Draught fan blade airfoil hysteresis loop calculation method, device and equipment and medium

Abstract

The invention provides a fan blade airfoil hysteresis loop calculation method and device, equipment and a medium. The method comprises the steps that input data are acquired; calling an aerodynamic parameter prediction model and performing prediction based on the input data to obtain aerodynamic parameters corresponding to the fan blade airfoil; and finally, drawing a hysteresis loop according to a prediction result of the aerodynamic parameters. Wherein the aerodynamic parameter prediction model is a deep neural network model trained based on an LM method in advance. Therefore, the technical problem of low aerodynamic parameter calculation efficiency in the prior art can be solved.

Description

Technical Field

[0001] This application relates to the field of wind turbine aerodynamic analysis technology, specifically to a method, apparatus, equipment, and medium for calculating hysteresis loops in airfoil wind turbine blades. Background Technology

[0002] In recent years, to improve wind energy capture efficiency and economic benefits, the wind power industry has been continuously promoting the development of larger wind turbines, with a significant increase in the spanwise length of turbine blades and a more prominent structural flexibility. This trend makes the blades more prone to large-scale elastic deformation under aerodynamic loads, thus inducing dynamic stall. Dynamic stall not only causes local flow separation in the blades, reducing wind energy utilization efficiency, but also, relative to the equilibrium point during stable operation, dynamic stall is often accompanied by multi-degree-of-freedom coupled vibrations of blade torsion, flapping, and swaying, leading to hysteresis loops with periodic changes in blade aerodynamic parameters. This force variation is also one of the important factors causing blade root fatigue damage and reducing the service life of blades and even the entire wind turbine. As a basic component of the blade, the dynamic stall performance of the airfoil directly affects the overall dynamic stall characteristics of the blade. Therefore, the study of airfoil dynamic stall problems has gradually become a focus of attention for blade design and load simulation departments in the wind power field.

[0003] Currently, the industry mainly relies on computational fluid dynamics (CFD) methods to simulate the dynamic stall of airfoils, thereby calculating relevant aerodynamic parameters and obtaining hysteresis loops. However, this method suffers from low computational efficiency and high requirements for equipment computing power, thus limiting its application in practical engineering. Summary of the Invention

[0004] In view of this, the embodiments of this application aim to provide a method, apparatus, equipment and medium for calculating the hysteresis loop of wind turbine blade airfoils, which can solve the technical problems of low calculation efficiency and high requirements for equipment computing power in the prior art.

[0005] In a first aspect, this application provides a method for calculating the hysteresis loop of a wind turbine blade airfoil, including: Acquire input data, which includes environmental parameters and motion change data of the wind turbine blade airfoil; The aerodynamic parameter prediction model is invoked and prediction is performed based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil. The aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method. Based on the predicted aerodynamic parameters, the hysteresis loop of the wind turbine blade airfoil is plotted.

[0006] In some embodiments, obtaining input data includes: Obtain environmental parameters and motion change data of the wind turbine blade airfoil.

[0007] In some embodiments, the method further includes: Obtain a sample set, which includes a training set and a test set. The training set and the test set each include environmental parameters of the blade airfoil, motion change data, and corresponding simulated aerodynamic parameters. Based on the training set and the LM method, the aerodynamic parameter prediction model to be trained is trained and updated to obtain the trained aerodynamic parameter prediction model. In some embodiments, obtaining the sample set includes: A sampling parameter space is constructed, which includes the sampling range corresponding to each sampling parameter. The sampling parameters include the environmental parameters and motion parameters of the wind turbine blade airfoil. Latin hypercube sampling is performed based on the sampling parameter space to obtain m sampling points, where m is a positive integer; Based on the motion parameters and numerical differential calculations of the m sampling points, the corresponding motion change data are obtained. Based on the airfoil of the wind turbine blades and the environmental parameters and motion change data, the simulated aerodynamic parameters are obtained through computational fluid dynamics simulation. The environmental parameters, motion change data, and corresponding simulated aerodynamic parameters of m sample points are determined as the sample set, and the sample set is divided into a training set and a test set.

[0008] In some embodiments, the environmental parameters include the wind speed of the wind turbine blade airfoil, the motion parameters include the angle of attack oscillation amplitude and oscillation reduction frequency of the wind turbine blade airfoil, and the motion change data includes the time series data of the angle of attack change and its corresponding angular velocity and angular acceleration data.

