Wind turbine generator set shutdown state blade dynamic inspection method and system

Abstract

This invention belongs to the field of wind power generation technology and provides a method and system for dynamic inspection of wind turbine blades during shutdown. The method includes: acquiring blade motion data of the wind turbine in a free-swinging state; constructing a blade swaying dynamic model based on the blade motion data to characterize the relationship between blade posture and time; predicting the blade posture in the future time period based on the blade swaying dynamic model to obtain blade posture prediction results, and generating an inspection route including dynamic waypoints based on the blade posture prediction results; controlling an inspection drone to perform dynamic inspection tasks according to the inspection route, and generating a dynamic inspection report. This solution improves the efficiency and safety of inspection operations, adapts to the actual operation and maintenance conditions of wind turbine blades swaying freely, has strong robustness, and can be easily integrated into the existing wind farm operation and maintenance system, providing reliable technical support for efficient, accurate, and intelligent inspection of wind turbine blades.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, and in particular to a method and system for dynamic blade inspection of wind turbine generator sets during shutdown. Background Technology

[0002] As the core component for wind energy capture, the blades of wind turbines directly determine the power generation efficiency and safety stability of the turbines. Therefore, conducting regular and precise inspections of the blades is a key link in ensuring the safe and efficient operation of wind turbines throughout their entire life cycle, and it is also a core task of wind farm operation and maintenance management.

[0003] However, existing blade inspection technologies have many limitations and are difficult to adapt to the actual operation and maintenance requirements of wind farms. Specifically, on the one hand, the existing mainstream blade inspection methods require the wind turbine to be shut down and the main shaft locked. The wind turbine cannot generate electricity during the inspection period, resulting in a significant loss of power generation. At the same time, the frequent locking and unlocking of the main shaft will aggravate the wear of the mechanical structure and easily increase the maintenance cost and failure risk of the equipment.

[0004] On the other hand, while non-stop inspection solutions for wind turbine operation can avoid power generation loss, this technology has extremely high requirements for the performance of inspection equipment and the accuracy of motion control algorithms, making it difficult to popularize in small and medium-sized wind farms. In addition, during the high-speed rotation of the blades, problems such as motion blur and blurring of the inspection images are easily caused, and the image quality cannot meet the requirements for refined defect detection, making it impossible to effectively identify tiny defects in the blades.

[0005] On the other hand, existing drone inspection technologies are all based on fixed route planning and design. Drones can only fly and take pictures according to preset waypoints and cannot perceive and track the swing trajectory of the blades in real time. This makes it very easy for the shooting posture to be mismatched with the actual posture of the blades, resulting in shooting failure or blurry images. It is difficult to obtain high-quality blade inspection images and cannot meet the fine inspection needs of minor defects such as blade surface corrosion, gel coat peeling, and local sand holes.

[0006] It is evident that existing blade inspection solutions suffer from technical problems such as low inspection efficiency and accuracy, as well as insufficient stability and reliability. Summary of the Invention

[0007] This invention provides a method and system for dynamic blade inspection of wind turbine generator sets under shutdown conditions, in order to solve the shortcomings of existing blade inspection schemes, such as low inspection efficiency and accuracy, and insufficient stability and reliability.

[0008] On one hand, the present invention provides a method for dynamic blade inspection of a wind turbine generator set under shutdown conditions, comprising: Acquire blade motion data of a wind turbine generator under free oscillation conditions; Based on the blade motion data, a blade oscillation dynamics model is constructed to characterize the relationship between blade pose and time. Based on the blade oscillation dynamics model, the blade pose in the future time period is predicted to obtain the blade pose prediction result, and based on the blade pose prediction result, an inspection route containing dynamic waypoints is generated. Control the inspection drone to perform dynamic inspection tasks according to the inspection route and generate a dynamic inspection report.

[0009] The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions provided by the present invention acquires blade motion data of the wind turbine generator set under free oscillation conditions, including: Acquire the original pose data of the wind turbine generator in a free-swinging state; The original pose data is initially screened to extract key feature points and their pose data, thus obtaining the blade motion data.

