Rocket pulsating pressure prediction method based on data fusion
By designing rocket models of different scales, conducting wind tunnel tests and aerodynamic simulations, fusing data and using long-short-term memory neural networks and active subspace mapping technology, we have solved the problem of difficulty predicting the pulsating pressure in the transonic flight section of the rocket and achieved Greater prediction accuracy and flight safety.
Patent Information
- Application Number
- CN202411503901.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-25
- Publication Date
- 2025-01-24
AI Technical Summary
Traditional methods cannot accurately predict the pulsating pressure of the rocket's transsonic flight segment, resulting in structural safety and flight stability being affected.
By designing rocket models with different shrinkage size, wind tunnel tests and pneumatic simulation, fusion test data and simulation data, use long and short -term memory neural networks and active sub -space mapping technology, and build a pulsating pressure predictive model.
The prediction accuracy of the pulsating pressure of the rocket's full-scale flight test is improved, and the safety of the rocket flight is enhanced.
Smart Images
Figure CN119358143A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a rocket pulsation pressure prediction method based on data fusion, belonging to the technical field of pneumatics. Background Art
[0002] During the transonic flight of a launch vehicle, the pulsating pressure problem will have a significant impact on the rocket's structural safety and flight stability. However, due to the significant differences between the sky and the earth in the pulsating pressure problem, the traditional research method based on ground wind tunnel test data cannot accurately predict the pulsating pressure in actual flight. Therefore, it is of great significance to effectively integrate ground wind tunnel test data with simulation data to improve the prediction accuracy of pulsating pressure in full-scale rocket flight tests. Summary of the invention
[0003] The technical problem solved by the present invention is: to overcome the shortcomings of the prior art and propose a rocket pulsation pressure prediction method based on data fusion, which accurately predicts the pulsation pressure characteristics of the rocket's transonic flight section in the design stage by fusing the wind tunnel test data of the scaled model and the simulation data of the scaled model and the full-size model.
[0004] The technical solution of the present invention is:
[0005] A rocket pulsation pressure prediction method based on data fusion, comprising:
[0006] S1: Design several rocket models of different scales based on the rocket model to be predicted within the use range of the wind tunnel equipment;
[0007] S2: For each scaled-down rocket model, a corresponding wind tunnel test model is manufactured, and a pulsating pressure wind tunnel test is carried out. The pulsating pressure time series data of the pulsating pressure sensor installed at each set position under each flight profile condition is recorded to form a test data set including the scaled-down size of the model, wind tunnel test conditions and pulsating pressure time series data;
[0008] Construct a rocket aerodynamic simulation model, simulate and obtain the flow field of rockets of various scaled sizes, record the pressure time series data at the same position as the pulsating pressure sensor, and obtain a simulation data set containing simulation conditions and pulsating pressure time series data;
[0009] S3: extracting features from the data of the test data set and the data of the simulation data set respectively, wherein the feature data include the pulsating pressure coefficient at the location of the pulsating pressure sensor and the pulsating pressure spatial correlation function at the locations of different pulsating pressure sensors, to form a test feature data set and a simulation feature data set;
[0010] S4: performing active subspace mapping on the feature data of the simulation feature data set to form a mapping of the simulation condition space to the low-dimensional active subspace;
[0011] S5: Construct a long short-term memory neural network, use the mapping relationship between the active subspace vector and the simulation condition as the mapping from the input layer of the long short-term memory neural network to the first layer of the intermediate layer, define the simulation data set as the input, and define the test data set as the output;
[0012] S6: training a long short-term memory neural network using the test data set and simulation data set obtained in step S2 to obtain a model for predicting the test pulsating pressure data based on the simulation pulsating pressure data;
[0013] S7: For the rocket model to be predicted, simulate and obtain the full-scale rocket flow field, record the pulsating pressure timing data at the same position as the pulsating pressure sensor, obtain simulation data including simulation conditions and pulsating pressure timing data, input the simulation data into the model for predicting the test pulsating pressure data based on the simulation pulsating pressure data, and obtain the full-scale rocket flight test pressure timing data.
[0014] Furthermore, in step S4, active subspace mapping is performed on the feature data of the simulation feature data set, and the specific method is as follows:
[0015] Construct the partial derivative matrix of the characteristic data variables to the simulation condition variables:
[0016]
[0017] Where m is the total number of simulation states considering different model scales and working conditions; φ is the characteristic data; are simulation conditions, including model scale L, Mach number Ma, flight angle of attack α, sideslip angle β, roll angle γ, pressure p and temperature T;
[0018] Perform singular value decomposition on the partial derivative matrix to obtain the eigenvectors d1, d2, …, d p Span the active subspace to form a mapping from the high-dimensional simulation condition space to the low-dimensional active subspace. The specific formula is:
[0019]
[0020] in, is the simulation condition of the active subspace after mapping.
[0021] Furthermore, the partial derivative matrix is decomposed into singular values:
[0022]
[0023] Where A is the skew variance matrix, Λ is the singular value matrix, and D is the right characteristic matrix;
[0024]
[0025] Λ1 and Λ2 are block diagonal matrices, and the specific formula is:
[0026] Λ1=diag(λ1,λ2,…,λ p )
[0027] Λ2=diag(λ p+1 ,…,λ m )
[0028] The basis of the block is λ p >>λ p+1 , λ is a singular value, and the expression of D1 is:
[0029] D1=[d1,d2,…,d p ]
[0030] That is, we get the eigenvectors d1, d2, …, d p Span the active subspace.
[0031] Furthermore, the test data set and simulation data set obtained in step S2 are randomly divided into a training set and a validation set; based on the data of the training set, a long short-term memory neural network is trained to obtain a model for predicting the test pulsating pressure data based on the simulated pulsating pressure data; the model is verified based on the data of the validation set, and if the maximum deviation between the pulsating pressure characteristic data output by the model and the corresponding test characteristic data obtained in step S3 exceeds a set threshold, the model is continuously optimized and the optimized model is verified again.
[0032] Furthermore, the set threshold is 10% to 20%.
[0033] Furthermore, in step S2, a Kalman filtering method is used to perform noise reduction processing on the obtained pulsating pressure time series data, and then a test data set including the model scale size, wind tunnel test conditions and pulsating pressure time series data is formed.
[0034] Furthermore, for a simulation data set including simulation conditions and pulsating pressure time series data, the simulation conditions include model scale, Mach number, flight angle of attack, sideslip angle, roll angle, pressure and temperature.
[0035] Furthermore, a rocket aerodynamic simulation model is constructed based on the large eddy simulation method, and the flow field of the rocket is simulated and obtained using the computational fluid dynamics method.
[0036] A computer-readable storage medium stores a computer program, and the computer program implements the steps of the above method when executed by a processor.
[0037] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.
[0038] The advantages of the present invention compared with the prior art are:
[0039] (1) The present invention solves the problem of difficulty in predicting the pulsating pressure of a rocket in full-scale flight state caused by the difference between heaven and earth by using a data fusion method based on a long short-term memory neural network. Compared with the traditional research method that does not consider the difference between heaven and earth, the present invention can more accurately predict the pulsating pressure of the actual flight of the rocket, which is beneficial to improving the safety of rocket flight.
[0040] (2) The present invention proposes a long short-term memory neural network with active subspace mapping. Aiming at the problem of rocket pulsation pressure prediction, simulations of different model scales and different working conditions are carried out to form a data set containing simulation conditions and pulsation pressure time series data. Feature extraction is performed on the data in the data set to form a feature data set. Active subspace mapping is performed on the data of the feature data set. The obtained low-dimensional mapping is used as the mapping from the input layer of the long short-term memory neural network to the first layer of the intermediate layer, thereby improving the physical constraints and interpretability of the neural network and improving the model confidence. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:
[0042] Figure 1 This is a flow chart of a rocket pulsation pressure prediction method based on data fusion according to an embodiment of the present invention;
[0043] Figure 2 A flow chart of a method for obtaining test data and simulation data of each scaled model according to an embodiment of the present invention;
[0044] Figure 3 A flow chart of a model method for predicting experimental pulsating pressure data based on simulated pulsating pressure data is established in accordance with an embodiment of the present invention. DETAILED DESCRIPTION
[0045] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0046] The present invention proposes a rocket pulsation pressure prediction method based on data fusion, such as Figure 1 As shown, specifically including:
[0047] 1) The length of the rocket is L0. Within the use range of the wind tunnel equipment, the number of rocket models with scaled sizes {L1, L2, …, L n}.
[0048] 2) For each scaled size, a wind tunnel test model is manufactured and a pulsating pressure wind tunnel test is carried out, such as Figure 2 As shown in the figure, the pulsating pressure time series data of the pulsating pressure sensor under various flight profile conditions are recorded, and the test data are denoised using the Kalman filter method to obtain a test data set {M1, M2, …, M n}.
[0049] 3) Based on the large eddy simulation method, a rocket aerodynamic simulation model is constructed. The flow field of the rocket of each scaled size is simulated and obtained using the computational fluid dynamics method. The flow field pressure time series data at the same position as the pulsating pressure sensor is recorded to obtain a data set containing simulation conditions and pressure time series data {S1, S2, …, S n}.
[0050] 4) Extract features from the wind tunnel test data. The feature data include the pulsating pressure coefficient at the location of the pulsating pressure sensor and the pulsating pressure spatial correlation function at different pulsating pressure sensor locations to form a wind tunnel test feature data set {CM1, CM2, …, CM n}.
[0051] 5) Extract features from the simulation data. The feature data also includes the pulsating pressure coefficient at the location of the pulsating pressure sensor and the pulsating pressure spatial correlation function at different pulsating pressure sensor locations to form a simulation feature data set {CS1, CS2, …, CS n}.
[0052] 6) Perform active subspace mapping on the feature data of the simulation feature data set. The specific method is: construct a partial derivative matrix of the feature data variable to the simulation condition variable. The specific formula is as follows:
[0053]
[0054] Where m is the total number of simulation states considering different model scales and working conditions, φ is the characteristic data, is the simulation condition, including model scale L, Mach number Ma, flight angle of attack α, sideslip angle β, roll angle γ, pressure p and temperature T. The specific formula is as follows:
[0055]
[0056] The partial derivative matrix is decomposed into singular values. The specific formula is:
[0057]
[0058] A=DΛD T
[0059] Where A is the skew variance matrix, Λ is the singular value matrix, and D is the right eigenvalue matrix. The singular value matrix and the right eigenvalue matrix are divided into blocks according to the size of the singular values. The specific formula is:
[0060]
[0061] Where Λ1 and Λ2 are block diagonal matrices, the specific formula is:
[0062] Λ1=diag(λ1,λ2,…,λ p )
[0063] Λ2=diag(λ p+1 ,…,λ m )
[0064] The basis of the block is λ p >>λ p+1 , λ is a singular value, and the expression of D1 is:
[0065] D1=[d1,d2,…,d p ]
[0066] Eigenvectors d1,d2,…,d p Span the active subspace to form a mapping from the high-dimensional simulation condition space to the low-dimensional active subspace. The specific formula is:
[0067]
[0068] where ζ is the simulation condition of the active subspace after mapping.
[0069] 7) Construct a long short-term memory neural network, such as Figure 3 As shown in the figure, the mapping relationship between the active subspace vector and the simulation condition is used as the mapping from the neural network input layer to the first layer of the intermediate layer (only acting on the simulation condition of the model input), and the simulation data set {S1, S2, …, S n} as the input of the model, and the experimental data set {M1,M2,…,M n} as the output of the model. According to the scaled size and working conditions, the input and output are randomly divided into the training set, validation set and test set of the long short-term memory neural network. The neural network is trained to obtain a model for predicting the experimental pulsating pressure data based on the simulated pulsating pressure data. The output of the model is tested by the validation set and the test set. The pulsating pressure characteristic data output by the model is obtained by feature extraction. If the pulsating pressure characteristic data output by the model is consistent with the characteristic data of the validation set {CM1, CM2, …, CM n If the maximum deviation of} exceeds 15%, continue to optimize the model;
[0070] 8) Simulate and obtain the full-scale rocket flow field, and obtain the pulsating pressure time series data S0 at the position of the pulsating pressure sensor;
[0071] 9) Through a model that predicts test pulsating pressure data based on simulated pulsating pressure data, the pressure time series data S0 at the position of the pulsating pressure sensor of the full-scale rocket is taken as input to obtain the flight test pressure time series data M0 of the pulsating pressure sensor of the full-scale rocket.
[0072] The present invention solves the problem of difficulty in predicting the pulsating pressure of a rocket in full-scale flight state caused by the difference between heaven and earth through a data fusion method based on a long short-term memory neural network. Compared with traditional research methods that do not consider the difference between heaven and earth, the present invention can more accurately predict the pulsating pressure of a rocket in actual flight, which is beneficial to improving the safety of rocket flight.
[0073] The above-described embodiments are only preferred specific implementations of the present invention. Common changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included in the protection scope of the present invention.
