Measuring point layout rapid determination method and system, storage medium, equipment and program

Through the feature space-based method, the POD method is used to extract low-dimensional features and data statistical features, and the layout of sparse measurement points on the aircraft surface is quickly determined, which solves the problems of slow convergence speed and complex factors in the traditional method, and improves the layout of measurement points. Accuracy and efficiency.

CN118940661APending Publication Date: 2024-11-12NORTHWESTERN POLYTECHNICAL UNIV
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202410984054.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, sparse reconstruction model requires multiple calls to calculate the objective function, resulting in slow convergence speed and the inability to quickly obtain the measurement point layout. In addition, traditional methods cannot fully consider complex factors when optimizing the sensor layout, which affects the reconstruction results. Accuracy and reliability.

Method used

The rapid determination method of measuring point layout based on feature space is adopted, and the low-dimensional characteristics of surface distribution load are extracted through the POD method, the number of sparse measurement points is determined, and the measurement point location is obtained through data statistical feature clustering, and the measurement point layout is quickly determined.

Benefits of technology

It improves the efficiency of optimized layout of measurement points, ensures the accuracy and representativeness of measurement points layout, reduces the error of sparse reconstruction of distributed loads, and is suitable for aircraft design and other similar data structures.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940661A_ABST
    Figure CN118940661A_ABST
Patent Text Reader

Abstract

The invention discloses a measuring point layout rapid determination method and system, a storage medium, equipment and a program, and belongs to the technical field of hydromechanics artificial intelligence. The method includes obtaining an airfoil pressure distribution dataset of a selected airfoil; based on a POD method, processing the airfoil pressure distribution data set to obtain the number of sparse measuring points; and taking the number of the sparse measuring points as a clustering number, selecting the measuring points, obtaining the positions of the sparse measuring points, and completing the confirmation of the layout of the measuring points. According to the method, the low-dimensional features of the surface distribution load are extracted through an intrinsic orthogonal decomposition method, the number of sparse measuring points is determined according to the spatial dimensions of the low-dimensional features, the data statistical features of the surface distribution load are represented through a fourth moment, the features are clustered to obtain the positions of the sparse measuring points, and the measuring point layout is rapidly obtained.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of fluid mechanics and artificial intelligence technology, and specifically relates to a method, system, storage medium, device and program for quickly determining measurement point layout based on feature space. Background Art

[0002] In the field of aircraft design, accurate aerodynamic distributed load data is crucial for verifying the aerodynamic performance of the aircraft, predicting flight status, and optimizing flight control strategies. However, due to technical limitations and cost considerations, it is often impossible to directly deploy a dense sensor network on the surface of the aircraft to obtain detailed aerodynamic data. Therefore, based on the sparse sensor measurements on the aircraft surface, fine-tuning the reconstruction to obtain complete aerodynamic distributed loads has become a challenging and highly practical topic.

[0003] In the implementation of sparse reconstruction technology, two core factors play a decisive role: the construction of the sparse reconstruction model and the optimal arrangement of sparse sensors. The two complement each other and jointly determine the accuracy and reliability of the reconstruction results.

[0004] First of all, the construction of a sparse reconstruction model is the basis of the entire process. Such models are usually built using a data-driven approach, relying on a large amount of known aerodynamic data and sensor measurements for training and learning. The quality of the model directly determines the accuracy of extracting and reconstructing aerodynamic distributed loads from sparse sensor measurements. However, building an efficient and accurate sparse reconstruction model is not easy. It requires a deep understanding of the aerodynamic flow field, as well as advanced mathematical modeling and algorithm design capabilities.

[0005] Secondly, the optimal arrangement of sparse sensors is also an issue that cannot be ignored. The number and location of sensors are crucial to the reconstruction results. Traditional methods are based on rigorous mathematical theories and use optimization algorithms to determine the optimal location and number of sensors. However, these methods are often limited by the one-sidedness of the evaluation criteria and cannot fully consider the impact of various complex factors on the reconstruction results. In addition, due to the high computational complexity and difficulty in solving the problem, traditional methods have certain limitations in practical applications.

