A scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions

Abstract

This invention provides a scaled-down test device for studying wind turbine blade flutter under natural wind conditions. Belonging to the field of wind turbine blade flutter testing, this scaled-down test device for studying wind turbine blade flutter under natural wind conditions is characterized by comprising a scaled-down wind turbine model consisting of a wind turbine body, a control system, and a monitoring device. The wind turbine body includes a foundation, tower, and blades. The control system controls the wind turbine speed, obtaining the speed of the scaled-down model corresponding to the incoming wind speed using a similarity criterion to achieve speed adjustment. The monitoring device monitors the natural wind speed and performs non-contact monitoring of the scaled-down blade flutter morphology. The control system includes a pitch control device, a yaw device, a servo motor, and a controller. The monitoring device includes an anemometer, a wind direction sensor, a high-speed camera, and a six-component balance. This invention solves the problem of not being able to reflect actual blade flutter changes under natural wind conditions.

Description

Technical Field

[0001] This invention pertains to wind turbine blade flutter testing, specifically, it relates to a scaled-down test device for studying wind turbine blade flutter under natural wind conditions. Background Technology

[0002] Using scaled-down models for testing is one of the main methods for studying wind turbine blade flutter. Existing scaled-down blade flutter studies almost always involve testing a single blade vertically fixed in a wind tunnel. This method has the following problems: First, the pre-twist of the wind turbine blade's cross-section is designed based on relative wind speed. Vertically fixed blades have no rotational motion; only the absolute wind speed of the airflow in the wind tunnel is present, resulting in significant differences in surface flow characteristics compared to actual blades. Second, the wind tunnel airflow cannot reflect the turbulence and randomness of natural wind.

[0003] Current flutter testing equipment for scaled-down models of wind turbine blades, such as Figure 1 As shown, the blade root is fixed in a wind tunnel, and the airflow velocity inside the tunnel is monitored using a hot-wire anemometer. A high-speed camera observes the blade flutter morphology. A six-component balance measures the blade root reaction force to reflect the flutter time-domain signal, which can be converted into a frequency-domain signal using Fourier transform. The airflow velocity of the blade under these wind tunnel test conditions is as follows: Figure 2 As shown, the speeds are of the same magnitude and direction. The airflow speed of the blades under natural wind conditions is as follows: Figure 3 As shown, the incoming wind speed v0 and the relative wind speed v generated by the rotation are... r The resulting wind speeds vary in both magnitude and direction. Therefore, existing experimental setups suffer from this limitation, resulting in an inability to accurately reflect actual blade flutter changes under natural wind conditions. Summary of the Invention

[0004] In view of this, the present invention provides a scaled-down test device for studying wind turbine blade flutter under natural wind conditions, which can solve the problem of not being able to reflect the actual blade flutter changes under natural wind conditions.

[0005] This invention is implemented as follows:

[0006] This invention provides a scaled-down test device for studying wind turbine blade flutter under natural wind conditions. The device includes a scaled-down model of a wind turbine consisting of a main body, a control system, and a monitoring device. The main body includes a foundation, a tower, and blades. The control system controls the wind turbine's rotational speed, using a similarity criterion to obtain the rotational speed of the scaled-down model corresponding to the incoming wind speed, thus achieving speed adjustment. The monitoring device monitors the natural wind speed and performs non-contact monitoring of the scaled-down blade flutter morphology.

[0007] Based on the above technical solution, the scaled-down test device for studying wind turbine blade flutter under natural wind conditions according to the present invention can be further improved as follows:

[0008] The blades used are scaled-down models of IEA 15MW wind turbine blades. To construct a scaled-down model with a response similar to the original blades, the following aerodynamic scaling criteria are established: the basic similarity ratios for wind turbine blade aerodynamic similarity include the Strauhal number St, Froude number Fr, Euler number Eu, and Reynolds number Re, and the following rules apply:

[0009] (1) If the only mass force is gravity, Fr and Eu can be equivalent;

[0010] (2) Fr and Re cannot be satisfied simultaneously;

[0011] (3) When Re is greater than 4×10 5 After that, it satisfies self-modeling property.