Claims
1. A rocket pulsation pressure prediction method based on data fusion, characterized in that: include: S1: Design several rocket models of different scales based on the rocket model to be predicted within the use range of the wind tunnel equipment; S2: For each scaled-down rocket model, a corresponding wind tunnel test model is manufactured, and a pulsating pressure wind tunnel test is carried out. The pulsating pressure time series data of the pulsating pressure sensor installed at each set position under each flight profile condition is recorded to form a test data set including the scaled-down size of the model, wind tunnel test conditions and pulsating pressure time series data; Construct a rocket aerodynamic simulation model, simulate and obtain the flow field of rockets of various scaled sizes, record the pressure time series data at the same position as the pulsating pressure sensor, and obtain a simulation data set containing simulation conditions and pulsating pressure time series data; S3: extracting features from the data of the test data set and the data of the simulation data set respectively, wherein the feature data include the pulsating pressure coefficient at the location of the pulsating pressure sensor and the pulsating pressure spatial correlation function at the locations of different pulsating pressure sensors, to form a test feature data set and a simulation feature data set; S4: performing active subspace mapping on the feature data of the simulation feature data set to form a mapping of the simulation condition space to the low-dimensional active subspace; S5: Construct a long short-term memory neural network, use the mapping relationship between the active subspace vector and the simulation condition as the mapping from the input layer of the long short-term memory neural network to the first layer of the intermediate layer, define the simulation data set as the input, and define the test data set as the output; S6: training a long short-term memory neural network using the test data set and simulation data set obtained in step S2 to obtain a model for predicting the test pulsating pressure data based on the simulation pulsating pressure data; S7: For the rocket model to be predicted, simulate and obtain the full-scale rocket flow field, record the pulsating pressure timing data at the same position as the pulsating pressure sensor, obtain simulation data including simulation conditions and pulsating pressure timing data, input the simulation data into the model for predicting the test pulsating pressure data based on the simulation pulsating pressure data, and obtain the full-scale rocket flight test pressure timing data.
2. The rocket pulsation pressure prediction method based on data fusion according to claim 1 is characterized in that: In step S4, active subspace mapping is performed on the feature data of the simulation feature data set, and the specific method is as follows: Construct the partial derivative matrix of the characteristic data variables to the simulation condition variables: Where m is the total number of simulation states considering different model scales and working conditions; φ is the characteristic data; are simulation conditions, including model scale L, Mach number Ma, flight angle of attack α, sideslip angle β, roll angle γ, pressure p and temperature T; Perform singular value decomposition on the partial derivative matrix to obtain the eigenvectors d1, d2, …, d p Span the active subspace to form a mapping from the high-dimensional simulation condition space to the low-dimensional active subspace. The specific formula is: Among them, ζ is the simulation condition of the active subspace after mapping.
3. The rocket pulsation pressure prediction method based on data fusion according to claim 2 is characterized in that: Perform singular value decomposition on the partial derivative matrix: Where A is the skew variance matrix, Λ is the singular value matrix, and D is the right characteristic matrix; D=[D1,D2] Λ1 and Λ2 are block diagonal matrices, and the specific formula is: Λ1=diag(λ1,λ2,…,λ p ) Λ2=diag(λ p+1 ,…,l m ) The basis of the block is λ p >>λ p+1 , λ is a singular value, and the expression of D1 is: D1=[d1,d2,…,d p ] That is, we get the eigenvectors d1, d2, …, d p Span the active subspace.
4. The rocket pulsation pressure prediction method based on data fusion according to claim 1 is characterized in that: The test data set and simulation data set obtained in step S2 are randomly divided into a training set and a validation set; based on the data of the training set, a long short-term memory neural network is trained to obtain a model for predicting the test pulsating pressure data based on the simulated pulsating pressure data; the model is verified based on the data of the validation set, and if the maximum deviation between the pulsating pressure characteristic data output by the model and the corresponding test characteristic data obtained in step S3 exceeds a set threshold, the model is continuously optimized and the optimized model is verified again.
5. The rocket pulsation pressure prediction method based on data fusion according to claim 4 is characterized in that: The set threshold is 10% to 20%.
6. The rocket pulsation pressure prediction method based on data fusion according to claim 1 is characterized in that: In step S2, the Kalman filtering method is used to perform noise reduction processing on the obtained pulsating pressure time series data, and then a test data set including the model scale size, wind tunnel test conditions and pulsating pressure time series data is formed.
7. The rocket pulsation pressure prediction method based on data fusion according to claim 1 is characterized in that: For a simulation data set including simulation conditions and pulsating pressure time series data, the simulation conditions include model scale, Mach number, flight angle of attack, sideslip angle, roll angle, pressure and temperature.
8. The rocket pulsation pressure prediction method based on data fusion according to claim 1 is characterized in that: A rocket aerodynamic simulation model is constructed based on the large eddy simulation method, and the flow field of the rocket is simulated and obtained using the computational fluid dynamics method.
9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.
10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.