[0006] In order to overcome the shortcomings of traditional methods, intelligent optimization algorithms have been introduced to the problem of sparse sensor optimization. These algorithms aim to search for the global optimal value of the objective function and find the best sensor layout through iteration and optimization. However, intelligent optimization algorithms also face some challenges. The selection of the objective function directly determines the accuracy and convergence of the optimization. A suitable objective function should be able to fully reflect the accuracy and reliability of the reconstruction results, while also considering the computational complexity and solution efficiency. However, intelligent optimization algorithms usually need to call the sparse reconstruction model multiple times to calculate the objective function, which leads to a slow convergence problem. Summary of the invention

[0007] In view of the problem that the existing technology usually needs to call the sparse reconstruction model multiple times to calculate the objective function, which leads to slow convergence speed and inability to quickly obtain the measurement point layout. The present invention provides a method for quickly determining the measurement point layout based on feature space, extracts the low-dimensional features of the surface distribution load by the intrinsic orthogonal decomposition method, determines the number of sparse measurement points according to the dimension of the low-dimensional feature space, characterizes the data statistical features of the surface distribution load by the fourth-order moment, clusters the features to obtain the sparse measurement point positions, and quickly obtains the measurement point layout.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] In a first aspect, the present invention provides a method for quickly determining a measurement point layout based on a feature space, comprising:

[0010] Obtaining a data set of airfoil pressure distribution on a selected airfoil;

[0011] Based on the POD method, the airfoil pressure distribution data set is processed to obtain the number of sparse measurement points;

[0012] The number of sparse measuring points is used as the cluster number to select measuring points, obtain the sparse measuring point locations, and complete the confirmation of the measuring point layout.

[0013] As a further improvement of the present invention, the step of obtaining a data set of airfoil pressure distribution of the selected airfoil includes:

[0014] Obtaining original flow field data for the selected airfoil;

[0015] According to the original flow field data of the selected airfoil, the airfoil pressure distribution data set is obtained.

[0016] As a further improvement of the present invention, the obtaining of original flow field data for the selected airfoil includes:

[0017] Determine the computational design space of the selected airfoil and obtain the operating parameters using Latin hypercube sampling;

[0018] The obtained working condition parameters are subjected to CFD simulation calculation to obtain the original flow field data.

[0019] As a further improvement of the present invention, the POD-based method is used to process the airfoil pressure distribution data set to obtain the number of sparse measurement points, including:

[0020] Using the POD method to reduce the order of the airfoil pressure distribution data set of the selected airfoil to obtain the basic mode of the low-dimensional space;

[0021] The fundamental mode in the low-dimensional space is truncated according to the energy proportion and reconstruction error to determine the fundamental mode order to be retained;

[0022] The fundamental mode order to be retained will be determined as the number of sparse measurement points.

[0023] As a further improvement of the present invention, the method of selecting measuring points by taking the number of sparse measuring points as the number of clusters to obtain the positions of sparse measuring points includes:

[0024] Calculate the statistical characteristics of the airfoil pressure distribution at different measuring points;

[0025] The number of sparse measurement points is used as the number of clusters to cluster the statistical characteristics of the data and obtain the clustering results;

[0026] In each class in the clustering result, the measurement point closest to the class center is selected, or a measurement point is randomly selected and combined as the sparse measurement point location.

[0027] As a further improvement of the present invention, the number of sparse measurement points is used as the clustering number to select measurement points and obtain sparse measurement point positions. Then, the sparse measurement values ​​in the sparse measurement point positions are used as input; the POD method is used to reduce the order of the wing surface pressure distribution data set to obtain the fundamental mode of the low-dimensional space, and the fundamental mode coefficients in the fundamental mode of the low-dimensional space are used as output to construct a BPNN model and train it.

[0028] In a second aspect, the present invention provides a system for quickly determining a measurement point layout based on a feature space, comprising:

[0029] Obtaining data set module: used to obtain the airfoil pressure distribution data set of the selected airfoil;

[0030] Measuring point number module: used to process the airfoil pressure distribution data set based on the POD method to obtain the sparse measuring point number;

[0031] Measuring point location module: It is used to select measuring points based on the number of sparse measuring points as the clustering number, form sparse measuring point locations, and complete the confirmation of the measuring point layout.

[0032] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of a method for rapidly determining a measurement point layout based on a feature space when executing the computer program.

[0033] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of a method for quickly determining a measurement point layout based on a feature space.

