Pneumatic correction method and device for wind turbine airfoil and medium

Abstract

The invention relates to the technical field of wind turbine generators, particularly provides a pneumatic correction method and device for a wind turbine airfoil and a medium, and aims to solve the technical problem that the existing method is difficult to obtain pneumatic data with higher precision through low-precision pneumatic data. In order to achieve the purpose, the pneumatic correction method for the wind turbine airfoil comprises the steps that first blade pneumatic information and second blade pneumatic information are obtained; performing feature extraction on the first blade aerodynamic information and the first incoming flow information to obtain a first blade aerodynamic feature; performing model training on a neural network fusion model by using the first blade aerodynamic characteristics and the second blade aerodynamic information to obtain a trained neural network fusion model; third blade pneumatic information corresponding to the second thickness type and the second incoming flow information is obtained; and according to the third blade pneumatic information and the second incoming flow information, blade pneumatic correction information corresponding to the second thickness type is obtained by using the trained neural network fusion model, and high-precision blade pneumatic data is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine units, and specifically provides a method, device and medium for aerodynamic correction of wind turbine airfoils. Background Art

[0002] High-precision aerodynamic data is the basis for the overall design of wind turbine airfoils and is the basic prerequisite and important guarantee for ensuring the success of airfoil development. Wind tunnel tests and physical models are the main research means for analyzing the aerodynamic characteristics of wind turbine airfoils. Wind tunnel tests can conduct high-precision aerodynamic characteristic analysis on airfoils and can provide direct and real aerodynamic characteristic data, but there are problems such as high cost and long test cycle. Physical models are based on assumptions and theoretical derivations. Although they can provide a preliminary understanding and prediction of aerodynamic characteristics, their results often depend on the simplification degree of the model and the rationality of the assumptions, and the accuracy of the calculated results is relatively low, which is suitable for the preliminary design stage and rapid evaluation. High-precision data (wind tunnel tests) has higher authenticity and reliability, but is costly and scarce; low-precision data (physical models) is low-cost and abundant, but may have deviations and errors.

[0003] Existing methods are difficult to obtain higher-precision aerodynamic data from low-precision aerodynamic data, which is difficult to meet actual needs. Summary of the Invention

[0004] In order to overcome the above defects, this application is proposed to provide a solution to or at least partially solve the technical problem that existing methods cannot obtain higher-precision aerodynamic data from low-precision aerodynamic data. This application provides a method, device and medium for aerodynamic correction of wind turbine airfoils.

[0005] In a first aspect, this application provides a method for aerodynamic correction of wind turbine airfoils based on a data fusion strategy, the method comprising:

[0006] Obtain first blade aerodynamic information and second blade aerodynamic information, where the first blade aerodynamic information and the second blade aerodynamic information are aerodynamic information corresponding to a first thickness type and a first oncoming flow information;

[0007] Extract features from the first blade aerodynamic information and the first oncoming flow information to obtain first blade aerodynamic features;

[0008] Use the first blade aerodynamic features and the second blade aerodynamic information to train a neural network fusion model to obtain a trained neural network fusion model;

[0009] Obtain third blade aerodynamic information corresponding to a second thickness type and a second oncoming flow information;

[0010] Based on the third blade aerodynamic information and the second oncoming flow information, the blade aerodynamic correction information corresponding to the second thickness type is obtained by using the trained neural network fusion model.

[0011] In an embodiment of the present application, the neural network fusion model includes L layers, each layer contains multiple neurons, the first layer is the input layer, the first layer contains n 1 neurons, the Lth layer is the output layer, the Lth layer contains n L neurons, and the neural network fusion model is a mapping from to where L is a positive integer greater than or equal to 3, and R is the set of real numbers.

[0012] In an embodiment of the present application, the training of the neural network fusion model by using the first blade aerodynamic characteristics and the second blade aerodynamic information includes:

[0013] Input the first blade aerodynamic characteristics into the neural network fusion model to obtain a prediction result;

[0014] Use the second blade aerodynamic information as a label, and calculate the model loss according to the prediction result and the second blade aerodynamic information;

[0015] Judge whether the model loss is less than a preset loss threshold, or whether the number of iterative training times reaches a preset number threshold;

[0016] If so, obtain the trained neural network fusion model.

[0017] In an embodiment of the present application, the mean square error function is used as the model loss function.