[0009] In some embodiments, the aerodynamic parameters include at least one of the lift coefficient, drag coefficient, and moment coefficient.

[0010] In some embodiments, training and updating the aerodynamic parameter prediction model to be trained based on the training set and the LM method includes: The environmental parameters, motion change data, and corresponding simulated aerodynamic parameters in the training set are input into the aerodynamic parameter prediction model to be trained in batches. The LM method is used to train and update the parameters of the aerodynamic parameter prediction model to be trained, so that the loss value between the predicted aerodynamic parameters and the corresponding simulated aerodynamic parameters in the training set is less than a preset loss threshold, thus obtaining the trained aerodynamic parameter prediction model.

[0011] Secondly, this application provides an apparatus for predicting aerodynamic parameters, comprising: The acquisition module is used to acquire input data, which includes environmental parameters and motion change data of the wind turbine blade airfoil; The processing module is used to call the aerodynamic parameter prediction model and make predictions based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil. The aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method. The processing module is also used to draw the hysteresis loop of the wind turbine blade airfoil based on the predicted aerodynamic parameters.

[0012] For any content not introduced or described in the embodiments of this application, please refer to the relevant descriptions in the foregoing method embodiments; they will not be repeated here.

[0013] Thirdly, this application provides a computer device, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the executable instructions to implement the steps of the above-described method for calculating the hysteresis loop of the wind turbine blade airfoil.

[0014] Fourthly, this application provides a computer-readable storage medium storing computer program instructions thereon, which, when executed by a processor, implement the steps of the above-described method for calculating the hysteresis loop of the wind turbine blade airfoil.

[0015] The technical solution provided in this application embodiment can include the following beneficial effects: This application acquires input data, which includes environmental parameters and motion change data of the wind turbine blade airfoil; it calls an aerodynamic parameter prediction model and performs prediction based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil, wherein the aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method; and based on the predicted aerodynamic parameters, it draws the hysteresis loop of the wind turbine blade airfoil. This allows for the efficient and accurate solution of the aerodynamic parameters involved in the hysteresis loop using the aerodynamic parameter prediction model, and the drawing of the corresponding hysteresis loop, which is beneficial for improving the efficiency and accuracy of hysteresis loop calculation. At the same time, by leveraging the simplicity and universality of the aerodynamic parameter prediction model, it also solves technical problems existing in the prior art, such as low computational efficiency and high requirements for equipment computing power.

[0016] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings.

[0018] Figure 1 A flowchart illustrating a method for calculating the hysteresis loop of a wind turbine blade airfoil, provided in an embodiment of this application.

[0019] Figure 2 A flowchart illustrating the model training process provided in this application embodiment.

[0020] Figure 3 This is a schematic diagram of the internal structure of a deep neural network model provided in an embodiment of this application.

[0021] Figure 4 A schematic diagram of the flow field mesh provided for an embodiment of this application.

[0022] Figures 5A-5D A comparison chart showing the results of hysteresis loop plotting under different calculation methods provided in the embodiments of this application.

[0023] Figure 6 This is a schematic diagram of the structure of an aerodynamic parameter prediction device provided in an embodiment of this application.

[0024] Figure 7 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] Unless otherwise defined, the technical or scientific terms used in the embodiments of this specification shall have the ordinary meaning understood by one of ordinary skill in the art to which this specification pertains. The terms "first," "second," and similar terms used in the embodiments of this specification do not indicate any order, quantity, or importance, but are merely used to avoid confusion of constituent elements.

[0027] Unless the context otherwise requires, throughout this specification, "a plurality of" means "at least two," and "including" is interpreted as open-ended or encompassing, that is, "including, but not limited to." In the description of this specification, terms such as "one embodiment," "some embodiments," "exemplary embodiment," "example," "specific example," or "some examples" are intended to indicate that a particular feature, structure, material, or characteristic associated with that embodiment or example is included in at least one embodiment or example of this specification. The illustrative representations of the above terms do not necessarily refer to the same embodiment or example.

[0028] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0029] In the process of filing this application, the applicant also discovered that existing technologies can utilize artificial intelligence algorithms to solve the dynamic stall hysteresis loop of wind turbine blade airfoils. However, these algorithms are mainly geared towards specific research problems, with relatively simple research parameters and poor engineering versatility. To address these issues, this application proposes a method, apparatus, equipment, and medium for calculating the hysteresis loop of wind turbine blade airfoils.