[0034] In a fifth aspect, the present invention provides a computer program product, characterized in that it includes computer instructions, which, when executed by a processor, implement the steps of the method for quickly determining the measurement point layout based on the feature space.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The invention discloses a method for quickly determining the layout of measuring points based on feature space. First, the POD method is used to process the airfoil pressure distribution data set, and the number of sparse measuring points can be quickly determined. This method is based on the statistical characteristics of the data and can capture the main change mode in the data, thereby effectively reducing the data dimension. The number of sparse measuring points obtained by the POD method is used as the cluster number, and the measuring points are clustered and selected, so that a set of better measuring point layouts can be quickly obtained. Compared with the traditional method based on the optimization algorithm, the complex process of repeatedly calling the sparse reconstruction model to calculate the objective function is avoided, and the efficiency of the optimal layout of the measuring points is greatly improved. Secondly, the POD method is a mathematically rigorous data reduction technology that can retain the main information in the data and remove noise and redundant information. Therefore, the number and position of the sparse measuring points determined based on the POD method can more accurately reflect the main characteristics of the airfoil pressure distribution. By clustering the statistical characteristics of the pressure distribution data at different measuring points, and selecting the measuring point closest to the class center in each category or randomly selecting a measuring point as the sparse measuring point, the representativeness and uniformity of the measuring point layout can be ensured, and the accuracy of the measuring point layout is further improved. Finally, the method for quickly determining the measurement point layout provided by the present invention is not only applicable to the optimization problem of airfoil measurement point layout, but can also be extended to other fields with similar data structures and optimization goals, such as aircraft design, wind tunnel testing, etc. By reducing the error of sparse reconstruction of distributed loads, this method can improve the accuracy of subsequent data analysis and processing, and provide basic support for subsequent scientific research and technological applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of a method for quickly determining measurement point layout based on feature space of the present invention;

[0038] Figure 2 It is a flowchart of a method for quickly determining measurement point layout based on feature space in the present invention;

[0039] Figure 3 Schematic diagram of the geometric shape of the airfoil in the present invention;

[0040] Figure 4 is a global grid schematic diagram of the airfoil aerodynamic grid in the present invention;

[0041] Figure 5 It is a schematic diagram of a local grid of an airfoil aerodynamic grid in the present invention;

[0042] Figure 6 A schematic diagram of sampling of airfoil calculation conditions in the present invention;

[0043] Figure 7 The distribution of cluster center measurement points on the airfoil geometry after clustering (a) in the present invention;

[0044] Figure 8 is a histogram of the central pressure distribution of type (b) in the present invention, the abscissa is the value of the airfoil pressure coefficient, and the ordinate is the frequency;

[0045] Fig. 9 The present invention compares the sparse reconstruction results of the wing surface pressure distribution of the test samples under four different working conditions with the CFD calculation results; wherein, (a) is a comparison between the CFD calculation results and the sparse reconstruction results of the wing surface pressure distribution under the test condition of Mach number 0.5476 and angle of attack 0.5357°; (b) is a comparison between the CFD calculation results and the sparse reconstruction results of the wing surface pressure distribution under the test condition of Mach number 0.4851 and angle of attack 1.8714°; (c) is a comparison between the CFD calculation results and the sparse reconstruction results of the wing surface pressure distribution under the test condition of Mach number 0.5683 and angle of attack 3.7124°; (d) is a comparison between the CFD calculation results and the sparse reconstruction results of the wing surface pressure distribution under the test condition of Mach number 0.5220 and angle of attack 3.0629°;

[0046] Fig.10 It is a structural schematic diagram of a system for quickly determining measurement point layout based on feature space of the present invention;

[0047] Fig.11 1 is a structural block diagram of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0049] The following detailed description is an exemplary description, which is intended to provide further detailed description of the present invention. Unless otherwise specified, all technical terms used in the present invention have the same meaning as those generally understood by those skilled in the art to which the present application belongs. The terms used in the present invention are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention.

[0050] In view of the problem that the prior art usually needs to call the sparse reconstruction model multiple times to calculate the objective function, which leads to a slow convergence speed and an inability to quickly obtain the measurement point layout. The present invention provides a method for quickly determining the measurement point layout based on the feature space, such as Figure 1 As shown, the method includes:

[0051] Obtaining a data set of airfoil pressure distribution on a selected airfoil;

[0052] Based on the POD method, the airfoil pressure distribution data set is processed to obtain the number of sparse measurement points;

[0053] The number of sparse measuring points is used as the cluster number to select measuring points, obtain the sparse measuring point locations, and complete the confirmation of the measuring point layout.