[0018] In an embodiment of the present application, the method further includes: when the model loss is not less than the preset loss threshold, or the number of iterative training times does not reach the preset number threshold, using the stochastic gradient descent method to optimize the model loss.

[0019] In an embodiment of the present application, the obtaining of the first blade aerodynamic information and the second blade aerodynamic information includes:

[0020] According to the first thickness type and the first oncoming flow information, perform aerodynamic data calculation through the physical model of potential flow theory to obtain the first blade aerodynamic information;

[0021] Obtain the second blade aerodynamic information by using a wind tunnel test.

[0022] In an embodiment of the present application, the feature extraction of the first blade aerodynamic information and the first oncoming flow information includes:

[0023] Eliminate the invalid information in the first blade aerodynamic information and the first oncoming flow information;

[0024] Convert the first blade aerodynamic information and the first oncoming flow information after eliminating the invalid information into feature vectors to obtain the first blade aerodynamic characteristics.

[0025] In one embodiment of the present application, the second thickness type is obtained by interpolating the thickness of the first thickness type;

[0026] The obtaining of the third blade aerodynamic information corresponding to the second thickness type and the second oncoming flow information includes: calculating aerodynamic data according to the second thickness type and the second oncoming flow information through the physical model of potential flow theory to obtain the third blade aerodynamic information.

[0027] In a second aspect, there is provided an electronic device, including:

[0028] At least one processor;

[0029] And a memory communicatively connected to the at least one processor;

[0030] Wherein, a computer program is stored in the memory, and when the computer program is executed by the at least one processor, the foregoing wind turbine airfoil aerodynamic correction method is performed.

[0031] In a third aspect, there is provided a computer-readable storage medium, in which multiple program codes are stored, and the program codes are adapted to be loaded and run by a processor to execute the wind turbine airfoil aerodynamic correction method described in any one of the foregoing.

[0032] One or more of the above technical solutions of the present application have at least one or more of the following Beneficial effects:

[0033] The wind turbine airfoil aerodynamic correction method in the present application specifically includes: obtaining first blade aerodynamic information and second blade aerodynamic information, where the first blade aerodynamic information and the second blade aerodynamic information are aerodynamic information corresponding to the first thickness type and the first oncoming flow information; performing feature extraction on the first blade aerodynamic information and the first oncoming flow information to obtain the first blade aerodynamic characteristics; using the first blade aerodynamic characteristics and the second blade aerodynamic information to train a neural network fusion model to obtain a trained neural network fusion model; obtaining third blade aerodynamic information corresponding to the second thickness type and the second oncoming flow information; and obtaining blade aerodynamic correction information corresponding to the second thickness type according to the third blade aerodynamic information and the second oncoming flow information by using the trained neural network fusion model. In this way, aerodynamic data of airfoils with different thicknesses can be quickly obtained, the dependence on traditional wind tunnel tests is reduced, and high-precision aerodynamic data are obtained. Brief Description of the Drawings

[0034] Referring to the accompanying drawings, the disclosure of the present application will become more readily understandable. It is easily understood by those skilled in the art that these drawings are only for illustrative purposes and are not intended to limit the scope of protection of the present application. In addition, similar numbers in the figures are used to represent similar components, where:

[0035] Figure 1 is a schematic diagram of the main process of the airfoil aerodynamic correction method in an embodiment of the present application;

[0036] Figure 2 is a schematic diagram of the structure of the neural network fusion model in an embodiment of the present application;

[0037] Figure 3 is a schematic diagram of the structure of the airfoil aerodynamic correction device in an embodiment of the present application;

[0038] Figure 4 is a schematic diagram of the structure of the electronic device in an embodiment of the present application. Detailed Description of the Embodiments

[0039] The following describes some embodiments of the present application with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principle of the present application and are not intended to limit the scope of protection of the present application.

[0040] In the description of the present application, "module" and "processor" may include hardware, software, or a combination of both. A module may include a hardware circuit, various suitable sensors, communication ports, memories, and may also include a software part, such as program code, or may be a combination of software and hardware. The processor may be a central processing unit, a microprocessor, an image processor, a digital signal processor, or any other suitable processor. The processor has data and / or signal processing functions. The processor may be implemented in software, in hardware, or in a combination of both. The non-transitory computer-readable storage medium includes any suitable medium for storing program code, such as a magnetic disk, a hard disk, an optical disk, a flash memory, a read-only memory, a random access memory, and so on. The term "A and / or B" represents all possible combinations of A and B, such as only A, only B, or A and B. The term "at least one A or B" or "at least one of A and B" has a meaning similar to "A and / or B" and may include only A, only B, or A and B. The singular terms "a" and "this" may also include the plural form.