[0010] According to the wind turbine generator blade dynamic inspection method under shutdown conditions provided by the present invention, based on the blade motion data, a blade oscillation dynamic model is constructed to characterize the relationship between blade attitude and time, including: Based on the blade motion data, a three-dimensional environment model of the wind turbine, including the blades, tower, and hub, is constructed. Based on the three-dimensional environment model of the wind turbine, the motion trajectory of the blade motion data is analyzed by feature point analysis, and a blade oscillation dynamic model is constructed to characterize the relationship between the blade posture and time.

[0011] According to the wind turbine generator blade dynamic inspection method provided by the present invention, an inspection route including dynamic waypoints is generated based on the blade attitude prediction results, including: Set an initial route that includes multiple dynamic waypoints; Based on the blade pose prediction results and the set optimal shooting pose, the position, altitude and orientation of each dynamic waypoint in the initial route are dynamically adjusted to obtain an inspection route that includes dynamic waypoints.

[0012] According to the present invention, a method for dynamic blade inspection of a wind turbine generator set under shutdown conditions includes controlling an inspection drone to perform dynamic inspection tasks according to the inspection route and generating a dynamic inspection report, including: Based on the blade pose prediction results, the three-axis stabilization gimbal of the UAV is controlled to adjust the shooting pose in real time. If the blade pose of the current dynamic waypoint matches the shooting pose of the UAV, the onboard camera is triggered to perform a shooting action to obtain the measured image data of the current dynamic waypoint. A dynamic inspection report is generated based on the measured image data.

[0013] The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions provided by the present invention generates a dynamic inspection report based on the measured image data, including: Defect detection is performed on the measured image data to obtain defect detection results; Based on the defect detection results, defect information is marked on the measured image data to obtain a defect-marked image; The defect detection results and related operation information of each dynamic waypoint in the inspection route, as well as the basic information of the wind turbine, are used as inspection log data. Based on the defect marker images and the inspection log data, a dynamic inspection report is generated using a pre-established standardized report template.

[0014] According to the dynamic blade inspection method for wind turbine generator sets under shutdown conditions provided by the present invention, defect detection is performed on the measured image data to obtain defect detection results, including: The measured image data is subjected to image denoising, standardized cropping, and image calibration respectively to obtain the preprocessed image; The preprocessed image is input into a pre-established defect recognition model to obtain the defect detection result output by the defect recognition model; The defect identification model is obtained by training a deep learning network based on leaf image samples and defect label data. The defect detection results include: defect type, defect location, and severity level.

[0015] According to the dynamic blade inspection method for wind turbine generator sets under shutdown conditions provided by the present invention, a dynamic inspection report is generated based on the defect marker image and the inspection log data, using a pre-established standardized report template, including: The inspection log data is integrated and preprocessed to obtain a structured dataset; Retrieve a pre-established standardized report template, wherein the standardized report template includes a defect display module and a data display module; The defect marker image is inserted into the defect display module, and the data of each type in the structured dataset is filled into the data display module according to the preset fields and formats of the template.

[0016] According to the method for dynamic blade inspection of a wind turbine generator set under shutdown conditions provided by the present invention, the method further includes: If the defect detection results indicate that the blade has a defect, then based on the defect detection results and in conjunction with the pre-established maintenance suggestion information database, blade maintenance suggestion information is generated. The defect detection results and blade maintenance recommendations are used as early warning content to generate and send blade abnormality early warning information to the maintenance management terminal.

[0017] On the other hand, the present invention also provides a dynamic blade inspection system for wind turbine generator sets in a shutdown state, including: a drone and a ground control station; The ground control station is used to acquire blade motion data of the wind turbine generator set under free swing state collected by the UAV; based on the blade motion data, a blade swing dynamic model is constructed to characterize the relationship between blade attitude and time; according to the blade swing dynamic model, the attitude of the blade in the future time period is predicted to obtain the blade attitude prediction result, and an inspection route including dynamic waypoints is generated based on the blade attitude prediction result; the inspection UAV is controlled to perform dynamic inspection tasks according to the inspection route and generate a dynamic inspection report.