[0054] The present invention extracts low-dimensional features of surface distributed loads through the intrinsic orthogonal decomposition method, determines the number of sparse measuring points according to the spatial dimension of the low-dimensional features, characterizes the data statistical features of the surface distributed loads through the fourth-order moment, clusters the features to obtain the sparse measuring point positions, and quickly obtains the measuring point layout.

[0055] The present invention is further explained below in conjunction with the specific drawings.

[0056] like Figure 2 The present invention provides a method for quickly determining the measurement point layout based on feature space, comprising:

[0057] S1: Generate a sample data set of fixed-shape deformation conditions. The specific steps are as follows:

[0058] S1.1: Select a certain type of airfoil and generate the aerodynamic calculation grid of the airfoil;

[0059] S1.2: Determine the computational design space of the airfoil and use Latin hypercube sampling to obtain a series of operating parameters;

[0060] S1.3: For the selected airfoil, perform CFD simulation calculation on the operating parameters obtained in S1.2 to obtain the corresponding original flow field data;

[0061] S1.4: Process the flow field data obtained in S1.3 to obtain a data set of airfoil pressure distribution;

[0062] S2: Determine the number of sparse measurement points by combining the Proper Orthogonal Decomposition (POD) method. The specific steps are as follows:

[0063] S2.1: Use the POD method to reduce the order of the pressure distribution data obtained in S1.4, truncate the fundamental mode in the low-dimensional space according to the energy proportion and reconstruction error, determine the fundamental mode order to be retained, and the corresponding sample data mean, fundamental mode matrix and fundamental mode coefficient;

[0064] S2.2: The fundamental mode order obtained in S2.1 is used as the number of sparse measurement points;

[0065] S3: Determine the initial measurement point location based on the statistical distribution characteristics of the data. The specific steps are as follows:

[0066] S3.1: Calculate the statistical characteristics of the airfoil pressure distribution at different measuring points;

[0067] S3.2: Use the k-means clustering method to cluster the statistical features obtained in S3.1, and the number of clusters is the number of sparse measurement points determined in S2.2;

[0068] S3.3: Select the measurement point closest to the center of each class obtained in S3.2, or randomly select a measurement point, and combine them as the sparse measurement point location;

[0069] S4: Construct a proxy model by combining the back propagation neural network BPNN, the input is the pressure distribution value corresponding to the sparse measurement point position obtained in S3.3, and the output is the fundamental mode coefficient obtained in S2.1;

[0070] S5: Use the POD model in S2 and the BPNN model in S4 for sparse reconstruction of the airfoil pressure distribution load. The specific steps are as follows:

[0071] S5.1: According to the measurement point positions in S3.3, the corresponding sparse measurement values ​​are obtained, and a set of fundamental mode coefficients are predicted using the BPNN model trained in S4;

[0072] S5.2: For the fundamental modal coefficients predicted in S5.1, combined with the fundamental modal matrix saved after POD reduction in S2 and the mean of the sample data, the fundamental modal coefficients are reconstructed back to the original space, that is, the complete airfoil pressure distribution data is obtained.

[0073] In the above S1.1, the certain type of airfoil is a NACA0012 airfoil, the aerodynamic calculation grid is a CH-type structured grid, the number of grid nodes on the upper and lower airfoils is 150, the first layer height of the boundary layer is 5e-6, and the total number of grids is 66554. The pressure far field boundary condition is adopted, and in order to reduce the influence of the far field condition, the far field is located at about 70 chord lengths from the airfoil.

[0074] In the above S1.2, the calculation conditions of the airfoil are the angle of attack and the Mach number. In the subsonic range, the angle of attack ranges from 0 to 5°, and the Mach number ranges from 0.3 to 0.6. The number of samples extracted using Latin hypercube is 300, of which 90% of the samples are used for training and the remaining 10% are used for testing.

[0075] In S1.3, the CFD calculation model adopts the SST k-omega turbulence model.