[0041] At present, it is difficult to obtain higher-precision aerodynamic data through low-precision aerodynamic data by traditional methods, which is difficult to meet the actual needs.

[0042] To this end, the present application provides a method, device, and medium for aerodynamic correction of a wind turbine airfoil.

[0043] Refer to the appendix Figure 1 , Figure 1 which is a schematic diagram of the main steps of the aerodynamic correction method for a wind turbine airfoil according to an embodiment of the present application.

[0044] As Figure 1 shown, the aerodynamic correction method for a wind turbine airfoil in the embodiment of the present application mainly includes the following steps S10 - S50.

[0045] Step S10: Obtain the aerodynamic information of the first blade and the aerodynamic information of the second blade. The aerodynamic information of the first blade and the aerodynamic information of the second blade are the aerodynamic information corresponding to the first thickness type and the first oncoming flow information.

[0046] Step S20: Extract features from the aerodynamic information of the first blade and the first oncoming flow information to obtain the aerodynamic features of the first blade.

[0047] Step S30: Use the aerodynamic features of the first blade and the aerodynamic information of the second blade to train the neural network fusion model to obtain a trained neural network fusion model.

[0048] Step S40: Obtain the aerodynamic information of the third blade corresponding to the second thickness type and the second oncoming flow information.

[0049] Step S50: According to the aerodynamic information of the third blade and the second oncoming flow information, use the trained neural network fusion model to obtain the blade aerodynamic correction information corresponding to the second thickness type.

[0050] Based on the above steps S10 - S50, first obtain the aerodynamic information of the first blade and the aerodynamic information of the second blade. The aerodynamic information of the first blade and the aerodynamic information of the second blade are the aerodynamic information corresponding to the first thickness type and the first oncoming flow information; extract features from the aerodynamic information of the first blade and the first oncoming flow information to obtain the aerodynamic features of the first blade; use the aerodynamic features of the first blade and the aerodynamic information of the second blade to train the neural network fusion model to obtain a trained neural network fusion model; obtain the aerodynamic information of the third blade corresponding to the second thickness type and the second oncoming flow information; according to the aerodynamic information of the third blade and the second oncoming flow information, use the trained neural network fusion model to obtain the blade aerodynamic correction information corresponding to the second thickness type. In this way, the aerodynamic data of airfoils with different thicknesses can be obtained quickly, reducing the dependence on traditional wind tunnel tests and obtaining high-precision aerodynamic data.

[0051] Next, the physical model based on the potential flow theory of the present application will be described in detail first.

[0052] The applicable object of this application can be any airfoil. To describe the aerodynamic correction method of the wind turbine airfoil of this application in more detail, taking the NREL S-series airfoils as an example, first select three airfoils with different thicknesses, namely S809, S816, and S823, and then use the Xfoil software to quickly obtain the low-precision aerodynamic data of the airfoils.

[0053] Xfoil is a numerical simulation tool based on the combination of potential flow theory and boundary layer theory, used to quickly calculate the aerodynamic performance of two-dimensional airfoils. The potential flow theory assumes that the fluid flow is incompressible, irrotational, and inviscid. Its core lies in describing the flow field through the velocity potential function φ:

[0054] The fluid is incompressible, and the potential function satisfies the Laplace equation:

[0055] The fluid is irrotational, and the velocity vector can be expressed as the gradient of the velocity potential:

[0056] Apply the impermeable condition on the airfoil surface, that is, the normal component of the fluid velocity is zero:

[0057] Meet the free-stream condition in the far field: (when r → ∞)

[0058] where n is the normal vector of the airfoil surface, V ∞ is the free-stream velocity, and r is the distance outside the airfoil.

[0059] The calculation process of the airfoil aerodynamic data in Xfoil is as follows: The panel method is used to solve the pressure distribution, lift, drag, and moment on the airfoil surface based on the potential flow theory. First, according to the shape of the airfoil, the airfoil surface is divided into a sufficient number of finite small elements, that is, panels; then a source or vortex is arranged on each panel to establish a linear equation system that satisfies the non-penetration boundary condition at the control point on each panel and the Kutta condition at the trailing edge of the airfoil; finally, the equation system is solved to determine the panel strength (dipole strength or vortex strength), and the perturbation velocity potential is obtained to further calculate the pressure distribution, lift, drag, and moment on the airfoil surface.