[0012] Therefore, the Re number must satisfy the self-modularity, and (1)

[0013] in:

[0014] (2)

[0015] Considering the relative velocity of wind turbine blades, the airfoils at different spanwise positions are designed with pre-twisting to achieve the optimal lift-to-drag ratio angle of attack under rated operating conditions. To ensure a constant blade pitch angle while maintaining the blade aerodynamic shape, the following relationship must be satisfied:

[0016] (3)

[0017] Where, λ ωrot Let be the blade speed ratio; Equation (1) is a commonly used criterion for blade aerodynamic similarity. The two equations contain four variables, which are uncertain systems; by introducing Equations (2) and (3), they are transformed into deterministic systems.

[0018] The control system includes a pitch control device, a yaw device, a servo motor, and a controller.

[0019] The beneficial effects of adopting the above-mentioned improvement scheme are: the controller controls the pitch device and yaw device based on the wind speed and wind direction measured by the anemometer and wind direction sensor.

[0020] Furthermore, the monitoring devices include an anemometer, wind direction sensor, high-speed camera, and six-component balance.

[0021] Furthermore, the pitch control device is used to control the blades to achieve the following control effects:

[0022] (1) 3m / s~6.98m / s: the minimum rotor speed. As the wind speed increases, the blade pitch angle gradually decreases to 0°.

[0023] (2) 6.98m / s~10.58m / s: the optimal tip speed ratio. As the wind speed increases, the blade pitch angle remains at 0°, and the rotational speed and torque increase proportionally.

[0024] (3) 10.58m / s~25m / s: The controller adjusts the blade pitch angle to increase, so that the blade lift remains unchanged, thereby maintaining a constant rotational speed;

[0025] (4) Greater than 25m / s: Brake to ensure the safety of the wind turbine.

[0026] Furthermore, the foundation is buried below ground level to serve as the supporting base for the entire device. The foundation is fixedly connected to the bottom of the tower with bolts to ensure the verticality and stability of the tower. The top of the tower is connected to the wind turbine hub via flanges, and blades are installed on the hub. The roots of the blades are connected to the hub with bolts, and the pitch control device is located inside the hub.

[0027] Furthermore, the servo motor is connected to the transmission shaft inside the tower via a coupling, driving the hub to rotate. The servo motor is connected to the controller via an electrical control signal line to receive speed commands.

[0028] Furthermore, the pitch control device drives the blade root to rotate via an electric actuator; the yaw device engages with the flange at the top of the tower via a gear set to adjust the horizontal angle of the hub.

[0029] Furthermore, the anemometer is fixedly installed at the top of the tower, the wind direction sensor is fixedly installed on the side of the tower, and the controller is used to receive real-time data from the anemometer and the wind direction sensor; and to send control commands to the servo motor, pitch control device, and yaw device via RS485 bus.

[0030] Furthermore, the six-component balance is embedded between the blade root and the hub flange, with its surface directly contacting the blade root.

[0031] Furthermore, the visual recognition method for non-contact monitoring of blade flutter morphology using the monitoring device includes the following steps:

[0032] S10: Records 1000Hz video using a high-speed camera and extracts the video frame by frame;

[0033] S20: Based on the neural network, the position of the blade in each frame of the image is identified and recorded to obtain the time domain response of the blade displacement. If the displacement is too large, it is determined that the blade is fluttering, and the wind speed and flutter frequency at this time are recorded.

[0034] S30: The time-domain response curves of velocity and acceleration can be obtained by taking the first and second derivatives of the time-domain response of the blade displacement;

[0035] S40: By performing a Fourier transform on the time-domain response curve of the acceleration, the frequency response of the blade and its natural frequency and other characteristics can be obtained.

[0036] Compared with existing technologies, the beneficial effects of the scaled-down test device for studying wind turbine blade flutter under natural wind conditions provided by this invention are as follows: By incorporating a pre-torsion angle in the aerodynamic shape design of the blade, the cross-sectional airfoil at each spanwise position under rated wind speed is guaranteed to have the optimal angle of attack. To ensure that the aerodynamic characteristics of the scaled-down blade test are similar to those of the actual blade, this invention establishes a similarity criterion; placing the test device under natural wind conditions ensures that the turbulence and randomness of the incoming wind are the same as in reality; by monitoring the natural wind speed with an anemometer, and based on the aerodynamic similarity criterion proposed in this invention, the controller matches the servo motor speed corresponding to the natural wind speed to realize the natural wind test. Compared with wind tunnel testing, the advantages are high realism and lower cost. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a schematic diagram of a wind tunnel test setup for a scaled-down model of a wind turbine blade in the background art.

[0039] Figure 2 This is a schematic diagram of the airflow velocity of the blades under wind tunnel testing conditions in the background art.

[0040] Figure 3 This is a schematic diagram of the airflow velocity of a blade under natural wind conditions in the background art;

[0041] Figure 4 This is a schematic diagram of a scaled-down test device for studying wind turbine blade flutter under natural wind conditions.

[0042] The attached diagram lists the components represented by each number as follows:

[0043] 10. Wind turbine body; 101. Tower; 102. Blades; 11. Pitch control device; 12. Yaw device; 13. Servo motor; 14. Controller; 15. Anemometer; 16. High-speed camera; 17. Six-component balance. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0045] like Figure 4 The image shows a first embodiment of a scaled-down test device for studying wind turbine blade flutter under natural wind conditions, provided by the present invention. In this embodiment, a scaled-down wind turbine model is included, consisting of a wind turbine body 10, a control system, and a monitoring device. The wind turbine body includes a foundation, a tower 101, and blades 102. The control system is used to control the wind turbine speed. The speed of the scaled-down wind turbine model corresponding to the incoming wind speed is obtained by using similarity criteria to achieve speed adjustment of the wind turbine. The monitoring device is used to monitor the natural wind speed and to perform non-contact monitoring of the scaled-down blade flutter morphology.

[0046] In the above technical solution, the blade uses a scaled-down model of the IEA 15MW wind turbine blade. To construct a scaled-down model with a response similar to the original blade, the aerodynamic scaling criteria are as follows: the basic similarity ratios for wind turbine blade aerodynamic similarity include the Strouhal number (St), Froude number (Fr), Euler number (Eu), and Reynolds number (Re), and the following rules apply:

[0047] (1) If the only mass force is gravity, Fr and Eu can be equivalent;

[0048] (2) Fr and Re cannot be satisfied simultaneously;

[0049] (3) When Re is greater than 4×10 5 After that, it satisfies self-modeling property.

[0050] Therefore, the Re number must satisfy the self-modularity, and (1)

[0051] in

[0052] (2)

[0053] Considering the relative velocity of wind turbine blades, the airfoils at different spanwise positions are designed with pre-twisting to achieve the optimal lift-to-drag ratio angle of attack under rated operating conditions. To ensure a constant blade pitch angle while maintaining the blade aerodynamic shape, the following relationship must be satisfied:

[0054] (3)

[0055] Where, λ ωrot Let L be the blade rotation speed ratio. Equation (1) is a commonly used criterion for blade aerodynamic similarity. The two equations contain four variables, making it an uncertain system. By introducing equations (2) and (3), it is transformed into a deterministic system. Based on this, the aerodynamic similarity ratio of the wind turbine can be determined, as shown in Table 1. Where L is the proportionality constant of the length ratio.

[0056] Table 1 Aerodynamic similarity ratio of wind turbines

[0057]

[0058] In the above technical solution, the control system includes a pitch device 11, a yaw device 12, a servo motor 13, and a controller 14.

[0059] Furthermore, in the above technical solution, the monitoring device includes an anemometer 15, a wind direction sensor, a high-speed camera 16, and a six-component balance 17.

[0060] Furthermore, in the above technical solution, the pitch control device is used to control the blades to achieve the following control effects:

[0061] (1) 3m / s~6.98m / s: the minimum rotor speed. As the wind speed increases, the blade pitch angle gradually decreases to 0°.

[0062] (2) 6.98m / s~10.58m / s: the optimal tip speed ratio. As the wind speed increases, the blade pitch angle remains at 0°, and the rotational speed and torque increase proportionally.

[0063] (3) 10.58m / s~25m / s: The controller adjusts the blade pitch angle to increase, so that the blade lift remains unchanged, thereby maintaining a constant rotational speed;

[0064] (4) Greater than 25m / s: Brake to ensure the safety of the wind turbine.

[0065] Furthermore, in the above technical solution, the foundation is buried below the ground surface as the supporting foundation for the entire device. The foundation is fixedly connected to the bottom of the tower with bolts to ensure the verticality and stability of the tower. The top of the tower is connected to the wind turbine hub through a flange, and blades are installed on the hub. The root of the blade is connected to the hub with bolts, and the pitch control device is located inside the hub.

[0066] Furthermore, in the above technical solution, the servo motor is connected to the transmission shaft inside the tower via a coupling to drive the hub to rotate. The servo motor is connected to the controller via an electrical control signal line to receive speed commands.

[0067] Furthermore, in the above technical solution, the pitch control device drives the blade root to rotate via an electric actuator; the yaw device engages with the flange at the top of the tower via a gear set to adjust the horizontal angle of the hub.

[0068] Furthermore, in the above technical solution, the anemometer is fixedly installed at the top of the tower, the wind direction sensor is fixedly installed on the side of the tower, and the controller is used to receive real-time data from the anemometer and the wind direction sensor; and to send control commands to the servo motor, pitch control device, and yaw device via RS485 bus.

[0069] The control commands include speed control adjustment commands, pitch angle control adjustment commands, and yaw angle control adjustment commands.

[0070] Furthermore, the anemometer and wind direction sensor are installed in accordance with the requirements of GB / T35221-2017 (Specification for Ground Meteorological Observation).

[0071] Furthermore, in the above technical solution, the six-component balance is embedded between the blade root and the hub connecting flange, and its surface directly contacts the blade root.

[0072] The above settings are used to achieve the goal of controlling the wind turbine's rotation speed via a servo motor. The rotation speed of the scaled-down model of the wind turbine is obtained by constructing a similarity criterion. Therefore, the wind turbine's rotation is not a free rotation under the lift of natural wind, but is controlled by a servo motor to simulate the actual blade flutter changes under natural wind conditions.

[0073] Furthermore, a high-speed camera is used to achieve non-contact monitoring of the flutter morphology of scaled-down blades. This allows the visual analysis results to be compared and complemented by a six-component balance located at the blade root, thereby ensuring the accuracy of the experimental results.

[0074] Furthermore, in the above technical solution, the visual recognition method for non-contact monitoring of blade flutter morphology by the monitoring device includes the following steps:

[0075] S10: Records 1000Hz video using a high-speed camera and extracts the video frame by frame;

[0076] The specific steps in S10 include:

[0077] The first step is to record video. While the blades are operating normally, start a high-speed camera and set the frame rate to 1000Hz to begin recording. The recording time is determined based on actual needs and can be several minutes or even longer to cover the blade movement under different operating conditions.

[0078] Step 2: After the video recording is complete, use video processing software (such as Adobe Premiere Pro, FFmpeg, etc.) to extract the video frame by frame. If using the FFmpeg command-line tool, extract the video into image files at a rate of 1 frame per second. You can set the extraction frame rate according to your actual needs.

[0079] S20: Based on the neural network, the position of the blade in each frame of the image is identified and recorded to obtain the time domain response of the blade displacement. If the displacement is too large, it is determined that the blade is fluttering, and the wind speed and flutter frequency at this time are recorded.

[0080] The specific steps in S20 include:

[0081] The first step involves training a neural network by collecting a large amount of image data containing different blade positions and motion states. This data should cover images of blades under different conditions, such as normal operation and flutter. The neural network architecture chosen is a convolutional neural network (CNN), such as AlexNet, VGGNet, or ResNet. Taking ResNet as an example, it has the characteristic of residual learning, which can effectively solve the gradient vanishing and gradient exploding problems in deep networks, improving the model's training effect. The collected image data is then used to train the neural network by setting appropriate training parameters, such as the learning rate and the number of iterations. Through training, the neural network can accurately identify the position of the blades in the images.

[0082] The second step is blade position recognition. The images extracted frame by frame in step S10 are input into the trained neural network. Through the calculation and analysis of the neural network, the position information of the blade in each frame image is determined. The position coordinates of the blade in each frame image (such as the x and y coordinates in the image coordinate system) are recorded, and a sequence of blade position changes over time is formed according to the time order of the frames, that is, the time domain response of blade displacement.

[0083] The third step is flutter detection. A displacement threshold is set, and when the blade displacement exceeds this threshold, flutter is determined to have occurred. For example, based on the blade's design parameters and the displacement range during normal operation, a displacement threshold can be set, such as a certain percentage of the blade's maximum allowable displacement. When flutter is detected, the wind speed and flutter frequency are recorded simultaneously. The wind speed can be measured in real time using an installed anemometer, and the flutter frequency can be obtained by analyzing the time-domain response of the blade displacement.

[0084] S30: The time-domain response curves of velocity and acceleration can be obtained by taking the first and second derivatives of the time-domain response of the blade displacement;

[0085] The specific steps of S30 include:

[0086] The first step is data preprocessing, which involves preprocessing the time-domain response data of the blade displacement, such as removing noise and filling in missing data. Filtering algorithms (such as Gaussian filtering, median filtering, etc.) can be used to smooth the data to reduce the impact of noise on the subsequent derivative results.

[0087] The second step is to calculate the velocity by first-order differentiation. Based on the numerical calculation method, the time-domain response of the blade displacement is differentiated by first-order differentiation.

[0088] Preferably, the finite difference method is used for discrete displacement data sequences. , its time velocity at that point It can be approximated as ,in The time interval is defined as . By performing the above calculations on the displacement data sequence, the time-domain response data sequence of velocity is obtained.

[0089] The third step is to calculate the acceleration by taking the second derivative of the obtained velocity time-domain response data sequence.

[0090] Preferably, the finite difference method is also used for acceleration. It can be approximated as The time-domain response data sequence of acceleration is obtained through calculation.

[0091] The fourth step is to plot the curves based on the time-domain response data sequence of velocity and acceleration.

[0092] Preferably, data visualization tools (such as Matplotlib, Origin, etc.) are used to plot the time-domain response curves of velocity and acceleration; when plotting the curves, the horizontal axis is time, and the vertical axis is velocity and acceleration, respectively, to intuitively show the changes of blade velocity and acceleration over time.

[0093] S40: By performing a Fourier transform on the time-domain response curve of the acceleration, the frequency response of the blade and its natural frequency and other characteristics can be obtained.

[0094] The specific steps of S40 include:

[0095] The first step is to process the time-domain response curve of the acceleration using the Fourier transform algorithm.

[0096] Preferably, for discrete acceleration time-domain response data sequences The Fast Fourier Transform (FFT) algorithm can be used; for example, the NumPy library in Python provides the `fft` function for convenient fast Fourier transform. Through Fourier transform, the time-domain acceleration data is converted into frequency-domain data, yielding the acceleration spectrum.

[0097] The second step is frequency response analysis. Based on the obtained acceleration spectrum, the frequency response of the blade is analyzed. The frequency response represents how the blade responds to acceleration at different frequencies. In the spectrum, the amplitude and phase information of different frequency components can be observed, thus understanding the vibration characteristics of the blade at different frequencies.

[0098] Step 3: Determine the natural frequency. In the frequency response, find the peak positions in the spectrum. The frequencies corresponding to these peaks are the natural frequencies of the blade.

[0099] Among them, the natural frequency is the vibration characteristic of the blade itself. By determining the natural frequency, we can assess whether the blade resonates, which provides an important reference for the design and operation of the blade.

[0100] Furthermore, the analysis results of visual recognition can be compared and complemented with the six-component balance located at the leaf root to ensure the accuracy of the results.