[0076] In S3.1, the statistical features are the first four moments, namely the mean, standard deviation, skewness, and kurtosis of the data.

[0077] In S4, the sparse measurement values ​​are taken as input and the fundamental modal coefficients are taken as output to construct the BPNN model and perform training.

[0078] In S5, the POD model in S2 and the BPNN model in S4 are used for sparse reconstruction of the airfoil pressure distribution load.

[0079] The present invention is further explained below in conjunction with specific embodiments.

[0080] Example

[0081] The present invention provides a method for quickly determining the measurement point layout based on feature space, comprising:

[0082] S1: Generate a sample data set of fixed-shape deformation conditions. The specific steps are as follows:

[0083] S1.1: In this embodiment, the NACA0012 airfoil is selected, and the airfoil geometry is as follows: Figure 3 The aerodynamic calculation grid is divided by Pointwise software, which is a CH-type structure grid. The number of grid nodes on the upper and lower wing surfaces is 150, the first layer height of the boundary layer is 5e-6, and the total number of grids is 66554. The pressure far field boundary condition is adopted. In order to reduce the influence of the far field condition, the far field is located at about 70 chord lengths from the wing surface. The global schematic diagram of the aerodynamic grid is shown in Figure 4 As shown, the local schematic diagram is as follows Figure 5 shown.

[0084] S1.2: Determine the design space of the calculation conditions of the airfoil. The calculation conditions are the angle of attack and the Mach number. In the subsonic range, the angle of attack ranges from 0 to 5°, and the Mach number ranges from 0.3 to 0.6. The number of samples extracted using Latin hypercube is 300, of which 90% are used for training and the remaining 10% are used for testing. The schematic diagram of the calculation condition sampling is as follows Figure 6 shown.

[0085] S1.3: Perform CFD simulation calculation on the operating parameters obtained in S1.2, wherein the CFD calculation model adopts the SSTk-omega turbulence model to obtain corresponding original flow field data.

[0086] S1.4: Process the flow field data obtained in S1.3 to obtain a wing surface pressure distribution data set with a dimension of 319×300.

[0087] S2: Determine the number of sparse measurement points by combining the proper orthogonal decomposition (POD) method. The specific steps are as follows:

[0088] S2.1: The POD method is used to reduce the order of the pressure distribution data in the training set obtained in S1.4. The fundamental mode in the low-dimensional space is truncated according to the energy proportion and reconstruction error. The fundamental mode order to be retained is determined to be 8, as well as the corresponding sample data mean (dimension is 319×1), fundamental mode matrix (319×8) and fundamental mode coefficient (8×270).

[0089] S2.2: Use the fundamental mode order obtained in S2.1 as the number of sparse measurement points, that is, select 8 sparse measurement points.

[0090] S3: Determine the initial measurement point location based on the statistical distribution characteristics of the data. The specific steps are as follows:

[0091] S3.1: Calculate the statistical characteristics of the airfoil pressure distribution at different measuring points, namely the first four moments, including mean, standard deviation, skewness, and kurtosis.

[0092] S3.2: Use the k-means clustering method to cluster the statistical features obtained in S3.1. The number of clusters is the number of sparse measurement points determined in S2.2. The location of the measurement point closest to the cluster center is as follows: Figure 7 The corresponding data distribution histogram is shown in Figure 8 shown.

[0093] S3.3: In each class obtained in S3.2, select the measurement point closest to the class center, or randomly select a measurement point, and combine them as the sparse measurement point location.

[0094] S4: Combine the back propagation neural network BPNN to build a proxy model. The input is the pressure distribution value corresponding to the sparse measurement point position obtained in S3.3, and the output is the base modal coefficient obtained in S2.1. Select a single hidden layer neural network, the number of hidden layer neurons is 12, and the activation function is tan-sigmoid and purelin.

[0095] S5: Use the POD model in S2 and the BPNN model in S4 for sparse reconstruction of the airfoil pressure distribution load. The specific steps are as follows:

[0096] S5.1: According to the measurement point positions in S3.3, the corresponding sparse measurement values ​​are obtained, and a set of fundamental mode coefficients are predicted using the BPNN model trained in S4.

[0097] S5.2: For the fundamental modal coefficients predicted in S5.1, combined with the fundamental modal matrix saved after POD reduction in S2 and the mean of the sample data, the fundamental modal coefficients are reconstructed back to the original space, that is, the complete airfoil pressure distribution data is obtained. Fig. 9 The comparison between the CFD calculation results and the sparse reconstruction results of the airfoil pressure distribution under four test conditions is given. The “real” represents the CFD calculation result, represented by “·”, and the “reconstructed” represents the result obtained by the reconstruction of the sparse measurement values ​​in the present invention, represented by “o”; wherein, Fig. 9 (a) a is the comparison between the CFD calculation results and the sparse reconstruction results of the airfoil pressure distribution under the test conditions of Mach number 0.5476 and angle of attack 0.5357°; Fig. 9 (b) Comparison between the CFD calculation results and the sparse reconstruction results of the airfoil pressure distribution under the test conditions of Mach number 0.4851 and angle of attack 1.8714°; Fig. 9 (c) Comparison between the CFD calculation results and the sparse reconstruction results of the airfoil pressure distribution under the test conditions of Mach number 0.5683 and angle of attack 3.7124°; Fig. 9 (d) Comparison of CFD calculation results and sparse reconstruction results of airfoil pressure distribution under test conditions with Mach number 0.5220 and angle of attack 3.0629°.

[0098] Can fit better.

[0099] In S2.1 above, the pressure distribution data is reduced using the intrinsic orthogonal decomposition method based on singular value decomposition (SVD). The essence of SVD is to orthogonalize a matrix to make it a diagonal matrix. Suppose the original data is A m×n , where m is the data dimension and n is the number of samples. After singular value decomposition, it can be expressed as follows:

[0100] A=USVT

[0101] Where U m×m and V n×n are all orthogonal matrices, V T represents the complex conjugate transpose of the matrix V. Each column vector in U is AA T The eigenvector of is called the left singular vector of A; each column vector in V is A T The eigenvector of A is called the right singular vector of A. The singular vector contains the basic modal information of data decomposition. S is an m×n matrix. The first r-order minor formula of S is a diagonal matrix. The elements on the main diagonal are all greater than zero. They are different singular values ​​arranged in order from large to small. Other than that, all other elements are zero. In this way, the matrix A can be represented by the product of three matrices of different sizes. The above decomposition can be represented in the following component form:

[0102]

[0103] in:

[0104] U=[u 1 u 2 ...u m ]

[0105] V=[v 1 v 2 ...v n ]

[0106]

[0107] Σ=diag([σ 1 ,σ 2 ,...,σ r ])

[0108] In the formula, u i (i=1,2,…,m) is the left singular vector of matrix A, v j (j=1,2,…,n) is the right singular vector of matrix A. σ k (k=1,2,…,r) is the singular value of matrix A, which can be regarded as the weighting factor of the decomposed matrix. The larger its value is, the greater its influence on the synthetic matrix in the reconstruction process. Experiments show that when using U, S, and V to reconstruct matrix A, ignoring some of its small information can still meet the reconstruction accuracy. If only the approximation of matrix A is needed, only the first-order singular values ​​with larger values ​​and their corresponding singular vectors can be used in the reconstruction process to meet the requirements, that is:

[0109]

[0110] The above is the basic principle of using the SVD method to construct a reduced-order model.

[0111] According to the sample flow field data reduction, the wing pressure distribution load A cp For example, the specific process is as follows:

[0112] 1) Standardization, that is, subtract the sample mean from each sample to obtain:

[0113]

[0114] In the formula, is the original data, which is an m-dimensional column vector, n is the number of samples, is the standardized data, is the mean value of the airfoil pressure distribution load, which can be calculated by the following formula:

[0115]

[0116] 2) For the standardized sample data matrix The singular matrix and the corresponding singular value matrix are calculated by SVD decomposition, and then the left singular vectors corresponding to the first q singular values ​​are taken through correlation comparison analysis. Each singular vector represents a fundamental mode of the pressure distribution, thus forming an m×q order fundamental mode matrix U cp :

[0117]

[0118] And calculate the basic modal coefficient matrix Z cp :

[0119]

[0120] 3) When q takes an appropriate value, the approximate pressure distribution data can be reconstructed by the following formula:

[0121]

[0122] After POD modeling, the original m-dimensional pressure distribution data can be reduced to q-dimensional. Since the low-dimensional feature space has interpretability for the physical characteristics of the flow field, it can be predicted that there is a certain correlation between the dimension of the low-dimensional feature space and the number of sparse measurement points. Therefore, the present invention determines the number of sparse measurement points based on the dimension of the low-dimensional feature space after POD reduction. In addition, the basic mode matrix U is saved cp , the basic modal coefficient matrix Z cp , sample data mean It is used to construct the subsequent sparse reconstruction model.