[0060] Use Xfoil to calculate the aerodynamic data of the S809, S816, and S823 airfoils at different Reynolds numbers and different angles of attack. The calculation results obtained by Xfoil are poor, which are used as low-precision aerodynamic data. At the same time, wind tunnel experiments are carried out to obtain the corresponding high-precision aerodynamic data.

[0061] The above steps S10 to S50 will be further described below.

[0062] In a specific embodiment of the present application for the above-mentioned step S10, the obtaining of the first blade aerodynamic information and the second blade aerodynamic information includes: calculating aerodynamic data according to the first thickness type and the first oncoming flow information through a physical model of potential flow theory to obtain the first blade aerodynamic information; obtaining the second blade aerodynamic information by using a wind tunnel test.

[0063] The first thickness type can be a blade airfoil of any thickness. Exemplarily, for the NREL S-series airfoils, the first thickness type can be three different thickness airfoils of S809, S816, and S823. The first oncoming flow information can include the angle of attack and Reynolds number of the airfoil. Among them, the first blade aerodynamic information and the second blade aerodynamic information can include lift coefficient and drag coefficient, etc.

[0064] Specifically, the first blade aerodynamic information under the first thickness type and the first oncoming flow information can be obtained through the aforementioned physical model based on potential flow theory.

[0065] The second blade aerodynamic information is obtained by conducting a wind tunnel test according to the first thickness type and the first oncoming flow information.

[0066] The above is a further description of step S10. Next, a further description of step S20 will be continued.

[0067] In a specific embodiment of the present application, the feature extraction of the first blade aerodynamic information and the first oncoming flow information includes: removing invalid information from the first blade aerodynamic information and the first oncoming flow information; converting the first blade aerodynamic information and the first oncoming flow information after removing invalid information into feature vectors to obtain the first blade aerodynamic features.

[0068] Specifically, for the first blade aerodynamic information and the first oncoming flow information, first, preprocessing can be performed on them, such as removing outliers and filling missing values, etc. Then, the preprocessed first blade aerodynamic information and the first oncoming flow information can be converted into feature vectors that can be recognized by the neural network fusion model. Exemplarily, for the first blade aerodynamic information and the first oncoming flow information, word embedding can be performed on the first blade aerodynamic information and the first oncoming flow information to obtain feature vectors that can be recognized by the neural network fusion model.

[0069] The above is a further description of step S20. Before elaborating on step S30 in detail below, the neural network fusion model of the present application will be introduced in detail first.

[0070] Since the high-precision (wind tunnel experiment) and low-precision (physical model) data of the wind turbine airfoil both come from the same physical system, there is a mapping relationship between the two. Assume that the data sets obtained by the airfoil through the physical model and the wind tunnel experiment are (X, Ylow ) and (X, Y high ), construct a mapping f(*) between the high-precision aerodynamic data obtained from wind tunnel experiments and the low-precision aerodynamic data obtained from physical models, then there is: Y high = f(X, Y low )

[0071] In the formula, Y low and Y high are the low-precision aerodynamic data set and the high-precision aerodynamic data set respectively, where the aerodynamic data mainly includes aerodynamic data such as lift coefficient (C L ), drag coefficient (C D ), etc.; X is the angle of attack and Reynolds number (Re) of the airfoil, etc.

[0072] The architecture of the neural network fusion model in this application consists of a neural network. Deep learning can automatically extract features from raw data. Therefore, using the airfoil flow information X and the low-precision aerodynamic data Y low as inputs, and adopting supervised machine learning techniques, where the high-precision aerodynamic data Y high is used as the label to train the neural network. Through the training of the neural network, it learns to "map" high-precision data from low-precision data, and models the relationship between low-precision aerodynamic data and high-precision aerodynamic data.

[0073] In a specific embodiment of this application, the neural network fusion model includes L layers, each layer contains multiple neurons. The first layer is the input layer, and the first layer contains n 1 neurons. The Lth layer is the output layer, and the Lth layer contains n L neurons. The neural network fusion model is a mapping from to , where L is a positive integer greater than or equal to 3, and R is the set of real numbers.

[0074] As Figure 2 shown, the neural network fusion model of this application includes an input layer, a hidden layer, and an output layer, where the total number of layers of the input layer, hidden layer, and output layer is L layers. Exemplarily, the number of propagation layers of the neural network can be set from 3 to 5 layers, where the first layer and the Lth layer represent the input layer and the output layer. Let the first layer have n l neurons, then n 1 is the number of neurons in the input layer, that is, the dimension of the input data. The neural network is a mapping from R n1 to R nL . Let represent the weight of the output value of neuron j in layer l acting on neuron k in layer l-1. The vector b [l] ∈R nlDenote the deviation of the l-th layer. Therefore, the deviation of neuron j corresponding to the l-th layer is Use the relu function as the activation function: σ(x) = max(0, x)

[0075] Then the final output result of the neural network is:

[0076] where denote the parameters of network training, u is the network output, and x is the network input. In the formula, denotes element-wise multiplication, also known as the Hadamard product. The Hadamard product calculates the product of corresponding elements of two matrices (or vectors) element by element, rather than performing matrix multiplication. In a neural network, usually represents the operation between layers.

[0077] In a specific embodiment of the present application, step S30 can be implemented by the following steps S301 to S303.

[0078] Step S301: Input the aerodynamic characteristics of the first blade into the neural network fusion model to obtain a prediction result.

[0079] Step S302: Use the aerodynamic information of the second blade as a label, and calculate the model loss according to the prediction result and the aerodynamic information of the second blade.

[0080] Step S303: Determine whether the model loss is less than a preset loss threshold, or whether the number of iterative training times reaches a preset number threshold. If so, obtain the trained neural network fusion model.

[0081] The preset loss threshold and the preset number threshold can be values obtained in advance according to experiments, and can be specifically adaptively modified and adjusted according to the actual usage scenario, and no specific limitation is made thereto.

[0082] Specifically, the first blade aerodynamic characteristics can be input into the neural network fusion model to perform forward propagation calculation and obtain the prediction result through the output layer. Error backpropagation uses an optimizer to backward transmit the model loss error through algorithms such as gradient descent and update the trainable parameters of the network. In one training iteration, a batch (i.e., a subset of the training set) is used for forward propagation and backpropagation. One iteration refers to the complete process of one forward propagation and one backpropagation of the entire training set. After forward feed calculation and error backpropagation, if the expected convergence is not achieved, the foregoing training steps are repeated. The end of training can be determined by a set preset loss threshold or the maximum number of iterations. Additionally, when the model loss is not less than the preset loss threshold, or the number of iterative training times does not reach the preset number threshold, the stochastic gradient descent method is further used to optimize the model loss.

[0083] The backpropagation of the neural network is an optimization process for learning unknown parameters by minimizing the loss function. Commonly used loss functions include mean square error (MSE), cross-entropy, classification hinge, etc. Since the mean square error function is more suitable for regression tasks, that is, tasks of predicting continuous values, this application can use the mean square error function as the model loss function, and the loss function of the training samples is defined as follows:

[0084] Where represents the true value of the label, y H is the output value of the neural network, and w is the weight of the network. N represents the number of high and low precision data sets, and the Adam stochastic gradient descent algorithm is used to optimize the loss function.

[0085] The above is a further description of step S30. Next, step S40 will be further described.

[0086] Regarding step S40, in a specific embodiment of the present application, the second thickness type is obtained by interpolating the thickness of the first thickness type; the obtaining of the third blade aerodynamic information corresponding to the second thickness type and the second oncoming flow information includes: performing aerodynamic data calculation according to the second thickness type and the second oncoming flow information through the physical model of potential flow theory to obtain the third blade aerodynamic information.

[0087] The second thickness type can be obtained by interpolating the thickness of the first thickness type. Exemplarily, when the first thickness type is airfoils with three different thicknesses of S809, S816, and S823, the second thickness type can be thickness types such as S812, S817, and S821 obtained by interpolating the thickness of the first thickness type.

[0088] Specifically, the third blade aerodynamic information corresponding to the second thickness type and the second oncoming flow information can be obtained through the physical model of the aforementioned potential flow theory.

[0089] The above is a further description of step S40. Next, step S50 will be further described.

[0090] In step S50, the third blade aerodynamic information and the second oncoming flow information can be further converted into feature vectors that can be recognized by the neural network fusion model, and the feature vectors are input into the trained neural network fusion model, and then the blade aerodynamic correction information can be obtained. The blade aerodynamic correction information is the high-precision blade aerodynamic information.

[0091] It should be noted that although the above embodiments describe the various steps in a specific order, those skilled in the art can understand that in order to achieve the effects of the present application, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders, and these variations are all within the protection scope of the present application.

[0092] Furthermore, the present application also provides a wind turbine airfoil aerodynamic correction device.

[0093] Refer to the attached Figure 3 , Figure 3 which is the main structural block diagram of the wind turbine airfoil aerodynamic correction device according to an embodiment of the present application.

[0094] As Figure 3 shown, the wind turbine airfoil aerodynamic correction device in the embodiment of the present application mainly includes a first acquisition module 11, a feature extraction module 12, a model training module 13, a second acquisition module 14, and a prediction module 15. In some embodiments, one or more of the first acquisition module 11, the feature extraction module 12, the model training module 13, the second acquisition module 14, and the prediction module 15 can be combined together into one module.

[0095] In some embodiments, the first acquisition module 11 can be configured to acquire the first blade aerodynamic information and the second blade aerodynamic information, and the first blade aerodynamic information and the second blade aerodynamic information are the aerodynamic information corresponding to the first thickness type and the first oncoming flow information.

[0096] The feature extraction module 12 can be configured to perform feature extraction on the first blade aerodynamic information and the first oncoming flow information to obtain the first blade aerodynamic feature.

[0097] The model training module 13 can be configured to use the first blade aerodynamic feature and the second blade aerodynamic information to train the neural network fusion model to obtain the trained neural network fusion model.

[0098] The second acquisition module 14 may be configured to acquire third blade aerodynamic information corresponding to the second thickness type and the second oncoming flow information.

[0099] The prediction module 15 may be configured to obtain blade aerodynamic correction information corresponding to the second thickness type according to the third blade aerodynamic information and the second oncoming flow information by using the trained neural network fusion model.

[0100] In one embodiment, the description of the specific implementation functions can be referred to in steps S10 - S50.

[0101] The above - mentioned wind turbine airfoil aerodynamic correction device is used to execute Figure 1 the embodiments of the wind turbine airfoil aerodynamic correction method shown. The technical principles, the technical problems solved and the technical effects produced are similar. Those skilled in the art can clearly understand that, for the convenience and conciseness of description, the specific working process and related descriptions of the wind turbine airfoil aerodynamic correction device can refer to the content described in the embodiments of the wind turbine airfoil aerodynamic correction method, which will not be elaborated here.

[0102] Furthermore, it should be understood that since the setting of each module is only for explaining the functional units of the device of the present application, the physical devices corresponding to these modules can be the processor itself, or a part of the software in the processor, a part of the hardware, or a part of the combination of software and hardware. Therefore, the number of each module in the figure is only illustrative.

[0103] Those skilled in the art can understand that the various modules in the device can be adaptively split or combined. Such splitting or combination of specific modules will not cause the technical solution to deviate from the principle of the present application. Therefore, the technical solutions after splitting or combination will all fall within the protection scope of the present application.

[0104] Those skilled in the art can understand that all or part of the processes in the method of the above - mentioned embodiment of the present application can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer - readable storage medium. When the computer program is executed by a processor, the steps of the above - mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer - readable storage medium can include: any entity or device, medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read - only memory, random access memory, electrical carrier signal, telecommunication signal, and software distribution medium, etc., that can carry the computer program code.

[0105] Further, the present application also provides an electronic device, which may include at least one processor; and a memory communicatively connected to the at least one processor; wherein, a computer program is stored in the memory, and when the computer program is executed by the at least one processor, the wind turbine airfoil aerodynamic correction method described in any of the above embodiments is implemented. Refer to Figure 4 as shown in Figure 4 FIG. shows an exemplary structure of an electronic device, which includes a processor 100 and a memory 200.

[0106] Further, the present application also provides a computer-readable storage medium. In an embodiment of a computer-readable storage medium according to the present application, the computer-readable storage medium may be configured to store a program for executing the wind turbine airfoil aerodynamic correction method in the above method embodiment, and this program may be loaded and run by a processor to implement the above wind turbine airfoil aerodynamic correction method. For the sake of convenience of description, only the parts related to the embodiments of the present application are shown. For the specific technical details not disclosed, please refer to the method part of the embodiments of the present application. The computer-readable storage medium may be a memory device formed by various electronic devices. Optionally, the computer-readable storage medium in the embodiments of the present application is a non-transitory computer-readable storage medium.