[0101] The specific steps are as follows:

[0102] The first step is to collect data using a six-component balance. This involves collecting real-time data on the forces and torques acting on the blades during operation using a six-component balance installed at the blade root. This includes forces along the three coordinate axes and torques around the three coordinate axes.

[0103] The second step is data comparison, which compares the information such as blade displacement, velocity, and acceleration obtained by visual recognition with the time-domain data collected by the six-component balance.

[0104] Preferably, the accuracy of the visual recognition results is verified by analyzing the relationship between the blade displacement and the force applied. If there is a significant difference between the two sets of data, possible causes are analyzed, such as measurement errors or model assumptions.

[0105] Step 3: Complementary Results. Based on the comparison results, the data from visual recognition and the six-component balance are complemented.

[0106] The following is a specific application example:

[0107] In a laboratory setting, to study the flutter characteristics of wind turbine blades under natural wind conditions, it is necessary to construct an experimental system that simulates the natural wind environment and reproduces the flutter state of real wind turbine blades. By introducing aerodynamic scaling criteria and controlling the rotor speed and pitch control device with a servo motor, the effect of natural wind on the blades can be reproduced on a scaled-down experimental model. This allows for safer and more efficient blade flutter monitoring research, providing a reference for the design and operation of actual wind turbines.

[0108] First, a scaled-down model of the wind turbine is constructed, and the rotation speed of the wind turbine (hub and blades) is controlled by a servo motor. The rotation speed of the wind turbine is determined according to the proposed aerodynamic scaling criteria, as shown in Table 1.

[0109] Furthermore, the model was applied to simulation tests under natural wind conditions, and included a pitch system that could reflect the actual operating conditions of the blades.

[0110] The pitch control device is mounted on the hub and is used to connect the blades. It adopts an electric pitch control system from Moog, with a pitch angle adjustment range of 0°-90°, an adjustment accuracy of 0.1°, and a maximum adjustment speed of 15° / s. It can quickly and accurately change the blade pitch angle and is equipped with a high-precision angle sensor to provide real-time feedback of pitch angle data.

[0111] The monitoring device uses a Phantom v211 high-speed camera with a frame rate of 1000Hz, a resolution of 1280×800, and a pixel size of 12μm×12μm. A Young 81000 high-precision ultrasonic anemometer is installed with a measurement accuracy of ±0.1m / s and a measurement range of 0-60m / s. A Kistler 9357B six-component balance is installed at the blade root with a force measurement accuracy of ±0.1%FS and a torque measurement accuracy of ±0.2%FS, used to measure the force and torque on the blade.

[0112] After determining the aerodynamic scaling parameters, the assembly of the experimental setup began:

[0113] The foundation is buried below ground level, serving as the support base for the entire device. It is bolted to the bottom of the tower to ensure its verticality and stability. The top of the tower connects to the wind turbine hub via flanges, and blades are mounted on the hub; the blade roots are connected to the hub via pitch control devices. A servo motor is connected to the internal drive shaft of the tower via a coupling, driving the hub to rotate. Electrical control signal lines are connected to the controller to receive speed commands. A pitch control device is installed, enabling it to drive the blade roots to rotate via an electric actuator. A yaw device is installed via a gear set meshing with the flange on the top of the tower, used to adjust the horizontal angle of the hub. The controller receives real-time data from an anemometer (mounted on the top of the tower) and a wind direction sensor (mounted on the side of the tower). Control commands are sent to the servo motor, pitch control device, and yaw device via an RS485 bus. The anemometer and wind direction sensor are installed around the test device according to GB / T35221-2017 (Specification for Ground Meteorological Observation). A six-component balance is embedded between the blade roots and the connecting flange of the hub, directly contacting the blade roots.

[0114] After the device is assembled, the motor is started to control the wind turbine speed to simulate natural wind speed:

[0115] Based on the collected wind speed information and the relationship between wind turbine rotation speed and wind speed determined by the aerodynamic scaling specifications in Table 1, the servo motor is started to rotate the wind turbine. The wind speed sensor data is monitored in real time. If the wind speed deviation exceeds ±0.1 m / s, the motor speed is finely adjusted via the speed controller in 10 rpm increments to stabilize the airflow speed generated by the wind turbine at the target value.

[0116] During the experiment, the pitch angle was adjusted using a pitch control device to simulate natural wind conditions.

[0117] Specifically, the pitch control unit uses data collected by the monitoring device to control the controller and implement the following commands:

[0118] (1) When the wind speed is between 3m / s and 6.98m / s: the minimum rotor speed is gradually reduced to 0° as the wind speed increases.

[0119] (2) When the wind speed is between 6.98m / s and 10.58m / s: This is the optimal tip speed ratio. As the wind speed gradually increases, the blade pitch angle is kept at 0°, and the rotational speed and torque are increased proportionally.

[0120] (3) When the wind speed is between 10.58 m / s and 25 m / s: the controller adjusts the blade pitch angle to increase so that the blade lift remains constant, thereby maintaining a constant rotation speed;

[0121] (4) When the wind speed is greater than 25 m / s: brake in time to ensure the safety of the wind turbine.

[0122] Furthermore, blade flutter is monitored and data is collected;

[0123] A high-speed camera records the blade movement video at a frame rate of 1000Hz, with the recording duration being the duration of each operating condition plus a 10-second buffer time. After recording, FFmpeg software is used to extract the video frame by frame and save it as a PNG image file.

[0124] The frame-by-frame images are input into a pre-trained ResNet-50 neural network model (trained on a dataset containing 20,000 images of different blade positions with a training accuracy of 98.5%) to identify the blade position and obtain the temporal response of the blade displacement.

[0125] The blade displacement data were preprocessed using a third-order Butterworth low-pass filter (cutoff frequency 100Hz) with Gaussian filtering to remove noise. Then, the first and second derivatives were calculated using the central difference method with a time step of 0.001s to obtain the time-domain response curves of velocity and acceleration.

[0126] A Fast Fourier Transform (FFT) was performed on the acceleration time-domain response curve. The Hanning window function was used to reduce spectral leakage, and the frequency response of the blade was analyzed to determine the flutter frequency and natural frequency of the blade.

[0127] The six-component balance data is collected synchronously at a sampling frequency of 1000Hz and transmitted to a computer for storage via an Ethernet interface.

[0128] Further processing and analysis of the data:

[0129] All collected data were processed using the Pandas library in Python. The Matplotlib library was used to plot the curves of blade displacement, velocity, and acceleration over time, as well as the frequency response curve. The x-axis time precision was 0.001s, the y-axis displacement precision was 0.1mm, the velocity precision was 0.1m / s, the acceleration precision was 0.1m / s², and the frequency precision was 0.01Hz.

[0130] The flutter characteristics of blades under different wind speeds and pitch angles were analyzed. The results of visual recognition were compared with the data from a six-component balance. The accuracy and reliability of the experimental results were evaluated using mean square error (MSE) and relative error (RE).

[0131] Based on experimental data, we summarized the patterns and established a mathematical model relating blade flutter characteristics to wind speed and pitch angle, providing a basis for wind turbine blade design and flutter prevention.

[0132] Specifically, the principle of this invention is as follows: a scaled-down test device for studying wind turbine blade flutter under natural wind conditions is designed to ensure aerodynamic similarity with actual wind turbine blades. This is achieved by matching the servo motor speed with the wind speed, distinguishing it from simply changing the tip speed ratio using a servo motor. By establishing an aerodynamic similarity criterion as a precise system, other similarity scales can be determined after the length similarity ratio is determined. Equation (1) is a commonly used criterion for blade aerodynamic similarity, and the two equations contain four variables, making it an uncertain system. This invention introduces equations (2) and (3) to transform it into a deterministic system. The wind turbine speed is controlled by a servo motor. The speed of the scaled-down wind turbine model corresponding to the incoming wind speed is obtained from the similarity criterion. Therefore, the wind turbine rotation is not a free rotation under the lift of natural wind, but a rotation controlled by a servo motor. Non-contact monitoring of the scaled-down blade flutter morphology is performed based on a high-speed camera. The analysis results of visual recognition can be compared and complemented with a six-component balance located at the blade root to ensure the accuracy of the results.