[0030] Please see Figure 1 This is a flowchart illustrating a method for calculating the hysteresis loop of a wind turbine blade airfoil, as provided in an embodiment of this application. Figure 1 The method shown can be applied to computer devices, and the method may include the following implementation steps: S101. Obtain input data, which includes environmental parameters and motion change data of the wind turbine blade airfoil.

[0031] The input data in this application refers to the data used to input into the pre-trained aerodynamic parameter prediction model, including environmental parameters and motion variation data of the wind turbine blade airfoil. The aforementioned environmental parameters may include, but are not limited to, wind speed (Mach number), aerodynamic viscosity (Reynolds number), turbulence intensity, temperature, density, pressure, or other custom environmental parameters. The aforementioned motion variation data refers to information describing the time-varying motion state of the aforementioned wind turbine blade airfoil under corresponding operating conditions (such as unsteady flow conditions). For example, when the wind turbine blade airfoil pitches (blade torsion), heaves (blade oscillation), or moves back and forth (blade flapping), the corresponding displacement or generalized displacement, generalized velocity, generalized acceleration, generalized angular displacement, generalized angular velocity, or generalized angular acceleration at each moment. For instance, using the angle of attack at each moment under single-degree-of-freedom oscillation of the airfoil as the generalized displacement, the aforementioned motion variation data may include the angle of attack and the angular velocity and angular acceleration data of the angle of attack variation.

[0032] S102. Call the aerodynamic parameter prediction model and make predictions based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil; wherein, the aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method.

[0033] The aforementioned aerodynamic parameter prediction model is a neural network model pre-defined by the system based on actual conditions. This application does not limit the internal structure of the aforementioned aerodynamic parameter prediction model; taking a deep neural network model as an example, please refer to [link to relevant documentation]. Figure 3 This is a schematic diagram of a possible internal structure of a model provided in an embodiment of this application. For example... Figure 3 The deep neural network model shown includes an input layer module 301, a hidden layer module 302, and an output layer module 303. The input layer module 301 is primarily responsible for receiving input feature data, including environmental parameters and motion variation data of the wind turbine blade airfoil. The hidden layer module 302, consisting of at least one fully connected layer, is primarily responsible for extracting deep features related to the output data from the input data. The output layer module 303 is primarily responsible for mapping the calculation results of the hidden layer module to the final predicted aerodynamic parameters.

[0034] The aerodynamic parameters involved in this application may refer to aerodynamic parameters with hysteresis effects, which may include, but are not limited to, any combination of one or more of the following: lift coefficient, drag coefficient, moment coefficient (e.g., pitching moment coefficient), other generalized force coefficients or dimensionless coefficients, etc.

[0035] By implementing the embodiments of this application, input data is obtained, including environmental parameters and motion variation data of the wind turbine blade airfoil; an aerodynamic parameter prediction model is invoked, and prediction is performed based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil. The aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method; based on the predicted aerodynamic parameters, the hysteresis loop of the wind turbine blade airfoil is plotted. This allows for the efficient and accurate solution of the aerodynamic parameters involved in the hysteresis loop using the aerodynamic parameter prediction model, and the plotting of the corresponding hysteresis loop, thus improving the efficiency and accuracy of hysteresis loop calculation. Simultaneously, the simplicity and versatility of the aerodynamic parameter prediction model also solves technical problems existing in prior art solutions, such as low computational efficiency and high requirements for equipment computing power.

[0036] The following describes some specific and optional embodiments related to this application.

[0037] In step S101, this application does not limit the implementation method for obtaining the above-mentioned input data. For example, this application can first obtain the motion parameters of the wind turbine blade airfoil, then combine the above-mentioned motion parameters to obtain basic motion change data (e.g., the angle of attack at each moment in the embodiment), and then perform numerical differentiation calculation on the above-mentioned basic motion change data to obtain its higher-order derivative information (e.g., the angular velocity and angular acceleration of the angle of attack change at each moment in the embodiment), thereby constructing complete motion change data. This application does not limit the specific implementation method of the above-mentioned numerical differentiation calculation.

[0038] Before step S102, this application also requires pre-training the aforementioned aerodynamic parameter prediction model. The relevant embodiments involving the training of the aerodynamic parameter prediction model are described below. Please refer to... Figure 2 This is a schematic diagram of a model training process provided in an embodiment of this application. For example... Figure 2 The illustrated process can be applied to training devices, which may include, but are not limited to, computer equipment, terminal devices, servers, or other devices with model training capabilities. The process may include the following implementation steps: S201. Obtain a sample set, which includes a training set and a test set. The training set and the test set each include environmental parameters and motion change data of the wind turbine blade airfoil, as well as simulated aerodynamic parameters obtained through computational fluid dynamics simulation.