[0018] The present invention provides a method and system for dynamic blade inspection of wind turbine generators under shutdown conditions. By constructing a blade oscillation dynamic model, it achieves accurate prediction of the future blade orientation and generates an inspection route with dynamic waypoints. This allows drones to dynamically track the blade oscillation trajectory to complete inspection and photography, effectively solving the problems of image failure and blurry images caused by blade oscillation during fixed-route inspections. This ensures high-quality acquisition of blade inspection images and lays the foundation for accurate defect identification. At the same time, relying on model prediction, it achieves automated control of drone inspection, significantly reducing the inspection time of a single wind turbine, improving inspection efficiency and operational safety. Moreover, the entire method is adaptable to the actual operation and maintenance conditions of wind turbine blades oscillating freely, has strong robustness, and can be easily integrated into the existing wind farm operation and maintenance system. It can provide reliable technical support for efficient, accurate, and intelligent inspection of wind turbine generator blades and help wind farms achieve refined operation and maintenance management. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the dynamic blade inspection method for wind turbine generator sets under shutdown conditions provided in this embodiment of the invention. Figure 2 This is a schematic diagram of the inspection route in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the construction principle of the defect identification model in this embodiment of the invention; Figure 4 This is a schematic diagram of the dynamic blade inspection system for wind turbine generator sets in a shutdown state, provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] The following is combined with Figures 1 to 4 This invention describes the detailed scheme of the dynamic blade inspection method and system for wind turbine generator sets under shutdown conditions provided in the embodiments of the present invention. This embodiment is designed for the working condition of wind turbine generator sets being shut down without braking, with the main shaft free and the blades swinging freely within ±5° under the action of natural forces. It combines UAV onboard sensing equipment, algorithm models and ground control stations to achieve dynamic blade inspection.

[0023] like Figure 1 As shown in the embodiment of the present invention, the method for dynamic blade inspection of a wind turbine generator set under shutdown conditions mainly includes the following steps: Step 110: Obtain blade motion data of the wind turbine generator set in a free swing state.

[0024] Understandably, by accurately acquiring the motion data of the blades swinging freely when the wind turbine is stopped and not braked, high-precision and timely raw data support is provided for subsequent modeling. This breaks through the technical limitations of traditional inspections that lack dynamic motion data acquisition of blades. It can truly capture the core characteristics of the natural swing of the blades, such as trajectory, amplitude, and period, avoiding deviations in subsequent model construction caused by missing or distorted data. At the same time, relying on the fusion acquisition method of airborne lidar and inertial measurement unit, the spatial accuracy and temporal continuity of the data are guaranteed, laying a solid data foundation for the implementation of the entire dynamic inspection method.

[0025] Step 120: Based on the blade motion data, construct a blade oscillation dynamic model to characterize the relationship between blade pose and time.

[0026] This embodiment constructs a oscillation dynamic model based on blade motion data to characterize the change of blade pose over time. This model quantifies and abstracts the free oscillation law of the blade, overcoming the technical challenge of traditional inspection methods failing to accurately describe the dynamic pose of blades. This blade oscillation dynamic model accurately reflects the dynamic changes in blade pose, providing core theoretical and model basis for subsequent prediction of future blade pose. It effectively avoids inspection tracking deviations caused by the inability to grasp the blade oscillation law. Furthermore, the model's construction closely matches the actual oscillation conditions of the blade, exhibiting good adaptability and accuracy, providing crucial model support for UAV dynamic flight path planning and collaborative tracking and shooting.

[0027] Step 130: Based on the blade oscillation dynamics model, predict the blade's pose in the future time period to obtain the blade pose prediction result, and generate an inspection route containing dynamic waypoints based on the blade pose prediction result.

