamic characteristic prediction method based on neural network |
| CN113901594A (en) * | 2021-12-09 | 2022-01-07 | 中国空气动力研究与发展中心计算空气动力研究所 | Intelligent prediction method for aerodynamic thermal environment on surface of aircraft |
| CN114996658A (en) * | 2022-07-20 | 2022-09-02 | 中国空气动力研究与发展中心计算空气动力研究所 | Projection-based hypersonic aircraft aerodynamic heat prediction method |
Non-Patent Citations (3)
| Title |
|---|
| " 基于神经网络的气动热关联研究";吴宁 等;《工业控制计算机》;第31卷(第11期);48-49、53 * |
| "Dual-convolutional neural network based aerodynamic prediction and multi-objective optimization of a compact turbine rotor";Yuqi Wang .etc;《Aerospace Science and Technology》;第116卷;1-13 * |
| "基于PIMTLNN的导弹气动特性快速预测技术研究";蔺佳哲 等;《北京航空航天大学学报》;1-15 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116029219A (en) | 2023-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116029219B (en) | Aerodynamic heat prediction method, device, equipment and storage medium for aircraft | |
| Dasgupta et al. | Simple method of diagnosing fast flavor conversions of supernova neutrinos | |
| CN112016167B (en) | Aircraft aerodynamic shape design method and system based on simulation and optimization coupling | |
| CN110222828B (en) | Unsteady flow field prediction method based on hybrid deep neural network | |
| CN102789539B (en) | Optimization design method for air vehicle lifting surface structure | |
| CN105115692B (en) | The pneumatic Forecasting Methodology that a kind of CFD numerical simulations are combined with wind tunnel test | |
| CN108959794A (en) | A kind of structural frequency response modification methodology of dynamics model based on deep learning | |
| CN103902782A (en) | POD (proper orthogonal decomposition) and surrogate model based order reduction method for hypersonic aerodynamic thermal models | |
| Wang et al. | Magnetohydrodynamics (MHD) numerical simulations on the interaction of the solar wind with the magnetosphere: A review | |
| CN115618498B (en) | Prediction method, device, equipment and medium for cross-basin flow field of aircraft | |
| CN112597702B (en) | Pneumatic modeling generation type confrontation network model training method based on radial basis function | |
| CN113777931B (en) | Icing wing type pneumatic model construction method, device, equipment and medium | |
| CN108536912A (en) | A kind of method that Transmission Tower mechanical analysis and its App make | |
| CN103885867B (en) | Online evaluation method of performance of analog circuit | |
| He et al. | Aerodynamic optimisation of a high-speed train head shape using an advanced hybrid surrogate-based nonlinear model representation method | |
| Peters et al. | On the construction of a mode based reduced order model for a moving store | |
| CN117933095B (en) | Earth surface emissivity real-time inversion and assimilation method based on machine learning | |
| CN115293424A (en) | New energy maximum power generation capacity calculation method, terminal and storage medium | |
| CN115099161A (en) | Low-scattering carrier shape construction method | |
| CN110826208A (en) | Pneumatic numerical simulation accelerated convergence method | |
| CN114638048A (en) | Three-dimensional spray pipe flow field rapid prediction and sensitivity parameter analysis method and device | |
| CN116718842B (en) | Electromagnetic environment monitoring airspace self-adaptive sampling method and device and computer equipment | |
| CN115470675A (en) | Prediction method and system for crack propagation path of heavy-load pressure vessel | |
| CN113722951B (en) | Scatterer three-dimensional finite element grid optimization method based on neural network | |
| CN109901164B (en) | Distributed back projection imaging method of synthetic aperture radar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |