Wind driven generator detection method
Through the image processing technology equipped with the drone, the damage of wind turbine blades is automatically detected, which solves the problems of detection limitations and complex operation in the existing technology, and achieves efficient and simple blade damage detection.
Patent Information
- Application Number
- CN202411023972.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-12
AI Technical Summary
The existing wind turbine blade damage detection methods have limitations, and multiple defects cannot be detected simultaneously, and the test time is long and the operation is inconvenient.
Using a drone equipped with a control unit, a gimbal camera and a wide-angle camera, the blade image is processed through an image processing model (such as the yolov5 model), and the damage position, area and depth of the blade are automatically detected, so as to realize multiple damage detection, and wirelessly The sending unit sends the detection result.
It realizes multiple damage detection of wind turbine blades, simplifies the detection process, improves detection efficiency, and reduces operational complexity and time cost.
Smart Images

Figure CN118934488A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of wind turbine generator detection, and in particular to a wind turbine generator detection method. Background Art
[0002] A wind turbine is an electric device that converts wind energy into mechanical work, which drives the rotor to rotate and finally outputs alternating current. Its working principle is that the blades of the wind turbine rotate under the action of wind force, thereby converting the kinetic energy of the wind into the mechanical energy of the wind rotor shaft. The generator rotates and generates electricity under the drive of the wind rotor shaft. Since wind power is a clean and pollution-free energy source, its application is becoming more and more extensive. In actual use, the blades are the most basic and most critical components of a wind turbine. Whether they are damaged during use is crucial to wind power generation, so it is often necessary to detect whether the blades are damaged in a timely manner.
[0003] At present, the common ways of blade damage in wind turbine generators are blade breakage, blade cracking and local surface erosion. Among them, acoustic wave detection is usually used to detect blade breakage or blade cracking, and infrared thermal imaging testing is used to detect local surface erosion. However, the current damage detection methods have the following defects:
[0004] 1. Existing damage detection methods generally use either acoustic wave detection or infrared thermal imaging testing, which cannot detect different damage conditions of wind turbine blades and has limitations;
[0005] Second: The existing damage detection methods, whether using acoustic wave detection or infrared thermal imaging testing, require the entire area of the wind turbine blade from the tip to the root of the blade to be tested. However, this large-scale regional testing using ultrasonic or infrared thermal imaging has the problems of long testing time, inconvenient operation and complex operation. Summary of the invention
[0006] In view of the shortcomings of the background technology, the present invention provides a wind turbine detection method, and all the technical problems to be solved are that the existing ultrasonic detection and infrared thermal imaging detection methods for detecting damage to wind turbine blades have limitations in actual use, cannot detect multiple defects at the same time, and have a long test time and inconvenient operation.
[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a wind turbine detection method, wherein the wind turbine includes three blades and is implemented by a drone, wherein the drone is provided with a control unit, a pan-tilt camera control unit, a pan-tilt camera and a wide-angle camera; the control unit controls the pan-tilt camera through the pan-tilt camera control unit, and the images captured by the pan-tilt camera and the wide-angle camera are sent to the control unit;
[0008] The method comprises the following steps:
[0009] S1: Control the drone to fly in front of the wind turbine, and let the control unit obtain an image of the wind turbine through the wide-angle camera. The control unit obtains the detection starting position, detection end position and detection direction of three blades based on the wind turbine image;
[0010] S2: Perform the following steps on the three leaves in sequence:
[0011] The control unit drives the drone to fly to the detection starting position of the currently detected blade, and then flies to the detection end position according to the detection direction of the currently detected blade. During the flight, the control unit controls the gimbal camera through the gimbal camera control unit to shoot the currently detected blade to obtain an image of the blade, in which the area of the blade accounts for more than 60%;
[0012] S3: After acquiring the image of each blade, the image of each blade is processed based on the image processing model to find out the loss position, damage area, damage depth, erosion position and erosion area of each blade.