[0028] Understandably, this embodiment relies on a blade oscillation dynamics model to accurately predict the future pose of the blade and generate an inspection route with dynamic waypoints accordingly. This completely solves the technical pain points of traditional fixed-route inspections, which cannot adapt to the free oscillation of the blade and are prone to mismatches in shooting pose. The dynamic waypoints are adjusted in real time according to the blade pose prediction results, ensuring that the inspection route is highly consistent with the actual oscillation state of the blade. This not only ensures the flight safety of the UAV inspection but also keeps the UAV in the optimal shooting pose of the blade, creating conditions for subsequent high-quality image acquisition. At the same time, the dynamic generation of the route enables intelligent planning of the inspection path, greatly improving the scientific nature and adaptability of the route planning.

[0029] Step 140: Control the inspection drone to perform dynamic inspection tasks according to the inspection route and generate a dynamic inspection report.

[0030] This embodiment achieves automated operation and standardized output of dynamic inspection results for wind turbine blades by controlling a drone to perform inspection tasks along a dynamic inspection route and generating dynamic inspection reports. This overcomes the limitations of traditional inspections, which involve significant manual intervention, low efficiency, and cumbersome result processing. The drone dynamically tracks the blade's sway along the route to capture images, ensuring high-quality image acquisition and providing a guarantee for accurate defect identification. Simultaneously, the automated inspection significantly reduces the inspection time for a single wind turbine, improving operational safety and efficiency. The generated dynamic inspection report systematically integrates defect information, inspection data, and environmental data, providing wind farm maintenance personnel with comprehensive and accurate inspection results. This provides a reliable basis for decision-making regarding subsequent blade repair, maintenance, and refined operation and maintenance management of the wind farm.

[0031] In one embodiment, acquiring blade motion data of a wind turbine generator in a free-swinging state specifically includes: First, acquire the original pose data of the wind turbine generator in a free-swinging state.

[0032] In practical applications, a drone equipped with an airborne lidar and inertial measurement unit can be controlled to fly to a safe distance of about 35 meters in front of the target wind turbine and hover. The lidar continuously scans the wind turbine blades at a high frequency of 20Hz, while the inertial measurement unit collects the drone's own attitude and position data. Combined with the lidar scanning data, the original pose data of the blades in a free swinging state is obtained. The original pose data includes the spatial coordinates, motion trajectory, swing amplitude, and other unfiltered raw information of the blades as a whole and in each area.

[0033] Then, the original pose data is initially screened to extract key feature points and their pose data, thus obtaining the blade motion data.

[0034] The onboard computing unit of the UAV can perform preliminary screening of the collected raw pose data, eliminating invalid and redundant data caused by environmental interference and slight equipment vibration, and accurately extracting three key feature points of the blade: the tip, the leading edge of the blade, and the root of the blade. At the same time, it extracts the real-time spatial coordinates, motion trajectory, pose change trend and other effective pose data of these three key feature points, and uses them as the final blade motion data. Subsequently, the blade motion data is transmitted back to the ground control station in real time, and the onboard computing unit simultaneously retains a backup.

[0035] In one embodiment, a blade oscillation dynamics model is constructed based on blade motion data to characterize the relationship between blade pose and time, specifically including: First, based on the blade motion data, a three-dimensional environmental model of the wind turbine, including the blades, tower, and hub, is constructed.

[0036] In this embodiment, the ground control station can process the blade motion data and construct a three-dimensional environment model of the wind turbine that includes the blades, tower, and hub. This three-dimensional environment model accurately restores the actual spatial position, size, and relative relationship of each component of the wind turbine, clarifies the spatial pose reference of the blades in the overall environment of the wind turbine, and can define the physical spatial boundary for subsequent trajectory analysis.

[0037] Then, based on the three-dimensional environmental model of the wind turbine, the motion trajectory of the feature points of the blade motion data is analyzed to construct a blade oscillation dynamic model to characterize the relationship between the blade pose and time.

[0038] In this embodiment, based on the three-dimensional environmental model of the wind turbine, the Kalman filter algorithm can be used to filter the motion trajectory of key feature points in the blade motion data, eliminate trajectory deviation, and restore the true free swing trajectory of the blade. Subsequently, the motion trajectory of the filtered feature points is analyzed in depth to uncover the core laws of blade swing. In this embodiment, the blade swing period is measured to be 5-8 seconds, with an amplitude within ±4°. Combining the dynamic principles, a blade swing dynamic model is constructed to characterize the relationship between blade pose and time. The model is then validated in a short period by comparing the predicted feature point pose within the next 200ms with the actual collected data to ensure the accuracy of the model.

[0039] In one embodiment, an inspection route including dynamic waypoints is generated based on the blade pose prediction results, specifically including: First, set an initial route that includes multiple dynamic waypoints.

[0040] In this embodiment, the ground control station can determine the safe distance and avoidance zone for UAV inspection in the three-dimensional environment model of the wind turbine according to the full coverage requirements of wind turbine blade inspection, avoid obstacles such as towers and hubs, and plan an initial route that covers all areas to be inspected on the blades and includes multiple dynamic waypoints. The dynamic waypoints in the initial route can be set as the optimal shooting position reference for each area to be inspected on the blades.

[0041] Then, based on the blade pose prediction results and the set optimal shooting pose, the position, altitude and orientation of each dynamic waypoint in the initial route are dynamically adjusted to obtain the inspection route containing the dynamic waypoints.

[0042] In practical applications, the ground control station can rely on the blade oscillation dynamics model to continuously predict the spatial pose of the blade in the future time period, and obtain the blade pose prediction results at each time node. Combined with the optimal shooting pose for blade inspection, it ensures that the camera optical axis is vertically aligned with the blade inspection area and the image is distortion-free. Based on the blade pose prediction results, the spatial position, flight altitude and orientation of each dynamic waypoint in the initial flight path are dynamically adjusted in real time, so that each dynamic waypoint is always accurately matched with the predicted pose of the corresponding blade inspection area. After the adjustment is completed, the final inspection flight path containing dynamic waypoints is obtained, and the flight path data is sent to the UAV onboard computing unit.

[0043] Figure 2 The diagram illustrates a segmented inspection route covering key locations such as the leading edge, trailing edge, windward side, leeward side, blade root, and blade tip. Figure 2 It shows the distribution of dynamic waypoints along the inspection route.

[0044] In one embodiment, the inspection drone is controlled to perform dynamic inspection tasks according to the inspection route and generate a dynamic inspection report, specifically including: First, based on the blade pose prediction results, the three-axis stabilization gimbal of the drone is controlled to adjust the shooting pose in real time.

[0045] In practical applications, after receiving the inspection route data, the UAV's onboard computing unit controls the UAV to fly autonomously along the route. At the same time, it continuously predicts the blade's pose based on the blade oscillation dynamics model. According to the blade pose prediction results, it controls the UAV's three-axis stabilization gimbal to adjust the shooting pose in real time, calculates the gimbal compensation angle, and completes the attitude calibration, thereby ensuring that the gimbal drives the camera to always be aimed at the blade inspection area.

[0046] Then, if the blade pose of the current dynamic waypoint matches the shooting pose of the UAV, the onboard camera is triggered to perform a shooting action to obtain the measured image data of the current dynamic waypoint.

[0047] During flight, the UAV continuously compares the predicted blade pose of the current dynamic waypoint with the actual shooting pose of the UAV. If the two match precisely, the onboard camera is automatically triggered to shoot. The camera parameters can be automatically optimized according to the ambient light. Multiple images are continuously taken at each dynamic waypoint, and the clear and unblurred images are selected and stored as the measured image data of the current dynamic waypoint. All subsequent measured image data are transmitted back to the ground control station in real time.

[0048] Finally, a dynamic inspection report is generated based on the measured image data.

[0049] In one specific implementation, a dynamic inspection report is generated based on measured image data, including: The first step is to perform defect detection on the measured image data and obtain the defect detection results.