[0123] In S3.1 above, the statistical characteristics of the pressure distribution data are described using the commonly used fourth-order moment parameters, which can measure the data distribution shape, concentration and dispersion.

[0124] For a set of data x (i) (i=1,2,...,n), the average value is Then the k-th order central moment is defined as:

[0125]

[0126] The commonly used central moments are the first four order moments, and the meanings of each order moment are:

[0127] (1) The first-order moment mean, which shows the concentration of the data;

[0128] (2) The second-order moment standard deviation describes the size of the data dispersion range;

[0129] (3) The third-order moment skewness is a measure of the direction and degree of skewness of the statistical data distribution, including normal distribution (skewness = 0), right-skewed distribution (skewness > 0), and left-skewed distribution (skewness < 0);

[0130] (4) The fourth-order moment kurtosis is a characteristic number that characterizes the height of the peak of the probability density distribution curve at the mean value, including normal distribution (kurtosis value = 3), thick tail (kurtosis value > 3) and thin tail (kurtosis value < 3).

[0131] Since different central moments have different sensitivities to different distribution types, the present invention comprehensively considers the fourth-order moment when analyzing data distribution.

[0132] In previous studies, the optimization layout of sensors mainly includes traditional mathematical methods and intelligent optimization algorithms. The traditional method has a rigorous mathematical theoretical basis, but due to the one-sidedness of the evaluation criteria, it has certain limitations; the purpose of the intelligent optimization algorithm is to search for the global optimal value of the objective function, so the selection of the objective function determines the accuracy and convergence of the optimization, and the sparse reconstruction model must be called multiple times to calculate the objective function, and the convergence speed is slow. Therefore, in order to improve efficiency, the present invention proposes a method for quickly determining the optimal measurement point layout based on the feature space.

[0133] Compared with the prior art, the present invention determines the sparsity of system data in a low-dimensional feature space based on the POD reduction method, and can determine the number of measuring points accordingly; the scheme clusters the statistical characteristics of the pressure distribution data at different measuring points on the wing surface, selects the measuring point closest to the class center in each category or randomly selects a measuring point, and combines them as the sparse measuring point position, so as to quickly obtain a set of better measuring point layouts; the method for quickly determining the measuring point layout based on the feature space proposed in the scheme can effectively improve the efficiency of the optimal arrangement of measuring points compared to the method based on the optimization algorithm, so as to reduce the error of the sparse reconstruction of the distributed load.

[0134] like Fig.10 As shown, the second object of the present invention is to provide a system for quickly determining the measurement point layout based on feature space, comprising:

[0135] Obtaining data set module: used to obtain the airfoil pressure distribution data set of the selected airfoil;

[0136] Measuring point number module: used to process the airfoil pressure distribution data set based on the POD method to obtain the sparse measuring point number;

[0137] Measuring point location module: It is used to select measuring points based on the number of sparse measuring points as the clustering number, form sparse measuring point locations, and complete the confirmation of the measuring point layout.

[0138] like Fig.11 As shown, the third object of the present invention is to provide an electronic device, the electronic device comprising: a processor, a memory and a display screen. The memory and the display screen are both connected to the processor, such as through a bus. Optionally, the electronic device may also include a transceiver. It should be noted that in actual applications, the transceiver is not limited to one, and the structure of the electronic device does not constitute a limitation on the embodiments of the present application.

[0139] The processor may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0140] The bus may include a path to transmit information between the above components. The bus may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc.

[0141] The memory can be ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or it can be EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to this.

[0142] The memory is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor. The processor is used to execute the application code stored in the memory to implement the content shown in the above method embodiment.

[0143] Fig.11 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0144] The fourth object of the present invention is to provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program is stored thereon, and when the program is executed by a processor, the aforementioned Figure 1 The various processes of the method embodiment shown in the figure include, for example, a memory including instructions, and the instructions can be executed by a processor of an electronic device to complete the method.