[0013] In a certain embodiment, in step S2, before the UAV flies, the control unit controls the gimbal camera to initially shoot the currently detected blade through the gimbal camera control unit to obtain an overall image of the blade, and the control unit obtains the radial direction of the currently detected blade based on the overall image;
[0014] During the flight, the control unit controls the zoom, pitch angle and yaw angle parameters of the gimbal camera through the gimbal camera control unit based on the radial direction of the currently detected blade, so that the area of the blade in the acquired blade image accounts for more than 60%.
[0015] In some embodiments, the processing model is a yolov5 model.
[0016] In some embodiments, the training process of the yolov5 model is as follows:
[0017] First, a sample image of a wind turbine blade with damage is acquired;
[0018] Then, the sample image is subjected to image filtering, image enhancement and normalization processing in sequence to obtain a preprocessed image;
[0019] Then, the defects in the preprocessed image are manually calibrated to generate a calibrated image;
[0020] Then all the calibrated images are divided into a training set, a validation set and a test set, the training set is used for training the yolov5 model, the validation set is used for validating the yolov5 model, and the test set is used for testing the yolov5 model.
[0021] In a certain embodiment, when performing the first wind turbine detection, after executing step S3, the control unit obtains the hub position of the wind turbine through the wind turbine image, and controls the drone to fly to the front of the hub center. At this time, the control unit records the current height position and records the current height position as the initial height position;
[0022] Thereafter, each time a wind turbine inspection is performed, after executing step S3, the control unit controls the UAV to fly to the front of the hub center based on the wind turbine image currently acquired, and records the current height position; the control unit determines whether the wind turbine has sunk based on the difference between the initial height position and the current height position.
[0023] In a certain implementation manner, the control unit is also electrically connected to a wireless transmission unit. After executing step S3, the control unit sends the detection result to the background terminal through the wireless transmission unit.
[0024] In a certain embodiment, the present invention further includes a protective box for placing a drone disposed on the periphery of the wind turbine, the top surface of the left wall and the top surface of the right wall of the protective box are respectively hinged with a rotating plate for covering the top surface of the protective box, the inner wall surface of the left wall and the inner wall surface of the right wall of the protective box are respectively hinged with a telescopic power device, and the telescopic end of the telescopic power device is hinged to the rotating plate;
[0025] A control device is also provided in the protective box, and the control device includes a power conversion unit, a second control unit, a second wireless transmission unit and a drone charging unit. The power conversion unit is used to convert the output voltage of the wind turbine into an operating voltage of the second control unit, the second wireless transmission unit and the drone charging unit. The second control unit is electrically connected to the telescopic power device and the second wireless transmission unit, respectively, and is used to control the telescopic power device to extend and retract and to send and receive data through the second wireless transmission unit. The drone charging unit is used to charge the drone.
[0026] In a certain embodiment, the background terminal also sends a telescopic instruction to the second control unit through the second wireless sending unit and sends an inspection instruction to the control unit through the wireless sending unit. After receiving the telescopic instruction, the second control unit drives the telescopic power device to telescope. When the control unit receives the inspection instruction, it controls the drone to fly in front of the wind turbine.
[0027] In a certain embodiment, the background terminal sends an inspection path to the control unit, and the control unit controls the drones to fly to the front of the wind turbines on the inspection path in turn for inspection according to the inspection path.
[0028] In a certain embodiment, before executing step S1, the drone communicates with the control system of the wind turbine and sends a shutdown signal to the control system, and the control system controls the wind turbine blades to stop rotating in response to the shutdown signal.
[0029] Compared with the prior art, the beneficial effect of the present invention is that in actual use, after obtaining the image of the wind turbine through a wide-angle camera, the present invention can obtain the detection starting point, detection end point and detection direction of each blade, so that the drone flies according to the detection information, and obtains the image of the blade through the gimbal camera during the flight. The damage condition of the blade can be obtained by processing the image of the blade through the processing model, thereby realizing wind turbine detection, and the whole process is simple and efficient, and can realize various damage detection of the blade. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 Schematic diagram of the connection between the control unit, the gimbal camera control unit, the gimbal camera and the wide-angle camera in the embodiment;
[0031] Figure 2 is a schematic structural diagram of a protection box in an embodiment;
[0032] Figure 3 is a connection diagram of the telescopic power device in the embodiment;
[0033] Figure 4 Schematic diagram of the structure of the control device in the embodiment. DETAILED DESCRIPTION
[0034] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.
[0035] In order to improve the detection speed of wind turbines, reduce the detection complexity and improve the detection efficiency, this embodiment provides a wind turbine detection method, the wind turbine includes three blades, the method is implemented by a drone, and the drone is provided with the following Figure 1 The control unit 1, the gimbal camera control unit 3, the gimbal camera 4 and the wide-angle camera 2 are shown; the control unit 1 controls the gimbal camera 4 through the gimbal camera control unit 3, and the images captured by the gimbal camera 4 and the wide-angle camera 2 are sent to the control unit 1;
[0036] The method comprises the following steps:
[0037] S1: Control the drone to fly in front of the wind turbine, and let the control unit 1 obtain the wind turbine image through the wide-angle camera 2. The control unit 1 obtains the detection starting position, detection end position and detection direction of the three blades based on the wind turbine image;
[0038] S2: Perform the following steps on the three leaves in sequence:
[0039] The control unit 1 drives the drone to fly to the detection starting position of the currently detected blade, and then flies to the detection end position according to the detection direction of the currently detected blade. During the flight, the control unit 1 controls the gimbal camera 4 through the gimbal camera control unit 3 to shoot the currently detected blade and obtain an image of the blade. In the image of the blade, the area of the blade accounts for more than 60%;
[0040] S3: After acquiring the image of each blade, the control unit 1 processes the image of each blade based on the image processing model to find out the loss position, damage area, damage depth, erosion position and erosion area of each blade.
[0041] In actual use, after obtaining the image of the wind turbine through the wide-angle camera 2, the present invention can obtain the detection starting point, detection end point and detection direction of each blade, so that the drone flies according to the detection information, and obtains the image of the blade through the gimbal camera 4 during the flight. The damage condition of the blade can be obtained by processing the image of the blade through the processing model, thereby realizing the wind turbine detection, and the whole process is simple and efficient, and can realize various damage detection of the blade.
[0042] In addition, although the gimbal camera 4 has a zoom function, which allows the gimbal camera 4 to capture the entire image of the blade when it is at the starting point of the detection position, due to the long length of the blade, if the gimbal camera 4 is only kept stationary to capture the image of the blade, there will be a problem of not being able to clearly capture the image of the blade. Therefore, in this embodiment, the drone drives the gimbal camera 4 to fly along the detection direction for shooting, thereby ensuring the clarity of the image of the blade, and then ensuring the accuracy of the detection result.
[0043] In actual use, the blades of the wind turbine need to be in a disabled state before the blades of the wind turbine are inspected. Based on this, in this embodiment, before executing step S1, the drone communicates with the control system of the wind turbine and sends a shutdown signal to the control system. The control system responds to the shutdown signal to control the wind turbine blades to stop rotating.
[0044] Specifically, in this embodiment, the shooting process of the pan / tilt camera 4 is as follows:
[0045] Before the UAV flies, the control unit 1 controls the gimbal camera 4 through the gimbal camera control unit 3 to initially shoot the currently detected blade to obtain the overall image of the blade. The control unit 1 obtains the radial direction of the currently detected blade based on the overall image.
[0046] During the flight, the control unit 1 controls the zoom, pitch angle and yaw angle parameters of the gimbal camera 4 through the gimbal camera control unit 3 based on the radial direction of the currently detected blade, so that the area of the blade in the acquired blade image accounts for more than 60%.
[0047] Specifically, in this embodiment, the processing model is the yolov5 model.
[0048] Among them, the training process of the yolov5 model is as follows:
[0049] First, a sample image of a wind turbine blade with damage is acquired;
[0050] Then, the sample image is subjected to image filtering, image enhancement and normalization processing in sequence to obtain a preprocessed image;
[0051] Then, the defects in the preprocessed image are manually calibrated to generate a calibrated image;
[0052] Then all the calibrated images are divided into training set, validation set and test set. The training set is used for training the yolov5 model, the validation set is used for validating the yolov5 model, and the test set is used for testing the yolov5 model.
[0053] Specifically, in this embodiment, when performing the first wind turbine detection, after executing step S3, the control unit 1 obtains the hub position of the wind turbine through the wind turbine image, and controls the drone to fly to the front of the hub center. At this time, the control unit 1 records the current height position and records the current height position as the initial height position;
[0054] Thereafter, each time a wind turbine inspection is performed, after executing step S3, the control unit 1 controls the UAV to fly to the front of the hub center based on the wind turbine image currently acquired, and records the current height position; the control unit determines whether the wind turbine has sunk based on the difference between the initial height position and the current height position.
[0055] Reference Figure 1 In this embodiment, the control unit 1 is also electrically connected to the wireless sending unit 5. After executing step S3, the control unit 1 sends the detection result to the background terminal through the wireless sending unit 5, thereby realizing remote detection and viewing.
[0056] Specifically, in this embodiment, the present invention further includes a protective box 6 for placing the drone outside the wind turbine. The structure of the protective box 6 is as follows: Figure 2 As shown, the top surfaces of the left side wall and the right side wall of the protective box 6 are respectively hinged with rotating plates for covering the top surface of the protective box, and the two rotating plates are respectively a left rotating plate 7 and a right rotating plate 8. The inner wall surface of the left side wall and the inner wall surface of the right side wall of the protective box 6 are respectively hinged with a telescopic power device 9, and the telescopic end of the telescopic power device 9 is hinged with the rotating plate; wherein the connection diagram of the telescopic power device 9 and the left rotating plate 7 is shown in Figure 3 As shown; in a certain embodiment, the telescopic power device 9 can be an electric push rod;
[0057] A control device is also provided in the protection box 6, such as Figure 4 As shown, the control device includes a power conversion unit 10, a second control unit 11, a second wireless transmission unit 12 and a UAV charging unit 13. The power conversion unit 10 is used to convert the output voltage of the wind turbine into an operating voltage of the second control unit 11, the second wireless transmission unit 12 and the UAV charging unit 13. The second control unit 11 is electrically connected to the telescopic power device 9 and the second wireless transmission unit 12 respectively, and is used to control the telescopic power device 9 to extend and retract and to send and receive data through the second wireless transmission unit 12. The UAV charging unit 13 is used to charge the UAV.
[0058] In actual use, drones can be stored and managed by setting up a protective box, so that the drones can be deployed near wind turbines and remotely controlled for inspection through the background terminal, greatly simplifying the inspection process.
[0059] Specifically, in this embodiment, the background terminal also sends a telescopic instruction to the second control unit 11 through the second wireless sending unit 12 and sends an inspection instruction to the control unit 1 through the wireless sending unit 5. After receiving the telescopic instruction, the second control unit 11 drives the telescopic power device 9 to telescope, thereby opening the rotating plate so that the drone can fly out of the protective box. When the control unit 1 receives the inspection instruction, it controls the drone to fly in front of the wind turbine.
[0060] When performing wind power generation, multiple wind turbines are set up in an area. In order to realize the detection of multiple wind turbines by one drone and reduce the detection cost and difficulty, in this embodiment, the background terminal sends an inspection path to the control unit 1, and the control unit 1 controls the drone to fly to the front of the wind turbines on the inspection path in turn for inspection according to the inspection path.
[0061] The above is based on the present invention as an inspiration. Through the above description, relevant staff can make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.
Claims
1. A method for detecting a wind turbine, wherein the wind turbine comprises three blades, characterized in that: This is achieved by a drone, wherein the drone is provided with a control unit, a pan-tilt camera control unit, a pan-tilt camera and a wide-angle camera; the control unit controls the pan-tilt camera through the pan-tilt camera control unit, and the images captured by the pan-tilt camera and the wide-angle camera are sent to the control unit; The method comprises the following steps: S1: Control the drone to fly in front of the wind turbine, and let the control unit obtain an image of the wind turbine through the wide-angle camera. The control unit obtains the detection starting position, detection end position and detection direction of three blades based on the wind turbine image; S2: Perform the following steps on the three leaves in sequence: The control unit drives the drone to fly to the detection starting position of the currently detected blade, and then flies to the detection end position according to the detection direction of the currently detected blade. During the flight, the control unit controls the gimbal camera through the gimbal camera control unit to shoot the currently detected blade to obtain an image of the blade, in which the area of the blade accounts for more than 60%; S3: After acquiring the image of each blade, the image of each blade is processed based on the image processing model to find out the loss position, damage area, damage depth, erosion position and erosion area of each blade.
2. A wind turbine detection method according to claim 1, characterized in that: In step S2, before the UAV flies, the control unit controls the gimbal camera to initially shoot the currently detected blade through the gimbal camera control unit to obtain an overall image of the blade, and the control unit obtains the radial direction of the currently detected blade based on the overall image; During the flight, the control unit controls the zoom, pitch angle and yaw angle parameters of the gimbal camera through the gimbal camera control unit based on the radial direction of the currently detected blade, so that the area of the blade in the acquired blade image accounts for more than 60%.
3. A wind turbine detection method according to claim 1, characterized in that: The processing model is the yolov5 model.
4. A wind turbine detection method according to claim 3, characterized in that: The training process of the yolov5 model is as follows: First, a sample image of a wind turbine blade with damage is acquired; Then, the sample image is subjected to image filtering, image enhancement and normalization processing in sequence to obtain a preprocessed image; Then, the defects in the preprocessed image are manually calibrated to generate a calibrated image; Then all the calibrated images are divided into a training set, a validation set and a test set, the training set is used for training the yolov5 model, the validation set is used for validating the yolov5 model, and the test set is used for testing the yolov5 model.
5. A wind turbine detection method according to claim 1, characterized in that: When performing the first wind turbine detection, after executing step S3, the control unit obtains the hub position of the wind turbine through the wind turbine image, and controls the drone to fly to the front of the hub center. At this time, the control unit records the current height position and records the current height position as the initial height position; Thereafter, each time a wind turbine inspection is performed, after executing step S3, the control unit controls the UAV to fly to the front of the hub center based on the wind turbine image currently acquired, and records the current height position; the control unit determines whether the wind turbine has sunk based on the difference between the initial height position and the current height position.
6. A wind turbine detection method according to claim 1, characterized in that: The control unit is also electrically connected to a wireless transmission unit. After executing step S3, the control unit sends the detection result to the background terminal through the wireless transmission unit.
7. A wind turbine detection method according to claim 6, characterized in that: It also includes a protective box for placing a drone arranged on the periphery of the wind turbine, the top surface of the left wall and the top surface of the right wall of the protective box are respectively hinged with a rotating plate for covering the top surface of the protective box, the inner wall surface of the left wall and the inner wall surface of the right wall of the protective box are respectively hinged with a telescopic power device, and the telescopic end of the telescopic power device is hinged to the rotating plate; A control device is also provided in the protective box, and the control device includes a power conversion unit, a second control unit, a second wireless transmission unit and a drone charging unit. The power conversion unit is used to convert the output voltage of the wind turbine into an operating voltage of the second control unit, the second wireless transmission unit and the drone charging unit. The second control unit is electrically connected to the telescopic power device and the second wireless transmission unit, respectively, and is used to control the telescopic power device to extend and retract and to send and receive data through the second wireless transmission unit. The drone charging unit is used to charge the drone.
8. A wind turbine detection method according to claim 7, characterized in that: The background terminal also sends a telescopic instruction to the second control unit through the second wireless sending unit and sends an inspection instruction to the control unit through the wireless sending unit. After receiving the telescopic instruction, the second control unit drives the telescopic power device to telescope. When the control unit receives the inspection instruction, it controls the drone to fly in front of the wind turbine.
9. A wind turbine detection method according to claim 7, characterized in that: The background terminal sends the inspection path to the control unit, and the control unit controls the drone to fly to the front of the wind turbines on the inspection path in turn for inspection according to the inspection path.
10. A wind turbine generator detection method according to claim 1, characterized in that: Before executing step S1, the drone communicates with the control system of the wind turbine and sends a shutdown signal to the control system. The control system controls the wind turbine blades to stop rotating in response to the shutdown signal.