[0050] In this embodiment, defect detection is performed on the measured image data to obtain the defect detection results, specifically including: First, image noise reduction, standardized cropping, and image calibration are performed on the measured image data to obtain the preprocessed image.

[0051] In this step, multi-level image preprocessing can be performed on the measured image data. By performing image denoising, environmental noise can be eliminated. By standardizing the cropping, the measured image data can be cropped to a uniform size to adapt to the model input requirements. By image calibration, image distortion can be corrected and the true shape of the blade can be restored, finally obtaining the preprocessed image.

[0052] Then, the preprocessed image is input into the pre-established defect recognition model to obtain the defect detection results output by the defect recognition model.

[0053] The defect identification model is obtained by training a deep learning network based on leaf image samples and defect label data; the defect detection results include: defect type, defect location, and severity level.

[0054] In this step, the preprocessed image is input into a pre-established defect recognition model, which automatically performs defect detection and outputs the defect detection results. In this embodiment, as shown... Figure 3 As shown, the defect identification model is built on a deep learning network. Specifically, it adopts a convolutional neural network model obtained by machine learning from a simple model with convolution operation capabilities. It is obtained through iterative training and optimization using a large number of blade image samples and corresponding defect label data labeled with defect type, location, severity, etc. It has high detection accuracy and detection speed. The output defect detection results clearly include defect types such as leading edge corrosion, gel coat peeling, lightning strike, and local sand hole, defect locations such as specific areas of the blade and relative positions of feature points, as well as severity levels such as slight, moderate, and severe.

[0055] The second step is to mark the defect information on the measured image data based on the defect detection results, and obtain the defect-marked image.

[0056] In this step, based on the defect detection results, image annotation tools can be used to mark defect information in the measured image data. Specifically, the defect location, type and severity level can be accurately marked in the measured image data to generate defect-marked images, and the defect-marked images can be classified and archived according to dynamic waypoints and blade regions.

[0057] The third step is to use the defect detection results, related operational information, and basic wind turbine information of each dynamic waypoint in the inspection route as inspection log data.

[0058] In this step, the data from the entire inspection process can be organized, integrating the defect detection results of each dynamic waypoint along the inspection route, relevant operational information including UAV operation information, operator information, environmental meteorological data, and basic wind turbine information, as inspection log data.

[0059] The fourth step is to generate a dynamic inspection report based on the defect marker images and inspection log data, using a pre-established standardized report template.

[0060] Specifically, based on defect marker images and inspection log data, a dynamic inspection report is generated using a pre-established standardized report template, including: First, the inspection log data is integrated and preprocessed to obtain a structured dataset.

[0061] In this step, by integrating and preprocessing the inspection log data, unstructured data can be transformed into structured data. Then, it can be classified according to defect information, operation information, fan information, and environmental information, invalid data can be removed, and missing basic information can be added to obtain a structured dataset.

[0062] Then, a pre-established standardized report template is retrieved, which includes a defect display module and a data display module.

[0063] In practical applications, a pre-established standardized report template can be retrieved from the integrated wind and solar power inspection platform. This standardized report template contains two core modules: a defect display module for displaying defect marker images and core defect information, and a data display module for displaying various structured inspection data. The template also presets the display fields, formats, and layout methods for each data item.

[0064] Finally, insert the defect marker image into the defect display module, and fill the data display module with the various types of data in the structured dataset according to the preset fields and formats of the template.

[0065] Specifically, the defect-marked images after classification and archiving can be inserted into the defect display module in sequence according to blade region and defect level, and the corresponding defect detection results can be marked next to the images; various types of data such as defect information, operation information, wind turbine information, and environmental information in the structured dataset can be accurately filled into the data display module according to the preset fields and formats of the template, completing the integration of data and template, automatically generating a standardized dynamic inspection report, and the report supports one-click export in multiple formats such as PDF and Excel.