[0039] This application does not limit the implementation method of obtaining the above sample set. For example, this application first determines the various research parameters (hereinafter also referred to as sampling parameters) of the wind turbine blade airfoil, such as the environmental parameters and motion parameters mentioned above. Then, a corresponding sampling parameter space is constructed. This sampling parameter space can include the sampling range corresponding to each sampling parameter, such as the sampling range corresponding to wind speed, angle of attack oscillation amplitude, and oscillation reduction frequency. This application does not limit or elaborate on this. Next, Latin hypercube sampling is performed based on the above sampling parameter space to obtain m sampling points, where m is a positive integer predefined according to the actual situation. Specifically, this application can use the Latin hypercube sampling algorithm to sample the above sampling parameter space and use the Morris-Mitchell criterion to evaluate the sampling result index. This application does not limit or elaborate on this. Furthermore, this application can obtain motion change data based on the motion parameters of the above m samples, which can be referred to in the relevant description in the aforementioned step S101, and will not be repeated here. This application utilizes computational fluid dynamics (CFD) simulation software to solve for the operating conditions corresponding to the aforementioned m sample points (specifically, the environmental parameters and blade airfoil motion change data corresponding to the aforementioned m samples), obtaining the corresponding m sets of simulated aerodynamic parameters. Finally, this application defines the environmental parameters and motion change data corresponding to the aforementioned m samples, along with the simulated aerodynamic parameters obtained through CFD simulation, as the aforementioned sample set, and divides the aforementioned sample set into corresponding training and testing sets according to a preset ratio. This preset ratio is a pre-defined and configured ratio based on actual conditions; it can be an empirical value set based on user experience, such as 5:1 or 9:1, etc. This application does not impose further limitations or details on this.

[0040] S202. Based on the training set and the LM method, the aerodynamic parameter prediction model to be trained is trained and updated to obtain the trained aerodynamic parameter prediction model.

[0041] This application uses the aforementioned training set and the Levenberg-Marquardt (LM) method to iteratively train the aerodynamic parameter prediction model to obtain a trained aerodynamic parameter prediction model. This application does not limit the specific implementation method of the model training. For example, this application can input the environmental parameters, motion change data, and corresponding aerodynamic parameters of the training set into the aerodynamic parameter prediction model to be trained in batches, and use the LM method to train and update the model parameters, so that the error between the predicted aerodynamic parameters output by the aerodynamic parameter prediction model corresponding to the training set and the simulated aerodynamic parameters corresponding to the training set is minimized. In practical applications, this application can pre-set model hyperparameters, which may include, but are not limited to, parameters such as the number of network layers involved in the aerodynamic parameter prediction model, the number of hidden layer neurons, the learning rate, the batch size, and the number of training steps. Then, based on the aforementioned model hyperparameters, the aforementioned training set, and the LM method, the aerodynamic parameter prediction model to be trained is learned and trained to obtain a trained aerodynamic parameter prediction model.

[0042] S203. Based on the test set, the performance of the trained aerodynamic parameter prediction model is tested to obtain the corresponding evaluation index.

[0043] This application allows the test set to be input in batches into the trained aerodynamic parameter prediction model for test set error calculation, thereby obtaining the corresponding evaluation index for the model. This evaluation index is used to assess the quality of model training and further reflects the accuracy of model calculations. This application does not limit the specific implementation of the evaluation index, which may include, but is not limited to, mean squared error (MSEloss), mean absolute error (MAEloss), relative error (REloss), or other custom-defined index parameters used to reflect the model training status.

[0044] S204. When the evaluation index is less than the corresponding preset index threshold, the trained aerodynamic parameter prediction model is obtained.

[0045] After obtaining the aforementioned evaluation indicators, this application can further determine whether these indicators are less than the corresponding preset threshold values. If so, the trained aerodynamic parameter prediction model can be identified as a well-trained aerodynamic parameter prediction model, and the model training process can be terminated. Conversely, if the aforementioned evaluation indicators are greater than or equal to the corresponding preset threshold values, it can be determined that the model training accuracy is not high. The model can be retrained by adjusting the aforementioned hyperparameters until the model's evaluation indicators meet the requirements of actual engineering applications. The aforementioned preset threshold values ​​are error thresholds pre-defined by the system based on actual conditions. They can be empirical values ​​customized based on user experience or statistical values ​​calculated from a series of experimental data. This application does not impose further limitations or details on these.

[0046] To aid in a better understanding of the embodiments of this application, examples are provided below. This application selects the classic NACA0012 wind turbine blade airfoil as an example, and studies its single-degree-of-freedom oscillation hysteresis loop characteristics under varying wind speeds, varying oscillation amplitudes, and varying oscillation reduction frequencies, and combines this with… Figure 4 This is a schematic diagram of the flow field grid for a wind turbine blade airfoil provided in an embodiment of this application. For example... Figure 4 As shown, the flow field grid corresponding to the above-mentioned wind turbine blade airfoil is given. This application will not impose further limitations or details on this.

[0047] In this example, the angle of attack of the wind turbine blade airfoil at various moments under single-degree-of-freedom oscillation is taken as the generalized displacement. Considering that the change in angle of attack is determined by two factors—oscillation amplitude and oscillation reduction frequency—and adding wind speed variation, a three-dimensional parameter space [wind speed, oscillation amplitude, oscillation reduction frequency] can be constructed. Furthermore, this application can use the Latin hypercube sampling method to sample points in the known parameter space. For example, this application can collect 100 sampling points and randomly sort and number them. For each sampling point, its corresponding oscillation amplitude and oscillation reduction frequency can determine a unique angle of attack change signal. Then, for each sampling point, this application can use, for example, a fourth-order precision numerical differentiation method to calculate the higher-order derivatives of the angle of attack change signal corresponding to each sampling point, such as obtaining the first and second derivatives to obtain the angular velocity and angular acceleration of the aforementioned angle of attack change.

[0048] This application can also utilize the unsteady flow solution method built into the CFD software to sequentially solve unsteady aerodynamic forces for angle-of-attack motion signals at different wind speeds, extracting the time-domain response curves of lift coefficient, drag coefficient, and moment coefficient as the corresponding simulated aerodynamic parameters. Then, for each sampling point, the wind speed, angle of attack (or angle-of-attack motion signal), angular velocity, angular acceleration, and the solved simulated aerodynamic parameters are mapped one-to-one over time. The data structure can be represented as [wind speed, angle of attack, angular velocity, angular acceleration, lift coefficient, drag coefficient, moment coefficient] at different times. In practical applications, this application can concatenate data along the time dimension in numerical order to construct the aforementioned training set. The aforementioned training set and the aforementioned test set can be stored separately, thereby constructing the database required for model training and validation.

[0049] Next, an aerodynamic parameter prediction model is constructed according to the input and output dimensions of the training set. To improve the representation performance of the network model, the neural network can be designed as a single-output form, and multiple neural networks can be used to solve the multi-output problem. In this embodiment, the model has an input dimension of 4, and the corresponding input data includes wind speed, angle of attack, angular velocity, and angular acceleration; the output dimension is 1, so three neural networks can be designed in the aerodynamic parameter prediction model, and the corresponding output data includes lift coefficient, drag coefficient, and torque coefficient. Assume that the model hyperparameters set in this embodiment include: 2 network layers, 16 hidden layer neurons, a learning rate of 0.0005, a batch size of 32, and 2000 training steps.

[0050] This application allows for the sequential training of three neural networks in an aerodynamic parameter prediction model. During training, the training set is divided into training and validation data in an 8:2 ratio. The application can load training data into the model in batches and simultaneously update the network parameters. In the current training step, after loading the training data, validation data can be loaded into the model. Mean squared error, mean absolute error, and relative error are selected as evaluation metrics to calculate the three errors between the predicted and simulated aerodynamic parameters output by the model. The training is considered complete based on whether the three errors meet the corresponding error thresholds. Training is terminated only when the corresponding error thresholds are met or the maximum number of training steps is reached.

[0051] After training the aforementioned model, this application can use a test set to verify the model's performance. For example, the input data from 10 test sets—that is, an input matrix composed of wind speed, angle of attack, angular velocity, and angular acceleration—can be sequentially input into the trained aerodynamic parameter prediction model. The corresponding predicted aerodynamic parameters are calculated, and then compared with the corresponding simulated aerodynamic parameters to calculate the aforementioned three types of errors. The average of the three errors is then taken. If the average error index of the above three errors is less than the corresponding error threshold, the model training is considered complete; otherwise, the process must return to the network design step, such as readjusting the model hyperparameters and retraining the model according to the aforementioned model training principles. Optionally, this application can save the aerodynamic parameter prediction model verified through the aforementioned test set, i.e., the trained aerodynamic parameter prediction model, for subsequent practical engineering applications. Please combine this with... Figures 5A-5D The diagram shows a comparison of the hysteresis loop plotting results under different calculation methods. In the diagram, A... m This represents the amplitude of the angle-of-attack oscillation, and k represents the oscillation reduction frequency. For example... Figure 5AAs shown, the horizontal axis represents the angle of attack, and the vertical axis represents the lift coefficient. The figure compares the proposed solution (deep neural network model (DNN) and LM method) with the traditional computational fluid dynamics solution (software CFD). It can be determined that the proposed solution can also accurately calculate and predict the corresponding lift coefficient of the wind turbine blade airfoil.

[0052] like Figure 5B As shown, the horizontal axis represents the angle of attack, and the vertical axis represents the moment coefficient. The figure compares the proposed scheme (DNN model and LM method) with the traditional computational fluid dynamics scheme, and it can be determined that the proposed scheme can also accurately calculate and predict the corresponding moment coefficient of the wind turbine blade airfoil.

[0053] like Figure 5C As shown, the horizontal axis represents the angle of attack, and the vertical axis represents the lift coefficient. The figure compares the prediction accuracy of the aerodynamic parameter prediction model DNN using the first-order gradient descent method and the second-order gradient descent method, respectively. It is clear that the second-order gradient descent method has higher accuracy.

[0054] like Figure 5D As shown, the horizontal axis represents the angle of attack, and the vertical axis represents the lift coefficient. The figure compares the prediction accuracy of the aerodynamic parameter prediction model DNN using the second-order gradient descent method and the LM method, respectively. It is clear that the LM method has higher accuracy; that is, the aerodynamic parameter prediction model that incorporates the LM method in this application has higher accuracy.

[0055] Based on the figures above, compared to traditional computational fluid dynamics methods, the proposed solution exhibits significant advantages in both efficiency and computational accuracy for predicting aerodynamic parameters. Furthermore, experiments have shown that the proposed solution requires less average time to solve for aerodynamic parameters, and the average time required for model training using the LM method is also shorter. Please refer to Table 1 below for the average solution time required for solving the aerodynamic parameters involved in the hysteresis loop using different methods.

[0056] Table 1 Table 2 below shows the average time required to train the model using gradient descent (gdx) and LM methods.

[0057] Table 2 As can be seen, this application establishes an aerodynamic parameter prediction model based on a neural network, replacing the traditional computational fluid dynamics method for rapidly solving aerodynamic parameters related to hysteresis loops. Leveraging the model's simplicity and adaptability to strongly nonlinear problems, it achieves the goal of high accuracy and high efficiency. This application establishes an aerodynamic parameter prediction model based on neural networks, which have easier-to-understand fundamental theories, simpler network structures, and greater versatility. It further improves the model by adjusting data as much as possible, reducing modifications to the model structure and the complexity of model parameters, thus lowering the difficulty of using the aerodynamic parameter prediction model. The Latin hypercube sampling method is used to treat all parameters involved in the hysteresis loop equally and uniformly, thereby constructing an scalable parameter space. Training and test sets are obtained from this parameter space, achieving the goal of simultaneously considering multiple research parameters. This results in a more comprehensive characterization of the hysteresis loop of the wind turbine blade airfoil, enhancing the engineering versatility of the solution. Incorporating wind turbine blade airfoil motion variation data into the model training input enables the model to better identify the nonlinear characteristics of the hysteresis loop, improving the model's prediction accuracy. Furthermore, considering the simplicity of the model structure and the limited training data, this application uses the LM method instead of the commonly used gradient descent method to train the model, thereby improving the training efficiency of the model.

[0058] By implementing the embodiments of this application, input data is obtained, including environmental parameters and motion variation data of the wind turbine blade airfoil; an aerodynamic parameter prediction model is invoked, and prediction is performed based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil. The aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method; based on the predicted aerodynamic parameters, the hysteresis loop of the wind turbine blade airfoil is plotted. This allows for the efficient and accurate solution of the aerodynamic parameters involved in the hysteresis loop using the aerodynamic parameter prediction model, and the plotting of the corresponding hysteresis loop, thus improving the efficiency and accuracy of hysteresis loop calculation. Simultaneously, the simplicity and versatility of the aerodynamic parameter prediction model also solves technical problems existing in prior art solutions, such as low computational efficiency and high requirements for equipment computing power.

[0059] Based on the above embodiments, please refer to Figure 6 This is a schematic diagram of the structure of an aerodynamic parameter prediction device provided in an embodiment of this application. Figure 6 The illustrated device can be applied to a computer device, and the device may include an acquisition module 601 and a processing module 602, wherein: The acquisition module 601 is used to acquire input data, which includes environmental parameters and motion change data of the wind turbine blade airfoil. The processing module 602 is used to call the aerodynamic parameter prediction model and make predictions based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil, and then draw the hysteresis loop of the wind turbine blade airfoil.

[0060] Please see Figure 7 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 7 The devices shown can be vehicles, mobile phones, computers, digital broadcasting terminals, messaging devices, game consoles, tablets, medical devices, fitness equipment, personal digital assistants, etc.

[0061] Reference Figure 7 The device 700 may include one or more of the following components: processing component 702, memory 704, power supply component 706, multimedia component 708, audio component 710, input / output interface 712, sensor component 714, and communication component 716.

[0062] Processing component 702 typically controls the overall operation of device 700, such as operations associated with display, telephone calls, data communication, camera operation, and recording. Processing component 702 may include one or more processors 720 to execute instructions to complete all or part of the steps of the aforementioned method for calculating the hysteresis loop of the wind turbine blade airfoil. Furthermore, processing component 702 may include one or more modules to facilitate interaction between processing component 702 and other components. For example, processing component 702 may include a multimedia module to facilitate interaction between multimedia component 708 and processing component 702.

[0063] Memory 704 is configured to store various types of data to support the operation of device 700. Examples of this data include instructions for any application or method operating on device 700, contact data, phonebook data, messages, pictures, videos, etc. Memory 704 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0064] Power supply component 706 provides power to various components of device 700. Power supply component 706 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to device 700.

[0065] Multimedia component 708 includes a screen that provides an output interface between the device 700 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 708 includes a front-facing camera and / or a rear-facing camera. When the device 700 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0066] Audio component 710 is configured to output and / or input audio signals. For example, audio component 710 includes a microphone (MIC) configured to receive external audio signals when device 700 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 704 or transmitted via communication component 716. In some embodiments, audio component 710 also includes a speaker for outputting audio signals.

[0067] Input / output interface 712 provides an interface between processing component 702 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, start buttons, and lock buttons.

[0068] Sensor assembly 714 includes one or more sensors for providing state assessments of various aspects of device 700. For example, sensor assembly 714 may detect the on / off state of device 700, the relative positioning of components such as the display and keypad of device 700, changes in the position of device 700 or a component of device 700, the presence or absence of user contact with device 700, the orientation or acceleration / deceleration of device 700, and temperature changes of device 700. Sensor assembly 714 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 714 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 714 may also include an accelerometer, a gyroscope, a magnetometer, a pressure sensor, or a temperature sensor.

[0069] Communication component 716 is configured to facilitate wired or wireless communication between device 700 and other devices. Device 700 can access wireless networks based on communication standards, such as WiFi, 2G, or 3G, or combinations thereof. In one exemplary embodiment, communication component 716 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 716 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0070] In an exemplary embodiment, device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the hysteresis loop calculation method for the wind turbine blade airfoil described above.

[0071] Understandably, the processor 720 in this application embodiment can be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method embodiment can be completed by integrated logic circuits in the processor's hardware or by software instructions. The processor described above can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0072] Understandably, the memory 704 in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0073] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 704 including instructions, which can be executed by the processor 720 of the device 700 to complete the above-described method for calculating the hysteresis loop of the upper-level wind turbine blade airfoil. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0074] The aforementioned device can be a standalone electronic device or a part of a standalone electronic device. For example, in one embodiment, the device can be an integrated circuit (IC) or a chip, wherein the integrated circuit can be a single IC or a collection of multiple ICs. The chip can include, but is not limited to, the following types: GPU (Graphics Processing Unit), CPU (Central Processing Unit), FPGA (Field Programmable Gate Array), DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), and SOC (System on Chip). The aforementioned integrated circuit or chip can be used to execute executable instructions (or code) to implement the aforementioned method for calculating the hysteresis loop of the wind turbine blade airfoil. The executable instructions can be stored in the integrated circuit or chip or obtained from other devices or equipment. For example, the integrated circuit or chip includes a processor, memory, and an interface for communicating with other devices. The executable instruction can be stored in the memory, and when the executable instruction is executed by the processor, it implements the above-mentioned method for calculating the hysteresis loop of the wind turbine blade airfoil; or, the integrated circuit or chip can receive the executable instruction through the interface and transmit it to the processor for execution to implement the above-mentioned method for calculating the hysteresis loop of the wind turbine blade airfoil.

[0075] In another exemplary embodiment, a computer program product is also provided, which includes a computer program executable by a programmable device, the computer program having a code portion for performing the hysteresis loop calculation method for the wind turbine blade airfoil described above when executed by the programmable device.

[0076] It should be noted that the descriptions of the above embodiments of storage media, devices, and equipment are similar to the descriptions of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the embodiments of storage media, devices, and equipment of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0077] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of this application. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed in this application. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0078] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications or equivalent substitutions made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for calculating the hysteresis loop of a wind turbine blade airfoil, characterized in that, include: Acquire input data, which includes environmental parameters and motion change data of the wind turbine blade airfoil; The aerodynamic parameter prediction model is invoked and prediction is performed based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil. The aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method. Based on the predicted aerodynamic parameters, the hysteresis loop of the wind turbine blade airfoil is plotted.

2. The method according to claim 1, characterized in that, The acquisition of input data includes: Obtain environmental parameters and motion change data of the wind turbine blade airfoil.

3. The method according to claim 1, characterized in that, The method further includes: Obtain a sample set, which includes a training set and a test set. The training set and the test set each include environmental parameters, motion change data and corresponding simulated aerodynamic parameters of the wind turbine blade airfoil. Based on the training set and the LM method, the aerodynamic parameter prediction model to be trained is trained and updated to obtain the trained aerodynamic parameter prediction model. The performance of the trained aerodynamic parameter prediction model is tested based on the test set to obtain the corresponding evaluation index. When the evaluation index is less than the corresponding preset index threshold, the trained aerodynamic parameter prediction model is obtained.

4. The method according to claim 3, characterized in that, The parameter training and updating of the aerodynamic parameter prediction model to be trained based on the training set and the LM method includes: The environmental parameters and motion change data in the training set are input into the aerodynamic parameter prediction model to be trained in batches; the LM method is used to train and update the parameters of the aerodynamic parameter prediction model to be trained, so that the loss value between the predicted aerodynamic parameters corresponding to the training set and the corresponding simulated aerodynamic parameters is less than a preset loss threshold, thereby obtaining the trained aerodynamic parameter prediction model.

5. The method according to claim 3, characterized in that, The acquisition of the sample set includes: A sampling parameter space is constructed, which includes the sampling range corresponding to each sampling parameter. The sampling parameters include the environmental parameters and motion parameters of the wind turbine blade airfoil. Latin hypercube sampling is performed based on the sampling parameter space to obtain m sampling points, where m is a positive integer; Based on the motion parameters and numerical differential calculations of the m sampling points, the corresponding motion change data are obtained. Based on the environmental parameters and motion change data of the aforementioned wind turbine blade airfoil, simulated aerodynamic parameters are obtained through computational fluid dynamics simulation. The environmental parameters, motion change data, and corresponding simulated aerodynamic parameters of m sampling points are determined as the sample set, and the sample set is divided into a training set and a test set.

6. The method according to any one of claims 1-5, characterized in that, The environmental parameters include the wind speed of the wind turbine blade airfoil, the motion parameters include the angle of attack oscillation amplitude and oscillation reduction frequency of the wind turbine blade airfoil, and the motion change data include the time series data of the angle of attack change and its corresponding angular velocity and angular acceleration data.

7. The method according to any one of claims 1-5, characterized in that, The aerodynamic parameters include at least one of the lift coefficient, drag coefficient, and moment coefficient.

8. A device for predicting aerodynamic parameters, characterized in that, include: The acquisition module is used to acquire input data, which includes environmental parameters and motion change data of the wind turbine blade airfoil; The processing module is used to call the aerodynamic parameter prediction model and make predictions based on the input data to obtain the predicted aerodynamic parameters corresponding to the wind turbine blade airfoil. The aerodynamic parameter prediction model is a deep neural network model pre-trained based on the LM method. The processing module is also used to draw the hysteresis loop of the wind turbine blade airfoil based on the predicted aerodynamic parameters.

9. A computer device, characterized in that, include: A processor, and a memory for storing processor-executable instructions; wherein the processor is configured to execute the executable instructions to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 7.

Images