Claims

1. A scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions, characterized in that, The system includes a scaled-down model of a wind turbine, comprising the turbine body, control system, and monitoring devices. The turbine body includes the foundation, tower, and blades. The control system controls the turbine's rotational speed, using similarity criteria to determine the turbine's rotational speed based on the corresponding speed of the scaled-down model under the incoming wind speed. The monitoring devices monitor natural wind speed and perform non-contact monitoring of the scaled-down blade flutter morphology. The blades are a scaled-down model of the IEA 15MW wind turbine blades. To construct a scaled-down model with a response similar to the original blades, the following aerodynamic scaling criteria are established: the basic similarity ratios for wind turbine blade aerodynamic similarity include the Strouhal number (St), Froude number (Fr), Euler number (Eu), and Reynolds number (Re), and the following rules apply: (1) If the only mass force is gravity, Fr and Eu can be equivalent; (2) Fr and Re cannot be satisfied simultaneously; (3) When Re is greater than 4×10 5 Then, it satisfies self-modeling property; Therefore, the Re number must satisfy the self-modularity, and ; in: ； Considering the relative velocity of wind turbine blades, the airfoils at different spanwise positions are designed with pre-twist to achieve the optimal lift-to-drag ratio angle of attack under rated operating conditions. To ensure a constant blade pitch angle while maintaining the blade aerodynamic shape, the following relationship must be satisfied: ； Where, λ ωrot Blade rotation speed ratio; This is a commonly used criterion for the aerodynamic similarity of blades. The two equations contain four variables, representing an uncertain system; by introducing equations... and Then, it is transformed into a deterministic system.

2. The scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 1, characterized in that, The control system includes a pitch control device, a yaw device, a servo motor, and a controller; the monitoring devices include an anemometer, a wind direction sensor, a high-speed camera, and a six-component balance.

3. The scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 2, characterized in that, The pitch control device is used to control the blades to achieve the following control effects: (1) Wind speed 3m / s~6.98m / s: the minimum rotor speed. As the wind speed increases, the blade pitch angle gradually decreases to 0°. (2) Wind speed 6.98m / s~10.58m / s: the optimal tip speed ratio. As the wind speed increases, the blade pitch angle remains at 0°, and the rotational speed and torque increase proportionally. (3) Wind speed 10.58m / s~25m / s: The controller adjusts the blade pitch angle to increase, so that the blade lift remains unchanged, thereby maintaining a constant rotation speed; (4) When the wind speed is greater than 25 m / s: brake to ensure the safety of the wind turbine.

4. The scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 3, characterized in that, The foundation is buried below the ground and serves as the supporting base for the entire device. The foundation is fixed to the bottom of the tower with bolts to ensure the verticality and stability of the tower. The top of the tower is connected to the wind turbine hub with flanges, and blades are installed on the hub. The roots of the blades are connected to the hub with bolts, and the pitch control device is located inside the hub.

5. A scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 4, characterized in that, The servo motor is connected to the drive shaft inside the tower via a coupling, driving the hub to rotate. The servo motor is connected to the controller via an electrical control signal line to receive speed commands.

6. A scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 5, characterized in that, The pitch control device drives the blade root to rotate via an electric actuator; the yaw device engages with the flange at the top of the tower via a gear set to adjust the horizontal angle of the hub.

7. A scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 6, characterized in that, The anemometer is fixedly installed at the top of the tower, and the wind direction sensor is fixedly installed on the side of the tower.

8. A scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 7, characterized in that, The controller is used to receive real-time data from the anemometer and wind direction sensor; and to send control commands to the servo motor, pitch control, and yaw control via RS485 bus.

9. A scaled-down test apparatus for studying wind turbine blade flutter under natural wind conditions according to claim 8, characterized in that, The six-component balance is embedded between the blade root and the hub flange, with its surface in direct contact with the blade root.

Images