[0066] In one embodiment, the above-mentioned dynamic blade inspection method for wind turbine generators in a shutdown state may further include: First, if the defect detection results indicate that the blade has a defect, then based on the defect detection results and in conjunction with a pre-established maintenance suggestion information database, blade maintenance suggestion information is generated.

[0067] In this embodiment, the ground control station judges the defect detection results. If the blade is determined to have a defect, whether it is minor, moderate or severe, the pre-established maintenance suggestion information database is retrieved. This maintenance suggestion information database contains standardized maintenance plans, processing procedures, consumable recommendations, maintenance cycles and other information corresponding to different severity levels of various blade defects. In practical applications, the corresponding maintenance information can be matched from the database based on the defect type, location and severity level detected in this test to generate targeted blade maintenance suggestion information.

[0068] Then, the defect detection results and blade maintenance recommendations are used as early warning content to generate and send blade anomaly early warning information to the maintenance management terminal.

[0069] In practical applications, the defect detection results and blade maintenance recommendations from this inspection can be integrated as the core early warning content. The integrated wind and solar power inspection platform generates blade anomaly early warning information according to the preset early warning format, and sends the early warning information to the wind farm's maintenance management terminal in real time via wireless communication, thereby reminding operation and maintenance personnel to deal with blade defects in a timely manner.

[0070] Based on the same general inventive concept, this invention also protects a dynamic blade inspection system for wind turbine generator sets under shutdown conditions. The dynamic blade inspection system for wind turbine generator sets under shutdown conditions provided by this invention will be described below. The dynamic blade inspection system for wind turbine generator sets under shutdown conditions described below can be referred to in correspondence with the dynamic blade inspection method for wind turbine generator sets under shutdown conditions described above.

[0071] like Figure 4 As shown, the wind turbine generator blade dynamic inspection system provided in this embodiment of the invention specifically includes: a drone 210 and a ground control station 220, with the ground control station 220 and the drone 210 being communicatively connected.

[0072] Ground control station 220 is used to acquire blade motion data of wind turbine generator set in free swing state collected by UAV 210; based on blade motion data, a blade swing dynamic model is constructed to characterize the relationship between blade attitude and time; based on the blade swing dynamic model, the attitude of the blade in the future time period is predicted to obtain the blade attitude prediction result, and based on the blade attitude prediction result, an inspection route including dynamic waypoints is generated; the UAV 210 is controlled to perform dynamic inspection tasks according to the inspection route and generate a dynamic inspection report.

[0073] In practical applications, it can be equipped with an unmanned hangar to achieve automated recovery, charging, and data export of drones. After completing the dynamic blade inspection task, the drone will autonomously return to the unmanned hangar according to the preset program. The hangar will automatically recover and dock through its precise positioning and docking mechanism. The hangar will then start an automatic charging process to intelligently replenish the drone's onboard battery, ensuring power supply for the next inspection operation. At the same time, the hangar and the drone's onboard storage unit will establish an automatic data connection to automatically export and back up all inspection data, including blade measurement image data, flight trajectory data, and operation parameter data collected during the inspection process. This data will be synchronously transmitted to the ground control station and the integrated wind and solar power inspection platform, achieving efficient collection and management of inspection data without human intervention, further improving the automation and intelligence level of wind turbine blade inspection.

[0074] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments of the relevant methods, and will not be elaborated further here.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic blade inspection of a wind turbine generator set under shutdown conditions, characterized in that, include: Acquire blade motion data of a wind turbine generator under free oscillation conditions; Based on the blade motion data, a blade oscillation dynamics model is constructed to characterize the relationship between blade pose and time. Based on the blade oscillation dynamics model, the blade pose in the future time period is predicted to obtain the blade pose prediction result, and based on the blade pose prediction result, an inspection route containing dynamic waypoints is generated. Control the inspection drone to perform dynamic inspection tasks according to the inspection route and generate a dynamic inspection report.

2. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to claim 1, characterized in that, Acquire blade motion data of a wind turbine generator under free oscillation conditions, including: Acquire the original pose data of the wind turbine generator in a free-swinging state; The original pose data is initially screened to extract key feature points and their pose data, thus obtaining the blade motion data.

3. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to claim 1, characterized in that, Based on the blade motion data, a blade oscillation dynamic model is constructed to characterize the relationship between blade pose and time, including: Based on the blade motion data, a three-dimensional environment model of the wind turbine, including the blades, tower, and hub, is constructed. Based on the three-dimensional environment model of the wind turbine, the motion trajectory of the blade motion data is analyzed by feature point analysis, and a blade oscillation dynamic model is constructed to characterize the relationship between the blade posture and time.

4. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to claim 1, characterized in that, Based on the blade pose prediction results, an inspection route including dynamic waypoints is generated, including: Set an initial route that includes multiple dynamic waypoints; Based on the blade pose prediction results and the set optimal shooting pose, the position, altitude and orientation of each dynamic waypoint in the initial route are dynamically adjusted to obtain an inspection route that includes dynamic waypoints.

5. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to claim 1, characterized in that, Control the inspection drone to perform dynamic inspection tasks according to the stated inspection route, and generate a dynamic inspection report, including: Based on the blade pose prediction results, the three-axis stabilization gimbal of the UAV is controlled to adjust the shooting pose in real time. If the blade pose of the current dynamic waypoint matches the shooting pose of the UAV, the onboard camera is triggered to perform a shooting action to obtain the measured image data of the current dynamic waypoint. A dynamic inspection report is generated based on the measured image data.

6. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to claim 5, characterized in that, Based on the measured image data, a dynamic inspection report is generated, including: Defect detection is performed on the measured image data to obtain defect detection results; Based on the defect detection results, defect information is marked on the measured image data to obtain a defect-marked image; The defect detection results and related operation information of each dynamic waypoint in the inspection route, as well as the basic information of the wind turbine, are used as inspection log data. Based on the defect marker images and the inspection log data, a dynamic inspection report is generated using a pre-established standardized report template.

7. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to claim 6, characterized in that, Defect detection is performed on the measured image data to obtain defect detection results, including: The measured image data is subjected to image denoising, standardized cropping, and image calibration respectively to obtain the preprocessed image; The preprocessed image is input into a pre-established defect recognition model to obtain the defect detection result output by the defect recognition model; The defect identification model is obtained by training a deep learning network based on leaf image samples and defect label data. The defect detection results include: defect type, defect location, and severity level.

8. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to claim 6, characterized in that, Based on the defect marker image and the inspection log data, a dynamic inspection report is generated using a pre-established standardized report template, including: The inspection log data is integrated and preprocessed to obtain a structured dataset; Retrieve a pre-established standardized report template, wherein the standardized report template includes a defect display module and a data display module; The defect marker image is inserted into the defect display module, and the data of each type in the structured dataset is filled into the data display module according to the preset fields and formats of the template.

9. The method for dynamic blade inspection of a wind turbine generator set under shutdown conditions according to any one of claims 6 to 8, characterized in that, The method further includes: If the defect detection results indicate that the blade has a defect, then based on the defect detection results and in conjunction with the pre-established maintenance suggestion information database, blade maintenance suggestion information is generated. The defect detection results and blade maintenance recommendations are used as early warning content to generate and send blade anomaly early warning information to the maintenance management terminal.

10. A dynamic blade inspection system for a wind turbine generator set under shutdown conditions, characterized in that, include: Unmanned aerial vehicles (UAVs) and ground control stations; The ground control station is used to acquire blade motion data of the wind turbine generator set in a free swinging state, collected by the UAV. Based on the blade motion data, a blade oscillation dynamic model is constructed to characterize the relationship between blade pose and time. Based on the blade oscillation dynamic model, the blade pose in the future time period is predicted to obtain the blade pose prediction result. Based on the blade pose prediction result, an inspection route containing dynamic waypoints is generated. Control the inspection drone to perform dynamic inspection tasks according to the inspection route and generate a dynamic inspection report.

Images