[0145] A computer-readable storage medium may be a tangible device that holds and stores instructions used by an instruction execution device. A computer-readable storage medium may be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, a computer-readable storage medium may be a portable computer disk, a hard disk, a USB flash drive, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a podium random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, an optical disk, a magnetic disk, a mechanical encoding device, and any combination thereof.

[0146] A fifth object of the present invention is to provide a computer program product, comprising computer instructions, which, when executed by a processor, implement the above Figure 1 The various processes of the method embodiment shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0147] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0148] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0149] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0150] These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operations S are performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 S of functions specified in one or more boxes.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for quickly determining measurement point layout based on feature space, characterized in that: include: Obtaining a data set of airfoil pressure distribution on a selected airfoil; Based on the POD method, the airfoil pressure distribution data set is processed to obtain the number of sparse measurement points; The number of sparse measuring points is used as the cluster number to select measuring points, obtain the sparse measuring point locations, and complete the confirmation of the measuring point layout.

2. The method for quickly determining the measurement point layout based on feature space according to claim 1, characterized in that: The step of obtaining a data set of airfoil pressure distribution of the selected airfoil includes: Obtaining original flow field data for the selected airfoil; According to the original flow field data of the selected airfoil, the airfoil pressure distribution data set is obtained.

3. The method for quickly determining the measurement point layout based on feature space according to claim 2 is characterized in that: The obtaining of original flow field data for the selected airfoil includes: Determine the computational design space of the selected airfoil and use Latin hypercube sampling to obtain the operating parameters; The obtained working condition parameters are subjected to CFD simulation calculation to obtain the original flow field data.

4. The method for quickly determining the measurement point layout based on feature space according to claim 1, characterized in that: Based on the POD method, the airfoil pressure distribution data set is processed to obtain the number of sparse measurement points, including: Using the POD method to reduce the order of the airfoil pressure distribution data set of the selected airfoil to obtain the basic mode of the low-dimensional space; The fundamental mode in the low-dimensional space is truncated according to the energy proportion and reconstruction error to determine the fundamental mode order to be retained; The fundamental mode order to be retained will be determined as the number of sparse measurement points.

5. The method for quickly determining the measurement point layout based on feature space according to claim 1 is characterized in that: The method of selecting measuring points by taking the number of sparse measuring points as the number of clusters to obtain the positions of sparse measuring points includes: Calculate the statistical characteristics of the airfoil pressure distribution at different measuring points; The number of sparse measurement points is used as the number of clusters to cluster the statistical characteristics of the data and obtain the clustering results; In each class in the clustering result, the measurement point closest to the class center is selected, or a measurement point is randomly selected and combined as the sparse measurement point location.

6. The method for quickly determining the measurement point layout based on feature space according to claim 1 is characterized in that: The method adopts the number of sparse measurement points as the number of clusters to select measurement points and obtain the sparse measurement point positions. Then, the sparse measurement values ​​in the sparse measurement point positions are used as input; the POD method is used to reduce the order of the wing surface pressure distribution data set to obtain the fundamental mode of the low-dimensional space, and the fundamental mode coefficients in the fundamental mode of the low-dimensional space are used as output to construct a BPNN model and train it.

7. A system for quickly determining measurement point layout based on feature space, characterized in that: include: Obtaining data set module: used to obtain the airfoil pressure distribution data set of the selected airfoil; Measuring point number module: used to process the airfoil pressure distribution data set based on the POD method to obtain the sparse measuring point number; Measuring point location module: It is used to select measuring points based on the number of sparse measuring points as the clustering number, form sparse measuring point locations, and complete the confirmation of the measuring point layout.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of a method for quickly determining a measurement point layout based on a feature space as described in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of a method for quickly determining a measurement point layout based on a feature space as recited in any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of a method for quickly determining a measurement point layout based on a feature space as claimed in any one of claims 1 to 6.

Citation Information

Cited By

  • Flow field reconstruction method based on test point position optimization and multi-source data fusion and related device

    CN120409335A

  • Load distribution machine learning method and system based on intrinsic characteristic modal enhancement

    CN121072808A

  • Online identification method and device for airfoil low Reynolds number laminar flow separation form

    CN121723405A