[0107] So far, the technical solutions of the present application have been described in conjunction with the specific embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present application is obviously not limited to these specific embodiments. Without departing from the principle of the present application, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present application.

Claims

1. A method for aerodynamic correction of a wind turbine airfoil, characterized in that: The method comprises: Acquire first blade aerodynamic information and second blade aerodynamic information, wherein the first blade aerodynamic information and the second blade aerodynamic information are aerodynamic information corresponding to the first thickness type and the first incoming flow information; Extracting features from the first blade aerodynamic information and the first incoming flow information to obtain aerodynamic features of the first blade; Using the aerodynamic characteristics of the first blade and the aerodynamic information of the second blade to perform model training on the neural network fusion model to obtain a trained neural network fusion model; Acquire aerodynamic information of a third blade corresponding to the second thickness type and the second incoming flow information; The blade aerodynamic correction information corresponding to the second thickness type is obtained according to the third blade aerodynamic information and the second incoming flow information by utilizing the trained neural network fusion model.

2. The method for aerodynamic correction of a wind turbine airfoil according to claim 1, characterized in that: The neural network fusion model includes L layers, each layer includes multiple neurons, the first layer is the input layer, the first layer includes n1 neurons, the Lth layer is the output layer, the Lth layer includes n L neurons, the neural network fusion model is from arrive , where L is a positive integer greater than or equal to 3, and R is the set of real numbers.

3. The method for aerodynamic correction of a wind turbine airfoil according to claim 2, characterized in that: The using the aerodynamic characteristics of the first blade and the aerodynamic information of the second blade to perform model training on the neural network fusion model includes: Inputting the aerodynamic characteristics of the first blade into the neural network fusion model to obtain a prediction result; Using the aerodynamic information of the second blade as a label, and calculating a model loss according to the prediction result and the aerodynamic information of the second blade; Determine whether the model loss is less than a preset loss threshold, or whether the number of iterative training times reaches a preset number threshold; If so, a trained neural network fusion model is obtained.

4. The method for aerodynamic correction of a wind turbine airfoil according to claim 3, characterized in that: The mean square error function is used as the model loss function.

5. The method for aerodynamic correction of a wind turbine airfoil according to claim 3, characterized in that: The method also includes: when the model loss is not less than a preset loss threshold, or the number of iterative training times does not reach a preset number threshold, using a stochastic gradient descent method to optimize the model loss.

6. The method for aerodynamic correction of a wind turbine airfoil according to claim 1, characterized in that: The obtaining of the first blade aerodynamic information and the second blade aerodynamic information comprises: According to the first thickness type and the first incoming flow information, aerodynamic data calculation is performed through a physical model of potential flow theory to obtain aerodynamic information of the first blade; The aerodynamic information of the second blade is obtained by using a wind tunnel test.

7. The method for aerodynamic correction of a wind turbine airfoil according to claim 1, characterized in that: The extracting features of the first blade aerodynamic information and the first incoming flow information includes: Eliminating invalid information in the first blade aerodynamic information and the first incoming flow information; The first blade aerodynamic information and the first incoming flow information after invalid information is eliminated are converted into a feature vector to obtain the first blade aerodynamic feature.

8. The method for aerodynamic correction of a wind turbine airfoil according to claim 1, characterized in that: The second thickness type is obtained by interpolating the thickness of the first thickness type; The obtaining of the aerodynamic information of the third blade corresponding to the second thickness type and the second incoming flow information includes: obtaining the aerodynamic information of the third blade by calculating aerodynamic data according to the second thickness type and the second incoming flow information through a physical model of potential flow theory.

9. An electronic device, characterized in that: include: at least one processor; and, a memory communicatively coupled to the at least one processor; Wherein, a computer program is stored in the memory, and when the computer program is executed by the at least one processor, the method for aerodynamic correction of a wind turbine airfoil according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium storing a plurality of program codes, characterized in that: The program code is suitable for being loaded and run by a processor to execute the wind turbine airfoil aerodynamic correction method according to any one of claims 1 to 8.

Images